0% found this document useful (0 votes)

90 views939 pages

32nd EBES CONFERENCE PROCEEDINGS VOLUME - I

Q: What are the primary entities interested in the RegTech phenomenon, and why are they involved?

The primary entities interested in the RegTech phenomenon are regulated entities, such as financial intermediaries, that need to comply with regulatory reporting requirements [FCA, 2016]. These entities use RegTech solutions to enhance their compliance, monitoring, risk management, and reporting processes . Additionally, non-regulated entities like tech companies, consulting firms, and fintech startups are also involved as they see opportunities in the development of RegTech solutions [Yang et Li, 2018]. Other stakeholders include industry associations, investors, non-profits, and consumers, who benefit from increased regulatory compliance, data-sharing capabilities, and enhanced consumer protection [Gurung et Perlman, 2018; Arner et al., 2016].

Q: Evaluate the mutual relationship between RegTech and traditional regulatory models and the potential for its evolution in the future.

The mutual relationship between RegTech and traditional regulatory models is characterized by the integration of advanced technologies that facilitate more efficient regulatory compliance and oversight processes . RegTech solutions enhance traditional models by providing data-driven insights, predictive analytics, and increased interaction between regulators and regulated entities [Baxter, 2016]. The potential for evolution includes transitioning towards primarily data-driven, predictive, proactive, and transparent regulation, facilitating enhanced compliance and reducing regulatory burdens. This evolution may lead to more dynamic and responsive regulatory environments [Yang et Li, 2018; Anagnostopoulos, 2018].

Q: Discuss the potential economic and social impacts of RegTech on a global scale, particularly in financial markets.

RegTech has the potential to significantly impact global economic and social dynamics by enhancing efficiency, accuracy, and transparency within financial markets [Quill et Lennon, 2019]. Economically, it may lower operational costs, improve risk management, and facilitate compliance, thus increasing the competitiveness of financial institutions [Capgemini, 2018]. Socially, RegTech can promote financial inclusion and protect consumer interests by ensuring more stringent compliance with regulations, thereby fostering trust and stability [Arner et al., 2016]. However, the global implementation of RegTech faces challenges such as varying regulatory standards and potential exacerbation of digital divides .

Q: In what ways does the financial sector leverage RegTech, and what are the associated benefits and challenges?

The financial sector leverages RegTech to manage compliance, monitoring, risk management, reporting, and operations more effectively . Benefits include increased efficiency and accuracy, reduced costs, and enhanced transparency and manageability [Baxter, 2016; Capgemini, 2018]. However, challenges arise with cyber risks, potential misinformation, and overdependence on technology, which might lead to complacency in risk assessment and management [Micheler et Whaley, 2019; Buckley et al., 2020].

Q: What role do non-profit organizations play in the development and implementation of RegTech solutions?

Non-profit organizations play a supportive role in RegTech development by funding initiatives that improve market supervision and policy analysis, especially in developing economies. For instance, the Bill & Melinda Gates Foundation, Omidyar Network, and USAID have partnered with financial authorities in countries like the Philippines, Ghana, and Mexico to develop tools for enhanced regulatory frameworks [Gurung et Perlman, 2018]. Their involvement ensures that RegTech solutions are not only profit-driven but also aligned with broader social objectives and financial inclusion goals .

Q: What challenges may arise from integrating predictive analytics and AI in RegTech solutions, and how can these be mitigated?

Integrating predictive analytics and AI in RegTech solutions may present challenges such as biases in AI algorithms, cyber risks, and potential over-reliance on technology, which could lead to inadequate oversight and risk management [Buckley et al., 2020]. To mitigate these challenges, stakeholders need to ensure robust cybersecurity measures, continuous evaluation and updating of AI models, and maintaining a balance between human oversight and technological reliance [Micheler et Whaley, 2019]. Furthermore, regulatory bodies should establish clear guidelines and ethical standards for the use of AI in regulatory processes [Currie et al., 2018].

Q: What are the potential risks associated with RegTech solutions, specifically regarding cybersecurity and information integrity?

Potential risks associated with RegTech solutions include cyber risks, which may compromise sensitive data through breaches [Buckley et al., 2020]. Over-reliance on RegTech could lead to a false sense of security, and inadequate long-term risk assessment might increase vulnerabilities [Micheler et Whaley, 2019]. Furthermore, RegTech solutions might enable malicious use by market participants, posing a threat to information integrity and potentially leading to regulatory challenges if information is incomplete or manipulated [Currie et al., 2018].

Q: How do RegTech solutions address different regulatory approaches and frameworks?

RegTech solutions are adaptable to various regulatory approaches such as insight-based, principle-based, and rules-based frameworks [Goul, 2019]. They aid in self-regulation, command-regulation, and meta-regulation by enabling interactive, predictive, and proactive methodologies [Micheler et Whaley, 2019; Yang et Li, 2018]. These solutions contribute to increased transparency and accountability by facilitating data-driven supervision, enhancing real-time data exchange between regulators and regulated entities, and supporting compliance with complex international regulations like MiFID II and UCITS [Financial Stability Board, 2017a; Anagnostopoulos, 2018].

Q: How does RegTech benefit consumer protection, and what technologies enable this protection?

RegTech benefits consumer protection by leveraging technologies such as Big Data and AI to analyze vast amounts of data, identify anomalies, and ensure compliance with consumer protection regulations [Financial Stability Board, 2017a]. This facilitates more informed decision-making for consumers through better access to accurate and timely information [Arner et al., 2016]. Additionally, the increased data transparency and rigorous monitoring capabilities of RegTech enhance consumer confidence by ensuring that financial products and services meet regulatory standards [FCA, 2016].

Q: How does RegTech contribute to the innovation of regulatory processes and systems?

RegTech contributes to the innovation of regulatory processes and systems by serving as an enabler and facilitator of smart, digital, and automated solutions, making compliance more efficient and transparent [Quill et Lennon, 2019]. It supports real-time operations and data-driven regulation, reshaping existing structures and encouraging interactive and responsive regulatory models [Baxter, 2016; Anagnostopoulos, 2018]. Furthermore, RegTech employs Big Data and AI technologies for improved analysis and processing [Financial Stability Board, 2017a].

EBES CONFERENCE PROCEEDINGS

Uploaded by

Iulia Elena Neagoe

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

90 views939 pages

32nd EBES CONFERENCE PROCEEDINGS VOLUME - I

EBES CONFERENCE PROCEEDINGS

Uploaded by

Iulia Elena Neagoe

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

32nd EBES CONFERENCE

PROCEEDINGS
VOLUME - I

ISTANBUL, TURKEY

AUGUST 5-7, 2020

(Online/Virtual Presentation Only)

ebes@[Link]
[Link]
EBES - Eurasia Business and Economics Society

EBES is a scholarly association for scholars involved in the practice

and study of economics, finance, and business worldwide. EBES was
founded in 2008 with the purpose of not only promoting academic
research in the field of business and economics, but also encouraging
the intellectual development of scholars. In spite of the term “Eurasia”,
the scope should be understood in its broadest term as having a global
emphasis.

EBES aims to bring worldwide researchers and professionals together

through organizing conferences and publishing academic journals and
increase economics, finance, and business knowledge through academic discussions. Any
scholar or professional interested in economics, finance, and business is welcome to attend
EBES conferences. Since our first conference in 2009, around 12,549 colleagues from 99
countries have joined our conferences and 7,091 academic papers have been presented.
EBES has reached 2,311 members from 87 countries.

Since 2011, EBES has been publishing two journals. One of those journals, Eurasian
Business Review - EABR, is in the fields of industrial organization, innovation and
management science, and the other one, Eurasian Economic Review - EAER, is in the fields
of applied macroeconomics and finance. Both journals are published quarterly by Springer
and indexed in Scopus. In addition, EAER is indexed in the Emerging Sources Citation
Index (Clarivate Analytics) and EABR is indexed in the Social Science Citation Index (SSCI).
EABR has an Impact Factor of 2.222 (2019 JCR Impact Factor).

Furthermore, since 2014 Springer has started to publish a new conference proceedings series
(Eurasian Studies in Business and Economics) which includes selected papers from the
EBES conferences. The 10th, 11th, 12th, 13th, 14th, 15th, 16th, 17th, 18th, 19th, 20th (Vol.2),
21st and 24th EBES Conference Proceedings have already been accepted for inclusion in the
Conference Proceedings Citation Index - Social Science & Humanities (CPCI-SSH).
Other conference proceedings are in progress.

On behalf of all EBES officers, I sincerely thank you for all your support in the past. We look
forward to seeing you at our forthcoming conferences. We very much welcome your comments
and suggestions in order to improve our future events. Our success is only possible with your
valuable feedback and support!

I hope you enjoy the conference!

With my very best wishes,

Klaus F. ZIMMERMANN
President

i
EXECUTIVE BOARD

• Klaus F. Zimmermann, UNU-MERIT & Maastricht University, The Netherlands

• Jonathan Batten, RMIT University, Australia
• Iftekhar Hasan, Fordham University, U.S.A.
• Euston Quah, Nanyang Technological University, Singapore
• John Rust, Georgetown University, U.S.A.
• Dorothea Schafer, German Institute for Economic Research DIW Berlin, Germany
• Marco Vivarelli, Università Cattolica del Sacro Cuore, Italy

OFFICERS
Klaus F. Zimmermann, PhD Mehmet Huseyin Bilgin, PhD
(President) (Vice President)
UNU-MERIT & Maastricht University, The Netherlands Istanbul Medeniyet University, Turkey
E-mail: zimmermann@[Link] E-mail: bilgin@[Link]

Hakan Danis, PhD Alina Klonowska, PhD

(Vice President) (Executive Secretary)
MUFG Union Bank, U.S.A. Cracow University of Economics, Poland
E-mail: danis@[Link] E-mail: klonowska@[Link]

Orhun Guldiken, PhD Ender Demir, PhD

(Treasurer) (Conference Coordinator)
Manhattan College, U.S.A. Istanbul Medeniyet University, Turkey
E-mail: guldiken@[Link] E-mail: demir@[Link]

Sofia Vale, PhD Jonathan Tan, PhD

(Conference Coordinator in Europe) (Conference Coordinator in Asia)
ISCTE - IUL, Portugal Nanyang Technological University, Singapore
E-mail: vale@[Link] E-mail: tan@[Link]

Ugur Can Aylin Akin

(Administrative Director of the EBES Office) (Assistant Editor of the EBES Publications)
E-mail: can@[Link] E-mail: akin@[Link]

Merve Erdemir
(Assistant of the Director)
E-mail: erdemir@[Link]

ii
ADVISORY BOARD

• Ahmet Faruk Aysan, Istanbul Sehir University, Turkey

• Michael R. Baye, Kelley School of Business, Indiana University, U.S.A.
• Mohamed Hegazy, School of Management, Economics and Communication, The
American University in Cairo, Egypt
• Cheng Hsiao, Department of Economics, University of Southern California, U.S.A.
• Noor Azina Ismail, University of Malaya, Malaysia
• Irina Ivashkovskaya, State University - Higher School of Economics, Russia
• Christos Kollias, Department of Economics, University of Thessaly, Greece
• Wolfgang Kürsten, Friedrich Schiller University Jena, Germany
• William D. Lastrapes, Terry College of Business, University of Georgia, U.S.A.
• Sungho Lee, University of Seoul, South Korea
• Justin Y. Lin, Peking University, China
• Brian Lucey, The University of Dublin, Ireland
• Rita Martenson, School of Business, Economics and Law, Goteborg University, Sweden
• Steven Ongena, University of Zurich, Switzerland
• Peter Rangazas, Indiana University - Purdue University Indianapolis, U.S.A.
• Peter Szilagyi, Central European University, Hungary
• Amine Tarazi, University of Limoges, France
• Russ Vince, University of Bath, United Kingdom
• Adrian Wilkinson, Griffith University, Australia
• Naoyuki Yoshino, Faculty of Economics, Keio University, Japan

iii
Welcome to the 32nd EBES Conference

We are excited to organize our 32nd conference, which will take place
on August 5th, 6th, and 7th, 2020. Due to the Covid-19 virus, the
conference presentation mode has been switched to "online/virtual
presentation only”. Participants will have two options for presentation:
PowerPoint Virtual Presentation and Online (Live) Presentation via
Zoom. Keep in mind that online conference is not intended to perfectly
replicate the in-person conference experience. However, this is the
only option to continue our research at this juncture. We will return to
our regular conference as soon as the world is safe to travel again.

We are honored to have received top-tier papers from distinguished scholars from all over the
world. We regret that we were unable to accept more papers. In the conference, 185 papers
will be presented and 355 colleagues from 49 countries will attend the conference. We are
pleased to announce that distinguished colleagues Asli Demirguc-Kunt (the Chief Economist
of Europe and Central Asia Region of the World Bank), Klaus F. Zimmermann (President of
the Global Labor Organization (GLO), EBES, UNU-MERIT & Maastricht University, the
Netherlands), Marco Vivarelli (Professor at the Catholic University of Milano), Dorothea
Schäfer (Research Director of Financial Markets at the German Institute for Economic
Research (DIW Berlin)), and Jonathan Batten (Professor at RMIT University, Australia) will join
the conference as keynote speakers.

Throughout the years, EBES conferences have been an intellectual hub for academic
discussion. Participants have found an excellent opportunity for presenting new research,
exchanging information and discussing current issues. We believe that our future conferences
will improve further the development of knowledge in our fields. In addition, based on the
contribution of the paper to the field, the EBES Award Committee has selected one of the
papers for the Best Paper Award. The Best Paper Award winner will be announced during the
conference.

On behalf of EBES, I would like to thank to all presenters, participants, board members, and
keynote speakers.

I am looking forward to meeting you in the conference and seeing you all again at the upcoming
EBES conferences. We hope that everything would be improved in a short period of time and
I could meet you in person in our conferences. We appreciate your patience, partnership,
support and understanding during this extraordinary times.

Stay safe and healthy!

Best regards,

Ender Demir, PhD

Conference Coordinator

iv
SCIENTIFIC COMMITTEE

• Sagi Akron, University of Haifa, Israel

• Hasan Fehmi Baklaci, Izmir University of Economics, Turkey
• Adam P. Balcerzak, Nicolaus Copernicus University, Poland
• Marco Bisogno, University of Salerno, Italy
• Gabor Bota, Budapest University of Technology and Economics, Hungary
• Laura Brancu, West University of Timisoara, Romania
• Taufiq Choudhry, University of Southampton, UK
• Andrzej Cieslik, University of Warsaw, Poland
• Joel I. Deichmann, Bentley University, USA
• Irene Fafaliou, University of Piraeus, Greece
• Clara García, Universidad Complutense de Madrid, Spain
• Tamara Jovanov, University Goce Delcev - Shtip, Macedonia
• Alexander M. Karminsky, National Research University, Russia
• Ashraf A. Khallaf, American University of Sharjah, UAE
• Tipparat Laohavichien, Kasetsart University, Thailand
• Gregory Lee, University of the Witwatersrand, South Africa
• Ivana Dražić Lutilsky, University of Zagreb, Croatia
• Roman Mentlik, University of Finance and Administration, Czech Republic
• Jan Jakub Michałek, University of Warsaw, Poland
• Veljko M. Mijušković, University of Belgrade, Serbia
• Alexander Redlein, Vienna University of Technology, Austria
• Nives Botica Redmayne, Massey University, New Zealand
• Liza Rybina, KIMEP University, Kazakhstan
• Hunik Sri Runing Sawitri, Universitas Sebelas Maret, Indonesia
• Irina Sennikova, RISEBA University, Latvia
• Pekka Tuominen, University of Tampere, Finland
• Manuela Tvaronavičienė, Vilnius Gediminas Technical University, Lithuania
• Meltem Ucal, Kadir Has University, Turkey
• Sofia de Sousa Vale, ISCTE Business School, Portugal
• Leszek Wincenciak, University of Warsaw, Poland

v
VOLUME 1 - CONTENTS
Page
Article Title & Authors
Numbers
1 The Effect of Personality Traits on Credit Score using BMTI Personality Types 1-13
Adnan Veysel Ertemel, Istanbul Commerce University, Turkey and Gokhan
Çaylak, Bahcesehir University, Turkey
2 Bank Lending and Small and Medium Sized Enterprises’ Access to Finance - 14-60
Effects of Macroprudential Policies
Aida Cehajic, University of Sarajevo and University of Ljubljana, Slovenia and
Marko Košak, University of Ljubljana, Slovenia
3 The Mediating Role of Emotional Stability Between Emotional Intelligence and 61-76
Overwork Behavior Ajtene Avdullahi, University of Mitrovica, Kosovo and
Osman Yildirim, Istanbul Arel University, Turkey
4 RegTech: a Comprehensive View from Academics, Authorities and 77-99
Practitioners
Alessandro Faes, Politecnico di Milano, Italy; Laura Grassi, Politecnico di
Milano, Italy; and Davide Lanfranchi, Politecnico di Milano, Italy
5 Empirical Modeling of International Banks’ Credit Risk: Assessment and 100-196
Comparison of Credit Ratings
Alexander M. Karminsky, National Research University Higher School of
Economics, MGIMO-University, Russia; Ella Khromova, National Research
University Higher School of Economics, Russia; and Roman Kudrov, National
Research University Higher School of Economics, Russia
6 Rethinking the Role of Credit Unions in the European Social Economy 197-208
Ana Ivanisevic Hernaus, University of Zagreb, Croatia and Ivana Biondic, The
Institute for Development and International Relations, Croatia
7 All We Need Are Taxes: A Systematic Review on Tax Avoidance 209-267
Andreia Magalhães, ISCTE-IUL, Portugal; Rogério Serrasqueiro, ISCTE-IUL,
Portugal; and Paulo Dias, ISCTE-IUL, Portugal
8 Turkey Direct Investments in Latvian Business: Agenda for Research 268-284
Andrejs Limanskis, RISEBA University, Latvia
9 Adoption of IoT technology among Aged NCD Patients in Malaysia: A 285-291
Conceptual Study based on the Theory of Planned behavior
Chinnasamy Agamudai Nambhi Malarvizhi, Multimedia University, Malaysia;
Srinivasan Jayashree, Multimedia University, Malaysia; and Shamima Raihan
Manzoor, Multimedia University, Malaysia
10 The Role of Innovation in Insurance and its Response to Covid-19 Pandemic 292-310
Emergency
Davide Lanfranchi, Politecnico di Milano, Italy; Marco Giorgino, Politecnico di
Milano, Italy; and Laura Grassi, Politecnico di Milano, Italy
11 Maximizing Customer Experience Using Display Retargeting 311-326
Denis Hyams Ssekasi, University of Bolton, United Kingdom; David Bamber,
University of Bolton, United Kingdom; and Roshan Panditharathna, OLC-
Europe LTD, United Kingdom
12 Big Data Analytics in Business: Evolving Knowledge, Skills and Roles of 327-338
Accounting and Finance Professionals
Eley Suzana Kasim, Universiti Teknologi Mara, Malaysia; Noryati Md Noor,
Universiti Teknologi MARA, Malaysia; Norlaila Md Zin, Universiti Teknologi
MARA, Malaysia; and Noor Sufiawati Khairani, Universiti Teknologi MARA,
Malaysia
13 Consequences of Technological Development on the Manifestation of the 339-343
Rebound Effect Phenomenon
Giani Ionel Gradinaru, The Bucharest University of Economic Studies, Institute
of National Economy, Romania
14 Income Distribution, Unemployment and Crime Relationship: Turkey Case 344-355
Haluk Yergin, Van Yuzuncu Yil University, Turkey and Mustafa Torusdag, Van
Yuzuncu Yil University, Turkey

vi
15 Sharing Economy in Practice: Intrinsic Motivation towards the Intention to 356-375
Participate with Empirical Evidence from Vietnam
Hang Thu Hoang, University of Economics Ho Chi Minh City, Vietnam; Tran
Quy Bao Bui, University of Economics Ho Chi Minh City, Vietnam; Mai Thi
Thanh Doan, University of Economics Ho Chi Minh City, Vietnam; Thy Phuong
Le, University of Economics Ho Chi Minh City, Vietnam; Nhi Le Thao Phan,
University of Economics Ho Chi Minh City, Vietnam; and Quynh Nhu Vo,
University of Economics Ho Chi Minh City, Vietnam
16 ICT and Social Media Influence on the MICE Market and the Event 376-384
Management
Igor Kovacevic, Faculty of Economics University of Belgrade, Serbia; Bojan
Zečević, Faculty of Economics University of Belgrade, Serbia; and Branislava
Hristov Stancic, Faculty of Economics University of Belgrade, Serbia
17 Selected Aspects of Nature Conservation Management at Local Level on the 385-403
Example of Communes in Poland
Jacek Witkowski, Lublin University of Technology, Poland
18 Uncertainty and Exchange Rates: Global Dynamics (Well, I Don't Quite Know 404-446
Anymore)
Jing Lian Suah, Central Bank of Malaysia, Malaysia
19 Leadership and Gender 447-458
Kezban Talak, Yeditepe University, Turkey
20 Optimization of Shareholder and Corporate Wealth Analysis: Effectiveness 459-463
Comparative Evaluation of SVA, EVA and MVA
Lukman Ayinde Olorogun, Istanbul Gelisim University, Turkey and Monsurat
Ayojimi Salami, International Islamic University Malaysia, Malaysia
21 The Market and Customer Orientation - Performance Relationship 464-469
Maria Gracner, University of Ljubljana, Slovenia
22 Experiences of Auditors and Tax Advisors with Accounting Errors: Empirical 470-482
Evidence from the Czech Republic
Marie Paseková, Tomas Bata University in Zlín, Czech Republic; Miroslava
Dolejšová, Tomas Bata University in Zlín, Czech Republic; Jana Helová,
Tomas Bata University in Zlín, Czech Republic; and Michal Šindelář, Tomas
Bata University in Zlín, Czech Republic
23 The Importance of Internationalization Strategy for Innovation in Portuguese 483-498
Firms
Mónica Isabel Lopes Azevedo, Universidade Portucalense Infante D.
Henrique, Portugal; Carla Azevedo Lobo, Universidade Portucalense,
Portugal; Carla Santos, Universidade Portucalense Infante D. Henrique,
Portugal; Natércia Durão, Universidade Portucalense Infante D. Henrique,
Portugal; and Isabel Maldonado, Universidade Portucalense, Portugal
24 Online Shopping Drivers and Barriers in the time of COVID-19 for Turkish 499-529
Young and Older Adults
Muge Akkor Koktekin, Yeditepe University, Turkey
25 The Influence of the European H2020 Program on the Development of 530-545
Regions: The Multiplier Impact of EU Funding in the Regions under Horizon
2020 in NUT III - Algarve - Portugal
Natacha Silva, Universidade Portucalense Infante D. Henrique, Portugal;
Ribeiro Diamantino, Universidade Portucalense Infante D. Henrique, Portugal;
and Ribeiro João, Universidade Portucalense Infante D. Henrique, Portugal
26 Reporting Expenses from the Standpoint of Accounting and Taxation 546-564
Requirements
Nexhmie Berisha Vokshi, University of Prishtina, Kosovo and Florentina Xhelili
Krasniqi, University of Prishtina, Kosovo
27 Accounting Ethics and Organizational Culture: A Study of the Points of 565-583
Contention and Similarities
Nida Turegun, Ozyegin University, Turkey
28 The Evolution of Tax Burden and Tax Effort in the OECD between 2000 and 584-652
2015
Paulo Dias, ISCTE-IUL, Portugal and Maria Cavadas, ISCTE-IUL, Portugal
29 The Determinants of Unemployment: A Case of South Africa 653-673
Priyanka Patel, North West University, South Africa and Ireen Choga, North
West University, South Africa
vii
30 The Moderating Role of Corporate Governance on the Effect of Covid-19 674-699
Pandemic on the Saudi Corporate Profitability
Raed Reda Obaid, King Abdulaziz University, Saudi Arabia and Rawia Reda
Obaid, King Abdulaziz University, Saudi Arabia
31 Institutional Determinants of Emerging Market Returns, Flows, and Risk 700-735
Premiums
Ralph Sonenshine, American University, U.S.A. and Bradley Erickson,
American University, U.S.A.
32 Investigation of the Effects of Non-Macroeconomic Indicators on the Exports 736-745
of Goods and Services in the OECD Member Countries
Sevgi Sezer, University of Balikesir, Turkey
33 Exploring the Dimensions forming the HEIs Image for the Current International 746-757
Students: Context of Malaysia
Shamima Raihan Manzoor, Multimedia University, Malaysia; Chinnasamy
Malarvizhi, Multimedia University, Malaysia; and Junainah Binti Mohd Mahdee,
Multimedia University, Malaysia
34 Innovation Capability and Logistics Service Quality in Improving the 758-773
Performance of Malaysian 3pl Service Providers
Siti Nur 'Atikah Zulkiffli, Universiti Malaysia Terengganu, Malaysia; Maisarah
Sebadak, Universiti Malaysia Terengganu, Malaysia; Siti Falindah Padlee,
Universiti Malaysia Terengganu, Malaysia; and Juhaizi Mohd Yusof, Universiti
Malaysia Terengganu, Malaysia
35 Credit Driven Household Demand Channel and Real Economic Activity with 774-812
Macroprudential Policy
Talnan Aboulaye Toure, Kobe University, Japan and Nakamura Tamotsu,
Kobe University, Japan
36 Financial Inclusion and Welfare: New Evidence from the Role of 813-837
Government
Talnan Aboulaye Toure, Kobe University, Japan and Nakamura Tamotsu,
Kobe University, Japan
37 Trade Credit, Trade Income Elasticity and the International Transmission of 838-879
Shocks
Anna Maria Watson, University of Cambridge, United Kingdom
38 Value at Risk Estimation of Sector Index Futures: Evidence from Thailand 880-896
Futures Exchange
Woradee Jongadsayakul, Kasetsart University, Thailand
39 Authenticity through the Eyes of the Host: The Complication of Presenting 897-906
Culture as Commodity
Yodmanee Tepanon, Kasetsart University, Thailand; Thanwa Benjawan,
Payap University, Thailand; and Chawan Maleehom, Payap University,
Thailand
40 Economic Development and Nationalism: In the Cases of Scotland and 907-930
Catalonia
Yusuf Omur Yilmaz, Mardin Artuklu Univeristy, Turkey

Note: No English-language editing and proofreading was done by the publisher therefore the quality of
language of papers is under the authors’ responsibility. The authors keep the copyright of their articles.

viii
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The Effect of Personality Traits on Credit Score using BMTI

Personality Types

Adnan Veysel Ertemel

Istanbul Commerce University
Turkey

Gökhan Çaylak
Bahcesehir University Turkey
Abstract
Technology advancements brings about the opportunity for individuals to have
access to financial services all around the world. Credit scoring plays a key role in
assessing the eligibility of individuals to use those financial services. Credit scoring
is also increasingly important in non-financial contexts including marketing and
human resources.

From personality traits perspective, an individual’s credit score can be said to be

theoretically affected from their choices and preferences. Therefore, we hypothesize
that human personality define their actions ultimately affecting their credit score.
Specifically, this study investigates the relationship of credit score and personality
traits. Although there are some studies on this topic using Big Five personality
traits, there isn’t any work done using another widely used instrument, Myers
Briggs Type Indicator (MBTI) which is a shorter and more practical for commercial
uses as opposed to Big Five model.

Through a survey administred to 181 participants, MBTI personality types, namely;

extraversion-introversion, thinking-feeling, judging-perceiving, sensing-intuitive
dimensions are measured. Credit score of participants was received from Kredi
Kayıt Burou (KKB) Turkey with their explicit consent. Correlation analysis was
carried out to test the hypothesis. Analysis results reveal that sensing and judging
dimensions positively correlate with credit score. The results partly conform to
those of previous studies that use Big Five model. That is, as Big Five
consiontesness dimension was found in previous studies to correlate with credit
score and consiontesness overlaps with BMTI sensing and judging dimensions
correlation between dimension, a correlation between credit score and sensing and
judging dimensions makes sense.

Keywords: Personality Traits, Credit Score, BMTI, Financial Inclusion

1. Introduction
It is evident that personality traits play a key role in predicting human behavior.
There is an extensive literature that studies personality traits in various contexts.
Credit scoring is not an exception to this. The extant literature demonstrates that
credit scoring is not mere a function of one’s economic data (Heck, 1980; Roberts et
al., 2007; Becker et al. 2012; Ladas [Link], 2015). Specifically, personality traits were
found to influence an individual’s economic preferences (Roberts et al., 2007; Ozer
and Benet-Martínez; 2006) and eventually credit score (Klinger, 2013; Rustichini,
2016; Wang, et al., 2011; Ottaviani and Vandone, 2011)

1
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Among many instruments that classify personality types, Big Five Model and Myers
Briggs Type Indicator (MBTI) are among the most widely used models for
describing individual personalities. Almost all of the studies on this subject use Big
Five model. On the other hand, there is a shortage of study that uses MBTI to
assess the effect of personality traits on individuals’ credibility score. As such, this
paper makes a contribution to the literature in this sense. In the remainder of the
paper, the literature on credit scoring and personality theory is reviewed first.
Existing studies based on Big Five on the subject is mentioned. Afterwards, the
analysis of the conducted study that uses BMTI instruments is made accompanied
by relevant discussion.

2. Conceptual Background

2.1 Credit Scoring

Credit Scoring aims at development of models for distinguishing good applicants

from bad applicants (Baesens et al., 2003). Credit scoring as we know it was not
introduced to the market as it is right now. Durand (1941) is the first to apply
statistical methods for selection of credit applicants. Later on, engineer Bill Fair and
mathematician Earl Isaac found an analytics company in 1956, acronym known as
today as FICO. FICO pioneered analytic solutions such as credit scoring that have
made credit more widely available, not just in the United States but around the
World (Kalmar & Blume, 2007). Credit score data consists of multiple data points
from banks information like payment history debt burden, length of credit history,
types of credit used etc. Credit scoring can be defined as the set of decision models
and their underlying methods that serve lenders in granting consumer credits (Zhang
et al. 2010).

Today, credit scoring systems are used in loan applications. Customers are
evaluated by their personal information and historical payment performance.
Traditionally some statistical methods like linear discriminant analysis and logistics
regression were being used. However new data mining techniques such as neural
networks can be used as alternatives (Wang et al. 2011). Credit scoring has been
one of the most successful applications of data mining (Finlay, S. M. 2011).

2.2 Personality Theory

Personality can be defined as the organization of the inner world of a person that is
consistent and predictable. (Piedmont, 1998). Kazdin refers personality to individual
differences in characteristic patterns of thinking, feeling and behaving. (Kazdin,
2000). The study of personality focuses on two broad areas: One is understanding
individual differences in particular personality characteristics, such as sociability or
irritability. The other is trying to perceive the reasons how the different parts of a
person come together as a whole. (Kazdin, 2000). Scholars attempt to develop
different measures for understanding human personality. Trait theory is one such
theory to underline the critical importance of stability parameters in human

2
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

personality. A trait is a cross-situational distinction of an individual (Ajzen, 2005).

It reflects the attributes of individuals that can be used to distinguish between
different persons. Carl Jung is known to be one of the pioneers in this area (Jung,
1971). Jungian theory inspired many scholars to develop new personality models
(Myers and McCaulley 1975; Keirsey and Bates, 1984; Rushton et al., 2007). Later
on, scholars elaborated on different theories and models to describe human
personality. Big Five Model is one of the most widely used personality trait model
in the last decades

2.2.1 Big Five Model

Big Five model, also called as Five Factor Model (FFM) is a well-known model that
describe five higher order personality traits to explain human behavior (Costa and
McCrae, 1992). It aims at characterizing humans by placing them in one or more of
five main traits. These traits are, extraversion, agreeableness, neuroticism, openness
to experience and conscientiousness (Costa and McCrae, 1992; Furnham 2002).
Extraversion refers to the tendency to be more assertive, highly activite, implying
preference of interpersonal relations. Agreeableness refers to friendly, trusting,
caring and modest behavior. Neuroticism refers to the tendency to be depressive,
anxious and fearful. Openness-to-experience refers to the tendency to get involved
in intellectual activities, be more creative and imaginative. Conscientiousness is
associated with disciplined work ethics, ambitiousness, determination and
responsibility.

Previous Studies in Personality Traits and Credit Scoring

Big Five model has been adopted by many scholars in personality theory in various
studies. The literature that relate personality traits to credit score are no exception.
The few studies found on this subject used the big five model to analyze personality
effects on credit scores (Rustichini et. al 2016; Klinger et. al 2013; Heck, 1980).

There are few studies that specifically investigate the effect of personality trait on
credit score of individuals. Among these, Rustichini et. al (2016) explored the effect
of personality traits and found negative correlation between neuroticism and credit
score. In another study that investigates the default risk of credits, Klinger et. al
(2013) similarly found neuroticism as negatively affecting credit score. Withdrawal
and volatility are the main facets of neuroticism. Among these, withdrawal was
similarly found by Heck (1980) to be one of the particular traits that characterize
those with low credit scores. In explaining this finding, we can say that attitudes
towards risk appear to be affected by Neuroticism. Previous evidence suggests that a
stronger sensitivity to ambiguity and unpredictable outcomes is associated with a
higher level of neuroticism (Hirsh et al., 2008). Conscientiousness, on the other
hand, is found to be positively affecting credit score in previous studies (Klinger et.
al 2013). This makes sense as conscientiousness represents the capacity and
propensity to exercise control over behavior, and to seek non-immediate [Link],
an individual with high score in Conscientiousness is expected to have high credit
score. Although there are studies on the subject on Big Five model, the MBTI,
which is also a widely used model, has not been used in literature for studying credit
score

2.2.2 Myers Briggs Type Indicator (MBTI)

3
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Myers-Briggs personality Type Indicator (MBTI) is an instrument developed by

Katharine C. Briggs and Isabel Briggs-Myers during World War II for job matching
and assignment (Myers and McCaulley 1975; Quenk, 2009). MBTI technique,
which can be viewed one based on Jungian theory, has been administered to
millions of individuals on various contexts. It involves a responding to a
questionnaire that allows the classification of a person’s traits according to four
main dimensions. (Furnham, 1996; Hammer and Barger 1996). As a measure that
consists of relatively fewer questions MBTI has been used in more commercial
settings and is regarded as the most commonly used non-clinical personality
measure (Davito 1985; Bayne 2003).

MBTI differentiates from Big Five in that instead of traits or continuous variables it
measures distinct personality types. It defines 16 personality types across four
dimensions that reveal personal preferences in explaining how people behave in
different circumstances. The four dimensions are namely: Extraverted
(E)/Introverted (I), Sensing (S)/ Intuitive (N), Thinking (T)/Feeling (F), and Judging
(J)/Perceiving (P). Each person belongs to one opposite on the scale.

The MBTI type theory encompasses two main assumptions. First, the equal
development of all four functions is not suitable (Bayne, 2003), and the prevalent
continuum should be the most advanced. Second, people in both dimensions can not
concurrently establish competing interests. This tends to lead to a dominant
preference and then to a cumulative cycle is further developed As such, each
individual slowly finds what they are best for his or her talents.

According to MBTI, extraverts are easier to relate to others, while intraverts focus
on the inner world of thoughts and ideas. Sensing and intuitive actions are ways to
understand the world. Sensing personalities sense through the five senses
whereas intuitive personality is rather unconscious-oriented. The two ways
of judging are the distinctions of thinking which highlight logic and impersonal
processes. Feeling based more on personal values and judgments. The final
dimension is a combination of perception and judgment. Judging type reflect
preferences to a scheduled, determined and organized way of life, whereas the
perceiving type prefers a flexible, spontaneous way of life. Different scholars
developed new models based on MBTI, including temperament types developed by
Keirsey and Bates (1984), which adopted the MBTI to examine psychological
preferences.

When we compare, Big Five and MBTI models, we can say that Big Five model is a
descriptive model. Myers-Briggs Type Indicator, however, places an individual to
one of the 16 predefined personality types. As opposed to the rather descriptive
nature of Big Five, BMTI places the individual to one of those predefined
personality types. Scholars have previously investigated the overlap between Big
Five and MBTI (Costa and McCrae1989; McDonald et al. 1994; Furnham et al.
(2002). Both Costa and McCrae(1989) and Furnham et al. (2002) studies confirm
that Extraversion-Introversion, Openness was correlated with Sensing-Intuition,
Agreeableness with Thinking–Feeling and Conscientiousness with both Judging-
Perceiving and Sensing-Intuition. This paper aims at filling a gap in the literature by
undertaking a study analyzing the relationship of credit score and MBTI personality
types.

4
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3. Research Model and Hypothesis Development

In light of the previous studies, the resarch question of this study is to test the effect
of the four MBTI dimensions, namely Extraverted vs. Introverted, Sensing vs.
Intuitive, Thinking vs. Feeling, and Judging vs. Perceiving on credit score. The
research model is shown in the Figure 3.1.

Figure 3.1: Research Model

Sensing vs.
Intuitive Thinking
[Link]

Extraverted vs. Judging vs.

Introverted Perceiving

Credit
Score

Based on the research model, following hypotheses are developed:

Hypothesis 1: There is a significant correlation between extraverted personality and

credit score.
Hypothesis 2: There is a significant correlation between sensing personality and
credit score.
Hypothesis 3: There is a significant correlation between thinking personality and
credit score.
Hypothesis 4: There is a significant correlation between judging personality and
credit score.

4. Research Metholodgy
A descriptive research was conducted in the study. Resources from past research
studies about credit scoring and psychometrics were probed into the study in
literature review part. In this part of the study subject selection, research
methodology, instrumentation, reliability, data collection procedures are explained.

4.1 Sampling Procedure

Research questions relating to the personality traits were explored with reference to
credit scores of 181 people. Convenience sampling method is used and an e-mail
was send to more than 300 individuals outlining the details about the study and its
purpose, and asked to participate. The study included both male and female

5
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

individuals in order to create a meaningful sample set for further discovery. Age
range from 18 to 60 individuals are included to the sample in order to understand
life stages of the participants. The information regarding their contact and
demographic details was received face to face, professions and background
information are excluded, most represented categories are age groups between 19 to
35, those are people with the most engagement with online tools. After the
approvals of individuals MBTI personality test is conducted. Together with
information collected from KKB(Kredi Kayit Bureau) via Kredico. results were
analyzed in SPSS statistics package. To ensure the anonymity of reported responses,
each individual name was anonymized.

4.2 Analysis Results

Reliability and Validity Analysis

Reliability of the scales was determined. To do this, cronbach’s alpha values were
calculated. The results are shown in table 4.1.

Table 4.1: Reability and validity results

Scale Cronbach Alpha

Extraverted vs. Introverted 0.87
Sensing [Link] 0.78
Thinking [Link] 0.75
Judging vs. Prospecting 0.82

As cronbach’s alpha values are bigger than the threshold value (0.7), the scales were
found to be reliable. Second step to determine validity using discriminant validity
analysis fort he four distinct scales namely; Extraverted vs. Introverted, Sensing vs
Intuitive, Thinking vs Feeling, Judging vs Perceiving. Table below indicates that
instrument scales’ coefficients are below the threshold which means they do not
have impact on each other.

Table 4.2 Coefficient Values of Personality Dimensions

Extraverted vs. Sensing vs. Thinking vs. Judging vs.

Introverted Intuitive Feeling Prospecting
Extraverted vs. -0.09 0.02 -0.01
Introverted
Sensing vs. -0.09 0.09 0.37
Intuitive
Thinking vs. 0.02 0.09 0.08
Feeling
Judging vs. -0.01 0.37 0.08
Prospecting

Preliminary Results
The survey was performed in Turkey using convenience sampling method. Data
was collected through email. Data collection processes lasted three months in the
second quarter of 2018. In the sample, 100 persons are mail and 81 persons are
female. Frequency and percentage of respondents by age, education and gender are
detailed in the table below.

6
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 4.3 Frequency Distribution of Participants

Frequency Percentage
AGE
20-29 70 39
30-39 59 32
40-49 38 21
50-59 14 8
Grand Total 181 100
Gender
Male 100 55,2
Female 81 44,8
Grand Total 181 100
EDUCATION LEVEL
Primary School 7 3,9
Highschool 40 22,1
University 104 57,5
Masters 22 12,2
PHD 8 4,4
Grand Total 181 100

In this study as seen from the table below, personality trait means were calculated
with SPSS. Extraverted profile has highest ratio with participant in the study with
minimum 10 and maximum 100 percent. Second highest trait is Judging trait with
minimum 10 and maximum 95 percent and the third highest trait is Assertive
personality trait with minimum 16 and maximum 100 percent. End results states
study participants have balanced ratio of Intuitive, Sensing, Thinking and Feeling
traits. As discussed in the limitations of study section study is limited to results for
the location it took place. There fore validity of results could be suspected as
subjective. A bigger sample size is needed to have a better understanding of
population traits distribution. Also honesty of survey participants whom participated
in this study have critical importance with the results.

Table 4.4 Descriptive Statistics of Personality Types

N Minimum Maximum Mean Std.

Deviation
EXTRAVERTED 181 11 100 71.88 15.712

INTROVERTED 181 0 89 28.12 15.712

SENSING 181 4 100 49.98 14.460

INTUITIVE 181 0 96 50.02 14.460

7
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

THINKING 181 14 98 49.48 14.828

FEELING 181 2 86 50.52 14.828
JUDGING 181 10 95 55.55 17.630
PROSPECTING 181 5 90 44.40 17.716
Valid (listwise) 181

[Link] Test Results

5.1 Effect of Extraversion (vs. Introversion) Personality Type on Credit Score

In this study correlation test technique is practiced for investigating the effects on
relationship between extraversion and introversion on credit score. Calculated results
are presented in the below chart as a summary.

Table 5.1: Correlation Test Results Of Extraversion vs Introversion On Credit Score

Credit score Extraverted

Extraverted Pearson .042 1
Correlation
Sig. (2- Tailed) .571

N 181 181
Credit score Pearson 1 .042
Correlation
Sig. .571
(2- Tailed)
N 181 181

According to analysis results no correlation between extraversion and introversion of

participants' according to credit score. (p=0.571, p<0,05). Since our p value is above 0,05
can say that there is no statistically significant association of being extraverted and
introverted on credit score. Our correlation coefficients are -.042 and, .042 which means
there is not really an association between extraversion and introversion on credit score.

5.2 Effect of Sensing (vs. Intuitive) Personality Type and Credit Score
In this study correlation test was carried out to test the relationship between sensing (vs.
intuitive) personality type and credit score segment. Calculated results are presented in
the below chart.
Table 5.2: Correlation Test Results of Sensing (vs. Intuitive) Personality Type and
Credit Score

Credit score Sensing

Sensing Pearson .288** 1.000**
Correlation

8
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Sig. (2- Tailed) .000 .000

N 181 181
Credit score Pearson 1 .288
Correlation
Sig. .000
(2- Tailed)
N 181 181

According to the results, correlation coefficient is 0.288 which means there is

significant correlation between Sensing (vs. Intuitive) personality type and credit
score. (p=0.000, p<0,05).

5.3 Effect of Thinking vs. Feeling Personality Type on Credit Score

In this study correlation test technique is practiced for investigating the effects
on relationship between thinking and feeling on credit score. Calculated results
are presented in the below chart as a summary.
Table 5.3: Correlation Test Results Of Thinking vs Feeling On Credit Score

Thinking Credit score

Thinking Pearson 1 .019

Correlation
Sig. (2- Tailed) .804

N 181 181
Credit score Pearson .019 1
Correlation
Sig. .804
(2- Tailed)
N 181 181

According to acquired results there no correlation between extraversion and

introversion of participants' according to credit score. (p=0.804, p<0,05). Since our p
value is above 0,05 can say that there is no statistically significant association of being
thinking and feeling on credit score. Our correlation coefficient is .019 which means
there is not really an association between extraversion (vs. introversion) personality
type and credit score.

5.4 Effect of Judging vs. Perceiving Personality Type on Credit Score

In this study correlation test technique is practiced for investigating the effects on
relationship between judging and perceiving on credit score. Calculated results are
presented in the below chart as a summary.

Table 5.4: Correlation Test Results of Judging vs Perceiving and Credit Score

Judging Credit score

9
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Judging Pearson 1 .187*

Correlation
Sig. (2- Tailed) .013
N 181 181
Credit score Pearson .187* 1
Correlation
Sig. .013
(2- Tailed)
N 181 181

According to acquired results there is a correlation between sensing (vs. intuitive) values of
participants' according to credit score. (p=0.013, p<0,05). Our correlation coefficient is 0.187
which means there is but weak association between judging and perceiving on credit score.
5.5 Hyphothesis Chart

Table 5.9: Hypothesis Chart

Hypotheses Result

1. There is a significant correlation between

Extraversion (vs. Intraversion) Personality
type and credit score Not Supported

[Link] is a significant correlation between

Sensing (vs. Intuitive) personality type Supported
and credit score.

[Link] is a significant correlation between Not Supported

Thinking (vs. Feeling) personality type
and credit score.

4. There is a significant correlation between

Judging (vs. Prospecting) personality type Supported
and credit score.

According to results of this study sensing and judging levels of individuals have significant
correlation with credit score. In this case, hypothesis 2 and 4 are supported. On the other hand,
extraversion and thinking personality types were found to have no correlation with credit score,
so hypothesis 1 and 3 are not supported

6. Findings and Discussion

For Sensing vs. Intuitive dimension the correlation coefficient is .288 which means there is a
significant correlation between sensing and on credit score. This result makes sense in that
sensing type of individuals are more concerned with real world and they have a better
management of tasks in order base (Ghauri and Usunier, 2003). Dealing with facts is a strong
function with this kind of personality. Credit score is affected from reality so it is understandable
that this type of individual score higher than intuitive types. Conversely, being intuitive is
10
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

associated with being imaginative (Jung, 1971), kind of detached from real world. Not
thouroughly thinking about the results of actions or make necessary preparation for an outcome
might negatively affect reality perception as well as credit score.

For Judging vs. Perceiving dimension the correlation coefficient is .187 which means there is but
weak association between judging and perceiving on credit score. The reason of this result is,
having a judging personality is associated with being prepared for the challenges. This type of
personality is known for their consistency and clarity, always being prepared means they have a
back up plan for things to come ahead (Jung, 1971). Also sense of responsibility is greater with
this type. From the financial behavior aspect, judging personalities plan ahead and even before
applying for a loan they devise a plan for repayment schedule. If they fail to comply with a
responsibility (most likely they do not) they have a plan B to make things back to normal
(Conger, 2005). Therefore, it is understandable to see this type has positive association with
credit score. They perceive the world as it is and not tries to hide or escape from it. If they are
using a financial product they are aware of the necessities to comply before hand. Financial
products by nature is strongly about being rational and factual. In contrast perceiving individuals
associated with relax attitude towards challenges. Their impulsive aspect of character makes
them kind of opportunist ther fore a trade off can be seen when comes to comply with
responsibilities. For example this type of individual can go for a second mortgage loan even
though the first one is not paid back.

The findings using BMTI personality types is partially in-line with those that overlap with the
Big Five model in that for consciountesness dimension of big five which correlates with both
sensing-intuition and judging-perceiving dimensions of MBTI, the results are parallel to Big
Five confirming the correlation in those dimensions. However, for neuroticisim dimension of
Big Five, which was found to have a weak correlation with extraversion-introversion BMTI
dimension, no significant correlation was found.

These results conform to the findings of the previous studies using Big Five traits. That’s, as Big
Five conscientiousness and credit score correlates (Rustichini et. al 2016; Klinger et. al 2013)
and there is an overlap of conscientiousness with BMTI sensing and judging (Furnham, 2002),
correlation between credit score and sensing and judging dimensions makes sense.

7. Conclusion

Personality traits explain human behavior to a great extent. Credit scoring is no exception to his.
Although this effect is previously studied in the literature, there is a lack of such a study using
widely accepted BMTI personality instrument. As such, this study fills in the missing gap in the
literature by investigating the relationship of BMTI personality type dimensions and credit score
of individuals. As a result of the hypothesis tests, correlations have been found between judging
(vs. perceiving) and sensing (vs. intuition) personality types and credit score. On the other hand,
no significant correlation was found for extraversion (vs. intraversion) and thinking (vs. feeling)
personality types.

The results have important managerial implications. Today’s highly competitive markets
demand for further segmentation of their customer base, improve profitability and decrease loss.
Better understanding customers’ financial behavior leads to important insights for companies. In
that regard, credit scores become a new instrument for companies to understand their portfolio’s
preferences and predictability. New ways of acquiring credit scores might be introduced even for
those with no previous payment behavior. The limitation of the study is that it was carried out
with a relatively small sample and just in Turkey. Future studies can be made with a larger
sample size and in different geographies.

11
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

REFERENCES

Ajzen, I. (2005). Attitudes, Personality and Behavior. 2nd Edition. UK: Open University
Press.

Baesens, B., Van Gestel, T., Viaene, S., Stepanova, M., Suykens, J., & Vanthienen, J. (2003).
Benchmarking state- of-the-art classification algorithms for credit scoring. Journal of the
Operational Research Society, 54, 627–635.

Bayne, R. (2003). Love, money and studying. The Psychologist, 16, 529–531.

Becker, A., T. Deckers, T. Dohmen, A. Falk, and F. Kosse 2012. “The relationship between
economic preferences and psychological personality measures.” Annual Review of
Economics 4, 453–478.

Conger, J. P. (2005). Jung & Reich: The body as shadow. North Atlantic Books.

Costa, P., & McCrae, R. (1992). Revised NEO Personality Inventory (NEO-PI-R) and NEO
Five Factor Inventory (NEO-FFI) Pro- fessional Manual. Odessa,

Davito, A. (1985). A review of the MBTI. In J. Mitchell (Ed.), North Mental Measurement
Yearbook.

Durand, D. (1941). Risk elements in consumer installment financing. National Bureau of

Economic Research, New York.

Finlay, S., 2011. Multiple classifier architectures and their application to credit risk
assessment. European Journal of Operational Research, 210(2), pp.368–378

Furnham, A. (2002). Personality at work. London: Routledge.

Ghauri, P. N., & Usunier, J. C. (Eds.). (2003). International business negotiations. Emerald
Group Publishing.

Hammer, A. L., & Barger, N. J. (1996). MBTI applications: A decade of research on the
Myers-Briggs Type Indicator. Palo Alto, CA: Consulting Psychologists Press.

Heck, R. K. Z., 1980, Identifying Insolvent Households, Journal of Housing Economics,

Winter: 14-17.

Hirsh, J.B. and M. Inzlicht, 2008. “The devil you know: neuroticism predicts neural response
to uncertainty.” Psychological Science 19, 962–967.

Jung, C. (1971). Psychological types. Princeton, NJ: Princeton University Press.

Kalmar, P., & Blume, M. (2007). FICO: Web person disambiguation via weighted similarity
of entity contexts. In Proceedings of the 4th International Workshop on Semantic
Evaluations (p. 149-152). Association for Computational Linguistics.

Kazdin, A. E. (2000) Encyclopedia of Psychology. New York: Oxford University Press.

12
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Keirsey, D., & Bates, M. (1984). Please understand me: Character and temperament types.
Del Mar, CA: Gnosology Books/Prometheus Nemesis.

Klinger, B., Khwaja, A. I., & Del Carpio, C. (2013). Enterprising psychometrics and poverty
reduction (Vol. 860). New York, NY: Springer.

Ladas, A., Ferguson, E., Aickelin, U., & Garibaldi, J. (2015). A Data Mining Framework to
Model Consumer Indebtedness with Psychological Factors. In 2014 IEEE International
Conference on Data Mining Workshop (pp. 150-157). IEEE

McDonald, D., Anderson, P., Tsagarakis, C., & Holland, J. (1994). Examination of the
relationship between the Myers-Briggs Type Indicator and the NEO personality inventory.
Psychological Reports, 74, 339–344.

Myers, I., & McCalley, M. (1975). Manual: A guide to the development and use of the
Myers-Briggs Type Indicator. Palo Alto, CA: Consulting Psychologists.

Ozer, D.J. and V. Benet-Martínez, 2006. Personality and the Prediction of Consequential
Outcomes. Palo Alto, CA, US: Annual Reviews.

Ottaviani, C., & Vandone, D. (2011). Impulsivity and household indebtedness: Evidence from
real life. Journal of Economic Psychology, 32, 754–761.

Piedmont, R. L. (1998). The Plenum series in social/clinical psychology. The revised NEO
Personality Inventory: Clinical and research applications. New York, NY, US.

Roberts, B.W., N.R. Kuncel, R. Shiner, A. Caspi, and L.R. Goldberg, 2007. “The power of
personality the comparative validity of personality traits, socioeconomic status, and
cognitive ability for predicting important life outcomes.” Perspectives on Psychological
Science 2, 313–345.

Rushton, S., Morgana, J., & Richard, M. (2007). Teacher’s Myers-Briggs personality profiles:
Identifying effective teacher personality traits. Teaching and Teacher Education, 23(4), 432–
441.

Rustichini, A., DeYoung, C. G., Anderson, J. E., & Burks, S. V. (2016). Toward the
integration of personality theory and decision theory in explaining economic behavior: An
experimental investigation. Journal of Behavioral and Experimental Economics, 64, 122-137.

Wang, L., Lu, L., & Malholtra, N.K. (2011). Demographics, attitude, personality and credit
card features correlate with credit card debt: A view from China. Journal of Economic
Psychology 32 (2011) 179–193

Zhang, D., X. Zhou, S. C. H. Leung, and J. Zheng (2010). Vertical bagging decision trees
model for credit scoring. Expert Systems with Applications 37(12), 7838– 7843.

13
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Bank lending and small and medium sized enterprises’ access to finance
– effects of macroprudential policies†
(Work-in-progress – not for publishing)

(June, 2020)

Aida Ćehajićab, Marko Košaka

a
School of Economics and Business, University of Ljubljana, Ljubljana, Slovenia
b
School of Economics and Business, University of Sarajevo, Sarajevo, Bosnia and Herzegovina

ABSTRACT: This paper examines the impact of macroprudential regulation on small and medium
sized firms’ access to bank finance, by using survey firm level data, covering European Union
countries over the period from 2009 to 2017, and newly published database on use of
macroprudential policies in European Union countries, MaPPED, which enabled us to take into
account both tightening and loosening actions of macroprudential policy separately. The findings
show significant relationship between macroprudential indices and firm’s access to finance. The
tightening actions of macroprudential policy are mostly associated with the restrained access to
bank loans for SMEs, while considering different firm and country level characteristics. While
usage of macroprudential regulation could result in higher capitalization and increase the resilience
of the financial sector, and decrease its risk exposures, this could also result in credit constrained
firms, especially smaller firms with fewer possibilities of finance and high dependence on bank
loans.

Keywords: Macroprudential regulation, bank lending, small and medium sized enterprises, access
to finance
JEL classification: E51, E58, G21
1. Introduction and literature review
The role of banks in financial intermediation as providers of external financing for firms is vital
for the healthy function of economic systems. This role is extremely important factor in firm

†
The authors would like to thank Matej Marinč, Miroslav Verbič, Saša Žiković, the participants of EBR Conference
2019: “Managing Business and Policies in a Changing Global Landscape”, School of Economics and Business,
University of Ljubljana and the participants of IFABS 2019 Angers Conference, France, for their valuable suggestions
and comments. We also thank European Central Bank for the permission to use the SAFE data. Lastly, the authors
thankfully ackonowledge valuable funding provided by Erasmus Mundus Project Green Tech WB. The views and
opinions expressed in this paper are entirely those of the authors. All errors remain our own.

Corresponding author at: University of Sarajevo, School of Economics and Business, Sarajevo 71 000, Bosnia and
Herzegovina; University of Ljubljana, Faculty of Economics, Ljubljana 1000, Slovenia; E-mail addresses:
[Link]@[Link] (A. Ćehajić); [Link]@[Link] (M. Košak)

14
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

growth, investments and ultimately economic growth. Perfect capital markets in Modigliani, &
Miller (1958) world make the capital structure of firm irrelevant to its value, which means that the
funding sources are not relevant for investments which firm undertakes. In real world on the other
hand, there are various market imperfections which make the capital structure decisions relevant
and give premium to external sources of funds. These imperfections include transaction costs
incurred by the finance provider, agency costs, tax preference of debt, and imperfect information
between the parties leading to the problem of information asymmetries. External financing for
firms, apart from transaction cost it carries, can also bring high costs and become unavailable to
firms due to existence of asymmetric information, making investors underinformed about the
firms’ future and wary of underling risk, and as put in Akerlof, (1970) not being able to distinguish
between a lemon and a good firm.
As access to funds and their price has the primary role in the small and medium sized firm’s
business, and considering their overall dependence on bank-based financing, the literature on their
financing obstacles and opportunities is a vital ingredient to both practitioners and academic
debates.
The financial crisis of 2008, followed by eurozone debt crisis has lowered the enterprises’ ability
to access bank financing, upon which the vast majority of small and medium sized firms in Europe
is highly dependent, (European Banking Authority, 2016; Campello, Graham, & Harvey, 2010;
Chava & Purnanandam, 2011; Aiyar, Al-Eyd, Barkbu, & Jobst, 2015; Acharya & Steffen, 2015;
Popov & Udell, 2012). Aiyar, et al. (2015) emphasize that small and medium sized enterprises
account for the substantial part of employment and production in Europe, while greatly rely on
bank loans for funding. Bank financing of Eurozone firm’s accounts for 80 percent of their debt,
(Berger & Udell, 2006). Majority of these firms have experienced increased cost of funding or
have been constrained in the aftermath of the financial crisis, (Santos, 2011). The crisis has induced
both: i) lowering of credit demand and ii) constrained loan supply to firms, especially small and
medium sized enterprises. While there are many previous studies which explained the impact of
the financial crisis on bank lending to firms (Jin, Luo, & Wan, 2018; Santos, 2011; and Tsuruta,
2015), and the determinants of the bank lending constraints on firms, (Behr, Foos, & Norden, 2017;
Becker & Ivashina, 2014; Beck, Demirgüç-Kunt, & Pería, 2011; Dell’Ariccia, Igan, & Laeven,
2008; Beck, Demirgüç-Kunt, Laeven, & Maksimovic, 2006), there are few research papers which
have analyzed post crisis regulatory reforms effects on bank lending to SMEs. It is safe to say that
majority of these reforms have been of macroprudential nature.
Macroprudential policy’s focus is on establishing a macro protection apparatus, whose primary
goal is establishing and maintaining the resilience of the financial sector and its agents, while also
smoothing bank procyclical behavior, (Lim et al., 2011). Emphasis on macroprudential policies
came as a result of regulators’ intention in strengthening of the financial system’s resilience, and
stabilizing credit cycles, (Claessens, Ghosh, & Mihet, 2013; ESRB, 2014). On the other side
consequences of regulatory changes can also be reflected through curtailment of lending supply,
erosion of economic activity and can have undesired effects of increased unemployment and

15
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

decline in investment activities, (Duygan-Bump, Levkov, & Montoriol-Garriga, 2015; Cosimano

& Hakura, 2011; Francis & Osborne, 2012).
The severity of the financial crisis of 2008, has resulted in the increased use of macroprudential
instruments in European countries, while in this period there were more policy actions which aim
to tighten credit conditions, than the loosening ones and previous literature has shown that
macroprudential regulation has substantially affected banking and its activities, (Cerutti,
Claessens, & Laeven, 2015; Claessens, Ghosh, & Mihet, 2013, Lim et al., 2011, Olszak,
Roszkowska, & Kowalska, 2018, Morgan, Regis, & Salike, 2018). Macroprudential policy is
designed and implemented through the use of different macroprudential measures which target
both, asset and liability side of bank balance sheets, thus impacting bank’s exposure to certain
risks. In a cross-country study Borio & Shim (2008), suggest that macroprudential policy is
efficient in decreasing loan growth in the subsequent years after the initiation of the instruments.
Crowe, Dell’Ariccia, Igan, & Rabanal (2013), investigate the effects of macroprudential measures
such as capital requirements, dynamic provisioning, loan-to-value and debt-to-income limits. They
conclude that loan-to-value measures can help in controlling the boom periods while dynamic
provisioning could moderate loan cycles. Other extensive cross-country studies on
macroprudential policy effects, suggest that macroprudential measures are successful in taming
credit growth and decreasing bank risk taking. The literature suggests that macroprudential
instruments can reduce the frequency of loan booms and lower the possibility of financial
instability and financial crisis, (Lim et al., 2011; Akinci & Olmstead-Rumsey, 2018; Crowe et al.,
2013; Cerutti, Claessens, & Laeven, 2017).
Our paper examines whether macroprudential policies are associated with small and medium sized
firms’ access to bank funding, and empirically examine this relationship, while controlling for
different firm level and macroeconomic determinants. This study is focused on the impact of
different macroprudential instruments implemented in European Union countries on the
availability of bank loans to small and medium sized enterprises, while considering firm level
characteristics, and different macroeconomic factors bound to countries of these firms, by using
survey level data on firm financing conditions (SAFE) conducted by European Central Bank, and
new macroprudential database, MaPPED, gathered and published by European Central Bank,
supplemented with macroeconomic data extracted from ECB and World Bank databases.
Our contribution to the firm’s financing and macroprudential literature is threefold. First, to our
knowledge, this is the first paper to test the impact of macroprudential regulation on the firms’
access to finance by using firm level survey data. Second, we combine this data with newly
published database on macroprudential policy and tools focused exclusively on European Union
countries, which provides detailed information on the scope of the instruments, policy actions, and
direction of the measures.1 We evaluate how macroprudential policies affect the extent to which
banks supply credit to borrowers (SMEs), and clarify which are the factors which cause firms to

1
See Budnik et al., 2018.

16
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

be credit constrained and how does bank regulation influence the willingness of banks to provide
financing to the enterprises. Third, considering the limited evidence in the literature, by using
micro level data, regarding the effects of macroprudential tools on firms, this paper also extends
this work by identifying the relationship between macroprudential regulatory instruments and
small and medium sized enterprises financing, by testing tightening and loosening actions of
macroprudential policy separately. Notwithstanding the vast literature on the financial obstacles
firm face, this paper bears additional contribution by using survey data which enables taking
information on financing constraints directly from firms’ loan applications. The information of
whether a firm applied, was declined or approved a loan and the characteristics associated with the
firm are highly informative and bear considerable predictive power.
2. Literature review and hypothesis development
The existence of asymmetric information means the complete information about the firm is only
available to firm managers and not the external agents. This can result from nondisclosure of
information which can have undesired effect of informing competitors or simply come as a result
of superior professional knowledge of management when compared to external investors, in terms
of what the information means for the firm, Myers and Majluf (1984). If a lender cannot
differentiate good and bad borrowers, this can lead to increase in loan rates or unavailability of
finance for some borrowers, (Fazzari, et al. 1987; Stiglitz, & Weiss, 1981). Furthermore, Calomiris
et al. (1986) show that credit rationing is more associated with firms reliant on external bank
financing, due to asymmetric information, dependence on collateral, sometimes not easily
available to firms, bank strength and borrower’s cash flow. As the asymmetry of information
grows, so do the costs of external financing and its availability. Seminal papers by Fazzari, et al.
(1987), & Gertler (1988) suggest that due to lack of relationship history with their creditors,
information asymmetry is more present with small and newer firms.
Later works have found that smaller firms have more difficulties in obtaining finance, than larger
firms, (Schiffer and Weder (2001); Beck et al. (2006); Ghosh et al. (1999), due to difficulties in
determining their risk of defaults, lack of collateral, and high monitoring costs for banks. In a
recent paper by Wang et al. (2019) authors stress that SMEs can deviate from sound behavior after
the loan has been granted, resulting in moral hazard problem and increase in credit risk for the
lender. Due all of these imperfections, the SME loan market has been struggling to obtain
necessary funds and due to its importance to the entire economy, is significant part of discussion
within both academic and institutional communities.
There are different strands in literature when measuring the firm access to funds and financing
constraints. Fazzari et al (1988) paper has proposed measuring firm constraints by using balance
sheet data and in particular the investment-cash flow sensitivity which serves as a way to sort
financially constrained firms, to the ones which are cash flow sensitive on a higher degree and the
ones which are not constrained, with less sensitivity in terms of cash flow. The logic behind this
approach is that financially constrained firms are dependent on internal source of funds. The paper
suggests that firms with low dividend payout ratio are more cash flow sensitive, and rely more on

17
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

internal financing, than higher payout ratio firms. On the other hand, Kaplan and Zingales (1997)
question this, and show that sensitivity of cash flow and investments could be heightened with
when financing constraints decrease. They also differentiate firms by using the managers’
statements in annual reports and using the information related wih difficulty in meeting the debt
payment owed by the firm, or liquidity issues. Similar findings were confirmed by Almeida and
Campello (2004) with emphasis on financial and macroeconomic circumstances in which firms
operate. This was followed by studies measuring firm financing constraint by differentiating firms
as in Fazzari et al (1988) to constrained and unconstrained based on different firm characteristics
such as age and size (Devereux & Schiantarelli (1990), sector, the market value of debt relative
to market value of firm, interest expense and bond rating (Whited, (1992)).
Unlike these studies, as in Beck et al (2006), we choose to use survey data, SAFE - collected by
ECB and European Commission directly from firms with information on loan applications and
difficulties in obtaining finance they face.
There have been several papers which have already examined the financing and borrowing
conditions of European firms by using SAFE survey data, but there were no papers developed
around the impact of macroprudential policy.2 Regarding availability of the bank loans to
enterprises, a paper by Casey & O’Toole (2014), shows that credit rationed firms will probably
turn to informal market for a loan. Bank loan constraints increase with firm’s size and decline with
firm’s age. More leveraged firms and the ones with impaired credit history are found to have a
bigger necessity to turn to loans from other companies or informal loans, as they cannot easily
obtain bank financing. Similarly, Ferrando, Popov, & Udell (2017), investigate the impact of
sovereign stress on SME’s access to finance. The authors find that sovereign stress has substantial
negative impact on availability of finance for SMEs, and that firms in stressed economies are likely
to use government subsidies as other mode of financing. The authors also point out that European
debt crisis has also had a strong negative impact on bank lending, which resulted in constrained
firms and employment cuts. The paper by Adelino & Ferreira (2016), suggests that when banks
face liquidity demands, the lending to companies may be impaired, even when firm’s
characteristics remain the same. Banks with credit rating which succeeds the sovereign
downgrades decrease their loan supply to greater extent than those banks which, due to credit

2
Bremus & Neugebauer (2018), investigate how the decline in bank cross border activities in the period between 2010
- 2014 has impacted the cost of financing of European enterprises in the aftermath of the financial crisis. The reduction
in cross-border banking has had a substantial effect on the likelihood of the rise of cost of funding for smaller
enterprises. Mascia & Rossi, (2017) investigate the impact of gender of the firm’ manager effect on the cost of
financing. The authors found that firms with female manager are more likely to experience higher financing costs,
while controlling for various firm and country characteristics, suggesting such firms as a result have worsened credit
conditions and lesser possibility of obtaining bank loans. Garcia-Posada Gomez (2018) paper examines credit
constraints, and their impact of firm’s investments and growth. By using survey data, the results show that firm’s
being credit constrained have important negative impact on firm investments and growth. In the paper by Mayordomo
& Rodríguez-Moreno (2018), it is suggested that SME’s supporting factor which enables banks to cut regulatory
requirements on their capital positions associated with their exposures to credit risk to small and medium sized firms,
decreased the credit rationing of medium sized enterprises but not micro firms.

18
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

rating policies, are not at the bound of sovereign rating. This lending cuts could be related with
difficulties in obtaining different funding sources, and consequently with higher funding costs.3
Ferrando, Popov, & Udell (2019), explore whether unconventional monetary policies introduced
by ECB increased the likelihood of SMEs to obtain financing. The authors find that unconventional
monetary policies announcements resulted in an increased likelihood of firms obtaining loans,
increasing firm’s profitability. Similar results on the impact of expansionary monetary policies
were found by Jin et al. (2018).
Albertazzi & Marchetti (2010), & Berger and Udell (1994) documented that larger banks with
lower capitalization ratios restricted loans to riskier firm borrowers, while the same effect was not
evident for smaller banks with lower capital levels. Similarly, Barth, et al. (2013), found that higher
capitalization regulatory demands are associated with lower level of non-performing loans in bank
assets thus decreasing bank riskiness, and improving asset quality. But in order to satisfy
regulatory requirements, financial institutions might cut excessive lending or charge the borrowers
higher loan rates, (Kahou and Lehar, 2016). In a study by Bole, Prašnikar, & Trobec (2014), for
Slovenian firms, authors show that post crisis regulatory interventions, in terms of increased capital
requirements in a short period of time, have resulted in a cut of bank lending to firms, heavily
increasing the firm bankruptcy numbers. Carlson, Shan, & Warusawitharana (2013), show that in
the period of downturns higher capitalization requirements substantially affect lending volumes,
with banks cutting lending to borrowers, while in normal times these requirements are not likely
to produce such effects. On the other hand, Bridges et al. (2014), show that higher capital
requirements cause lending reduction to firms and households and that loan growth normalizes
only after three years. Authors also point out the necessity of investigating potential similarities
to these effects resulting from macroprudential measures and how they might affect lending. These
contributions to the literature provide an insightful discussion and point out the necessity of
investigating the potential trade-off between financial stability and economic development.
Our paper extends the study by Ayyagari, Beck, & Martinez Peria (2017), which represents one of
the first attempts to measure the extent of the impact of macroprudential regulation on firm’s loan
availability. By using firm level data for 59 countries, in the period 2002 to 2011, the authors found
that younger and smaller enterprises are substantially affected by macroprudential tools usage.
They also point out that borrower-based measures have stronger effects than financial institutions
targeted measures. The authors stress the importance of the understanding of the trade-off in the
usage and deepening of financial regulation by usage of macroprudential tools.
Unlike the previously mentioned paper (Ayyagari et al., 2017), we use survey based data, focused
on EU countries, and macroprudential data published by the ECB researchers, which contains vast
information on the use of macroprudential policies and policies with macroprudential nature in EU
countries with availability of data from 1995 up to 2017. We merge these two datasets, while also
controlling for several macroeconomic characteristic for the period of 2009-2017. This period

3
See also Correa, Sapriza, & Zlate (2013).

19
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

reflects the period when macroprudential policies were most actively used, (Cerutti et al., 2017).
We differentiate borrower related instruments, such as limits on credit growth, loan to value cap,
debt to income cap, caps on foreign currency loans, liquidity targeting measures, such as reserve
requirements, caps on maturity and currency mismatches, and capital tools, covering profit
allocation limits, dynamic provisioning, or countercyclical capital requirements. The rest of
macroprudential instruments captured by the database, we group together in index Other.
We develop four hypotheses that we will examine in this paper. Firstly, we test whether
macroprudential measures curtail the credit availability for SMEs.
H1. Macroprudential measures are associated with deterioration in bank lending to SMEs.
Secondly, we test whether the effects of macroprudential policy on availability of loans for SMEs
depend on the direction of these measures. Hence, we form two hypotheses:
H2a. When macroprudential policy is tightened, SMEs are less likely to receive bank funding.
H2b. When macroprudential measures are loosened, SMEs are more likely to obtain bank
financing.
By following previous papers which determined the importance of size of firms when applying for
a bank loan, we form the last hypothesis:
H3. Macroprudential policy effects on firms’ access to bank credit decrease as the firm’s size
grows.
In order to test these hypotheses, we exploit data provided by SAFE which enables us to control
for various firm characteristics such as firm’s size, age, use of debt, turnover level, credit history
and ownership. The choice of these variables is supported by the literature, (Berger & Udell, 2006;
& Mascia & Rossi, 2017). The MaPPED database gives us a possibility to construct different
macroprudential indices based on the policy actions captured in the database. We measure the
impact of macroprudential policies, first by overall macroprudential index, by summing up policy
actions over waves of the survey. We further divide this index by the target of the macroprudential
instruments, so for the instruments which target borrowers, we construct Borrower index, if the
instruments are targeting liquidity positions of financial institutions, we construct Liquidity index.
For macroprudential policy actions targeting capital positions of the banks, we construct Capital
index and for the remaining macroprudential policy actions, we construct index Other.4 In this way,
we are able to evaluate through which channels macroprudential policy implementation can affect
lending to SMEs.

Our findings support the hypothesis that SMEs are less likely to get bank financing, when
regulators enforce macroprudential policies. We find the strongest relationship for overall
macroprudential index and capital macroprudential index. Our findings, when testing the direction

4
See section 2.2 for more information on the construction of the indices, and the Appendix A, Table 3. for more
information on the macroprudential policies which are covered by MaPPED database.

20
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of macroprudential measures, show that when macroprudential instruments are tightened, small
and medium sized firms are less likely to receive funding. On the contrary, when macroprudential
instruments are loosened, SMEs are more likely to access bank financing. Our results hold when
performing different robustness checks.

The remainder of this study is organized as follows. Section 2 describes the data and the empirical
framework of the study. Section 3 represents the main results and finding of the paper, while fourth
section represents further extensions and robustness checks. The last, fifth section serves as a
conclusion of the study.

2. Data and methodology

Our firm level data comes from the restricted Survey on the access to finance of enterprises (SAFE)
obtained by the permission of European Central Bank. SAFE was initiated in 2009, in the
aftermath of the financial crisis, and is conducted every six months, for a number of euro area
countries, the surveys known as ECB rounds. The larger survey, conducted once a year, is run with
the European Commission, the Common round, which covers all EU countries and some other
European countries. Firms in the database are randomly selected from the Dun & Bradstreet
database and stratified by country, economic activity and companies’ size. The survey provides a
vast amount of qualitative data on the financial access for enterprises in euro area, and European
union as well. In this paper, we focus only on countries which are part of EU, for which we have
available information.5 Apart from SAFE, in order to capture the information on implementation
and changes in macroprudential regulations, we use a newly collected and publicly available
database on the implementation of macroprudential policies, focused on EU countries, with
information over a long time period, covering the scope and intensity of the instruments.
Additionaly, we control for different macroeconomic determinants by using data available at World
Bank and IMF’s International Financial Statistics.
2.1. Firm level data
SAFE has been initiated in 2009, with two waves conducted in that year. Wave 1 of 2009 covers
the period from January 2009 – June 2009, while wave 2 covers the period from July 2009 –
December 2009. Wave 3 has the reference period from March-September 2010 and wave 4 from
September 2010-February 2011. For all other waves (wave 5 – wave 18), first wave of each survey
covers the second and third quarter of the current year, and second wave of each survey covers the
fourth quarter of the current year and first quarter of the following year.6 We adapt our
macroeconomic variables to these reference periods in order to be able to measure their impact
accurately. We use all available waves, both ECB and Common rounds, and since we want to

5
Further information on the survey are available at:
[Link]
6
For more information on data reference period, see Appendix, Table 1.

21
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

measure access to bank financing, we include in the analysis firms which have applied for bank
loan or a credit line, by following their responses to the questionnaire. We end up with 52,189
observations for 36,990 firms, with most of firms in our dataset surveyed only once, although we
capture and analyze the panel component of the firms, by analyzing the subset of firms with at
least two observations in the database. The SAFE provides data stratified by country, and firm size
measured by number of employees, following this division: a) micro (from 1 employee to 9
employees), b) small (from 10 employees to 49 employees), c) medium (from 50 employees to
249 employees), and d) large (250 employees or more).7 The sample is further stratified by main
activity, age and ownership. Additionally, SAFE provides weights based on the size of the firms,
measured by number of employees through firm’s size, activity and country, which we include in
majority of our regressions.
2.2. Macroeconomic and macroprudential data
Apart from macroeconomic data obtained from International Financial Statistics, ECB data
warehouse, and partly from World Bank and national sources, we also use the data set based on
detailed survey on macroprudential policy actions in EU countries, provided by European Central
Bank, (Budnik, & Kleibl, 2018). The database, titled Macroprudential Policies Evaluation database
(MaPPED)8, is based on a survey conducted by ECB, and the data is collected with the assistance
of national central banks and regulators in EU countries. It provides the information on prudential
policies and tools of macroprudential nature in all EU member states from 1995 to 2017. The
dataset covers information on 1700 prudential actions, nested in 11 categories of instruments by
their target and purpose, and subdivided in 53 sub-categories of policy instruments.9 It captures
the introduction of an instrument, the changes in the intensity the instrument in the following
period and the cancelation of an instrument. These are policy actions, captured as loosening,
tightening or neutral. We are able to measure these changes in policy actions over time and
possibility to construct cumulative indices. In this process we follow several approaches in the
existing literature, (see Kuttner & Shim, 2016; Altunbas, Binici, & Gambacorta, 2018; Akinci &
Olmstead-Rumsey 2018; Lim et al., 2011).
Considering the MaPPED dataset has the information on the legal character of the measure:
whether a policy action is installed as a legally binding measure or recommendation of the
regulator, we choose to exclude recommended policy actions, and include in our analysis only
binding measures, in order to gain more accurate estimates. We are able to capture the information
on the intensity of macroprudential tools with three possible responses on the direction of the
policy action: 1) policy tightening, 2) policy loosening, and 3) other and with ambiguous impact.
As the database has quarterly information of any change in the policy instrument, we start by
denoting the policy actions with dummy variables: -1 for loosening actions, 1 for tightening actions

7
The number of large firms in SAFE is limited, and these firms were added in our sample in order to provide an
opportunity for comparison of conditions for large versus SME enterprises.
8
Database available at: [Link]
9
In order to see a list of instruments covered by the MaPPED database, see Appendix, Table 3.

22
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and 0 for no change. (Budnik et al., 2018; Altunbas et al., 2018). We mark “other and ambiguous”
with 0 as well, as we do not have enough information on the scope and intensity of the change,
and we want to measure the effects of loosening and tightening of macroprudential instruments on
credit availability in the small and medium firms’ market. These action changes are captured at
quarterly level, and we start by coding quarterly changes in policy actions over our sample period
2009-2017.
Since the frequency of firm level data used in our study is semiannual, in order to capture the
correct information and be able to merge the data, we semi-annualize the data from
macroprudential database in order to capture the changes in the period which corresponds to
waves. We do this by summing policy actions over the quarters. This also means that policy actions
with different signs can sum up to 0, or as specified in Altunbas et al., 2018: “cancel each other,
leaving no net change”. We do the same for macroeconomic variables, for which we took annual
data based on averaged quarterly data corresponding to each survey (wave) round (see Appendix
1), as in (Ferrando et al., 2017; Mascia & Rossi, 2017; Bremus & Neugebauer, 2018). Firstly, we
differentiate macroprudential tools with same target and divide them into categories: a) capital
related instruments (minimum capital requirements, capital buffer, risk weights, provisioning), b)
borrower targeted instruments (lending caps, reserve requirements), c) liquidity related
instruments (loan to deposit ratios, funding requirements, liquidity coverage ratios) , and d) other
measures (taxes, exposures, other). Secondly, we construct 4 indices based on this subdivision,
which are summed across all countries in the sample. This means the summed index can take value
of -1, 0, 1, -2, 2 or up to -6 or 6 or more. We also construct an index which captures all
macroprudential instruments: MPP, which can take value, in our database, from -3 to 6.
Additionally, by following the approach in (Cerutti, Correa, Fiorentino, & Segalla, 2016) we
construct cumulative indices, based on overall intensity of macroprudential policy in our sample,
which means we sum up the dummies denoting the instruments, by country and wave. In the case
of these indices, all groups of instruments are adjusted to have maximum and minimum changes
of 1 and -1. Finally, we construct two indicators, which are based on the direction of the measures.
We measure tightening and loosening separately, by constructing dummies which equal 1 for
overall stance of macroprudential policy as being tightening or loosening, and 0 otherwise, (see
Altunbas et al., 2018).
2. 3. Empirical model
To examine the effects of macroprudential policies on small and medium sized firms financing,
we specify the following probability model:
∗
Pr⁡(𝑦𝑖,𝑘,𝑡 = 1) = 𝐹(𝛽0 + 𝛽1𝐹𝐼𝑅𝑀𝑖,𝑘,𝑡 + 𝛽2 𝑀𝑃𝑃𝑘,𝑡 +𝛽3 𝑀𝐴𝐶𝑅𝑂𝑘,𝑡 + 𝛽4 𝐶𝑅𝐼𝑆𝐼𝑆𝑘,𝑡 + 𝜀𝑖,𝑘,𝑡 )

⁡where index i refers to the firm, k to the country and t to the time period. Yi,k,t is the loan
application outcome, which is a dichotomous variable denoted with dummy variable which equals
1 if a firm applied for a loan and was approved, and 0 otherwise. FIRMi,k,t includes firm level
characteristics: size, age, leverage, profit and ownership. MPPk,t represents a vector of

23
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

macroprudential policy index; MACROk,t is a vector of macroeconomic variables which include

real GDP growth rate, unemployment rate in a respective country, while the CRISISk,t denotes a
dummy variable which captures the effects of the financial crisis, which takes value of 1 for the
waves 1-7 (2009-2012) and 0 otherwise; εj,k,t represents the error term. As our estimation method
we use probit with robust standard errors, and for a subset of firms with available panel for more
than two waves, we use panel probit with random effects and robust standard errors. In majority
of our regressions we also include weights10, and country, wave, and sector control variables, in
order to capture unobserved heterogeneity and time effects in our data.11

2. 4. Variables definition

Access to finance (y*) is the dependent variable. As a measure of access to finance, we use a
dummy variable which equals 1 if the firm has applied and was approved a loan or a credit line by
the bank, otherwise 0. We follow the approach by (Moro, Wisniewski, & Mantovani, 2017). Firm
level determinants include the information on the firm’s size captured by two variables, Small
which takes value of 1 if the firm has from 1 to 49 employees, and zero otherwise. 12 Since small
firms usually are more financially constrained than large firms (Petersen & Rajan, 1994; Beck et
al., 2005; Beck & Demirguc-Kunt, 2006; Mascia & Rossi, 2017), in order to capture that effect,
we also include Large dummy, taking value of 1 if firm has more than 250 employees. We measure
the effects of firm’s age by including the variable Age, which has the value of 1 if the firm is older
than 10 years and 0 otherwise. Beck et al., (2006) find that financing constraints are inversely
associated to firm’s age. As age increases financing obstacles are decreased.

In order to control for firm’s risk taking and credit history, by following (Petersen & Rajan, 1994;
Petersen et al, 1995; & Casey & O’Toole, 2014) we also control for several other firm
characteristics. In order to capture the firm’s level of use of debt financing, we include Leverage
dummy, equaling 1 if firm has increased the use of debt in the last six months. The survey provides
the information on the credit history of the firms, by inclusion on question whether the firm’s credit
history improved, or deteriorated. We capture this information by inclusion of a dummy Credit
history which equals 1 if firms credit history improved in the last six months. Additionally, in
order to control for profitability and economic prospects of the firm, we include variable Profit,
which informs whether the firms’ profits increased in the last six months. The variable equals 1 if

10
Weight variable which restores the economic weight is given in SAFE, and economic weight is constructed by
economic activity, country to which the firm belongs, and its size. The number of employees serves as a measures of
economic weight. For more information on weighting variable, see Methodological information on the survey and
user guide for the anonymised micro dataset, available at:
[Link]
58739e26506dfb36e24899f5c0
11
We follow the approach in Beck et al., 2011; Mascia & Rossi, 2017; Bremus & Neugebauer, 2018.
12
The SAFE provides the size decomposition in the following way: 1-9 employees - micro firm, 10-49 employees-
small firm, 50 – 249 employees – medium firm, and over 250 emloyees – large firm. The 0 in our dummy SMALL
means that a firm is of medium size.

24
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the firm’s profit increased and 0 if the firm’s profit decreased in the last six months. In line with
(Beck et al., 2006 & Mayordomo & Rodríguez-Moreno, 2018), in order to assess the impact of the
firm ownership on the availability of bank finance, we include two variables, namely Autonomous,
which equals 1 if the firm in an autonomous profit-oriented enterprise, and 0 if the firm is a
subsidiary, branch or part of another enterprise, and Ownership whose value of 1 indicates a firm
is owned by an individual, family, or entrepreneurs, and zero otherwise, meaning the owned by
shareholders or other businesses. Additionally, we include two variables which are available only
for those firms which have applied for bank loan: Collateral, which is a dummy variable equal to
1 if the bank decreased the amount of collateral required by the firms when applying for bank loan,
and 0 otherwise, and Maturity, which is an indicator equal to 1 if bank increased the maturity of
loans, and 0 otherwise.

Macroprudential variables include several indices which enable the prediction of probability of
firms obtaining a bank loan. The indices include MPP, Capital, Borrower, Liquidity, and Other,
which are our base measures for macroprudential policy stance.13 We construct indices by their
target, by following the similar approach as in several prior studies which examined the effects of
macroprudential policies, (Lim et al., 2011; Altunbas et al., 2018; Akinci & Olmstead-Rumsey,
2018). These indices represent the sum of macroprudential instruments introduced or loosened in
a certain country for a certain wave. We additionally include cumulative indices capturing the
overall macroprudential stance, which can take value of -1, 0 and 1, namely MPP_cumulative,
Capital_cumulative, Borrower_cumulative, Liquidity_cumulative, and Other_cumulative. Finally,
we include indices through which we capture whether the overall stance of macroprudential policy
in a country in a certain wave is overall regulatory easing, or overall tightening. By following
(Ferrando, Popov, & Udell, 2019) for macro level controls we include real GDP growth, as a proxy
for credit demand and economic conditions, while we also control for Unemployment rate in a
country. Finally, we include the Crisis dummy which equals 1 for waves over 2009-2012, in order
to control for periods of stress.
Table 1
Summary statistics of the main regression variables.

(1) (2) (3) (4) (5)

Variables Observations Mean SD Min Max

Dependent variable

Access to finance 52,189 0.795 0.403 0 1

Macroprudential indices
MPP 52,189 0.251 0.766 -3 6
Capital 52,189 0.165 0.540 -2 3
Borrower 52,189 0.024 0.192 -1 3

13
For more detail on the construction of indices and which instruments are included, please consult the Appendix,
Table 3.

25
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Liquidity 52,189 0.026 0.182 -1 2

Other 52,189 0.035 0.474 -3 3
MPP_cumulative 52,189 0.162 0.484 -1 1
Capital_cumulative 52,189 0.120 0.401 -1 1
Borrower_cumulative 52,189 0.021 0.154 -1 1
Liquidity_cumulative 52,189 0.023 0.154 -1 1
Other_cumulative 52,189 0.028 0.380 -1 1
MPP_loosening 52,189 0.050 0.217 0 1
Capital_loosening 52,189 0.028 0.164 0 1
Borrower_loosening 52,189 0.002 0.042 0 1
Liquidity_loosening 52,189 0.001 0.026 0 1
Other_loosening 52,189 0.059 0.235 0 1
MPP_tightening 52,189 0.211 0.408 0 1
Capital_tightening 52,189 0.147 0.354 0 1
Borrower_tightening 52,189 0.023 0.149 0 1
Liquidity_tightening 52,189 0.023 0.151 0 1
Other_tightening 52,189 0.087 0.281 0 1

Firm determinants
Small 52,189 0.435 0.496 0 1
Large 52,189 0.380 0.485 0 1
Age 52,189 0.812 0.391 0 1
Leverage 52,189 0.300 0.458 0 1
Profit 52,189 0.329 0.470 0 1
Credit history 52,189 0.307 0.461 0 1
Autonomous 52,189 0.862 0.345 0 1
Ownership 52,189 0.697 0.460 0 1
Collateral 52,189 0.048 0.213 0 1
Maturity 52,189 0.099 0.298 0 1

Macroeconomic determinants

GDP growth (%) 52,189 0.908 2.579 -14.79 23.04

Unemployment (%) 52,189 10.371 5.718 2.88 27.83

HHI index 52,142 0.066 0.049 0.021 0.388

Crisis 52,189 0.365 0.481 0 1

Source: Authors’ calculations based on estimation sample of our base regressions, which includes summary statistics
for all variables used in our estimations. The summary statistics is based on estimates which include weights available
in SAFE, based on size of the firm (number of employees).14

3. Empirical results and findings

14
See Ferrando et al. (2017) & Mayordomo & Rodriguez-Moreno (2017).

26
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

This section of the paper represents the base results of our estimations. First, we test the effects of
macroprudential policy on SMEs access to bank finance, by grouping different macroprudential
measures into 5 indices based upon the target of these measures: MPP, Capital, Borrower,
Liquidity and Other. Furthermore, we include country, year and sector dummies in order to control
for heterogeneity within our sample. Finally, we test different indices of macroprudential policies,
based on different construction of the measures and their directions, and check whether there is a
substantial change in our results.

3.1. Base results

Table 2 represents our main findings, estimated with pooled probit with robust standard errors on
a sample of 36,990 firms, with majority of the firms surveyed only once. By following previous
studies, we focus on marginal effects (columns 6-10) of probability of firm obtaining a bank loan,
which are more convenient to interpret. We start with interpretation of our key variables, and we
find that small and medium sized firms in countries with greater use of macroprudential policies
are more likely to be rejected when applying for a bank loan. Marginal effect of MPP index,
aggregate macroprudential index, indicates that firms are 1.4% less likely to obtain bank loan. The
results are similar for sub-indices as well, with Capital index’ marginal effect indicating that firms
are around 1.5% less likely to receive bank funding. We find even stronger results with borrower
and liquidity targeting measures, with Borrower index marginal effect indicating that firms are less
likely to access bank financing at 5% level, relative to Liquidity index indicating marginal effects
of 4.5%. These results are statistically significant at 1% level. Index Other is not statistically
significant in this specification. These findings support our first hypothesis, by showing that
macroprudential measures are associated with bank lending behavior towards SMEs. Similar
findings were established in Ayyagari et al. (2017).

Considering firm level characteristics, we first interpret the size of the firm measured with two
variables: small and large. As expected, and similar to (Beck, Demirguc-Kunt, Laeven, & Levine,
2008; Beck & Demirguc-Kunt, 2006; Mayordomo & Rodríguez-Moreno, 2018), in table 3, we
find that small firms are 7.6% less likely to obtain bank loan. On the contrary we find that large
firms are around 4% more likely to obtain bank funds and the results are statistically significant
on 1% level. Considering the Age dummy, we find that firms which are older than 10 years or
more are around 5% more likely to get the loan approval. If a firm increased use of debt in the last
six months, this indicates that the firm is less likely to access bank funds at 2% level. If firm’s
turnover increased, the firm is around 2% more likely to obtain loan financing. Looking at the
variable indicating improvement in firm’s credit history, we find that there is 4.4% likelihood that
a bank loan for a firm will be approved. Both our variables for ownership, namely autonomous
and ownership, are not statistically significant. If collateral amount was decreased by the bank in
the last six months, the firms are 4.5% more likely to receive bank funding. When maturity of the
loan is increased by the bank, the firms are 5.6% more likely to be approved for a bank loan.
Moving on to the interpretation of our macroeconomic variables, we find that firms in countries

27
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

with an increase in GDP growth rate are at 1.3% more likely to obtain bank loan financing, while
the increase in unemployment rate could result in firms being less likely at 1% to have successful
loan application. Crisis dummy in this specification is not statistically significant.

28
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2
The effect of macroprudential policy on SMEs access to finance

Dependent variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Access to finance MPP Capital Borrower Liquidity Other MPP Capital Borrower Liquidity Other
Coefficients Marginal effects

Small -0.291*** -0.290*** -0.291*** -0.291*** -0.291*** -0.0759*** -0.0759*** -0.0758*** -0.0758*** -0.0758***
(0.0178) (0.0178) (0.0178) (0.0178) (0.0178) (0.00483) (0.00483) (0.00484) (0.00484) (0.00484)
Large 0.141*** 0.141*** 0.142*** 0.141*** 0.141*** 0.0381*** 0.0381*** 0.0382*** 0.0382*** 0.0381***
(0.0277) (0.0277) (0.0277) (0.0277) (0.0277) (0.00774) (0.00774) (0.00774) (0.00774) (0.00774)
Age 0.196*** 0.196*** 0.196*** 0.195*** 0.196*** 0.0480*** 0.0480*** 0.0479*** 0.0477*** 0.0478***
(0.0214) (0.0214) (0.0214) (0.0214) (0.0214) (0.00565) (0.00565) (0.00565) (0.00565) (0.00565)
Leverage -0.0929*** -0.0926*** -0.0922*** -0.0939*** -0.0935*** -0.0234*** -0.0234*** -0.0234*** -0.0235*** -0.0235***
(0.0192) (0.0192) (0.0192) (0.0192) (0.0191) (0.00504) (0.00504) (0.00504) (0.00504) (0.00504)
Profit 0.0666*** 0.0669*** 0.0671*** 0.0682*** 0.0672*** 0.0217*** 0.0219*** 0.0217*** 0.0218*** 0.0217***
(0.0220) (0.0220) (0.0220) (0.0220) (0.0220) (0.00591) (0.00591) (0.00590) (0.00590) (0.00591)
History 0.177*** 0.177*** 0.177*** 0.176*** 0.177*** 0.0437*** 0.0437*** 0.0438*** 0.0438*** 0.0438***
(0.0219) (0.0219) (0.0219) (0.0219) (0.0219) (0.00582) (0.00582) (0.00582) (0.00582) (0.00582)
Autonomous 0.0185 0.0191 0.0164 0.0189 0.0194 0.00442 0.00441 0.00427 0.00439 0.00435
(0.0327) (0.0327) (0.0327) (0.0327) (0.0327) (0.00870) (0.00869) (0.00869) (0.00869) (0.00869)
Ownership 0.0106 0.00808 0.0107 0.00724 0.00777 -0.00109 -0.00117 -0.00106 -0.00112 -0.00110
(0.0253) (0.0253) (0.0253) (0.0253) (0.0253) (0.00681) (0.00681) (0.00681) (0.00681) (0.00681)
Collateral 0.156*** 0.157*** 0.156*** 0.157*** 0.157*** 0.0446*** 0.0443*** 0.0444*** 0.0444*** 0.0444***
(0.0471) (0.0470) (0.0471) (0.0470) (0.0470) (0.0127) (0.0126) (0.0127) (0.0127) (0.0127)
Maturity 0.192*** 0.192*** 0.191*** 0.194*** 0.193*** 0.0563*** 0.0563*** 0.0563*** 0.0564*** 0.0564***
(0.0324) (0.0325) (0.0325) (0.0325) (0.0325) (0.00862) (0.00863) (0.00862) (0.00862) (0.00862)
MPP -0.0523*** -0.0139***
(0.0114) (0.00302)
Capital -0.0572*** -0.0152***
(0.0171) (0.00453)
Borrower -0.208*** -0.0552***
(0.0413) (0.0110)
Liquidity -0.167*** -0.0445***
(0.0380) (0.0101)
Other -0.000200 -0.0000533
(0.0187) (0.00497)
GDP growth 0.0505*** 0.0510*** 0.0501*** 0.0511*** 0.0506*** 0.0134*** 0.0136*** 0.0133*** 0.0136*** 0.0135***
(0.00414) (0.00412) (0.00412) (0.00412) (0.00412) (0.00110) (0.00110) (0.00110) (0.00110) (0.00110)
Unemployment -0.0197*** -0.0193*** -0.0194*** -0.0194*** -0.0192*** -0.00523*** -0.00514*** -0.00517*** -0.00516*** -0.00511***
(0.00151) (0.00150) (0.00150) (0.00150) (0.00150) (0.000391) (0.000390) (0.000389) (0.000389) (0.000389)
Crisis -0.0228 -0.0197 -0.0242 -0.0152 -0.0179 -0.00607 -0.00524 -0.00645 -0.00403 -0.00478
(0.0212) (0.0212) (0.0212) (0.0212) (0.0213) (0.00565) (0.00564) (0.00565) (0.00565) (0.00566)
Observations 52,189 52,189 52,189 52,189 52,189 52,189 52,189 52,189 52,189 52,189

29
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Pseudo R squared15 0.0617 0.0614 0.0618 0.0614 0.0609 0.0617 0.0614 0.0618 0.0596 0.0591
Note: This table reports pooled probit with robust standard errors estimation results for different macroprudential indices. The estimation period is January 2009-September 2017 (wave 1 – wave 17).
Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or a credit line and was approved, and 0 if a firm applied and was rejected. MPP is an index which represents a sum of
all policy changes over waves in a particular country. Capital index is a sum of all macroprudential measures targeting capital positions of banks. Borrower index is a sum of all macroprudential measures
targeting borrowers. Liquidity is an index which sums up all policy actions over the course of waves whose primary target is the enhancement of liquidity positions in bank. Other is an index, which is a
sum of the remaining macroprudential measures. See Table 3 in the Appendix for more information on the construction of macroprudential indices. See Table 4 in the Appendix for all variable’s definitions.
Robust standard errors in parentheses. All regressions include sampling weights. ***indicate p<0.01, **indicate p<0.05, and * indicates p<0.1.

15
McFadden's pseudo R squared values are usually substantially lower compared to ordinary R squared index, and should not be looked on with the same norm of
good fit, as in common regression analysis, (McFadden, 1978).

30
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In all subsequent regressions in Tables 3 to 7 we include control dummies for countries, time
effects, and sector to which the firm belongs. Sector denotes the main activity of the firm, the firm
belonging to industry, construction, trade or services. Table 4 shows our main results after we enter
control dummies. The MPP and Capital index are both significant and of the expected sign as in
our base results, while Borrower and Other index are no longer significant with respect to SME’s
access to finance. The MPP index is statistically significant with marginal effects of 1%, while
Capital index with 1.8%. This indicates that capital measures are most significant in their
association with bank funding available to small and medium sized firms.

Table 3.
The effects of different macroprudential policy indices on SMEs access to finance – controlling
for country, time and sector.
Dependent variable (1) (2) (3) (4) (5)
Access to finance MPP Capital Borrower Liquidity Other

Small -0.0759* -0.0759* -0.0758* -0.0758* -0.0758***

(0.00483) (0.00483) (0.00484) (0.00484) (0.00484)
Large 0.0381*** 0.0381*** 0.0382*** 0.0382*** 0.0381***
(0.00774) (0.00774) (0.00774) (0.00774) (0.00774)
Age 0.0480*** 0.0480*** 0.0479*** 0.0477*** 0.0478***
(0.00565) (0.00565) (0.00565) (0.00565) (0.00565)
Leverage -0.0234*** -0.0234*** -0.0234*** -0.0235*** -0.0235***
(0.00504) (0.00504) (0.00504) (0.00504) (0.00504)
Profit 0.0217*** 0.0219*** 0.0217*** 0.0218*** 0.0217***
(0.00591) (0.00591) (0.00590) (0.00590) (0.00591)
History 0.0437*** 0.0437*** 0.0438*** 0.0438*** 0.0438***
(0.00582) (0.00582) (0.00582) (0.00582) (0.00582)
Autonomous 0.00442 0.00441 0.00427 0.00439 0.00435
(0.00870) (0.00869) (0.00869) (0.00869) (0.00869)
Ownership -0.00109 -0.00117 -0.00106 -0.00112 -0.00110
(0.00681) (0.00681) (0.00681) (0.00681) (0.00681)
Collateral 0.0446*** 0.0443*** 0.0444*** 0.0444*** 0.0444***
(0.0127) (0.0126) (0.0127) (0.0127) (0.0127)
Maturity 0.0563*** 0.0563*** 0.0563*** 0.0564*** 0.0564***
(0.00862) (0.00863) (0.00862) (0.00862) (0.00862)
MPP -0.0104***
(0.00330)
Capital -0.0178***
(0.00530)
Borrower -0.00800
(0.0117)
Liquidity -0.00871
(0.0115)
Other -0.00178
(0.00561)
GDP growth 0.00489*** 0.00472*** 0.00509*** 0.00510*** 0.00511***
(0.00152) (0.00152) (0.00152) (0.00152) (0.00152)
Unemployment -0.00589*** -0.00591*** -0.00578*** -0.00577*** -0.00583***
(0.00146) (0.00146) (0.00146) (0.00147) (0.00146)
Crisis 0.0137 0.00979 0.0103 0.00942 0.0101
(0.0165) (0.0164) (0.0165) (0.0164) (0.0165)
Country dummies YES YES YES YES YES

31
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Wave dummies YES YES YES YES YES

Sector dummies YES YES YES YES YES
Observations 52,189 52,189 52,189 52,189 52,189
Pseudo R squared 0.0834 0.0835 0.0831 0.0831 0.0831
Note: This table reports marginal effects after pooled probit estimations for different macroprudential indices. The estimation period is January
2009-September 2017 (wave 1 – wave 17). Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or a credit line and
was approved, and 0 if a firm applied and was rejected. MPP is an index which represents a sum of all policy changes over waves in a particular
country. Capital index is a sum of all macroprudential measures targeting capital positions of banks. Borrower index is a sum of all macroprudential
measures targeting borrowers. Liquidity is an index which sums up all policy actions over the course of waves whose primary target is the
enhancement of liquidity positions in bank. Other is an index, which is a sum of the remaining macroprudential measures. See Table 3 in the
Appendix for more information on the construction of macroprudential indices. See Table 4 in Appendix for all variable’s definitions. Robust
standard errors in parentheses. All regressions include sample weights, and country, time, and sector dummies. *** indicate p<0.01, **indicate
p<0.05, and * indicates p<0.1.

3.2. Different indices of macroprudential policies

In order to test whether the construction of our indices impacts the results, we construct different
indices based on our base indices used in aforementioned regressions. We construct cumulative
index which equals to -1 if an overall level of MPP aggregate macroprudential policy index in a
country in a specific wave is negative, indicating dominance of loosening macroprudential actions,
and +1 if an overall level of MPP aggregate index is of positive sign, suggesting the overall stance
of macroprudential regulations in the country in the observed period was to a large extent
dominated with the tightening actions. Zero indicates no change. Since all our control variables
are of similar signs and significance level, we comment only macroprudential indices. In this
specification our main MPP index is statistically significant, but on 10% level, and marginal effects
indicate that an increase in this index leads to firms being less likely at 9.3% to obtain bank loan.
We find similar results for capital-based index, which is statistically significant on 5% level, and
marginal effects indicate that when this index is increased by one unit, the firms are less likely by
1.5% to obtain bank loan. Other indices of macroprudential instruments are not statistically
significant. This reaffirms our base findings, as we again confirm that macroprudential measures
which target capital positions of banks are strongly related with SMEs access to finance.

Table 4
The impact macroprudential policy on access to finance for SMEs - cumulative indices
Dependent variable (1) (2) (3) (4) (5)
Access to finance MPP Capital Borrower Liquidity Other

Small -0.0758* -0.0758* -0.0758* -0.0758* -0.0758***

(0.00484) (0.00483) (0.00483) (0.00484) (0.00484)
Large 0.0381*** 0.0381*** 0.0382*** 0.0382*** 0.0381***
(0.00774) (0.00774) (0.00774) (0.00774) (0.00774)
Age 0.0479*** 0.0479*** 0.0479*** 0.0478*** 0.0479***
(0.00565) (0.00565) (0.00565) (0.00565) (0.00565)
Leverage -0.0234*** -0.0234*** -0.0234*** -0.0235*** -0.0235***
(0.00504) (0.00504) (0.00504) (0.00504) (0.00504)
Profit 0.0217*** 0.0218*** 0.0218*** 0.0218*** 0.0217***
(0.00591) (0.00591) (0.00590) (0.00591) (0.00591)
History 0.0437*** 0.0437*** 0.0438*** 0.0438*** 0.0438***
(0.00583) (0.00582) (0.00582) (0.00582) (0.00582)
Autonomous 0.00448 0.00440 0.00425 0.00438 0.00436

32
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

(0.00869) (0.00869) (0.00869) (0.00869) (0.00869)

Ownership -0.00117 -0.00117 -0.00103 -0.00111 -0.00111
(0.00681) (0.00681) (0.00681) (0.00681) (0.00681)
Collateral 0.0445*** 0.0443*** 0.0444*** 0.0444*** 0.0444***
(0.0127) (0.0126) (0.0127) (0.0127) (0.0127)
Maturity 0.0564*** 0.0563*** 0.0563*** 0.0564*** 0.0564***
(0.00862) (0.00863) (0.00862) (0.00862) (0.00862)
MPP_cumulative -0.00913*
(0.00508)
Capital_cumulative -0.0150**
(0.00694)
Borrower_cumulative -0.0213
(0.0151)
Liquid_cumulative -0.00837
(0.0146)
Other_cumulative -0.00342
(0.00698)
GDP growth 0.00511*** 0.00501*** 0.00508*** 0.00510*** 0.00513***
(0.00152) (0.00152) (0.00152) (0.00152) (0.00152)
Unemployment -0.00565*** -0.00563*** -0.00571*** -0.00579*** -0.00581***
(0.00147) (0.00146) (0.00146) (0.00147) (0.00146)
Crisis 0.0121 0.00956 0.00987 0.00949 0.0105
(0.0165) (0.0164) (0.0164) (0.0164) (0.0165)
Country dummies YES YES YES YES YES
Wave dummies YES YES YES YES YES
Sector dummies YES YES YES YES YES
Observations 52,189 52,189 52,189 52,189 52,189
Pseudo R squared 0.0832 0.0832 0.0832 0.0831 0.0831
Note: This table reports marginal effects after pooled probit estimations for different macroprudential indices. The estimation period is January
2009-September 2017 (wave 1 – wave 17). Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or a credit line and
was approved, and 0 if a firm applied and was rejected. MPP_cumulative is an index which represents a sum of all policy changes over waves in a
particular country, standardized to a minimum of -1 and maximum of 1. Capital_cumulative index is a sum of all macroprudential measures targeting
capital positions of banks, standardized to a minimum of -1 and maximum of 1. Borrower_cumulative index is a sum of all instrument actions
targeting borrowers, standardized to a minimum of -1 and maximum of 1. Liquidity_cumulative is an index which sums up all policy actions over
the course of waves whose primary target is the enhancement of liquidity positions in bank, standardized to a minimum of -1 and maximum of 1.
Other_cumulative is an index, which is a sum of the remaining measures, standardized to a minimum of -1 and maximum of 1. See Table 3 in the
Appendix for more information on the construction of macroprudential indices. See Table 4 in Appendix for all variable’s definitions. Robust
standard errors in parentheses. All regressions include sample weights, country, time, and sector dummies. *** indicate p<0.01, **indicate p<0.05,
and * indicates p<0.1.

Table 5 shows the results obtained with pooled probit regression by inclusion of macroprudential
policy indices which are obtained by measuring the overall stance of macroprudential policy. We
establish the overall stance of macroprudential policy in a country by constructing policy indices
in the following way: the dummy variables for these indices indicated with a “tightening” equal 1
if the sum of overall actions in the respective countries observed over each wave is of positive sign
indicating the dominance of tightening actions, leaving 0 otherwise. On the contrary, dummies
marked with “loosening” equal 1 if the sum of overall actions is of negative sign, indicating the
dominance of easing actions in the observed period, and 0 otherwise. The estimation results
indicate, as we expected, the negative coefficient with the tightening actions, but only significant
for capital index. The tightening of capital macroprudential measures decreases the probability of
firms access to bank funding by 1.5%. We received positive marginal effects with the loosening
dummies in the case of MPP_loosening (2%), Capital_loosening (3.4%) and Other_loosening
(2.2%) index. This means when macroprudential regulations are tightened, SMEs are less likely
to access bank financing. On the contrary, when macroprudential actions are loosened, small and

33
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

medium sized firms are more likely to be obtain bank loan. This supports our H2 and H3
hypotheses.

34
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 5
Macroprudential policy effects on financing for SMEs – tightening policy Macroprudential policy effects on financing for SMEs – loosening policy
actions actions
Dependent variable (1) (2) (3) (4) (5) Dependent variable (6) (7) (8) (9) (10)
Access to finance MPP Capital Borrower Liquidity Other Access to finance MPP Capital Borrower Liquidity Other

Small -0.0758*** -0.0758*** -0.0758*** -0.0758*** -0.0758*** Small -0.0758*** -0.0758*** -0.0758*** -0.0758*** -0.0758***
(0.00484) (0.00483) (0.00483) (0.00484) (0.00484) (0.00484) (0.00483) (0.00484) (0.00484) (0.00483)
Large 0.0381*** 0.0381*** 0.0382*** 0.0382*** 0.0382*** Large 0.0382*** 0.0382*** 0.0381*** 0.0382*** 0.0382***
(0.00774) (0.00774) (0.00774) (0.00774) (0.00774) (0.00774) (0.00774) (0.00774) (0.00774) (0.00774)
Age 0.0479*** 0.0479*** 0.0479*** 0.0478*** 0.0478*** Age 0.0478*** 0.0478*** 0.0478*** 0.0478*** 0.0479***
(0.00565) (0.00565) (0.00565) (0.00565) (0.00565) (0.00565) (0.00565) (0.00565) (0.00565) (0.00565)
Leverage -0.0234*** -0.0234*** -0.0234*** -0.0235*** -0.0234*** Leverage -0.0234*** -0.0235*** -0.0234*** -0.0235*** -0.0234***
(0.00504) (0.00504) (0.00504) (0.00504) (0.00504) (0.00504) (0.00504) (0.00504) (0.00504) (0.00504)
Profit 0.0217*** 0.0218*** 0.0218*** 0.0218*** 0.0218*** Profit 0.0217*** 0.0218*** 0.0217*** 0.0218*** 0.0217***
(0.00591) (0.00590) (0.00591) (0.00591) (0.00591) (0.00590) (0.00591) (0.00590) (0.00590) (0.00590)
History 0.0438*** 0.0438*** 0.0438*** 0.0438*** 0.0439*** History 0.0437*** 0.0437*** 0.0439*** 0.0438*** 0.0438***
(0.00582) (0.00582) (0.00582) (0.00582) (0.00582) (0.00582) (0.00582) (0.00582) (0.00582) (0.00582)
Autonomous 0.00439 0.00440 0.00420 0.00439 0.00436 Autonomous 0.00452 0.00434 0.00431 0.00435 0.00451
(0.00869) (0.00869) (0.00869) (0.00869) (0.00869) (0.00869) (0.00869) (0.00869) (0.00869) (0.00869)
Ownership -0.00112 -0.00116 -0.001000 -0.00111 -0.00110 Ownership -0.00120 -0.00113 -0.00109 -0.00108 -0.00114
(0.00681) (0.00681) (0.00681) (0.00681) (0.00681) (0.00681) (0.00681) (0.00681) (0.00681) (0.00681)
Collateral 0.0445*** 0.0444*** 0.0444*** 0.0444*** 0.0443*** Collateral 0.0444*** 0.0443*** 0.0444*** 0.0444*** 0.0446***
(0.0127) (0.0127) (0.0127) (0.0127) (0.0127) (0.0126) (0.0126) (0.0127) (0.0127) (0.0127)
Maturity 0.0564*** 0.0564*** 0.0562*** 0.0564*** 0.0564*** Maturity 0.0563*** 0.0561*** 0.0564*** 0.0564*** 0.0563***
(0.00862) (0.00862) (0.00862) (0.00862) (0.00862) (0.00862) (0.00863) (0.00862) (0.00862) (0.00862)
MPP_tightening -0.00827 MPP_loosening 0.0203*
(0.00675) (0.0111)
Capital_tightening -0.0146* Capital_loosening 0.0342*
(0.00827) (0.0182)
Borrower_tightening -0.0259 Borrower_loosening -0.0195
(0.0161) (0.0381)
Liquidity_tightening -0.0101 Liquidity_loosening -0.0437
(0.0150) (0.0576)
Other_tightening 0.00775 Other_loosening 0.0220*
(0.0103) (0.0121)
GDP growth 0.00511*** 0.00507*** 0.00508*** 0.00509*** 0.00508*** GDP growth 0.00509*** 0.00497*** 0.00511*** 0.00508*** 0.00521***
(0.00152) (0.00152) (0.00152) (0.00152) (0.00152) (0.00152) (0.00152) (0.00152) (0.00152) (0.00152)
Unemployment - - - - - Unemployment - - - - -
0.00568*** 0.00568*** 0.00571*** 0.00578*** 0.00579*** 0.00577*** 0.00570*** 0.00583*** 0.00579*** 0.00573***
(0.00147) (0.00146) (0.00146) (0.00147) (0.00146) (0.00146) (0.00146) (0.00146) (0.00146) (0.00146)
Crisis 0.0121 0.0101 0.0102 0.00947 0.00719 Crisis 0.00891 0.00829 0.00970 0.00946 0.00932
(0.0166) (0.0164) (0.0164) (0.0164) (0.0167) (0.0164) (0.0164) (0.0164) (0.0164) (0.0164)
Country dummies YES YES YES YES YES Country dummies YES YES YES YES YES
Wave dummies YES YES YES YES YES Wave dummies YES YES YES YES YES
Sector dummies YES YES YES YES YES Sector dummies YES YES YES YES YES
Observations 52,189 52,189 52,189 52,189 52,189 Observations 52,189 52,189 52,189 52,189 52,189
Number of firms 36,990 36,990 36,990 36,990 36,990 Number of firms 36,990 36,990 36,990 36,990 36,990
Pseudo R-squared 0.0831 0.0832 0.0832 0.0831 0.0831 Pseudo R-squared 0.0832 0.0832 0.0831 0.0831 0.0832

35
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Note: This table reports marginal effects after pooled probit estimations for macroprudential indices based on their intensity. The columns 1-5 represent the estimations with tightening
indicators, which is equal to 1 if the overall stance of macroprudential policy is dominantly tightening based on the sign of the base index. The tightening indicators are also separately
estimated for each index. The columns 6-10 represent the estimations with loosening indicator, which is equal to 1 if the overall stance of macroprudential policy is dominantly
loosening based on the sign of the base index. The loosening indicators are also separately estimated for each index. The estimation period is January 2009-September 2017 (wave
1 – wave 17). Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or a credit line and was approved, and 0 if a firm applied and was rejected.
MPP_tightening is an index which represents a sum of all policy changes over waves in a particular country, and equal to 1 if the sign of the base index is positive, and 0 otherwise.
Capital_tightening index is a sum of all macroprudential measures targeting capital positions of banks, and equal to 1 if the sign of the base index is positive, and 0 otherwise.
Borrower_tightening index is a sum of all instrument actions targeting borrowers, and equal to 1 if the sign of the base index is positive, and 0 otherwise. Liquidity_tightening is an
index which sums up all policy actions over the course of waves whose primary target is the enhancement of liquidity positions in bank, and equal to 1 if the sign of the base index
is positive, and 0 otherwise. Other_tightening is an index, which is a sum of the remaining measures, and equal to 1 if the sign of the base index is positive, and 0 otherwise.
MPP_loosening is an index which represents a sum of all policy changes over waves in a particular country, and equal to 1 if the sign of the base index is negative, and 0 otherwise.
Capital_loosening index is a sum of all macroprudential measures targeting capital positions of banks, and equal to 1 if the sign of the base index is negative, and 0 otherwise.
Borrower_loosening index is a sum of all instrument actions targeting borrowers, and equal to 1 if the sign of the base index is negative, and 0 otherwise. Liquidity_loosening is an
index which sums up all policy actions over the course of waves whose primary target is the enhancement of liquidity positions in bank, and equal to 1 if the sign of the base index
is negative, and 0 otherwise. Other_loosening is an index, which is a sum of the remaining measures, and equal to 1 if the sign of the base index is negative, and 0 otherwise. See
Table 3 in the Appendix for more information on the construction of macroprudential indices. See Table 4 in Appendix for all variable’s definitions. Robust standard errors in
parentheses. All regressions include sample weights, country, time, and sector dummies. *** indicate p<0.01, **indicate p<0.05, and * indicates p<0.1.

36
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.3. Marginal effects by size and age

Due to significance of size and age when it comes to credit access for firms, evident in literature
(Beck et al., 2006; Casey & O’Toole, 2014), as well as in our results, we choose to additionally
analyze the effects of macroprudential policies with respect to firms’ size and age. Considering the
heterogeneity of firms in our data, we expect to find different degree of macroprudential policy
effect, considering different firm groups based on size and age.16

In Table 6, we show the results of our marginal effects of estimating the effect of macroprudential
policy on access to finance for different size and age characteristics of the firms in our sample. We
expect that effect on smaller firms should be larger, since smaller firms are more reliant on bank
loans and have fewer alternative sources for finance, (Chava & Purnanandam, 2011; Beck et al.,
2006). As expected, we find some support for our fourth hypothesis, as we find that the effect is
larger for smaller firms. The marginal effect is around 1.3% for small firms, while the effect for
large firms is significantly smaller around 0,8%. It is also evident that the effect of macroprudential
policy diminishes with firm size. When it comes to age, we find small variation in terms of effects
of macroprudential policy, but we can confirm that the effect is shrinking as the firms’ age
increases. The marginal effect for start-ups is 1.09%, while for firms older than 10 years 1.03%.

Table 6
Marginal effects of macroprudential policy on access to bank finance by firms’ size and age
Dependent variable
Access to finance
MPP index
Size Age
Micro -0.01325*** <=10 years or more -0.01033***
(0.0042075) (0.0032771)
Small -0.01173*** 5-9 years -0.01052***
(0.0037239) (0.0033403)
Medium -0.010*** 2-4 years -0.01072***
(0.003173) (0.003404)
Large -0.0082*** >2 years -0.01092***
(0.0026054) (0.0034679)
Note: This table reports marginal effects after probit estimations with random effects for different macroprudential
aggregate index (MPP) with different classes of firms’ age and size. The estimation period is January 2009-September
2017 (wave 1 – wave 17). Robust standard errors in parentheses. All regressions include sample weights, country,
time, and sector dummies. *** indicate p<0.01, **indicate p<0.05, and * indicates p<0.1

16
The firms in SAFE are divided by size following this division: a) micro (from 1 employee to 9 employees), b) small
(from 10 employees to 49 employees), c) medium (from 50 employees to 249 employees), and d) large (250 employees
or more).
We also inspected grups of firms with respect to ownership, but these results were not significant and we do not report
them.

37
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4. Extensions and robustness tests

4.1. Heckman correction

As our sample includes only firms which have applied for bank loan, we are concerned about
potential selection bias arising from excluding those firms which did not apply for bank loan, either
because they were discouraged because of the rejection expectations, or due to the fact that some
firms did not have the need for loan application as they have enough funds. We resolve this issue
by using Heckman correction (Heckman, 1979), which is able to reduce any bias arising from
truncated data. Heckman selection procedure consists of two equations, namely outcome equation
which includes all our independent variables as in our base regressions, while the selection
equation includes one more additional variable, which serves as an exclusion restriction. The
exclusion restriction is the variable which explains the probability of being observed, in our case,
credit demand, but does not affect the outcome, credit supply.
We follow Ferrando et al., 2017, & Bremus & Neugebauer, 2018, we choose Competition as our
selection variable. The variable is available in SAFE database and equals 1 if the competition is
the most pressing problem the firm is facing, and 0 otherwise. We follow the same logic as in
aforementioned studies: if the firm is facing high competition, and lower sales and profitability,
followed by insufficient liquid funds, it might have higher need for bank financing. We expect,
that the need for bank financing does not affect the bank’s decision to approve the loan, so it has
effect on credit demand, but not on credit supply, which fulfils the exclusion restriction
requirements.
Table 6 shows results for probit and probit with heckman correction. Column 1 represents the main
result, the probability of credit access based on our base probit estimation. Column 2 represents
the probability of credit access in probit with sample selection setting. Column 3 represents the
estimations with heckman selection and the probability of access to credit conditional upon the
dependent variable being observed. Column 4 represents the probability of dependent variable
being observed. Our exclusion restriction variable, Competition is significant and of positive sign,
indicating that the firms having considerable competition problems, display positive demand for
bank financing. The variable, as we expected, is not significant in estimations capturing the
probability of credit access (column 3). The impact of the size of the firm is similar for both,
access and demand for credit, although in the case of the credit demand we observe this effect is
stronger. Smaller firm are 11% more likely to demand credit, than other firms. We also observe
that large firms are 7.3% more likely to obtain bank funding. The age is not significant for credit
demand, while it stays significant in our outcome equation with older firms being 4% more likely
to get credit access. Leverage enters significantly in both, access and demand equations, but with
different signs. The firms with an increase in leverage are 20% more likely to display credit
demand, while they are also 2% less likely to have successful credit access. Increased profit and
improved credit history enter significantly and with positive signs for credit access as well as
demand, while our ownership variables are only significant in credit demand setting. The
autonomous and firm owned by family or entrepreneurs are more likely on 11% and 3% to display
higher credit demand. MPP index is significant and of expected sign in all columns. We also

38
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

confirm that in countries with tightening macroprudential policies, the SMEs are around 1% more
likely to be credit constrained, and as well to be discouraged to apply. The significance and signs
of macroeconomic determinants are similar to our base results, while these variables do not enter
significantly for credit demand.
Table 6
The effects of macroprudential policy on access to funds for SMEs – Probit with sample correction
Dependent variable (1) (2) (3) (4)
Probit Probit with Probit with Probit with
Access to finance Access sample selection sample selection sample selection
Access Access Demand17

Competition 0.000622 0.0140***

(0.000922) (0.00501)
Small -0.0759*** -0.0663*** -0.0758*** -0.111***
(0.00483) (0.0134) (0.00484) (0.00402)
Large 0.0381*** 0.0326*** 0.0381*** 0.0730***
(0.00774) (0.00984) (0.00775) (0.00654)
Age 0.0480*** 0.0448*** 0.0480*** -0.000456
(0.00565) (0.00678) (0.00565) (0.00487)
Leverage -0.0234*** -0.0298*** -0.0232*** 0.197***
(0.00504) (0.00943) (0.00501) (0.00468)
Profit 0.0217*** 0.0197*** 0.0217*** 0.0147***
(0.00591) (0.00614) (0.00591) (0.00476)
History 0.0437*** 0.0389*** 0.0437*** 0.0461***
(0.00582) (0.00818) (0.00582) (0.00480)
Autonomous 0.00442 -0.000572 0.00437 0.112***
(0.00870) (0.0103) (0.00870) (0.00692)
Ownership -0.00109 -0.00226 -0.00108 0.0303***
(0.00681) (0.00656) (0.00681)
Collateral 0.0446*** 0.0417*** 0.0446***
(0.0127) (0.0126) (0.0127)
Maturity 0.0563*** 0.0526*** 0.0563***
(0.00862) (0.00944) (0.00862)
MPP -0.0139*** -0.00915*** -0.0103*** -0.0123***
(0.00302) (0.00346) (0.00330) (0.00274)
GDP growth 0.0134*** 0.00461*** 0.00489*** -0.000975
(0.00110) (0.00148) (0.00152) (0.00125)
Unemployment -0.00523*** -0.00547*** -0.00588*** -0.000620
(0.000391) (0.00148) (0.00146) (0.00125)
Crisis -0.00607 0.0128 0.0138 0.00275
(0.00565) (0.0155) (0.0165) (0.0141)
Year dummies YES YES YES YES
Country dummies YES YES YES YES
Sector dummies YES YES YES YES
Observations 52,189 136,764 136,764 136,764
Note: This table reports marginal effects after probit and heckprobit estimations with MPP index as our variable of
interest. The estimation period is January 2009-September 2017 (wave 1 – wave 17). Access to finance is our
dependent variable, equal to 1 if a firm applied for a loan or a credit line and was approved, and 0 if a firm applied

17
Variables Collateral and Maturity are available only for those firms which applied for bank loan, and as such are
not part of estimation for the demand, (Column 4).

39
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and was rejected. MPP is an index which represents a sum of all policy changes over waves in a particular country.
Competition is our exclusion restriction variable which equals 1 if the firm report the competition as its most pressing
problem and 0 otherwise. See Table 3 in the Appendix for more information on the construction of macroprudential
indices. See Table 4 in Appendix for all variable’s definitions. Robust standard errors in parentheses. All regressions
include country, time, and sector dummies. ***indicate p<0.01, **indicate p<0.05, and * indicates p<0.1.

4.2. Panel component

SAFE database provides firm identifier for each wave of the survey, which we used to identify
firms which were surveyed at least twice in the observed period. We exclude all firms which have
only one observation, and we obtain an unbalanced panel of 9,712 firms over 17 waves. The
following results (Table 7) were obtained with panel probit with random effects with robust
standard errors. The results show statistically significant coefficients for MPP and Capital index,
which indicates that a tightening of macroprudential policy causes SMEs less likely to obtain credit
at 1% (MPP) and 1.4% (Capital) level. These results again confirm that credit access reacts
strongly to capital macroprudential measures being imposed by the regulators. Results for our
other firm and macroeconomic variables remain similar as to our main results, with expected signs
and significance. Borrower, Liquidity and Other index are not significant in this specification.

Table 7
The effects of macroprudential policy on access to funds for SMEs – panel component
Dependent variable (1) (2) (3) (4) (6)
Access to finance MPP Capital Borrower Liquidity Other

Small -0.0666* -0.0666* -0.0667* -0.0667* -0.0666***

(-8.86) (-8.84) (-8.86) (-8.86) (-8.85)
Large 0.0382** 0.0381** 0.0379** 0.0381** 0.0379**
(3.03) (3.02) (3.00) (3.02) (3.00)
Age 0.0501*** 0.0502*** 0.0499*** 0.0497*** 0.0500***
(6.49) (6.50) (6.46) (6.44) (6.47)
Leverage -0.0220*** -0.0221*** -0.0221*** -0.0221*** -0.0221***
(-3.91) (-3.92) (-3.93) (-3.93) (-3.93)
Profit 0.0167** 0.0168** 0.0169** 0.0169** 0.0168**
(2.73) (2.75) (2.76) (2.76) (2.76)
History 0.0558*** 0.0559*** 0.0560*** 0.0560*** 0.0560***
(9.02) (9.03) (9.05) (9.05) (9.05)
Autonomous -0.00189 -0.00188 -0.00208 -0.00205 -0.00201
(-0.18) (-0.18) (-0.20) (-0.19) (-0.19)
Ownership -0.0127 -0.0129 -0.0128 -0.0128 -0.0128
(-1.58) (-1.60) (-1.59) (-1.59) (-1.58)
Collateral 0.0417** 0.0419** 0.0421** 0.0418** 0.0419**
(3.13) (3.15) (3.16) (3.15) (3.15)
Maturity 0.0439*** 0.0440*** 0.0444*** 0.0443*** 0.0443***
(4.72) (4.73) (4.77) (4.77) (4.76)
MPP -0.0102**
(-3.03)
Capital -0.0142**
(-2.88)
Borrower -0.00831
(-0.61)
Liquidity -0.0175
(-1.47)

40
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Other -0.00558
(-0.99)
GDP growth 0.00298 0.00294 0.00294 0.00298 0.00294
(1.83) (1.82) (1.82) (1.84) (1.81)
Unemployment -0.00951*** -0.00960*** -0.00938*** -0.00920*** -0.00945***
(-5.27) (-5.32) (-5.20) (-5.08) (-5.23)
Crisis 0.00479 0.00251 -0.000184 -0.00119 0.000660
(0.28) (0.15) (-0.01) (-0.07) (0.04)
Country dummies YES YES YES YES YES
Wave dummies YES YES YES YES YES
Sector dummies YES YES YES YES YES
Observations 26,901 26,901 26,901 26,901 26,901
Number of firms 9,712 9,712 9,712 9,712 9,712
Note: This table reports marginal effects after panel probit estimations with random effects for different macroprudential indices. The estimation
period is January 2009-September 2017 (wave 1 – wave 17). Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or
a credit line and was approved, and 0 if a firm applied and was rejected. MPP is an index which represents a sum of all policy changes over waves
in a particular country Capital index is a sum of all macroprudential measures targeting capital positions of banks. Borrower index is a sum of all
instrument actions targeting borrowers. Liquidity is an index which sums up all policy actions over the course of waves whose primary target is the
enhancement of liquidity positions in bank. Other is an index, which is a sum of the remaining measures. See Table 3 in the Appendix for more
information on the construction of macroprudential indices. See Table 4 in Appendix for all variable’s definitions. Robust standard errors in
parentheses. All regressions include country, time, and sector dummies. ***indicate p<0.01, **indicate p<0.05, and * indicates p<0.1.

41
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4.3. Subset of countries present in all waves

In order to check whether our results are affected by choice of sample which includes all waves of
survey, we choose to reestimate our model in a setting which includes only firms from countries
which were surveyed in all waves (both ECB rounds and Common rounds). By doing so, we are
able to check the strength of our results when choosing the countries within the database with
highest number of firms and degree of information available. These countries include eleven
eurozone countries: Austria, Belgium, Germany, Spain, Finland, France, Greece, Ireland, Italy,
Netherlands and Portugal. The results of these estimations are given in Table 9. As we received
very similar results, as in our main regressions, for our control variables, and macroeconomic
determinants, we comment only on macroprudential variables. We again confirm the significance
of MPP and capital index. Firms in these countries are 1.2% (MPP) and 2% (Capital) more likely
to be credit constrained due to implementation of macroprudential measures, especially capital
targeting macroprudential instruments. Furthermore, we analyze tightening and loosening
macroprudential policies separately. We receive similar results as with total sample of firms.
Capital_tightening and Borrower_tightening indices marginal effects indicate that when these
macroprudential measures are tightened, the firms are 1.5% and 3.3% less likely to access bank
credit. In the case of loosening actions, we find that if macroprudential measures are loosened, the
SMEs are 3.1% (MPP_loosening), 5.5% (Capital_loosening) and 2.4% (Other_loosening) less
likely to access bank funding.

We further analyze the panel of these countries, by limiting sample only to firms which have at
least two observations in the covered period. The results remain very similar, as expected, as the
sample is dominated by the firms from these countries. The results are stronger in panel component
of the sample, with firms being 1% more likely to be credit constrained due to implementation of
macroprudential measures and 1.5% due to increase in index of capital macroprudential measures,
which target capital positions of banks.

42
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 8
The effects of macroprudential policies on access to funding for SMEs - subset of countries present
in all waves
Dependent variable (1) (2) (3) (4) (5)
Access to finance MPP Capital Borrower Liquidity Other

Small -0.0753* -0.0752* -0.0751* -0.0751* -0.0751***

(0.00535) (0.00535) (0.00535) (0.00535) (0.00535)
Large 0.0375*** 0.0374*** 0.0375*** 0.0376*** 0.0375***
(0.00852) (0.00852) (0.00852) (0.00852) (0.00852)
Age 0.0468*** 0.0470*** 0.0468*** 0.0467*** 0.0468***
(0.00622) (0.00622) (0.00622) (0.00622) (0.00622)
Leverage -0.0266*** -0.0266*** -0.0266*** -0.0267*** -0.0267***
(0.00545) (0.00544) (0.00544) (0.00544) (0.00544)
Profit 0.0209*** 0.0211*** 0.0209*** 0.0210*** 0.0209***
(0.00661) (0.00661) (0.00661) (0.00661) (0.00661)
History 0.0461*** 0.0462*** 0.0463*** 0.0464*** 0.0463***
(0.00647) (0.00647) (0.00647) (0.00647) (0.00647)
Autonomous 0.00296 0.00284 0.00259 0.00288 0.00281
(0.00975) (0.00974) (0.00975) (0.00975) (0.00975)
Ownership 0.00400 0.00389 0.00415 0.00399 0.00401
(0.00761) (0.00761) (0.00762) (0.00761) (0.00761)
Collateral 0.0592*** 0.0589*** 0.0591*** 0.0590*** 0.0590***
(0.0142) (0.0142) (0.0142) (0.0142) (0.0142)
Maturity 0.0495*** 0.0495*** 0.0494*** 0.0496*** 0.0495***
(0.00950) (0.00952) (0.00951) (0.00951) (0.00951)
MPP -0.0122***
(0.00359)
Capital -0.0194***
(0.00574)
Borrower -0.0171
(0.0136)
Liquidity -0.00910
(0.0127)
Other -0.00205
(0.00661)
GDP growth 0.00603*** 0.00566*** 0.00614*** 0.00620*** 0.00622***
(0.00170) (0.00170) (0.00170) (0.00170) (0.00171)
Unemployment -0.00548*** -0.00555*** -0.00544*** -0.00538*** -0.00546***
(0.00161) (0.00161) (0.00161) (0.00162) (0.00161)
Crisis 0.0135 0.00830 0.0103 0.00845 0.00929
Country dummies YES YES YES YES YES
Wave dummies YES YES YES YES YES
Sector dummies YES YES YES YES YES
Observations 41,848 41,848 41,848 41,848 41,848
Number of firms 25,956 25,956 25,956 25,956 25,956
Pseudo R-squared 0.0820 0.0821 0.0816 0.0816 0.0816
Note: This table reports marginal effects after pooled probit estimations for different macroprudential indices. The estimation period is January
2009-September 2017 (wave 1 – wave 17). Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or a credit line and
was approved, and 0 if a firm applied and was rejected. MPP is an index which represents a sum of all policy changes over waves in a particular
country. Capital index is a sum of all macroprudential measures targeting capital positions of banks. Borrower index is a sum of all instrument
actions targeting borrowers. Liquidity is an index which sums up all policy actions over the course of waves whose primary target is the enhancement
of liquidity positions in bank. Other is an index, which is a sum of the remaining measures. See Table 3 in the Appendix for more information on
the construction of macroprudential indices. See Table 4 in Appendix for all variable’s definitions Robust standard errors in parentheses. All
regressions include country, time, and sector dummies. All regressions include sampling weights. ***indicate p<0.01, **indicate p<0.05, and *
indicates p<0.1.

43
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 9

Macroprudential policy effects on financing for SMEs – tightening policy actions Macroprudential policy effects on financing for SMEs – loosening policy actions
Dependent variable (1) (2) (3) (4) (5) Dependent variable (6) (7) (8) (10)
Access to finance MPP Capital Borrower Liquidity Other Access to finance MPP Capital Borrower Other

Small -0.0751*** -0.0751*** -0.0751*** -0.0751*** -0.0750*** Small -0.0751*** -0.0751*** -0.0751*** -0.0751***
(0.00535) (0.00535) (0.00535) (0.00535) (0.00535) (0.00535) (0.00535) (0.00535) (0.00535)
Large 0.0375*** 0.0374*** 0.0375*** 0.0376*** 0.0375*** Large 0.0376*** 0.0375*** 0.0376*** 0.0376***
(0.00852) (0.00852) (0.00852) (0.00853) (0.00852) (0.00852) (0.00852) (0.00852) (0.00852)
Age 0.0469*** 0.0469*** 0.0468*** 0.0467*** 0.0468*** Age 0.0466*** 0.0468*** 0.0469*** 0.0468***
(0.00622) (0.00622) (0.00622) (0.00622) (0.00622) (0.00622) (0.00621) (0.00622) (0.00622)
Leverage -0.0267*** -0.0266*** -0.0266*** -0.0267*** -0.0267*** Leverage -0.0266*** -0.0267*** -0.0267*** -0.0267***
(0.00545) (0.00544) (0.00544) (0.00544) (0.00544) (0.00544) (0.00544) (0.00544) (0.00545)
Profit 0.0209*** 0.0210*** 0.0210*** 0.0210*** 0.0210*** Profit 0.0209*** 0.0211*** 0.0209*** 0.0209***
(0.00661) (0.00661) (0.00661) (0.00661) (0.00661) (0.00661) (0.00661) (0.00661) (0.00661)
History 0.0462*** 0.0463*** 0.0463*** 0.0464*** 0.0464*** History 0.0461*** 0.0461*** 0.0463*** 0.0462***
(0.00647) (0.00647) (0.00647) (0.00647) (0.00647) (0.00647) (0.00647) (0.00647) (0.00647)
Autonomous 0.00297 0.00285 0.00259 0.00286 0.00277 Autonomous 0.00305 0.00274 0.00282 0.00300
(0.00975) (0.00975) (0.00975) (0.00975) (0.00975) (0.00975) (0.00974) (0.00975) (0.00975)
Ownership 0.00397 0.00394 0.00417 0.00399 0.00401 Ownership 0.00379 0.00393 0.00399 0.00394
(0.00761) (0.00761) (0.00762) (0.00761) (0.00761) (0.00761) (0.00761) (0.00761) (0.00761)
Collateral 0.0591*** 0.0590*** 0.0590*** 0.0590*** 0.0589*** Collateral 0.0589*** 0.0588*** 0.0590*** 0.0592***
(0.0142) (0.0142) (0.0142) (0.0142) (0.0142) (0.0142) (0.0142) (0.0142) (0.0142)
Maturity 0.0495*** 0.0496*** 0.0493*** 0.0496*** 0.0496*** Maturity 0.0495*** 0.0491*** 0.0496*** 0.0495***
(0.00950) (0.00951) (0.00951) (0.00951) (0.00951) (0.00951) (0.00952) (0.00951) (0.00951)
MPP_tightening -0.0115 MPP_loosening 0.0315**
(0.00735) (0.0130)
Capital_tightening -0.0147* Capital_loosening 0.0549**
(0.00879) (0.0214)
Borrower_tightening -0.0328* Borrower_loosening 0.0757
(0.0180) (0.0542)
Liquidity_tightening -0.00826
(0.0171)
Other_tightening 0.00497 Other_loosening 0.0244*
(0.0112) (0.0138)
GDP growth 0.00627*** 0.00619*** 0.00615*** 0.00620*** 0.00617*** GDP growth 0.00617*** 0.00565*** 0.00610*** 0.00645***
(0.00170) (0.00170) (0.00170) (0.00170) (0.00170) (0.00170) (0.00170) (0.00170) (0.00171)
Unemployment -0.00529*** -0.00528*** -0.00536*** -0.00541*** -0.00541*** Unemployment -0.00520*** -0.00525*** -0.00546*** -0.00518***
(0.00162) (0.00161) (0.00161) (0.00162) (0.00161) (0.00162) (0.00161) (0.00161) (0.00162)
Crisis 0.0121 0.00910 0.00947 0.00853 0.00712 Crisis 0.00776 0.00591 0.00830 0.00873
(0.0179) (0.0177) (0.0177) (0.0177) (0.0182) (0.0177) (0.0177) (0.0177) (0.0177)
Country dummies YES YES YES YES YES Country dummies YES YES YES YES
Wave dummies YES YES YES YES YES Wave dummies YES YES YES YES
Sector dummies YES YES YES YES YES Sector dummies YES YES YES YES
Observations 41,848 41,848 41,848 41,848 41,848 Observations 41,848 41,848 41,848 41,848
Number of firms 25,956 25,956 25,956 25,956 25,956 Number of firms 25,956 25,956 25,956 25,956
Pseudo R-squared 0.0838 0.0838 0.0838 0.0837 0.0837 Pseudo R-squared 0.0839 0.0840 0.0837 0.0838

Note: This table reports marginal effects after pooled probit estimations for macroprudential indices based on their direction. The columns 1-5 represent the estimations with tightening indicators, which is equal to 1 if the overall
stance of macroprudential policy is dominantly tightening based on the sign of the base index. The tightening indicators are also separately estimated for each index. The columns 6-10 represent the estimations with loosening indicator,
which is equal to 1 if the overall stance of macroprudential policy is dominantly loosening based on the sign of the base index. The loosening indicators are also separately estimated for each index. The estimation period is January
2009-September 2017 (wave 1 – wave 17). Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or a credit line and was approved, and 0 if a firm applied and was rejected. MPP_tightening is an index

44
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

which represents a sum of all policy changes over waves in a particular country, and equal to 1 if the sign of the base index is positive, and 0 otherwise. Capital_tightening index is a sum of all macroprudential measures targeting
capital positions of banks, and equal to 1 if the sign of the base index is positive, and 0 otherwise. Borrower_tightening index is a sum of all instrument actions targeting borrowers, and equal to 1 if the sign of the base index is positive,
and 0 otherwise. Liquidity_tightening is an index which sums up all policy actions over the course of waves whose primary target is the enhancement of liquidity positions in bank, and equal to 1 if the sign of the base index is
positive, and 0 otherwise. Other_tightening is an index, which is a sum of the remaining measures, and equal to 1 if the sign of the base index is positive, and 0 otherwise. MPP_loosening is an index which represents a sum of all
policy changes over waves in a particular country, and equal to 1 if the sign of the base index is negative, and 0 otherwise. Capital_loosening index is a sum of all macroprudential measures targeting capital positions of banks, and
equal to 1 if the sign of the base index is negative, and 0 otherwise. Borrower_loosening index is a sum of all instrument actions targeting borrowers, and equal to 1 if the sign of the base index is negative, and 0 otherwise.
Other_loosening is an index, which is a sum of the remaining measures, and equal to 1 if the sign of the base index is negative, and 0 otherwise. See Table 3 in the Appendix for more information on the construction of macroprudential
indices. See Table 4 in Appendix for all variable’s definitions. Robust standard errors in parentheses. All regressions include sample weights, country, time, and sector dummies. *** indicate p<0.01, **indicate p<0.05, and * indicates
p<0.1.

45
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 10
The effects of macroprudential policies on access to funding for SMEs - subset of countries present
in all waves – panel component.

Dependent variable (1) (2) (3) (4) (5)

Access to finance MPP Capital Borrower Liquidity Other

Small -0.0672* -0.0671* -0.0672* -0.0673* -0.0672***

(-8.29) (-8.28) (-8.29) (-8.30) (-8.29)
Large 0.0315* 0.0314* 0.0312* 0.0314* 0.0312*
(2.31) (2.30) (2.28) (2.30) (2.28)
Age 0.0494*** 0.0495*** 0.0492*** 0.0490*** 0.0493***
(5.98) (5.99) (5.96) (5.93) (5.97)
Leverage -0.0239*** -0.0239*** -0.0240*** -0.0240*** -0.0240***
(-3.99) (-4.00) (-4.01) (-4.01) (-4.01)
Profit 0.0176** 0.0178** 0.0178** 0.0178** 0.0178**
(2.68) (2.70) (2.71) (2.71) (2.70)
History 0.0610*** 0.0611*** 0.0612*** 0.0612*** 0.0612***
(9.17) (9.18) (9.19) (9.20) (9.19)
Autonomous -0.00340 -0.00337 -0.00361 -0.00361 -0.00352
(-0.30) (-0.29) (-0.31) (-0.31) (-0.31)
Ownership -0.0112 -0.0114 -0.0113 -0.0113 -0.0112
(-1.29) (-1.31) (-1.29) (-1.30) (-1.29)
Collateral 0.0440** 0.0443** 0.0444** 0.0442** 0.0442**
(3.10) (3.12) (3.13) (3.11) (3.11)
Maturity 0.0409*** 0.0411*** 0.0415*** 0.0414*** 0.0414***
(4.12) (4.14) (4.18) (4.18) (4.17)
MPP -0.0107**
(-3.08)
Capital -0.0149**
(-2.86)
Borrower -0.00972
(-0.67)
Liquidity -0.0158
(-1.26)
Other -0.00767
(-1.29)
GDP growth 0.00281 0.00280 0.00275 0.00277 0.00276
(1.67) (1.67) (1.64) (1.65) (1.64)
Unemployment -0.00964*** -0.00975*** -0.00954*** -0.00936*** -0.00959***
(-5.17) (-5.23) (-5.12) (-5.00) (-5.14)
Crisis -0.000912 -0.00313 -0.00577 -0.00710 -0.00461
(-0.05) (-0.17) (-0.31) (-0.39) (-0.25)
Country dummies YES YES YES YES YES
Wave dummies YES YES YES YES YES
Sector dummies YES YES YES YES YES
Observations 24,521 24,521 24,521 24,521 24,521
Number of firms 8,629 8,629 8,629 8,629 8,629
Note: This table reports marginal effects after panel probit estimations with random effects for different macroprudential indices. The estimation
period is January 2009-September 2017 (wave 1 – wave 17). Access to finance is our dependent variable, equal to 1 if a firm applied for a loan or
a credit line and was approved, and 0 if a firm applied and was rejected. MPP is an index which represents a sum of all policy changes over waves
in a particular country. Capital index is a sum of all macroprudential measures targeting capital positions of banks. Borrower index is a sum of all
instrument actions targeting borrowers. Liquidity is an index which sums up all policy actions over the course of waves whose primary target is the
enhancement of liquidity positions in bank. Other is an index, which is a sum of the remaining measures. See Table 3 in the Appendix for more
information on the construction of macroprudential indices. See Table 4 in Appendix for all variable’s definitions Robust standard errors in
parentheses. All regressions include country, time, and sector dummies. ***indicate p<0.01, **indicate p<0.05, and * indicates p<0.1.

46
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5. Conclusion

We examine the impact of macroprudential policy on the availability of bank financing to small
and medium sized enterprises. Our results suggest the implementation of macroprudential policy
is associated with a smaller likelihood of SMEs obtaining bank financing. The effects are
confirmed when testing in different subsets, while we also test different direction of these
measures, namely the tightening versus loosening actions of macroprudential policy. Our results
suggest that when macroprudential measures are enforced and tightened, it is less likely that the
banks will be willing to lend to SMEs, while the opposite result is evident for estimations with
loosening macroprudential actions in the respective period. We extend the empirical part by testing
the results for subsample of countries present in all waves within SAFE database and with subset
of firms which appear in the covered period more than once. The overall results remain the same,
while the effects are even stronger for panel of countries present in all waves. The strongest
association between macroprudential policy and SMEs financing is evident in estimations with
capital index, which captures macroprudential policies whose target is improving capital position
of banks.

These results confirm the trade-off related with the implementation of macroprudential policies,
and strengthening regulatory activity, which can also affect availability of finance for small firms,
and slow down economic activity and credit supply. We argue that macroprudential policies should
be cautiously chosen and calibrated, with respect to macroeconomic and overall financial sector
conditions of the respective country. The policymaker should take into consideration the banks’
response to the implementation of macroprudential policies, and how banks adjust their balance
sheet, overall riskiness of their asset channels, and how they select their borrowers in response to
regulatory interventions. We find potential in further research in addressing the other possible
factors which might explain the borrowing channels of banks, and how does the adoption of
macroprudential policies affect the cost of financing for firms. We are also aware of limitations of
using survey-based data, as we did not employ the real financial firm level data, which represents
another potential for further research in the field.

47
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

REFERENCES

1. Acharya, V. V., & Steffen, S. (2015). The “greatest” carry trade ever? Understanding
eurozone bank risks. Journal of Financial Economics, 115(2), 215–236.
[Link]
2. Adelino, M., & Ferreira, M. A. (2016). Bank ratings and lending supply: Evidence from
sovereign downgrades. Review of Financial Studies, 29(7), 1709–1746.
[Link]
3. Aiyar, S., Al-Eyd, A., Barkbu, B., & Jobst, A. A. (2015). Revitalizing Securitization for
Small and Medium-Sized Enterprises in Europe. IMF Staff Discussion Note. Retrieved
from [Link]
4. Aiyar, S., Calomiris, C. W., & Wieladek, T. (2016). How does credit supply respond to
monetary policy and bank minimum capital requirements? European Economic Review,
82, 142–165. [Link]
5. Akerlof, A. G. (1958). The Market for "Lemons": Quality Uncertainty and the Market
Mechanism, The Quarterly Journal of Economics, Vol. 84, No. 3 (Aug., 1970), pp. 488-
500, URL: [Link]
6. Akinci, O., & Olmstead-Rumsey, J. (2018). How effective are macroprudential policies?
An empirical investigation. Journal of Financial Intermediation, 33(1136), 33–57.
[Link]
7. Albertazzi, U., & Marchetti, D. J. (2010). Credit supply, flight to quality and evergreening:
an analysis of bank-firm relationships after Lehman. Bank of Italy Working Paper.
[Link]
8. Almeida, H., Campello, M., & Weisbach, M. S. (2004). The cash flow sensitivity of cash.
Journal of Finance, 59(4), 1777-1804. [Link]
9. Altunbas, Y., Binici, M., & Gambacorta, L. (2018). Macroprudential policy and bank risk.
Journal of International Money and Finance, 81, 203–220.
[Link]
10. Ayyagari, M., Beck, T., & Martinez Peria, M. S. (2017). Credit Growth and
Macroprudential Policies: Preliminary Evidence on the Firm Level. BIS Papers No. 91,
(91), 15–34.
11. Barth, J. R., Lin, C., Ma, Y., Seade, J., & Song, F. M. (2013). Do bank regulation,
supervision and monitoring enhance or impede bank efficiency? Journal of Banking and
Finance, 37(8), 2879–2892. [Link]
12. Beck, T., & Demirguc-Kunt, A. (2006). Small and medium-size enterprises: Access to
finance as a growth constraint. Journal of Banking and Finance, 30(11), 2931–2943.
[Link]
13. Beck, T., Demirguc-Kunt, A., Laeven, L., & Levine, R. (2008). Finance, firm size, and
growth. Journal of Money, Credit and Banking, 40(7), 1379–1405.
[Link]
14. Beck, T., Demirgüç-Kunt, A., Laeven, L., & Maksimovic, V. (2006). The determinants of
financing obstacles. Journal of International Money and Finance, 25(6), 932–952.
[Link]
15. Beck, T., Demirgüç-Kunt, A., & Pería, M. S. M. (2011). Bank Financing for SMEs:
Evidence Across Countries and Bank Ownership Types. Journal of Financial Services

48
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Research, 39(1–2), 35–54. [Link]

16. Beck, T., & Demirgüç-Kunt, A., & Maksimović, V. (2005). Financial and Legal
Constraints to Growth : Does Firm Size Matter ? Journal of Finance,
[Link]
17. Becker, B., & Ivashina, V. (2014). Cyclicality of credit supply: Firm level evidence.
Journal of Monetary Economics, 62(1), 76–93.
[Link]
18. Behr, P., Foos, D., & Norden, L. (2017). Cyclicality of SME lending and government
involvement in banks. Journal of Banking and Finance, 77, 64–77.
[Link]
19. Berger, A. N., & Udell, G. F. (1994). Did Risk-Based Capital Allocate Bank Credit and
Cause a "Credit Crunch" in the United States?, Journal of Money, Credit and Banking,
Vol. 26, No. 3, [Link]
20. Berger, A. N., & Udell, G. F. (2006). A more complete conceptual framework for SME
finance. Journal of Banking and Finance, 30(11), 2945–2966.
[Link]
21. Bole, V., Prašnikar, J., & Trobec, D. (2014). Policy measures in the deleveraging process:
A macroprudential evaluation. Journal of Policy Modeling, 36(2), 410–432.
[Link]
22. Borio, C. E. V., & Shim, I. (2008). What Can (Macro-) Prudential Policy do to Support
Monetary Policy? Ssrn, (242). [Link]
23. Bremus, F., & Neugebauer, K. (2018). Reduced cross-border lending and financing costs
of SMEs. Journal of International Money and Finance, 80, 35–58.
[Link]
24. Bridges, J., Gregory, D., Nielsen, M., Pezzini, S., Radia, A., & Spaltro, M. (2014). The
Impact of Capital Requirements on Bank Lending. Ssrn, (486).
[Link]
25. Budnik, K., & Kleibl, J. (2018). Macroprudential regulation in the European Union in
1995-2014: introducing a new data set on policy actions of a macroprudential nature,
European Central Bank, Working Paper Series
26. Calomiris, C. W. & Hubbard, R. G. & Stock, J. H. (1986). "The Farm Debt Crisis and
Public Policy," Brookings Papers on Economic Activity, Economic Studies Program, The
Brookings Institution, vol. 17(2), pages 441-486
27. Campello, M., Graham, J. R., & Harvey, C. R. (2010). The real effects of financial
constraints: Evidence from a financial crisis. Journal of Financial Economics, 97(3), 470–
487. [Link]
28. Carlson, M., Shan, H., & Warusawitharana, M. (2013). Capital ratios and bank lending: A
matched bank approach. Journal of Financial Intermediation, 22(4), 663–687.
[Link]
29. Casey, E., & O’Toole, C. M. (2014). Bank lending constraints, trade credit and alternative
financing during the financial crisis: Evidence from European SMEs. Journal of Corporate
Finance, 27, 173–193. [Link]
30. Cerutti, E, Correa, R., Fiorentino, E., & Segalla, E. (2016). Changes in Prudential Policy
Instruments, (16/110), IMF Working Paper
31. Cerutti, Eugenio, Claessens, S., & Laeven, L. (2017). The use and effectiveness of
macroprudential policies: New evidence. Journal of Financial Stability, 28, 203–224.

49
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[Link]
32. Chava, S., & Purnanandam, A. (2011). The effect of banking crisis on bank-dependent
borrowers. Journal of Financial Economics, 99(1), 116–135.
[Link]
33. Claessens, S., Ghosh, S. R., & Mihet, R. (2013). Macro-prudential policies to mitigate
financial system vulnerabilities. Journal of International Money and Finance, 39, 153–
185. [Link]
34. Correa, R., Sapriza, H., & Zlate, A. (2013). Liquidity Shocks, Dollar Funding Costs, and
the Bank Lending Channel During the European Sovereign Crisis. Ssrn, (1059).
[Link]
35. Cosimano, T. F., & Hakura, D. (2011). Bank Behavior in Response to Basel III: A Cross-
Country Analysis. Ssrn. [Link]
36. Crowe, C., Dell’Ariccia, G., Igan, D., & Rabanal, P. (2013). How to deal with real estate
booms: Lessons from country experiences. Journal of Financial Stability, 9(3), 300–319.
[Link]
37. Dell’Ariccia, G., Igan, D., & Laeven, L. (2008). “Credit Booms and Lending Standards :
Evidance from the Subprime Mortgage Market.” Journal of Money, Credit and Banking,
44(2), 367–384. [Link]
38. Devereux, M., Schiantarelli, F., 1990. Investment, financial factors and cash flow from UK
panel data. In: Hubbard, G. (Ed.), Information, Capital Markets and Investment. University
of Chicago Press, Chicago, IL, [Link]
39. Duygan-Bump, B., Levkov, A., & Montoriol-Garriga, J. (2015). Financing constraints and
unemployment: Evidence from the Great Recession. Journal of Monetary Economics, 75,
89–105. [Link]
40. ESRB. (2014). Flagship Report on Macro-prudential Policy in the Banking Sector, 1–26.
Retrieved from [Link]
41. European Banking Authority. (2016). EBA Report on SMEs and SME Supporting Factor,
1–137.
42. Ferrando, A., Popov, A., & Udell, G. F. (2017). Sovereign stress and SMEs’ access to
finance: Evidence from the ECB’s SAFE survey. Journal of Banking and Finance, 81, 65–
80. [Link]
43. Ferrando, A., Popov, A., & Udell, G. F. (2019). Do SMEs Benefit from Unconventional
Monetary Policy and How? Microevidence from the Eurozone. Journal of Money, Credit
and Banking, 51(4), 895–928. [Link]
44. Francis, W. B., & Osborne, M. (2012). Capital requirements and bank behavior in the UK:
Are there lessons for international capital standards? Journal of Banking and Finance,
36(3), 803–816. [Link]
45. García-Posada Gómez, M. (2018). Credit Constraints, Firms Investment and Growth
Evidence from Survey Data. Ssrn, (2126). [Link]
46. Gertler, M. (1988). Financial Structure and Aggregate Economic Activity: An Overview,
Journal of Money, Credit and Banking, vol. 20(3), pages 559-588, August, DOI:
10.2307/1992535
47. Ghosh, P., Mookherjee, D., Ray, D. (1999). Credit Rationing in Developing Countries: An
Overview of the Theory, Blackwell, London
48. Heckman, J. J. (1979) Sample Selection Bias as a Specification Error, Econometrica, Vol.
47, No. 1, pp. 153-161 [Link]

50
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

49. Jin, Y., Luo, M., & Wan, C. (2018). Financial constraints, macro-financing environment
and post-crisis recovery of firms. International Review of Economics and Finance,
55(January), 54–67. [Link]
50. Kahou, E. M., & Lehar, A. (2017). Macroprudential policy: A review. Journal of Financial
Stability, 29, 92–105. [Link]
51. Kaplan, S. N., & Zingales, L. (1997). Do Investment-Cash Flow Sensitivities Provide
Useful Measures of Financing Constraints? The Quarterly Journal of Economics, 112(1),
169–215. [Link]
52. Kuttner, K. N., & Shim, I. (2013). Can non-interest rate policies stabilize housing markets?
evidence from a panel of 57 economies, NBER Working Paper Series
53. Lim, C., Columba, F., Costa, A., Kongsamut, P., Otani, A., Saiyid, M., … Wu, X. (2011).
Macroprudential Policy : What Instruments and How to Use Them ? Lessons from Country
Experiences. IMF Working Papers,
[Link]
54. Mascia, D. V., & Rossi, S. P. S. (2017). Is there a gender effect on the cost of bank
financing? Journal of Financial Stability, 31, 136–153.
[Link]
55. Mayordomo, S., & Rodriguez-Moreno, M. (2017). “Support Is Appreciated”: On the
Effectiveness of the SME Supporting Factor. SSRN Electronic Journal.
[Link]
56. Mayordomo, S., & Rodríguez-Moreno, M. (2018). Did the bank capital relief induced by
the Supporting Factor enhance SME lending? Journal of Financial Intermediation, (May),
1–13. [Link]
57. McFadden, D. (1978). Quantitative Methods for Analyzing Travel Behaviour of
Individuals: Some Recent Developments. Behavioural Travel Modelling, Croom Helm, pp.
279-318
58. Myers, S. C., & Majluf, N. S. (1984). Corporate financing and investment decisions when
firms have information that investors do not have, Journal of Financial Economics,
Volume 13, Issue 2, Pages 187-221, DOI: [Link]
59. Modigliani, F., & Miller, M. H. (1958). The American economic Revlew. The American
Economic Review, 48(3), 261–297. Retrieved from [Link]
60. Morgan, P. J., Regis, P. J., & Salike, N. (2018). LTV policy as a macroprudential tool and
its effects on residential mortgage loans. Journal of Financial Intermediation, 1–15.
[Link]
61. Moro, A., Wisniewski, T. P., & Mantovani, G. M. (2017). Does a manager’s gender matter
when accessing credit? Evidence from European data. Journal of Banking and Finance,
80, 119–134. [Link]
62. Noss, J., & Toffano, P. (2016). Estimating the impact of changes in aggregate bank capital
requirements on lending and growth during an upswing. Journal of Banking and Finance,
62, 15–27. [Link]
63. Olszak, M., Roszkowska, S., & Kowalska, I. (2018). Macroprudential policy instruments
and procyclicality of loan-loss provisions – Cross-country evidence. Journal of
International Financial Markets, Institutions and Money, 54, 228–257.
[Link]
64. Petersen, M. A., & Rajan, R. G. (1994). The Benefits of Lending Relationships : Evidence
from Small Business Data, The Journal of Finance, Vol. 49, No. 1, pp. 3-37.

51
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[Link]
65. Petersen, M. A., Rajan, R. G. (1995). The Effect of Credit Market Competition on Lending
Relationships, The Quarterly Journal of Economics, Vol. 110, No. 2, pp. 407-443,
[Link]
66. Popov, A., & Udell, G. F. (2012). Cross-border banking, credit access, and the financial
crisis. Journal of International Economics, 87(1), 147–161.
[Link]
67. Heckman, J. J. (1979) Sample Selection Bias as a Specification Error, Econometrica, Vol.
47, No. 1, pp. 153-161 [Link]
68. Santos, J. A. C. (2011). Bank corporate loan pricing following the subprime crisis. Review
of Financial Studies, 24(6), 1916–1943. [Link]
69. Schiffer, M., Weder, B., 2001. Firm size and the business environment: worldwide survey
results. Discussion Paper No.43. International Finance Corporation, Washington, DC.
70. Stiglitz, J. E., & Weiss, A. (1981). Credit rationing in markets with imperfect information,
The American Economic Review, Vol. 71, Issue 3, pp. 393-410, URL:
[Link]
71. Tsuruta, D. (2015). Bank loan availability and trade credit for small businesses during the
financial crisis. Quarterly Review of Economics and Finance, 55, 40–52.
[Link]
72. Wang, X., Han, L., & Huang, X. (2020). Bank market power and SME finance: Firm-bank
evidence from European countries. Journal of International Financial Markets, Institutions
and Money, (xxxx). [Link]
73. Whited, T., 1992. Debt, liquidity constraints, and corporate investment: evidence from
panel data. Journal of Finance 47 (4), 1425e1460. DOI, [Link]
6261.1992.tb04664.x

52
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

APPENDIX
Table 1: Reference periods
Survey round Round Wave Reference period - last 6 months Macroprudential data Macroeconomic data

2009H1 Common 1 January-June 2009 1 & 2 Q of 2009 1 & 2 Q of 2009

2009H2 ECB round 2 July-December 2009 3 & 4 Q of 2009 3 & 4 Q of 2009
2010H1 ECB round 3 March-September 2010 1, 2 & 3 Q of 2010 2 & 3 Q of 2010
2010H2 ECB round 4 September 2010-February 2011 4Q of 2010 & 1Q of 2011 4Q of 2010 & 1Q of 2011
2011H1 Common 5 April-September 2011 2 i 3 Q of 2011 2 i 3 Q of 2011
2011H2 ECB round 6 October 2011-March 2012 4Q of 2011 & 1Q of 2012 4Q of 2011 & 1Q of 2012
2012H1 ECB round 7 April-September 2012 2 i 3 Q of 2012 2 i 3 Q of 2012
2012H2 ECB round 8 October 2012-March 2013 4Q of 2012 & 1Q of 2013 4Q of 2012 & 1Q of 2013
2013H1 Common 9 April-September 2013 2 i 3 Q of 2013 2 i 3 Q of 2013
2013H2 ECB round 10 October 2013-March 2014 4Q of 2013 & 1Q of 2014 4Q of 2013 & 1Q of 2014
2014H1 Common 11 April-September 2014 2 i 3 Q of 2014 2 i 3 Q of 2014
2014H2 ECB round 12 October 2014-March 2015 4Q of 2014 & 1Q of 2015 4Q of 2014 & 1Q of 2015
2015H1 Common 13 April-September 2015 2 i 3 Q of 2015 2 i 3 Q of 2015
2015H2 ECB round 14 October 2015-March 2016 4Q of 2015 & 1Q of 2016 4Q of 2015 & 1Q of 2016
2016H1 Common 15 April-September 2016 2 i 3 Q of 2016 2 i 3 Q of 2016
2016H2 ECB round 16 October 2016-March 2017 4Q of 2016 & 1Q of 2017 4Q of 2016 & 1Q of 2017
2017H1 Common 17 April-September 2017 2 i 3 Q of 2017 2 i 3 Q of 2017

53
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Firms by country

Wave 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Country Total
1 AT 54.0 58.0 61.0 145.0 141.0 177.0 155.0 172.0 166.0 146.0 128.0 205.0 140.0 229.0 132.0 185.0 141.0 2,435.0
2 BE 55.0 30.0 48.0 171.0 130.0 164.0 154.0 165.0 152.0 164.0 157.0 239.0 133.0 283.0 157.0 245.0 145.0 2,592.0
3 BG 29.0 0.0 0.0 0.0 127.0 0.0 0.0 0.0 148.0 0.0 118.0 0.0 109.0 0.0 135.0 0.0 121.0 787.0
4 CY 31.0 0.0 0.0 0.0 28.0 0.0 0.0 0.0 27.0 0.0 16.0 0.0 25.0 0.0 16.0 0.0 29.0 172.0
5 CZ 42.0 0.0 0.0 0.0 144.0 0.0 0.0 0.0 119.0 0.0 134.0 0.0 110.0 0.0 111.0 0.0 88.0 748.0
6 DE 250.0 271.0 300.0 288.0 263.0 309.0 292.0 341.0 335.0 335.0 322.0 331.0 345.0 337.0 330.0 352.0 319.0 5,320.0
7 DK 29.0 0.0 0.0 0.0 90.0 0.0 0.0 0.0 112.0 0.0 84.0 0.0 74.0 0.0 80.0 0.0 85.0 554.0
8 EE 19.0 0.0 0.0 0.0 16.0 0.0 0.0 0.0 10.0 0.0 17.0 0.0 15.0 0.0 19.0 0.0 19.0 115.0
9 ES 377.0 421.0 435.0 416.0 396.0 422.0 421.0 418.0 401.0 379.0 496.0 540.0 498.0 564.0 444.0 554.0 410.0 7,592.0
10 FI 10.0 17.0 19.0 74.0 86.0 90.0 92.0 105.0 106.0 98.0 121.0 132.0 136.0 105.0 138.0 141.0 119.0 1,589.0
11 FR 263.0 290.0 382.0 416.0 447.0 437.0 409.0 432.0 451.0 434.0 620.0 556.0 558.0 589.0 604.0 627.0 556.0 8,071.0
12 GB 86.0 0.0 0.0 0.0 225.0 0.0 0.0 0.0 213.0 0.0 231.0 0.0 239.0 0.0 232.0 0.0 166.0 1,392.0
13 GR 92.0 70.0 69.0 163.0 143.0 158.0 83.0 112.0 117.0 124.0 115.0 187.0 109.0 134.0 100.0 165.0 106.0 2,047.0
14 HR 17.0 0.0 0.0 0.0 34.0 0.0 0.0 0.0 29.0 0.0 92.0 0.0 77.0 0.0 79.0 0.0 90.0 418.0
15 HU 30.0 0.0 0.0 0.0 103.0 0.0 0.0 0.0 143.0 0.0 123.0 0.0 127.0 0.0 96.0 0.0 115.0 737.0
16 IE 17.0 23.0 33.0 125.0 127.0 142.0 147.0 128.0 125.0 121.0 130.0 119.0 106.0 136.0 112.0 107.0 94.0 1,792.0
17 IT 363.0 376.0 396.0 483.0 381.0 445.0 454.0 463.0 466.0 444.0 576.0 604.0 599.0 615.0 566.0 582.0 557.0 8,370.0
18 LT 26.0 0.0 0.0 0.0 92.0 0.0 0.0 0.0 101.0 0.0 68.0 0.0 66.0 0.0 66.0 0.0 79.0 498.0
19 LU 20.0 0.0 0.0 0.0 31.0 0.0 0.0 0.0 36.0 0.0 26.0 0.0 19.0 0.0 34.0 0.0 25.0 191.0
20 LV 12.0 0.0 0.0 0.0 49.0 0.0 0.0 0.0 22.0 0.0 53.0 0.0 30.0 0.0 40.0 0.0 38.0 244.0
21 MT 20.0 0.0 0.0 0.0 17.0 0.0 0.0 0.0 28.0 0.0 29.0 0.0 25.0 0.0 25.0 0.0 24.0 168.0
22 NL 55.0 39.0 36.0 80.0 83.0 94.0 93.0 78.0 105.0 88.0 132.0 152.0 115.0 154.0 125.0 132.0 118.0 1,679.0
23 PL 130.0 0.0 0.0 0.0 327.0 0.0 0.0 0.0 291.0 0.0 367.0 0.0 380.0 0.0 342.0 0.0 332.0 2,169.0
24 PT 84.0 60.0 73.0 150.0 134.0 125.0 129.0 126.0 130.0 139.0 121.0 224.0 159.0 252.0 158.0 234.0 139.0 2,437.0
25 RO 48.0 0.0 0.0 0.0 142.0 0.0 0.0 0.0 115.0 0.0 151.0 0.0 153.0 0.0 142.0 0.0 143.0 894.0
26 SE 46.0 0.0 0.0 0.0 74.0 0.0 0.0 0.0 79.0 0.0 89.0 0.0 88.0 0.0 64.0 0.0 75.0 515.0
27 SI 44.0 0.0 0.0 0.0 45.0 0.0 0.0 0.0 54.0 0.0 79.0 0.0 64.0 0.0 75.0 0.0 71.0 432.0
28 SK 24.0 0.0 0.0 0.0 110.0 0.0 0.0 0.0 100.0 0.0 124.0 123.0 141.0 153.0 145.0 133.0 151.0 1,204.0
Total 2,273.0 1,655.0 1,852.0 2,511.0 3,985.0 2,563.0 2,429.0 2,540.0 4,181.0 2,472.0 4,719.0 3,412.0 4,640.0 3,551.0 4,567.0 3,457.0 4,355.0 55,162.0
Note: The table represents the information on number of firms by country which applied for bank loan, across all waves, and includes both ECB &
Common rounds.

54
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 3: Macroprudential instruments in MaPPED and construction of base indices

Instrument group Instruments Index

Capital adequacy ratio (CAR)
Tier 1 capital ratio
Minimum capital requirements Common Equity Tier 1 capital ratio
(CET 1)
Core Tier 1 capital ratio
Countercyclical capital buffer
Capital conservation buffer
Systemic risk buffer
G-SII capital buffer
Capital buffers O-SII capital buffer
Other capital requirements targeting Capital =
most important institutions (Minimum capital requirements +
Other capital surcharges and own funds Capital Buffers + Risk weights +
requirements Loan loss provisioning)
Profit distribution restrictions
Risk weights for loans backed by
residential property
Risk weights Risk weights for loans backed by
commercial property
Other sectoral risk weights
Loan classification rules
Loan loss provisioning Minimum specific provisioning
General provisioning
Capital treatment of loan loss reserve
Loan-to-value (LTV) limits
Lending standards restrictions Loan-to-income (LTI) limits
Debt-to-income (DTI) limits
Debt-service-to-income (DSTI) limits
incl. interest rate stress testing
Limits on interest rates on loans
Maturity and amortisation restrictions Borrower = MPP =
Other income requirements for loan (Lending standards restrictions + (Capital + Borrower + Liquidity
eligibility Limits on credit growth and volume) + Other)
Limits on the volume of personal loans
Other restrictions on lending standards

55
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Limits on credit growth and volume Reserve requirements related to banks’

liabilities
Asset-based reserve requirements
Loan to deposit (LTD) ratios
Other stable funding req. incl. Net
Stable Funding Requirement
Short-term liquidity coverage ratios Liquidity =
Liquidity requirements and limits on incl. Liquidity Coverage Ratio (Liquidity requirements and limits on
currency and maturity mismatch Liquidity ratios and deposit coverage currency and maturity mismatch)
ratios
Limits on FX mismatches
Other liquidity requirements
Leverage ratio Leverage ratio
Tax on assets/liabilities
Levies/taxes on financial institutions Tax on financial activities
Single client exposure limits
Limits on large exposures and Intragroup exposures limits
concentration Sector and market segment exposure
limits
Funding concentration limits
Limits on qualified holdings outside
financial sector Other =
Other exposure and concentration (Leverage ratio + Levies on financial
limits institutions + Limits on large
Structural measures exposures and concentration + Other
Margin requirements measures)
Other regulatory restrictions on
Other measures financial activities
Limits on deposit rates
Debt resolution policies
Crisis management tools
Changes in regulatory framework
Other
Source: Budnik, K., Kleibl, J. (2018). Macroprudential regulation in the European Union in 1995-2014: introducing a new data set on policy actions of a
macroprudential nature. Macroprudential Policies Evaluation Database (MaPPED).

56
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 4: Variables sources

Variable Description Source
Dependent variable

Dummy variable equal to 1 if the firm

applied for a bank loan or a credit line in
past six months and was approved, and 0 ECB, SAFE
Access to finance otherwise.

Macroprudential indices

An indicator which represents a sum of all ECB, MaPPED

policy changes over waves in a particular
country with the scope from -3 to 6. The
higher the index, the tighter
MPP macroprudential policy in a country.
An index with represents the sum of all ECB, MaPPED
macroprudential measures targeting capital
positions of banks, with the scope from -2
Capital to 3.
An index which is a sum of all instrument ECB, MaPPED
actions targeting borrowers, with the scope
Borrower from -1 to 3.
An index which sums up all policy actions ECB, MaPPED
over the course of waves whose primary
target is the enhancement of liquidity
positions in bank, with the scope from -1 to
Liquidity 2.
An index, which is a sum of the remaining ECB, MaPPED
macroprudential measures, with the scope
Other from -3 to 3.
An index which represents a sum of all ECB, MaPPED
MPP_cumulative policy changes over waves in a particular

57
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

country, standardized to a minimum of -1

and maximum of 1.
An index which is a sum of all ECB, MaPPED
macroprudential measures targeting capital
positions of banks, standardized to a
Capital_cumulative minimum of -1 and maximum of 1
And index which is equal to a sum of all ECB, MaPPED
instrument actions targeting borrowers,
standardized to a minimum of -1 and
Borrower_cumulative maximum of 1
An index which sums up all policy actions ECB, MaPPED
over the course of waves whose primary
target is the enhancement of liquidity
positions in bank, standardized to a
Liquidity_cumulative minimum of -1 and maximum of 1.
An index, which is a sum of the remaining ECB, MaPPED
macroprudential measures, standardized to
Other_cumulative a minimum of -1 and maximum of 1.
An index equal to 1 if the sign of the base ECB, MaPPED
MPP_loosening index (MPP) is negative, and 0 otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
index (Capital) is negative, and 0
Capital_loosening otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
index (Borrower) is negative, and 0
Borrower_loosening otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
index (Liquidity) is negative, and 0
Liquidity_loosening otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
Other_loosening index (Other) is negative, and 0 otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
MPP_tightening index (MPP) is positive, and 0 otherwise.

58
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

An index equal to 1 if the sign of the base ECB, MaPPED

Capital_tightening index (Capital) is positive, and 0 otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
index (Borrower) is positive, and 0
Borrower_tightening otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
index (Liquidity) is positive, and 0
Liquidity_tightening otherwise.
An index equal to 1 if the sign of the base ECB, MaPPED
Other_tightening index (Other) is positive, and 0 otherwise.

Firm determinants

Dummy variable takes value of 1 if the ECB, SAFE

firm has from 1 to 49 employees, and 0
Small otherwise
Dummy variable takes value of 1 if the ECB, SAFE
firm has over 250 employees, and 0
Large otherwise
Dummy variable which has the value of 1 ECB, SAFE
if the firm is older than 10 years and 0
Age otherwise
Dummy variable which equals 1 if firm has ECB, SAFE
increased the use of debt in the last six
Leverage months.
Dummy variable equals 1 if the firm’s ECB, SAFE
profit increased and 0 if the firm’s profit
Profit decreased in the last six months.
Dummy variable which equals 1 if firms’ ECB, SAFE
credit history improved in the last six
Credit history months.

59
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Dummy variable which equals 1 if the firm ECB, SAFE

in an autonomous profit-oriented
Autonomous enterprise.
Dummy variable which equals 1 if the firm ECB, SAFE
is owned by an individual, family, or
Ownership entrepreneurs, and 0 otherwise.
Dummy variable equal to 1 if the bank ECB, SAFE
decreased the amount of collateral required
by the firms when applying for bank loan,
Collateral and 0 otherwise.
Dummy variable which is an indicator ECB, SAFE
equal to 1 if bank increased the maturity of
Maturity loans, and 0 otherwise.

Macroeconomic determinants
The annual growth rate of GDP based on World Bank Database
averages of quarterly data corresponding to
GDP growth (%) each wave.
The annual rate of unemployment in a World Bank Database
country based on averages of quarterly data
Unemployment (%) corresponding to each wave.
The dummy variable equal to 1 for waves Own elaboration
Crisis waves 1-7 (2009-2012) and 0 otherwise.

60
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

THE MEDIATING ROLE OF EMOTIONAL STABILITY BETWEEN EMOTIONAL

INTELLIGENCE AND OVERWORK BEHAVIOR

Ajtene Avdullahi1, Osman YILDIRIM2

1
University of Mitrovica "Isa Boletini", Faculty of Economics, Kosovo
2
Istanbul Arel University, Engineering Department, Istanbul Turkey

Abstract

The purpose of this research is to explore the mediating role of emotional stability between the
regulation of emotion and working excessively. For this purpose, a research survey was designed by
using the scales related to the variables of emotional stability, regulation of emotion and working
excessively in the literature. The research data was collected by applying survey to 320 participants
who voluntarily accepted participation in the survey by easy sampling method. Reliability, combined
reliability, convergence validity, decomposition validity and average variance extracted values were
investigated with the obtained research data, respectively. In addition, factor analyzes of the scales were
made and the research model was confirmed by conducting a mediator variable effect search with the
AMOS program. It is concluded that the research model is significant since the model test values x2
(397.166), x2 / df (2.206) and (p <0.05) in the path analysis model with observed variables for all data
(n = 300). It is understood that the model is valid because the fit index values of the research model are
in the acceptable fit limits of GFI (.911), CFI (.961), SRMR (.0702), RMSEA (.071).

Key Words: Emotional Stability, Overwork Behavior, Structural Equation Modeling

INTRODUCTION

The organization will achieve its goals largely depending on the employees of the organization.
For this reason, the top management of the organizations deal with the issues that will affect
their employees individually for the goals of the organization. In the review of the related
literature, it can be suggested that there are variables such as emotional stability, regulation of
emotion, working excessively and working compulsively, which are considered important for
the employees and presents quite a subject of research.

61
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Emotional Stability: McCrae and Costa (1989) defined individuals' attitudes and
behaviours that show continuity according to changing situations and the way they
interact. In his study, Goldberg (1990) expressed personality types in five
dimensions: openness to experience, responsibility, extraversion, compatibility and
neuroticism (Andreassen, Hetland & Pallesen, 2010; Barrick & Mount, 1991).

Burke (2006) searched the relationship of personality factors (Neuroticism, Extraversion,

Conscientiousness, Agreeableness, Openness to new experience) with workaholism defined by
Spence and Robbins (1992) and pointed out that personality factors were statistically correlated
to workaholism.

Besides, Aziz & Tronzo (2011) investigated the relationship of the five-factor model of
personality with the workaholism. According to their research, conscientiousness and
agreeableness were positively correlated to work involvement; conscientiousness and openness
to experience were positively correlated to work drive. Their research findings showed that
agreeableness, conscientiousness, and openness to experience were positively correlated to
work enjoyment, whereas neuroticism was negatively correlated to work enjoyment. On the
other hand, Jackson et al. (2016) formulated the relationships between personality traits (Big
Five Models) and workaholics and showed that workaholics are not related to an addictive
personality.

Regulation of Emotion: First, Thorndike and Stein defined the concept of social intelligence
in their study in 1937. Wechsler (1939) added the emotional intelligence dimension to this
concept in his study. In his research in 1983, Gardner put forward the idea of "multiple
intelligence". In his "Multiple Intelligence" theory, Gardner (1983) put forward the difference
between "knowing one's inner world" and "social skill". This research later resulted in
emotional intelligence theories (Gardner, 1983). Salovey and Mayer (1990), on the other hand,
revealed the connection of emotional intelligence with empathy.

Emotional intelligence is expressed as the individual's being aware of his own emotions,
empathizing, organizing his feelings and behaviours and managing his emotions (Salovey &
Mayer, 1990; Charbonneau & Nicol, 2002).

Goleman (1998) research has made the concept of emotional intelligence very popular (Kun &
Demetrovics, 2010; Schutte, Malouff, Thorsteinsson, Bhullar, & Rooke, 2007). While Schutte
et al. (2007) examined the relationship between sensory intelligence and health in their study,

62
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Yildirim (2007) revealed the relationship between salespeople and information technology
staff's performance with emotional intelligence. Lopes et al. (2002) in their study, revealed the
relationships between emotional intelligence and measured personality traits. Law [Link]. (2014)
reviewed the emotional intelligence and discussed how it differs from big five personalities.

It is possible to empathize, to understand the person opposite, to show the ability to reduce
negative emotions and increase positive emotions, to manage emotions. Therefore, emotional
balance is directly related to managing emotions (De Raad, Sullot & Barelds, 2008).

Working Excessively and working compulsively: Overwork means that the individual does
business without waiting for a response or works above the level specified in the employment
contract. Because of their loyalty to the business, the employees do extra work to avoid giving
up half, or they choose to work hard to avoid the stresses and problems of everyday life.
Business owners or top managers are satisfied with the attitude of the employees, provided that
they do not overdo it. Scott et al. (1997) expressed workaholics overwork, as a dedication to
working spiritually at all times.

The original a 25-item self-report inventory, work addiction test, was developed by Robinson
(1999). Later, as a result of the work done on the work addiction test or the workaholism test,
two-dimensional DUWAS scale has emerged (Taris & Schaufeli, 2003). Taris et al. (2005)
proposed the validation of a Dutch version of Robinson’s (1999) scale (Work Addiction Risk
Scale). It is called the Dutch Work Addiction Scale which has two dimensions such as working
excessively and working compulsively.

In many types of research, workaholicism is considered as two dimensions. These dimensions

are; (1) overwork (the individual's overwork or obsession to work), (2) the tendency to work
(the individual perceives work as a necessity) (Schaufeli, Taris and Bakker, 2006; Schaufeli,
Bakker, van der Heijden and Prins, 2009). Ying-Wen & Chen-Mıng (2009) investigated the
interactions among antecedents of workaholism and determined the antecedents linked to
workaholism.

63
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

METHODOLOGY

Data Collection Tools

Aiming to gather information about the emotional balances of the participants, De Raad et al.
(2008), the emotional stability dimension of the personality types scale included in the study
was included in the research questionnaire. However, as a result of the statistical analysis, the
statements such as "Is dependent on taking decisions" and "Is affectionate towards family and
friends" in the original scale dimension were excluded from the evaluation.

In accordance with the research hypotheses, the Dutch Work Addiction Scale in the Taris and
Bakker (2006) study was used in the research questionnaire to reveal the working excessively
and working compulsively aspects of the participants. The sub-dimensions of the emotional
intelligence scale were adapted from the research of Law et al. (2014).

Research Model

Emotional
Stability (ES)

Regulation of Emotion (ROE)- Working Excessively (WE)-

Others-Emotions Appraisal (OEA) Working Compulsively (WC)

Fig.1 Research Model

Research Hypotheses:

Hypothesis 1: Emotional stability has a mediating role in the relationship between the regulation
of emotion and working excessively.

Hypothesis 2: Emotional stability has a mediating role in the relationship between the regulation
of emotion and working compulsively.

64
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Hypothesis 3: Emotional stability has a mediating role in the relationship between others-
emotions appraisal and working excessively.

Hypothesis 4: Emotional stability has a mediating role in the relationship between others-
emotions appraisal and working compulsively.

Respondents

Interviews were conducted with 329 people for face-to-face survey work. However, as a result
of the examination, the completion rate of 9 questionnaires was very low and it was excluded.
Thus, the number of participants was determined as 320. Of the participants who participated
in the research, 76.7% were female and 23.3% were male. The highest age group among the
respodents is 30-34 years old with 59.7%. Followed by 25-29 age group with a rate of 35.3%.
Concerning the educational status of our participants, there are 54.3% primary school graduates,
25.7% high school graduates and 20% university graduates. In the occupational groups, the
administrator was in the rate of 33.7%, while the ratio of doctor-nurse was in the ratio of 17.7%.
The rate of other personnel is determined as 16.3%.

Statistical Techniques

The collected date in the study were analyzed and interpreted using SPSS for Windows 22.00
and AMOS 22.0 program. Confirmatory factor analyzes related to the scales used in the study
were made in the AMOS program and cronbach's Alpha, AVE and CR values were also
calculated. In addition, the mediating effects of structural equation modeling were analyzed in
the AMOS program using the bootstrap method.

Findings

Confirmatory Factor Analysis (CFA) investigated the significance of the measurement models
for each scale and the good fit criteria required for both confirmatory factor analysis and the
suitability of the measurement model with the AMOS 22.0 package program.
As it is known, as the sample grows, the Chi-Square (x2) value also increases and the statistical
significance level of the Chi-Square (x2) test decreases (Bollen, 1989; Fornel & Larcker, 1981;
Bagozzi and Yi, 2012). Both the confirmatory factor analysis and the suitability of the tested
model are determined by taking into account the values in the chi-square (x2) value corrected

65
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

by the degree of freedom, other goodness-of-fit indices and the values in the standardized
residual covariance matrix (Schermelleh-Engel, Moosbrugger & Müller, 2003).

Table 1. Goodness of Fit Indices and Conformity Values Used in Confirmatory Factor
Analysis

Indexes Good Fit Acceptable Fit

χ2 / df 0 ≤ χ2/df ≤ 2 2 < χ2/df ≤ 3

GFI ≥ 0.90 0.85-0.89

CFI ≥ 0.97 ≥ 0.95

SRMR ≤0.05 0.06 ≤ SRMR ≤ 0.08

RMSEA ≤ 0.05 0.06 ≤ RMSEA ≤ 0.08

From the 16-item Big four (bf) scale, the 4-item emotional stability (ES) dimension was
eliminated from the analysis because the factor load was low (FL <0.50). From the 17-item
DUWAS scale, the 6-item factor load was low (FL <0.50). There are 4 items in regulation of
emotion (ROE) and others-emotions appraisal (OEA) sub-dimensions, which are sub-
dimensions of emotional intelligence (ei) scale. In all of the scale items, factor load values are
in the range (0.543; 0.895) (Kline, 2005; Schermelleh-Engel, Moosbrugger & Müller, 2003).

Table 2. Confirmatory factor analysis for measurement model

Items Dimension Estimate Standart C.R. P

Estimate
bf13 (Is controlled…emotional situations) ES 1.000 0.895
bf15 (Remains…chaotic situations) ES 0.745 0.649 6.056 ***
dws4 (I find myself …called it quits) WE 1.000 0.778
dws8 (I stay busy … irons in the fire) WE 1.085 0.706 10.67 ***

dws10 (I seem to be …against the clock) WE 1.310 0.757 11.12 ***

dws12 (I spend …on leisure activities) WE 0.874 0.77 11.45 ***
dws13 (I feel guilty …on something) WE 1.294 0.67 6.995 ***
dws15 (I find myself …on the telephone) WE 1.075 0.643 6.164 ***

66
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Items Dimension Estimate Standart C.R. P

Estimate
dws17 (It is hard …I am not working) WE 1.425 0.611 6.637 ***
dws7 (I overly commit… than I can chew) WC 1.000 0.686
dws11 (I feel … drives me to work hard) WC 1.296 0.708 8.210 ***
dws14 (I feel …when it is not enjoyable) WC 1.249 0.651 7.750 ***
dws16 (I feel guilty… time off work) WC 1.573 0.781 8.712 ***
ei5 (I always know…from their behavior) OEA 1.000 0.631

ei6 (I am a good … others’ emotions) OEA 1.138 0.751 10.367 ***

ei7 (I am sensitive…emotions of others) OEA 1.114 0.783 8.878 ***

ei8 (I have good…people around me) OEA 1.379 0.859 9.106 ***

ei13 (I am able…difficulties rationally) ROE 1.000 0.818

ei14 (I am quite… my own emotions) ROE 0.968 0.795 11.406 ***

ei15 (I can always calm…I am very angry) ROE 0.853 0.660 9.369 ***
ei16 (I have good…own emotions) ROE 0.799 0.700 10.030 ***
***p<0.001 **p<0.01 *p<0.05; Bigfive (bf); Emotional Stability (ES); Dutch Work Addiction Scale (dws);
Working Excessively (WE); Working Compulsively (WC); Emotional Intelligence (ei); Others-Emotions Appraisal
(OEA); Regulation of Emotion (ROE)

It is understood that Confirmatory Factor Analysis is significant for the Measurement model
since model test values (P <0.05), x2 (318.078), x2 / df (1.797) are found in confirmatory factor
analysis. It is understood that the Confirmatory Factor analysis of the Measurement model is
valid since the fit index values of the model are within the acceptable limits of GFI (0.911), CFI
(0.941), SRMR (0.0631), RMSEA (0.060).

Reliability, combined reliability, convergence validity, decomposition validity and

average variance values

In order to collect data, a research questionnaire was designed using the scales in the literature.
The reliability of the scales used for the analysis (Cronbach Alpha ≥ 0.7), composite reliability
(CR ≥ 0.7) and the average variance extracted (AVE ≥ 0.7) values for the convergence validity

67
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

were tested. Decomposition validity values are compared with the correlation value for each
dimension and are expected to be greater than it (Raykov, 1997; Fornell and Larcker, 1981).

Table 3. Correlation, reliability and dissociation validity values of the scales used in the
research

Sub Dimensions MEAN SD ES WE WC OEA ROE

Emotional
3.65 1.113 (0.781)
Stability (ES)

Working
3.54 0.891 0.287** (0.707)
Excessively (WE)

Working
Compulsively 3.47 1.031 0.286** 0.795** (0.715)

(WC)
Others-Emotions
4.27 0.739 0.145* 0.215** 0.153* (0.760)
Appraisal (OEA)

Regulation of
3.93 0.930 0.400** 0.239** 0.211** 0.311** (0.746)
Emotion (ROE)

Reliability
Coefficient - - 0.719 0.848 0.789 0.823 0.821

Cronbach’s Alpha
Composite
- - 0.754 0.875 0.800 0.844 0.833
Reliability (CR)

Average Variance
- - 0.611 0.501 0.512 0.578 0.557
Extracted (AVE)

***p<0.001 **p<0.01 *p<0.05 WE: Working Excessively WC: Working Compulsively ES : Emotional Stability
OEA: Others-Emotions Appraisal , ROE: Regulation of Emotion

68
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

All sub-dimensions of research scales are very reliable according to the research findings (See
Table 3). The Cronbach Alpha, CR and AVE values are enough high for all subdimensions of
research scales such as the others-emotions appraisal (OEA), the regulation of emotion (ROE),
the working excessively (WE), the working compulsively (WC) and the emotional stability
(ES).

It can be stated that the unified reliability condition is fulfilled, since the unified reliability
values are in all CR values (CR> 0.70). As the mean variance values (AVE> 0.50) for all
variables are found, it is seen that the necessary condition is met in the convergence validity.
The square root values of the calculated AVE values are given in brackets in the table in order
to see the validity of decomposition. Since these values are higher than all correlation values in
that column, the separation validity is assumed to be provided for all variables.

Structural Equation Modeling

The research model given in Figure 1 was tested using the path analysis with the observed
variables with the help of the AMOS program version 20.0 (Kline 2005). The statistical
significance of the indirect effects present in the model was tested by the bootstrapping method
used instead of the Sobel test (Preacher and Hayes 2004, 2008).

69
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 2. Testing the mediator model with path analysis by latent variables

It is understood that the model is significant since the model test values x2 (397,166), x2 / df
(2,206) and (p <0.05) in the path analysis model with the observed variables for all data (n =
320). It is understood that the model is valid because the fit index values of the model are
within the acceptable fit limits of GFI (.911), CFI (.961), SRMR (, 0702), RMSEA (, 071).
Presence of meaningless (p> 0.05) path values in the model is predicted to slightly weaken the
model fit index values. Direct and indirect relationships tested with bootstrap to determine
mediator variables in the model are given in the Table 4.

70
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 4. Direct and indirect effect values between independent, dependent and
mediator variables

Hypotheses Direct Direct Indirect Indirect Result

effect effect p effect effect p

OEAESWC 0.645 0.204 -0.560 0.250 Not Mediator

OEAESWE 0.667 0.174 -0.517 0.243 Not Mediator

ROEESWC -1.835 0.001 2.022 0.001 Partial Mediator

ROEESWE -1.704 0.001 1.867 0.001 Partial Mediator

***p<0.001 **p<0.01 *p<0.05 WE: Working Excessively WC: Working Compulsively ES : Emotional Stability
OEA: Others-Emotions Appraisal , ROE: Regulation of Emotion

In the research conducted with 320 people, the following hypotheses were investigated: These
are; (a) mediator role of emotional stability (ES) variable in the effect of others-emotions
appraisal (OEA) variable on working compulsively (WC) variable, (b) emotional stability (ES)
variable in the effect of regulation of emotion (ROE) variable on working compulsively (WC)
variable mediator role, (c) the effect of others-emotions appraisal (OEA) on the working
excessively (WE) variable, the mediator role of the emotionalstability (ES) variable, (d) the
effect of the regulation of emotion (ROE) variable on the working excessively (WE) variable
The mediating role of the variable).

The findings obtained in the analysis made by Bootstarp (n = 2000) method are as follows;

 In the model, the direct effect of others-emotions appraisal (OEA) variable on the
working compulsively (WC) variable (β = 0.645; p> 0.05), but also the emotional
stability (OC) variable of the others-emotions appraisal (OEA) variable to the working
compulsively (WC) variable. The indirect effect (β = -0.560; p> 0.05) through the ES)
variable was not significant. Therefore, it was understood that emotional stability (ES)
variable does not have a mediator role in the effect of others-emotions appraisal (OEA)
variable on working compulsively (WC) variable.
 In the model, the direct effect of others-emotions appraisal (OEA) variable on the
working excessively (WE) variable is not significant (β = .667; p> 0.05), but also the
others-emotions appraisal (OEA) variable to the working excessively (WE) variable.

71
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The indirect effect via the emotional stability (ES) variable was not found significant (β
= -. 517; p> 0.05). Therefore, it is understood that emotional stability (ES) variable does
not have a mediator role in the effect of others-emotions appraisal (OEA) variable on
working excessively (WE) variable.
 In the model, the direct effect of the regulation of emotion (ROE) variable on the
working compulsively (WC) variable is significant (β = -1,835; p <0,05), but also the
emotional of the regulation of emotion (ROE) variable to the working compulsively
(WC) variable. The indirect effect via the stability variable (ES) was found significant
(β = 2.022; p <0.05). Therefore, it can be said that emotional stability (ES) variable
plays a partial mediator role in the effect of regulation of emotion (ROE) variable on
working compulsively (WC) variable.
 In the model, the direct effect of the regulation of emotion (ROE) variable on the
working excessively (WE) variable is significant (β = -1,704; p <0.05), but also
emotional to the regulation of emotion (ROE) variable's working excessively (WE)
variable. The indirect effect via the stability (ES) variable was found significant (β =
1,867; p <0.05). Therefore, it can be said that emotional stability (ES) variable has a
partial mediator role in the effect of regulation of emotion (ROE) variable on working
excessively (WE) variable.

CONCLUSION

The emotional balance of individuals, both in their daily and business life, is a very important
feature and necessity. While the emotional balance of the individual is suppressed by his
emotional intelligence, on the other hand, his good mood deteriorates and even affects his / her
work performance. When the related literature is examined, it is possible to find many
qualitative and quantitative studies in this context.

A new questionnaire was not studied for this research, but a research questionnaire was created
using the scales in the literature. Reliability and validity (Cronbach Alpha, AVE, CR) for the
scales in the research questionnaire were analyzed statistically. In addition, confirmatory factor
analysis was performed for the original scales. Structural equation modeling was used to test
the hypotheses created according to the research model.

72
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

According to the findings, emotional stability has no madiating role in the relationship between
both regulation of emotion working excessively and regulation of emotion-working
compulsively (Hypotheses 1 and 2: Rejected).

On the other hand, research findings say that emotional stability has madiating role in the
relationship between both others-emotions appraisal-working excessively and others-emotions
appraisal-working compulsively (Hypotheses 3 and 4: Accepted).

REFERENCES

Andreassen, C. S., Hetland, J., & Pallesen, S. (2010). The relationship between “workaholism”,
basic needs satisfaction at work and personality. European Journal of Personality, 24(1), 3–17.
doi:10.1002/per.737

Aziz, S. and Tronzo, C.L. (2011). Exploring the Relationship Between Workaholism Facets
and Personality Traits: A Replication in American Workers, The Psychological Record, 61,
269–286. DOI: 10.1007/BF03395760

Bagozzi, P. R., & Yi, Y. (2012). Specification, Evaluation, and Interpretation of Structural
Equation Models. Academy of Marketing Science, 40, 8-34. [Link]
011-0278-x

Barrick, M. R., & Mount, M. K. (1991). The Big Five Personality Dimensions and Job
Performance: A Meta-Analysis. Personnel Psychology, 44(1), 1–26.
[Link]

Benet-Martínez, V., & John, O. P. (1998). Los Cinco Grandes across cultures and ethnic groups:
Multitrait-multimethod analyses of the Big Five in Spanish and English. Journal of Personality
and Social Psychology, 75(3), 729–750. [Link]

Bollen K.A. (1989). Structural equations with latent variables, New York; Wiley.

Burke, R. J., Matthiesen, S. B., & Pallesen, S. (2006). Personality correlates of workaholism.
Personality and Individual Differences, 40(6), 1223–1233. doi:10.1016/[Link].2005.10.017

Charbonneau, D. and Nicol, A. A. M. (2002). Emotional intelligence and leadership in

adolescents. Personality and Individual Differences. 33, 1101-1113. DOI: 10.1016/S0191-
8869(01)00216-1

73
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

De Raad, B., Sullot, E., & Barelds, D. P. H. (2008). Which of the Big Five factors are in need
of situational specification? European Journal of Personality, 22(4), 269–289.
doi:10.1002/per.668

Fornell, C.,&Larcker . D.(1981) . Evaluating Structural Equation Models with Unobservable

Variables and Measurement Error. Journal of Marketing research, 39-50.

Gardner, H (1983). Frames of Mind: The Theory of Multiple Intelligences (New York: Basic
Books).

Goleman, D. (1998). Working with emotional intelligence. New York: Bantam Books.

Jackson, S. S., Fung, M.-C., Moore, M.-A. C., & Jackson, C. J. (2016). Personality and
Workaholism. Personality and Individual Differences, 95, 114 120.
doi:10.1016/[Link].2016.02.020

Kline, R.B. (2005) Principles and Practice of Structural Equation Modelling. New York: The
Guilford Press.

Kun, B., & Demetrovics, Z. (2010). Emotional Intelligence and Addictions: A Systematic
Review, Substance Use & Misuse, 45(7-8), 1131–1160,
[Link]

Law, K. S., Wong, C.-S., & Song, L. J. (2004). The Construct and Criterion Validity of
Emotional Intelligence and Its Potential Utility for Management Studies. Journal of Applied
Psychology, 89(3), 483–496. [Link]

Lopes, P. N., Salovey, P., & Straus, R. (2003). Emotional intelligence, personality, and the
perceived quality of social relationships. Personality and Individual Differences, 35(3), 641–
658. doi:10.1016/s0191-8869(02)00242-8

McCrae, R. R., & Costa, P. T., Jr. (1989). More reasons to adopt the five-factor model.
American Psychologist, 44(2), 451–452. [Link]

Preacher KJ, Hayes AF (2008) Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behav Res Methods 40: 879-91. DOI:
10.3758/BRM.40.3.879

Raykov,T. (1997). Estimation of composite reliability for congeneric measures. Applied

Psychological Measurement, 21, 173-184. [Link]

74
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Robinson, B.E. (1999), The Work Addiction Risk Test: Development of a tentative measure of
workaholism. Perceptual and Motor Skills, 88, 199-210. PMID: 10214644, DOI:
10.2466/pms.1999.88.1.199.

Salovey, P., & Mayer, J. D. (1990). Emotional Intelligence: Imagination, Cognition, and
Personality, 9, 185-211. [Link]

Schaufeli, W.B., Bakker, A.B., Van der Heijden, F.M.M. and Prins, J.T. (2009). Workaholism
among medical residents: it is the combination of working excessively and compulsively that
counts. International Journal of Stress Management, 16(4), 249–272.
[Link]

Scott, K.S., Moore, K.S. and Miceli, M.P. (1997).An Exploration of the Meaning and
Consequences of Workaholism, Human Relations 50(3). DOI: 10.1023/A:1016986307298

Spence, J.T. & Robbins, A.S. (1992). Workaholism: definition, measurement and preliminary
results. Journal of Personality Assessment, 58, 160-178. DOI: 10.1207/s15327752jpa5801_15

Schutte, N. S., Malouff, J. M., Thorsteinsson, E. B., Bhullar, N., & Rooke, S. E. (2007). A meta-
analytic investigation of the relationship between emotional intelligence and health. Personality
and Individual Differences, 42(6), 921–933. [Link]

Schermelleh-Engel, K., Moosbrugger, H., & Müller, H. (2003). Evaluating The Fit Of
Structural Equation Models: Tests Of Significance And Descriptive Goodness- Of- Fit
Measures, Methods Of Psychologial Research Online, 8(2), 23-74.

Schaufeli, W.B., Taris, T.W., & Bakker, A. (2006). Dr. Jekyll and Mr. Hide: On the differences
between work engagement and workaholism. In R. Burke (Ed.): Research companion to
working time and work addiction (pp. 193-217). Edward Elgar: Northampton, MA.

Taris, T.W. & Schaufeli, W.B. (2003). Werk, werk, en nog eens werk : De conceptualisering,
oorzaken en gevolgen van werkverslaving. De Psycholoog, 38, 506-512.

Taris, T.W., Schaufeli, W.B. & Verhoeven, L.C. (2005). Internal and external validation of the
Dutch Work Addiction Risk Test: Implications for jobs and non-work conflict. Journal of
Applied Psychology: An international Review, 54, 37-60.

Thorndike, R. L., Stein, S., An Evaluation of the Attempts to Measure Social Intelligence,
Psychological Bulletin, 34, pp. 275–284, 1937. Google Scholar | Crossref

75
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Wechsler, D. (1939). The Measurement and Appraisal of Adult Intelligence (Baltimore:

Williams & Wilkins Company.

Yildirim, O. (2007). Discriminating emotional intelligence‐based competencies of IT

employees and salespeople, Journal of European Industrial Training, Vol. 31 No. 4, pp. 274-
282. [Link]

Ying-Wen Lıang, & Chen-Mıng Chu. (2009). Personality Traits and Personal and
Organizational Inducements: Antecedents of Workaholism. Social Behavior & Personality: An
International Journal, 37(5), 645–660. [Link]

76
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

RegTech: a comprehensive view from academics, authorities, and

practitioners
Davide Lanfranchi*, Laura Grassi*, Alessandro Faes*,
* Politecnico di Milano, Department of Management Engineering, Italy

Objective
RegTech is gaining interest from academics, authorities, and practitioners, as a relevant
phenomenon promising higher efficiency and effectiveness (FCA, 2016). What is RegTech? How
can be characterized? Is a part of Fintech? Which are the interested stakeholders? Which are the
benefits and the risks of those solutions? Academic literature providing a comprehensive view of
RegTech phenomenon is scarce, and at the moment there is no universal understanding of RegTech
in research. This work aims to make insights into how academics, authorities and practitioners
understand and frame RegTech and provide, building on this, a conceptual framework to organize
and present the main knowledge and results on the concept of RegTech, in order to contribute to an
objective understanding of RegTech.
Data and Method
To identify all the relevant contributions in academic literature studying the phenomenon of
RegTech, we rely on a Systematic Literature Review of academic literature, as well of authorities
and practitioners’ contributions. Content analysis is run on the overall 143 results, collecting all the
codes relative to RegTech phenomenon.
Results
The conceptual framework resulting provides a comprehensive view of the concept of RegTech.
RegTech concept is rationalized in terms of regulation addressed, technologies and data involved,
characteristics, stakeholders, sectors interested, applications for regulated entities and
regulators/authorities, risk and benefits, and impacts on regulatory models.
Relevance
From an academic point of view, relevance and value are clear, being this work aiming at providing
a comprehensive view of the relatively new concept of “RegTech”, a relevant one due to the
potential effects that it may have on the financial system.
From the point of view of authorities and regulators, relevance is clear: giving this work a view on
the opportunities and risk deriving from RegTech both for themselves and for the regulated entities,
authorities can better understand and employ those solutions, along with better regulate their
adoption by regulated entities.
From a practitioner point of view, value is clear, giving this research a comprehensive view of the
benefits that RegTech solutions can provide, fostering their adoption, as well with a view on risk
that they must control.

KeyWords: RegTech, Technology, Regulation, Fintech, Suptech

77
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Introduction

Covid-19 marked a defining moment for the global financial system and the world economy. Financial crisis
of 2008 as well (IMF, 2020), with severe consequences on world economy and on trust in the financial sector
(Baxter, 2016).
Financial regulator responded timely, with measures comparable only to that in the aftermath of the Great
Depression, aiming mainly at enhancing the shock-absorbing capacity of the system (IMF, 2020).
Consequently, cost of compliance increased dramatically.
In the same time span, financial markets, services, and institutions started facing technology-driven
innovations (Arner et al., 2016). Fintech and the application of new technological solutions in financial
sector opened new opportunities and business model possibilities for financial institutions, empowering at
the same time people (Zavolokina et al., 2016). Fintech, however, brought in new risks for the financial
system (Buckley et al., 2020) and this required regulators to better study the phenomenon and evaluate
themselves the adoption of new solutions and new regulatory approaches (Buckley et al., 2020).

In this context, regulated entities and regulators started looking with interest to RegTech, defined by the
Financial Conduct Authority (FCA, 2016) as “a sub-set of FinTech that focuses on technologies that may
facilitate the delivery of regulatory requirements more efficiently and effectively than existing capabilities”.
Interest for RegTech solution increased, and according to CBInsight (2018), since the 2008 financial crisis
17% of the firms have implemented a RegTech solution.

What is RegTech and which are its main characteristics?

Which is the role of technology and data?
Which are the interested parties and how can they employ RegTech?
Which are the risks and benefits of RegTech?
How RegTech will influence regulatory models?

Despite academics started studying RegTech (see for instance Arner et al. (2016), Baxter (2016) or Yang et
Li (2018)), academic literature providing a comprehensive view of RegTech phenomenon is scarce.
This work aims to make insights into how academics, authorities and practitioners understand and frame
RegTech and provides, building on this, a conceptual framework to organize and present the main
knowledge and results on the concept of RegTech, in order to contribute to its objective understanding.
The rest of this work is organized as follows: next section provides the background on which this research is
conducted, third section describes the methodology, fourth section provides main results and the fifth section
discusses them. Last section concludes, highlighting implications and suggesting further research steps.

78
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Setting the scene

Covid-19 marked a defining moment for the global financial system and the world economy. Financial crisis
of 2008 as well (IMF, 2020).
According to World Bank, world level GDP decreased from 63.61 trillion of dollars in 2008 to 60.33 trillion
of dollars in 2009. In the US, where crisis originated, U.S. unemployment rate increased to 9.5% by June
2009, nearly twice the pre-crisis level, while average hours per work week declined to 33 (Nastase et al.,
2009).
It emerged in a severe loss of trust in the financial sector, with high popular anger about the damage of the
economy and individual welfare (Baxter, 2016).
The response of financial regulator was concrete, with a response comparable only to that in the aftermath of
the Great Depression even if with an essential role of international cooperation (IMF, 2020).
The post-crisis reforms aimed mainly at enhancing the shock-absorbing capacity of the system, trying to
improve timeliness, quantity, and quality of the resources to support financial stability (IMF, 2020).
A 492% increase in regulatory changes between 2008-2015 characterized developed markets (Translatlantc
institute, 2017). Consequently, compliance activity increased dramatically.
An example comes from the European context: as resulting by a survey of the European Commission (2019)
on the economic effect of key EU financial legislation, firms pointed out the number of requirements these
legislations introduced, the limited extent to which compliance could be automated, the timing of the
legislative changes, the short transition periods set for businesses to comply with some legislation and the
inconsistency or lack of clarity of the requirements introduced.
On average, 2% to 4% of total operating costs were for compliance to the EU Framework (European
Commission, 2019), and fines exceeding in post-crisis period US$200 billion (Arner et al., 2016).
In the same time span, financial markets, services, and institutions started facing technology-driven
innovations (Arner et al., 2016). “Fintech” grew rapidly after 2008 crisis with investment passing from 930
millions of dollars in 2008 to 4.1 billion of dollars in 2013, tripling during 2014 alone to 12.2 billion of
dollars and nearly doubling in 2015 to 22.3 billion of dollars (Anagnostopoulos, 2018).
Fintech and the application of new technological solutions in financial sector opened new opportunities and
business model possibilities for financial institutions, empowering at the same time people (Zavolokina et al.,
2016). In a similar fashion, technology have been increasingly relevant by financial institutions as well in
meeting their compliance obligations with lower cost (FCA, 2016), showing during time benefits as well in
terms of customer experience (EY, 2016) or business model improvement (Financial Stability Board,
2017b).
Fintech, however, is not free of risks, sometimes providing advantages to financial institutions through
regulatory arbitrage or regulatory avoidance (Buckley et al., 2020).
As a result, regulators started studying this phenomenon to keep pursuing their objectives, with academics
even suggesting new regulatory approaches (Buckley et al., 2020).
These above-mentioned phenomena have been a driving factor in the adoption and use of new technologies
that aid compliance with regulation, i.e. RegTech (Buckley et al., 2020).
According to Arner et al. (2016), the emergence of RegTech is attributable to post-crisis regulation changes
requiring huge volumes of additional data disclosure from supervised entities, developments in data science
allowing the structuring of unstructured data, economic incentives for participants to minimize rising
compliance costs and regulators’ efforts to enhance the efficiency of supervisory tools to foster competition
and pursuit their mandates of financial stability and market integrity.
Indeed, it entails a potential decrease cost and complexity of compliance for regulated entities (Quill et
Lennon, 2019), as well as to enable regulators to better conduct their regulatory and supervisory tasks
(Financial Stability Board, 2017a).

For those reasons as well, academics started studying the phenomenon of RegTech, applying time by time
different lenses. Different dimensions were analysed by academic literature time by time, such as what
RegTech is and its characteristics, the interested actors, technologies involved, possible uses for regulators
and regulated entities until benefits and risks.
Dealing with what RegTech is, one of the first debates were about the relationship between Fintech and
RegTech, present in particular in contributions such as Anagnostopoulos (2018), Soloviev (2018) or
Sangwan et al. (2019), that evaluated the impact of Fintech – and consequently RegTech, considered as a
sub-set of Fintech – on regulators and banks.

79
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The identified characteristics of RegTech solutions however are far more various along the different papers,
considered to be seamless (Kavassalis et al., 2018), non-invasive (FCA, 2016) and customizable (Capgemini,
2018), intuitive (FINRA, 2018), agile (Capgemini, 2018), sustainable and scalable (Quill et Lennon, 2019)
among others.
Different are the viewpoints to approach Regtech. In 2016, Baxter (2016) explained the increasing
complexity in supervising modern financial institutions faced by regulators, suggesting that regulators should
develop their own sophisticated methods of automated supervision, with a clear role of RegTech solutions.
On the other way around, Currie et al. (2018, p.304) focused more on financial institution adoption of
RegTech solutions to meet their compliance obligation, finding how dialectic tensions could arise as the
pursuit of “transparency, surveillance and accountability in compliance mandates is simultaneously
rationalized, facilitated and obscured by regulatory technology“. With a wider point of view, trying to merge
both regulator and regulated point of view, Arnet et al. (2017) described the evolution of RegTech, studying
the use of technology by financial institutions and financial industry to meet regulatory requirements, and the
use by regulators, concluding with a view on new challenges of FinTech and the need for RegTech in
meeting them.
Other contributions focused more on specific RegTech solutions or on more vertical applications of
RegTech, studying time by time specific technologies or specific applications. Examples are Kavassalis et al.
(2018), that proposed a blockchain-based solution aiming at transforming supervisors’ capacity to monitor
risk in the financial system, basing on data preserving informational content of financial instruments at a
granular level; Seppala et al. (2017), that presented a visualization technique aiming at assisting legal staff in
formalising their interpretation of legal texts in terms of regulatory requirements; or Quill et Lennon (2019),
proposing a preliminary solution for automating the generation of compliance documentation.
Looking at benefits and risks, an example comes from Micheler et Whaley (2019), that reviewed the
advantages and disadvantages of combining technology with regulatory law.

As different contributions apply different lenses and analyse different dimensions, to the best of our
knowledge, a wide and comprehensive understanding of Regtech is scarce, which often implies
misunderstandings among actors and non-univocal approaches to regulation.
Building on previous contributions, this research aims at providing a conceptual framework to organize and
present the main knowledge and results on the concept of RegTech, to contribute to an objective
understanding of this phenomenon and its implications for academics, authorities and practitioners.

80
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Methodology

Despite a growing number of contributions studying RegTech phenomenon, little attempt have been found
trying to translate and organize these findings systematically into a comprehensive review of existing
knowledge, aiming at making insights into how academics, authorities and practitioners understand and
frame RegTech. The complexity and the novelty of the issues involved require a systematic review exploring
all the dimensions of the existing literature and empirical evidence.
Following Pittaway et al. (2004) work, the review strategy adopted in this work has several stages
focused on providing a systematic and explicit method for the review.
The different steps taken in this study are presented in two sections, namely Keyword and Eligibility criteria
selection stage and Consistency selection stage.
Keywords and eligibility: Under this stage, the first step was the selection of relevant keywords. “RegTech”
was identified as natural keyword, but to consider research dealing with the same objective of our work, we
relied on the fact that RegTech derives from the contraction of words “regulatory” and “technology” (Arner
et al., 2016), therefore including the Keyword “regulatory” and “technology”.
Consequently, we organized the identified keywords in a string that could be used for a search in the main
databases, defining as a suitable string “RegTech OR (regulatory AND technology).
We searched for citations using this string in Scopus database, as referring to English language papers and
consistently with recent systematic literature review studies (e.g. Alam et al., 2020). Search on Scopus
resulted in in 39556 citations.
In order to deal with the divergent number of results at this point, we identified hence two set of eligibility
criteria, depending on whether results were deriving from the keyword “RegTech” or from the keywords
“regulatory AND technology” AND NOT “RegTech” in order to avoid overlapping.
Figure 1 shows the set of eligibility criteria for both cases. In particular, we selected, for the second set of
keywords, 2009 as starting reference period, due to the attention of this work of post-crisis evolution of the
studied phenomenon. The overall number of results was 1058.

Figure 1: Eligibility Criteria

Consistency selection: To select meaningful results, we processed the Title and Abstract of each contribution
following the principle of minimizing the probability of false negatives. Results not related to the purpose of
the current research project were excluded. The number of results after this filter application was 58. By
reading Introduction and Conclusions sections, following the principle of minimizing false positives, we
obtained a final number of results of 30.
Procedure is summarized in figure 2:

81
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 2: Funnel summarizing selection phases

Financial Authorities are committed as well on the topic, and their multiple contributions result in agendas
for the system. Hence, to understand RegTech with the different views, Authorities reflections must be
considered. As such, we included their papers in our review.
Keywords and eligibility. In details, we searched for contributions on “RegTech” in the websites of the main
financial authorities at world level namely European authorities (European Central Bank, European Banking
Authority, European Securities and Markets Authority, European Insurance and Occupational Pensions
Authority, European Systemic Risk Board), United States authorities (Federal Reserve, Securities and
Exchange Commission, Commodity Futures Trading Commission, The Federal Deposit Insurance
Corporation, Financial Industry Regulatory Authority, Consumer Financial Protection Bureau), British
authorities (Financial Conduct Authority, Bank of England, Prudential Regulation Authority), Chinese
authorities (China Securities Regulatory Commission, China Banking Regulatory Commission, China
Insurance Regulatory Commission), Japanese authorities (Financial Services Agency, Securities and
Exchange Surveillance Commission) and supranational authorities (International Monetary Fund, Financial
Stability Board, Bank for International Settlements).
Consistency selection. Having their own specific structure, we opted for reading the whole resulting
documents, with 58 relevant contributions emerging.

Furthermore, due to the novelty and dynamism of the topic, and to the scarcity of contributions, we relied as
well on industry reports and contributions from practitioners, as previously done in literature (e.g. see Eling
et Lehmann (2018)).
Keywords and eligibility. To identify finally contributions from practitioners, we searched with the keyword
“RegTech” all the results on the global websites of the following firms: McKinsey, Bain & Company, BCG,
PwC, Deloitte, EY, KPMG, Oliver Wyman Accenture and Thomson Reuters.
Consistency selection. Having their own specific structure, we opted for reading the whole resulting
documents, with 55 relevant contributions emerging.

To create a conceptual framework to organize and present the main knowledge and results on the concept of
RegTech, to contribute to an objective understanding of RegTech, we used an inductive approach and a
qualitative process analysis.
We started collecting form contributions all the codes related to the concept of RegTech.
Figure 3 presents the Codebook involved in the analysis.
By applying an iterative process, involving all the authors to reduce the personal bias, we analyzed the
various codes, clustering and conceptualizing them in a four-level conceptual tree, with codes as leaves and
with an increasing level of abstraction.

82
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Topic Item Example

Authors
Year of pubblication
Sources recorded information Title
Journal
Abstract
Regulations Buckley et al., 2019
Technology Arner et al., 2017
Data role Yang et Li, 2018
Relationship with Fintech Anagnostopoulos, 2018
Company or solution FCA, 2018a
Characteristics Baxter, 2016
Stakeholders Buckley et al., 2019
Inducted dimensions of analysis
Sectors involved Yang et Li, 2018
Uses for Regulators FCA, 2017
Uses for Regulated entities Currie et al., 2018
Benefits Barberis et Douglas, 2016
Risks Buckley et al., 2019
New regulatory models O'Riain et al, 2012
Improvment or disruption Kavassalis et al., 2017

Figure 3: Codebook

83
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Results

From our analysis of aforementioned literature, we identified the most common concepts involved when
dealing with RegTech and a conceptual framework explaining in a comprehensive way the concept of
RegTech as seen by academics, authorities and practitioners.
Different keywords emerged from our analysis, trying to label a specific aspect or application of RegTech:
first of all, SupTech, which, according to BIS (2019, p.20), aims to “use machine learning or artificial
intelligence to more efficiently monitor the financial industry".
Other contributions argue that RegTech include use of technology for enhancing operations (Operations
RegTech), for increasing compliance controls (ComplianceTech), for influencing the legislature
(PolicyTech) (Buckley et al., 2020), and for improving risk management practices, including using risk
management proactively (RiskTech) (IMF, 2018).
Focusing on the most frequent codes (see fig.4), we see that RegTech is mainly associated to the concept fo
Regulation (by 73% of contributions) and Compliance (61%).
“Technology” is the other main concept, appearing in the 55% of contribution (further contributions however
report the role of specific technologies, such as Artificial Intelligence (27%), Big Data (19%) or Blockchain
(14%)). Together, “Technology” and “Regulation” concepts appear in 47% of contributions, while
“Technology” and “Compliance” in 36%.
RegTech is often associated to concepts “Efficiency” (40%) and “Effectiveness” (29%), concepts present
since the definition of FCA (2016) “RegTech is a sub-set of FinTech that focuses on technologies that may
facilitate the delivery of regulatory requirements more efficiently and effectively than existing capabilities”.
Together, “Efficiency” and “Effectiveness” appear in 22% of contributions.
Moreover, “Cost reduction” and “Facilitator” are concepts very common when dealing with RegTech (37%
and 39% respectively), highlighting once more how increasing complexity and cost of compliance after 2008
crisis fostered RegTech emergence (Arner et al., 2016).

Most frequent codes in analysed contributions

73%
61% 55%
40% 39% 37% 35% 34% 30% 29%

Figure 4: Most frequent codes in analysed contributions

From the analysis of the contribution, we derived the conceptual framework presented in fig. 5.

84
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 5: RegTech conceptual framework deriving from literature analysis

First, the resulting framework presents in an organized way the characteristics of RegTech as emerging from
literature, whether RegTech should be considered as part of Fintech or not, if RegTech refers to companies
or to solutions in general, and if RegTech consists in an improvement or in a disruption of current
capabilities and models.
Regulations addressed and technologies involved are then two main factors constituting RegTech.
Actors interested in RegTech are several and belonging to different sectors. In particular, regulated entities
and regulators/authorities may find relevant application of RegTech, with implications on the whole system.
The role of data in all of this is crucial, being a structural component of how above-mentioned actors can
leverage RegTech to achieve the benefits, as shown in the right-side of the framework.
However, risks related to RegTech must be considered as well.
Finally, on the left side of the framework is clear how exists a mutual relation between RegTech and
regulatory models.
Each block and relation in the framework will be deeply described in the discussion section.

85
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5. Discussion

In present section, we discuss the resulting conceptual framework presented before, building on contributions
analysed.

5.1 What is RegTech?

The block at the centre of the framework presents in a organized way the main results emerging from
literature discussing what RegTech is, in terms of characteristics, relationship with Fintech, whether it should
be considered a company or a solution, an improvement or a disruption.
5.1.1 The characteristics of RegTech
Academic literature and other analysed contributions provide a wide set of common characteristics that
RegTech has (or should have).
First of all, RegTech is an innovation and a driver to innovate ("“RegTech” provides a valuable opportunity
for the FCA to innovate, as a regulator, by using technology to augment and support all that we do" (FCA,
2017, p.5).
RegTech is considered both an enabler (it can “enable regulators to keep up with the very rapid evolution of
markets and their underlying technological development” as suggested by Baxter (2016, p.572)) and a
facilitator, for instance in the delivery of regulatory requirements (Anagnostopoulos, 2018).
RegTech is often considered a smart (Bain&Co, 2016), digital solution, for instance in the context of digital
reporting (IMF, 2019a), automated ("Cost can be reduced due to automated mapping of tasks to risks and on
to compliance requirements" (Quill et Lennon, 2019, p.410)) and mainly a B2B solution ("We can therefore
expect RegTech to focus more on business-to-business (B2B)" (Arner et al., 2016, p.384).
RegTech enables a view that is wide, "across the organization to conduct enterprise-level reviews" (FINRA,
2018, p.6) and holistic (for instance, "Some market participants are also seeking to combine data obtained
directly from customers with data from external sources, and then processing this data using sophisticated
data analytics to create a more holistic view of the customer” (FINRA, 2018, p.4)).
Furthermore, RegTech solutions are considered to be seamless (Kavassalis et al., 2018), non-invasive (FCA,
2016) and customizable (Capgemini, 2018), intuitive (FINRA, 2018), agile (Capgemini, 2018), sustainable
and scalable (Quill et Lennon, 2019), modular (EY, 2018a) and reliable (EY, 2018b).
RegTech promises to be dynamic (Sangwan et al., 2019), quick and to enable real-time operations (examples
are "real-time risk analysis tools to help institutions spot fraud more quickly" (Michaels et Homer, 2018,
p.340) or RegTech solutions making “compliance easier for regulated entities” by supplying regulators with
“more accurate and real time information" (Micheler et Whaley, 2019, p.8).
RegTech solutions are furthermore integrated, standardized (see for instance the framework proposed by
Alrabiah (2018) integrating a private cloud computing network with standardised, automated and integrated
features aiming at enhancing the regulatory reporting system) and interoperable (FCA, 2016).
Finally, "RegTech provides senior executives with an opportunity to introduce new capabilities that are
designed to leverage existing systems" (Deloitte, 2016, p.3), but a cultural change is nevertheless required, as
highlighted by Accenture (2018). For conscious adopters, RegTech could be a strategic differentiator
(Accenture, 2017a).
5.1.2 Fintech or not Fintech?
There is no consensus in literature on whether RegTech should be considered a sub-set of Fintech or not.
The discussion is not merely theoretical: indeed, even though several authors consider RegTech as an
example of Fintech (see for instance Anagnostopoulos (2018, p.13) assessing that "RegTech refers to a sub-
division of the fintech sector that focuses on technologies that may facilitate the delivery of regulatory
requirements more efficiently and effectively than existing capabilities", or FCA (2016) definition as well
“RegTech is a sub-set of FinTech that focuses on technologies that may facilitate the delivery of regulatory
requirements more efficiently and effectively than existing capabilities”), others highlight how it is important
to differentiate the two concepts, due for instance to different underlying causes, to a different development
path (Fintech growing as a bottom-up movement led by startups and IT firms, RegTech as a response to top-
down institutional demand) and to possible application also in non-financial sectors of RegTech (Arner et al.,
2016).
Moreover, Arner et al. (2016, p.382) highlight further how RegTech cannot be considered as part of Fintech,
being “more than just an efficiency tool and rather is a pivotal change leading to a paradigm shift in
regulation”

86
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.1.3 Solution or company?

There are authors labelling with the keyword “RegTech” specific companies developing solutions employing
technology to address compliance (see for instance McKinsey (2017, p.96): "RegTechs (start-ups with a
focus on issues surrounding financial regulation)"), but from our analysis emerges how “RegTech” refers
more widely to a solution (e.g. “"The Authorities intend, subject to applicable domestic laws and regulations,
to share further information on regulatory technology (regtech) developments, including experiences with
trialling regtech solutions for regulatory purposes" (FCA, 2018a, p.6).
5.1.4 Improvement or disruption?
From the analysis of contributions, emerges how RegTech can be an improvement of existing capabilities
(for instance RegTech can improve efficiency or effectiveness (FCA, 2016), or, as highlighted by Kavassalis
et al. (2018, p.41), "improve the monitoring of the behaviour of financial institutions in almost real time")
but as well have the potential to disrupt pre-existing structures, re-shaping current regulatory processes and
systems (FCA, 2018b)
Arner et al. (2016) suggest that RegTech developments are at present incremental, but it is ultimately moving
towards a paradigm shift requiring a reconceptualization of financial regulation.

5.2 Reg and Tech: the role regulation and technology

As clear also from the origin of the buzzword, regulation and technologies have a main role in fostering and
enabling RegTech solutions, as clear also from fig. 4, affecting what RegTech is.
5.2.1 Regulations addressed by RegTech
RegTech solutions resulted to address several different regulations, in a broad sense (i.e. including for
example privacy issues), both at national and international level. It could focus on a single regulation or on
multiple at the same time.
Considering RegTech solutions dealing with a specific regulation, common examples are RegTech solutions
dealing with Know-Your-Customer directives (Buckley et al., 2020).
Considering on the other hand RegTech solutions and technologies that can help addressing several
regulations simultaneously, Financial Stability Board (2017a) suggests how Natural Language Processing
could support for instance investment managers in comply with different requirements, from Markets in
Financial Instruments Directive (MiFID II) to the Alternative Investment Fund Managers Directive
(AIFMD).
However, not all RegTech solutions address specifically regulations, an example being certain RegTech
solutions for internal reporting (Deloitte, 2017a).
5.2.2 Technology for RegTech
RegTech solutions involve a wide set of technologies, such as Big Data, Artificial Intelligence (Financial
Stability Board, 2017a), API (relevant in encouraging integration and interoperability between systems
(FCA, 2016), or cloud computing (Yang et Li, 2018).
The US Securities and Exchange Commission (SEC) leverages Big Data to develop text analytics and
machine learning algorithms to detect fraud and misconduct (Financial Stability Board, 2017a).
Artificial intelligence play as well a central role (for instance Central banks can use AI to assist with
monetary policy assessments (Financial Stability Board, 2017a)) and in a similar way do machine learning
(an example comes from ASIC, using machine learning software to identify misleading marketing in certain
sub-sectors, such as unlicensed accountants in the provision of financial advice (Financial Stability Board,
2017a) and Deep Learning (Arner et al., 2016). A relevant application of Artificial Technology is Natural
Language Processing (NLP): for instance, "EIOPA has already experimented in the past with some of these
new technologies and their application to supervision, like the social media monitoring tool which used
Natural Language Processing (NLP) technologies to assess the sentiment of comments about insurance in
social media" (EIOPA, 2020, p.3).
Furthermore, modelling/visualisation technology could allow “the simulation of actions and interactions to
assess their effects on the system as a whole" (FCA, 2016, p.8).
Technologies for automation are relevant, enabled by suitable algorithms (Baxter, 2016): indeed "Increased
automation of compliance processes is one of the most widely used forms of RegTech and offers many
potential benefits. For example, use of robotics process automation (RPA) may allow firms to minimize the
need to perform repetitive tasks" (FINRA, 2018, p.6).
A pre-requisite for automation of certain tasks is the possibility to have machine-readable regulation, paying
attention however to risks: as highlighted by Micheler et Whaley (2019, p.16), "It is, for example, not yet

87
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

clear how easy it will be to ‘map’ machine readable code or new regulatory software tools onto existing IT
systems".
Blockchain and Distributed Ledger Technology are discussed dealing with RegTech: IIF highlighted in its
RegTech Report that “[n]ear real-time settlement could be achieved through automation and global
consensus on the blockchain. These capabilities could automate compliance aspects in use cases including
cross-border payments, syndicated loans, and repo markets” (Arner et al., 2016, p.379).
Other technologies playing an important role in RegTech are biometrics, that “measures people’s
physical characteristics and could allow more efficient and robust ways to verify identity"(PwC, 2016, p.5)
and Internet of Things (Goul, 2019). Mobile phones themselves can be involved in RegTech solutions:
“Using this input along with other data, perhaps collected automatically from mobile phones (geo-location
and transaction data, for example), regulators could identify issues and providers that warrant increased
scrutiny" (Michaels et Homer, 2018, p.340)).

5.3 Grounding RegTech: interested actors and applications

RegTech is not relevant only from a theoretical point of view but is showing strong implications for a wide
set of actors, along with for the whole system. In this paragraph, we will discuss which are the main
stakeholder of RegTech, from which sectors, and which are the main identified RegTech application for each
of them.
5.3.1 RegTech stakeholder
From literature and contributions, it is clear how RegTech is of interest not only for regulators and regulated
entities.
Stakeholders are different and consist both of public and private entities.
Considering public entities, regulators and policymakers are clearly interested, being new technologies
enabling them to respond to the growing trend toward digitization among regulated entities (Yang et Li,
2018). Furthermore, Central banks are strongly interested in RegTech, as highlighted by Financial Stability
Board (2017a), showing how they can use Artificial Intelligence to help with monetary policy assessments
and how Central banks themselves expect an increasing use of big data for macroeconomic and financial
stability goals.
Supervisors in general can benefit from RegTech (FCA, 2018b).
Finally, governments as well should look at RegTech, even though their “long suffered shortages in IT skills
and literacy, and the complex skill sets required for big data analytics pose an acute challenge” as
highlighted by Yang et Li (2018, p.3259).
On the other hand, it is even more wide the set of private entities interested in RegTech phenomenon.
Potential of RegTech solutions is clearly present for regulated entities (Micheler et Whaley, 2019) and those
firms in general that must provide their regulatory data (FCA, 2016).
In the financial sector, intermediaries in general are affected by RegTech solutions (Buckley et al., 2020).
Not regulated entities, however, should be considered as well interested parties of RegTech phenomenon:
Tech companies, consulting firms and Fintech startups can identify opportunities form RegTech
development (Yang et Li, 2018), as well as service providers (Goul, 2019).
Employees and managers of firms themselves should look with interest at how RegTech solutions are
evolving: indeed, “regulatory software could monitor other data files, email and voice communication
carried out by employees of regulated entities" (Micheler et Whaley, 2019, p.8) and "Technology mediates
the information processes at the inter and intra-organizational levels which may lead to a ‘moral hazard’
(Greenwald and Stiglitz, 1990) with incomplete or inaccurate information used by senior managers to
obfuscate, mislead, distort, or confuse regulators and investors" (Currie et al., 2018, p.306).
Not only firms are interested in RegTech: industry association and investors should participate to data-
sharing with regulators to create basis for integrated technology-driven regulation (Yang et Li, 2018).
Non-profit organizations and donors participate as well at the development of RegTech, as done for instance
by the Bill & Melinda Gates Foundation, Omidyar Network and USAID, that in 2016 backed an initiative in
partnership with selected financial authorities in Philippines, Ghana and Mexico to develop tools and
techniques for improve market supervision and policy analysis (Gurung et Perlman, 2018).
Last but not least, consumers benefit from RegTech development, for instance to increase their protection
relying on increasing amount of information and new technologies (Arner et al., 2016).
5.3.2 Sectors

88
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The main sector involved in contributions dealing with RegTech is the financial sector, a heavily regulated
sector, in particular after the 2008 crisis (see for instance Transatlantic institute (2017)).
In this direction, RegTech is relevant for several sub-sectors, from banking (Oliver Wyman, 2018), asset
management (Financial Stability Board, 2017a), payments and securities market (Buckley et al., 2020),
lending and trading (Arner et al., 2016) until insurance (Deloitte, 2018a).
However, other authors suggest how other sectors as well could benefit from RegTech solutions adoption,
for instance in monitoring corporations for environmental compliance, trucking companies for speeding
infractions until tracking the global location of airliners, as suggested by Arner et al. (2016).
5.3.3 RegTech for Regulated Entities
We identified five main functions in which RegTech can support firms, and in particular those subject to
regulation: compliance, monitoring, risk management, reporting and operations.
Considering compliance function, RegTech solutions can “enable firms to keep track of regulatory updates in
a timely and efficient manner"(Capgemini, 2018, p.3), along with "assisting with regulatory intelligence
typically provide a catalog of regulatory requirements in a user-friendly manner" (FINRA, 2018, p.5)
Know-your-customer and Customer Due Diligence is one of the main fields where RegTech solutions are
emerging (Buckley et al., 2020), providing for instance new ways to verify identity with biometrics (Deloitte,
2017b).
In this direction Anti-Money Laundering compliance as well is under the lens of RegTech solutions: as
reported by Arner et al. (2016, p.375) “"AML and KYC has been a particular focus of RegTech spending
and development within major financial institutions and in IT and advisory firms and start-ups, such that the
majority of RegTech solutions to date have focused on KYC compliance".
With RegTech solutions, customer can be better profiled (FINRA, 2018, p.6), complying for instance with
MiFID II requirements for European institutions.
Furthermore, privacy is considered by RegTech, as a "customer-led RegTech approach to GDPR can
include: Adopting privacy-preserving analytics such as privacy-preserving linkages to obtain customer
insights while still complying with GDPR" (Accenture, 2017a)
RegTech solutions focus as well on Counter Terrorist Financing regulation, as they “may improve cross-
border payments, including by offering better and cheaper services, and lowering the cost of compliance with
anti-money laundering and combating the financing of terrorism (AML/CFT) regulation" (IMF, 2017a, p.5).
Capital and liquidity ratios requirement (Deloitte, 2017c) and disclosure (Accenture, 2017b) are as well
considered by RegTech.
These solutions moreover support entities as well in detecting conflict of interest (Deloitte, 2017b) and inside
trading events (Buckley et al., 2020).
In future, authors such as Arner et al. (2016, p.411) highlight a "progressive alignment underway in how
FinTech and RegTech are evolving, both sharing data-centricity as a common denominator. This
represents a paradigm shift from a KYC approach towards a KYD (KnowYour-Data) paradigm, which,
while profound, remains a few years away."
In monitoring function, RegTech support in controlling potential risky situations, providing alerts (FINRA,
2018). Monitored variables are internal communications (vocal (KPMG, 2017a) or textual (Das et Kim,
2017)), suspicious activities (FED, 2016) or behaviours of employees (EY, 2016) and client activities
(FINRA, 2018).
Risk management is the third function where RegTech can be applied, from identification (FED, 2019),
monitoring (KPMG, 2017b) to analysis of risks (Michaels et Homer, 2018).
Different risks are considered, from compliance risk (Capgemini, 2018), to cyber risk (FINRA, 2018) or
operational risk (Kavassalis et al., 2018).
Furthermore, RegTech “may be deployed to gather and analyse information on capital and liquidity for
use in internal models" (FINRA, 2018, p.5) or "offer synthesized visualization of complex analytics and
intuitive tools for end users to extrapolate different scenarios" (FINRA, 2018, p.7).
RegTech can be then applied for reporting, both to supervisors ("RegTech in Europe developed rapidly
with the introduction of extensive, purely digital, reporting from intermediaries to regulators" (Buckley et al.,
2020, p.2) and internal (Deloitte, 2017a).
Finally, RegTech can support as well in day by day operations, for instance in onboarding processes of
customer by "removing friction" ([Link]
[Link]), of employees ("RegTech is helps ensure that all branches of the organization are running of the
same standards and all reaching the correct laws and regulations in order to protect both the company and
staff. The onboarding of new staff is also enhanced as they become aware of the guidelines and

89
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

standards to follow" (Quill et Lennon, 2019, p.410) and of vendors ("A best-in-class, investigative regtech
asset has become indispensable for financial companies seeking to mitigate counterparty risk in the diligence
and onboarding phases of business, vendor and customer relationships"
([Link]
Moreover, RegTech technologies can help “ build a “balanced scorecard” for HR decisions" (Deloitte,
2017b, p.25).
5.3.4 RegTech for Regulators / Authorities
Authorities can leverage on RegTech solutions in their functions as well: as highlighted by FCA (2017, p.5),
““RegTech” provides a valuable opportunity for the FCA to innovate, as a regulator, by using technology to
augment and support all that we do (such as making policy, authorising, supervising and enforcing
activities)".
A further suggestion is given by Micheler et Whaley (2019, p.7), highlighting how "software with a ‘smart
contract’ functionality that identifies breaches, makes suggestions for sanctions or perhaps even
automatically issues fines".
In general, as reported by EIOPA (2020), RegTech can be used to automate and streamline administrative
and operational procedures.
From the analysis of contribution emerges clearly how RegTech can “enhance efficiency and effectiveness
of supervision and surveillance” (Financial Stability Board, 2017a, p.20), supporting them in monitoring and
controlling both macro and micro variables.
Regarding macro variables, "Using data to monitor and safeguard the system is essential" (Bank of England,
2019, p.10).
Examples are monetary policy, systemic risk analysis and risk propagation channels (Financial Stability
Board, 2017a). Moreover, "NLP tools may help authorities to detect, measure, predict, and anticipate, among
other things, market volatility, liquidity risks, financial stress, housing prices, and unemployment" (Financial
Stability Board, 2017a, p.21).
A concrete example comes from Bank of Italy, that, with the aim of studying the redistributive effects of
fiscal policy over different municipalities, relied on Artificial Intelligence (Financial Stability Board, 2017a).
Authorities can however supervise as well micro variables, controlling, thanks to technology, the entire
population ([Link]
“The use of RegTech can add value to regulators by helping them understand, in closer to real-time,
innovative products and complex transactions, market manipulation, internal fraud, and risks" (Arner et al.,
2016, p.389).
Authorities can supervise the conduct of the market participants (FINRA, 2018), detecting suspicious
activities such as money laundering or terrorist financing (IMF, 2017b), preventing frauds (Micheler et
Whaley, 2019) and identifying red flags (EIOPA,2016).
Authorities can control as well the state of health of regulated entities, monitoring “insolvency, ability,
liquidity, and other risk factors in real time” with the benefit of enhance “market stability and
competitiveness" (Yang et Li, 2018, p.3262).
Ethical sphere is affected as well, due to the "role of regulatory technology to enhance surveillance of
financial activities and encourage better ethical practices” (Currie et al., 2018, p.307).

5.4 The role of Data

It is clear from analysis how Data play a central role in RegTech solutions.
Indeed, Data are fundamental to enable and empower entities to apply proficiently RegTech, constituting an
essential raw material for making RegTech work in the real world.
In this section we will provide a vertical view on how Data are collected, managed and analysed in the
context of RegTech phenomenon.
First, Data involved can be of various type, both structured data (such as trade orders and cancels, market
data, customer portfolio) and unstructured data (e.g. emails, voice recordings, social media profiles or others
communications) (FINRA, 2018), qualitative and quantitative data(EIOPA, 2020), granular (Deloitte, 2017a)
and aggregated (Yang et Li, 2018).
Sources from which data can be collected are heterogeneous, including regulatory databases (Michaels et
Homer, 2018), public record database ([Link]
suspicious-activity-reports), internal sources such as internal communications (Das et Kim, 2017) until social
media (an example is provided by Currie et al. (2018), reporting how studies on innovations have explored
the use of blogs, social media), or surveillance video and satellite positioning (Yang et Li, 2018).

90
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Data collection can be automated (Yang et Li, 2018) and RegTech solutions “seeks to provide configurable,
reliable, easy to integrate, secure and cost-effective regulatory solutions as an option to extract as much data
as possible" (EY, 2018b, p.24). In this context, regulators can evaluate as well to pulling data directly from
firms’ systems (Bank of England, 2019), and as well "Some market participants are also seeking to combine
data obtained directly from customers " (FINRA, 2018, p.4).
Different type of data can be collected by different sources, but “Data integrity and control is of paramount
importance for many RegTech tools" (FINRA, 2018, p.9): indeed, "technology-driven regulation focuses on
data aggregation, big data processing and interpretation, modeling analysis and forecasting, which all rely on
high-quality data” (Yang et Li, 2018, p.3262). For this reason, many RegTech solutions focus on the
improvement of data quality (Arner et al., 2016, p.375).
Once collected, Data must be properly managed. RegTech can “support the technical handling of large
amounts of data" (Buckley et al., 2020, p.3), storage (Kavassalis et al., 2018) and consolidate information
(PwC, 2018) but as well control and protect them (as reported by KPMG (2017, p.10) "For financial
institutions, implementing RegTech solutions that lead to improved data protection and quality, customer
suitability standards and audit trail may engender a favourable supervisory response").
RegTech solutions focus then on Data analysis, automatizing it (Arner et al., 2016), processing Big Data
(Yang et Li, 2018) and performing data mining (e.g. Currie et al. (2018, p.305) talk about " technologies
used by regulators to identify market manipulation by data mining") in order to obtain Data-driven outcomes
(Deloitte, 2018b).
In all of this, is gaining importance the concept of data ecosystem (O'Riain et al, 2012), where data are
shared “among regulators, industry associations, and investors, which is the basis for integrated technology-
driven regulation" (Yang et Li, 2018, p.3263).

5.5 Benefits and Risks

From literature emerges how RegTech promises to bring benefits for regulated entities, regulators and for the
whole system as well. However, risks related to RegTech solutions are as well evaluated in literature, and
hereby presented.
5.5.1 Benefits of RegTech
From our analysis emerge several benefits coming from adoption of RegTech solutions, both for single
entities and for the whole system.
Focusing on benefits for the single entity, we found benefits that are mainly for regulated entities, mainly for
regulators and benefits equally relevant for both.
Focusing on the latter class, higher efficiency and efficacy are among the most cited benefits (for instance by
FCA (2016)).
RegTech may increase accuracy (Micheler et Whaley, 2019), transparency and manageability (Baxter, 2016)
while driving down costs (see for instance Quill et Lennon (2019, p410): "Cost can be reduced due to
automated mapping of tasks to risks and on to compliance requirements.") and times ("Intelligent automation
can also enable firms to keep track of regulatory updates in a timely and efficient manner, resulting not only
in lower risk of non-compliance but also in time savings to focus on more value-adding tasks" (Capgemini,
2018, p.3).
RegTech allows to liberate significant amounts of resources (see for instance "the Monetary Authority of
Singapore (MAS)” that “is exploring the use of AI and machine learning in the analysis of suspicious
transactions to identify those transactions that warrant further attention, allowing supervisors to focus their
resources on higher risk transactions" (Financial Stability Board, 2017a, p.23) reducing at the same time the
errors (Currie et al., 2018), the false alerts (FINRA, 2018) and the risks, such as operational one (Kavassalis
et al., 2018).
RegTech can improve governance (Currie et al., 2018) and coordination inside the entity in general (KPMG,
2017a), protecting all the staff (as reported by Quill et Lennon (2019, p.410), "RegTech is helps ensure that
all branches of the organization are running of the same standards and all reaching the correct laws and
regulations in order to protect both the company and staff").
RegTech enables a smarter analysis of information (Deloitte, 2018a), enabling a better cost-benefits analysis
(Enriques, 2017) and improving consequently the decision-making process (Arner et al., 2016).
While providing a higher degree of flexibility (FCA, 2016), RegTech is consider to increase robustness as
well (for instance in the verification of identity (PwC, 2016)).

91
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Focusing more on benefits mainly for regulated entities, we see a reduction in the overall compliance
process, with a reduction of the regulatory burden (Deloitte, 2017b) and regulatory complexity (Currie et al.,
2018), resulting in a better interpretation of regulation (FCA, 2017), better understanding of the implications
(ESMA, 2017) and lower risk on non-compliance (Capgemini, 2018).
Moreover, RegTech can participate to the creation of value of the firm (Deloitte, 2016), that is able
furthermore to transfer part of this value to consumers
([Link] enhancing their experience
(EY, 2016) and protection.
All of this can improve the performances of the firm, increasing profits
([Link] and providing a competitive
advantage (Accenture, 2018).
Regulators finally can better understand, thank to RegTech, innovative products ("The use of RegTech can
add value to regulators by helping them understand, in closer to real-time, innovative products and complex
transactions, market manipulation, internal fraud, and risks" (Arner et al., 2016, p.389).
Focusing on the benefits for the whole system, our analysis highlighted different suggestions from literature.
RegTech is considered to contribute to the economic growth (Arner et al., 2019) and stability (Buckley et al.,
2020) of the system, reducing for instance information asymmetry (Currie et al., 2018) and enabling a better
regulation (Barberis et Arner, 2016) with different authorities collaborating (IMF, 2019b).
The increased competition fostered by RegTech (FCA, 2017) and the enhance financial inclusion (Arner et
al., 2019) imply benefits for the final consumers, that are more protected (Arner et al., 2016) thank as well to
a higher security of digital services (Accenture, 2018) and reduction of financial crimes (Oliver Wyman,
2018). Markets, finally, may become, thanks to RegTech, more trustable (Goul, 2019), integer (Buckley et
al., 2020), stable (PwC, 2017), open (Michaels et Homer, 2018), safe (FINRA, 2018) and agile (EY, 2016).
5.5.2 Risks of RegTech
Literature provides several benefits of RegTech solutions, but risks are discussed and considered by several
authors.
One of the main risks is cyber risk (Buckley et al., 2020) due to the high level of technology involvement.
In a similar way, RegTech solutions may “pose potential risks associated with customer data privacy,
particularly where customer data is shared with a third-party vendor" (FINRA, 2018, p.9).
Relying on algorithms brings with it the problem of algorithmic biases (Gurung et Perlman, 2018, p.26).
Regulator itself may have serious challenges if not able to properly regulate RegTech: EBA (2019, p.13)
reported “a lack of provisions in the current legal framework dealing with RegTech solutions, which
means that different standards are applied by different solutions;"
The way in which human resources and technology will coexist may pose serious challenges: firms’ over‐
reliance on information technology solutions could “lead to a loss of human professional expertise and
judgement in monitoring processes" (EBA, 2019, p.13), firms themselves could suffer a lack of
understanding of new technologies" (EBA, 2019, p.13) and even if skilled people is available, different
competences could have difficulties in communicating: as highlighted by Buckley et al. (2020, p.11)
"Regulatory technology takes advantage of computer science tools that are relatively
new. Our knowledge and understanding of any new technology is initially and invariably limited. In
addition, it combines computer science with law. Those trained in law do not normally know about the
characteristics and limitations of computer software. Those trained in computer science are not normally
familiar with the scope and subtleties of legal terms. Neither group is well placed to anticipate problems
that may arise when the two are combined. They may not even be in a good position to appreciate what it is
they do not know. This makes it difficult for either group of experts to at least ask the right questions.".
There is furthermore the risk that RegTech solutions could not only not work properly, but even increase the
problems they should be meant to solve: as reported by EBA (2018. P.28) “"there is a risk that these
RegTech solutions could potentially weaken ML/TF safeguards, if applied unthinkingly".
Systemic risk could increase as well according to (Micheler et Whaley, 2019, p.25): "Regulatory technology
could cause regulation to become more granular, leading to inflexibility, technical rather than functional
compliance and increasing standardisation. This could generate systemic risk".
Transparency and accountability could be reduced (Currie et al., 2018)
Old risk could be replaced by new ones: " antitrust risks and the risks for markets resulting from extremely
swift transmission of information will increase and require further investigation" (Buckley et al., 2020, p.8)
and RegTech could pose “unique risk such as resiliency and confidentiality" (FED, 2019, p.6).

92
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In general, risks could increase if over reliance on RegTech solutions leads regulator and regulated entities
in a false sense of security (Micheler et Whaley, 2019) or if system focuses too much on short-term risks,
neglecting long-term ones: Buckley et al. (2020, p.18) assessed that "When the regulator receives real-time
transactional information its systems can respond in real-time. There is a risk that this encourages regulated
entities to orient themselves towards impressing the regulator in real-time. They could become too focused
on real-time reporting, orient their business model accordingly and inadvertently overlook longer-term
risks".
RegTech could increase furthermore risks if certain market participants apply it in a malicious way (Anti-
RegTech as called by Packin (2017)): "One risk that authorities should bear in mind when developing
automated detection tools is the possibility that malicious agents may learn to frustrate the tools by adapting
their behaviour. For instance, market participants could, in theory, learn what types of behaviours are likely
to cause a flag in a SupTech monitoring system. Using such information, firms might be able to structure
their regulatory returns in such a way as to remain undetected." (ESMA, 2019, p.46).
Adopting RegTech solutions could be not so easy for entities, for instance due to high initial investments
required (Packin, 2017), and the ability to adopt those technologies could be different between regulators and
regulated entities: for instance, “the potential risk to financial institutions is that if the regulator has access to
technological capacity far in advance of the organization itself, the regulator could predict risk areas that the
institution does not see coming" ([Link]
[Link]) but if the opposite happens, other risk (such as the above-mentioned Anti-RegTech) could arise.
Those participants not able to adopt new solutions could find themselves “with platforms ill suited for the
current regulatory framework” (ESMA, 2017, p.6).
Those actors could outsource solutions, but "when firms are outsourcing all or part of their activities to
RegTech providers without proper oversight and governance arrangements in place, it may lead to:
difficulties with accessing customer data owing to RegTech providers’ potentially short lifespan and with
establishing the ownership of that data; questions about the reliability of records held owing to unsound and
unsafe record‐keeping practices put in place by the RegTech provider; a lack of transparency in the
allocation of responsibilities between firms and RegTech providers, particularly when the processes are
outsourced to providers that are not obliged entities under the AMLD." (EBA, 2019, p.13).
Moreover, problems that “can emerge are exacerbated by the fact that there is a potential for the oligopolistic
market that is currently dominating data analysis to move into the realm of regulation" (Micheler et Whaley,
2019, p.3).

5.5.3 RegTech and regulatory models

RegTech is considered to fit and contribute to different possible regulatory approaches, and the different
approaches may influence the evolution of RegTech as well.
Goul (2019) highlighted the role of RegTech in three different approaches: insight-based approach,
principle-based approach and rules-based approach:
"insight-based approach is seen as the most flexible. Regtech to support this approach is increasingly
data-driven and is designed to reduce friction between involved parties because information is more readily
shared and can be acted upon by both regulators and businesses using advanced analytics and AI"
(Goul, 2019, p.220);
"the principles-based approach seeks to increase dialog between regulators and business entities. Standards
are often set for desired outcomes, and there are consequences for non-compliance. There is more flexibility
in this approach than with a rules-based approach as it provides some flexibility and freedom to interpret
standards as fintech services rapidly adapt, scale and improve. This aspect of uncertainty and unpredictability
is more appropriate for Fintech where services provided are
constantly changing and evolving. Of course, Regtech implementors have more difficulty devising solutions
that have the desired flexibility and freedom than in the rulesbased approach, and they need to be partner to
ongoing dialogue between parties as changes occur" (Goul, 2019, p.220);
“Regtech is evolving along three dimensions: rules-based, principles-based and insight-based [...] a rules-
based approach is most rigid; it implies there are detailed rules, transparency is required and strong audit
trails are needed to document for possible litigation.” (Goul, 2019, p.220).
In a similar fashion, Micheler et Whaley (2019) analysed the role of RegTech in a different framework of
possible regulatory approaches: self-regulation, command-regulation, meta-regulation.
"Self-regulatory approaches are likely to resurface particularly for regulatory technology. An argument could
be made that the availability of new technological tools makes it possible for the regulator to step back and

93
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

leave it to the market, now equipped with regulatory technology, to create appropriate frameworks."
(Micheler et Whaley, 2019, p.20);
"The essence of a command regulatory strategy is in its control of the achievement of certain outcomes by
imposing sanctions where outcomes are not met. The government is in charge. It writes rules and designs
sanctions through primary or secondary legislation and a regulatory body enforces them. The regime that
was put in place after the financial crisis can in large parts be characterised as a command
regime. A command approach could integrate regulatory technology. The government could control the
development, its maintenance and updates of the technology. It could issue software requiring regulated
entities to run that software on their systems." (Micheler et Whaley, 2019, p19);
"Meta-regulation has been described as ‘the state’s oversight of self-regulatory arrangements’, and also as
‘interactions between different regulatory actors or levels of regulation’. It occupies a middle ground
somewhere between command regulation with a high level of government involvement and self-regulation
with a minimal amount of government involvement. The regulator delegates risk control to the regulated
entities themselves, giving them primary responsibility for the risk management systems, while the regulator
audits, monitors and incentivises the systems. The regulator steers, the regulated entities row. The regulator
would not design or maintain regulatory software itself but oversee and validate its production. The regulator
could specify requirements leading to the creation of a distributed reporting ledger leaving the development
and maintenance of the system to regulated entities or their providers. The regulator could issue a machine-
readable version of the rules for reporting specific data points" (Micheler et Whaley, 2019, p22).
Buckley et al., (2020, p.3) suggest that "As laid out in previous research, the new risks created by FinTech
can be addressed by new approaches to regulation (which we have termed Smart Regulation) which
incorporate an ecosystem design approach involving regulatory and supervisory technologies (collectively
referred to as RegTech)".
Furthermore, a machine-readable regulation (O'Riain et al, 2012) could enable further inbuilt-compliance,
where “compliance Regulatory requirements can be coded into automated rules applied when relevant.
A system that can automatically apply the regulatory ‘programme code’ would improve compliance,
reducing regulatory and staff costs" (FCA, 2016, p.9).
More in general, from our analysis emerges how RegTech can transform regulation and supervision.
Arner et al. (2016, p.375) highlighted how "RegTech provides the means to move towards a proportionate
risk-based approach where access to and management of data enables more granular and effective
supervision of markets and market participants.”
Yang et Li (2018, p.3258) highlighted how “Technology-driven regulation emphasizes oversight through
real-time, dynamic, information-sharing mechanisms involving all regulators and regulated institutions,
which will reduce regulatory costs, solve the problem of data validity, and build a real-time, predictive, top-
down, and transparent regulatory system based on technology".
Furthermore, supervision and regulation are expected to become more data-driven (Buckley et al., 2020),
interactive between regulators and regulated entities (Baxter, 2016), predictive (Arner et al., 2016),
proactive, transparent (Yang et Li, 2018) and responsive (Anagnostopoulos, 2018).

94
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Conclusion

The resulting framework from our analysis presents in an organized way the characteristics of RegTech as
emerging from literature, whether RegTech should be considered as part of Fintech or not, if RegTech refers
to companies or to solutions in general, and if RegTech consists in an improvement or in a disruption of
current capabilities and models. Regulations addressed and technologies involved are then two main factors
constituting RegTech.
Actors interested in RegTech are several and belonging to different sectors. Regulated entities and
regulators/authorities may find relevant application of RegTech, with implications on the whole system. The
role of data in all of this is crucial, being a structural component of how above-mentioned actors can leverage
RegTech to achieve the benefits, but risks related to RegTech must be considered as well. Finally, is clear
how exists a mutual relation between RegTech and regulatory models.
We briefly summarize here the main results.
What is RegTech?
RegTech is considered by contributions to be an innovation and a driver to innovate (FCA, 2017), both an
enabler (Baxter (2016) and a facilitator (Anagnostopoulos, 2018).
RegTech solutions are often considered a smart (Bain&Co, 2016), digital solution, automated (Quill et
Lennon, 2019), seamless (Kavassalis et al., 2018), non-invasive (FCA, 2016) and customizable (Capgemini,
2018), intuitive (FINRA, 2018), agile (Capgemini, 2018), sustainable and scalable (Quill et Lennon, 2019),
modular (EY, 2018a) and reliable (EY, 2018b), dynamic (Sangwan et al., 2019), quick and to enable real-
time operations (Michaels et Homer, 2018) among others.
There is no consensus in literature on whether RegTech should be considered a sub-set of Fintech or not.
Some authors (e.g. Anagnostopoulos (2018)) assess that RegTech is a sub-division of the fintech sector, but
others highlight how it is important to differentiate the two concepts, due for instance to different underlying
causes, to a different development path and to possible application also in non-financial sectors of RegTech
(Arner et al., 2016).
There are authors labelling with the keyword “RegTech” specific companies developing solutions employing
technology to address compliance (McKinsey (2017), but from our analysis emerges how “RegTech” refers
more widely to a solution (FCA, 2018a).
From the analysis of contributions, emerges how RegTech can be an improvement of existing capabilities
(Kavassalis et al., 2018) but as well have the potential to disrupt pre-existing structures, re-shaping current
regulatory processes and systems (FCA, 2018a).
Reg and Tech: the role regulation and technology
RegTech solutions address several regulations, both at national and international level, such as MiFID II or
UCITS (Financial Stability Board, 2017a). However, not all RegTech solutions address specifically
regulations, an example being certain RegTech solutions for internal reporting (Deloitte, 2017a).
RegTech solutions involve a wide set of technologies, such as Big Data, Artificial Intelligence (Financial
Stability Board, 2017a), API (FCA, 2016), cloud computing (Yang et Li, 2018) and so on.
Grounding RegTech: interested actors and applications
From literature and contributions, it is clear how RegTech is of interest not only for regulators and regulated
entities. Stakeholders are different and consist both of public (from regulators and policymakers to Central
banks and governments) and private entities (regulated entities, Tech companies, consulting firms, startups
until consumers). The main sector involved in contributions dealing with RegTech is the financial sector, but
some authors suggest how other sectors as well could benefit from RegTech solutions adoption (Arner et al.,
2016).
We identified five main functions in which RegTech can support firms, and in particular those subject to
regulation: compliance, monitoring, risk management, reporting and operations.
On the other way around, authorities can leverage on RegTech solutions in their functions as well, for
instance for supporting them in monitoring and controlling both macro and micro variables.
The role of Data
It is clear from analysis how Data play a central role in RegTech solutions, with various type or data
involved, different sources from which data can be collected and sophisticated ways to analyse them (Yang
et Li, 2018).
Benefits and Risks
From our analysis emerge several benefits coming from adoption of RegTech solutions, both for single
entities and for the whole system. Focusing on benefits for the single entity, we found benefits that are

95
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

mainly for regulated entities, mainly for regulators and benefits equally relevant for both, such as higher
efficiency and efficacy (FCA, 2016). Among other benefits, RegTech may as well increase accuracy
(Micheler et Whaley, 2019), transparency and manageability (Baxter, 2016) while driving down costs (see
for instance Quill et Lennon (2019) and times (Capgemini, 2018, p.3).
Literature provides several benefits of RegTech solutions, but risks are discussed and considered by several
authors: one of the main risks is cyber risk (Buckley et al., 2020), but RegTech solutions may pose risk
related to dehumanization or lack of skills (EBA, 2019).
In general, risks could increase if over reliance on RegTech solutions leads regulator and regulated entities
in a false sense of security (Micheler et Whaley, 2019) or if system focuses too much on short-term risks,
neglecting long-term ones (Buckley et al. (2020). RegTech could increase furthermore risks if certain market
participants apply it in a malicious way (Anti-RegTech as called by Packin (2017).
RegTech and regulatory models
RegTech is considered to fit and contribute to different possible regulatory approaches.
Goul (2019) highlighted the role of RegTech in three different approaches: insight-based approach,
principle-based approach and rules-based approach:
In a similar fashion, Micheler et Whaley (2019) analysed the role of RegTech in a different framework of
possible regulatory approaches: self-regulation, command-regulation, meta-regulation.
More in general, supervision and regulation are expected to become more data-driven (Buckley et al., 2020),
interactive between regulators and regulated entities (Baxter, 2016), predictive (Arner et al., 2016),
proactive, transparent (Yang et Li, 2018) and responsive (Anagnostopoulos, 2018).

From an academic point of view, relevance and value are present, being this work aiming at providing a
comprehensive view of the relatively new concept of “RegTech”, a relevant one due to the potential effects
that it may have on the financial system (Arner et al., 2016)
From the point of view of authorities and regulators, relevance is clear: as highlighted by Arner et al. (201, p.
371) “Regulating rapidly transforming financial systems requires increasing the use of and reliance on
RegTech”.
Giving this work a view on the opportunities and risk deriving from RegTech both for themselves and for the
regulated entities, authorities can better understand and employ those solutions, along with better regulate
their adoption by regulated entities.
In this direction, indeed, regulators are acquiring knowledge from the ecosystem about RegTech (see FCA
Model driven machine executable regulatory reporting TechSprint in 2017 or European Commission
consultation on a new digital finance strategy for Europe).
From a practitioner point of view, value is clear, giving this research a comprehensive view of the benefits
that RegTech solutions can provide, fostering their adoption, as well with a view on risk that they must
control.

Further research could focus on RegTech phenomenon will evolve in time, understanding if and how the
model presented in current research should be modified.
We suggest furthermore future research to focus on the understanding of best practices in the adoption of
RegTech solutions, in order to maximize the benefits and minimize the risks highlighted in this work.
Authorities should furthermore investigated more deeply whether RegTech should be considered or not part
of Fintech, in order to optimally regulate the phenomenon, increasing benefits and controlling risks.

96
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References

Accenture (2017a). RegTech: Keeping the customer first to gain a competitive advantage
Accenture (2017b). Investment Banks: Beware bespoke
Accenture (2018). REGTECH FOR REGULATORS. RE-ARCHITECT THE SYSTEM FOR
BETTER REGULATION
Alam, M. Z., Nasir, N., & Rehman, C. A. (2020). Intrapreneurship concepts for engineers: a
systematic review of the literature on its theoretical foundations and agenda for future research. Journal of
Innovation and Entrepreneurship, 9, 1-21.
Alrabiah, A. (2018). Optimal regulation of banking system’s advanced credit risk management by
unified computational representation of business processes across the entire banking system. Cogent
Economics & Finance, 6(1), 1486685.
Anagnostopoulos, I. (2018). Fintech and regtech: Impact on regulators and banks. Journal of
Economics and Business, 100, 7-25.
Arner, D. W., Barberis, J., & Buckey, R. P. (2016). FinTech, RegTech, and the reconceptualization
of financial regulation. Nw. J. Int'l L. & Bus., 37, 371.
Arner, D. W., Zetzsche, D. A., Buckley, R. P., & Barberis, J. N. (2019). The identity challenge in
finance: from analogue identity to digitized identification to digital KYC utilities. European business
organization law review, 20(1), 55-80.
Bain & Company (2016). Banking Regtechs to the Rescue?
Bank of England (2019). New economy, new finance, new Bank
Barberis, J., & Arner, D. W. (2016). FinTech in China: From shadow banking to P2P lending.
In Banking Beyond Banks and Money (pp. 69-96). Springer, Cham.
Baxter, L. G. (2016). Adaptive financial regulation and RegTech: a concept article on realistic
protection for victims of bank failures. Duke LJ, 66, 567.
BIS (2019): BIS Working Papers | No 811 | Embedded supervision: how to build regulation into
blockchain finance
Buckley, R. P., Arner, D. W., Zetzsche, D. A., & Weber, R. H. (2020). The road to RegTech: the
(astonishing) example of the European Union. Journal of Banking Regulation, 21(1), 26-36.
Capgemini (2018). Financial services firms are leveraging RegTech solutions to keep regulatory
compliance costs in check
CBInsight (2018). Regtech 102: The Evolution Of Regtech And The Future Of Regulatory
Compliance
Currie, W. L., Gozman, D. P., & Seddon, J. J. (2018). Dialectic tensions in the financial markets: a
longitudinal study of pre-and post-crisis regulatory technology. Journal of Information Technology, 33(4),
304-325.
Das, S. R., Kim, S., & Kothari, B. (2019). Zero-Revelation RegTech: Detecting Risk through
Linguistic Analysis of Corporate Emails and News. The Journal of Financial Data Science, 1(2), 8-34.
Deloitte (2016). ‘RegTech is the new FinTech’ How agile regulatory technology is helping firms
better understand and manage their risks
Deloitte (2017a). The RegTech universe on the rise
Deloitte (2017b). Managing conduct risk. Addressing drivers, restoring trust
Deloitte (2017c). The promise of RegTech
Deloitte (2018a). InFocus | Insurance regulation and technology: Adding business value to
compliance
Deloitte (2018b). Inside magazine - Edition 2018 | Using RegTech to transform compliance and risk
from support functions into business
EBA (2018). THE EBA’S FINTECH ROADMAP | CONCLUSIONS FROM THE
CONSULTATION ON THE EBA’S APPROACH TO FINANCIAL TECHNOLOGY (FINTECH)
EBA (2019). JOINT OPINION OF THE EUROPEAN SUPERVISORY AUTHORITIES ON THE
RISKS OF MONEY LAUNDERING AND TERRORIST FINANCING AFFECTING THE EUROPEAN
UNION’S FINANCIAL SECTOR
EIOPA (2016). Joint Committee Discussion Paper on the Use of Big Data by Financial Institutions
EIOPA (2020). SUPERVISORY TECHNOLOGY STRATEGY
Eling, M., & Lehmann, M. (2018). The impact of digitalization on the insurance value chain and the
insurability of risks. The Geneva papers on risk and insurance-issues and practice, 43(3), 359-396.

97
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Enriques, L. (2017). Financial supervisors and Regtech: Four roles and four challenges. Revue
Trimestrielle de Droit Financier, 53.
ESMA (2017). The Adoption of RegTech within the Financial Services Industry: Ten years from the
Start of the ‘Great Financial Crisis’ |RegTech: A Practical Guide to Harnessing the Benefits of the
Compliance Function, London England
ESMA (2019). ESMA Report on Trends, Risks and Vulnerabilities
European Commission (2019). Study on the costs of compliance for the financial sector
EY (2016). Innovating with RegTech | Turning regulatory compliance into a competitive advantage
EY (2018). Global M&A themes 2018 | Banking & Capital Markets
EY (2018). Reimagining the tax function | Find your path forward
FCA (2016). Call for input on supporting the development and adopters of RegTech
FCA (2017). European Commission consultation: “FinTech: A More Competitive and Innovative
European Financial Sector”: The UK Financial Conduct Authority’s response
FCA (2018a). Innovation Hubs | Enhanced Co-operation Agreement
FCA (2018b). Call for Input: Using technology to achieve smarter regulatory reporting
FED (2016). The Opportunities and Challenges of Fintech
FED (2019). Report to the Congress on the Use of the ACH System and Other Payment Mechanisms
for Remittance Transfers to Foreign Countries
Financial Stability Board (2017a). Artificial intelligence and machine learning in financial services |
Market developments and financial stability implications
Financial Stability Board (2017b). Financial Stability Implications from FinTech | Supervisory and
Regulatory Issues that Merit Authorities’ Attention
FINRA (2018). Technology Based Innovations for Regulatory Compliance (“RegTech”) in the
Securities Industry
Goul, M. (2019, July). Services Computing and Regtech. In 2019 IEEE World Congress on Services
(SERVICES) (Vol. 2642, pp. 219-223). IEEE.
Gurung, N., & Perlman, L. (2018). Use of Regtech by Central Banks and Its Impact on Financial
Inclusion. Available at SSRN 3285985.
IMF (2017a). Fintech and Financial Services: Initial Considerations
IMF (2017b). RECENT TRENDS IN CORRESPONDENT BANKING RELATIONSHIPS—
FURTHER CONSIDERATIONS
IMF (2018). FINTECH EXCHANGE | IMF/World Bank Spring Meetings 2018
IMF (2019A). THE BAHAMAS FINANCIAL SECTOR ASSESSMENT PROGRAM
TECHNICAL NOTE ON FINANCIAL INCLUSION, RETAIL PAYMENTS, AND SME FINANCE
IMF (2019B). Fintech in Latin America and the Caribbean: Stocktaking
IMF (2020). BIS Working Papers | No 859 | Post-crisis international financial regulatory reforms: a
primer
Kavassalis, P., Stieber, H., Breymann, W., Saxton, K., & Gross, F. J. (2018). An innovative RegTech
approach to financial risk monitoring and supervisory reporting. The Journal of Risk Finance.
KPMG (2017a). Reimagining the tax function Find your path forward
KPMG (2017b). Frontiers in Finance
McKinsey (2017). Digital Australia: Seizing the opportunity from the Fourth Industrial Revolution
Michaels, L., & Homer, M. (2018). Regulation and supervision in a digital and inclusive world. In
Handbook of Blockchain, Digital Finance, and Inclusion, Volume 1 (pp. 329-346). Academic Press.
Micheler, E., & Whaley, A. (2019). Regulatory technology: replacing law with computer
code. European Business Organization Law Review, 1-29.
NĂSTASE, M., CREŢU, A. Ş., & Stanef, R. (2009). Effects of global financial crisis. Revista de
Management Comparat International/Review of International Comparative Management, 10(4), 691-699.
Oliver Wyman (2018). REGTECH ON THE RISE | TRANSFORMING COMPLIANCE INTO
COMPETITIVE ADVANTAGE
O'Riain, S., Curry, E., & Harth, A. (2012). XBRL and open data for global financial ecosystems: A
linked data approach. International Journal of Accounting Information Systems, 13(2), 141-162.
Packin, N. G. (2018). RegTech, compliance and technology judgment rule. Chi.-Kent L. Rev., 93,
193.
Pittaway, L., Robertson, M., Munir, K., Denyer, D., & Neely, A. (2004). Networking and innovation:
a systematic review of the evidence. International journal of management reviews, 5(3‐4), 137-168.

98
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

PwC (2016). The changing landscape | How to use RegTech and make regulatory compliance your
strategic advantage
PwC (2017). Top financial services issues of 2018
PwC (2018). Regulatory brief | A publication of PwC’s financial services regulatory practice
Quill, T., & Lennon, R. (2019, April). Automating Legal Compliance Documentation for IoT
Devices on the Network. In 2019 IEEE 5th World Forum on Internet of Things (WF-IoT) (pp. 408-412).
IEEE.
Sangwan, V., Prakash, P., & Singh, S. (2019). Financial technology: a review of extant
literature. Studies in Economics and Finance.
Seppälä, S., Ceci, M., Huang, H., O'Brien, L., & Butler, T. (2017). SmaRT Visualisation of Legal
Rules for Compliance. In TERECOM@ JURIX (pp. 73-85).
Soloviev, V. (2018, October). Fintech ecosystem in Russia. In 2018 Eleventh International
Conference" Management of large-scale system development"(MLSD (pp. 1-5). IEEE.
Transatlantic institute (2017). The future of regtech for regulators
Yang, D., & Li, M. (2018). Evolutionary approaches and the construction of technology-driven
regulations. Emerging Markets Finance and Trade, 54(14), 3256-3271.
Zavolokina, L., Dolata, M., & Schwabe, G. (2016). FinTech–What's in a Name?.

Sitography

[Link]
[Link]
[Link]

99
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

«EMPIRICAL MODELING OF INTERNATIONAL BANKS’ CREDIT RISK: ASSESSMENT AND

COMPARISON OF CREDIT RATINGS»

Alexander M. Karminsky1, Ella Khromova2, Roman Kudrov3

The School of Finance, National Research University Higher School of Economics,
Research group “Innovations in banking sector, its financial stability and prudential regulation”
11 Pokrovsky Bulvar, Pokrovka Complex, Moscow 109028, Russia
Abstract
The work is devoted to credit risk modeling of international banks by constructing
ordered logistic models of credit ratings assigned by the agencies: Moody's, Standard & Poor's
and Fitch Ratings and mapped into the base scale. Using a random sample of 478 banks from
more than 40 countries for the period of 2007-2019, we obtain a credit rating model that can be
effectively used to assess credit risk using public information, which is demonstrated with a
help of out-of-sample forecasts of the obtained regression. The resulting model contains
financial indicators of the bank's sustainability, macroeconomic indices and indicators of the
state governance level. The quality of the model prediction is significantly improved by
including interaction terms and applying the Principal Component Analysis (the share of
correct forecasts increases by 9% for the class model and by 6% for the grade model).
Additionally, using marginal effects calculation, the empirical evidence of the
importance of maintaining a high level of liquidity by banks in times of crisis is found. At the
same time the highest marginal effects of the whole sample time horizon are reached in factors
of asset quality and bank’s size among all financial indicators included in the model.
To calibrate a credit rating into a quantitative measure of credit risk, we develop a
dynamic transmission scale for converting credit ratings into probability of default (PD) by
evaluating the average historic default frequencies of Russian banks. The analysis brings us to
an investment advice: to achieve a higher return from an investment in banks with highly
speculative rating grades, it is optimal to choose a long run investment 1 or 2 years after the
rating assignment.

© 2020 A.M. Karminsky, E. Khromova., R. Kudrov

JEL: G21, G33

Keywords: Banks; credit ratings; probability of default; logit and probit models; rating agencies

1
e-mail: karminsky@[Link] tel: +79037254937
2
e-mail: epkhromova@[Link] tel: +79168039208
3
e-mail: kudrovroman@[Link] tel: +79169904898

100
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Introduction
Sustainability of a country’s financial system primarily depends on the performance of
financial institutions. The key financial institutions are banks and credit organizations. The
assessment of banks’ credit risk is an important issue for the government, regulators and
investors. All the economic agents are interested in well-functioning of banks, which serve as the
main financial intermediaries on the market. Banking organizations should be evaluated not only
in terms of individual characteristics (different financial ratios) but also in terms of overall
environment on the market (characteristics of the region where an enterprise operates). The
possible influence of regulatory framework and administrative barriers should be taken into
account. This creates practical difficulties for ordinary economic agents, that is why the impact
and extension of the credit ratings issued by credit rating agencies (CRAs) is growing each year.
Credit ratings are the opinions of CRAs on the financial performance and stability of a bank.
They are of very important practical significance both for investors and banks themselves.
The activities of CRAs are finite, hence not all the banks are assigned a credit rating.
Credit rating modeling is wildly used in the academic literature, but there are several limitations
in the practical implementation of these models. Firstly, the credit rating models are usually
constructed for a sample of banks from one country. This does not allow to use such models for
the banks from other regions and challenge the interpretation of significant factors of banks’
financial sustainability. Secondly, there are many CRAs operating around the world, their
assessment of credit risk may not match each other. Due to difference in methodologies and
possible subjectivity, we may not rely on the forecasted credit rating of a single CRA. This
prevents the practical use of models obtained in the previous papers. Thirdly, a forecasted
symbolic rating may be misunderstood and wrongly interpreted by an investor or bank’s internal
management. This can lead to incorrect investment decisions and possible losses. To solve these
problems, in this paper, we construct a credit rating model for assessing bank’s credit risk, namely
an ordered logit model with a panel data structure. This type of model is chosen because it allows
to evaluate the impact of each indicator included in the model and to interpret its influence on
the credit risk.
Despite the existence of many academic papers on the subject of credit risk assessment
of financial institutions, the problems of financial stability analysis in the banking sector do not
lose their relevance. The topic of our paper may be of particular importance in the current
situation, which is critically close to the global economic crisis. Nowadays, only a small number
of banks has a credit rating, hence our research is useful for the government, banks and investors.
The aim of this paper is to develop the technique for credit risk assessment using a
uniform credit rating model. The uniform model simultaneously includes the credit ratings of

101
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

several CRAs and allows to forecast a rating score on the base scale. Several econometric
techniques are used to increase the accuracy power of the obtained models. The forecasted score
can be used to evaluate the credit risk of a bank using a dynamic transmission scale which relates
a rating score to average default frequencies of Russian banks.
In order to achieve the aim of the paper the following tasks are set and successfully
completed:
1. To analyze existing literature on credit risk assessment of financial enterprises and to
determine the most efficient econometric technique which can be used in practice;
2. To summarize the potential indicators on bank’s credit risk on the basis of academic articles.
This is helpful for both our analysis and further research on the topic;
3. To create a random and representative sample of international banks and to prepare the set of
explanatory variables;
4. To construct the credit rating model for different CRAs and to find out the possible
noncoincidences in their methodologies of rating assignment;
5. To construct uniform credit rating models (both class and grade models) with credit scores
mapped into the base scale;
6. To calculate the marginal effects of each indicator included in the final model specification.
This helps us to find out the most significant and important factors of influence on credit risk;
7. To evaluate the in-sample accuracy of the obtained credit rating models. The distribution of
forecast errors is provided in order to test the normality of the errors;
8. To check the robustness of the final model specification. The results show the possibility of
practical usage of the obtained models;
9. To create the dynamic transmission scale of credit score and probability of default (PD) using
the average historic default frequencies. This makes the credit risk valuation more accurate
and provides investors with a quantitative measure of credit risk.
The object of our research is the international banks. The subject is the financial
sustainability of banks taking into account the macroeconomic situation and level of state
governance in the country of residence.
The credit rating models are built for 478 international banks from 46 countries. The
financial informational is extracted from S&P CAPITAL IQ platform. Macroeconomic indexes
and indicators of state governance are provided by the World Bank and [Link]. The
unbalanced panel is constructed for 49 time periods: from the first quarter of 2007 to the first
quarter of 2019.
The paper is structed as follows. In chapter 1 we introduce the literature review on credit
risk modeling. An extensive number of international and Russian articles is analyzed by us in

102
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

order to cover all the aspects of banks’ credit risk assessment. Afterwards, the main hypotheses
are formulated. In chapter 3 you may find the detailed data description and the methodological
issues. The main econometric results are provided in the fourth chapter, where the results of
regression estimation of credit risk and interpretation of factors of influence are represented.
Chapter 5 contains the calibration of credit rating scores on PD on the example of Russian banks.
The paper ends up with conclusion.

103
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chapter 1. Review of related academic literature

1.1. Credit risk assessment methods

The econometric techniques used for credit risk assessment belong to one of the following
groups: company’s bankruptcy prediction and credit rating analysis. The fist paper which
addressed the probability of company’s default estimation was Altman (1968), where the linear
discriminant analysis was applied, and Z-score was introduced. The analysis divided the firms
into stable and bankrupt on the basis of balance sheet information in the previous periods. Later,
Altman [Link] (1977) developed the ZETA model which had a stronger accuracy power than the
previous bankruptcy model. Although these models have been introduced more than 50 years
ago, they still have practical implementation, which is demonstrated in recent researches (Elliott
et al., 2014; Almanny et al., 2016; Ko et al., 2017).
The linear discriminant analysis recommended by Altman did not allow to estimate the
probability of failure, it only helped to assess the financial position of a company (good/bad).
Martin (1977) firstly used the logit model to forecast the American banks failure. This method
helped to identify and correctly interpret the coefficients of the model. Nowadays, in the majority
of articles, the binary choice models (logit/probit) models are used (Godlewski, 2007; Lin, 2009;
Karminsky & Khromova, 2018; Westgaard & Wijst,2001). Lin (2009) demonstrated that logit
models were preferred to the standard discriminant analysis. Artificial Intelligence techniques
are also wildly used to forecast the PD of companies (Lin, 2009; Kvamme et al., 2018).
PD forecasting is not the only method which is used as a tool for company’s bankruptcy
prediction. The expected loss (EL) is determined using the fundamental risk parameters: PD,
LGD (loss given default) and EAD (exposure at default) according to the formula:
EL  PD  LGD  EAD . (1)
LGD represents the share of credit which is not recovered in the case of default. This
indicator is calculated by internal management of banks in order to evaluate the losses the bank
would fail in case of client’s default. EAD is added to equation (1) to indicate the expected
amount of exposure in case of default. Modeling techniques for evaluation of these credit risk
parameters are presented in the papers (Spuchľakova & Cug, 2015; Krüger & Rösch, 2017; Leow
& Mues, 2012; Tanoue et al., 2017).
Altman & Rijken (2004) demonstrated that the models of bankruptcy prediction worked
accurately in the short run period (1 year), but for long run (around 6 years) credit risk estimation
credit rating models were preferred. Later, the models of credit ratings were applied in the works
(Hwang et al., 2010; Bellotti et al., 2011; Le & Viviani, 2018). Onyiriuba (2016) defines credit
rating as an opinion of external expert about the probability that an issuer will not default on
investment’s obligation. The author emphasized the role of credit ratings in developed countries.

104
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The credit rating may help an investor to assess the risk of default and to make a correct
investment decision. Individuals prefer to rely on judgments of independent parties, that is why
the role of credit rating agencies (CRAs) is important in the financial systems around the world.
The experience of credit risk modeling shows that for correct prediction of company’s
sustainability it is not enough to account for financial indicators only. Banks’ credit risk depends
on the regulations and norms implemented in a country. Ashraf et al. (2020) showed that strong
and open capital regulation reduced the probability of bank failure. The analyzed sample
contained banks from 111 countries. Different time periods were analyzed, including times of
financial crises. Strict capital regulation in the pre-crisis period helps the banks to survive and to
fully meet their obligations during the crisis period. The results of econometric modeling showed
the importance of Basel III capital accord implementation. Bermpei et al (2018) demonstrated
the influence of institutional quality on the sustainability of banks and the credit risk.
Good credit risk management helps banks to be stable to various financial and business
crises. Basel II introduced the internal ratings based (IRB) method as a tool for credit risk
measurement by the banks themselves. The importance of IRB was demonstrated by Cucinelli et
al. (2018). The authors analyzed the risk management of European banks that used IRB method
and those that did not. It was shown that IRB models helped banks to adopt accurate lending
practices and to evaluate the creditworthiness of their clients correctly.

1.2. Various approaches to credit rating analysis and modeling

Credit rating agencies (CRAs) are sometimes criticized for their subjectivity and incorrect
approach to assessing the financial sustainability. There are a lot of studies investigating the
possible inappropriate assessment of company’s financial performance by the credit rating
agency. Kashyap & Kovrijnykh (2016) analyzed the behavior of the CRAs depending on the
buyer of their products. It was shown that the errors in ratings usually occurred when the analyzed
firm payed for it. The model was built to find out how and in what circumstances a credit rating
might be helpful for investors. An investor was proven to be an optimal customer and
compensator of CRA’s rating. Only under this compensation scheme, the credit rating could
really help an investor to make a choice of a firm to finance.
Ponce (2012) proposed the empirical research which showed that the quality of agencies’
work was below the optimal quality standards under issuer-pays model. The author provided
several reasons why the return to investor-pays model looked impossible. Firstly, an investor
could sell the received information to others which would be not socially optimal due to free-
rider problem. Secondly, several difficulties with taxation of such relations might occur. It seems
that these constraints remain relevant in the modern time.

105
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Park & Lee (2018) investigated the influence of competition on the quality of credit
ratings on the example of Korean bonds. The authors showed that increased competition
(measured by the number of agencies hired) increased the probability of credit rating being
upgraded. The results depicted that the newly hired rating agency gave a higher rating than those
already existed. However, the favor was given by the bond issuers to the CRAs assigning lower
credit ratings than its competitors.
Hu et al. (2019) examined the effects of increased competition between CRAs on the
example of Chinese bonds market. It was demonstrated that a combination of investor-paid
model and public utility model might be more socially efficient that traditional schemes. A
combined model had also several advantages for regulators. The entry of an agency with new
model of providing its services may cause other CRAs to review their work.
Karminsky & Khromova (2016) built econometric models to predict the rating of three
international agencies. The authors obtained that CRAs assigned different credit ratings to the
firms with the same financial performance. This fact indicated the difference in methodologies
and possible subjective evaluations of these CRAs. Some evidence for conservatism of agencies
was found. Standard & Poors issued the most conservative ratings.
Despite all that, the investors and internal management still use credit rating as a signal
of possible failure. The recent paper by Driss et al.(2019) depicts the importance and relevance
of ratings. The authors found out that firms with watchpreceded rating confirmations had better
long-term financing and this helped them with their investment activities. Companies with
confirmed ratings had better indicators of their operational activities and profitability than their
competitors after the credit watch period. It served as a proof to the idea that credit agencies were
respected by the market players and could simplify the firm’s course to credit market.
Ali & Durdu (2012) provided the list of spheres of economy which benefit from CRAs
existence. The main market players noticed the opinions of CRAs about the financial behavior
of an enterprise although some subjectivity in the behavior of CRAs was depicted. The authors
recommended some political measures to support the activities of rating agencies and to create
favorable conditions to increase the competition between them.
Gu et al. (2018) conducted an econometric analysis of the role of credit ratings in
supplying the public debt. Credit ratings acted as an information channel between a firm and its
shareholders. The services of CRAs provided the shareholders with efficient monitoring of a
company. This information channel played an important role in firms’ operational activities, it
helped them to increase the capital expenditures only when creditors had strong rights. From the
terms of policy implications, these ideas might encourage firms to strengthen creditor rights for
them to be able to diversify their sources of credit.

106
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The usage of credit ratings is seriously limited as not all the economic agents are assigned
a credit rating. It also takes time for the rating agency to update the rating (Peresetsky, 2012).
The services provided by CRAs are usually expensive and imply significant costs for an
enterprise. All these circumstances forced the researches to build the econometric models for
credit rating prediction.
In the paper by Bellotti et al. (2011) the comparison between ordered choice models and
support vector machines (SVM) in credit rating forecasting was provided. The authors used the
credit ratings of Fitch Ratings issued to 681 international banks from 2000 to 2007. SVMs
produced more accurate in-sample predictions than ordered choice models. It was also
demonstrated that a credit rating served as an indicator of banks’ financial sustainability and
could be used as a measure of credit risk.
Hwang et al. (2010) complemented the standard ordered probit model with a
semiparametric function. The methodology was tested on the sample of 779 companies with S&P
long-term credit ratings. More than 20 potential explanatory indicators were chosen according to
the previous articles. The outcomes depicted the effectiveness and usefulness of such approach
in credit ratings modeling.
Karminsky et al. (2019) examined the behavior of different credit rating models for
industrial companies from BRICS countries. The statistical models included linear discriminant
analysis (LDA), ordered logit regression (OLR). They were compared to SVM, ANN and random
forest. Machine learning techniques were preferred by the authors due to higher predictive power
in both in-sample and out-of-sample forecasts.
Some papers present the way of measuring the credit risk with a help of the probability
of default (PD) models. Lee & Choi (2013) built bankruptcy models of Korean companies using
back-propagation neural network (BNN) and Multivariate Discriminant Analysis (MDA). BNN
method performed better than MDA in bankruptcy prediction in all the built models for the
analyzed industries. It was shown that the quality of the models increased when controlling for
industry-specific characteristics.
Le & Viviani (2018) compared the traditional models with modern machine learning
techniques. They used the most popular statistical models (Discriminant analysis and Logistic
regression) and three machine learning techniques (Artificial neural network [ANN], Support
Vector Machines and k-nearest neighbors [k-NN]). A detailed of all these models was provided
in the paper. On the basis of econometric results, the authors showed that the difference in
accuracy prediction of banks’ failures between Logistic regression, ANN and k-NN is not very
big. SVM did not perform better than statistical models.

107
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Karminsky & Khromova (2018) aggregated the previous experience in credit ratings and
PD modeling in their work on the sample of Russian banks from 2007 to 2017 year. The authors
used ordered probit and logit models. Their econometric results demonstrated that credit rating
models overestimated the financial instability of banks, while the PD models underestimated it.
After monotonic transformations of these two models, they were combined into a single
logarithmic model with high predictive power (44% of correct predictions).

1.3. Potential indicators of banks’ credit risk

There is a variety of factors that can be used in the models of credit risk. The researches
try to explain the variations of some particular factor with a help of different indicators. We have
examined a considerable set of previous academic articles in credit risk modeling and in the Table
1 we present the results of our analysis. It was noticed by us that the majority of the authors tend
to use very similar factors of influence on credit risk, but the indicators may differ. We have
decided to group these factors according to three main groups: Macro Profile, Financial Profile
and Qualitative Judgments. These groups represent the Baseline Credit Analysis (BCA) used by
Moody’s agency in credit rating determination. BCA is generated by the CRA in order to evaluate
the strength of an organization without any external support. The methodology of Moody’s is
depicted in the Appendix 2 (see Figure 21).
Credit rating modeling is primarily aimed at BCA, as this component of the overall
analysis measures the credit risk of a bank itself without any support. The evaluation of BCA
factors helps to measure the bank’s probability of default, which is of primary interest for both
bank’s internal management and outside experts. Basing on BCA, Moody’s assigns a preliminary
rating to a bank and then adjusts it assessing the results of Support& Structural analysis. Macro
profile helps to measure the conditions in which a bank operates and makes an assessment of its
economic environment. Financial profile determines the financial sustainability of a bank. This
profile mostly consists of different financial ratios. Qualitative judgements profile includes some
non-financial judgments which can be helpful in measuring the overall risk profile.
The factors and potential indicators used in the previous academic articles are grouped
by us to the groups according to BCA system: Financial profile (F1, F2, …. , F10), Macro profile
(M1, …, M4) and Qualitative judgments (Q1,…,Q4) in

Table 1 below.

108
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1: Possible indicators of financial organizations’ credit risk (cont’d)

Indicators Sources
F1: Capital Adequacy
 (Book Equity capital - Bad  Customer Deposits / Shareholders’ Thomson (1991); Karminsky
Loans)/Total Assets Equity & Peresetsky (2007); Pagratis
 Shareholders’ Equity / Total  Equity / (Total or Net) Loans & Stringa(2007); Pasiouras,
Assets  (Equity + Loan Loss Reserve) / Gaganis & Doumpos (2007);
 Tier 1 and Tier 2 Capital / Risk Loans Männasoo & Mayes (2009);
Weighted Assets  Non-performing Assets / Total Peresetsky (2009); Sahut &
 Shareholders Equity / Total Assets Mili (2011); Betz et al. (2014);
Liabilities  Total Capital Ratio (Tier 1,2,3 to Alrabadi & Hamarneh (2016);
 Tier 1 capital / Risk Weighted Risk Weighted Assets) Le & Viviani (2018);
Assets  Total Equity / Total Liabilities Karminsky & Khromova
 Equity / Customer and Short-term  Capital funds / Deposit & Short- (2016); Karminsky &
funding term funding Khromova (2018)
 Capital funds / Total Assets  Total Equity/ Total Assets
 Total Equity/Total Debt  Capital funds / Liabilities
F2: Asset quality
 Net charge offs / Total loans  Impaired loans / Gross loans Thomson (1991); Karminsky &
 Loan portfolio Herfindahl index  Yield on earning assets Peresetsky (2007); Pagratis &
 Net Loans and Leases / Total  Long-term assets / Total assets Stringa(2007); Peresetsky
Assets  Reserves for Loan Losses / Non (2009); Sahut & Mili (2011) ;
 Loan loss reserves / Gross loans Performing Assets Alrabadi & Hamarneh (2016);
 Loan loss reserves / Net Interest  Provisions for Loan Losses / Total Le & Viviani (2018); Gogas et
Income Average Loans al. (2014); Betz et al. (2014);
 Loan loss provisions/Net Interest  Net charge off / Average Gross Karminsky & Khromova
Revenue Loans (2016)
 Impaired Loans / Total Equity  Loans growth
F3: Management quality
 Overheads (+provisions) / Total  Operating Costs / Operating Thomson (1991); Sahut & Mili
Assets Income (2011); Betz et al. (2014); Jiao
 EBIT/Total Revenue  Average sale growth rate during et al. (2007); Karminsky &
 EBIT/Total Assets the last years Petrov (2004); Le & Viviani
 Loans to Insiders / Total Assets  Total Expenses / Total Revenue (2018); Petropoulos et al.
 Overheads /Income before  Dummy for private ownership (2020)
provisions  Total Assets turnover
 DEA Score Efficiency  Net Operating Income / Assets
F4: Earnings/Profitability

109
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

 Retained Earnings / Total Assets  Net interest Income / Earning Altman (1968); Thomson
 EBIT / Total assets Assets (1991); Altman & Rijken
 Return on Assets (ROA)  Dividends paid / Net Income (2004); Karminsky &
 Interest Expense / Average  Net Income / Total Assets Peresetsky (2007); Pagratis &
interest-bearing liabilities  Net Income / Shareholders' equity Stringa(2007);Pasiouras,
 Net interest margin  Return on Average Equity Gagasis & Doumpos (2007);
 Overheads / (Net interest income  (Interest income + other operating Peresetsky (2009); Sahut &
+ other operating income) income) / Total assets Mili (2011);Gogas et al.
 Intereste Expense / Interest  Interest expense /Total liabilities (2014); Le & Viviani (2018);
income  Net Interest Income/Total Revenue Ashbaugh-Skaife et al. (2006);
 Return on Average Assets  Dividend Payout ratio Betz et al. (2014); Alrabadi &
 Securities gains (losses) as % of  Recurring Earning power Hamarneh (2016); Jiao et al.
total interest income  Return on Net Worth before tax (2007); Karminsky &
 Inc. Net of Dist / Average Equity  Interest Expense / Net Sales Khromova (2018)
Table 1: Possible indicators of financial organizations’ credit risk (cont’d)
Indicators Sources
F5: Liquidity
 Working Capital/Total Assets  (Total or Net) Loans / Total Assets Altman (1968); Thomson
 Nondeposit Liabilities / Cash and  Liquid assets /Customer deposits (1991); Altman & Rijken
Investment Securities and short-term funds (2004); Karminsky &
 Net working capital/ Total Assets  Net loans / Customer deposits and Peresetsky (2007); Pagratis &
 Customer accounts / short-term funds Stringa(2007); Pasiouras,
Shareholders’ equity  Liquid assets / Total deposits and Gaganis & Doumpos (2007);
 Net loans / Total deposits and borrowings Männasoo & Mayes (2009);
borrowings  (Deposits and short-term funding - Sahut & Mili (2011); Betz et al.
 Bank deposits / Customer Customer deposits) / Liquid assets (2014); Alrabadi & Hamarneh
deposits  Interest expenses / Total liabilities (2016); Le &Viviani (2018);
 Total deposits / Total Assets  Total Deposits / Funding Jiao et al.(2007)
 Short-term borrowing-Cash /  Current Assets / Current Liabilities
Total Liabilities  Net Loans / Total Assets
 Current Assets-  Liquid Assets/Deposit & Short
Inventories/Current Liabilities term Funding

F6: Sensitivity to market risk

 Trading Income / Operating  Asset Fair Value Betz et al. (2014); Petropoulos
Income et al. (2020)
F7: Solvency/Leverage
 Market value equity/Book value  Cash flow / Total debt Altman (1968); Altman &
of Total Debt  Total Debt/ Total Assets Rijken (2004); Sahut & Mili
 Market value of interest/ Book  Market Value of Equity/Book (2011); Ashbaugh-Skaife et al.
value of total liabilities value of Total Liabilities (2006); Attig et al. (2013)
F8: Efficiency

110
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

 Sales / Total assets  Other Operation income / Average Altman (1968); Peresetsky
 Personnel expenses / Operation assets (2009); Le & Viviani (2018);
income  Non-interest expense / Average Ashbaugh-Skaife et al. (2006);
 Net interest margin assets Jiao et al.(2007); Karminsky &
 Net interest revenue / Average  Pre-tax Operating Income / Khromova (2018)
assets Average Assets
 Operating income before  Non-Operating Items & taxes /
dep'n/Interest Expense Average Assets
 Net Credit Sales/Average  Cost of goods sold/Average
Account Receivables inventory
 Operating Expenses/Total  Net Sales/Average Total Assets
Revenues
F9: Size
 Logarithm of Total Assets  Logarithm of Total Equity Thomson (1991); Altman &
 Total Liabilities normalized by Rijken (2004); Schaeck &
the total value of the equity Čihák (2007); Karminsky &
market Peresetsky (2007); Peresetsky
(2009); Sahut & Mili (2011) ;
Gogas et al. (2014); Alrabadi &
Hamarneh(2016); Ashbaugh-
Skaife et al. (2006); Karminsky
& Khromova (2016);
Karminsky & Khromova
(2018) ; Attig et al. (2013)
Table 1: Possible indicators of financial organizations’ credit risk (cont’d)
Indicators Sources
F10: Age
 Number of years since a firm was Alrabadi & Hamarneh (2016);
first rated by an agency Altman & Rijken (2004)
 Logarithm of Years of bank age
M1: Macroeconomic indexes
 Output Herfindahl index  Real interest rate Thomson (1991); Schaeck &
 Unemployment rate  GDP per capita Čihák (2007); Karminsky &
 Change in state-level  (Export + Import) / GDP Peresetsky (2007); Männasoo
personal income  Savings / GDP & Mayes (2009); Peresetsky
 Dun and Bradstreet's  Country ceiling (2009); Sahut & Mili (2011);
state-level failure rate  Private lending / GDP Betz et al. (2014); Karminsky
 GDP growth  Exchange rate & Khromova (2016)
 Inflation rate (CPI)  Total Banking Loans / GDP
 EURIBOR 3 months  Growth rate of the HICP index
 Corruption Perceptions Index  General government debt (%GDP)
(CPI)  Trade balance
 10-year government bond yield  Growth rate of the stock price
 Net International Investment index
Position (%GDP)  Growth rate of the house price
 Private sector credit flow (% index
GDP)
 GDP per capita
M2: Market structure and Market discipline

111
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

 Money lent to other banks /  Foreign owned banks assets/ Total Schaeck & Čihák (2007);
Money borrowed from other banking sector assets Pagratis & Stringa(2007);
banks  EBRD banking sector reform index Pasiouras, Gaganis &
 Market share held by the  Stock market index Doumpos (2007); Männasoo &
institution  State-owned banks assets/ Total Mayes (2009); Betz et al.
 Log(Number of banks / banking sector assets (2014); Karminsky &
Population of a country) Khromova (2016); Karminsky
 Herfindahl-Hirschman index & Khromova (2018)
 H-Statistic
 The Heritage Banking and
Finance score
M3: Banking competition
 The growth rate of (domestic  H-statistics Schaeck & Čihák (2007)
credit to the private sector/ GDP)
M4: Health of banking system
 Non- performing Loans / Total Schaeck & Čihák (2007)
loans
Q1: Corporate governance
 Number of institutional  % of shares held by insiders Pasiouras, Gaganis &
shareholders (officers and directors) Doumpos (2007); Ashbaugh-
 Whether the bank is listed or not  % of independent directors on the Skaife et al. (2006); Attig et al.
in a stock exchange board (2013)
 Number of holders with 5%  % of independent directors that
ownership level hold seats on other firms’ boards

Table 1: Possible indicators of financial organizations’ credit risk (cont’d)

Indicators Sources
Q1: Corporate governance (cont’d)
 % of shares held by  % of directors that own stock in the Pasiouras, Gaganis &
institutional investors firm Doumpos (2007); Ashbaugh-
 Dummy for a formal  % of insiders on the finance Skaife et al. (2006); Attig et al.
governance policy committee (2013)

Q2: Earnings diversification

 The number of subsidiaries Pasiouras, Gaganis &
 Logarithm of Average Deposits Doumpos (2007); Thomson
per Banking Office (1991)
Q3: Financial Transparency
 Total fees paid by the firm to its  Dummy for independent financial Ashbaugh-Skaife et al. (2006)
auditor/Total revenues of the expert on the audit
audit firm committee
 % of audit committee made up of
independent directors

Q4: Geographical deversification

 Share of investment portfolio Karminsky & Khromova
across geographic regions (2016)
 Operation of a bank in foreign
countries (# countries)

112
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Source: Author’s calculations

The presented indicators were used in the literature to evaluate the credit risk. The
majority of the papers used credit ratings and probability of default models, but it was decided to
include some works with other econometric methods for credit risk valuation in order to extend
the list of indicators. Financial indicators are crucial in determination of the financial
performance of financial organizations. The majority of the factors from the Financial profile
constitute the components of the CAMEL(S) methodology.
Capital Adequacy helps to evaluate the value of banks’ deposits and whether this value
is enough for depositor’s protection in the failure times. The bank needs capital in order to fulfill
its obligations in the event of both expected and unexpected losses. Betz et al. (2014) used Tier
1 Capital Ratio and Total Capital Ratio (Equity/Assets) as proxies for bank capitalization. The
estimated model showed the significance of Total Capital Ratio, but insignificance of Tier 1
Capital Ratio. This could happen as the author inserted both these indicators in the model, which
could lead to multicollinearity and weakened the significance of one of the indicators. Pagratis
et al. (2007) examined the influence of different indicators on the Moody’s credit ratings. Tier 1
Capital Ratio appeared to be significant only in the cross-term with the sub-investment rating
grades. The authors concluded that this reflected the methodology of Moody’s and its view on
capital ratios.
Asset quality factor measures the capacity of bank’s assets. They include both bank’s
accumulated finances and various accounts receivables. The operational activity of banks implies
receivables of many times including loans to customers and other enterprises. Sahut & Mili
(2011) estimated logit models for banks distress and found out the significance of Loan Loss
Reserves/Gross Loans and Loans growth as proxies of asset quality. Their effect was positive on
the likelihood of banks’ distress. Le & Viviani (2018) analyzed the influence of 5 proxies of
asset quality on banks failure: Loan Loss reserve/Gross Loans, Loan Loss provision/Net interest
revenue, Impaired Loans/Gross Loans, Net charge off/Average Gross Loans, Impaired
Loans/Equity. The ratio Impaired Loans/Gross Loans played the most significant role in the
group.
Management quality estimates the compliance with law and various instructions by the
bank’s administration. This factor is used for credit risk assessment in the academic papers (Sahut
& Mili, 2011; Betz et al., 2014; Jiao et al.,2007; Petropoulos et al, 2020).
Profitability factor show the value of banks’ earnings over a certain period of time. Banks
usually earn from their operational activities (credit and deposit operations, provision of
guarantees, etc.) and non-operational activities (dividends received, fines, income from property
leasing, etc.) Betz et al. (2014) demonstrated the negative association with bank’s profitability

113
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and possible distress. Return on Equity (ROE) and net interest margin were used as proxies. Both
variables appeared to be significant in the econometric models.
Liquidity is analyzed in order to assess the banks’ possibilities in meeting its financial
obligations full and in time. Sometimes it may show whether a bank can quickly and without
losses transfer its assets into cash, which might be essential in difficult times. Alrabadi &
Hamarneh (2016) studied the following liquidity variables: Loans/ Total Assets, Loans/
Customer and short-term funding, Loans/ Total deposits and borrowings, Liquid assets/
Customer and short-term funding. All these variables were included in ordered probit models
and appeared to be highly significant.
Sensitivity to market risk factors evaluated the impact of major market risks on the
activities of banks. This factor can be hardly measured, but sometimes the authors try to see the
effect of market conditions on banks’ performance. The effect of this component on banks’ credit
risk can be found in the papers (Betz et al., 2014; Petropoulos et al.,2020).
The factors from the CAMEL(S) system do not cover all the aspects of companies’
financial sustainability. Le & Viviani (2018) considered 6 indicators of efficiency and concluded
that this group of variables was one of the most important in predicting the banks failure. The
proxies for efficiency (Other Operation Income/Average Assets, Net interest revenue/Average
Assets, Non-Operation Items&taxes/Average Assets) appeared to be significant in the obtained
models. The indicators for banks’ efficiency can be also met in the articles (Altman,1968;
Peresetsky, 2009; Le & Viviani,2018; Ashbaugh-Skaife et al.,2006;
Jiao et al.,2007; Karminsky & Khromova,2018). The level of debt is sometimes measured by
the solvency ratios. Extremely high levels of debt may harm banks in the times of failure. Altman
& Rijken (2004) used the variable Market Value of Equity/Book value of Total Liabilities as a
proxy for market leverage. The index appeared to be highly significant and increased the quality
of default-prediction model.
Age and size of banks also influence on the credit risk. Old and large banks are financially
more stable, and probability of their failure is minor (Ayuso et al., 2004). People trust these banks
and CRAs assign higher credit ratings to them. It was proven in several academic papers
(Alrabadi & Hamarneh, 2016; Altman & Rijken,2004; Sahut & Mili, 2011; Gogas et al.,2014).
Credit risk is related not only to financial performance of banks, but also to market
conditions. Pasiouras et al. (2007) in his work tried to develop a model and replicate the credit
ratings of Fitch Ratings agency on Asian banks. The authors considered various financial and
non-financial variables for the development of the model. Firstly, 19 financial variables were
analyzed. As a result, only 5 factors explained 83% of the total variance in a sample and these
variables (Equity/Customer and Short-Term Funding, Net interest margin, Return on Average

114
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Equity, Liquid Assets/ Total Deposits and Borrowings, Net loans/ Total Deposits and
Borrowings). The first indicator was used as a measure of capital adequacy (or capital strength);
the second and third indicators depicted the profitability of a bank; the fourth and the fifth
indicators were used to evaluate liquidity. However, it was mentioned that it was not enough to
use financial variables only. So, non-financial factors (the auditor’s opinion on banks’ financial
statements, the number of subsidiaries, the number of institutional shareholders, the Heritage
Banking and Finance score, and dummy for whether the bank is listed or not in a stock exchange)
were also used in developing the model, which helped to increase the model’s forecasting power.
The positive effect of macroeconomic variables on the models’ predictive power was also
demonstrated in papers (Peresetsky, 2009; Sahut & Mili,2011; Betz et al.,2014; Karminsky &
Khromova,2016).
The effects of business environment and market conditions were presented by Schaeck
& Čihák (2007). The authors empirically proved that banks operating in competitive financial
system held higher capital ratios. The models controlled for many factors, including
concentration, inter-industry competition, financial system characteristics. Other indicators for
market structure can be found in the academic articles (Betz et al.,2014; Karminsky & Khromova,
2016; Karminsky & Khromova,2018).
Among the qualitative judgments the most important one is corporate governance. Good
corporate governance may be a signal for a rating agency to assign a higher credit rating. The
majority of indicators measuring corporate governance were tested by Ashbaugh-Skaife et al.
(2006). It was proven that firms with high level of corporate governance received higher credit
ratings that their competitors. The author found out the following significant indicators in credit
rating modeling: the number of holders with at least 5% ownership share in the firm (-), weaker
shareholder rights in terms of takeover defenses (+), degree of financial transparency (+),
company’s board independence (+) and CEO power on the board (-). The level of quality of the
corporate governance was evaluated from the viewpoint of the shareholders and it was concluded
that if they considered the level of corporate governance.

1.4. The interdependence of credit ratings and other risk parameters

Credit rating may help economic agents to evaluate the financial stability of a bank. Since
credit rating is not a quantitative figure, it is usually compared to the fundamental risk ratios (PD,
LGD and EAD). The CRAs also provide the summary tables where each rating score is related
to the probability of an enterprise’s failure.
Shi et al. (2019) proposed a credit risk model that connected the credit ratings with LGD.
The non-linear relation between the ratings and LGD was demonstrated on the sample of Chinese

115
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

firms. Under certain macroeconomic circumstances, higher credit ratings were associated with
higher LGD, which contradicted to the credit rating definition. The same tendency was noticed
by Shi et al. (2020). The proposed approached helped to explain the mismatch between ratings
and LGD. The correct interpretation of credit risk is that for higher credit ratings, the recoveries
in case of default are high (low LGD).
Godlewski (2007) investigated the consistency between the Moody’s credit ratings and
PD in emerging countries. Scoring and mapping techniques were applied in order to examine the
relation of low ratings and increasing PD. It was shown that credit ratings could serve as good
substitutes of direct PD estimation, as they correctly determined the credit risk. The accurate
quantification of ratings was demonstrated in the paper.
Pomasanov & Vlasov (2008) introduced the model of credit ratings calibration on PD
for Russian banks. The following calibration function was tested:
1
PD  , (2)
exp  A  Re  B 

where Re is the rating grade, A and B are unknown parameters. The annual transmission scale of
credit ratings into PD was prepared and compared to the estimated provided by Standard&
Poor’s. This method can be applied to different credit rating scales, including the national ones,
which was shown in the paper.
The credit scoring model for PD calculation was also used by Chan-Lau (2006). Two
methods were mentioned: cohort analysis (when credit ratings were available to a large number
of companies in a sample) and duration analysis. Similar methods were compared by
Schuermann& Hanson, 2004. The authors used the credit ratings issued by Standard& Poor’s for
estimating PDs and comparing the different ways of confidence intervals calculations.
Hamalinsky & Pomazanov (2012) offered the models for credit ratings and PD calibration
in the samples with small number of bankrupt firms. The proposed method was based on the idea
of benchmarking and genetic algorithms.
Karminsky & Khromova (2018) used the transmission scale of credit ratings and PD
developed by Pomasanov & Vlasov (2008) for ratings calibaration. The non-linear
correspondence of credit ratings and PD was demonstrated. It was more efficiently to use a
combination of an exponential and a polynomial quadratic function than a standard linear
function of dependence.

116
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chapter 2: Hypotheses development

The comprehensive analysis of academic literature helps us to identify the potential
spheres in credit risk estimation. We formulate the following list of hypotheses to be tested in
our paper:
Hypothesis 1: Subjectivity of CRAs’ assessment of international banks’ financial
sustainability can be reduced by using multiple ratings simultaneously in the base scale.
The econometric analysis of our paper begins with constructing credit rating models for
Moody’s, S&P and Fitch Ratings separately. Afterwards, we create a base scale by mapping all
the available credit ratings in our database and build the model on the basis of mapped rating
scores. We compare these approaches and select the most appropriate for the credit risk valuation.
Additionally, this hypothesis is tested by us in order to determine the list of potential indicators
to be included in the final uniform model specification and to assess the possibility of accurate
credit risk assessment with a help of traditional credit rating models (Jivaykina &
Peresetsky,2017; Karminsky & Khromova, 2016; Karminsky &Sosurko, 2010).
Hypothesis 2: Transparency of the public authority and government effectiveness benefit
banks in credit risk management and help them to obtain higher credit ratings.
The inclusion of macroeconomic indicators increases the forecasting power of the credit
rating models (Karminsky & Khromova, 2016; Sahut & Mili, 2011; Betz et al., 2014). However,
macroeconomic factors do not fully incorporate the market conditions in a country. We have
gathered a broad list of indexes for the state governance level in order to measure the quality of
administration, legal compliance and anticorruption laws. These factors are believed to have a
significant impact on credit risk. The hypothesis will be tested by the significance analysis of the
indexes from this set of factors. We believe that banks from the countries with transparent public
authorities tend to get higher credit ratings, which helps them to create a desirable image of their
creditworthiness and to attract more clients.
Hypothesis 3: The share of a factor of liquidity in the credit rating assessment increases
during the periods of financial crises and distress.
It is extremely important for banks to keep adequate level of liquidity during the distress
time (Zhang et al., 2020). Considerable amount of liquidity helps banks to survive and not to go
bankrupt. The hypothesis is tested by marginal effects calculation for the proxy of liquidity in
the time period before, during and after the financial crisis. The comparison of obtained marginal
effects helps us to determine the power of influence of a factor of liquidity in different time
periods.
Hypothesis 4: Banks with recently issued credit ratings tend to be less stable just after the
rating assignment than banks with even lower credit ratings issued a long time ago.

117
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

An analysis of default frequencies of Russian banks is analyzed in order to find evidence

for this hypothesis. We suppose that investors should take into account not only the rating grade
of a bank, but also the time period during which the bank remains in a rating class. Answer to
the stated hypothesis would help economic agents in investment decisions making.

118
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chapter 3. Data description and model specification

3.1. Credit ratings issued by international rating agencies

We collect the credit ratings assigned by the Big-3 international rating agencies: Standard
Poor’s (S&P), Moody’s and Fitch Ratings. The credit ratings are collected from the website
[Link] which provides financial information about the international companies. Additionally,
a part of credit ratings assigned by S&P are extracted from the portal S&P CAPITAL IQ which
is a part of S&P Global Market Intelligence. We build econometric models for the most famous
and respected credit rating agencies. In the models we include the following credit ratings:
Standard& Poor’s Issuer credit ratings (foreign currency LT), Moody’s LT Bank Deposits
(Foreign), Moody’s LT Issuer Rating (Foreign) and Fitch Long Term Issuer Default Rating.
We analyze a time period from the first quarter of 2007 to the first quarter of 2019 year.
This time interval includes the worldwide financial crisis (2008/2009 years) which allows our
econometric calculations to be persistent to shocks. In total, there are 49 quarters. All the data is
considered at the end of the corresponding quarter.
We gather the list of international banks available on S&P CAPITAL IQ. According to
the industry classification provided on the portal, the list consists of diversified banks (diversified
commercial banks, foreign banking and branches of foreign banks, private and industrial
banking), regional banks (national and state commercial banks, regional commercial banks) and
thrift and mortgage finance. Our analysis is based upon the experience of international banks, as
banks from different geographic regions are chosen. The data is conducted of the banks from the
developed and emerging markets from the following regions: Asia/Pacific, European and United
States/Latin America/Caribbean. It is impossible to analyze all the available entities, that is why
the random selections of 500 banks is created for each geographic group using code 1 (see
Appendix 3).
Afterwards, we verify the availability of the credit ratings of either of the three credit
rating agencies (CRA) assigned to the banks. For all randomly chosen 1 500 banks the fact of
rating assignment during the analyzed time period is checked. In total, we are able to obtain the
sample consisting of 478 banks. If a rating agency has assigned a credit rating to a bank at the
end of a time quarter, we assume that this rating remained relevant until the quarter at which the
agency had upgraded it. The final sample contains of 8684 observations (bank- quarters) for
Moody’s agency, 7195 for S&P and 6261 for Fitch rating agency. We notice that each
observation corresponds to the credit rating assigned to a bank by the CRA at the end of the
quarter.

119
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The sample consists of 46 countries: 14 developed and 32 developing ones. The list of
countries presented in the selection may be found in the Table 23 (see Appendix 1). Figure 1
illustrates the number of observations (bank-quarters) for each geographic region.

Figure 1: Observations per geographic region, the sample

8000

2797
6000

3104 5106
Emerging markets
4000
Developed markets
5372
2000
3340
2421

0
Asian European American

Source: Author's calculations

The selection includes 6 261 credit ratings assigned by Fitch, 7 195 ratings given by S&P
and 8 684 credit ratings assigned to the entities by Moody’s. The sample represents different
geographic regions, just as the initial data extracted from the portal, which makes it
representative. The representation of the ratings included in the model is presented on Figure 2
below.

Figure 2: Distribution of credit ratings across geographic regions, the sample

10000

8000 1304
594 1512 Emerging (American)
6000 899 2226 Developed (American)
1438
1634 Emerging (Asian)
4000 985 1356
681 Developed (Asian)
815 1169
2000 Emerging (European)
2267
1565 1274 Developed (European)
667 947 807
0
Fitch S&P Moody's
Source: Author's calculations

Figure above demonstrates that the ratings in the sample are randomly distributed among
the regions, so there is no bias towards either of the geographic groups used in the analysis, which
helps us to capture different country-specific characteristics of financial institutions. It is also
important that the representative sample contains banks with credit ratings from different classes

120
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

assigned by the CRAs in order not to cause survivorship bias. As there are some defaulted banks
included, credit classes “Highly Speculative” and “Default” are present in the selection. Figure
3 illustrates the proportion of ratings from each class.

Figure 3: Distribution of credit ratings between the classes, the sample

3000
Fitch S&P Moody's
2500

2000

1500

1000

500

0
High Upper Lower Speculative Highly Extremely
Medium Medium Speculative Speculative
/Default
Source: Author's calculations

Figure above depicts that the most common rating classes presented in the sample are the
Upper Medium and Lower Medium ones. A greater amount of credit ratings in the speculative
level are issued by Moody’s than by S&P or Fitch. In all other credit rating levels, the distribution
is similar for all three rating agencies. The review of the data shows that we were able to generate
a random and representative sample of banks from different countries. The comparison and
correspondence of credit ratings of different CRAs with numerical equivalents are provided in
Table 24 (see Appendix 1).

Table 2: Descriptive statistics of credit ratings, the sample

Observations Mean Std. Dev.
S&P ratings 7 195 13.69457 3.423604
Moody's ratings 8 684 12.84178 4.448751
Fitch ratings 6 261 13.33847 4.022095
Source: Author’s calculations

The credit ratings of these three rating agencies are highly correlated. All the Spearman
rank-order correlations are highly signification (see the table below). There appears to be a strong
positive correlation (more than 90%) between the credit ratings of different CRAs. The high
correlation level between the ratings is reasonable and expected, as it is usually believed that the

121
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

rating agencies analyze very similar set of indicators of bank’s financial performance in order to
assign the rating.

Table 3: Spearman’s rank correlation coefficients (credit ratings)

S&P Moody's Fitch

S&P 1
Moody's 0.9278*** 1
Fitch 0.9257*** 0.9398*** 1
*** - statistically significant at 1%
Source: Author’s calculations

3.2. Selected indicators of the banks’ sustainability

There is no universal algorithm that would allow to objectively assess the bank’s credit
risk. A full analysis of the financial condition of a credit institution should be carried out with
the help of a large set of indicators responsible for a particular factor. In order to form an opinion
about the organization, CRAs periodically conduct on-site inspections, which assesses the
financial stability of the organization, its compliance with the law and the requirements of
regulatory bodies, as well as the quality of management and internal control system (Karminsky,
2015).
In this work we focus our analysis primarily on the indicators which were previously used
in the academic literature for measuring the financial risk. We have explicitly discussed them in
Chapter 1. On the basis of our review of previous researches, we are able to gather financial ratios
which can help in measuring the financial stability of a banking organization. We use the portal
S&P CAPITAL IQ as a source of companies’ financial information. It provides independent
reports, financial and statistical indexes for international companies from different spheres and
industries. All the indicators are collected for a time period from the first quarter of 2007 to the
first quarter of 2019. At the moment of collecting the information, it is the most recent data
available on the site. In the cases when the value of some indicator is absent, it is averaged by
the nearest time periods. We summarize all the found variables potentially responsible for credit
risk in the table below. Although not all the collected variables are used in modeling, we find it
important to mention all of them, as it might be helpful for further research on this topic.

122
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 4: Selected indicators of banks’ financial performance

Factor Indicator Abbreviation

Tier 1 capital ratio TIER 1
Total Equity to Total Assets EQ/AS
Capital adequacy
Total Equity to Total Debt EQ/DE
Total Capital ratio CAP
Provision for Loan Losses to Gross Loans PROV/G
Asset quality
Nonperforming Loans to Total Loans NONL/L
Allowance for Credit Losses to Nonperforming Loans ALL/NONL
Management Interest Expenses to Total Deposits IE/DE
quality Total Debt to Total Deposits DEB/DE
Operating Expenses to Total Revenues OE/REV
Return on Assets (ROA) ROA
Earnings
Return on Equity (ROE) ROE
Net Loans to Total Deposits NL/DE
Total Cash and Equivalents to Total Deposits CASH/DE
Total Deposits to Total Equity DE/EQ
Liquidity
Liquid Assets to Non-Liquid Assets LA/NLA
Net Loans to Total Assets NL/A
Current Assets to Current Liabilitites CUR
Net Interest Income to Total Revenues NII/REV
Net Interest Income to Average Assets NII/AA
Efficiency Operating Expenses to Operating Income OE/OI
Operating Expenses to Net Interest and Non-Interest Revenues EFF
Total Non-Interest Expenses to Total Assets NONIE/A
Natural logarithm of Total Assets LOGA
Size
Natural logarithm of Total Equity LOGEQ
Growth rate of Net Income NI
Growth rates
Growth rate of Total Revenues REV
Leverage Tier 1 Capital to Adjusted Average Total Assets LEV
Age Natural logarithm of bank's age AGE

We have managed to collect 29 indicators of influence on banks’ financial sustainability.

Such an extensive set of potential variables can help us to choose the most influential and
significant in credit rating forecasting. We try to find several indicators associated with each
factor of financial sustainability in order to deeply analyze each of them and include in the final
model specification only those that have a significant effect on the credit ratings. The first five
factors are collected according to the CAMEL methodology discussed in Chapter 1. The previous
researches show that it is sometimes difficult to account for banks’ efficiency, that is why five
possible indicators are chosen by us. Size of the bank measures its resistance and reliance
(Pasiouras et al., 2007; Alrabadi and Hamarneh, 2016; Cogas et al., 2014). We analyze LOGA

123
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and LOGEQ as potential indicators of banks’ size and expect their positive influence on banks’
credit ratings. As a measure of banks’ development growth, we analyze the effect of NI and REV.
The dynamics of these two components may be also helpful in the rating determination. Tier 1-
Leverage ratio LEV may help to assess the possibility of bank to meet its obligations and to
determine the overall financial position. This index is used for banks and considered to be a more
accurate proxy than a Debt to Equity ratio due to operational activities of banking organizations.
It is not enough to include the financial variables only. Business risk is also important to take
into account while modeling the credit rating model. Karminsky and Peresetsky (2007) were able
to demonstrate that the accuracy of the model for international banks increases if we account for
macroeconomic factors. We collect the macroeconomic ratios that are usually used in the
academic literature. The source of information is [Link] and The World Bank. All the
variables are collected from the first quarter 2007 to the first quarter 2019. We represent the full
list of macroeconomic variables in the Table 5.

Table 5: Indicators of macroeconomic situation

Indicator Abbreviation
Inflation(CPI) growth rate (%) CPI
GDP per capita growth rate (%) GDP
Ratio of Domestic credit provided by financial sector to GDP (%) DOMC
Ratio of Current account balance to GDP (%) CABAL
Ratio of Gross savings to GDP (%) GROSS
Ratio of Exports of goods and services to GDP (%) EXP

CPI has an expected negative effect on credit ratings. An increase in the growth of
inflation leads to a market risk which is characterized by an increased level of non-fulfillment of
financial obligations and inaccessibility of financial assets. GDP is used as a measure of
economic growth and wealth of residents inside the country. This indicator is expected to have a
positive effect on credit ratings and bank’s financial performance. DOMC measures the credit
provided by financial sector. Credit is one of the key earning assets of banks as it brings them
interest which is a part of interest income. DOMC has a potential positive influence on the credit
ratings assigned by CRAs. CABAL measures the balance of payments and may have different
effect on the financial sustainability of banking organizations. The higher people save, the more
money they put on their banking accounts and hence GROSS is expected to have a positive effect
on banks’ earnings. EXP affects the credit ratings both positively and negatively. High level of
exports implies a favorable economic situation in a country; hence the financial position of a
bank is generally better. However, higher exports level may also indicate that large amounts of
money are sent to foreign banks, which is harmful for the banks inside the country.

124
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

We may not include all the macroeconomic indicators which were selected in the final
model specification as the correlation coefficients between some factors may lead to
multicollinearity in the model. The table with correlation coefficients of macroeconomic
variables is presented in Table 25 (see Appendix 1). As we can see, correlation coefficient is very
high (more than 60%) between the following pairs of variables: DOMC and GDP, GROSS and
CABAL. Correlation is not so high, but still reasonable (more than 40%) between GDP and CPI,
DOMC and CPI, GROSS and GDP, GROSS and DOMC, EXP and CABAL, EXP and GROSS.
The pairs of variables mentioned above cannot be included in the model together. If we want to
measure the risk associated with macroeconomic situation in a country, we need to make sure
that the model is not suffering from multicollinearity, as this may lead to poor results.
Macroeconomic indicators cannot fully explain the business risk associated with a
particular country. That is why we collect the indicators responsible for the state governance
quality. This is the first time when such set of variables is used in credit rating modelling. We
have gathered these variables from The World Bank. These indexes are the components of the
database “Worldwide Governance Indicators”. According to Moody’s Methodology (2016), the
quality and transparency of public institutions are essential in influencing stability of banking
system.

Table 6: Indicators of state governance quality

Indicator Abbreviation
Control of Corruption (Estimate) CORRC
Government Effectiveness (Estimate) GOVER
Regulatory Quality (Estimate) REGUL
Political Stability and Absence of Violance/Terrorism (Estimate) POLIT
Voice and Accountability (Estimate) VOICE
Rule of Law (Estimate) LAW

CORRC measures the usage of corruption schemes by the public power for private gain.
It also captures the legitimacy of the public power, i.e. whether it is grasped by some elite groups
inside the country. GOVER is responsible for the quality of public services provided by the
government. REGUL measure the transparency and clarity of the actions and policies pursued
by the public power. POLIT captures the likelihood of exploitation of politically motivated
violence towards the people inside the country. It shows the possibility of using illegal actions
(such as terrorism) by the government. VOICE measures the observance of the following human
rights: freedom to choose power, freedom of expression, freedom of association, and a free
media. LAW measures the overall situation with compliance with the law inside the country:

125
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

independence of the court, lawfulness of the police, etc. All the above-mentioned indexes vary
from -2.5 (the most difficult situation) to 2.5 (favorable environment). The methodology of
estimating each country’s score is explained in the working paper (Kaufmann et al., 2010).
All the indexes responsible for state governance level are heavily correlated between each
other (see Table 26 in Appendix 1), so even two indicators cannot be included in the model
simultaneously. As we do not know exactly what indicators might be useful in credit rating
forecasting, some techniques are needed to be applied to explain as much variation as possible.

3.3. Construction of the base scale of credit rating scores

International rating agencies use different methodologies and evaluate different

components while assigning a rating. Methodological differences in the approaches of credit
agencies (different scales, etc.) make it difficult for us to use these credit ratings in the model. In
order to solve all the misperceptions, we try to consolidate the credit agencies’ assessments into
one scale. This will help us to construct a single rating space (Karminsky, 2015), at which we
will be able to compare the credit ratings of different agencies more accurately. In order to build
a single rating scale, we form a system of displaying different ratings of all considered agencies
into the base scale. The detailed comparison of various credit ratings is presented in the article
(Karminsky and Sosyurko, 2011). The authors developed a methodology for matching the rating
scales of international and Russian agencies. It was demonstrated that the class of linear-
logarithmic transformations was useful and efficient in ratings mapping. The constructed models
allowed for comparison of national and international ratings. The emphasis was made on Russian
banks, but database with international companies was also used for validation of the models.
The paper by Ayvazyan et al. (2011) presented a method of credit ratings mapping of 10
different agencies. The method was to compare the ratings of Russian banks on a basis of
econometric models. The main hypothesis tested was that the objective comparison of rating
scales should not be based only on direct correlation of rating pairs to each other. The author
provided a detailed analysis of their econometric methodology and tested it empirically on
Russian banks.
The unification of credit ratings of different agencies to the common scale is crucial as it
will help us to compare the differences between the credit ratings accurately. Before applying the
logarithmic transformation for credit ratings modelling, we need to evaluate the dependence
between the credit ratings of international banks used in the analysis. We have the ratings
available for the most famous and reliable rating agencies – Moody’s, S&P and Fitch. Although
these rating agencies are believed to apply different methodologies, their ratings are usually
considered to be similar as the long-time experience of these agencies allows them to assess the

126
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

financial performance of a bank more objectively. We have shown the similarity of these ratings
at Table 3. The analysis of graphical representation of rating matches of different rating agencies
can be helpful in determination of the method for credit ratings mapping. From Figures 22 and
23 (see Appendix 2), it becomes clear that the relation between Moody’s ratings and S&P ratings
is close to linear and the dependence between these credit ratings can be built with a linear
regression. The similar relation is seen for the pair of credit ratings Moody’s – Fitch for the
international banks. In the Table 28 and Table 29 (see Appendix 1) we also provide the pairwise
tables of ratings correspondences. From these tables it can be clear that the majority of differences
between the credit ratings appears in high grade and speculative classes. In the field of the upper
and lower medium grade, the credit agencies tend to give similar credit ratings.
In our work we will use the linear logarithmic model which was successfully applied in
the papers (Karminsky& Sosyurko, 2011; Karminsky et al, 2013) in order to obtain the pairs of
coefficients for displaying each of analyzed scale into the base one:
ln  M   ai ln  Ri   bi . (3)

For each rating scale i the equation (1) is built, where M – Moody’s rating in numerical scale; Ri
– rating of an agency i in the numerical scale; ai and bi are the found coefficients of the model.
We use the Moody’s international scale as a base one as it has the greatest number of ratings-
pairs with other scales. The results of logarithmic modeling and obtained coefficients ai and bi
for each of the rating agencies are presented in the Table 7. All the coefficients are significant at
the 1% significance level.

Table 7: Coefficients of logarithmic model of ratings mapping

Rating scale ai bi Number of observations R2

Standard & Poor’s 1.156 -0.402 4784 0.86

Fitch Ratings 1.019 -0.058 4487 0.86

Source: Author’s calculations
Based on the resulting coefficients, which determine the display of scales into the base
one, we can assess the correspondence of ratings of considered agencies with international
Moody’s scale. The graphical representation of comparison of rating scales is provided below.
The table form of ratings correspondence is provided in the Appendix.

127
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 4: Correspondence of rating scales

AAA AA+ AA AA- A+ A A- BBB+BBB BBB- BB+ BB BB- B+ B B- CCC+ CCC CCC- D

AAA AA+ AA AA- A+ A A- BBB+ BBB BBB- BB+ BB BB- B+ B B- CCC+ CC C D

Aaa Aa1 Aa2 Aa3 A1 A2 A3 Baa1 Baa2 Baa3 Ba1 Ba2 Ba3 B1 B2 B3 Caa1 Caa2 Caa3 Ca C

23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
Base scale
Moodys SP Fitch
Source: Author's calculations

From Figure 6 we can see what value a particular symbolic rating takes in the sample.
For example, Moody’s Caa2 symbolic rating is represented in the data as 3, S&P’s AA- rating is
transformed into the number 18.5. The numeric scale varies from 1 (the lowest ratings: D, C) to
21.5 (the highest rating, AAA, assigned by S&P). Rating agencies do not often change their
ratings when the financial performance of bank is stable, hence we assume that the ratings are
unchanged until the moment when the new rating is assigned. The figure also depicts the
differences in the scales of CRAs.

3.4. Description of ordered probit/ logit models

In this paper we build the models to predict a categorical variable yit (a credit rating),
hence standard logit and probit models cannot be used. Ordered logit/probit models will be
applied. Standard binary logit models were extended to categorical outcomes estimation in Snell
(1964). For the first time a probit model for categorical variables was described in the article
(Aitchison and Silvey, 1957). The authors analyzed the responses with three or more outcomes.
Afterwards, this kind of models was wildly used in researches in various fields, including credit
rating modeling. Now these model specifications are considered as “traditional techniques” for
credit rating modeling due to their common usage.
Consider that observed variable has an order with more than 2 levels (credit ratings may
take several rating scores). Categorical outcome yit is ordered, i.e. yit  1,2,..., h , where h

denotes the total number of categories. The order 1 is the lowest credit rating assigned, order h
is the highest rating. For each agent i , the relationship between the rating class yit and set of

covariates xit is defined using unobserved continuous latent variable yit* :

yit*  xit  it , (4)

128
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

where it  
iid , 2 is the error component which has a logistic distribution with mean   0


and standard deviation   or the standard normal distribution N(0,1);
3

i is the number of observations, 1 i  n ;

t is the number of periods (quarters), 1 t   ;

xit  is a set of parameters (covariates) which helps to explain the dependent variable;
 is the vector of coefficients estimated by a model.

The observed variable yit takes the value yit  1,2,..., h depending on the latent variable

crossing some particular threshold value:

1 ,if yit*  m1;

2 ,if m1  yit*  m2 ;
yit   (5)
...

h ,if mh 1  yit ,
*

where m   m1, m2 ,..., mh 1  is the vector of cut-off values estimated by a model.

Since yit is a continuous and non-deterministic variable we can only measure the
probability of it belonging to a particular class. For example, the probability that a particular bank
will be assigned a unit credit rating yit  1 is calculated in the following way:

     
Pr( yit  1)  Pr yit*  m1  Pr xit  it  m1  Pr it  m1  xit   F (m1  xit ) . (6)

The estimated models calculate the probabilities for a bank to get each credit rating form
the range from 1 to h. Probabilities are calculated in the same way as above for each credit rating.
In general, for h ordered alternatives, the probabilities that bank i is assigned a credit rating
yit  1,2,..., h are:

Pr( y  1)  F (m1  xit );

 it
Pr( y  2)  F ( m2  xit )  F ( m1  xit );
 it
... (7)

Pr( yit  h  1)  F (mh 1  xit )  F ( mh  2  xit );

Pr( yit  h)  1  F ( mh 1  xit ),

where F is the cumulative distribution (cdf) function of the standard normal distribution (for
ordered probit model):

129
32nd EBES Conference Proceedings - Volume I August 5-7, 2020


1  2 
F ()  ()     2 d
2 
exp  (8)
 
or the cdf of logistic distribution (for ordered logit model):
exp()
F ()  ()  . (9)
1  exp()
The only parameters estimated on the real data by the model are the vectors of coefficients
 and the vector of cut-off points m. After their estimation, forecasted values of probabilities are

calculated, attributing each bank i to each credit rating from the class yit  1,2,..., h . It is

reasonable to assign the bank to a credit rating for which the estimated probability is the highest
one. The parameters are found using the maximum likelihood estimation and the interpretation
of coefficients is possible using their marginal effects.
We have estimated both ordered probit and logit models, but ordered logit was preferable
due to higher predictive power and greater number of significant variables. In all further chapters
we present the results of ordered logit models estimation.

130
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chapter 4. Building a credit rating model for international banks

In Chapter 1 we provided the review of the academic articles on the research topic and
made a list of potential factors of influence on credit ratings that are often included in the models.
According to this list, we gathered the explanatory variables which could be used in credit rating
modeling and described them in Chapter 3. Now we can proceed to the empirical modeling of
credit ratings. The plan of modeling procedure used in the work is provided below.

Table 8: The modeling procedure

- Detailed review of previous works on credit risk assessment of financial organizations

- Creation of a list of potential indicators of influence on company's credit rating

- Traditional credit rating modeling for Big-3 rating agencies

- Determination of significant factors of influence on banks' rating
- Finding out problems and difficulties in traditional rating modeling procedures

- Building a uniform credit rating model with errors close to Normal ones
- Solving the multicolinearity problems using PCA
- Robustness check of the obtained model

- Estimation of PD by historic default frequences on the example of Russian banks

- Calibration of credit ratings on probability of default

We begin with traditional models of credit ratings for each CRA separately. Afterwards,
all the ratings are aggregated into one scale and a uniform model is built with a help of several
interaction variables and Principal Component Analysis (PCA). PCA helps us to include the
indicators of state governance level, which constitute an expert assessment of overall state
administration and are highly correlated. The resulting models are checked for their robustness.
In the last chapter of the paper we present the estimation of PD using default frequencies of
Russian banks. With the help of an independent database, we provide the calibration of credit
ratings and PD for Russian banks, which can be useful for banks and counterparties (investors,
experts, etc.).

131
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4.1. Credit rating models for different agencies

In this section we present the models aimed at credit rating forecast on the basis of public
information for three agencies: Moody’s, Standard & Poor’s and Fitch Ratings. It takes some
time for the credit rating agency to prepare and change the credit rating of a bank, hence it is
important for us to include the explanatory variables into the model with a lag. The authors in
the article (Karminsky and Sosyurko, 2010) analyzed the lag which should be used for credit
rating modeling. They took different values of a time lag (from 3 to 30 months) and found out
that the results were quite similar; the minor differences in the results from the lag selection were
confirmed. In our work we decided to use the time lag of 2 quarters. In order to control for
heteroscedasticity, we use White-Huber standard errors.
While constructing the model we control for correlation level between the explanatory
variables included (correlation is less than 40%). We try to include the indicators for all the
factors from the CAMEL(S) model, as they are believed to explain the bank’s overall internal
financial condition. Table 27 (see Appendix 1) demonstrates that the financial indicators for the
same factor are usually highly correlated, hence we include only one index per each factor in the
model. The choice of a particular indicator was made taking into account its expected economic
interpretation, significance (p-value) and its influence on the forecasting power of the model
(Akaike information criterion and Schwarz information criterion).
We have built both random effects ordered logit and probit models, but the ordered logit
model has more significant variables and better forecasting power, that is why we show the
results of ordered logit model estimation. The models are built for unbalanced panel data, as
banks might not have credit ratings during any time period. In order not to cause a possible bias,
we have decided not to cut the data.
We start with the first model specification (models m(1), sp(1) and f(1)) including
financial variables (with interaction terms) only:
yits  1CAPit + 2 NONL/Lit +3IE/DEit + 4 ROAit +5LA/NLAit + 6 EFFit +
+ 7 LOGAit  8LEV*EUit   9 LEV*CISit  10CUR*Y2008it  11CUR*Y2009it  (10)
 12CUR*Y2010it  it .

The second model specification (models m(2), sp(2) and f(2)) includes the country-
specific ratios (macroeconomic and state governance indexes):
yits  1CAPit + 2 NONL/Lit +3IE/DEit + 4 ROAit +5LA/NLAit + 6 EFFit +
+ 7 LOGAit  8LEV*EUit   9 LEV*CISit  10CUR*Y2008it  11CUR*Y2009it  (11)
 12CUR*Y2010it  13GROSSit  14CABALit  15POLITit  it ,

where yit is the credit rating grade, s is the index for the CRA (Moody’s, S&P or Fitch).

132
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 9: Ordered logistic models for different credit rating agencies

Moody’s S&P Fitch

Indicators m(1) m(2) sp(1) sp(2) f(1) f(2)

CAP 0.031* 0.049* 0.051* 0.065* 0.029 0.046*

(0.006) (0.006) (0.011) (0.013) (0.011) (0.011)
NONL/L -0.077*** -0.053*** -0.049*** -0.029*** -0.010 -0.049***
(0.005) (0.005) (0.007) (0.008) (0.007) (0.007)
IE/DE -8.236*** -4.223** -1.873* -2.194** -23.353*** -9.986**
(1.715) (1.756) (0.989) (1.012) (4.242) (4.267)
ROA 3.413*** 2.974** 3.542** 3.930** 0.717 0.076
(1.217) (1.296) (1.570) (1.754) (1.268) (1.205)
LA/NLA 0.021*** 0.018*** 0.019*** 0.019*** 0.003 0.005**
(0.003) (0.003) (0.005) (0.005) (0.002) (0.002)
EFF -0.041*** -0.039*** -0.013** -0.016*** -0.009 -0.025**
(0.007) (0.007) (0.006) (0.006) (0.009) (0.010)
LOGA 0.736*** 1.094*** 0.822*** 1.163*** 0.816*** 1.465***
(0.067) (0.075) (0.144) (0.149) (0.104) (0.118)
(LEV)*(EU) 0.324** 0.291** 0.754*** 0.822** 0.653*** 0.589**
(0.116) (0.111) (0.244) (0.249) (0.174) (0.174)
LEV*(CIS) 0.249*** 0.254*** 0.078 0.023** 0.114* 0.051***
(0.056) (0.057) (0.056) (0.007) (0.066) (0.077)
(CUR)*(Y2008) 0.013*** 0.013*** 0.011* 0.011* 0.728*** 0.727***
(0.003) (0.003) (0.006) (0.006) (0.103) (0.102)
(CUR)*(Y2009) 0.027* 0.036*** 0.069*** 0.084*** 0.249** 0.251**
(0.014) (0.014) (0.019) (0.019) (0.08) (0.08)
(CUR)*(Y2010) 0.013* 0.013** 0.026* 0.028** 0.229** 0.226**
(0.007) (0.007) (0.01) (0.01) (0.07) (0.07)
GROSS 0.024* 0.092*** -0.021
(0.014) (0.028) (0.023)
CABAL -0.055*** -0.094*** -1.369*
(0.013) (0.024) (0.792)
POLIT 0.251** 0.775*** 0.673***
(0.107) (0.186) (0.160)
Sample size 8,684 8,671 7,186 7,181 6,261 6,256
AIC 24732.98 24425.01 11121.29 10960.23 13214.53 13072.24
BIC 24937.99 24686.51 11313.93 11207.88 13403.31 13314.93
Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1
Source: Author’s calculations

133
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In the first model specification (m(1), sp(1) and f(1)) we include only financial variables
for the following factors: CAMELS model, efficiency and leverage. Also, we add the natural
logarithm of bank’s total assets (LOGA) as a measure of size. Models m(2), sp(2) and f(2) are
built for each СRA but with two macroeconomic indexes (GROSS and CABAL) and one
indicator of state governance level (POLIT). It can be seen that adding country-specific variables
increases the forecasting power of the model. Hence, we will interpret the second specification
of the models.
Total capital ratio (CAP) stands for capital adequacy and represents the bank’s Total
Capital as a percentage of Total Risk-Weighted Assets. CAP has a positive effect on bank’s credit
ratings of each CRA and is highly significant, which is expected. If the bank has enough capital
compared to the risk it is taking, then it is believed to be financially strong and its credit rating is
higher. On the contrary, the credit rating is negatively affected by the situation when the bank
increases its risky assets without capital extension. NONL/L measures the bank’s asset quality.
It is significant at 1% level for each agency in the second specification of the model.
Nonperforming loans usually comprise of non-accrual loans and restructured loans. If the share
of bad loans among total loans is high, the bank’s asset quality is poor, and it is possibly close to
default. That is why NONL/L has a negative influence on a rating. As a measure of management
quality, we include IE/DE; it is significant in the model for each agency and stands there with
negative sign. IE/DE can be interpreted as the price of funds as the bank’s total deposits represent
its total funding. Banks with lower price of funds have the better management quality as their
costs are lower and, as a result, they receive higher ratings. Despite experienced major banks
usually offer deposit rate lower than emerging commercial banks, people still prefer to deposit
money in them as they believe in its management level. ROA is an indicator for bank’s
profitability, and it is significant for Moody’s and S&P. Fitch may take into account different
variables as a measure of earnings, but we could not find ones in our selection. Return on Assets
shows the ratio of Operating Assets to bank’s Total Assets. Banks with stable profitability have
higher Operating Assets and this has a positive influence on their credit ratings. Our results are
consistent with those obtained by Thomson (1991). As a proxy for liquidity we use LA/NLA,
which is significant for three rating agencies. If the ratio LA/NLA is high, then the amount of
liquid assets is greater than non-liquid ones. This indicates the ability of banks to convert cash
quickly to meet its obligations. It is necessary in arduous situations as liquid assets may prevent
banks from bankruptcy. Banks with greater amount of liquid assets tend to have higher ratings
of three rating agencies. Efficiency is measured in our model by the bank efficiency ratio (EFF),
which is significant in credit rating models and has a negative effect. It helps to depict the
effectiveness of internal resources usage. This index is used by banks’ internal management but

134
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

is also taken into account by СRAs (as results of model estimation show) to assess the
performance of bank’s operations expenses to generate revenue. Banks with higher EFF are less
efficient as the amount of operations overhead does not generate enough revenue; such banks get
lower credit ratings.
Natural logarithm of bank’s total assets (LOGA) is significant in all the reported models.
The higher the total assets, the larger the bank is believed to be. Our models show that the CRA’s
methodology assumes that large banks have smaller probability of default and generally more
financially stable. Hence, LOGA has a positive influence on the credit ratings assigned to the
banks in the sample.
It is essential to assess the effect of country-specific variables on the bank’s ratings.
GROSS is significant for Moody’s and S&P with positive affect on banks’ ratings. It is reported
with negative sign in model f(2) for Fitch, but it is insignificant there. Amount of deposits put in
the banks directly depends on national gross savings. People with higher savings will deposit
their money in the banks. CABAL depicts the Current account balance and is significant for all
CRAs. The balance of payments (BP) occurs when the sum of Current Account (Cu) and Capital
Account (Ca) equals zero. When Cu is positive, Ca should be in deficit to equilibrate BP.
Financial account deficit causes capital outflow from the country. This is harmful for banks inside
the country as people withdraw their money from them. That is why CABAL index is reported
with negative sign in all the columns in Table 9.
Due to high correlation between indicators of state governance level, we are able to
include only one index – POLIT, which is highly significant for all the СRAs we build the models
for. This serves as a proof of the Hypothesis 2. Political stability may influence the banks’
financial strength both directly and indirectly. Good political situation affords people to earn and
save money, putting them on bank accounts. What is more, banks in the countries with stable
politics tend to concentrate more on its operational activities than on solving different
bureaucratic problems. We would like to include other variables from this selection, but in order
to avoid multicollinearity, we include only one. However, the methodology of СRAs may include
other variables to evaluate the level of governance in the country where the bank operates.
We include the cross terms (LEV)*(EU) and (LEV)*(CIS) to see the difference in the
effect of leverage on credit risk of banks from emerging (CIS) and developed (EU) countries.
CIS and EU are the dummy variables which equal to 1 if the bank operates in one of the countries
from CIS or EU organizations respectively. CIS organization includes emerging countries
(Russia, Ukraine, Belarus, etc.), while the members of EU are the developed countries (France,
Germany, etc.). We can see that LEV ratio positively affects the credit rating of each agency.
LEV ratio shows whether a bank can effectively resist to financial crises. High LEV ratio implies

135
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

a sufficient amount of Tier 1 capital, a bank’s core capital that is considered to be very stable and
liquid. LEV is positive for banks from CIS and EU countries and significant in all the final
models. However, we may not compare the power of the influence as the coefficients in the
ordered logit models cannot be interpreted.
In order to see the influence of liquidity (CUR index) we include the cross terms:
(CUR)*(Y2008), (CUR)*(Y2009) and (CUR)*(Y2010). Variables Y2008, Y2009 and Y2010
equal 1 if the rating was assigned in 2008, 2009 and 2010 years respectively and 0 otherwise.
These interaction terms are significant in m(2), sp(2) and f(2) models and stand with positive
sign. This means that CUR ratio positively influences on banks’ financial sustainability in the
periods before, during and after financial crisis happened in 2009 (as lagged variables are
included). We may not completely measure the effect, as the significance level is different for
different CRAs. Hence, no particular interpretation of the influence on credit risk can be
presented.
Overall, we are able to demonstrate that it is possible to predict the rating scores issued
by different CRAs using the same set of explanatory variables. However, we fail to reject the
first hypothesis stated in chapter 2. CRAs may have different methodologies and various external
factors (subjectivity, competitiveness, regulations, etc.) may lead to biased credit scores. The
significance levels of indicators are different for CRAs and some indexes are significant for
Moody’s and S&P but insignificant for Fitch. Hence, we cannot accurately interpret the influence
of these explanatory variables on the credit risk, which proves Hypothesis 1. With a help of such
models we can determine the influence of indexes on the credit rating of any CRA, but not on
overall bank’s credit risk.

4.2. Principal Component Analysis

Principal Component Analysis (PCA) may help to capture as much information about the
data as possible in the minimum number of variables. PCA is also helpful in solving the
multicollinearity problem. This method allows to reduce the number of variables by selecting the
most volatile ones.
Mathematically, this method enables us to switch to a new set of representative
explanatory variables, not correlated with each other.
Each of the principal component (PC) represents a linear combination of explanatory

 
variables Z1, Z 2 ,..., Zq with particular weights:

PC j  1 j Z1  2 j Z 2  ...  qj Z q . (12)

136
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The principal components are built in such a way that the sum of squared loadings
(weights) equals 1:


q
2 j  1 . (13)
1

Principal components are functional and the scores (factor scores) are calculated for each
variable and observation in the sample. For accurate calculation of the scores it is better to
normalize the explanatory variables for them to have zero mean. It is a useful procedure as the
variables may have different measurement units and some outliers may present. The scaling of
each feature xit is done according to the formula:

xit  E  x 
zit  . (14)
Var  x 

Afterwards, the scores for each principal component are calculated. Given the panel data,
the linear combination of the sample values of explanatory variables represents the factor scores
(James et al., 2013):
pcitj  1 j zit1  2 j zit 2  ...  qj zitq (15)

subject to the constraint (13).

In (12), (13), (14) and (15) the following notations are used:
i is the number of observations, 1 i  n ;
t is the number of periods (quarters), 1 t   ;
j is the number of a principal component;
 is the number of features (explanatory variables), 1    q ;

 is the weight (loading) for each feature xit .

The first principal component is built is such a way that it has the largest variance. The
scores of the first PC have a form:
pcit1  11zit1  21zit 2  ...  q1zitq . (16)

The weights of the first principal component are found from the variance maximization.
Since the explanatory variables zit are centered (have zero mean), the average of pcit1 is zero
as well. The loading vector of the first principal component solves the problem:
2
  
1 n   q 
 1zit   max (17)
n i 1 t 1 1  11 ,...,q1


q
s.t. 2 1  1 .
1

137
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

This is solved by eigen decomposition and the weights of the features for the first
principal component are found.
Macroeconomic situation in a country can be depicted by a various set of indicators. We
have 6 macroeconomic indicators (see Table 5) which are significant if we include them by one
into the model. The inclusion of too many variables may harm the forecasting power and lead to
overfitting of the model. What is more, some macroeconomic variables are highly correlated
between each other, which is a source of multicollinearity problem. Running PCA on
macroeconomic variables (CPI, GDP, DOMC, CABAL, GROSS and EXP) results in the
principal components (MAC1, MAC2, …, MAC6) with the loading vectors (eigenvectors)
presented in the table below. In this table we can see how much variation of each explanatory
variable is caught by a principal component.

Table 10: Loading vectors of PCs (macroeconomic situation)

Variable MAC1 MAC2 MAC3 MAC4 MAC5 MAC6

CPI 0.3269 -0.4471 0.2602 0.7859 0.0543 0.0703
GDP -0.5451 0.2539 0.2262 0.2518 -0.2402 0.6827
DOMC -0.4855 0.3394 -0.1526 0.4592 0.4267 -0.4825
CABAL 0.3291 0.559 -0.0363 0.2624 -0.6632 -0.2632
GROSS 0.4312 0.3678 -0.5138 0.1288 0.4142 0.476
EXP 0.2569 0.4157 0.7697 -0.1503 0.382 -0.0147
Source: Author’s calculations
The first principal component (MAC1) has large association in absolute terms with GDP,
DOMC and GROSS. This component primarily measures the standard of living (GDP) and
financial sustainability of the population (DOMC and GROSS). Alternatively, MAC2 hase large
association with CPI (-44%), CABAL (56%) and EXP (41%). It would be reasonable to say that
MAC2 is responsible for the price level (CPI) and export opportunities (CABAL and EXP) inside
the country.
There are two possible ways to choose the optimal number of principal components to be
included in the model. In our paper we will consider two methods in detail. It is important to
examine the proportion of variance explained (PVE) by each of the six principal components.
The total variance of the features in a panel data is calculated as:

  
1 q n 
Var  zit2 (18)
n 1 i 1 t 1

and the variance explained by jth principal component is defined by:

2
    
1 n  1 n   q 
Var j  2
pcitj   j zit   . (19)
n i 1 t 1 n i 1 t 1 1 

138
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Hence, the PVE of the jth principal component is given by:

Var j
PVE j  . (20)
Var
Equivalently, it may be written as:
2
i1 
n   q 
 j zit  
t 1 1  .
PVE j  (21)
 1i1t 1
q n  2
zit

The first principal component (MAC1) explains 40% of the total variations inside the
group, whereas MAC2 and MAC3 explain 30% and 13% of the variance respectively. Remaining
17% are explained by three other components (MAC4, MAC5 and MAC6). Together first two
principal components explain 70% of the variance in the data, which is significant. Analysis
shows that at least two components should be included in the model.

Figure 5: PVE by each PC (macroeconomic situation)

50% 100%

40%
80%

30%
60%
20%

40%
10%

0% 20%
MAC1 MAC2 MAC3 MAC4 MAC5 MAC6
Individual Cumulative
Source: Author's calculations

Another way to select the number of principal components to be included in the model is
to assess the value of the eigenvalues. The Kaiser criterion (Kaiser, 1960) states that we need to
retain the eigenvectors (principal components; the columns in Table 10) with corresponding
eigenvalues greater than one. Principal component with an eigenvalue of unity explains the same
amount as a single dependent variable.

139
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 6: Scree plot of eigenvalues for PCs (macroeconomic situation)

0
MAC1 MAC2 MAC3 MAC4 MAC5 MAC6

Source: Author's calculations

From the figure above we can see that the eigenvalues of the first two principal
components are significantly greater than zero. The shape of the line highlights the eigenvalues
that are located at the steep curve. The third PC (MAC3) starts the delimited trend line; this tells
us that MAC3 does not considerably augment the share of variation explained. Both approaches
lead us to the conclusion that two principal components (MAC1 and MAC2) should be used in
credit rating model to explain the macroeconomic situation.
Indicators of state governance level are heavily correlated inside the group (see Table 26
in the Appendix 1). Running PCA on six indicators (CORRC, GOVER, REGUL, POLIT, VOICE
and LAW) we get six principal components (GOV1, GOV2,…,GOV6). The resulting loading
vectors are provided in each column of the table below.

Table 11: Loading vectors of PCs (state governance level)

Variable GOV1 GOV2 GOV3 GOV4 GOV5 GOV6

CORRC 0.4367 -0.0652 -0.0452 -0.363 -0.382 -0.7248
GOVER 0.4297 -0.2073 -0.2356 -0.3502 -0.3825 0.6693
REGUL 0.4302 -0.1362 -0.2478 -0.2132 0.8292 -0.0433
POLIT 0.3809 -0.5893 0.5025 0.5048 -0.0142 0.0058
VOICE 0.3977 0.4022 -0.4886 0.6528 -0.12 -0.0297
LAW 0.3691 0.6521 0.6243 -0.1385 0.0746 0.1549
Source: Author’s calculations

The first principal component (GOV1) has large association with all the variables
responsible for state governance. It is associated with CORRC (44%), GOVER (43%), REGUL
(43%), POLIT (38%), VOICE (40%) and LAW (37%). As we can see, GOV1 has approximately
equal coefficients of the same sign for these variables. The first component relates to the quality

140
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of the government body as a whole. The second principal component (GOV2) has large
association with POLIT (59%), VOICE (40%) and LAW (65%). This component measures the
respect for human rights inside the country, as this is primarily responsible for absence of
violence, freedom of speech and supremacy of law.
PVE explained by GOV1 is 83%, whereas the second principal component (GOV2)
explains only around 8% of the variation. The remaining 9% are explained by the other
components. The analysis of PVE depicts that it is enough to include only the first PC, as the
variance explained by the remaining ones is not significant.

Figure 7: PVE by each PC (state governance level)

100% 100%

80%

90%
60%

40%
80%

20%

0% 70%
GOV1 GOV2 GOV3 GOV4 GOV5 GOV6

Proportion Cumulative

Source: Author's calculations

From the figure below it can be clearly seen that the eigenvalue of the first principal
component is greater than one, it equals 5, hence it certainly should be retained. The eigenvalue
of the second PC is less than 1 and equals 0.4, so we should not include it in the model.
Figure 8: Scree plot of eigenvalues for PCs (state governance level)
6

0
GOV1 GOV2 GOV3 GOV4 GOV5 GOV6
Source: Author's calculations

141
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

According to the abovementioned analysis, we try to include the first two principal
components for macroeconomic situation (MAC1 and MAC2) and one for state governance level
(GOV1) and examine their significance in credit rating forecasting and influence on the overall
quality of the model. The principal components in every group of features are built in such a way
that the correlation between them is zero. This helps us to eliminate the multicollinearity problem.
The significance and economic interpretation of each principal component in credit rating
forecasting are provided in the next section.

4.3. Uniform credit rating model

Rating agencies use different methodologies to assign the credit rating. While mapping
the credit ratings we showed that one bank may get different ratings from agencies. This tells us
that the factors and indicators taken into account by СRAs may vary. In real life, credit rating
may be used as an estimator of credit risk and the agency that provides this rating is not always
important. This brings us to the idea of joint modeling of different credit ratings. We will take
the minimum value of single scale numeric credit rating (numbers were assigned according to
the mapping provided above). In total, we have 30 different credit values from the lowest rating
(1) to the highest one (21.5). We decided to take the minimum value as this would help us to
estimate credit risk correctly. Taking the average value would result in numerous non-integer
numerical ratings.
Different external factors influence the behavior of any CRA, and the rating assigned may
not always objectively reflect the credit risk and bank’s financial strength. By taking the
minimum rating issued by any of the three rating agencies, we try to minimize the subjectivity
and the dependent variable in the model becomes more impartial and fairer. We attempt at
smoothing out the external factors affecting the assessment of credit risk provided by CRA.
The procedure of choosing the explanatory variables remains the same as explained
above. However, in this section we try to increase the forecasting power of traditional credit
rating models and include the principal components for macro and state governance factors,
which help us to control for external environment correctly. Also, a set of interaction variables
is used in uniform credit risk modeling. We build two types of models: class model and grade
model. The classes and numerical grades from are provided in Table 30 (see Appendix 1).

The first model specification is constructed including financial variables only:

yit  1CAPit + 2 NONL/Lit +3OE/REVit + 4ROEit +5LA/NLAit   6 EFFit 
(22)
  7 NIit +8LOGAit  it .

142
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The second model specification is constructed including financial, interaction terms and
principal components. For the class model the regression estimated is:
yit  1CAPit + 2 NONL/Lit +3OE/REVit + 4 ROEit +5LA/NLAit   6 EFFit 
  7 NIit +8LOGAit2  9CUR*Y2008it  10CUR*Y2008it +
(23)
 11CUR*Y2009it  12CUR*Y2010it  13LEV*CISit  14 LEV*EUit 
 15MAC1it  16 MAC2it  17GOV1it  it .

For the grade model additional financial interaction terms are included:
yit  1CAPit + 2 NONL/Lit +3OE/REVit + 4 ROEit +5LA/NLAit   6 EFFit 
  7 NIit +8LOGAit2  9CUR*Y2008it  10CUR*Y2008it +
 11CUR*Y2009it  12CUR*Y2010it  13LEV*CISit  14LEV*EUit  (24)
 15PROV/G*IE/DEit  16 ALL/NONL*OE/OIit  17 ROA*AGEit
 18MAC1it  19 MAC2it   20GOV1it  it .

 
where yit is the minimum class of ratings for a class model yit  min classitm , classitsp , classitf or

 
the minimum credit rating grade yit  min gradeitm , gradeitsp , gradeitf assigned by a CRA: Moody’s

(m), S&P (sp) or Fitch (f).

143
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 12: Uniform ordered logit credit rating model

Class model Grade model

Indicators class(1) class(2) grade(1) grade(2)

CAP 0.016* 0.016* 0.016* 0.016*

(0.002) (0.002) (0.001) (0.001)
NONL/L -0.077*** -0.069*** -0.083*** -0.069***
(0.006) (0.007) (0.005) (0.005)
OE/REV -0.004*** -0.003*** -0.003*** -0.003***
(0.001) (0.001) (0.0004) (0.0004)
ROE 0.056 0.110* 0.125*** 0.119**
(0.060) (0.062) (0.046) (0.048)
LA/NLA 0.000*** 0.000*** 0.003** 0.003**
(0.000) (0.000) (0.001) (0.001)
EFF -0.032*** -0.031*** -0.017*** -0.017***
(0.006) (0.006) (0.004) (0.004)
NI 0.007 0.006 0.007** 0.008**
(0.005) (0.005) (0.003) (0.003)
LOGA 0.535*** 0.247***
(0.087) (0.057)
LOGA2 0.032*** 0.081***
(0.004) (0.011)
(CUR)*(Y2008) 0.018*** 0.023**
(0.004) (0.011)
(CUR)*(Y2009) 0.078*** 0.016***
(0.021) (0.004)
(CUR)*(Y2010) 0.033*** 0.019***
(0.011) (0.007)
(LEV)*(CIS) 0.026*** 0.026***
(0.008) (0.008)
(LEV)*(EU) 0.235*** 0.0145*
(0.047) (0.0077)
(PROV/G)*(IE/DE) 0.00002*
(0.00001)
(ALL/NONL)*(OE/OI) -0.016***
(0.003)
(ROA)*(AGE) 0.199*
(0.673)
MAC1 0.874*** 0.886***
(0.074) (0.052)
MAC2 -0.316*** -0.953***
(0.059) (0.121)
GOV1 2.863*** 1.603***
(0.128) (0.092)
Sample size 12,350 12,091 12,350 12,174
AIC 12369.24 11567.20 39330.51 38009.13
BIC 12480.46 11737.41 39619.95 38394.29
Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1
Source: Author’s calculations

144
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Models class(2) and grade(2) are the final model specifications which are built by us for
credit risk valuation. We provide several intermediate steps with inclusion of financial indicators
into the grade model in Table 31 (see Appendix 1). We include in the models the indicators which
showed their significance and intuitive economic interpretation in the previous chapter when the
models for different CRAs were built. Interpretation of this set of variables remains unchanged
due to the same signs, hence we will focus on new findings only. We notice that the chosen
financial indicators remain significant in both models, hence, their influence on banks’ credit risk
is validated.
Additionally, we include OE/REV to capture the management quality of a bank.
Revenues of banks usually consists of loan and deposit operations. Low OE/REV ratio depicts
low operational costs compared to the earnings of the bank. This indicates good level of
management and positively influences on assigned credit rating. We also substitute the previous
proxy for earnings (ROA) on a similar one (ROE). It is significant and has a positive sign in the
both class(2) and grade(2) models, hence banks with higher ROE tend to have higher credit
ratings as earnings directly stands for financial performance. In order to explain the change in
bank’s earnings, we decided to include NI index. The banks with stable net income growth are
assigned higher ratings, according to the grade(2) model, where this indicator is significant at 5%
significance level. NI is calculated as the net income change comparing to the previous quarter
(in %). NI may serve as an indicator of bank’s overall sustainability. If bank’s profit quarterly
increases (NI is positive), the bank is considered to be in good position and its rating will be
higher. Notice than NI is not significant in class model.
In the first model specification we decided to include LOGA as a measure of bank’s size
because this variable was highly significant for all CRAs (see models m(2), sp(2) and f(2)). In
the second model specification we included the squared natural logarithm of bank’s total assets
LOGA2 to capture the diminishing effect of LOGA on bank’s financial performance. It means
that when a bank reaches some size value, the influence of its size becomes smaller. This is
intuitive as small banks during their operational development grow in size and this has a strong
effect on their sustainability. However, the change in size of large banks is not very dramatic.
The grade model has more credit rating grades to be forecasted (the scale is broader),
hence, for accurate training of the model, we include several financial cross-terms. Explanatory
term (PROV/G)*(IE/DE) shows the interdependence of asset and management quality. The
banks with good management quality may afford low rate on deposits (low IE). This will attract
only conscientious clients (due to honest business conduct) and banks’ asset quality increases.
Because of good asset quality, the bank may not have high level of provision for loan losses and
this decreases PROV/G ratio. (PROV/G)*(IE/DE) has a positive influence on credit rating of a

145
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

bank. Another cross-product (ALL/NONL)*(OE/OI) shows the interdependence between

management quality and efficiency. Banks with high levels of allowance for credit losses (with
better management quality) typically have better efficiency as reserves may help banks in
difficult situations. Better efficiency suggests lower operating expenses (OE/OI ratio is smaller)
and this causes an increase in credit rating. Interaction variable (ROA)*(AGE) captures the
dependence of credit rating from bank’s age along with its profitability. Older banks are more
experienced, and their earnings are higher because of high customer base, which positively
impacts on their credit ratings.
In the class(2) and grade(2) models we also include some cross terms with dummy
variables. We can see that the current ratio (CUR) has a positive influence on credit rating as it
captures the bank’s ability to pay its obligations. CUR positively affects the credit rating in 2008,
2009 and 2010 year. Both interaction variables (LEV)*(CIS) and (LEV)*(EU) are significant
and stand with positive signs. LEV positively influences the credit rating of banks from
developed and emerging countries, which was also demonstrated on the models for different
CRAs in the previous section.
The principal components for macroeconomic situation (MAC1 and MAC2) and for the
state governance level (GOV1) are significant at 1% significant level, which depicts the correct
choice of the PCs. GOV1 explains more than 80% of the variation in a group of the state
governance indicators. High level of significance of this PC supports the second hypothesis. The
government of a country should not only monitor the conditions of the financial system, but also
to carry out transparent and legitimate politics. Favorable conditions inside a country benefit
banks in several ways. Firstly, the credit standing of country’s citizens is better, they save more
and put money on bank deposits. Secondly, the possibility of external shocks (political, social,
etc.) is limited, and banks will not go bankrupt because of unexpected bank run. Thirdly, the
CRAs favor the banks from transparent and progressing countries. Higher credit ratings allow
banks to attract more clients and investors.

4.3.1. Marginal effects calculation

The estimation results of ordered logit model help us to understand the sign of influence
of a particular covariate on credit risk. The coefficients themselves do no show the power of the
influence. To evaluate it, we need to calculate the marginal effects. We will do it for the class (2)
model only, as in the rating grade (2) model the number of cut points and the values takes by
dependent variable is too large. The interpretation of marginal effects for class(2) model is
provided below.

146
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The marginal effect of an increase in some continuous covariate xc on the probability of

selecting some arbitrary rating j is given by:

MPE jc ( x) 
 Pr( yit  j ) 
xc    
 f m j 1  xit  f m j  xit  c ,
  (25)

where function f denotes the probability distribution function (pdf). For standard normal
distribution it equals to:

1  2 
f ()  ()  exp    . (26)
2  2 
 
When ordered logit models are built the pdf of logistic distribution should be assumed:
exp()
f ()  ()  . (27)
1  exp()2
(Note: notations are the same as in Chapter 3)

The abovementioned formulas are used for continuous variables only. If an explanatory
variable xb is binary (e.g. dummy variable) the marginal effect is calculated as:

MPE jb ( x)  Pr  yit  j | xb  1  Pr  yit  j | xb  0  . (28)

The marginal effects in mean for the significant indicators from model class(2) are
presented in the Table 32 (see Appendix 1). The marginal effects are calculated for the probability
to get each rating class. Columns Pr(j) indicates the marginal effect of a covariate on probability
of assigning class j. We can see that the principal components included in the model have very
high marginal effects, which was expected since they were constructed to explain as much
variation as possible. For us it is interesting to analyze and compare the marginal effects of
financial variables included in the model.
The figure below illustrates only significant marginal effects (at any reasonable
significance level). All the variables have insignificant marginal effects of influence on
probability to get class 4 (Pr(4)), this is clearly depicted in Table 32; that is why we do not show
these effects on the figure and do not analyze them. Also, marginal effects for Pr(1), Pr(2) and
Pr(6) are insignificant for ROE, we do not analyze them as well. This observation implies the
importance of marginal effects calculation. Indicator ROE is significant in model class(2) at 10%
significance level, but the significance of this variable is doubtful for several rating classes
included in the model. This can be understood only by analyzing marginal effects, which is so
rarely met in academic papers.

147
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 9: Marginal effects of financial indicators, class(2) model

Pr(1) Pr(2) Pr(3) Pr(5) Pr(6)

0,28%

0,21%
0,18%
0,15%
0,13%
0,09%
0,07%
0,1%
0,02% 0,03%
0,01% 0,007% 0,01%

-0,005% -0,01% -0,005%

-0,02%
-0,04%
-0,05%
-0,07% -0,07%

-0,13% -0,11%
-0,17%

-0,248%

-0,38%

CAP NONL/L OE/REV ROE LA/NLA EFF LOGA2

Source: Author's calculations

Figure 9 illustrates that the largest marginal effects are caused by indicator NONL/L
(proxy for asset quality), hence in our sample the credit risk is mostly associated with poor asset
quality of a bank. The internal management of a bank should take into account the share of bad
loans. The great share of bad loans may not harm banks much during the ordinary times, but in
crises this may lead to failure of banks.
The size of a bank is also influential for credit risk. It is a good idea to include the squared
logarithm of total assets LOGA2 as this variable is more descriptive than LOGA. Larger banks
tend to obtain higher credit ratings as they are more sustainable to credit risk.
Other important indicators are: CAP, EFF and ROE (only for two classes). Proxies for
management quality (OE/REV) and liquidity (LA/NLA) have significant but very weak marginal
effects.
To evaluate the influence of Tier-1 leverage ratio for banks in developed and emerging
countries on credit ratings and credit risk, we have included two cross-terms in the model. The
figure below demonstrates the comparison of the marginal effects depending on the geographical
location of banks.

148
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 10: Marginal effect of LEV in emerging and developed countries

Pr(1) Pr(2) Pr(3) Pr(5) Pr(6)
1,30%

0,38%

0,15%
0,04%

-0,01% -0,06%
-0,07% -0,11%

-0,50%

-0,96%
(LEV)*(CIS) (LEV)*(EU)
Source: Author's calculations

For both emerging and developed countries the marginal effects of leverage on rating
classes are significant. However, it can be clearly seen that for banks from developed countries
(EU) this ratio affects the credit ratings stronger than for banks from emerging countries (CIS).
Tier 1 capital may be classified as the source of banks’ financing. In EU there are more banks
that are traded on financial markets. In developed countries there are fewer private banks because
of advanced market structure. For banks that go public via an Initial Pricing Offering (IPO), there
occurs a periodic revaluation of their accounting value to the market value, hence CRAs pay a
greater attention to LEV in developed countries than in emerging ones. In EU, this indicator
shows the market attitude to banks.

Figure 11: Marginal effect of CUR in 2008, 2009 and 2010 years

Pr(1) Pr(2) Pr(3) Pr(5) Pr(6)

0,43%

0,18%
0,13%
0,10%
0,03% 0,05%

-0,01% -0,01%
-0,02% -0,04%
-0,07% -0,08%
-0,14%
-0,17%

-0,32%

(CUR)(Y2008) (CUR)(Y2009) (CUR)*(Y2010)

Source: Author's calculations

149
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 11 shows that CUR as a proxy of liquidity has the largest marginal effects in 2009
year. We fail to reject the third hypothesis. During the financial crisis in 2009 it was important
for banks to keep high liquidity in order to survive and decrease the credit risk. Before crisis and
after crisis CUR is not so influential on credit ratings and credit risk. In 2010 the marginal effects
of CUR decreases comparing to 2009 year, but do not return to pre-crisis values. During financial
crises banks may face liquidity management problems even if they have adequate level of capital
and earnings (Yeddou, 2020). Hence, for banks it is crucial to have enough liquid assets in order
to withstand the risks. We have managed to show that liquidity proxy is significant in different
model specifications, but only with a help of marginal effects calculations we are able to depict
the power of influence.

4.3.2. Arising econometric issues

To get a representative sample we have randomly chosen 500 banks from each geographic
group using the Python code (see code 1 in Appendix 3). This was done in order to get banks
from every region, otherwise we could get a not representative distribution. A lot of banks
dropped out from this selection due to the lack of rating, but this occurred randomly as well.
Also, it is important to assess the size and the age of banks in the sample. Table 13 shows that
the sample contains banks of different age and size.

Table 13: Descriptive statistics, the sample

Observations Mean Min Max
Total Assets 23 765 $185 119mm $10.6mm $3 867 171mm
Total Equity 23 765 $12 518mm $0.063mm $308 275mm
Age 21 952 60 years 6 years 134 years
Source: Author’s calculations

Total Assets and Total Equity are used as proxies for the bank’s size, it can be seen from
the table that the size and the age of banks vary in the obtained sample. Hence, the sample is
random and represents banks of different size and of different age. Our sample is an unbalanced
panel data and we have decided not to cut it down to the balanced panel data for not to causing
possible sample bias.
There are three methods of panel data estimation: pooled OLS method (POLS), random-
effects method (RE) and fixed-effects method (FE). Firstly, RE and FE models are compared
using Durbin-Wu-Hausman test. The hypotheses of the test are:
H0: individual unobserved effects are uncorrelated with included regressors
H1: individual unobserved effects are correlated with included regressors

150
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

There is no standard command to estimate ordered logit model FE in Stata (Gayle &
Lambert, 2018), however some academic articles present the programs, which can be used. We
have managed to find one and use in for FE model estimation (see code 2 in Appendix 3). The
resulting p-values are 0.52 for class(2) model and 0.46 for grade(2) model. We fail to reject the
null hypothesis and conclude that individual effects are exogenous, RE models are more efficient
and are therefore preferred.
RE model is needed to be compared to POLS model. There is no common way to estimate
POLS ordered logit model in Stata. It is sometimes recommended to assess the LR test. We obtain
chibar(1)=22577. Our model specification has p-value =0.000, we reject the null of no
autocorrelation and conclude that RE model is preferred over POLS due to existence of
unobserved effects and autocorrelations.

4.4. In-sample classification accuracy

In previous chapters we have provided not only the final models, but also the results of
intermediate estimations with different sets of variables. The influence of particular features on
model quality is assessed by information criteria: Akaike’s information criterion (AIC) and
Schwarz’s Bayesian information criterion (BIC). They are provided in each table with
estimations to compare the quality of obtained models. AIC and BIC are calculated as:
AIC  2log L  2k ; (29)
BIC  2log L  k ln N , (30)
where logL is the maximized log-likelihood of the estimated model; k is the number of parameters
included in the model; N is the sample size.
To evaluate the prediction quality of the obtained models we use the Wald test and the
distribution of forecast errors. From the analysis of AIC and BIC we conclude that the most
accurate models are: model m(2) for Moody’s ratings, sp(2) for S&P ratings, f(2) for Fitch
ratings, class(2) for rating classes and grade(2) for rating grades. Wald test is aimed at assessing
the significance of the model as a whole. We test the null hypothesis of zero coefficients against
the alternative that at least one coefficient is different from zero. The chi-squared values
generated by the Wald test are provided for each final model in the table below.

Table 14: Wald test statistics

m(2) sp(2) f(2) class(2) grade(2)
Chi-squared 866.4 302.85 394.39 723.38 1323.49
p-value 0.00 0.00 0.00 0.00 0.00
Source: Author’s calculations

151
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

According to the p-values presented in Table 14 we able to reject the null hypothesis and
conclude that each model is statistically significant.
Besides, the adequacy of the econometric models is determined by the goodness-of-fit.
We will calculate the deviations from the actual rating as a difference of actual rating (
Ractual ) and a forecasted one ( Rmodel ) :

  Ractual  Rmodel . (31)

As statistical measures we look precisely at the probability of correct forecasting P    0  ,

deviation of no more than one numerical rating P |  | 1 and the probability of no more than

two numerical ratings deviation P |  | 2  . This helps us to evaluate the forecasting performance

of the models inside the sample.

First of all, we verify the predictive power of the obtained models for different CRAs.
We compare the first modification of the models with the second modification (macro and state
governance indexes included). The information criteria (AIC,BIC) are provided in the Table 9. It
is seen that the second specification of the models is preferable for all rating agencies. Model
training with a help of macroeconomic factors and indicators of state governance increases the
forecasting power of each model. Hence, macroeconomic situation and overall administration
level inside the bank’s country of residence are significant in credit rating estimation. Now let us
refer to the distribution of forecast errors.

Table 15: Forecast errors of the models for three CRAs

Model m(2) Model sp(2) Model f(2)
∆=0 0.13 0.14 0.20
|∆|≤1 0.42 0.45 0.56
|∆|≤2 0.66 0.71 0.74
Source: Author’s calculations

Probability of observing the credit rating being equal to the actual one P    0  is 13%,

14% and 20% for m(2), sp(2) and f(2) respectively. The distribution of deviations of these models
is not very identical, however, we can see that the difference of probabilities of models mistake
of no more than two numerical scores P |  | 2  is not great. The most accurate model is f(2)

which is built to forecast the credit ratings of Fitch agency.

152
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 12: Distribution of forecast errors of models for three CRAs

0,25 m(2) sp(2) f(2)

0,20

0,15

0,10

0,05

Value of ∆
0,00
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author's calculations

The analysis demonstrates that there is a negative asymmetry with longer left tail. The
distribution is left skewed. The ratings of obtained models are less conservative than the actual
ones. Skewness of the models may be explained by the influence of banks from developed
countries in the randomized sample. The similar results were obtained in the papers
(Karminsky& Sosurko,2010; Karminsky& Khromova,2016). In order to address this issue more
country-specific variables are needed to be included as this would allow the model to distinguish
the banks from emerging countries. The error distributions of the models bring us to the idea of
pooling the ratings of CRAs into one uniform model. Looking at Figure 12, we anticipate that
credit ratings pooling would allow us to get model errors close to Normal, as the distribution
curves for CRAs have different modes and different right tails.
After aggregating the credit ratings into one database, we estimate the uniform models
class(2) and grade(2) and class(1) and grade(1). At each step of modelling the information criteria
decreases which implies that adding different interaction variables increases the forecasting
power of the credit rating model. Inclusion of principal components for macroeconomic and state
governance indexes increases the forecasting power of the models dramatically. We will compare
the distribution of the forecast errors of the first pooling models specifications and the final ones
to measure the effect occurred from cross-term variables and principal components. We start with
the comparison of the class models.

Table 16: Forecast errors of class models

Model class(1) Model class(2)
∆=0 0.28 0.38
|∆|≤1 0.70 0.88
|∆|≤2 0.93 0.99
Source: Author’s calculations

153
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 13: Distribution of forecast errors of class models

0,4 38%
class(1) class(2)
0,35
0,3 28% 28% 29%

0,25
21%
0,2
0,15 12%
13%
10%
0,1
7%
6%
0,05 4%
1%
1% 0,4% Value of ∆
0
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author's calculations

The final model specification class(2) becomes more accurate than class(1), hence the
interaction terms and principal components help us to increase the quality of the obtained model.
Class(2) predicts correct rating classes in 38% cases and makes the mistake of no more than two
classes in 99% cases. The final model specification class(2) can be efficiently used to forecast
rating classes of international banks.
Now let us refer to the model which forecasts rating grades. We compare the first model
specification grade(1) and the final one grade(2).

Table 17: Forecast errors of grade models

Model grade(1) Model grade(2)
∆=0 0.17 0.23
|∆|≤1 0.42 0.56
|∆|≤2 0.58 0.78
Source: Author’s calculations

Figure 14: Distribution of forecast errors of grade models

0,25
23% grade(1) grade(2)

0,2
17% 17%
16%
0,15
12% 13%
12%
10% 10% 10%
0,1

6% 6%
5% 5% 5%
0,05 3%
4%
3%
3% 2% 3%
2% 2%
0,5% 1% 1%
1%
1%
2%
1%
1% 1% 1%
1% Value of ∆
0
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author's calculations

154
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

We can see that model grade(1) predicts correctly only in 17% cases, while the share of
correct forecasts of model grade(2) is 23%. The errors distribution of grade(1) is disproportional,
we can see significant number of errors in the left tail, so the skewness remains similar to the
models of different CRAs. With the help of various explanatory variables included in the model
grade(2) we are able to smooth the distribution of forecast errors. The final grade model has the
highest forecasting power. Our analysis provides us with the model with almost 80% of rating
grades predictions with an error less than two rating grades. Only 12% of forecasted ratings by
grade(2) deviate from the actual ones by more than three rating grades. The distribution tails are
smoothed, but slight overestimation of credit ratings still presents in the ordered logistic model.

4.5. Robustness check of the obtained results

4.5.1. Sensitivity to changing the type of the model

The reported estimations are the results of ordered logit model; however, one more
similar method is available- ordered probit model. Ordered probit model is used for credit rating
modeling (Karminsky & Khromova, 2018). The results of our estimation may differ if we change
the model specification from logit to probit, but overall outcomes and economic interpretations
should remain unchanged. We estimate class(2) and grade(2) model using the ordered probit
approach. The results of ordered probit estimation with the same set of explanatory variables are
provided in Table 33 (see Appendix 1). Some variables become insignificant, which was
expected. Nevertheless, the majority of explanatory variables remains significant in the model.
The principal components for the state governance level are significant, hence hypothesis 2 is
not rejected even if we change the type of the model. It can be seen that the signs of some features
have changed compared to the ordered logit model, but such variables are insignificant in ordered
probit model.
Additionally, we calculate the marginal effects of CUR variable in 2008, 2009 and 2010
years for class(2) model estimated by ordered probit. The results of estimation are provided in
the table below.

Table 18: Marginal effects of CUR in class(2) ordered probit model

155
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Pr(1) Pr(2) Pr(3) Pr(4) Pr(5) Pr(6)

(CUR)*(Y2008) -4.56E-08 -0.0000558 -0.001451 9.52E-06 0.0014782 0.0000192
(2.10E-05) (0.0000156) (0.0003593) (0.0001114) (0.000366) (5.38E-06)
(CUR)*(Y2009) -1.20E-07 -0.0001469 -0.0038232 0.0000251 0.0038947 0.0000505
(6.38E-08) (0.0000672) (0.0016797) (0.0002937) (0.0017108) (0.0000232)
(CUR)*(Y2010) -5.98E-08 -0.0000731 -0.0019009 0.0000125 0.0019365 0.0000251
(0.000022) (0.0000336) (0.0008407) (0.000146) (0.0008565) (0.0000116)
- insignificant marginal effects
Standard errors in parentheses. Source: Author’s calculations

The tendency remains the same as for the ordered logit model. We can see that the highest
marginal effects of CUR proxy are in 2009 year, hence the third hypothesis is not rejected even
if we change the type of the model.
Distribution of in-sample errors of the ordered probit estimation are provided in the
Appendix 2 (see Figure 26 for class(2) and Figure 27 for grade(2)). We can see that the in-sample
forecasting power of ordered probit models remains quite strong, although it is worse than the
original ordered logit model. The distribution of errors still has a negative tail. We can conclude
that both final credit rating models are robust to change in the estimation method.

4.5.2. Sensitivity to changing the credit rating classification

Models class(2) and grade(2) are estimated on the minimum credit rating classes and
minimum credit rating grades assigned by 3 CRAs respectively. This is done for the models to
predict the credit risk correctly. However, it is possible to estimate the ordered logit credit rating
models taking maximum credit ratings. This estimation may be of interest for internal
management of banks in the situations of assessing their performance and effectiveness. We
expect that the economic interpretation of the models would remain unchanged even if we change
the classification of dependent variable (from minimum to maximum ratings by any of the
CRAs). We estimate equations (23) and (24), but now the dependent variable is


yit  max gradeitm , gradeitsp , gradeitf  for grade(2) model and 
yit  max classitm , classitsp , classitf  for
class(2) model. The results of ordered logit estimation with maximum credit rating are depicted
in Table 33 (See Appendix 1) in the last columns. We can see that principal component for the
governance level is significant, so we still fail to reject the second hypothesis. The distribution
of errors is depicted in the Appendix 2 (see Figure 28 and Figure 29). We can see that the grade
model is not conservative enough comparing with CRAs, but the right tail is more distinct. Again,
we estimate the marginal effects of CUR in 2008, 2009 and 2010 years with new classification

156
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of credit ratings and check their sensitivity. The obtained marginal effects are calculated in the
table below.

Table 19: Marginal effects of CUR in class(2) with maximum rating classes
Pr(1) Pr(2) Pr(3) Pr(4) Pr(5) Pr(6)
(CUR)*(Y2008) -0.0000531 -0.000493 -0.0004044 -0.0008539 0.0007933 0.0010111
(0.0000182) (0.0001797) (0.000174) (0.0004883) (0.0002103) (0.0002423)
(CUR)*(Y2009) -0.0000608 -0.0005642 -0.0004628 -0.0009772 0.0009079 0.0011571
(0.0000243) (0.0001218) (0.000231) (0.000701) (0.0004118) (0.0005269)
(CUR)*(Y2010) -0.0000303 -0.0002815 -0.0002309 -0.0004875 0.0004529 0.0005772
(0.000012) (0.0001318) (0.0001098) (0.00047) (0.0002111) (0.0002381)
- insignificant marginal effects
Standard errors in parentheses. Source: Author’s calculations

We can see that the estimation results remain consistent even if we change the credit
ratings specification. The third hypothesis is not rejected even if we change the classification of
the dependent variable, the proxy for liquidity (CUR) has the marginal effects in 2009 year. The
forecasting power of the models decreases, but still helps to examine bank’s credit risk. Thus,
the obtained credit rating models are robust to the change of a dependent variable.

4.5.3. Out-of-sample forecasting power

The number of credit ratings assigned by CRAs is limited and only a part of banks has
the credit rating. Credit rating modeling is aimed at rating forecasting primarily for banks without
a credit rating available. Our final unified models may assign credit classes and rating scores to
banks on the basis of public information. Out-of-sample forecasting may yield information on
the effectiveness of the estimated models and check their robustness. We will apply two out-of-
sample tests for the models: testing the results for different time periods and testing the results
for different banks.
For the out-of-sample forecast for different time periods we divide the database on two
parts. Part A contains the data from the first quarter of 2007 to the fourth quarter of 2016 (40
time segments). Part B contains the data from the first quarter of 2017 to the first quarter of 2019
(9 time segments). We estimate the model for part A and then use the resulting coefficients to
predict the credit ratings for part B. As the time periods from part B are excluded from modeling,
their estimation occurs out-of-sample. Below we present the distribution of the errors of the
forecasted values.
Table 20: Forecast errors of model class(2) (out-of-sample)

157
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

class(2) (out-of-sample)
∆=0 0.36
|∆|≤1 0.85
|∆|≤2 0.97
Source: Author’s calculations

Figure 15: Distribution of forecast errors of model class(2) (out-of-sample)

0,4 36%
0,35
0,3
25% 24%
0,25
0,2
0,15
10%
0,1
0,05 2% 2% 1%
Value of ∆
0
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author's calculations

The results depict that the class model is sustainable to the sample change. It predicts
correct rating classes in 36% cases and the probability of getting an error more than two classes
is only 3%.

Table 21: Forecast errors of model grade(2) (out-of-sample)

grade(2) (out-of-sample)
∆=0 0.20
|∆|≤1 0.54
|∆|≤2 0.73
Source: Author’s calculations
Figure 16: Distribution of forecast errors of model grade(2) (out-of-sample)
0,25
20%
0,2 18%
16%
0,15
11%
0,1 8%
6%
5%
0,05 4%
2% 3%
0,7% 1% 1% 1% 1% 1% 0,4% Value of ∆
0
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11
Source: Author's calculations

158
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Our analysis shows that the probability of obtaining the correct credit rating grade (the
same as the one assigned by the credit rating) is 20%. About 73% of the estimated rating deviate
from an actual one by no more than 2 numerical points. What is more, only 16% of the estimated
values have a deviation of more than 3 rating grades. The obtained results bring us to the
conclusion that the credit rating model grade (2) can be used in practice to forecast the banks’
rating in order to assess the credit risk.
Another way to test the obtained credit rating models for accuracy prediction outside the
sample is to divide the banks on two groups. Group A contains the randomly chosen 428 banks
from the initial database. Group B contains the remaining 50 banks. We estimate the models on
the information of banks from group A, hence the prediction of credit ratings model for group B
occurs out-of-sample. We present the distribution of the errors of the forecasted values for class
and grade models in Appendix 2 (see Figure 30 and Figure 31). It is clearly seen that the share
of correct predictions decreases for both models comparing to the first our-of-sample test, which
was expected. However, we can see only the minor differences. The share of deviations of no
more than 2 points is 92% for class(2) model and 71% for grade(2) model. The models remain
accurate and can be used for credit rating prediction of banks out of the sample used for models
construction.
We have conducted four robustness checks and conclude that both credit rating models
are robust to the changes in the sample, in the method of estimation and in the dependent variable
classification. The economic interpretation of the models’ results remains unchanged; hence the
models are reliable.

159
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chapter 5: Calibration of credit ratings and PD on the example of Russian

banks
In this chapter we present the results of constructing a dynamic scale of credit rating
conversion to probability of default (PD). Investors are interested in a quantitative measure of
banks’ credit risk. This goal is achieved with a help of default frequencies estimation for each
group of credit rating grades. This scale is built by us on the basis of an extensive sample of
Russian banks and can be used by both investors and internal management in credit risk
assessment. Besides, while constructing the calibration scale, we notice several important
patterns and try to explain their possible reasons and origins.

5.1. Sample description of Russian banks

The list of Russian banks used for the analysis is gathered from the “Mobile” agency. For
the purpose of correct credit risk estimation, we exclude all the banks with more than 50%
government share. As a source of credit ratings, we use the Cbonds platform, which helps us to
obtain the credit ratings from the first quarter of 2007 to the first quarter of 2019. In total, the
sample consists of 338 banks which have credit ratings at any of the quarters. The activity of
international CRAs is limited in Russia, hence we have also decided to include the credit ratings
assigned by national CRAs. Сbonds platform provides us with both credit ratings of international
CRAs (Moody’s, S&P and Fitch Ratings) and national CRAs (Rus-Rating, RAEX, AK&M, NRA
and Ria-Rating). Inclusion of national CRAs’ assessment allows us to increase the database and
to represent the situation on the Russian market more accurately. In order to maintain
consistency, we include the long run (LR) credit ratings. Statistics on defaulted banks is extracted
from [Link] platform. By “default” in our paper we use the definition provided by Karminsky
& Kostrov (2014). The characteristics of bank’s default are: the level of bank’s capital sufficiency
is lower than 2%; the value of bank's resources drops lower than the allowed minimum;
foundation by the Deposit Insurance Agency (DIA) or inability of banks to meet its obligations
to creditors. There are 95 banks that had been assigned a credit rating and went bankrupt from
the first quarter of 2007 to the first quarter of 2019. The credit ratings are mapped on the base
scale. International scale is provided in Table 30 and national scale is provided in Table 34 in
Appendix 1. Hence, the credit ratings forecasted with a help of models from Chapter 4 and those
used for PD calibration are mapped on a base unified scale.

160
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.2. Creating a scale of credit rating conversion to PD

The credit rating is used by investors primarily for credit risk estimation. Credit risk is
associated with a probability of default. Investors refer to PD as a valuation of possible losses
from an investment decision. PD is summarized on a numerical scale; hence investors can use
this value to assess the expected losses (EL) in a dollar value. The assessment of credit risk by
CRAs allows to judge about the creditworthiness and financial stability of banks especially in
emerging countries. In emerging countries, the amount of services provided by CRAs has
increases in recent times. The dynamics of rating services in Russia has doubled from 2003 to
2007. The largest increase in the number of given credit ratings has occurred for the Moody’s
agency (Karminsky & Peresetsky, 2007). The authors found out that 70% of Russian banks with
credit ratings were the clients of Moody’s agency. For ordinary consumers of financial services,
the symbolic figures have probably little to say (especially in emerging countries). The
econometric models which can be used for interpretation of credit ratings with a help of PD are
used in articles (Pomasanov &Vlasov, 2008; Tasche, 2013; Rudakova & Ipatyev, 2015;
Karminsky & Khromova, 2018; Volk, 2012). The tables with credit ratings and implied PDs are
provided by CRAs themselves: S&P (2019 Annual Corporate Default and Rating Transition
Study), Moody’s (Corporate Default and Recovery Rates, 1920-2008) and Fitch (2019 Transition
and Default Studies). However, several limitations are faced while applying these scales to
Russian banks:
o Firstly, CRAs do not provide the corresponding scales of credit rating conversion for different
countries and for different geographic groups. The data used by the CRAs to prepare these
calculations mostly include their home country (the USA) and the countries of major
shareholders (developed countries such as Canada and the UK). Russian banks are not
representative in such databases;
o Secondly, the values of annual historical default frequencies (estimates of default
probabilities) for various credit ratings are calculated by each CRA empirically (based on
default statistics of banks with credit ratings of a particular CRA), which leads to an
inadequate comparison of the level of creditworthiness for the same rating class in different
time periods;
o Thirdly, the scales provided by each CRA are not dynamic in nature, i.e. they provide only
annual frequencies. An investor may not evaluate the possible losses which can occur in the
short run (in a month/quarter after the investment is made).
Taking into account all the above-mentioned limitations, we aim to construct a uniform
dynamic scale of the credit rating score conversion to PD. The credit rating scores used for
calibration is calculated according to the base scale obtained after the credit rating mapping in

161
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chapter 3. To prepare a scale of credit rating score and PD compliance we take the following
steps.
Step 1: we calculate the matrix for each credit rating score which shows the default
frequency for the banks which were assigned a particular credit rating in each of the available
time periods. To estimate the default frequencies, we create R code (see code 3 in Appendix 3),
which helps us to calculate the following matrices for each rating score. Each cell in a matrix
represents the default frequency (DF) which is calculated as:

Defaultl(,rq)
DFl(,qr )  , (32)
Banksl( r )

where r is a rating score;

l is the time quarter of a credit rating assignment, l=(1,2,…,48);
q is the time quarter of a bank’s default, q=(2,3,…, 49);

Default is the number of bankrupt banks, Defaultl(,rq) is the number of banks that got credit rating

r in period l and defaulted in period q;

Banks is the total number of banks, Banksl( r ) is the total number of banks that got credit rating r

in period l.
Consider Table 35 and Table 36 (see Appendix 1) which present the default frequencies
for credit scores r =4.5 and r =7 from the second quarter of 2014 to the first quarter of 2019. In
the columns, the periods of ratings assignment are presented (l=29,…,48). In the rows, the
periods of default with ratings r =4.5 and r =7 are shown (q=30,…,49).
Step 2: We do not fix the quarter when the credit rating r was assigned. We estimate the
period after which a bank goes bankrupt starting from the moment of rating assignment over the
entire time horizon. Thus, to estimate the PD we take the average values of the cells diagonally.
For example, the PD after one period is found as an average default frequency:

DF ( r ) (k  1)  Average DF1,2
(r ) (r )
; DF2,3 
(r )
;...; DF48,49 , (33)

where k is the number of quarters after which the bank went bankrupt.
Alternatively, the PD after two periods (quarters) is calculated as:

DF ( r ) (k  2)  Average DF1,3 
(r ) (r )
; DF2,4 (r )
;...; DF47,49 
. (34)

Hence, the default frequencies after  periods are found as:


DF ( r ) (k  )  Average DF1,(
r) (r ) (r )

1; DF2,  2 ;...; DF49 ,49 . (35)

Step 3: we summarize the obtained results for each rating score r presented in the sample
of Russian banks. The intermediate tables are constructed where we present the default

162
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

frequencies which are used to estimate the PD in k time periods (quarters) of a bank with credit
rating r.
Step 4: the estimated default frequencies for a set of rating grades are averaged for each
rating class. This is done for more logical representation of the obtained results and for keeping
an approximately equal number of bank-periods in each class. We divide the rating scores on 5
rating classes: BBB[14-12], BB[12.5-8.5], B[8-6], CCC[5.5-5] and C[4.5-1.5] based on
international scale. For example, the default frequency for a bank with credit rating from class
CCC[5.5-5] after  quarters is calculated as:

DF (5.5) (k  )  n(5.5)  ...  DF (5)  k     n(5)

DF (CCC )
(k  )  , (36)
n(5.5)  n(5)

where n ( r ) is the total number of bank-periods with rating r.

Table 22: Dynamic transmission scale of credit ratings and DF

BBB [14-12] BB [12.5-8.5] B [8-6] CCC [5.5-5] C [4.5-1]
1 quarter 0.40% 2.20% 5.30% 56.10%
2 quarters 0.30% 1.30% 2.50% 1.80%
3 quarters 0.30% 1.30% 2.50% 1.70%
4 quarter 0.40% 1.40% 2.30% 1.80%
Cum. DF in 1 year - 1.30% 6.30% 12.60% 61.40%
5 quarters 0.40% 1.40% 2.40% 1.80%
6 quarters 0.30% 1.40% 2.50% 1.30%
7 quarters 0.20% 1.60% 2.80% 1.20%
8 quarters 0.20% 1.60% 2.50% 1.30%
Cum. DF in 2 years - 2.40% 12.30% 22.90% 67.00%
9 quarters 0.20% 1.60% 2.60% 0.70%
10 quarters 0.30% 1.40% 2.70% 0.50%
11 quarters 0.30% 1.30% 2.40% 0.70%
12 quarters 0.30% 1.30% 2.40% 1.20%
Cum. DF in 3 years - 3.50% 18.00% 33.00% 70.00%
13 quarters 0.30% 1.20% 2.40% 1.80%
14 quarters 0.30% 1.20% 2.00% 2.70%
15 quarters 0.50% 0.90% 1.60% 2.80%
16 quarters 0.60% 0.80% 2.00% 0.60%
Cum. DF in 4 years - 5.20% 22.00% 41.00% 77.80%
17 quarters 0.90% 0.60% 1.90% 0.70%
18 quarters 1.10% 0.20% 1.40% 1.60%
19 quarters 1.60% 0.40% 1.20% 1.50%
20 quarters 1.80% 0.20% 1.1% 1.40%
Cum. DF in 5 years - 10.60% 23.40% 46.60% 83.00%
Source: Author’s calculations

163
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

With a help of Table 22 an economic agent can understand the quantitative estimate of
credit risk associated with a particular credit rating of Russian bank on the basis of default
frequencies analysis. Each credit rating can be easily converted into a clear PD estimated by
default frequency. The dynamic scale allows to evaluate the credit risk both annually and
quarterly. We present in the table the annual cumulative DFs and an incremental DF per quarter.
We notice that banks with newly assigned credit ratings tend to be less stable than banks with
unchanged ratings during a long period of time. Consider that a bank with credit rating B receives
credit rating CCC, which is not a very dramatic downgrade. The probability that this bank
defaults after one period is now 5.30%. The probability of default of the bank is significantly
higher than of a similar one that has received credit rating CCC a year ago (it has PD after one
quarter of 2.40%). This effect can be seen for every rating class presented in the Table, but it
increases as credit ratings become worse. The observed tendency does not allow us to reject the
fourth hypothesis formulated in Chapter 2 for the highly speculative rating grades (starting with
class B[8-6]).

Figure 17: Comparison of quarterly default frequency increment for credit classes

0%
1Q 2Q 3Q 4Q 5Q 6Q 7Q 8Q 9Q 10Q 11Q 12Q 13Q 14Q 15Q 16Q 17Q 18Q 19Q 20Q
Year 1 Year 2 Year 3 Year 4 Year 5

BB [12.5-8.5] B [8-6] CCC [5.5-5] C [4.5-1]

Source: Author's calculations

Figure 17 illustrates the tendency explained above. We notice that banks with junk ratings
(from class C) have very high probability of failure after the first quarter of credit rating
assignment (the default frequency is about 56%). However, such banks that can survive after the
first quarter have probabilities of default even lower than banks with ratings from a better class
(class CCC) and for some periods the probabilities of default are even less than for credit ratings
from class B. The junk credit ratings are usually assigned to banks with very poor financial

164
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

sustainability (these will hardly survive for more than two quarters) and small expanding banks
(these have great chances to survive for a long period of time). Small banks have lower ratios
than large and mature banks, but this does not mean that the probability of failure is extremely
large for them. If they survive during the first quarters after they receive a credit rating, their
sustainability can be even better than those with ratings from a better class. Hence, we conclude
that investors should take into account not only the rating grade of a bank, but also the time period
during which the bank remains in a rating class. From the figure above we also notice that an
increase in default frequencies is growing beginning from year 3 for almost each rating class.
This pattern may be explained in two pays. Firstly, it is possible that a credit rating has become
outdated and the banks’ financial sustainability has not been revalued. Secondly, internal
management of banks with unchanged credit rating for several years may put less effort into
development and innovations, which makes such banks less stable to external shocks.

Figure 18: Distribution of annual cumulative Default Frequencies for credit rating classes

90% 83%
78%
80%
70%
70% 67%
61%
60%
50% 47%
41%
40% 33%
30%
22% 23% 23%
20% 18%
11% 12% 13%
10% 5% 6%
1% 2% 4%
0%
BB [12.5-8.5] B [8-6] CCC [5.5-5] C [4.5-1]
DF in 1 year DF in 2 years DF in 3 years DF in 4 years DF in 5 years

Source: Author's calculations

Figure 18 illustrates the graduate cumulative increase in annual default frequencies for
each rating class. As expected, the lower probabilities of default are associated with a rating class
BB[12.5-8.5] (the best class of ratings presented), while the PDs in the High Speculative Grade
are larger for poorer rating classes. However, we can observe an interesting pattern: cumulative
annual growth rate (CAGR) of default frequencies decreases as credit ratings become lower.
CAGR is calculated for each rating class as:
1/5
 DFyear 5 
CAGR    (37)
 DFyear1 
 

165
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

From the figure above, we can see that in the class C[4.5-1], although it contains junk
credit ratings, a bank which is able to survive for one year after the rating issue has a lower
incremental PD on the horizon of the next five years (as CAGR value shows). This tendency
does not allow us to reject the fourth hypothesis for the banks with highly speculative credit
ratings. However, we cannot depict this for banks with credit ratings from class BB[12.5-8.5] in
our sample of Russian banks. We conclude that the better the financial sustainability of a bank,
the lower CAGR of PD is. Analysis of default frequencies shows us that PD increases in time at
a faster rate in the better rating classes. We are able to formulate an investment advice. Investment
in banks with better credit ratings is optimal right after the rating issue and is efficient to be held
for the short run period. However, to achieve a higher return from an investment in banks with
highly speculative rating grades, it is optimal to choose a long run investment 1 or 2 years after
the rating assignment. It is demonstrated with a help of CAGR values in classes BB[12.5-8.5]
and C[4.5-1]. In class BB, CAGR of PD equals to 62%, which is more than ten times higher than
the value of CAGR in class C (6%). If a bank is assigned a better credit rating (from lower
speculative class), it operates in a more competitive environment and its default brings large
losses to investors, while the failure of a bank with junk credit rating may not imply huge losses.

Figure 19: Average cumulative Default Frequencies for credit rating classes

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%
0 DF in 1 year DF in 2 years DF in 3 years DF in 4 years DF in 5 years

BB [12.5-8.5] B [8-6] CCC [5.5-5] C [4.5-1]

Source: Author's calculations

Figure 19 illustrates the graduate difference between the default frequencies of banks with
ratings from classes BB, B and CCC. PD is much larger for credit class C, we notice that banks
with credit ratings on the range from 4.5 to 1 are extremely unstable during the first year after
the credit rating assignment comparing even to the banks with very similar ratings from class

166
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

CCC. However, if such banks can survive during this period, an incremental PD for them is not
substantial and lower than for banks from better rating classes.

Thereby, in our paper we present the process which can be useful for investors for credit
risk evaluation. Calibration of credit rating score on PD allows us:
o to assess the quality of information enclosed in the credit rating scores obtained from the
uniform credit rating model grade (2). This proves the significance of the obtained credit
rating model and shows the possibility of its practical use, as the forecasted credit ratings on
a base scale are interdependent with estimated PD. We show empirically that lower credit
scores are associated with higher PD, while the highest credit scores (from 14 to 12 for
Russian banks) imply almost zero PD. We have created the base scale of credit ratings which
is more objective than the credit ratings of a single CRA because of mapping process. With
a help of default frequencies, we show the degree of information that the forecasted credit
scores bring to Russian economic agents;
o to make the credit ratings scores more intuitive, meaningful. The forecasted credit ratings
calibrated on default frequencies represent a quantitative measure that can be used by
investors for calculation of the expected loss (EL) in dollar value. The majority of investors
do not comprehend the credit risk from a symbolic credit rating. We provide investors with
the method of assessing the credit risk more accurately. The introduced method of credit risk
assessment is depicted on the figure below.

Figure 20: Method of credit risk assessment presented in the article

Publicly avaliable
information on bank's Uniform credit rating Credit rating score on the
performance and market model grade(2) base scale from 1 to 21.5
environment

Quantitative credit risk Planning an investment Checking the default

evaluation, calculation of period taking into account frequencies associated
possible expected losses the dynamic default with the corresponding
(EL) in dollar value frequencies credit rating score

Investors obtain not only the numeric credit rating, but also the quantitative measure of
credit risk, which is more comprehensive. This helps them to plan their investment strategy and
to calculate the expected losses (EL) in dollar value.

167
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Conclusion
In our work we present the method of credit risk estimation for banks. The proposed
ordered logit model is helpful for internal management of banks in IRB approach and for external
investors. The extensive analysis of academic literature allows us to collect valuable indicators
of the banks’ financial sustainability and to find out the most significant among them. The
indexes included in the final model specification cover all the aspects of bank’s credit risk, which
makes our model more accurate than those proposed in the academic literature. The final
specification of the grade model controls for the following factors: capital adequacy, asset
quality, management quality, profitability and its growth rate, liquidity, efficiency, bank’s size,
leverage, macroeconomic conditions and level of state governance. Additionally, interaction
terms are included into the model. We are able not only to find the significant proxies for the
abovementioned factors but also to evaluate the power of their influence on the credit risk using
the marginal effects calculation.
We fail to reject all the stated hypotheses. It is found out that the same set of explanatory
variables can be used to forecast the credit ratings of different CRAs (Moody’s, Fitch and
Standard& Poor’s), but these models are not accurate in credit risk forecasting. We are able to
demonstrate that mapping the credit ratings into a base scale helps to decrease the possible
subjectivity of CRAs. The second hypothesis indicates the importance of inclusion of state
governance indicators into the model. Using the constructed models, we observe that the
transparent politics of the government can lead to increased credit ratings of international banks.
This helps them in credit risk management and in the client’s base extension. We also reveal the
importance of maintaining an adequate liquidity level for banks in terms of financial crisis.
If we compare the final model with previous studies, we should pay attention to the work
of Karminsky & Khromova (2018), where the credit rating model for Russian banks was
developed. The share of correct forecasts of our credit rating model (grade(2)) is 5 percent higher
(23% compared to 18%), while the share of deviations of no more than one rating grade 21
percent higher (56% compared to 35%). What is more, the model proposed by the author does
not include the indicators for all the factors from the CAMEL(S) approach, which may lead to
imprecise credit risk valuation. The recommended model can be hardly used for international
banks, as only Russian banks were included in the sample. This, in turn, confirms the relevance
and necessity of our research.
Our paper introduces the approach to credit risk valuation, which is not limited by credit
rating estimation only. The models proposed in the previous academic articles help economic
agents to forecast the numerical credit rating that could have been assigned by a CRA;
interpretation of this rating score remained for the interested party itself. In our work we calibrate

168
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the credit rating on the PD using the historic default frequencies of Russian banks. Using the
obtained dynamic transmission scale, it is possible to value the credit risk associated with a
particular grade, which makes credit ratings more comprehensive and intuitive for investors and
may help them in investment decision making. The analysis does not allow us to reject the fourth
hypothesis for banks with credit ratings from highly speculative class.
The novelty of the article is the derivation of a credit risk estimation method on new
extensive sample of international banks. We are able to show that in order to assess the credit
risk correctly it is not enough to forecast the credit rating of any CRA, as the forecasted scores
may vary. We present the model of credit rating forecasting on a base scale, which helps us to
eliminate the subjective and unjust assessments of rating agencies. Moreover, we find out new
significant indicators on banks’ credit risk and evaluate their power of influence. We also present
the calibration of credit scores on PD, which allows to assess the information incorporated in the
forecasted credit ratings and to make these numeric scores more comprehensive. In further
research, it is possible to calibrate the credit ratings on the historic default frequencies of a
developed country and to compare the transition scale with the Russian one. Moreover, more
significant indicators on banks’ credit risk may be found. The obtained ordered logit regression
can be also compared to the AI methods of credit risk forecasting on the basis of the same set of
indicators.

169
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References
1. Ayvazyan, S., Golovan, S., Karminsky, A. & Peresetsky, A. (2011). About the approaches to
the comparison of rating scales. Applied econometrics, 3, 13-40. (in Russian)
2. Jivaykina, A. & Peresetsky, A. (2017). Credit ratings of Russian banks and reviews of
banking licenses 2012-2016. Journal of the New economic Association, 4, 49-80. (in
Russian)
3. Karminsky, A. (2015). Credit ratings and their modeling. National Research University
Higher School of Economics.-Moscow: HSE Publishing House. (in Russian)
4. Karminsky, A. & Peresetsky, A. (2007). Models of ratings of international agencies. Applied
econometrics,1, 3-19. (in Russian)
5. Karminsky, A. & Petrov, A. (2004). Ratings of dynamic financial stability for banks and
enterprises, Analytical Banking Journal, 12, 74-78. (in Russian)
6. Karminsky, A. & Sosurko, V. (2010). Features of modeling of the international ratings of the
banks. Financial risk management, 4, 292-305. (in Russian)
7. Peresetsky, A. (2009). Measuring the external support components of Moody's ratings.
Applied econometrics, 2, 3-23. (in Russian)
8. Peresetsky, A. (2012). Econometric approach to off-site analysis of Russian banks. Publ.
House of Higher School of Economics.-235,[5]p. (in Russian)
9. Pomasanov, M., Vlasov, A. (2008).Calibration of national rating systems.
Rynokcennichbumag (Security market),74-79. (in Russian)
10. Hamalinsky, A. & Pomasanov, M. (2012). The rating model calibrating for sectors with a
low number of defaults. Financial risk management, 2, 82-94. (in Russian)
11. Aitchison, J. & Silvey, S. (1957). The generalization of probit analysis to the case of multiple
responses. Biometrika, 44(1-2), 131-140.
12. Ali, A. & Durdu, K. (2012). An assessment of strategic importance of credit rating agencies
for companies and organizations. Procedia - Social and Behavioral Sciences, 58, 1628-1639.
13. Almamy, J., Aston, J. & Ngwa, L. (2016). An evaluation of Altman's Z-score using cash flow
ratio to predict corporate failure amid the recent financial crisis: Evidence from the UK.
Journal of Corporate Finance, 36, 278-285.
14. Alrabadi, D. & Hamarneh, R. (2016). Determinants of Banks' Credit Ratings: Evidence from
Jordan. Jordan Journal of Business Administration,12(4), 919-933.
15. Altman, E. (1968). Financial Ratios, Discriminant Analysis and the Prediction of Corporate
Bankruptcy. The Journal of Finance, 23(4), 589-609.
16. Altman, E., Haldeman, R. & Narayanan, P. (1977). ZETATM analysis. A new model to
identify bankruptcy risk of corporations. Journal of Banking & Finance, 1(1), 29-54.

170
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

17. Altman, E. & Rijken, H. (2004). How rating agencies achieve rating stability. Journal of
Banking & Finance, 28(11), 2679–2714.
18. Ashbaugh-Skaife, H., Collins, D. & LaFond, R. (2006). The effects of corporate governance
on firms’ credit ratings. Journal of Accounting and Economics, 42, 203–243.
19. Ashraf, B., Zheng, C., Jiang, C. & Qian, N. (2020). Capital regulation, deposit insurance and
bank risk: International evidence from normal and crisis periods. Research in International
Business and Finance, 52, 101-188.
20. Attig, N., Ghoul, S., Guedhami, O. & Suh, J. (2013). Corporate Social Responsibility and
Credit Ratings. Journal of Business Ethics, 117(4), 679-694.
21. Ayuso, J., Perez, D. & Saurina, J. (2004). Are capital buffers procyclical? Evidence from
Spanish panel data. Journal of Financial Intermediation,13(2),249–264.
22. Baetschmann, G., Staub, K. & Winkelmann, R (2011). Consistent Estimation of the Fixed
Effects Ordered Logit Model. IZA Discussion Paper No. 5443. Available at
[Link] .
23. Bellotti, T., Matousek, R. & Stewart, C. (2011). Are rating agencies’ assignments opaque?
Evidence from international banks. Expert Systems with Applications,38(4), 4206-4214.
24. Bermpei, T., Kalyvas, A. & Nguyenc, T. (2018). Does institutional quality condition the
effect of bank regulations and supervision on bank stability? Evidence from emerging and
developing economies. International Review of Financial Analysis, 59, 255-275.
25. Betz, F., Oprica, S., Peltonen, T. & Sarlin, P. (2014). Predicting distress in European banks.
Journal of Banking & Finance, 45, 225–241.
26. Chan-Lau, J. (2006). Fundamentals-Based Estimation of Default Probabilities: A Survey.
IMF Working Papers, 2006(149).
27. Cucinelli, D., Battista, M., Marchese, M. & Nieri, L. Credit risk in European banks: The
bright side of the internal ratings based approach. Journal of Banking & Finance, 93, 213-
229.
28. Driss, H., Massoud, N. & Roberts, G. (2019). Are credit rating agencies still relevant?
Evidence on certification from Moody's credit watches. Journal of Corporate Finance,
59,119-141.
29. Elliott, R., Siu, T. & Fung, E. (2014). A Double HMM approach to Altman Z-scores and
credit ratings. Expert Systems with Applications, 41(4/2), 1553-1560.
30. Godlewski, C. (2007). Are Ratings Consistent with Default Probabilities?: Empirical
Evidence on Banks in Emerging Market Economies. Emerging Markets Finance and Trade,
43(4), 5-23.

171
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

31. Gogas, P., Papadimitriou, T. & Agrapetidou, A. (2014). Forecasting bank credit ratings. The
Journal of Risk Finance, 15(2), 195-209.
32. Gu, X., Kadiyala, P. & Mahaney-Walter, X. (2018). How creditor rights affect the issuance
of public debt: The role of credit ratings. Journal of Financial Stability, 39,133-143.
33. Hu, X., Huang, H., Pan, Z. & Shi, J. (2019). Information asymmetry and credit rating: A
quasi-natural experiment from China. Journal of Banking & Finance, 106, 132-152.
34. Hwang, R-C., Chung, H. & Chu, C. (2010). Predicting issuer credit ratings using a
semiparametric method. Journal of Empirical Finance,17(1),120-137.
35. James, G., Witten, D. & Hastie, T. & Tibshirani, R. (2013). An Introduction to Statistical
Learning with Application in R. Springer-Verlag New York.
36. Jiao, Y., Syau, Y. & Lee, E. (2007). Modelling credit rating by fuzzy adaptive network.
Mathematical and Computer Modelling,45, 5–6.
37. Kaiser, H. (1960). The application of electronic computers to factor analysis.
Educational and Psychological Measurement, 20,141–151.
38. Karminsky, A., Hainsworth, R. & Solodkov, V. (2013). Arm’s Length Method for
Comparing Rating Scales. Eurasian Economic Review, 3, 114-135.
39. Karminsky, A., Grishunin, S., Dyachkova, N. & Bisenov, M. (2019). The comparison of
empirical methods for modeling credit ratings of industrial companies from BRICS countries.
Eurasian Economic Review, 10, 333–348.
40. Karminsky, A. & Khromova, E. (2016). Extended modeling of banks’ credit
ratings. Procedia Computer Science, 91, 201-210.
41. Karminsky, A. & Khromova, E. (2018). Increase of banks’ credit risks forecasting power by
the usage of the set of alternative models. Russian Journal of Economics, 4(2), 155-174.
42. Karminsky, A. & Khromova, E. (2016). Modelling banks’ credit ratings of international
agencies. Eurasian Economic Review,6, 341-363.
43. Karminsky, A. & Kostrov, A. (2014). The probability of default in Russian banking. Eurasian
Economic Review, 4(1), 81-98.
44. Kashyap, A. & Kovrijnykh, N. (2016). Who should pay for credit ratings and how? The
Review of Financial Stidies, 29(2), 420-456.
45. Kaufmann, D., Kraay, A. & Mastruzzi, M. (2010). The Worldwide Governance Indicators.
Policy Research working paper; no. WPS 5430. World Bank.
46. Ko, Y., Fujita, H. & Li, T. (2017). An evidential analysis of Altman Z-score for financial
predictions: Case study on solar energy companies. Applied Soft Computing, 52, 748-759.
47. Krüger, S. & Rösch, D. (2017). Downturn LGD modeling using quantile regression. Journal
of Banking & Finance, 79, Pages 42-56.

172
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

48. Kvamme, H., Sellereite, N., Aas, K. & Sjursen, S. (2018). Predicting mortgage default using
convolutional neural networks. Expert Systems with Applications, 102, 207-2017.
49. Le, H. & Viviani, J. (2018). Predicting bank failure: An improvement by implementing a
machine-learning approach to classical financial ratios. Research in International Business
and Finance, 44, 16-25.
50. Lee, S. & Choi, W. (2013). A multi-industry bankruptcy prediction model using back-
propagation neural network and multivariate discriminant analysis. Expert Systems with
Applications, 40( 8), 2941-2946.
51. Leow, M. & Mues, C. (2012). Predicting loss given default (LGD) for residential mortgage
loans: A two-stage model and empirical evidence for UK bank data. International Journal of
Forecasting, 28(1), 183-195.
52. Lin, T.-H.(2009). A cross model study of corporate financial distress prediction in Taiwan:
Multiple discriminant analysis, logit, probit and neural networks models. Neurocomputing,
72, 3507-3516.
53. Männasoo, K. & Mayes, D. (2009). Explaining bank distress in Eastern European transition
economies. Journal of Banking & Finance, 33, 244–253.
54. Onyiriuba, L. (2016). Bank Risk Management in Developing Economies. 1st edition.
Academic press.
55. Pagratis, S. & Stringa, M. (2007). Modelling bank credit ratings: A structural approach to
Moody’s credit risk assessment. Working paper. Bank of England, London.
56. Park, G. & Lee, H. (2018). Opportunistic behaviors of credit rating agencies and bond issuers.
Pacific-Basin Finance Journal,47, 39-59.
57. Pasiouras, F., Gaganis, C. & Doumpos, M. (2007). A multicriteria discrimination approach
for the credit rating of Asian banks. Annals of Finance,3(3),351-367.
58. Petropoulos, A., Siakoulis, V., Stavroulakis, E., Vlachogiannakis, N. (2020). Predicting bank
insolvencies using machine learning techniques. International Journal of Forecasting, In
Press. [Link]
59. Ponse, J. (2012). The quality of credit ratings: A two-sided market perspective. Economic
Systems ,36(2), 294-306.
60. Sahut, J-M. & Mili, M. (2011). Determinants of Banking Distress and Merger as Strategic
Policy to Resolve Distress. Economic modelling, 28(1/2), 138-146.
61. Schaeck, K. & Čihák, M. (2007). Banking Competition and Capital Ratios. IMF Working
Paper 07/216, International Monetary Fund.
62. Schuermann, T. & Hanson, S. (2004). Estimating Probabilitites of Default. Federal Reserve
Bank of New York Staff Reports, 190.

173
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

63. Shi, B., Zhao, X., Wu, B. & Dong, Y. (2019). Credit rating and microfinance lending
decisions based on loss given default (LGD). Finance Research Letters, 30, 124-129.
64. Shi, B., Chi, G. & Li, W. (2020). Exploring the mismatch between credit ratings and loss-
given-default: A credit risk approach. Economic Modelling, 85, 420-428.
65. Snell, E. (1964). A Scaling Procedure for Ordered Categorical Data. Biometrics, 20(3),592-
607.
66. Spuchľakova, E. & Cug, J. (2015). Credit Risk and LGD Modelling. Procedia Economics
and Finance, 23, 439-444.
67. Rudakova, O. & Ipatyev, K. (2015). Some Approaches to the Calibration of Internal Rating
Models. Review of European Studies, 7(10).
68. Tanoue, Y., Kawada, A. & Yamashita, S. (2017). Forecasting loss given default of bank loans
with multi-stage model. International Journal of Forecasting, 33(2), 513-522.
69. Tasche, D. (2013). The art of probability-of-default curve calibration. Journal of Credit Risk,
9(4), 63-103.
70. Thomson, J. (1991). Predicting Bank Failures in the 1980s. Economic Review, Federal
Reserve Bank of Cleveland, 27(1),9–20.
71. Volk, M. (2013). Estimating Probability of Default and Comparing It to Credit Rating
Classification by Banks. Economic and Business Review, 14(4).
72. Westgaard, S. & Wijst, N. (2001). Default probabilities in a corporate bank portfolio: A
logistic model approach, 135(2), 338-349.
73. Yeddou, N. & Pourroy, M. (2020). Bank liquidity creation: Does ownership structure matter?
The Quarterly Review of Economics and Finance. Available 7 June, 2020 at
[[Link]
74. Zhang, J., He, L. & An, Y. (2020). Measuring banks’ liquidity risk: An option-pricing
approach. Journal of Banking & Finance,111, 105703.
75. [Link] website. The source of Russian defaulted banks for the time period 2007-2019.
[[Link]
76. [Link] website. [[Link]
77. Moody’s Investors Service: Bank Rating Methodology. Available 20 May, 2020 at
[[Link]
78. Moody’s Investors Service: Corporate Default and Recovery Rates, 1920-2008. Available 19
May, 2020 at
[[Link]
79. Moody’s Rating Methodology: Banking. (2016). Available 21 May, 2020 at
[[Link]

174
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

80. OECD data website. [[Link]

81. S&P CAPITAL IQ website. [[Link]
82. S&P Global Direct: 2019 Annual Corporate Default and Rating Transition Study. Available
20 May, 2020 at
[[Link]
83. OECD data website. [[Link]
84. Fitch Ratings: 2019 Transition and Default Studies. Available 20 May, 2020 at
[[Link]
27-03-2020]
85. World Bank website. [[Link]

175
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Appendix 1: Supplementary tables

Table 23: List of countries from the sample

# Country # Country
1 Austria 24 Chile
2 Bulgaria 25 China
3 Czech Republic 26 Georgia
4 Estonia 27 Hong Kong
5 France 28 India
6 Germany 29 Indonesia
7 Hungary 30 Japan
8 Latvia 31 Kazakhstan
9 Lithuania 32 Malaysia
10 Malta 33 Mexico
11 Netherlands 34 Pakistan
12 Poland 35 Panama
13 Romania 36 Peru
14 Slovakia 37 Philippines
15 Slovenia 38 Russia
16 Sweden 39 Singapore
17 United Kingdom 40 South Korea
18 Argentina 41 Sri Lanka
19 Armenia 42 Switzerland
20 Australia 43 Taiwan
21 Azerbaijan 44 Thailand
22 Brazil 45 Ukraine
23 Canada 46 United States

Source: Author’s calculations

176 77
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 24: Numerical equivalence assigned to the symbolic ratings

Grade Moody’s Fitch S&P Numerical equivalence

Prime Aaa AAA AAA 21
Aa1 AA+ AA+ 20
High Aa2 AA AA 19
Aa3 AA- AA- 18
A1 A+ A+ 17
Upper medium A2 A A 16
A3 A- A- 15
Baa1 BBB+ BBB+ 14
Lower medium Baa2 BBB BBB 13
Baa3 BBB- BBB- 12
Ba1 BB+ BB+ 11
Speculative Ba2 BB BB 10
Ba3 BB- BB- 9
B1 B+ B+ 8
Highly speculative B2 B B 7
B3 B- B- 6
Substantial risk Caa1 CCC CCC+ 5
Extremely Caa2 CC CCC 4
speculative
Default with little Caa3 C CCC- 3
prospect to
Ca - - 2
recovery
In default C D D 1

Source: Karminsky & Khromova (2016) and Author’s calculations

177
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 25: Correlation matrix for macro variables

Variable (1) (2) (3) (4) (5) (6)

(1) CPI 1
(2) GDP -0.47 1
(3) DOMC -0.49 0.74 1
(4) CABAL -0.09 -0.13 -0.05 1
(5) GROSS 0.01 -0.46 -0.47 0.63 1
(6) EXP -0.02 -0.06 -0.13 0.49 0.41 1
- high correlation level (greater than 0.4)
Source: Author’s calculations

Table 26: Correlation matrix for state governance variables

Variable (1) (2) (3) (4) (5) (6)

(1) CORRC 1
(2) GOVER 0.96 1
(3) REGUL 0.94 0.95 1
(4) POLIT 0.82 0.82 0.80 1
(5) VOICE 0.83 0.82 0.84 0.62 1
(6) LAW 0.78 0.70 0.71 0.59 0.75 1
- high correlation level (greater than 0.4)
Source: Author’s calculations

178
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 27: Correlation matrix for financial indicators

Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27) (28) (29)
(1) TIER1 1
(2) EQ/AS 0.30 1
(3) EQ/DE 0.10 0.02 1
(4) CAP 0.77 0.19 -0.01 1
(5) PROV/G 0.01 -0.01 0.00 0.02 1
(6) NONL/L 0.03 0.08 0.00 0.14 -0.02 1
(7) ALL/NONL 0.00 -0.01 -0.01 0.05 -0.01 -0.03 1
(8) IE/DE 0.07 0.03 0.00 0.09 0.03 -0.06 -0.03 1
(9) DEB/DE 0.20 0.01 0.00 0.01 0.00 0.02 0.10 0.15 1
(10) OE/REV 0.00 -0.01 0.00 0.00 0.51 -0.01 0.20 0.18 0.05 1
(11) ROA 0.01 0.02 0.00 0.02 0.00 -0.02 0.00 0.07 0.00 0.00 1
(12) ROE 0.02 0.04 0.02 0.01 0.00 -0.01 0.01 0.03 0.00 0.00 0.61 1
(13) NL/DE 0.00 0.01 0.15 0.01 0.00 0.02 0.00 0.00 0.89 0.00 0.00 0.00 1
(14) CASH/DE -0.01 -0.02 0.00 0.00 0.00 0.05 0.00 0.01 0.10 0.00 0.04 0.00 0.00 1
(15) DE/EQ -0.11 -0.29 0.03 -0.11 0.00 -0.08 0.00 -0.08 -0.03 -0.01 0.22 -0.03 -0.03 0.00 1
(16) LA/NLA -0.04 -0.07 0.00 -0.06 0.00 -0.08 -0.01 0.01 0.00 0.00 0.01 0.01 0.00 0.00 0.08 1
(17) NL/A -0.04 0.03 0.01 0.01 0.02 -0.06 -0.02 0.07 -0.01 0.03 0.02 0.00 -0.01 0.01 0.01 -0.35 1
(18) CUR 0.14 0.03 0.00 0.15 0.00 0.00 0.00 0.01 0.00 0.00 -0.02 -0.01 0.14 0.20 -0.11 -0.06 0.00 1
(19) NII/REV -0.01 -0.01 0.03 0.01 0.00 0.04 0.00 0.00 0.02 0.00 0.10 -0.01 0.00 0.12 0.02 -0.01 0.01 0.00 1
(20) NII/AA 0.09 0.21 0.01 0.13 -0.01 0.19 0.04 -0.09 0.01 -0.01 -0.02 -0.02 0.01 -0.01 -0.12 -0.21 0.06 0.01 0.02 1
(21) OE/OI 0.00 0.03 0.05 0.00 0.02 0.01 0.00 0.01 0.00 0.00 0.20 0.00 0.10 0.02 0.23 0.04 0.03 0.00 -0.01 0.00 1
(22) EFF 0.05 0.10 0.00 -0.03 0.00 0.10 -0.04 -0.06 0.04 0.00 -0.01 -0.02 0.04 0.00 -0.09 -0.08 0.00 0.01 0.00 0.10 0.01 1
(23) NONIE/A -0.01 -0.01 0.00 0.00 0.42 -0.01 0.00 0.00 0.00 0.84 0.00 0.00 0.00 0.00 -0.01 0.00 0.04 0.00 0.00 -0.01 0.00 0.00 1
(24) LOGA -0.17 -0.31 -0.02 -0.21 0.01 -0.32 -0.04 0.05 -0.02 0.01 0.03 0.04 -0.02 0.01 0.20 0.27 -0.16 -0.02 -0.02 -0.56 0.00 -0.28 0.01 1
(25) LOGEQ -0.12 -0.15 -0.02 -0.19 0.01 -0.30 -0.04 0.04 -0.02 0.00 0.01 0.04 -0.01 0.00 0.08 0.27 -0.16 -0.02 -0.03 -0.52 0.00 -0.27 0.00 0.97 1
(26) NI 0.00 0.00 0.00 -0.01 0.00 0.02 -0.01 0.06 0.04 0.12 0.00 0.00 0.00 0.01 0.00 0.12 -0.01 0.04 0.06 -0.00 0.02 0.01 0.00 -0.01 -0.01 1
(27) REV 0.03 0.09 0.00 0.01 0.00 0.03 0.00 0.01 0.10 0.00 0.01 0.01 0.00 0.00 -0.01 -0.01 -0.01 0.01 -0.01 0.02 0.01 0.01 0.00 -0.04 -0.03 0.21 1
(28) LEV 0.07 0.19 0.00 0.02 0.00 -0.04 -0.03 -0.01 0.00 0.00 0.00 -0.01 0.00 0.00 -0.09 0.00 0.02 0.00 0.00 0.08 0.00 0.08 0.00 -0.19 -0.17 0.01 0.01 1
(29) AGE -0.10 -0.10 0.00 -0.16 0.00 -0.23 -0.10 0.09 -0.01 0.00 0.00 0.04 -0.01 0.01 0.01 0.11 -0.03 -0.02 -0.02 -0.25 0.00 0.01 0.00 0.40 0.40 -0.04 -0.02 -0.02 1
- high correlation level (higher than 0.4) Source: Author’s calculations

179 80
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 28: Pairwise correspondence of credit ratings Moody’s – S&P

Moody's
S&P
Categorical scale C Ca Caa3 Caa2 Caa1 B3 B2 B1 Ba3 Ba2 Ba1 Baa3 Baa2 Baa1 A3 A2 A1 Aa3 Aa2 Aa1 Aaa
Total
Categorical
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
scale Numerical scale
D 1 12 2 4 3 1 22
CC 4 5 5
CCC+ 5 1 12 1 1 36 51
B- 6 4 1 10 38 80 6 8 8 155
B 7 30 1 31 71 37 13 11 18 212
B+ 8 17 23 76 27 68 26 8 10 6 261
BB- 9 4 4 39 74 125 28 5 279
BB 10 5 87 37 20 31 1 1 4 186
BB+ 11 48 28 102 33 1 4 216
BBB- 12 4 40 90 67 20 13 234
BBB 13 29 66 147 36 17 9 7 311
BBB+ 14 4 19 30 91 267 202 66 10 7 696
A- 15 8 91 75 183 198 79 6 640
A 16 7 67 176 328 95 17 690
A+ 17 18 8 42 125 139 32 364
AA- 18 23 46 114 180 61 424
AA 19 16 5 21
AA+ 20 58 58
AAA 21 0
Total 46 2 11 26 92 179 158 127 327 165 123 305 355 536 513 526 588 361 254 68 63 4825

Source: Author’s calculations

180 81
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 29: Pairwise correspondence of credit ratings Moody’s - Fitch

Moody's
Fitch
Categorical scale C Ca Caa3 Caa2 Caa1 B3 B2 B1 Ba3 Ba2 Ba1 Baa3 Baa2 Baa1 A3 A2 A1 Aa3 Aa2 Aa1 Aaa
Total
Categorical
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
scale Numerical scale
D 1 38 38
CCC- 3 0
CCC 4 17 4 6 11 41 34 37 1 11 162
B- 6 34 14 16 33 37 40 20 1 8 203
B 7 15 1 7 31 39 21 23 54 191
B+ 8 5 31 53 58 83 11 241
BB- 9 22 18 48 77 44 19 228
BB 10 15 92 13 7 8 135
BB+ 11 2 5 47 54 6 33 56 203
BBB- 12 3 33 33 131 96 9 1 3 309
BBB 13 1 29 94 126 9 259
BBB+ 14 8 25 76 115 102 42 3 371
A- 15 20 158 117 95 90 480
A 16 7 41 89 127 193 191 32 680
A+ 17 22 16 61 137 109 29 374
AA- 18 1 44 65 123 152 61 446
AA 19 33 13 52 17 115
AA+ 20 0
AAA 21 13 13
Total 66 4 7 70 93 161 166 190 387 156 82 233 383 519 373 438 519 277 233 78 13 4448

Source: Author’s calculations

181
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 30: Base credit rating scale (international scale)

Base scale Moody’s S&P Fitch Class

21.5 AAA 6
21 Aaa 6
20.5 AA+ 6
20 Aa1 AAA 6
19.5 AA 6
19 Aa2 AA+ 6
18.5 AA- 6
18 Aa3 AA 5
17 A1 A+ AA- 5
16 A2 A A+ 5
15.5 A 5
15 A3 A- 5
14.5 A- 5
14 Baa1 BBB+ BBB+ 4
13 Baa2 BBB BBB 4
12 Baa3 BBB- BBB- 4
11 Ba1 BB+ BB+ 3
10 Ba2 BB BB 3
9 Ba3 BB- BB- 3
8 B1 B+ B+ 2
7.5 B 2
7 B2 B 2
6.5 B- 2
6 B3 B- 2
5.5 CCC+ 1
5 Caa1 CCC+ 1
4.5 CCC 1
4 Caa2 CC 1
3.5 CCC- 1
3 Caa3 C 1
2 Ca 1
1 C D D 1

Source: Author’s calculations

182 83
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 31: Intermediate steps in credit rating modeling (Uniform Grade model)

Indicators grade(1) grade(1.1) grade(1.2) grade (2)

CAP 0.016* 0.016* 0.016* 0.016*

(0.001) (0.001) (0.001) (0.001)
NONL/L -0.083*** -0.083*** -0.083*** -0.069***
(0.005) (0.005) (0.005) (0.005)
OE/REV -0.003*** -0.003*** -0.004*** -0.003***
(0.0004) (0.0004) (0.0004) (0.0004)
ROE 0.125*** 0.127*** 0.128*** 0.119**
(0.046) (0.047) (0.046) (0.048)
LA/NLA 0.003** 0.003*** 0.004*** 0.003**
(0.001) (0.001) (0.001) (0.001)
EFF -0.017*** -0.017*** -0.018*** -0.017***
(0.004) (0.004) (0.004) (0.004)
NI 0.007** 0.008** 0.008** 0.008**
(0.003) (0.003) (0.003) (0.003)
LOGA 0.247***
(0.057)
LOGA2 0.054*** 0.048*** 0.081***
(0.012) (0.012) (0.0112)
(CUR)*(Y2008) 0.016*** 0.018*** 0.023**
(0.004) (0.004) (0.011)
(CUR)*(Y2009) 0.043*** 0.046*** 0.044***
(0.014) (0.013) (0.014)
(CUR)*(Y2010) 0.0165** 0.0188*** 0.019***
(0.007) (0.007) (0.007)
(LEV)*(CIS) 0.040*** 0.026***
(0.008) (0.008)
(LEV)*(EU) 0.026*** 0.0145*
(0.0079) (0.0077)
(PROV/G)*(IE/DE) 0.00002* 0.00002*
(0.00001) (0.00001)
(ALL/NONL)*(OE/OI) -0.016*** -0.016***
(0.003) (0.003)
(ROA)*(AGE) 0.227** 0.199*
(0.115) (0.114)
MAC1 0.886***
(0.052)
MAC2 -0.953***
(0.121)
GOV1 1.603***
(0.092)
Sample size 12,350 12,350 12,350 12,174
AIC 39,330.51 39,307.16 39,066.53 38,009.13
BIC 39,619.95 39,626.28 39,422.76 38,394.29
Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1

Source: Author’s calculations

183
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 32: Marginal effects of each covariate (Model class(2))

Pr(1) Pr(2) Pr(3) Pr(4) Pr(5) Pr(6)

CAP -0.00005 -0.00035 -0.00067 -0.00010 0.00091 0.00027
(0.0000265) (0.0000984) (0.0001413) (0.0002862) (0.0001458) (0.000108)
NONL/L 0.00021 0.00148 0.00281 0.00044 -0.00382 -0.00112
(0.0001077) (0.0004012) (0.0005329) (0.0012032) (0.0005058) (0.0004395)
OE/REV 0.00001 0.00007 0.00014 0.00002 -0.00018 -0.00005
(0.000005) (0.00002) (0.00002) (0.00005) (0.00003) (0.00002)
ROE -0.00033 -0.00236 -0.00448 -0.00069 0.00609 0.00178
(0.0002556) (0.0014638) (0.0014255) (0.0019585) (0.001199) (0.0012221)
LA/NLA -0.00000 -0.00000 -0.00000 -0.00000 0.00000 0.00000
(0.00000) (0.00000) (0.00000) (0.00000) (0.00000) (0.00000)
EFF 0.00010 0.00067 0.00128 0.00020 -0.00173 -0.00051
(0.0000518) (0.0002169) (0.0003304) (0.000548) (0.00039) (0.0002212)
LOGA2 -0.00010 -0.00068 -0.00130 -0.00020 0.00176 0.00052
(0.0000497) (0.000185) (0.0002763) (0.0005576) (0.000297) (0.0002056)
(CUR)*(Y2008) -0.00006 -0.00040 -0.00076 -0.00012 0.00103 0.00030
(0.0000313) (0.0001282) (0.0001953) (0.0003262) (0.0002315) (0.000131)
(CUR)*(Y2009) -0.00024 -0.00168 -0.00319 -0.00049 0.00434 0.00127
(0.0001369) (0.0006056) (0.0009822) (0.0013745) (0.0012209) (0.0005902)
(CUR)*(Y2010) -0.00010 -0.00072 -0.00136 -0.00021 0.00185 0.00054
(0.000061) (0.0002863) (0.0004787) (0.0005862) (0.0006076) (0.0002678)
(LEV)*(CIS) -0.00008 -0.00057 -0.00108 -0.00017 0.00146 0.00043
(0.0000473) (0.0002226) (0.0003664) (0.0004635) (0.0004566) (0.0002069)
(LEV)*(EU) -0.00071 -0.00504 -0.00958 -0.00148 0.01301 0.00381
(0.0003962) (0.0016061) (0.0023564) (0.0041243) (0.0028701) (0.0016755)
MAC1 -0.00267 -0.01886 -0.03581 -0.00555 0.04865 0.01424
(0.0013673) (0.0048498) (0.0063986) (0.0153449) (0.0060012) (0.0055256)
MAC2 0.00093 0.00659 0.01251 0.00194 -0.01700 -0.00497
(0.0005157) (0.0020855) (0.003103) (0.0053704) (0.0035969) (0.0021494)
GOV1 -0.00870 -0.06145 -0.11670 -0.01809 0.15854 0.04640
(0.0044787) (0.0154598) (0.0192284) (0.049922) (0.01503) (0.0177526)
- insignificant marginal effect
Standard errors in parentheses.
Source: Author’s calculations

184
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 33: Robustness check of credit rating models

Class model Grade model
class(2) class(2) grade(2) grade(2)
Indicators
(ordered probit) (maximum classes) (ordered probit) (maximum ratings)
CAP 0.005*** 0.002 0.006*** 0.005***
(0.001) (0.001) (0.001) (0.002)
NONL/L -0.031*** -0.015*** -0.038*** -0.052***
(0.003) (0.003) (0.002) (0.005)
OE/REV -0.000* -0.000 -0.000*** -0.000
(0.000) (0.000) (0.000) (0.000)
ROE 0.011 0.103*** 0.008 0.094*
(0.033) (0.038) (0.026) (0.053)
LA/NLA 0.001 0.003*** 0.001* 0.000
(0.001) (0.001) (0.001) (0.001)
EFF -0.014*** -0.019*** 0.013*** -0.011***
(0.002) (0.003) (0.002) (0.004)
NI 0.001 -0.001 0.001 0.004
(0.002) (0.003) (0.002) (0.003)
LOGA2 0.021*** 0.028*** 0.034*** 0.053***
(0.001) (0.002) (0.004) (0.003)
(CUR)*(Y2008) 0.010*** 0.014*** 0.03* 0.014***
(0.002) (0.003) (0.002) (0.003)
(CUR)*(Y2009) 0.034*** 0.026** 0.031*** 0.044***
(0.012) (0.0105) (0.010) (0.015)
(CUR)*(Y2010) 0.015** 0.015** 0.012** 0.017**
(0.006) (0.006) (0.006) (0.006)
(LEV)*(CIS) 0.008* 0.012*** -0.017*** -0.044***
(0.004) (0.004) (0.003) (0.009)
(LEV)*(EU) 0.130*** 0.06* 0.003 -0.047
(0.036) (0.032) (0.009) (0.032)
(PROV/G)*(IE/DE) 0.00002** 0.00012
(0.000) (0.0001)
(ALL/NONL)*(OE/OI) -0.000** -0.000***
(0.000) (0.000)
(ROA)*(AGE) -0.094 0.913***
(0.139) (0.333)
MAC1 0.481*** 0.261*** 0.548*** 0.200***
(0.036) (0.039) (0.023) (0.047)
MAC2 -0.136*** -0.148*** -0.306** -0.174***
(0.029) (0.031) (0.102) (0.040)
GOV1 1.356*** 1.123*** 0.906*** 1.612***
(0.051) (0.061) (0.020) (0.088)
Sample Size 12,091 12,091 12, 174 12, 174
AIC 12,151.35 11,749.59 41,031.3 39,123.32
BIC 12,312.15 11,912.4 41,070.24 39,765.45
Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1
Source: Author’s calculations

185
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 34: Base credit rating scale (national scale)

Base scale S&P Fitch Moody's RAEX Rus-Rating AK&M NRA Ria
21.5
21 Aaa
20.5
20 Aa1
19.5
19 Aa2
18.5
18 Aa3
17 A1
16 A2
15.5
15 A3
14.5
14 Baa1
13 ruAAA Baa2 AAA
12.5 AAA(rus)
12 ruAA+ AA+(rus) Baa3 A++
11.5 AA(rus) AA+
11 ruAA AA-(rus) Ba1
10.5 ruAA- A+(rus) AA
10 ruA+ A(rus) Ba2 AAA
9.5 ruA A-(rus) AA-
9 ruA- BBB+(rus) Ba3
8.5 ruBBB+ BBB(rus) A+ A+ AA+
8 ruBBB BBB-(rus) B1 A A+ AA
7.5 ruBBB- BB+(rus) A- AA
7 ruBB+ BB(rus) B2 A BBB+ AA- AA-
6.5 ruBB BB-(rus) BBB A A+ A+
6 ruBB- B+(rus) B3 BBB- A A
5.5 B(rus) BB+ B++ A-
5 ruB+ B-(rus) Caa1 B++ BB BBB+ A-
4.5 B+ BBB BBB+
4 ruB Caa2 B+ BBB- BBB
3.5 B BB+ BB+
3 ruB- Caa3 B B C++ BB C
2.5 B- C+ BB-
2 ruCCC- Ca C++ CC
1.5
1 ruD D(rus) C E C C

Source: Karminsky & Sosurko (2011) and Author’s calculations

186
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 35: Intermediate matrix with default frequencies for credit score 4.5 (for periods 29-49)
quarter of credit rating assignment (l)
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
30 0.00%
31 0.00% 1.14%
32 1.15% 1.14% 2.20%
33 1.15% 1.14% 1.10% 3.30%
34 1.15% 1.14% 1.10% 1.10% 3.23%
35 0.00% 0.00% 0.00% 1.10% 1.08% 4.35%
36 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 3.19%
quarter of default (q)

37 4.60% 4.55% 3.30% 2.20% 2.15% 2.17% 3.19% 3.33%

38 1.15% 1.14% 1.10% 1.10% 1.08% 1.09% 1.06% 1.11% 4.40%
39 3.45% 3.41% 3.30% 3.30% 3.23% 3.26% 3.19% 3.33% 3.30% 4.30%
40 5.75% 5.68% 5.49% 6.59% 6.45% 6.52% 6.38% 5.56% 5.49% 5.38% 6.67%
41 3.45% 3.41% 3.30% 2.20% 2.15% 2.17% 2.13% 2.22% 2.20% 3.23% 3.33% 5.81%
42 1.15% 1.14% 1.10% 1.10% 1.08% 1.09% 1.06% 1.11% 1.10% 1.08% 1.11% 1.16% 6.98%
43 3.45% 3.41% 4.40% 4.40% 4.30% 4.35% 3.19% 3.33% 3.30% 4.30% 5.56% 5.81% 5.81% 8.75%
44 3.45% 3.41% 4.40% 4.40% 5.38% 5.43% 5.32% 5.56% 6.59% 6.45% 6.67% 6.98% 5.81% 6.25% 12.33%
45 1.15% 1.14% 2.20% 2.20% 2.15% 2.17% 1.06% 1.11% 1.10% 1.08% 1.11% 1.16% 1.16% 1.25% 1.37% 9.86%
46 1.15% 1.14% 1.10% 1.10% 2.15% 2.17% 2.13% 2.22% 1.10% 1.08% 1.11% 1.16% 2.33% 2.50% 2.74% 2.82% 4.62%
47 0.01% 0.01% 0.2% 3.00% 0.00% 1.00% 0.00% 3.00% 0.00% 1.00% 0.00% 3.40% 0.20% 0.00% 1.00% 0.00% 4.00% 1.69%
48 4.60% 4.55% 5.49% 5.49% 5.38% 5.43% 5.32% 5.56% 4.40% 4.30% 4.44% 4.65% 4.65% 5.00% 5.48% 4.23% 4.62% 5.08% 4.84%
49 1.15% 1.14% 1.10% 1.10% 1.08% 1.09% 1.06% 1.11% 1.10% 1.08% 2.22% 2.33% 2.33% 2.50% 2.74% 2.82% 3.08% 3.39% 3.23% 8.33%

Source: Author’s calculations

187 88
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 36: Intermediate matrix with default frequencies for credit score 7 (for periods 29-49)
quarter of credit rating assignment (l)
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
30 0.10%
31 1.20% 2.00%
32 0.96% 0.96% 0.97%
33 0.96% 0.96% 0.97% 2.04%
34 0.96% 0.96% 0.97% 1.02% 2.22%
35 1.92% 1.92% 1.94% 1.02% 1.11% 1.11%
36 1.00% 0.30% 1.30% 1.50% 1.60% 2.00% 2.10%
quarter of default (q)

37 2.88% 2.88% 3.88% 5.10% 5.56% 4.44% 3.45% 3.41%

38 0.96% 0.96% 1.94% 2.04% 2.22% 2.22% 2.30% 2.27% 4.55%
39 0.96% 0.96% 0.97% 1.02% 1.11% 1.11% 1.15% 1.14% 1.14% 2.50%
40 0.96% 0.96% 0.97% 0.00% 0.00% 0.00% 0.00% 1.14% 1.14% 1.25% 2.50%
41 0.96% 0.96% 2.91% 3.06% 3.33% 3.33% 2.30% 2.27% 3.41% 3.75% 2.50% 3.85%
42 2.88% 2.88% 2.91% 3.06% 3.33% 3.33% 3.45% 2.27% 2.27% 2.50% 2.50% 2.56% 1.32%
43 2.88% 2.88% 1.94% 2.04% 2.22% 2.22% 3.45% 3.41% 3.41% 2.50% 1.25% 1.28% 1.32% 2.53%
44 2.88% 2.88% 2.91% 3.06% 2.22% 2.22% 2.30% 2.27% 1.14% 1.25% 1.25% 1.28% 2.63% 2.53% 3.70%
45 0.96% 0.96% 0.00% 0.00% 0.00% 0.00% 1.15% 2.27% 2.27% 2.50% 2.50% 1.28% 1.32% 1.27% 1.23% 3.85%
46 1.92% 1.92% 1.94% 2.04% 1.11% 1.11% 1.15% 1.14% 1.14% 1.25% 2.50% 2.56% 2.63% 2.53% 2.47% 2.56% 3.90%
47 1.92% 1.92% 1.94% 2.04% 2.22% 2.22% 2.30% 2.27% 2.27% 2.50% 2.50% 2.56% 2.63% 2.53% 2.47% 2.56% 2.60% 3.95%
48 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 1.14% 1.25% 1.25% 1.28% 1.32% 1.27% 1.23% 1.28% 1.30% 1.32% 4.35%
49 0.96% 0.96% 0.97% 1.02% 1.11% 1.11% 1.15% 1.14% 1.14% 1.25% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 1.41%

Source: Author’s calculations

188
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Appendix 2: Supplementary figures

Figure 21: Moody’s approach to rating banking organizations

Source: Moody’s Investors Service: Bank Rating Methodology

189 90
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 22: Relationship between ratings Moody’s – S&P, the sample

24
22
20
18
16
Moody's (y)

14
12
10
8 y = 1,0505x - 0,7598
R² = 0,8775
6
4
2
0
0 2 4 6 8 10 12 14 16 18 20 22 24
S&P (x)

Note: the relationship between the numerical credit scores (not logarithmic) is depicted
Source: Author’s calculations

Figure 23: Relationship between ratings Moody’s – Fitch

24
22
20
18
16
Moody's (y)

14
12
10
8
6 y = 0,8974x + 0,9217
R² = 0,8875
4
2
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Fitch (x)

Note: the relationship between the numerical credit scores (not logarithmic) is depicted
Source: Author’s calculations

190
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 24: Loading plots for the macro PCs

Source: Author’s calculations

Figure 25: Loading plot for the state governance PCs

Source: Author’s calculations

191
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 26: Distribution of forecast errors of model class(2) (ordered probit)

Model class(2) (ordered probit)
∆=0 0.33
|∆|≤1 0.81
|∆|≤2 0.92

0,35 33%

0,3
26%
0,25 22%

0,2

0,15

0,1
6%
5%
0,05 3% 3%
2%
Value of ∆
0
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author’s calculations

Figure 27: Distribution of forecast errors of model grade(2) (ordered probit)

Model grade(2) (ordered probit)
∆=0 0.19
|∆|≤1 0.53
|∆|≤2 0.71

0,25

19%
0,20
18%
16%
0,15

9% 10%
0,10 8%
6%
0,05 3%
2% 2% 2% 2%
0% 0% 1% 1% 1% 1% 1%
Value of ∆
0,00
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author’s calculations

192
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 28: Distribution of forecast errors of model class(2) (with maximum rating classes)
Model class(2) (with maximum rating grades)
∆=0 0.30
|∆|≤1 0.79
|∆|≤2 0.90

0,35
30%
0,3
25%
24%
0,25

0,2

0,15

0,1
6%
5%
0,05 3% 3%
2% 2%
Value of ∆
0
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author’s calculations

Figure 29: Distribution of forecast errors of model grade(2) (with maximum rating grades)
Model grade(2) (with maximum rating grades)
∆=0 0.16
|∆|≤1 0.47
|∆|≤2 0.70

0,20 18%
0,18
16%
0,16
0,14 13%
13%
0,12
10%
0,10
8%
0,08
6%
0,06
3%
0,04 3% 2%
2%
0,02 1% 1% 1% 1% Value of ∆
0% 0% 0% 1% 0% 0% 0%
0,00
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author's calculations

193
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 30: Distribution of forecast errors of model class (2) (out-of-sample of banks)
Model class(2) (another sample of banks)
∆=0 0.33
|∆|≤1 0.83
|∆|≤2 0.92

0,35 33%

0,3 27%

0,25 23%

0,2

0,15

0,1
5% 5%
4%
0,05 3%
Value of ∆
0
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author’s calculations

Figure 31: Distribution of forecast errors of model grade(2) (out-of-sample of banks)

Model grade(2) (another sample of banks)
∆=0 0.17
|∆|≤1 0.48
|∆|≤2 0.71

0,18 17%
16%
0,16 15%
0,14
12%
0,12 11%
0,10
0,08 7% 7%
0,06 5%

0,04
2% 2% 2%
0,02 1% 1% 1% 1%
0,2% Value of ∆
0,00
-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10 11

Source: Author’s calculations

194
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Appendix 3: Codes

Code 1: Python code to generate random sample of 500 banks

data = pd.read_csv('data_geographic_region.csv')
random_sample = [Link](list(data['bank ID']), k = 500)
print(random_sample)
Source: Author’s calculations

Code 2: Stata program to estimate ordered logit model fixed effect (FE)

capture program drop feologit_buc

program feologit_buc, eclass
version 10
gettoken gid 0: 0
gettoken y x: 0
tempvar iid id cid gidcid dk
qui sum `y'
local lk= r(min)
local hk= r(max)
bys `gid': gen ìid'=_n
gen long ìd'=`gid'*100+ìid'
expand `=`hk'-`lk''
bys ìd': gen `cid'=_n
qui gen long `gidcid'= `gid'*100+`cid'
qui gen `dk'= `y'>=`cid'+1
clogit `dk' `x', group(`gidcid') cluster(`gid')
end
feologit_buc ivar yvar xvars

Note: This program was used to estimate the fixed effect logistic regression, as Stata does not
have a uniform command for it.
ivar is the individual identifier;
yvar is the ordered dependent variable;
xvars is the list of explanatory variables
Source: Baetschmann et al. (2014).

195
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Code 3: R code for default frequencies matrix calculations

getwd()
setwd("/Users/romanmacbook/Desktop")
[Link]("openxlsx")
library(openxlsx)
file <- read.csv2("[Link]")

file_matrix <- function(r) {

m <- matrix(nrow = 40, ncol = 40)
colnames(m) <- seq(1,40)
rownames(m) <- seq(2,41)

for (i in 1:40){
for (j in i:40){
a <- subset(file, bank %in% file$bank[file$rating == r & file$quarter ==
i])
sum_defaulted_quarter <- sum(a$fact_of._default[a$quarter == j])
sum_rated_quarter <- sum(ifelse(file$rating == 21
& file$quarter == i, 1, 0))
m[j,i] <- round(sum_defaulted_quarter/sum_rated_quarter, 2)
}
}
return(m)
}
View(file_matrix(21
))
[Link](file_matrix(21
), '[Link]', [Link] = T, colnames = T)

Source: Author’s calculations

196
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

RETHINKING THE ROLE OF CREDIT UNIONS IN THE EUROPEAN SOCIAL ECONOMY

Ana IVANISEVIC HERNAUS

(corresponding author)
University of Zagreb, Faculty of Economics and Business, Croatia
[Link]@[Link]

Ivana BIONDIC
The Institute for Development and International Relations, Zagreb, Croatia
ibiondic@[Link]

Global financial crisis of 2007/2008 has brought a lot of rethinking, not just in terms of financial sector
regulation, but also of the very role of financial institutions in supporting the society's needs.
Motivated by opinion of some finance academics and professionals, that banking should be small-scale
and boring (e.g. see Gilmann, 2017), we reexamine the role of small, bank-like financial institutions, by
focusing on credit unions. Although popular in some parts of the world, credit unions have in certain
countries been ignored in financial sector development. The aim of this review paper is to rethink the
role of credit unions in social economy and asses their future prospects. In our theoretical research
focused on credit unions in the European Union countries, we systematically approach their
organizational and institutional features, analyse significance in serving society's needs and identify
major obstacles to their development. We conclude that credit unions have conditional potential in
contributing to the European social economy.

JEL Classification: G21, G23

Keywords: credit unions, cooperatives, Europe.

1. INTRODUCTION

Credit unions, as not-for-profit financial institutions, have traditionally operated at small scale,
providing primarily basic banking services of accepting deposits and making loans. These financial
cooperatives, managed by voluntary boards of directors, have a significant role in enabling access to
affordable financing for small businesses, consumers with limited resources, and underserved
communities. Diversity of the financial system, to which these institutions contribute, is essential for
the European economy. As Castello et al. (2018) emphasized, an effective complementation of
shareholder-value banks by stakeholder-value financial institutions is the only way to secure a stable
and effective financial ecosystem that makes its full contribution to the development of the real
economy (Castello et al., 2018).

Although small in terms of assets under their control in comparison to other financial institutions,
credit unions have a unique and important role in the financial system. Moreover, the emergence of
microfinance in the mid-2000s as a sector of responsible finance pointed to their potential, as one of

197
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the institutions supporting this financial practice. In addition, Castello et al. (2018) note that, in a
way, such, cooperative financial institutions have also been an oddity in the financial system exactly
due to the above mentioned reasons of their market share and their specialisation, social nature and
internal workings. Furthermore, today's financial institutions are fewer in number and bigger in size,
as a result of the mega-merger trend in financial sector of developed economies, which have been
transformed by consolidation, convergence and competition (Santomero, 1999; Ralston et al., 2011).
Ralston et al. (2011) pose an interesting question of where does this leave small financial institutions,
having their advantages over global financial conglomerates (primarily in terms of information
advantages for a certain class of borrowers) and their role in serving the financial needs of a
significant segment of the society?

From a research perspective, Fried, Lovell and Eeckaut (1993) interestingly note that credit unions,
perhaps because of their small size and small share in most services in the financial intermediation
sector, and perhaps because of their unconventional organizational structure, have not been the
subject of much theoretical or empirical research. Their sustainable development trends and prospects
are rarely analyzed in contemporary scientific publications and more detailed studies are lacking in
these topics (Dubauskas, 2012). Many authors (e.g. Karafolas, 2016) indicate the importance of such
knowledge, on different credit cooperative systems, stating it is highly useful and of significant
scientific and practical interest, especially for policy makers.

Due to the loss of trust in (large) financial institutions caused by the emergence of the global financial
and economic crisis 2007/2008, recently one witnesses increasing efforts of researchers and
professionals in addressing the issue of the role of financial institutions in supporting the society's
needs. Motivated by opinion of some finance academics and professionals, that banking should be
small-scale and boring (e.g. see Gilmann, 2017), we rethink the role of small, bank-like financial
institutions, by focusing on credit unions. The aim of the paper is to reexamine the role and the
potential of these institutions in the European Union and their contribution the European social
economy. We provide an overview and analysis of conceptual basis and development of credit unions,
highlight current major challenges the credit union movement is facing and provide reflection on their
future perspective.

Our contribution to the literature is three-fold. First, we add to the scarce literature on credit union
sustainable development trends and prospects by analyzing the impact of significant regulatory,
legislative and structural changes these institutions are facing nowadays. Second, we also contribute
to the literature on socially responsible and sustainable financing, as the credit union concept is
founded on principles of such financing. Third, in light of a broader context of the European social
economy and the view that the social economy is a crucial part of the EU socio-economic landscape
(Social Economy Europe, 2018), our insights and reflection could be useful for increasing the visibility
of the economic and social impact of the analysed cooperative institutions.

2. THEORY, REGULATORY ASPECTS AND MARKET OVERVIEW OF CREDIT UNIONS

2.1. The development of the credit union concept

Credit unions are small, member-owned cooperatives and not-for-profit institutions, which
distinguishes them from other financial intermediaries (Fried, Lovell and Eeckaut, 1993). They are
governed by a board of members who usually serve on a voluntary basis without receiving any
remuneration for the time and resources they dedicate to the credit union (European Network of
Credit Unions, 2013). They are designed as, and in many countries have kept this track through their

198
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

evolution, financial institutions offering affordable and easily understandable financial products. Credit
unions usually provide basic financial services, such as savings, loans and, at times, payment and
insurance products, to their members at a reasonable price. In balancing the interest of members,
credit unions may be borrower- or saver-oriented, or neutral. Several defining characteristics of
cooperatives, as the European Commission (2020) states are: an open and voluntary association, a
democratic structure with each member having one vote and an equitable and fair distribution of
economic results according to the volume of operations made through the cooperative.

The concept of a credit union is based on the cooperative banking model. This model has specific core
values – solidarity, self-help, fighting against exclusion, social and environmental concerns, resilience,
proximity, trust and governance. And these values are embedded in the ways cooperatives do banking,
explaining their commitment to societal responsibility (European Association of Co-operative Banks,
2020). Therefore, neoclassical theories of the firm, which assume firms maximize profits, are
inadequate for understanding the economic behavior of cooperative organizations (e.g. McKillop and
Wilson, 2011). Those researching cooperatives contend that these organizations embody multiple
values and objectives (McKillop and Wilson, 2011), such as presented above. Stoffman (2017)
concludes that, in credit unions, as opposed to their non-cooperative competitors, there is a deeper
level of connection to the corporate social responsibility. He states that credit unions are able to stake
a true strategic claim and point of differentiation (Stoffman, 2017). The development of the credit
union movement, and the unique organizational and institutional features of credit unions are
systematically presented and analysed below.

The credit union movement traces its roots to Germany in the middle of the 19th century, when
Friedrich Wilhelm Raiffeisen and Hermann Schultze-Delitzsch independently looked for ways to
improve the situation of the impoverished rural population as well as the afflicted urban middle class.
Following the principles of self-help, self-administration and self-responsibility, first credit unions were
supporting primarily farmers, small craftsmen and traders (World Council of Credit Unions, 2020). The
idea expanded to North America in the early 20th century. The purpose of the movement was also to
attack usury on a global scale and to provide a simple, yet effective solution to improve people's
economic situations. In the 1950s, international credit union development emphasized community
development, so the programs had broad social, as well as economic objectives (World Council of
Credit Unions, 2020). The importance of credit unions grew after the Second World War due to their
participation in the reconstruction efforts (Karafolas, 2016). Over the next few decades credit unions
were organized in nearly all parts of the world (for a more detailed overview of the history of credit
unions see World Council of Credit Unions, 2020). Today there is a great diversity within the credit
union movement and development across different countries and parts of the world, which can be
explained by various economic, historic and cultural contexts (see e.g. McKillop and Wilson, 2011).

Over time, and especially with the beginning of 21st century, the business of credit unions has in many
countries taken advantage of loosening regulatory restrictions on their fields of membership and some
of them expanded their offer from, for example, holding uncollateralized, short-term loans to often
devoting large shares of their assets to credit cards, auto loans, residential mortgage, and increasingly,
business loans (Wilcox, 2011). As the overviews of Walter (2006) and Wilcox (2006) concluded, many
of the traditionally important differences between commercial banks and credit unions have eroded,
but some remain (see Wilcox, 2011). As credit unions today operate in a highly competitive financial
market, a number of studies (e.g. Feinberg, 2001; Hannan, 2003; Jackson, 2006) researched the
competitive disciplines provided by credit unions in local markets for financial services, suggesting the
credit unions play an important role also in affecting the behaviour of commercial banks and other
profit-maximizing financial institutions and in disciplining the exercize of market power by banks (see

199
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

e.g. McKillop and Wilson, 2011). Hannan (2003) found that banks and thrifts offer higher rates on
deposits in markets where there is a significant credit union presence (Hannan, 2003 in McKillop and
Wilson, 2011). Furthermore, although contemporary financial sector environment favours larger
depository institutions, credit unions have in some countries traditionally gained on popularity and are
the only depository institution that manages to resist the competition from institutional investors
increasingly taking over the management of citizens' savings through various forms of contractual
savings and investment (e.g. Leko and Stojanovic, 2018).

Theory, the history of credit union movement and credit unions' operating principles suggest that
these institutions aim to balance interests of their members, having this as a priority over
commercial objectives. And this is especially so where they are not in significant direct competition
with other financial institutions (McKillop and Wilson, 2011).

2.2. Regulatory aspects

As regarding the regulatory aspects of credit union business, we build on research of McKillop and
Wilson (2011) who provide a systematic account of the related issues. First, a credit union may be
established by a group of persons who share a common interest and act jointly with a view to promote
and protect mutual interests on the principle of financial reciprocity (e.g. Croatian National Bank,
2015), limiting its services to members only and therefore making the common bond one of its defining
characteristics. The common bond is based on a pre-existing social connection such as belonging to a
particular associational, community, industrial or geographic group and it is debated whether it limits
credit union growth by restricting the pool of (potential) members to a relatively finite group or
whether it is useful in the development of nascent credit union movements where it reduces the cost
of gathering credit information on members (for more detail on this debate see e.g. McKillop and
Wilson, 2011).

Second, in most countries credit unions are tax exempt (McKillop and Wilson, 2011). One of the views
offered in support of tax exemption is that credit unions provide subsidized services to members, many
of whom are of modest means, and a tax levy would create pressure to eliminate some of these
subsidized services (McKillop and Wilson, 2011).

Third, what is discussed is the question of deposit insurance. The introduction of deposit insurance
may increase the probability of credit unions engaging in risk-shifting behavior (at the expense of the
insurance fund) although the special characteristics of credit unions reduce the probability of this
happening. However, less risk-taking behaviour by credit unions is more likely. This is because they
depend on internally generated capital to fuel expansion and their managers may act in a risk-averse
manner because they cannot be rewarded with stock-based compensation packages enabling them to
participate in the profitable performance of the organization (Llewellyn and Holmes, 1991 and
Rasmusen, 1988 in McKillop and Wilson, 2011).

Fourth, capital regulation is now applied to credit unions in most countries (McKillop and Wilson,
2011). For credit unions, any capital raised is via retained earnings and is tax exempt. Credit unions do
not have an option to raise new capital in the form of equity, and so are more likely to manage their
capital cautiously over the course of the business cycle (McKillop and Wilson, 2011).

2.3. Market overview

According to the newest data published by the World Council of Credit Unions (2019), in year 2018, a
number of 85,400 credit unions were operating worldwide in 188 countries. They had total of

200
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2,191,086 USD million assets under control, with membership of 274 million which equated to a
population penetration of 9.4%.

The geographical region of Europe, which is the focus of our paper, comprised credit unions in 17
countries, including Albania, Armenia, Belarus, Croatia, Estonia, Great Britain, Ireland, Latvia,
Lithuania, North Macedonia, Moldova, Netherlands, Poland, Romania, Russia, Turkey and Ukraine
(World Council of Credit Unions, 2019). Out of these countries, the largest number of credits unions
operates in Turkey (1,625), Great Britain (429), Ukraine (358), Ireland (331), Moldova (268) and Russia
(244).

Table 1. Global market overview of credit unions

Geographical region Number of Membership Savings & Shares Loans Total Assets Reserves Penetration
(Number of countries) credit unions (millions) (USD millions) (USD millions) (USD millions) (USD millions) (%)

Africa (27) 39,447 35.78 9,595.8 8,132.6 10,779.9 1,059.8 13.8

Asia (25) 33,004 57.45 147,223.1 138,186.9 180,826.2 5,772.5 4.3
Carribean (17) 374 3.43 6,273.1 4,996.3 7,662.5 774.2 65.2
Europe (17) 3,491 9.10 24,095.3 11,647.8 32,959.4 4,068.7 9.2
Latin America (17) 2,891 35.81 59,840.1 54,212.0 90,863.9 17,209.7 14.6
North America (2) 6,010 127.97 1,485,177.5 1,326,185.3 1,786,602.2 170,218.7 48.9
Oceania (13) 183 4.68 70,025.7 66,763.8 81,392.4 6,423.7 3.9
World Total (118) 85,400 274.20 1,802,240.5 1,610,124.8 2,191,086.3 205,527.3 9.4

Source: World Council of Credit Unions 2019. Statistical Report 2018. Retrieved from:
[Link] (20 June 2020)

When comparing our market overview with a similar one done by McKillop and Wilson for year 2009,
we find that the statistics for credit unions has increased significantly in the last decade. At the global
level, the increase in the number of credit unions equals 73% and in membership 49%. Savings and
shares have increased by 57%, loans by 77%, assets under control by 62% and reserves by 72% . The
penetration rate increased from 7.6 to 9.4%.

The region of Europe recorded increases, but of a smaller extent in comparison to global trends in the
analysed decade. An increase of 44% was recorded in the number of credit unions, 7% in membership
and 5% in savings. Loans has a contrary trend and decreased by -16%. Other indicators recorded
positive trends – total assets increased by 25%, reserves by 18% and the penetration rate increased
from 3.6% to 9.2%.

3. THE EUROPEAN COOPERATIVE STATUS

European Commission (2020) defines cooperative as an autonomous association of persons united to

meet common economic, social, and cultural goals, who achieve their objectives through a jointly-
owned and democratically-controlled enterprise. Cooperatives are enterprises that serve the needs of
their members who contribute to their capital. When looking at the EU economy, cooperatives are
important because there are 250,000 of them in the EU, owned by 163 million citizens (one third of EU
population) and employing 5.4 million people. They hold substantial market shares in same industries.
For example, in banking sector this indicator amounts to 50% in France, 37% in Cyprus, 35% in Finland,
31% in Austria and 21% in Germany (European Commission, 2020). The presence of cooperatives and
their recognition varies greatly among Member States. While in most EU countries cooperatives are
present in almost all sectors of the economy, they are lagging behind in some, lacking the necessary
recognition and support (European Commission, 2015). As Castello et al. (2018) emphasized, in many

201
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

European cities cooperatives form part of the community’s identity, contribute to economic well-being
and therefore that model’s essence needs to be guaranteed and its survival promoted. Credit unions
are organized as financial cooperatives (where the owners are the members) (McKillop and Wilson,
2011). In some countries, “credit union” concept replaces terms such as cooperative financial
institutions, micro-credit or self-help organizations, or cooperative banks (Dubauskas, 2012).

At a historical glance, in European countries credit unions were established as institutions that saved
people from usurers and loan sharks, and operated iniatially as agricultural lenders, founded by
medium and small-sized farmers to improve their access to banking services (Karafolas, 2016). Their
further development in most European countries had a somewhat different direction in comparison
with their American counterparts. While in North America credit unions remained more local and kept
autonomous structure, based on their unique defining characteristics, in most European countries they
evolved to cooperative banks – dominant consumer financial providers that serve the general public
(Kaupelyte and McCarthy, 2004). Indeed, in some countries of Western Europe, credit cooperatives
have grown into large cooperative banks that provide full banking services (Dubauskas, 2012).

Furthermore, when taken a closer look at Europe, two credit union movements can be noted – the one
in Western European- and in Eastern European countries, as they evolved differently. Most of the
Western European countries have well established credit union systems with long tradition. Countries
in Eastern Europe had to abandon old cooperative systems and create new ones after the planned
economy period (see e.g. Karafolas, 2016).

Today, there is a European cooperative status (Castello et al., 2018; Kuc and Teply, 2018), with
characteristics shares by all cooperatives: ownership rights and, consequently, governance using the
„one member – one vote“ rule. However, this different evolution or structure of credit cooperative
systems accross EU countries resulted in, by cases, sharp contrasts across the credit cooperative
systems throughout the EU (e.g. Karafolas, 2016). Differences among European countries exist not only
in their terminology as a credit union, cooperative bank, or credit cooperative (Karafolas, 2016) but in
some cases also in terms of defining characteristics of credit unions, such as the common bond
principle, „one member – one vote“ principle, geographical and product offerings etc. Moreover, in
some countries, for example, a minimum initial capital is required for the creation and function of a
cooperative, while in others, there is no such requirement.

In addition to such various practice, differences may exist also within the same country accross credit
cooperative institutions. While in some cases, the credit cooperative system has a pyramid structure
with local, regional, and national central institutions, in other, credit cooperative systems are
characterized by independent local cooperative banks (Karafolas, 2016). In some countries (e.g.
Finland), during the second half of the 20th century, the management of credit unions professionalized
and their legal status was converted to banks (see Kalmi, 2016). To conclude, each country has its
specific legislative framework and traditions. A detailed insight into this topic can be found in a valuable
and succesfull attempt of Karafolas and his associates (2016), who gave a comparative presentation of
the credit cooperative systems in 23 European countries (EU member countries and one candidate-
country), being one of the most recent and international collection of essays on the credit cooperative
systems in Europe. Karafolas (2016) notes these differences within the systems can provoke problems
in the application of common policies within the EU, and we deal with this issue in the next part of our
paper.

202
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4. FUTURE PERSPECTIVE OF CREDIT UNIONS IN THE EUROPEAN UNION

4.1. Challenges in development of credit unions in the European Union

There is a wide variety of issues that credit unions are facing today, such as the appropriate EU
regulatory framework, the identification of barriers for further development at national level, or their
internationalization. First, in analysing the challenges credit union come across, we start off by
presenting criticism of the credit union model as such. Then we move on to the analysis of the response
of credit unions in the recent global financial and economic crisis of 2007/2008. Finally, we discuss the
impact of current regulatory trends on credit union business.

The credit union model is also criticized for some aspects, coming out of the specific defining
characteristics of these institutions. Some authors argue (e.g. Leko and Stojanovic, 2018) that the
drawbacks of the model are the impossibility of applying economies of scale, being too protected,
creating unfair competition to banks, having too many bad loans due to a lack of professionalism, and
high possibility of corruption among its administration. If taken the issue of the participatory and
democratic structure of cooperatives, for example, this requires specific competences to assume board
responsibility in order to ensure effective governance, yet the elected officials to the board of directors
often lack the required training and skills (European Commission, 2015). Credit cooperatives in some
countries, e.g. Romania, have in recent years been faced with the problem of internal fraud, as Pirvu
and Banica (2016) report. The example of other countries, such as the Czech Republic, as Kuc and Teply
(2018) report, show that a poor legislative framework, which did not reflect trends and developments
in cooperative banking of the second half of the 20th century, from which credit unions in this country
suffered from, turned into a severe crisis of the sector in year 1999. Therefore, presented issues should
be addressed, discussed and dealt with in an appropriate manner. For example, in order to manage
information risk more effectively and prevent the problem of internal fraud, the mentioned authors
(Pirvu and Banica, 2016) propose that all transactions of secondary credit union locations shall obtain
approval from the head office, by authorized persons.

Moreover, recent years, marked with financial crises and regulatory changes that followed in many EU
countries, brough a number of structural changes to credit unions (see e.g. Karafolas, 2016). In
addition, differences in credit unions among EU countries made consequences of the recent financial
crises manifest in different ways. In some countries, e.g. Cyprus, extra shaken by the crisis of 2013,
credit unions were forced to merge and most part (99%) of their ownership was transferred to state
in exchange for a capital injection (Kleanthous and Hadjimanolis, 2016). These authors strongly point
out that the trend and efforts in making several or only one large financial institution at disposal of a
foreign bank, out of many, previously operating, credit unions, might result in changing of a true not-
fot-profit nature of cooperative financial institutions and might have negative consequences for their
members and for wider society. A positive dealing of credit unions with contemporary challenges can
be found, for example, in Portugal, where the strong customer deposit base of these institutions and
prudential credit management policy have proved to be crucial in confronting current hard times (see
Cabo and Rebelo, 2016).

However, as the competitiveness of the business environment that credit unions are nowadays facing
increases, the fear of many academics and professionals (see e.g. Fajardo-Garcia and Soler-Tormo,
2016; Castello et al., 2018) remains, that this could jeopardize the valuable principles which inspire the
cooperative movement and which are the basis of the idea and their concept. This would mean loosing
the special nature of cooperatives, thereby depriving society of a major social asset that has been built
up over several centuries (Castello et al., 2018). Under the above discussed condition that obstacles to
credit unions development should be discussed and dealt with, our conclusions go in line with the ones

203
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of Fajardo-Garcia and Soler-Tormo (2016) who emphasize, that economic viability can and must be
compatible with the democratic principle of „one member – one vote“, with enchancing ownership
participation, with neighbourhood and community, with transparency and with their combination of
social and financial objectives.

4.2. Credit unions and the global financial and economic crisis of 2007/2008

The recent financial and economic crisis affected almost all European economies and financial sectors,
including the cooperative credit systems (e.g. Karafolas, 2016). Credit unions are considered not to be
the cause of the financial crisis, and, in comparison with their banking counterparts, have better been
able to resist these turbulences because of their focus on the retail banking activity and less
dependence on networks of the global financial markets (see e.g. Richez-Battesti and Leseul, 2016).
Since cooperative credit systems have not been exposed to the same high-risk speculative activities
one might expect that they have performed better than their banking counterparts (e.g. Poprawa,
2009; Karafolas, 2016). Castello et al. (2018), taking a broader perspective, gave one of the
explanations for the resilience of the cooperative banking model during times of financial crisis. That
is, while commercial banks are more inclined towards short-term profit and maximising value for
shareholders, cooperative institutions advocate sustainable development over the long term and at
local level.

In the extent to which credit cooperatives were affected by the financial crisis and experienced its
consequences, their response to these challenges comprised cleaning up their balance sheets, cutting
operating costs and launching merger processes (Castello et al., 2018). Interestingly, Rauterkus and
Ramamonjiarivelo (2010) shed a whole new light on the importance of credit unions (in McKillop and
Wilson, 2011). They found that credit union deposits increased in times of economic uncertainty and
suggested that this indicated that there is a group of people that consider credit unions a safe haven
during an economic crisis. According to Statistical Reports published by the World Council of Credit
Unions (2008; 2013) savings in credit union increased by 31% in a period from year 2007 to 2012.

4.3. Impact of regulatory trends on credit unions

Several key issues relate to the global regulatory trends regarding credit unions. The key challenges,
as systematically presented, for example by Bakani (2014), are an increased regulatory burden,
payments innovation, young adults membership growth and small credit union sustainability. In light
of the increasing need for regulatory compliance globally since the global financial crisis of 2007/2008,
it is crucial for credit unions that legislative reforms be made to enable appropriate and relevant
regulation and supervision by the Regulator (Bakani, 2014). Pushing for credit unions to come under
close scrutiny is good for credit union credibility but could damage the movement, as Bakani (2014)
warns. The total financial cost of regulations on credit unions is enormous, and has grown substantially
over the past several years (see e.g Credit Union National Association, 2017).

As far as it regards the second area, related to payments innovation, being one of the challenges driving
the current regulatory changes, practice has shown that credit unions venture into partnerships with
private sector to improve their payment systems. Furthermore, there is a trend in addressing
membership growth of credit unions (as well as other financial institutions) through networking
workshops, educational awareness and financial inclusion expositions. Finally, ensuring the
sustainability of smaller credit unions is also a question, taking into consideration the related cost of
the increasing regulatory burden (Bakani, 2014).

Bakani (2014) stresses that making reforms to regulation is necessary, but the capacity for regulators
to regulate, and for credit unions to comply with them is an ongoing challenge where finding the

204
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

balance between the two remains an open question. If taken a closer look at the question of the
increased regulatory burden, European Network of Credit Unions (ENCU), comprising eight European
associations and the World Council of Credit Unions, formed and published its attitude towards this
regulatory trend. While ENCU agrees that credit unions should comply with the principles set out in
the Basel Framework, it asks that the European Commission and the Basel Committee on Banking
Supervision focus on the implementation of proportionality contained in the Basel Framework and
ensure that included is the ability to tailor rules based on the size, risk, and complexity of the
institution. As ENCU empahizes, credit unions are unique in structure and are organized and operate
on a not-for-profit, member-owned and controlled cooperative basis. As other community-based
mutual depository institutions, they typically operate using a non-complex community-banking model
and are often subject to investment portfolio-shaping rules that limit their investments primarily to
loans to their members, government-guaranteed debt, and deposits held by banks or other credit
unions (European Network of Credit Unions, 2019). Taking this specific characteristics of credit unions
into consideration, the requests of the ENCU can be summerized as follows: further request that the
EU maintain the ability to provide exemptions from the Capital Requirements Directive for member
states and to allow for the continuation of national level regulation of credit unions in order to ensure
the success of these institutions, further urge the European Commission to consider measures that will
reduce the administrative burdens associated with regulatory programs and tailor them towards a
specific industry that has an important role in providing credit and depository services to the under-
banked and un-banked communities, and the inclusion of clearer instructions to the national-level
authorities.

ENCU observes a tendency for the application of Basel standards without regards for proportional
tailoring to the risk or complexity of a credit union or other community-based mutual depository
institutions, and therefore fears that the regulatory burden and compliance burdens could drive
consolidation or elimination of these types of institutions from the marketplace. If many smaller credit
unions and other financial institutions with a social purpose close or merge into larger institutions, this
will leave many of the more vulnerable sections of European society without access to basic financial
services. It should also be taken into account that neither credit unions nor other community-based
cooperative depository institutions are systemically important on a global or domestic level and that
their operations are far less complex than those of Globally Systemically Important Banks, Domestic
Systemically Important Banks or internationally active banks (European Network of Credit Unions,
2019).

5. CONCLUSIONS

The credit union model contributes to a creation and empowerment of the link of the financial system
with the real economy. At financial markets of some EU countries, credit unions are the key social
enterprises and microfinance institutions. A considerable potential of cooperatives has also been
recognised by the European Commission due to their both economic and social role. Credit unions
contribute to the European social economy and to a diversity of the financial system. Our conclusions
go in line with Castello's findings, that the characteristic features of cooperatives’ contribution to the
philosophy of the social economy can be summarised as follows: their contribution to local economic
growth, the principle of financial non-exclusion and the large-scale social work derived from those
initiatives. Therefore, the development of credit unions should be encouraged, and be done in two
complementary ways. One is addressing the challenges coming out of their specific characteristics, by
for example, promoting staff professionalism, training and development, enhancing internal risk-
management systems, fostering transparency and good corporate governance, encouraging

205
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

investment in technological innovation, etc. And the other way is dealing with regulation trends and
tailoring them towards this specific industry, to enable it to continue with providing financing for small
businesses, consumers with limited resources, and underserved communities. The overall message is
that the role credit unions in the European social economy (still) counts. Therefore, the policy makers,
when bringing regulatory and supervisory decisions, should take into account the very unique features
of these institutions to preserve their value and prevent their potential dissapearance from the
marketplace. Credit unions are an essential component of the European financial system.

References

Bakani, L.M. 2014. The future of credit unions given the global regulatory trends. BIS central bankers’
speeches. Retrieved from [Link] (30 June 2020).

Cabo, P. and Rebelo, J. 2016. The Portuguese Cooperative Credit System. In: Karafolas, S. (ed.) Credit
Cooperative Institutions in European Countries. Springer International Publishing Switzerland, 191-
212.

Castello, E., Trias, C. and Arribas, A. 2018. Europe’s cooperative banking models (Revised edition).
European Economic and Social Committee, Brussels.

Credit Union National Association 2017. Regulatory Impact Study conducted by Cornerstone
Advisors. Retrieved from:
[Link]
df (30 June 2020).

Croatian Natioanal Bank 2015. Core function. Authorisation of credit unions. Retrieved from
[Link] (30 June 2020).

Dubauskas, G. 2012. Sustainable growth of the financial sector: The case of credit unions. Journal of
Security and Sustainability Issues 1(3): 159-166.

European Association of Co-operative Banks 2020. Co-operative Banks. Key values. Retrieved from
[Link] (30 June 2020).

European Commission 2015. Report of the discussions of the Cooperative working group "Fostering
cooperatives' potential to generate smart growth & jobs" held between June 2013 to November 2014
containing ideas and proposals for action. Retrieved from
[Link]
growth-and-jobs-0_en (30 June 2020).

European Commission 2020. Internal Market, Industry, Entrepreneurship and SMEs. Cooperatives.
Retrieved from [Link] (30 June
2020).

European Network of Credit Unions 2013. Position on the proposal for a Directive on the
comparability of fees related to payment accounts, payment account switching and access to
payment accounts with basic features. Retrieved from
[Link] (30 June 2020).

European Network of Credit Unions 2019. Engagement with the EU. Re: Inception Impact
Assessment: Amendments to the Capital Requirements Regulation and the Capital Requirements
Directive. Retrieved from [Link] (30 June 2020).

206
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Fajardo-Garcıa, G. and Soler-Tormo, F. 2016. The Credit Cooperative System in Spain. In: Karafolas, S.
(ed.) Credit Cooperative Institutions in European Countries. Springer International Publishing
Switzerland, 213-232.

Feinberg, R.M. 2001. The competitive role of credit unions in small local financial services markets.
Review of Economics and Statistics 83(3): 560-563.

Fried, H.O. Lovell, C.A.K. and Eekaut, P.V.. 1993. Evaluating the performance of US credit unions.
Journal of Banking & Finance 17(2-3): 251-65.

Gilmann, S. 2017. Bank should be small and boring – prof Eckhard Hein. Horizon – the EU Research &
Innovation Magazine, 4 January. Retrieved from [Link]
[Link] (30 June 2020).

Hannan, T.H. 2003. The impact of credit unions on the rates offered for retail deposits by banks and
thrift institutions. Unpublished paper, Board of Governors of the Federal Reserve System.

Jackson, W.E. 2006. A comparison of the deposit and loan pricing behavior of credit unions and
commercial banks. Wisconsin: Filene Research Institute.

Kalmi, P. 2016. Co-operative Banks in Finland. In: Karafolas, S. (ed.) Credit Cooperative Institutions in
European Countries. Springer International Publishing Switzerland, 43-54.

Karafolas, S. 2016. (ed.) Credit Cooperative Institutions in European Countries. Springer International
Publishing Switzerland.

Kleanthous, A. and Hadjimanolis, A. 2016. Co-operative Credit Institutions in Cyprus. In: Karafolas, S.
(ed.) Credit Cooperative Institutions in European Countries. Springer International Publishing
Switzerland, 19-42.

Kuc, M. and Teply, P. 2018. A financial performance comparison of Czech credit unions and European
cooperative banks. Prague Economic Papers 27(6): 723-742.

Leko, Stojanovic, 2018. Financijske institucije i trzista. Sveuciliste u Zagrebu, Ekonomski fakultet.

Llewellyn, D.T. and Holme, M.J. 1991. In defense of mutuality: A redress to an emerging conventional
wisdom. Annals of Public and Cooperative Economics 62(3) :319-54.

Mavrenko, T. 2016. Cooperative Savings and Credit Unions in Latvia. In: Karafolas, S. (ed.) Credit
Cooperative Institutions in European Countries. Springer International Publishing Switzerland, 309-
326.

McKillop, D. and Wilson, J.O.S. 2011. Credit Unions: A Theoretical and Empirical Overview. Financial
Markets, Institutions & Instruments, 20(3), 79-123.

Pirvu, D. and Banica, L. 2016. The Credit Cooperative System in Romania. In: Karafolas, S. (ed.) Credit
Cooperative Institutions in European Countries. Springer International Publishing Switzerland, 363-
378.

Poprawa, A. 2009. Regulation and Legislation of Cooperative Banks and Credit Unions. Prepared for
United Nations Expert Group Meeting on Cooperatives April 28-30, 2009 New York. Retrieved from:
[Link] (30 June 2020).

Ralston, D., Wright, A. and Garden, K. 2000. Can mergers ensure the survival of credit unions in the
thrid millenium? Journal of Banking and Finance 25 (12): 2277-2304.

207
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Rasmusen, E. 1988. Mutual banks and stock banks. Journal of Law and Economic 31(2): 395-421.

Rauterkus, A and Ramamonjiarivelo, Z. 2010. Why choose a credit union? Determinants of credit
union deposits.” Retrieved from [Link] (30 June 2020).

Richez-Battesti, N. and Leseul, G. 2016. Cooperative Banks in France: Emergence, Mutations and
Issues. In: Karafolas, S. (ed.) Credit Cooperative Institutions in European Countries. Springer
International Publishing Switzerland, 55-82.

Santomero, A.M. 1999. Bank mergers: what’sa policymaker to do? Journal of Banking &
Finance, 23(2-4): 637-643.

Social Economy Europe 2018. The Future of EU Policies for the Social Economy: Towards a European
Action Plan. Brussels.

Stoffman, M. 2017. Credit Union Social Responsibility: Their Mission-Driven Difference. Banking &
Finance Law Review 33(1): 41-55.

Walter, J.R. 2006. Not Your Father’s Credit Union. Federal Reserve Bank of Richmond Economic
Quarterly 92(4): 353-377.

Wilcox. J.A. 2006. Credit Union Conversions to Banks: Facts, Incentives, Issues, and Reforms. Filene
Research Institute 1-104.

Wilcox, J.A. 2011. The Increasing Importance of Credit Unions in Small Business Lending. Office of
Advocacy, the United States Small Business Administration.

World Council of Credit Unions 2020. Our History. Retrieved from

[Link] (30 June 2020).

World Council of Credit Unions 2008. Statistical Report 2007. Retrieved from
[Link] (30 June 2020).

World Council of Credit Unions 2013. Statistical Report 2012. Retrieved from
[Link] (30 June 2020).

World Council of Credit Unions 2019. Statistical Report 2018. Retrieved from
[Link] (30 June 2020).

208
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

All we need are taxes: A systematic review on tax avoidance

Andreia Magalhães
Department of Accounting
ISCTE-IUL
Portugal
andreiamagalhaes88@[Link]

Rogério Marques Serrasqueiro

Department of Accounting
ISCTE-IUL
Portugal
[Link]@[Link]

Paulo Jorge Dias

Department of Accounting
ISCTE-IUL
Portugal
[Link]@[Link]

Although tax avoidance is a widely studied matter and a unavoidable issue in

public policy discussions about tax planning, there is no clear knowledge about its
determinants and consequences, at company’s and its shareholders levels, and the
way they are combined. This systematic review aims to identify the determinants
and consequences of tax avoidance and understand the different perspectives
adopted in academic research. 123 identified papers were analysed using a
computer-assisted qualitative data analysis software, to typify four main research
domains concerning tax avoidance: Quantification, endogenous determinants,
exogenous determinants and consequences. Additionally, subdivisions were found
within each sections of the determinants. Main findings are that research has
dominantly been performed in the USA, the more frequent metrics are ETR,
CashETR and BTD, the main research methods are OLS and Logistic regressions.
The endogenous factors that may explain tax avoidance are the most analysed
probably due to a real need to understand the more controllable variables
influencing tax avoidance and for the difficulty in gathering and analysing data
concerning external factors. In this group we identified five subdivisions, namely
related to the drivers of profit shifting, the ownership structure and corporate
governance, the CSR, the auditor-provided tax services and the information quality
and the human resources.

209
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

For the exogenous determinants affecting tax avoidance we identified two

subdivisions, formal and informal factores. Finally, we find that the main
consequences include the integrity of the tax system, through loss of tax revenue,
the loss of reputation and increased scrutiny at company level, and the increased
level of risk and associated costs (e.g.: increased financing and equity costs).
Future research should address other settings besides the USA and new topics like
the impact of new technologies and artificial intelligence on tax avoidance.

Keywords: Tax, tax avoidance, tax planning, tax aggressiveness, tax sheltering.

JEL Classiﬁcations: M41 – Accounting; M48 – Government Policy and

Regulation

“Just stop talking about philanthropy and start talking about taxes. … We can invite Bono
once more, but we’ve got to be talking about taxes. That’s it. Taxes, taxes, taxes”.

Rutger Bregman in World Economic Forum in Davos (2019)

1. Introduction

Tax avoidance led to a significant erosion of country tax bases and to profit shifting to

countries with more favourable tax regimes, inducing substantial losses, estimated in

about 4% to 10 % of the global tax revenue, yielding high losses for States, and also

inducing loss of confidence, as well as competitive distortions among companies (OCDE,

2015).

Empirical research examining this activity had progressively increase since the

90s. Shackelford and Shevlin (2001) made a review of empirical tax research in the

beginning of 2000 where they called attention for unexplored areas such as the

determinants of tax aggressiveness. Later, Hanlon and Heitzman (2010), also in a review,

called for an explanation why some companies avoid more taxes than others. More

recently, Wilde and Wilson (2018) published a survey where they highlighted the

210
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

difficulty in analysing all the studied determinants together and extract inferences from

them.

This paper goes beyond a mere tax avoidance literature review, since we attempt

to answer the question of what the determinants and consequences of tax avoidance are,

identifying, selecting and critically assessing previous research dimensions to ensure

transparency and integration in discussing research findings. Research gaps are singled

out and potential directions for future research are identified.

Our work contributes to the literature by reviewing papers from 2003 to 2018, and

addressing an identified gap about the relation and trends among all the determinants and

its consequences.

2. Methodology

The main objective of this review is to understand the ‘state of the art’ on the topic of tax

avoidance, namely its definition and assessment, and the main determinants and the

consequences arising from it.

Based on the structure proposed by Fink (2010), Petticrew and Roberts (2006) and

Tranfield, Denyer and Smart (2003), we defined seven tasks for developing our

systematic review: (i) the selection of the research question, (ii) bibliography, (iii)

database, (iv) the choice of appropriate terms, (v) selection criteria, (vi) the performance

of the review and (vii) synthesis of the results.

The following steps were performed according to a protocol previously defined.

2.1 Research questions and PICO’s analysis.

Papers addressing the issue of tax avoidance do so dominantly seeking the identification

of determinants and interactions, that may be endogenous or exogenous, aiming to build

or confirm theories that could help explaining the different levels of tax avoidance

211
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

observed and assess its consequences. To perceive the whole panorama that involves tax

avoidance we conceive the main research question in a broad way: What determines tax

avoidance and what are its consequences? Secondary issues were also defined, like the

definition of tax avoidance and the metrics used to assess the level of tax avoidance.

Following Petticrew and Roberts (2006), after selecting the research questions we

planned the topic analysis under a more detailed perspective in accordance to PICO

(Population , Intervention, Comparison and Outcomes) method (Table 1).

Table 1: PICO analysis

Population: What population am I interested in?
Scientific papers using tax avoidance.
Intervention: What intervention exactly am I interested in reviewing?
In the ways of assessing, in obtaining determinants and consequences of tax
avoidance.
Comparison/Control: What is the intervention being compared to?
Studies using tax avoidance.
Outcomes: Which results are relevant for the research question?
Obtaining the models, determinants and consequences that allow evaluating
the tax avoidance, in order to understand its applicability and impact.

2.2 Bibliography, databases and keywords.

In the research phase, the main objective is to only find information deemed relevant and

the choice of sources and the definition of keywords allow to outline the path to be

followed.

We search the Digital Library B-On ([Link]), the Business Source

Complete and the Science Direct database ([Link]).

Following Hanlon and Heitzman (2010), we define tax avoidance as the explicit

reduction of taxes. Accordingly, we used only ‘tax avoidance’ as keyword, considering

that other words, such as ‘aggressiveness’ or ‘sheltering’, serve to characterize a very

212
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

specific type of tax avoidance not allowing to capture different interpretations present in

different studies.

2.3 Applying practical screening criteria

As inclusion criteria scientific papers quoted in a journal up to the 3rd quartile of Scimago

Journal & Country Rank ([Link]) were chosen (Table 2). For the analysis

of Quartiles, the latest rankings available at the date of the research were used

(December/2018). Finally, the English language criterion was adopted to avoid the bias

caused by different languages or by certain preferences.

As exclusion criteria, papers analyzing tax avoidance in sectors such as

banking/insurance and public companies were excluded once they reflect very specific

areas, with their own rules and regulations, making the comparison with other sectors

difficult. We also excluded those papers exploring tax avoidance only in a theoretical

way, those whose focus was on individuals (and not companies) and those exploiting

taxes other than corporate income tax. Finally, papers that did not use statistical methods

were also excluded.

Table 2: Inclusion and exclusion criteria

Inclusion Criteria Type

1) Scientific papers up to the 3rd quartile of Journal
Scimago (reference year: 2018)
2) English Publication language

Exclusion Criteria Type

3) Studies of tax avoidance in sectors such as Research Design (sample)
banking, insurance, public companies and
agriculture.
4) Studies that only refer to tax avoidance in a Content
theoretical way.
5) Studies that do not focus on the use of the Content
concept of tax avoidance associated with
corporate income tax
6) Studies that do not use statistical methods of Method
data analysis.

213
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2.4 Performance of the review

To avoid bias or human error, the process of gathering data should be previously defined.

For the collection of information, Tranfield et al. (2003) recommend elements such as

sources of information, any characteristics, the context of the study, the evaluation of the

methodological quality of the study, the identification of the emerging themes, key results

and additional notes.

A data extraction form was constructed (Table 3), divided into five parts, where

the first records the bibliographic data of each article, namely the author, year, title and

journal, the second is the characteristics of the population, as is the case of the country

and the selection criteria, the third is the objective and theoretical framework, the fourth

the methodology and methods used and, lastly, the fifth part shows the findings.

The extraction form was applied to all papers and allowed the elaboration of a

synthetic framework.

Data have been systematized in tabulation tables, one of the most important steps

for narrative synthesis in order to summarize the information, making it accessible and

more transparent (Petticrew and Roberts, 2006)

214
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 3: Data extraction form

1) Bibliographic data

Author: Who is the author of the publication?

Year of study: In what year was the work published?
Title: What is the title of the publication?
Journal: In what Journal was it published?

2) Characteristics

Population Which population was chosen for analysis? (e.g. listed and/or
non-listed companies)
Sample Selection criteria: What was the selection criterion for the final sample? (e.g.
criteria for eliminating observations)
Country under study: Which country is under study?
Years in study: What years have been studied?

3) Context

Research objectives What is the research question?

Fundamental theoretical aspects or What is the theoretical framework of the study (summary)
framework given by the author: given by the authors?
Does the study relate tax avoidance Do the authors relate tax avoidance with other research fields
with other research fields? If so, (interactions)?
what fields?

4) Methodology and methods

Adopted method: What kind of method was used for data analysis? (E.G.:
Correlation and/or regression)
Hypotheses in study: Are there any hypotheses under study?
Equations and regressions What equations and regressions were developed, as well as
developed for the study of the the variables and proxies used to measure the tax avoidance
avoidance tax: and its determinants?
Control variables: What are the control variables used in the models?

5) Results:

Main conclusions What conclusions have been obtained?

Identified limitations What are the limitations identified by the authors of the
study?
Future research/ recommendations What are the recommendations for future research made by
the authors?
Additional notes Space to withdraw notes about possible concepts present in
the paper or space to annotate definitions that do not come in
the paper, but which are extracted from the Internet and which
are useful for it interpretation. (e.g: Concept of rent).

3. Data analysis, results and discussion

This chapter is the 7th task of the review. First, we performed the bibliographic analysis.

215
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

After, we analyzed the sample’s characteristics as well as the statistical methods used

and, finally, the findings.

3.1 Bibliographical data analysis

A total of 951 papers were obtained, 284 manually from digital library B-on database

(once this database did not allow the search through keyword), 583 papers (61%) through

keyword, from Business Source Complete and 84 (9%) from Science Direct.

The criterion that led to the greatest exclusion of papers was the criteria that

consider tax avoidance in relation to corporate income tax and not the other taxes. Papers

that adopt mathematical models were excluded, however, we decided to include those

that were composed of both mathematical and empirical models.

At the end, a total of 123 scientific papers were selected (see Figure 1).

Figure 1: Papers selection diagram

Publications go from 2003 to 2018, being this last year the one with most papers

216
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

published. The growing trend in the number of publications over time shows the interest

that the topic has been gaining within the academic and scientific communities (see Figure

2).

Figure 2: Publications by year

30 26 28
25
20 17
14 13
15
9
10 5 6
5 1 2 1 1
0
2003 2006 2008 2010 2011 2012 2013 2014 2015 2016 2017 2018

Number of publications by year

As for the geographic origins, based on the first author, 78 papers were from the USA, 14

from Australia, 4 from each Canada and China, 3 from each Germany, UK and Singapore,

and the remaining countries with only 2 or 1 paper. It has to be highlighted the only 13

papers from Europe (only 10 have European Union countries as a sample).

As for universities, the Duke University and the University of Kansas, both from

the USA, have the highest number of publications (5), followed by Curtin University of

Australia and Texas A&M University of the USA, (4 ). Only one first author (Hasan, Hoi,

Wu and Zhang, 2014) does not have university affiliation, suggesting that the topic is

mainly researched by academics.

Papers were published in 38 different scientific journals, most of which from the

accounting and taxation research fields. Almost half (highlighted in Table 4) were

published in just four journals (American Taxation Association, Accounting Review,

Journal of Accounting & Economics, Journal of Financial Economics and National Tax

Journal). As for the remaining, journals like Applied Economics, Decisions Support

217
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Systems, Journal of Business Ethics and Journal of Financial Crime, suggest that tax

avoidance is transversal to other research areas.

The breakdown by ranking quality shows 19 journals in the 1st quartile, 16 in the

2nd quartile and only 3 are in the 3rd quartile.

218
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 4: Journals and publications per year

Journals 2003 2006 2008 2010 2011 2012 2013 2014 2015 2016 2017 2018 Total
ABACUS 1 1
ACCOUNTING AND BUSINESS
RESEARCH
1 1
ACCOUNTING AND FINANCE 1 1 2
ACCOUNTING FORUM 1 1
ACCOUNTING HORIZONS 1 1
ACCOUNTING REVIEW 1 1 4 1 2 1 3 4 1 18
ADVANCES IN ACCOUNTING 1 1
APPLIED ECONOMICS 4 4
BRITISH ACCOUNTING REVIEW 1 1
CONTEMPORARY ACCOUNTING
RESEARCH
1 1
CHINA JOURNAL OF ACCOUNTING
RESEARCH
1 1
CORPORATE GOVERNANCE-AN
INTERNATIONAL REVIEW
1 1
CRITICAL PERSPECTIVES ON
INTERNATIONAL BUSINESS
1 1
DECISION SUPPORT SYSTEMS 1 1
ECONOMIC MODELLING 1 1
EURASIA BUSINESS REVIEW 1 1
INTERNATIONAL REVIEW OF
FINANCIAL ANALYSIS
1 1
INTERNATIONAL TAX AND PUBLIC
FINANCE
1 1
JOURNAL OF ACCOUNTING AND
PUBLIC POLICY
1 1
JOURNAL OF ACCOUNTING,
AUDITING AND FINANCE
1 1
JOURNAL OF ACCOUNTING &
ECONOMICS
2 4 1 1 8
JOURNAL OF BANKING & FINANCE 1 1 1 3
JOURNAL OF BUSINESS ETHICS 2 1 1 1 5
JOURNAL OF BUSINESS FINANCE &
ACCOUNTING
3 1 4
JOURNAL OF BUSINESS RESEARCH 1 1 2
JOURNAL OF CORPORATE FINANCE 1 1 2 4
JOURNAL OF CONTEMPORARY
ACCOUNTING AND ECONOMICS
1 1 1 1 4
JOURNAL OF FINANCIAL ECONOMICS 1 1 1 1 2 6
JOURNAL OF FINANCIAL CRIME 1 1
JOURNAL OF PUBLIC ECONOMICS 1 1
JOURNAL OF INTERNATIONAL
ACCOUNTING, AUDITING AND 2 1 3
TAXATION
JOURNAL OF THE AMERICAN
TAXATION ASSOCIATION
2 4 3 3 3 3 3 21
MANAGEMENT SCIENCE 1 1
NATIONAL TAX JOURNAL 1 4 1 6
PACIFIC-BASIN FINANCE JOURNAL 1 1
REVIEW OF ACCOUNTING STUDIES 1 3 2 6
REVIEW OF QUANTITATIVE FINANCE
AND ACCOUNTING
1 1 1 3
SUSTAINABILITY 3 3
Total 1 2 1 1 5 6 14 9 13 17 26 28 123

219
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Histcite, a bibliometric software used for the analysis and visualization of information,

allows us to evaluate the most cited authors outside and within the perimeter of the

review. From the 123 research works, 85 (69%) cite other publications in the sample. The

most cited authors are Dyreng, Hanlon and Maydew (2008) with 57 citations, Desai and

Dharmapala (2006) with 48 citations, Dyreng, Hanlon and Maydew (2010) with 34 and

Rego (2003) with 30 citations (see Table 5).

The first two most cited works both explore ways to measure tax avoidance.

Dyreng et al. (2008) proposed CashETR and Desai and Dharmapala (2006) created

discretionary BTD, both used to quantify tax avoidance, which may explain why these

two authors are so widely cited. Rego (2003) is also cited several times to justify the

choice of the effective tax rate (ETR) as a metric for tax avoidance level.

Table 5: Papers analysed in HistCite.

Notes: LCR: Local Cited References, number of times the article cited local papers. LCS: Local Citation Score, number of times the
article was quoted locally. CR: Cited References, Number of references cited by article. GCS: Global Citation Score, number of
citations of the article in all sources reported in the Web of Science or Scopus.

Year ID Author(s) LCR LCS CR GCS

2003 1 Rego SO 0 30 43 162
2006 2 Desai MA, Dharmapala D 0 48 51 283
3 Desai MA, Foley CF, Hines JR 0 3 25 141
2008 4 Dyreng SD, Hanlon M, Maydew EL 1 57 55 264
2010 5 Dyreng SD, Hanlon M, Maydew EL 3 34 46 208
2011 6 Taylor G, Tower G, Van Der Zahn M 0 0 51 5
7 Donohoe MP, McGill GA 0 4 56 18
8 Krishnan GV, Visvanathan G 0 2 55 26
9 Lim Y 1 1 50 16
10 Kim JB, Li YH, Zhang LD 2 20 55 169
2012 11 Lim Y 0 1 43 5
12 Blaylock B, Shevlin T, Wilson RJ 1 5 36 35
13 McGuire ST, Omer TC, Wang DC 4 14 65 74
14 Wahab NSA, Holland K 3 1 70 15
15 Cheng CSA, Huang HH, Li YH, Stanfield J 5 13 45 45
16 Atwood TJ, Drake MS, Myers JN, Myers LA 4 4 59 24
2013 17 Chen E, Gavious I, Yosef R 0 0 60 3
18 Richardson G, Taylor G, Lanis R 0 0 59 5
19 Taylor G, Richardson G 0 0 58 5
20 Chang L-L, Hsiao FD, Tsai Y-C 0 0 35 2
21 Abernathy JL, Davenport SA, Rapley ET 0 1 43 9

220
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

22 Boone JP, Khurana IK, Raman KK 0 4 59 38

23 Khurana IK, Moser WJ 0 6 45 26
24 Robinson LA, Schmidt AP 0 0 60 10
25 Dyreng SD, Lindsey BP, Thornock JR 0 3 42 25
26 Hill MD, Kubick TR, Lockhart GB, Wan HS 3 0 42 4
27 Hoi CK, Wu Q, Zhang H 2 7 68 73
28 Hope OK, Ma M, Thomas WB 6 7 86 37
29 Badertscher BA, Katz SP, Rego SO 4 7 72 33
30 Makoto Hasegawa, Jeffrey L. Hoopes, Ryo Ishida, Joel Slemrod n/a n/a n/a n/a
2014 31 Taylor G, Richardson G 0 0 52 2
32 Inger KK 0 2 54 7
33 Francis BB, Hasan I, Wu Q, Yan M 0 2 57 21
34 Gupta S, Mills LF, Towery EM 0 4 58 10
35 Crabtree AD, Kubick TR 0 0 34 4
36 De Simone L, Robinson JR, Stomberg B 0 0 20 16
37 Graham JR, Hanlon M, Shevlin T, Shroff N 2 18 53 83
38 McGuire ST, Wang DC, Wilson RJ 6 7 44 29
39 Hasan I, Hoi CK, Wu Q, Zhang H 9 10 60 58
2015 40 Salihu IA, Annuar HA, Sheikh Obid SN 0 0 73 3
41 Richardson G, Taylor G, Lanis R 5 0 69 5
42 Donohoe MP 5 1 98 10
43 Kubick TR, Lynch DP, Mayberry MA, Omer TC 4 4 66 12
44 Lanis R, Richardson G 5 2 96 20
45 Dyreng SD, Lindsey BP, Markle KS, Shackelford DA 1 0 31 10
46 Taylor G, Richardson G, Taplin R 6 0 69 5
47 Armstrong CS, Blouin JL, Jagolinzer AD, Larcker DF 3 13 53 62
48 Gallemore J, Labro E 3 5 66 30
49 Watson L 5 1 39 8
50 Jackson M 2 0 26 7
51 Mayberry MA, McGuire ST, Omer TC 3 0 47 0
52 Ravenda D, Argiles-Bosch JM, Valencia-Silva MM 5 0 78 0
2016 53 Habib A, Hasan MM 3 0 65 2
54 Col B, Patel S 0 0 0 2
55 Evertsson N 0 0 47 3
56 Davis AK, Guenther DA, Krull LK, Williams BM 5 1 62 16
57 Olsen KJ, Stekelberg J 6 3 70 7
58 Finley AR, Stekelberg J 3 0 54 1
59 Brooks C, Godfrey C, Hillenbrand C, Money K 8 0 84 4
60 Richardson G, Wang B, Zhang XM 5 0 75 2
61 McGuire ST, Neuman SS, Olson AJ, Omer TC 2 1 40 1
62 Goh BW, Lee J, Lim CY, Shevlin T 6 2 71 9
63 Kubick TR, Lynch DP, Mayberry MA, Omer TC 7 2 72 10
64 Huang HH, Lobo GJ, Wang C, Xie H 9 0 64 4
65 Gao L, Yang LL, Zhang JH 7 0 61 4
66 Green DH, Plesko GA 2 1 38 1
67 Gaertner FB, Laplante SK, Lynch DP 4 1 40 1
68 Henry E, Massel N, Towery E 7 2 25 4
69 Clausing KA 0 0 47 11
2017 70 Chen LH, Gramlich J, Houser KA 0 0 0 1

221
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

71 Alstadsaeter A, Jacob M 1 0 29 0
72 Baumann F, Buchwald A, Friehe T, Hottenrott H, Weche J 0 0 9 0
73 Young A 3 0 20 0
74 Schimanski C 0 0 9 0
75 Gao Z, Lu LY, Yu Y 0 0 0 0
76 Bird A, Karolyi SA 6 1 47 7
77 Guenther DA, Matsunaga SR, Williams BM 4 4 42 13
78 Cen L, Maydew EL, Zhang LD, Zuo L 7 0 72 2
79 Khan M, Srinivasan S, Tan L 8 0 45 7
80 Tang T, Mo PLL, Chan KH 0 0 50 0
81 Law KKF, Mills LF 7 1 118 3
82 Kubick TR, Lockhart GB, Mills LF, Robinson JR 5 0 53 2
83 Huseynov F, Sardarli S, Zhang W 4 0 44 0
84 Austin CR, Wilson RJ 4 0 33 2
85 Li OZ, Liu H, Ni CK 7 0 72 0
86 Kubick TR, Lockhart GB 9 1 72 1
87 Francis BB, Ren N, Wu Q 6 0 29 1
88 Dyreng SD, Hanlon M, Maydew EL, Thornock JR 7 2 70 12
89 Shevlin T, Thornock J, Williams B 3 0 39 0
90 Col B 1 0 47 1
91 Cook KA, Moser WJ, Omer TC 8 1 54 2
92 Hanlon M, Maydew EL, Saavedra D 6 0 63 3
93 Moore JA, Suh S, Werner EM 9 0 92 1
94 Seidman JK, Stomberg B 4 2 44 2
95 Huang HH, Sun L, Yu T 14 0 89 1
2018 96 Stewart J 0 0 57 0
97 Bradshaw M, Liao G, Ma MS 0 0 94 0
98 Lanis R, Richardson G, Liu C, McClure R 0 0 0 1
99 Thomsen M, Watrin C 0 0 69 0
100 Gul FA, Khedmati M, Shams SMM 0 0 97 1
101 Chyz JA, Gaertner FB 7 0 64 1
102 Lee N 1 0 38 0
103 Malik S, Mihm B, Timme F 0 0 51 0
104 McClure R, Lanis R, Wells P, Govendir B 7 0 84 0
105 DeZoort FT, Pollard TJ, Schnee EJ 4 0 80 0
106 Inger KK, Meckfessel MD, Zhou M, Fan WG 6 0 55 0
107 Frank MM, Lynch LJ, Rego SO, Zhao R 5 0 81 0
Lismont J, Cardinaels E, Bruynseels L, De Groote S, Baesens B,
108 7 0 44 1
et al.
109 Isin AA 6 0 139 0
110 Hsieh TS, Wang ZH, Demirkan S 7 0 67 0
111 Jimenez-Angueira CE 7 0 93 0
112 Duan TH, Ding R, Hou WX, Zhang JZ 9 0 43 0
113 Hardeck I, Wittenstein PU 1 0 73 0
114 Beladi H, Chao CC, Hu M 4 0 39 0
115 Jiang C, Kubick TR, Miletkov MK, Wintoki MB 8 0 54 0
116 Karamshahi B, Azami Z, Salehi T 5 0 47 0
117 Henry E, Sansing R 7 0 33 1
118 Hsu PH, Moore JA, Neubaum DO 7 0 70 0
119 McGuire ST, Rane SG, Weaver CD 1 0 52 0

222
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

120 Park S 2 0 24 0
121 Khurana IK, Moser WJ, Raman KK 8 0 48 0
122 Kanagaretnam K, Lee J, Lim CY, Lobo G 4 0 66 0
123 Gulzar MA, Cherian J, Sial MS, Badulescu A, Thu PA, et al. 11 0 46 1

A more in-depth analysis of the cross citations allowed to identify four clusters in the tax

avoidance research, suggesting clues about the way studies would be analyzed (see Figure

3). The first cluster refers to publications developing ways to quantify the tax avoidance,

the second and third cluster deal with, respectively, endogenous and exogenous

determinants and how they explain the different levels of tax avoidance and, lastly, the

fourth cluster refers to works identifying results explained by tax avoidance. This

structure will be adopted for the analysis of our sample.

223
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 3: Thematic areas on the study of tax avoidance.

Notes: Clusters: 1st red: Measures of tax avoidance. 2nd blue: Endogenous determinants. 3rd green: Exogenous determinants. 4th
Orange: Consequences

1st Red- 3rd Green:

2nd Blue: 4th Orange:
Measures of tax Exogenous Other papers
Endogenous determinants Consequences
avoidance determinants
1 (Rego, 2003) 5 (Dyreng et al., 2010) 7 (Donohoe & McGill, 36 (Graham, Hanlon, 10 (Kim, Li, & Zhang 2011)
2 (M. A. Desai & 13 (Mcguire, Omer, & 2011) & Shevlin, 2014) 3 (Desai et al., 2006)
Dharmapala, 2006) Wang, 2012) 28 (Hope, Ma, & 38 (Hasan, Hoi, Wu, 8 (Krishnan & Visvanathan,
4 (Dyreng et al., 2008) 15 (Cheng, Huang, Li, & Thomas, 2013) & Zhang, 2014) 2011)
12 (Blaylock, Shevlin, Stanfield, 2012) 33 (Gupta, Mills, & 60 (Mcguire, Olson, 16 (Atwood, Drake, Myers, &
& Wilson, 2012) 22 (Boone, Khurana, & Towery, 2014) & Omer, 2016) Myers, 2012)
Raman, 2013) 42 (Kubick, Lynch, 61 (Goh, Lee, Lim, & 31 (Inger, 2014)
23 (Khurana & Moser, 2013) Mayberry, & Omer, Shevlin, 2016) 87 (Dyreng, Hanlon, Maydew,
25 (Dyreng, Lindsey, & 2015) 76 (Guenther, & Thornock, 2017)
Thornock, 2013) 62 (Kubick, Lynch, Matsunaga, &
27 (Hoi, Wu, & Zhang, Mayberry, & Omer, Williams, 2017)
2013) 2016) 90 (Cook, Moser, &
29 (Badertscher, Katz, & 67 (Henry, Massel, & Omer, 2017)
Rego, 2013) Towery, 2016)
32 (Francis, Hasan, Wu, &
Yan, 2014)
37 (Mcguire, Wang, &
Wilson, 2014)
43 (Lanis & Richardson,
2015)
46 (Armstrong, Blouin,
Jagolinzer, & Larcker, 2015)
47 (Gallemore & Labro,
2015)
48 (Watson, 2015)
55 (Davis, Guenther, Krull,
& Williams, 2016)
56 (Olsen & Stekelberg,
2016)
93 (Seidman & Stomberg,
2017)

3.2 Samples and methods

Regarding the settings, 117 papers used data from a single country, from which 86 are

the USA, highlighting the dominance of this country when investigating tax avoidance

related issues. As for the remaining works, 7 used data from Australia, 6 from China and

224
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4 from Korea. Only 6 papers relayed on samples from more than one country (EU and

non-EU).

The time slot varies widely, Jackson (2015), the one with the greatest time span,

encompasses 34 years. On average, the time span is around 13 years, with the most

common being 10 and 18 years, with 9 publications each.

Due to the screening criteria, all the selected papers use statistical methods.

Nevertheless, statistical methods vary among works. Over 90% use multivariate statistical

methods, the most common being Multiple or Logistic Linear Regressions. Methods such

as the Two-Stage model of Heckman (1979), used in 4 papers, the Discontinuous

Regression (3 papers), the Propensity Score Match (2 papers), the Linear Probability

Model (2 papers) and the Quantile regression (1 paper) are also used. In another strand,

there were 2 experimental studies, 6 quasi-experimental studies, 3 works through only

Descriptive Analysis and a single article that used surveys to understand incentives and

disincentives for tax avoidance activities.

3.3 Tax avoidance definition and measurement

3.3.1 The concept of tax avoidance

Tax avoidance concept remained relatively stable over the years, with some nuances

resulting from the way authors intend to characterize or quantify it. The need to adjust the

concept to the subject under study led to different approaches, sometimes broadly, other

times in a stricter sense, limited to a certain kind of tax avoidance, and still, in some cases,

indirectly, when the concept is evoked through the variables used in the empirical

analysis. Regarding the broad definition of tax avoidance (Table 6), some are original

proposals from the authors but, in most cases, are quotations from other seminar authors

like Dyreng et al. (2008) and Hanlon and Heitzman (2010). These authors have taken an

225
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

important role in creating a generalized and accepted definition of tax avoidance which

is of great importance because, like Hanlon and Heitzman (2010, p.137) refer: ‘The

challenge for the area is that there are no universally accepted deﬁnitions of, or constructs

for, tax avoidance or tax aggressiveness; the terms mean different things to different

people’.

Table 6: Broad tax avoidance definitions

Author(s) Definition of Tax Avoidance

Rego (2003, Effective tax planning (a.k.a. tax avoidance) reduces the present value of tax payments and
p.808) generally increases the after-tax rate of return to investors in a firm.
It is important to emphasize that tax avoidance does not necessarily imply that firms are
engaging in anything improper. There are numerous provisions in the tax code that allow and/or
encourage firms to reduce their taxes. In addition, in practice there are many areas in which the
Dyreng et al.
law is unclear, particularly for complex transactions, and firms may take positions on their
(2008, p. 62)
returns in which the ultimate tax outcome is uncertain. For purposes of this study, we define tax
avoidance broadly as anything that reduces the firm’s cash effective tax rate over a long
time period, i.e., ten years.
Taylor et al. Tax avoidance is defined as any activity or strategy that reduces a firm’s taxes relative to its
(2011, p. 34) pre-tax accounting income.
I define tax avoidance as a tax savings that arises from both the general tax reduction methods
and tax shelters that are occasionally of questionable legality to minimize tax liability. In other
Lim (2011, p.456)
words, the tax avoidance measure conceptually captures the cumulative number of
transactions to minimize tax liabilities (Desai and Dharmapala, 2006).
Hanlon and Heitzman (2010, p. 137) state, ‘'If tax avoidance represents a continuum of tax
Kim et al. (2011, planning strategies where something like municipal bond investments are at one end, then terms
p.640) such as "noncompliance", "evasion", "aggressiveness", and "sheltering" would be closer to the
other end of the continuum'. (p. 640)
Corporate tax avoidance is defined in this study as any transaction or event ('passive' or
'aggressive') that leads to a reduction in the amount of corporate taxes paid by a firm (see, e.g.,
Dyreng et al., 2008). Tax avoidance may be achieved through legitimate methods in accordance
Taylor and
with tax legislation provisions. In fact, tax reduction methods may be either passive (complying
Richardson (2014,
with tax provisions) or aggressive (structuring transactions or activities with one of the principle
p.1)
objectives to decrease the amount of corporate taxes). Tax avoidance may alternatively be
achieved through illegal means or means that are not in compliance with tax legislation
provisions. These particular methods constitute tax evasion.
We define tax avoidance broadly as firms’ activities to report less taxable income and to reduce
Li et al. (2017, p.
taxes paid per unit of accounting earnings. Consistent with the literature (Chen et al., 2010; and
697)
Cheng et al., 2012), we do not distinguish between legal and illegal tax avoidance activities.
The tax literature consistently distinguishes between tax avoidance and tax evasion. The
Internal Revenue Service (IRS 2014) indicates that tax avoidance refers to legal (and therefore
DeZoort et al. legitimate) efforts to reduce or minimize tax liability, while tax evasion refers to illegal efforts to
(2018, p.88) reduce tax liability by deceit, subterfuge, or concealment. Our study focuses on legal tax
avoidance methods rather than illegal tax evasion methods because the literature clearly
establishes that illegal acts are unfair and unethical.

The definition of Dyreng et al. (2008), although broad in scope, is conceptually limited

by the notion of CashETR (cash effective tax rate), a measure of tax avoidance that results

226
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

from the taxes actually paid and not from the tax amount that is recognized in the financial

statements as an expense.

The definition proposed by Hanlon and Heitzman (2010, p. 137) (“If tax

avoidance represents a continuum of tax planning strategies where something like

municipal bond investments are at one end (lower explicit tax, perfectly legal), then terms

such as ‘noncompliance’, ‘evasion’, ‘aggressiveness’, and ‘sheltering’ would be closer to

the other end of the continuum. A tax planning activity or a tax strategy could be

anywhere along the continuum depending upon how aggressive the activity is in reducing

taxes”) is the most comprehensive and presents several advantages because, on the one

hand, it standardizes the concept of tax avoidance through a broadly accepted definition

and, on the other hand, it allows researchers to justify the different level of tax avoidance

aggressivity they find in their researches, making it possible to create a comparable basis

among different research works.

There are conceptual differences regarding tax evasion practices and tax

avoidance. Both activities aim to reduce the amount of tax payable, but the way this is

achieved may raise more or less doubts on its legality. Using tax avoidance practices, the

way to reduce the tax payable is to perform transactions that are legal or, in the worst

case, dubious, when in the grey zone of tax legislation. On the contrary, tax evasion

consists in conducting operations always considered illegal and sanctionable.

These two practices are often referred in literature, but few times are differentiated

given the difficulty in delimiting the legality/illegality (only determined ex-post by the

courts) borders of an operation through the use of indicators or models. An exception

found is the research by DeZoort, Pollard, and Schnee (2018), who, by means of an

experimental study, distinguish the two concepts within students.

In the impossibility of a rigorous differentiation, researchers often choose to use

models or indicators that set different levels of tax avoidance that may range from the less

227
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

to the most aggressive strategies, assuming the latter as being the threshold between

legality and lawlessness.

3.3.2 Measurement of tax avoidance

Tax avoidance measurement consists in choosing the metric(s) or model(s) that best fits

the research objectives. The use of different metrics/models in the same study may lead

to inconsistent results.

As activities of tax avoidance respect only those resulting from the book-tax gap1

(designated non-conforming tax avoidance) the use of metrics/models present this

limitation. The non-conforming tax avoidance is the result of activities that substantiate

the realization of tax planning operations in a continuous manner, and which are

accounted in different ways for accounting and tax purposes, with the objective of

reducing the taxes to pay (Hanlon & Heitzman, 2010). Most of the indicators used cannot

capture the effect of tax planning strategies that directly reduce the accounting income

(conforming tax avoidance).

Over the years, several metrics have been developed to quantify the non-

conforming tax avoidance, mainly ratios, and currently there are two large groups: the

global indicators of tax avoidance activities (Table 7) and the specific indicators of certain

tax avoidance activities (Table 8).

One of the most used metrics for tax avoidance, is the effective tax rate (ETR), a

ratio that allows capture the average rate of tax per monetary unit of income. Depending

on the numerator used, there are different ways to compute this rate. Some authors

consider the total income tax expense (GAAP ETR - captures the effect of accounting

1
The Book-Tax gap or Book-Tax Differences (BTD) is ‘…the difference between profits reported to tax
authorities and those reported to capital markets’. (Desai, 2005, p. 172)

228
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

earnings), others adopt the current tax (Current ETR – does not capture the effect of

deferred taxes) and others still consider the amount actually paid (CashETR or Lon-run

CashETR – The value per year or the sum of several years may be used). In all cases the

denominator used is a pre-tax accounting income.

The numerator may or may not reflect the impacts of the accounting earnings and

deferral strategies, and for both national or global scales, only a few numerators allow to

combine the effects of taxes on the income in all jurisdictions in which the company

carries out business (Dyreng et al., 2008). In both cases, the difference that is captured

essentially reflects the permanent differences, and in the case of the Current ETR

indicator, it measures only the permanent differences, that is, it measures a tax avoidance

normally associated with the most aggressive strategies. However, both measures capture

the effects of the differences generated by Book-Tax Differences (BTD) and the existence

of income transfer to jurisdictions with the lowest tax level.

Regarding the use of the ETR indicator, Rego (2003) succeeded in demonstrating

that companies that avoid taxes tend to reduce their taxable income while maintaining the

accounting income, which generates smaller ETR. Thus, this indicator is useful since it

manages to capture all the activity of non-conforming tax avoidance, that is, it allows to

evaluate the effectiveness of the strategies of tax planning.

Alternatively to ETR, some researchers choose to use the CashETR ratio, that has

the advantage of not being skewed by changes in tax accounting accruals. Moreover,

CashETR reflects any tax avoidance activity that reduces the amount of taxes paid in the

current period, including those that differ from the payment of taxes resulting from

temporary differences (Dyreng et al., 2008).

Some measures apply to a short period, usually one-year, other measures seek to

address more extended temporal periods, such as five or ten years. Dyreng et al. (2008)

used the Long-Run CashETR measure, that corresponds to the sum of cash taxes paid

229
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

over a long period of time, and concluded that companies behaviour is better captured

when the time period is extended (e.g. companies that tend to have low CashETR keep

these values persistently over time, compared to companies that have high CashETR that

end up reversing the situation).

As for effective rates, Dyreng et al. (2017) concluded that there was a progressive

decrease in effective tax rates in the last 25 years in the USA, and this is not only observed

in multinational companies but also in domestic companies, where tax rates sometimes

verified a higher decrease, contrary to expectations.

One of the main drawbacks of measures so far presented concerns the difficulty

of addressing companies with negative pre-tax income and/or negative tax expense, due

to the difficulty of interpretation of data obtained. The exclusion of firms with losses was

appointed as a limitation by Henry and Sansing (2018), and led to the development of the

adjusted CashETR, which allows to include all companies, making the interpretation

possible of the ratio. The adjusted CashETR ratio allows to perceive whether a company

is favoured or disadvantaged, when the amount of taxes paid is respectively less than, or

greater than, the pre-tax book income multiplied by the statutory rate. With this measure,

Henry and Sansing (2018) was able to replicate the study of Dyreng et al. (2017) with all

companies and conclude that, contrary to previous findings, there is no global tax-

favoured, over the last 27 years. The same conclusion was reached by the separation

between domestic and multinational firms and it was concluded that domestic firms are

the most disadvantaged. For the authors, this disadvantage is consistent with the

asymmetric treatment that the tax system gives to profits and losses so the tax analysis of

the whole system should include all firms and not only profitable firms.

An alternative to ratios may be the BTD metric, which is given by the difference

between the pre-tax book income and taxable income. A high BTD is associated with tax

avoidance and with the income manipulation and this indicator may be a good clue to the

230
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

analysis of the company's future earnings (Blaylock, Shevlin, & Wilson, 2012; Jackson,

2015). Jackson (2015) deepened the study of BTD through the analysis of its components,

namely the temporary BTD, which refers to deferred taxes, and the permanent BTD,

which results from the difference between the total and temporary BTD, seeking to find

evidence about the significance that each component has to the company's future

economic performance. He concluded that temporary BTD serves as a signal for future

changes in economic performance (earnings persistence), while the permanent BTD

contains information on future changes in tax expenses, being this relationship stronger

in the presence of firms with tax avoidance practices.

Desai and Dharmapala (2006) developed a model, by means of a regression, called

the total discretionary BTD or the abnormal BTD, that accounts the part of it not

explained by the differences between accounting and taxation. This model shows the

effects of accounting practices flexibility that allow managers to take advantage from it,

favouring tax planning when in presence of incentives to do so.

For Frank, Lynch, and Rego (2009), the use of discretionary BTD, CashETR and

the total BTD have some disadvantages once they are affected by the temporary effect

from the non-discretionary BTD which, sporadically, results from activities of tax shelter

and, as such, does not reflect aggressive tax reporting. Accordingly, the authors developed

an alternative measure called DTAX, which results from the regression of the total

permanent differences of non-discretionary items (which are known as responsible for

permanent differences) and legal adjustments that are not related to aggressive tax

reporting. The regression error reflects the discretionary permanent differences. For

Hanlon and Heitzman (2010), the measure created by Frank et al. (2009) is no more than

the differences between rates and can be calculated by the difference between the legal

tax rate and the effective tax rate (Differential ETR). They defend that DTAX has some

disadvantages once it fails to capture the conforming tax avoidance, the tax deferral

231
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

strategies and the global tax avoidance (the latter conditioned by different country rates

because, as Schimanski (2017) refers, there may exist different rates within a country due

to individual agreements between companies and States being its use very restrictive).

Thomsen and Watrin (2018) used Differential ETR and found that, although the

average value of ETR is higher when compared to the legal rate, in the USA, the

difference between rates is higher in USA than in countries of Europe. This means that

ETR in European countries has declined due to the decline in the legal rate of each country

and not due to higher tax avoidance.

Kim et al. (2011) adopted a different strategy for measuring tax avoidance by

using the combination of three indicators previously described: The total BTD, the

Differential ETR and the Abnormal BTD. Through factorial analysis, it was possible to

withdraw the common factor to all and capture the trend of companies that avoid taxes.

Finally, as a metric for very extreme tax avoidance practices, there are the

equations of Wilson (2009) and Lisowsky (2010), which measure the probability of the

companies resorted to the tax shelters. Wilson (2009) used data from companies with trial

in USA courts due to the use of tax shelters and Lisowsky (2010) used confidential data

from the Internal Revenue Service, intending to expand Wilson’s model by the inclusion

of a greater number of predictors. Both models have as a disadvantage the fact that their

application is limited to USA companies, and, in the case of Lisowsky’s model, this

limitation extends to the data needed for predictors that are partly confidential.

232
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 7: Global tax avoidance metrics

Measure Formula Description

𝑇𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡 Total tax expense per

GAAP ETR1,2 𝐺𝐴𝐴𝑃 𝐸𝑇𝑅𝑖𝑡 = monetary unit of pre-tax
𝑃𝑟𝑒𝑡𝑎𝑥 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡 book income

𝐶𝑎𝑠ℎ 𝑡𝑎𝑥 𝑝𝑎𝑖𝑑𝑖𝑡 Total taxes paid per

CashETR 1,2,3 𝐶𝑎𝑠ℎ 𝐸𝑇𝑅𝑖𝑡 = monetary unit of pre-tax
𝑃𝑟𝑒𝑡𝑎𝑥 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡 book income

𝐶𝑎𝑠ℎ 𝑡𝑎𝑥𝑒𝑠 𝑝𝑎𝑖𝑑𝑖𝑡 Total taxes paid per

𝐶𝑎𝑠ℎ𝑅𝑎𝑡𝑖𝑜 =
Cash Ratio4 𝐸𝑥𝑡𝑟𝑎𝑜𝑟𝑑𝑖𝑛𝑎𝑟𝑦 𝑖𝑡𝑒𝑚𝑠 𝑎𝑛𝑑 monetary unit of pre-tax
(𝑃𝑟𝑒𝑡𝑎𝑥 𝑜𝑝𝑒𝑟𝑎𝑡𝑖𝑛𝑔 𝑐𝑎𝑠ℎ 𝑓𝑙𝑜𝑤𝑠𝑖𝑡 − ) operating cash-flows
𝑑𝑖𝑠𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑒𝑑 𝑜𝑝𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑠 𝑖𝑡

∆
𝐶𝑎𝑠ℎ𝐸𝑇𝑅𝑎𝑑𝑗 − 𝜏 =
𝑃𝑟𝑒 𝑡𝑎𝑥 𝑖𝑛𝑐𝑜𝑚𝑒
Adjusted Represents the extent to
CashETR5 and 𝐶𝑎𝑠ℎ 𝑡𝑎𝑥𝑒𝑠 which the CashETR differs
𝑆𝑡𝑎𝑡𝑢𝑡𝑜𝑟𝑦 𝑡𝑎𝑥 𝑃𝑟𝑒𝑡𝑎𝑥 𝑏𝑜𝑜𝑘
Delta MVA6 ( − ∗ ) from the tax rate.
𝑝𝑎𝑖𝑑 𝑖𝑡 𝑟𝑎𝑡𝑒 𝑖𝑡 𝑖𝑛𝑐𝑜𝑚𝑒 𝑖𝑡
𝐷𝑒𝑙𝑡𝑎𝑀𝑉𝐴 =
𝑀𝑎𝑟𝑘𝑒𝑡 𝑉𝑎𝑙𝑢𝑒 𝐴𝑠𝑠𝑒𝑡𝑠𝑖𝑡

𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡 + 𝐹𝑜𝑟𝑒𝑖𝑔𝑛 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡 The total differences

Total BTD 𝑇𝑜𝑡𝑎𝑙𝐵𝑇𝐷𝑖𝑡 = 𝑃𝑟𝑒𝑡𝑎𝑥 𝑏𝑜𝑜𝑘 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡 − [ ]
𝑆𝑡𝑎𝑡𝑢𝑡𝑜𝑟𝑦 𝑡𝑎𝑥 𝑟𝑎𝑡𝑒𝑖𝑡 between book and taxable
(Hanlon)7
− ∆ 𝑁𝑒𝑡 𝑜𝑝𝑒𝑟𝑎𝑡𝑖𝑛𝑔 𝑙𝑜𝑠𝑠 𝑖𝑡 income

BTD 𝐷𝑜𝑚𝑒𝑠𝑡𝑖𝑐 𝐷𝑜𝑚𝑒𝑠𝑡𝑖𝑐 𝑆𝑡𝑎𝑡𝑒 𝑂𝑡ℎ𝑒𝑟 𝐸𝑞𝑢𝑖𝑡𝑦 𝑖𝑛

( − − − − )
(Manzon e 𝑖𝑛𝑐𝑜𝑚𝑒 𝑖𝑡 𝑇𝑎𝑥𝑎𝑏𝑙𝑒 𝑖𝑛𝑐𝑜𝑚𝑒 𝑖𝑡 𝑖𝑛𝑐𝑜𝑚𝑒 𝑡𝑎𝑥𝑒𝑠𝑖𝑡 𝐼𝑛𝑐𝑜𝑚𝑒 𝑡𝑎𝑥𝑒𝑠𝑖𝑡 𝑒𝑎𝑟𝑛𝑖𝑛𝑔𝑠 𝑖𝑡
Plesko)8 𝐵𝑇𝐷𝑖𝑡 =
𝐴𝑠𝑠𝑒𝑡𝑠𝑡−1

Total taxes that the

𝑝𝑟𝑒𝑡𝑎𝑥 𝑒𝑎𝑟𝑛𝑖𝑛𝑔𝑠 𝑏𝑒𝑓𝑜𝑟𝑒 ℎ𝑜𝑚𝑒𝑐𝑜𝑢𝑛𝑡𝑟𝑦 𝑠𝑡𝑎𝑡𝑢𝑡𝑜𝑟𝑦 company managed to avoid
Tax avoidance ( ∗ ) − 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑡𝑎𝑥𝑒𝑠 𝑝𝑎𝑖𝑑𝑖𝑡
𝑒𝑥𝑐𝑒𝑝𝑡𝑖𝑜𝑛𝑎𝑙 𝑖𝑡𝑒𝑚𝑠 𝑖𝑡 𝑡𝑎𝑥 𝑟𝑎𝑡𝑒 𝑖𝑡 by monetary unit of the
model9 𝑇𝑎𝑥𝐴𝑣𝑜𝑖𝑑 =
𝑝𝑟𝑒𝑡𝑎𝑥 𝑒𝑎𝑟𝑛𝑖𝑛𝑔𝑠 𝑏𝑒𝑓𝑜𝑟𝑒 𝑒𝑥𝑐𝑒𝑝𝑡𝑖𝑜𝑛𝑎𝑙 𝑖𝑡𝑒𝑚𝑠𝑖𝑡 pretax earnings before
exceptional items.
Sum of total tax expense
GAAP ETR ∑𝑁
𝑡=1 𝑇𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒 𝑖𝑡 over n years, divided by the
𝐺𝐴𝐴𝑃 𝐸𝑇𝑅𝑖𝑡 =
(> 1 year) ∑𝑁
𝑡=1 𝑃𝑟𝑒𝑡𝑎𝑥 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡
sum of the pre-tax book
income over n years
Sum of cash taxes paid over
Long-run CASH ∑𝑁
𝑡=1 𝐶𝑎𝑠ℎ 𝑡𝑎𝑥 𝑃𝑎𝑖𝑑𝑖𝑡 n years divided by the sum
ETR 𝐶𝑎𝑠ℎ 𝐸𝑇𝑅𝑖𝑡 =
(> 1 year)10 ∑𝑁
𝑡=1 𝑃𝑟𝑒𝑡𝑎𝑥 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡
of pre-tax book income over
n years
Notes:
1
The taxes GAAP ETR and CETR may take other denominators, for example, the use of operational cash flows. (See Salihu, Annuar,
Normala, & Obid, 2015).
2
There are authors who divide both ratios (GAAP ETR and CETR) by the rates applicable to the companies under analysis since not all
countries apply a single tax, which may vary according to industry, region, etc. (see Tang, Lai, & Mo (2017)). Another alternative may be
the adjustment of ratios in relation to the portfolio of companies located in the same quintile of total assets and the same industry. (see
Guenther et al. 2017).
3.
Some authors choose to change the variable in situations where the result before taxes is negative or in situations where there are refunds
for taxes. In these cases the authors choose to equalize CashETR: (1) to zero for companies receiving refunds, (2) to one for companies
with positive paid taxes and the result before negative taxes and (3) to one for companies whose ratio value is greater than one in order to
mitigate the distortions created by small denominators (See Jiménez-Angueira (2018)). There are other authors who choose to transform
the variable into a rate, by multiplying by-1 (See Gao, Yi, and Yangxin (2017)).
4
Indicator used by Cen, Maydew, Zhang, & Zuo (2017)
5
Measure developed by Henry and Sansing (2018) that represents the difference between the adjusted CETR of the tax refunds claims and
the statutory tax rate (τ). The variable (Δ) represents the cash tax avoidance measure scaled by pre-tax book income
6
Indicator developed by Henry et al. (2016), inspired by the adjusted CashETR of Henry e Sansing (2018).
7.
The indicator BTD is sometimes calculated without the variation of losses. However, the variation of losses allows to capture changes in
taxable income that are not reflected in the amount of current tax expense (Hanlon & Heitzman, 2010).
8.
Indicator developed by Manzon and Plesko (2012)
9.
Indicator developed by Atwood et al. (2012) that intends to capture the effect of the existence of different tax rates and its possible
management.
10
Measure created by Dyreng et al. (2008). In the analyzed papers, the time period used varies normally between 3 and 5 years. Some
authors choose to deflate the pre-tax income by adjusting special items.

233
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 8: Metrics for specific tax avoidance practices

Measure Formula Description

𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡 Total current tax expense

Current ETR1 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝐸𝑇𝑅𝑖𝑡 = per monetary unit of pre-tax
𝑃𝑟𝑒𝑡𝑎𝑥 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡 book income.
Sum of the total current tax
Current ETR1 ∑𝑁
𝑡=1 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒 𝑖𝑡 expense over n years divided
𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝐸𝑇𝑅𝑖𝑡 =
(> 1 year) ∑𝑁
𝑡=1 𝑃𝑟𝑒𝑡𝑎𝑥 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡
by the sum of pre-tax book
income over n years.
It consists of the factorial
The main component extracted from three different BTD measures: analysis of three indicators
BTD Factor2
Total BTD, ETR differential and Abnormal BTD simultaneously, where the
common factor is removed.
Reported Tax liability accrued for
Unrecognized Tax Value withdrawn from financial statement notes after FIN-48 taxes not yet paid on
Benefit (UTB) uncertain positions.
𝑃𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑 𝑈𝑇𝐵 = 0,004 + 0,011 ∗ 𝑃𝑇𝑅𝑂𝐴 + 0.001 ∗ 𝑆𝐼𝑍𝐸 + 0.010 ∗ 𝐹𝑂𝑅_𝑆𝐴𝐿𝐸 + 0.092
Predicted UTB3 ∗ 𝑅&𝐷 + UTB forecasting models.
0.002 ∗ 𝐷𝐼𝑆𝐶_𝐴𝐶𝐶 + 0.003 ∗ 𝐿𝐸𝑉 + 0.014 ∗ 𝑆𝐺&𝐴 − 0.018 ∗ 𝑆𝐴𝐿𝐸𝑆_𝐺𝑅
The difference between the
ETR differential4 𝐸𝑇𝑅𝑑𝑖𝑓 = 𝑆𝑡𝑎𝑡𝑢𝑡𝑜𝑟𝑦 𝑇𝑎𝑥 𝑅𝑎𝑡𝑒 − 𝐺𝐴𝐴𝑃 𝐸𝑇𝑅 country’s tax rate and the
company’s GAAP ETR.
𝐷𝑒𝑓𝑒𝑟𝑟𝑒𝑑 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡
Temporary BTD 𝑇𝑒𝑚𝑝𝐵𝑇𝐷𝑖𝑡 =
𝑆𝑡𝑎𝑡𝑢𝑡𝑜𝑟𝑦 𝑡𝑎𝑥 𝑟𝑎𝑡𝑒
Difference between the total
Permanent BTD 𝑃𝑒𝑟𝑚𝐵𝑇𝐷𝑖𝑡 = 𝑇𝑜𝑡𝑎𝑙𝐵𝑇𝐷𝑖𝑡 − 𝑇𝑒𝑚𝑝𝐵𝑇𝐷𝑖𝑡 BTD and the temporary
BTD
Calculation of probable tax
Current tax 𝐶𝑎𝑠ℎ 𝑃𝑎𝑖𝑑 𝑇𝑎𝑥 𝑏𝑒𝑛𝑒𝑓𝑖𝑡 𝐶ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑖𝑛𝑐𝑜𝑚𝑒 liabilities related to tax
Cushion Tax5 ∆ 𝐶𝑢𝑠ℎ𝑖𝑜𝑛𝑖𝑡 = ( − − − )
expense 𝑖𝑡 𝑓𝑜𝑟 𝑇𝑎𝑥𝑒𝑠 𝑖𝑡 𝑓𝑟𝑜𝑚 𝑠𝑡𝑜𝑐𝑘 𝑜𝑝𝑡𝑖𝑜𝑛 𝑖𝑡 𝑡𝑎𝑥𝑒𝑠 𝑝𝑎𝑦𝑎𝑏𝑙𝑒 𝑖𝑡 positions that may be
annulled.
The measure corresponds to
𝐵𝑇𝐷 the part that is not explained by
Abnormal BTD6 The residue of: = 𝛽𝑇𝐴𝑖𝑡 + 𝛽𝑚𝑖 + 𝜀𝑖𝑡
𝑇𝐴𝑖𝑡 the differences between
accounting and taxation.
The regression error reflects
The error of: 𝑃𝐸𝑅𝑀𝐷𝐼𝐹𝐹𝑖,𝑡 = 𝛼0 + 𝛼1𝐼𝑁𝑇𝐴𝑁𝐺𝑖,𝑡 + 𝛼2 𝑈𝑁𝐶𝑂𝑁𝑖,𝑡 + 𝛼3 𝑀𝐼𝑖,𝑡 + 𝛼4 𝐶𝑆𝑇𝐸𝑖𝑡 + the discretionary permanent
DTAX7
𝛼5 ∆𝑁𝑂𝐿𝑖,𝑡 + 𝛼6 𝐿𝐴𝐺𝑃𝐸𝑅𝑀𝑖,𝑡 + 𝜀𝑖,𝑡 differences.

𝑃𝑡𝑎𝑥 𝑠ℎ𝑒𝑙𝑡𝑒𝑟
𝑇𝑆𝑝𝑟𝑜𝑏 = 𝑙𝑛 = 𝛼 + 𝛽𝑋 + 𝜀
Prediction model of 1 − 𝑃𝑡𝑎𝑥 𝑠ℎ𝑒𝑙𝑡𝑒𝑟
Measure for the use of tax
tax sheltering de
shelter
Wilson8 𝑆ℎ𝑒𝑙𝑡𝑒𝑟 = −4,86 + 5,20 ∗ 𝐵𝑇𝐷 + 4,08 ∗ |𝐷𝐴𝑃| − 1,41 ∗ 𝐿𝐸𝑉 + 0,76 ∗ 𝐴𝑇 +
3,51 ∗ 𝑅𝑂𝐴 + 1,72 ∗ 𝐹𝑜𝑟𝑒𝑖𝑔𝑛 𝑖𝑛𝑐𝑜𝑚𝑒 + 2,43 ∗ 𝑅&𝐷

Prediction model of
expanded tax 𝑃𝑡𝑎𝑥 𝑠ℎ𝑒𝑙𝑡𝑒𝑟 Measure for the use of tax
𝑇𝑆𝑝𝑟𝑜𝑏 = 𝑙𝑛 = 𝛼 + 𝛽𝑋 + 𝜀
sheltering of 1 − 𝑃𝑡𝑎𝑥 𝑠ℎ𝑒𝑙𝑡𝑒𝑟 shelter
Lisowsky9

Notes:
1
Some authors choose to deflate the result before taxes by adjusting the special items.
2
Technique used by Kim et al. (2011).
3.
Model developed by Rego and Wilson (2012)
4
The authors Mcguire, Rane, and Weaver (2018)considered the difference between statutory tax rate and the foreign effective tax rate
in order to measure the incentives for shift income.
5
Indicator developed by Blouin and Tuna (2007)
6
The formula developed by Desai and Dharmapala (2006) uses the total accruals (TA) as a way to isolate the BTD component that is
affected by earnings management. Lim (2011) modified the formula using, instead of the TA, the discretionary accruals and the
performance-matched discretionary accruals.
7
The DTAX model was developed by the authors Frank et al. (2009). The PERMDIFF variable represents the difference between the
total BTD and the temporary differences and is calculated as follows:

𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑓𝑒𝑑𝑒𝑟𝑎𝑙 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡 + 𝐶𝑢𝑟𝑟𝑒𝑛𝑡 𝑓𝑜𝑟𝑒𝑖𝑔𝑛 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡 𝐷𝑒𝑓𝑒𝑟𝑟𝑒𝑑 𝑡𝑎𝑥 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑖𝑡
𝑃𝑒𝑟𝑚𝐷𝑖𝑓𝑓 = 𝑃𝑟𝑒 𝑡𝑎𝑥 𝑏𝑜𝑜𝑘 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡 − ( )− ( )
𝑆𝑡𝑎𝑡𝑢𝑡𝑜𝑟𝑦 𝑇𝑎𝑥 𝑅𝑎𝑡𝑒𝑖𝑡 𝑆𝑡𝑎𝑡𝑢𝑡𝑜𝑟𝑦 𝑇𝑎𝑥 𝑅𝑎𝑡𝑒𝑖𝑡

8
Measure developed by Wilson (2009) to measure the probability of a company being associated with a tax shelter. Where:

𝛽𝑋 = 𝛽1𝐵𝑇𝐷𝑖𝑡 + 𝛽2 𝐷𝐴𝑃𝑖𝑡 + 𝛽3 𝐿𝑒𝑣𝑖𝑡 + 𝛽4 𝑆𝑖𝑧𝑒𝑖𝑡 + 𝛽5 𝑅𝑜𝑎𝑖𝑡 + 𝛽6 𝐹𝑜𝑟𝑒𝑖𝑔𝑛 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡 + 𝛽7 𝑅&𝐷𝑖𝑡

9
Measure used by the authors Austin and Wilson (2017), developed by Lisowsky (2010) which, through the confidential data of the
Internal Revenue Service, allowed to create a model that measures the probability of a company being associated with a tax shelter.
The model was expanded from the model of Wilson, in which:

𝛽𝑋 = 𝛽1 𝐵𝑇𝐷𝑖𝑡 + 𝛽2 𝐷𝐴𝑃𝑖𝑡 + 𝛽3 𝐿𝑒𝑣𝑖𝑡 + 𝛽4 𝑆𝑖𝑧𝑒𝑖𝑡 + 𝛽5 𝑅𝑜𝑎𝑖𝑡 + 𝛽6 𝐹𝑜𝑟𝑒𝑖𝑔𝑛 𝑖𝑛𝑐𝑜𝑚𝑒𝑖𝑡 + 𝛽7 𝑅&𝐷𝑖𝑡 + 𝛽8 𝑇𝑎𝑥 𝐻𝑎𝑣𝑒𝑛𝑠𝑖𝑡 + 𝛽9 𝐿𝑎𝑔𝐸𝑇𝑅𝑖𝑡 +
18 26

𝛽10 𝐸𝑞𝐸𝑎𝑟𝑛𝑖𝑡 + 𝛽11 𝑀𝑒𝑧𝑧𝐹𝑖𝑛𝑖𝑡 + 𝛽12 𝐵𝑖𝑔 5𝑖𝑡 + 𝛽9 𝐿𝑖𝑡𝑖𝑔𝑎𝑡𝑖𝑜𝑛𝑖𝑡 + 𝛽14𝑁𝑂𝐿𝑖𝑡 + ∑ 𝛽𝑦 𝑦𝑒𝑎𝑟𝑖𝑡 + ∑ 𝛽𝑖 𝑖𝑛𝑑𝑖𝑡
𝑦=15 𝑖=19

234
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.4 Determinants and Consequences of Tax Avoidance

This chapter intends to analyse in greater depth factors that may explain (or result from)

tax avoidance, the three remaining clusters previously identified: endogenous

determinants, exogenous determinants and consequences.

As determinants we refer to internal or external characteristics that are associated

with tax avoidance as independent variables in statistical explanatory models that have

tax avoidance as dependent variable.

Most research works relates to the analysis on endogenous determinants which

can be explained not only by a real need to understand which internal factors explain

different levels of tax avoidance, but also for the difficulty in gathering and analysing

data concerning external factors (Figure 4).

As consequences we refer to all kind of factors as dependent variables in statistical

explanatory models that have tax avoidance as an independent variable. Only 29 works

of our sample addressed the consequences of tax avoidance practices and, as for the

remaining papers selected, only 7 analyse exclusively measures of tax avoidance.

Figure 4: Research works addressing determinants and consequences of tax avoidance

235
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Endogenous Determinants Exogenous Determinants

The Company Formal Factors

Drivers of Profit Shifting 1 3 18 25 42 45 46 Tax System and Tax Obligations 6 7 16 17 21 28 30

65 96 102 113 120 34 36 40 63 66 67 68
82 89 92 103 104 107 120
Ownership Structure and Corporate 2 15 23 29 31 38 47
Governance 60 73 76 79 80 83 85
93 94 97 111 118

Corporate Social Responsability 27 44 49 52 56 95 123

The Auditor-Provided Tax Services 8 13 48 51 55 58

and the Information Quality

Human Resources Informal Factors

Characteristics 5 22 33 55 57 70 71 Surrounding Enviroment 22 41 43 45 64 75 78

81 86 100 105 108 110 112 84 87 105 116 122
115

Tax Avoidance
Concept and Measures
1 2 4 10 12 50 74 88
99 117

Consequences

9 10 11 14 19 20 24 26
32 35 37 39 53 54 59 61
62 69 72 77 90 91 98 101
106 109 114 119 121

Notes:

Papers numbers correspond to table 5

Papers highlighted are used in more than one analysis

3.4.1. Endogenous Determinants

We found 59 papers with firm's internal characteristics that may explain tax avoidance

from which 44 papers with focus on the firms’ characteristics and the remaining 15 with

focus on the characteristics of human resources.

[Link]. Drivers of Profit Shifting The use of tax havens allows companies to shift

their income to low-tax jurisdictions, where companies can reduce their tax burden

(Dyreng, Lindsey, & Thornock, 2013; Stewart, 2018), or minimize dividend withholding

taxes and dividend taxes.

All these devices allow multinational companies to move their dividends between

the subsidiary companies and their parent companies through fiscally efficient channels

236
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

(Dyreng, Lindsey, Markle, & Shackelford, 2015), that allow exploring the legislative

differences and taking advantages in agreements with host countries (Rego, 2003).

Regarding the profile of the companies that do it, Desai et al. (2006) analysed the

affiliated companies of U.S. multinationals, concluding that the largest companies and

the most active internationally, as well as those with the highest R&D intensity and

significant volumes of intrafirm exchanges are those who are more likely to use tax

havens. Taylor, Richardson and Taplin (2015) also confirm this incidence in Australian

companies, where they found a positive association between the use of tax havens and

the variables intangible assets, transfer pricing, the term of interaction between transfer

pricing and intangible assets, withholding taxes, performance-based remuneration and

multinationalism.

The use of transfer prices does not always lead to artificially shift profits.

However, it is in companies that uses it that a higher level of tax avoidance was found

compared to others that do not make related party transactions (Park, 2018). Richardson,

Taylor, and Lanis (2013), when analysing transfer pricing aggressiveness, found that

company size, profitability, leverage, intangible assets and multinationalism are

positively associated with the tax avoidance activities.

As an alternative or complement to transfer prices, some multinationals use hybrid

combinations that allow them to exploit the differences in tax treatment given to an entity

or instrument, thus achieving a double non-taxation (Hardeck & Wittenstein, 2018), and

others use financial instruments, such as derivatives, which also allow companies to take

advantage of the complexity and tax law ambiguity that apply to them in order to decrease

their effective tax rates (Donohoe, 2015).

Intangible assets may also explain tax avoidance practices given its

characteristics, in particular, their intangibility and the inherent difficulty in valuing,

which favours profit shifting to jurisdictions with low tax rates or to jurisdictions with

237
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

incentives or tax benefits, such as tax credits or patent box regimes2 (Gao, 2016). Also,

combined with intangible assets, the investment in R&D recognized as expense, is usually

associated with the practices of tax avoidance (Lee, 2018).

Therefore, regarding the main characteristics of companies that adopt tax

avoidance strategies involving profit shifting, there are common factors, namely being

multinationals, having a high activity to benefit from different tax regimes in relation to

their intangible assets, transfer prices and withholding taxes.

[Link]. Ownership Structure and Corporate Governance: There are two

theoretical approaches to the way taxes are analysed concerning shareholders.

For the traditional one the payment of taxes is seen as an expense to the

shareholders reducing the value of the company and the return of their investments, so

the activities of tax avoidance represent a way to maximize the value of the company. In

these situations, managers are encouraged to promote tax planning activities, aligning

their interests with those of shareholders. Examples are Hedge Funds3, and SOEs4, where

there is an incentive to influence companies to pursue tax avoidance strategies, in order

to increase their cash flows after taxes (Cheng, Huang, Li, & Stanfield, 2012; Li, Liu, &

Ni, 2017) or the amount of dividends distributed (Tang et al., 2017).

Also, higher market exposition promotes higher tax avoidance given the pressure

of investors in obtaining greater returns. Analysing the move of companies from the

2
The ‘patent box’ regime consists in allocating tax benefits to income from the exploitation of industrial
property assets.
3
Consist of highly speculative investments that aim to monetize the invested capital.
4
Companies held mostly by the local or central government where taxes paid by these companies used to
be allocated to the company's holding government. From 2002, taxes are now being distributed equally
regardless of the government that hold them.

238
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

bottom of Russell 1000 index5 to the top of the Russell 2000 index, Khan, Srinivasan, and

Tan (2017) verified that this migration is accompanied by an increase of exposition and

institutional ownership, which yield an increase in tax avoidance activities. Similar

conclusion is that of Bird and Karolyi (2017) who also concluded that tax avoidance

activities is higher in companies with poor governance and with higher initial tax rates,

suggesting that the increase in institutional ownership leads companies towards a

common effective tax rate.

Likewise, the entrance in other indexes, as S&P 500, impact tax avoidance.

Huseynov, Sardarli, and Zhang (2017) concluded that there are asymmetric effects on tax

avoidance practices, decreasing when companies previously had high levels of tax

avoidance, and increasing when it was previously smaller, in the same sense with the

conclusion of Bird and Karolyi (2017). Companies reacts to the effects they anticipate,

being that for too high levels, there is a concern to mitigate possible future negative

impacts and, for lower levels, there is a willingness to explore some potential benefits

(Armstrong, Blouin, Jagolinzer, & Larcker, 2015; Bird & Karolyi, 2017).

Tang et al. (2017) concluded that when local Chinese governments have a high

control, they encourage tax avoidance activities to avoid sharing taxes with the central

government.

The alternative approach defends that tax avoidance is harmful to shareholders

because it allows protective shields that lead to situations of managerial opportunism and

diversion of rents, making the company opaque, with less control of managers

performance (Khurana & Moser, 2013). In this situation, the agency problem is

5
Stocks at the bottom of the Russell 1000 have a small weight in the portfolio as they represent the smallest
firms in the index, while, inversely, stocks at the top of the Russell 2000 have a high weight. Therefore,
investors prefer to invest at the top of Russell 2000 instead of the base of Russell 1000, even in the case of
similar stocks in terms of capitalization of the market in general.

239
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

compounded, leading to the manager seeking to extract benefits for himself to the

detriment of the shareholder's will6. This agency problem can be mitigated through a

greater entrenchment between shareholders and managers7. Mcguire et al. (2014),

through the analysis of dual class companies8, concluded that the detachment of voting

rights and the rights of cash flow of managers gives managers greater control and a greater

entrenchment, removing from them the pressure to increase the value of the company

through tax avoidance practices. However, in structures with dual entrenchment, as is the

case of the classified board structure and family firms, there is only a negative association

with tax avoidance, being one of them absent (Moore, Suh, and Werner, 2017).

In a long-term perspective, Badertscher et al. (2013) and Khurana and Moser

(2013) demonstrated that the greater the proximity between shareholders and managers

in decisions and in the interest in the company in the long term, the lower the probability

of tax avoidance due to increased restrictions imposed on the manager through increased

monitoring.

Under the agency theory, we can find other relationships that go beyond the

relationship between manager and shareholder. There are situations where shareholder

concentration allow big owners to have the effective power, so the agency conflicts are

often between the controlling owners and minority shareholders, as is the case of some

Chinese companies (Richardson, Wang, & Zhang, 2016). In these cases, the relationship

between shareholder concentration and tax avoidance practices assumes a non-linear U

form. At the lowest level, the increase in the owner’s concentration is positively

associated with tax avoidance due to the entrenchment effect. However, beyond the

6
The agency's theory was formalized by the authors Jensen and Meckling (1976).
7
The assertion is based on the theory of property separation and control of Fama and Jensen (1983).
8
The dual class companies are characterized by the existence of a separation between the exercise of power
over cash flows and voting power.

240
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

minimum level required for effective control, there is a negative association with tax

avoidance due to the alignment effect.

Therefore, it is possible to conclude that do not always exist a direct relationship

between corporate structure and tax avoidance practices, and that there are different

factors affecting it. One of them relates to the objectives of shareholders and the control

they exert as investors, once, more or less interventionist, they can exert an influence on

the decisions on the company corporate governance structure, avoiding agency problems

(Young, 2017).

Desai and Dharmapala (2006) gave an important seminal contribution for

understanding the impact that incentives have on managers. They argue that the increase

in equity incentives tends to reduce the level of tax avoidance in a way that it is consistent

with tax avoidance and the extraction of rents being complementary activities. This

relationship is more visible in poorly governed companies, which means that the link

between compensation incentives and tax avoidance practices is mediated by the

corporate governance characteristics.

Contrary to this conclusion, Taylor and Richardson (2014), from the analysis of

listed Australian companies, concluded that the equity incentives are positively associated

with the tax avoidance activities. Also Armstrong et al. (2015), starting from the findings

of Desai and Dharmapala (2006), refer that the reasons pointed out by these authors are

not linear, and concluded that, in presence of equity incentives, corporate governance

operates essentially at the extreme levels of tax avoidance. Seidman and Stomberg (2017)

also contest Desai and Dharmapala's theory considering that their findings may be

241
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

associated with tax exhaustion9 and not with the extraction of rents from high-powered

incentives on poorly governed companies.

Beyond equity incentives, there may be other incentives that could explain the

levels of tax avoidance as is the case of the need of maintaining the position (Bradshaw,

Liao, & Shuai, 2018), the characteristics of corporate boards, such as financial

sophisticated and independent boards. Armstrong et al. (2015) concluded that, the more

financial sophisticated and independent the boards are, the higher is their influence on

extreme positions of tax avoidance (they influence its increase when it is low and its

decrease when it is high). However for Hsu, Moore, and Neubaum (2018), the way some

boards work in their advisory and monitoring roles of tax planning strategies depends on

the company's business strategy.

In presence of these mixed results, Jiménez-Angueira (2018) also added the tax

enforcement to their study, and concluded that, in periods of high regulation, there was a

reduction of tax avoidance practices for companies that had a weak corporate governance

in the periods of weak regulation, supporting the positive association between tax

avoidance and weak corporate governance depending on the external monitoring

environment.

Albeit the relationship between institutional ownership, corporate governance and

tax avoidance are not linear, an alignment between shareholders and managers lead the

company to a certain level of tax avoidance that is accepted by both parties. When this

alignment doesn’t exist, in presence of a weak corporate governance structure or in a

weak regulatory environment, the manager is given a margin to use tax avoidance

9
The tax exhaustion theory argues that as the taxable income approaches zero and the marginal benefits
of tax avoidance decrease, taxpayers engage in less incremental tax avoidance (Seidman and Stomberg,
2017)

242
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

strategies for personal purposes. In these situations, literature shows mixed conclusions

about the effect of equity incentives on tax avoidance practices.

[Link]. Corporate Social Responsibility: Corporate Social Responsibility (CSR)

has been used in some research as a way of understanding the link between the voluntary

contribution of companies to increase social welfare and the payment of taxes as part of

this same responsibility. The analysis between the two variables was performed in seven

studies and was carried out based on three theoretical trends.

For a positive relation are evoked: the risk management theory10 and the slack

resource theory. The risk management theory considers that companies increase their

CSR activities to protect themselves from possible reputational risks or negative events,

thus focusing their activities on the main interests of shareholders and not on social

responsibility. The slack resource theory11 defends that, when experiencing a strong

performance, firms can allocate their resources in satisfying the needs of all stakeholders,

which may result in greater CRS contributions and tax payments.

For a negative relation between the CSR and the tax avoidance authors highlight

the stakeholder theory12 suggesting that companies have obligations to society and,

accordingly, should pay fair taxes to ensure the satisfaction of collective needs.

Consistent with the risk management theory, Gulzar et al. (2018), regarding

Chinese companies, found that there is a negative relationship between Corporate Tax

Payments (less corporate tax payments means high tax avoidance) and CSR activities.

Supporting slack resource theory, Watson (2015) found evidence that the

relationship between tax avoidance and CSR is moderate by earnings performance. In

10
See authors Godfrey (2005) and Minor and Morgan (2011)
11
See author Penrose (1959)
12
See author Clarkson (1995)

243
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

presence of lower profitability, companies tend to not allocate their resources to CSR

activities and use tax avoidance to pay less taxes. However, for Davis et al. (2016) CSR

and tax avoidance activities are not complementary but rather substitute.

According to stakeholder theory, Hoi et al. (2013) and Lanis and Richardson

(2015) through the analysis of the performance of CSR activities, concluded that

companies with lower levels of CSR have the highest tendency to avoid taxes. The

reputational concern is also evidenced in CSR activities, since the most socially

responsible companies are the ones less likely to expatriate the parent company for a low

or non-tax country (Huang & Sun, 2017), while the most irresponsible companies are the

ones which pay less taxes (Ravenda, Argilés-Bosch, and Valencia-Silva, 2015).

Mixed results and, in some cases, the absence of direct relations make it clear that

the relation of CSR and Tax Avoidance is complex, influenced by other intermediate

factors and can hardly be achieved only by its intersection.

[Link] The auditor-provided tax services and the information quality: Companies

with audited financial statements with the same audit firm providing audit and tax services

get benefits as it allows for cost savings and the concentration of knowledge into a single

entity13 albeit it also allows the loss of independence of the auditor that may impact the

audit quality.

Krishnan and Visvanathan (2011) analysed the relationship between this

concentration and tax avoidance practices, concluding that auditor-provided tax services

contribute to the increased quality of audit work, as accumulated knowledge allows to

13
Authors designate this accumulation of knowledge as knowledge spillover. ‘The knowledge spillover
argument suggests that by performing tax services auditors become more familiar with clients’ strategic
decisions regarding tax planning: a feature that beneﬁts the auditors in uncovering tax avoidance
policies…’ (Habib & Hasan, 2016, p.53).

244
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

detect aggressive strategies of tax avoidance and act immediately, so a direct impact is

not evident.

Investigating a real case involving the auditor KPMG, in which the auditee firm

was the subject of sanctions for having been caught in an improper conduct involving tax

shelters, Finley and Stekelberg (2016) documented that, after the sanctioning, the

probability that clients acquire auditor-provided tax services declined. In addition, they

found that the sanctioned companies had higher levels of tax avoidance comparing to

other companies, suggesting that, in some cases, the choice of an audit firm to provide

tax consultancy is for tax avoidance practices and can help in the elaboration of schemes

(Evertsson, 2016; Mcguire, Omer, & Wang, 2012).

Gallemore & Labro (2015) found that a high quality of internal information is

associated with a higher tax avoidance especially, in situations related to income shifting,

inversely, to external information which is written in an ambiguous manner in order to

conceal tax planning practices from tax authorities.

Mayberry, Mcguire, and Omer (2015) analysed the impact of the volatility of

taxable income on the ability to inform, having concluded that when there is a taxable

income smoothness, it has less value relevance14, suggesting that managers in tax

avoidance situations make taxable income less informative.

[Link]. Human Resources: Individual factors related to people also play a role

when explaining tax avoidance and the choice of the executives may reflect a desired tax

behaviour by its shareholders.

Some research focus on the impact caused by the admission of executives with

previous background related to tax avoidance activities. These entrances are usually

14
The term is used to designate the relevance and capacity of information of the result in the evaluation of
the company.

245
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

accompanied by an increase in tax avoidance and later, when they leave, by a decrease

(Dyreng et al., 2010; Jiang, Kubick, Miletkov, & Wintoki, 2018; Lismont, Cardinaels,

Bruynseels, & Groote, 2018). Thus, executives may have an important role in the level

of company's tax avoidance, depending on their previous experience and on how they

perceive and accept these practices from the ethical and legal perspectives (DeZoort et

al., 2018; Evertsson, 2016; Jiang et al., 2018).

Individual characteristics of executives are also associated with tax avoidance

behaviours due to their ability to influence company management15. Factors such as

overconfidence, powered by public recognition and media exposure (Duan, Ding, Hou,

& Zhang, 2018; Hsieh, Wang, & Demirkan, 2018; Kubick & Lockhart, 2017), the

existence of a high narcissism (Olsen & Stekelberg, 2016), a high tendency for managerial

acquisitive16 (Gul, Khedmati, & Shams, 2018), a prior connection to business via training

or through parents (Alstadsæter & Jacob, 2017) and the fact that they are men, translates

into a lower fear towards risk (Francis, Hasan, Wu, & Yan, 2014), which favours a higher

tax avoidance. Inversely, women, due to a more risk aversion willingness, are less likely

to be associated with fiscal aggressiveness (Francis et al., 2014), as well as executives

with military experience (Law & Mills, 2017) or belonging to the catholic religion

(Boone, Khurana, & Raman, 2013), because they consider that paying taxes is important

for social welfare.

Chen et al. (2017) concluded that a gender-diverse board is also negatively

associated with tax avoidance.

15
Studies that tend to focus on the personality of top leaders are based on the theory of upper echelons. The
central idea of this theory is that the organization is the reflection of its top managers and, as such, the
strategies of the organization and the results are strongly influenced by them, notably by their personal
characteristics.
16
An acquisitive manager carries out mergers and acquisitions with high frequency.

246
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.4.2. Exogenous Determinants

Besides internal factors, external factors to may also explain firms’ tax avoidance levels.

They may be formal or informal, being the former all situations of fiscal nature, that can

go from the tax system to tax enforcement, and the latter, circumstances related to the

surrounding environment. In our review 21 papers fit the first group and 12 papers the

second.

[Link]. Formal Factors - Tax system and tax enforcement: Differences between

the accounting and tax system, called BTD, allow managers to take advantage of these

differences to promote tax avoidance, leading some authors to propose an increase in

conformity (Book-Tax Conformity17) as a way to reduce tax avoidance (Atwood, Drake,

Myers, & Myers, 2012) while others argue that this conformity does not have to be so

high, since the decrease of just a few differences can be enough to reduce managers

opportunistic behaviour (Chen, Gavious, & Yosef, 2013; Simone, Robinson, & Stomberg,

2014).

The perception of tax enforcement and other mechanisms of control and

monitoring are seen as being relevant in the relationship between BTD and tax avoidance

(Atwood et al., 2012; Frank, Lynch, Rego, & Zhao, 2018). Kubick, Lockhart, Mills, and

Robinson (2017), found that companies that are situated near the IRS industry specialist

services present lower tax avoidance levels as a result of tax enforcement exercised by

these services. On the other hand, when that perception is weak, as is the case of tax

amnesties, managers interpret it as a decrease in tax authorities action, favouring the

increasing of fiscal aggressiveness (Shevlin, Thornock, & Williams, 2017)

With regard to monitoring mechanisms, provided by the scrutiny of Sarbanes-

17
See (Desai, 2005; Tang, 2015)

247
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Oxley Act of 2002 (SOX) and Securities and Exchange Commission (SEC), in the USA,

Frank et al., (2018) and Kubick et al. (2016) concluded that greater scrutiny led to a

decrease of the levels of tax avoidance, because it limits the ability to aggressive reporting

practices (Hope, Ma, & Thomas, 2013; Kubick et al., 2016).

In this way, the engaging fiscal system and the perception that managers have

concerning actual monitoring and tax enforcement are key points in choosing a given

location for taxing corporate profits. In countries where this perception is weak there are

greater incentives for shifting income, because there are greater opportunities for the

payment of fewer taxes (Hope et al., 2013; Salihu, Annuar, Normala, & Obid, 2015).

The existence of specific policies and laws, such as the payment of dividends with

tax credits, in Australia, and the creation of the Unfair Related Party Transactions Tax

Law, in Korea, allowed to restrict tax avoidance activities (McClure, Lanis, Wells, &

Govendir, 2018; Park, 2018). However, some laws may have a perverse effect, because

they lead to the adoption of more aggressive alternatives (e.g tax evasion) to achieve the

intended goal (Malik, Mihm, & Timme, 2018).

The creation of an enhanced informative environment requiring more tax

information disclosure is an important instrument of tax enforcement (Hasegawa,

Hoopes, Ishida, & Slemrod, 2013). In the USA, the introduction of new disclosure

obligations, such as Schedule M-3 in 2004, the FIN48 in 2006 and Schedule UTP in 2010,

led some authors to study their impacts on company’s tax avoidance behaviour. Schedule

M-3 consists of a three-part statement where the first party reconciles worldwide tax

payments with the accounting income and the second and third parties provide a detailed

reconciliation of accounting income of the entities with taxable income and require

taxpayers to disclose whether each accounting-tax difference is temporary or permanent.

Research on the impact of this measure yield mixed results. Donohoe and McGill

(2011) found that, even before its implementation, markets anticipated that it would

248
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

increase tax burden and compliance costs, and this perception was greater for companies

that had permanent and discretionary BTD. Hope et al. (2013) also found that, after the

implementation, differences between levels of tax avoidance decreased for all the

companies that had higher levels of tax avoidance.

Furthermore, Henry et al. (2016) demonstrated that, in the long term, the

introduction of Schedule M-3 is associated with high levels of tax avoidance for domestic

companies, but this relationship is weak for multinational companies concluding that, for

these companies, there is a perception of the usefulness of the information provided by

Schedule M-3.

However Gaertner, Laplante, and Daniel (2016) studied the specific source of

BTD, during the 2004-2013 years, and concluded that there was an increase in the BTD,

not resulting from tax planning but from companies’ operating and financing structures,

tax law changes or economy-wide events. Green and Plesko (2016) also concluded that

the BTD increased in value although its weight in the accounting income remained

unchanged.

The introduction of FASB interpretation Accounting for Uncertainty in Income

Tax Positions – FIN 48, aimed to respond to tax planning strategies. With this

requirement, companies became obliged to register liabilities for uncertain tax positions

to be sustained if challenged by tax authority that is presumed to have the full of

knowledge of all relevant information. Research on the effects of introducing this

requirement concluded that it contributed to decrease tax avoidance, showing that

companies are aware that those disclosures provide additional information about their tax

avoidance activities (Gupta, Mills, & Towery, 2014; Henry et al., 2016), although this

awareness may depend on manager’s evaluation (Simone et al., 2014).

After analysing the values declared in accordance with FIN48, Hanlon, Maydew,

and Saavedra (2017) found that, in both domestic and multinational companies, the

249
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

recognition of uncertain tax liabilities, derived from the tax avoidance, is associated with

a high level of cash balances, needed for any future payments of tax corrections.

After FIN48, Schedule UTP was created, requiring companies to disclose the level

of their uncertain tax positions and all uncertain positions accumulated to date. Analysing

the market impact of this requirement, Abernathy, Davenport, and Rapley (2013) found

a negative reaction, at an early stage although, with the final model, a positive reaction

was observed, allowing investors to become more relaxed. This conclusion is consistent

with that of Henry et al. (2016), who concluded that Schedule UTP is associated with

high levels of tax avoidance, so companies do not consider this regime to be informative

of tax avoidance activities.

Accordingly, one may concluded that, albeit tax disclosure requirements are

important instruments of tax enforcement, their effectiveness will always be conditioned

by the perception taxpayers have on their efficacy, and, besides, taxpayers can always

avoid or hinder disclosure, making information less transparent and informative (Gaertner

et al., 2016; Hasegawa et al., 2013).

In addition to local requirements, there are international tax requirements

demanding companies to provide a high level of information as a condition for obtaining

certain tax deductions as is the case of withholding taxes or thin capitalization regime

(Taylor, Tower, & Van Der Zahn, 2011).

Thus, an environment that requires companies to provide information to their

stakeholders should be constantly reinforced given the ability companies have to adapt

creating disclosure protection mechanisms.

[Link]. Informal factors - Surrounding Environment: Informal factors may also

explain tax avoidance, alone or together with formal factors. Social capital elements,

such as a culture of trust, religion and the way tax avoidance is perceived by the general

250
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

public, influence the way companies deals with tax avoidance making it less aggressive

as these factors have stronger expression (Boone et al., 2013; DeZoort et al., 2018;

Dyreng et al., 2015; Kanagaretnam, Lee, Lim, & Lobo, 2018).

Even so, there may be factors that reduce this relationship, such as the existence of high

earnings and capital market pressure on managers (Kanagaretnam et al., 2018;

Karamshahi, Azami, & Salehi, 2018), situations of financial constraints or the existence

of financial incentives for undertake risk (Gao et al., 2017) .

Moreover, the way companies interacts with their customers and suppliers affects the

level of tax avoidance. Huang, Lobo, Wang, and Xie (2016) demonstrated that a higher

concentration of customers leads to higher levels of tax avoidance due to cash flow needs.

Cen et al. (2017) has shown that the relationships between companies and its main

customers and dependent suppliers affect the levels of tax avoidance, due to the fact that

specific customer-supplier arrangements facilitate tax avoidance strategies.

The market relevance of a specific product, resulting from its unique characteristics, leads

companies to have greater incentives for the use of tax avoidance strategies that do not

represent a high risk, given their competitive advantage compared to other companies.

On the other hand, shareholders, having knowledge of this advantage, require fewer return

to maintain their shares (Kubick, Lynch, Mayberry, & Omer, 2015). Inversely, companies

with valuable brands are those that most avoid tax avoidance, thus also avoiding

reputational costs (Austin & Wilson, 2017).

Economic factors such as the economic crisis of 2008, led companies to use tax avoidance

to obtain the necessary cash resources. Bankruptcy Risk encouraged both shareholders

and managers to accept more aggressive tax practices (Richardson, Taylor, & Lanis,

2015). However, Francis, Ren, and Wu (2017) did not find different behaviours in

situations where the constrain is lower and argue that tax avoidance behaviour in context

of ﬁnancial constraints is not totally clear.

251
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.4.3. Consequences

In 29 papers of the sample, researchers used tax avoidance as an explanatory factor for

different topics under study. In the framework that we adopt in this review we present

those topics as consequences of tax avoidance. Tax avoidance practices entail

consequences that extends at various levels, namely, tax revenue, companies themselves

and their stakeholders.

Baumann et al. (2017) documented that tax avoidance levels explain the profit

shifting to subsidiaries located in more favourable tax regimes and weak enforcement

systems. It is estimated that the worldwide losses of tax revenues, might reach the $280

billion dollars and, in the USA, around $77 to $111 billion dollars, corresponding to 30%

of USA Corporate income tax revenues (Clausing, 2016).

At firm level, consequences are diverse and can go beyond tax issues. One of them

is the effect on the reputation of the company and its managers. On a survey, tax

executives manifested concern about the impact of tax avoidance on company's

reputation. This concern was the second main cause for not adopting aggressive strategies

of tax avoidance, which was more evident in publicly traded companies, larger

companies, more profitable companies and companies in retail industry (Graham,

Hanlon, & Shevlin, 2014). In the case of companies that choose to open tax haven

affiliates, the negative impact on image is offset by the increase in CSR activities after

two years (Col & Patel, 2016).

At management level reputational consequences is not clear. There are situations

in which tax avoidance practices are beneficial for directors and managers who see their

reputation improved, which translates into the maintenance of their jobs (Lanis,

Richardson, Liu, & Mcclure, 2018). But it may also be harmful, once Chyz & Gaertner,

(2018) verified a forced CEO turnover, albeit it could also be linked to an increased

252
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

scrutiny of aggressive tax practice by regulators and tax authorities.

Firms valuation could be affected by the tax avoidance, depending on the

perception that shareholders have about their benefits or harm (Inger, 2014), having an

impact on companies and shareholders. There are cases in which tax planning negatively

weighted valuation (Wahab & Holland, 2012) and consequently increase the stock price

crash risk (Kim et al., 2011). Those situations are probably due to information

asymmetries when investors fear opportunistic behaviours by managers (Crabtree &

Kubick, 2014; Inger, Meckfessel, & Fan, 2018; Kim et al., 2011; Mcguire, Rane, &

Weaver, 2018; Wahab & Holland, 2012), and that can be mitigated by the increase of the

financial reporting quality. For Habib and Hasan (2016), one way is through the existence

of auditors who provide non-audit tax services (NATS).

On the contrary, there are cases where the tax avoidance positively weighted

value. Tax loss carry forwards, when shareholders value the company's ability to

generate, through tax planning, taxable income capable of absorbing them (Mcguire,

Olson, & Omer, 2016), stock option tax deductions, when investors value the low risk

and the transparent nature of the method for tax reduction (Inger, 2014), the financing of

the activity of lobbying fiscal matters (Hill, Kubick, Brandon Lockhart, & Wan, 2013),

and he source of income (operating or non-operating) and their origin (domestic vs

international). Operating domestic income and non-operating income, regardless the

origin, both increase the value of the company (Chang, Daniel, & Tsai, 2013) and in the

presence of proprietary costs (Robinson & Schmidt, 2013).

These cases are in accordance with Inger (2014), who argues that the positive

impact of tax avoidance on valuation depends on the activity that originated it, varying in

function of the tax risk, permanence of tax savings, tax planning cost, implicit taxes and

contrasts in disclosures of tax reduction in the financial statements.

Thus, the valuation of companies depends on the trade-off between risks (agency

253
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

costs) and benefits made by investors (Col, 2017; Kim et al., 2011; Wahab & Holland,

2012)

An example of this weighting was obtained by Col (2017), who, through an

international sample of mergers and acquisitions, concluded that the changing of the tax

residence to tax havens, following these operations, has implications that go beyond tax

issues, such as the loss of shareholder rights and the reduction of mandatory disclosures.

These implications are reflected in the company valuation, showing that shareholders

weighted tax avoidance beyond tax issues and feared its impacts at different levels,

namely the loss of company transparency (Kim et al., 2011).

Nevertheless, Brooks, Godfrey, Hillenbrand, & Money (2016) consider that

investors are not concerned with the impact of tax avoidance on valuation once it will be

reversed in the long term due to short memory concerning these activities. Reinforcing

this conclusion, Guenther et al. (2017), analysing the effective tax rates, concluded

companies with low tax rates (a proxy for tax avoidance) tend to maintain it over the time.

This suggests that corporate tax avoidance is usually accompanied by strategies that are

persistent and do not increase the risk of the company.

Khurana et al. (2018), reported that the increase in tax avoidance, associated with

high managerial ability and strong governance, lead to greater efficiency of the company's

investments, meaning that, in these situations, managers use tax avoidance to increase

shareholders value and not for the extraction of rents.

Besides shareholders, there are other stakeholders who also consider tax

avoidance activities and reflect them in their decisions concerning the company, such as

the state by creating new tax rules, debt holders influencing the interest rate and investors

influencing the cost of equity.

The use of debt to finance investments has advantages due to the deductibility of

interest, which translates into a decrease in the taxable income. Some multinational

254
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

companies, subject to different tax rates, use this advantage to do intercompany debts.

Taylor and Richardson (2013), concluded that thinly capitalized18 companies, in

Australia, use this practice and take advantage from the absence of debt limit rules.

Some states decided to restrict the use of tax shields limiting the interest tax deductions

or excessive debt, although the intended effect may be contrary to the desired by the

increase of tax avoidance activities (Lim, 2012).

For the debt holders, a high tax avoidance from the financed companies can result

in an increase or a decrease of the spreads (Hasan et al., 2014) depending on negative or

positive impact they anticipate (Lim, 2011). For companies, the negative impact of

increased spreads, may be overcame through new lending mechanisms and alternative

contractual designs, or through an improvement in corporate management and

transparency of the company (Beladi, Chur, & Hu, 2018; Isin, 2018).

Tax avoidance also impacts the cost of equity. Goh et al. (2016) concluded that due to

expectations of future cash flows generated by the tax avoidance, investors demand lower

rates of return which is more evident for companies that have a high external monitoring.

However, in extreme situations (high or low tax avoidance), there is always an increase

in the cost of equity (Cook, Moser, & Omer, 2017).

18
A thinly capitalized entity is one whose assets are funded by a high level of debt and relatively little
equity ([Link]

255
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4. Conclusions

This paper provides a systematic review on tax avoidance over the period between 2003

and 2018 and extends the extant literature in some dimensions.

Our bibliographic analysis shows that research on tax avoidance has significantly

increased in recent years, is mainly performed in USA context, which reflect the

importance that the topic assumes as a social and economic concern and is published in

high quality journals.

Our analysis allowed to identify the tax avoidance metrics, endogenous

determinants, exogenous determinants and consequences as the four main clusters of

research.

Concerning tax avoidance measurement, over the years several metrics have been

developed to quantify non-conforming tax avoidance, mainly ratios, and currently there

are global indicators and specific indicators, the two main groups that are used in different

research contexts. Finely, we found ETR, CashETR and BTD, all of them global

measures, as the more frequent metrics.

Internal factors that may explain tax avoidance are the most analysed probably

due to a real need to understand the more controllable variables influencing tax avoidance

and for the difficulty in gathering and analysing external factors data. We identified five

main groups, namely related to the drivers of profit shifting, the ownership structure and

corporate governance, the CSR, the auditor-provided tax services, the information quality

and the human resources.

The drivers of profit shifting used to explain tax avoidance include factors such as

the existence of a strong international activity by companies that allows them to take

advantage of several international mechanisms (e.g. prices of transfer rates and more

favourable tax regimes).

256
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Ownership structure allows the identification of two different approaches (the

traditional and the alternative) and the type of company affects the way shareholders face

and influence tax avoidance. The characteristics of corporate governance systems affect

the way managers influence tax avoidance and mechanisms adopted to mitigate agency

costs show mixed results.

CSR activities relation with tax avoidance show mixed results and is not clear due

to the existence of intermediate factors that influence it, such as the company's

profitability levels.

Auditors-provided tax services can contribute to the optimization of the

company's resources, although the most recent studies point to an active participation of

auditors-provided tax services in the development of tax avoidance strategies. The quality

of the information for internal use tends to be higher in order to facilitate the development

of tax avoidance strategies and for external use tends to be lower in order to conceal tax

avoidance information and avoid unwanted attentions (e.g. tax authority).

Individual characteristics of managers, the agents responsible for the conception

and approval of tax policies, that may influence tax avoidance are the professional

background, overconfidence, high narcissism, a tendency for managerial acquisitive and

gender.

External factors affecting tax avoidance may be tax system and tax enforcement

related and surrounding related. National tax enforcement influences the activities of tax

avoidance to the extent that companies act on the perception of the effectiveness of the

tax system, whereby, in countries where this perception is strong, there are fears that

aggressive practices will be detected. The surrounding environment characteristics, such

as culture of trust and religion, a high concentration of consumers and the power of

specific product in the market contributes to a greater tax avoidance. For financial

constraints (e.g. economic crisis) there are mixed results.

257
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Factors that may be explained by tax avoidance (consequences) include the

integrity of the tax system, through loss of tax revenue, the loss of reputation and

increased scrutiny at company level, and the increased level of risk and associated costs

(e.g.: increased financing and equity costs).

Future research on tax avoidance should address other settings besides the USA

companies once tax environments differ along the world and should address the most

controversial factors found in previous research, like managerial incentives, corporate

governance characteristics and CSR activities and the impact of tax avoidance on

company value. Due to the importance of new technologies and artificial intelligence

nowadays, future research should also address these topics, namely the use of

technologies by tax authorities and how Base Erosion and Profit Shifting (BEPS)

initiatives, such as Country by Country Reporting, can influence tax avoidance.

258
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References

Abernathy, J. L., Davenport, S. A., & Rapley, E. T. (2013). Schedule UTP: Stock Price
Reaction and Economic Consequences. Journal of the American Taxation
Association, 35(1), 25–48.
Alstadsæter, A., & Jacob, M. (2017). Who participates in tax avoidance ? Evidence from
Swedish microdata. Applied Economics, 49(28), 2779–2796.
Armstrong, C. S., Blouin, J. L., Jagolinzer, A. D., & Larcker, D. F. (2015). Corporate
governance, incentives, and tax avoidance. Journal of Accounting and Economics,
60(1), 1–17.
Atwood, T. J., Drake, M. S., Myers, J. N., & Myers, L. A. (2012). Home Country Tax
System Characteristics and Corporate Tax Avoidance: International Evidence.
Accounting Review, 87(6), 1831–1860.
Austin, C. R., & Wilson, R. J. (2017). An Examination of Reputational Costs and Tax
Avoidance: Evidence from Firms with Valuable Consumer Brands. Journal of the
American Taxation Association, 39(1), 67–93.
Badertscher, B. A., Katz, S. P., & Rego, S. O. (2013). The separation of ownership and
control and corporate tax avoidance. Journal of Accounting and Economics, 56(2–
3), 228–250.
Baumann, F., Buchwald, A., Friehe, T., & Hottenrott, H. (2017). Tax enforcement and
corporate profit shifting. Applied Economics Letters, 24(13), 902–905.
Beladi, H., Chur, C., & Hu, M. (2018). Does tax avoidance behavior affect bank loan
contracts for Chinese listed firms? International Review of Financial Analysis, 58,
104–116.
Bird, A., & Karolyi, S. A. (2017). Governance and Taxes: Evidence from Regression
Discontinuity. Accounting Review, 92(1), 29–50.
Blaylock, B., Shevlin, T., & Wilson, R. J. (2012). Tax avoidance, large positive temporary
book-tax differences, and earnings persistence. Accounting Review, 87(1), 91–120.
Boone, J. P., Khurana, I. K., & Raman, K. K. (2013). Religiosity and Tax Avoidance.
Journal of the American Taxation Association, 35(1), 53–84.
Bradshaw, M., Liao, G., & Shuai, M. (2018). Agency costs and tax planning when the
government is a major Shareholder. Journal of Accounting and Economics, 1–23.
Brooks, C., Godfrey, C., Hillenbrand, C., & Money, K. (2016). Do investors care about
corporate taxes? Journal of Corporate Finance, 38, 218–248.
Cen, L., Maydew, E. L., Zhang, L., & Zuo, L. (2017). Customer–supplier relationships
and corporate tax avoidance. Journal of Financial Economics, 123(2), 377–394.
Chang, L., Daniel, F., & Tsai, Y. (2013). Earnings, institutional investors, tax avoidance,
and firm value: Evidence from Taiwan. Journal of International Accounting,
Auditing and Taxation, 22(2), 98–108.
Chen, E., Gavious, I., & Yosef, R. (2013). The relationship between the management of
book income and taxable income under a moderate level of book-tax conformity.
Journal of Accounting, Auditing and Finance, 28, 323–347.
Chen, L. H., Gramlich, J., & Houser, K. (2017). The effects of board gender diversity on

259
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

a firm’s risk strategies. Accounting and Finance.

Cheng, C. S. A., Huang, H. H., Li, Y., & Stanfield, J. (2012). The Effect of Hedge Fund
Activism on Corporate Tax Avoidance. Accounting Review, 87(5), 1493–1526.
Chyz, J. A., & Gaertner, F. B. (2018). Can Paying “Too Much” or “Too Little” Tax
Contribute to Forced CEO Turnover? Accounting Review, 93(1), 103–130.
Clarkson, M. B. E. (1995). A stakeholder framework for analyzing and evaluating
corporate social performance. Academy of Management Review, 20(1), 92–117.
Clausing, K. A. (2016). The effect of profit shifting on the corporate tax base in the United
States and beyond. National Tax Journal, 69(4), 905–934.
Col, B. (2017). Agency Costs of Moving to Tax Havens: Evidence from Cross-border
Merger Premia. Corporate Governance: An International Review, 25(4), 271–288.
Col, B., & Patel, S. (2016). Going to Haven ? Corporate Social Responsibility and Tax
Avoidance. Journal of Business Ethics.
Cook, K. A., Moser, W. J., & Omer, T. C. (2017). Tax avoidance and ex ante cost of
capital. Journal of Business Finance and Accounting, 44, 1109–1136.
Crabtree, A. D., & Kubick, T. R. (2014). Corporate tax avoidance and the timeliness of
annual earnings announcements. Review of Quantitative Finance and Accounting,
51–67.
Davis, A. K., Guenther, D. A., Krull, L. K., & Williams, B. M. (2016). Do Socially
Responsible Firms Pay More Taxes? Accounting Review, 91(1), 47–68.
Desai, M. (2005). The Degradation of Reported Corporate Profits. Journal of Economic
Perspectives, 19(4), 171–192.
Desai, M. A., & Dharmapala, D. (2006). Corporate tax avoidance and high-power
incentives. Journal of Financial Economics, 79, 145–179.
Desai, M. A., Foley, C. F., & Hines, J. R. (2006). The demand for tax haven operations.
Journal of Public Economics, 90, 513–531.
DeZoort, F. T., Pollard, T. J., & Schnee, E. J. (2018). A Study of Perceived Ethicality of
Low Corporate Effective Tax Rates. Accounting Horizons, 32(1), 87–104.
Donohoe, M. P. (2015). The economic effects of financial derivatives on corporate tax
avoidance. Journal of Accounting and Economics, 59(1), 1–24.
Donohoe, M. P., & McGill, G. A. (2011). The Effects of Increased Book-Tax Difference
Tax Return Disclosures on Firm Valuation and Behavior. The Journal of the
American Taxation Association, 33(2), 35–65.
Duan, T., Ding, R., Hou, W., & Zhang, J. Z. (2018). The burden of attention: CEO
publicity and tax avoidance. Journal of Business Research, 87, 90–101.
Dyreng, S. D., Hanlon, M., & Maydew, E. L. (2008). Long-Run Corporate Tax
Avoidance. The Accounting Review, 83(1), 61–82.
Dyreng, S. D., Hanlon, M., & Maydew, E. L. (2010). The Effects of Executives on
Corporate Tax Avoidance. Accounting Review, 85(4), 1163–1189.
Dyreng, S. D., Hanlon, M., Maydew, E. L., & Thornock, J. R. (2017). Changes in
corporate effective tax rates over the past 25 years. Journal of Financial Economics,
124(3), 441–463.

260
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Dyreng, S. D., Lindsey, B. P., Markle, K. S., & Shackelford, D. A. (2015). The effect of
tax and nontax country characteristics on the global equity supply chains of U.S.
multinationals. Journal of Accounting and Economics, 59(2–3), 182–202.
Dyreng, S. D., Lindsey, B. P., & Thornock, J. R. (2013). Exploring the role Delaware
plays as a domestic tax haven. Journal of Financial Economics, 108(3), 751–772.
Evertsson, N. (2016). Is the top leadership of the organizations promoting tax avoidance.
Journal of Financial Crime.
Fama, E. F., & Jensen, M. C. (1983). Separation of Ownership and Control. Journal of
Law and Economics, 26(2), 301–325.
Fink, A. (2010). Conducting Research Literature Reviews: From the Internet to Paper
(3rd ed.). London: SAGE publications.
Finley, A. R., & Stekelberg, J. (2016). The Economic Consequences of Tax Service
Provider Sanctions: Evidence from KPMG’s Deferred Prosecution Agreement.
Journal of the American Taxation Association, 38(1), 57–78.
Francis, B. B., Hasan, I., Wu, Q., & Yan, M. (2014). Are Female CFOs Less Tax
Aggressive? Evidence from Tax Aggressiveness. Journal of the American Taxation
Association, 36(2), 171–202.
Francis, B. B., Ren, N., & Wu, Q. (2017). Banking deregulation and corporate tax
avoidance. China Journal of Accounting Research, 10(2), 87–104.
Frank, M. M., Lynch, L. J., & Rego, S. O. (2009). Tax Reporting Aggressiveness and Its
Relation to Aggressive Financial Reporting. Accounting Review, 84(2), 467–496.
Frank, M. M., Lynch, L. J., Rego, S. O., & Zhao, R. (2018). Are Corporate Risk-Taking
Practices Indicative of Aggressive Reporting Practices? Journal of the American
Taxation Association, 40(1), 31–55.
Gaertner, F. B., Laplante, S. K., & Daniel, P. (2016). Trends in the sources of permanent
and temporary book-tax differences during the schedule M-3 era. National Tax
Journal, 69, 785–808.
Gallemore, J., & Labro, E. (2015). The importance of the internal information
environment for tax avoidance. Journal of Accounting and Economics, 60(1), 149–
167.
Gao, L. (2016). Corporate patents, R&D success, and tax avoidance. Review of
Quantitative Finance and Accounting, 47, 1063–1096.
[Link]
Gao, Z., Yi, L., & Yangxin, L. (2017). Local Social Environment , Firm Tax Policy , and
Firm Characteristics. Journal of Business Ethics.
Godfrey, P. C. (2005). The Relationship between Corporate Philanthropy and
Shareholder Wealth: A Risk Management Perspective. Academy of Management
Review, 30(4), 777–798.
Goh, B. W., Lee, J., Lim, C. Y., & Shevlin, T. (2016). The Effect of Corporate Tax
Avoidance on the Cost of Equity. Accounting Review, 91(6), 1647–1670.
Graham, J. R., Hanlon, M., & Shevlin, T. (2014). Incentives for Tax Planning and
Avoidance: Evidence from the Field. Accounting Review, 89(3), 991–1023.
Green, D. H., & Plesko, G. A. (2016). The relation between book and taxable income

261
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

since the introduction of the schedule M-3. National Tax Journal, 69, 763–784.
Guenther, D. A., Matsunaga, S. R., & Williams, B. M. (2017). Is Tax Avoidance Related
to Firm Risk? Accounting Review, 92(1), 115–136.
Gul, F. A., Khedmati, M., & Shams, S. M. M. (2018). Managerial acquisitiveness and
corporate tax avoidance. Pacific-Basin Finance Journal, 1–27.
Gulzar, M. A., Cherian, J., Sial, M. S., Badulescu, A., Thu, P. A., Badulescu, D., &
Khuong, N. V. (2018). Does Corporate Social Responsibility Influence Corporate
Tax Avoidance of Chinese Listed Companies? Sustainability.
Gupta, S., Mills, L. F., & Towery, E. M. (2014). The Effect of Mandatory Financial
Statement Disclosures of Tax Uncertainty on Tax Reporting and Collections: The
Case of FIN 48 and Multistate Tax Avoidance. Journal of the American Taxation
Association, 36(2), 203–229.
Habib, A., & Hasan, M. M. (2016). Auditor-provided tax services and stock price crash
risk Auditor-provided tax services and stock price crash risk. Accounting and
Business Research, 4788, 51–82.
Hanlon, M., & Heitzman, S. (2010). A review of tax research. Journal of Accounting and
Economics, 50(2–3), 127–178.
Hanlon, M., Maydew, E. L., & Saavedra, D. (2017). The taxman cometh: Does tax
uncertainty affect corporate cash holdings? Review of Accounting Studies, 22, 1198–
1228.
Hardeck, I., & Wittenstein, P. U. (2018). Assessing the tax benefits of Hybrid
Arrangements — Evidence from the Luxembourg Leaks. National Tax Journal, 71,
295–334.
Hasan, I., Hoi, C. K. (Stan), Wu, Q., & Zhang, H. (2014). Beauty is in the eye of the
beholder: The effect of corporate tax avoidance on the cost of bank loans. Journal
of Financial Economics, 113(1), 109–130.
Hasegawa, M., Hoopes, J. L., Ishida, R., & Slemrod, J. (2013). The effect of public
disclosures on reported taxable income: Evidence from individuals and corporations
in Japan. National Tax Journal, 66, 571–608.
Henry, E., Massel, N., & Towery, E. (2016). Increased tax disclosures and corporate tax
avoidance. National Tax Journal, 69(4), 809–830.
Henry, E., & Sansing, R. (2018). Corporate tax avoidance: data truncation and loss firms.
Review of Accounting Studies, 23, 1042–1070.
Hill, M. D., Kubick, T. R., Brandon Lockhart, G., & Wan, H. (2013). The effectiveness
and valuation of political tax minimization. Journal of Banking and Finance, 37(8),
2836–2849.
Hoi, C. K., Wu, Q., & Zhang, H. (2013). Is Corporate Social Responsibility (CSR)
associated with Tax Avoidance? Evidence from irresponsible CSR Activities.
Accounting Review, 88(6), 2025–2059.
Hope, O. K., Ma, M. S., & Thomas, W. B. (2013). Tax avoidance and geographic earnings
disclosure. Journal of Accounting and Economics, 56(2–3), 170–189.
Hsieh, T., Wang, Z., & Demirkan, S. (2018). Overconfidence and tax avoidance: The role
of CEO and CFO interaction. Journal of Accounting and Public Policy, 37, 241–
253.

262
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Hsu, P., Moore, J. A., & Neubaum, D. O. (2018). Tax avoidance, financial experts on the
audit committee, and business strategy. Journal of Business Finance and
Accounting, 2, 1293–1321.
Huang, H. H., Lobo, G. J., Wang, C., & Xie, H. (2016). Customer concentration and
corporate tax avoidance. Journal of Banking and Finance, 72, 184–200.
Huang, H. H., Sun, L., & Yu, T. R. (2017). Are Socially Responsible Firms Less Likely
to Expatriate? An Examination of Corporate Inversions. Journal of the American
Taxation Association, 39(2), 43–62.
Huseynov, F., Sardarli, S., & Zhang, W. (2017). Does index addition affect corporate tax
avoidance? Journal of Corporate Finance, 43, 241–259.
Inger, K. K. (2014). Relative Valuation of Alternative Methods of Tax Avoidance.
Journal of the American Taxation Association, 36(1), 27–55.
Inger, K. K., Meckfessel, M. D., & Fan, W. P. (2018). An Examination of the Impact of
Tax Avoidance on the Readability of Tax Footnotes. Journal of the American
Taxation Association, 40(1), 1–29.
Isin, A. A. (2018). Tax avoidance and cost of debt : The case for loan-specific risk
mitigation and public debt financing. Journal of Corporate Finance, 49, 344–378.
Jackson, M. (2015). Book-Tax Differences and Future Earnings Changes. Journal of the
American Taxation Association, 37(2), 49–73.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the Firm: Managerial Behavior ,
Agency Costs and Ownership Structure. Journal of Financial Economics, 3(4), 305–
360.
Jiang, C., Kubick, T. R., Miletkov, M., & Wintoki, M. B. (2018). Offshore expertise for
onshore companies: Director connections to island tax havens and corporate tax
policy. Management Science, 64(7), 2973–3468.
Jiménez-Angueira, C. E. (2018). The effect of the interplay between corporate
governance and external monitoring regimes on firms’ tax avoidance. Advances in
Accounting, 41, 7–24.
Kanagaretnam, K., Lee, J., Lim, C. Y., & Lobo, G. (2018). Societal trust and corporate
tax avoidance. Review of Accounting Studies, 23(4), 1588–1628.
Karamshahi, B., Azami, Z., & Salehi, T. (2018). The association between competition
power in markets and tax avoidance: evidence from Tehran stock exchange. Eurasia
Business Review, 8(3), 323–339.
Khan, M., Srinivasan, S., & Tan, L. (2017). Institutional Ownership and Corporate Tax
Avoidance: New Evidence. Accounting Review, 92(2), 101–122.
Khurana, I. K., & Moser, W. J. (2013). Institutional Shareholders’ Investment Horizons
and Tax Avoidance. Journal of the American Taxation Association, 35(1), 111–134.
Khurana, I. K., Moser, W. J., & Raman, K. K. (2018). Tax Avoidance, Managerial
Ability, and Investment Efficiency. ABACUS, 54(4), 547–575.
Kim, J. B., Li, Y., & Zhang, L. (2011). Corporate tax avoidance and stock price crash
risk: Firm-level analysis. Journal of Financial Economics, 100(3), 639–662.
Krishnan, G. V, & Visvanathan, G. (2011). Is There an Association between Earnings
Management and Auditor-Provided Tax Services? Journal of the American Taxation

263
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Association, 33(2), 111–135.

Kubick, T. R., & Lockhart, G. B. (2017). Overconfidence, CEO Awards, and Corporate
Tax Aggressiveness. Journal of Business Finance and Accounting, 44(5–6), 728–
754.
Kubick, T. R., Lockhart, G. B., Mills, L. F., & Robinson, J. R. (2017). IRS and corporate
taxpayer effects of geographic proximity. Journal of Accounting and Economics,
63(2–3), 428–453.
Kubick, T. R., Lynch, D. P., Mayberry, M. A., & Omer, T. C. (2015). Product Market
Power and Tax Avoidance: Market Leaders, Mimicking Strategies, and Stock
Returns. Accounting Review, 90(2), 675–702.
Kubick, T. R., Lynch, D. P., Mayberry, M. A., & Omer, T. C. (2016). The Effects of
Regulatory Scrutiny on Tax Avoidance: An Examination of SEC Comment Letters.
Accounting Review, 91(6), 1751–1780.
Lanis, R., & Richardson, G. (2015). Is Corporate Social Responsibility Performance
Associated with Tax Avoidance ? Journal of Business Ethics, 127, 439–457.
Lanis, R., Richardson, G., Liu, C., & Mcclure, R. (2018). The Impact of Corporate Tax
Avoidance on Board of Directors and CEO Reputation. Journal of Business Ethics.
Law, K. K. F., & Mills, L. F. (2017). Military experience and corporate tax avoidance.
Review of Accounting Studies, 22(November 2016), 141–184.
Lee, N. (2018). R&D Accounting Treatment, R&D State and Tax Avoidance: With a
Focus on Biotech Firms. Sustainability.
Li, O. Z., Liu, H., & Ni, C. (2017). Controlling Shareholders ’ Incentive and Corporate
Tax Avoidance : A Natural Experiment in China. Journal of Business Finance and
Accounting, 44(June), 697–727.
Lim, Y. (2011). Tax avoidance, cost of debt and shareholder activism: Evidence from
Korea. Journal of Banking and Finance, 35(2), 456–470.
Lim, Y. (2012). Tax avoidance and underleverage puzzle : Korean evidence. Review of
Quantitative Finance and Accounting, 39, 333–360.
Lismont, J., Cardinaels, E., Bruynseels, L., & Groote, S. De. (2018). Predicting tax
avoidance by means of social network analytics. Decision Support Systems, 108, 13–
24.
Lisowsky, P. (2010). Seeking Shelter : Empirically Modeling Tax. Accounting Review,
85(5), 1693–1720.
Malik, S., Mihm, B., & Timme, F. (2018). An experimental analysis of tax avoidance
policies. International Tax and Public Finance, 25(1), 200–239.
Mayberry, M. A., Mcguire, S. T., & Omer, T. C. (2015). Smoothness and the Value
Relevance of Taxable Income. Journal of the American Taxation Association, 37(2),
141–167.
McClure, R., Lanis, R., Wells, P., & Govendir, B. (2018). The impact of dividend
imputation on corporate tax avoidance: The case of shareholder value. Journal of
Corporate Finance, 48, 492–514.
Mcguire, S. T., Olson, A. J., & Omer, T. C. (2016). Do Investors Use Prior Tax Avoidance
when Pricing Tax Loss Carryforwards? Journal of the American Taxation

264
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Association, 38(2), 27–49.

Mcguire, S. T., Omer, T. C., & Wang, D. (2012). Tax Avoidance: Does Tax-Specific
Industry Expertise Make a Difference? Accounting Review, 87(3), 975–1003.
Mcguire, S. T., Rane, S. G., & Weaver, C. D. (2018). Internal Information Quality and
Tax-Motivated Income Shifting. Journal of the American Taxation Association,
40(2), 25–44.
Mcguire, S. T., Wang, D., & Wilson, R. J. (2014). Dual Class Ownership and Tax
Avoidance. Accounting Review, 89(4), 1487–1516.
Minor, D., & Morgan, J. (2011). CSR as Reputation Insurance: Primum Non Nocere.
California Management Review, 53(3), 40–60.
Moore, J. A., Suh, S., & Werner, E. M. (2017). Dual entrenchment and tax management:
Classified boards and family firms. Journal of Business Research, 79, 161–172.
Olsen, K. J., & Stekelberg, J. (2016). CEO Narcissism and Corporate Tax Sheltering.
Journal of the American Taxation Association, 38(1), 1–22.
Park, S. (2018). Related Party Transactions and Tax Avoidance of Business Groups.
Sustainability, (10), 1–14.
Penrose, E. (1959). The Theory of the Growth of the Firm. Oxford, UK: Blackwell.
Petticrew, M., & Roberts, H. (2006). Systematic Reviews in the Social Sciences. Malden,
MA: Blackwell Publishing.
Ravenda, D., Argilés-Bosch, J., & Valencia-Silva, M. M. (2015). Labor Tax Avoidance
and Its Determinants: The Case of Mafia Firms in Italy. Journal of Business Ethics,
132, 41–62.
Rego, S. O. (2003). Tax-Avoidance Activities of U.S. Multinational Corporations.
Contemporary Accounting Research, 20(4), 805–833.
Richardson, G., Taylor, G., & Lanis, R. (2013). Determinants of transfer pricing
aggressiveness : Empirical evidence from Australian firms q. Journal of
Contemporary Accounting & Economics, 9(2), 136–150.
Richardson, G., Taylor, G., & Lanis, R. (2015). The impact of financial distress on
corporate tax avoidance spanning the global financial crisis: Evidence from
Australia. Economic Modelling, 44, 44–53.
Richardson, G., Wang, B., & Zhang, X. (2016). Ownership structure and corporate tax
avoidance: Evidence from publicly listed private firms in China. Journal of
Contemporary Accounting and Economics, 12(2), 141–158.
Robinson, L. A., & Schmidt, A. P. (2013). Firm and Investor Responses to Uncertain Tax
Benefit Disclosure Requirements. Journal of the American Taxation Association,
35(2), 85–120.
Salihu, I. A., Annuar, H. A., Normala, S., & Obid, S. (2015). Foreign investors ’ interests
and corporate tax avoidance: Evidence from an emerging economy. Journal of
Contemporary Accounting & Economics, 11(2), 138–147.
Schimanski, C. (2017). Earnings shocks and tax-motivated income- shifting: evidence
from European multinationals – revisited. Applied Economics Letters, 24(21), 1558–
1566.
Seidman, J. K., & Stomberg, B. (2017). Equity compensation and tax avoidance:

265
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Disentangling managerial incentives from tax benefits and reexamining the effect of
shareholder rights. Journal of the American Taxation Association, 39(2), 21–41.
Shackelford, D. A., & Shevlin, T. (2001). Empirical tax research in accounting. Journal
of Accounting and Economics, 31, 321–387.
Shevlin, T., Thornock, J., & Williams, B. (2017). An examination of firms ’ responses to
tax forgiveness. Review of Accounting Studies, 577–607.
Simone, L. De, Robinson, J. R., & Stomberg, B. (2014). Distilling the reserve for
uncertain tax positions : the revealing case of black liquor. Review of Accounting
Studies, 19, 456–472.
Stewart, J. (2018). MNE tax strategies and Ireland. Critical Perspectives on International
Business, 14(4), 338–361.
Tang, T., Mo, P. L. L., & Chan, K. H. (2017). Tax Collector or Tax Avoider? An
Investigation of Intergovernmental Agency Conflicts. Accounting Review, 92(2),
247–270.
Tang, T. Y. H. (2015). Does Book-Tax Conformity Deter Opportunistic Book and Tax
Reporting? An International Analysis. European Accounting Review, 24(3), 441–
469.
Taylor, G., & Richardson, G. (2013). The determinants of thinly capitalized tax avoidance
structures: Evidence from Australian firms. Journal of International Accounting,
Auditing and Taxation, 22(1), 12–25.
Taylor, G., & Richardson, G. (2014). Incentives for corporate tax planning and reporting:
Empirical evidence from Australia. Journal of Contemporary Accounting and
Economics, 10(1), 1–15.
Taylor, G., Richardson, G., & Taplin, R. (2015). Determinants of tax haven utilization:
evidence from Australian firms. Accounting and Finance, 55, 545–574.
Taylor, G., Tower, G., & Van Der Zahn, M. (2011). The influence of international
taxation structures on corporate financial disclosure patterns. Accounting Forum,
35(1), 32–46.
Thomsen, M., & Watrin, C. (2018). Tax avoidance over time: A comparison of European
and U.S. firms. Journal of International Accounting, Auditing and Taxation, 33, 40–
63.
Tranfield, D., Denyer, D., & Smart, P. (2003). Towards a Methodology for Developing
Evidence-Informed Management Knowledge by Means of Systematic Review.
British Journal of Management, 14, 207–222.
Wahab, N. S. A., & Holland, K. (2012). Tax planning , corporate governance and equity
value. The British Accounting Review, 44(2), 111–124.
Watson, L. (2015). Corporate Social Responsibility, Tax Avoidance, and Earnings
Performance. Journal of the American Taxation Association, 37(2), 1–21.
Wilde, J. H., & Wilson, R. J. (2018). Perspectives on corporate tax planning: Observations
from the past decade. Journal of the American Taxation Association, 40(2), 63–81.
Wilson, R. J. (2009). An Examination of Corporate Tax Shelter Participants The
University of Iowa, 84(3), 969–999.
Young, A. (2017). How does governance affect tax avoidance ? Evidence from

266
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

shareholder proposals proposals. Applied Economics Letters, 24(17), 1208–1213.

267
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Turkey Direct Investments in Latvian Business: Agenda for Research

Andrejs Limanskis, PhD (Econ), RISEBA
Abstract
Topicality of the research is determined by increasing uncertainty and structural change in
foreign direct investments (FDI) as well as by potential role of investments from Turkey in
Latvian development. Latvia is a small EU member state. Turkey is the 13th global giant by
GDP.
The aim of this research is to examine basic aspects of the direct investments (DI) from Turkey in
Latvia as sustainability driver for fine tuning of FDI policy.
The tasks fulfilled are: literature review of FDI sustainability in Latvia, research design,
identification of the state of Turkey DI volume and structure in Latvia, calculation of Sigma for
Turkey DI flows in Latvia before and after joining the EU, discussion of possible future
feasibility of Turkey DI in Latvia, formulation of conclusions and recommendations.
The paper consists of an introduction, the theoretical and the empirical parts, conclusions as
well as references. The theoretical part is a literature review on sustainable FDI and evaluation
of relevant methods of research. Empirical part presents the results of investigation of the
research question.
The scientific novelty consists of proposing for scientific circulation of Sigma for Turkey DI in
Latvia based on analysis of registration data on FDI deals for all the years after restoration of
Latvian state independency and data gained as result of interviewing of experts.
Hypothesis: the need for investments in sustainable growth in Latvia is partially satisfied by
Turkey DI. The hypothesis is proven based on quantitative and qualitative methods of analysis
applied by the author.
Keywords: Foreign Direct Investment (FDI), Turkey, Latvia, Sustainability, International
Business
INTRODUCTION
Foreign direct Investment (FDI) belongs to topical research themes. Unfortunately, recent years
witnessed dramatic fall in global FDI, even before corona virus crisis. By regions the process is
not so univocal. The actual flows of FDI are different in countries ruled by specific factors.
1

A. Limanskis
Faculty of Business, RISEBA University of Business, Arts and Technology, Riga, Latvia
e-mail: [Link]@[Link]

268
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Latvia is an Eurozone country with typically much bigger inflow of FDI as its outflow. Turkey
proved an influential donor. First Turkey DI in Latvia was registered in 1991. Now, Turkey
ranks 34th among investors in Latvia, next after Uzbekistan. There are processes in the Turkey DI
in Latvia that demand investigation by many considerations.
Sustainable growth proves to be new phenomena with great potential. Investment responding to
the sustainability challenges constitutes an ever-growing trend both on national DI level and
internationally. Accenture (2018) and other think tanks like Boston Consulting Group,
McKinsey, etc. speak about digital disruption, new revolution and intelligent enterprise changing
economy and whole life. Alarmists prioritize the world is to be preserved for future generations
through environment and social awareness in investments.
As soon as cross-border investments come, compete and interact with local ones, the future
depends on their character. Research and education play growing role in securing that
sustainability go hand in hand with them in the investment process.
The aim of this research is to examine basic aspects of the direct investments (DI) from Turkey
in Latvia as sustainability driver for fine tuning of FDI policy.
The tasks to be fulfilled are: literature review of FDI sustainability in Latvia, research design,
identification of the state of Turkey DI volume and structure in Latvia, calculation of Sigma for
Turkey DI flows in Latvia before and after joining the EU, discussion of possible future
feasibility of Turkey DI in Latvia, formulation of conclusions and recommendations.
The scientific novelty consists of introduction in the scientific circulation of complex data,
including Sigma, for Turkey DI in Latvia based on collection and analysis of registration data on
DI deals for all the years after restoration of Latvian state independency, periodization of the DI
process and disclosure of structure by recipients. Primary data stems from interviews and
personal observation by author.
Hypothesis: the need for investments in sustainable business in Latvia is partially satisfied by
Turkey direct investments. The hypothesis is proven based on quantitative and qualitative
methods of analysis done by the authors.
The paper consists of an introduction, the theoretical and the empirical parts, conclusions and
recommendations as well as references. The theoretical part is a literature review on FDI and
methods of research. Empirical part presents the results of investigation of two research
questions.
LITERATURE REVIEW OF ESSENCE AND NEW TRENDS IN FDI
Foreign Direct Investment, further FDI, is one of the key forms of international business activity.
Wild J.J. et al (2000) indicate that foreign direct investment is “the purchase of physical assets or
a significant amount of the ownership of a company in another country to gain measure of

269
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

management control”. Similar point of view by Griffin et al (1996) is that FDI are investments
made for the purpose of actively controlling property, assets, or companies located in host
countries. To make traditional definition sharp, FDI is acquisition of foreign assets for getting
profit. A number of sources investigate factors of greater profit potential abroad against limited
opportunities for getting profit at home.
However, there is a change in approach. Sustainable investing according to World Resources
Institute (2018) “is the new black: essential, ubiquitous and a subject of forward-looking
discussion. But this is no passing fashion. What was once a niche investment approach is
becoming mainstream... Increasingly, investors acknowledged that the economy is closely
connected to social and environmental realities. They realized that ignoring this link would be a
liability”.
Author supports this notion and sticks to the sustainability triangle: profitable, ecological and
socially responsible investment.
Recent research in FDI draws not only sustainability in the mainstream. Digital investments are
intensively researched by FDI intelligence (2018) since “they create Digital Economies of the
Future. Attracting major technology companies from abroad and digital jobs creation are
regarded as new FDI flows to be traced and area of competition among recipient countries”.
Covid crisis made distant work a reality and gave a push to sustainability vector.
One more point highlighted by Covid crisis is sectorial risks. Civil aviation fell one of the first
victims to closure under emergency. Investments from abroad in this and other industries
sensible to freedom of movement of people ran in the red.
Latvia, having regained state independency after collapse of the Soviet Union, started change to
market model of economy, and immediately, in 1991, first foreign investors arrived. In line with
the definitions above, investment was buying assets for getting profit under own control. First
publication in Latvian language by Praude (2010) generalized FDI experience within traditional
investment concept. Sustainability and anti-crisis vector was raised later by Sauka (2020).
Nowadays, business gets internationally mobile, and not only immense global companies can
engage in FDI, but also small entrepreneurs. One will be able to see internationalization
tendency in this paper through Latvian enterprises with Turkey capital which are not big as a
rule.
Economic reforms in Latvia made many industries attractive for FDI due to low wage level,
reduction of trade barriers and growing efficiency of export of goods and services. A lot of
international companies opened their subsidiaries in Latvia.
Nowadays Turkey is the world's thirteenth largest economy by GDP (purchasing power parity)
according to the holder of economic indicators for over 200 countries (GlobalEconomy 2018).

270
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Economic growth may overcome international benchmark. Turkey companies continue overseas
investments, as well as participate in mergers and acquisitions of foreign companies. Turkey
develops innovations-driven economy with high value added industries. It focuses on
infrastructure investment, education, and science besides HORECA much spoken about. A part
of efforts promotes a sustainability model.
According to Czinkota (1996) major determinants of FDI are marketing factors, barriers to trade,
cost factors, investment climate and other vectors. All these influenced FDI inflows and outflows
which is noticeable in Turkey DI in Latvia. According to the world investment report 2017, FDI
are boosted by ongoing technological change and the digital economy; global urbanization; and
offshoring, as well as longer-term trends in prospective mega groups. Geopolitical uncertainty,
debt concerns, terrorism and cyber threats are almost universally considered in a negative light
and as likely to dampen FDI activity (UNCTAD, 2017). Covid introduced additional risks.
Investors from Turkey strive to exploit their competitive advantages by accessing Latvia. For
recipient enterprises, the critical role of foreign direct investment includes jobs, taxes and
technology transfer, which means the transfer of a combination of hardware, software, and skills
for production processes.
In Latvia, traditional theoretical approaches to FDI can be traced. There exist partisans of
different international investment theories on why FDI occur. According to Griffin and Pustay
(1996) there are: the ownership advantage theory, internalization theory, eclectic theory.
However, as per Wild (2000), there are four main theories that attempt to explain why companies
engage in foreign direct investment: international product life cycle, market imperfections
(internalization), eclectic theory, and market power. Sustainability investment overlaps all
previous theories. However, published attempts to investigate FDI in Latvia are few. It still is the
norm to overlook the relationship between investments and sustainability issues. It was only in
the last decade or so that the idea of sustainability started to penetrate this sphere. Limanskis
(2020) elaborated methology of research and applied it to the DI from the USA in Latvia.
Unfortunately, role of Statistical classification of Economic Activities in the European
community known as NACE codes in sustainability research is not commonly recognized yet.
No enterprises in Latvia as recipients of DI from majority of states of origin are ever analyzed in
the triangle of profitability, ecology and social responsibility.
Turkey is regarded as DI donor for Latvia in the World Investment Report Investment and New
Industrial Policies (2018). Some investors should care for accusations of alleged failure to
comply with anti-money laundering and terrorism financing regulations. Covid-19 claims are
arising out of the Emergency imposed by the government in Latvia in March 2020.

271
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Turkey DI in Latvia up to now did not constitute subject of published research. It is known only,
according to Ministry of Foreign Affairs (2020) that DI flows from Turkey are welcome. No
deeper information has ever been published.
Update of FDI policy lags behind the needs. The appeals to step up the FDI policy are emerging.
Ministry of Economy of Latvian Republic (2019) set an aim to elaborate a profile of recipient
enterprise. However, the appeals are too common, like raise the volume, perfect the structure,
increase the mutual benefits, etc. It is under the sustainability of DI concept that the process can
be investigated in detail and instrumental changes in policy proposed.
DATA AND METODOLOGY OF RESEARCH
Both primary and secondary data constitute the basis for the research. Following methods for
gaining primary data were used: interviews disclosing opinions of investors from Turkey in
Latvia and local experts, and personal observation through research and participation in
conferences and business panel discussions.
Secondary data was gained from the electronic statistical database of the Enterprise Register of
Latvian Republic (Lursoft), Central Statistical Bureau, the Bank of Latvia, Investment and
Development Agency of Latvia (LIAA), Office of Citizenship and Migration Affairs at the
Ministry of Interior of Republic of Latvia, academic databases Emerald, EBSCO, Crediweb
database, books and articles.
Author used qualitative, quantitative and mixed methods of research. The qualitative research
method is adequate for understanding many aspects of Turkey DI in Latvia. Turkey investors’
opinions are disclosed via in-depth interviews. Questionnaire for a sample of 139 Turkey
investors in 68 industries proved impossible due to several reasons and author focused on
interviews with the 10 Turkey investors in Latvia who invested not less than 100 000 EUR each
as well as three Latvian practitioners dealing with FDI.
The following quantitative data analysis processes are used: time series statistical data analysis,
calculation and interpretation, plotting of linear and pie-type charts, trend analysis and
calculation of Sigma. The charts on Turkey DI in Latvia are plotted by author on the basis of
Register of enterprises of Latvian Republic big data with Lursoft permission. The
abovementioned data complex of DI from Turkey has never been singled out and published
before. Structure of DI from Turkey by NACE codes has not been researched in Latvia as well.
Time span is covering all the years after regained statehood of Latvia up to 01.06.2020.
COVID19 emergency in Latvia 13.03-09.06.2020 is partially covered although deserves
additional research.
RESULTS: STATE OF TURKEY DI IN LATVIA

272
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

FDI for Latvia is a new path of economic behavior. Before regaining political independency in
the collapse of the USSR, Latvia was a part of the centrally planned economy. In the system,
foreign investments were strictly canalized by government. After Latvian Republic restored its
independency, in 1991 the first investors came and the deal with Turkey first investor was
registered.
Since then many investment projects proved success, a part liquidated, but the process gained
momentum. In 2020, Turkey direct investment in Latvia ranks 34th of all investing countries. It
has exceeded 8 million EUR, particularly 8332368 EUR. Turkey FDI does play significant role
in Latvian economy, despite exaggeration of risks by some politicians and mass media.
Dynamics of the flow is controversial. The general trajectory in non-monetary terms is
implemented in the following chart (see fig. 1)
200

150 Investors from

Turkey
100 Recipient
Enterprises
50
Deals in Latvia

0
1991
1994
1997
2000
2003
2006
2009
2012
2015
2018

Figure 1. FDI from Turkey to Latvia in absolute terms, 1991- 2019 (author, based on Lursoft
data, 2020)
Three important indicators in the same axes are represented: number of Turkey direct investors,
of Latvian recipient companies with Turkey capital and of projects (deals registered). One can
find out that from 1991 till 2017 there was exponential growth in numbers of investors and
recipient enterprises. By opinion of the author, shared by majority of interviewees, one can
single out two periods in the Turkey DI in Latvia, namely the first 14 years 1991-2004 of
“reconnaissance” (”try and error”) and then acceleration since 2005 connected to EU entry by
Latvia 1.05.2004. The jump, unfortunately, is not without reverse change in the recent years. The
tendency to grow is breaking through in spite of discrepancy in numbers of projects.
After record level of 2019, the number of projects declined dramatically. One needs to research
this situation deeper, analyze it in monetary terms (see Fig. 2)

273
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

10000000.00

8000000.00

6000000.00 FDI from Turkey

Cumulative, EUR
4000000.00
FDI Flow, EUR
2000000.00

0.00

-2000000.00

-4000000.00

-6000000.00

Figure 2. DI from Turkey to Latvia yearly flow and cumulative1991-2019, EUR (author, based
on Lursoft data, 2020)
In contrast to non-financial indicators, the EU entry of Latvia happened rather to stop dramatic
withdrawal of capital than to accelerate the process. Since 2002, outflow occurred (minus 4.3
mln euro in 2004) after record 4 mln euro inflow. It was remedied slightly in 2005 and 2006 just
to run again in the negative area in 2007 and especially in 2009.
In 2011, cumulative 5.5 million euro level was approached for the second time. Unfortunately, in
2012 the second deep decline occurred. And increase of the sum resumed in 2013-2018. In 2018,
a peak in the cumulative indicator was recorded, namely above eight million euro. After this,
unfortunately, a decline can be seen.
The recent decline in Turkey FDI in Latvia was not so dramatic as could be expected against the
background of the 2019 US sanctions against Turkey and political maneuvers by EU and Latvia.
The interviewees underline long-term intentions in DI from Turkey in Latvia. Nevertheless a part
of projects fall victims, and flows of investment turned negative in 2018. The Covid crisis effect
in the DI in Latvia demands special investigation in future.
The scale of monthly fluctuations of Turkey FDI in Latvia can be seen on Figure 3

274
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1200000.00

1000000.00

800000.00

600000.00

400000.00 Monthly Flow, EUR

200000.00

0.00
2018 1

2019 1

2020 1
3
5
7
9

3
5
7
9

3
11

11
-200000.00

-400000.00

Figure 3. Turkey monthly FDI flows to Latvia 2018-2020*, EUR (author, based on Lursoft data,
2020)
*stand 1/04/2020
Starting from March 2018 monthly flow of direct investments from Turkey became negative
with rare exclusions. Under COVID19 emergency no deal was registered in April 2020.
To generalise, six years of 30 registered negative DI flows, namely 2001, 2004, 2007, 2009,
2012 and 2019. In 2001 the fall was not dramatic. But in 2004 the fall -4361272 EUR was the
biggest in history of Turkey investments in Latvia. It was unexpected since Latvia joined the EU
on May 1, 2004 and the event could be an immediate driver for additional investments from
Turkey. Notably, the crisis year 2009 resulted in visible fall, but nevertheless the inflow of
capital recovered next year. 2012 fall was second record after 2004.
To through light on uncertainty of DI flows from Turkey in Latvia, author calculated Standard
Deviation (Sigma) for Turkey yearly DI flows to Latvia 1991-2019. First, an average (mean)
yearly flow was calculated for 29 years. It is 287 323 EUR. Then the differences were calculated,
summed up, and a square root drawn. The Sigma proved 1501296 EUR. The 3 Sigma rule reads
that the next year DI flow from Turkey in Latvia would be within 3 Sigma with probability
99.73%. Sigma is 5,23 times bigger than the mean, which is too big.
Second, author deepened the uncertainty research for DI from Turkey in Latvia in financial and
non-financial indicators. The period was subdivided in two periods by criteria of Latvia joining
the EU, namely 1991-2004 and 2005-2019. The Sigmas can be observed in the following table
(see table 1).

275
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1
Sigmas for financial and non-financial indicators of DI from Turkey flow in Latvia
Sigma Flow, EUR Deals Investors Recipients
1991-2004 1613812,21 11,13 12,28 9,51
2005-2019 1356952,44 17,98 46,21 42,19
Times 0,84 1,62 3,76 4,44
1991-2019 1501296,94 22,60 56,30 51,97
Average 287 323,05 30,48 70,31 60,00
Sigma/Average 5,23 0,74 0,80 0,87

The financial indicator demonstrated comparable Sigmas for two periods with no visible
financial DI from Turkey effect of Latvia joining the EU. However the non-financial indicators
demonstrate growing uncertainty after Latvia joined the EU.
For the calculation, 100% of time series data covering whole years of the period of observation
was taken. The conclusion is that the uncertainty of the DI flow from Turkey to Latvia is
immense. Sigma for DI flows for the whole history is dramatically bigger than the mean yearly
flow. In future research, this Sigma can be compared to the Sigmas of DI flows to Latvia from
other countries. Investors from Turkey interviewed mentioned geopolitical risks as major reason
for uncertainty of DI flows in Latvia.
Important peak of 5491999 euro in 2011 was achieved due to record 4 416 473 euro of flow.
Since 1999, all times low of 516246 euro of accumulated investment was registered in 2009.
Recent cumulative level of above 8.3 million euro is under threat. By opinions of interviewees
shared by author, the level of DI from Turkey in Latvia is disproportionally small against the
economic might of Turkey. Its prospects are to great extent connected to the structure.
Structural analysis of DI from Turkey in Latvia is based on NACE codes. Following is the
selection of top 10 investments by industries (complete table of 68 industries for research can be
requested from author).

276
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2
Top 10 investments from Turkey in Latvia by industries by NACE Codes
Recipient
NACE Direct Turkey
Economic Activity (by NACE) Enterprise Investors
Code Investments, EUR
s

52.23 Support activities for air transport 3 354 152,00 2 2

68.20 Renting and operating of own or leased real estate 1 405 045,00 4 8
01.42 Raising of other cattle and buffaloes 1 008 491,49 5 8
56.10 Restaurant and mobile food service activities 635 359,46 20 18
46.39 Non - specialized wholesale of food, beverages and tobacco 500 000,00 1 1
47.11 Retail sale in non - specialized stores with food, beverages or tobacco 426 861,40 1 1
16.10 Sawing, planing and impregnation 369 626,00 1 1
46.16 Agents involved in the sale of textiles, clothing, footwear and leather go 146 546,87 2 2
46.73 Wholesale of wood, construction materials and sanitary equipment 85 372,31 1 1
73.11 Activities of advertising agencies 50 100,00 2 2

The top 10 industries represent 96,25% of total direct investments from Turkey in Latvia.
Support activities for air transport is the leading line followed by renting and operating real
estate, cattle raising, restaurants, whole- and retail sale, sawing, planing and impregnation of
wood, etc.
Top five investments are presented in the pie-type chart below (see Figure 4).

17%
52.23

40% 68.20
6%
01.42
8% 56.10
46.39
12% Other

17%

Figure 4. The structure of FDI from Turkey in Latvia by NACE codes

The already mentioned Support activities for air transport (NACE 52.23) constitute 40% of total.
Interviewing investors disclosed that investments are supporting Turkish Airlines on the EU
market of passenger and cargo transportation. Interviewees sit in Marupe administrative unit
known for Riga airport. Author summed up data on all enterprises with Turkey roots in this
address and the capital reached 5 183 858 euro, i.e. 62,51% of total.
Covid crisis in 2020 introduced stop of passenger flights. Vacations in Turkey were sacrificed.
Rethinking the sustainability aspect of related investments seems to result in excluding flights

277
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

from the list of counter-sustainability industries. People in Latvia need vacations in Turkey for
being physically and mentally fit.
Certain part of Turkey investment growth in 2011-2014 was attributed by researcher (Limanskis
2014) to the Latvian governmental program of foreign investment attraction through amending
Latvian Immigration Law in 2010, so that investment of certain sum created claim on Temporary
residence permit for parents and up to two of their children. Precise figures were calculated by
author and published.
In the beginning of 2020, Turkey investment in real estate reached 1 459 570 EUR, i.e. 17,60%
of total. The figure resulted by adding data from three other sub-industries of 68 NACE. There
were total 884 deals recorded in Latvian Register of enterprises with Turkey entrepreneur
investors, excluding data of Turkey investors who invested by buying real estate as physical
persons and are counted in statistics by the Office of Citizenship and Migration Affairs (OCMA).
What is included in this research is the entrepreneurial part of such purchases. A limited liability
company is grounded and registered in Latvia as a rule to facilitate commercial use of real estate.
The significance to Latvian economy of the Turkey investment as real estate direct buying in the
period from 2010 till 2014 was noticeable. Since then the process was curbed down.
Investments in IT as part of sustainable growth in Latvia up to now have not met the problematic
of FDI research. Three investors from Turkey were discovered by author in the IT industry (61-
63 NACE), but their share in total was slightly above 0,01%. One investor was active in
Recycling of sorted materials (38.32 NACE), an example of sustainability.
Average Turkey direct investment per deal (project) can be observed on Fig.5
300000.00

250000.00

200000.00

150000.00

100000.00 Investment per Deal,

EUR
50000.00
Investment per
0.00 Recipient, EUR
1991
1994
1997
2000
2003
2006
2009
2012
2015
2018

-50000.00

-100000.00

-150000.00

-200000.00

Figure 5. Average Turkey direct investment per deal and per recipient enterprise in Latvia 1991-
2019, EUR (author, based on Lursoft data, 2020)

278
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The average value of Turkey DI in Latvia is fluctuating around scarce 9500 euro per project and
4800 euro per recipient enterprise. Such sums are too small to revolutionize digital economy and
sustainable growth in Latvia. No reliable trend for growth can be seen.
The investment creates turnover and jobs. Following is the table characterizing top 10 recipient
enterprises with Turkey capital in Latvia according to official reports for 2018.
Table 3
Turnover and jobs Turkey capital in Latvia created in recipient enterprises (2018)

The data in the table showed that turnover in 2018 was ten times higher than investment for all
years. Average turnover per employee is 100 834 EUR.
Deepening of research in sustainability is possible by one by one analysis of deals. The
following chart introduces the case of the ecology friendly and socially responsible enterprise
Jauda-Koks in sawing, planning and impregnation of wood.

Figure 6. Taxes paid by Jauda-Koks SIA (author, based on Lursoft data, 2020)
The chart discloses social responsibility of the enterprise with 100% Turkey capital in Latvia. To
deepen financial analysis, return on assets is 7.96% and return on sales is 6.19%. History of
paying taxes was perfect. Unfortunately, the enterprise was hit by Covid crisis and the State

279
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Revenue Service reported 6 277 EUR debt in taxes 07.07.2020. Against the typical profit 231
266 EUR (2018) this debt proves temporary.
To summarize, enterprises with DI from Turkey, in general, are drivers of sustainability in Latvia
They are present in basic cells of Latvian economy. There are pre-conditions for the increase and
diversification of investments from Turkey to Latvia. But much depends on future political
factors and outcome of Covid 19 crisis.
CONCLUSIONS
For 2 million people strong Latvian economy, 82 million populated Turkey proves significant
investor. 139 investors from Turkey set foot on the ground in Latvia in 119 enterprises in 68
industries. DI relations started in 1991 and nowadays Turkey takes 34th place by volume of DI in
Latvia (EUR 8 332 368). This rank, however, proves disproportionally small against the
economic might of Turkey as the 13th economy in the world. Analysis disclosed uncertainty in
DI flows from Turkey in Latvia. Sigma for DI flows for the whole history is 5.23 times bigger
than the average yearly flow. Cumulative DI figure stagnates at 8.3 million euro. The average
value of Turkey DI in Latvia is fluctuating around scarce 9500 euro per project and 4800 euro
per recipient enterprise. Such sums are too small to secure sustainability in Latvia. No reliable
trend for growth can be seen.
Structural analysis disclosed Support activities for air transport (NACE 52.23) constitute 40% of
total, with Marupe administrative unit known for Riga airport concentrating 62,51% of total DI
from Turkey in Latvia.
For Top 10 recipient enterprises turnover in 2019 is ten times higher than all-time investment.
Average turnover per employee is 100 834 EUR. Return on sales 6,19% and return on assets
7.96% disclosed for Jauda-Koks SIA are typical for majority of enterprises with DI from Turkey
in Latvia. One by one analysis of all enterprises stands in the agenda.
There are a lot of possibilities to attract investment for new business fields, namely digital
economy and sustainable growth despite the fact COVID crisis hit recipient enterprises.
The first Joint Economic and Trade Commission (JETCO) meeting organized on-line 2020 by
Latvian investment and development agency in cooperation with The Board of foreign economic
relations is to be turned into regular body canalizing sustainable DI from Turkey in Latvia. Fine
tuning of the FDI policy should be based on sustainable investment stumulation.
Results of this research can be used in both Latvia and Turkey, e.g. Turkish Industrialists and
Businessmen’s Association, MUSIAD, Chambers of Commerce, Ministries and other bodies
facilitating mutual economic cooperation. Educational value of the research is evident, too.

280
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Feasibility of Turkey investments in Latvia can be questioned if innovation-driven economy is

delayed. The most frequent constraint faced by potential investors in sustainable development
projects is the lack of concrete proposals of sizeable, impactful, and bankable projects.
At the same time it is important to analyse further the structure of Turkey investments in Latvia
by business field of activity, especially digital economy, socially and environmentally oriented
profitable projects, to understand its prospects and needs, to evaluate the development
perspectives, thereby facilitating to the development and support of mutually beneficial
cooperation. New ideas can be connected to 2023, the 100th anniversary of the Republic of
Turkey.
REFERENCES
Accenture Technology Vision 2018. Redefine your company based on the company you keep.
Intelligent Enterprise Unleashed. Accenture.
Agreement Between The Government Of The Republic Of Latvia And The Government Of The
Republic Of Turkey Concerning The Reciprocal Promotion Article II Promotion and protection
of investments of 18.02.1997 [Link] (Retrieved 16.04.2020)
Aharoni, Y., 1999. The Foreign Investment Decision Process. In: P.J. Buckley ed., 2003.
International Business. Dartmouth: Ashgate Publishing [Link]. 87-98.
Business Session of the First JETCO Meeting (2020) [Link]
Latvia-to-Turkey-Latvijas-v%C4%93stniec%C4%ABba-Turcij%C4%81-332135123631580/
(Retrieved 31.07.2020)
Czinkota, M.R., Ronkainen, I.A., Moffett, M.H., 1996. International Business. 4th ed. Orlando:
Harcourt Brace College Publishers.
Dettoni J. (2020) FDI activity hits record lows in April
[Link]
&utm_source=emailCampaign&utm_medium=email&utm_content= (Retrieved 16.06.2020)
Dunning, J.H., Narula, R., eds. 1996. Foreign Direct Investment and Governments. Catalysts for
Economic Restructuring. London: Routledge.
Eiche, D., ed., 1996. Investment in Latvia. Riga: KPMG Latvia.
El-Agraa, A.M., 2001. The European Union. Economics & Policies. Essex: Pearson Education
Limited.
Electronic source of Latvian Republic Register of Enterprises, LURSOFT, 2018. Turkey Foreign
Direct Investment in Latvia. [online]. Available at
˂[Link] 20.11.2019].
Eurofound, European Foundation for the Improvement of Living and Working Conditions, 2018.
Fric Karel, Statutory minimum wages 2018. [online]. Available at

281
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

˂[Link]
eu-2018 ˃[Accessed 2019].
Greenfield investment monitor fDi Markets. Digital Economies of the Future 2018/19 – the
results [online]. Available at ˂[Link]
Pacific/Singapore/Digital-Economies-of-the-Future-2018-19-the-
results?utm_campaign=October+2018+e-
news+2&utm_source=emailCampaign&utm_medium=email&utm_content=˃ [Accessed
20.11.2019].
Global Economy. Turkey [Link] [Accessed
14.04.2020
Graham, E.M., 1997. Should there be Multilateral Rules on Foreign Direct Investment? In: J.H.
Dunning, ed., 1997. Governments, Globalization, and International Business. Oxford: Oxford
University Press. Pp. 481-506.
Griffin, R.W., Pustay, M.W., 1996, International Business. New York: Addison-Wesley
Publishing Company.
Hofstede Insights, 2018. National Culture. Country Comparison Tool. [online] Available at:
<[Link] [Accessed 10 March 2020].
Hill, C.W.L., 2003. International Business: Competing In The Global Marketplace. 4th ed. New
York: McGraw Hill Irwin.
UNCTAD Investment Dispute Settlement Navigator. Latvia as respondent State
[Link] [Accessed 10 November 2019].
Investment and Development Agency of Latvia, LIAA, 2018. Investment Opportunities
Database. [online]. Available at ˂[Link]
opportunities-database˃[Accessed 2020].
Kim, T., 1998. Investing in the Baltic States. Estonia, Latvia, Lithuania. London: Euromoney
Publications PLC.
Investment Policy Framework for Sustainable Development.
[Link] [Accessed 2020]
New Latvian Ambassador in Turkey (1998), in Latvian [Link]
[Accessed 2020]
Legal acts of the Republic of Latvia, 2018. Immigration Law. [online]. Available at
˂[Link] 2020].
Limanskis A. (2014) Non-EU Eurasia Investment for Residence: the Latvian Way.10th EBES
Conference Proceedings 23-25 May 2013. Thomson Reuters Conference Proceedings Citation

282
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Index [Link] ISBN

978-605-64002-1-6
Limanskis A.(2020) US Direct Investments in Latvia under Sustainability Perspective American
International Journal of Business Management (AIJBM) ISSN-2379-106X, Volume 3, Issue 7
(July 2020), PP 20-28 [Link]
Ministry of Foreign Affairs Republic of Turkey [Link]
[Link] [Accessed 16.04.2020]
Penrose, E.T., 1956. Foreign Investment and the Growth of the Firm. In: P.J. Buckley ed.,
[Link] Business. Dartmouth: Ashgate Publishing [Link]. 33-48.
Praude, V., 2010. Investīcijas. Rīga.
Turkey share in the global GDP [Link]
[Accessed 18.03.2020]
Turkey's Strategic Vision 2023. [Link]
[Link] [Accessed 18.04.2020]
Turkey. Foreign Direct Investment in Turkey. Bank of Latvia.
[Link] (retrieved 18.01.2020)
Sauka A. (2020) Foreign Investors on the Impact of Covid-19: Latvia Needs A Strategic Long-
term Plan to Minimise the Negative Effect. Research report [Link]
negative-impact-of-covid-19-on-business-sector-in-latvia-needs-a-strategic-long-term-recovery-
plan-by-the-policymakers/ (retrieved 2.07.2020)
Sekaran, U., & Bougie, R. (2010). Research Methods for Business (5th Edition ed.). Haddington:
Scotprint.
Sorainen, A., ed. 2003. A Foreign Investor in the Baltics. 3 ed. Stockholm: Association of
Swedish Chambers of Commerce and Industry.
Tax debt by Jauda-Koks SIA [Link]
[Accessed 2020].
The Central Bank of Latvia, Bank of Latvia, 2017. Statistics of FDI in Latvia of Bank of Latvia.
[online]. Available at ˂[Link] 2020].
The Foreign Investors' Council in Latvia, FICIL, 2017. Progress report of FICIL’s
recommendations on Investment Protection 2016. [online]. Available at
˂[Link] 2020].
The Office of Citizenship and Migration Affairs, the Ministry of Interior of Republic of Latvia,
2018. Statistic of Turkey Residence Permits in Latvia. [online]. Available at ˂
[Link] 2020].

283
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

United Nations Conference on Trade and Development, 2017. UNCTAD Handbook of Statistics
2016. [online]. Available at ˂[Link]
[Accessed 2020].
United Nations Conference on Trade and Development, 2017. Key Statistics and Trends in
International Trade 2016. [online]. Available at
˂[Link] ˃[Accessed 2020].
United Nations Conference on Trade and Development, 2014. Foreign Direct Investments in
Latvia, Statistics. [online]. Available at
˂[Link] 2020].
Wild, J.J., Wild, K.L., Han, J.C.Y., 2000. International Business, an Integrated Approach. New
Jersey: Prentice Hall.
World Investment Report. Investment and New Industrial Policies. Available at
˂[Link] 2020].
World Investment Report 2018. Country sheet Latvia
˂[Link] ˃[Accessed 2020].
World Resources Institute, 2018., Learning by Doing: Lessons from WRI’s Sustainable
Investing Journey. World Resources Institute. [online]. Available at
˂[Link] 2020].
World Trade Organization, WTO, 2018. Annual Report 2018. [online]. Available at
˂[Link] ˃[Accessed 2020].

284
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Adoption of IoT technology among Aged NCD Patients in Malaysia: A conceptual study
based on the Theory of planned behavior

Abstract

Non-Communicable diseases (NCDs) are rising as the prime cause of death worldwide, as well
as in Malaysia. This is the result of alterations occurring in social and economic factors in
health. Some of these factors include- proliferation of business and promotion, improved style
of living, change in socio-demographics, enhanced financial affordability, effortless travel and
mobility, financial evolution and import-export of unhealthy goods, causing high risk
behavioural deviations and spike metabolic hazards. Elderly people are amid the high chance
gathering of having NCDs because of poor sickness opposition, the progressing impact of way
of life, terrible eating routine, and reduced amount of physical activities. At the end of year
2020, individuals matured 60 years of age and above in Malaysia are anticipated to be 9.7% of
entire populace. With the arrival of numerous wearable gadgets and smart devices, the IoT
(Internet of Things) gadgets are transforming the conventional medical management system
into a further customized and smart one. Hence the objective of this study is to investigate the
factors related to the “theory of planned behaviour” and their impact on the behavioural
intention of IoT adoption among elderly NCD patients in Malaysia. The outcome of this study
is expected to provide meaningful understandings in creating awareness on the use of IoT
among the elderly NCD patients in Malaysia and understanding their behavioural intentions
that can help to maximise the utilisation of the available necessary skills and tools in the health
clinics.

Keywords: Non-Communicable Disease, Internet of Things, Malaysia, Elderly People, Theory

of Planned Behaviour

285
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1. Introduction

Irrespective of revenue, NCDs are substantial and an increasing worldwide communal health
issue that is impacting every nation (Swinburne, Sacks & Hall, 2019).

In the year 2011, The Malaysian National Health and Morbidity Survey which observes NCD
hazards, revealed that the number of obese individuals aging 18 and above, had risen three
times, since 1996; from 4.4% to a whopping 15.1% (Tahir & Noor Ani, 2012). This indicates
that in 2011, nearly 2.5 million Malaysian citizens matched the obesity criterion. One of the
major causes for 60% of disease in Malaysia is due to NCDs. Moreover, there is a notable raise
in number to 72% in 2013 (The Star, 2018).

Every year, NCDs cause over 36 million deaths which is 63% of the sum of deaths worldwide;
this includes 14 million untimely deaths below 70 years. 90 out of 100 of these early demises
take place in nations where the income range is middle to low, which could have been avoided.
Besides poor diet and lack of exercise, abuse of alcohol and tobacco are mutual hazardous
aspects that connect to most of these untimely demises.

Because of enhancements in clinical sciences and advances (Azimi, Rahmani and Liljeberg,
2017), the old total populace is assessed to increment from 8.5 percent of the absolute populace
in 2015 to 12 percent and 16.7 percent in 2030 and 2050, separately (He, Goodkind, & Kowal,
2016). Consequently, added pressure has been experienced by the healthcare management
system due to the rising number in aged population and their medical needs; like- the increasing
amount in medical expenditure as well as higher occurrence of long-term illnesses (Nguyen,
Mirza & Naeem, 2017).

However, human resource and financial constrictions could bound medical management
services otherwise their standard for elderly individuals. Medical care can be significantly
revolutionized through the use of IT. It has the capability to cut down hospital expenses though
delivering good quality, cost-efficient regular care that can be done at home; easing the pressure
on medical management services as well as its resources (Nguyen et al. 2017; Ni Scanaill,
Carew & Barralon, 2006). The wide-ranging term IoT encompasses gadgets and mechanisms
related to sensing, some of which are radars, RFID, GPS, infrared as well as various wearable
technologies.

Older are defenceless against incessant medical issues, also NCDs (Ayernor, Sazlina, Zaiton,
Nor Afiah and Hayati, 2012) These individuals are amid the peak-hazard gathering of having
NCDs because of poor sickness obstruction, the continuous impact of way of life, less than

286
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

stellar eating routine and less physical exercises. By 2020, man matured 60 and over in
Malaysia are anticipated to be 9.7% of all out populace (Samsudin, Abdullah and Applanaidu,
2016). This demonstrates the potential ascent in medicinal services use later on and the fast
procedure of urbanization and country advancement may fuel it.

The advent of NCDs as the principal reason of death in Malaysia as well as worldwide were
the result of numerous alterations taking place in social and economic factors in health. Some
of these factors include- proliferation of business and promotion, improved style of living,
change in socio-demographics, enhanced financial affordability, effortless travel and mobility,
financial evolution and import-export of unhealthy goods, causing high risk behavioural
deviations and spike metabolic hazards. Hence the objective of the current research study is to
examine the factors related to the “theory of planned behaviour” and their impact on the
behavioural intention of IoT adoption for elderly NCD patients in Malaysia. The subsequent
sections of this paper will discuss about the theoretical framework, proposed research method
and conclusion along with the necessary indications for the new research directions.

2. Literature Review and Theoretical Framework

Internet of Things (IoT)

The idea of “Internet of Things” (IoT) comprises of a collection of inspections and control
applications. These rely on a system of detecting and impelling gadgets. Thus it can act
logically arranged conditions and governed from far away through the use of Internet (Li et al.,
2011; Solima et al., 2016).

With the arrival of numerous wearable gadgets and smart devices, the IoT (Internet of Things)
gadgets are transforming the conventional medical management system into a further
customized and smart one. As a result, the current healthcare system is addressed otherwise as
Personalized Healthcare System (PHS).

Besides a significant cutback on costs because of improved sustainability, a notable

development in patient centric practice can be empowered if IoT technology were to be
accompanied by cloud computing. In order to observe health issues, numerous researches have
been conducted and ‘Smart Wearable Devices (SWH)’ have emerged in the past few years
(Chan et al., 2012).

287
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The medical fitness and overall welfare of aged people can be monitored by the utilization of
IoT tools and technology (Mirza, 2020). Ambient assisted living, active ageing, therapy and
entertainment are some of the IoT implementation instances. The core purpose of this set of
applications are group observation, enhancement of lifestyle, encouragement to independent
and safe living.

Theory of Planned Behaviour (TPB)

In the “Theory of Planned Behavior” (TPB) context, the primary study has been on the
utilization within the marketing, promotions and PR framework (Ferdous, 2010). Not many
studies have touched on the implementation of IoT from the standpoint of multiple theory, such
as- the “Theory of Planned Behavior (TPB)”, the “Theory of Reasoned Action (TRA)” (Mital
et al., 2017).

TPB is considered most suitable for this study to understand the behaviour towards the adoption
of IoT. Expectations, which are considered to show the measure of exertion any particular
person is probably going to dedicate to playing out a conduct, are thusly dictated by
perspectives, a general assessment of the conduct, emotional standards (SN), an assessment of
whether a person feels critical other people (he/she) think ought to take part in the conduct and
“perceived behavioral control (PBC)”, which speaks to a person's view of authority over that
conduct. Past researchers have found positive associations among the antecedents and intention
to use certain products.

The capability of the TPB as a model to explore the wellbeing practices was affirmed by an
audit led by McEachan et al. (2009). Their study found that the hypothesis represented
somewhere in the range of 14 and 24 percent of study fluctuation in conduct. Hence the
proposed hypotheses for this study are as follows:

H1(a): There is a positive association between “Intention to use IoT” and attitude of
Aged NCD patients in Malaysia.

H1(b): There is a positive association between “Intention to use IoT” and subjective
norm of Aged NCD patients in Malaysia.

H1(c): There is a positive association between “Intention to use IoT” and perceived
behavioral control of Aged NCD patients in Malaysia.

288
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Attitude

Subjective norm Behavioral Intention

(to use IoT)

Perceived
Bewhavioral Control

Fig. 1. Conceptual Framework based on Theory of Planned Behavior (TPB)

3. Proposed Research Method

The survey will contain three segments. Segment one will have questions related to the
participants’ socio-demographic info. Segment two questions will focus on the participants’
knowledge, attitude and practice regarding NCDs e.g. heat diseases, hypertension, stroke,
diabetes, cervical cancer etc. The last segment will contain queries on attitudes, perceived
barriers and practice and behavioural intentions associated with the use of IoT among the
elderly NCD patients in Malaysia. A “7-point Likert scale (1- strongly disagree; 7 – strongly
agree)” will be used in this questionnaire.

The four items of attitude towards adoption will be adapted from Madden et al. (1992); three
items of subjective norm adapted from Madden et al.(1992); three items of perceived
behavioural control is adapted from Hair et al. (2013); The four items of adoption intention
will be adapted from Mathieson (1991).

The SEM strategy is valuable since it expects that singular factors are transformed individually
with the other parts in the research model. Afterwards, the resultant model fit files are checked
in the estimation part of the model. As the investigation explains itself and the appropriation
of IOT remains a developing stage in Malaysia in case of NCD patients, the study intends to
use sample of 100 elderly NCD patients as respondents.

4. Conclusions and Future Research

Clinics or health care centres are usually the initial place where patients look for medical
attention. In such case, the shift of demography and epidemiology of these centres can impose
major challenges. For individual as well as population level activities; chronic NCDs like heart

289
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

diseases, hypertension and diabetes are challenging to handle. They also demand intensive
labour from every component of the medical system. Identifying the disease early and primary
risk factor management can help prevent NCDs.

Through creating awareness on the use of IoT among the elderly NCD patients in Malaysia and
understanding their behavioural intentions can help to maximise the utilisation of the available
essential services as well as equipment in the medical centres. Moreover, further plan should
be outlined to allocate services and equipment to communal level. Studies in the future should
concentrate further on the clinical aspect and utilization of IoT wearables. The usability and
cost efficiency of these are also to be thoroughly analysed in respect to the elderly population
in Malaysia so that, a meaningful usage is ensured. Future research should focus on the
empirical outcomes based on the model in this study to get more deeper insights on the
behavioural intentions of the elderly NCD patients as well.

References:

Ajzen I (1991). The theory of planned behavior. Organ Behav Hum Dec. 1991;50(2): 179–211.
doi:10.1016/0749-5978(91)90020-T

Ajzen I (1985). From intentions to actions: A theory of planned behavior. In: Kuhl J, Beckman
J, editors. Action-control: From cognition to behavior Heidelberg: Springer; 1985. p. 11–39

Azimi I, Rahmani AM, Liljeberg P et al (2017) Internet of things for remote elderly monitoring:
a study from user-centered perspective. J Ambient Intell Humanized Comput 8:273–289.

Ayernor P. (2012). Diseases of ageing in Ghana. Ghana Med J. 2012;46(2 Suppl):18-22

Brownsell S, Hawley M. (2004) Fall detectors: do they work or reduce the fear of falling? Hous
Care Support 7:18–24

Chan, M., Estève, D., Fourniols, J.-Y., Escriba, C., Campo, E., (2012). Smart wearable
systems: current status and future challenges. Artif. Intell. Med. 56 (3), 137–156

Fishbein M, Ajzen I. (1975). Belief, attitude, intention and behaviour: an introduction to theory
and research. Reading MA: Addison-Wesley; 1975.

290
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Goodkind D, Kowal PR (2016) International population reports. US Census Bureau.

[Link]
[Link]. Accessed 19 Sep 2019

Hair Jr, J.F., Hult, G.T.M., Ringle, C., Sarstedt, M., (2016). A Primer on Partial Least Squares
Structural Equation Modeling (PLS-SEM). Sage Publications

Hazriani, Yasmin B. H. Ooi, Lee and Patricia Matanjun. Non-communicable diseases among
low income adults in rural coastal communities in Eastern Sabah, Malaysia

Li, X., Lu, R., Liang, X., Shen, X., Chen, J., Lin, X., (2011). Smart community: an internet of
things application. IEEE Commun. Mag. 49 (11), 68–75.

MacCallum, R.C., Austin, J.T., (2000). Applications of structural equation modeling in

psychological research. Annu. Rev. Psychol. 51 (1), 201–226

McEachan RRC, Conner M, Taylor NJ, Lawton RJ. Prospective prediction of health-related
behaviours with the Theory of Planned Behaviour: a metaanalysis. Health Psychol Rev.
2011;5(2):97–144. doi:10.1080/17437199.2010. 521684

Madden, T.J., Ellen, P.S., Ajzen, I., 1992. A comparison of the theory of planned behavior and
the theory of reasoned action. Personal. Soc. Psychol. Bull. 18 (1), 3–9.

Máirtín S. McDermott, Oliver, Svenson, Simnadis, Eleanor J. Beck, Coltman, Iverson, Caputi
and Sharma. The theory of planned behaviour and discrete food choices: a systematic review
and meta-analysis

291
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The role of innovation in insurance and its response to Covid-19 pandemic

emergency

Davide Lanfranchi, Marco Giorgino, Laura Grassi*

* Politecnico di Milano, Department of Management Engineering, Italy

Abstract
The role of insurance as a social protection mechanism is relevant, mitigating the effects of
uncontrollable events and allowing individuals and corporations to recover from unfortunate events,
including Covid-19. The relevance of the initiatives undertaken by insurance companies responding
to Covid-19 emerge clearly, with implications on health issues, economic and financial issues,
work-related issues (e.g. risks from remote working and teleworking), and finally fostering
everyday-life continuity.
Insurance companies have traditionally struggled with innovating, despite the nature of their
products and services. However, technological improvements and market trends are fostering
innovation. Digitalization, socio-economic changes, consumer demand for a better offer as well as
protection from new risks generate additional innovation impulses.
As the likelihood of infectious diseases was considered limited, the response of insurance
companies to Covid-19 gives us the unique opportunity to understand innovation mechanisms in the
industry, which of their main activities are the most affected by these innovations and which are the
societal benefits.
We adopted an inductive approach and a qualitative process analysis, focusing on the initiatives to
respond to Covid-19 emergency of the most impactful insurance companies at world level.
Four rationales of initiative responding to Covid-19 pandemic emerge, discriminated in terms of
relevance of the technology and innovation of the portfolio of covered risks. Similarly to other
contexts, market demand from unsatisfied customer needs (i.e. Covid-19 implications) opened the
way for new processes, services and products despite the complexity resulting from insuring new
risks in conditions of scarce historical data and no models to accurately forecast scenarios.
This work contributes to the understanding of how insurance initiatives are affecting society by
addressing increasing health, economic and financial and work-related issues.
Moreover, this work can be the starting point for further research, in particular about the study of
the short- versus medium-long term societal benefit of those initiatives and the perdurance of
innovation impulse in the industry.

Keyword: Insurance, Innovation, Covid-19, Risk, Insurtech

292
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1. Introduction
The role of insurance as a social protection mechanism is one of the most important when dealing with its
benefits, mitigating the effects of uncontrollable events and allowing individuals to recover from unfortunate
events by relieving or at least limiting the financial burden (The Geneva Association, 2012).
Through insurance, economic and financial responsibility for several risks, such as illness, death, or damage
is mutualized by spreading it across a group (McFall and Moor, 2018). In fact, the uncertain loss related to
events of unknown magnitude and time span is exchanged for a known cost, namely the premium (Zweifel et
Eisen, 2012). In 2018, global direct premiums were above USD 5.2 trillion (Swiss Re, 2019).
“In a world of increasing uncertainty and dynamics, the economic and social importance of being insured
seems undisputed” (Stoeckli et al., 2018, p.287). From an economic point of view, insurance is fundamental
since it deals with the negative consequences of economic activity that would occur in its absence (Zweifel et
Eisen, 2012), while from a social point of view, insurance is crucial, since it provides a social protection
mechanism (The Geneva Association, 2012).

Insurance companies have traditionally struggled with innovating and changing due to their conservative
attribute (Nam, 2018): for instance, even if the intangible nature of their products and services could enable
them to become digital leaders with clear digital business cases (EY Global Insurance, 2013), the
transformation of the insurance industry has come rather late (Eling et Lehmann, 2018).
Moreover, certain innovations, such as the use of self-tracking data to assess and price individual risk, are
fraught with practical, regulatory and reputational obstacles (McFall et Moor, 2018).
However, despite the above-mentioned issues, we maintain insurance industry has to innovate, to consider
socio-economic changes, to which it is connected in several ways (Bohnert et al., 2019). New developments,
trends and changes in the society and in the economy affect indeed the demand for insurance (Bohnert et al.,
2019).

Insurance, by transferring the risk of a loss (Dorfman and Cather, 2012), can play an important role in
protecting people from losses arising from the current pandemic.
Due to the above-mentioned linkage between insurance and socio-economic changes, insurance companies
need to change and innovate to properly respond to changing situations and needs. Hence, it is important to
understand how insurance companies are innovating to answer to Covid-19 pandemic emergency.

How are insurance companies responding in an innovative way to Covid-19? Which are the main activities
affected by these innovations? Which are the societal benefits? How can we learn for the future? Examining
their response can provide insight into their way to innovate.

To address these issues, we explored the innovations set by such insurance companies having serious impact
on the society by conducting an inductive analysis of the 30 most representative insurance companies at
world level in the first quarter of their response to the Covid-19 pandemic. Our study builds on technological
and market innovation research to offer a conceptualization of potential innovation ways of insurance
companies (Section 2). We carefully describe our sample and methodological approach (Section 3). Main
outcomes are then presented (Section 4) and discussed, proposing a response to Covid-19 innovation
framework in the insurance industry (Section 5).

293
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2. Theoretical background
Literature about innovation processes is flourish and different models emerged during time (Du Preez and
Louw, 2008).
Technology and Market demand have been recognized as two main impulses to innovation (Brem and Voigt,
2009; Voss, 1984) and widely studied by academics (Maier et al., 2016). Technology, often resulting from
internal or external research (Maier et al.,2016) can enable the creation of commercialized innovative
products (Du Preez and Louw, 2008; Maier et al., 2016) as well as enable innovation in services (Geum et
al., 2016) and processes (Brem and Voigt, 2009). In the same way, the market itself and customer needs can
be source for new ideas (Du Preez and Louw, 2008) aiming at satisfying consumer demands (Nicolov and
Badulescu, 2012).
These considerations are valid also in the context of insurance Industry, where nevertheless literature
studying the phenomenon is scarce (Kose et al., 2018). Technology and Market demand roles in fostering
innovation are hereby presented for the specific insurance context.

2.1 The role of Technology

Digital transformation has become an important enabler of innovation (Urbinati et al., 2020) and insurance
sector is interested as well. Eling and Lehmann (2018) analyzed the impact of digitalization on the insurance
value chain: the main areas appear to be the way insurance companies interact with their customers and adapt
to their behavior, the automatization of business processes and decisions and improvements in the existing
products and new products offerings. In this direction, Insurtech, a new “phenomenon comprising
innovations of one or more traditional or non-traditional market players exploiting information technology to
deliver solutions specific to the insurance industry” (Stoeckli et al., 2018, p.289) is gaining interest, driven
by increased customer satisfaction and efficiency (McKinsey, 2018).
The theme has a high importance at societal level as well: Insurtech can bring both opportunities, such as
higher insurance inclusiveness (Altamirano and van Beers, 2018) and public health improvement (Yamasaki
and Hosoya, 2018), and threats, such as privacy concerns (Banerjee et al., 2018).

2.2 The role of the market

Insurers can provide protection, addressing the concerns of potential investors who consider insurance a
prerequisite (McAlea et al., 2016), they can reduce public anxieties and concerns, understanding risks - long-
term in particular (McAlea et al., 2016), and can help entrepreneurs, individuals or corporation to handle risk
and support further gradual advancement by proposing new products, also to cover new risk (Śliwński et al.,
2017).
But in order to play a central role for the society, to create value for customers by transferring the risk of a
loss from one entity to another in exchange for a payment (Dorfman and Cather, 2012), they have to
promptly serve the current and perspective needs of the market. Consumers are in fact demanding more and
more convenience, quality, and suitability of the offer (Kose et al., 2018), as well as protection from new
risks, generating additional innovation impulses. Therefore, insurance can innovate its products and
processes in a twofold way: dealing with market demand and customers’ needs and addressing the rise of
new risks.
The plethora of different emerging risks is getting increasingly important, with risks deriving from changing
business environment, disruptive environmental patterns, evolving social and demographic trends,
technological advancements and new medical and health concerns (Capgemini and Efma, 2019).
Should the forecast of their likelihood (and impact) be straightforward, it’s not. While risks such as cyber
risk or extreme weather risk were particularly under the attention of industry, considered in the Top 10, the
risks related to infectious diseases were seriously less considered, ranking out of Top20 in terms of
likelihood (World Economic Forum, 2020).

294
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3. Methodology
We used an inductive approach and a qualitative process analysis. We consider case study as an adequate
approach considering the exploratory nature of our work and the deep understanding we would like to reach
when it turns to the societal response of insurance companies to the Covid-19 pandemic.
By focusing on top insurance companies at world level we had the possibility to study all the different
responses to the Covid-19 pandemic whose societal impact applies at large, with the aim of sharing
meaningful experiences so that they can be useful both for the current crisis and for future ones.
This study focuses on innovations in the insurance industry with a clear societal outcome, with the unit of
analysis being each initiative undertaken. Companies of the sample were retrieved from Orbis database,
ranking them by Net Written Premium. An overview of each company is reported in Appendix A.
Initiatives from all these insurance companies and groups where systematically mapped, for a final number
of 101 initiatives.
We selected insurance companies’ websites as Primary source, searching for results with the following
characteristics: (i) real initiatives of the company, not opinion papers or suggestions for the industry and (ii)
initiatives developed for facing Covid-19 emergency.
Data were supplemented by a deep analysis of secondary sources such as news (e.g. CNBC), articles from
industry-specific and business magazines (e.g. Forbes), interviews reported by the press (e.g. Dirk
McMahon, UnitedHealth Group CEO) and, where possible, directly testing the tools under study (e.g.
Humana Chatbot).
By applying an iterative process, involving all the authors to reduce the personal bias, we
analysed the various sources searching for information or details useful for addressing the objective of this
work.

4. Results
Overall, 101 initiatives were mapped, ascribable to 20 insurance companies. Distribution of initiatives
among insurance companies is shown in Fig. 1.
Most of insurance companies show specific initiatives for facing Covid-19 emergency, with some companies
clearly stating their commitment in the pandemic emergency response (more than 5 initiatives in a couple of
months).
Initiatives ranged in different areas and accordingly, we decided to initially map their impact on the different
activities of the insurance value chain framework (Rahlfs, 2007) following Eling and Lehmann (2018)
presentation of results.
Table 1 presents exemplificative initiatives for each activity of the insurance value chain (where possible)
while Fig. 2 shows their distribution by activity.
Concerning Primary activities, initiatives spread in any activity, from Product innovation, such as creation of
solution aiming at facing Covid-19 infection risk (e.g. Artificial Intelligence-based symptoms checkers by
Allianz for better sorting patients) to innovations regarding services and processes (e.g. remote claim invoke
by phone, web, e-mail or app, by Assicurazioni Generali) (see Fig. 2).
Support activities were impacted as well, at least, where nevertheless in the HR practice, with many
companies putting in place teleworking solutions (e.g. UnitedHealth and Credit Agricole Assurances).

295
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5. Discussion
Examining the 101 initiatives revealed an important role of technology and market needs in the insurance
companies’ innovative responses to Covid-19, with reference to all the activities of the insurance value
chain, as represented in Fig. 2.

5.1 The role of technology

Focusing on technology, insurance companies responded to the current pandemic emergency in two main
ways.
On one side, insurance companies leveraged their existing technological features, empowering their use or
opting for the adoption of new ones. In the U.S, Humana trained its already existing Chatbot solution for a
timely detection of suspicious Covid-19 contagion. On top of Marketing effect, it also suggests adequate
Humana products to be covered. Several insurance companies, such as Aetna, have made telemedicine
solutions available to their customers, including for instance live-video conferencing with medical experts,
with the aim of limiting potential exposure to COVID-19 in physician offices.
On the other hand, technology was not always central in insurance companies’ response to Covid-19. For
instance, several insurance companies, such as UnitedHealth, extended grace periods to delay premiums
payment (especially for those who used to pay in cash in a physical agency or those suffering temporary
financial difficulties). Others, such as National Mutual Insurance Federation of Agricultural Cooperatives or
Anthem, extended the existing health insurance coverages to explicitly include coronavirus infections.

5.2 The role of the market

In the current pandemic emergency, insurance companies have found themselves dealing with new risks, but
at the same time they had to not forget the already existing ones.
Risk of contagion by Covid-19 is clearly central in the extension of already existing products. Customers
benefitted of extended coverage on their policy while doctors similarly on their Professional Civil Liability
insurance (e.g. Credit Agricole Assurances’ subsidiary La Médicale) with reference to the practice of
telemedicine and for all acts they perform beyond their usual sphere of expertise while combating the
epidemic.
Collateral psychological issues related to lockdown measures raised concern. Responses ranged from 24/7
hotline resources for living the crisis (e.g. Aetna), to COVID-19 microsite and emotional support (e.g.
Aetna), or free Netflix and Spotify subscriptions (e.g. Assicurazioni Generali in Turkey).
Nutrition and wellness issues were faced with free advice and consultations with certified personal trainers,
chefs and dieticians, in addition to discounts for grocery home delivery (e.g. Assicurazioni Generali in
Turkey).
Insurance companies have developed initiatives related to many primary activities of the value chain,
affected as well by the consequence of the Covid-19. For instance, considering the Sales activity, Prudential
PLC increased efforts towards digital sales, with 25 per cent of their products now sold virtually in Hong
Kong. Considering the claim management activity, an example comes from Zurich with its supplier Prism,
able to establish the cause of a house leak through a ‘drive-by’ survey.

5.3 A response to Covid-19 innovation framework

From previous considerations, it emerges how cases can be discriminated in terms of relevance of the
technology: in one class of cases, there is the empowerment of existing technologies or even technological
innovation in order to answer to the emergency; in the other, technology does not play a major role, and no
empowerment or innovation is needed.
Moreover, cases can be classified depending on the innovation of the portfolio of risks covered by insurance
companies; in the first class, insurance companies innovate their portfolio of covered risks, launching
initiatives in order to create value for the customers addressing new perils; in the second class of cases,
insurance companies adopt solutions aiming at dealing with already existing risks.

Building on the evidence from the cases, four different rationales of initiatives responding to Covid-19
pandemic emerge (Fig. 3). Some examples are reported in Appendix B.

296
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.3.1 Adaptation
Initiatives addressing needs related to pre-existing risk and where technology does not play a major role.
Many of those initiatives are related to the Underwriting activity (7 out of 17): examples are flexibility in
premium payments (e.g. CNP Assurances) or the opening of new special enrolment periods to allow
corporations to let employees who previously declined health benefit be covered (e.g. UnitedHealth).
In Contract Administration & Customer Service activity (7 initiatives out of 17) insurance companies ex-post
adapted the premium, by reducing it, as a consequence of the reduction in claims (e.g. Liberty Mutual, 15%
refund on two months of customers’ car premium).

5.3.2 Expansion
Solutions addressing new risks, still without requiring a major role of technology. Initiatives in this class are
mainly in the Contract Administration & Customer Service activity (11 out of 21): several insurance
companies (e.g. CVS Health Group’ subsidiary Aetna, Anthem or UnitedHealth) are expanding the covered
events of their health insurance policies, such as waving copays, coinsurance and deductibles for the
diagnostic test related to COVID-19 or treatment of COVID-19 or health complications associated with
COVID-19. Moreover, insurance companies such as Aetna are enabling expedited access to treatment: in
States like New York and Washington with a strongest prevalence of COVID-19 cases, hospitals no longer
need advance approval from Aetna for members requiring hospitalization for COVID-19.
Expansion initiatives are present as well concerning the Product Development activity (7 out of 21): for
instance, Assicurazioni Generali is expanding its portfolio by offering an insurance to all Chinese Medical
Experts in Italy, to fully protect their safety and health.

5.3.3 Reaction
Initiatives addressing pre-existing risks, with the aim to react to the difficulties arising from the pandemic
emergency, to keep serving customers in an effective and efficient way, by empowering or adopting new
technologies.
Some of those initiatives concern the Contract Administration & Customer Service activity (8 out of 30),
with several insurance companies (e.g. Humana) temporarily waiving member out-of-pocket costs for
telehealth visits (also known as telemedicine). Similarly, due to the lockdown restrictions and the need of
reducing the movement of people, Aetna expanded coverage of telehealth and offered all telehealth visits
with their network providers at no cost to all members.
Claim management activity as well sees an important number of Reaction initiatives (7 out of 30), with many
insurance companies (such as Liberty Mutual Insurance or Zurich Insurance) stopping all in-home damage
inspections to avoid having employees entering homes and conducting remotely the claim assessment,
through video chat or video collaboration tools.
Concerning Sales activity (3 out of 30), an example comes from Life Insurance India that, under the
restrictions imposed by various local authorities, reduced the availability at physical touch points like
branches and call-center, requiring people to leverage more the internet and opting for online purchase,
advertising on its cheaper cost.
Consistently with Eling and Lehmann (2018), it is clear how the alternative responses in the market build on
the digital interaction between customers and insurance companies since the beginning during the sales
process (e.g. Life Insurance India) and for all the time of validity of the policy, whether a claim occurs (e.g.
Zurich Insurance) or not (see for instance wellness advices provided by Prudential PLC).

5.3.4 Aggression
Initiatives empowering or leveraging new technologies to address new risks.
Many insurance companies started leveraging technologies to develop products addressing new risks.
Medi24, a Switzerland-based telehealth subsidiary of Allianz, is exploring the use of Artificial Intelligence-
based symptom checkers for sorting patients more efficiently, wearables and diagnostics to better understand
the needs of patients and steer them to their nearest healthcare facilities. Similarly, Anthem is working to
accelerate the availability of a Coronavirus Assessment on the Sydney Care mobile app, which members can
download at no cost.
Covid-19 pandemic brought together collateral risk, somehow related to the lockdown. UnitedHealth
established a navigation support program for its members who are under home isolation due to COVID-19
diagnosis or exposure.

297
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

This is again consistent with results by Eling and Lehmann (2018), highlighting how digitalization can affect
existing products and new products offerings: in this context, it is clear how digitalization enables insurance
companies to offer enrichment of traditional products centered on protection, such as guiding members under
home isolation due to COVID-19 diagnosis or understanding their real exposure to Covid-19 (e.g.
UnitedHealth).
Several insurance companies focused on Marketing initiatives: an example comes from Humana,
empowering its Chatbot tool to allow people to take a pre-assessment of a possible Covid-19 infection,
suggesting a health insurance policy in specific cases.
Regarding Human Resources support activity, many insurance companies started making their employees
teleworking, to reduce the spread of the disease and protect them (e.g. Credit Agricole Assurances).

Fig. 4 shows the distribution of initiatives, depending on the cluster they belong to, along the insurance value
chain.
Most of Aggression initiatives are related to Product Development activity, with the creation of technological
enabled products addressing new risks. Expansion initiatives are more related to Contract Administration &
Customer Service activity, being often extension of already existing (and purchased) products to cover also
new risks. A good number of Reaction initiatives concern Claim Management activity, being insurance
companies leveraging technologies to keep serving customers. Finally, Adaptation initiatives are frequent in
the Underwriting activity, with discounts or extension for the payment of premiums.
While Contract Administration & Customer Service are proposing several initiatives to manage the daily
routine, we noticed something similar in Product Development. Therefore, market demand from unsatisfied
customer needs showed potential room for new products (Maier et al., 2016) despite the difficulties that from
a practical point of view insurance companies face due to the complexity resulting from insuring new risks in
conditions of scarce historical data and models to accurately quantify risks (Śliwński et al., 2017).

The relevance of such initiatives on the society emerges, from several point of view.
Some innovations respond to increasing health issues, from Covid-19 infection risk to psychological issues
due to the lockdown (see for instance 24/7 hotline resources for living the crisis by Aetna). Other initiatives
respond more to economic and financial issues, as the decision by UnitedHealth of allowing grace periods
for employees and individuals to pay premiums. Certain initiatives aim at solving work related issues in
insurance companies, in particular by allowing teleworking, but as well by extending coverages to risk
coming from increase in teleworking in other industries: in this direction, for instance, AXA extended
guarantees to business clients for specific needs, such as the coverage of cyberattacks for companies, since
most of their workers are working remotely. Finally, other initiatives aim at guaranteeing the continuity of
everyday life of customers, for instance by running in a digital way home inspection in case of damages (see
Zurich Insurance). Table 2 shows some examples of initiatives having different benefits on the society.

298
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

6. Conclusions
Covid-19 pandemic emergency gave rise to several important issues for society. Due to their socio-economic
importance, insurance companies are eligible to play an important role in facing all those problems.
Many insurance companies have already supported the population, for instance by donating important
amount of resources to the health system (e.g. Allianz donating 350,000 surgical masks to hospitals) or
supporting people suffering economic and financial distress (e.g. Generali France taking part of the € 200
million insurance Federation contribution to the € 1 billion solidarity fund created by the French government
dedicated to VSEs, SMEs and self-employed workers in difficulty).
At the same time, even if the insurance industry traditionally struggled with innovating and changing due to
its conservative attribute (Nam, 2018), we found insurance companies innovating to answer to Covid-19
pandemic emergency.

This research provides important insight into the innovations of insurance companies as a response to Covid-
19 situation. With the emerging of a new risk, insurance companies had the opportunity to re-conceptualize
their value chain and develop new products and processes, leveraging technology and customers’ needs.
Sudden but unexpected innovations which opened the way for new business opportunities for such market
operators, highlighting once more their role in providing a social protection mechanism (The Geneva
Association, 2012).

Indeed, the relevance of such initiatives on the society emerges, from several point of view. Some
innovations respond to increasing health issues, others more to economic and financial issues and others to
work related issues in insurance companies, in particular by allowing teleworking, but as well by extending
coverages to risk coming from increase in teleworking in other industries, and finally by fostering everyday
life continuity.

Due to the novelty of the phenomenon and the potential emergence of new initiatives, further research could
focus on how this framework should be validated or modified due to future developments.
Moreover, it would be interesting to study how Covid-19 pandemic emergency will affect insurance value
chain and the profitability and economic results of insurance companies. Further, some research is needed in
evaluating such innovative initiatives, understanding their short- versus medium-long term societal benefit
the perdurance of innovation impulse in the industry.
Finally, focusing on the role of technology, it is interesting to study how Covid-19 pandemic emergency is
affecting the interest and adoption of Insurtech innovations in the insurance industry.

299
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

7. References

Altamirano, M. A., & van Beers, C. P. (2018). Frugal innovations in technological and institutional
infrastructure: Impact of mobile phone technology on productivity, public service provision and
inclusiveness. The European Journal of Development Research, 30(1), 84-107.
Banerjee, S., Hemphill, T., & Longstreet, P. (2018). Wearable devices and healthcare: Data sharing
and privacy. The Information Society, 34(1), 49-57.
Bohnert, A., Fritzsche, A., & Gregor, S. (2019). Digital agendas in the insurance industry: the
importance of comprehensive approaches. The Geneva Papers on Risk and Insurance-Issues and Practice,
44(1), 1-19.
Brem, A., & Voigt, K. I. (2009). Integration of market pull and technology push in the corporate
front end and innovation management—Insights from the German software industry. Technovation, 29(5),
351-367.
Capgemini and Efma (2019). World Insurance Report 2019
Dorfman and Cather (2012). Introduction to Risk Management and Insurance
Du Preez, N. D., & Louw, L. (2008, July). A framework for managing the innovation process. In
PICMET'08-2008 Portland International Conference on Management of Engineering & Technology (pp.
546-558). IEEE.
Eling, M., & Lehmann, M. (2018). The impact of digitalization on the insurance value chain and the
insurability of risks. The Geneva Papers on Risk and Insurance-Issues and Practice, 43(3), 359-396.
EY Global Insurance (2013). Insurance in a digital world: the time is now
Geum, Y., Jeon, H., & Lee, H. (2016). Developing new smart services using integrated
morphological analysis: integration of the market-pull and technology-push approach. Service Business,
10(3), 531-555.
Kose, I., Guner, S., Isguzerer, B., & Sisli, M. E. (2018, September). A case study of the extended
interactive innovation management model in insurance company. In International Conference on Innovation
and Entrepreneurship (pp. 375-XVI). Academic Conferences International Limited.
Maier, M. A., Hofmann, M., & Brem, A. (2016). Technology and trend management at the interface
of technology push and market pull. International Journal of Technology Management, 72(4), 310-332.
McAlea, E. M., Mullins, M., Murphy, F., Tofail, S. A., & Carroll, A. G. (2016). Engineered
nanomaterials: risk perception, regulation and insurance. Journal of Risk Research, 19(4), 444-460.
McFall, L., & Moor, L. (2018). Who, or what, is insurtech personalizing?: persons, prices and the
historical classifications of risk. Distinktion: journal of social theory, 19(2), 193-213.
McKinsey (2018). Digital Insurance in 2018
Nam, S. (2018). How Much Are Insurance Consumers Willing to Pay for Blockchain and Smart
Contracts? A Contingent Valuation Study. Sustainability, 10(11), 4332.
Nicolov, M., & Badulescu, A. D. (2012, October). Different types of innovations modeling. In
Annals of DAAAM for 2012 & Proceedings of the 23rd International DAAAM Symposium (Vol. 23, No. 1,
pp. 1071-1074).
Rahlfs, C. (2007) Redefinition der Wertschopfungskette von Versicherungsunternehmen, Gabler
Edition
Wirtschaft. Wiesbaden: Deutscher Universitats-Verlag."
Śliwński, A., Karmańska, A., & Michalski, T. (2017). European Insurance Markets in Face of
Financial Crisis: Application of Learning Curve Concept as a Tool of Insurance Products Innovation–
Discussion. Journal of Reviews on Global Economics, 6, 404-419.
Stoeckli, E., Dremel, C., & Uebernickel, F. (2018). Exploring characteristics and transformational
capabilities of InsurTech innovations to understand insurance value creation in a digital world. Electronic
Markets, 28(3), 287-305.
Swiss Re (2019). Sigma 3/2019: World insurance: the great pivot east continues
The Geneva Association (2012). The Social and Economic Value of Insurance
Urbinati, A., Chiaroni, D., Chiesa, V., & Frattini, F. (2020). The role of digital technologies in open
innovation processes: an exploratory multiple case study analysis. R&D Management, 50(1), 136-160.
Voss, C. A. (1984). Technology push and need pull: a new perspective. R&D Management, 14(3),
147-151.
World Economic Forum (2020). The Global Risks Report 2020

300
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Yamasaki, K., & Hosoya, R. (2018, August). Resolving Asymmetry of Medical Information by
using AI: Japanese People's Change Behavior by Technology-Driven Innovation for Japanese Health
Insurance. In 2018 Portland International Conference on Management of Engineering and Technology
(PICMET) (pp. 1-5). IEEE.
Zweifel, P., & Eisen, R. (2012). Insurance economics. Springer Science & Business Media.

301
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figures

List of Figures
Figure 1: distribution of the number of initiatives undertaken by each insurance company
Figure 2: distribution of initiatives along the insurance value chain (insurance value chain from Rahlfs
(2007))
Figure 3: response to Covid-19 innovation framework. Each initiative is mapped considering the role of the
technology and the risk at which it refers. Four rationales emerge: Adaptation, Expansion, Reaction,
Aggression. Numbers in each sector refer to our coding of each initiative. Encircled data refer to the
numerousness of initiatives.
Figure 4: distribution of initiatives along the insurance value chain depending on the rationale (insurance
value chain from Rahlfs (2007))

Figure 1: distribution of the number of initiatives undertaken by each insurance company

Figure 2: distribution of initiatives along the insurance value chain (insurance value chain from Rahlfs (2007))

302
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 3: response to Covid-19 innovation framework. Each initiative is mapped considering the role of the technology and the risk
at which it refers. Four rationales emerge: Adaptation, Expansion, Reaction, Aggression. Numbers in each sector refer to our coding
of each initiative. Encircled data refer to the numerousness of initiatives.

Figure 4: distribution of initiatives along the insurance value chain depending on the rationale (insurance value chain from Rahlfs
(2007))

303
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Tables
List of Tables
Table 1: Exemplificative initiatives for each value chain activity
Table 2: Exemplificative initiatives with different benefits on the society

Table 1: Exemplificative initiatives for each value chain activity

Tasks (Eling
Primary
and Organizatio
activity Description Source
Lehmann, n Name
(Rahlfs, 2007)
2018)
Market and
customer
research:
researching
Humana Chatbot starts the conversation asking to Human
ideas for
the person whether he/she is suffering life- a
product
threatening emergency that requires immediate Chatbot
development;
attention. In case his / her first concern in that : covid-
Analysing
moment is Coronavirus, the Chatbot starts an healthb
target groups;
Marketing HUMANA assessment process, ending with suggesting [Link].
Development
telehealth support or simple steps for preventing [Link]
of pricing
disease. Depending on the path, chatbot can end [Link]
strategy for
up for instance asking: “Would you like to learn m/?utm
product sales;
more about insurance coverage related to =camp
Designing of
Coronavirus?” aign2-a
advertisement
and
communicatio
n strategies
“Anthem is working to accelerate the availability
of a Coronavirus Assessment on the Sydney Care
mobile app, which members can download at no www.a
cost. The Coronavirus Assessment is designed nthemc
‘‘Manufacturi
based on guidelines from the Centers for Disease orporat
ng’’ the
Control and Prevention (CDC) and National erespon
products;
Institutes of Health (NIH) to help individuals sibility.
Product
Product quickly and safely evaluate their symptoms and com/as
pricing ANTHEM
Development assess their risk of having COVID-19. Inputs sets/im
(actuarial
provided by individual users include symptoms, ages/pd
methods);
recent travel and potential contact with anyone f/pressr
Check legal
with the disease. Based on the results, Anthem elease_
requirements
members will be able to connect directly to a covid_
board-certified doctor via text or secure two-way [Link]
video via the Sydney Care app who can then
recommend care options.”
“[…] we have been working closely with www.p
regulators to allow more products to be sold rudenti
Customer virtually. Our Mainland China business has long [Link]
acquisition, been a heavy user of digital platforms, such as m/news
PRUDENTI
Sales consultation; WeChat, but in other markets many of our /news-
AL PLC
Product sale; products have traditionally required face-to-face releases
After-sales meetings as part of the sales process. That is now /2020/1
changing. In Hong Kong, 25 per cent of our 5-04-
products can now be sold virtually. And in 2020

304
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Indonesia and Malaysia most products can be

sold without the need for the customer and agent
to be in the same room.”
www.u
Application nitedhe
handling; althgro
Risk [Link]
assessment; /newsro
Assessment of om/pos
UNITEDHE “Allowed grace periods for employees and
Underwriting the final ts/2020
ALTH individuals to pay premiums.”
contract -04-15-
details, if uhg-
necessary ask covid-
for more 19-
information respons
[Link]
“For 90 days, Anthem’s affiliated health plans
www.a
will waive any member cost share for telehealth
nthemc
Change of visits, including visits for mental health, for our
orporat
contract data; fully insured employer plans, Individual plans,
erespon
Answering Medicare plans and Medicaid plans, where
Contract sibility.
customer permissible. Cost sharing will be waived for
administration com/as
requests ANTHEM members using Anthem’s telemedicine service,
& customer sets/im
regarding the LiveHealth Online, as well as care received from
service ages/pd
contract or other telehealth providers delivering virtual care.
f/pressr
other Access to LiveHealth Online as well as virtual
elease_
purposes care via text is available to members through the
covid_
Sydney Care app. Self-insured plan sponsors will
[Link]
have the choice to participate in this program.”
www.z
urich.c
“97 year-old customer John needed help because om/en/
his roof was leaking. He was vulnerable due to media/
his age and because his wife was recently moved magazi
into a care home due to dementia. Normally, ne/202
home repairs is a simple process. […] But these 0/coron
are not normal times. John’s claim came at the avirus-
end of March just as the UK entered its stories-
Investigation
ZURICH coronavirus lockdown. helping
Claim of fraud;
INSURANC […] Zurich supplier, Prism Network […] were -a-97-
management Claim
E able to arrange for a ‘drive-by’ survey of the year-
settlement
customer’s home and establish the leak was old-
caused by a damaged chimney stack. They were vulnera
able to fix the external damage and stop the leak ble-
without entering the property or coming into custom
close contact with the vulnerable customer. They er-
plan to return after the lockdown to inspect and during-
repair any potential internal damage.” the-
lockdo
wn
Asset “Investment in Health Bonds Issued by the Asian www.d
DAI-ICHI
allocation; Development Bank (“ADB”) - Supporting the ai-ichi-
LIFE
Asset & risk Asset liability Health Sector in Asia and the Pacific, including [Link].j
INSURANC
management management; responding to Covid-19 - p/englis
E
Analysis and The Dai-ichi Life Insurance Company, Limited h/news
COMPANY
management […] is pleased to announce that it has purchased _releas

305
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of all a total of JPY 10 billion in Health Bonds issued

e/2020/
risks by the Asian Development Bank (“ADB”) the pdf/ind
proceeds of which support ADB’s projects in the
ex_001
health sector, a key theme for the Company.”
.pdf
www.u
nitedhe
althgro
Planning HR [Link]
development; /newsro
Job “Within days, 90% of our non-clinical workforce om/pos
Human UNITEDHE
interviews; was transitioned to a work at home status while ts/2020
Resources ALTH
Job market maintaining all service levels.” -04-15-
advertisement uhg-
; Job training covid-
19-
respons
[Link]

306
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Exemplificative initiatives with different benefits on the society

Organization
Description Source
Benefit Name
"Anthem’s affiliated health plans will continue to
waive copays, coinsurance and deductibles for
the diagnostic test related to COVID-19. In
addition, this will be extended to include waiver
of copays, coinsurance, and deductibles for visits
associated with in-network COVID-19 testing, [Link]
whether the care is received in a physician’s [Link]/as
office, an urgent care center or an emergency sets/images/pdf/pressr
Health ANTHEM department. " elease_covid_19.pdf
"We have automatically stopped charging late
fees and have temporarily paused personal auto
Financial & home coverage cancellations due to non-
and payment from March 23 through June 1, 2020, or
Economic LIBERTY later as directed by your state. We are also [Link].c
wealth MUTUAL extending payment dates if needed." om/covid-19
[Link]/le-
groupe-cnp-
assurances/cnp-
Job CNP "No employee is partially unemployed and 98% assurances-mobilisee-
continuity ASSURANCES of the head office employees are teleworking." contre-le-covid-19
"We are maintaining industry-leading service [Link]
Everyday levels and our wholesalers are helping advisers m/news/news-
life PRUDENTIAL develop new digital solutions so they can interact releases/2020/15-04-
continuity PLC with their clients without meeting face-to-face." 2020

307
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Appendices

Appendix A: List of analysed Insurance companies. Source: Orbis - Bureau Van Dijk; Consolidated financial statements 2018

Name Hedquarters Net Written

Premium
(thousands of
US dollars) -
2018
UNITEDHEALTH United States 163,160,844
AXA France 103,056,922
PING AN China 102,165,332
CHINA LIFE INSURANCE China 92,672,701
KAISER FOUNDATION GROUP OF HEALTH PLANS United States 92,054,688
JAPAN HEALTH INSURANCE ASSOCIATION Japan 90,774,480
ANTHEM United States 85,048,000
ALLIANZ Germany 83,222,077
ASSICURAZIONI GENERALI Italy 72,634,257
THE PEOPLE'S INSURANCE CO. OF CHINA (PICC) China 68,434,994
STATE FARM United States 65,468,992
BERKSHIRE HATHAWAY United States 59,216,000
HUMANA United States 55,139,807
NIPPON LIFE INSURANCE Japan 54,319,124
MUNCHENER RUCKVERSICHERUNGS Germany 53,480,687
NATIONAL MUTUAL INSURANCE FEDERATION OF Japan 50,242,159
AGRICULTURAL COOPERATIVES
LIFE INSURANCE CORP. OF INDIA India 48,746,429
CENTENE United States 48,274,104
DAI-ICHI LIFE HOLDINGS Japan 48,179,014
PRUDENTIAL PLC Great Britain 44,431,000
CHINA PACIFIC INSURANCE CO. China 44,408,581
CVS HEALTH CORP (Aetna) United States 43,956,803
METLIFE United States 43,840,000
PRUDENTIAL United States 42,707,423
ZURICH INSURANCE Switzerland 41,230,000
LIBERTY MUTUAL United States 39,100,000
CREDIT AGRICOLE ASSURANCES France 37,637,314
THE ALLSTATE CORP. United States 37,451,000
CNP ASSURANCES France 35,995,383
HEALTH CARE SERVICE CORP United States 37,586,890

308
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Appendix B: Examples of initiatives according to their the rationale

Rationale Company Name Short Description Fonte Technology Risk

https://
"UnitedHealth Group’s move www.u
to accelerate claim payments nitedhe
to medical and behavioral althgro
care providers applies to [Link]
UnitedHealthcare’s fully /newsro
insured commercial, Medicare om/202
Advantage and Medicaid 0/2020-
businesses. The decision to 04-07-
accelerate claims and uhg-
incentive payments builds on acceler
previously adopted measures ates-
taken by the company to paymen
streamline processes for ts-
health care professionals and support
facilities, as well as to help -
UNITEDHEALT members more easily access provide
Adaptation H the care they need." [Link] \ Health
https://
www.c
a-
assuran
[Link]
m/en/C
hannels
/Covid-
19-
prevent
ion-
measur
es/Cred
it-
Agricol
e-
establis
hes-a-
mutuali
"Crédit Agricole has decided st-
to support professionals and support
help them make it through this -
unprecedented crisis by scheme
implementing an unparalleled -to-
support system for all help-
policyholders who have taken professi
out professional onal-
comprehensive insurance with policyh
operating loss cover. This olders-
mutualist support scheme will cope-
lead to the payment of a sum with-
corresponding to a flat-rate their-
CREDIT estimate of the income lost by operati Pandemic- due
AGRICOLE the economic sector ng- economic
Expansion ASSURANCES concerned during this period." losses \ losses

309
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

“97 year-old customer John

needed help because his roof
was leaking. He was
vulnerable due to his age and
because his wife was recently
moved into a care home due
to dementia. Normally, home
repairs is a simple process. https://
[…] But these are not normal www.z
times. John’s claim came at urich.c
the end of March just as the om/en/
UK entered its coronavirus media/
lockdown. magazi
[…] Zurich supplier, Prism ne/2020
Network […] were able to /corona
arrange for a ‘drive-by’ virus-
survey of the customer’s stories-
home and establish the leak helping
was caused by a damaged -a-97-
chimney stack. They were year-
able to fix the external old-
damage and stop the leak vulnera
without entering the property ble-
or coming into close contact custom
with the vulnerable customer. er-
They plan to return after the during-
lockdown to inspect and the-
ZURICH repair any potential internal lockdo “Drive-by” Home risks
Reaction INSURANCE damage.” wn survey (Leak)
"To combat the impact of
COVID-19, our offices are https://
operating with skeletal staff, [Link]
whenever feasible, within the cindia.i
restrictions imposed by the n/Home
local administration. All other /Covid-
employees are working from 19-
LIFE home and are available on Informa
INSURANCE telephone and electronic tion-to-
CORP. OF means of communication to Policyh
Aggression INDIA attend to all queries" olders Telework Covid-19

310
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

MAXIMIZING CUSTOMER EXPERIENCE USING

DISPLAY RETARGETING
Dr. Denis Hyams-Ssekasi
Senior Lecturer in Business Management
Institute of Management
University of Bolton
E-mail: [Link]-ssekasi@[Link]

Dr. David Bamber

Director of PhD Studies
Institute of Management
University of Bolton
E-mail: [Link]@[Link]

Dr. Roshan Panditharathna

Lecturer - Business
OLC (Europe) Ltd,
726-728 Seven Sisters Road, London N15 5NH
E-mail: roshanp@[Link]

ABSTRACT
The paper explains the effect of display retargeting on customer experience where it improves the
customer satisfaction and repurchase intention. The special feature of this study is that paper
combines the thoughts of both academics and practitioners. The paper is the first to focus on
enhancing customer experience by using display targeting. The main contribution of this paper is
to introduce a model that explains how the different types of display targeting mechanisms improve
the elements of customer experience, customer satisfaction and repurchase intention. The paper
provides managerial implications for marketing managers who must devise display advertising
for generating leads and sales. Future research will need to test the validity of conceptual model
in an online environment.

Keywords: Customer Experience, Display Retargeting, Customer Satisfaction, Repurchase

Intention, Conceptual Development

311
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

INTRODUCTION
The unprecedented growth of digital advertising has transformed the nature of how organizations
advertise offerings, how customers seek and make purchases (Nam and Kannan, 2020). As a result,
organizations are able to cater to customer preferences better than before. On the other hand,
customers build interactions between organizations through touchpoints (Nam and Kannan, 2020)
that improve customer experience. Hence, the paper explains the effect of display retargeting on
customer experience where it enhances customer satisfaction and repurchase intention. Many
customers do online shopping and leave online platforms without making any purchase; effectively
just “window shopping”. In such circumstances, display retargeting is an important tool for
marketers. So display retargeting, that shows advertisements established from the user’s previous
online activity, has become a popular alternative for organizations to bring back their customers
to make a purchase or to provide a lead (Eigenbrod, Janson and Leimeister, 2018; Sahni,
Narayanan and Kalyanam, 2019). Arguably, it is an uncommon form of promoting for a brand to
target customers who are already familiar with the brand (Sahni et al., 2019). Jain, Aagja and
Bagdare (2017) claim that there is a need to develop new tools to enhance customer experience
and Virtual Reality (VR) has been promoted “to create novel shopping experiences that combine
the advantages of e-commerce sites and conventional stores” (Speicher et al., 2018). Moreover,
most of the contribution to customer experience has come from the industry and academic research
that shows “lack of progressing” (Jain et al., 2017). Hence, there is an important shortcoming in
the literature: investigations have not attempted to maximize customer experience and its
dominants consequences using display targeting. Therefore, this paper shows how such a
phenomenon, based on a literature review, leads to a conceptual development, discussion,
conclusion, and future research avenues.

Literature Review and Conceptual Development

Display Retargeting

In 2019, the global digital advertising market was $333.25 billion worth by claiming a 17.6% rise
compared to the previous year (Enberg, 2019). This becomes a milestone of advertisement history
as the digital market segment claims nearly half of global advertisement market for the first time

312
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

in the history (Enberg, 2019). Google which is the top digital ad seller in 2019 has a 31.1%
($103.73 billion) market share on ad spending (Enberg, 2019). Facebook becomes the second
highest ad seller with $67.37 billion followed by Alibaba which claims $29.20 billion (Enberg,
2019). AdRoll is one of the most popular retargeting organizations which serve more than 20,000
businesses globally (AdRoll, 2020). The AdRoll retargeting caters almost every industry which
operates online for customizing and attracting customer offering across platforms and devices
(AdRoll, 2020). Gordon et al. (2020) claim that organizations tend to promote brands by targeting
potential buyers via finely designed commercials in specific locations and times in a variety of
formats. In the sense of digital advertising, display retargeting is inevitable. Rutz and Bucklin
(2012) claim that researching the display advertising has become an integral aspect of both
academics and practitioners. Thus, display retargeting is an innovative
online marketing mechanism that can provide consumer specific advertising content based on
consumers’ browsing behavior that meets consumers’ preferences and interests (Eigenbrod et al.,
2018; Lambrecht and Tucker, 2013). When surfing the Internet, consumers who previously viewed
products on the firm’s website are shown ads with images of those same products (Lambrecht and
Tucker, 2013). Sahni et al. (2019) claim that certain product categories may have a 50% of the rise
of sales within the first two days launching the retargeting. Drèze and Hussherr (2003) explain that
display retargeting focuses on positioning personalized adverts together with the website’s main
content. Morrissey (2013) shows that an average Internet user is targeted more than 1,700 banner
ads per month. However, only effective online adverts grab the attention of customers who buy
through follow up targeting whilst ineffective ads incur sunk cost and customer dissatisfaction
(Wang and Hong, 2019). Moreover, table 1 explains that different types of retargeting strategies
used by industry practitioners.

From the practitioners’ perspective, display retargeting is a cookie-based mechanism that uses
JavaScript codes to track user behavior on the web (Yang, Huang and Tsai, 2015). Web developers
may include an inoffensive code on the website where each time a new visitor comes, the code
puts an anonymous cookie (Yang et al., 2015). Subsequently, the visitor returns to use the Internet,
the cookie provides the browsing history to the retargeting provider to target potential products
(Reddy, 2016). Nevertheless, display retargeting is an effective way to call back customers to

313
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

return to make a purchase, it may increase the violation of privacy concern of the customers who
also become annoyed and frustrate about the organization (Eigenbrod et al., 2018).

Table 1: Display Retargeting Strategy

Display retargeting strategy Application

Sequential retargeting The user goes through a sequence of ad experiences after they leave the
company's website.
Offer change retargeting When customers upset seeing the repetitive ads, it is a good strategy to provide
an offer to convert them into a buyer.
Up sell retargeting Display a fresh offer related to a product which has already been purchased by
the customer.

Cross sell retargeting Similar to up selling retargeting, cross-selling retargeting offer valued added
products for those who customer already owns.

Bulk email retargeting Advertisers may the opportunity to target customers based on the email list
precisely through Google AdWords, Facebook, or Twitter
(Sources: Khadjav, 2017; Reddy, 2016)

Facebook Ads Manager provides advertisers to engage sequential retargeting which shows vivid
adverts passage of the time. Criteo (2010) finds that customized retargeted adverts are four times
more productive than generic retargeted adverts. Offer change retargeting is also effective in many
situations such as Chtourou, Chandon, and Zollinger (2002) claim that promotional based ads
banners boost higher click through rates (CTRs) than those that lack incentive offers. Moreover,
Lambrecht and Tucker (2013) reveal that display retargeting is effective when consumers have
strong preferences that have been stimulated with incentives and offers. In reality, display
retargeting hardly offers discounts: as customers are targeted to purchase original prices (Hecht,
2014).

Drèze and Hussherr (2003) showed that although people actively avoid looking at display ads,
such ads still have a positive effect on brand awareness and advertising recall and indeed the savy
computer user can use specialized software to remove ads. They analyze display ad effects for
users at different stages of the company’s purchase funnel: nonvisitor, visitor, authenticated user,
or converted customer; and find that display advertising positively affects visitation to the firm’s
website for users in most stages of the purchase funnel (Hoban and Bucklin, 2015). Hupfer and

314
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Grey (2005) show that banner ads offer a free sample achieved higher CTRs than banner ads with
information only. Braun and Moe (2013) also find that the effects of creative content in banner ads
can differ, even though their data do not ascribe substantive meaning to these contents.

Customer Experience

Providing a great customer experience is one of the main concerns of today’s organizations (Jain,
Aagja and Bagdare, 2017). In the earlier studies, Holbrook and Hirschman (1982) reveal that
customer experience involves playful leisure activities, sensory pleasures, daydreams, esthetic
enjoyment and emotional responses. They also comment that consumption as an experience to be
seen as involving a steady flow of fantasies, feelings, and fun (Holbrook and Hirschman, 1982).
Pine and Gilmore (1998) identify experiences as inherently personal responses occurring only in
the mind of an individual who has been engaged on an emotional, physical, intellectual or even
spiritual level.

In their generalized theory-based framework, DeKeyser et al. (2015, p. 1) define customer

experience as “the cognitive, emotional, physical, sensorial, and social elements that mark the
customer’s direct or indirect interaction with a [set of] market actor[s]”. Meyer and Schwager
(2007) divide customer experience as direct contact and indirect contact where unplanned or
automated communications are exposed to the customer. Customer experience is measured by
using touchpoints where customers interact with brands (Lemon and Verhoef, 2016; Meyer and
Schwager, 2007). To date, no single study has yet to address the impact of display retargeting tools
for maximizing customer experience and its consequences. Table 2 shows different elements of
customer experience and their meaning.

315
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Customer Experience Elements

Elements Meaning
Cognitive Mental and invisible interactions, referring to imaginary speculation, goal
orientation, prior expectations

Emotional Joy and surprise, delight, satisfaction

Physical Manmade, firm-controllable surroundings within which service experience

occurs, consumer-friendly shopping interfaces, uncluttered screens, and fast
presentations, ambiance (of space)
Sensorial Ease of use, customer meaning-making, background music/sound

Social The customer’s social environment and the opinions and experiences of other
people
(Source: Addis and Holbrook, 2001; Arnold et al., 2005; Fisk et al., 2011; Griffith, 2005; Helkkula, Kelleher and
Pihlström, 2012; Holbrook, 2006; Kleijnen, de Ruyter and Wetzels, 2007; Oliver, Rust and Varki, 1997;
Parasuraman et al., 2016; Rose, Hair and Clark, 2011; Verhoef et al., 2009)

Berry and Carbone (2007) claim that organization needs to create a cohesive, authentic and
sensory-stimulating total customer experience that resonates, pleases and differentiates
organization from the competition to build an emotional connection with customers. The customer
experience in contemporary large enterprises comprises both human and non-human encounters,
with the large firms’ main focus being on managing the interactions between the company’s
employees, customers, and machines (Bolton et al., 2018; Gilboa, Seger-Guttmann, and Mimran,
2019; Lemon and Verhoef, 2016). In practice, “large firms manage customer experience by
employing advanced technologies based on virtual, non-human interaction with customers”
(Gilboa, Seger-Guttmann and Mimran, 2019, p. 152). Thus, it is interesting to discover how
maximizing customer experience responses using display retargeting. For this conceptual paper,
authors adopt the definition of DeKeyser et al. (2015) for building model.

Consequences (Customer Satisfaction and Repurchase Intention)

Customer experience “needs to be considered and managed as a holistic strategic process for
creating customer value, differentiation, customer satisfaction, loyalty and competitive advantage”
(Jain et al., 2017, pp. 642). There could be many outcomes (consequences) of customer experience.
For the conceptual purposes of this paper, two dominants consequences are proposed such as
customer satisfaction and repurchase intention. Literature claim that display retargeting may

316
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

enhance advertisement awareness, brand awareness, purchase intention, and site visits (Dreze and
Hussherr, 2003; Ilfeld and Winer, 2002). Academics and industry experts examine brand
awareness, brand attitudes, and purchase intentions as a function of exposure (as against the effects
of a banner advertising on actual purchase behavior) (Cho, Lee, and Tharp, 2001; Dahlen, 2001;
Dreze and Hussherr, 2003; Gallagher, Foster, and Parsons, 2001). Thus, growing attention has
escalated to develop and test reliable methods to establish and measure the links between online
display advertising and consumer behavior (Manchanda et al., 2006).

In this conceptual study, the consequences of customer experience considered as customer

satisfaction and repurchase intention. Customer satisfaction which can be defined as a positive
emotional reaction after the purchase of a product or service (Mei, Li and Li, 2017) is an important
metric for the future survival of the organizations. In order to improve the customer satisfaction,
organizations may promise to deliver the advertised satisfaction and ensure that their products and
services are of excellent quality (Othman, Hamzah and Hassan, 2020). Heidemann et al. (2013)
claim that customer satisfaction can be enhanced by using a higher degree of service quality.
Thakur (2019) researches the moderating effect of customer engagement experience between
customer satisfaction and loyalty. However, no research so far enables to maximize customer
experience using display retargeting that lead to higher customer satisfaction.

Previous research finds that customer experience may have a positive relationship among loyalty,
word of mouth, and repurchase intentions (Rageh Ismail et al., 2011; Verhoef et al., 2009, Verhoef
et al., 2016). Repurchase intention means whether the customer may seek the service provided by
the same supplier in the future (Wang et al., 2019). This is different from purchase intention where
the customer tends to make the initial purchases that do not recurrent with the time (Wang et al.,
2019). The repurchase intention in an online environment becomes complicated when compared
with bricks and mortar mechanism (Griffis et al., 2012; Wen et al., 2011). Wells et al. (2011) found
that the website influences the perception of the quality and subsequently repurchase intention but
fails to identify from the display retargeting point of view. Therefore, a conceptual model has been
derived based on the existing knowledge as shown in Figure 1.

317
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Conceptual Development
Sahni et al. (2019) claim that display retargeting enhances the chances of attracting web users to
become a potential buyer. Eventhough Sahni et al. (2019) have studied the impression of
retargeting ads into consumer behavior, they fail to identify the customer experience point of view.
Gentile, Spiller and Noci (2007) consider that customer experience is a multidimensional
construct. As an antecedent to customer experience, Edvardsson et al. (2010) introduced the notion
of “value in pre-use” to explain the pre-purchase service experience which provided a vivid image
of the real consumption experience. In this particular research, pre-purchase behavior is largely
considered as the customer experience. Bleier and Eisenbeiss (2015) claim that personalized ad
banners that displays previously browsed products have a better chance to click by the customers.
Moreover, Johnson, Lewis, and Nubbemeyer (2017) reveal that display retargeting has a positive
impact on sales of the organization. Moriguchi, Xiong and Luo (2016) investigate the effects of
retargeting at different stages of the consumer’s buying behaviour. As a result, Lambrecht and
Tucker (2013) find that a generic banner works better when individuals are in their advanced stage
of the purchase funnel. Manchanda et al. (2006) confirm that online advertising influences online
repurchase decisions in positive and negative ways. As a result, the following model has been
derived to find out how display retargeting effects on customer experience where it improves
customer satisfaction and repurchase intention.

“Customer experience is regarded as a holistic interactive process, facilitated through cognitive

and emotional clues, moderated by customer and contextual characteristics, resulting into unique
and pleasurable/un-pleasurable memories” (Jain et al., 2017, pp. 642). Conventional retargeting
tools have been surpassed by digital tools. Therefore, display retargeting acts effectively for getting
customers back to the company’s site to make a purchase or become a lead when they otherwise
would have been bounced traffic (Barnhart, 2020).

318
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1: Conceptual Development

Display Retargeting

Sequential retargeting
Offer change retargeting
Up sell retargeting
Cross sell retargeting
Bulk email retargeting

Customer Experience
Consequences
Cognitive
Customer satisfaction
Emotional
Repurchase intention
Physical
Customer dissatisfaction
Sensorial
Avoidance behaviour
Social
(De Keyser et al., 2015; Rageh Ismail et al., 2011; Reddy, 2016; Rose, Hair and Clark, 2011;
Verhoef et al., 2009)

DISCUSSION

Compared with traditional advertising, digital ads provide much control over targeting metrics
such as personalization, content diversification, real-time updates and higher conversion rates
(Gordon et al., 2020). However, customers may be opened for many digital touchpoints that end
up with frustration, dissatisfaction, lower confidence, rejection and in extreme cases even product
sabotage. Therefore, Broniarczyk, Hoyer and McAlister (1998) find that many customers postpone
the purchase and stop searching further information. Moreover, renown retargeting organizations
such as AdRoll and Marketingsherpa claim that only about 4% online shoppers buy from an online
store while the other 96% are engaging only window shopping (Yeo et al., 2017). Therefore, in
order to motivate window shoppers to covert actual buyers, display retargeting plays one vital (Yeo
et al., 2017).

To date, most studies dealing with aspects relating to customer experience in the context of digital
channels. However, in this paper shows that display retargeting works as a digital technique that
moderates between the customer experience and its consequences. Online adverts generate a
platform for organizations to gain new customers and urge existing customers to make a purchase

319
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

by delivering marketing messages through the Internet (Wang and Hong, 2019). Therefore,
display retargeting has become a vital tool for marketers to target viewers who have not yet been
converted to customers at the initial stage of purchase (Estrada-Jiménez et al., 2017). Therefore, it
supports organizations to introduce dynamic offerings that are based on their previous online
behaviors (Yu, Hudders and Cauberghe, 2017). The insights of this conceptual model should work
“for managers who must make budget allocation decisions across multiple types of devices in
consumers’ online path to purchase while accounting for differences in the product category and
customer experience” (De Haan et al., 2018, p. 2). Moreover, Lambrecht and Tucker (2013)
suggest that in evaluating how best to reach consumers through adverts, managers should be aware
of the multistage nature of consumers’ decision processes and vary advertising content along those
stages.

LIMITATIONS AND FUTURE RESEARCH AVENUES

The paper shows the effect of display retargeting on customer experience where it potentially
improves the customer satisfaction and repurchase intention. However, there are a few limitations
that prevent explaining research gaps are identified. The first limitation is that paper is purely
developed based on the existing knowledge, then, does not contain primary data; just like display
retargeting uses past data and cannot well predict customers new future needs. Though, this paper
considers one of the universally used definitions of customer experience, there are other well-
known definitions were not considered as O’Loughlin, Szmigin and Turnbull (2004) proposed
three levels of customer experience: brand experience, transactional experience and relational
experience. On the other hand, display retargeting which has many variations such search
retargeting, behavioral research targeting, social retargeting does not consider proposing the said
model.

Even though the paper is non-empirical nature, it opens up new research avenues. Since the
construct of customer experience varies with the industry (Jain et al., 2017). the model validity
should be empirically tested in an online environment. However, checking the validity of the
conceptual model, from the organizational (advertisee) perspective may be highly technical and
expensive (Sahni et al., 2019). The model can be tested from the customer perspective. Some web

320
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

users are concerned about recurrent retargeting which they found to be an unpleasant experience
and react against their online behavior being tracked and their privacy being invaded (Chae, Bruno
and Feinberg, 2019). Therefore, in future, research may be able to study how retargeting may also
mitigate against the customer experience and lead to adverse consequences.

REFERENCES

Addis, M. and Holbrook, M. B. (2001) On the conceptual link between mass customisation and
experiential consumption: an explosion of subjectivity. Journal of Consumer Behaviour, 1(1), pp.
50-66.

AdRoll (2020) About us [Online] Available from: [Link]

[Accessed 7th May 2020].

Arnold, M. J., Reynolds, K. E., Ponder, N. and Lueg, J. E. (2005) Customer delight in a retail
context: investigating delightful and terrible shopping experiences. Journal of Business
Research, 58(8), pp.1132-1145.

Barnhart, B (2020) 41 must-have digital marketing tools to help you grow [Online] Available
from: <[Link] [Accessed 15th April
2020].

Berke, A., Fulton, G. and Vaccarello, L. (2014) The retargeting playbook: how to turn web-
window shoppers into customers. John Wiley & Sons.

Berry, L. L. and Carbone, L. P. (2007) Build loyalty through experience management. Quality
progress, 40(9), p.26.

Bleier, A. and Eisenbeiss, M. (2015) The importance of trust for personalized online
advertising. Journal of Retailing, 91(3), pp.390-409.

Bolton, R. N., McColl-Kennedy, J. R., Cheung, L., Gallan, A., Orsingher, C., Witell, L. and
Zaki, M. (2018) Customer experience challenges: bringing together digital, physical and social
realms. Journal of Service Management, 29(5), pp. 776-808.

Braun, M. and Moe, W. W. (2013) Online display advertising: modeling the effects of multiple
creatives and individual impression histories. Marketing Science, 32(5), pp. 753-767.

Broniarczyk, S. M., Hoyer, W. D. and McAlister, L. (1998) Consumers’ perceptions of the

assortment offered in a grocery category: the impact of item reduction. Journal of Marketing
Research, 35(2), pp.166-176.

321
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chae, I., Bruno, H. A. and Feinberg, F. M. (2019) Wearout or weariness? measuring potential
negative consequences of online ad volume and placement on website visits. Journal of Marketing
Research, 56(1), pp. 57-75.

Cho, C. H., Lee, J. G. and Tharp, M. (2001) Different forced-exposure levels to banner
advertisements. Journal of Advertising Research, 41(4), pp.45-56.

Chtourou, M. S., Chandon, J. L. and Zollinger, M. (2002) Effect of price information and
promotion on click-through rates for internet banners. Journal of Euromarketing, 11(2), pp. 23-
40.

Criteo (2010) Targeting & retargeting interview with Criteo. [Online] Available
[Link] [Accessed 7th May 2020].

Dahlen, M. (2001) Banner advertisements through a new lens. Journal of Advertising

Research, 41(4), pp.23-30.

De Haan, E., Kannan, P. K., Verhoef, P. C. and Wiesel, T. (2018) Device switching in online
purchasing: examining the strategic contingencies. Journal of Marketing, 82(5), pp. 1-19.

De Keyser, A., Lemon, K. N., Klaus, P. and Keiningham, T. L. (2015) A framework for
understanding and managing the customer experience. Marketing Science Institute working
paper series, 15(121), pp. 1-48.

Drèze, X. and Hussherr, F. X. (2003) Internet advertising: Is anybody watching?. Journal of

Interactive Marketing, 17(4), pp. 8-23.

Edvardsson, B., Enquist, B. and Johnston, R. (2010). Design dimensions of experience rooms for
service test drives. Managing Service Quality: An International Journal.

Eigenbrod, L., Janson, A. and Leimeister, J. M. (2018) How digital nudges influence consumers-
the role of social and privacy nudges in retargeting. Academy of Management, 1, p. 11298.

Enberg, J. (2019) Global digital ad spending 2019 [Online] Available from:

<[Link] [Accessed 17th April
2020].

Estrada-Jiménez, J., Parra-Arnau, J., Rodríguez-Hoyos, A. and Forné, J. (2017). Online

advertising: analysis of privacy threats and protection approaches. Computer
Communications, 100, pp. 32-51.

Fisk, R. P., Patrício, L., Rosenbaum, M. S. and Massiah, C. (2011) An expanded servicescape
perspective. Journal of Service Management. 22(4), pp. 471-490.

322
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Gentile, C., Spiller, N. and Noci, G. (2007). How to sustain the customer experience: An overview
of experience components that co-create value with the customer. European Management Journal,
25(5), pp. 395-410.

Gilboa, S., Seger-Guttmann, T. and Mimran, O. (2019) The unique role of relationship marketing
in small businesses’ customer experience. Journal of Retailing and Consumer Services, 51, pp.
152-164.

Gordon, B. R., Jerath, K., Katona, Z., Narayanan, S., Shin, J. and Wilbur, K. C., (2020
forthcoming). Inefficiencies in Digital Advertising Markets. Journal of Marketing, pp. 1-19.

Griffis, S. E., Rao, S., Goldsby, T. J. and Niranjan, T. T. (2012) The customer consequences of
returns in online retailing: an empirical analysis. Journal of Operations Management, 30(4), pp.
282-294.

Griffith, D. A. (2005) An examination of the influences of store layout in online

retailing. Journal of Business Research, 58(10), pp.1391-1396.

Hecht, D. (2014) A beginner’s guide to retargeting ads. [Online] [Link] hubspot.

com/marketing/retargeting campaigns beginner guide. [Accessed 8th May 2020].

Heidemann, J., Klier, M., Landherr, A. and Zimmermann, S. (2013) The optimal level of CRM
IT investments. Electronic Markets, 23(1), pp. 73-84.

Helkkula, A., Kelleher, C. and Pihlström, M. (2012) Characterizing value as an experience:

Implications for service researchers and managers, Journal of Service Research, 15 (1), pp. 59-
75.

Hoban, P. R. and Bucklin, R. E. (2015) Effects of Internet display advertising in the purchase
funnel: Model-based insights from a randomized field experiment. Journal of Marketing
Research, 52(3), pp. 375-393.

Holbrook, M. B. (2006) Consumption experience, customer value, and subjective personal

introspection: an illustrative photographic essay. Journal of Business Research, 59(6), pp. 714-
725.

Holbrook, M. B. and Hirschman, E. C. (1982) The experiential aspects of consumption:

consumer fantasies, feelings, and fun. Journal of Consumer Research, 9(2), pp. 132-140.

Hupfer, M. E. and Grey, A. (2005) Getting something for nothing: the impact of a sample offer
and user mode on banner ad response. Journal of Interactive Advertising, 6(1), pp.105-117.

Ilfeld, J. S. and Winer, R. S. (2002) Generating website traffic. Journal of Advertising

Research, 42(5), pp.49-61.

323
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Jain, R., Aagja, J. and Bagdare, S. (2017) Customer experience - a review and research
agenda. Journal of Service Theory and Practice. 27(3), pp. 642-662.

Johnson, G. A., Lewis, R. A. and Nubbemeyer, E. I. (2017) Ghost ads: Improving the economics
of measuring online ad effectiveness. Journal of Marketing Research, 54(6), pp. 867-884.

Khadjavi, R. (2017) How to improve your Facebook advertising funnel with sequential
retargeting [Online] Available from: [Link]
retargeting [Accessed 12th April 2020].

Kleijnen, M., de Ruyter, K. and Wetzels, M. (2007) An assessment of value creation in mobile
service delivery and the moderating role of time consciousness. Journal of Retailing, 83(1), pp.
33-46.

Lambrecht, A. and Tucker, C. (2013) When does retargeting work? Information specificity in
online advertising. Journal of Marketing research, 50(5), pp. 561-576.

Lemon, K. N. and Verhoef, P. C. (2016) Understanding customer experience throughout the

customer journey. Journal of Marketing, 80(6), pp. 69-96.

Manchanda, P., Dubé, J. P., Goh, K. Y. and Chintagunta, P. K. (2006) The effect of banner
advertising on Internet purchasing. Journal of Marketing Research, 43(1), pp. 98-108.

Mei, J., Li, K. and Li, K. (2017) Customer-satisfaction-aware optimal multiserver configuration
for profit maximization in cloud computing. IEEE Transactions on Sustainable Computing, 2(1),
pp. 17-29.

Meyer, C. and Schwager, A. (2007) Understanding customer experience. Harvard Business

Review, 85(2), p. 116.

Moriguchi, T., Xiong, G. and Luo, X. (2016) Retargeting ads in the upper and lower purchase
funnel: online field experiments. Working Paper, Waseda University.

Morrissey, B. (2013) 15 alarming stats about banner ads [Online] Available from:
[Link] [Accessed 6th May 2020].

Nam, H. and Kannan, P. K. (2020) Digital environment in global markets: cross-cultural

implications for evolving customer journeys. Journal of International Marketing, 28(1), pp. 28-
47.

O’Loughlin, D., Szmigin, I. and Turnbull, P. (2004) From relationships to experiences in retail
financial services. International Journal of Bank Marketing. 22(7), pp. 522-539.

Oliver, R. L., Rust, R. T. and Varki, S. (1997) Customer delight: foundations, findings, and
managerial insight. Journal of Retailing, 73(3), p.311.

324
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Othman, A. K., Hamzah, M. I. and Hassan, L. F. A. (2020) Modeling the contingent role of
technological optimism on customer satisfaction with self-service technologies. Journal of
Enterprise Information Management, 33(3), pp. 559-578.

Parasuraman, A. P., Ball, J., Aksoy, L. and Keiningham, T. L. (2016) From practicality to
paradox: examining the customer experience of delight, Working Paper, University of Miami.

Pine, B. J. and Gilmore, J.H., 1998. Welcome to the experience economy. Harvard Business
Review, 76, pp.97-105.

Rageh Ismail, A., Melewar, T. C., Lim, L. and Woodside, A. (2011) Customer experiences with
brands: Literature review and research directions. The Marketing Review, 11(3), pp. 205-225.

Reddy, S. (2020) All you need to know about retargeting campaigns [Online] Available from:
<[Link]
[Accessed 10th April 2020].

Rose, S., Hair, N. and Clark, M. (2011) Online customer experience: a review of the business‐to‐
consumer online purchase context. International Journal of Management Reviews, 13(1), pp. 24-
39.

Rutz, O. J. and Bucklin, R. E. (2012) Does banner advertising affect browsing for brands?
clickstream choice model says yes, for some. Quantitative Marketing and Economics, 10(2), pp.
231-257.

Sahni, N. S., Narayanan, S. and Kalyanam, K. (2019) An experimental investigation of the

effects of retargeted advertising: the role of frequency and timing. Journal of Marketing
Research, 56(3), pp. 401-418.

Speicher, M., Hell, P., Daiber, F., Simeone, A. and Krüger, A. (2018). A Virtual Reality
Shopping Experience using the Apartment Metaphor. Proceedings of the International
Conference on Advanced Visual Interfaces, pp 1-9.

Thakur, R. (2019) The moderating role of customer engagement experiences in customer

satisfaction - loyalty relationship. European Journal of Marketing. 53(7), pp. 1278-1310.

Verhoef, P. C., Lemon, K. N., Parasuraman, A., Roggeveen, A., Tsiros, M. and Schlesinger, L.
A. (2009) Customer experience creation: Determinants, dynamics and management
strategies. Journal of Retailing, 85(1), pp. 31-41.

Wang, H. and Hong, M. (2019) Online ad effectiveness evaluation with a two-stage method
using a Gaussian filter and decision tree approach. Electronic Commerce Research and
Applications, 35, p. 100852.

325
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Wang, Y., Anderson, J., Joo, S. J. and Huscroft, J. R. (2019) The leniency of return policy and
consumers repurchase intention in online retailing. Industrial Management & Data Systems.
120(1), pp. 21-39.

Wells, J. D., Valacich, J. S. and Hess, T. (2011) What signal are you sending? How website
quality influences perceptions of product quality and purchase intentions, MIS Quarterly, 35(2),
pp. 373-396.

Wen, C., Prybutok, V. R., Xu, C. and Chao, R. (2011) An integrated model for customer online
repurchase intention. Journal of Computer Information Systems, 52(1), pp. 14-23.

Yang, K. C., Huang, C. H. and Tsai, C. W. (2015) Applying reinforcement theory to implementing
a retargeting advertising in the electronic commerce website. In: Proceedings of the 17th
International Conference on Electronic Commerce, pp. 1-5.

Yeo, J., Kim, S., Koh, E., Hwang, S. W. and Lipka, N. (2017) Predicting online purchase
conversion for retargeting. In: Proceedings of the Tenth ACM international conference on web
search and data mining, pp. 591-600.

Yu, S., Hudders, L. and Cauberghe, V. (2017) Targeting the luxury consumer. Journal of
Fashion Marketing and Management.

326
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Big Data Analytics in Business: Evolving knowledge, skills and roles of

Accounting and Finance Professionals

Eley Suzana Kasim

Accounting Research Institute, Universiti Teknologi MARA (UiTM)
40450 Shah Alam, Malaysia
Email: ekasim@[Link]

Noryati Md Noor
Faculty of Accountancy, Universiti Teknologi MARA (UiTM)
Cawangan Negeri Sembilan, 70300 Seremban, Malaysia
Email: noryatimdnoor@[Link]

Norlaila Md Zin
Institute of Leadership Development, Universiti Teknologi MARA (UiTM)
71760 Nilai, Malaysia
Email: norlaila249@[Link]

Noor Sufiawati Khairani

Faculty of Accountancy, Universiti Teknologi MARA (UiTM)
Cawangan Johor, 85000 Segamat, Malaysia
Email: noors771@[Link]

Abstract

The accounting and finance professionals in organisations have the potential to contribute
significantly in the digital era by incorporating Big Data Analytics (BDA) into their existing
reporting framework. However, the nature of their contribution in BDA is still largely
unexplored. Moreover, it is still unclear whether these professionals have the required digital
skills to cope with BDA. Given the lack of evidence on how BDA impact on the profession,
this study seeks to examine the knowledge and skills required and the change on the role of
accounting and finance professionals in Big Data era. This study adopts a qualitative approach
by utilizing semi-structured interviews, focus group discussions and document reviews to
address the main research issues. Findings from this study suggest that while accounting and
finance professionals need to keep abreast with new skills such as BDA, their involvement are
more critical in interpreting the results and providing useful insights for the business. With Big
Data, the profession shifts to a more “enhanced” role as advisors at strategic level within the
organisation. Findings from the study will provide a valuable input for higher institutions and
professional bodies in developing their accounting and finance curriculum for quality
education.

327
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Keywords: Technology, Big Data, Big Data Analytics, Accounting, Finance, case study

1. Introduction

The digital era is often characterised by a new era that builds and extends the impact of
digitization in new and unanticipated ways (Davis, 2016). Technologies such as robotics,
artificial intelligence and biotechnology are just some examples of new developments in the
Industrial Revolution 4.0. While these technologies are very promising, the patterns of
consumption, production and employment created by it requires proactive adaptation by
corporations, governments, and individuals (World Economic Forum, 2016). In line with this,
businesses currently need to deal with enormous amounts of complex data, both structured and
unstructured, known as the Big Data. Big Data permits applications of large-scale data mining
and data analytics which can allow businesses to make more informed decisions. The real
challenge of data analytics comes from the timely informed decisions that management make
(Chen, Chiang & Storey, 2012). Since data is now regarded as an essential asset for an
organization’s success, when processed and analysed appropriately these voluminous data
could enable a competitive advantage over rivals (Taylor-Sakyi, 2016).
Despite the ability of modern technologies to process and analyse Big Data, human
intervention is still critical. In particular, the accountants and finance leaders in organisations
have the potential to contribute significantly in the digital era by incorporating Big Data into
the existing reporting framework. A report by PwC (2015) noted that the need for appropriate
technology-related knowledge and skills for accounting and finance professionals in a data-
driven business environment must be addressed. In particular, the digital skills required by
these professionals to handle Big Data have become increasingly important to add new insights
into decision making and business operations (Wang & Wang, 2016; Brands & Holtzblatt,
2015). Since Big Data is a new phenomenon, the ability of accountants to deal with the
complexities surrounding Big Data is still at its infancy (Brands, 2014). Moreover, it is still
unclear which technological knowledge and skills are required of accountants and finance
leaders working with Big Data. Another issue that arises is the question of how BDA impacts
the role of accounting and finance professionals in the digital era. A report by the Chartered
Global Management Accountants (CGMA) in 2014 noted that these professionals should be
able to leverage on BDA to increase operating efficiencies, assess risks and identify advantages
and weaknesses through analysis and be competent to utilize BDA to position themselves as
strategic business partners instead of their more traditional accounting role (CGMA, 2014).

328
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The accounting and finance professionals are therefore increasingly challenged to

demonstrate their relevance in the industry and their ability to evolve and face new challenges
have become critical. Given the call for concern by professional bodies to equip accounting
and finance graduates with appropriate technology-related knowledge and skills in a data-
driven business environment, the lack of evidence on how Big Data Analytics impact on the
profession needs to be addressed. Hence, this study seeks to examine the impact of Big Data
on the accounting and finance profession, both in terms of knowledge and skills required in
interpreting results from BDA and also the change on the role of the accounting and finance
professionals in the digital era.

2. Literature Review
2.1 Big Data and Big Data Analytics Knowledge and Skills
The term Big Data (BD) has been defined by Gartner (2016) as high-volume, high-
velocity and high-variety type of data. This is similar to the description by Davis (2014)
whereby BD is described as consisting of expansive collections of data that are updated quickly
and frequently and that exhibit a huge range of different formats and content. In addition, a
more comprehensive characterisation of BD was proposed as having five “Vs” viz., volume,
velocity, variety, veracity and value (Akter, et al., 2016) as shown in Table 1.

Table 1: The 5 “Vs” characteristics of Big Data

Characteristics Description
Volume large size of data generated in every second
Velocity real-time data access
Variety various sources of data location received
Veracity reliability and accuracy of data
Value the cost-benefit of collecting data

Gartner (2016) suggests that BD benefits organisations since it can reflect the useful
information that are cost effective, facilitates successful decision-making process through
effective transfer of information into knowledge, and with the help of process automation. Big
Data management and analytics involves the process starting from data extraction, data
representation, analysis of data and ends with data interpretation (Gandomi & Haider, 2015).
Connolly (2012) views BD as the interconnection between three key elements namely

329
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

transactions, interactions and observations. Kho (2016) refer to the conversion of messy data
into meaning as Big Data Analytics.
Knowledge and skills on Big Data Analytics (BDA) has become increasingly important
for accounting and finance professionals (McLeod, Bliemel & Jones, 2017). BDA has been
defined as comprehensive way to retrieve, process and analyze the Big Data in order to create
valuable intelligent insights for continual value delivery, performance assessment and forging
competitive advantages (Bumblauskas [Link]., 2017). Palem (2014) succintly described the dual
scope of BDA as a storage platform and as a solution enabler. In other words, not only that
BDA enables reliable storage of voluminous data, it also offers distributed computing using a
large number of networked machines to reduce the total "time-to-solution" (Palem, 2014).
BDA can take several forms viz., descriptive, diagnostic, discovery, predictive and prescriptive
(Joshi & Marthandan, 2018). BDA differs from traditional data analytics in many ways. For
instance, Bumblauskas, Nold, Bumblauskas and Igou, (2016) and Chen [Link]. (2012) suggest
that BDA entails huge data assets that require advanced and unique data storage, management,
analysis, visualization technologies and statistical analysis. As a result, BDA will affect how
businesses use data (McAfee & Brynjolfsson, 2012) which leads to a new era for global
commerce (Shah, Horne, & Capella, 2012). Kaya and Akbulut (2018) further noted that BDA
can be used as a solution provider that handles the gap between what the companies’ perceived
needs and level of comprehension about its scope and the actual solution to handle.

2.2 Role of Accountants in Big Data

Janvrin and Watson (2017) and Capriotti (2014) contend that the traditional role of
accounting and finance as provider of accounting information would remain even though they
have to deal with Big Data. This involves recording, collecting and analyzing accounting
information to internal and external decision makers. However, several scholars noted that the
emergence of Big Data has fundamentally changed how accounting data are understood and
reported (Gamage, 2016; Griffin & Wright, 2015). For instance, Kaya and Akbulut (2018)
suggests that with BD, the nature of accounting and financial reporting do not alter but the
methods of recording, collecting and analyzing accounting information will change. Similarly,
Warrant [Link]. (2015) suggests that Big Data could significantly influence the future of
accounting particularly in the area of financial reporting.
Palem (2014) further noted that in forensic accounting field, with big data techniques,
fraud detection that usually requires hours or even days to complete can be accomplished in
much shorter time. From the management accounting perspective, it is suggested that

330
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

management accountants are required to evolve as technology evolves including in response

to BD (Gray & Alles, 2015). In particular, Gray and Alles (2015) claim that management
accountants need to play a more forward-looking data analytics users to fully evaluate the
potential benefits and consequences of alternative actions and decisions. On the other hand, in
auditing perspective, Big Data has been found to be an essential part of their assurance practices
(Alles & Gray, 2016). These findings suggest that the role of accounting and finance
professionals are expected to evolve as a result of the availability of BD and the use of BDA
for various accounting and finance functions.
BDA revolve around three main characteristics, i.e., the data itself, the analytics applied
to the data, and the presentation of results (Gantz & Reinsel, 2012). While the BD and its
analytics are mainly of concern to data analysts, the presentation of results and its interpretation
in a way that allows the creation of business value remains to be the forte of accounting and
finance professionals. Thus, the role of these professionals need to evolve to cater for the new
way of presenting and interpreting business results to enable more informed decision making.

3. Methodology

This study used a qualitative methodology to address the main research questions. The research
process consisted of four stages: (1) Literature review, identification of research variables and
preparation of interview protocol; (2) Pilot interviews followed by review and revision of
interview protocol; (3) Main study, and (4) Data analysis and report writing

During stage 1 of the research process, an extensive review of the literature was conducted to
identify the current state of knowledge in the field, followed by relevant frameworks and
factors to be considered in the current study. Subsequently, the identification of research
variables as well as the appropriate research methodology and procedures were conducted. An
interview protocol was prepared as a guide for semi-structured interviews conducted in this
study. Stage 2 involved conducting several pilot interviews with academicians particularly in
the accounting and computing/artificial intelligence/Big Data areas. This was done in order to
obtain better understanding of Big Data and the link with accounting as well as to acquire
practical input in determining the feasibility of the study. Several practitioners from the
industry were also interviewed. These practitioners consisted of practicing accountants in audit
firms. At this stage, consistent with Yin (2003), a pilot study to test the applicability of factors

331
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

identified in stage 1 was conducted. Hence, the pilot interview serves as a means of revising
the interview protocol and to achieve external validity of the research constructs.

During stage 3, the main study was conducted whereby the data was collected mainly through
semi-structured interviews and focus group discussions. The objective of the interviews and
focus group discussions was to address the following main research issues:
i. What types of digital skills and knowledge are required by accountants to embrace the
Big Data developments?
ii. How does Big Data Analytics change the role of accounting and finance profession?

Participants of the study included academicians/lecturers in accounting/finance and Computer

Science Faculties, practitioners from accounting and audit firms as well as accounting and
finance personnel in multinational corporations. The qualitative data was recorded in mp3
format and later transcribed. Additionally, review of documentations such as websites,
newspaper clippings, and archival records were also performed. Finally, in stage 4, the data
was transcribed and coded. Using Yin’s (2009) pattern matching logic, the data was analysed
together with more specific analysis as suggested by Bloomberg & Volpe (2008) and Miles &
Huberman (1994). The final stage of the study was concluded with report writing.

4. Findings and discussions

4.1 Accounting and Finance knowledge and Skills for Big Data Era

Findings in this study suggest four key skills that accountants and finance professionals need
to have in handling Big Data. Firstly, they need to posses somewhat basic understanding of Big
Data and Big Data Analytics. This is because they need to know what the different types of
data are, what data might be useful and where and how it can be obtained. Secondly, accounting
and finance professionals are required to have relevant skills to exploit Big Data in order to
add value to the business. A close working relationship with systems developers and data
scientists may help them identify how big data can add value to the business. Thirdly, these
professionals need to be able to interpret results from Big Data Analytics. This is consistent
with predictive analytical skills of Big Data Analytics. In interpreting analytics results, these
professionals need to exercise their judgement in filtering relevant information to avoid
information overload. This is because with enormous amount of data, information overload can

332
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

easily occur which can lead to poor interpretation of business results. Finally, they should also
be able to present the results from Big Data Analytics effectively. One way to achieve this is
by making use of visualization techniques such as digital dashboards.

Findings also highlight that acquiring new knowledge and skill in Big Data analytics is
important for accounting and finance graduates. As such, universities need to equip their
graduates with relevant technical knowledge and skills in Big Data analytics. One of the ways
in which universities and other higher learning institutions can achieve this is by integrating
Big Data Analytics within existing accounting curriculum. By doing this, the universities will
help students in preparing themselves with up to date technology, updated information, and
mould them to be internationally minded future accountant. This is because acquiring Big Data
knowledge and skills would provide some exposure on something ‘big’ in the industry globally
that is Big Data. Hence, this BDA is viewed as a powerful and useful tool for the graduates.

Furthermore, findings indicated that future accountants and finance professionals do not need
to acquire advanced qualification in BDA to enable them to perform their tasks. Nevertheless,
they must posses somewhat basic knowledge and skills in the field. This is consistent with the
view of an interviewee who commented: “Maybe they do not need in depth BDA knowledge
but at least they know how can they make use of that data, know what is the point of having
the data and what are the things that they can do with the data.”

These professionals also need to be adaptive towards the environment. In other words, when
they use their skills and detailed understanding of the business and the environment in which
it operates, these professionals are able to ask challenging questions in providing sound advice.
Furthermore, accounting and finance professionals are increasingly needed in many industries
and they need to be an “all rounder” rather than strictly having accounting and finance
qualification. Along similar lines, universities have also recognised the need for accounting
and finance graduates to have more non-technical skills which would indirectly be useful for
their profession. One of the interviewees from a local university mentioned that:

“Nowadays companies are not just looking for accounting graduates with excellent
academic qualification. Rather, employers look for graduates who have good extra
curricular activities...they look for people who have climbed mountains or joined
kayak expeditions...basically, they want resilient people (to join the profession).”

333
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The above findings highlight the need for accounting and finance graduates to exhibit strong
soft skills besides acquiring technical skills to secure employment.

4.2 Evolving role of Accounting and Finance professionals

Findings from this study indicated that there have been significant changes to the role of
accounting and finance professionals over recent years. In particular, with the emergence of
Big Data, their role has shifted from mere input provider to systems developer to a deeper
involvement in assisting business decision makings. More specifically with the current
development, there is higher expectations on what these professionals can do in harnessing
voluminous data and convert them into business insights. However, this does not mean that
their traditional role in financial reporting has changed. Instead, their role needed to expand
considering the multiples sources of data they need to handle. As noted by an interviewee:
“...they still need to do the same as before but now because the sources of data is many and
varied, they need to discover, visualize and display patterns”.
This findings is consistent with the suggestions by Chartered Institute of Management
Accountants (CIMA) that amongst others, accounting professionals need to be able to identify
which data point i.e., the factors that are useful in driving the business forward. They also need
to have a clear sense, able to embrace new forms of data and creative ways while staying
comfortable with uncertainty including Big Data.

Additional findings of the study indicated that accounting and finance professionals play
enhanced function in decision making whereby they need to play good advisory role to the top
management. This is because these professionals are increasingly recognised to hold positions
more on the strategic level, and less of the transaction processing and reporting role previously
noted. Another important finding of this study is the role of accounting and finance
professionals in shaping the insights developed using Big Data Analytics. Figure 1 shows the
process of converting Big Data into business insights of which contribution from accounting
and finance professionals are essential.

334
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Input Process Output

Big Data Analytics Insights

Basic role Enhanced Role

Figure 1: Expected involvement of Accounting and Finance Professionals in BDA

As shown in Figure 1, contribution of accounting and finance professionals are evident in the
process and output stage. Within the process stage, the professionals play important role in
providing inputs to system developers for Big Data analytics on what and how the data needs
to be analysed as well as what type of output is required. They are tasked with giving relevant
advice to the developers about the rule of the business, how the data needs to be stored, and
presented. Although systems developer are mostly staff with computer science background and
trained to conduct business analytics, advice on how the output needs to be presented must be
provided by the accounting and finance professionals. Since the professionals are involved at
more strategic levels, they are the most appropriate individuals in the organisation to determine
what data they want to support management function such as planning, decision making and
control. However, as the analysed data is converted at the output stage, more contribution from
the accounting and finance professionals are required. The focus is now more on “decision
usefulness” rather than “faithful representation” of business transactions. This is because, at
the output stage, relevant interpretations of the Big Data analytics are essential to build
meaningful insights of the business. As commented by an interviewee: “With Big Data, the
role of accountants and finance personnel are enhanced and magnified.”

However, with voluminous nature of Big Data the professionals face bigger challenge to filter
the information since data can also be of unstructured nature. Findings also suggest that in Big
Data environment, the flow from data to information may also be accompanied by the reverse
flow of information back to data. Apart from that, nowadays accounting and finance

335
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

professionals are expected to be involved in predictive analytics. This means that from data
mining process and subsequent analytics, the accounting and finance professionals need to go
beyond mere interpretation of results to providing useful prediction models for the business.

5. Conclusions
This study aimed at examining the impact of Big Data on the accounting and finance
profession, both in terms of knowledge and skills and also the change on the role of the
accounting and finance professionals in the digital era. Findings suggest that in response to the
new BD era, they must posses somewhat basic knowledge and skills in BDA in order to
effectively generate business insights that could add value to the business. In addition, findings
of this study highlighted that Big Data has an impact on the future role of accounting and
finance professionals. More specifically, the evolving nature of the role of accounting and
finance professionals from mere input provider to system developers to strategic advisory role
has to be emphasised. As such, this study recommends that Big Data Analytics should be
embedded in existing accounting and finance courses at institutions of higher learning to
prepare twenty-first-century accounting professionals with skills related to Big Data analytics.
This study will provide a valuable input for higher institutions and professional bodies in
developing their accounting and finance curriculum for quality education that meets the
challenges of the new digital era.
Acknowledgement

The authors would like to acknowledge the financial support received from the Malaysian
Ministry of Higher Education under the Fundamental Research Grant Scheme (FRGS) (Ref:
FRGS/1/2018/SSO 1/UITM/02/22) as well as the support from Universiti Teknologi MARA,
Malaysia.

References

Akter, S., Wamba, S. F., Gunasekaran, A., Dubey, R. & Childe, S. J. (2016). How to improve
firm performance using big data analytics capability and business strategy alignment?
International Journal of Production Economics, 182, 113-131.

Alles, M. & Gray, G. L. (2016). Incorporating big data in audits: Identifying inhibitors and a
research agenda to address those inhibitors. International Journal of Accounting
Information Systems, 22. DOI: 10.1016/[Link].2016.07.004

Brands, K. (2014). Big Data and Busine ss Intelligence for Management Accountants. Strategic
Finance. 64-65.

336
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Brands, K. & Holtzblatt, M. (2015). Business Analytics: Transforming the Role of

Management Accountants. Management Accounting Quarterly, 16 (3), 1-12.

Bloomberg, L. D. & Volpe, M. (2008). Completing your qualitative dissertation: A road map
from beginning to end. Thousand Oaks, CA: Sage.

Bumblauskas, D., Nold, H., Bumblauskas, P. & Igou, A. (2017). Big data analytics:
transforming data to action. Business Process Management Journal, 23 (3), 703-720.
[Link]

Capriotti, R. J. (2014). Big data: Bringing big changes to accounting. CPA Journal, 85 (2), 36.

Chartered Global Management Accountant (2014). CGMA BRIEFING: Readying business for
the big data revolution. 1-12.

Chen, H., Chiang, R.H.L. & Storey, V.C. (2012). Business intelligence and analytics: From
Big Data to big impact. MIS Quarterly, 36 (4), 1165-1188.

Connolly, S., (2012), “7 Key Drivers for the Big Data Market”. [Link]
key-drivers-for-the-big-data-market/

Data driven: What students need to succeed in a rapidly changing business world. (2015). PwC.
Retrieved from Janvrin, D.J. & Watson, M.W. (2017). ‘‘Big Data”: A new twist to
accounting. Journal of Accounting Education, 38, 3-8.
[Link]
[Link]

Davis, N. (2016). What is the fourth industrial revolution? Retrieved 2020, from World
Economic Forum: [Link]
industrial-revolution/
Davis CK (2014) Beyond data and analysis. Commun ACM, 57(6), 39–41.

Gamage, P. (2016). Big Data: are accounting educators ready? Accounting and Management
Information Systems, 15 (3),588-604.

Gandomi, A. & Haider, M. (2015). Beyond the hype: Big data concepts, methods, and
analytics. International Journal of Information Management, 35 (2), 137-144.

Gantz, J. & Reinsel, D. (2012). The digital universe in 2020: Big data, bigger digital shadows,
and biggest growth in the Far East. IDC iView IDC Anal Future 2012:1–16.

Gartner (2012) “The Importance of ‘Big Data’: a Definition” available online at

[Link] (accessed on 20 August 2019)

Gray, G. L. & Alles, M. (2015). Data Fracking Strategy: Why Management Accountants Need
It. Management Accounting Quarterly, 16 (3), 22-33.

Griffin, P. A., & Wright, A. M. (2015). Commentaries on Big Data’s importance for accounting
and auditing. Accounting Horizons, 29(2), 377–379.

337
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Igou, A., & Coe, M. (2016). Vistabeans coffee shop data analytics teaching case. Journal of
Accounting Education, 36, 75–86. [Link]

Janvrin, D. J., & Watson, M. W. (2017). “Big Data”: A new twist to accounting. Journal of
Accounting Education, 38, 3-8. [Link]

Joshi, P. L. & Marthandan, G. (2018). The Hype of Big Data Analytics and Auditors. Emerging
Markets Journal, 8(2), 1-5.

Kaya, I. & Akbulut, D.H. (2018). Big data analytics in financial reporting and accounting.
PressAcademia Procedia (PAP), 7, 256-259.

Kho, N. D. (2016). The state of big data. EContent 39(1).

McAfee, A. & Brynjolfsson, E. (201 2). Big Data: The Management Revolution. Harvard
Business Review, October issue. [Link]
revolution.

McLeod, A. J., Bliemel, M. & Jones, N. Examining the adoption of big data and analytics
curriculum. Business Process Management Journal, 23 (3), 506-517.

Miles, M. B., & Huberman, A. M. (1994). Qualitative data analysis: An expanded sourcebook
(2nd ed.). Thousand Oaks, CA: SAGE Publications.

World Economic Forum (2016). The Future of Jobs: Employment, Skills and Workforce
Strategy for the Fourth Industrial Revolution. 1-167.

Palem, G. (2014). Formulating an Executive Strategy for Big Data Analytics. Technology
Innovation Management Review, 4(3):25-34.

Shah, S., Horne, A. & Capellá, J. (2012). Good Data Won’t Guarantee Good Decisions.
Harvard Business Review, 23-25.

Taylor-Sakyi, K. (2016). Big data: Understanding big data. arXiv preprint arXiv:1601.04602.

Technology is disrupting accountancy profession, said MIA. (2017, September 7). News Straits
Times. Retrieved from [Link]
disrupting-accountancy-profession-said-mia

The future of accountancy in Malaysia – An interview with the MIA. (2016, December 12).
Retrieved from [Link]
Malaysia

Wang, Y. & Wang, Z. (2016). Integrating Data Mining into Managerial Accounting System:
Challenges and Opportunities. Chinese Business Review, 15 (1), 33-41. doi:
10.17265/1537-1506/2016.01.004

Yin, R. K. (2003). Case Study Research (3rd ed.). London, England: Sage Publications.

338
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Consequences of technological development on the manifestation of the

rebound effect phenomenon
Giani Gradinaru, The Bucharest University of Economic Studies, Institute of
National Economy – Romanian Academy, Romania
Iulia Neagoe, The Bucharest University of Economic Studies, Romania

Abstract
We are now part of a world surrounded by innovation. From the first spark of fire, people
tended to develop, learn and create a better life. Over time, there have been several industrial
revolutions that have changed the way mankind understood comfort and well-being. If in the
past, simple artificial light was a luxury, today we can say that we have reached the
economic and social level where dreams are not a barrier. No one could have thought in the
past that in less than a moment, all the information would be a click away. At present, it is
not enough to have the information, but it is crucial to understand it and use it for the good of
all. It is important for everyone to know the advantages and disadvantages of technology, so
that the benefits do not backfire on each of us. Currently, we are facing a rapid evolution of
technology and a significant depletion of natural resources, which forces us to take action in
this regard.
The paper aims to present the consequences of using technology in inappropriate ways and to
provide an overview of rational consumption. The motivation comes from understanding the
phenomenon of rebound effect and from the desire to highlight the potential of technology
when used carefully.
Keywords: rebound effect, industry 4.0, improving efficiency

1. Introduction

The first industrial revolution arose when the power of water and steam was used to
mechanize manufacturing. The emergence of electricity, followed by conveyor belts and
mass production recommended the second revolution, and the third industrial revolution was
dominated by automation and the introduction of information technology in production
processes. When it comes to Industry 4.0, concepts such as the smart factory or “smart city”
appear, assuming that the modular ecosystem of physical-cybernetic systems to monitor
physical processes, to create a virtual copy of the situations encountered. and make a
decentralized decision (following a new situation).
By explaining the analyzed phenomenon and by bringing to light the actions at
international level to stop it, we want to rationalize individuals and encourage their
information when it comes to measures taken in support of the environment.
A large number of natural resource protection initiatives are supported by the
authorities, while rational consumption in an age of consumerism is a step forward.
Romania is one of the countries where, although there are developed cities and
counties, there are estimates for the next decade that by 2030 can lose up to 62% of jobs,
because there is the possibility of adapting the economy and society to the changes caused of
the fourth industrial revolution, also called industrialization 4.0.

2. The historical, social, economic and technological moment of the emergence of

Industry 4.0.

339
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The first meeting between Mankind and Industry 4.0 took place in Hanover,
Germany, where, through a government initiative, the adoption of ITC-type technologies was
prioritized. Traditional manufacturing industries are increasingly using Big Data, artificial
intelligence or IoT (Internet of Things) technologies. Despite the fact that this program
started at the local level, Germany has quickly become a model to follow and a leader in the
use of new technologies and industrial automation when it comes to Europe. A concrete
example regarding the evolution and level of robotization in Germany is the data obtained at
the beginning of 2018, when there were already more than 300 generic robots allocated for
every 10,000 employees. Among them, co-robots have been identified, which interact with
humans in defined physical spaces, the autonomous ones as well as those that require
minimal human intervention. At that time, Germany's main opponents were South Korea,
which registered 631 robots for every 10,000 employees, and Singapore with 488 robots. At
the beginning of 2018, the global average was 74 robots for every 10,000 employees. If we
refer to Romania, it was much below average, counting only 15 robots, which belonged
mostly to foreign companies that carry out their production here. (International Federation of
Robotics, 2018)
Most of the merits of Industry 4.0 belong to interconnected systems, less visible than
robots of various sizes. We are currently witnessing how a factory can become a unitary
entity, controlled through digital interfaces, where supply and logistics are automatically
controlled and in which traditional engineers are no longer relevant, but IT specialists.
The transition between the last two industrial periods has been rapid and facilitated
by technological developments, but at the same time much more expensive. The moment
when it was wanted to robotize most of the production lines coincided with an economically
favorable period. In recent years, the prices of robots used in industries have fallen by 30-
40%, Chinese labor market prices have risen by 150%, and Germany has reportedly had the
opportunity to bring in some of China's production lines.
Making a comparison between Romania and China, we can see that here too there
have been increases in the minimum wage, forced by government decisions, without a
program to encourage industrial automation and digitization, and the operating cost of a co-
robot is approximates the cost of an employee when it comes to hourly pay. This is one of the
biggest fears of employees today. Undoubtedly, the possible imbalances will be overcome by
redistributing jobs, as demonstrated by the latest industrial revolutions.

3. The importance of data for the rapid evolution of industry 4.0

For the rapid evolution of Industry 4.0, data collection and analysis play an extremely
important role. Through modernization, we can generate and interpret a very large volume of
data that plays an essential role both for building the future and for understanding the past.
The condition of current machines and people can be assessed and behavioral typologies can
be intuited.
Data that can be analyzed from several points of view and with the help of several
scenarios will play an important role in managerial decisions and in building strategies.
Articial intelligence and machine learning algorithms will make their place in every
economic branch, within each visionary company with a desire to develop. The
recommendation for a successful future comes down to training staff when it comes to using
new technologies and understanding data. There are already several operating systems
dedicated to factories that collect and interpret all information, both from the environment
and from inside mechanical and electronic equipment, in order to create a complete digital
experience. In the near future, systems like Mindsphere will enjoy the success that Oracle and
SAP enjoy. (Iceberg, 2018)

340
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Every customer wants a quality product at a convenient price and to be obtained in the
shortest possible time, but it is complicated to get such a product. Industry 4.0 aims to solve
all the mentioned issues, as it is about precision, speed and dexterity.
When we talk about benefits, we mention improving efficiency, increasing
productivity, gaining flexibility and agility in terms of production, reducing costs and a better
customer experience. These could be final arguments for investors and traders in the market,
even if there are disadvantages both for the company and among the employees. Automation
is an advantage in the case of standardized products, while there are also traders who want to
provide products that are as creative and original as possible. The number of employees will
decrease considerably, which encourages the specialization of people who practiced basic
activities in order to avoid the mass increase of unemployment. Legal regulations will also be
amended so that new automated activities do not exceed pollution rules. The very high initial
costs cannot be neglected either, because in addition to automation, in addition to mechanical
robots, the appropriate systems that are part of a range of advanced technology must be
implemented, while also requiring regular maintenance and specialization courses for those
who use (Greensoft, 2019).
With the advent of technology that facilitated the installation of Industry 4.0, there have
been several cases in which the rebound effect has manifested itself.

4. Rebound effect and Industry 4.0

When we refer to industry 4.0, we also think of revolutionary methods by which to

protect the environment and to make efficient the consumption from various energy sources.
Most of us believe that when we use energy-efficient systems, we slow down or reduce
pollution and the effects of global warming. Since the middle of the 19th century, it has been
proven that when we have the opportunity to consume less for the processes we normally
perform, most of us consume more than before.
In the book by the British economist Wiliam Stanley Jevons, The Coal Question (The
Coal Question. An Inquiry Concerning the Progress of the Nation and the Probable
Exhaustion of Our Coal Mines) mentioned the observation of coal consumption, through
which it was once created. A way to extract energy from fossil fuels more efficiently, the
economy has increased and so has the demand for coal.
After it was enunciated, the paradox took the name of Javons. Any of us can be
confused, but fuel efficiency is nothing more than speeding up consumption, no matter what
it is.
In the Javons paradox we talk about the use of energy services, which change as we
evolve technologically. Theoretically, when we make consumption more efficient, its use
should decrease, in the theoretical situation by which we will use the same energy consumers,
in the same quantities. If the theoretical model were followed, then there would be many
benefits in environmental protection.
The rebound effect is the event that occurs when the theory of economic efficiency is
applied in real life. The lower costs of energy services and the increase in energy efficiency
put their mark on the economic behavior of individuals and society. This effect represents the
way in which investments for energy efficiency manage to save the use of the resource for
various events, and consumers choose to practice the same type of event more often or
replace it with similar, higher quality and more costly activities (Herring & Roy, 2007).
In the most aggressive forms of the rebound effect, it can be perceived as a
boomerang effect, when energy savings are much lower than subsequent consumption.

5. Manifestations of the rebound effect

341
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The direct effect

The direct effect is manifested when there is a decrease in the price of energy in the
market, and consumers are tempted to consume more energy without increasing their usual
costs (Gillingham, 2013).
A concrete case is found in the case of hybrid cars, where the engine is designed to
reduce fossil fuel consumption, but owners are attracted to the possibility of using the car for
a longer period and distance at the same price. On the contrary, they often end up consuming
more fuel than before, due to the advantageous price.
In 1996, a study was conducted at Yale University that estimated the hours of work
required for a person to enjoy "artificial" light (from the ordinary fire of cavemen, to
electricity for a light bulb / LED) . Thus, from 40-50 hours of rough labor, to less than half a
second, in most cases, light has become an accessible and in many cases an indispensable
resource, and darkness a rarity (Nordhaus, Bresnahan, & Gordon, 1996).
Refrigeration is a model of direct rebound, because in its early days, it offered very
little storage space and low efficiency. During the evolution of technology, refrigeration
spaces began to be more efficient, more accessible and higher quality. Currently, refrigerators
are among the most common electronics and owners do not care about recycling them. This
simple ordinary device, present in the life of each of us, causes immense damage to the
environment. From the production of refrigerators through several polluting processing
operations to their storage and use, which involve the use of energy from various sources and
the long-term storage of food, which we allow ourselves to store and replace. Food disposal
involves large investments in agriculture (fuel for agricultural machinery, herbicides,
processes for processing basic food, etc.).
A study conducted in the US in 2009 showed how more than a quarter of drinking
water was used to produce food that ends up in landfills (Owen, 2010).

Indirect effect
The indirect effect occurs when the savings in a given area of consumption are
invested in other appliances or activities that consume fuel and energy. When the expenses
for the maintenance and use of the devices decrease, it appears the possibility to purchase
other devices or experiences (holidays, various recreational activities).
The effect of direct or indirect rebound refers to individual behavior, while there are
also effects on society as a whole (Gillingham, 2013).

The effect created by the decrease in the macroeconomic price

The rebound effect created by the decrease in the macroeconomic price is one of the
situations in which there are changes in society as a whole. Lower petrol prices, for example,
due to lower demand in one region, may lead to higher consumption in another region
(Gillingham, 2013).

Macroeconomic growth
Macroeconomic growth is another factor that can set in motion the rebound effect. We
take into account a growing economic efficiency, which creates an environment conducive to
rapid economic development at least in the economic areas and sectors of interest. In this case
we have as an example the development of aeronautical equipment and parts, which are
becoming lighter and more efficient, which increase the production and use of aviation
systems and equipment (Gillingham, 2013).
Making a summary of the four types of rebound effect that are found in the economy,
we can see that they are related to both the economic environment and the technological

342
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

environment. The strongest effects are felt in developing countries, where total satisfaction is
not yet reached, while in developed countries there is a relative balance and very few unmet
needs.
Several econometric studies have been carried out so far, which classify the sum of all
manifestations of the rebound effect contributing in proportion of 20% to 60% to changes in
the economy (Barker, Dagoumas, & Rubin, 2009).

6. Conclusions

Economic growth is a desirable phenomenon among the political leaders of each state,
but it brings with it the consequences of the setback effect. At the same time, with the
application of environmental measures, in order to reduce greenhouse gas emissions,
innovative systems are being developed, less polluting, but with greater success in the
market.
Current climate policies contain several main ideas designed to stop the emission of
carbon dioxide and methane in the first phase.
The strategies are divided into two approaches. On one hand, there is talk of
increasing the volume of renewable energy and reducing the use and production of fossil
fuels, and on the other hand, there is a decrease in energy consumption in general.
The energy consumption equation can be expressed according to two factors when it
comes to the second strategy, it follows the size of the economy, expressed in GDP and the
energy intensity required for each unit of GDP.
In this equation, if the growth rate of the economy is higher than the rate at which the
energy intensity decreases, the final energy consumption will be higher than before the
increase of energy efficiency. In the last economic century, the evolution of technology and
the decrease of energy intensity have led to a continuously sustained economic growth and to
a higher energy consumption than ever before.
In order to reduce total energy consumption, energy intensity should decrease at a rate
higher than the rate of economic growth. This result could be achieved if we reduce global
energy intensity or if we reduce global economic growth.

References
Barker, T., Dagoumas, A., & Rubin, J. (2009). The macroeconomic rebound effect and the
world economy, Energy Efficiency.
Gillingham, K. (2013). The rebound effect is overplayed, Nature.
Greensoft. (2019, iulie). INDUSTRIA 4.0, Nu este o revoluție, este o evoluție! Preluat de pe
[Link]
Herring, H., & Roy, R. (2007). Technological innovation, energy efficient design and the
rebound effect, Technovation;.
Iceberg. (2018). Industria 4.0. Preluat de pe [Link]: [Link]
4-0/
International Federation of Robotics. (2018). International Federation of Robotics. Preluat de
pe International Federation of Robotics: [Link]
IFR-Press_Release_Robot_density_EN.pdf
Nordhaus, W. D., Bresnahan, T. F., & Gordon, R. J. (1996). The Economics of New Goods.
University of Chicago Press.
Owen, D. (2010). The Efficiency Dilemma – If our machines use less energy, will we just use
them more. The New Yorke.

343
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

INCOME DISTRIBUTION, UNEMPLOYMENT AND CRIME RELATIONSHIP:

TURKEY CASE

Assoc. Prof. Haluk YERGIN

Van Yüzüncü Yıl University, Faculty of Economics and Administrative
Sciences, Department of Economics, halukyergin@[Link], orcid id:0000-0002-8168-9115.

Asst. Prof. Mustafa TORUSDAĞ

Van Yüzüncü Yıl University, Faculty of Economics and Administrative
Sciences, Department of Economics, mustafatorusdag@[Link], orcid id: 0000-0002-8839-
0562.

Abstract
The increase in crime rates in developed and developing countries causes attention to
be turned into this field. Whether the notion of crime is related to economic variables
has been discussed for many years. Economic structure, poverty, unemployment, wage,
inequality in income distribution and education level are among the determinant factors
of crime. The notion of crime, explained by socio-economic and cultural factors, is one
of the most important problems in social, economic and political terms and it has the
effect of disrupting social ties and order. Distorted income distribution of a country
leads to an increase in unemployment and has a detrimental effect on social and social
order. The effect of changes in economic indicators and unemployment on the crime
rate in the society is the subject of the study. In this study for the period 1990-2019, the
relationship between income distribution, unemployment and crime analyzed with
Bayer-Hanck (2012) cointegration and Hacker-Hatemi (2006) causality analysis tests
are examined for Turkey. As a result of the analysis findings, it was found that there is
a one-way causality relationship from income distribution to crime rates, and from
crime rates to the unemployment variable. It was also found that there is a bidirectional
causality relationship between income distribution and unemployment.
Keywords: Crime, Inequality in Income Distribution, Unemployment, Bayern-Hanck (2012)
Cointegration Test, Hacker-Hatemi (2006) Causality Test

344
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

INTRODUCTION

The change in the income distribution of individuals affects the social events as well as
the development level of the countries. When the factors determining inequality in
income distribution are examined, it is seen that the stability of the country's economy,
the distribution of wealth in the country, the demographic structure of the country and
the phenomenon of immigration have an impact. In addition, increasing population
rates, inflation and devaluation among the factors that cause uneven income distribution
in a country also increase unemployment. Tax policies implemented in the country also
increase unemployment and cause the income distribution to deteriorate (Yiğit and
Teker, 2018: 54, 55).

The pioneering work on crime economy belongs to Becker (1968) and Ehrlich (1973).
Both Becker (1968) and Ehrlich (1973) expressed the assumption that labor market
conditions were effective on crime and made inferences accordingly. According to the
theory, in the labor market, if the opportunities are scarce, people can turn to illegal
activities that are considered a substitute for this. From the same perspective, it can be
said that the unemployed person's marginal benefit from the legal activity is falling and
he has more free time to engage in illegal activities. Another factor that leads to illegal
activity is seen as income. The income (wages) that the person receives in the labor
market and the income he / she receives from illegal activities are evaluated by
participating in the job and the risk of being caught and punished by the individual and
a choice is made accordingly. If the individual's legal income is illegal and is more than
the cost it will incur, he will not turn to criminal activity.

As a result, both Becker (1968) and Ehrlich (1973) stated in their study that there was
a positive relationship between unemployment and crime rate, while they stated that
there was a negative interaction between wages and crime rate (Ata, 2009: 126).

Apart from these studies, many researches on crime economy have been carried out
until today. However, there is no consensus on how crime is affected by economic
variables. (Pazarlıoğlu and Turgutlu, 2007: 64; Mustard, 2010: 4).

Although a common vision for increasing crime rate that the phenomenon of
immigration aspect of this relationship is not working with a lot of empirical methods
to investigate to Turkey, and aims to contribute to the literature.

345
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Literature review
Hale and Sabbagh (1991) found that in a joint study on England, there is a positive
relationship between unemployment and crime rate. In this study, they achieved this
result by using time series analysis of England's annual data from 1949-1987. Similarly,
Papps and Winkelman (2002) achieved similar results in a joint study on New Zealand.

Witt, Clarke and Fielding (1999) aimed to reveal the economic factors that determine
crime with panel data method by using the data of 11 different regions between the
years 1986-1996 in England. As a result of their work, they concluded that the existence
of economic conditions such as unemployment and low wages increased the crime.

Burdett, Lagos and Wright (2003), in a joint study they conducted, found that there was
an interaction between these variables and criminal activity in the context of elements
that lost individuals, such as unemployment and wages. According to this study, high
unemployment increases crime while high wages reduce crime.

Cömertler ve Kar (2007); Turkey's 81 provinces, using data of 2000, the question of
socio-economic factors that determine what crimes have tried to seek answers with a
horizontal cross-sectional method. As a result of this collaboration, they have found
that there is a positive relationship between unemployment and crime rate.

Doyle, Ahmed and Horn (1999) examined the effects of wages and unemployment on
crime using the panel data method of 1984-1993 data of the United States. According
to their findings, both wages and unemployment have an impact on the crime
phenomenon. But the effect of wage on crime is higher than unemployment.

Grogger (1998), concluded that a 20% drop in wages led to a 20% increase in crime
rate. Likewise, Gould, Weinberg and Mustard (2002) found in a joint study that a
change in the drop in wages increased the crime rate.

In addition to these studies, Edmark (2005), Arvanites and Defina (2006), Lin (2008),
Narayan and Smyth (2004), Corman and Mocan (2005), Raphael and Winter-Ebmer
(2001), Baharom and Habibullah (2008 ), Huang, Laing and Wang (2004), in their

346
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

study, have revealed that labor market indicators such as unemployment and wages
have an impact on the crime phenomenon.

Using the panal data analysis method, Machin and Meghir (2004) analyzed England's
1975-1996 period and found that there was no statistically significant relationship
between unemployment and the crime phenomenon. However, the authors concluded
in the same study that there was a negative and statistically significant relationship
between wages and crime.

Cantor and Land (1985) concluded in a study on the United States that there is a
negative relationship between unemployment and extreme. In their study, they reached
this result by using time series analysis by using the annual data of the USA between
1946-1982.

Poutvaara and Priks (2007), Cantor and Land (1985) achieved similar results with their
findings. As a result of a joint study in İmrohoroğlu, Merlo and Rupert (2001), they did
not find a significant relationship between the unemployment rate and the crime rate.

Tang (2009) in this paper attempts to examine the linkages among inflation,
unemployment and crime rates in Malaysia. The sample period covered annual data
from 1970 to 2006. The estimated cointegrating vector revealed that inflation and
unemployment are positively related to crime rate. In this study investigates the
linkages among inflation, unemployment and crime rates in Malaysia through
cointegration and causality analyses. The Johansen’s test reveals that the
macroeconomic variables, inflation and unemployment were coalescing with crime rate
to achieve their steady-state equilibrium in the long run, although deviations may occur
in the short run. In this study, the normalised coefficients for inflation and
unemployment rate are positively related to crime rate in Malaysia over the sample
period of 1970 to 2006.

Gillani, Rehman, etc. (2009) in this study is designed to investigate the relationship
between crime and various economic indicators such as unemployment, poverty and
inflation in Pakistan. The study covers the period for 1975-2007. The stationary
properties of the time series data are examined by using Augmented Dickey-Fuller

347
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

(ADF) test. Johansen Maximum Likelihood Cointegration and Granger Causality tests
are applied to find out long-run relationship along with causality among the variables.
When, unemployment rate increases the opportunities for earning income decreases
which instigate the individuals to commit crime. The results of Granger causality
through Toda-Yamamoto procedure.

Tang, Lean (2007) in hhis paper employs the modified Wald (MWALD) causality test
to re-examine the relationship between crime and its determinants (inflation and
unemployment) in the United States from 1960 to 2005. Bounds test approach is
employed to investigate the existence of a long-run relationship. The empirical
evidence suggests that inflation and crime rates are cointegrated with a positive
relationship. Moreover, the causal link is from inflation and unemployment to crime.
The objective of this paper is to re-examine the impact of inflation rate on crime in the
United States over the period of 1960-2005. In order to ascertain a reliable relationship
between inflation (by controlling unemployment rate) and crime rates, we employ the
newly developed econometrics techniques, i.e. bounds testing approach for
cointegration test and TYDL causality test. Using annual data from 1960 to 2005, all
the results consistently indicate that the crime rate is cointegrated with inflation and
unemployment rates.
Although there are many empirical studies on the interaction between crime and
unemployment and wage as the factors that determine it, the findings obtained from
these studies cannot reveal clear results regarding the direction and degree of relations
between crime, unemployment and wage.

3. Data and Method

In the study of income distribution for the period 1990-2019, unemployment and crime
relationship using annual data for Turkey-Khatami Hacker (2006) examined the
causality analysis. Income distribution and unemployment variables Data from the
World Bank database, crime rates are variable from TUIK (Turkey Statistical Institute)
was obtained. Econometric analyzes were analyzed using Eviews 10.0, Stata 14.0 and
Gauss 10.0 econometric programs.

348
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.1. Analysis Findings

Table 1: Unit Root Test for Turkey

Variables ADF Unit Root Test Result

I(0) I(1)
Unemployment 0.6698 0.0047**
Income Distibution 0.2107 0.0001**
Crime 0.9368 0.0000**
Note: ***, **, * indicate that the variables are stationary at 1%, 5% and 10% significance levels,
respectively.

In Table 1, stationarities of variables are examined. Units, income and crime variables
variables were found to be stationary at I (1) level. As it is known, variables should be
stationary at I (1) level in order to investigate for cointegration relationship.

Table 1: Türkiye için Bayern-Hanck (2012) Cointegration Test Results

Model 1: Unemployment = f(Income, Crime)

Fisher Type Test statistics, Bayer Hanck Test
EG-J: 5.5320 10% kritik değer: 7.486
EG-J-Ba-Bo: 17.403 10% kritik değer: 33.694
Model 2: Crime = f(Income, Unemployment)
Fisher Type Test istatistiği, Bayer Hanck Test
EG-J: 11.239 10% kritik değer: 7.486
EG-J-Ba-Bo: 42.90848 10% kritik değer: 33.694
Model 3: Income = f(Crime, Unemployment)
Fisher Type Test istatistiği, Bayer Hanck Test
EG-J: 12.0670 10% kritik değer: 7.486
EG-J-Ba-Bo: 41.438 10% kritik değer: 33.694

Whether there is a cointegration relationship between time series Engle and Granger
(1987), Johansen (1991), Boswijk (1994), Banerjee et al. (1998) combining the
probability (p) values of the tests with the combination of Bayer and Hanck (2012) with
Fisher (1932) formulas. This test provides an important advantage by allowing
researchers who make econometric analysis to decide without coincidence in case of
being unstable between cointegration tests (Kızılkaya et al., 2017). Fisher (1932)
formulas are given in equations 1, 2 and 3 (Kızılkaya, 2018: 64):

349
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2
𝑋̃𝑖 = -2∑İ∈𝔩 ln⁡(𝑃𝑖) (1)

EG-JOH= -2[ln(𝑃𝐸𝐺) + ln⁡(𝑃𝐽𝑂𝐻)] (2)

EG-JOH-BO-BDM = -2 [ln(𝑃𝐸𝐺) + ln(𝑃𝐽𝑂𝐻) + ln(𝑃𝐵𝑂) + ln⁡(𝑃𝐵𝐷𝑀)] (3)

EG, Engle and Granger (1987), JOH, Johansen (1988), BO, Boswijk (1994), BDM,
Banerjee et al. (1998) define probability values of cointegration tests. The aspect that
makes Bayer and Hanck (2012) cointegration test more unique than other tests provides
the opportunity to obtain the results of the cointegration tests as a whole. The calculated
Fisher test statistics in Bayer and Hanck (2012) refuse the basic hypothesis, which
indicates that there is no cointegration relationship if the test statistic is greater than the
critical value. Therefore, the test results are interpreted as having a cointegration
relationship (Kızılkaya, 2018: 64).

In Table 2, where the results of Bayern-Hanck cointegration test, which is a

combination of Engle-Granger Johansen, Banerjee and Boswijk (EG-J-Ba-Bo) tests of
three different models, in which unemployment, crime and income variables are taken
as dependent variables, respectively, are presented in crime and income. In models 2
and 3, where the variables are taken as dependent variables, the basic hypothesis that
there is no cointegration relationship between the variables is rejected because Bayern-
Hanck cointegration test statistics are greater than 10% critical value. Therefore, it can
be said that there is a cointegration relationship for models 2 and 3 in the long run.

Hacker-Hatemi (2006) causality test is created by the resampling bootstrap simulation

developed by Efron (1979) because the residues are not normally distributed and the
Toda-Yamamoto test statistics do not fit the⁡𝜒 2 distribution when ARCH effect occurs.
With the bootstrap simulation, more reliable table critical values can be achieved.
Hatemi-J (2003) established the Hatemi-J (HJC) information criterion based on the
average of SIC and Hannan-Quinn (HQ) information criteria and expressed in equation
4. In equation 4, defining the variance-covariance matrix of the error terms of the VAR
model estimated based on the length of the delay, n the number of equations in the VAR
model, the number of T observations (Pata, 2018: 104). ⁡

350
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

𝟐 𝒍𝒏𝑻+𝟐𝒏𝟐 𝐥𝐧⁡(𝒍𝒏𝑻)
̂ | + j (𝒏
𝐇𝐉𝐂 = 𝐥𝐧|𝛀 ), j=0.......,k (4)
𝟐𝑻

Table 3: Hacker- Hatemi-J (2006) Causality Analysis for Turkey

Causality Direction w-stat (MWald) Critical Value Değerler

stat. %1 (***) %5 (**) %10 (*)
Crime⁡⁡⁡⁡→ Income 5.998 16.638 9.875 7.583
Income →⁡⁡⁡ Crime 11.905** 14.818 9.189 7.047
Crime → 40.689** 0.000 0.000 0.000
Unemployment
Unemployment ⁡→ 0.218 10.941 4.593 2.931
Crime
Income ⁡⁡⁡⁡→ 301.965** 17.959 4.510 2.601
Unemployment
Unemployment ⁡→ 0.006* 0.650 0.017 0.002
Income
Note: Bootstrap critical values are obtained in 10,000 cycles. The appropriate delay length was
determined according to AIC (Akaike Information Criterion). ***, **, * indicate that the variables are
stationary at 1%, 5% and 10% significance levels, respectively.

In Table 3, since the w-stat test statistic value is greater than the critical value, the
⁡𝐻0⁡ hypothesis - 𝐻0 : Granger is not the reason - has been rejected. Therefore, it has been
found that there is a one-way causality relationship from income distribution to crime
rates, from crime rates to unemployment variable at 5% significance level. It was also
found that there is a bidirectional causality relationship between income distribution
and unemployment.

Result

Unemployment is a phenomenon with economic social and psychological dimensions.

Being in the work environment plays an important role in self-realization, which is one
of the human needs. For someone who has been unemployed for a long time, this
deficiency may be tried to be provided in illegal ways since the will to self-actualization
will be incomplete. Providing justice in personal income distribution is one of the
macroeconomic objectives, such as keeping unemployment stable at reasonable levels.

351
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Income distribution injustice has political as well as economic social and psychological
consequences.

Income distribution justice is an important phenomenon in terms of equality of

opportunity and stability such as prosperity. Increasing injustice in income distribution
is an indicator of poverty in one aspect and an important factor for social peace.
ensuring a fair distribution of the revenue distribution is important in Turkey. Because
the increase in injustice in income distribution causes economic, political and socio-
cultural problems. Therefore, the issue should be handled from a wider perspective.
Increasing investment, employment and production comes to the fore.

In this study, the relationship between income distribution, crime rates and
unemployment was examined by using the data of 1990-2019. According to the
empirical results obtained from Hacker-Hatemi (2006) causality analysis, it was found
that there is a one-way causal relationship from income distribution to crime rates,
crime rates to unemployment variable. . It was also concluded that there is a
bidirectional causality relationship between income distribution and unemployment.

References
Arvanites T. M. and Defina R. H. (2006), “Business Cycles and Street Crime”,
Criminology, 44(1), 139-164.

Ata A. Y. (2009), Opportunities and Motivations of Corruption in the Frame of

Institutional Economics: A Study on EU Countries, Ünal Aysal Thesis Evaluation
Competition Series, Economic Research Foundation Publications, Istanbul.

Baharom, A., H. and Habibullah, M., S. (2008), “Is Crime Cointegrated with Income
and Unemployment?: A Panel Data Analysis on Selected European Countries”, MPRA
Paper, No.11927.

Becker, G., S. (1968), “Crime and Punishment: An Economic Approach”, Journal of

Political Economy, 76, 169-217.

Burdett K., Lagos R. and Wright R., (2003), “Crime, Inequality and Unemployment”,
American Economic Review, 93(5), 1764-1777.

352
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Cantor D. and Land K. C. (1985) “The Unemployment and Crime Rates in the Post-
World War II United States: A Theoretical and Empirical Analysis”, American
Sociological Review, 50(3), 317-332.

Corman H. and Mocan N. (2005), “Carrots, Sticks and Broken Windows”, Journal of
Law and Economics, 48(1), 235-266.

Cömertler, N. and Kar, M., (2007) "Socioeconomic Determinants of Crime Rate in

Turkey: Cross Section Analysis", Ankara University Faculty of Political Science
Journal, 62 (2), 1-17.

Doyle J. M., Ahmed E. and Horn R. N. (1999), “The Effects of Labor Markets and
Income Inequality on Crime: Evidence From Panel Data”, Southern Economic Journal,
65, 717-738.

Ehrlich, I. (1973), “Paticipation in Illegitimate Activities: A Theoretical and Empirical

Investigation”, The Journal of Political Economy, 81(3), 521-565.

Gillani Syed Yasır Mahmood; Hafeez Ur Rehman; Abid Rasheedgill,

“Unemployment, Poverty, Inflation and Crime Nexus: Cointegration and Causality
Analysis Of Pakistan”, Pakistan Economic and Social Review Volume 47, No. 1
(Summer 2009), pp. 79-98.

Gould E. D., Weinberg B. A. and Mustard D. (2002), “Crime Rates and Local Labor
Opportunities in the United States: 1979-1995”, Review of Economic and Statistics,
84(1), 45-61.

Grogger J. (1998), “Market Wages and Youth Crime”, Journal of Labor Economics,
16(4), 756-791.

Hale, C. and Sabbagh, D. (1991), “Testing The Relationship Between Unemployment

and Crime: A Methodological Comment and Empirical Analysis Using Time Series
Data From England and Wales”, Journal of Research in Crime & Delinquency, 28(4),
400-417.

353
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Huang C., Laing D. and Wang P. (2004), “Crime and Poverty: A Searchtheoretic
Approach”, International Economic Review, 45(3), 909-938.

Imrohoroglu A. Merlo A. and Rupert P. (2001), “What Accounts for the Decline in
Crime?”, Federal Reserve Bank of Cleveland Working Papers, No. 0008.

Kızılkaya, Oktay; Kızılkaya, Fatma; Dağ, Mehmet; Demez, Selim, "Current Account
Deficit Sustainability: In Turkey Case Cointegration Analysis", 1st International Social
Sciences and Education Research Symposium, Antalya, 2017.

Kızılkaya, O. (2018), “Energy Consumption and Growth Relationship in Turkey:

Cointegration and Causality Analysis”, International Journal of Economic and
Administrative Studies, 59-72.

Kuştepeli Y, Halaç U, (2004), “Analysis and Improvement of Income Distribution in

Turkey “, Dokuz Eylül University Social Sciences Institute Journal, 6 (4), 143-160.

Lin M.J. (2008), “Does Unemployment Increase Crime? Evidence From The U.S. Data
1974-2000”, Journal of Human Resources, 43(2), 413-436.
Machin S. and Meghir C. (2004), “Crime and Economic Incentives”, Journal of Human
Resources, 39(4), 958-979.
Mustard, D. B. (2010), “How Do Labor Markets Affect Crime? New Evidence on an
Old Puzzle”, IZA Discussion Paper, No. 4856.

Narayan P. K. and Smyth R. (2004), “Crime Rates, Male Youth Unemployment and
Real Income in Australia: Evidence from Granger Causality Tests”, Applied
Economics, 36(18), 2079-2095.

Papps, K. and Winkelmann (2002), “Unemployment and Crime: New Evidence For an
Old Question”, New Zealand Economic Papers, 34(1), 53-72.

Pata, U., K. (2018), "Inflation in Turkey, Savings and Symmetric and Asymmetric
Causality Analysis of the Relationship Between Economic Growth", Journal of Finance
January-June 174: 92-111.

354
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Pazarlıoğlu, M., V. and Turgutlu, T. (2007), "Income, Unemployment and Crime: A

Review on Turkey", Finance, Political & Economic Reviews, 44 (513), 63-70.

Raphael S. and Winter-Ebmer R. (2001), “Identifying the Effect of Unemployment on

Crime”, Journal of Law and Economics, 44(1), 259-283.

Tang Chor Foon (2009), “The Linkages among Inflation, Unemployment and Crime
Rates in Malaysia”, Int. Journal of Economics and Management 3(1): 50 – 61.

Tang Chor Foon, Lean Hooi Hooi, “Will Inflation Increase Crime Rate? New Evidence
from Bounds and Modified Wald Tests”, Global Crime Volume 8 Number 4
(November 2007).

Witt, R., Clarke, A. and Fielding A., (1999), “Crime and Economic Activity: A Panel
Data Approach”, British Journal of Criminology, 39(3), 391-400.

Yiğit, A., G. and T., Teker (2018), “The Effect of Income and Income Distribution on
Crime Rates”, II. International Multidisciplinary Academic Studies Symposium, Full
Text Book, 53-67.

355
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Sharing Economy in Practice: Intrinsic Motivation towards the Intention to

Participate with Empirical Evidence from Vietnam

Thu Hang Hoang* , Phuong Thy Le , Nhu Quynh Vo , Thi Thanh Mai Doan , Le Thao Nhi
Phan , Quy Bao Tran Bui
School of International Business & Marketing, University of Economics Hochiminhcity

Abstract:
A sharing economy model appears to be the best approach in a post-pandemic world with
an economy already being on a fragile, unsustainable path with regularly diminishing assets,
excessively populated urban areas, and increasing costs. Otherwise called “shareconomy”,
“collaborative consumption”, “collaborative economy”, or “peer economy”, it is a market model
of peer-to-peer (P2P) trade, conventionally exhilarated by an online platform. Although much of
previous empirical research in the shared economy deals with the tradition of business-to-
customers (B2C), there is limited empirical work which studies how players in these shared
economies integrate into their own practices and interactions different aspects of conflicting
market exchanges and pro-social norms that thrusts for more scholarly attention. Drawn from
both self-determination theory (SDT) and the theory of reasoned action (TRA) model, this paper
explores and verifies four intrinsic motives to explain people’s intention to participate in sharing
economy acts.
Using a survey sample of 355 young adults in Ho Chi Minh City, Vietnam, the results from
multiple regression analysis support all hypotheses, indicating that the intention to acquire
sharing economy services is positively affected by one’s level of environmental awareness, social
responsibility, degree of socialization, and hedonism factors. The relevance of the research is also
discussed, include a contribution towards a better comprehension of the fundamental antecedents
influencing citizens’ participation in sharing economy acts. Secondly, it extends the literature on
the inner driving forces for the intention to participate in sharing economy practices from the
perspective of young adults (Gen Z) in an emerging economy such as Vietnam. Finally, this
research provides more in-depth insights and suggests practical implications to enhance the
efficiency of promotional marketing messages by CC platforms (for instance, Airbnb, Uber/Grab,
or XtayPro) and public agencies who wish to promote such practices.

Keywords: sharing economy, collaborative consumption, sustainability, hedonism, socialization,

intrinsic motivation

JEL classification: D16 (Collaborative Consumption); O35 (Social Innovation); Q56

(Sustainability); M31 (Marketing)

356
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1. Introduction
Nowadays, people increasingly turn to sharing assets instead of private ownership, as
resource scarcity is becoming more severe in parallel with the booming population. With the
advancement of digital technologies, a new business model, named sharing economy, has risen
and gained its popularity worldwide. Grasping with the global trend, Vietnam has been excessively
promoting collaborative consumption applications, represented by the growing services in Home-
sharing activities (Airbnb, Luxstay), Car-sharing (Grab, Be, FastGo), Book-sharing (GAT),
Goods-sharing (DOBODY), Electronic-repair services (Rada); Financial services ([Link],
[Link], [Link]), or Travel-sharing ([Link]). Most recent studies on sharing economy
investigate 'collaborative consumption' (CC) as a pillar of this modern business model. The
concept of CC can be commonly understood as a form of sharing economy based mainly on
customer-to-customer trading and predicted that not only can advance underutilized assets but also
connect people through digital communication (Ernst & Young, 2016). Although there is countless
international research on collaborative consumption practices, little has been found in the context
of developing countries, let alone Vietnam, which studies how players in these shared economies
integrate into their own practices and interactions different aspects of conflicting market exchanges
and pro-social norms that thrust for more scholarly attention; thus, this research serves a great
purpose in terms of ideology and practicability.
Much of previous empirical research in the shared economy deals with the tradition of
business-to-customers (B2C), while this research focuses primarily on collaborative sharing
practices that thrust for more academic attention (Kumar et al., 2018). Skimming through multiple
practice theory and market studies (Araujo et al., 2010; Geiger et al., 2014), it is noted that the
performance of sharing practices is not dependent solely on the internal performance of the
platform provider (e.g., the usability of the website as well as technical functionality), but also and
much more on the interactions of peer companies, who are external resources outside the direct
control of the platform provider (Benoit et al., 2017; Kumar et al., 2018). It is evident that the field
of understanding the variation in the collaborative consumer practices by participant procedures,
understanding, and engagements remains under-developed (Schau et al., 2009; Hamari et al.,
2016). Moreover, to date, there are limited numbers of research efforts to understand the
phenomenon of sharing economy practices in developing countries; notably, there have been no
attempts to do so in the context of Vietnam – an emerging market. Hopefully, the research will

357
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

shed light on the mindsets of Vietnamese young adults' cognitive process when it comes to
participation in sharing economy acts.
In the attempt to provide a solid theoretical basis for examining the intention to acquire a
CC service, this paper draws on two theoretical underpinnings from (1) the Theory of Reasoned
Action (TRA) by Fishbein & Ajzen (1975), and (2) the Self-determination theory (SDT) by Deci
and Ryan (1985). This study proposes integrating the four intrinsic motives, including one's level
of environmental awareness, social responsibility, degree of socialization, and hedonism, as
discussed previously with the TRA/SDT to explain and predict customer intention to participate
in sharing economy acts. The main focus of this study is that Vietnamese young adults (aged 18-
35-year-old) account for nearly 32% of the world population and approximately 35% of the
population of Vietnam (Kantar, 2018). This generation is the dominant group in the innovative
global workforce and the primary consumer target, and therefore, it will represent a truthful portrait
of how the sharing economy phenomenon works in a developing country such as Vietnam.
The research aims to accomplish the following objectives:
Firstly, to understand the young adults' participation regarding the collaborative
consumption model in Vietnam.
Secondly, to identify the intrinsic motives affecting the intention to join in collaborative
consumption in Vietnamese young adults.
Thirdly, to determine the level of importance of each intrinsic motive and clarify which
factor would have the most influence on young adults' intention to participate.
Finally, to propose managerial recommendations for enterprises to increase the efficiency
of their platforms.
This paper proceeds in the following structure: Section 2 introduces the theoretical
underpinnings and hypothesis development; Section 3 details the methodology and process of data
analysis; Section 4 discusses hypothesis testing results and findings; Section 5 provides
implications and concludes with this paper's limitations and potential topics for future research.
2. Literature Review
2.1. Background theories
[Link]-determination theory (SDT)
Self-determination theory (SDT), developed by Richard Ryan and Edward Deci in 2000,
refers to a framework of human motivation for growing and integrating through three vital needs:

358
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

competence, relatedness, and autonomy. SDT emphasizes that self-determination occurs when
these vital needs are satisfied. Besides, this theory introduces three different types of motivation
based on autonomous level.
The lowest level of autonomous feeling is amotivation, the state of lacking the intention to
act. Amotivation is considered as a nonself-determination state of a person. The next level is
extrinsic motivation. According to Ryan and Deci (2000), extrinsic motivation presents that the
performance of an action in order to achieve an outcome is not relevant to the action. The highest
level is intrinsic motivation, which indicates a close relation to the high satisfaction level of the
needs for autonomy, relatedness, and competence. Intrinsic motivation is associated with
performing behavior by visual interests and proceeding to action, not by an irrelevant result.
Because the research focuses on intrinsic motives behind the intention of acquiring
collaborative consumption of Vietnamese young adults, the self-determination theory is studied in
order to help the authors have insights into intrinsic motivation and how it is explained.

Figure 2.1: Self-determination theory (SDT) (Ryan and Deci, 2000)

(Source: Self-determination theory and the facilitation of intrinsic motivation, social

development, and well-being. American Psychologist, 2000)
2.2.2. Theory of Reasoned Action (TRA)
TRA model was first developed in 1975 by Fishbein and Ajzen for sociological and
psychological research. In this model, any human behaviour is predicted and explained through
three main cognitive components including Attitudes (unfavourable or favourable level of person’s

359
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

feeling towards a behaviour), Social norms (social influence), and Intentions (individual’s decision
do or don’t do a behaviour) (Taherdoost, 2018). According to TRA, if humans examine the
suggested conduct as positive (attitude) and if they supposthers need them to carry out the conduct
(subjective norm), this outcomes in a better intention (motivation) and they may be much more
likely to perform the behavior. TRA is mentioned in the research for the purpose of framing a
conceptual model for the study objective.
Figure 2.2: Theory of Reasoned Action (TRA) (Ajzen, 1975)

(Source: Belief, attitude, intention and behavior: An introduction to theory and research.
Reading, MA: Addison Wesley, 1975)

2.2. Sharing economy & Collaborative consumption

Belk (2014) insists that ‘sharing’ is as old as humankind. Traditional sharing, defined as
practices by which individuals use, occupy, or enjoy something with others (Schor and
Fitzmaurice, 2014), is one of the primate instincts which mankind is endowed with since their
forefathers (Botsman, 2010). Anthropologically and academically, it is accentuated as a non-
market-mediated process occurring between two parties for the aim of ‘trade up’ and more likely
to take place within family, kin, and friends (Gell, 1986; Benoit, 2017), and does not involve the
transfer of ownership (Belk, 2010).
According to several previous academic researches, ‘collaborative consumption’ is
synthesized and defined as a product-service system in conjunction with sharing economy
(Botsman, 2010) (1) refers to peer-to-peer commerce and sharing activities, including renting,
lending, and swapping of goods and services, solutions, space or money (Hamari et al., 2015;

360
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Möhlmann, 2015) (2) in short-term practices (Ertz et al., 2017), which is (3) facilitated by an
intermediary, as a digital platform, (4) for a fee or other compensation (Belk, 2014).
Researches about collaborative consumption is being made popularly because the recently
incremental concerns about climate change, over-consumption, and planned obsolescence have
fostered a seismic leap from humongous hangover of emptiness and waste to a more sustainable
system (Botsman, 2010); thus, dwellers are craving for an alternative solution for social
embeddedness by localness and communal consumption (Botsman and Rogers, 2010) has fortified
the value of collaborative consumption. Recently, ‘collaborative consumption’ is viewed as an
emerging consumer culture, yet it is not just consumption but an activity where both the
contribution and use of resources are intertwined through peer-to-peer networks (Hamari et al.,
2015). Another previous scholarly definition relates the collaborative consumption to ‘the set of
resource circulation systems’ which consumers are enabled to both temporarily or permanently
obtain and provide valuable resources or services is mentioned by Ertz et al. (2017).
However, the definition is not as simple but is continuously contemporized. Subsequently,
collaborative consumption is broadly defined as a peer-to-peer- based activity of obtaining, giving,
sharing, or trading goods and services, anchored in digital networks (Möhlmann, 2015; Hamari et
al., 2015). Together with Guyader (2018), these authors mutually emphasize the importance of
having a ‘digital network’ or ‘online platform’ as a 'mediator’ in defining collaboration
consumption.
2.3. Hypothesis development
After thoughtful deliberation, the researchers narrow down to mainly focus on four well-
proved intrinsic motivations, namely Environment Awareness, Socialization, Social
Responsibility, and Hedonism. These factors are reflected as the antecedents of satisfaction and
behavioral intention in CC. Besides, one new factor is discovered and added as a determinant
affecting the CC participating motivation. Indeed, empirical work has supported the fundamental
understanding that these factors psychologically influence customer retention. Therefore, they are
as well conceptualized in this study as intrinsic incentives of behavioral intention to use CC
services in the future and will be discussed further in this section. The table below will briefly
summarize the definition of mentioned factors:

361
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2.1: Definition of key concepts

Concepts Definition References
Concern about the environmental, social, and economic
Environment Luchs et al.
consequences of consumption in order to meet the needs of
Awareness (2011)
both current and future generations
Social A sense of responsibility (duty or obligation) to help others Wasko and Faraj
Responsibility within the collective on the basis of shared membership (2005)
The state or process of taking pleasure in meeting new Amirkiaee
Socialization
people, talking to and being with others (2018)
The increased arousal, heightened involvement, perceived Babin et al.
Hedonism
freedom, fantasy fulfillment, and escapism (1994)
Amirkiaee and
An indication of a person’s readiness to participate in
Intention Evangelopoulos
ridesharing
(2018)
Source: Authors' compilation
2.3.1. Environment Awareness
Ecological sustainability is illustrated as an intrinsic motivation as alternative forms of
green, ethical or sustainable consumption become increasingly desirable. (Hamari et al., 2015).
Sharing solutions are generally considered to have a positive environmental impact (Barnes, 2017;
Mohlmann, 2015). Rather than buying new, re-circulation of goods involves a lower ecological
impact including cacbon and eco-footprints footprint (Schor, 2014). The participants’
understanding and perception of environmental benefits will be reflected by their former behaviors
(Barnes, 2017). Participants are reported to be more environmentally driven and resonate with
Böcker and Meelen (2017), who observe the group of intrinsically motivated sharing idealists.
That the intense environmental pressure also turns environmentally friendly strategy to be a
potential offer to users (Guyader, 2018; Herbert and Collin-Lachaud, 2017; Belk, 2014). As a
result, the researchers propose the following hypothesis:
H1: Environment Awareness (EV) has a positive influence on the intention to acquire CC.
2.3.2 Social Responsibility
Referring to prior studies, the Social responsibility factor is reported as an underlied
motivation affecting sharing economy participation. Herbert and Collin-Lachaud (2017) revealed
social commitment as an ambivalent motivation underpinning such CC practices. As in their
studies, Amirkiaee and Evangelopoulos (2018) also highlighted that social commitment fueled CC
practices. In reality, community members might want to help others in order to enhance

362
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

belongingness, which is related to commitment to the community, for the simple reason that they
are part of the same group and follow a collective goal. The researchers, thus, hypothesize the
following:
H2: Social Responsibility (RE) has a positive influence on the intention to acquire CC.
2.3.3 Socialization
Since social networks and collaboration fuel CC, direct peer-to-peer interactions and the
sharing of personal experiences allow participants to create and establish social connections with
others. Long or short, one is spending time with others in a small, shared space. Therefore, the
researchers consider Socialization as a factor for intention to participate in CC. This view is
supported by several studies. Increasing social connections and building social networks was found
as a common motivation among users as it generates a high level of socialisation (Herbert and
Collin-Lachaud, 2017; Schor, 2014). The aspiration to be part of a group or community is a
fundamental stimulating role in acquiring CC activity (Amirkiaee and Evangelopoulos, 2018;
Böcker and Meelen, 2017; Möhlmann, 2015). Several other studies also investigate socialization
as a determinant of intrinsic motivation for participation in group activities and collaborative
consumption. Participating in CC platforms is an opportunity to make new friends and to develop
meaningful connections (Botsman and Rogers, 2010), which not only help interact with nearby
strangers, but also allow individuals and communities to meet physically by jointly engaging in a
consumption activity (Zhang, 2019).Therefore, The researches introduce a related hypothesis as
follows:
H3: Socialization (SC) has a positive influence on the intention to acquire CC.
2.3.4 Hedonism
According to Benoit et al. (2017), CC and access-based consumption may provide
customers with hedonic value. Ertz et al. (2017) named hedonic aspects as an intrinsic
encouragement for consumers to participate in these practices. Hwang (2017) and Ertz et al. (2017)
confirmed that consumers could be motivated by hedonic value that impacts CC acquisition.
Enjoyment, as a hedonic value, has been regarded as an essential factor also in other sharing-
related activities, such as information system use (Zhang, 2019). A study on the continued use of
social networking services established that enjoyment is a primary factor, followed by the number
of peers and usefulness. Thus, it is hypothesized that:
H4: Hedonism (HE) has a positive influence on the intention to acquire CC.

363
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3. Methodology and Data

3.1. Research methodology
3.1.1. Type of study
Exploratory research is conducted through qualitative methods in order to discover factors
that impact the acquisition of CC products. There are indeed many articles and research papers
proposing different conceptual models; however, very few of them are related to developing
countries, Vietnam as in this case. Therefore, this study aims to test the existing theories and also
explore new aspects to this emerging market.
3.1.2. Research tool
A questionnaire is developed based on previous research materials and reports in the field
of collaborative consumption to obtain consistent survey results. As the surveyors are well-
exposed to social networks, the questionnaire can be easily spread to the targeted population, in
this case, young adults, through online utilities.
3.1.3. Scale development
In the questionnaire, a filtered question was used at first to group respondents who have
used CC and who have not. The next part was designed to collect accessible basic information of
each respondent. The nominal and ordinal scale questions were used in this part.
The main part of the questionnaire which was focused on the intrinsic motives behind the
decision to acquire CC of Vietnamese young adults employed psychometric measurement. The
researchers measured each variable with at least four items which were all on a 5-point Likert scale
including: (1) Strongly disagree; (2) Disagree; (3) Neutral; (4) Agree; (5) Strongly agree.
Table 3.1: Latent construct & Observed variables
Environment Awareness (EV)
EV1 I think the use of collaborative consumer services can protect the Hwang et al. (2017)
environment.
EV2 I think participating in collaborative consumption helps to use energy Amirkiaee and
resources more effectively. Evangelopoulos (2018);
Hamari et al. (2015)
EV3 For me, participating in collaborative consumption is an
environmentally friendly option.

EV4 I think collaborative consumption is a sustainable way of Hamari et al. (2015)

consumption (which can be sustained in the long run).
EV5 I feel I am helping the environment by using collaborative Barnes (2017)

364
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

consumption services.
Social Responsibility (RE)
Amirkiaee and
RE1 I feel the need to take care of society.
Evangelopoulos (2018)
RE2 I want to contribute a part to the society.
Oreg and Nov (2008);
RE3 My responsibility is to fulfill my social obligations.
Wasko and Faraj (2005)
RE4 Whenever I have the opportunity, I will help the society.
Socialization (SC)
SC1 I got to know others through collaborative consumption services.
Through collaborative consumption services, I create good
SC2 Hawlitschek et al. (2018)
relationships.
SC3 I met interesting people through collaborative consumption services.
Möhlmann (2015);
Using collaborative consumption services helps me find a group of
SC4 Lamberton and Rose
people with similar interests.
(2012)
I think participating in collaborative consumption services brings
SC5 Barnes (2017)
people closer together.
Hedonism (HE)
HE1 I think using collaborative consumption is an interesting experience.
Compared to other products, I enjoy using collaborative consumption
HE2
time.
In the course of collaborative consumption use, I felt the excitement
HE3 Babin et al. (1994)
of experiencing a new one
HE4 I enjoy participating in a new form of consumption.
I think I will continue using it because I like it, not because I'm
HE5
forced.
Intention (IT)
IT1 I plan to use collaborative consumer services in the future.
I look forward to continuing to use the collaborative consumer
IT2
services in the future.
Bhattacherjee (2001)
I find myself participating in collaborative consumer services more
IT3
often in the future.
IT4 I plan to use regular collaborative consumer services in the future.
If possible, I will engage in collaborative consumer activities more
IT5
often.
Venkatesh et al. (2012)
It is likely that I will participate regularly in collaborative consumer
IT6
communities in the future.
Source: Authors' compilation

365
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.1.4. Sampling design

In this study, the target population is young adults who are currently living, studying,
and/or working in Ho Chi Minh City. Being one of the largest megacities in Vietnam, HCMC with
a nearly nine-million population originated from all different parts of the country (2019) is an ideal
choice as a survey site. As mentioned, this study is primarily about the intrinsic incentives of
the intention to acquire collaborative consumption of young adults. Therefore, observations by
those who are outside the 18-35 year old range or those are not aware of the investigated practice
(i.e. CC) would be rejected.
Henceforth in the study, Exploratory Factors Analysis and Regression Analysis would be
focused on. To firmly produce intuitive and reliable results, the minimum sample for each analysis
method would be respectively discussed. Based on the study of Hair et al. (2006) and Tabachnick
and Fidell (2007), the proposed minimum sample size would be 200 to prevent any losses or faulty
that might occur during the sample collecting and data processing procedures.
Convenience sampling was nominated in the study. Despite its high error possibility which
might lead to inaccurate results, this method is still preferred for its low budget requirement in a
short amount of time. Besides this sampling technique, as its name implies, allows the researchers
to choose the samples from the population, which is the available and accessible source.
3.2 Data collection
Facing social distancing order due to the spread of COVID-19 pandemic, an online survey
using a web-based structured questionnaire was conducted in one-week period targeted at
Vietnamese young adults living in HCMC. From the 416 survey responses received, 355 were
retained as valid for final analysis, corresponding to 85.34%. The table 3.2 below describe the
overall statistics of survey participants based on percentage value. The general variables used in
the study were Gender and Age group.
Table 3.2: Sample characteristics

Group profile characteristic Frequency Percentage(%)

Male 253 71
Gender
Female 102 29

18-20 56 16
Age
21-23 50 14

366
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

24-30 217 61

31-35 32 9

3.3. Reliability Test

Table 3.3: Reliability testing results

Factors Cronbach’s Alpha

Environment Awareness (EV) .845

Social Responsibility (RE) .824

Socialization (SC) .882

Hedonism (HE) .861

After testing Cronbach’s alpha reliability coefficient and implementing factor analysis with
varimax rotation, 4 factors have high Cronbach’s α coefficient (over 0.8). Moreover, all the
observed variables are higher than 0.3, which means there are no eliminated variables. To
conclude, the measurements of this study are acceptable in reliability.
3.4. Exploratory Factor Analysis
Table 3.4: KMO and Bartlett's Test for Independent variables

Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .877

Approx. Chi-Square 3313.387

Bartlett's Test of
df 153
Sphericity
Sig. .000

The table analyses the result of the last time researchers ran the EFA test. Factor analysis
appropriate with research data because KMO coefficient = 0.877 (higher than 0.5), Bartlett's Test
is 3313.387 and significance level Sig = 0.00 < 0.05.

367
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

After running EFA, all items in matrix components rotate larger than 0.5. This means that
all accepted items for validation and independent variables will be sorted into 4 groups: EV, RE,
SC, and HE. According to the last Rotated Component Matrix, we have new groups of variables:
Table 3.5: Results from EFA testing

Variables Observed variables Type

EV EV1, EV2, EV3, EV5 Independent

RE RE1, RE2, RE3, RE4 Independent

SC SC1, SC2, SC3, SC4, SC5 Independent

HE HE1, HE2, HE3, HE4, HE5 Independent

The exploratory factor analysis (EFA) was also applied to verify the validity and reliability
of the items comprosing of dependend variable (i.e. Intention to acquire) and showed satisfactory
result.
3.5 Pearson Correlation Analysis
The linear correlation helps to analyze the correlation between the dependent and
independent variables. The results show that Socialization has closed correlation with Hedonism
(coefficient = 0.506) and Hedonism also affects most strongly on Intention with coefficient equal
to 0.645.
In conclusion, there is a correlation between factors EV, RE, SC, HE and IT so authors can
continue putting all these variables in linear regression.
3.6. Linear Regression
Table 3.6: Linear Regression

Model Summary

Adjusted R Std. Error of the Durbin-

Model R R Square Square Estimate Watson

1 .694a .482 .476 .53126 2.065

368
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ANOVA

Model Sum of df Mean F Sig.

Squares Square

Regression 91.845 4 22.961 81.354 .000b

Residual 98.783 350 .282

Total 190.628 354

Coefficients

Unstandardized Standardized
Coefficients Coefficients

Std.
Model B Error Beta t Sig.

(Const
.152 .235 .649 .517
ant)

EV .085 .043 .086 1.976 .049

1
RE .227 .047 .196 4.806 .000

SC .089 .040 .103 2.220 .027

HE .552 .049 .516 11.231 .000

After running Linear Regression, as a result in table, four variables EV, RE, SC, HE have
Adjusted R Square 0.482. Therefore, 4 independent variables show that the regression affects
49.5% of the variance of the remaining dependent variable 50.5% due to extrinsic variables and
random errors.
The purpose of the F test in the ANOVA is to test whether the linear regression model is
generalized and applicable to the whole. For instance, in this study, Table introduces that the sig
value of F test is 0.000 < 0.050. So that, the linear regression model constructed is consistent with
the overall and significant.

369
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Sig. values of EV, RE, SC, and HE are lower than 0.05 so these variables have a
relationship with the dependent variable (IT - Intention). From the Table 3.6, Regression
Coefficients standardized Beta of variables: EV is .086, RE is .196, SC is .103, and HE is .516,,
which means HE is the most influential variable to user participation in CC, next is RE and SC,
and the last is EV.
Thus, based on Beta column, variables can be explained by the following linear equation:
IT = .086EV + .196RE + .103SC + .516HE
4. Results and Discussions
After conducting all above testings, the researchers summarize the results hypothesis that
are assumed as following below:
Table 4.1: Hypothesis testing

Hypothesis Result
H1 Environment Awareness (EV) has a positive influence on the intention to Accept
acquire CC.
H2 Social Responsibility (RE) has a positive influence on the intention to Accept
acquire CC.
H3 Socialization (SC) has a positive influence on the intention to acquire CC. Accept
H4 Hedonism (HE) has a positive influence on the intention to acquire CC. Accept
Source: Author’s compilation

From the analysis result, all suggested motivations, namely Environment Awareness,
Socialization, Social Responsibility, and Hedonism, are confirmed as important psychological
drivers for the intention of the discussed topic, as to reflect positively on the behavioral intention
(H1, H2, H3, and H4 all accepted).
The analysis outcome has indicated that Hedonism has the most significant influence on
the intention formation for CC activities. It might be explained that some people take part in CC
simply because the service pushes their emotions to the climax as participants can feel the novelty,
content, freedom, and meaningfulness. Making a request and receiving feedback in seconds, being
pre-informed with the driver's information and the costs, or sharing your verdicts after the drive,
such experience is unfamiliar and unprecedented in Vietnam before the CC-era, which obviously

370
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

bring the enjoyment and excitement of sustaining a new technology to the youngsters. This
conclusion is in agreement with the result of Hamari et al. (2015) as enjoyment is also proved and
confirmed to play an essential role in attitude formation and use intentions.
Social Responsibility, the second hypothesis, examined CC services as a means for people
to help others who belong to the same community and gain social capital. Our results indicate
sufficient evidence to prove that individuals with a higher commitment to the community show
more positive attitudes towards CC participation. Considering the context of Vietnam, since their
entries, CC services like Grab or GoViet have given the temporary solution to unemployment for
the country by creating jobs and millions of income-opportunities for locals. Besides, that monthly
sharing and donating campaigns held by these services often taking place have attracted the
advocates of participants, including Vietnamese students. Responsible communities of sharers will
not only create communal benefits but also stimulate a sense of belonging that contributes to
creating a pleasant experience.
Furthermore, Socialization, as expected, is proved to positively influence the intentions of
using CC platforms. Vietnamese people often spend much time hanging out, widening their
connection, and making new friends. Thus, CC applications, which include plenty of people with
daily engagement on a pervasive setting as online platforms with the support of social media, are
thought of as great sources and an ideal environment for finding new mates who can share their
interests and hobbies. This result of our study is also in line with Bucher et al. (2016), who
concluded that some of the strongest influencers of sharing attitudes are social-hedonic motives
and sociability.
Finally, Environment Awareness is shown to have a positive effect on the intention of CC
participation, but to quite a smaller degree in comparison with other observed constructs.
Considering the context of an emerging country like Vietnam, there is obviously less diversity of
CC services presenting which makes the perspective of Vietnamese youngsters and students on
this economic trend narrowed down to ridesharing applications like Grab and Grab-like services
only. Referring to the dense traffic condition in Vietnam with motorbikes as the most popular
means of transportation, using the CC services would make some differences as they help reduce
the number of vehicles traveling around and lower the congestion rate on Vietnamese streets; thus,
is considered a quite sustainable method for the environment. Besides, emerging issues such as
global warming and air degradation which serve as a reminder of a race against time have formed

371
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

major shifts in attitude and behavior and raise the awareness of this environmental crisis of citizens
worldwide, including the Vietnamese.
5. Conclusion & Limitation
Applying exploratory research with 355 valid samples, the researchers statistically
conclude that all proposed variables are important psychological drivers of the acquisition of CC.
Also, based on the regression model, Hedonism has the greatest influence on the intention,
illustrating that some people take part in CC simply because it pushes their emotions to the climax
as participants can feel the contentment, freedom, and meaningfulness. Social Responsibility, the
third hypothesis, examined CC services as a way for people to help others who belong to the same
community and gain social capital. Indeed, this also dominates consumers' disposition regarding
sharing and green behavior. Responsible communities of sharers will not only create communal
benefits but also stimulate a sense of belonging that contributes to creating a pleasant experience.
As expected, Socialization also positively influences the use intentions as considering CC an ideal
setting for finding someone who can share their interests and hobbies. Lastly, Environment
Awareness has the least impact on the intention since individuals are becoming more concerned
with emerging issues such as global warming and air degradation.
In terms of practical implications, regarding policy makers, there should be more incentives
for start-ups to thrive in the context of digital transformation and CC has now become a worldwide
trend. Indeed, governments and local authorities should join hands in order to create a more
friendly and more attractive business environment by offering tax reduction, simplifying formality
and bureaucracy and proposing investment packages to boost the entrepreneurial spirit.
In addition, enterprises may position their promotional campaign as a joint consumer
activity that can offer enjoyment to consumers when one participates. To be more specific, an
adapted system, a middle platform that connects peer-to-peer activity, would have great support to
the increase in the hedonic values of consumers. With a simple ‘achievements’ add-in, managers
can make much of a difference in boosting motivation and monitoring user behavior with their
partake in the activity. Furthermore, the research also provides insights for companies aiming to
do CC as their base business. It guides enterprises on how to better target different consumer
segments for their CC campaigns. Strong social commitment motivation also encourages people
to liquidate their possessions and therefore, stimulates the activity within the collaborative
consumption. The importance of building a community-based platform is that it would help the

372
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

user feel the commitment to society, since our research confirmed the relation between social
responsibility and the intention to use CC services. Thus, holding events and promotions
highlighting the opportunities for helping others may be prudent for managers, which should be
included to motivate CC behavioral intention.
Although this study showed positive and interesting insights, it should be interpreted in
light of some limitations. First, this study does not cover all valid segments. The researchers mainly
focused on Vietnamese young adults with the age ranging from 18 to 25. To expand the theory,
various technological behavioral groups should be taken into consideration for further studies.
Second, because the research was launched online, the generalization of our result should be
approached with caution. Many participants responded quite superficially and unfocused, which
minorly influenced the research results. In addition, all research models and questionnaires can
only be formed from other existing models of studies that researchers refer to, which resulted in a
lack of factors. There are more components that might affect consumers' intention in CC
acquisition, but the authors cannot list all of them in this study.
Despite merits results, our research still draws some boundaries. Nevertheless, practical
implications for enterprises as well as recommendations for future research should also be
considered thoroughly as it mainly focuses on analyzing the inner factors of each person, instead
of external factors such as economic benefits, especially in the context of the burning-money race
is increasingly fierce and nonstop. It is expected that the paper might serve as a good reference for
further research in terms of providing a broad perspective of the motivational factors involved in
CC and the issues relative to their involvement.

373
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References

1. Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human
Decision Processes, 50, 179-211.
2. Amirkiaee, S. Y., & Evangelopoulos, N. (2018). Why do people rideshare? An experimental
study. Transportation Research Part F: Traffic Psychology and Behaviour, 55, 9-24.
3. Babin, B. J., Darden, W. R., & Griffin, M. (1994). Work and/or Fun: Measuring Hedonic and
Utilitarian Shopping Value. Journal of Consumer Research, 20(4), 644.
4. Barnes, S. J., & Mattsson, J. (2017). Understanding collaborative consumption: Test
5. of a theoretical model. Technological Forecasting and Social Change, 118, 281-292.
6. Belk, R. (2014). You are what you can access: Sharing and collaborative consumption online.
Journal of Business Research, 67(8), 1595–1600.
7. Benoit, S., Baker, T. L., Bolton, R. N., Gruber, T., & Kandampully, J. (2017). A triadic
framework for collaborative consumption (CC): Motives, activities and resources &
capabilities of actors. Journal of Business Research, 79, 219–227.
8. Bhattacherjee, A. (2001). Understanding Information Systems Continuance: An Expectation-
Confirmation Model. MIS Quarterly, 25(3), 351.
9. Böcker, L., & Meelen, T. (2017). Sharing for people, planet or profit? Analysing motivations
for intended sharing economy participation. Environmental Innovation and Societal
Transitions, 23, 28-39.
10. Botsman, R., & Rogers, R. (2010). What's mine is yours: the rise of collaborative consumption.
London: Collins.
11. Bucher, E., Fieseler, C., & Lutz, C. (2016). What’s mine is yours (for a nominal fee)–Exploring
the spectrum of utilitarian to altruistic motives for Internet- mediated sharing. Computers in
Human Behavior, 62, 316–326.
12. Eckhardt, G.M., Belk, R., & Devinney, T.M. (2010). Why don’t consumers consume ethically?.
Journal of Consumer Behaviour, 9(6), 426–436.
13. Ertz, M., Lecompte, A., & Durif, F. (2017). Dual Roles of Consumers: Towards an Insight into
Collaborative Consumption Motives. International Journal of Market Research, 59(6), 725–
748.
14. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to
theory and research. Reading, MA: Addison Wesley
15. Guyader, H. (2018). No one rides for free! Three styles of collaborative consumption. Journal
of Services Marketing, 32(6), 692–714.
16. Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., & Tatham, R. L. (2006). Multivariate
Data Analysis. New Jersey: Pearson University Press.
17. Hamari, J., Sjöklint, M., & Ukkonen, A. (2015). The sharing economy: Why people participate
in collaborative consumption. Journal of the Association for Information Science and
Technology, 67(9), 2047–2059.

374
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

18. Hawlitschek, F., Teubner, T., & Weinhardt, C. (2018). Trust in the Sharing Economy: A
Behavioral Perspective on Peer-to-Peer Markets. Die Unternehmung – Swiss Journal of
Business Research and Practice (70:1), pp. 26-44.
19. Herbert, M., & Collin-Lachaud, I. (2017). Collaborative practices and consumerist habitus: An
analysis of the transformative mechanisms of collaborative consumption. Recherche Et
Applications En Marketing (English Edition), 32(1), 40–60.
20. Hwang, J., & Griffiths, M. A. (2017). Share more, drive less: Millennials value perception and
behavioral intent in using collaborative consumption services. Journal of Consumer
Marketing, 34(2), 132-146.
21. Lamberton, C. P., & Rose, R. L. (2012). When is Ours Better than Mine? A Framework for
Understanding and Altering Participation in Commercial Sharing Systems. SSRN Electronic
Journal.
22. Luchs, M., Naylor, R. W., Rose, R. L., Catlin, J. R., Gau, R., Kapitan, S., ... Weaver, T. (2011).
Toward a sustainable marketplace: Expanding options and benefits for consumers. Journal of
Research for Consumers, 19, 1.
23. Möhlmann, M. (2015). Collaborative consumption: determinants of satisfaction and the
likelihood of using a sharing economy option again. Journal of Consumer Behaviour, 14(3),
193–207.
24. Oreg, S., & Nov, O. (2008). Exploring motivations for contributing to open source initiatives:
The roles of contribution context and personal values. Computers in Human Behavior, 24(5),
2055-2073.
25. Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic
motivation, social development, and well-being. American Psychologist, 55(1), 68-78.
26. Schau, H.J., Muñiz, A.M., and Arnould, E.J. (2009), “How Brand community practices create
value”, Journal of Marketing, Vol. 73 No. 5, pp. 30-51.
27. Schor, J. B., & Fitzmaurice, C. J. (2014). Collaborating and connecting: the emergence of the
sharing economy. Handbook of Research on Sustainable Consumption, 410–425.
28. Tabachnick, B. G., & Fidell, L. S. (2007). Using Multivariate Statistics (5th ed.). New York:
Allyn and Bacon.
29. Venkatesh, Thong, & Xu. (2012). Consumer Acceptance and Use of Information Technology:
Extending the Unified Theory of Acceptance and Use of Technology. MIS Quarterly, 36(1),
157.
30. Wasko, & Faraj. (2005). Why Should I Share? Examining Social Capital and Knowledge
Contribution in Electronic Networks of Practice. MIS Quarterly, 29(1), 35.
31. Zhang, Y., Phang, C. W., Gu, R., & Zhang, C. (2019). Antecedents and role of individual
sociability on participation in mobile collaborative consumption. Internet Research, 29(5),
1064-1089.

375
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ICT and social media influence on the MICE market and the
event management
Igor Kovacevic
University of Belgrade, Faculty of Economics
Belgrade, Serbia
[Link]@[Link]

Bojan Zecevic
University of Belgrade, Faculty of Economics
Belgrade, Serbia
[Link]@[Link]

Branislava Hristov Stancic

University of Belgrade, Faculty of Economics
Belgrade, Serbia
[Link]@[Link]

Abstract: Information-communication technology (ICT) and social media became a “new

normal process” of destination management in all tourism aspects, including the meetings and
events industry (MICE) and the event management. Although medical crisis caused by covid-
19, showed all the potential and reach of the various ICT solutions, it is inevitable that face-to-
face meetings will remain as the main form of the meetings industry. Paper focuses on the
theoretical and practical aspects of new economy paradigm caused by ICT in the field MICE
industry, how social media influenced market changes, and what is the framework for using
social media in the event management.
Keywords: event management, social media, meetings, MICE, destination management
JEL: L83, Z33, L82

1. INTRODUCTION

ICT influences new solutions that creates and supports new business environment making
difference between a traditional approach and a “new economy approach”. This changes are
visible in all business segments and all industries, including meetings industry and event
management. However, this difference is not only based on the historical transformation of
business model through time, but represents also the current (nowadays) possible orientation
that companies and destinations rely on, making difference among passive and old-fashioned
on one side, and proactive and contemporary business models, on the other side. Paper
focuses on the essence of change that created new economy and implications on using the
ICT and social media in destination and event management.

2. UNDERSTANDING OF THE NEW ECONOMY APPROACH

Although ICT represents the major game-changing factor and, besides know-how, critical
infrastructure element of all destinations and companies, it still questionable if it is fully
implemented in nowadays operations. Following table shows clear difference among
approaches (Table 1).

376
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1. Traditional VS Contemporary approach to business models

Scope of business Traditional approach Contemporary approach
 Business targets Domestic market Global market
 Orientation Mass production Innovation and technology
 Key resources Capital Know-how and information
 Type of the company Stabile, large enterprises Dynamic, small enterprises
 Organizational design Centralized, clear hierarchy Decentralized, flexible
 Information flow Top-down Bottom-up
Source: Kotler, P. et al. (2008). Marketing Places (kindle version). The Free Press. NY. USA. pp. 15, on the basis
of Kanter, R.M. (1983). The Change Masters. Simon & Schuster. NY. USA, chapter 2

It is possible, also, to determine, several changing aspects caused by ICT in marketing

philosophy of companies and destinations performing in today’s competitive market (Bakic,
2008):

 From mass marketing to one-on-one data base marketing

 From traditional media to usage of interactive media in communication
 From market share to enriching customer relationships
 From sales towards the service
 From most creative to most technology sophisticated marketing
ICT, internet, interned and mobile based services, are changing relations among destination
and clients. Internet is allowing, for proactive oriented destinations and companies,
personalization of communication. By doing this, companies are becoming strategically
oriented towards not-the-average product, but highly personalized. Internet and database
technology have great advantage in creating unique and personalized “conversations” with
each individual clines (consumer), and are foundations for development of customer
relationship management (Schultz, 2000). In order to have real benefits from the two-way
communication, company (and destination) needs to follow and analyses data and sources
through database systems, and turn into large number of useful data that is applicable for long-
term relationship development (Preston, 2000). Unfortunately, great challenge is to create the
best possible client profile, in order to have base for future customization of products and
communications. Data on motivation, lifestyle, values and priorities are often neglect or not
part of data bases, putting demographic data and buying history as the prime data that is put
in the database (Peltier et al., 2002). This makes important negative effect on operational
procedures within tourism companies and destinations.

Essence is that ICT makes new media more interactive in relations with clients and consumers,
but the base for this interactivity is that companies or destinations needs to gather useful data,
and then use that processed information in creating useful strategy of interactive
communication (Davenport et al., 2001). Social media and ICT increased possibility of
interaction between the users and it open a new era of participation and transparency
(Qualman, 2009).

Being interactive is a prerogative for proactive oriented destinations, that are putting clients on
the first place, making sure to communicate personalized massages and experience. Having
this in mind, contemporary and proactive business models of destinations and companies in
tourism are bringing changes in daily communication, usage of data-base management and to
relationship dynamics, as shown in following table (Table 2).

Table 2. Traditional VS Contemporary approach in communication

Component of change Traditional approach in Contemporary approach in
communication communication
Data base
Importance Medium High

377
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Methodology Traditional research Traditional + qualitative

quantitative methods methods
Type of data Demographic, with only small Psychographic, behavioral and
portion of behavioral demographic
Purpose Choosing right communication Understanding individual needs
channels of clients and relationship
development
Communication challenges
Media selection Market coverage There were our clients are
Reach Mass communication Targeted, customized
Communication flow One-way communication from Dialog, continuous
company to the market
Nature of message One, the best message The best one with targeted
massage
Relationship dynamics
Members Company is leading Clients and company together
Directors One-way Mutual, two-way
Measuring
ROI Investments and results Value and return
KPI Number of new clients Keeping the clients
Value Effective Effectiveness and life-time
value
Source: Peltier, J.W., et al. 2003. Interactive integrated marketing communication: combining the power of IMC,
the new media and database marketing. International Journal of Advertising, 22, pp. 98

It is logical to conclude that successful destinations will create their success around
relationship marketing. Relationship marketing is focused on long-term development of win-
win relationship with individual consumers (clients), while value is created as a joint effort of
involved stakeholders (Gummesson, 2002). CRM (as the base element of the relationship
marketing) is, both, business philosophy and coordination strategy, that is based on the ICT,
two-way communication, and a need to determine needs, wants and decision making factors
of individual clients (Lun et al., 2008). For the company (and destination) to build up position
within chosen market segments, it is necessary, not just to have client orientation and to
understand their needs, but also to clearly defines what are the goals of using CRM (Rababah
et al., 2011).

3. WHAT SOCIAL MEDIA BROUGHT TO THE NEW ECONOMY

Social media includes various internet sources of information that are created, initiated,
circulated and controlled by the users themselves, with the aim of mutual exchange and
education about products, brands, services, persons and reasons for certain decision making
(Blackshaw & Nazzaro, 2004). Of course, social media today are the key platforms for
information and experience sharing of and among destinations. By analyzing comments and
posts on social media, destinations and services providers can achieve better understanding
of clients and users’ desires, demands and interests (Leung et al., 2013), but destinations need
to understand that information acquisition is an initial step in planning process of delegates
and travelers (Chen & Gursoy, 2000).
The key thing about social media are experiences that are generated in situations when
companies managed to incorporated reach, personalization and involvements of social media
strategies into the total marketing strategies (Hanna et al., 2011). Social media creates on-line
communities that have joint interests and activities, or better to say networks of users (and
clients) that are interested in exploring interests and activities of others (Nusair, 2012). In
defining social media strategy, it is necessary to understand that purpose of internet and social

378
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

media is the influence and not the control (Safko & Brake, 2009). Therefore, destination
management organizations need to understand that it is not possible to control how visitors,
delegates and clients will share their experience and activities on social media, but with proper
implemented strategy it is possible to influence destination positioning. In addition, social
media made possible that local events and activities of local companies (and destinations)
became global, with marketing and other affects that also went global, and that users and
clients became creators of products (Berthon, et al., 2012). It is the reason why in new
economy we need to understand shift towards user-generated products and user-generated
content.
4. INFLUENCE OF SOCIAL MEDIA ON THE MEETINGS (MICE) MARKET

it is not possible to manage user-generated content or to control form and direction of

communication about destination and destination experiences and products, destination
management organization and individual stakeholders are looking at balancing various social
media channels in order to stimulate and be part of positive communication. Social media and
ICT made a huge impact on consumer buying process as well as on distribution and travel
related information (Xiang & Gretzel, 2010), and on the nature of communication among
decision makers, particularly travelers (Litvin et al., 2008), making a strong effect on marketing
and electronic commerce (Casalo et al. 2010).
Since evolution of social media is happening almost every day, it is possible to note several
forms of communicating message that destinations can use on the meetings market. Most
common is social networking (i.e. Facebook, LinkedIn), publishing (i.e. WordPress), photo
sharing (i.e. Instagram), video platforms (i.e. YouTube), microblogs (i.e. Twitter) (Safko &
Brake, 2009). Or better to say, social media reflects a mix of different types of ICT tools, .such
as podcasts, messaging, applications or blogs (Stillman & McGrats,2008). For example,
TikTok represents latest social media platform based on creating and sharing short video
format, with the content possible to be share on other social media as well, and destination
management organizations just started to using it.
Social media in tourism represents digitalized “word of mouth”, but besides friends, relatives
and colleague’s recommendations, clients are also relying on reactions and experiences of
unknown users, that are making total destination information more objective (Pan & Crotts,
2012). Quality of content depends on the level of engagement and the excitations from
invested resources for managing social media.
Social media has meditating role in storytelling, reconstruction and repeated feel of destination
experiences (Tussyadiah & Fesenmaier, 2009). Content is user-generated in order to assist
other network members to better understand a destination.

5. USING SOCIAL MEDIA IN EVENT MANAGEMENT

Events (congresses, conferences, receptions) organizers, in the field of associations,

corporations or government events, are more oriented towards using social media channels in
process of communication with delegates and an event partners. Event organizers are looking
at real-time communication, that helps them avoid late responses and registrations, which
makes social media a great platform. In the same time, social media represents a tool that is
used to increase event awareness among more targeted potential delegates and to attract new
delegates. Different social media has different role in event organization and in event
management. Through social media, event organizers can engage with delegates, involve
them in topics and speakers’ selection, content form and other event elements.

379
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Social media allows event organizers opportunities that traditional media, due to high costs,
low response rate and time gap in answers, could not provide, and that is a situation that
through social media delegates are sharing their experience and thoughts on relevant content,
evaluations, marking, comparing qualities, and all that in real time (Everton, 2007). This makes
a benefit for, both, organizers and delegates.
Event organizers found possibility to use various social media channels before, during and
after the events. Before the event, on social media (i.e. Facebook, LinkedIn) delegates can
rang sessions according ot individual preferences. During the events, various blogs (i.e.
Twitter) allow constant and live interaction among delegates, speakers and organizers. After
the events, experiences are being user-generated in forms of comments, photos and videos.

Benefits from using social media during event management and organizations are as following
(ICCA, 2013):
- More intensive contact and communication with target segments before, during and
after the vents, helps organizers to identify interests, needs and wants of the target
market and delegates
- Increased level and quality of feed-back information is the base line for learning and
customizing future events content of the organizers, and can be used as the
reference and PR of events
- Increase number of potential delegates and total market, and increased reach of
messages
- Real-time communication and problem solving

Following table (Table 3) show major activities on social media during event management
process of a certain event
Table 3. Event management activities on social media

380
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Social Before the event During the event After the event
media

Facebook Creation of separate page Clearly defined plan of posts Statistics analyses,
with the title of the event and information flow defining parameters,
(independent from the page understanding key
Influence on delegates to
of a destination or event audience, reporting and
share their impressions and
organizer), as well as measuring results
experiences
mapping using Facebook
Posting and sharing video
locations Supporting conversations,
and photo content
replying to questions,
suggestions and critics Influencing delegates to
tag themselves and
colleagues on photos and
videos

Twitter Creation of separate Creating a Twitter wall – Posting and sharing video
hashtag (#) for the event allow live projection of and photo content
impressions, questions and
delegates thought as the
Initial launching of the way of communication with Analyses of the statistics,
hashtag of the event and speakers and organizers defining trends, and
continuity in the usage improvement measures

Motivate delegates to
Creating a separate event
comment about the panes
profile using the Tweet My
contents, speakers and
Events option
presenters, as well as a tool
of announcing changings in
the agenda of the Event

Youtube Creation of separate Live streaming of selected Posting video content after
channel of the association parts of the event (such as the event
or the company opening sessions, some
plenaries…)

In case company or
association is organizing
larger number of events, it
is recommended that each
event has it own channel

Google+ Creating of separate Creating Collections and Measuring success of the

Google+ profile posting video content event

Live streaming of the

events on Hangout options

381
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Pinterest Creating a separate event Photo sharing and video Photo sharing and video
profile clips sharing clips sharing

Source: Amiando, Event registration and ticekting. Social Media & Events Report 2012: How Is The Event
Industry Using Social Networks .and International Congress and Convention Association ICCA. 2013. Social
Media for Meetings

Described activities allow strengthening relations among event itself, the organizer and the
delegates, while the content generates in this process is considered as more objective
(Everton, 2007). Event’s organizers take a roll of communication administrators, but more
proactive organizers will be active in communication and answering all questions and doubts.
Through social media organizers are trying to achieve higher visibility of the event, that will in
long term period lead to maintaining or increasing delegates number and delegates
satisfaction. This long-term goals should be dividing into several operational sub-goals, while
certain researches also identifies reasons why organizers of events use social media (Lee &
Tyrrell, 2012): communication with other events organizers as a form of networking, getting a
feedback from the delegates, promotion of the conference and sharing conference education
program, destination research for the organization of future congresses and events, research
and communication with other service providers and similar.
Users of social media can be classified into several categories: watchers, sharers, commenters
and producers (Li, 2010). This segmentation has it foundation in behavior in the process of
creating and exchanging content on social media, while users can change their role, depending
on the event organization or company lifecycle.
Watchers represent a category that use social media as one of sources during decision making
process and destination and service providers selection. This category is the least active in
producing a content and in communication with destination and service providers. Sharers,
besides being active in following various social media channels, are looking at quality content
to share among own networks of contacts. Sharing is done in real-time, and represent and
instant reaction on the content they liked and support. Commenters are looking at open and
transparent communication with destination and service providers. They are loyal to a
destination or to a brand, and will keep insisting on service quality consistency. They will initiate
a communication. Producers are very active in producing a content and in communication.

6. CONCLUSION

In the new economy, ICT and social media became inevitable part of destination and event
management. In order to use them, destination management organization and event
organizers needs to understands that shift from traditional to contemporary approach is not
just based on the used technology as infrastructure, but rather represents change in
organization approach to the business model and the market. We need to understand that all
working destinations and event organizers are part of global working environment, and are
globally competing for clients and delegates, but not all are using benefits and tools that social
media and ICT brought

REFERENCES

382
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Amiando, Event registration and ticketing. Social Media & Events Report 2012: How Is the
Event Industry Using Social Networks. Available at
[Link] [Accessed November 2013]
Bakić, O. 2008. Marketing u turizmu, Čigoja štampa. Beograd.
Berthon, P.R., Pitt, L.F., Plangger, K. & Shapiro, D. 2012. b Marketing meets Web 2.0, social
media, and creative consumers: Implications for international marketing strategy. Business
Horizons. Vol 55, pp. 261-271
Blackshaw, P., & Nazzaro, M. 2004. Consumer-Generated Media (CGM) 101: Word-of-mouth
in the age of the Webfortified. Report issued by Intelliseek. USA, pp 2. Available at:
[Link] [Accessed 1 May 2020]
Casaló, L. V., Flavián, C., & Guinalíu, M. (2010). Determinants of the intention to participate in
firm-hosted online travel communities and effects on consumer behavioral intentions. Tourism
management, 31(6), 898-911.
Chen, J. S., & Gursoy, D. (2000). Cross-cultural comparison of the information sources used
by first-time and repeat travelers and its marketing implications. International Journal of
Hospitality Management, 19(2), 191-203
Davenport, T.H., Harris, J.G. & Kohli, A. 2001. How do they know their customers so well?
Sloan Management Review, 42(2), pp. 59-69
Everton, R. 2007. How Web 2.0 creates trust for you and your events. Available at:
[Link] [Accessed 15 March
2020]
Gummesson, E. 2002. Relationship Marketing in the New Economy. Journal of Relationship
Marketing, 1(1), pp. 37-57
Hanna, R., Rohm, A., Crittenden, V.L. 2011. We’re all connected: The power of the social
media Ecosystem. Business Horizons. Vol 54, pp. 265-273
International Congress and Convention Association ICCA. 2013. Social Media for Meetings.
Available at: [Link] [Accessed November 2013]
Kotler, P. et al. (2008). Marketing Places (kindle version). The Free Press. NY. USA.
Lee, W. & Tyrrell, T.J. 2012. Arizona Meeting Planners’ Use of Social Networking Media, pp.
128, In Sigala, M., Christou, E. & Gretzel, U. 2012. Social Media in Travel, Tourism and
Hospitality Theory, Practice and Cases. Ashgate Publishing Limited. UK.
Leung, D., Law, R., van Hoof, H., & Buhalis, D. 2013. Social media in tourism and hospitality:
A literature review. Journal of Travel & Tourism Marketing, 30(1-2), 3-22
Li, C. 2010. Open Leadership: How Social Technology Can Transform the Way You Lead.
First Edition. Jossey-Bass. USA
Litvin, S. W., Goldsmith, R. E., & Pan, B. (2008). Electronic word-of-mouth in hospitality and
tourism management. Tourism management, 29(3), 458-468
Lun, Z., Jinlin, L., & Yingying, W. 2008. Customer relationship management system framework
design of Beijing Rural Commercial Bank. In proceedings of IEEE International Conference on
Service Operations and Logistics, and Informatics, pp. 97-101
Nusair, K., Erdem, M. & Okumus, F. 2012. Users’ Attitudes Toward Online Social Networks in
Travel and Anil Bilgihan, pp. 207, In Sigala, M., Christou, E. & Gretzel, U. 2012. Social Media
in Travel, Tourism and Hospitality Theory, Practice and Cases. Ashgate Publishing Limited.
UK.

383
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Qualman E (2009) Socialnomics: How Social Media Transforms the Way We Live and Do
Business. Hoboken, NJ: Wiley
Pan, B. & Crotts, J.C. 2012. Theoretical Models of Social Media, Marketing Implications, and
Future Research Directions, pp. 76, In Sigala, M., Christou, E. & Gretzel, U. 2012. Social
Media in Travel, Tourism and Hospitality Theory, Practice and Cases. Ashgate Publishing
Limited. UK.
Peltier, J.W., et al. 2003. Interactive integrated marketing communication: combining the power
of IMC, the new media and database marketing. International Journal of Advertising, 22, pp.
93-115
Peltier, J.W., Schibrowsky, J., Schultz, D. & Davis, J. 2002. Interactive psychographics: cross-
selling in the banking industry. Journal of Advertising Research, 42(2), pp. 7-22
Preston, C. [Link] problem with micro-marketing. Journal of Advertising Research, 40(4),
pp. 55-58
Rababah, K., Mohd, H. & Ibrahim, H. 2011. Unified definition of CRM towards the successful
adoption and implementation. Academic Research International. Volume 1, Issue 1, pp. 220-
228
Safko, L. & Brake, D.K. 2009. The Social Media Bible - Tactics, Tools & Strategies for Business
Success
Schultz, D.E. 2000. Customer/brand loyalty in an interactive marketplace. Journal of
Advertising Research, 40(3), pp. 41-52
Stillman L and McGrath J (2008) Is it Web 2.0 or is it better information and knowledge that we
need? Australian Social Work 61(4), pp. 421–428
Tussyadiah, I. & Fesenmaier, D. R. 2009. Mediating tourist experiences: access to places via
shared videos. Annals of Tourism Research, 36(1), pp. 24-40
Xiang, Z., & Gretzel, U. (2010). Role of social media in online travel information search.
Tourism management, 31(2), 179-188.

384
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

SELECTED ASPECTS OF NATURE

CONSERVATION MANAGEMENT AT LOCAL
LEVEL ON THE EXAMPLE OF COMMUNES IN
POLAND
JACEK WITKOWSKI
Department of Economy and Economic Management
Faculty of Administration
Lublin University of Technology
Poland
[Link]@[Link]

Abstract: The purpose of the present article is to characterize the

impact of the lowest level self-governments in Poland on the protection
and environmentally friendly use of local natural values by means of
some available management instruments. The research sample
consisted of a group of mayors performing management functions in 49
communes i.e. administrative units characterized by the smallest
territory in Poland. Diagnostic poll method based on a questionnaire
containing 34 closed questions was applied as the research method.
Additionally, the analysis of selected statistical data was carried out for
the communes encompassed by the research. It has been found that the
part of self-governments is still relying on the planning documents which
have been adopted before the entry into force of currently applicable
regulations in the scope of nature protection and before the end of the
process of creation of the areas of Natura 2000. Not all local
development plans wholly encompass the surface of the areas subjected
to protective regimes which may pose a risk of their insufficient
protection. The fiscal instruments and the legal administrative tool in the
form of right to place selected areas and objects under legal protection
are marginally used by local authorities.

Keywords: Commune, Nature Conservation, Sustainable Development,

Local Self-government, Management, Fiscal Instruments

1 Introduction

Local governments, managing their subordinate areas, achieve many

social and economic goals, and their main task should be caring for the

385
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

interests of the local community. In the era of the popularity of the concept
of sustainable development, which was born in response to a significant
degradation of the natural environment, local authorities should also take
care of the protection and environmentally friendly use of local natural
values, which is also enshrined in the legal acts of many countries. With a
range of competences and instruments at their disposal, they may attempt
to directly or indirectly affect the state of valuable ecosystems, and their
activities in this field may be complementary to actions taken at higher
levels of power.

The issue of environmental management is becoming more and

more the subject of interest of scientists, however, the publications are
dominated by the issue of influence on the state of the environment of
transnational and governmental organizations. The latter, despite some
differences in national environmental protection systems (Mazur, 2011),
still have relatively the broadest competences to develop and then
implement various pro-environmental solutions. However, some papers
from the turn of the twentieth and twenty-first centuries also emphasize
the role of local actors in the development process taking into account
socio-economic and ecological needs (Harris, 2000; Kates, Parris and
Leiserovitz, 2005; Organisation for Economic Co-operation and
Development, 2002). One of the factors that has caused increased
interest in nature protection management at the local level in recent years
was the implementation of Natura 2000 in European countries and
delineation of European ecological network areas within it, which have
become another legal form of nature protection. Thus, local governments
were put in a new situation in which they were forced to carry out their
tasks in the conditions of restrictions on the use of some parts of their
subordinate areas (sometimes significant). This thread is touched upon by
many authors, however, they most often emphasize the problem of the
attitude of local officials to the new form of protection and the concerns
related to it. At the same time, there is still a lack of research on the use
by local authorities of competences and impact instruments to increase
the effectiveness of protection of both Natura 2000 areas and other
natural values subject to legal protection requirements.

The purpose of the article is to characterize activities undertaken

by the lowest level self-governments to strengthen the protective functions
of legally protected areas and facilities using selected economic, legal and
planning instruments. This problem was presented and discussed on the
example of the smallest territorial self-government units in Poland called
communes. In particular, attempts were made to find answers to the
following research questions:

386
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

- can local government documents regarding spatial planning be

considered reliable from the point of view of nature protection needs and
thus promising in terms of future prospects for the implementation of
conservation tasks?

- do local authorities use tax and fiscal instruments to stimulate behavior

to protect nature in restricted areas?

- to what extent do local governments use the competence to cover new,

valuable natural ecosystems with legal forms of protection?

2 Literature review

Among the research threads presented above, the topic of local spatial
planning in the context of nature protection requirements is relatively often
discussed in the literature. In some publications, the creation of the
European ecological network was an impulse to consider this issue. For
example, Beunen (2006, pp. 2, 16), writing about the implementation of
the Habitats and Birds Directive in the Netherlands, states, among other
things, that this caused a delay in the planning processes themselves and
points out that in the initial phase these plans took insufficient account of
the recommendations contained in the directives. In another study
prepared in 2017 for the needs of the European Commission, attention is
drawn to the need for spatial planning that will ensure connectivity
between individual protected areas and their surroundings. In addition, the
authors postulate the use of integration strategies or strategies for the
division of conflicting social and economic functions of Natura 2000 areas
in development plans (Simeonova et al., 2017, pp. 51-53). Woźniak writes
about conflicts that may occur when planning sustainable development,
distinguishing between conflicts on the line: individual interest - individual
interest, individual interest - public interest, and conflict within the public
interest itself, when the state (regional) interest does not coincide with the
municipal (Woźniak, 2011, pp. 140-141). According to Proebstl,
something that can protect against similar conflicts and at the same time
increase the effectiveness of nature protection is involvement in the
planning processes of local interest groups (Proebstl, 2003, p. 345).
Private owners' understanding of the needs of nature protection and
informing them about long-term development plans and including them in
the process of creating these plans seems to be crucial. Otherwise, as the
authors of a study in Turkey state, for example, the local community,
losing benefits due to imposed restrictions on use, could continue to
degrade the environment (Alkan, Korkmaz and Tolunay, 2009). The
thread of the participation of local entities in the creation and
implementation of sustainable development plans, which is obviously

387
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

closely related to the issue of development, is also taken up in work

related to the realities of Lithuania. In this case, the authors note that local
authorities have a special responsibility for the success of the process, as
they should support and stimulate the enthusiasm of social participants
during the creation of the strategy, as well as monitor progress in this area
(Čiegis and Gineitiene, 2008, p. 111). The decentralization process in
planning and environmental management sometimes goes even further.
In the publication about the situation in the Scandinavian countries one
can read, among others, that as a result of the reorganization carried out
in Sweden in 1991, the role of local self-governments with traditionally
strong powers in the field of spatial planning was marginalized in favor of
the newly created local structures associating private official organizations
(Hongslo et al., 2016, pp. 7, 13). Hoffman raises an interesting aspect
related to spatial planning in the context of nature conservation
management. Based on a case study, he comes to the conclusion that the
challenge in the process of spatial planning, if it is to lead to the protection
of natural resources on private land, is to ensure the mutual cooperation
of their owners with each other when negotiating their commitment to a
shared vision of the future. Planners should therefore take care of
participatory aspects of planning so that the process of creating plans is
also a forum for public discussion (Hoffman, 2017, p. 562). Finally, it is
worth mentioning several publications in which the authors analyze
whether local planning documents meet the requirements for nature
protection on the example of specific Polish municipalities (Gałecka-
Drozda et al., 2019; Mastalska-Cetera and Krajewski, 2015; Antolak,
2011; Giedych, 2017).
In order to implement the idea of sustainable development at the
local level, it is necessary for local governments to finance pro-
environmental activities, which can also be done through an appropriate
tax policy. For example, in a study prepared by the Directorate-General
for Economic and Financial Affairs operating at the European
Commission, it is noted that in virtually all EU countries property taxes are
imposed at the local level, and in some they are even the main tax base
for local authorities. On the other hand, there is a widespread view that
they are partly a kind of fee for goods and services provided locally
(European Commision, 2012, p. 23). Meanwhile, concessions in this and
other local taxes can be oriented in a way that serves the purposes of
nature protection. The authors of the report issued by the Institute for
European Policy claim that such possibilities exist in the fiscal systems of
various countries, but are currently not used on a larger scale (Kettunen
and Illes (Eds), 2017, p. 62). As for direct funding, although the example
of Australia (Wild River, 2006) shows that local environmental expenditure
may be higher than that borne by central authorities (measured relative to
other expenses), but in Europe this is rare and local governments often

388
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

expect funding from central governments. This problem is raised in the

work discussing the solutions used in Portugal in recent years, in which
case the authors focus on the issue of compensation to local authorities
for restrictions on the use of protected areas. The conclusions state,
among other things, that the regulations introduced in this country and the
related transfers of funds were of importance above all for those self-
governments that managed units with the largest share of protected area
(Santos et al., 2010, p. 21). As the study by D. Guzal-Dec shows, also
heads of communes in Poland would expect compensation for the
presence of legally protected areas in their territory (Guzal-Dec, 2015).
The involvement of local authorities in conservation protection of
nature can also be an important element of eco-development
management of a given area, because it not only builds a specific image,
but also creates the basis for developing local entrepreneurship based on
the use of natural values for tourism in an environmentally friendly way.
However, this thread is rarely discussed in the literature, especially when
it comes to the European continent. Considering the research area
proposed in this paper and the specificity of Poland's legislative
conditions, two publications can be cited in this context in which the issue
of creating new forms of legal protection by local governments is raised.
Both Wolańska-Kamińska and Ratajczyk (2014) as well as Dawid and
Deska (2014) on the example of selected areas generally indicate low
activity of officials in this area. In another document constituting the report
of the Supreme Chamber of Control, after auditing selected local
government units in Poland, many irregularities are found regarding the
creation and functioning of protected areas and objects, such as outdated
legal provisions that have not been adjusted in a timely manner, the lack
of rules, standards or procedures for the protection of valuable natural
objects, and no reviews of their status (Supreme Chamber of Control,
2018).

3 Materials and methods

To achieve the research objective, a study was conducted on a group of

communes located in the Lubelskie Voivodeship in the south-eastern part
of Poland. The diagnostic survey method based on the questionnaire
containing 34 closed questions was used, as well as statistical data
obtained and analyzed.
The survey was conducted in the first half of 2019, and the
respondents were mayors, i.e. persons managing the work of 49 local
governments. The respondents completed the questionnaire forms by
checking one or more of the available options in each of the points. The
communes represented by the persons covered by the survey were

389
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

selected in such a way that within each of them protected areas and
objects were located (including areas of the European ecological network
Natura 2000 delineated in 29 communes). Information on the occurrence of
legal forms of nature protection in individual areas was taken from the
database of the General Directorate for Environmental Protection GDOŚ),
an institution which in the Polish legal system is responsible for the
implementation of tasks in the field of environmental protection and nature
protection.
In addition, the data of the Central Statistical Office (GUS), i.e. the
central office of state administration in Poland dealing in the collection and
access to statistical data, was analyzed. The materials used in the work
refer to the 49 communes surveyed and relate to the percentage of areas
of municipalities covered by development plans and the area of legal forms
of protection occurring in each of the communes. In the second case, the
appropriate conversion into indicators showing the percentage share was
made by the Author comparing the given values with the total area of
communes. Data on the forms of protection created under the decisions of
municipal bodies were obtained from local government documents
(resolutions of municipal councils) shared by GDOŚ.

4 Results

4.1 Spatial development plans and legal forms of nature protection

In Poland, communes authorities are required to prepare and to
implement the document called the study of land use conditions and
directions of spatial development. Furthermore, communes authorities
may optionally adopt the local spatial development plans containing the
provisions which, in contrary to the study, are legally binding. The purpose
of the above plans is to determine the land use conditions and methods
development thereof while the legislator clearly provides that it is
mandatory to determine, on this occasion, the principles of environment,
nature and landscape protection (Planning and Spatial Development Act
of 27 March, 2003, art. 14-15). Already in the 1990s, the obligation has
been imposed on the communes to prepare the study under the pain of
loss of binding force of previously applicable local development plans
(Spatial Development Act of 7 July, 1994). Nevertheless, the research
carried out by the Author many years later in Lubelskie Voivodeship
demonstrated that this task was not completed by the significant part of
commune authorities within prescribed time limit and that there were still
communes without this basic planning study (Witkowski, 2008).
Currently, all communes encompassed by the research are in the
possession of the study of land use conditions and directions of spatial
development. However, from answers given in questionnaires it appears
that only a part of them have relatively new documents prepared during

390
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the period of recent 5 years (22 indications – 45% of total number) but the
study implemented more than 15 years ago is used as the basis in almost
every third commune i.e. at the time when the outlining of Natura 2000
areas was commenced and when the provisions of already expired Act on
Nature Protection from 1991 were binding. When it comes to the local
plans of spatial development, the degree of coverage of surface area of
individual communes with them is obviously diversified. In case of 28
communes (57% of their total number subjected to analysis) their whole
surface area is covered by above plans or this coverage is almost equal
to 100%. However, on the other hand, the local development plan in 10
communes covers less than 1% of their territory (Table 1).

Table 1. The share of areas covered by local spatial development plans

in the total area of examined communes.
% of total area Number of communes
0-1 10
1-10 5
10-20 2
20-40 3
70-80 1
90-99 2
99-100 26
Source: Own preparation based on: GUS, 2019

As already mentioned above, the local development plans are the

acts of local law. Therefore, their provisions should directly determine
the admissibility of planned undertakings. This is very important, among
others, in case of legally protected areas with obligatory limitations in
use. This applies, to different extent, to each commune encompassed by
the research, with the following forms of nature protection: areas of
Natura 2000, ecological sites, protected landscape areas, landscape
parks and, less commonly, nature reserves and national parks (GDOŚ,
2019). In 50% of communes, the total percentage of the above area is
not higher than 20%. However, the research sample encompassed the
units with significantly higher percentage of protected areas (Table 2).

Table 2. The share of areas covered by legal forms of nature protection

in the total area of examined communes
% of total area Number of communes
0-20 24
20-40 12
40-60 8
60-80 3
80-100 2

391
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Source: Own preparation based on: GUS, 2019

Based on the indicator illustrating the percentage of areas under
nature protection and areas covered by local spatial development plans
in the entire area of individual municipalities one can found that land use
has been determined in some communes in a legally binding manner
only for a part of areas encompassed by legal protection. Such situation
takes place in every third self- government unit (16 communes).
Additionally, it should be stated that the disproportions in this sco pe are
really significant, among others in the communes which are known in the
region for the presence of many natural values.
In order to correctly determine the methods of management in terms
of nature protection needs, it is necessary to thoroughly identify the
occurring naturally valuable elements through the execution of natural
stocktaking encompassing the diagnosis of environment condition as
well as the characteristics of its resistance to anthropopressure as well
as evaluation of environment transformations status (Koreleski, 2009, p.
38-39). In an item contained in the questionnaire, the respondents have
been asked whether such type of action has been carried out in the
territory of their commune and, in case of positive answer, how long ago
this action took place. The answer that the stocktaking was carried out
more than 10 years ago (20 indications – 41%) was the most frequently
selected option and the answer that such procedure was never carried
out (13 indications – 26%) was indicated more seldom. Due to the fact
that, in the year 2015, new solutions have been introduced in the scope
of execution of nature stocktaking (The Act of 9 October amending the
act on providing information on the environment and its protection,
public participation in environmental protection and on environmental
impact assessments and certain other acts, 2015) from the research it
appears that these solutions could be applied only by less than every
fifth self-government from among the units subjected to analysis (9
communes).

4.2 Fiscal instruments and their role in stimulation of pro-ecological

behaviours
Although, the capabilities of local self-governments in the scope of
formulation their own fiscal policy are limited in Poland, certain
instruments are available to them and can be used by the local self-
governments in order to stimulate pro-ecological behaviours. The scope
of the above instruments encompasses the local taxes and expenditures.
Pursuant to applicable Polish laws, there are some taxes administered
by commune authorities e.g. real estate tax, tax on means of transport
and agricultural tax. The possibilities of the impact of commune council
(commune legislative body) on the taxable base as well as on the value of

392
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

applied tax reliefs are clearly determined in applicable acts governing

these issues (Local Taxes and Fees Act of 12 January, 1991; Agricultural
Tax Act of 15 November, 1984). In case of real estate tax and tax on
means of transport, only tax exemptions can be applied by commune
councils. However, in case of agricultural tax commune councils are
additionally entitled to adopt tax reliefs (Etel, 2004, p. 82). The
respondents had the opportunity to indicate in the form reliefs and
exemptions in the above-mentioned taxes previously or currently used to
support the following activities: environmental education, development of
ecotourism services and exclusion of some areas from normal agricultural
activity for conservation purposes. The answers show that such measures
of influence were not used at all by the vast majority of the self-
governments studied (42 indications – 85% of total number). Other local
governments applied tax incentives in real estate tax and tax on means
of transport (4 units and 3 units correspondingly) as well as agricultural
tax preferences (1 commune) (Figure 1). If pro-ecological activities were
supported, then preferably those related to the promotion of knowledge
about local natural values and their protection.

45 42
40
35
30
25
20
15
10
4 3
5 1
0
no tax property tax tax on agricultural
solutions means of tax
transport

Figure 1. The use of tax solutions for supporting selected activities related
to nature protection by the number of communes
Source: Own preparation

From obtained results it appears that commune offices are slightly

more ready to participate in the projects contributing to nature protection
when their participation takes the forms other than application of tax
reliefs and exemptions. First of all, from declarations of commune heads it
appears that the majority of their offices used or use the external sources
for financing of undertakings in the scope of environment and nature
protection. Most frequently the funds for such undertakings are raised by

393
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

communes from European funds (37 indications – 75% of total number)

and from national ecological funds i.e. from the National and Voivodeship
Fund for Environmental Protection and Water Management (42
indications – 86%). The above funds are mainly used for co-financing of
projects e.g. construction of onsite sewage facilities and for energy
harvesting from renewable sources. Moreover, about every fourth
respondent (12 persons from 49) in the item concerning the projects
implemented with the whole or partial participation of self-government
indicated also the direct activities intended to protect the local natural
resources.

4.3 Creation of legally protected forms

The local self – governments in Poland are legally empowered to create
some legal forms of nature protection (Nature Conservation Act of 16
April, 2004). The scope the above forms encompasses the following:
nature monuments, nature and landscape complexes, documentation
sites and ecological sites. Although their functioning might lead to the
occurrence of certain impediments for the owners of the areas
encompassed by such form, on the other hand, however, this may be an
essential factor encouraging for greater attention to valuable elements of
natural environment and can be also used for economic purposes, for
instance for development of environment friendly tourism. It should be
concluded that the introduction the above forms leads to the best results
in case of the protection of endangered areas, objects or species
(Symonides, 2008, p. 396).

40 37
35
29
30
25
20
15
10
5 2
0
nature ecological sites nature and
monuments landscape
complexes

Figure 2. The number of legally protected areas and objects established

by the decision of local authorities

394
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Source: Own preparation based on: GDOŚ, 2019

From the data of the General Directorate for Environmental Protection it

appears that up to the present time, in 49 self-government entities
subjected to the research, the total number of the new legally protected
forms i.e. area - oriented as well as object-oriented forms created on the
basis of decisions made by the local officials (commune councils) is equal
to 68 (Figure 2). The majority of them are nature monuments (37) with
significant part of this number occurring in one commune (21).
Furthermore, 29 ecological sites have been created in this area. This case
is also characterized by certain asymmetry because one third of them (10)
has been created as a result of resolutions adopted the council of one
commune. Moreover, in two cases the creation of a nature and landscape
complex has been adopted by communes councils. In 30 communes (61%
of total number) there was no activity of self-governments in the scope of
coverage of selected objects or areas with legal protection (GDOŚ, 2019).
At this point, it is worth mentioning that the number of nature monuments
per 100 square kilometres as well as percentage of land covered by legally
protected areas are significantly less in examined communes than national
average values (Table 3).

Table 3. Selected indicators for Poland and the area of 49 examined

communes
Indicator Poland 49
communes
Number of nature 11,2 5,0
monuments per 100 sq km
Percentage of area 32,6 23,7
covered by legal protection
Source: Own preparation based on: GUS, 2019

Relatively poor engagement of self-governments in this type of activity

will be probably not changed in near future. Only less than every fifth head
of commune (9 indications) said that in his/ her office there are any plans to
adopt future resolutions concerning the creation of new protection forms.
Such lack of readiness to increase the activity in this scope seems to be
confirmed by other answers which, among others, indicate to the fact that
any increase of the land covered by area-oriented protection would be
accepted by 20 respondents (41% of total number) only and that,
additionally, in the group of heads managing the communes with Natura
2000 areas (29 such units have been considered in the research) almost
everybody represented critical attitude towards the present range of
occurrence of the European Ecological Network expressing their reduction
will or even declaring their protest against setting – out thereof.

395
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5 Conclusions
The engagement of self- governments in the scope of environment
management on the planning level is mainly characterized by relatively
frequent occurrence of low attention to the consideration of the present
condition of natural resources which might lead to dangers in the scope of
effectiveness of their protection as well as in the scope of the methods of
use of „green potential” for communes needs associated with development.
Numerous studies of local development prepared long time ago and,
therefore, potentially nonconforming with currently applicable regulations
and the lack of thorough natural stocktaking may result, for instance, in
wrong decisions associated with investments made by self-governments
which, obviously, could contribute to hampering of local development and/
or to environment destruction. The potential effects of this situation are
described by Beunen (2006), and to some extent they are presented in the
case study of the Stęszew commune (Gałecka-Drozda et al., 2019).
Antolak (2011) and Giedych (2017) also point out the imperfections of
municipal development plans in selected local government units and
insufficient consideration of the requirements for nature protection in these
documents. On the other hand, the conclusions drawn on the basis of the
results presented above do not seem to be confirmed in the publication by
Mastalska-Cetera and Krajewski. These authors conclude that the
documents of the communes of the Bolesławiec poviat in western Poland
that they analyze give a good basis for proper management of local Natura
2000 sites (Mastalska-Cetera and Krajewski, 2015). From the questionnaire
it appears that communes are managed by significant part of self-
governments on the basis of planning studies implemented many years ago
which, in combination with the results of the research carried out by the
Author previously (Witkowski, 2008), indicates to the fact that the above
documents have been often prepared by external entities and therefore
they may be less suitable for local circumstances. A similar problem occurs
in other countries of Central and Eastern Europe, which is indicated in
particular in the article on the situation in Lithuania (Čiegis and Gineitiene,
2008, p. 111).
After the completion of the research in the territory of Lubelskie
Voivodeship, the conclusion can be drawn that it is probable that at least a
part of local planning documents (in particular, the studies of land use
conditions and directions of spatial development) could be created with
relatively limited participation or even without participation of local
communities. For example, it is indicated by the answers from which it
appears that no registers of the owners of Natura 2000 areas are available
to local offices as well as the fact that a part of studies has been adopted,
in accordance with information obtained from respondents, before the entry
into force of the act governing the issue of participation of private entities in
the processes associated with the assessment of environmental impact as

396
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

well as disclosure of environmental information to public (The Act of 3

October on providing information on the environment and its protection,
public participation in environmental protection and on environmental
impact assessments, 2008). The need to ensure the proper participation for
the local actors not only at implementation but also already in the
development plans creation phase is strongly emphasized in many
publications (e.g. Proebstl, 2003; Čiegis and Gineitiene, 2008; Hoffman,
2017; Hongslo et al., 2016; Cent et al., 2007). Obviously, the admittance of
private owners to participation in works associated with local development
plans should not be regarded as a warranty ensuring any provisions
favourable from nature protection point of view. On the other hand,
however, the exclusion of this group may cause an opposite effect i.e.
intensification of environment degrading activity in the face of the loss of
benefits stemming from imposed limitations in use (Alkan, Korkmaz and
Tolunay, 2009).
The analysis of activities initiated by self-governments in the scope of
economy indicates that tax reliefs and exemptions encouraging the local
entities for behaviours promoting to the performance of protective tasks
are not used. This is despite the potential that exists in this area,
especially in the case of real estate tax, because, as note by Podstawka
and Rudowicz (2010), revenues from this tax to local budgets are the
largest in Poland, and most often municipal authorities use exemptions
precisely in this tax. Similar conclusions for the whole European continent
are drawn by the authors of the report issued by the Institute for European
Policy (Kettunen and Illes [Eds], 2017). As regards direct financing, the
self – governments in Lubelskie Voivodship are engaged in various
projects mainly in undertakings co-financed from ecological special
purpose funds i.e. public funds. The surveyed local governments
therefore expect financed state participation in implemented pro-
ecological projects, which would partly confirm the statements contained
in the work related to Portuguese conditions (Santos et al., 2010).
Furthermore, it was found that the group of supported actions is
dominated by those that can be classified as indirectly contributing to the
improvement of the state of the environment (sewage treatment plants,
obtaining energy from renewable sources).
The activity of local authorities in establishing legal forms of nature
protection may constitute an essential element of management of
specified area in a manner considering the principles of sustainable
development because it is important for building of positive image and
creates the basis for development of local enterpreneurship based on use
of natural values for eco - friendly tourism. Data analysis indicates that the
legal possibilities for the protection of valuable elements of the natural
environment are used only to a small extent by the local offices, despite
the fact that the presence of such forms in the analyzed area is generally

397
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

smaller compared to other Polish regions. These results confirm those

obtained by the authors who conducted the research in other regions of
Poland (Wolańska-Kamińska and Ratajczyk, 2014; Dawid and Deska,
2014).
In summary, the results of the research carried out on the group of
persons managing the works of selected communes offices in Lubelskie
Voivodeship as well as the data analysis indicate that local authorities are
trying to take action to support the protection of valuable ecosystems.
However, the scale and sometimes the choice of available impact
instruments raises certain doubts and may lead to situation where the
protective functions will be not performed successfully and valuable
environmental resources may be used in a manner not always conforming
with the idea of sustainable development.

References
Agricultural Tax Act of 15 November, 1984. [online] Available at:
<[Link]
[Link]> [Accessed 12 April 2020].

Alkan, H., Korkmaz, M. and Tolunay, A., 2009. Assessment of primary

factors causing positive or negative local perceptions on protected areas.
Journal of Environmental Engineering and Landscape Management, 17(1).
pp. 20-27. doi: 10.3846/1648-6897.2009.17.20-27

Antolak, M., 2011. Praktyka ochrony środowiska w sporządzaniu gminnych

dokumentów planistycznych – na wybranych przykładach. Problemy
Ekologii Krajobrazu, [Link], 6-1, pp. 103-105. [online] Available at:
<[Link]
9b43-4715-8548-491052a473ad/c/vol31_12_Antolak.pdf> [Accessed 24
March 2020].

Beunen, R. (2006) European Nature Conservation Legislation & Spatial

Planning: for Better or for Worse? Journal of Environmental Planning and
Management, 49 (4), pp. 605-619.
[Link]

Cent, J., Kobierska, H., Grodzińska-Jurczak, M. and Bell, S., 2007. Who is
responsible for Natura 2000 in Poland?—a potential role of NGOs in
establishing the programme. International Journal of Environment and
Sustainable Development, 6, pp. 422–435. doi:
10.1504/IJESD.2007.016245.

398
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Central Statistical Office, 2019. Bank Danych Lokalnych. Warszawa:

Główny Urząd Statystyczny. [online] Available at:
<[Link] [Accessed 30 March 2020]

Čiegis, R. and Gineitiene, D., 2008. Participatory aspects of strategic

sustainable development planning in local communities: Experience of
Lithuania. Technological and Economic Development of Economy, 14:2,
pp. 107-117. [Link]

Dawid, L. and Deska, K., 2014. Wybrane problemy tworzenia użytków

ekologicznych na przykładzie powiatu koszalińskiego i miasta Koszalin.
Chrońmy Przyrodę Ojczystą, 70 (5), pp. 445-451.

European Commision, 2012. Possible reforms of real estate taxation:

Criteria for successful policies. European Economy. Occasional Papers
119. [online] Available at:
<[Link]
2/pdf/ocp119_en.pdf> [Accessed 3 April 2020].

Etel, L., 2004. Uchwały podatkowe samorządu terytorialnego. Białystok:

Temida 2.

Gałecka-Drozda, A., Raszeja, E., Szczepańska, M. and Wilkaniec, A.,

2019. Land cover changes in Natura 2000 areas located in suburban
zones: planning problems in the context of environmental protection. Polish
Journal of Environmental Studies, Vol. 28, No. 2, pp. 587-595. doi:
10.15244/pjoes/80894.

General Directorate for Environmental Protection (2019). Centralny Rejestr

Form Ochrony Przyrody. Generalna Dyrekcja Ochrony Środowiska.
Warszawa. [online] Available at: <[Link]
[Accessed 14 April 2020].

Giedych, R., 2017. Funkcjonowanie lokalnych form ochrony przyrody w

miastach, na przykładzie Warszawy, Krakowa, Łodzi, Wrocławia i
Poznania. Prace Komisji Krajobrazu Kulturowego, 38, pp. 43-55. [online]
Available at:
<[Link]
8b54d9cc-0311-43a0-bfaa-e7ee640fc74d/c/PKKK_2017_38-[Link]>
[Accessed 12 May 2020].

Guzal-Dec, D., 2015. Samorząd gminny w kreowaniu zrównoważonego

399
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

rozwoju obszarów przyrodniczo cennych województwa lubelskiego. Biała

Podlaska: Państwowa Szkoła Wyższa im. Papieża Jana Pawła II w Białej
Podlaskiej.

Harris, J. (2000). Basic Principles of Sustainable Development. Medford:

Tufts University. [online] Available at:
<[Link]
[Link]> [Accessed 4 May 2020].

Hongslo, E., Hovik, S., Zachrisson, A. and Lundberg, A. K. A., 2016.

Decentralization of conservation management in Norway and Sweden—
different translations of an international trend. Society & Natural Resources,
Vol 29, pp. 1-17. doi: 10.1080/08941920.2015.1086456.

Hoffman, M., 2017. The role of public land use planning in facilitating
conservation on private land. Natural Areas Journal, 37(4), pp. 556-563.
[Link]

Kates R. W., Parris T. M. and Leiserovitz A. A., 2005. What is sustainable

development? Goals, Indicators, Values, and Practice. Environment:
Science and Policy for Sustainable Development, Volume 47, No. 3, pp. 8–
21. doi: 10.1080/00139157.2005.10524444.

Kettunen, M. and Illes, A. (Eds.), 2017. Opportunities for innovative

biodiversity financing: ecological fiscal transfers (EFT), tax reliefs, marketed
products, and fees and charges. A compilation of cases studies developed
in the context of a project for the European Commission (DG ENV) (Project
ENV.B.3/ETU/2015/0014). Brussels / London: Institute for European Policy
(IEEP). [online] Available at:
<[Link]
en_2017_financing_biodiversity_case_studies.pdf> [Access 9 May 2020].

Koreleski, K., 2009. Problematyka ochrony i kształtowania środowiska w

dokumentach służących realizacji zrównoważonego rozwoju gmin.
Infrastruktura i ekologia terenów wiejskich, 4, pp. 31-42. [online] Available
at: <[Link]
76f82b20-ac36-44c9-a409-c816dccdec6a/c/Koreleski_3.pdf> [Accessed 2
May 2020].

Local Taxes and Fees Act of 12 January, 1991. [online] Available at:
<[Link]
[Link]> [Accessed 24 March 2020].

400
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Mastalska-Cetera, B. and Krajewski, P., 2015. Obszary Natura 2000 jako

uwarunkowanie planowania rozwoju regionalnego. Studia KPZK, No. 161,
148-155. [online] Available at:
<[Link]
00_jako_uwarunkowanie_planowania_rozwoju_regionalnego> [Accessed
11 May 2020].

Mazur, E., 2011. Environmental Enforcement in Decentralised Governance

Systems: Toward a Nationwide Level Playing Field. OECD Environment
Working Papers, No. 34. Paris: OECD Publishing.. doi:
10.1787/5kgb1m60qtq6-en.

Nature Conservation Act of 16 April, 2004. [online] Available at:

<[Link]
[Link]> [Accessed 23 April 2020].

Organisation for Economic Co-operation and Development, 2002.

Sustainable Development Strategies: A Resource Book. Paris-New York:
United Nations Development Programme. [online] Available at:
<[Link]
Clayton,%20Bass%20(2002)%20-
%20Sustainable%20Development%20Strategies%20-
%20A%20Resource%[Link]> [Accessed 14 May 2020].

Planning and Spatial Development Act of 27 March, 2003. [online]

Available at:
<[Link]
[Link]> [Accessed 9 March 2020].

Podstawka, M. and Rudowicz, E., 2010. Wykorzystanie wybranych

podatków lokalnych w kreowaniu polityki fiskalnej gmin. Zeszyty Naukowe
Polityki Europejskie, Finanse i Marketing, 4(53), pp. 79-87. [online]
Available at: <[Link]
[Accessed 11 May 2020].

Proebstl, U., 2003. NATURA 2000 – The influence of the European

directives on the development of nature-based sport and outdoor recreation
in mountain areas. Journal for Nature Conservation, 11 (4), pp. 340-345.
doi: 10.1078/1617-1381-00066.

Santos, R., Ring, I., Antunes, P. Clemente P., 2010. Fiscal transfers for

401
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

biodiversity conservation: The Portuguese Local Finances Law. UFZ

Discussion Paper, No. 11/2010. Leipzig: Helmholtz-Zentrum für
Umweltforschung. doi: 10.1016/[Link].2011.06.001.

Simeonova, V., Bouwma, I., van der Grift, E., Sunyer, C., Manteiga L.,
Külvik, M., Suškevičs, M., Dimitrov, S. and Dimitrova, A., 2017. Natura
2000 and Spatial Planning. [online] Available at:
<[Link]
_spatial_planning_final_for_publication.pdf> [Accessed 20 April 2020].

Spatial Development Act of 7 July, 1994. [online] Available at:

<[Link]
[Link]> [Accessed 29 March 2020].

Supreme Chamber of Control, 2018. Lokalne formy ochrony przyrody.

Informacje o wynikach kontroli. Warszawa: Najwyższa Izba Kontroli.
[online] Available at: <[Link]
nik/pobierz,ksi~p_17_049_201707280641451501224105~01,typ,[Link]>
[Accessed 29 April 2020].

Symonides, E., 2008. Ochrona przyrody. WUW: Warszawa.

The Act of 3 October on providing information on the environment and its

protection, public participation in environmental protection and on
environmental impact assessments, 2008. [online] Available at:
<[Link]
[Link]> [Accessed 15 April 2020].

The Act of 9 October 2015 amending the act on providing access to

information on the environment and its protection, public participation in
environmental protection and on environmental impact assessments and
certain other acts. [online] Available at:
<[Link]
[Link]> [Accessed 15 April 2020].

Wild River, S., 2006. The role of local government in environmental and
heritage management. Article prepared for the 2006 Australia State of the
Environment Committee. Canberra: Department of Environment and
Heritage.

Witkowski, J., 2008. Zaangażowanie samorządów gminnych w działania na

rzecz ochrony przyrody na przykładzie gmin województwa lubelskiego.

402
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Ekonomia i Środowisko, 1(33), pp. 118-129.

Wolańska – Kamińska, A. and Ratajczyk, N., 2014. Powoływanie lokalnych

form ochrony przyrody przez samorządy gmin wiejskich. Woda-
Środowisko-Obszary Wiejskie, Vol. 14. Z. 1 (45), pp. 129-142. [online]
Available at:
<[Link]
ska%[Link]> [Accessed 18 May 2020].

403
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Uncertainty and Exchange Rates: Global Dynamics

(Well, I Don’t Quite Know Anymore)
Jing Lian Suah∗
July 24, 2020

Abstract
This paper offers two points on the impact of uncertainty and exchange rate shocks
on output. (1) A conceptual model where aggregate inefficiencies stem from rational
inattentiveness and bounded expectations amongst agents. Interaction with uncer-
tainty generates output losses. Central banks can target these macro-behaviourial
frictions to stabilise output and prices. (2) Empirical findings from a panel of ad-
vanced and emerging economies. Output and inflation slow in response to uncertainty
shocks. Long-term government bond yields moderate and exchange rates depreciate,
suggesting within-country and between-country flight-to-safety respectively. Exchange
rate appreciation shocks generate similar responses. The Malaysia-specific analysis also
finds Balassa-Samuelsson effects, where divergent responses in productive tradable and
less productive non-tradable sectors lead to differing employment and output dynam-
ics. In a panel fixed effects and quantile regression setting, I find indicative interaction
between output, exchange rate and uncertainty, and a distributional dimension.

Keywords: Uncertainty, Rational Inattention, Bounded Rationality, VAR

JEL Classification: E0, E7

∗
Address: Bank Negara Malaysia, Jalan Dato’ Onn, 50480 Kuala Lumpur, Malaysia Phone: +603 2698
8044 (ext 8652). Email: suahjinglian@[Link]. Any views expressed are solely mine and should not
be taken to represent those of Bank Negara Malaysia. Special thanks to Boon Hwa Tng, Ahmad Othman
Amrul Aaz and Thevesh Thevanathan from Bank Negara Malaysia for the useful feedback and comments.

404
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1 Introduction
Since the US-China trade war in the second half of 2018, trade dynamics have been shaped
by an oscillation between escalation and de-escalation. For export-dependent economies,
especially in East Asia, which is highly integrated in the manufacturing global value chain
(GVC), macroeconomic resilience is brought to questioned as trade growth slowed to its
lowest level since the 2008-09 Global Financial Crisis (GFC). As the COVID-19 pandemic
escalated into a severe global recession, economic uncertainty, stemming from the pandemic,
as well as a range of geopolitical and country-specific political risks, had risen steeply in the
first half of 2020.
The literature has broadly two finding. The first — the response of real economic activity
in response to uncertainty shocks — is well established, though largely on the US and some
advanced economies, and rarely considered in a panel data setting. The second — macro-
financial responses to uncertainty shocks, such as exchange rates and bond yields — is
more ambiguous and given less attention to. In light of these gaps, this paper expands
the both points to consider a wider universe of advanced and emerging economies. On the
second point, I analyse also the responses of currency and bond markets, as well as the
broader uncertainty-macroeconomic-macrofinancial nexus. Specifically, this paper considers
the impact of exchange rates, a determinant of the well-being of small open economies, which
is better documented in the exchange rate and capital flows literature.
Two questions will anchor the discussion. Firstly, what are the dynamics between un-
certainty, the exchange rate and aggregate output? Secondly, what may be the frictions
and channels underlying these effects? The empirical section focuses on the former. The
theoretical framework addresses the latter.
Empirical findings from the uncertainty literature suggest that positive uncertainty shocks
on real economic activity correspond to near-term slowdown, followed by minor corrections.
Bloom (2009)[11] used a calibrated general equilibrium model for the US to examine the
impact of adverse shocks to the Chicago Board of Exchange Volatility Index (CBOE VIX), a
measure of realised volatility over a rolling 20-day period in US equity markets, on investment
made by firms. Subsequently, Baker, Bloom and Davis (2016)[8] proposed a news-based index
of economic policy uncertainty. This included a Purchasing Power Parity (PPP)-weighted
global index.
With a richer addition of indicators in the literature, the degree of persistence and the
presence of a correction varies on the measure of uncertainty used. Macroeconomic uncer-
tainty, as estimated in Jurado, Ludvigson and Ng (2015) [25], generate persistent output
slowdowns without rebounds. This was examined in Jurado, Ludvigson and Ng (2015) itself

405
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[25]. Compared to output responses to realised volatility in financial markets or subjective

uncertainty, output slowdown is persistent and without noticeable rebounds. Alessandri and
Mumtaz (2019) [5] used a threshold VAR to assess differences in the impact of macroeco-
nomic uncertainty shocks between periods with financial crises and periods without. They
find similar responses in Jurado, Ludvigson and Ng (2015) [25]. Responses during periods
of financial crisis are found to be sharper, albeit takes a similar duration to converge to its
steady state.
The literature offers two strands of theoretical underpinnings. The first strand points
to a build-up of precautionary savings amongst households and firms. Risk averse agents
defer expenditure, deterred by the possibility of downside surprises to real income. This
subsequently leads to near-term losses in output. However, there are multiple points of
controversies within the literature. This includes a lack of consensus over the intensity and
duration of such motives, as well as the appropriateness of selected uncertainty measures,
as documented by Lugilde, Bande and Riveiro (2017) [30]. The second strand points to real
options effects, where an expected need to adjust factors of production, which incurs a real
cost of adjustment, leads firms to ”wait-and-see”. Hence, firms defer investment, production
and employment decisions. Bloom (2000) details the long and short-run impacts of the real
options channel.
In extension to the two established views, Leduc and Liu (2016) [28] posit that uncertainty
shocks can be interpreted as aggregate demand shocks. This contributed to theoretical and
quantitative research that linked sustained shortfalls in demand and output responses to
transitory uncertainty shocks. Fajgelbaum, Schaal and Taschereau-Dumouchel (2017) [20]
viewed uncertainty as a shortfall of information flow on the economy essential to investment
decisions. Using a calibrated quantitative model, an unanticipated rise in uncertainty leads
to a prolonged slowdown in output. Persistence is sustained by the endogeneity between
uncertainty, the flow of economic information, and aggregate demand.
Kamber, Karagedikli, Ryan and Vehbi (2016) [26] offers a contrarian view. Using a
structural Factor-Augmented Vector Autoregression (FAVAR) model on US data, they find
that whilst macroeconomic uncertainty shocks lead to a shortfall in output as expected, the
exchange rate appreciates and corporate bond yields increase, a direction that juxtaposes
an aggregate demand shock. However, the study also examines spillovers to other advanced
and emerging economies. Notably, for emerging economies, in response to uncertainty shocks
originating from the US, exchange rates depreciate, whilst the opposite is observed for cur-
rencies of advanced economies typically classified as safe havens, such as the Japanese Yen
and the Swiss Franc, suggesting flight-to-safety.
The real effects of exchange rates, and capital flows, is well documented, though only

406
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

a selection consider uncertainty and exchange rates concurrently. Veerhoogen (2008) [37]
used the 1994 Mexican Peso Crisis as a ‘natural’ experiment, which found that the depre-
ciation shock induced quality-upgrading responses in the manufacturing sector, especially
amongst exporters, and consequently widening within-industry wage inequality. Cravino
and Levchenko (2017) [15] documented further the distributional effects of the 1994 Peso
depreciation shock on prices, which inflationay pressures to be larger for households in lower
income deciles. Saffie, Varela and Yi (2020) [19] and Benigno, Fornaro and Wolf (2020)
[9] further studied the effect of capital flows on sectoral-level productivity and output in
Hungary and the US respectively.
I offer a theoretical framework to examine the effects of uncertainty shocks on the real
economy, specifically output and inflation, through macro-behavioural frictions. My concep-
tual model has two key ingredients. Firstly, firms and households are rationally inattentive.
I simplify Sims (2003)’s [34] conjecture that agents incur costs when processing information,
leading to incomplete usage of the full set of information made available to them. Specif-
ically, my model posits that households and firms are only able to process information up
to k-periods prior. This limited information span informs expectations of future output and
prices, as well as influences the degree of risk aversion. This time-varying parameter then
influences both the quantum and allocation of savings. Secondly, expectations formation are
rational but bounded, drawing from principles underlying Ding (2018)’s [18] κ-augmented
Phillips Curve. In Ding (2018), price expectations are forward-looking, albeit limited, and
anchored by the central bank’s forecasts. I posit that macroeconomic uncertainty influences
the degree in which agents dis-anchor from credible public information, simplified by the cen-
tral bank’s communicated forecast, and resort to agent-specific heuristics that are subject to
innate biases.
These behavioural frictions have three implications. Firstly, there is immediate output
losses as risk aversion rises. Secondly, subsequent expectations formed for real output are
biased downwards, due to a high weight assigned to the immediate past, thereby influencing
price-setting and expenditure-savings decisions. I conjecture that this element contributes to
the persistent shortfall in output relative to the steady state that recent literature on macroe-
conomic uncertainty found, such as in Jurado, Ludvigson and Ng (2015) [25] and Kamber,
Karagedikli, Ryan and Vehbi (2016) [26]. Finally, monetary policy eases more aggressively,
whilst the certainty over the efficacy of monetary policy transmission diminishes due to be-
havioural changes. Nevertheless, this framework opens up the possibility for central banks to
influence price and output expectations during periods of elevated uncertainty as a potential
countercyclical policy tool. Multi-country extensions to the theoretical model, which details
a single economy, can explicitly describe capital flows and exchange rate dynamics.

407
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

My empirical strategy is five-fold, each corroborating my attempt to answer the following

question — “what is the impact of uncertainty shocks on the aggregate economy for open
economies in tandem with exchange rate shocks?”.
The first is closely related to Jackson, Klieson and Owyang (2019) [23], who employed a
non-linear vector autoregression (VAR) with a max-uncertainty variable, drawn from Hamil-
ton (1996)’s [22] max-oil variable. I depart from the extant literature that focuses on firm-
level investment and aggregate-level private sector investment, and focus on aggregate real
economic activity open economies. The second draws extends the analysis to a panel of
advanced and emerging economies with a panel VAR model, with General Method of Mo-
ments (GMM)-style instruments as per Abrigo and Love (2016) [1]. The third estimates a
Bayesian Hierarchical Panel VAR to account for cross-sectional heterogeneity in the panel,
as per Jarocinski (2010) [24] and documented by Dieppe, Legrand and van Roye (2016)
[17]. The fourth departs from macroeconomic structural analysis and analyses the nonlin-
ear relationship between economic uncertainty, the exchange rate and other macroeconomic
variables in a fixed effects setting. The final borrows from the empirical macro-finance lit-
erature. Using a panel quantile regression model, building on Adrian, Boyarchenko and
Giannone (2019) [2], and as per Machado and Silva (2018) [31], I extend the preceding
analysis with a distributional dimension.
The remainder of the paper will be structured as follows. Section 2 proposes the theo-
retical framework. Section 3 describes the data. Section 4 discusses the empirical strategies,
followed by the findings in section 5. Finally, section 6 concludes.

2 Theoretical Framework
The conceptual model contains three major building blocks - households, firms and policy-
makers. With a continuum of economies of such structure, the model is generalisable to a
wider global model with multilateral flows, both real goods and services, and financial, to
which exchange rate dynamics can be modelled explicitly.

2.1 Household
Households maximise expected-utility, subject to a budget constraint. Here, households are
rationally inattentive. Expectations formation is rational but bounded. Both of these are
key elements underpinning my explanation of the empirical findings later. Decision-making
is forward-looking. Written sequentially, consumption and labour supply decisions are de-
termined to maximise the household’s expected lifetime utility, where β is the subjective

408
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1
discount factor. Conventionally, this is the inverse of the long term real interest rate 1+r̄
.
∞
X
max β t E0 (U (ct , ht )), β < 1 (1)
c,h
t=0

subject to

A
X A
X
(a) (a) (a) (a) (a)
ct + dt (1 + rt ) + (zt bt ) ≤ wt ht + (zt bt−1 (1 + rt )) (2)
a=1 a=1

The budget constraint has three components. At equality, households’ consumption

expenditure, debt repayments and savings over a class of A assets, are balanced by their
contemporaneous wage income, as well as the cash value of all assets held at the start of the
period. In summary, this is the household’s balance sheet.
The utility function is increasing and concave in consumption, whilst decreasing and
concave in labour supply. The former captures non-satiation and diminishing marginal
utility. The latter allows for disutility from working, in lieu of leisure. As work hours
increase, the pace at which disutility rises similarly increases.

U 0 (c) > 0, U 00 (c) < 0 (3)

U 0 (h) < 0, U 00 (c) < 0 (4)

The A classes of assets that households can invest in differ across two aspects. The first is
(a) (a)
risk profile, particularly the variability in prices zt . The interest rate rt offered increases
with the underlying degree of risk, reflecting the higher premium placed by investors to
offset the disutility generated from the uncertainty in price outturns. The second is time-
to-maturity. Returns risks and maturity terms are independent. This yields two concerns
- households’ tolerance of price uncertainty and liquidity shocks. Households invest savings
between safe and risky assets, while considering their liquidity buffers. To generalise, this
decision is defined by the euler equation for these savings instruments, where the marginal
costs and benefits of investing in any asset type a are held at equality.

(a) (a) (a)

U 0 (bt ) = β(1 + rt )Et U 0 (bt+1 ) (5)

The empirical study focuses on the sustained shortfall of output in response to a transitory
shock to uncertainty, without a noticeable temporary overshoot. As such, the theoretical
segment should include an element that maintains expenditure decisions below the baseline.

409
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

To this end, I formally consider that risk preferences may be time-varying and expectations on
returns are informed by the immediate past, rather than the full set of information available
at period t. The latter yields rational inattentiveness, particularly in savings decisions.
Parameters in the savings euler equation can be made time-varying, influencing the expected
marginal costs and benefits of each class of assets.

(a) (a) (a)

U 0 (bt ; γt , λa ) = β(γt )(1 + rt )Et U 0 (bt+1 ; Et γt+1 , λa ) (6)

γt is the risk preference of the household in period t, which determines the allocation of
savings across all A classes of assets. This preference differs across time, and is influenced by
two factors, the degree of economic uncertainty and the household’s expected future income
at the start of period t.

γt = f (Ut , γt−k ; ρ(t)) where ρ(t) < 1 and ρ0 (t) < 0 (7)

A key channel in which uncertainty shocks affect output is through the temporary build-
up of precautionary savings, as reflected by the risk preference parameter entering the subjec-
tive discounting factor. Crucially, the change in risk preference affects not only the allocation
of savings, but the quantum of savings relative to total household budget inflows. I rewrite
the household’s optimisation problem, allowing for this generalisation.
∞ Y
X ∞
max (β(E0 γt ))E0 (U (ct , ht )) , β(γt ) < 1 (8)
c,h
t=0 t=0

subject to

A
X A
X
(a) (a) (a) (a) (a)
ct + dt (1 + rt ) + (zt bt ) ≤ wt ht + (zt bt−1 (1 + rt )) (9)
a=1 a=1

λa characterises the risk profile of the A classes of assets considered. An asset class
a is considered a ’safe/hedge’ asset if returns are inversely related with aggregate output
growth and inflation, and, conversely, a ’risk’ asset if this relationship is positive. This
classification guides the asset allocation decision of households when the risk appetite pa-
rameter γt changes. For simplicity, during risk-on periods, households substitute away from
’safe/hedge’ assets and towards ’risk’ assets. The converse occurs during risk-off periods.
Allocation between the two classes of assets is broadly balanced during risk-neutral periods.

λa = b if κa > 0 and λa = −b if κa < 0, where b > 0 (10)

410
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

(a) (a) (a) (a)

κ(a) = g(cov(gt , zt ), cov(πt , zt )), where g 0 (·) > 0 and g 00 (·) ≤ 0 (11)

Risk preferences form only one part of household response to uncertainty. Both the
quantum and allocation of savings also depend on expected asset value. Rational inattention
can imply a prolonged shortfall in output, similar to the conjectures in Leduc and Liu (2016)
[28], Fajgelbaum, Schaal and Taschereau-Dumouchel (2017) [20], and Ludvigson, Ma and
Ng (2018) [29]. In this strand of the uncertainty literature, the build-up of precautionary
savings and real options effects first lead to a shortfall in economic activity, leading to lower
information flow. Agents then perceive the state of low information as a persistent state
of uncertainty, further delaying investment and consumption. Ultimately, the shortfall in
output is prolonged and uncertainty remain elevated.

(a) (a)
Et (zt+1 bt+1 ) = h(Et (gt+1 ), Et (πt+1 )) (12)

Household expectation of asset value depends on expected economic growth and infla-
tion. Sims (2003) [34] posits that costs arising from processing the full span of information
flow leads to a deviation from the full-information rational expectations behaviour. Sim’s
(2003) [34] seminal paper focuses on a constraint on the rate of entropy of information flow.
whereas I posit a simple representation. Households form expectations using only limited
information from the immediate past k months, rather than the full span of information It .
By assigning larger weights, ρ(t), to the immediate past, post-shock, households treat the
projected economic growth and inflation trajectories to be materially different from the true
long-term average. We can treat ρ(t) as an inverse of the cost of processing or acquiring
information from that past. Adding a contemporaneous dimension on signal-processing costs
may provide greater depth to further work. We can treat this decreasing weight on informa-
tion sets as the revealed preference of the household having considered the signal-processing
cost minimisation problem. Where macroeconomic conditions undershoot expectations due
to unanticipated shocks, agents behave as if the impulse of shocks do not dissipate, leading to
prolonged undershooting of macroeconomic outturns. Expectations on macroeconomic con-
ditions are procyclical here, which then provide justification to use policy interventions aimed
at augmenting expectations when adverse uncertainty shocks materialise. Nevertheless, this
is conceptually generalisable to any adverse macroeconomic shock.

Et (gt+1 ) = Lg (I(t, ..., t − k); ρ(t)), where 0 < ρ(t) < 1 and ρ0 (t) < 0 (13)

Et(πt+1) = Lπ (I(t, ..., t − k); ρ(t)), where 0 < ρ(t) < 1 and ρ0(t) < 0 (14)

411
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

I propose further details on the formation of inflation expectations, drawing from the
bounded expectations literature, particularly that of Ding (2018) [18]. Ding (2018) proposed
a Neo-Keynesian Philips Curve with rationally bounded agents (κ-augmented NKPC), where
inflation expectations are anchored by the central bank’s projection. I weave this principle
with findings from surveys conducted by various central banks, such that inflation expecta-
tions are typically biased upwards. I model households’ expectations of inflation as biased
upwards from the long-term mean, but anchored by the central bank’s forecast. The weight
on the latter’s projection decreases on two counts. The first comes from Ding (2018) [18],
where the central bank’s credibility gravitates expectations formation closer to the rational
expectations case. The second comes from the degree of prevailing economic uncertainty
in period t. Uncertainty erodes the precision of macroeconomic indicators, reducing the
reliability of central bank forecast. As the span of information set considered is rife with
imprecision, agents resort to judgment heuristics. Though not essential, we may generalise
this conjecture to the formation of economic growth expectations.

t,CB
Et (πt+k ) = (π̄ + χt )4t + (1 − 4t )π̂t+k , where 4t ≤ 1 (15)

Formally, the household’s inflation expectation is a weighted average of its own pro-
jection, formed through judgment heuristics with an upward bias χt in period t, and the
central bank’s forecast produced in the same period. 4t is a increasing and convex function
of macroeconomic uncertainty. Uncertainty erodes agents’ reliance on the central bank’s
forecast. This yields an inflation response that overshoots the baseline projection. Absent
of shocks, expectations are anchored by the central bank. Once uncertainty shocks mate-
rialise, the upward bias pushes price-setting behaviour into an acceleration, despite output
moderating.

d4t
>0 (16)
dUt

d2 4t
>0 (17)
dUt2

2.2 Firms
A continuum of {i}∞ 0 firms employ labour Lgit and capital Kgit to produce an amount Ygit
of goods of type g , priced at pgit , in period t. The production function satisfies constant
returns to scale. A workhorse model would be the Cobb-Douglas production function, with

412
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0
Ut

Figure 1: Aggreggate Uncertainty and Agents’ Reliance on Own Projection

labour input shares αgit and firm-goods-level total factor productivity (TFP) Agit .

Ygit = Agit F (Kgit , Lgit ; αgit ) (18)

(K)
ygit = Agit f (lgit ) (19)

(L)
ygit = Agit f (kgit ) (20)

Firms maximise expected profits. Hence, employment and investment, and similarly sep-
aration and dis-investment, decisions depend on profit expectations. Similar to the house-
hold’s problem, this depends on the firms’ expectations for inflation and economic growth.
While rational inattentiveness and bounded expectations similarly apply, parameters may
differ. The degree of behavioural frictions amongst firms may be less binding than amongst
households. Firms may have longer records-keeping and inertia in institutional behaviour.
Hence, the information considered for the average firm may span a longer period than for
the average household.

(D) (D)
{Lgit , Kgit , } = max{Et (Φ)i } (21)
L,K

∞
X E0 (φit )
Φi = (22)
t=0
(1 + rt )t

413
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

From the expected profit equation, written sequentially, employment and investment
decisions are directly dependent on expected prices and output, and, by extension, expected
wages and rent. By requiring the rate at which the information weights for period t decreases
to be smaller amongst firms, we can capture the asymmetry in attentiveness between firms
and households.

G
X (K)
E0 (φit ) = E0 {(pgit Ygit − wgit Lgit − rgit Kgit )} (23)
g=1

Et (gt+1 ) = Lg (I(t, ..., t − k); ρ(t)), where 0 < ρ(t)(i) < 1 and ρ0 (t)(i) < 0 (24)

Et (πt+1 ) = Lπ (I(t, ..., t − k); ρ(t)), where 0 < ρ(t)(i) < 1 and ρ0 (t)(i) < 0 (25)

ρ0 (t)(f irm) > ρ0 (t)(household) (26)

A key deviation in expectations formation amongst firms, notably that of inflation, from
the household is the lack of a systematic upward bias. The empirical literature on firm-
level expectations finds that biases are broadly firm-specific, informed by their experience.
Presence of upward biases, similar to those that of household inflation expectations, is not
uniform. For instance, Richards and Verstraete (2016) [32] finds that whilst upward biases
are found in the Bank of Canada’s Business Outlook Survey (BOS), this is generally absent
amongst firms surveyed in the Business Confidence Survey (BCS). Unsurprisingly, profes-
sional forecasters, with expertise in processing macroeconomic data and likely to utilise such
data at a more intensive rate than the average firm or household, do not exhibit such biases.
Together with the well-established evidence on upward inflation expectation biases amongst
households, this gives rise to an upshoot of inflation away from its steady state in response
to an uncertainty shock.
Uncertainty clouds the reliability of expectations, leading to the build-up of precautionary
savings and the exercise of real options by delaying investment and employment decisions.
Drawing from the literature that discusses uncertainty traps, notably Fajgelbaum, Schaal
and Taschereau-Dumouchel (2017) [20], economic growth first undershoots the steady state
as uncertainty shocks materialise, and firms subsequently take that as the ex post anchor due
to inattentiveness to the full information span. Ultimately, a supposedly transitory downturn
posited, such as Bloom (2009) [11] become prolonged and amplified, akin to that of Leduc
and Liu (2016) [28] and Fajgelbaum, Schaal and Taschereau-Dumouchel (2017) [20].

414
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2.3 Policy
The economy has two policymakers - the treasury (fiscal policy authority) and the central
bank (monetary policy authority). The former decides on public expenditure and taxes,
whilst the latter determines the nominal interest rate through the policy rate. As with
standard theoretical and quantitative macroeconomic , one may assume exogeneity in the
way government expenditure and taxes are determined. However, this framework should be
generalisable to the case of endogenous fiscal policy.
The central bank minimises a quadratic Barro-Gordon loss function, due to Kydland and
Prescott (1977). The central bank suffers value losses Lt that are increasing in the deviation
of inflation πt and the output gap xt from their respective targets, π ∗ andx∗ , in period t.
The relative importance of the inflation target relative to the output gap target, λ, is pre-
determined. In practice, this may be informed by the central bank’s institutional memory,
mandates and preferences of its monetary policy committee.

Lt = λ(πt − π ∗ )2 + (1 − λ)(xt − x∗ )2 , where 0 < λ < 1 (27)

Due to lags in the collection and publication of macroeconomic data by statistical agen-
cies, central banks are unlikely to have access to the necessary span of contemporaneous
data. GDP, industrial production, and likewise major labour market data such as wages,
employment and unemployment, typically follow a substantial lag of up to a quarter. While
inflation data are published with a shorter lag, incomplete information motivates the use
of an expected loss function rather than a realised loss function. Uncertainty that con-
cerns macroeconomic conditions now affects policy decisions, where the contemporaneous
estimates of the output gap and inflation enter the loss function as second moments.

Et (Lt ) = λEt (πt − π ∗ )2 + (1 − λ)Et (xt − x∗ )2 , where 0 < λ < 1 (28)

Rather than solving for the optimal policy rule, which may be subject to a nonstandard
Philips Curve and IS function due to the presence of macro-behavioural frictions outlined in
the previous subsections, I consider a forward-looking Taylor Rule policy response function,
due to Taylor (1993) [35] and Romer and Romer (2004) [33], expressed in terms of inflation
and the output gap. Taking first differences then yields the evolution rule for the nominal
interest rates in response to movements in inflation rate and in the output gap. Adhering to
the Taylor’s Principle requires βπ,t > 1.

it = β0,t + βπ,t Et (πt+1 − π ∗ ) + βx,t Et (xt+1 − x∗ ) + t (29)

415
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4it = βπ,t Et (4πt+1 ) + βx,t Et (4xt+1 ) + 4t (30)

To relate uncertainty with policymaking, I posit that central banks face cognitive limita-
tions on forming contemporaneous expectations, which builds on two observations. Firstly,
macroeconomic data crucial to central banks, such as GDP and industrial production, are
released with noticeable lags. Secondly, Central Banks are generally non-participants, or, at
best, limited participants, in the real economy and financial markets. The ability to access
information is likely less comprehensive that agents deeply integrated within market oper-
ations. In this model, central banks do not know what πt and xt are. Rather, they form
expectations of inflation and the output gap, both of which are subject to estimation er-
rors, contemporaneously. This is not to say that firms and households have an idea of what
aggregate output and prices are contemporaneously. Rather, firms and households know
their respective output and prices. The central bank’s estimates are, amongst all economic
agents, the ’best guesses’ of aggregate economic variables contemporaneously. The central
bank’s Et πt and Et xt may differ from actual outturns. To simplify, I write the central bank’s
projection as a function of uncertainty, which multiplicatively interacts with the projection
error.

Et (πt ) = πt + κ(Ut )π,t (31)

Et (xt ) = xt + κ(Ut )x,t (32)

κ = p(Ut ), where p(0) = 1, p0 (·) < 0, p00 (·) < 0 (33)

Uncertainty adversely affects the precision in evaluating macroeconomic conditions cru-

cial for policy decisions and the efficacy of monetary policy transmission channels. The
latter draws from behavioural changes established in earlier subsections, where agents’ price,
output and expenditure decisions deviate from a baseline where uncertainty is absent. With
higher uncertainty having a multiplicative effect on the precision of inflation and output gap
estimates, we expect the second moment of the deviation of inflation and output gap from
respective targets to rise. Expected loss, therefore, increases with uncertainty, prompting
central banks to react more aggressively to a shortfall in the output gap. Even if the cen-
tral bank possesses full information rational expectations, without any behavioural frictions,
central banks can respond to a shortfall in economic growth and hence the output gap, but
to a different degree of aggressiveness. This differs from an increase in uncertainty in policy

416
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

parameters, drawing from the literature on monetary policy uncertainty, notably Brainard
(1967) [12], where policy reaction become more modest, contrary to the type of uncertainty
here that concerns the first moment estimates in the loss function. Discussion of interac-
tions with policy parameter uncertainty, as well as implications to monetary policy with
quantitative applications may be steps for further research.
In sum, policy has a role in stabilising inflation and the output gap. Importantly, Et (πt )
and Et (xt ), as well as k periods-ahead forecasts, πt+k ˆ and xCB
CB ˆ
t+k may be made known to
the public to anchor expectations. Policy efficacy depends on the central bank’s ability to
influence the reliance of agents on official projections, 4t , when uncertainty rises. Central
d4t
banks can provide a credible anchor to lower . Social welfare may be relative higher if
dUt
the upward bias of households and firms χt is shut off when uncertainty rises, which could
coincide with a shortfall in the output gap, as well as a multiplicative increase in the second
moment terms Et (·)2 in the loss function.

3 Data
This paper uses a panel of macroeconomic data from 16 advanced and emerging economies
— the United States of America (USA), Euro Area (EUR), Japan (JPN), United Kingdom
(GBR), Canada (CAN), PR China (CHN), South Korea (KOR), Chinese Taipei (TWN), Sin-
gapore (SGP), Malaysia (MYS), Thailand (THA), Philippines (PHP), India (IND), Mexico
(MEX), Brazil (BRA) and South Africa (ZAR).
This sample is sufficiently representative of the global economy, as they account for
71.1% of global real GDP in purchasing power parity (PPP) terms in 2018, based on the
IMF’s October 2019 World Economic Outlook (WEO) Database. Moreover, it covers the
major advanced and emerging economies, as well as with additional granularity in East
Asia. All variables used — industrial production (IPI), consumer price index (CPI), 10-
year government bond yields, nominal effective exchange rate (NEER) and economic policy
uncertainty (EPU) — were commonly available between June 2013 and February 2020 at a
monthly frequency.
The variables here are chosen as proxies for specific macroeconomic factors, considering
also their availability at a monthly frequency. IPI proxies for aggregate output, CPI for
prices, NEER for the exchange rate and bond yields for the degree of risk aversion. EPU
proxies for uncertainty. The Malaysia-specific study is conducted on quarterly data spanning
1Q 2001 to 3Q 2019. I used GDP growth as a proxy of output, instead of IPI growth, and
also included employment growth as a measure of labour market conditions.

417
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.1 Uncertainty
The main measure of uncertainty, EPU, was prepared by Baker, Bloom and Davis (2016)
[8]. Country-specific indices were additionally prepared by Baker, Bloom, Davis and Wang
(2013) [7], Arbatli, Davis, Ito and Miake (2019) [6] and Davis (2016) [16]. For economies
without a country-specific index, the global PPP-weighted index was used, prepared by Davis
(2016) [16]. These economies are small open economies, whose economic conditions likely
depend on global developments. All major economies in the sample had respective indices.

3.2 Macroeconomic Data

The main measures used are IPI, CPI, 10-year government bond yields and NEER. All data
is published by the statistical agencies of respective countries, except for the NEER, which
is published by the Bank for International Settlements (BIS). All data were prepared by
respective national authorities, and made available by Haver Analytics and CEIC.
Bond yields underwent year-on-year difference transformation, while the rest of are trans-
formed into year-on-year growth rates. While this addresses non-stationarity issues in the
respective time series, the transformation allows for seasonality adjustments by directly com-
paring the same period from the preceding year. One source of seasonality may be holidays
that fall on the same period every year, such as Lunar New Year for several of the Asian
economies covered. Additionally, this transformation is commonly used by policymakers,
market participants and professional forecasters.
In the Malaysia-specific study, a different set of transformation was applied on the data
set, which spanned a longer period and is available at a quarterly frequency. Additionally,
world GDP growth in this segment is estimated as a weighted average of real GDP growth
in local constant prices of the following economies — United States, Euro Area, United
Kingdom, Japan, Canada, Australia, New Zealand, Switzerland, Sweden, Denmark, Norway,
Singapore, Chinese Taipei, Hong Kong SAR, South Korea, PR China, India, Indonesia,
Philippines, Thailand, Brazil, Russia, Mexico, Colombia, Chile, Czech Republic, Poland,
Romania, South Africa and Turkey. Malaysia is excluded from this measure. These countries
accounted for 83.1% of global real GDP in PPP terms in 2018, as indicated in the IMF’s
October 2019 WEO. Country weights are taken as respective shares of global GDP in PPP
terms from the same issue of the IMF’s WEO.

418
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4 Empirical Strategy
Five stages of empirical analysis is conducted, each to address a variant of the question
outlined in section 1 — ‘what are the dynamics between uncertainty, the exchange rate and
aggregate output?’ Five strategies will be followed.
Firstly, the country-level max-uncertainty VAR (MUVAR), following Jackson, Kliesen
and Owyang (2020) [23] sheds light on the dynamics between output, uncertainty and the
exchange rate specifically for Malaysia, accounting for nonlinear uncertainty effects during
upward jumps in uncertainty. The later analyses trade this detail off for country coverage,
and against the wider global economy, in a panel data setting.
Secondly, I estimate a panel VAR, with GMM-style instruments, following Abrigo and
Love (2016) [1]. In this case, the paper focuses on the impulse response functions from EPU
and NEER shocks, structurally estimated via Cholesky Decomposition. Specifically, these
are the estimated average within-country dynamics. As this is akin to a pooled estimator,
which estimates the average within-country effects, I have also estimated a Bayesian version
of the model.
Thirdly, the data is fitted to a Bayesian Hierarchical Panel VAR, which allows for cross-
sectional heterogeneity, such that estimated country-specific dynamics differ from sample
‘within’ estimates, as per Canova and Ciccarelli (2006) [13], due to Jarocinski (2010) [24]
and documented in Dieppe, Legrand and van Roye (2016) [17]. Compared to the second
approach, this approach allows for heterogeneous VAR estimates and estimated dynamics
between countries. The Gibbs Sampling formulation further overcomes restrictions from
limited degrees of freedom posed by the data set.
The fourth approach departs from the structural analysis in the previous approaches.
I analyse the nonlinear relationship between economic uncertainty, the exchange rate and
other macroeconomic variables in a fixed effects setting. Additionally, I apply a judgment-
free approach to determine any nonlinearities implied by the data by estimating a panel
regression augmented with the Least Absolute Shrinkage Selection Operator (LASSO), fol-
lowing Tibshirani (1996) [36] and Ahrens, Hansen and Schaffer (2019) [4]. The sensitivity
parameter is estimated through a k-fold cross validation.
Finally, I estimate a panel quantile regression model, as per Machado and Silva (2018)
[31]. This borrows from Adrian, Boyarchenko and Giannone (2019) [2]. Here, the research
question is viewed from a distributional dimension. I examine if there are material differences
in the relationship between uncertainty, the exchange rate and output across various points
of the output distribution.

419
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4.1 Country-Level VAR

The Max-Uncertainty VAR is adapted from Jackson, Kliesen and Owyang (2019)[23]. Yt
contains real GDP growth in period t. Xt is a vector of CPI inflation, NEER growth,
M2 growth, annual difference of 10-year government bond yield in period t, and global
PPP-weighted real GDP growth. Zt is the measure of uncertainty — the global PPP-
weighted EPU. Finally, Ẑt−1 is the max-uncertainty measure, which is the percentage change
of uncertainty relative to its highest level in the past 12 months. Ẑt−1 takes the value of zero
if this change is nonpositive. Additionally, sign restrictions are imposed on the Malaysian
variables in the global equations to reflect Malaysia’s position as a small open economy.
Shocks from the domestic economy should not affect the global economy. In the global GDP
and EPU equations, the coefficients of Malaysian variables are restricted to zero.
        
Yt β11 β11 β11 Yt−1 γyz Y,t
Xt  = β21 β22 β23  Xt−1  + γxz  Ẑt−1 + X,t  (34)
        

Zt β31 β32 β33 Zt−1 0 Z,t

100 ∗ (Zt−1 − max{Zt−2 , ..., Zt−13 })

Ẑt−1 = max{0, } (35)
max{Zt−2 , ..., Zt−13 }
The merits of this approach is of two-fold. Firstly, we distinguish between an elevation
in uncertainty and spikes in uncertainty that are ‘abnormal’ relative to the near-term. A
12-month period could reflect possible rational inattention of agents who may have limited
retrospective vision. However, this choice of backward visibility may be arbitrary and could
vary. This representation is a simplification of Sims (2003)’s [34] principles, where agents face
signal-processing costs in consuming the span of informational set available to them, leading
to inefficient consumption of information and, hence, producing decisions that deviate from
the rational expectations case. Kamdar (2019) [27] proposes a suite of structural models,
augmenting workhorse models with these behavioural properties. We may interpret the limit
on the backward-looking abilities of agents, as well as the decreasing weight of longer lags, as
the revealed preference of the information-processing problem faced by the respective agents.
Secondly, the convenient structure of the max-uncertainty VAR allows for a linear estimation,
as well as standard computation of the IRFs, allowing for a simple, non-parametric and data-
driven procedure.
Nevertheless, I note two key shortcomings, in reference to the theoretical underpinnings
of this paper’s analysis. Firstly, while allowing for substantial jumps in uncertainty to have
an added effect, parameters of the model remain linear. Secondly, similar to a regular VAR,
responses to a downward shock in uncertainty are treated as symmetric.

420
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4.2 Panel VAR

Taking the estimation exercise to a panel of 16 economies outlined in section 3, the following
model estimates the average within-country dynamics of IPI growth, CPI inflation, M2
growth, annual difference of bond yields, NEER growth and uncertainty growth. The reduced
form within-effect estimates β are estimated with GMM-style instruments, as per Abrigo and
Love (2016) [1]. Impulse responses are structurally estimated with Cholesky Decomposition,
which are reported in section 5.2.
In this model, Yit is the IPI growth of country i in month t, Xit the vector of CPI inflation,
M2 growth and bond yields difference, and Zit the vector EPU and NEER growth, the shock
variables of interest. αiY , αiX and αiZ are endogenous variable-specific fixed effects for country
i. All variables in the model here are treated as endogeneous. The fixed effects control for
variable-country-specific heterogeneity that are time-invariant but unaccounted for in the
system of macroeconomic variables in the reduced form VAR.
   Y     Y
Yit αi β11 β11 β11 Yi,t−1 it
   X    X
Xit  = αi  + β21 β22 β23  Xi,t−1  + it  (36)


Zit αiZ β31 β32 β33 Zi,t−1 Zit

4.3 Bayesian Panel VAR

In this segment, I estimate two versions — (i) pooled estimator, as documented in Dieppe,
Legrand and van Roye (2016) [17], and (ii) a hierarchical model, due to Jarocinski (2010)
[24], with impulse response functions structurally estimated with sign restrictions.
While not covered here, further work may consider the structural factor models in Canova
and Ciccarelli (2013) [14], which allows also for static and dynamic interdependencies. The
reduced form errors are correlated across units, capturing static [Link]
from the lagged endogenous variables of all countries i 6= j affect variables in country i
equation, capturing dynamic interdependencies. A model with less macroeconomic variables,
but one of wider coverage may be useful in characterising more intricate global linkages.
The Bayesian hierarchical panel VAR extends the GMM-style model in section 4.2 by
allowing for cross-sectional heterogeneity. Instead of β, we estimate βi for each country i,
as per Jarocinski (2010) [24]. Section 5.3 reports the impulse response functions of interest.
Compared to the static GMM-styled panel VAR, we can extract country-specific responses
to their own shocks, providing a richer characterisation of heterogeneous behaviour in the

421
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

global economy in response to uncertainty and exchange rate shocks.

   Y     Y
Yit αi βi,11 βi,11 βi,11 Yi,t−1 it
   X    X
Xit  = αi  + βi,21 βi,22 βi,23  Xi,t−1  + it  (37)


Zit αiZ βi,31 βi,32 βi,33 Zi,t−1 Zit

4.4 Fixed Effects Regression

This section deviates from macroeconomic structural analysis. Specifically, we are interested
in the possible heterogeneous relationship between output, prices, risk aversion, the exchange
rate and uncertainty. This approach entails two steps — (i) model selection with LASSO
and (ii) estimation of the selected fixed effects ‘within’ model.
Firstly, I implemented the LASSO on a fixed effects model, as per Tibshirani (1996)
[36] and Ahrens, Hansen and Schaffer (2019) [4], with a triple interaction of risk aversion,
uncertainty and the exchange rate. This shrinks the coefficient estimates β towards zero. By
design, variables that are estimated to be close to zero in the fixed effect model are dropped
(zero-rised). The purpose of this segment is to drop noise interaction terms.

n p
1X 0 2
X
min (yi − xi β) + λ |βj | (38)
n i=1 j=1

Secondly, I estimate the fixed effects model with interaction terms selected in step one.
Yit is the IPI growth for country i in month t, Xit the vector of selected variables amongst
bond yields and inflation, and Zit that of NEER and EPU growth. Xi t · Zit is the vector of
selected interaction terms. it is the unexplained variation in IPI growth.

Yit = αi + Xit β1 + Zit β2 + Xit · Zit β3 + it (39)

4.5 Panel Quantile Regression

Finally, to add a distributional dimension to the analysis, which is lacking in the previous
four sections, I estimate a panel quantile regression model of the same variables chosen by
the LASSO in section 4.4. We are interested in βkq , the average within-country effect of
variable k on IPI growth for country at quantile q of IPI growth, as per Machado and Silva
(2018) [31].

Yit = αiq + Xit β1q + Zit β2q + Xit · Zit β3q + it (40)

422
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5 Empirical Findings
5.1 Country-Level Max-Uncertainty VAR
Zooming into Malaysia, the estimated impulse response functions point to findings that are
broadly in line with the wider literature. Aggregate output growth moderates in response to
an unanticipated increase in economic policy uncertainty for up to 2-years, with a correction
absent. Inflation moderates slightly but persistently. At this juncture, this finding deviates
from the theoretical model’s conjecture, and will be analysed together in section 5.2 with
the panel VAR findings. The exchange rate tends towards depreciation in a persistent
manner. In line with the negative output response, the response of employment growth is
similarly negative. Risk premia, as measured by the 10-year yield spread, relative to 10-year
US treasury yields, closes, suggesting within-country flight-to-safety. Kamber, Karagedikli,
Ryan and Vehbi (2016) [26] finds that bond yields in advanced economies rise in response to
uncertainty shocks. For Malaysia, an EME, the opposite response may be expected.

423
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Responses to a 1 S.D. Shock in Global Economic Policy Uncertainty Growth

0.2
0
Real GDP Growth

CPI Inflation
−0.2
−0.5
−0.4

−1 −0.6

0 2 4 6 8 10 12 0 2 4 6 8 10 12
Quarters Quarters

Figure 2: Real GDP Growth Figure 3: CPI Inflation

0.4
Employment Growth

0.2 10Y Bond Yields

0
0

−0.2
−0.1
−0.4

−0.6
−0.2
−0.8
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Quarters Quarters

Figure 4: Employment Growth Figure 5: 10-Year Bond Yields

0.5
World Real GDP Growth

0
0
NEER Growth

−0.2
−0.5
−0.4
−1

−1.5 −0.6

−2
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Quarters Quarters

Figure 6: NEER Growth Figure 7: World Real GDP Growth

424
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

30
Change in EPU

0 2 4 6 8 10 12
Quarters

Figure 8: Change in World EPU

In response to an upward (appreciation) shock in the exchange rate, output moderates,
as expected for a trade-dependent small open economy. Inflation moderates, likely due to
cheaper imports. Bond yields fall, as prices of portfolio assets increase. Viewed against other
economies, a higher valued exchange rate may attract portfolio investors seeking to retain
value, hence the higher demand and subsequently lower yields in Malaysian bonds.
While the upward response in employment growth may not be expected, this may arise
from compositional effects, as more labour-intensive, less trade-dependent, sectors may ex-
perience real income windfalls from the stronger exchange rate, leading to stronger labour
demand. The net employment gains can be reconciled with a net output loss if sectors that
benefited from a stronger exchange rate, namely the labour-intensive non-tradable sectors,
are less productive than capital-intensive tradable sectors, such as manufacturing. This set
of estimates for Malaysia suggest that employment gains in non-tradable sectors outpace
that of employment losses in tradable sectors, the productivity gap is sufficiently large such
that the net output response is negative in the near-term. To corroborate, Benigno, Fornaro
and Wolf (2020) [9] argues that capital flows into the US, which strengthened the USD, had
prompted an faster expansion in the less-productive services sector relative to manufactur-
ing, hence dampening global productivity growth. Using detailed firm-level data spanning
1992-2008 in a ‘natural’ experiment — Hungary’s full capital account liberalisation in 2001
ahead of its ascension to the European Union (EU), — Saffie, Varela and Yi (2020) [19] found
faster expansion in the services sector, whose estimated real productivity is lower as access
to international capital markets is enabled. In these sectors, relative to the manufacturing
sector whose real productivity is estimated to be higher, the number of firms also expanded
faster, aligned with the estimated net employment gains in this paper.

Responses to a 1 S.D. Shock in NEER Growth

425
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0.2
0.1
0.1
Real GDP Growth

CPI Inflation
0

−0.1 −0.1

−0.2
−0.2
−0.3

0 2 4 6 8 10 12 0 2 4 6 8 10 12
Quarters Quarters

Figure 9: Real GDP Growth Figure 10: CPI Inflation

5 · 10−2
0.6
Employment Growth

10Y Bond Yields

0.4 0

0.2
−5 · 10−2
0
−0.1
−0.2
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Quarters Quarters

Figure 11: Employment Growth Figure 12: 10-Year Bond Yields

3 1
World Real GDP Growth

0.5
NEER Growth

0
1

−0.5
0

0 2 4 6 8 10 12 0 2 4 6 8 10 12
Quarters Quarters

Figure 13: NEER Growth Figure 14: World Real GDP Growth

426
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0.5
Change in EPU

−0.5

0 2 4 6 8 10 12
Quarters

Figure 15: Change in World EPU

5.2 Panel VAR

Output contracts in the short-run in response to a positive shock in uncertainty growth,
before normalising after approximately 2 years. However, there is no positive correction
in the aftermath, in contrast to earlier country-specifc studies, such as Bloom (2009) [11].
This may be due to (i) the different transformation conducted to handle nonstationarity in
the levels of some country-specific EPU series, and (ii) the differential dynamics of output
and uncertainty in the later periods of the 2010s, where the extant literature focuses on the
pre-GFC period and the early half of the 2010s.
Inflation moderates persistently, nearly converging to its steady state only after four years.
This contrasts the theoretical mechanism proposed, where the upward bias of household
expectations dominate during times of uncertainty. The sign here suggests that the aggregate
demand effect, outlined in Leduc and Liu (2016) [28], may dominate. Output and prices move
in the same direction.
Bond yields accelerate, suggesting stronger risk aversion behaviour on average across the
panel of countries. The overall findings reflects that of Kamber, Karagedikli, Ryan and
Vehbi (2016) [26], where bond yields were estimated to rise in response to an uncertainty
shock. Finally, the effective exchange rate depreciates, in line with the rise in domestic
money supply.

Responses to a 1 S.D. Shock in Economic Policy Uncertainty Growth

427
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 16: IPI Growth Figure 17: CPI Inflation

Figure 18: 10-Year Bond Yields Figure 19: NEER Growth

Figure 20: EPU Growth

Output moderates persistently in response to an upward exchange rate shock (unantic-
ipated appreciation). The response of NEER growth to its own shock suggest a slow con-
vergence in prices relative to that of trade partners over a period of 2 years. As a stronger
exchange rate raises purchasing power relative to other countries, and reduces import prices,
inflation moderates persistently. In line with this conjecture, money supply slows gradually
but persistently. Likewise, bond yields slow persistently. as prices of portfolio bond assets

428
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

would now be more expensive in terms of the trade partners’ basket of currencies. EPU
experiences a minor fall, likely as the news of higher purchasing power for domestic agents
are treated as positive economic news, hence reducing the prevailing level of uncertainty
modestly, although temporarily before correcting.

Responses to a 1 S.D. Shock in NEER Growth

Figure 21: IPI Growth Figure 22: CPI Inflation

Figure 23: 10-Year Bond Yields Figure 24: NEER Growth

Figure 25: EPU Growth

429
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.3 Bayesian Panel VAR

5.3.1 Pooled Model

The Bayesian pooled estimates are similar to that of the GMM-style panel VAR, except for
the response of bond yields and inflation to uncertainty shocks. Responses to exchange rate
shocks are similar to that of section 5.2.
Where the earlier estimates showed a temporary acceleration in yields, the estimates here
suggest yields to moderate before converging back to its steady state after approximately 2
years. This suits the prediction of the model in section 2, where risk aversion and flight-to-
safety behaviour raises the demand and subsequently prices of safe assets, such as long-term
sovereign bonds, leading to lower yields in the near-term.
Inflation is estimated to accelerate here, fitting the prediction of the theoretical model.
While I make no claim here on the driver of this behaviour, one possible channel is macro-
behaviourial. Uncertainty dislodges inflation expectations, leading to a surge in price as
the upward biases in households expectations feed into aggregate price-setting behaviour.
Alternatively, this can be corroborated by that the exchange rate faces depreciation pressures,
which raises import prices, and hence accelerating inflation.

Responses to a 1 S.D. Shock in EPU Growth

·10−2
0.1

0 4
IPI Growth

CPI Inflation

−0.1 2

−0.2 0

−0.3 −2

0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 26: IPI Growth Figure 27: CPI Inflation

430
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

·10−2
0.1
2
NEER Growth

Yields Change
0
−0.1

−0.2 −2

−0.3 −4
0 10 20 30 40 50
0 10 20 30 40 50
Months
Months

Figure 28: NEER Growth Figure 29: 10-Year Bond Yields Change

40
EPU Growth

0
0 10 20 30 40 50
Months

Figure 30: EPU Growth

431
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Responses to a 1 S.D. Shock in NEER Growth

0.2 0

−5 · 10−2
0.1

CPI Inflation
IPI Growth

−0.1
0
−0.15

−0.1 −0.2

−0.25
−0.2
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 31: IPI Growth Figure 32: CPI Inflation

1 0
−2
0 −2 · 10
Yields Change
EPU Growth

−1 −4 · 10−2

−6 · 10−2
−2
−8 · 10−2
−3
−0.1
−4
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 33: EPU Growth Figure 34: 10-Year Bond Yields Change

2
NEER Growth

0
0 10 20 30 40 50
Months

Figure 35: NEER Growth

5.3.2 Hierarchical Model

As the hierarchical random effects model allows for cross-sectional heterogeneity in responses,
such the country-specific responses may differ, the section here will report the output re-
sponses to uncertainty and exchange rate shocks for all 16 economies. The remainder of the

432
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

responses, which is similar qualitatively to the pooled model, will be shown separately.
While the magnitude and degree of persistence differs across countries, uncertainty shocks
generate a dip in output growth in the near-term, similar to estimates from the pooled model.
Corrections are absent in the responses of all countries, except for the US. Exchange rate
shocks generate heterogeneous output responses across economies. This may be in part due
to varying trade intensity, in which aggregate output then responds differentially, or the ex-
pected exchange rate policy responses. Economies such as Korea, Singapore, Philippines and
the eurozone experience weaker output growth in the event of an unanticipated appreciation
shock, as exports become more expensive. On the other hand, Japan, Canada and Mexico
benefit from a stronger exchange rate. Stronger domestic purchasing power translating into
higher demand for goods and services may be able to explain the response. Finally, J-curve
responses were estimated for the United Kingdom, Thailand, Malaysia and India. Stronger
purchasing domestic power may first dominate, but is eventually offset by adjustments in
export demand.

Responses to a 1 S.D. Shock in EPU Growth

0.4
0
0.2
IPI Growth

IPI Growth

0 −0.5

−0.2
−1
−0.4

−0.6
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 36: USA Figure 37: EUR

0
0
IPI Growth

IPI Growth

−1

−0.5
−2

−1
−3
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 38: JPN Figure 39: GBR

433
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0
0
−0.5
IPI Growth

IPI Growth
−0.5 −1

−1.5
−1
−2

0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 40: CAN Figure 41: CHN

0 0

−1 −1
IPI Growth

−2 IPI Growth −2

−3 −3

−4 −4

0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 42: KOR Figure 43: TWN

0
0
IPI Growth

IPI Growth

−2 −0.5

−1
−4
−1.5

−6 −2
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 44: SGP Figure 45: MYS

434
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0
0

−0.5
IPI Growth

IPI Growth
−2
−1
−4
−1.5

−2 −6
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 46: THA Figure 47: PHP

0
0
IPI Growth

−1 IPI Growth
−1

−2
−2

−3
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 48: IND Figure 49: MEX

1
0
0
IPI Growth
IPI Growth

−1 −1
−2
−2
−3

−4
−3
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 50: BRA Figure 51: ZAR

435
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Responses to a 1 S.D. Shock in NEER Growth

95% confidence intervals omitted to illustrate clearer the IRFs

·10−2 ·10−2

2 0
IPI Growth

IPI Growth
1.5
−2
1
−4
0.5
−6
0
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 52: USA Figure 53: EUR

·10−2
0.15
0
0.1
IPI Growth

IPI Growth

−1
5 · 10−2

−2
0

0 10 20 30 40 50
0 10 20 30 40 50
Months
Months

Figure 54: JPN Figure 55: GBR

·10−2 ·10−2
4
8
3
6
IPI Growth

IPI Growth

2 4

1 2

0
0
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 56: CAN Figure 57: CHN

436
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

·10−2
0

−2 · 10−2
4
IPI Growth

IPI Growth
−4 · 10−2

−6 · 10−2 2
−2
−8 · 10

−0.1 0

0 10 20 30 40 50
0 10 20 30 40 50
Months
Months

Figure 58: KOR Figure 59: TWN

·10−2 ·10−3

5
0
IPI Growth

IPI Growth

−1 0

−2 −5

0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 60: SGP Figure 61: MYS

·10−2 ·10−2
1
4
0.5
IPI Growth

IPI Growth

0 2

−0.5
0
−1
−2
−1.5
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 62: THA Figure 63: PHP

437
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

·10−2 ·10−2
6
1
4
IPI Growth

IPI Growth
0
2
−1
0

−2
−2
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 64: IND Figure 65: MEX

·10−2 ·10−2

0
0
IPI Growth

IPI Growth
−2 −1

−2
−4
−3
−6
−4
0 10 20 30 40 50 0 10 20 30 40 50
Months Months

Figure 66: BRA Figure 67: ZAR

5.4 Fixed Effects Regression

Table 1 column 1 shows the LASSO estimates of the selected variables. Not shown here
are the dropped interactions between NEER growth, EPU growth and the 10-year bond
yields, each proxies for the exchange rate, uncertainty and risk aversion, drawing from the
theoretical framework in section 2.
In line with the previous sections, uncertainty has a negative relationship with output
growth. Corroborating findings in section 5.3, NEER has a positive relationship with output
growth, likely capturing the initial stages of the inverted J-curve effect. Additionally, the
interaction effect between the exchange rate and bond yields is estimated to be a small and
statistically insignificant negative. This suggests that the positive relationship between the
exchange rate and output is drawn smaller and towards negative territory as bond yields rise
faster, likely during periods of rising risk aversion. We cannot take these results as causal.
These findings should be interpreted only as statistical co-movements as implied by the
data. Nevertheless, this may motivate further structural or quantitative analysis that ac-

438
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

(LASSO) (FE)
VARIABLES IPI IPI

Inflation -0.108
(0.300)
10-Year Bond Yields 0.410 0.595
(0.505)
NEER 0.0102
(0.0546)
EPU -0.00264 -0.00401
(0.00393)
10-Year Bond Yields * NEER 0.0359 0.0447**
(0.0159)

10-Year Bond Yields * EPU

NEER * EPU

10-Year Bond Yields * EPU * NEER

Observations 1,296 1,296

Number of countries 16 16
FE Yes Yes
R-squared 0.021
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

Table 1: LASSO and Fixed Effects Model Estimates

count for nonlinearities along the dimensions of risk aversion and the exchange rate in an
environment with economic uncertainty.

5.5 Panel Quantile Regression

Next, building on the model selected in section 5.4, table 2 reports the quantile regression
estimates. Figures 68 and 69 show the estimated coefficient on EPU growth and NEER
growth, for the case in which the annual change of bond yields is zero, respectively, along
quantiles of IPI and their 95% confidence intervals. Coefficient estimates on inflation and
bond yields are reported in figures 70 and 71.
Of interest here are the coefficients on EPU and NEER; other variables are included

439
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

as controls but are nevertheless reported. The estimates here suggest that uncertainty is
negatively associated with output along the entire distribution of output. However, the
point estimate is larger at lower quantiles than at higher quantiles. The relationship between
the exchange rate and output is positive up to the 70th percentile of output growth. The
quantile estimates decrease along the distribution. At the highest quantile, the estimate
is negative. The interaction term between NEER and bond yields is statistically different
from zero except for the upper quantiles. Extending the findings from section 5.4, future
structural analysis may consider distributional dimensions of output dynamics in response
to uncertainty and exchange rate shocks.

Figure 68: Quantile coefficient estimates on EPU growth

Figure 69: Quantile coefficient estimates on NEER growth (Bond yields change = 0)

440
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 70: Quantile coefficient estimates on CPI inflation

Figure 71: Quantile coefficient estimates on bond yields change

441
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19)
32nd EBES Conference Proceedings - Volume I

VARIABLES 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

Inflation -0.251 -0.217 -0.195 -0.181 -0.167 -0.152 -0.137 -0.124 -0.113 -0.102 -0.0930 -0.0829 -0.0732 -0.0629 -0.0505 -0.0360 -0.0188 0.00161 0.0341
(0.345) (0.283) (0.246) (0.222) (0.200) (0.180) (0.163) (0.151) (0.144) (0.141) (0.140) (0.143) (0.148) (0.157) (0.170) (0.188) (0.213) (0.246) (0.304)
10-Year Bond Yields 1.123** 0.997** 0.918*** 0.866*** 0.814*** 0.758*** 0.704*** 0.656*** 0.616*** 0.574*** 0.541*** 0.504** 0.469** 0.430* 0.385 0.332 0.268 0.193 0.0736
(0.488) (0.399) (0.347) (0.314) (0.283) (0.254) (0.230) (0.213) (0.204) (0.199) (0.199) (0.202) (0.210) (0.222) (0.240) (0.266) (0.301) (0.348) (0.429)
NEER 0.0358 0.0297 0.0259 0.0234 0.0208 0.0181 0.0155 0.0132 0.0112 0.00919 0.00760 0.00581 0.00408 0.00222 2.88e-06 -0.00258 -0.00566 -0.00930 -0.0151
(0.0479) (0.0392) (0.0341) (0.0308) (0.0278) (0.0249) (0.0225) (0.0209) (0.0200) (0.0195) (0.0195) (0.0199) (0.0206) (0.0218) (0.0236) (0.0261) (0.0296) (0.0342) (0.0421)
EPU -0.00552 -0.00516 -0.00493 -0.00478 -0.00463 -0.00447 -0.00432 -0.00418* -0.00407* -0.00394* -0.00385* -0.00374 -0.00364 -0.00353 -0.00340 -0.00325 -0.00306 -0.00285 -0.00251
(0.00568) (0.00465) (0.00404) (0.00365) (0.00330) (0.00295) (0.00267) (0.00248) (0.00237) (0.00231) (0.00231) (0.00235) (0.00244) (0.00258) (0.00279) (0.00309) (0.00351) (0.00405) (0.00499)

442
10-Year Bond Yields * NEER 0.0770 0.0693* 0.0645* 0.0612** 0.0581** 0.0546** 0.0514** 0.0484** 0.0460** 0.0434** 0.0414** 0.0392** 0.0370* 0.0346 0.0319 0.0286 0.0247 0.0201 0.0128
(0.0470) (0.0384) (0.0334) (0.0302) (0.0273) (0.0244) (0.0221) (0.0205) (0.0196) (0.0191) (0.0191) (0.0195) (0.0202) (0.0213) (0.0231) (0.0256) (0.0290) (0.0335) (0.0413)

Observations 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296 1,296
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

Table 2: Panel Quantile Regression Estimates

August 5-7, 2020
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

6 Conclusion
Building on the growing literature on uncertainty and the macroeconomy, this paper offers
two contributions.
Firstly, this paper proposed a conceptual framework to analytically view uncertainty
shocks. Macro-behavioural frictions, specifically agents with rational inattention and bounded
expectations, produce a prolonged slowdown in output, as unanticipated uncertainty shocks
materialise. Moreover, the model indicates that central banks could employ communication-
based unconventional monetary policy to anchor expectations when aggregate uncertainty
rises, thereby providing countercyclical balances against output shortfalls. This framework
appends the literature on uncertainty through a macro-behavioural lens, whereas the extant
literature mainly views uncertainty in a neokeynesian framework, such as Leduc and Liu
(2016) [28] and Jurado, Ludvigson and Ng (2015) [25] that characterises uncertainty shocks
akin to demand shocks. From this paper’s perspective, uncertainty shocks generate output
responses through behavioural frictions, and also allows for policy responses that exploit
these frictions to generate countercyclical responses.
Secondly, this paper provides an empirical analysis on the impact of uncertainty and
exchange rate shocks in open economies.
The country-level exercise, focusing on Malaysia, in a max-uncertainty VAR framework
due to Jackson, Kliesen and Owyang (2019) [23], finds that output, price and bond yield
responses are broadly in line with the extant literature. Specifically, output and inflation
moderate while yields fall, suggesting flight-to-safety in response to an uncertainty shock. In
response to an exchange rate appreciation shock, output moderates, an expected response as
exports become more expensive. Inflation moderates, reflecting likely cheaper imports from
a stronger exchange rate. Bond yields fall, likely an implication of higher priced portfolio
assets relative to that of other economies. Net employment gains, together with, net output
losses indicate heterogeneity at the sectoral-level. Low productivity labour-intensive non-
tradable sectors experience a boom from positive real income effects, prompting a larger
expansion in labour hiring, but output gains here are offset by output losses in the higher
productivity capital-intensive tradable sectors.
Taking the analysis to a panel of advanced and emerging economies, I find similar re-
sponses in a Bayesian panel VAR setting. The estimated impulse responses differ somewhat
from the extant literature, likely due to two things. Firstly, key variables had undergone
different transformation to overcome country-specific non-stationarity, which is likely absent
in other country-level empirical studies. Secondly, the period covered focuses heavily on the
later half of the 2010s, whereas the bulk of the uncertainty literature focuses on the pre-GFC

443
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

period and the early half of the 2010s. Dynamics may differ over the two periods.
Departing from time series methods, I took the data to a fixed effects setting to examine
possible nonlinearities in the relationship between the exchange rate, uncertainty and out-
put. The choice of interaction terms was selected with a shrinkage algorithm, the L1 norm —
the LASSO, as per Tibshirani (1996) [36] and Ahrens, Hansen and Schaffen (2019) [4]. The
model motivates the possibility of nonlinearities between risk aversion, the exchange rate and
output. Finally, I introduced a distributional dimension to the analysis in a panel quantile
regression setting, using the method in Machado and Silva (2018) [31]. While the interpreta-
tion of the findings were limited to statistical co-movements and have neither structural nor
causal implications, they motivate further analysis to examine the distributional dimension
of output, uncertainty and exchange rate dynamics.

References
[1] Abrigo, M. R. M., and Love, I. (2016), ‘Estimation of panel vector autoregression in
Stata’, The Stata Journal, Vol. 16 (3), pp. 778-804.

[2] Adrian, T., Boyarchenko, N., and Giannone, D. (2019), ‘Vulnerable growth’, American
Economic Review, Vol. 109 (4), pp. 1263-1289.

[3] Ahir, H., Bloom, N., and Furceri, D. (2018), ‘The world uncertainty index’, mimeo.

[4] Ahrens, A., Hansen, C. B., and Schaffer, M. E. (2019), ‘lassopack: model selection and
prediction in regularized regression in stata’, IZA Discussion Paper, No. 12081.

[5] Alessandri, P. and Mumtaz, H. (2019), ‘Financial regimes and uncertainty shocks’, Jour-
nal of Monetary Economics, Vol. 101, pp. 31-46.

[6] Arbatli, E., Davis, S. J., Ito, A., and Miake, N. (forthcoming), ‘Policy uncertainty in
Japan’, The Economic Challenges of Japan’s Aging and Shrinking Population, Interna-
tional Monetary Fund.

[7] Baker, S. R., Bloom, N., Davis, S. J., and Wang, X. (2013), ‘Economic policy uncertainty
in China’, Working Paper.

[8] Baker, S. R., Bloom, N., and Davis, S. J. (2016), ‘Measuring economic policy uncertainty’,
The Quarterly Journal of Economics, Vol. 131 (4), pp. 1593-1636.

[9] Benigno, G., Fornaro, L., and Wolf, M. (2020), ‘The global financial resource curse’,
Federal Reserve Bank of New York Staff Reports, No. 915, February 2020.

444
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[10] Bloom, N. (2000), ‘The real options effect of uncertainty on investment and labour
demand’, The Institute for Fiscal Studies Working Papers, Nov. 2000.

[11] Bloom, N. (2009), ‘The impact of uncertainty shocks’, Econometrica, Vol. 77 (3), pp.
623-685.

[12] Brainard, W. C. (1967), ‘Uncertainty and the Effectiveness of Policy’, The American
Economic Review, 57(2), pp. 441-425.

[13] Canova, F., and Ciccarelli, M. (2006), ‘Estimating multi-country VAR models’, ECB
Working Papers, No. 603.

[14] Canova, F., and Ciccarelli, M. (2006), ‘Panel vector autoregressive models: a survey’,
ECB Working Papers, No. 1507.

[15] Cravino, J., and Levchenko A. A. (2017), ‘The distributional consequences of large
devaluations’, American Economic Review, Vol. 107(11), pp. 3477-3509.

[16] Davis, S. J. (2016), ‘An index of global economic policy uncertainty’, Macroeconomic
Review, October.

[17] Dieppe, A., Legrand, R., and van Roye, B. (2016), ‘The BEAR toolbox’, ECB Working
Paper, 1934.

[18] Ding, S. (2018), ‘Bounded rationality in the rules of price adjustments and the phillips
curve’, pp. 1-53.

[19] Saffie, F., Varela, L., and Yi, K. (2020). ‘The micro and macro dynamics of capital
flows’, National Bureau of Economic Research (NBER) Working Paper No. 27371.

[20] Fajgelbaum, P.D., Schaal, E., and Taschereau-Dumouchel, M. (2017), ‘Uncertainty

Traps’, The Quarterly Journal of Economics, pp. 1641-1692.

[21] Golob, J. E. (1994), ‘Does inflation uncertainty increase with inflation’, Federal Reserve
Bank of Kansas Economic Review, 3Q 1994, pp. 27-38.

[22] Hamilton, J.D. (1996), ‘This is what happened to the oil price-macroeconomy relation-
ship’, Journal of Monetary Economics, 38, 2, pp. 127-157.

[23] Jackson, L. E., Kliesen, K. L., and Owyang, M. T. (2019), ‘The nonlinear effects of
uncertainty shocks’, Federal Reserve Bank of St. Louis Working Papers, Oct. 2019.

445
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[24] Jarocinski, M. (2010). ‘Conditional forecasts and uncertainty about forecast revisions
in vectorautoregressions’. Economics Letters, 108(3), pp. 257–259.

[25] Jurado, K., Ludvigson, S. C., and Ng, S. (2015), ‘Measuring uncertainty’, The American
Economic Review, Vol. 105 (3), pp. 1177-1216.

[26] Kamber, G., Karagedikli, O., Ryan, M., and Vehbi, T. (2016), ‘International spill-
overs of uncertainty shocks: evidence from a FAVAR’, Centre for Applied Macroeconomic
Analysis Working Paper.

[27] Kamdar, R. (2019), ‘The inattentive consumer: sentiment and expectations’.

[28] Leduc, S., and Liu, Z. (2016), ‘Uncertainty shocks are aggregate demand shocks’, Jour-
nal of Monetary Economics, 82, pp.20-35.

[29] Ludvigson, S. C., Ma, S., and Ng, S. (Forthcoming), ‘Uncertainty and business cycles:
exogenous impulse or endogenous response?’, American Economic Journal: Macroeco-
nomics.

[30] Lugilde, A., Bande, R., and Riveiro, D. (2017), ‘Precautional saving: a review of the
theory and the evidence’.

[31] Machado, J. A. F., Silva, J. M. C. S. (2019), ‘Quantiles via moments’, Journal of

Econometrics, Vol. 213 (1), pp. 145-173.

[32] Richards, S., and Verstraete, M. (2016), ‘Understanding firms’ inflation expectations
using the Bank of Canada’s Business Outlook Survey’ Bank of Canada Staff Working
Paper 2016-17.

[33] Romer, C. D., and Romer, D. H. (2004), ‘A new measure of monetary shocks: derivation
and implications’, The American Economic Review, pp. 1055-1084.

[34] Sims, C. A. (2003), ‘Implications of rational inattention’, Journal of Monetary Eco-

nomics, Vol. 50, pp. 665-690.

[35] Taylor, J. B. (1993), ‘Discretion versus policy rules in practice’, Carnegie-Rochester

Conference Series on Public Policy, Vol. 39(1), pp.195-214.

[36] Tibshirani, R. (1996), ‘Regression shrinkage and selection via the lasso’, Journal of the
Royal Statistical Society. Series B (Methodological), Vol. 58 (1), pp. 267-288.

[37] Verhoogen, E. (2008), ‘Trade, quality upgrading, and wage inequality in the Mexican
manufacturing sector’, The Quarterly Journal of Economics, Vol. 123(2), pp. 489-530.

446
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

LEADERSHIP AND GENDER

*Kezban Talak

Abstract

Leadership phenomenon in many societies around the world; It depends on various factors
such as education, experience, culture, gender and personality. However, the extent to which
leadership is influenced by gender and other factors is uncertain.

Despite significant increases in women's employment in recent years, women are still
underrepresented in management positions. In addition, their earnings are lower than men.
The main reasons for the low employment of women in this managerial position are
explained by their thinking structures such as the fact that men are superior to women in
terms of power and masculine characteristics. For leadership positions across society, men
are seen as more appropriate than women. However, a debate has recently been raised in the
academic literature and in the popular press about the potential existence of women's
leadership advantage.

In this context, this study aims to start a discussion about whether the underrepresentation
of women in leadership positions has different leadership qualities of women and men, or
whether leadership is affected by gender, and whether there are gender differences in
leadership and potential sources. A detailed literature analysis was conducted to present the
subject from various angles.

For the purpose of the study, answers to the following questions will be sought:

Is gender a dominant factor for leadership position?

What kind of differences come to the fore among male and female managers in the leadership
position?

Inadequate representation of women in management positions, identifying the best potential

candidates in leadership positions leads to wasted opportunity. As a result of this study, he
concluded that women do not reflect less leadership style than men for leadership

Dr., Yeditepe University, International Management (DE)

447
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

position. First of all, it was concluded that gender is not a dominant factor in leadership style.
However, according to the information obtained from the current literature, it is concluded
that women have adopted a more democratic, tolerant and understanding style than men
compared to men.

As a result of the study, it was concluded that the leadership style of women is different from
that of men, but men can learn and adopt from the leadership style of women. From another
point of view, effective leadership is not the exclusive area of both genders, and both can
learn leadership from another. According to another result of the study, the assessment that
a woman's leadership style is less effective than a man is not based on facts, rather it is
directed towards an ongoing perception through socialization.

Keywords: Leadership, Gender, Management, Organization.

448
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Introduction
Leadership research is one of the most studied topics in management and organizational
literature in our age. Leadership is considered as a process (Koçel, 2001), which is analyzed
from different perspectives (Hodgetts, 1999) and has different meanings by different people
(Nicholls, 1994). Leadership studies are kept on the agenda with both empirical and
theoretical studies thanks to many different areas in which the leadership phenomenon
interacts (Hoyt, Goethals, Forsyth, 2008).

Although many studies have been conducted on leadership and leadership styles, the effect
of gender factor on leadership position has not been studied much. However, in line with the
development of technology and the requirements of the era of globalization, the issue of
women's leadership in the popular press has come to the fore. Researchers doing research in
this area have raised a number of research questions related to this topic, but they all raised
a larger query. These questions can be listed as follows:

Why is there a gender difference in the top leadership positions? What is the determination
of the gender factor in the leadership position?

What are the differences that arise between male and female managers in leadership
positions?

In the context of these questions, in this study, leadership phenomenon is examined in the
axis of gender factor. In this context, this study is intended to contribute to the relevant
literature.

Gender

Gender is handled in academic literature with two different perspectives, namely gender and
biological gender, consisting of social, historical and cultural variables and components
(Goktope and Schneier, 1988).

Since the 1970s, researchers have pointed out that a distinction must be made between
biological sex and gender (Bem, 1974; Spence et al., 1975).

449
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Johnson and Repta characterize the entirety of species' anatomical, genetic, physiological
and hormonal aspects as biological sex. They explain the roles, responsibilities and
experiences of individuals in society as a gender (Johnson and Repta).

Bem (1974), on the other hand, explains the gender phenomenon as the psychosocial
situation or preference that arises as a result of qualifying the individual as male or female.

Gender is a complex phenomenon. In different contexts, gender differences stand out in

many aspects. These aspects, gender schemes and stereotypes; It covers gender roles,
attitudes and values (Bem, 1993).

Gender also includes how individuals interact with each other and the social roles expected
to fulfill in a society (Ayman & Koranik, 2010). In addition, gender-related ideas are
culturally and temporally specific and subject to change. Historically, men's higher social
status in many cultures meant that women had more opportunities, strength and resources
than they had, and as a result men); more power and influence were recognized. Various
gender roles (regardless of biological sex), along with views, self-identity and relatively new
changes, make gender and leadership worth discussing (Applebaum et al., 2002).

Leadership
Leadership and management are two indispensable elements. In order for these two skills to
coexist, the person in the administrative role must have the ability to manage daily tasks,
present results, and at the same time seize opportunities. Showing good leadership skills
without management skills will cause that person to fail to fulfill what the vision requires.
Likewise, being a manager without good leadership skills will cause constant difficulties in
motivating his team and producing the necessary results. In this respect, it is considered as
an important skill to combine leadership and management skills.

To define the leadership phenomenon; It is possible to define a definition that “under certain
conditions, a person's process of influencing and directing others' activities to achieve pre-
determined personal or group goals” (Koçel, 2001).

According to another definition; "Leadership is to know yourself, to have a well-

communicated vision, to build trust among colleagues, and to take effective steps to realize
your own leadership potential" (Catalyst, 2009).

450
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The leader first recognizes himself and then the group, and as a result, determines his
superiorities and weaknesses, and then determines and executes his goals and activities
according to possible opportunities and threats in the environment. It is seen that Drucker
uses the terms "listening" and "looking out the window" as a similar method, according to
him, it means that it should seize opportunities in the environment and take action against
threats (Ruble et al., 1984).

In the studies on leadership, it is mentioned that there are some determining features that
distinguish a leader from his followers (Aydemir, 2010). These features are divided into
three as physical, social and personal. Physical features include being young or middle-aged,
energetic, looking great, tall. As social features, characteristics such as having a good
education and having a status are mentioned. As for personality traits, features such as
compatibility, emotional balance, self-confidence and courage are mentioned (Durukan,
2013).

Other elements that constitute the defining characteristics of a leader are explained by
behavioral leadership approach. The main idea of the behavioral leadership approach is that
the issues that make the leader successful, rather than the characteristics of the leader, are
the behaviors that the leader shows while leading the characteristics of their relations with
the people (followers) that form the leader group (Koçel, 2011).

Leadership and Gender

A leader is distinguished by his style, character and behavior, but gender is the main factor
in the effectiveness of his leadership style. From the past to the present, there was only one
gender that was accepted and considered appropriate for leadership. Men were accepted as
a recognized leader profile in business environment, offices, politics, academic institutions
and government worldwide. Research examining the relationship between leadership
position and gender factor broadens the idea that an individual's gender is connected only to
the person and moves to a more general concept of gender role in order to relate leadership
effectiveness or leadership emergence to features or behaviors that are typically considered
male or female. The various consequences can be described very simply as the gender role
is a better predictor of the emergence of the leader relative to gender (Kent & Moss, 1994).

451
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

According to Kiamba, the phenomenon of leadership is described as a male-specific position

from past to present, and the view that good leaders consist of men in this context is widely
accepted. As a matter of fact, the first gender that comes to mind for leadership position in
this context is men (Kiamba, 2008).

Gumbi (2002), by expressing a view that supports Kiamba, states that women are excluded
for leadership position in the stereotypes of many societies, and positions such as high-level
leadership and management are considered as masculine areas.

In the current literature, biological / genetic differences are considered to significantly affect
gender orientation. In the context of this view, Gelman et al. (1981); they acknowledge that
men and women experience the world differently than their hormones. They believe that
hormones do not only contribute to external sexual characteristics, but also affect the view
of the individual to the world.

Differences Between Female and Male Managers in Leadership

Position

Although there are many similarities and differences between male and female managers in
the current literature, the main reason for these differences between men and women is the
fact that women are less represented in the business world than men. In this context, it has
been observed that women apply some strategic differences in order to become a leader in
the business world where women are minorities (Eagly Johannesen-Schmidt, 2001).

In the studies conducted in the literature, it is argued that men who are in the leadership
position have a more business-oriented and mechanical attitude compared to female
managers and that they do not treat the members of the organization personally and
emotionally (Gray Stewart, 2002).

It was claimed by Smith (1992) that female managers take a confident position in leadership
position, but they come to the fore with more passive and silent leadership features than male
managers. However, leading female managers are more advanced in terms of emotional and
social intelligence than male managers, and it is claimed that they are more careful in
establishing healthy communication with members of the organization (Brackett et al.,
2006).

452
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In the studies carried out in the literature, it is accepted that women in leadership place pay
more attention to communication than men and in this context, they display a
“transformative”, “participant” and “inclusive” profile (Chandler, 2011).

In parallel with this view, Eagly (1991, 1995) argues that women in leadership positions
value communication, cooperation and commitment more. In general, it has been observed
that while men use language to improve their social domination, women use communication
to develop social connections and relationships (Mulac et al. 2001).

In studies that examine the characteristics of women and men in the leading position, it has
been revealed that there are great stylistic differences in speech styles between women and
men, but no gender has a superiority over the other in terms of leadership (Chodorow, 1978).
However, it is suggested that women use a more impressive language and are good listeners
in conflict situations than men (Merchant, 2012).

In their research, Kolb (1999) and Shimanoff and Jenkins (1991) show that there is much
more similarity and equally effective than differences in leadership behavior of men and
women. Therefore, the biological gender approach leads to wider studies, with few findings
that rely on biological sex as a valid research hypothesis to differentiate female leadership
of men.

Interestingly, although the learning process continues, the idea behind the biological
approach continues to exist. Although many researchers find that there are few differences
in the innate talents of male and female managers (Oakley, 2000; Dobbins & Platz, 1986;
Powell, 1993), stereotypes depict women as less talented leaders than men.

Research in the current literature broadens the idea that an individual's gender is connected
only to the individual, and transitions to a more general concept of gender role in order to
relate leadership effectiveness or the emergence of leadership to characteristics or behaviors
that are typically considered male or female. The various consequences can be described
very simply as the gender role is a better predictor of the emergence of the leader relative to
gender (Kent & Moss, 1994).

Selected factors that potentially weaken a woman's leadership effectiveness are:

women's attitude, women's self-confidence, women's previous work experience,
work environment.

453
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

According to the gender role theory, being a man or woman means playing a general role as
a function of one's gender. However, this theory also uses the words masculine and feminine,
claiming that especially the feminine character is produced for the role of women through
socialization. According to this approach, women have to learn a lot of gender roles early in
their lives, and this work can lead to a mental attitude that later creates difficulties throughout
their lives. This is a kind of "culture trap". Therefore, the roles taught by women by society
and the attitudes they encourage to assume seem to point to a certain `` second class ''. This
is even more important in the group environment, because group members will choose a
leader who can represent the group's interests. Attitude towards leadership is an important
predictor for emerging leader evaluated by the group (Kolb, 1997).

Conclusion

When an assessment is made in general, the leadership position continues with an

appropriate and accepted mentality for male gender. However, women have degrees
equivalent to men in every field. Discrimination and male domination over organizations
due to stereotypes that exist throughout society cause women to be neglected in high jobs
and not considered as suitable for managers and leadership positions as men. In the context
of these stereotypes, men are embarrassed to work under them in organizations where
women are leaders, so they try to overcome them despite their qualities. However, women
are paid approximately 20% to 30% less than men, regardless of whether they are managers
or leaders.

Although there are many executives in the world, very few actually embody the
characteristics of a leader. As a result, both women and men must have two sets of skills to
be effective and effective in management roles. First, management skills can be easily
learned and trained; this learning process does not differ between genders. On the other hand,
leadership skills are not acquired as easily as management skills and are difficult to master,
and it should not be forgotten that women's leadership skills and styles are different from
those of men.

Undoubtedly, women have come a long way from discrimination and struggles in the
previous century in education and employment. Nowadays, it is no longer surprising to see
a woman as a school principal, a corporate department manager or a dean of a university

454
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

college. Women overcame most of the obstacles in these positions and almost reached the
top of the pyramid. However, there is still a fair inequality when it comes to salary,
promotion, guidance in certain areas or entering certain areas. We hope that the inequality
and stereotypes in society disappear as soon as possible.

KAYNAKÇA

Appelbaum, S.H. and Shapiro, B.T. (1993), ``Why can’t men lead like women?’’,
Leadership & Organization Development Journal, Vol. 14 No. 7, pp. 28-34.

Applebaum, S., L. Audet, and J. Miller(2002). Gender and leadership? Leadership and
gender? A journey through the landscape of theories. Leadership and Organization
Development Journal, 2002. 24(1): pp. 43–51.

Aydemir, N. M. (2010). İşletmelerdeki Liderlik Yaklaşımları İle Türk Silahlı

Kuvvetleri’ndeki Liderlik Yaklaşımlarının Karşılaştırılması.(Yüksek Lisans Tezi).
Afyon Kocatepe Üniversitesi Sosyal Bilimler Enstitüsü, Afyon.

Ayman, R. and K. Koranik, Leadership: Why gender and culture matter. American
Psychologist, 2010. 65(3): pp. 157–170.

Bem, B.M. (1974), ``The measurement of psychological androgyny’’, Journal of Consulting

and Clinical Psychology, Vol. 42, pp. 155-62.

Bem, S., The Lenses of Gender: Transforming the Debate on Sexual Inequality.1993, New
Haven, CT: Yale University Press.

Bem, S.L., Gender schema theory: A cognitive account of sex typing. Psychological Review,
1981. 88(4): p. 354.

Bem, S.L., Gender schema theory and its implications for child development: Raising
gender‐ aschematic children in a gender‐schematic society. Signs, 1983. 8(4): pp.
598–616.

Brackett, M. A., Rivers, S. E., Shiffman, S., Lerner, N., & Salovey, P. (2006). Relating
emotional abilities to social functioning: a comparison of self-report and performance
measures of emotional intelligence. Journal of personality and social psychology,
91(4), 780

455
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Chandler, D. (2011). What women bring to the exercise of leadership? Journal of Strategic
Leadership, 3(2), 1- 12.

Catalyst. (2009). Statistical Overview of Women in the Workplace. Retrieved

09.07.2020from[Link]
_in_the_workplace.pdf.

Chodorow, N. (1978). The reproduction of mothering: Psychoanalysis and Sociology of

Gender. Berkeley: University of California Press

Chodorow, N. J. (1989). Feminism and psychoanalytic theory. Yale University Press

Dobbins, G.H. and Platz, S.J. (1986), ``Sex differences in leadership: how real are they?’’,
Academy of Management Review, Vol. 11 No. 1, pp. 118-27.

Durukan, G. (2013). Liderlik Yaklaşımları Algısının İşe İlişkin Duygusal İyilik Durumu
Üzerine Etkisi. (Yüksek Lisans Tezi). Marmara Üniversitesi Sosyal Bilimler
Enstitüsü. İstanbul.

Eagly, A. H., Karau, S. J., & Makhijani, M. G. (1995). Gender and the effectiveness of
leaders: a meta-analysis.

Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. (1992). Gender and the evaluation of
leaders: A meta-analysis.

Eagly, Alice H. and Johannesen-Schmidt, Mary C. (2001). The Leadership Styles of Women
and Men. Journal of Social Issues 57 (4): 781–97.

Goktope, J.R. and Schneier, C.E. (1988), “Sex and gender effects in evaluating emergent
leadership in small groups”, Sex Roles, 19 (1-2), pp. 29-36.

Gray, J., & Stewart, J. (2002). Mars and Venus in the workplace: A practical guide for
improving communication and getting results at work. HarperCollins Publishers.

Gumbi, R.V. (2002). Tracing differentiation in gendered leadership: an analysis of

differences in gender composition in top management in business, politics and the
civil service. Gender, work and organisation 9(1): 15-38.

456
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Hodgetts, R. M.(1999)., Yönetim Teori, Süreç ve Uygulama, (Çev. Canan Çetin ve Esin Can
Mutlu), 5. Baskı, İstanbul: Der Yayınları.

Nıcholls, J.(1994)., "The 'Heart, Head and Hands' of Transforming Leadership", Leadership
& Organization Development Journal, 15(6), 1994, 8-15.

Hoyt, C. L., Goethals, G. R., & Forsyth, D. R. (2008a). A contemporary social psychology
of leadership. In C. L. Hoyt, G. R. Goethals & D. R. Forsyth (Eds.) Leadership at the
Crossroads: Leadership and Psychology (Vol. 1, pp. 1–10). Westport, CT: Praeger.

Jackson, L.A. (1989), ``Relative deprivation and the gender wage gap’’, Journal of Social
Issues, Vol. 45, pp. 117-33.

Jacobs, J.A. (1992), ``Women’s entry into management: trends in earnings, authority, and
values among salaried managers’’, Administrative Science Quarterly, Vol. 37, pp.
282-301

Johnson, J.L. and R. Repta, Sex and Gender. in Designing and Conducting Gender, Sex, and
Health Research, SAGE Publications Los Angeles J.L. Oliffe 37.

Kent, R.L. and Moss, S.E. (1994), ``Effects of sex and gender role on leader emergence’’,
Academy of Management Journal, Vol. 37 No. 5, pp. 1335-47.

Koçel, T.(2001), İşletme Yöneticiliği, [Link]ı, Istanbul: Beta Basım Yayım Dağıtım A.Ş,
2001.

Kolb, J. (1997), ``Are we still stereotyping leadership? A look at gender and other predictors
of leader emergence’’, Small Group Research, Vol. 28 No. 3, pp. 370-93.

Kolb, J. (1999), ``The effect of gender role, attitude toward leadership, and self-confidence
on leader emergence: implications for leadership development’’, Human Resource
Development Quarterly, Vol. 10 No. 4, pp. 305-20.

Merchant, K. (2012). How men and women differ: Gender differences in communication
styles, influence tactics, and leadership styles. / Claremont McKenna College.

Mulac. A.. Bradac. J. J.. & Gibbons. P (2001). Empirical support for the gender-as Gender
Differences in Leadership culture hypothesis: An intercultural analysis of
male/female language differences’’ / Human Communication Research. 27. 121-152.

457
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Oakley, J.G. (2000), ``Gender-based barriers to senior management positions: understanding

the scarcity of female CEO’s’’, Journal of Business Ethics, Vol. 27 No. 4, pp. 321-
34.

Powell, G.N. (1993), Women and Men in Management, 2nd ed., Sage, Newbury Park, CA.

Ruble, T.L., Cohen, R. and Ruble, D.M. (1984), “Sex stereotypes: occupational barriers for
women”, American Behavioral Scientist, 27, pp. 339-56.

Shimanoff, S.B. and Jenkins, M.M. (1991), ``Leadership and gender: challenging
assumptions and recognizing resources’’, in Cathcart, R.S. and Samovar, L.A. (Eds),
Small Group Communication: A Reader, 6th ed., W.C. Brown, Dubuque, IA, pp.
504-22.

Smith, J.S. 1992. Politeness & directives in Japanese women’s speech. Language in Society
21 (1), 59–82.

Wicks, D. and Bradshaw, P. (1999), ``Gendered organizational cultures in Canadian work

organizations: implications for creating an equitable workplace’’, Management
Decision, Vol. 37 No. 4, pp. 372-80.

458
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Optimization of Shareholder and Corporate Wealth analysis: effectiveness comparative

evaluation of SVA, EVA and MVA

Olorogun, [Link]; Monsurat Ayojimi Salamib

ab
Department of Economics and Finance
Faculty of Economics, Administrative, and Social Sciences
Istanbul Gelisim University, Avcila/Istanbul, Turkey
a
Corresponding email: yinluk2000@[Link]; laolorogun@[Link]
b
Institute of Islamic Banking and Finance (IIiBF)
International Islamic University Malaysia

Abstract

This comparative evaluation study endeavors to establish which of the value-added techniques have
direct positive impact relationship on the shareholder value creation using Julphar Inc. as a case study.
In addition to this, it pursues to discover salient significant features position of Julphar Inc, in creation
of shareholder added-value under the assumptions of Shareholder added-value (SVA), economic value
added (EVA), and market added-value (MVA). The data for this study consists of Julphar’s financial
reports data covering from 2013–2019. SVA, MVA, and EVA were employed respectively to study the
relationship. The study reveals SVA and EVA were directly lean to shareholder value-added creation
compare to MVA. Furthermore, SVA and EVA showed the extent Julphar’s investors were rewarded.
Particularly, SVA results were positive from 2013-2017. Similarly, EVA results were positive from
2014-2017 albeit 2013 was negative indicating that financial increment might not amount to value-added
because SVA of 2013 was positive. Another alarming element is that Julphar recorded net operating
profit after tax (NOPAT) in 2013 not net operating loss after tax (NOPLAT). However, SVA and EVA
were negative in year 2018and 2019; showing a consistent relationship between SVA, EVA and
shareholder value creation partly cause by NOPLAT recorded. In a twist MVA has recorded positive
results throughout i.e. 2013-2019 which pointed to a lack relatedness between MVA and shareholder
value-added. This was regardless of NOPLAT and high expenditures recorded in year 2018 and 2019.
A positive MVA is counterintuitive even after NOPLAT for two years consecutively. It indicates that
there is other element inducing positive MVA. Thus, MVA is an outward insight which measures
managerial acumen of the executive management in deployment of capital for realization of the
shareholder objective. SVA and EVA notwithstanding are inward or direct measure of shareholder value
creation. Other findings are lower profitability, high market risk, inefficient resource management, high
leverage, higher expenditures and weak market capitalization. The primary limitations of this study are
uses of Julphar as a case which might not resolute with other firms, estimation of beta, and risk free rate.
Thus, an industry-wide study is recommended. Optimization of shareholder value has become vital
modern corporate objective globally, and a measure of corporate executives and managers.

Keywords: Shareholder Value Added, Julphar pharmaceutical, Corporate, WACC, Wealth

Optimization, comparative evaluation

1. Introduction

Over the course of past decades, development in theory of shareholders’ rights in organization advances
theory of incentives improved our understanding of relationship between owners and managers of
corporate entities. In the same vain, for decades, optimization of shareholders’ wealth has been studied
extensively using Shareholders’ Value Added (SVA), Economic Value Added (EVA) and Market Value

459
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Added (MVA) (Chartered Institute of Managerial Accountant, 2004; Petravičius & Tamošiūniene 2008;
Shukla 2009; Largani et al. 2012). These tools serve as standard of measuring corporate management’s
performance in large corporations. Unlike for small scale business, establishing individual stakeholders’
interest of large corporations is a very complex task. Axiomatically, the government’s stake is overall
economy interest, while investors such as shareholders need return on their investment, the creditors are
concern if the corporate entity would be able fulfills its financial obligations, and managers’ interest
hangs onto their jobs. Significant added value therefore lies in the interest of the shareholders, who are
the owners of the corporation in order to maximize their wealth pledged in the hands of the managers.
Triggered by the popular corporate governance theory based on asymmetrical relationship between
ownership and management; it is imperative duty of the managers and other stakeholders such as
academic, market players etc. to establish a crystal methods of measuring corporate and management
performance.
Therefore, it is veracious to state that SVA, EVA, and MVA are synonyms by function i.e. to
measure maximization of Shareholder wealth albeit from different approaches. However, creating
shareholder value is subject to periodic review, in Madden (2007) view, corporate boards and
management must come into terms on the content of a periodic shareholder value review. The board can
determine several tactics which differ, nonetheless, they must be relevant to estimate the salient
measures for adjusted profit, total operating assets, rate of return on investment, and cost of capital. By
and large, our goal in this paper is to determine the best and effectiveness of tools of measuring
maximization of shareholders’ wealth among three popular methods i.e. SVA, EVA, MVA. We address
this question in the application of these three methods to a real corporate entity i.e. Julphar Inc. The
study analyzes 2013-2019 financial performance in relations to value added to the shareholders,
economic, and market. We value Julphar for good three reasons: it is a developing economies based
Multinational Corporation; a pharmaceutical products and services, with a view that healthcare industry
is one of the major global powers; lastly, a regional non-state actor in the middle-east particularly the
GCC.
The study applies each evaluation method one after the other considering different theoretical
assumptions and argument for and against each. It places priority on SVA as deontology and a more
effective means of evaluation of corporate performance given its sensitiveness to the continuity of the
firm. It however, did not disregard EVA and MVA because they serve as important indicators measuring
success and failure of corporate organizations, and as well national and global economies.

2. Comparative Analysis of Results

A cordial relationship exists between financing decision and value of corporate organization. This is not
without crinkum-crankum [elaborate detail or something full of twists and turns] of value-based
management (VBM) which plays a vital role in modern corporate discourse. Table 1, depicts inconclusive

460
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

debates on credibility of all VBM over traditional accounting performance measurement. Perhaps, these
results are because Julphar’s capital structure throughout the year understudy, mimic Modigliani and
Miller (henceforth referred as M&M) (1958) propounded value model that a firm’s value is independent
of its capital structure. M&M assumes a world without taxation, agency cost and information
asymmetry. In the same vain, Julphar a UAE-based firm operates in a tax haven nation and other factors
to be elaborated later.

Table 1: Comparative Results of Value Added Estimations of Julphar Inc.

Year SVA MVA EVA

2013 213097.88 23744698 (40582.32)

2014 215289 26713971 170222.92

2015 4896 23339166 188688.54

2016 189990.92 24877400 147662

2017 65676.6 243575 25399.44

2018 (664547.95) 5403900 (693745.01)

2019 (439618.25) 185450 (459013.52)

The results above, were derived in consideration of cost of capital (WACC), net operating profit after
tax (NOPAT) and total capital invested. This provides a level ground of comparing SVA, MVA, and
EVA. In 2013 where Julphar’s capital was full equity SVA and MVA were positive 213097 and
23744698 respectively indicating added value whereas EVA indicate loss (40582). It indicates, that
SVA’s return on capital exceeds cost of capital which does not necessarily translate to economic added
value (EVA). Similarly, market perception through ‘MVA’ the stock price might not indicate value
added. It perhaps justifies Froud et al. (2000) conclusion that anxious profitability focus could lead to a
company value destruction. Likewise, Fernández (2004) concluded that EVA does not measure value-
added. In another words, a company can add financial value without economic value which evident in
2013 SVA, MVA, and negative EVA. Fernández (2004) further resolved that a positive EVA might be
motivated by high NOPAT, decrease in cost of capital, or decrease in assets deployed. Additional
reasonable cause and effects is Narang & Kaur (2014) impact of industry specific on value-added
measurement.
The trend is encouraging from 2014-2017 where SVA, MVA, and EVA are positive but with
varying figures. MVA results were more interesting between 2014-2016, followed by EVA in 2014-
2015, and SVA has lesser added value. Nonetheless, SVA in 2016 is higher than EVA 189990 and

461
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

147662 respectively. Amazingly, the three measurements have almost equal range in 2017 which stood
less than 70,000. Furthermore, SVA and EVA are negative in 2018 and 2019 indicating that no
shareholder and economic added values were recorded rather losses. In contrast, with MVA for 2018-
2019 which were positive at 5403900 and 185450 respectively.
Table 1 show resemblance in performance evaluation between SVA and EVA which indicates
that SVA and EVA leans towards the shareholder value system than MVA. It portrays the vital link
between Julphar’s corporate objective of maximizing shareholder wealth and theoretical basic valuation
or value drivers. This is in line with Ameels (2002) and Largani et al. (2012) affirmation on SVA’s and
as well as Laksana & Hersugondo (2016) EVA link with corporate objective and value drivers. It is
important to reiterate that throughout the period understudy; the effects of capital structure are not felt.
Confirming M&M proposition that firm’s value is independent of its capital structure. This is in
consideration of 2013 negative EVA when total capital employed was equity. Instinctively, positive
EVA is expected resulting from NOPAT as against net operating loss after tax (NOLAT). Moreover,
SVA and EVA show negative results consecutively in 2018 and 2019 which is justify from theoretical
perspective under NOLAT assumptions. Julphar recorded NOLAT for 2018-2019 which is not
surprising from the financial statements, it increases in expenditures and high debt financing due from
bank loans with higher borrowing rates. Despite these evidences, MVA for 2018 and 2019 are positive.
This is an indication that MVA is an external measurement of performance. It measures the
management’s acumen and wisdom in the deployment of capital and amplifying equivalent or higher
value than the capital employed in the market. According to Atiyet (2012) a negative MVA indicates
that managerial activities are less than the value of capital contributed to the market. Julphar Inc. must
have been rewarded by positive MVA throughout the years’ understudy by higher research and
development (R&D) and as a dominant pharmaceutical firm in MENA region. This was in addition to
continuous increases in working capital as pointed out by Olorogun et al. (2020). As evident in its
financial reports of 2018 and 2019 no dividends were distributed which means the investors gain nothing
for the tradeoff or opportunity cause of investing in similar organization in the industry. Another
assumption is Julphar might have indulged in share buyback which minimizes it external dividend
commitments. Thus, it suffices to conclude that MVA is an indirect value-added measurement. The
results are in contrast to Fernández (2004) stance that EVA, CVA etc. are surrogates of MVA. Arguably,
a positive MVA would encourage intending and existing investors of making right choice on where to
put their funds. Equally, it sends positive signal to the market about the position of the firm in the
industry. It also supports macroeconomic policy of the government.

References
Ameels A. (2002). Value –based management control processes to create value through integration a
literature review.; P 21.

462
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Atiyet, B. A. (2012). The Impact of Financing Decision on the Shareholder Value Creation. Journal of
Business Studies Quarterly, Vol. 4 (1), 44-63
Fernández, P. (2004). "EVA, economic profit and cash value added do NOT measure shareholder value
creation", ICFAI Journal of Managerial Economics. Retrieved 26 March 2020,
[Link]
Froud J, Haslam C, Johal S, Williams K. (2000). Shareholder value and financialization: consultancy
promises, management moves. Economy and Society, 29(1):80-110.
Laksana, R. D. & Hersugondo (2016). Does Economic Value Added Influence the Shareholder Value
in Indonesia? International Journal of Applied Business and Economic Research Volume 14,
Issue 3, 2016, 1547-1560
Largani, M. S.; Kaviani, M.; Abdollahpour, A. (2012). A review of the application of the concept of
Shareholder Value Added (SVA) in financial decisions. Procedia - Social and Behavioral
Sciences 40, 490 – 497.
Modigliani, F. & Miller, M. (1958). The cost of capital corporate finance and the theory of investment.
American Economic Review, 48, pp. 261-297.
Narang, S. & Kaur, M. (2014). Impact of Firm-specific Attributes on Shareholder Value Creation of
Indian Companies: An Empirical Analysis. Global Business Review, Vol. 15 (4), pp. 847-866
Olorogun, L.A.; Yalcine, K.; Al Jaberi, A.S. (2020). Symbiotic Relationship between a Healthy
Population and a Sustainable Economy: Financial Performance of Julphar Gulf Pharmaceutical
Industries. International Journal of Economics and Financial Issues, 10(2), 148-155
Petravičius, T. & Tamošiūniene, R. (2008). Corporate performance and the measures of value added.
Transport, 23:3, 194-201
Shukla, H. J. (2009). Creating and Measuring Shareholder Value: A Study of CadiIa Healthcare Limited.
Paradigm, Vol. 13 (1), 66-72.
The Chartered Institute of Managerial Accountants (2004). Maximising Shareholder Value Achieving
clarity in decision-making: Technical Report. Retrieved, 26 March 2020,
[Link]

463
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

THE MARKET AND CUSTOMER ORIENTATION -

PERFORMANCE RELATIONSHIP

PhD student Maria Gračner

Ljubljana, 2020

464
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

THE MARKET AND CUSTOMER ORIENTATION -

PERFORMANCE RELATIONSHIP

Maria Gracner
University of Ljubljana, School of Economics and Business, Kardeljeva ploscad 17, 1000
Ljubljana, Slovenia
Abstract
The paper introduces a new insight into market orientation’s measurement looking at the different levels of
market orientation applied in the different markets. Firms are not simple entities, but have multiple markets,
diverse customers and activities divided across regions, continents, political systems, etc. One strategy does not
fit all. We presume that multinational companies possess the flexibility to apply market orientation (MO)
strategy differently across the markets they serve (export vs domestic). We claim that one market (e.g. highly
risky distant market) may require completely different investment in MO than another (e.g. a neighboring market
with a very low psychic distance). Consequently, we are moving away from a simplified description of firms by
either being market oriented or not, as we believe different foreign market conditions will require different levels
of market orientation applied. Most of papers up to today (Narver, J. C., & Slater, S. F., 1990; Jaworski, B. J., &
Kohli, A. K., 1993; Deshpandé, R., Farley, J. U., & Webster, F. E. Jr., 1993; Diamantopoulos, A., & Hart, S.,
1993; Gounaris, S. P., 2006; Keskin, H., 2006; Ellis, P. D., 2010; Felix R., 2015) measured MO as perception
of managers how much their companies invest in market orientation (Appendix: table 1), but not really as actual
company’s investments.

Keywords: Export, export-oriented firms, export performance, international business, customer

orientation.

1. INTRODUCTION
Firms are not simple entities, they have multiple markets, different customers and activities
divided across regions, continents, political systems, etc. One strategy can’t be equally efficient
for all. Many companies operate in multiple markets, and those markets are not always equal
in terms of their complexities. To optimize performance advantage (gained from the market
orientation) it is possible that different levels of the customer orientation may be needed in
different markets to ensure that market orientation is suitable for those markets. Also, the
businesses become more dynamic, the organization which truly excels in the future will be the
one that discovers how to increase knowledge of the consumer base in the realization of their
highest aspirations (Lee and Tsai, 2005).

Thereby, if it is feasible for multinational companies to purposefully manage their customer

orientation strategy so that they enact different customer orientation behaviors and consumer
behaviours in the different markets that they serve, then a situation is possible that some
companies may gain from having completely different customer orientation levels across their
diverse operations.

In its general sense, we focus our attention on the export activities of businesses, and note that
this level of analysis allows for clear distinctions between domestic and export activities to be
drawn, and for the customer orientation to be viewed from the perspective of the levels of
analysis. Specifically, the domestic market is the prevailing market for many companies, and is
certainly the primary focus of assessment of customer orientation levels when it comes to
empirical work with customer orientation (Ellis, 2007).

465
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The literature indicates that firms may not always get it right in terms of allocating resources
and effort to business orientations and marketing strategies. For instance, the literature suggests
that ‘value induced asymmetry” can affect how firms allocate their resources, so that certain
functional or operational groups and business units may become so valued by executives that,
eventually, they focus more and “more attention, status, and resources on the celebrated group
or function, often at the expense of other areas of business” (Le Bretton-Miller and Miller 2015,
p. 400).

At its heart, then, the paper will addresses critical theoretical question:

“Does customer orientation related to export performance in export oriented firms?”

2. LITERATURE REVIEW

Customer orientation.
Baes on literature analysis, firms which focus their activities on the needs of their customers
perform better than those companies that do not (Donavan et al., 2004). The studies by Narver
and Slater (1990) and Jaworski and Kohli (1993) empirically substantiate the economic
potential of a firm’s customer orientation. Jaworski and Kohli in1993 and Narver and Slater in
1990 support the assumed or implied relationship between a customer orientation and business
performance, the fundamental question as to how a customer orientation influences perceived
performance from a customer’s perspective has yet to be addressed.

Performance.
Because the market orientation helps firms monitor and respond to changing customer needs,
it should be associated with business performance. Literature suggests that firms manage their
relationship with the environment in order to maximize their performance (Shoham et al.,
2005). Resource Based View Theory postulates that differential resources of a firm give rise to
superior strategy and performance (Barney, 1991). Because market orientation helps firms to
improve their resources and is a market differential, the investments on this strategy should
result in superior performance (Perin, Sampaio, & Henriqson, 2005).

Profitability is an indicator of economic efficiency of a business. Literature review provides

evidence that the MO and profitability are interlinked. Narver and Slater shows positive results
between the MO and probability, they stated that “the greater the business’s market orientation,
the greater the business’s profitability” and supported this hypothesis (Narver, Slater, 1990,
2000). But literature findings are very contradictory (Cadogan, 2020). Thus, in our study we
will test the relationship between the MO and profitability at the level of domestic markets and
at the level of export markets as well. Following Agarwal et al. (2003), five items including
service quality, customer satisfaction, employee satisfaction, gross profit margin, and market
share were used in the measurement of business performance.

466
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3. THEORETICAL FRAMEWORK AND DEVELOPMENT OF THE

HYPOTHESES

The practical take out from the study should be a better understanding of how exporting firms
can manage customer orientation to enhance export performance.

Fig 1 presents a preliminary model of how customer orientation at the level of firms’ export
ventures could shape the performance of those ventures, and identifies that there may be
features of the firm design and structure, as well as management factors, that shape the fit
between customer orientation and the export markets (export ventures) the firm operates in.

Export market

HQ investments
in subsideries

Export Export
Autonomy of Customer Performance
subsideries Orientation

Weight of
subsideries

Domestic market

Customer
Orientation

467
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4. Research methodology
As the exploratory phase pre-study was made in Skype with the top manager of export-oriented
firms in Slovenia and with their ventures during June-July 2020. The study consisted of 10
structured depth interviews lasting one to two hours each. Respondents were between 26 and
67 years of age, highly educated. The participants work for export-oriented firms and
predominantly have subsidiaries on the foreign market. The interviews were audiotaped.
As the second stage I developed an online questionnaire in order to assist us with our research
questions. The questionnaire consisted of several items measuring customer orientation and
export performance. All variables were measured with a 5-point Likert scale ranging from
“strongly disagree” (1) to “strongly agree” (5).

5. Results
The analysis shows us that customer orientation is related to export performance of the firms.
The preliminary study is the first step of analysis in the such sphere and in the future qualitative
analysis will take a place.

6. Conclusion
The impact of customer orientation on export performance is enormous. Pre-study based on the
interviews is just a first step for the assessment of the whole picture, further qualitative analysis
will be provided, questionnaires for data collection will be send and then data will be analized.

7. REFERENCES

Abidemi, B. T., Halim F., & Alshuaibi, A. I. (2018). The Relationship Between Market Orientation
Dimensions and Performance of Micro Finance Institutions. Journal of Marketing Management
and Consumer Behavior, 2(3), 1-14.
Cadogan, J. W., Diamantopoulos, A., & Mortanges, C. P. (1999). A measure of export market
orientation: Scale development and cross-cultural validation. Journal of International Business
Studies, 30(4), 689-707.
Cadogan J. W., Paul, N.J., Salminen, R.T, Puumalainen, K., & Sundqvist, S. (2001). Key antecedents
to “export” market-oriented behaviors: a cross-national empirical examination. International
Journal of Research in Marketing, 18(3), 261-282.
Cadogan, John W. (2003). Multiple Perspectives on Market Orientation's Domain Specification:
Implications for Theory Development and Knowledge Accumulation, chapter in Hart, Susan
(ed), Marketing Changes, Thomson Business Press, London, 95-123; 325-328.

468
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Cooper, R. G., & Edgett, S.G. (2001). Portfolio Management for New Products. Product Development
Institute.
Deshpandé, R., Farley, J. U., & Webster, F. E. Jr. (1993). Corporate culture, customer orientation, and
innovativeness in Japanese firms: A quadrad analysis. Journal of Marketing, 57(January), 23-27.
Diamantopoulos, A., & Hart, S. (1993). Linking market orientation and company performance:
preliminary evidence on Kohli and Jaworski’s framework. Journal of Strategic Marketing, 1 (2),
93-121.
Dong, X., Andrew Hinsch, C., Zou, S., & Fu, H. (2013). The effect of market orientation dimensions on
multinational SBU's strategic performance: An empirical study. International Marketing Review,
30(6), 591-616
Ellis, P. D. (2007). Distance, dependence and diversity of markets: effects on market orientation, Journal
of International Business Studies, 38, 374–386.
Ellis, P. D. (2010). Is Market Orientation Affected by the Size and Diversity of Customer
Networks? Management International Review, 50(3), 325-345.
Kohli, A. K., Jaworski, B. J., & Kumar, A. (1993). MARKOR: A measure of market orientation. Journal
of Marketing Research, 30(4), (November), 467-477.
Laukkanen, T., Nagy, G., Hirvonen, S., Rejonen, H., & Pasanen, M. (2013). The effect of strategic
orientations on business performance in SMEs: A multigroup analysis comparing Hungary and
Finland. International Marketing Review, 30(6), 510-535.
Morgan, N. A., Vorhies, D. W., & Mason, C. H. (2009). Market orientation, marketing capabilities, and
firm performance. Strategic Management Journal, 30(8).
Mahmoud, A. M., Blankson, C., Owusu-Frimpong, N., Nwankwo, S., & Trang, T. P. (2016). Market
orientation, learning orientation and business performance: The mediating role of innovation.
International Journal of Bank Marketing, 34(5), 623-648.
Molander S., Fellesson M., & Friman M. (2018). Market Orientation in Public Service—A Comparison
Between Buyers and Providers. Journal of Nonprofit & Public Sector Marketing, 30(1), 74-94.
Narver, J. C., & Slater, S. F. (1998). Additional thoughts on the measurement of market orientation: a
comment on Deshpande and Farley. Journal of market-focused management, 2(3), 233-236.
Narver, J. C., & Slater, S. F. (1990). The effect of a market orientation on business profitability. Journal
of Marketing, (October), 20-34.

469
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

EXPERIENCES OF AUDITORS AND TAX ADVISORS WITH

ACCOUNTING ERRORS: EMPIRICAL EVIDENCE FROM THE
CZECH REPUBLIC

Marie Paseková, Miroslava Dolejšová, Jana Helová, Michal Šindelář

Abstract
Flawless and properly prepared accounting documentation is necessary for companies for
issuing meaningful and reliable financial statements. Results of auditors’ and tax advisors’ work
are further important for other institutions; therefore, it is necessary to minimize the possibility
of errors realized by misunderstanding or overlooked.
The aim of this paper is to evaluate the view of auditors and tax advisors how they perceive
errors in accounting. We compare the opinions of auditors and tax advisors and find out whether
there are some significant differences between the group of auditors and tax advisors. We
created a questionnaire and received 102 completed answers - it means the overall response rate
was 33.4%. Our research was conducted in 2019 in the Czech Republic. We verify the data
with the help of the tests of proportions and the Chi-square test of independence. We confirmed
that more than 46% of auditors and tax advisors find the accounting errors frequently. 6% of
them find the accounting errors constantly. 62.7% of auditors and tax advisors believe that the
usual cause of accounting errors is poor compliance with accounting rules.
Our research can provide important insights into further developments in the field of errors
made by accountants in financial statements and attitude of auditors and tax advisors towards
errors reported or not reported in their final reports.

Keywords: auditors, tax advisors, errors, accounting

JEL Classification: M41

1 INTRODUCTION
The general purpose of annual reports, including financial statements, is to represent the
economic phenomena in numbers and words. Accounting information presented in the annual
reports is based on many requirements. The fundamental requirements are e.g. accrual basis,
true and fair view, entity principle etc. Users of accounting information expect that the annual
report represents the economic situation of the entity faithfully. Then they can make important
qualified economic decisions based on this accounting information. To ensure the possibility
of actively make economic decisions, users require disclosure of financial information that such
information should be complete, neutral and free of error. Avoiding errors is impossible, but
the potential existence of errors can be minimized, e.g. ensuring a quality internal control
system (Icerman and Hillison 1990; Knechel 1985) or using external verification of accounting
information (Jensen and Meckling 1976; K. Hung Chan 1996; Lobo and Yuping Zhao 2013).
In addition to the management itself, auditors and tax advisors are usually the first external
persons who can analyze the accounting information of the entity. Although the primary
objective of both is not to detect errors (for auditors see International Auditing Standard (ISA)
200), they are persons who can find errors in the accounting information. The goal of this paper
is to analyze whether the accounting errors occurred in the companies, how often they occurred,
how large the errors were on average and how management of companies responded to this

470
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

error detection. The analysis is based on the experiences of auditors and tax advisors in the
Czech Republic.
The structure of this paper is the following: The introduction, which defines the motivation and
main objective of the article, is followed by an analysis of international research e.g.
characterize the concept of accounting, characteristics of auditors and tax advisors and what the
detection of accounting errors means. The next part establishes research hypotheses, data and
the suitable methods for data analysis. Chapter 4 will be devoted to the interpretation of the
results and finally, the last part will include a summary of the results and the suggestions for
further research.

2 LITERATURE BACKGROUND
Accounting is main source of all information relating to certain entity. The key to true and fair
view of financial reporting is to tell the investors information about the enterprise that should
influence their decisions. No rule measures what really is material to investors, but internal and
external accountants are supposed to apply general guidelines and their own judgment to decide
whether something is material. In this environment, enterprise management must publish
financial reports that are accurate within the boundaries of materiality. (Cendrowski, H. & Mair,
C.W., 2015). That is one of purposes, which makes profession of auditor, tax advisor and
accountant so closely related. The auditor verifies the financial statements prepared by the
accountant and the tax advisor provides tax services or advices to the entity.
2.1 Accountant
The accountant is a professional party in the relationship and is therefore required to ensure that
the parties are cooperated on the relevant contractual obligations. Accountants lend credibility
to financial statements by the association of their name with those statements. They take all
reasonable measures to protect themselves against a liability to a third party including the use
of appropriately worded disclaimers (Woolf, E., Hindson, M. & Mewi, M., 2015). The issues
surrounding the quality of accounting information amount to no more than a question of
whether this information true and fair. Accountant has to ensure that this information accurately
correspond with relevant aspects of the world that the financial statements are purported to
describe. However, under conventional “accrual” accounting, there are two crucial sorts of
judgment involved. First, although the scope for choice is a little bit restricted, different
methods can be used to combine and process the real transactions. Second, accounting for the
past entails myriad assumptions about the future. (Cowton, C.J., 2019)
In last decades a shift of balance has occurred in the accounting profession — a shift from a
focus on work that facilitates business and economic development toward an emphasis on
satisfying regulatory requirements. One result is that accountants are increasingly trained to
satisfy these regulatory requirements instead of being provided with the skills they need to
become financial professionals who contribute as performance advisors and strategic business
partners (Sharman, P.A., 2007).
2.2 Tax advisor
Profession of tax advisor represents a person registered in the list of tax advisors maintained by
the Chamber of Tax Advisors of the Czech Republic in accordance with the European law. Tax
advisory is based on contract concluded between tax advisor and client. The tax advisor is
remunerated for the provision of tax advice, the amount of which is agreed by both parties.
The profession of tax advisor in the Czech Republic is a profession regulated by a special act
No. 523/1992 Coll., On Tax Advisory Services and the Chamber of Tax Advisors of the Czech

471
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Republic, as amended. This act establishes the fundamental rights and obligations of a tax
advisor in the exercise of tax advisory and liability for breach of obligations. The most serious
impact associated with the counsellor’s unlawful conduct is, of course, the breach of criminal
law (Kuchta, J., 2009).
2.3 Auditor
An auditor is a qualified person who in general carries out the verification or review of
something in an objective manner. An example of an auditor's work is to verify the accuracy of
bookkeeping or compliance of business processes with a given standard (Dufils, P., 2004).
Among external auditors’ responsibilities in auditing of financial statements, belong for
example:
 the role of independent auditors in detecting fraud,
 independent auditors’ opinion on internal control system,
 independent auditors’ opinion on going concern assumption.
The auditor's report on financial statements usually covers the company's financial statements
- balance sheet, income statement, statement of cash flow, statement of owners’ equity and of
course the notes to the financial statement (Rezaee, Z. & Riley, R., 2015). The main
responsibility of auditor is to express an opinion on the financial statements. Audits are
performed in a client‐controlled environment. Corporate fraud is likely to present substantial
challenges to both auditors and forensic accounting investigators. Auditing cannot prevent
financial reporting fraud or prevent employees from looting corporate assets, but it can
significantly reduce these efforts (Gerson, J.S., Brolly, J.P. & Skalak, S.L., 2015).
In periods of financial distress, management may attempt to suppress unfavorable information
to creditors and investors using undisclosed changes in accounting methods, estimates and
procedures. This can significantly reduce the quality of the information contained in the
financial statements. The auditor's role in this context is to ensure that such behavior does not
occurred. If the auditor does not permit such accounting treatments, the company may choose
to change the auditor to future accounting periods (Kluger, B.D. & Shields, D., 1989). An
external audit does not provide absolute assurance that the financial statements are true and fair.
This is because an audit contains inherent limitations, such as the use of judgements in the
auditing process, the use materiality etc. The overall objective of the auditor is to express an
opinion as to whether the financial statements are true and fair. (Collings, S., 2016).
Companies and their investors expect auditors to detect material misstatements and prevent the
issuance of misleading financial statements. By certifying the annual reports, the independent
auditor assumes a public responsibility transcending any employment relationship with the
client (Aronow, G. & Harris, H., 2015). Audit firms claim that they are used as the “whipping
boy” when something goes wrong, either because of the public's poor knowledge of the auditing
function or because financial incentives exist to blame the audit firm. An alternative explanation
is that audit firms' active communication of their assurance provider role causes blaming
behavior. (Hoos, F, Saad, EB. & Lesage, C., 2018) Audit performance require the use of
professional judgement and other audit procedures that inherently cannot guarantee that they
detected all misstatements in financial statements. (Howieson, B., 2013)
Act No. 93/2009 Coll., On Auditors, as amended regulates auditors’ activities in Czech
Republic. The audit is also regulated by a number of related laws and implementing regulations,
including Code of Ethics, International Auditing Standards and it´s application clauses issued
by the Chamber of Auditors of the Czech Republic (Müllerová, L. et al., 2017).
2.4 Errors in accounting

472
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Accounting errors can occur in double entry bookkeeping for a number of reasons. Accounting
errors are not the same as fraud. The errors usually happen unintentionally, whereas fraud is
provided intentionally to falsify the bookkeeping entries (Venter, ER, Gordon, EA & Street,
DL., 2018). Accounting professionals play two important roles in any forensic investigation:
 as leaders of financial investigators and
 potentially, as expert witnesses in any subsequent civil or criminal trials.
In the first instance, they are the key people in any fraud investigation because they understand
accounting systems and internal controls and know how to track the flow of funds into, through
and out of the entity. They are also in a position to provide an independent, objective, and
unbiased critique of the entity organization. As experts assisting in case strategy and testimony,
accounting professionals must be knowledgeable about court proceedings, rules of evidence
etc. The good financial investigator must be knowledgeable about fraudulent practices both in
general and within a specific industry (Silverstone, H. 2015).
Unintentional accounting errors are common if the accountant is not careful when fulfilling
own duties. The discovery of such errors usually occurs when companies conduct their month-
end book closings. Most errors, if not all, can be corrected instantly (Abreu, R., 2015). There
are ranges of personal incentives that may act as motivation to indulge in accounting fraud.
Managers may wish to use creative accounting to increase reported earnings so that they will
receive increased salaries, greater bonuses or more lucrative share options. In the other words,
they may wish to maximize their own personal remuneration (Drábková, Z., 2017). Second,
there are pressures from the external users to use creative accounting. Managers may wish to
meet analysts’ expectations, both in the short term by last‐minute earnings adjustments and in
the long term through profit smoothing. There may also be a range of special circumstances
that act as motivations to indulge in creative accounting and fraud (Jones, M., 2015). When
correcting an error in prior period financial statements, the term restatement is commonly used.
Preparers of financial statements need to have control procedures to reduce the risk of
accounting errors being committed and not detected. It is required that auditors examine the
financial statements so that there is reasonable assurance of detecting material misstatements
(Flood, J. M., 2015).
The attempt to analyse errors in the financial statements is a relatively frequently investigated
problem in the international context, but in the Czech Republic only a fraction of studies deals
with this problem (Kolářová, 2014). Acito, Burks, and Johnson (2019) analyse the accounting
errors by examining SEC comment letter correspondence and find out, that managers deem the
error immaterial despite its exceeding the traditional “5 percent of earnings” rule of thumb,
often in multiple periods and by a large degree. Instead of attempting to conceal these overages,
managers tend to forthrightly acknowledge them, often asserting that the benchmark (yearly
earnings) is abnormally low during the violation period. This result means, that managers are
aware of errors in accounting information. Moreover Fang, Huang, and Wang (2017) highlight
the imperfectness of accounting using several number of items. They find out that the errors are
ubiquitous and when errors are more prevalent, the market reacts less to firms’ earnings
surprises and bias is more difficult to detect. How to correct accounting errors is analysed by
Steliana (2013).
In the international research audit is often associated with the analysis of errors in the financial
statements. Shin, Randall Zhaohui Xu, and Lacina (2011) analysed auditor attributes that may
help determine the time it takes auditors to detect and have clients correct financial statement
misstatements in pre-Sarbanes Oxley Act period and post-Sarbanes-Oxley Act period. Their
results conclude that Sarbanes Oxley Act improve the timely detection and correction of
improper accounting by auditors. The research of Kent and Weber (1998) is based on evaluation

473
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of the quality of auditor performance in comparison with author´s estimation of dollar error in
the inventories. They calculated the sum of dollar error by program simulation using the
company´s internal control and accounting system. Adalı and Kizil (2017) investigate 36
accounting professionals residing in Turkey by the questionnaire concerning on the ethical
dimensions of accounting professionals related to accounting errors and frauds. They conclude
two important conclusions. One of them states that the effective tool to prevent errors and fraud
is indicated as external audit. This is in line with the economic justification for the audit within
the agency theory (Fama and Jensen 1983; Cardinaels and Jia 2016; Tillema and ter Bogt 2016).
Second conclusion contains assertion that accounting most errors occur due to incorrect data
received from clients and as a result of recording. Other study try to determine whether auditors'
knowledge of basic accounting principles and error frequencies improves over the course of
their early careers so as to enhance performance of a common analytical procedure (Nelson
1993).
On the other hand, accounting errors with an impact on profit or loss may affect the tax base
and income tax paid. According to this, revenues and expenses should be checked by the tax
advisor. Bloom and Weinstein (2004) emphasize that accounting textbooks are invariably weak
and unclear on how to deal with the income tax consequences of accounting changes and error
corrections. Martins and Sa (2018) focus on motives showing a connection to errors or fraud in
the recognition of operations by the financial accounting system in Portugal. Gurău and Grigore
(2016) analysed the impact of accountancy errors on financial and tax reports in Romania.

3 RESEARCH OBJECTIVE, METHODOLOGY AND DATA

For the purpose of the analysis, the authors proposed a questionnaire that was sent to auditors
and tax advisors for completion. Our research was made in 2019 within the Czech Republic.
We received 102 completed questionnaires, of which 45 were the auditors and 57 were the tax
advisors. Tax advisors were selected by random sampling from the list of tax advisors of the
Czech Republic. The auditors were selected by random sampling from the register of auditors
of the Czech Republic. In the Czech Republic is registered about 4 800 tax advisors and about
1 200 auditors. We contacted 157 tax advisors and 148 auditors by email and asked them for
completing an online questionnaire The response rate was 33.4%.
Our data was verified through the tests of proportions and the Chi-square test of independence.
When the Chi-square test of independence confirmed the statistical dependence, the adjusted
residuals were also applied to interpret our results. If the Chi-square test of independence could
not be used due to failure to meet their prerequisites for its use, this fact was highlighted in the
text of this paper. The data were analyzed in the statistical applications XLStatistics and SPSS
at a significance level of 5%.
The aim of this research paper is to evaluate the view of auditors and tax advisors how they
perceive errors in accounting. We are interested in how large the errors are on average, whether
there is any relationship between the average size of errors and their frequency of occurrence,
and whether poor compliance with accounting rules or their deficiencies are a common cause
of accounting errors.
To analyze our data obtained, we have identified the following research questions:
H1: How often do accounting errors occur?
H2 How large are the average errors in accounting?
H3: Is there any relationship between the frequency of accounting errors and the average size
of errors?

474
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

H4: Is the poor accounting compliance with accounting rules or their shortcomings, i.e. lack of
procedure, opacity, incomprehensibility etc., common cause of accounting errors?
These four research questions were statistically evaluated through the test of proportions and
the Chi-square tests of independence. When the Chi-square test of independence confirmed the
statistical dependence, the adjusted residuals were also applied to interpret our results. If the
Chi-square test of independence could not be used due to failure to meet the prerequisites for
its use, this fact was highlighted in the text of this paper. The data were analyzed in the statistical
applications XLStatistics and SPSS at a significance level of 5%.

4 RESULTS AND DISCUSSION

The first research question H1: How often do accounting errors occur?
Table 1 shows the frequencies of answers to this question.
Tab. 1 – How often do accounting errors occur? Source: own research
Answer Frequency %
Never 0 0.0%
Sporadically 49 48.0%
Often 47 46.1%
Permanently 6 5.9%
Total 102 100.0%
No auditor or tax advisor stated that there were no errors in the accounting at all. Errors occur
sporadically in 48% of clients. We were surprised that more than 46% of auditors and tax
advisors find mistakes frequently in the accounting. Only 6% of auditors and tax advisors find
errors in the accounting permanently.
The pivot table (Table 2) shows the answers of auditors and tax advisors to this question:
Tab. 2 – How often do accounting errors occur? Source: own research
Never Sporadically Often Permanently Total
Auditor 0 12 30 3 45
Tax advisor 0 37 17 3 57
Total 0 49 47 6 102
We can see that almost 65% of tax advisors and almost 27% of auditors find accounting errors
sporadically. On the other hand, almost 67% of auditors and almost 30% of tax advisors often
find errors in the accounting. Almost 7% of auditors and more than 5% of tax advisors find
errors in the accounting of their clients constantly.
The statistical hypothesis HA11: More than 50% of auditors and tax advisors believe that the
errors in accounting occur frequently. The one-sample test of proportions did not confirm the
majority of auditors and tax advisors that the errors in accounting occur frequently (Z = -0.792,
p = .786). In other words, majority of most auditors and tax advisors agree that the accounting
errors do not occur too often.
The statistical hypothesis HA12: The auditors are more confident that accounting errors occur
more often than the tax advisors. A two-sample test of proportions did not confirm the opinion
of auditors that the accounting errors occur more often than tax advisors (Z = 3.706, p <.001).

475
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The proportion of auditors (66.7 %) is significantly higher than the proportion of tax advisors
(29.8 %).
The statistical hypothesis HA13: The tax advisors more believe that errors in accounting occur
sporadically than the auditors. A two-sample test of proportions confirmed the opinion of tax
advisors that the accounting errors occur sporadically (Z = -3.84, p < .001). The proportion of
tax advisors (64.9%) is significantly higher than the proportion of auditors (26.7%).
The statistical hypothesis HA14: There is a statistical dependence between the frequencies of
accounting errors and whether they we ask the auditors or the tax advisors. The existence of
statistical dependence could not be confirmed by this independence test due to fact that two
cells (33.3%) in the pivot table had the expected frequencies less than 5.
The second research question H2: How large are the average errors in accounting?
Table 3 shows the frequencies of answers to this question:
Tab. 3 – How large are the average errors in accounting? Source: own research
Answer Frequency %
Insignificant errors - crown character 17 16.7%
Minor errors (0.1 to 0.9% of total annual turnover) 79 77.5%
Significant errors (more than 1% of the company's
6 5.9%
total annual turnover)
Total 102 100.0%
Almost 78% of auditors and tax advisors find minor accounting errors. 6% of auditors and tax
advisors find significant accounting errors. Only 17% of them find insignificant accounting
errors.
The pivot table (Table 4) shows the answers of auditors and tax advisors to this question:
Tab. 4 – The frequencies of answers. Source: own research
Answer Auditors Tax advisors Total
Insignificant errors - crown character 1 16 17
Minor errors (0.1 to 0.9 % of total annual turnover) 40 39 79
Significant errors (more than 1 % of the company's
4 2 6
total annual turnover)
Total 45 57 102
Table 4 shows that 88.9% of auditors and 68.4% of tax advisors find minor accounting errors.
Surprisingly, 28.1% of tax advisors and 2.2% of auditors find insignificant accounting errors.
8.9% of auditors and only 3.5% of tax advisors find significant accounting errors.
The statistical hypothesis HA21: More than 50% of auditors and tax advisors believe that minor
errors occur in accounting. A one-sample test of proportions confirmed the opinion of auditors
and tax advisors that minor errors occur in accounting (Z = 5.545, p < .001).
The statistical hypothesis HA22: The auditors more believe that there are minor errors in
accounting than the tax advisors. A two-sample test of proportions confirmed the auditors'
opinion that there are minor errors in accounting (Z = 2.456, p < .001). The proportion of
auditors (88.9%) is significantly higher than the proportion of tax advisors (68.4%).

476
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The statistical hypothesis HA23: The tax advisors more than auditors believe that minor errors
appear in accounting. The two-sample test of proportions confirmed the opinion of tax advisors
that there are minor errors in accounting (Z = -3.478, p < .001). The proportion of tax advisors
(22.2%) is significantly higher than the proportion of auditors (2.2%).
The statistical hypothesis HA24: There is a statistical dependence between the average size of
accounting errors and whether we ask auditors or tax advisors. The existence of statistical
dependence could not be confirmed by this independence test due to fact that two cells (33.3%)
in the pivot table had the expected frequencies less than 5.
The third research question H3: Is there any relationship between the frequency of accounting
errors and the average size of errors? The pivot table (Table 5) shows the answers of auditors
and tax advisors to this question:
Tab. 5 – The frequency of accounting errors and average size of errors. Source: own research
Answer Sporadically Often Permanently Total
Insignificant errors - crown character 14 3 0 17
Minor errors (0.1 to 0.9% of total annual
35 40 4 79
turnover)
Significant errors (more than 1% of the
0 4 2 6
company's total annual turnover)
Total 49 47 6 102
82.4% of insignificant accounting errors occur sporadically. 50.6% of minor accounting errors
occur frequently and 66.7% of significant errors occur frequently in accounting.
The statistical hypothesis HA31: There is a statistical dependence between the average size of
accounting errors and frequencies of their occurrence. The existence of statistical dependence
could not be confirmed by this independence test due to fact that five cells (56%) in the pivot
table had the expected frequencies less than 5.
The fourth research question H4: Is the poor accounting compliance with accounting rules or
their shortcomings a common cause of accounting errors?
Table 6 shows the frequency of answers to this question:
Tab. 6 – The frequencies of answers. Source: own research
Answer Frequency %
Certainly not 8 7.8%
Rather not 30 29.4%
Rather yes 56 54.9%
Definitely yes 8 7.8%
TOTAL 102 100.0%
62.7% of auditors and tax advisors believe that the usual cause of accounting errors is due to
poor compliance with accounting rules or their shortcomings (i.e. lack of procedure, opacity,
incomprehensibility) and 37.3% of auditors and tax advisors take the opposite view.
The pivot table (Table 7) shows the answers of auditors and tax advisors to this question:
Tab. 7 – The frequency of answers to this question. Source: own research

477
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Certainly not Rather not Rather yes Definitely yes Total

Auditor 5 11 24 5 45
Tax advisor 3 19 32 3 57
Total 8 30 56 8 102
More than 64% of auditors and more than 61% of tax advisors agree that the usual cause of
accounting errors is due to poor compliance with accounting rules or their deficiencies. 36% of
auditors and 39% of tax advisors take the opposite view.
The statistical hypothesis HA41: More than 50% of auditors and tax advisors believe that the
usual cause of accounting errors is poor compliance with accounting rules. A one-sample test
of proportions confirmed the opinion of majority auditors and tax advisors that the usual cause
of errors in accounting is poor compliance with accounting rules (Z = 2.574, p < .001).
The statistical hypothesis HA42: The tax advisors more believe that the usual cause of accounting
errors is poor compliance with accounting rules than the auditors. A two-sample test of
proportions confirmed the opinion of tax advisors that the usual cause of accounting errors is
poor compliance with accounting rules (Z = -2,592, p < .001). The proportion of tax advisors
(61.4%) is significantly higher than the proportion of tax advisors (35.6%).
The statistical hypothesis HA43: There is a statistical dependence between the fact that the usual
cause of accounting errors is poor compliance with accounting rules and whether we ask the
auditors or the tax advisors. The test of independence did not confirm a statistical dependence
between the fact that the usual cause of accounting errors is poor compliance with accounting
rules and whether we ask the auditors or the tax advisors: χ2 (1) = 0.012, p = .913.
The Fisher's exact test (p = .838) did not confirm this statistical dependence either.

5 CONCLUSION
Conclusions relating to considering of errors in accounting brought interesting results.
According to results of given research, most auditors and tax advisors believe that minor errors
occur in accounting. We also confirmed that auditors more believe that there are minor errors
in accounting than the tax advisors. According to given research, we found out that 82.4% of
minor accounting errors occur sporadically. 50.6% of minor accounting errors occur frequently
and 66.7% of significant errors occur frequently in accounting.
We were somewhat surprised that more than 46% of auditors and tax advisors find accounting
errors frequently. Only 6% of auditors and tax advisors find the accounting errors constantly.
62.7% of auditors and tax advisors believe that the usual cause of accounting errors is poor
compliance with accounting rules and 37.3% of auditors and tax advisors take the opposite
view.
We cannot say that the accounting errors occurred due to oversight or initial misstatements that
caused the differences in the reported documents presented by accounting staff. In our research,
we were not interested in what specific accounting errors were and how they occurred. This
may be the subject of further research.
Even fraud commitment can be firstly coming from few small not discovered mistakes resulting
in fraudulent issues. First not discovered error can persuade the accountant to try more hidden
surprises inside presented documentation and that is why auditors and tax advisors need to
detect such a commitment and try to predict and protect negative consequences leading to other
errors and mistakes in accounting documentations.

478
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Our research can provide important insights into further developments in the field of errors
made by accountants in financial statements and attitude of auditors and tax advisors towards
errors reported or not reported in their final reports.

Acknowledgement
This article has been processed as one of the outputs of the research project VSE F1/25/2019 -
"Market failures and their impact on audit quality in the Czech Republic".

References
1. Abreu, R. (2015). Accounting for Citizenship: The Role of Accountant. Procedia
Economics and Finance, 26(1), 933-941. doi: 10.1016/S2212-5671(15)00909-0
2. Acito, Andrew A., Jeffrey J. Burks, and W. Bruce Johnson. 2019. ‘The Materiality of
Accounting Errors: Evidence from SEC Comment Letters’. Contemporary
Accounting Research 36 (2): 839–68. [Link]
3. Act No. 93/2009 Coll., On Auditors, as amended
4. Act No. 523/1992 Coll., On Tax Advisory Services and the Chamber of Tax Advisors
of the Czech Republic, as amended
5. Adalı, Semanur, and Cevdet Kizil. 2017. ‘A Research on the Responsibility of
Accounting Professionals to Determine and Prevent Accounting Errors and Frauds:
Edirne Sample’. SSRN Scholarly Paper ID 3179567. Rochester, NY: Social Science
Research Network. [Link]
6. Aronow, G. and Harris, H. (2015). Auditor Responsibilities and the Law. In A Guide
to Forensic Accounting Investigation (eds T.W. Golden, S.L. Skalak, M.M. Clayton
and J.S. Pill). doi:10.1002/9781119200048.ch4
7. Bloom, Robert, and Gerald P. Weinstein. 2004. ‘Income Tax Aspects of Accounting
Changes and Error Corrections: An Instructional Approach’. Advances in Accounting
Education; Bingley 6: 87–115.
8. Cardinaels, Eddy, and Yuping Jia. 2016. ‘How Audits Moderate the Effects of
Incentives and Peer Behavior on Misreporting’. European Accounting Review 25 (1):
183–204. [Link]
9. Cendrowski, H. & Mair, C.W. (2015). Accounting Policies and Procedures. In
Enterprise Risk Management and COSO (eds H. Cendrowski and W.C. Mair).
doi:10.1002/9781119203780.ch14
10. Collings, S. (2016). The Auditor's Report. In UK GAAP Financial Statement
Disclosures Manual, S. Collings (Ed.). doi:10.1002/9781119283393.ch8
11. Cowton, C. J. (2019). Making a contemporary contribution using old data:
Reflections on delayed doctorates. International Journal of Management Education
(Elsevier Science), 17(1), 77–84. [Link]
12. Drábková, Z. (2017). Kreativní účetnictví a účetní podvody: řízení rizika účetních chyb
a podvodů. Prague, Czech Republic: Wolters Kluwer.
13. Dufils, P. (2004). Finanční audit. Praha, Czech Republic: HZ.
14. Fang, Vivian W., Allen H. Huang, and Wenyu Wang. 2017. ‘Imperfect Accounting
and Reporting Bias’. Journal of Accounting Research (John Wiley & Sons, Inc.) 55
(4): 919–62. [Link]

479
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

15. Fama, Eugene F., and Michael C. Jensen. 1983. ‘Separation of Ownership and
Control’. The Journal of Law & Economics 26 (2): 301–25.
16. Flood, J. M. (2015). ASC 250 Accounting changes and error corrections. In GAAP
2016 Interpretation and Application of Generally Accepted Accounting Principles,
J.M. Flood (Ed.). doi:10.1002/9781119216698.ch09
17. Gerson, J.S., Brolly, J.P. & Skalak, S.L. (2015). The Roles of the Auditor and the
Forensic Accounting Investigator. In A Guide to Forensic Accounting Investigation
(eds T.W. Golden, S.L. Skalak, M.M. Clayton and J.S. Pill).
doi:10.1002/9781119200048.ch3
18. Gurău, Mariana, and Maria Zenovia Grigore. 2016. ‘The Influence of Accountancy
Errors on Financial and Tax Reports’. Global Economic Observer 4 (2): 123–28.
19. Hoos, F, Saad, EB. & Lesage, C. (2018). Why are auditors blamed when something
goes wrong? Experimental evidence. Int J Audit. 22: 422– 434.
[Link]
20. Howieson, B. (2013). Quis Auditoret Ipsos Auditores? Can Auditors Be Trusted?.
Australian Accounting Review, 23: 295-306. doi:10.1111/auar.12008
21. Icerman, Rhoda C., and William A. Hillison. 1990. ‘Distributions of Audit-Detected
Errors Partitioned by Internal Control’. Journal of Accounting, Auditing & Finance 5
(4): 527–43. [Link]
22. Jensen, Michael C., and William H. Meckling. 1976. ‘Theory of the Firm:
Managerial Behavior, Agency Costs and Ownership Structure’. Journal of Financial
Economics 3 (4): 305–60. [Link]
23. Jones, M. (2015). Motivations to Indulge in Creative Accounting and Fraud. In
Creative Accounting, Fraud and International Accounting Scandals, M.J. Jones (Ed.).
doi: 10.1002/9781119208907.ch3
24. K. Hung Chan. 1996. ‘Estimating Accounting Errors in Audit Sampling: Extensions
and Empirical Tests of a Decomposition Approach’. Journal of Accounting, Auditing
& Finance 11 (2): 153–61. [Link]
25. Kent, Pamela, and Ron Weber. 1998. ‘Auditor Expertise and the Estimation of
Dollar Error in Accounts’. Abacus 34 (1): 120–39. [Link]
6281.00025.
26. Kluger, B.D. & Shields, D. (1989), Auditor changes, information quality and
bankruptcy prediction. Manage. Decis. Econ., 10: 275-282.
doi:10.1002/mde.4090100404
27. Knechel, W. Robert. 1985. ‘An Analysis of Alternative Error Assumptions in
Modeling the Reliability of Accounting Systems’. Journal of Accounting Research
(Wiley-Blackwell) 23 (1): 194–212. [Link]
28. Kolářová, Eva. 2014. Vypovídací schopnost účetních dat. Available at:
[Link]
29. Kuchta, J. (2009). Kurs trestního práva, Trestní právo hmotné, Zvláštní část. Praha,
Czech Republic: C. H. Beck.
30. Lobo, Gerald J., and Yuping Zhao. 2013. ‘Relation between Audit Effort and
Financial Report Misstatements: Evidence from Quarterly and Annual

480
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Restatements’. Accounting Review 88 (4): 1385–1412. [Link]

50440.
31. Martins, António, and Cristina Sa. 2018. ‘The Computation of Taxable Income
When Accounting Numbers Are Not Reliable’. International Journal of Law &
Management 60 (2): 543–62. [Link]
32. Müllerová, L. a kol. Auditing. Prague: Oeconomica, 2017. P. 387. ISBN 978-80-245-
2233-3.
33. Nelson, Mark W. 1993. ‘The Effects of Error Frequency and Accounting Knowledge
on Error Diagnosis in Analytical Review’. Accounting Review 68 (4): 804–24.
34. Rezaee, Z. & Riley, R. (2015). Role of External Auditors. In Financial Statement Fraud
(eds Z. Rezaee and R. Riley). doi:10.1002/9781119198307.ch11
35. Sharman, P.A. (2007), Rebalancing the accounting profession. J. Corp. Acct. Fin.,
18: 45-50. doi:10.1002/jcaf.20291
36. Silverstone, H. (2015). The Role of the Accounting Professional. In Forensic
Accounting and Fraud Investigation for Non‐Experts (eds H. Silverstone, M. Sheetz,
S. Pedneault and F. Rudewicz). doi:10.1002/9781119200635.ch6
37. Steliana, Busuioceanu. 2013. ‘Correcting Accounting Errors and Acknowledging
Them in the Earnings to the Period’. Annals of ‘Constantin Brancusi’ University of
Targu-Jiu. Economy Series, no. 5 (September): 18–23.
38. Tillema, Sandra, and Henk J. ter Bogt. 2016. ‘Does an Agency-Type of Audit Model
Fit a Stewardship Context? Evidence from Performance Auditing in Dutch
Municipalities’. Financial Accountability & Management 32 (2): 135–56.
[Link]
39. Venter, ER, Gordon, EA & Street, DL. (2018). The role of accounting and the
accountancy profession in economic development: A research agenda. J Int Financ
Manage Account. 29, 195– 218. [Link]
40. Woolf, E., Hindson, M. & Mewi, M., (2015). Professional Pitfalls for Accountants. In
Audit and Accountancy Pitfalls (eds E. Woolf, M. Hindson and M. Mewi).
doi:10.1002/9781119209287.ch4
Contact information
Associate prof. Marie Paseková, Ph.D.
Tomas Bata University in Zlín
Faculty of Management and Economics
Department of Finance and Accounting
E-mail: pasekova@[Link]
ORCID ID: [Link]

Assistant prof. Miroslava Dolejšová, Ph.D.

Tomas Bata University in Zlín
Faculty of Management and Economics
Department of Statistics and Quantitative Methods
E-mail: dolejsova@[Link]
ORCID ID: [Link]

Assistant Jana Helová

481
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Tomas Bata University in Zlín

Faculty of Management and Economics
Department of Finance and Accounting
E-mail: helova@[Link]
ORCID ID: 0000-0003-3780-2907

Assistant prof. Michal Šindelář, Ph.D.

University of Economics, Prague University
Faculty of Finance and Accounting
Department of Financial Accounting and Auditing
E-mail: [Link]@[Link]
ORCID ID: [Link]

482
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The importance of Internationalization strategy for Innovation in Portuguese firms

Mónica Azevedo1, Carla Azevedo Lobo2, Carla Santos Pereira3, Natércia Durão3
e Isabel Maldonado2
1
Universidade Portucalense; REMIT- Research on Economics, Management and
Information Technologies;
2
Universidade Portucalense; REMIT- Research on Economics, Management and
Information Technologies; IJP- Instituto Jurídico Portucalense, Porto, Portugal;
GOVCOPP, Aveiro, Portugal
3
Universidade Portucalense; REMIT- Research on Economics, Management and
Information Technologies; IJP- Instituto Jurídico Portucalense, Porto, Portugal
monicaa@[Link]
cadsa@[Link]
carlasantos@[Link]
natercia@[Link]
ianm@[Link]

Abstract
Innovation and internationalization seem to be vital strategies for the survival and
growth of companies facing an increasingly competitive global environment. The two-
way link between these two factors has become a topic of interest among researchers.
Although innovation and internationalization are highly related activities, the role
played by innovation in the internationalization process of firms has been analysed by
two streams of literature: while some authors state that innovation assists companies’
internationalization, others consider that innovation can be a consequence of
internationalization process. Given the great importance of small and medium-sized
enterprises (SMEs) for the growth of economies, part of literature has paid special
attention to this kind of enterprises. The aim of this study is to explore the relationship
between innovation and internationalization within Portuguese firms, in particular to
analyse whether Portuguese entrepreneurs consider innovation as an important factor
in the process of internationalisation. To achieve this goals, we will use descriptive and
inferential data analysis techniques.

Keywords: Internationalization, Innovation, SMEs

JEL classifications: D22, F13, H32, L25, L26, M16

Introduction
The development of the business world has been breaking all boundaries, being the
increasing internationalization of companies taken for granted regardless of size, age,

483
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

international or domestic experience, slightly contradicting the traditional theories of

internationalization. Companies need to find new ways to develop competitive
advantages which comprises seeking and acquiring new skills, resources and
capabilities. Several are the traditional theories that focus attention on the determining
factors in the internationalization process. Sapienza, Autio, George and Zahra (2006)
argue that companies have to continually adapt to the dynamic changes in the
environment so the innovation process is fundamental and product innovation is
essential. According to Lobo, Fernandes, Ferreira and Peris-Ortiz (2019), SMEs face
specific challenges such as a lack of human, financial, technological and informational
resources. The importance of innovation seems to increase in a global market context.
Zucchella and Siano (2014) underline that innovative resources and capabilities are
essential for firms’ growth in both domestic and foreign markets. Genc, Dayan and Genc
(2019) and Saridakis, Idris, Hansen and Dana (2019) address the product life cycle: on
the one hand, to stress that as these cycles are getting shorter, it is imperative that
companies continually explore new products (or services) to offer; on the other hand,
Vernon's Product Life Cycle Theory is mentioned in order to link innovation and
internationalisation. Additionally, it seems to be consensual that internationalization
and R&D investment / innovation are two important strategic decisions related to the
success of firms (Chang, Chang, Hsu and Yang, 2018; Gjergji, Lazzarotti, Visconti and
Saha, 2019). Moreover, their joint effect is considered vital for companies’ success and
survival in the global markets (Onetti, Zucchella, Jones and McDougall-Covin, 2012). This
relationship is even described as a dynamic virtuous circle given that internationalization
and innovation reinforce each other (Golovko and Valentini, 2011).
Therefore, it’s our intention to investigate the influence of internationalization on
innovation, in particular to analyse whether Portuguese entrepreneurs consider
innovation as an important factor or a motivation in the process of internationalization.

Literature Review
Several are the traditional theories that focus attention on the determining factors in
the internationalization process. Sapienza et al. (2006) argue that companies have to
continually adapt to the dynamic changes in the environment so the innovation process

484
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

is fundamental and product innovation is essential. According to Lobo et al. (2019),

SMEs face specific challenges such as a lack of human, financial, technological and
informational resources. Czinkota et al. (1999) provide a list of key factors that influence
the strategic approach to business internationalisation. These factors are grouped into
proactive factors (potential advantages in terms of profits, technology, unique products,
information management, tax benefits and economies of scale) and reactive factors
(competitive pressure, excess capacity, saturation of the domestic market and proximity
to customers). Mathews (2006) argues that companies might internationalise for
relational motives (by responding to competitors, following up on customers’
internationalisation processes or acting upon approaches of foreign companies), access
to foreign resources (finding lower production costs abroad or accessing technological
knowledge through branches) or government incentives (through the support of the
home or host government). Also, as reported by Dunning and Lundan (2008),
internationalization has four motivations, including Resource-seeking, which defines a
search for resources, such as raw material, labour and obtaining technological,
management or marketing know-how, which given the scarcity or non-existence in the
country of origin, lead to a need for investment in foreign countries. And Strategic asset-
seeking, where MNEs constantly seek to exploit gains from market imperfections, or
originate them for their own benefit. To do so, acquiring physical goods, specialized
labour, know-how, in R&D, through mergers or joint ventures, in order to achieve long-
term goals.
Empirical and theoretical research have underlined the importance of innovation for the
survival and growth of firms (e.g. Audretsch, Coad and Segarra, 2014; Baumol, 2002;
Cefis and Marsili, 2006; Schumpeter, 1942; Zucchella and Siano, 2014). Within this
framework Schumpeter’s work is considered pioneer in recognizing innovation as
fundamental for the growth and survival of firms competing in the market (Audretsch
et al., 2014; Cefis and Marsili, 2006). By allowing cost reduction and/or product
differentiation, innovation can be an important tool for the competitiveness of
companies, ensuring them a competitive advantage (Genc et al., 2019). Furthermore,
Zucchella and Siano (2014) stress that the creation and development of SMEs are
commonly based on innovations.

485
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In this discussion, the distinction between different types of innovation cannot be left
out. Schumpeter (1934) proposed five types of innovations: product innovation, process
innovation, market innovation, input innovation and organizational innovation. OECD
(2005) consider the following types of innovation: product innovation, process
innovation, marketing innovation and organisational innovation. The success of firms
can differ according to innovation type. Heunks (1998), for example, concludes that any
kind of innovation contributes to small firms’ growth but only process innovation
increases their productivity. However, the author underlines that profits tend to be low
in innovating small firms due to the price of innovative investments and/or because that
was the main reason for innovation. For medium enterprises the results found by
Heunks (1998) for the relation between innovation types and firms success are not
statistically significant (at 0.05 level). Moreover, his study shows that medium-sized
firms exhibit higher levels of all types of innovation than small firms which can be
pointed as an explanation for why only in small enterprises success depends on
innovation. Also the classification of innovations as to the degree of novelty should be
considered: incremental vs. radical. An incremental innovation refers to an innovation
with low level of novelty while an innovation is considered radical when it comprises a
high level of novelty. Furthermore, radical innovations imply big changes (correspond to
something completely new) and incremental innovations are related to a process of
continuously improvement, for example, step-by-step improvements of existing
products (Eiriz, Faria and Barbosa, 2013; Sarkar, 2014; Tidd et al., 2018). Thus, radical
innovations can be associated with the Schumpeterian term of ‘creative destruction’,
that is, from Schumpeter’s original view incremental innovation does not exist
(Schumpeter, 1934, 1942). The contribution of innovations to firms’ growth can be
different if strategies are based on radical or incremental innovations (e.g. Eiriz et al.,
2013).

In a global market context, innovation seems to be increasingly important. Innovative

resources and capabilities are considered essential for firms’ growth in both domestic
and foreign markets (Zucchella and Siano, 2014). Analysing the product life cycle can
help to understand the importance of innovation for business performance: on the one
hand, as these cycles are getting shorter, it is imperative that companies continually

486
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

explore new products (or services) to offer; on the other hand, according to Vernon's
Product Life Cycle Theory, firms' internationalization process is related to the product
life cycle (Genc et al., 2019; Saridakis et al., 2019). Furthermore, there is broad
consensus among the researchers that internationalization and R&D investment /
innovation are two key strategic decisions to the success of firms (Chang et al., 2018;
Gjergji et al., 2019). In addition, the combined effect of innovation and
internationalization is viewed as a driving force of companies’ success and survival in the
global markets (Onetti et al., 2012). The relationship between these two factors is even
described as a dynamic virtuous circle given that internationalization and innovation are
mutually reinforcing (Golovko and Valentini, 2011).

Studies about the link between innovation and internationalization are relatively recent.
Additionally, although most studies examine large multinationals, a stream of literature
has analysed the relationship between innovation and internationalization with a special
focus in SMEs. In particular, the innovation contribution to SMEs internationalization
process has been arousing the interest of researchers in recent years (Saridakis et al.,
2019). However, the innovation–internationalisation link is not a straightforward issue
and the referred relationship is not a matter of consensus among researchers. If, on the
one hand, internationalization can be considered in itself a process of innovation and it
allows the access to diverse sources of knowledge stimulating innovation, on the other
hand, the development of innovations can create or improve market opportunities
(Zucchella and Siano, 2014). Some of published studies focus on the impact of
innovation on companies’ internationalization while other address the reverse causality
and in each case can be discussed if there is a direct or an indirect influence. There are
also works that explore a bidirectional link and/or the complementary effect of the two
for example on firm’s growth (Gjergji et al., 2019). According to Genc et al. (2019) most
of the existing research about this relationship examine the influence of innovation on
internationalization and only few studies investigates the opposite impact. Recent
studies (e.g. Genc et al., 2019 and Saridakis et al., 2019) emphasize the lack of consensus
among the literature that have studied this relationship, namely the effect of innovation
on internationalization, given that empirical studies present mixed results about this
relationship: they find a positive, negative or even a not significant effect between these

487
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

two factors. The arguments pointed out as justification for the different results found
include the industry context, the innovation measurement, especially when the focus
are the SMEs, and how the studies address the type of innovation. Gjergji et al. (2019)
present the main empirical findings about the relationship between internationalization
and innovation distinguishing the different branches of research that have addressed
this relationship. They point out that the most investigated are the influence of
innovation on internationalization and the reverse direction of the relationship.
Nevertheless, regarding the sign of the relationship the authors conclude that there is a
broad consensus among researchers with respect to the positive impact of innovation
on internationalization but the results on the reverse causal effect are controversial.
Empirical studies on innovation-internationalization reciprocal causal relationship find a
positive sign for the relationship. Evidence on innovation-internationalization
complementarity and simultaneity impact on firm’s performance and/or growth also
shows a positive sign. An important note is that most of the studies analysed measure
internationalisation through exports (in particular, export propensity).

As previously mention, some researchers study how innovation contributes to

internationalization. Focusing on SMEs in the textile and clothing industry in the
Campania Region (Italy), Zucchella and Siano (2014) analyse the role played by
innovation in driving export performance. Meliá, Pérez and Dobón (2010) investigate
the influence of innovation orientation on the internationalization of Spanish SMEs
within the service sector concluding that an innovative orientation contributes for an
early foreign market entry and allows them to choose higher-control entry modes.
Williams and Shaw (2011) discuss the innovation – internationalization relationship in
tourism sector, emphasizing, for instance, that successful internationalization requires
innovation. Using data from SMEs in the United Kingdom, Saridakis et al. (2019) study
the role of innovation on SMEs internationalization paying special attention to
innovation’s degree of novelty and the different types of innovation. They find that
although innovative SMEs are more likely to internationalise than non-innovative ones,
the results differs according to the type of innovation (for example, the propensity to
export is higher in case of goods innovation than in case of service or process innovation)
and the degree of novelty of innovation (for example, radical innovations or the

488
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

combination of radical and incremental innovations have a higher effect on the

probability of SMEs exporting than incremental innovations).

On the other hand, another set of studies investigate the influence of

internationalization on innovation. For example, Boermans and Roelfsema (2015)
examine how various internationalization modes affect innovation in ten Eastern
European and Central Asian transition economies and Abubakar, Hand, Smallbone and
Saridakis (2019) investigate the influence of internationalization modes on SME
innovation in Sub-Saharan least developed countries. Chang et al. (2018) and Genc et
al. (2019), for instance, analyse the impact of internationalization on innovation
considering multinational enterprises and SMEs, respectively. The latter conclude that,
in general, there is a positive impact of internationalization on innovation and
demonstrate that this impact is indirect through market orientation and entrepreneurial
orientation, which are considered by the authors as key factors to internationalized
SMEs obtain better innovation performance. Using a sample of 2420 Spanish private
firms, Gjergji et al. (2019) find not only a positive effect of the level of exports on the
number of innovative products of firm but also a positive impact of family management
on innovation, underlining that family firms have particular characteristics that can
boost the impact of internationalization on innovation.

Additionally, other authors, such as Bagheri, Mitchelmore, Bamiatzi and Nikolopoulos

(2019) and Golovko and Valentini (2011), highlight the importance of combining
innovation and international activities.

Notwithstanding, it seems to be consensual that one of the advantages providing by

internationalization is the access to/acquisition of new resources (Bagheri et al., 2018;
Genc et al., 2019; Saridakis et al., 2019). Aiming to corroborate the positive influence of
internationalization on innovation, Gjergji et al. (2019) underline the importance of the
resources acquired and developed during the internationalization process, in particular,
the access to new technological know-how and knowledge achieved from new alliances
(resulting from interaction with foreign partners). Thus, the authors believe that
exporting firms can acquire and develop resources and skills which are essential for
innovation.

489
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

SMEs often lack resources and capabilities (Bagheri et al., 2018; Genc et al., 2019).
Consequently, it may be difficult to invest in research departments (Sarkar, 2014). In this
context, sharing innovation activities can be essential and partnerships developed in
international market can play an important role.

Methodology
Aiming to explore the relationship between innovation and internationalization within
Portuguese firms, in particular to analyse whether Portuguese entrepreneurs consider
innovation as an important factor in the process of internationalization, an online
questionnaire survey was conducted with several variables, based on the literature
review. The questionnaire was sent to all companies registered in the AICEP database of
Portuguese internationalized firms, by sending a link via e-mail and using the Google
Forms tool between May 2019 and January 2020.

Data collected from the 238 valid responses (Portuguese internationalized firms) were
treated by IBM SPSS Statistics 26.0 software through a quantitative approach based on
descriptive and inferential analysis.

This 238 Portuguese firms are from different sectors (mostly firms are from sectors
Manufacturing industries / Wholesale and retail trade; car and motorcycle repair /
Consulting, scientific, technical and similar activities) and different sizes. In terms of size:
36.1% are micro firms (< 10 persons), 40.3% small (10-50 persons),17.6% medium (50-
250 persons) and the remaining are large firms (>250 persons) 1.

The various determining factors of internationalization referred to as External Market

Characteristics have been classified on a 5-Likert point scale: "1-not important", "2-not
very important", "3-important", "4-very important" and "5-extremely important" and
two more options (term unknown and not applicable (NA)). It should be noted that the
most important factor is "Favourable perspectives of growth in a new market" with an
average of 3.86 (median 4), followed by "Follow up customers" with an average of 3.57
(median 4), so they are considered as very important factors. Considered as important

1
Classification according to Commission Recommendation of 6 May 2003 available at [Link]
[Link]/legal-content/PT/ALL/?uri=CELEX%3A32003H0361

490
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

factors are "Follow partners" (average 3.11, median 3), "Allow access to new
technologies or resources" (average 2.70, median 3) and "Weak competition in the
market" (average 2.68, median 3). It should also be noted that the "Follow competitors"
factor is considered to be of little importance (average 2.57, median 2). We can see some
of these results in Figure 1.

Figure 1- Median of the factors (characteristics of external market)

It should be noted that among all the factors mentioned above, "Allow access to new
technologies or resources" is the factor that characterises innovation that drives the
internationalisation of Portuguese companies, with 83.5% of respondents to the
questionnaire considering it to be at least important. Since 94% of the firms that
answered the questionnaire are micro firms or small/medium firms, we decided to
consider only these two groups, thus having a sample of 224 firms. Analysing now the
possibility of differences in the degrees of importance attributed to the factors
determining the characteristics of the external market (depending on the size of the firm
and/or the sector) we carried out non-parametric tests.

After applying the Mann-Whitney test, we can conclude that there are no significant
differences between the size of the firm (Micro/SMEs) versus the factor "Allow access
to new technologies or resources" (p-value=0.419>0.05). In a final remark, the size of
the company in which the respondents are included does not influence the importance
that they attach to the factor characterizing innovation.

491
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Regarding the sector in which the companies operate, we found that out of the 238
companies in the sample the majority (199 companies corresponding to 83.6% of the
total) belong to one of the 3 sectors: (A) Manufacturing industries (124), (B) Wholesale
and retail trade; car and motorcycle repair (51) and (C) Consulting, scientific, technical
and similar activities (24). Therefore, we now work with the recoded sector variable in
these 3 categories (setor_rec). Figure 2 presents the median for the several Factors of
External Market by sector.

Figure 2- Median of external market factors by sector

The Kruskal Wallis test, allows us to conclude at a level of significance of 10%, that are
significant differences between the sector of the firm versus the factor "Allow access to
new technologies or resources" (p-value=0.08<0.10). However, this test does not allow
us to conclude which sectors are different from the others. Therefore, for this
identification, we proceeded with multiple comparison tests.

492
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Multiple Comparisons
Dependent Variable: Rank of Fator
LSD
95% Confidence
Mean Interval
Difference (I- Lower Upper
(I) setor_rec (J) setor_rec J) Std. Error Sig. Bound Bound
Manufacturing Wholesale and retail 23,087761 11,056356 ,038
*
1,28307 44,89246
industries trade; car and
motorcycle repair
Consulting, scientific, 18,786290 14,821881 ,206 -10,44455 48,01713
technical and similar
activities
Wholesale and retail Manufacturing -23,087761* 11,056356 ,038 -44,89246 -1,28307
trade; car and industries
motorcycle repair Consulting, scientific, -4,301471 16,452384 ,794 -36,74790 28,14495
technical and similar
activities
Consulting, scientific, Manufacturing -18,786290 14,821881 ,206 -48,01713 10,44455
technical and similar industries
activities Wholesale and retail 4,301471 16,452384 ,794 -28,14495 36,74790
trade; car and
motorcycle repair
*. The mean difference is significant at the 0.05 level.

Table 1- Multiple comparisons between sector and factor “Allow access to new
technologies or resources”

Analysing the output of multiple comparisons (Table 1) it is then possible to conclude

that for a significance level of 5%, the statistically significant differences in the degrees
of importance assigned to the factor "Allow access to new technologies or resources"
occur between sectors (B)-Wholesale and retail trade; car and motorcycle repair and
(A)-Manufacturing industries (p-value=0.038).

Manufacturing industries attach a greater degree of importance to this factor that

characterizes innovation. This is because 63.7% of these industries consider this factor

493
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

at least important, while in the Wholesale and retail trade; car and motorcycle repair
this percentage drops to 49%. Moreover, 51% of the industries in the latter sector
consider the innovation factor to be of little or no importance (see Table 2). Finally, more
detailed information that better visualises the behaviour of the factor "Allow access to
new technologies or resources" in the three most highlighted sectors in the study can
also be seen in Figure 3.

Wholesale and retail

trade; car and
Manufacturing motorcycle repair
industries (A) (B)
Allow access to Not important Count 18 12
new % within setor_rec 14,5% 23,5%
technologies or Not very important Count 27 14
resources % within setor_rec 21,8% 27,5%
Important Count 44 17
% within setor_rec 35,5% 33,3%
Very important Count 29 7
% within setor_rec 23,4% 13,7%
Extremaly important Count 6 1
% within setor_rec 4,8% 2,0%
Total Count 124 51
% within setor_rec 100,0% 100,0%

Table 2- Contingency table between factor “Allow access to new technologies or

resources and sectors (A) and (B)”

494
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 3- Pyramid count of factor ““Allow access to new technologies or resources” by

sector.

Conclusion

This study explores the relationship between innovation and internationalization within
Micro/SMEs Portuguese firms, in particular it analyses whether Portuguese
entrepreneurs consider innovation as an important factor in the process of
internationalisation. To achieve these goals, a sample of 224 micro firms or
small/medium firms was used and descriptive and inferential data analysis techniques
were applied.

Consistent with the literature, the results show that Portuguese entrepreneurs
(Micro/SMEs) consider innovation as an important factor in the process of
internationalisation.

In particular, the findings reveals that "Allow access to new technologies or resources"
is the third most important factor (of those specified as characteristics of the external
market) after "Favourable perspectives of growth in a new market" and "Track
customers". Focusing on that factor that characterises innovation, "Allow access to new

495
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

technologies or resources", it is important to underline that 83.5% of respondents to the

questionnaire considering it to be at least important.

Moreover, considering different firm sizes it is possible to conclude that there are no
significant differences regarding the factor "Allow access to new technologies or
resources", which means that the size of the company in which the respondents are
included does not influence the importance that they attach to the factor characterizing
innovation.

When differentiating according to sectors in which the companies operate, and

considering the three sectors present in the sample ((A) Manufacturing industries (124),
(B) Wholesale and retail trade; car and motorcycle repair (51) and (C) Consulting,
scientific, technical and similar activities (24)), this study suggests that there are
statistically significant differences in the degrees of importance assigned to the factor
"Allow access to new technologies or resources" for sectors (B)-Wholesale and retail
trade; car and motorcycle repair and (A)-Manufacturing industries (p-value=0.038).
Manufacturing industries attach a greater degree of importance to this factor that
characterizes innovation.

This work is supported by FEDER funds from COMPETE 2020 and Portuguese funds - PORTUGAL
2020. Project IEcPBI - Interactive Ecosystem for Portuguese Business Internationalization - POCI-
01-0145-FEDER-032139.

References
Abubakar, Y. A., Hand, C., Smallbone, D., & Saridakis, G. (2019). What specific modes of
internationalization influence SME innovation in Sub-Saharan least developed countries
(LDCs)?. Technovation, 79, 56-70.

Audretsch, D. B. , Coad, A. & Segarra, A. (2014). Firm growth and innovation. Small business
economics, 43(4), 743-749.

Bagheri, M., Mitchelmore, S., Bamiatzi, V., & Nikolopoulos, K. (2019). Internationalization
orientation in SMEs: The mediating role of technological innovation. Journal of International
Management, 25(1), 121-139.

Baumol, W. J. (2002). Free Market Innovation Machine: Analyzing the Growth Miracle of
Capitalism. Princeton Univ Press.
Boermans, M. A., & Roelfsema, H. (2015). The effects of internationalization on innovation: Firm-
level evidence for transition economies. Open Economies Review, 26(2), 333-350.

496
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Cefis, E., & Marsili, O. (2006). Survivor: The role of innovation in firms’ survival. Research
policy, 35(5), 626-641.
Chang, C. H., Chang, C. H., Hsu, P. K., & Yang, S. Y. (2019). The catalytic effect of
internationalization on innovation. European Financial Management, 25(4), 942-977.
Czinkota, M., Ronkainen, I. and Moffett, M. (1999). International Business, 5th ed., The Dryden
Press, Orlando.

Dunning, J., & Lundan, S. M. (2008). Multinational Enterprises and The Global Economy, 2ª
edição. Cheltenham: Edward Elgar.

Eiriz, V., Faria, A., & Barbosa, N. (2013). Firm growth and innovation: Towards a typology of
innovation strategy. Innovation, 15(1), 97-111.
Genc, E., Dayan, M., & Genc, O. F. (2019). The impact of SME internationalization on innovation:
The mediating role of market and entrepreneurial orientation. Industrial Marketing
Management, 82, 253-264.
Gjergji, R., Lazzarotti, V., Visconti, F., & Saha, P. (2019). Internationalization and innovation
performance: the role of family management. Economia Aziendale Online-, 10(2), 321-343.
Golovko, E., & Valentini, G. (2011). Exploring the complementarity between innovation and
export for SMEs’ growth. Journal of international business Studies, 42(3), 362-380.
Heunks, F. J. (1998). Innovation, creativity and success. Small Business Economics, 10(3), 263-
272.
Lobo, C., Fernandes, C., Ferreira, J. & Peris-Ortiz, M. (2019). Factors affecting SMEs’ strategic
decisions to approach international markets. European Journal of International
Management. Inderscience Publisher. DOI: 10.1504/EJIM.2020.10018550

Mathews, J. (2006) ‘Dragon multinationals: new players in 21st century globalization’, Asia
Pacific Journal of Management, Vol. 23, No. 1, pp.5–27.

McDougall, P., & Oviatt, B. M. (2000). International entrepreneurship: the intersection of two
research paths. Academy of Management Journal, vol. 43, nº 5, 902–906.

McDougall, P., Shane, S., & Oviatt, B. M. (1994). Explaining the formation of international new
ventures: The limits of theories from international business research. Journal of Business
Venturing, vol. 9, nº 6, 469-487.

Meliá, M. R., Pérez, A. B., & Dobón, S. R. (2010). The influence of innovation orientation on the
internationalisation of SMEs in the service sector. The Service Industries Journal, 30(5), 777-
791.
OECD (2005). The Measurement of Scientific and Technological Activities: Guidelines for
Collecting and Interpreting Innovation Data: Oslo Manual, Third Edition prepared by the
Working Party of National Experts on Scientific and Technology Indicators, OECD, Paris, para.
146.
Onetti, A., Zucchella, A., Jones, M. V., & McDougall-Covin, P. P. (2012). Guest editor’s
introduction to the special issue: entrepreneurship and strategic management in new
technology based companies. Journal of Management & Governance, 16(3), 333-336.

497
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Sapienza, H. J., Autio, E., George, G., & Zahra, S. A. (2006). A Capabilities Perspective on the
Effects of Early Internationalization on Firm Survival and Growth. Academy of Management
Review, 31, 914-933.

Saridakis, G., Idris, B., Hansen, J. M., & Dana, L. P. (2019). SMEs' internationalisation: When does
innovation matter?. Journal of Business Research, 96, 250-263.
Sarkar, S. (2014). Empreendedorismo e inovação. Third Edition. Escolar Editora.
Schumpeter, J. (1934). The Theory of Economic Development, Harvard University Press,
Cambridge, Massachusetts.
Schumpeter, J. A. (1942). Capitalism, Socialism and Democracy. New York: Harper & Row.

Tidd, J., & Bessant, J. R. (2018). Managing innovation: integrating technological, market and
organizational change. John Wiley & Sons.

Williams, A. M., & Shaw, G. (2011). Internationalization and innovation in tourism. Annals of
Tourism Research, 38(1), 27-51.

Zucchella, A., & Siano, A. (2014). Internationalization and innovation as resources for SME
growth in foreign markets: a focus on textile and clothing firms in the Campania
Region. International Studies of Management & Organization, 44(1), 21-41.

498
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ONLINE SHOPPING DRIVERS AND BARRIERS IN THE TIME OF COVID-19

FOR TURKISH YOUNG AND OLDER ADULTS

MUGE AKKOR KOKTEKIN

YEDITEPE UNIVERSITY

INSTITUTE OF SOCIAL SCIENCE

5/5/2020

499
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ABSTRACT

COVID-19 is a global outbreak which has spread all over the world. During the time of

COVID-19, young and older consumers are forced to modify their lifestyle habits until things

take a turn for the better. Among Turkish older consumers, this pandemic increases challenges

to adopt the use of e-commerce as they are quarantined and enter lockdowns. Therefore, older

adults are becoming more and more serious potential market for online shopping.

Unfortunately, previous practitioners and researchers have been paying less attention to the

online shopping behavior of older adults. To bridge the gap, the aim of this study is to enhance

a better understanding of the differences between the online shopping drivers and barriers of

both young and older adults based on the Unified Theory of Acceptance and Use of Technology

and the Innovation Resistance Theory particularly in the time of COVID-19. Unlike the

previous studies, this study not only focuses on drivers and barriers but also differentiates itself

based on the integrated model depending on composite theories.

Keywords: Online shopping, Unified Theory of Acceptance and Use of Technology (UTAUT),

Innovation Resistance Theory, COVID-19

500
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1 - INTRODUCTION

The World Health Organization declared a novel coronavirus, COVID-19, outbreak a

Public Health Emergency of International Concern on January 30th, 2020 (WHO, 2020).

Governments worldwide were prompt to implement a variety of actions. As part of the

precautions against the coronavirus pandemic, the Ministry of Interior in Turkey imposed a

partial curfew for the citizens aged 65 and over, and the ones with chronic illnesses, effective

as of March 22, 2020. Among Turkish older consumers, this pandemic increases challenges to

adopt the use of e-commerce as they are quarantined and enter lockdowns. Older adults are

becoming more and more serious potential market for online shopping. Unfortunately,

practitioners and researchers have paid less attention to the older adults’ online behaviors, and

rather they have focused on young adults.

In addition, in order to prevent the spread of the virus, the Ministry of Interior in Turkey

imposed not only a 48-hour curfew for all citizens in 31 cities across the country on April 11-

12, 2020 and on April 17-18, 2020, but also a 4-day curfew on April 23-26, 2020 and a 3-day

curfew on May 1-3, 2020 in the same 31 cities. Therefore, consumers at every age group are

forced to modify their lifestyle habits until things take a turn for the better. In this sense, there

is a great online selling opportunity especially for FMCG manufacturers and retailers since

curfews seem to continue in the near future. Turkey's B2C e-commerce market revenue

amounts to USD 10,657 million in 2020 and is expected to reach USD 21,769 million by 2024,

with an annual growth rate of 19.5% (Statista, 2020). In order to get a share from this growth,

FMCG manufacturers and retailers urgently need to come up with a more dynamic e-commerce

strategies considering online shopping drivers and barriers in the time of COVID-19 for young

and older adults. Therefore, the aim of this study is to enhance a better understanding of the

online shopping drivers and barriers influencing both young and older Turkish consumers in

the time of COVID-19.

501
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Unlike this study, most of the previous studies had concentrated on driver factors instead

of the barriers. Furthermore, this study differentiated itself based on the integrated model

depending on composite theories since most of the previous studies had applied single theory

instead of composite theories. To this end, this study is developed by integrating the Unified

Theory of Acceptance and Use of Technology and the Innovation Resistance Theory.

502
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2 - LITERATURE SYNTHESIS

E-commerce, also known as internet commerce or electronic commerce, is a commercial

transaction conducted over the internet and value added networks. Instead of going to a brick

and mortar store, looking for a parking lot, wandering in the aisles of the malls, searching for

the right products, waiting in lines to make a payment, shopping online saves one a lot of time

and effort. Some examples of the most popular websites are Amazon, eBay, Alibaba in the

world, and Hepsiburada, n11, Trendyol in Turkey.

2.1 Definition of e-Commerce

E-commerce is as the purchasing and selling of products, services and information

through utilizing any one of the thousands of computer networks forming the internet as defined

by Lawrence et al. in 1998 (Steve Lawrence, 1998). In addition, online banking, buying or

selling goods and services, arranging and paying for travel and accommodation, online share

purchase, soliciting investment advice, and the use of Automatic Teller Machines (ATMs) are

among typical e-commerce activities (Tatnall & Lepa, 2003).

2.2 The Origin of e-Commerce

For the first time, e-commerce has started in the 1960s through electronic data

interchange on value added networks. Next, ATMs were used by customers in order to do

banking by credit cards and point of sale terminals, which was followed by inter-organizational

systems. Organizations were able to conduct business and exchange information by these

systems (Molla & Licker, 2001). Amazon introduced a book shipping business in 1995 in the

garage of Jeff Bezos. In the same year, eBay began online auctions. After the above mentioned

first online shopping experiences, e-commerce continued growing. For instance, the online

sales turnover in the U.S.A. was $601.75 billion in 2019, up by 14.9% from $523.64 billion in

2018, according to U.S. Department of Commerce quarterly e-commerce figures on February

503
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

19, 2020, which was a higher growth rate than 2018, when online sales increased by 13.6%

year over year (Young, n.d.).

In Turkey, e-commerce activities began in the late 1990s. The e-commerce industry

grew by 35.5% year on year from 2008 to 2012 as the number of companies had been shifting

to e-commerce (Bafra, 2019). According to the annual report prepared by Informatics Industry

Association and Deloitte Turkey (Informatics Industry Association (TUSIAD), 2019), Turkish

e-commerce sector reached to TRL 59.9 billion in 2018 with 42% annual growth and the

average annual growth rate of Turkish e-commerce sector was 33% between the years 2014 and

2018.

In 2019, the global B2C e-commerce market was estimated to reach approximately $3.5

trillion (Informatics Industry Association (TUSIAD), 2019). China and USA are the leaders in

the market while opening the gap with other countries. Although Turkey does not have a

considerable share in the global e-commerce market yet, it has an important potential thanks to

relatively young population, individual internet penetration, increase in mobile broadband

subscriptions and the advantage of Turkey’s geographical location.

2.3 Use of e-Commerce among Different Age Groups: World vs. Turkey

All around the world, according to Verto Analytics data (Hwong, 2018), the majority of

consumers who shop online are under the age of 75. 95% of online shoppers are between the

ages of 18 and 74 and only 5% of online shoppers are older than 75 years old. Millennials (also

called Generation Y born between 1980 and 2000) are slightly underrepresented, accounting

for only 30% of the online shopping population. Meanwhile, members of Generation X (the

ones born between 1965 and 1979) account for 34% of the online shopping population,

followed by Baby Boomers (the ones born between 1946 and 1964), who are 31% of the online

shopping population.

504
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In Turkey, the highest rate of the online shoppers are between the ages of 25 and 34.

36% of people between 25 and 34 years choose to shop online based on the data from the

Interbank Card Center (Taylan, 2015). This figure is 32% for the ones between 18 and 24 years

old, 24% for the ones between 35 and 44, 18% for the ones between 45 and 54, and 11% for

the ones older than 55 years old.

Consequently, when the use of e-commerce across different age groups in Turkey is

compared with the world statistics, the major difference is the rate of use by the members of

the Generation Y. Although Turkish people ages between 18 and 35 account for 68% of the

online shopping population, Generation Y in the world accounts for only 30% of the online

shopping population. The sum of Generation X and Baby Boomers in the world (ages between

41 and 74) account for 65% of the online shopping population whereas Turkish people aged 45

and above account for only 22% of the online shopping population.

2.4 The Driving Factors to Use e-Commerce

Research from the “drivers” perspective is summarized below. First of all, Kwon and

Noh investigated the effect of prior experience and age on mature American consumers’ (the

ones born before 1964) online clothing shopping perceptions and intentions (Kwon & Noh,

2010). The results of their research indicated that perception of consumer regarding price

discounts, product benefits, and financial risk would influence their intention to use e-

commerce for clothing. Furthermore, the authors added that preceding e-commerce experiences

of the mature consumers would affect their perceived benefits and risks. On the other hand, it

was indicated that online shopping intentions were not affected by online experience and age.

Secondly, McCloskey used and modified the Technology Acceptance Model to

investigate the impact attitudes regarding usefulness, ease of use, and trust on the usage of e-

commerce by older Americans (McCloskey, 2006). The user behavior was affected positively

505
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

by trust for the website and usefulness of the website. Moreover, ease of use had important

effect on the perception of usefulness. In addition to this, users’ perception of usefulness and

ease of use are affected by trust.

Finally, Lian and Yen used both Innovation Resistance Theory and the UTAUT model

in order to find the drivers and barriers influencing the intention of older consumers to shop

online (Lian & Yen, 2014). Meanwhile, they made a comparison of online shopping drivers

and barriers for the older adults (ages between 56 and 70) and the younger adults (ages between

21 and 25). The most significant factors driving both older and younger adults to do online

shopping were social influence (The friends of the user and/or a person who is very important

to the user think that the user should shop online and many of the friends of the user shop online

overall.) and performance expectation (Online shopping is helpful for the users and allows the

users to buy more quickly and efficiency.) (Lian & Yen, 2014).

2.5 Barriers Preventing Adults to Use e-Commerce

Studies regarding the “barriers” preventing adults from adopting e-commerce are

summed up as below.

First of all, Molesworth and Suortti studied the adoption of the web for high-

involvement and high-cost purchases including buying car online by using the Innovation

Resistance Theory (Molesworth & Suortti, 2002). The authors used a sample of adults aged

between 20 and 57 years old. As a result of the study, the barriers were indicated as risk,

tradition, usability and image (Molesworth & Suortti, 2002).

Next, Laukkanen et al. investigated the causes restraining the adoption of mobile

banking for the mature consumers (older than 55 years old) and the younger consumers

(younger than 55 years old) based on Innovation Resistance Theory (T. Laukkanen, Marke,

Laukkanen, & Sinkkonen, 2007). The most intense barrier was found to be the value barrier

506
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

both for the mature and the younger consumers. The authors concluded that the barriers

regarding the adoption of mobile banking were value, usage, tradition, risk, and image for the

mature consumers whereas value and usage for the younger consumers (T. Laukkanen et al.,

2007).

Furthermore, Lian and Yen also found the similar results for the young and older adults

as Laukkanen et al. However, the authors indicated less barriers for the users compared to

Laukkanen et al. This could be due to different samples, location of the study, different e-

commerce type or 7 years of differences between 2 studies. Lian and Yen indicated that value

(buying online, which is economical, increases the ability of the user to control payment details

and product information), tradition (when the user needs to buy, s/he likes the staff to provide

services in a physical store), and risk (the user fears that while buying online, the connection

will be lost, and/or the user might type the product information incorrectly, and/or username

and password may be lost and end up in the wrong hands) are major barriers preventing older

adults from online shopping. On the other hand, the major barrier is only value for younger

adults (Lian & Yen, 2014). Besides, Kwon and Noh found that financial risk shows an important

effect onto e-commerce perception of the mature consumers (Kwon & Noh, 2010).

Finally, Laukkanen et al. analyzed the innovation resistance through dividing non-

adopters of internet banking (ages between 18 and 65) into three groups including opponents,

rejectors and postponers depending on their intentions to adopt the innovation (P. Laukkanen,

Sinkkonen, & Laukkanen, 2008). The result s of the study stated that the postponers presented

no significant resistance to adopt internet banking (P. Laukkanen et al., 2008). However, the

barriers to use the internet banking were tradition, risk and image for the opponents (P.

Laukkanen et al., 2008). In addition to this, the resistance of the rejectors was much more

diverse and intense than that of the two groups mentioned above, with tradition and risk to be

the most diverse and intense barriers (P. Laukkanen et al., 2008).

507
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

To sum up, the literature on drivers and barriers for young and older adults were outlined

in this chapter. Next, a theoretical background will be provided regarding the UTAUT model

and the Innovation Resistance Theory.

508
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3 - THEORETICAL BACKGROUND

This study is developed by integrating the Unified Theory of Acceptance and Use of

Technology and the Innovation Resistance Theory.

3.1 Unified Theory of Acceptance and Use of Technology

Technology Acceptance Model (TAM), Diffusion of Innovation Theory, Motivation

Theory, Theory of Reasoned Action (TRA), Theory of Planned Behavior (TBP), Information

Systems (IS) Success Model by DeLone and McLean, and Resource-Based Theory are widely

used theoretical models to understand the user behavior and technology acceptance. Venkatesh

and many other scholars consolidated the previous research including the understanding of user

intentions and behaviors for the acceptance and usage of new technology (Venkatesh, Morris,

Davis, & Davis, 2003). Therefore, they proposed the Unified Theory of Acceptance and Use of

Technology (UTAUT) aiming to demonstrate the user intentions to adopt an information

system and consequent usage behavior (Venkatesh et al., 2003). This model was tested and

verified by a longitudinal study, where the explanatory power of the UTAUT model was found

as high as 70% (the variance in Behavioral Intention to Use) (Venkatesh et al., 2003). The

theory states that there are four antecedent variables (Venkatesh et al., 2003) as follows:

1. Performance expectation indicates the degree to which users’ job performance will be

improved by IT.

2. Effort expectation is the level to which users expect the new product will be easy to

use.

3. Social influence indicates the degree to which users are expected to use the new

technology by their peers.

4. Facilitating conditions is the state to which users will be helped by the organizational

and technical infrastructure to use the innovation.

509
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1 Unified Theory of Acceptance and Use of Technology (Venkatesh et al., 2003)

As shown in Figure 1, the first three antecedent variables are direct determinants of the

behavioral intention and the use behavior. In addition, the fourth antecedent variable is a direct

determinant of the use behavior. Gender, experience, age and voluntariness of use are

positioned as a moderator to influence the four key constructs on the behavioral intention and

the use behavior.

3.2 Innovation Resistance Theory

The Innovation Resistance Theory (IRT) was proposed by Ram in 1987 (Ram, 1987).

This theory aims to clarify the reasons of users not accepting an innovation while focusing on

the characteristics of the innovation, marketing mechanisms, and users’ characteristics. The

conflicts between innovation and traditions create barriers for the adoption of the new

technology, and also increase resistance towards the innovation (Ram, S. & Sheth, 1989). There

are two types of barriers:

510
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1. Functional Barriers include usage, value and risk (Ram, 1987).

a) Usage: The innovative product inconsistency with the users’ general experiences and

values lead to a longer time to adopt the innovation.

b) Value: If the innovative product does not provide a higher or an equal value than does

the current product, the consumer will resist to accept the change.

c) Risk: If the new technology in the innovative product is not completely understood by

the user, the user will have uncertainties after using the product, and this will lead to

resistance for the innovation.

2. Psychological Barriers include tradition and image (Ram, 1987).

a) Tradition: When the current culture of the user is changed by the innovation and is

conflicted with the innovation, the resistance occurs. The stronger the conflict, the

greater the resistance will be.

b) Image: When the impression of the brand, side effects, industry, or the image of the

country of innovation are unfavorable, the image based barrier occurs.

Usage barrier
Functional
Value barrier
barriers
Consumer
Risk barrier
Resistance to
Innovation Tradition barrier
Psychological
barriers
Image barrier

Figure 2 Innovation Resistance Theory

After reviewing the UTAUT model and the IRT that this study is based on, the research

method is explained below.

511
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4 - RESEARCH METHOD

4.1 Research Model

Based on the above theoretical background, the research model is proposed below.

Drivers Barriers
1) Performance 1) Usage
expectation 2) Value
2) Effort expectation 3) Risk Intention
3) Social influence 4) Tradition
4) Facilitating conditions 5) Image

Figure 3 Research Model

4.2 Research Hypothesis

This study combines the UTAUT model as driver perspective and the Innovation

Resistance Theory as barrier perspective. The proposed drivers based on the UTAUT model

are performance expectation, effort expectation, social influence, and facilitating conditions.

On the other hand, the proposed barriers based on the Innovation Resistance Theory are usage,

value, risk, image, and tradition. This study aims to understand the differences between young

and older consumers as far the online shopping drivers and barriers of both groups are

concerned in the time of COVID-19. Therefore, the following hypotheses are proposed based

on the above discussion.

H1: The mean of the performance expectation (online shopping is helpful for the users,

and allows the users to buy more quickly and efficiently) is higher for young adults than for

older adults.

H2: The mean of the effort expectation (online shopping websites are easy to

understand, online shopping is easy to use and learning how to shop online is easy) is higher

for young adults than for older adults.

512
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

H3: The mean of the social influence (the friends of the user and a person who is very

important to the user think that the user should shop online, and many friends of the user shop

online overall) is higher for young adults than for older adults.

H4: The mean of the facilitating conditions (the user has not only the hardware and

software, but also the skills and the knowledge for online shopping, the experience of using

online shopping and internet is similar) is higher for young adults than for older adults.

H5: The mean of the usage barrier (the user believes that online shopping is easy,

convenient fast and clear, changing password and personal information in online shopping

environments is convenient) is lower for young adults than for older adults.

H6: The mean of the value barrier (buying online, which is economical, increases the

ability of the user to control payment details and product information) is lower for young adults

than for older adults.

H7: The mean of the risk barrier (the user fears that while buying online, the connection

will be lost, and/or the user might type the product information incorrectly, and/or username

and password may be lost and end up in the wrong hands) is lower for young adults than for

older adults.

H8: The mean of the tradition barrier (when the user needs to buy, the user likes the

staff to provide services in a physical store) is lower for young adults than for older adults.

H9: The mean of the image barrier (new technology is often too complicated to be useful

and online shopping services are difficult to use) is lower for young adults than for older adults.

H10: The mean of the intention (intention of shopping online in the future and/or

continuing to shop online to improve convenience) is higher for young adults than for older

adults.

The rate of new technology adoption and use of these technologies are higher in the

countries with young population compared to the countries with older population (Informatics

513
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Industry Association (TUSIAD), 2019). Since Turkish population is reasonably young meaning

that the median age of the population is at 30.9 years of age (“Turkey Population 2020

(Demographics, Maps, Graphs),” 2020) and since the share of the young internet users in the

population of Turkey by age group is more than 60% as shown in Table 1, the mean of the

drivers is expected to be higher for young adults than for older adults; whereas the mean of the

barriers is expected to be lower for young adults than for older adults.

4.3 Respondents

In order to validate this proposed research model, a survey study was conducted. In this

study, two groups of subjects are included: Young and older Turkish adults. In order to

distinguish the young and older adults, the percentage of the internet users in the population of

Turkey was investigated (“Internet usage: users by age group Turkey 2018 | Statista,” 2019)

and a clear boundary at age 45 was found (Table 1).

Table 1
Share of the internet users in the general population of Turkey by age group.

The young adults, younger than 44 years old and major internet users, were drawn the

Generation Y living in Turkey. Although different dates and names have been used to define

generational cohorts by historians and researchers, the literature has confirmed some standard

behavior patterns and attitudes of these segments (Markert, 2004; Norum, 2003). The

Generation Y has grown up with the internet. Therefore, the Generation Y, also called “digital

natives” is expected to be more tech-savvy (Prensky, 2001).

The older adults in this study, on the other hand, are people over the age of 45 living in

Turkey and able to use computer. Therefore, the older adults considered in this study are

514
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

belonging to the cohorts of the Generation X and the Baby Boomers since they are over 45

years old.

Consequently, this study compares two groups: older adults (Generation X and Baby

Boomers) and young adults (Generation Y), with respect to their online shopping drivers and

barriers in the time of COVID-19.

4.4 Measurements

In total, this research included ten variables (Table 2). Facilitating conditions, social

influence, effort and performance expectation were the four drivers, for which the measurement

items were adjusted from Venkatesh et al. (2003). On the other hand, image, tradition, risk,

value and usage were the five barriers for which the measurement items were modified from

Laukkanen et al. (2007). Besides, the measurement item of online shopping intention was

developed by Venkatesh et al. (2003). The survey including altogether 31 questions was used

in a study by Lian and Yen (Lian & Yen, 2014). The same measurement tools were used in this

study.

Furthermore, a five point Likert scale ranging from 1 (strongly disagree) to 5 (strongly

agree) was used for all measurements. A higher value not only shows that the respondent has

higher drivers and lower barriers but also indicates that the intention of respondents to shop

online is stronger. In addition, each variable was labeled and abbreviations were illustrated in

Table 2.

Finally, the independent t-test is used in this study in order to determine whether there

is a statistically significant difference between the means of drivers and barriers in two groups:

young and older Turkish consumers.

515
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2
Measurements.

Item
Dimension Variable Abbreviation Source
Number
Performance
Driver Expectation PE 3
Effort Expectation EE 4 Venkatesh et al.,
Social Influence SI 3 2003
Facilitating
conditions FC 4
Barrier Usage barrier UB 5
Value barrier VB 2
P. Laukkanen et al.,
Risk barrier RB 3
2008
Tradition barrier TB 2
Image barrier IB 2
Online shopping Venkatesh et al.,
Intention intention I 3 2003

516
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5 - DATA ANALYSIS AND RESULTS

Data collection was conducted from March 29th, 2020 to April 6th, 2020. Convenience

Sampling Method is used to choose the online sample in this study since the survey is conducted

through a website. In total, 176 participants responded to the online survey. After invalid

surveys were removed, this study received a sum of 172 valid surveys.

5.1 Demographics

Out of 172 respondents, 63% of them are between the ages 18 and 44 and 37% of them

are between the ages 45 and 74 (Table 3). 57% of the respondents are female whereas 43% of

the respondents are male (Table 4).

Table 3
Age distribution of respondents.

Age Number %
18-44 109 63%
45-74 63 37%
Total 172 100%

Table 4
Gender distribution among different age groups.

Gender
Age Female Male Total
18-44 53 56 109
% 48.6% 51.4% 100%
45-74 45 18 63
% 71.4% 28.6% 100%
Total 98 74 172
% 57.0% 43.0% 100%

When the online shopping experience had been asked to the respondents, 93% of them

had shopped online before compared with 7% having no online shopping experience (Table 5).

Among the different age groups, 96.3% of the young group (ages between 18 and 44) had online

shopping experience whereas 87.3% of the older group (ages between 45 and 74) had online

shopping experience before.

517
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 5
Online shopping experience among respondents.

Online Shopping Experience

Age Yes No Total
18-44 105 4 109
% 96.3% 3.7% 100%
45-74 55 8 63
% 87.3% 12.7% 100%
Total 160 12 172
% 93.0% 7.0% 100%

Males had a relatively more online shopping experience compared to their female

counterparts (Table 6).

Table 6
Online shopping experience among different genders.

Online Shopping Experience

Gender Yes No Total
Female 90 8 98
% 91.8% 8.2% 100%
Male 70 4 74
% 94.6% 5.4% 100%
Total 160 12 172
% 93.0% 7.0% 100%

5.2 Reliability

The survey including altogether 31 questions (Appendix A) was used before in 2014 by

Lian and Yen who tested both validity and reliability, conducted a pilot test for the questionnaire

and found the results as acceptable since measurements in this study were not only modified

but also extended from previous studies (Lian & Yen, 2014). Although the same measurements

were used in this study, Cronbach’s alpha of tradition barrier variable was found below 0.5.

Therefore, the second question of the tradition barrier scale was not used in calculations as

Nunnally explained that the minimum threshold of Cronbach’s alpha should be 0.5 or 0.6

(Nunnally, 1978). In addition, in order to increase the Cronbach’s alpha of facilitating

conditions variable above 0.70, the fourth question of facilitating conditions’ scale was deleted.

In this regard, the threshold for Cronbach’s alpha in this study was over 0.70 since the

518
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

acceptable range for Cronbach’s alpha test of reliability was above 0.70 (Cortina, 1993).

Therefore, the variables in this study were found as reliable (Table 7).

Table 7
Reliability.

Variables Cronbach's Alpha

Performance Expectation 0.76
Effort Expectation 0.81
Social Influence 0.72
Facilitating conditions 0.78
Usage barrier 0.84
Value barrier 0.70 '

Risk barrier 0.77

Image barrier 0.76
Online shopping intention 0.89

5.3 Results

The variables were measured by using a five point Likert scale. Therefore, the values of

all variables were between 1 (strongly disagree) and 5 (strongly agree). A higher value not only

showed that the respondent had higher drivers and lower barriers; but also indicated that the

respondents’ intention to shop online was stronger. The proposed hypotheses were tested by an

independent sample t test. The results were summarized in Table 8.

Table 8
Hypotheses testing results.

Hypotheses Result
H1: The mean of the performance expectation is higher for young adults than for
Supported
older adults.
H2: The mean of the effort expectation is higher for young adults than for older
Supported
adults.
H3: The mean of the social influence is higher for young adults than for older
Supported
adults.
H4: The mean of the facilitating conditions is higher for young adults than for older
Supported
adults.
H5: The mean of the usage barrier is lower for young adults than for older adults. Supported
H6: The mean of the value barrier is lower for young adults than for older adults. Supported
H7: The mean of the risk barrier is lower for young adults than for older adults. Rejected
H8: The mean of the tradition barrier is lower for young adults than for older adults. Supported
H9: The mean of the image barrier is lower for young adults than for older adults. Supported
H10: The mean of the intention is higher for young adults than for older adults. Rejected

519
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 9 shows that young adults had significantly high drivers and lower barriers than

do the older adults (p is smaller than 0.05) other than the risk barrier and intention. Therefore,

hypotheses 7 and 10 were not supported (Table 8).

The analysis results present that both young and older adults consider tradition as the

strongest barrier due to its mean being the lowest between other variables (Table 9). Therefore,

both young and older adult still believe that when they need to buy, they like the staff to provide

services in a physical store. However, the analysis results show that young and older adults

agree on facilitating conditions as the strongest driver due to its mean being the highest among

other variables (Table 9). This means that both young and older adults have not only the

hardware and software, but also the skills and the knowledge for online shopping. Plus, both

subject groups agree that the experience of using online shopping and internet is similar.

Table 9
Independent sample t test between different age groups.

Standard
Variables Age N Mean Significance
Deviation
Young 109 4.3028 0.62120
PE 0.000 ***
Older 63 3.8307 0.73048

Young 109 4.0872 0.60799

EE 0.014 ***
Older 63 3.8413 0.64805

Young 109 3.7859 0.67514

SI 0.004 ***
Older 63 3.4339 0.87300

Young 109 4.3853 0.67155

FC 0.004 ***
Older 63 4.0635 0.75457

Young 109 4.1376 0.58922

UB 0.002 ***
Older 63 3.8222 0.66828

Young 109 3.8028 0.80513

VB 0.030 ***
Older 63 3.5238 0.80537

Young 109 3.2385 0.82414

RB 0.054
Older 63 2.9683 0.97038

Young 109 2.9817 0.96207

TB 0.022 ***
Older 63 2.6349 0.92111

Young 109 3.7248 0.86999

IB 0.032 ***
Older 63 3.4286 0.85127

Young 109 4.3333 0.70711

I 0.254
Older 63 4.2063 0.69155

***p<0.05.

520
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

About the gender differences among young adults, the analysis results show that male

young adults have significantly higher intention of shopping online compared to their female

counterparts (Table 10).

Table 10
Independent sample t test between different genders among young adults (18-44).

Standard
Variables Gender N Mean Significance
Deviation
Female 53 4.2956 0.68144
PE 0.908
Male 56 4.3095 0.56446

Female 53 4.1226 0.53848

EE 0.556
Male 56 4.0536 0.67034

Female 53 3.6855 0.61816

SI 0.132
Male 56 3.8810 0.71754

Female 53 4.3711 0.70305

FC 0.830
Male 56 4.3988 0.64642

Female 53 4.2000 0.52035

UB 0.284
Male 56 4.0786 0.64687

Female 53 3.6981 0.88979

VB 0.188
Male 56 3.9018 0.70980

Female 53 3.2453 0.68298

RB 0.934
Male 56 3.2321 0.94478

Female 53 3.1321 0.89952

TB 0.113
Male 56 2.8393 1.00502

Female 53 3.8774 0.61163

IB 0.071
Male 56 3.5804 1.04349

Female 53 4.1887 0.81259

I 0.037 ***
Male 56 4.4702 0.56417

***p<0.05.

Likewise, about the gender differences among older adults, the analysis results show

that male older adults have significantly higher intention of shopping online compared to their

female counterparts (Table 11). In addition, about the gender differences among older adults,

male older adults have significantly higher social influence driver compared to female older

adults. Therefore, male older adults believe in the thought of their friends and/or a beloved

person that they should shop online. In addition to this, many friends of male older adults shop

online overall.

521
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 11
Independent sample t test between different genders among older adults (45-74).

Standard
Variables Gender N Mean Significance
Deviation
Female 45 3.8000 0.81464
PE 0.602
Male 18 3.9074 0.46870

Female 45 3.8389 0.70742

EE 0.964
Male 18 3.8472 0.48612

Female 45 3.2889 0.91453

SI 0.036 ***
Male 18 3.7963 0.64816

Female 45 4.0000 0.76541

FC 0.295
Male 18 4.2222 0.72310

Female 45 3.8267 0.70659

UB 0.934
Male 18 3.8111 0.57995

Female 45 3.5222 0.74586

VB 0.981
Male 18 3.5278 0.96211

Female 45 2.9259 1.06337

RB 0.521
Male 18 3.0741 0.70066

Female 45 2.6222 0.93636

TB 0.864
Male 18 2.6667 0.90749

Female 45 3.4222 0.89160

IB 0.926
Male 18 3.4444 0.76483

Female 45 4.0889 0.70853

I 0.032 ***
Male 18 4.5000 0.56302

***p<0.05.

522
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

6 - CONCLUSION

After reviewing the theoretical background and the literature, this study combines the

UTAUT model as driver perspective and the Innovation Resistance Theory as barrier

perspective to investigate a better understanding of the differences between the online shopping

drivers and barriers of both young and older adults in the time of COVID-19. In order to fulfill

the aim of this study by testing the hypotheses, the independent sample t-test is used to

determine whether there is a statistically significant difference between the means of drivers

and barriers in two subject groups living in Turkey.

The first finding of this research is that young adults have significantly high drivers and

lower barriers than do the older adults in the time of COVID-19 other than the risk barrier and

intention. Both young and older adults consider tradition as the strongest barrier. When they

need to buy, they like the staff to provide services in a physical store. Tradition, including the

staff serving them in the brick and mortar stores, is such a high barrier that even the closed

stores and shopping malls in the time of COVID-19 cannot overcome this barrier. On the other

hand, both young and older adults agree on facilitating conditions as the strongest driver. This

means that they have not only the hardware and software, but also the skills and the knowledge

for online shopping. Plus, both subject groups agree that the experience of using online

shopping and internet is similar.

The second finding of this research is that among both young and older adults, male

adults have significantly higher intention of shopping online compared to their female

counterparts.

The third finding is that about the gender differences among older adults, male older

adults have significantly higher social influence driver compared to female older adults.

523
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Therefore, male older adults pay more attention to their friends and family’s experiences on

shopping online.

Finally, according to Eurostat data of the EU Statistical Office, the percentage of online

shoppers (among all individuals, aged between 16 and 74 years old, who shopped online at least

one within a year), which was 3% in 2009, increased to 30% in 2019 (Eurostat, 2020). In 2020,

the pandemic of COVID-19 has an effect on each aspect of life including what we buy, how we

shop, and where we shop. In Turkey, the online sales have been increasing due to home office

way of studying especially in the private sector (Gunyol, 2020). Many consumers have switched

to digital channels, services, and products. Thinking about the “new normal”, consumers remain

doubtful whether they will carry the same activities which were part of their life before the

beginning of the outbreak. The pandemic will definitely have consequences that last longer.

524
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ACKNOWLEDGMENTS

The author would like to thank to Prof. Dr. Tulin Ural, Associate Prof. Dr. Yusuf Can

Erdem and Mr. Emrah Koktekin for their valuable support, great motivation, consistent

feedback and outstanding proof reading skills.

525
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

APPENDIX A: MEASUREMENT ITEMS

Dimension Variable # Questions
Performance 1 Shopping online is helpful to me.
Expectation 2 Online shopping allows me to buy more quickly.
(PE) 3 Online shopping allows me to buy more efficiency.
When I interact with online shopping websites, they
4
Effort are always clear and easy to understand.
Expectation 5 I am familiar with online shopping and find it easy.
(EE) 6 I feel that online shopping is easy to use.
7 Learning how to shop online is easy.
8 My friends think that I should shop online.
Driver Social A person who is very important to me believes that I
9
Influence (SI) should shop online.
10 Overall, my of my friends shop online.

11 I have the hardware and software for online shopping.

Facilitating 12 I have the skill and knowledge for online shopping.
Conditions The experience of using online shopping is similar to
(FC) 13
using the internet.
When I have problems shopping online, someone can
14
help me solve them.
15 In my opinion, shopping online is easy.
16 In my opinion, shopping online is convenient.

Usage Barrier 17 In my opinion, online shopping services are fast.

(UB) 18 In my opinion, the online shopping process is clear.
The system for changing my password and personal
19 information in online shopping environments is
convenient.
20 Buying online is economical.
Value Barrier In my opinion, buying online increases my ability to
(VB) 21 control payment details and product information on my
own.
Barrier I fear that while I am buying online , the connection will
22
be lost.
Risk Barrier I fear that while I am buying online , I might type out
23
(RB) the information of the product incorrectly.
I fear that my username and password may be lost and
24
end up in the wrong hands.
When I need to buy, I like the staff to provide services
Tradition 25
in a physical store.
Barrier (TB)
26 When I need to buy, I like online self-service.
In my opinion, new technology is often too complicated
Image 27
to be useful.
Barrier (IB) I have the impression that online shopping services are
28
difficult to use.
29 I intend to shop online in the future.
Online
Intention Shopping 30 I predict I would shop online in the future.
Intention (I) I intend to continue shopping online to improve
31
convenience.

526
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

REFERENCES
Bafra, T. (2019). E-Commerce in Turkey: Potential for Development and Logistics Processes.
Istanbul.

Cortina, J. M. (1993). What is coefficient alpha? An examination of theory and applications.

Journal of Applied Psychology, 78(1), 98–104.

Eurostat. (2020). E-banking and e-commerce - Eurostat. Retrieved May 4, 2020, from
[Link]

Gunyol, A. (2020). Koronavirüs ile birlikte Türkiye’de e-ticaret satışları arttı. Retrieved April
29, 2020, from [Link]
ticaret-satislari-artti-/1772734

Hwong, C. (2018). Chart of the Week: E-commerce demographics - who shops online?
Retrieved February 21, 2020, from Consumer Insights website:
[Link]

Informatics Industry Association (TUSIAD), D. (2019). E-Ticaretin Gelişimi, Sınırların

Aşılması ve Yeni Normlar. 1–114. Retrieved from [Link]
bultenleri/item/download/9232_e0200c4e2deaf513f571f854f11f6192

Internet usage: users by age group Turkey 2018 | Statista. (2019). Retrieved February 21, 2020,
from [Link]

Kwon, W.-S., & Noh, M. (2010). The influence of prior experience and age on mature
consumers’ perceptions and intentions of internet apparel shopping. Journal of Fashion
Marketing and Management, 14(3), 335–349.

Laukkanen, P., Sinkkonen, S., & Laukkanen, T. (2008). Consumer resistance to internet
banking: Postponers, opponents and rejectors. International Journal of Bank Marketing,
26(6), 440–455. [Link]

Laukkanen, T., Marke, K., Laukkanen, P., & Sinkkonen, S. (2007). Innovation resistance
among mature consumers. Journal of Consumer Marketing, 24(7), 419–427.

Lian, J. W., & Yen, D. C. (2014). Online shopping drivers and barriers for older adults: Age
and gender differences. Computers in Human Behavior, 37, 133–143.
[Link]

527
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Markert, J. (2004). Demographics of Age: Generational and Cohort Confusion. Journal of

Current Issues & Research in Advertising, 26(2), 11–25.

McCloskey, D. (2006). The importance of ease of use, usefulness, and trust to online
consumers: An examination of the technology acceptance model with older consumers.
Journal of Organizational and End User Computing, 18(3), 47–65.
[Link]

Molesworth, M., & Suortti, J.-P. (2002). Buying cars online: The adoption of the web for high-
involvement, high-cost purchases. Journal of Consumer Behaviour, 2(2), 155–168.
Retrieved from
[Link]
n_of_the_web_for_high-involvement_high-cost_purchases

Molla, A., & Licker, P. (2001). E-commerce systems success: An attempt to partially extend
and respecify the Delone and Mclean model of IS success. Journal of Electronic
Commerce Research, 2(4), 131–141.

Norum, P. S. (2003). Examination of Generational Differences in Household Apparel

Expenditures. Family & Consumer Sciences, 32(1), 52–75.

Nunnally, J. C. (1978). Psychometric Theory. New York: McGraw-Hill.

Prensky, M. (2001). Digital natives, digital immigrants part 1. On the Horizon, 9(5), 1–6.

Ram, S. & Sheth, J. N. (1989). Consumer resistance to innovations: The marketing problems
and its solutions. The Journal of Consumer Marketing, 6(2), 5–14.

Ram, S. (1987). A Model of Innovation Resistance. NA - Advances in Consumer Research, Vol.

14, pp. 208–212. [Link]

Statista. (2020). eCommerce - Turkey | Statista Market Forecast. Retrieved May 3, 2020, from
[Link]

Steve Lawrence, C. L. G. (1998). Searching the World Wide Web. Computer Science,
280(5360), 98–100.

Tatnall, A., & Lepa, J. (2003). The Internet, e‐commerce and older people: an actor‐network
approach to researching reasons for adoption and use. Logistics Information Management,
16(1), 56–63. [Link]

528
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Taylan, E. (2015). BKM verilerine göre online alışveriş yapanların yüzde 70’i 34 yaşın altında
- Webrazzi. Retrieved March 1, 2020, from [Link]
verileri-online-alisveris-yas/

Turkey Population 2020 (Demographics, Maps, Graphs). (2020). Retrieved March 25, 2020,
from [Link]

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User Acceptance of
Information Technlogy: Toward a Unified View. MIS Quarterly, 27(3), 425–478.

WHO. (2020). Rolling updates on coronavirus disease (COVID-19). Retrieved May 2, 2020,
from [Link]
happen

Young, J. (n.d.). US ecommerce sales grow 14.9% in 2019 | Digital Commerce 360. Retrieved
February 29, 2020, from [Link]
sales/

529
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The influence of the European H2020 Programme on the development

of regions
The multiplier effect of EEUU funding in the regions under Horizon 2020 in NUT
III – Algarve – Portugal

N. Jesus-Silva¹; Ribeiro, Diamantino²; Ribeiro, João³

¹²³ Instituto Jurídico Portucalense, Universidade Portucalense Infante D. Henrique
¹natachajsilva@[Link]; ²diamantinojtribeiro@[Link]; ³joao217@[Link]

Abstract

The partnership agreement established between the European Union and the member
countries for the application of the European Structural and Investment Funds (ESIF), in
the period between 2014 and 2020, is in its final phase.
Other partnerships have also been established. These partnerships will be analysed during
the article. It is therefore important to understand the impact that the partnership
agreement has on public and private investment in the different regions and,
consequently, in the member countries.
This study analyses the multiplier effect on regional investment, of European funds made
available for the Algarve - Portugal - NUTS III region, under the H2020 programme, up
until March 2020 and aims to answer the following questions:
a) What is the amount invested in the regional economy for each euro of support
allocated by the EU through the H2020 programme.
b) What is the percentage distribution of community support versus investment per area
of intervention?
After collecting the information available on the H2020 managing entity's website for the
Algarve region (data until 31/3/2020), the analysis was structured regarding the areas of
intervention, thematic objectives and Investment Priority. After collecting, organising and
processing the data, we obtained the results necessary to answer the above questions.
956 projects were supported, which represented a total investment of 467,671,827.76
euros, having been supported with community funds in the amount of 249,966,698.38
euros, which represents a support of 53.4%. In other words, for each euro of support from
the European Union, 1.87 euros were invested in the Algarve region.
It can be concluded that, in the framework of Horizon 2020, and according to the data
available until 31 March 2020, 956 projects were supported.
These projects represent a total investment of 467,671,827.76€, having been supported
with community funds in the amount of 249,966,698.38€ (53.4% community support). In
other words, for each euro of support from the European Union, 1.87 euros were invested
in the Algarve region.
Considering that the initial objectives of the European Union foresaw support of
318,676,488.00 euros, providing a global investment in the region of 437,289,055.00, that
is to say, each euro of support should generate a regional investment of 1.37€ , we can
conclude that the Algarve Region has already exceeded the estimated investment value
for the region by 30 million euros. We still need to apply more than 68 million EU grants.

Keywords: Decentralisation; Investments; NUTS III; Public Value Capture; Property

Value; Europe.

530
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Introduction

On 12 June 1985, Portugal signed the Treaty of Accession to the European Economic Community
(EEC). Prime Minister Mário Soares led the party that formalised, at the Jerónimos Monastery,
the country's entry into the European project.

Portugal was going through a serious financial crisis, accentuated by the recession of the world
economy. After the revolution of 25 April 1974 and the loss of the colonial market, Portugal
maintained a great external dependency. It was in this context that the country approached the
European market, making the application for EEC membership in 1977. Only in the next decade
did this application came to fruition (1986), simultaneously with Spain, in what was the third
enlargement of the European Community.

The European Economic Community is part of the process of forming what today is the European
Union, which was originally intended to foster economic progress, freedom and a lasting peace
between Europe's neighbouring states. It started in 1950 with the European Coal and Steel
Community (ECSC) by six founding countries: Germany, Belgium, Italy, France, Luxembourg
and the Netherlands, which took the first steps towards the union of many of the European
countries.

Mateus (1992) states that the changes brought about by Portugal's full accession to the EEC in
1986, in a context of eliminating tariff, technical and fiscal barriers required by the completion of
the European internal market in the horizon of 1993, constituted a fundamental milestone. It
precisely situated the different experiences, policies and actions to promote development, whether
at the sectorial, regional or national level.

This evolution represented the culmination of a long process of European integration and was part
of a broader movement of economic, social and political restructuring.

In fact, Portugal's full accession to the European Communities triggered a significant process of
modernisation of structures and economic behaviour in parallel with the global liberalisation of
markets. Public intervention started to be clearly polarised by the promotion of important
infrastructure investments, namely, road, rail and port, telecommunications, energy, scientific and
technological, education and professional training, among others, which were boosted by access
to structural funds of community origin, aimed towards the regional development of the country.

The Portuguese Economy after joining the European Communities:

transformations and challenges.

The 2007-2013 Community financing framework

The National Strategic Reference Framework (QREN) was the local framework for the
implementation of the Community’s policy for economic and social cohesion in Portugal in the
period between 2007-2013.
The QREN took on the qualification of Portuguese men and women as a great strategic goal,
valuing knowledge, science, technology and innovation, as well as promoting high and sustained
levels of economic and socio-cultural development and territorial qualification, within a

531
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

framework of valuing equal opportunities and, also, increasing the efficiency and quality of
public institutions.
The pursuit of this great strategic plan was anchored on three major Thematic Operational
Agendas:
• Operational Agenda for Human Potential;
• Operational Agenda for Competitiveness Factors;
• Operational Agenda for Territorial Enhancement.
The execution of the NSRF (QREN) was made possible by the mobilisation of approximately
21.5 thousand M€, the use of which respected three main guidelines:
• Reinforcement of appropriations for the Qualification of Human Resources;
• Reinforcement of the financing directed to the Promotion of the Sustained Growth of
the Portuguese Economy;
• Reinforcement of the financial relevance of the Continent's Regional Operational
Programmes.
Depending on the Cohesion Policy Objectives:
• Convergence objective;
• Regional Competitiveness and Employment Objective;
• Objective of European Territorial Cooperation.
The operationalisation of these three Thematic Agendas was divided by the following QREN
Operational Programmes:
• Thematic Operational Programmes;
• Regional Operational Programmes for the Continent;
• Operational Programmes of the Autonomous Regions;
• Operational Territorial Cooperation Programmes;
• Operational Technical Assistance Programmes.
The governance of the NSRF (QREN) was based on the following organic structure:
• A political steering body - the QREN Ministerial Coordination Commission;
• A technical body responsible for the respective coordination and strategic monitoring -
the QREN Technical Coordination Commission;
• Two technical bodies for the coordination and financial monitoring of the Cohesion
Fund and the Structural Funds (ESF and ERDF) - IFDR and IGFSE, which, with the
General Inspection of Finance, also exercised control and audit responsibilities.

The 2014-2020 Community financing framework

The Partnership Agreement that Portugal proposed to the European Commission, called Portugal
2020, adopts the programming principles of the Europe 2020 Strategy (H2020) and enshrines
the economic, social, environmental and territorial development policy to stimulate growth and
job creation (in the coming years) in Portugal.
The correction of budgetary and external imbalances in the Portuguese economy (the Economic
and Financial Adjustment Plan adopted by Portugal in 2011) had negative social consequences
and asymmetric impacts on the development of the various regions. Thus, public policies,

532
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

namely those co-financed by community funds, should promote growth and employment, aiming
at reducing poverty and correcting the remaining external imbalance.
The programming and implementation of Portugal 2020 was organised into four thematic
domains - competitiveness and internationalisation, social inclusion and employment, human
capital, sustainability and efficiency in the use of resources - also considering the transversal
domains related to the reform of Public Administration and the territorialisation of interventions.
The identification of the main constraints and potentialities of these areas allowed for the
intervention priorities of the Union funds for the period 2014-2020 to be defined.
The existing constraints in the field of competitiveness and internationalisation are due to the
fact that productive specialisation is based on activities with low added value, low technological
intensity and knowledge, the weak skills and strategies of companies inherent to the weaknesses
in the qualification of employers and employees and their reduced propensity for more
sophisticated business strategies; the difficult contextual conditions of the business activity,
namely the financing conditions of the companies and the increased costs and transport times,
taking into account the geographical position of Portugal in Europe, and the distance from the
main export destinations, are other constraints felt.
Thus, the following public policy instruments are foreseen: direct incentives for business
investment, especially in R&D, qualification of SME’s focused on internationalisation
strategies; indirect support for business investment to train companies to pursue more advanced
business strategies; support for qualified and creative entrepreneurship and enhancement of more
dynamic business opportunities and in innovation domains; support for the production and
dissemination of scientific and technological knowledge, promoting international connections
of the national and regional R&D systems, as well as the transfer of knowledge and technology
between companies, R&D centres and higher education institutions; support for business
training to train companies' human resources for innovation and internationalisation processes;
investments in transport infrastructures, focused on reducing transport time and costs for
companies, especially in the context of international connectivity; and support for administrative
modernisation and training for Public Administration, aiming at reducing public costs.
As for the social inclusion and employment domain, there is a high level of unemployment, due
to the recent economic crisis. This is aggravated by the fact that long-term unemployment affects
more than half of the unemployed, and the number of young people who are unemployed or
studying or in training has increased significantly.
Another constraint is the strong segmentation of the labour market between a more qualified
segment with a tendency for greater conditions of employability and quality of employment, and
a less qualified segment or with qualifications that are not suited to the needs of the productive
fabric, with the prospect of unemployment risks, structural change or access to precarious jobs.
In this domain, and with these associated factors, a high level of poverty and social exclusion
persists in Portugal. Thus, with a view to promoting employment and social inclusion,
community support for the following policy instruments is undertaken: qualification of assets,
for the development of certified skills for the labour market; transition between situations of
inactivity or unemployment and employment, as well as net job creation and maintenance in the
labour market; consolidation and requalification of the network of collective equipment and
services; specific interventions in favour of target territories or groups where situations or risks
of poverty are cumulative with those of social exclusion; promotion of gender equality, non-
discrimination and accessibility, interventions that complement those mentioned above;
combating failure and early school dropouts. These public policy instruments will be properly

533
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

coordinated with the numerous instruments that, by their nature, will not benefit from the
Union’s funding.
The so-called Human Capital domain shows: a lag compared to the more developed countries in
the average level of qualifications of the adult and young population (associated with the late
schooling of the Portuguese population, the reduced participation of the adult population in
certified education and training activities and the early school dropout of young people); the
lack, despite the improvement, of a quality and efficient education and training system; and the
mismatch between the qualifications produced and those sought by the labour market. It is
therefore important to provide for direct interventions to reduce school dropout and promote
educational success; promote professional training offers for young people; guarantee school
social action (in basic, secondary and higher education); and provide higher level training.
The main constraints that Portugal still faces in the field of sustainability and efficiency in the
use of natural resources are summarised in the high energy intensity of the Portuguese economy,
in the inefficient use and management of resources, in the vulnerabilities in the face of various
natural and technological risks and in the weaknesses in the protection of environmental values.
The approach to respond to these constraints is structured in three vectors that will be the basis
for the mobilisation of community funds for the next cycle: the transition to a low carbon
economy, mainly associated with the promotion of energy efficiency and the production and
distribution of renewable energy; risk prevention and adaptation to climate change; protection
of the environment and promotion of resource efficiency, structured around the intervention
areas: waste management; water management (urban water cycle and water resources
management); management, conservation and enhancement of biodiversity; recovery of
environmental liabilities; qualification of the urban environment, mainly as a result of the urban
regeneration and revitalisation process.
The Partnership Agreement was subject to an ex-ante evaluation that consisted of an interactive
process of reflection that led to the consideration of the incorporation of the recommendations
that occurred in the successive versions. One of the main conclusions was that the diagnosis that
supports the Partnership Agreement and its recommendations and the identification of the
thematic areas to be considered, are convergent and consistent with the objectives and goals of
the Europe 2020 Strategy and with the National Reform Programme. As the intervention logic
of the funds is organised around the above-mentioned thematic areas and with two dimensions
of a transversal nature, thematic objectives selected from those provided for in the EEUU
regulations, the investment priorities mobilised for the intended purposes, as well as the specific
objectives and main expected results for each of them.
To complete the description of the programming logic, the operational delimitation (intervention
of each Fund and Financing Operational Programme) and the territorial scale of intervention
(articulation of national and regional scope) are presented.
The Partnership Agreement applied the principle of partnership, embodied in the work of the
Government, Public Administration, various entities of civil society and the public in general. It
is important to highlight the involvement of institutional, economic and social partners in the
preparatory work for the 2014-2020 cycle, in particular the Assembly of the Republic, the
Economic and Social Council, the Permanent Commission for Social Coordination (which
involves unions and business organisations) and the National Association of Portuguese
Municipalities. Initiatives for the dissemination and consultation of civil society were also
promoted. The principle of equality between men and women, non-discrimination and
accessibility, and the principle of sustainable development were also guaranteed. The

534
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

governance model of the Partnership Agreement and Operational Programmes 2014-2020 and
its institutional architecture, aimed at four objectives: the simplification of the governance
model, favouring, on the one hand, the segregation of responsibilities and institutional support
for the exercise of political and technical orientation functions, and, on the other hand, valuing
the involvement of partners; the orientation towards results, implemented through the
valorisation of the results in the financing decisions and its evaluation and consequences
resulting in the payments of the final balance of the projects; the establishment of common rules
for financing, which not only ensure conditions of equity and transparency, but also competition
between beneficiaries; and simplifying beneficiaries' access to financing and reducing their
administrative costs. This programming exercise presents the assessment of compliance with the
ex-ante conditionalities applicable at a national level. For cases in which, according to the
aforementioned assessment, the conditionalities are not met, action plans have been established
in accordance with Article 19 of Regulation (EU) 1303/2013. The principle of additionality was
also verified.
In order to ensure the performance analysis of all instruments, applying the mechanism provided
for in the regulatory funds framework, called the Performance Framework, initiatives were
developed to ensure consistency in the selection of indicators for programming.
Finally, the territorial development strategies adopted that will contribute to the territorial
strengthening of the Europe 2020 Strategy are clear, ensuring that the specificities and the
different degrees of development of the sub-regions are taken into account, guaranteeing the
involvement of sub-regional entities and regional and local authorities in planning and
implementing their programmes and projects.
Portugal 2020 is the Partnership Agreement adopted between Portugal and the European
Commission, which brings together the activities of the 5 European Structural and Investment
Funds - ERDF, Cohesion Fund, ESF, FEADER and FEAMP. It incorporates the programming
principles that conserve the economic, social and territorial development policies in Portugal
between 2014 and 2020.
These programming principles were aligned with Smart, Sustainable and Inclusive Growth,
continuing Europe’s 2020 Strategy.
It was forecasted that Portugal should receive around 25 billion euros by 2020; to this end, it
defined the Thematic Objectives to stimulate growth and job creation, the interventions
necessary to achieve them and the achievements and results expected from this financing:
stimulating the production of tradable goods and services; increase in exports; transfer of results
from the scientific system to the productive fabric; compliance with compulsory education up to
18 years; reducing the levels of early school dropouts; integration of people at risk of poverty
and combating social exclusion; promoting sustainable development from the perspective of
resource efficiency; strengthening territorial cohesion, particularly in cities and in low-density
areas; rationalisation, modernisation and training of Public Administration were the main
objectives of the policies to be pursued in Portugal up until 2020.
The programming and implementation of Portugal 2020 was organised into four thematic
domains: - Competitiveness and Internationalisation; - Social Inclusion and Employment; -
Human capital; - Sustainability and Efficiency in the Use of Resources. It also considered the
transversal domains related to the reform of Public Administration and the territorialisation of
interventions.

535
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In terms of eligibility for European Investment Funds (ERDF, FC, ESF, FEADER and FEAMP),
the 7 regions of Portugal were divided into:
• Less developed regions (GDP per capita <75% EU average): North, Centre, Alentejo and
Autonomous Region of the Azores (RAA)
Funds co-financing rate: 85%
• Regions in transition (GDP per capita between 75% and 90%): Algarve
Funds co-financing rate: 80%
• Most developed regions (GDP per capita> 90%): Lisbon and Madeira Autonomous
Region (RAM)
Funds co-financing rate: 50% (Lisbon) and 85% (RAM)
Portugal should receive 25 billion euros by 2020, for all the funds, which will be allocated under
each of the 16 Operational Programmes, thematic and regional, as we can see in the following
figure:

Graphic 1 - Operational, thematic and regional programmes

Source: European Union

The Algarve Region

The Algarve is a sub-region and historical province (or natural region) of Portugal, and is the most
southern of all regions in the country. Its capital is the city of Faro.

It coincides perfectly with the District of Faro, having an area of 5,412 km² and a permanent
population of 451,005 inhabitants (Census 2011), (0.06% of the population of Europe and 6.27%
of the population of Portugal).

It is the most important tourist region in Portugal and one of the most important in Europe. Its
Mediterranean climate, characterised by mild and short winters and long, hot and dry summers,

536
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the warm and calm waters that bathe its south coast, its natural landscapes, the historical and
ethnographic heritage and the renowned and healthy gastronomy, are attributes that attract
millions of national and international tourists every year and make the Algarve the most visited
region and one of the most developed in the country.

In 2020, the Algarve is the second region of Portugal with the highest purchasing power, only
behind the Metropolitan Area of Lisbon, with a GDP per capita of 83% of the European Union
average.

Due to the above-mentioned factors, the Algarve has become one of the Portuguese regions with
the highest number of foreign residents, mainly from other European countries. In 2018, 69,000
of the inhabitants were not Portuguese.

Algarve Regional Operational Programme 2014-2020

With an allocation of funds exceeding 318 million euros from the European Regional
Development Fund (ERDF - 224,321,248.00 €) and the European Social Fund (ESF -
94,355,240.00 €) for the period 2014-2020, the Algarve programme aims to promote the
competitiveness of the regional economic development, sustainable development and internal
cohesion in the region, as well as the region's capacity to contribute to the achievement of the
main national and European development objectives. The European Union's objectives would be
to provide a global investment in the region of €437,289,055.00 over the years, as shown in Figure
2 below. In other words, each euro of support should generate a regional investment of € 1.37:

Graphic 2 - Estimated investment over the period

537
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Source: European Union

The Programme focuses on eight main priorities and one technical assistance:

1- Promote regional research and innovation - ERDF - 12.19% of total EU resources to

improve research and development (R&D) infrastructures, promote centres of competence
and develop synergies between companies and research centres.

2- Support internationalisation, competitiveness of companies and qualified entrepreneurship

- ERDF - 26.89% of the total EU resources to improve the competitiveness of SMEs through
the promotion of entrepreneurship, business incubators and internationalisation.

3- Promote sustainability and resource efficiency - ERDF - 6.3% of EU resources to develop

sustainable resources and energy efficiency (companies, public infrastructure, social housing)
and promote strategies based on a low level of carbon emissions, namely in urban areas,
including multimodal mobility.

4- Strengthening territorial competitiveness - ERDF - 7.8% of EU resources, to improve the

actions concerning the conservation, protection and promotion of natural and cultural
heritage, to improve in a sustainable and integrated way the urban environment through the
revitalisation of cities, to requalify former industrial facilities, reduce air pollution and
promote noise reduction measures;

5- Investing in employment - ESF and ERDF, 10.78% and 3.8%, respectively, of EU

resources to create employment, improve access to employment for job seekers and inactive
people, develop local initiatives and support labour mobility;

6- Ensuring social and territorial cohesion - ESF and ERDF, 9.9% and 3.9%, respectively, of
EU resources to promote social inclusion, invest in health infrastructure, fight poverty and
discrimination, support disadvantaged communities in rural and urban areas, improve the
integrated approach to social intervention with local community-oriented strategies;

7- Improve skills - ESF and ERDF, 6.6% and 3.3%, respectively, of EU resources to invest
in education, training, vocational training and lifelong learning, develop and modernise
school infrastructure, reduce and prevent early school dropouts, increase the quality of
education and improve the adaptation of the provision of educational and training to the needs
of the labour market;

8- Modernise and empower the Administration - ESF and ERDF, 3.1% and 2.5%,
respectively, of EU resources for the OP to improve access to information and communication
technologies, invest in institutional capacity and efficiency administrations and public
services;

9- Technical Assistance - ERDF, 3.1% of EU funds.

538
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Graphic 3 - Thematic priorities

Source: European Union

Results on 31 March 2020

At the date of the last update, 31/03/2020, a total of 956 projects were approved, generating a
total investment of 467 million euros, which had a community support of 250 million euros
(53%), as can be seen in table 1:
Table 1- Approved projects

The investment in R&D and innovation supported 115 projects with a value of 25,978,407 euros,
which generated an investment of 45,104,829 euros.
In the area of business competitiveness, the 377 business projects (47.1% of the total)
represented an investment of approximately 220 million euros with a European contribution of
86 million euros.
91 projects in the area of the environment were supported for a total value of 31.6 million euros,
which generated a regional investment of 57.9 million euros.
In the area of employment and mobility, 119 projects were supported in the amount of 35.3
million euros, generating a regional investment of 46 million euros.
For social inclusion and the fight against poverty, 148 projects were supported for a total of 28
million euros, generating a regional investment of 38.7 million euros.

539
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In the education and enhancement of skills and lifelong learning, 63 projects were supported for
a total of 22.4 million euros, generating a regional investment of 34.3 million euros.
To promote information technologies in public services, 31 projects were supported, coming to
a total of 11 million euros, which generated a regional investment of 13.7 million euros.
Finally, 12 projects in the area of technical assistance were supported for a total of 9.6 million
euros, which generated a regional investment of 11.7 million euros.

Results
The results indicate that each euro of support granted by the European Union through the H2020
Programme has the following impacts on investment in the Algarve Region: reinforcement of
R&D and innovation OT1 = 1.74 euros; improvement of the competitiveness of SMEs OT3 =
2.55; support for the transition to a low-carbon economy OT4 = 1.54 euros; protection of the
environment and promotion of resource efficiency OT6 = 1.86 euros; promotion of employment
and support for labour mobility OT8 = 1.30 euros; promotion of social inclusion and the fight
against poverty OT9 = 1.38 euros; education, skills and lifelong learning OT10 = 1.54 euros;
promotion of health technologies in administration and public services axis 8 = 1.25; technical
assistance, axis 9 = 1.22.
956 projects were supported, representing a total investment of 467,671,827.76 euros, having
been supported with community funds in the amount of 249,966,698.38 euros, which represents
a support of 53.4%. That is to say, for each euro of support from the European Union, 1.87 euros
were invested in the Algarve region.

Regarding the percentage distribution (base 100) of support granted versus investment made by
each of the areas, we arrive at the following indicators: reinforcement of R&D and innovation,
10.4% support / 9.6% investment; improvement of the competitiveness of SMEs, 34.5% support
/ 47.1% investment; transition to a low-carbon economy, 1.1% support / 0.9% investment;
environment and promotion of resource efficiency, 11.5% support / 11.5% investment;
promotion of employment and labour mobility, 14.1% support / 9.9% investment; promotion of
social inclusion and the fight against poverty, 11.2% support / 8.4% investment; education, skills
and lifelong learning, 8.9% support / 7.3% investment; promotion of information technologies
in administration and public services, 4.4% support / 2.9% investment; technical assistance,
3.8% support / 2.5% investment.

Conclusion
We can conclude that in the framework of Horizon 2020, and according to the data available up
until 31 March 2020, 956 projects were approved, which represented a total investment of
467,671,827.76 euros, having been supported with EEUU funds in the amount of €
249,966,698.38, which represents a support of 53.4%, that is, for each euro of support from the
European Union, € 1.87 was invested in the Algarve Region.
If we take into account that the initial objectives of the European Union estimated
318,676,488.00 euros, providing a global investment in the region of 437,289,055.00, that is to
say, each euro of support should generate a regional investment of 1.37 euros, we can conclude
that the Algarve Region, (lacking more than 68 million EU grants) has already exceeded the
estimated investment value for the region by 30 million euros.

540
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Therefore, for each euro of EU support whose initial multiplier expectation was 1.37 euros, the
region has so far achieved a higher impact: 1.87 euros for each euro of EU support.

Future studies
In future studies we will carry out tests on the impact of EEUU funds in the remaining 6
Portuguese regions (North, Centre, Lisbon, Alentejo, Azores and Madeira), making comparisons
between them. Simultaneously, tests will be carried out on the impact of these investments on
regional and national GDP.

541
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References
ABREU, Cancela de; CORREIA, Teresa Pinto; OLIVEIRA, Rosário (2004) - Contributos para a
identificação e caracterização da paisagem em Portugal Continental, 5 vol., Ed. DGOT-DU,
Lisbon.

Agência para o Desenvolvimento e Coesão, IP (September 2016) “Boletim Informativo dos

Fundos da União Europeia” [Link]

AGGERI, Gaell (2004) La nature sauvage et champêtre dans les villes : Origine et construction
de la gestion différenciée des espaces verts publics et urbains. Le cas de la ville de Montpellier.
Sciences of the Universe. ENGREF (AgroParisTech).

BALEIRAS, Rui N. (2009a), Intervenção do Secretário de Estado do Desenvolvimento Regional

no painel O futuro da política de coesão, conferência ―Cooperação Transfronteiriça de Segunda
Geração‖, 4 and 5 February, Guimarães: Eixo Atlântico do Noroeste Peninsular e Instituto
Financeiro para o Desenvolvimento Regional.

BALEIRAS, Rui Nuno (2009b), Cohesion Policy in Low-Density Cross-Border Territories,

apresentação realizada na conferência ―Territorial Co-operation: Working Together Across
Borders‖ integrada nos Open Days do Desenvolvimento Regional — European Week of Regions
and Cities, 6 de outubro, Bruxelas: Comissão Europeia e Comité das Regiões; access at:
[Link]
Ponencia_Rui_Baleiras_open_days_09.pdf.

BARCA, Fabrizio (2009), An Agenda for a Reformed Cohesion Policy: a Place-based Approach
to Meeting European Union Challenges and Expectations, relatório independente preparado para
a Comissária da Política Regional, Abril, s/ local; also accessible at:
[Link]
Boletim Informativo QREN no. 23 (2014). Indicadores Conjunturais de Monitorização, 31
March 2014. Access at:
[Link]
BRONZINI Fabio; BEDINI M. Angela; MARINELLI Giovanni (2014) Nuovi paradigmi per
uno sviluppo alternativo di contrasto alla crisi: politiche e strategie per riscoprire le potenzialità
economiche, sociali e ambientali di connettività città-campagna, in Urbanistica Informazione, nº
257, INU Ed, pp. 23-25.
CAVACO, Carminda (2005) As paisagens rurais: do “determinismo natural “ ao “determinismo
político? in Finisterra, no. 79, CEG, Lisbon.

CCDR (2009) Plano Regional de Ordenamento do Território da Região Norte (proposta), pol.
Porto.

Charbit, C. (2011). Governance of Public Policies in Decentralised Contexts: The Multilevel

Approach. OCDE Regional Development Working Papers, 2011/04. OCDE Publishing.

Charbit, C. e Michalun, M. (2009). Mind the Gaps: Managing Mutual Dependence in Relations
amog Levels of Government. OCDE Working Papers, Número 14. Access at:
[Link]
mutual-dependence-in-relations-among-levels-ofgovernment_221253707200#page1.

Comissão Europeia “O que é o FSE?” [Link]

542
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Comissão Europeia “Fundo Europeu de Desenvolvimento Regional”

[Link]

Comissão Europeia “Fundo de Coesão” [Link]

fund/ .

Comissão Europeia (2010). Investing in Europe’s Future [Relatório], November 2010. 5º

Relatório sobre a coesão económica, social e territorial. Bruxelas. Access at:
[Link]
5cr_part1_en.pdf.

COMPETE: Programa Operacional Fatores de Competitividade “Estrutura”

[Link] [15-10-2016].

COMPETE: Programa Operacional Fatores de Competitividade “Incentivos às Empresas”

[Link]

COOPER, T.; HART, K.; BALDOCK, D. (2009) The Provision of Public Goods Through
Agriculture in the European Union, Report Prepared for DG Agriculture and Rural Development,
Contract No 30-CE-0233091/00-28, Institute for European Environmental Policy, London.

COVAS, António; COVAS, Maria das Mercês (2012) A caminho da 2ª ruralidade. Uma
introdução à temática dos sistemas territoriais, Ed. Colibri, Lisbon.

DELGADO, Carlos (2010) Expansão urbana e fragmentação de áreas com forte aptidão agrícola:
o caso de estudo da bacia leiteira primária de Entre- Douro e Minho, Dissertação de Mestrado,
FLUP, Porto.

DONADIEU, Pierre (2012) Construction et déconstruction des identités paysagères dans les
régions urbaines, in Paysages européens et mondialisations, dir A. Bergé, M. Collot et J.
Mottet, Seyssel, Champ Vallon, "Pays/Paysages" pp. 179-198.

DONADIEU, Pierre (2012) As paisagens agro-urbanas: uma utopia realista? in Filosofia e

arquitectura da paisagem , A. Veríssimo Serrão (coord); Ed Centro de filosofia da Universidade
de Lisboa, pp. 281-290.

DONADIEU, Pierre (2007) Le paysage. Un paradigm de mediation entre l`espace et la société,

Èconomie rural, no. 297-298, pp 5-9.

DONADIEU, Pierre (2005) Campagne urbane, una nuova proposta di paesaggio della città, cura
di M. V. Minini, Donzelli ed, Rome.

DONADIEU, Pierre (2003) La construction actuelle des villes-campagnes. De l`utopie aux

réalités, Rev. Histoire urbaine, no. 8, pp. 157-170.

FLEURY, André ; DONADIEU, Pierre (1997) De l`agriculture péri-urbaine á l` agriculture

urbaine, in Le Courrier de L` environnement, no. 31.

GRECO, Francesca ; TRIBUZIO, Francesca (2014) Dalla riqualificazione delle matrici agricole
tradizionali alla valorizzazione del paesaggio. Un caso di studio nell’area metropolitana di Bari,
in Urbanistica Informazione, no. 257, INU Ed, pp. 89-93.

543
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

GONZÁLEZ; Roman; MARQUES, Helder (2009) Policentrismo e complementariedade do

sistema urbano, in Complementaridade para competir. Eixo Atlântico: uma estratégia pelo
território; Ed. Eixo Atlântico do Noroeste Peninsular, pp. 9 – 33.

INE: Instituto Nacional de Estatística (2016) Produto interno bruto (B.1*g) a preços correntes
(Base 2011 - €); Trimestral.
[Link]
823&contexto=bd&selTab=tab2.

KAYSER, Bernard (1990) La renaissance rurale. Sociologie des campagnes du monde

occidental, ED. Armand Colin, Paris.

KAYSER, Bernard et alii (1994) Pour une ruralité choisie, DATAR / Editions de L’Aube.
Paris

MARQUES, Helder (2008) - Porquê (e razões para) a mitificação do campo, VII Colóquio
Ibérico de Estudos Rurais - Cultura, Inovação e Território, Coimbra.

MARQUES, Helder; MENDONÇA, João; SANTOS SOLA, Xosé (2004) Permanências e

mutações nos territórios vitícolas da Galiza e do Noroeste de Portugal. Uma abordagem
geográfica in Actas do III Simpósio da Associação Internacional de história da Vinha e do
Vinho, Ed. Centro de Estudos de História do Atlântico, Funchal, pp. 229 – 247.

MARQUES, Helder; MENDES, Américo (2002) – Sector agro-florestal: um desafio para o

desenvolvimento rural, in Sistema urbano Nacional -Rede Complementar, Coord. Teresa Sá
Marques, Ed. Direcção Geral do Ordenamento do Território e Desenvolvimento Urbano, Lisbon,
pp. 150-167.

MARQUES, Helder (2000) Modernidade e inovação na ruralidade do Noroeste de Portugal,

FLUP, pol., Porto.

MARQUES, Teresa; QUEIROZ J. P. ; ALVES, Paulo (coord) (2014) AMP 2020 Estratégia de
Base territorial, Ed. AMP, 44 pp.

MARTINS, Natalino, FIGUEIREDO, Carlos, PROENÇA, Manuela e RIBEIRO, José F. (2008),

PROVERE, das Ideias à Acção: Visão e Parcerias, Ministério do Ambiente, do Ordenamento do
Território e do Desenvolvimento Regional, abril, Lisboa: Departamento de Prospectiva e
Planeamento e Relações Internacionais; also accessible at:
[Link] (Portuguese version) e [Link]
[Link]/pages/files/PROVERE_EN.pdf (English version).

Mcvittie, E. & Swales, JK. (2007). Constrained Discretion in UK Monetary and Regional Policy.
University of Plymouth Business School, Drake Circus, Plymouth. Department of Economics
Fraser of Allander Institute, University of Strathclyde, Glasgow, and Centre for Public Policy
for Regions.

Observatório do QREN (2010). Abordagens Integradas de Base Territorial - Relatório Final.

Observatório do QREN (2013). A contratualização como expressão da governação multinível:

as lições do QREN e os desafios 2014-2020. Braga: 19º Congresso da APDR, 20 June 2013.

OCDE (2007). Linking Regions and Central Governments. Contracts for Regional
Development.

OCDE (2009). Governing Regional Development Policy. The Use of Performance Indicators.

544
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

OECD (2011). Making the Most of Public Investment in a Tight Fiscal Environment: Multi-
level Governance. Lessons from the Crisis. OECD Publishing.

OCDE (2013). Governança multinível para políticas de desenvolvimento regional mais eficazes,
2013. Relatório Territorial da OCDE, Brasil 2013, Capítulo 2. OCDE Publishing. Access at:
[Link]
development/relatorio-territorial-da-ocde-brasil_9789264189058- pt#page1.

Orientações para a contratualização com subvenção global entre as autoridades de gestão dos
POR e as associações de municípios baseadas em NUTS III (2008). Deliberação aprovada em
19 de março de 2008. Access at:
[Link]

Portugal 2020 - Acordo de Parceria 2014-2020. July 2014. Access at:

[Link] .pdf.

Quadro de Referência Estratégico Nacional – Portugal 2007-2013. Observatório do QCA III,

September 2007. Access at:
[Link]

REGULAMENTO (CE) n.o 1083/2006, do Conselho, de 11 de Julho de 2006, que estabelece

disposições gerais sobre o Fundo Europeu de Desenvolvimento Regional, o Fundo Social
Europeu e o Fundo de Coesão, e que revoga o Regulamento (CE) n.o 1260/1999, Jornal Oficial
da União Europeia, série L, n.o 210, de 31 de Julho de 2006, pp. 25–78, com as alterações
introduzidas pelo Regulamento (CE) n.o 284/2009 do Conselho, de 7 de Abril, Jornal Oficial da
União Europeia, série L, n.o 94, de 8 de Abril de 2009, p. 10–12; also accessible at:
[Link]
regulation/pdf/2007/general/ce_1083(2006)_pt.pdf.

Relatório Anual do QREN V 2012. Comissão Técnica de Coordenação do QREN (2013). Access
at: [Link]

Relatório de Execução do Norte 2012. ON2: O Novo Norte. Programa Operacional Regional do
Norte. Access at: [Link]

RIBEIRO, José; FERRÃO João (coord) (2014) Noroeste Global, Ed. Fundação Calouste
Gulbenkian, Lisbon.

SAMECKI, Pawel (2009), Orientation Paper on Future Cohesion Policy, December, mimeo,
access at: [Link]
orientation_paper.pdf.

TRATADOS (2008), Versões Consolidadas do Tratado da União Europeia e do Tratado sobre

o Funcionamento da União Europeia, Jornal Oficial da União Europeia, series C, no. 26 115, 9
May; also accessible at: [Link]
uri=OJ:C:2008:115:SOM:PT:HTML.

545
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

REPORTING EXPENSES FROM THE STANDPOINT OF ACCOUNTING AND

TAXATION REQUIREMENTS

NEXHMIE BERISHA VOKSHI1

FLORENTINA XHELILI KRASNIQI2

ABSTRACT

Financial reporting provides very important information to businesses, ensuring a clear overview of how the business
is conducted, and it is also fundamental for decision-making by the unit's managers and not only. However, businesses
should also report fiscal data, which include all relevant information about the businesses’ financial standing during a
fiscal year, operations and net business value, which are needed for use by tax authorities. Consequently, accounting
rules and fiscal rules are drafted by different authorities and have different purposes, but both provide important
information to third parties.

Based on this fact, this study aims to present the current regulatory and practical situation of accounting and taxation
in Kosovo, in terms of acknowledging, recording and reporting of expenses. The study was conducted through data
from the questionnaire made in the 264 businesses with turnover of over one million€ (addressed to accountants, one
employee in each business), in order to obtain information on the orientation of these businesses to the use of
accounting rules, fiscal rules or both regulations together with recognition, registration and reporting of expenses. The
research result currently reflects a correlated link between financial reporting on expenditures and their fiscal
reporting, which implies that the businesses are initially oriented towards the use of accounting regulations when
recognizing, recording and reporting the expenses and then making adjustments for fiscal needs.

Keywords: accounting, taxation, financial statements, financial reporting, fiscal reporting, expenses.

JEL classification: M 40; M 41

INTRODUCTION
In order for financial reporting to be useful, it should provide qualitative, transparent and comparable financial
information for the reporting periods of the businesses. The truthfulness and accuracy of reporting will be a key
element for proper and adequate decision making by the users of that information and by the businesses. Qualitative
features of financial information apply to the elements presented in the financial statements, including those on
expenses (2018). The information should be prepared under the regulatory basis, and should faithfully represent the
elements presented in the financial statements.

1
Associate Professor, PhD, Faculty of Economic, University of Prishtina, Kosovo
[Link]@[Link]

2
Full Professor, PhD, Faculty of Economic, University of Prishtina, Kosovo
[Link]@[Link]

546
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Since expenses are the focus of this study, as one of the elements of financial statements, we will present the current
regulation regarding their financial and fiscal reporting, as well as empirical research on the use of these regulations by
the drafters of financial statements, specifically accountants.

The contribution of the study will be in identifying the relationship between accounting and taxation, with a focus on
expenses, where research results are considered to be relevant and may serve in the practical field of accounting. This
research paper highlights the most important components that affect the accounting-taxation relationship and will
contribute to enhancing the effectiveness of business reporting through the experiences and concerns that arise during
business management practices, during the preparation and use of financial reports, during decision-making process and
in achieving the performance and objectives of businesses in Kosovo. It is also considered that the paper will be a useful
reference for future researchers of this issue in various countries, thus enabling us to draw comparative analyzes
regarding the relationship between financial and fiscal reporting.

LITERATURE REVIEW
Financial reporting by accounting standards makes accounting information accessible to a wide range of users. The
unification of the accounting language affects the development of the capital market, attracting investors, overcoming
group interests and eliminating business barriers. Consequently, different countries apply IFRS to make financial
reporting the basis of business decision making, and for fiscal purposes the necessary adjustments are made under the
applicable legal framework.

The financial statements prepared under accounting system are used for decision making, while for tax purposes
specific modifications should be made as required by applicable law. According to Gavana, Guggiola and Marenzi,
codified in Italian Tax Code 12, the derivation rule states that taxable income is measured on the basis of accounting
results, with specific and limited modifications required by tax law when accounting requirements are not appropriate
for tax purposes (Gavana, Guggiola, & Marenzi, 2015, p.82). This shows that there is a correlation between accounting
and taxation, where accounting system dominates.

According to Susana Aldeia, the accounting principles are the overriding principles which are integrated in both
business and tax law. In the Portugese and Spanish case studies, she identified the corporation income tax law as
recognized in both laws by inherent acceptance of general accounting principles (Aldeila, 2019). In addition, the
taxable income could be further adjusted, all while keeping the accounting principles as the overriding principles for
both accounting and taxation purposes. According to Daniel N. Shaviro’s proposal, reaching the optimal relationship
between taxable income and financial accounting income would require a 50% taxable income adjustment for large,
publicly traded companies which would substantially improve the current law if adopted (Shaviro, 2008, p. 65).

Fiscal reporting based entirely on financial reporting rules would enable the use of financial statements for accounting
and fiscal purposes, which would consequently eliminate the need for double reporting. Article 1 2019 According to
Plaksiienko, Melikhova, Yermolaieva, Chernenko and Lipskyi, there is a close relationship between accounting policy
and tax system, where the organization of accounting is used by businesses for tax purposes. Therefore, the efficiency
of the tax system is influenced by the development of accounting policies. Furtherrmore, the accounting policy is a
tool used for increasing efficiency on implementation of company’s tax liabilities. This taxation adjustment under the
framework of accounting principles would result in an increase of efficiency in the economic activity of the business
without violating the statutory provisions (Plaksiienko, Melikhova, Yermolaieva, Chernenko, & Lipskyi, 2019).

However, the development of IFRS is a process that leads to the change or updating of existing standards, this affects
the emergence of the need for changes and updates of fiscal laws. Since fiscal adjustments depend heavily on different
business environments in different countries, then this can be a problem in bringing fiscal adjustments closer to those
of accounting that are constantly being updated and adopted. In the relationship between financial and fiscal reporting,

547
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

different authors have been identified, identifying the factors that determine this relationship and analyzing the degree
of impact. According to Cuzdriorean and Matiş, the implementation of IAS/IFRS in the European Union has had a
major impact on the accounting system in countries where there is a decreasing binding link between accounting and
taxation (Cuzdriorean & Matiş, 2012). According to Fernandes, Cerqueira, Bandao, there is evidence that there is a
link between tax and financial aggressiveness in firms based in Europe. This link between aggressive financial and
tax reports is proved to get weaker after the IFRS implementation in Europe (Fernandes, Cerqueira, & Brandao, 2017,
p.39).

Therefore, the implementation of IAS/IFRS is important in terms of unifying the reporting language into an accounting
one and increasing effectiveness of decision making by users of accounting information. Furthermore, we should be
aware of the fact that accounting purposes are different from those of taxation. While the purpose of accounting is to
provide information relevant to internal and external users of information, the purpose of taxation is to provide revenue
for budgetary purposes. This contributes to different requests for financial and taxation reporting. According to
Adeniran Samuel, Olusola Samuel and Obiamaka, the connection between tax and accounting is a complex topic. One
reason for this is changes in global accounting. This presents a tax law dilemma and a need to re-examine the
theoretical and practical basis for using accounting as a starting point for corporate taxation (Adeniran Samuel, Olusola
Samuel, & Obiamaka, 2013, p.169).

ACCOUNTING AND FISCAL REGULATORY FRAMEWORK FOR ACKNOWLEDGING,

RECORDING AND REPORTING EXPENSES
Businesses in Kosovo are obliged to prepare financial statements for general purposes in accordance with international
accounting standards (IAS) and financial reporting standards (IFRS) [1]. This indicates the fact that expenses should
also be acknowledged, recorded and reported according to the standards set out in the regulatory framework. In their
income statements, businesses must present expenses classified by their nature or function, always providing accurate
and useful information to the information users.

Acknowledging expenses results directly from acknowledging and measuring assets and liabilities. A business should
acknowledge expenses in the income statement when there is a reduction in future economic benefits associated with
a decrease in the asset or an increase in liabilities that can be reliably measured [2].

Fiscal regulations in Kosovo classify expenses as allowed expenses and disallowed expenses (2019). The Personal
Income Tax (PIT) law allows for deductions from total income, expenses paid or incurred during the tax period that
are wholly, directly or indirectly related to such revenue generating activities, including premiums for health
insurance, payable in the interests of employees and dependents who should be included in the policy of the employee.
Allowed expenses can be deducted by businesses with annual gross income of € 50,000 or more. These expenses
include: depreciation, representation, bad debt, travel, repair and improvement, amortization, research and
development expenses and educational expenses. Disallowed expenses include: fines and penalties, amusement and
recreation expenses, expenses for gifts other than those that represent the business, tax losses, manufacturing and
storage damages, grants, subsidies and donations, undocumented expenses etc.

According to the Corporate Income Tax Law (CIT), in determining taxable income, taxpayers are allowed deductions
from the taxable income, paid or incurred only during the tax period as a whole and related to its economic activity,
including premiums for health insurance, paid in the interest of employees and dependents who should be included in
the employee's policy. These expenses include: educational expenses, training, rent, representation, advertising and
promotion, bad debt, upgrade and repair expenses, research and development expenses. Disallowed expenses include:
expenses for gifts other than those representing the business, land acquisition and improvement expenses; upgrading,
refurbishing and reconstructing expenses of assets that have been capitalized and depreciated, pension contributions

548
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

over the maximum amount allowed by the relevant legislation on pensions in Kosovo, tax losses, manufacturing or
storage damages, benefits in kind in the form of meals and transportation, unless organized by the businesses, etc.

ANALYSIS OF THE RELATIONSHIP BETWEEN THE ACCOUNTING AND FISCAL

RULES IN TERMS OF EXPENSE REPORTING

-Data, research questions, methodology and raised hypotheses-

The main objective of this research is the issue of acknowledging, recording and reporting expenses according to both
accounting and fiscal rules. The research goal is to obtain an overview whether businesses, namely the drafters of
financial statements, only use accounting rules, only fiscal rules or both regulatory frameworks simultaneously when
acknowledging, recording and reporting their expenses, thus establishing a relationship between expense accounting
and their fiscal reporting.

The above purpose is accomplished by asking the following research questions to accountants:

- Do you use accounting rules, fiscal rules, or both rules to acknowledge, record and report allowed business expenses?

- Do you use accounting rules, fiscal rules, or both rules to acknowledge, record and report disallowed business
expenses?

Research questions also raise the hypotheses to be tested, as follows:

H01 = Allowed and disallowed business expenses are reported by businesses according to the accounting regulatory
framework only;

H02 = Allowed and disallowed business expenses are reported by businesses according to the fiscal regulatory
framework only;

H03 = Allowed and disallowed business expenses are initially reported by businesses under the accounting regulations
and then if needed adjustments for fiscal purposes are made.

To test the hypotheses raised, in addition to the secondary data as above, provided by the legal regulations, the primary
data were also obtained, which were provided by the questionnaire filled out by 264 businesses with a turnover of
over million €, targeting the accountants employed in these businesses. The methodology used for data processing is
that of descriptive statistic and quantitative analytical method with cross-tabs, in order to analyze whether independent
variables representing the characteristics of respondents are of significance to the research questions, as dependent
variables.

-Analysis of data and research results-

Below we have a statistical analysis of the data related to the reporting of allowed business expenses, such as: lease
expenses, pension contribution expenses, bad debts, representation expenses, advertising and promotional expenses,
as well as disallowed business expenses, such as: land acquisition and enhancement expenses, as well as fines and
penalties. We also present the analysis of significance between the dependent variables (research questions) and
separate independent variables that represent the characteristics of respondents, such as their education, work
experience, and the degree of knowledge of the regulatory framework.

Regarding education, out of 264 respondents, 141 or 53.4% have undergraduate degrees, 120 or 45.5% have obtained
MA or PhD degrees, while only 3 persons, or 1.1%, do not have university degrees.

Allowed business expenses

549
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1 When acknowledging, recording and reporting allowed business expenses "Rental Expenses", what rules
do you use?

70.0 59.1
60.0
50.0
40.0
30.0 25.4
20.0 15.5
10.0
0.0
Accounting rules, then fiscal rules Fiscal rules Accounting rules

Source: Research conducted from author on a questionnaire to accountants

Figure 1 When acknowledging, recording and reporting allowed business expenses "Rental Expenses", what rules
do you use?

Valid
Frequency Percent
Accounting rules, than fiscal rules 156 59.1
Fiscal rules 41 15.5
Accounting rules 67 25.4
Total 264 100.0

The data show that around 60.0% answered that they use accounting rules, then fiscal adjustments are made, while
25.4% only use accounting rules, and 15.5% responded that they only use fiscal rules when recognizing and recording
lease expenses.

Table 1 Acknowledging, recording and reporting of rent expenses versus education (in percentage)
What is your education level?
Faculty/BA MA or PhD High School Total
Accounting rules, then
58.2 61.7 59.1
fiscal rules
Fiscal rules 17.0 14.2 15.5
Accounting rules 24.8 24.2 100.0 25.4
Total 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

The data show that the variable educational background leads respondents to state that they use accounting rules and
then fiscal adjustments related to these expenses are made. Of those with a MA or PhD degree, 61.7% made the first
choice, same as 58.2% of those with a BA degree, compared to employees without university degree who unanimously
declare that they only use the accounting rules.

Table 2 Acknowledging, recording and reporting of rent expenses versus work experience (in percentage)

550
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

How many years have you practiced your profession?

1-3 years 4-6 years Over 6 years Less than a year Total
Accounting rules,
65.6 40.5 60.9 100.0 59.1
then fiscal rules
Fiscal rules 18.8 35.7 10.9 15.5
Accounting rules 15.6 23.8 28.3 25.4
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Thus, professional experience leads accountants and auditors to choose the first option, which is that rent expenses
are recorded and reported according to accounting rules, and then necessary adjustments or disclosures on fiscal issues
are made.

Table 3 Acknowledging, recording and reporting of rent expenses versus knowledge of the legal basis (in percentage)

How much do you know about the legal framework

I have an average I have enough I have little I have great
knowledge knowledge knowledge knowledge Total
Accounting rules,
46.0 62.8 20.7 84.7 59.1
then fiscal rules
Fiscal rules 14.3 20.4 20.7 5.1 15.5
Accounting rules 39.7 16.8 58.6 10.2 25.4
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Thus, knowledge of the legal framework leads accountants to point out that in the case of rent expenses recording and
reporting, they use accounting rules and then adjustments or disclosures on fiscal matters are made. Of those with
great and sufficient knowledge of the law, 84.7% and 62.8 respectively made this choice, while 10.2%% and 16.8%,
respectively, stated they only use accounting rules, and 5.1% and 20.4% only fiscal rules.

Figure 2 When acknowledging, recording and reporting allowed business expenses for "Pension Contribution
Expenses", what rules do you use?

Valid
Frequency Percent

Accounting rules, than fiscal rules 128 48.5

Fiscal rules 71 26.9
Accounting rules 65 24.6
Total 264 100.0

551
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 2 When acknowledging, recording and reporting allowed business expenses for "Pension Contribution
Expenses", what rules do you use?

60.0
48.5
50.0
40.0
26.9 24.6
30.0
20.0
10.0
0.0
Accounting rules, then fiscal rules Fiscal rules Accounting rules

Source: Research conducted from author on a questionnaire to accountants

To this research question, 48.5% of respondents answered that they use accounting rules, and then fiscal adjustments
are made, 26.9% only use fiscal rules and 24.6% only use accounting rules.

Table 4 Acknowledging, recording and reporting of expenses for retirement contributions versus education (in
percentage)

What is your education level?

High
Faculty/BA MA/PhD School Total
Accounting rules, then
46.8 51.7 48.5
fiscal rules
Fiscal rules 29.8 24.2 26.9
Accounting rules 23.4 24.2 100.0 24.6
Total 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

When this research question is cross-tabbed with the education level of the respondents, the school education variable
leads them to state in greater proportions that they use accounting rules and then fiscal adjustments are made. Of those
with a MA or PhD degree, 51.7% made the first choice, same as 46.8% of those with a BA degree, compared to
employees without university degree who unanimously declare that they only use the accounting rules.

Table 5 Acknowledging, recording and reporting of expenses for retirement contributions versus professional
experience (in percentage)

How many years have you practiced your profession?

1-3 years 4-6 years Over 6 years Less than a year Total
Accounting rules, then
56.3 45.2 46.2 100.0 48.5
fiscal rules
Fiscal rules 28.1 42.9 23.9 26.9
Accounting rules 15.6 11.9 29.9 24.6
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Thus, professional experience leads accountants to choose the first option, which is that pension contributions are
recorded and reported according to accounting rules, and then necessary adjustments or disclosures on fiscal issues

552
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

are made. Of those with 1 to 3 years of work experience, 56.3% have claimed this, versus 16.6% who said they only
use fiscal rules and 28.1% who only use accounting rules. The situation is the same in other cases in terms of the
duration of professional experience.

Table 6 Acknowledging, recording and reporting of expenses for retirement contributions versus knowledge of legal
basis (in percentage)

How much do you know about the legal framework?

I have an
average I have enough I have little I have great
knowledge knowledge knowledge knowledge Total
Accounting rules, then
49.2 44.2 20.7 69.5 48.5
fiscal rules
Fiscal rules 19.0 38.9 20.7 15.3 26.9
Accounting rules 31.7 16.8 58.6 15.3 24.6
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Thus, knowledge of the legal framework leads accountants to point out that in the case of pension contribution
expenses recording and reporting, they use accounting rules and then adjustments or disclosures on fiscal matters are
made. Of those with great, sufficient and average knowledge of the law, 84.7%, 44.2 and 49.2% respectively made
this choice, while 15.3%, 16.8% and 31.7%, respectively, stated they only use accounting rules, and 15.3%, 38.9%
and 19.0% only fiscal rules.

Figure 3 When acknowledging, recording and reporting allowed business expenses for "Bad Debt Expenses", what
rules do you use?

Valid
Frequency Percent

Accounting rules, than fiscal rules 168 63.6

Fiscal rules 42 15.9
Accounting rules 54 20.5
Total 264 100.0

Figure 3 When acknowledging, recording and reporting allowed business expenses for "Bad Debt Expenses", what
rules do you use?

553
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

70.0 63.6
60.0
50.0
40.0
30.0 20.5
20.0 15.9
10.0
0.0
Accounting rules, then fiscal rules Fiscal rules Accounting rules

Source: Research conducted from author on a questionnaire to accountants

Regarding the question that when acknowledging, recording and reporting allowed business expenses, more
specifically in case of bad debts, what kind of rule you use, 63.6% answered that they first use the accounting rules
and then fiscal adjustments are made, 20.5% answered that they use accounting rules and the rest, 15.9%, use fiscal
rules.

Table 7 Acknowledging, recording and reporting of bad debt expenses versus education (in percentage)

What is your education level?

Faculty/Bach. MA/PhD High School Total
Accounting rules, then
63.1 65.8 63.6
fiscal rules
Fiscal rules 21.3 10.0 15.9
Accounting rules 15.6 24.2 100.0 20.5
Total 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

When the question on acknowledging, recording and reporting bad debts using the accounting rules and then making
fiscal adjustments, fiscal rules or accounting rules, is cross-tabbed with the question on education level, the latter leads
them to state that they use the accounting rules and then fiscal adjustments are made. Of those with a MA or PhD
degree, 65.8% made the first choice, 63.1% of those with a BA degree, compared to employees without university
degree, who unanimously declare that they only use the accounting rules.

Table 8 Acknowledging, recording and reporting of bad debt expenses versus professional experience (in percentage)

How many years have you practiced your profession?

1-3 years 4-6 years Over 6 years Less than a year Total
Accounting rules, then
71.9 59.5 62.0 100.0 63.6
fiscal rules
Fiscal rules 18.8 21.4 14.7 15.9
Accounting rules 9.4 19.0 23.4 20.5
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Professional experience leads accountants to state that bed debt expenses are recorded and reported according to
accounting rules, and then the necessary adjustments or disclosures on fiscal issues are made. Of those with 1 to 3
years of work experience, 71.9% have claimed this, versus 9.4% who said they only use fiscal rules and 18.8% who

554
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

only use accounting rules. The situation is the same with the professionals with 4 to 6 years of work experience, of
whom 59.5% use the rules as in the first case, versus 19.0% who only use accounting rules and 21.4% only fiscal
rules. Of those with more than 6 years of work experience, 62.0% made the first choice.

Table 9 Acknowledging, recording and reporting of bad debt expenses versus knowledge of the law (in percentage)

How much do you know about the legal framework?

I have an I have
average enough I have little I have great
knowledge knowledge knowledge knowledge Total
Accounting rules, then
58.7 66.4 20.7 84.7 63.6
fiscal rules
Fiscal rules 14.3 18.6 31.0 5.1 15.9
Accounting rules 27.0 15.0 48.3 10.2 20.5
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

The level of knowledge of the law also leads accountants to stress that in the case of bad debt expenses recording and
reporting, they use accounting rules and then adjustments or disclosures on fiscal matters are made. Of those with
great and sufficient knowledge of the law, 84.7% and 66.4 respectively made this choice, while 10.2%% and 15.0%,
respectively, stated they only use accounting rules, and 5.1%, 18.6% only fiscal rules.

Figure 4 When acknowledging, recording and reporting allowed business expenses “Representation Expenses", what
rules do you use?

Valid
Frequency Percent

Accounting rules, than fiscal rules 171 64.8

Fiscal rules 51 19.3
Accounting rules 42 15.9
Total 264 100.0

Figure 4 When acknowledging, recording and reporting allowed business expenses “Representation Expenses", what
rules do you use?

70.0 64.8
60.0
50.0
40.0
30.0 19.3
20.0 15.9
10.0
0.0
Accounting rules, then fiscal rules Fiscal rules Accounting rules

Source: Research conducted from author on a questionnaire to accountants

555
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

To the question “When acknowledging, recording and reporting allowed business expenses, representation expenses,
what rules you use”, 64.8% responded that they first use the accounting rules then fiscal adjustments are made
regarding the registration and reporting of these expenses, while 19.3% stated that they only use fiscal rules and 15.9%
stated that they only use accounting rules.

Table 10 Acknowledging, recording and reporting of representation expenses versus education (in percentage)

What is your education level?

Faculty/Bach. MA/PhD High School Total
Accounting rules, then fiscal
65.2 65.8 64.8
rules
Fiscal rules 19.1 20.0 19.3
Accounting rules 15.6 14.2 100.0 15.9
Total 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

The education level variable leads them to state that they use accounting rules and then fiscal adjustments are made.
Of those with a MA or PhD degree, 65.8% made the first choice, same as 65.2% of those with a BA degree, compared
to employees without university degree, all of whom state that they only use the accounting rules. Thus, education has
e dependent relationship with the use of accounting and fiscal regulations by professionals regarding the recording
and reporting of representation expenses.

Table 11 Acknowledging, recording and reporting of representation expenses versus professional experience (in
percentage)

How many years have you practiced your profession?

1-3 years 4-6 years Over 6 years Less than a year Total
Accounting rules, then
71.9 52.4 65.2 100.0 64.8
fiscal rules
Fiscal rules 28.1 28.6 16.3 19.3
Accounting rules 19.0 18.5 15.9
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Professional experience leads accountants to state that representation expenses are recorded and reported according
to accounting rules, and then they make necessary adjustments or disclosures on fiscal issues. Of those with 1 to 3
years of work experience, 71.9% have claimed this, versus 28.1% who said they only use fiscal rules. The situation is
the same with the professionals with 4 to 6 years of work experience, of whom 52.4% use the rules as in the first case,
versus 19.0% who only use accounting rules and 28.6% only fiscal rules. Of those with more than 6 years of work
experience, 62.5% made the first choice, and also all respondents with one year of experience.

Table 12 Acknowledging, recording and reporting of representation expenses versus knowledge of the
legal basis (in percentage)

How much do you know about the legal framework?

I have an
average I have enough I have little I have great
knowledge knowledge knowledge knowledge Total

556
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Accounting rules, then fiscal

63.5 61.1 31.0 89.8 64.8
rules
Fiscal rules 14.3 29.2 20.7 5.1 19.3
Accounting rules 22.2 9.7 48.3 5.1 15.9
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

The level of knowledge of the law also leads accountants and auditors to stress that in the case of representation
expenses recording and reporting, they use accounting rules and then adjustments or disclosures on fiscal matters are
made. Of those with great and sufficient knowledge of the law, 89.9% and 61.1 respectively made this choice, while
5.1% and 9.7%, respectively, stated they only use accounting rules, and 5.1% and 29.2% only fiscal rules.

Figure 5 When acknowledging, recording and reporting allowed business expenses, "Advertisement and promotion
expenses", what rules do you use?

Valid
Frequency Percent

Accounting rules, than fiscal rules 157 59.5

Fiscal rules 59 22.3
Accounting rules 48 18.2
Total 264 100.0

Figure 5 When acknowledging, recording and reporting allowed business expenses, "Advertisement and promotion
expenses", what rules do you use?

70.0 59.5
60.0
50.0
40.0
30.0 22.3
18.2
20.0
10.0
0.0
Accounting rules, then fiscal rules Fiscal rules Accounting rules

Source: Research conducted from author on a questionnaire to accountants

Regarding advertising and promotional expenses, 59.5% of respondents answered that they use accounting rules and
then fiscal adjustments are made, 22.3% answered that they only used fiscal rules and the remaining 18.2% answered
that they only use accounting rules.

Table 13 Acknowledging, recording and reporting of advertisement expenses versus education (in percentage)

557
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

What is your education level? Total

Faculty/Bac
h. MA/PhD High School
Accounting rules, then fiscal
54.6 66.7 59.5
rules
Fiscal rules 23.4 19.2 100.0 22.3
Accounting rules 22.0 14.2 18.2
Total 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

When the question on acknowledging, recording and reporting advertisement and promotional expenses using the
accounting rules and then making fiscal adjustments, fiscal rules or accounting rules, is cross-tabbed with the question
on education level, the latter leads them to state that they use the accounting rules and then fiscal adjustments are
made. Of those with a MA or PhD degree, 66.7% made the first choice, same as 54.6% of those with a BA degree,
compared to employees without university degree, all of whom (100%) state that they only use accounting rules.

Table 14 Acknowledging, recording and reporting of advertisement expenses versus professional experience (in
percentage)

How many years have you practiced your profession?

Over 6
1-3 years 4-6 years years Less than a year Total
Accounting rules, then
71.9 52.4 59.2 50.0 59.5
fiscal rules
Fiscal rules 28.1 21.4 22.3 22.3
Accounting rules 26.2 18.5 50.0 18.2
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Work experience leads accountants to state that advertisement and promotional expenses are recorded and reported
according to accounting rules, and then the necessary adjustments or disclosures on fiscal issues are made. Of those
with 1 to 3 years of work experience, 71.9% have claimed this, versus 28.1% who said they only use fiscal rules. The
situation is the same with the professionals with 4 to 6 years of work experience, of whom 52.4% use the rules as in
the first case, versus 26.2% who only use accounting rules and 21.4% only fiscal rules. Of those with more than 6
years of work experience, 62.5% made the first choice, while 50.0% of respondents with one year of experience have
made the same choice and the remaining 50.0% only use accountingrules.

Table 15 Acknowledging, recording and reporting of advertisement expenses knowledge of the legal
framework (in percentage)

How much do you know about the legal framework?

I have an I have
average enough I have little I have great
knowledge knowledge knowledge knowledge Total
Accounting rules, then
54.0 59.3 10.3 89.8 59.5
fiscal rules
Fiscal rules 23.8 25.7 41.4 5.1 22.3
Accounting rules 22.2 15.0 48.3 5.1 18.2
Total 100.0 100.0 100.0 100.0 100.0

558
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Source: Research conducted from author on a questionnaire to accountants

The level of knowledge of the law also leads accountants to stress that in the case of advertisement and promotional
expenses recording and reporting, they use accounting rules and then adjustments or disclosures on fiscal matters are
made. Of those with great and sufficient knowledge of the law, 89.9% and 59.3 respectively made this choice, while
5.1% and 15.0%, respectively, stated they only use accounting rules, and 5.1% and 25.7% only fiscal rules.

Disallowed business expenses

Figure 6 When acknowledging, recording and reporting disallowed business expenses, "Land acquisition and
enhancement expenses", what rules do you use?

Valid
Frequency Percent
Accounting rules, than fiscal rules 162 61.4
Fiscal rules 27 10.2
Accounting rules 75 28.4
Total 264 100.0

Figure 6 When acknowledging, recording and reporting disallowed business expenses, "Land acquisition and
enhancement expenses", what rules do you use?

70.0 61.4
60.0
50.0
40.0
28.4
30.0
20.0 10.2
10.0
0.0
Accounting rules, then fiscal rules Fiscal rules Accounting rules

Source: Research conducted from author on a questionnaire to accountants

Professionals were also asked about the rules they use when acknowledging, recording and reporting disallowed
expenses, more specifically on land acquisition and enhancement costs. As a result, 61.4% responded that they use
accounting rules, then fiscal adjustments are made, while 28.4% only use accounting rules and 10.2% only use fiscal
rules.

Table 16 Acknowledging, recording and reporting of land acquisition and enhancement expenses versus education (in
percentage)

What is your education level?

Faculty/Bach. MA/PhD High School Total
Accounting rules, then
58.9 65.8 61.4
fiscal rules
Fiscal rules 12.8 7.5 10.2

559
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Accounting rules 28.4 26.7 100.0 28.4

Total 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

The variable educational background leads respondents to declare that they use accounting rules and then fiscal
adjustments are made. Of those with a MA or PhD degree, 65.8% made the first choice, same as 58.9% of those with
a BA degree, compared to employees without university degree who unanimously declare that they only use the
accounting rules.

Table 17 Acknowledging, recording and reporting of land acquisition and enhancement expenses versus professional
experience (in percentage)

How many years have you practiced your profession?

Over 6
1-3 years 4-6 years years Less than a year Total
Accounting rules, then
56.3 78.6 58.7 50.0 61.4
fiscal rules
Fiscal rules 28.1 7.1 8.2 10.2
Accounting rules 15.6 14.3 33.2 50.0 28.4
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Professional experience leads accountants to state that land acquisition and enhancement expenses are recorded and
reported according to accounting rules, and then the necessary adjustments or disclosures on fiscal issues are made.
Of those with 1 to 3 years of work experience, 56.3% have claimed this, versus 15.6% who said they only use fiscal
rules and 28.1% who only use accounting rules. Of the professionals with 4 to 6 years of work experience, 78.6% of
them use the rules as in the first case, versus 14.3% who use only accounting rules and 7.1% only fiscal rules. Of those
with more than 6 years of work experience, 58.7% made the first choice.

Table 18 Acknowledging, recording and reporting of land acquisition and enhancement expenses versus
knowledge of legal framework (in percentage)

How much do you know about the legal framework?

I have an average I have enough I have little I have great
knowledge knowledge knowledge knowledge Total
Accounting rules,
63.5 51.3 37.9 89.8 61.4
then fiscal rules
Fiscal rules 14.3 10.6 20.7 10.2
Accounting rules 22.2 38.1 41.4 10.2 28.4
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Thus, knowledge of the legal framework leads accountants to point out that in the case of land acquisition and
enhancement expenses recording and reporting, they use accounting rules and then adjustments or disclosures on fiscal
matters are made. Of those with great and sufficient knowledge of the law, 89.8% and 51.3%, respectively, have
chosen the first option, while 10.2% and 38.1%, respectively, said they only use the accounting rules. Of the
professionals with average knowledge, 63.5% made the first choice, while 22.2% of them use the accounting rules
and 14.3% use fiscal rules when recording this category of expenses.

560
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 7 When acknowledging, recording and reporting disallowed business expenses, "Fines and penalties", what
rules do you use?

Valid
Frequency Percent
Accounting rules, than fiscal rules 149 56.4
Fiscal rules 41 15.5
Accounting rules 74 28.0
Total 264 100.0

Figure 7 When acknowledging, recording and reporting disallowed business expenses, "Fines and penalties", what
rules do you use?

56.4
60.0
50.0
40.0
28.0
30.0
20.0 15.5
10.0
0.0
Accounting rules, then fiscal rules Fiscal rules Accounting rules

Source: Research conducted from author on a questionnaire to accountants

When professionals were asked what rules they use when acknowledging, recording and reporting disallowed fines
and penalties, around 56.4% responded that they use accounting rules, then fiscal adjustments are made, 28.0% only
use accounting rules and 15.5% only use fiscal rules.

Table 19 Acknowledging, recording and reporting of fines and penalties versus education (in percentage)

What is your education level?

Faculty/Bach. MA/PhD High School Total
Accounting rules, then fiscal rules 57.4 56.7 56.4
Fiscal rules 19.1 11.7 15.5
Accounting rules 23.4 31.7 100.0 28.0
Total 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

The variable educational background leads respondents to declare that they use accounting rules and then fiscal
adjustments are made. Of those with a MA or PhD degree, 56.7% made the first choice, same as 57.4% of those with
a BA degree, compared to employees without university degree who unanimously declare that they only use the
accounting rules.

Table 20 Acknowledging, recording and reporting of fines and penalties versus professional experience (in
percentage)

561
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

How many years have you practiced your profession?

1-3 years 4-6 years Over 6 years Less than a year Total
Accounting rules, then fiscal
56.3 66.7 54.3 50.0 56.4
rules
Fiscal rules 28.1 21.4 12.5 15.5
Accounting rules 15.6 11.9 33.2 50.0 28.0
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Professional experience leads accountants to state the expenses for fines and penalties are recorded and reported
according to accounting rules, and then they make necessary adjustments or disclosures on fiscal issues. Of those with
1 to 3 years of work experience, 56.3% have claimed this, versus 15.6% who said they only use fiscal rules and 28.1%
who only use accounting rules. Of the professionals with 4 to 6 years of work experience, 66.7% of them use the rules
as in the first case, versus 11.9% who only use accounting rules and 21.4% only fiscal rules. Of those with more than
6 years of work experience, 54.3% made the first choice.

Table 21 Acknowledging, recording and reporting of fines and penalties versus knowledge of the legal framework
(in percentage)

How much do you know about the legal framework?

I have an I have
average enough I have little I have great
knowledge knowledge knowledge knowledge Total
Accounting rules, then
58.7 44.2 41.4 84.7 56.4
fiscal rules
Fiscal rules 19.0 20.4 20.7 15.5
Accounting rules 22.2 35.4 37.9 15.3 28.0
Total 100.0 100.0 100.0 100.0 100.0
Source: Research conducted from author on a questionnaire to accountants

Thus, knowledge of the legal framework leads accountants to stress that in the case of fines and penalties expenses
recording and reporting, they use accounting rules and then adjustments or disclosures on fiscal matters are made. Of
those with great and average knowledge of the law, 84.7% and 57.5%, respectively, have chosen the first option, while
15.3% and 22.1%, respectively, said they only use the accounting rules.

-Summary of research results and confirmation of hypotheses-

The results of the data analysis reveal the current situation of the accounting regulatory framework and fiscal reporting
of expenses. Moreover, from the data analysis we obtained information that businesses acknowledge, record and report
expenses under the accounting regulatory framework, then the necessary adjustments for fiscal requirements are made.
This regulatory applies to all alowed and disallowed expenses by fiscal categorization, and are used as a case study in
this research.

The research results enabled us to prove the third hypothesis (H3) that allowed and disallowed business expenses are
initially reported by businesses under the accounting regulations and then if needed adjustments for fiscal purposes
are made. Meanwhile, the other two hypotheses raised in this study, (H1) allowed and disallowed business expenses
are reported by businesses according to the accounting regulatory framework only, and (H2) allowed and disallowed
business expenses are reported by businesses according to the fiscal regulatory framework only, are rejected.

562
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

On the other hand, the results from the quantitative analysis of the cross-tabulation method showed a significance link
between dependent variables (research questions) and the separate independent variables of respondent features. This
also shows a dependent relationship between financial reporting of expenses and their fiscal reporting.

We summarize the research results in the table below, to reflect the relationship between accounting and taxation
when acknowledging, recording and reporting expenses:

Table 22 Allowed and Disallowed Business Expenses: Accounting Rules & Fiscal Rules

Allowed and disallowed business expenses Accounting rules, Accounting Fiscal rules
then fiscal rules rules only only
Rent expenses 
Expenses on pension contributions 
Bad debt expenses 
Representation expenses 
Advertisement and promotion expenses 
Land acquisition and enhancement expenses 
Fines and penalties 

SUMMARY AND CONCLUSION

This study included an analysis of financial and fiscal reporting for acknowledging, recording and reporting expenses
by businesses in Kosovo.

Initially, the study presented the regulatory framework with regard to expense accounting, as well as the respective
fiscal regulation. The accounting legal framework requires businesses to acknowledge, record and report expenses in
accordance with accounting standards, while the fiscal legal framework classifies expenses for reporting purposes as
allowed and disallowed business expenses.

Further, the study provides an analysis backed by evidence of financial and fiscal reporting of expenses, with data
analysis results demonstrating a satisfactory level of knowledge of legal (financial and fiscal) frameworks, which
gives good hopes that professionals in businesses award proper importance to reporting issues, thus providing
appropriate, qualitative and reliable information on the financial situation of the businesses vis-à-vis multiple users of
this information, including tax authorities. Likewise, the research results proved that the drafters of the financial
statements acknowledge, record and report allowed and disallowed business expenses in accordance with the
accounting rules, then the necessary adjustments for fiscal requirements are made. This also identifies the fact that
there is a dependency relationship between accounting and taxation in terms of the businesses’ reporting of expenses.

REFERENCES

(2015). Retrieved from [Link] No. 05/L -028 ON PERSONAL INCOME TAX.

(2018). Retrieved from [Link]

(2019). Retrieved from [Link] NO.06/L-105 ON CORPORATE INCOME TAX.

(n.d.). Retrieved from [Link]

563
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Adeniran Samuel, Olusola Samuel, & Obiamaka. (2013). The Impact of International Financial Reporting Standards
on Taxation, 169.

Aldeila, S. (2019). The "true and fair view" concept in business and tax law: The Portuguese and Spanish
case studies. Academy of Accounting and Financial Studies.

Cuzdriorean, & Matiş. (2012). The relationship between accounting and taxation insight the european union: the
influence of the international accounting regulation, 28.

Fernandes, Cerqueira, & Brandao. (2017). Tax and Financial Reporting Aggressiveness: Evidence from Europe, 39.

Gavana, Guggiola, & Marenzi. (2015). Aligning taxation and international financial reporting standards:
evidence from italian listed companies, 82.

Plaksiienko, Melikhova, Yermolaieva, Chernenko, & Lipskyi. (2019). Formation of accounting and tax
policy of the company. Academy of Accounting and Financial Studies, 3.

Shaviro, D. N. (2008). The Optimal Relationship between Taxable Income and Financial Accounting
Income: Analysis and a Proposal. New York University School of Law, 65.

564
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Accounting Ethics and Organizational Culture: A Study of the Points of Contention and

Similarities

NIDA TÜREGÜN

Assistant Professor

Ozyegin University – School of Applied Sciences

[Link]@[Link]

ABSTRACT

The current study is motivated by the recent corporate failures that are associated with

accounting malpractices. Accounting ethics comprise of the moral elements that guide

accounting practices. They ensure that there are standards and principles that the accountants

would have to follow. Thus, the study aims to understand the approaches that could be used to

prevent accounting failures by improving accounting practices. Ethics in accounting and

organizational culture could help to improve the situation. This study is essential and helps to

understand organizational culture and accounting ethics and how both concepts affect

operations in organizations. The two are able to improve the productivity of companies by

ensuring that the right and moral behaviors are in place. Additionally, the study indicates the

role of management in ensuring that both accounting ethics and organizational culture are in

place to ensure effective operations in the company. The use of the findings of this paper could

assist to come up with the necessary measures that would improve organizational culture and

also establish the appropriate accounting ethics.

Key Words: Accounting Ethics, Organizational Culture, Contention, Similarities

JEL: M41, M14

565
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Accounting ethics defines the moral values and principles that are applied in accounting.

Mainly, accounting services demand high ethical standards to be maintained by accountants as

they dispense their duties (Duska et al., 2018). Notably, in recent years, there have been several

corporate collapses that have occurred as a result of failure to observe the necessary accounting

principles. Global scandals have been witnessed in various parts of the world, including

companies like Polly Peck, Commerce International, Bank of Credit, and others (Mabil, 2019).

Such cases have raised questions on how accounting ethics could affect the operations of a

company. Thus, it has become critical to come up with standards that need to be adhered to by

people working in this profession. With ethical regulations, governments and organizations

would be able to prevent and combat fraudulent accounting practices.

The nature of job positions of accountants places them in areas where several players rely on

them to get crucial and accurate information that would affect decisions in different ways. The

public, clients, and employers would use the information that accountants provide to make

certain decisions, and it is essential that these parties can trust this information. Accountants

are expected to operate ethically and professionally, following the laid down procedures,

principles, and rules (Duska et al., 2018). Adherence to ethics and moral standards would ensure

that scandals and criminal accounting activities do not occur.

On the other hand, organizational culture refers to various values, assumptions, beliefs, and

principles. In other words, culture defines how an organization engages in various activities.

The organizational culture guides the behavior of employees in the organization, and it

influences how people and groups think, feel, and interact. Organizational culture gives the firm

a particular identity that the stakeholders can associate themselves with. Organizational culture

has several implications for the company and could influence several elements. It affects

566
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

performance, effectiveness, and various processes in the company (Warrick, 2017). It would

affect the level of motivation among workers, and thus the extent to which they deliver to the

organization. Eventually, this could have a significant effect on the productivity and

profitability of a firm. Hence, it would be critical to align the production goals and the

organization culture. Organizational culture can improve the level of ethics in a company.

Therefore, it is possible to come up with accounting practices by establishing the correct culture

in the organization. The appropriate culture would make sure that the accountants and managers

prepare the accounting documents in the right way and engage in legal and ethical accounting

activities. The organizational culture would ensure people follow the principles that have been

laid down are followed and also ensure compliance with accounting standards.

The current study is motivated by the recent corporate failures that are associated with

accounting malpractices. Thus, the study aims to understand the approaches that could be used

to prevent accounting failures by improving accounting practices. Ethics in accounting and

organizational culture could help to improve the situation. The current study would be essential

in improving the understanding of how the two concepts could enhance financial improvements

and make sure accounting practices improves. The information would enrich the literature, and

make sure the situation changes.

LITERATURE REVIEW

Accounting Ethics

There exists literature that has examined ethics in accounting cases (Cohen et al., 1996; Elias,

2002; Smith, 2003; Bernardi and Bean, 2006; Uşurelu & Loghin, 2010; Gong, 2017; West 2018;

Gunz and Thorne, 2018; Payne et al., 2018). Accounting ethics focus on moral values,

567
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

judgments, human and business ethics, and how all these aspects apply in accounting (Duska

et al., 2018). In general, the primary aim of ethical values and accounting ethics is to uphold

accounting professionalism and also ensure that their good practice (Akadakpo and Enofe,

2013). Ethical responsibility and behavior are critical aspects of the business world. In most

organizations, the largest ethical issues involve accounting processes and practices in the

organization. Mainly, the breach of moral and ethical values and rules that govern corporate

financial reporting would jeopardize the primary aims of financial reporting (Enobong, 2017).

Usually, unethical behaviors would happen through financial misstatements, and it could have

significant effects on organizations. It would affect the level of trust that investors have in the

organization and also it influences the level of satisfaction among the clients (Baud et al, 2019).

Importantly, unethical accounting practices would damage the reputation of the company, and

thus its publicity.

Primarily, ethics investigates the moral principles and human behavior and aims to distinguish

bad from good, and vice versa (Smith and Smith, 2003). In businesses and organizations, the

code of conduct would guide the ethical behavior of the employees and stakeholders. The code

of ethics helps organizations to maintain a particular moral dimension. Ethical codes would

secure the fidelity of financial processes and business transactions (Gunz and Thorne, 2018).

The credibility of an organization, relationships in the organizations, and the performance

workers could be influenced by the ethical code that an organization has.

The accountants have a role to prepare accurate and timely financial reports that would indicate

the position of the organization’s transactions. Mainly, these reports are crucial to senior

managers and investors and guide them in making particular decisions in organizations (Gong,

2017). Thus, it is essential that accountants engage in ethical activities that would help them to

568
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

come up with the correct reports. In accounting, adherence to ethics assists in ensuring that

employees comply with ethical standards and internal controls (Todorović, 2018). With such

measures, it is possible for accountants to engage in activities that would ensure that they

perform, investigate, identify, and measure various activities, and identify frauds that could

occur in the organization. Mainly, unethical behaviors would lower the credibility of a person,

and also degrade the reputation of the company. Unethical practices would create room for

criminal activities that may affect the profit levels of an organization (Duska et al., 2018). Thus,

to avoid such cases, the accountants would have to observe and maintain professional ethics.

The private sector and the government have a commitment to instill ethical accounting

practices. It is for this reason that rules and regulations have been developed to regulate and

control accounting professionals (Chelariu et al., 2014). There is a professional code of ethics

that the accountants must observe in the line of their duties. Importantly, legislation has been

put in place to further enforce ethical behavior among accounting professionals. The Sarbanes-

Oxley Act of 2002 is one of the legislations in place to improve accounting ethics (Gunz and

Thorne, 2018). Mainly, the law was put in place to safeguard investors from fraudulent

activities relating to accounting. It aimed to enhance the reliability and accuracy of corporate

disclosures, and also improve compliance with the existing regulations and laws. Mainly, there

should be approaches that could help to detect and prevent fraud in accounting reports, and the

current legislation aimed to make this possible. According to this Act, the management must

assess the accuracy of financial reporting by having an effective internal control system. With

the Act, it was possible to ensure that the accountants and managers do not manipulate financial

information. It issues penalties for people who are responsible for fraudulent financial reports

and thus assisting to prevent unethical and criminal accounting activities.

569
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

According to Jaijairam (2017), in the accounting profession, financial ethics is one of the

greatest problems. Mostly, fraudulent activities would happen when managers and accountants

fail to adhere to particular ethical standards and change financial information in the reports. The

alterations lead to different and false results that deceive various players about the organizations

and their financial position. Studies (Cameron and O’Leary, 2015; Payne et al., 2018; Christy

and Dinah, 2018; Butcher, 2019; Ionescu, 2019; Baud et al., 2019) indicate that there are several

principles that are necessary for the accounting profession, and adherence to these principles

would ensure that unethical behaviors reduce. Integrity and objectivity are among the primary

principles that accountants should adhere to. Moreover, employees in the accounting field have

to adhere to professional competence and also indicate increased levels of confidentiality. The

behaviors of these people must also be professional, as the code of conduct for accountants’

dictates (Sepasi, 2019). With such principles and standards, it will be possible to realize

improved accounting ethics and ensure that accounting operations are carried in the correct

ways.

Moreover, as Ahinful et al. (2017) note in their article, society has high expectations on

accountants, and it is critical that these expectations are met. Mainly, ethics in this profession

dictates that financial reports should be beneficial to the end-users for them to make appropriate

financial decisions. Thus, the business society anticipates that accounting professionals will

comply with all ethical values and make sure that they provide transparent, accurate, and timely

information to all users. Primarily, any attempt to willingly provide false financial reports could

have significant implications for the business.

Poor ethical standards would increase the rate of fraud and criminal activities in organizations

(Idowu, 2018). The level of control and oversight decreases when there are no ethical

570
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

considerations in place. Such a situation creates weaknesses in the system that the auditors can

use to engage in unethical practices and also conceal evidence. Lack of ethics also creates room

for considerable data manipulations, and this could lead to cases of fraud and tax evasion.

Moreover, information indicates that the lack of adherence to ethical values in accounting could

lead to the damaged reputation of an organization (Todorović, 2018). Unethical activities by

accountants would negatively influence the trustworthiness of an organization and its reputation

to the stakeholders. Lack of trust that arises from unethical accounting activities would taint the

identity of the firm. It would also affect the business operations of the company.

Organizational Culture

The idea of organizational culture has developed significant attention in various circles over the

last decades (Tagiuri and Litwin, 1968; Frost et al.,1985; Ott, 1989; Denison, 1990; Schein,

2010). Several studies have been carried out on how organizations can develop and utilize

organizational culture to realize improved productivity (Nasiripour et al., 2009; Mathew, 2007;

Chegini, 2010; Uddin et al., 2013; Akbari et al., 2015; Kelepile, 2015; Nikpour, 2017). In

particular, organizational culture can affect the behavior of the workers and other stakeholders

in an organization (Zeyada, 2018). The behavior that the culture develops would have

significant implications on various aspects of the organization.

Mainly, each company has a unique aspect of the way in which it operates and carries its

activities. The operational uniqueness of the company and the assumptions, values, and beliefs

constitute organizational culture. Importantly, the uniqueness of the company that is based on

its culture differentiates from other organizations and gives it an identity (Odor, 2018). Culture

would influence how individual workers perform, and it will improve the overall success of the

organization.

571
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Zhang and Li (2013) noted that organizational culture is among the core competencies that a

company has. Though it is intangible, organizational culture various groups and individual act

based on the culture. Importantly, organizational culture is associated with the effectiveness

that the company shows (Zhang and Li, 2013). Thus, the organization needs to put measures in

place that would ensure that there is a beneficial organizational culture that would improve the

performance of the employees and the entire organization. Culture is also able to improve the

satisfaction of workers. It affects the psychological aspects of the people and improves their

morale and attachment to the organization.

The level of commitment of employees could improve the appropriate culture is in place

(Nikpour, 2017). The managers must adopt various approaches that would establish the right

culture, improve organizational commitment, and increase the performance in the company. In

fact, Ahmed and Shafiq (2014) argued that the most important thing that leaders in an

organization would do is to build and manage culture. They have to establish the appropriate

culture and manage it in a way that would improve the performance of the company (Ahmed

and Shafiq, 2014). Eventually, the profits of the firm would improve, and it will be possible for

the company to realize growth.

Most of the studies have focused on organizational culture and business ethics (McDaniel,

1995; Cohen, 1995; Dion, 1996; Warren et al., 2014; Khalşd et al., 2018; Manggai et al, 2018).

It is only a few studies that have investigated organizational culture and accounting ethics and

how the two concepts are interrelated (Douglas et al., 2001; Chen, 2001; Oseni, 2011; Zulfanef,

2015; Jaijairam, 2017). According to Zulganef (2015), the ethical values that accountants

should have would be derived from the culture of the organization. Thus, a culture that enforces

572
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

good morals would result in accountants engaging ethical practices. As information providers,

organizational culture ensures that the accountants have the right behaviors and apply ethics in

their activities (Zulganef, 2015). Such information would be useful in ensuring that improved

strategic decision making.

CONTENTION and SIMILARITIES

There are several areas of disagreement between accounting ethics and organizational culture.

Mainly, accounting ethics deal with behaviors that would ensure that moral standards are

followed when engaging in accounting activities (Melé et al., 2017). With accounting ethics, it

is possible to distinguish what is wrong accounting practices from those that are bad. It would

ensure that accountants engage in the right activities. Accounting ethics are concerned with

enforcing beliefs and behaviors that people have adopted as they go on with the duties. The

ethical principles that accountants can adopt would guide the behavior of these professionals

and ensure that they carry out their roles diligently (Jaijairam, 2017). It would result in

improved financial reporting and also lower cases of financial fraud. They guide personal and

professional behavior in organizations and among accountants to ensure improved accounting

practices.

On the contrary, organizational culture is concerned with beliefs and norms that the entire

organizations observe. It is not limited to a given profession but deals with all people in the

organization. It includes all the operations of the organization and the activities that the

employees and stakeholders engage in. Unlike accounting ethics, organizational culture gives

the company identity (Ilies and Metz, 2017). Usually, there are no particular rules that enforce

the culture of the organization compared to the situation of accounting ethics. However, the

management has to put measures that will ensure that there is a constructive culture. Essentially,

573
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

this is unlike the situation with the accounting ethics where several rules and regulations are in

place to ensure that everything is complied with. With organizational culture, it develops as a

result of activities and processes in the organization.

Notably, accounting ethics and organizational culture have several similarities. In particular,

both concepts help to instill and guide behaviors in the organization. Accounting ethics makes

sure that accountants have the right behavior that would make them comply with several

requirements in the organization (Duska et al., 2018). Such behaviors would ensure that

professionals engage in the right activities. Organizational culture makes sure that people can

have particular behaviors that would improve their productivity and that of the organization.

Additionally, the two concepts would require management for their effective realization in the

company. The management must put procedures and processes in place that make sure that

people develop the right behaviors, attitudes, and beliefs about the organization. With the right

culture, the management would lead to an improved culture in the organization. Importantly, to

have the right accounting ethics, the management would need to come up with policies and

develop a code of conduct for accountants (Avdeev et al., 2019). Effective leadership results in

improved ethics by the professionals and thus better practices.

Moreover, both aspects affect the rate of performance in the organization. In particular,

accounting ethics result in better account reporting (Oseni, 2011). Eventually, this would lead

to a case where there is reduced fraudulent activities and criminal activities. Ethical practices

would also ensure that the company has a sufficient number of investors. All these elements

would result in a situation in which the performance of the firm improves. On the other hand,

organizational culture improves the performance of the company by ensuring that there is the

574
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

correct behavior in the organization. It cultivates and builds the nature of people of the people

and ensures that they work hard towards the goals of the company. With increased commitment

and satisfaction among workers, the rate of productivity improves.

FINDINGS

The literature also indicated a possible relationship between organizational culture and

accounting ethics. In an article, Lozano (1998) indicated a possible relationship between ethics

and organizational cultures. The use of the two concepts can help to avoid manipulations that

happen in organizations. They would have significant impacts on accounting and ensure that

workers engage in their activities ethically, and reduce risks involved in accounting (Lozano,

1998). Through organizational culture and accounting ethics, it is possible to ensure that the

financial information that is given to investors, clients, employees, and managers is accurate.

Such information would be relied upon to make the right decisions by the various parties. Thus,

managers must ensure that the organization has the right culture that would ensure that there

are accounting ethics in the organization.

Additionally, organizational culture brings an ethical approach to do business among

organizations. Culture forms a foundation on which corporate social responsibility can be

actualized, and business ethics can be observed (Militaru and Zanfir, 2012). It would help the

employees and other stakeholders to engage in ethical activities. In essence, accountants would

be compelled by the organizational culture to engage in ethical accounting activities, and it

could improve the financial reporting in the organization.

Moreover, the information provided indicated that the leadership of an organization has a role

to play in developing accounting ethics and establishing organizational culture. In particular,

575
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the controller, chief financial officer, and chief executive officer are critical managers who need

to work closely with auditors and accountants to come up with ethical standards. The leaders

must ensure that there are no deficiencies and weaknesses in the design and operation of internal

controls, especially concerning financial statement reporting. All these parties have to work

towards improving processes concerned with summarization, reporting, and recording of

financial data.

The leadership also needs to be concerned with the identification and control of fraudulent

practices that the workers may engage in, both at low and high levels. Mainly, to realize this, it

would be critical that the management involves all the workers in the development of the

internal controls. The leaders must ensure that the design of the internal control system and its

implementation meets the requirement of the Sarbanes-Oxley Act. Constant improvements in

the internal control system are critical, and the management must ensure that ethical accounting

standards are met.

The accountants are also essential in ensuring that there are necessary ethical accounting

principles in the organization. They need to develop financial reporting in accordance with the

laid low rules, policies, and code of ethics. Their motivation should be to ensure accurate data

is provided for the stakeholders for the appropriate decision making. The auditors and

accountants have to assist in the collection of evidence that could help to confirm, test, and

compare fairness and validity of financial statements. Auditors must also help in the evaluation

and understanding of the internal control system, and come up with approaches that would help

to develop ethical accounting practices.

CONCLUSION

576
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The current study indicates that accounting ethics comprise of the moral elements that guide

accounting practices. They ensure that there are standards and principles that the accountants

would have to follow. Lack of ethics in this profession could lead to various cases of crimes

and fraud in the accounting world. In worst case situations, accounting malpractices would

result in the collapse of the organizations. Thus, it is essential that ethics are maintained in this

profession and within organizations. Importantly, organizational culture indicates the principles

and beliefs that an organization practices. It defines the unique way in which the organization

operates and how the employees behave. It is essential and it improves the behavior of the

workers and also enhances the productivity of the company. Thus, it is essential that a firm

establish an appropriate culture that would positively shape its operations.

The concepts have several differences and similarities as indicated in the above part. While one

dictates the behavior of members of a given profession, the other one guides the behavior of

employees in a particular organization. Importantly, accounting ethics are sustained by rules

and regulations, while no particular laws are necessary for organizational culture. However, the

two have the capacity to improve the performance of employees, and also that of a given

organization. Thus, organizations should strive to ensure that ethics in accounting are adhered

to, and a positive culture is in place.

The current study is essential and helps to understand organizational culture and accounting

ethics and how both concepts affect operations in organizations. The two are able to improve

the productivity of companies by ensuring that the right and moral behaviors are in place.

Additionally, the study indicates the role of management in ensuring that both accounting ethics

and organizational culture are in place to ensure effective operations in the company. The use

577
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of the findings of this paper could assist to come up with the necessary measures that would

improve organizational culture and also establish the appropriate accounting ethics. In the end,

the organization would be able to attract investors, improve employee performance, and

enhance productivity.

REFERENCE LIST

Ahinful, G., S. Addo, F. Boateng, and J. Danquah. 2017. Accounting Ethics and the

Professional Accountant: The Case of Ghana. International Journal of Applied

Economics, Finance and Accounting 1 (1): 30-36.

Ahmed, M., and S. Shafiq. 2014. The Impact of Organizational Culture on Organizational

Performance: A Case Study on Telecom Sector. Global Journal of Management and

Business Research 14 (3): 20-30.

Akadakpo, B. A., and A. O. Enofe. 2013. Impact of Accounting Ethics on the Practice of

Accounting Profession in Nigeria. IOSR Journal of Business and Management 12 (1):

45-51.

Akbari, P., K. Nazari, and A. Mahmoudi, 2015. Analysis of the Impact of Organizational

Culture on Employee Productivity. Case Study: Bistoon Petrochemical

Company. Global Journal of Multidisciplinary and Applied Sciences 3 (1): 5-9.

Avdeev, V., S. Nassiripour, and H. Wong. 2019. Case Study: Ethical Considerations of an

Accounting Professional. Journal of Leadership, Accountability and Ethics 16 (2): 10-

14.

Baud, C., M. Brivot, and D. Himick. 2019. Accounting Ethics and the Fragmentation of

Value. Journal of Business Ethics, 1-15.

578
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Bernardi, R. A., and D. F. Bean. 2006. The Growth of Accounting Research in Ethics Journals.

Research on Professional Responsibility and Ethics in Accounting. Bingley: Emerald

Group Publishing Limited.

Butcher, D. 2019. Adapting Accounting Ethics to New Technology. Strategic Finance 101 (1):

19-20.

Cameron, R. A. and C. O'Leary. 2015. Improving Ethical Attitudes or Simply Teaching Ethical

Codes? The Reality of Accounting Ethics Education. Accounting Education 24 (4):

275-290.

Chegini, M. G. 2010. The Relationship Between Organizational Culture and Staff Productivity

Public Organizations. Journal of Social Sciences 6 (1): 127-129.

Chelariu, A., E. Horomnea, and T. Florentin. 2014. Education Regarding Ethics in the

Accounting Profession–A Literature Review. EIRP Proceedings 9: 372-379.

Chen, T. T. 2001. Ethics Control Mechanisms: A Comparative Observation of Hong Kong

Companies. Journal of Business Ethics 30 (4): 391-400.

Christy, C., and P. Dinah. 2018. Accounting Ethics: Principles to Work by. International

Journal of Advances in Management and Economics 1 (5): 175-182.

Cohen, D. V. 1995. Ethics and Crime in Business Firms: Organizational Culture and the Impact

of Anomie. Advances in Criminological Theory 6, 183-206.

Cohen, J. R., L. W. Pant, and D. J. Sharp. 1996. A Methodological Note on Cross-Cultural

Accounting Ethics Research. The International Journal of Accounting 31 (1): 55-66.

Denison, D. R. 1990. Corporate Culture and Organizational Effectiveness. John Wiley & Sons

Dion, M. 1996. Organizational Culture as Matrix of Corporate Ethics. The International

Journal of Organizational Analysis 4 (4): 329-351.

579
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Douglas, P. C., R. A. Davidson, and B. N. Schwartz. 2001. The Effect of Organizational Culture

and Ethical Orientation on Accountants' Ethical Judgments. Journal of Business

Ethics 34 (2): 101-121.

Duska, R. F., B. S. Duska, and K. W. Kury. 2018. Accounting ethics. New Jersey: Wiley-

Blackwell.

Elias, R. Z. 2002. Determinants of Earnings Management Ethics Among Accountants. Journal

of Business Ethics 40 (1): 33–45.

Enobong, U. 2017. Implications of Accountants’ Unethical Behavior and Corporate Failures.

International Journal of Business 4 (4): 82-94.

Frost, P. J., L. F. Moore, M. R. E. Louis, C. C. Lundberg, and J. E. Martin. 1985. Organizational

Culture. Sage Publications, Inc.

Gong, J. J. 2017. Ethics in Accounting: A Decision-Making Approach. Journal of Business

Ethics 142: 621–623.

Gunz, S., and L. Thorne. 2018. Thematic Symposium: Accounting Ethics and Regulation: SOX

15 Years Later. Journal of Business Ethics 158 (2): 293-296.

Idowu, A. 2018. Accounting Ethics and Professionalism on Fraud Management: Nigerian

Experience. Sumerianz Journal of Economics and Finance 1 (2): 49-59.

Ilies, L., and D. Metz. 2017. Organizational Culture: Key Issues. A Literature Review. Annals

of Faculty of Economics 1 (1): 797-805.

Ionescu, L. 2019. Communicating Accounting Ethics. Analele Universităţii Spiru Haret. Seria

Jurnalism 20 (1): 53-57.

Jaijairam, P. 2017. Ethics in Accounting. Journal of Finance and Accountancy 23: 1-13.

Kelepile, K. 2015. Impact of Organizational Culture on Productivity and Quality Management:

A Case Study in Diamond Operations Unit, dtc Botswana. International Journal of

Research in Business Studies and Management, 2 (9): 293–296.

580
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Khalid, K., S. E. Eldakakand, and A. T. Bon. 2018. Exploring the Dynamic Organizational

Culture on Ethics and Compliance: Engineering Perspective. International Journal of

Engineering & Technology, 7 (3.20): 104-108.

Lozano, J. M. 1998. Ethics and Corporate Culture: A Critical Relationship. Ethical

Perspectives 5 (1): 53-70.

Mabil, A. N. 2019. Investigating Effects of Accounting Ethics on Quality of Financial

Reporting of an Organization: Case of Selected Commercial Banks in South

Sudan. Mediterranean Journal of Social Sciences 10 (1): 177-191.

Manggai, B. A., K. Thukiman, M. F. Othman, and M. K. Abdul Majid. 2018. Organizational

Culture and Ethics in Decision-Making. International Journal of Engineering &

Technology. 7 (2.29): 257-259.

Mathew, J. 2007. The Relationship of Organisational Culture With Productivity and Quality: A

Study of Indian Software Organisations. Employee Relations 29 (6): 677-695.

McDaniel, C. 1995. Organizational Culture and Ethics Work Satisfaction. The Journal of

Nursing Administration. 25 (11): 15-21.

Melé, D., J. M. Rosanas, and J. Fontrodona. 2017. Ethics in Finance and Accounting: Editorial

Introduction. Journal of Business Ethics 140 (4): 609-613.

Militaru, C., and A. Zanfir. 2012. The Influence of Organizational Culture Over the Ethical

Principles in International Businesses. International Journal of Academic Research in

Accounting, Finance and Management Sciences 2 (1): 26-33.

Nasiripour, A. A., P. Raeisi, and S. P. Hedayati. 2009. The Relationship Between

Organizational Cultures and Employees Productivity. Journal of Health

Administration 12 (35): 17-24.

581
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Nikpour, A. 2017. The Impact of Organizational Culture on Organizational Performance: The

Mediating Role of Employee’s Organizational Commitment. International Journal of

Organizational Leadership 6: 65-72.

Odor, H. O. 2018. Organisational Culture and Dynamics. Global Journal of Management and

Business Research 18 (1): 22-29.

Oseni, A. I. 2011. Unethical Behavior by Professional Accountant in an Organization. Research

Journal of Finance and Accounting 2 (2): 106-111.

Ott, J. S. 1989. The organizational culture perspective. Chicago: Dorsey Press.

Payne, D. M., C. M. Corey, and C. Raiborn. 2018. A Model Code of Ethics for Decision Making

in Accounting Professions. Southwest Academy of Management Proceedings Annual

Meeting 195: 215–235.

Schein, E. H. 2010. Organizational Culture and Leadership. New Jersey: John Wiley & Sons.

Sepasi, S. 2019. Accounting Ethics. International Journal of Ethics and Society 2 (1): 23-29.

Smith, K. T., and L. M. Smith. 2003. Business and Accounting Ethics. T & M University:

Texas.

Smith, L. M. 2003. A Fresh Look at Accounting Ethics. Accounting Horizons 17 (1): 47-49.

Tagiuri, R., and G. H. Litwin. 1968. Organizational Culture: A Key to Financial

Performance. Organizational Climate and Culture. San Francisco: Jossey-Bass.

Todorović, Z. 2018. Application of Ethics in the Accounting Profession with an Overview of

the Banking Sector. Journal of Central Banking Theory and Practice 7 (3): 139-158.

Uddin, M. J., R. H. Luva, and S. M. M. Hossian. 2013. Impact of Organizational Culture on

Employee Performance and Productivity: A Case Study of Telecommunication Sector

in Bangladesh. International Journal of Business and Management 8 (2): 63-77.

582
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Uşurelu, V. I., and D. Loghin. 2010. Accountıng Ethıcs-Responsıbılıty Versus

Creatıvıty. Annals of the University of Petrosani Economics 10 (3): 349-356.

Warren, D. E., J. P. Gaspar, and W. S. Laufer. 2014. Is Formal Ethics Training Merely

Cosmetic? A Study of Ethics Training and Ethical Organizational Culture. Business

Ethics Quarterly 24 (1): 85-117.

Warrick, D. D. 2017. What Leaders Need to Know About Organizational Culture. Business

Horizons 60 (3): 395-404.

West, A. 2018. After Virtue and Accounting Ethics. Journal of Business Ethics 148 (1): 21-36.

Zeyada, M. 2018. Organizational Culture and its Impact on Organizational Citizenship

Behavior. International Journal of Academic Research in Business and Social

Sciences 8 (3): 418-429.

Zhang, X., and B. Li. 2013. Organizational Culture and Employee Satisfaction: An Exploratory

Study. International Journal of Trade, Economics and Finance 4 (1): 48-54.

Zulganef, Z. 2015. The Roles of Organizational Culture and Ethics in Shaping the Behavior of

Accountant from the Perspective of Strategic Management. International Journal of

Economics, Commerce and Management 3 (7): 441-455.

583
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

THE EVOLUTION OF TAX BURDEN AND TAX EFFORT IN

THE OECD BETWEEN 2000 AND 2015

Maria João Barros Alves Cavadas

Department of Management, ISCTE-University Institute of Lisbon, Lisbon, Portugal

mjcavadas@[Link]

Master in Management from ISCTE-University Institute of Lisbon in 2018.

Paulo Jorge Varela Lopes Dias

Department of Accounting, ISCTE-University Institute of Lisbon, Lisbon, Portugal

[Link]@[Link] (corresponding author)

PhD in Accounting and MBA in Management, by ISCTE-University Institute of Lisbon, Post-

Graduate in Finance, by CEMAF-ISCTE and Graduate in Management, by UML. He is
coordinator of Courses of Taxation, Business Taxation, Company Tax Management, Principles
and Procedures of Tax Litigation and Taxation of Non-Residents. He is Director of the MsC in
Management at ISCTE-University Institute of Lisbon and Director of Post-Graduation in Tax
Management at INDEG-ISCTE.

584
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ABSTRACT

Fiscal policies are one of the main concerns of any government. Balancing the decisions
regarding public expenditure and how that expenditure is to be financed is paramount to
allow for a favourable level of welfare in a society. The expenditure side is dependent
upon the collection of revenues that can finance the government’s expenses. These
revenues come mostly from taxes. Therefore, taxation plays a pivotal role in a society,
and it is the focus of this dissertation. Greater attention is paid to what can be considered
measures of the level of taxation of a country – tax burden and tax effort. There is great
deal of concepts attached to taxation, hence the necessity to have them defined and
explained in order to make any further analyses. A systematic literature review is
performed to study the investigation on taxation since 1972.

A database of data for 35 of the 36 countries of the OECD is used to compute the three
aforementioned metrics, which are then analysed and compared between them. An
alternative measure of tax effort (the World Tax Index) is briefly described, since it is a
breakthrough in the field. The development of new metrics in future investigation is
proposed.

KEYWORDS: tax burden; tax effort; taxation; tax systems

JEL Classification: H200; H220

585
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1. INTRODUCTION

“Taxation has always been a central issue in political economy because it is one of the
main activities of all states and a necessary condition for everything else states do. It is
the core feature of state capacity.” (Kiser and Karceski, 2017: 76).

The aim of this dissertation is to assess the evolution of the level of taxation as measured
by the tax burden and the tax effort in the OECD during the period from 2000 until 2015.
An introductory part (section 2) aims at defining and explaining introductory and
fundamental concepts related to taxation. In section 3, a systematic literature review is
performed in order to analyse the investigation on taxation in literature.

In section 4, three metrics for the calculation of the tax burden and the tax effort of
countries are explained and computed: Frank’s index, Bird’s index, and the tax-to-GDP
ratio. Furthermore, an alternative metric for the measurement of the tax effort is
summarised. In section 5, a literature review is performed in order to gather previously
studied determinants of tax revenue. The main conclusions of the dissertation, along with
proposals for future research and limitations to the study, are presented in section 6.

2. DEFINITIONS OF TAX BURDEN AND TAX EFFORT

The most general concept underlying this dissertation is that of taxation. Winer and
Hettich (2008) summarise the importance of taxation and of its study. Firstly, taxation is
fundamental to finance essential services and activities of the public sector, namely the
courts, the legal system, national defence, and police protection. Secondly, it allows for
the development of social programs, among which the authors highlight public health
services, education, and welfare. Finally, the authors mention the “distributional goals”
of a community, which can be achieved through taxation. Hence the two sides of the
government’s fiscal decisions: the revenue side, which mainly depends on taxation
(Macek, 2018; Nisha, 2018), and the expenditure side. The focus of this dissertation is
taxation.

Taxes have an economic origin and a political origin (Silva, 2015). The economic origin
is related to what has been established about satisfying the financial needs of a political

586
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

organisation. The political origin of taxes, on the other hand, regards the political
legitimacy of the coerciveness of a specific tax, which is based on political choices
concerning both the tax base and the associated tax rate (Silva, 2015).

As Gur (2014) asserts, tax decisions are the result of a bargaining process established
between the government and its citizens. On one side of this process, there is the need to
levy taxes to finance the government expenditure. On the other side of the bargaining
process are the citizens, who wish to participate in the process and have a say on how
their payments in taxes are used (Gur, 2014).

The collection of taxes contrasts with the distortion of the behaviour of taxpayers and the
decline of economic efficiency that result from increased taxes (Koch, et al., 2005; Xing
and Zhang, 2018). For this reason, it is important to investigate and to develop literature
taxation-wise, in order for authorities to make the most accurate and economically and
socially beneficial decisions with regards to fiscal policies, particularly tax policies.

In order to further develop this subject, one must firstly define what the author considers
to be the most basic concept of this dissertation: tax. What is a tax? The literature on any
tax-related subject seldom defines tax. The assumption is made that the reader is
familiarised with the concept. The author considers it fair to assume that most taxation
literature readers have a basic knowledge of the subject under analysis. It is important,
however, to explicitly provide the reader of this dissertation with a definition of tax – and
every other related concept –, as one of its main aims is to serve as academic literature
support for future work.

According to the OECD (2017) a tax is “a compulsory unrequited payment to the

government”. The “unrequited” denomination is due to the fact that the benefits the
government provides to the taxpayers are not necessarily proportional to the tax payments
made.

Some important observations must be made in order not to lead to the misinterpretation
of the tax concept as defined by the OECD (2017). In first place, taxes do not include any
fines, penalties, or compulsory loans paid to the government. Secondly, compulsory
social security contributions are considered to be part of tax revenues. Finally, there are
other payments made to the government whose nature can generate some doubt regarding

587
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

whether or not to be considered taxes, and therefore as generating of tax revenue. The
author believes, however, that the provided definition and notes suffice for the
development of the subject in the present dissertation.

A more polished definition of the concept of tax is provided by Xavier (1974), as cited
by Silva (2015) – “a patrimonial provision that is established by law in favour of an entity
responsible for the exercise of public functions, and whose main aim is to obtain the
means for the financing of such entity”1.

Two other important and basic taxation-related concepts ought to be briefly addressed:
tax systems and tax policies. The tax system established in a state is the reflection of the
communal values of that state and of those who hold the political power (Nisha, 2018).

According to Winer and Hettich (2008), the underlying structure of tax systems is
identical among different jurisdictions, despite the great deal of variations encountered.
Le et al. (2012) affirm that the existence of an effective tax system is a crucial factor for
the economic development of both developing and developed countries. Not only do tax
systems have a great influence on investment decisions, but they can also promote a lower
aid dependency in low-income countries if they result in higher tax revenues.
Furthermore, the effectiveness of tax systems encourages good governance, strengthens
state building, and promotes government accountability (Le et al., 2012).

Tax policy is a type of fiscal policy. Winer and Hettich (2008) provide us with a practical
definition of tax policy - the “manipulation of some aspect, or a combination of
characteristics” of what they refer to as the “tax skeleton”. It is the government’s strategy
for influencing the performance of the economy (Nisha, 2018). In summary, this tax
skeleton concerns the various bases taxes are levied on, the tax rates structures, which can
be simple or more complex, and other special arrangements which affect how those tax
bases are defined, and determine tax rates for specific components of the tax base.

Every tax policy is applied with the main objective of generating the revenue required to
fund public services and the state’s fundamental expenses in order to enhance the public
interest (Arif and Rawat, 2018; Kim and Lim, 2018). To meet such crucial aim, the
government and the tax authorities must ensure revenue adequacy and an adequate level

1
Free translation by the author. See Silva (2015: 24).

588
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of tax burden (Kim and Lim, 2018). Understanding countries’ tax effort levels is
paramount for tax policy makers to know which countries have the capability of
increasing their tax revenue and which ones do not (Fenochietto and Pessino, 2010).

One important and noteworthy duality in what regards taxation is that of vertical equality
and horizontal equality (Kim and Lim, 2017). On one hand, the principle of vertical
equality concerns how the tax system affects every taxpayer and their families, from the
bottom to the top of the income spectrum. In this sense, tax systems should be established
in such way that those who are more able to pay ought to contribute more in taxes than
taxpayers who have less ability to pay (Cronin et al., 2012). On the other hand, horizontal
equality measures whether taxpayers within similar conditions of income, family
structures, and age pay identical amounts of tax (Kim and Lim, 2017).

The literature review performed to write this dissertation led the author to conclude that
great deal of authors in tax literature mention both concepts of “tax burden” and “tax
effort” without providing the reader with precise and clear definitions. “Tax burden” and
“tax effort” are frequently assumed to be self-defining concepts. Vallés-Giménez and
Zárate-Marco (2017), for instance, claim that there is not a universally satisfactory
approach of tax effort in the literature. Furthermore, as noted by Fedyshyn (2013), there
are several interpretations of the term “tax burden” in the literature.

In order to investigate the evolution of both the tax burden and the tax effort of countries,
it is paramount one establishes accurate definitions of these concepts. The analysis of the
literature regarding taxation provides us with several important, however, disparate ones.
Furthermore, these two terms are frequently used interchangeably.

Semantic-wise, a burden is “a load, typically a heavy one”2, and in particular it can be “a

duty or misfortune that causes worry, hardship, or distress”3; effort can be defined as a
“strenuous physical or mental exertion”4. Therefore, in simple terms, both tax burden and
tax effort would represent the weight borne by taxpayers.

2
Oxford dictionary online
3
Oxford dictionary online
4
Oxford dictionary online

589
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The most commonly found representation of tax burden in the literature is that of the
percentage of tax revenues over a measure of income, normally GDP, e.g. Bird (1964),
Schneider (2005), Švaljek (2005), Vasiliauskaite and Stankevicius (2009), Lago-Peñas &
Lago-Peñas (2010), Dubauskas (2016). To Wang (2007: 279), the total tax revenue over
GDP is “a standardised measurement of tax burden”. According to Reed and Rogers
(2006: 404), tax burden is “the ratio of total tax revenues over personal income”. All of
these definitions, despite some differences, measure the tax burden as the simple ratio
between tax revenue and a measure of income. Nevertheless, the most widely used
variable in the literature is the tax-to-GDP ratio (Machová and Kotlán, 2013), which
Mahdavi (2008) refers to as the level of taxation.

The measurement of the tax burden of a country is an attempt to measure the importance
of the public sector of that country’s economy in contrast with its private sector (Frank,
1959). Furthermore, tax burden is frequently considered to be a crucial determinant of a
country’s shadow economy (Schneider, 2005; Buehn et al., 2018). Reed and Rogers
(2006), however, argue that tax burden can be an inaccurate measure of tax policy, which
the authors claim is acknowledged by the literature. Consequently, this lack of accuracy
of the variable is frequently omitted from research discussions. Reed and Rogers (2006)
justify the use of tax burden in studies with both the fact that it is easy to compute, due to
data availability, and the fact that the literature lacks more qualified alternatives. These
reasons are, however, unsatisfying: not only is there the possibility of “misinterpretation
of empirical results”, but also, and more importantly, of low quality tax policy advice. In
order to evaluate the preciseness of tax burden as a proxy for tax policy, Reed and Rogers
(2006) quantify the relationship between changes in tax burden and changes in tax policy.
They conclude that better measures should be constructed and applied.

Despite the fact that a great deal of authors agree with the aforementioned definition of
tax burden, to Bird et al. (2008) and Thornton (2014) the ratio between tax revenues and
GDP corresponds to the definition of “tax effort”. Rabiei and Balagetabi (2013) agree,
and go one step further by defining a tax effort index, which corresponds to the ratio
between the actual tax effort and the estimated tax effort. The difference between the
actual and the estimated tax effort is simple. Both are computed with GDP as the
denominator, but the actual tax effort is the ratio between the total tax revenues in one
year and GDP.

590
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Some authors define tax effort differently. According to Fenochietto and Pessino (2010)
and to Bingyang and Qingwang (2012), tax effort is “the ratio between actual revenue
and tax capacity”. This tax capacity is “the maximum level of tax revenue that a country
can collect given its economic, social, institutional, and demographic characteristics”
(Fenochietto and Pessino, 2010). In accordance with these authors, Le et al. (2012) assert
that the usage of tax effort without the weighting of these characteristics is not effective
for cross-country comparison “due to different economic structures, institutional
arrangements, and demographic trends” (Le et al., 2012: 2). To overcome this issue, some
authors favour the usage of the tax effort index, which is the index of the ratio between
the share of the actual tax collection in GDP and taxable capacity. Here, taxable capacity
is the “predicted tax-to-GDP ratio that can be estimated empirically, taking into account
 a country’s specific macroeconomic, demographic, and institutional features, which all
change through time”.

Fenochietto and Pessino (2010) provide us with another important concept, potential tax
collection, which “represents the maximum revenue that could be obtained through the
law tax system” (Fenochietto and Pessino, 2010: 66). Furthermore, the authors refer to
the difference between this potential tax collection and a country’s actual revenue - the
tax gap. This is “a function of tax capacity and the extent to which, by tax laws and
administration, a society wishes to mobilise resources for public use” (Fenochietto and
Pessino, 2010: 66).

There is some misconception around these concepts, which can lead to some confusion
in their study and empirical analysis. The author proposes the agreement of clear and
unanimous definitions of the concepts in future literature.

Finally, in spite of the focus of this dissertation being the level of taxation, it is
acknowledged that there are other important factors in tax policy-making, such as the
components of tax revenue (Mahdavi, 2008).

591
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3. TAXATION IN LITERATURE: A SYSTEMATIC LITERATURE REVIEW

3.1 The study of taxation

As stated by Winer and Hettich (2008), both a positive and a normative analysis are
required for a comprehensive approach to taxation. The reasoning here is that apart from
allowing for the analysis of the tax systems in existence, it leads one to debate possible
improvements of such systems.

According to Winer and Hettich (2008), the study of taxation is divided in two extensive
approaches, which shall be analysed to the extent that is relevant for the development of
this dissertation. The first approach is related to the works of Wicksell (1896), Lindahl
(1919), and, more recently, Buchanan (e.g. 1968, 1976). The second approach Winer and
Hettich (2008) refer to is, according to these authors, based on the works of Edgeworth
(1925), Ramsey (1927), and Pigou (1951), and, more recently, of Mirrlees (1971,
amongst others. These approaches are different both in their emphases and in the results
that they arrive at. Nevertheless, both deal with the same essential problem: the separation
of taxing and spending.

The goods and services provided by governments, i.e. public goods, are different from
those of private markets. Public goods are consumed equally by all members of a
collectivity, and they cannot be rationed according to price, which is the case of private
goods. Preference revelation and free-riding problems can therefore arise, as those who
do not pay taxes voluntarily are not withheld from consuming those public goods or
services. In order to overtake these two significant economic drawbacks, taxation is
coercive in most collectivities. As the authors expose it, the coerciveness of taxation is
done through the creation in such collectivities of “tax systems in which there is only a
diffuse and distant link between additional consumption of publicly provided goods and
increases in tax liability” (Winer and Hettich, 2008: 393).

The separation of taxing and spending bears some problems. In order to reduce their tax
liability, taxpayers will adjust their activities. This adjustment behaviour leads to a loss
of economic welfare in comparison to a situation where such trade-off would not be
elicited. In the literature, such loss (the excess burden or deadweight cost of taxation) is
a measure of the inefficiency generated by a specific tax (Winer and Hettich, 2008).

592
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The second problem with the separation of taxing and spending is concerned with
redistribution. The allocation of public goods and their costs among their users in a
collectivity, and the determination of the level of production of those public goods cannot
be done through markets. Therefore, there is the need to employ other mechanisms in
order to make decisions. On one hand, all the available collective decision processes
establish their own incentives for redistribution between the members of a collectivity.
On the other hand, the separation of taxing and spending makes understanding the
distributional implications of the numerous ways of providing and financing public
programs more complicated. This can result in the management of public resources for
personal purposes by those who are in a position to do so (Winer and Hettich, 2008).

It is important to note that, despite considering the same essential issue – the separation
of taxing and spending –, the two approaches differ in the way they deal with their
implications. Essentially, the first approach regards the need for collective decision
processes, for fiscal structures that would allow for the reduction of the gap between
taxing and spending. This approach can also focus on the limitation of coercion, by
considering the institutional and fiscal constraints that would allow for that (Winer and
Hettich, 2008).

The second approach considers that the decision processes are taken as exogenous. The
assumption is made that there is a social planner whose role is to make decisions on behalf
of the collectivity based on a exogenously-defined welfare function. The emphasis is on
the welfare of the collectivity, and tax systems are designed to maximise it, under an
assumed analytical framework (Winer and Hettich, 2008).

Given that the tax systems in existence are built upon the decisions made through
collective choice processes, the comprehensive approach to taxation which introduced
this part can only be based on the first of these two approaches. The reasoning here is that
in order to analyse tax systems, one must firstly model those collective choice processes,
and compare the results predicted by the models to what is actually observed in the
context of those systems. The second approach, which assumes the “social planner”, only
allows for a normative analysis (Winer and Hettich, 2008).

593
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.2 Systematic literature review

A systematic literature review was performed to analyse the incidence of our main
taxation-related variables in past studies. Firstly, the relevant paperwork was gathered.
The academic paper database used to conduct this search was b-on5.

1. Looking up for publications with at least one of these keywords: tax revenue, tax
burden, tax effort. This retrieved a total of 3,286 results.
2. These were narrowed down to the 432 which were part of the Scopus® database.
3. Of these, 378 were published in academic magazines.
4. Articles published in 2018 were excluded, due to the fact that the year is not over.
This resulted in a total of 346 being included in this systematic literature review.

The variables analysed are: year of publication, number of authors per article, country of
affiliation of the authors, keyword (or keywords) of the article – tax burden, tax effort,
and tax revenue –, and methodology used in the article. This review only considers
paperwork published between 1972 and 2017.

Some notes ought to be made regarding these results:

 Keywords are not necessarily just “tax burden”, for instance. Keywords like “tax
burden level” or “corporate tax burden” belong to the set of keywords.
 Some keywords can be in plural form, e.g. “tax revenues”.
 Some authors have more than one country of affiliation.

3.2.1 Year

The first aspect of the investigation in taxation that will be addressed is the year of
publication of the articles. Has the investigation in taxation increased or decreased over
time? Has it been fairly constant frequency-wise?

In graph 1, depicted below, the number of articles published in each year is presented:

5
As of September 16, 2018

594
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0
1972
1982
1989
1990
1992
1996
1997
1998
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
Graph 1: Number of articles published per year

The first important note is that the four last years alone account for exactly half of the
total number of articles, with 173 articles published since 2014. There was a peak in the
number of articles published from 2009 to 2010.

Another important remark the author would like to address is the fact that it cannot be
concluded from these results that the investigation in taxation has increased over the
years. In order for this literature review to be carried out, limits to the search had to be
set. It is clear from the detailed search description above what these limits are. For
instance, a great deal of keywords other than the ones mentioned could have been used to
filter the results of the search. The keywords used in this review are the ones considered
to be significant for this dissertation. The conclusions are of value to the extent that they
only regard the 346 articles under analysis.

3.2.2 Country of affiliation

The second variable in this analysis of the investigation on taxation is the country of
affiliation of the authors. The results are represented in graph 2 below:

595
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

14% Czech Repubic

China
33% Russia
10%
Spain
Germany
9% Ukraine
Brazil
2% 7%
UK
3% 7% Australia
3% 6%
3% Switzerland
3%

Graph 2: Authors of each country of affiliation (% of total)

It is evident that the most represented countries of affiliation are the US (14%, with 94
authors), the Czech Republic (10%, with 67 authors), and China (10%, with 62 authors).
The author reiterates the fact that some authors have more than one country of affiliation.
Therefore, the percentages presented are in a total of 691. The full list of countries and
number of authors affiliated is found in the annexes.

3.2.3 Authorship

The third variable to be analysed is the number of authors of each article. This
information is presented in both table 1 and graph 3 below:

Number of authors Number of articles

1 126
2 129
3 70
4 15
5 4
6 2
Table 1: Number of articles per number of authors

596
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1% 1%

4%
1
20% 37% 2
3
4
5
6
37%

Graph 3: Number of articles per number of authors

Most articles were written by 1 or 2 authors (74%). One fifth of the 346 articles were
written by 3 authors, and only the remaining 6% by 4 or more. Nevertheless, one
important idea is that most articles (63%) were written by more than one author.
Therefore, most authors prefer to work with peers, insofar as our sample is limited.

3.2.4 Keywords

The fourth feature of this study is the frequency of use of each of the keywords tax burden,
tax effort and tax revenue. Both table 2 and graph 4 present the results of the search:

Keyword Number of articles

Tax burden 177
Tax effort 15
Tax revenue 161
Total 353
Table 2: Number of articles with each keyword

597
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Tax burden
46%
Tax effort
50%
Tax revenue

Graph 4: Percentage of use of each keyword

By analysing the data in table 2, the sum of the number articles (177+15+161) is not 346,
but 353. This is explained by the fact that some articles have more than one of this three
keywords.

“Tax burden” is the most used keyword out of the three, representing half of the keywords
used. “Tax revenue” is also very common, and the difference between the frequency of
use of “tax burden” and “tax revenue” is very small. “Tax effort”, however, only exists
in 15 of the 346 articles. This goes in hand with what has been concluded with regards to
the use of the concepts of “tax burden” and “tax effort” in the literature: either they are
used interchangeably, and the preferred denomination is that of “tax burden”; or “tax
effort” is interpreted with the other referred meanings or as a sacrifice (this meaning will
be explained further in section 4).

Regarding the keywords used, the frequency per year was also analysed. The results are
presented in graph 5 below:

598
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0
1972
1982
1989
1990
1992
1996
1997
1998
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010

2012
2013
2014
2015
2016
2017
2011
Tax burden Tax effort Tax revenue
Graph 5: Evolution of the usage of each keyword (1972-2017)

In general, the number of times “tax burden” and “tax revenue” were used increased
throughout the period between 1972 and 2017. Interestingly, in some periods one
keyword was preferred over the other. For instance, between 2005 and 2007, “tax burden”
was more frequently used than “tax revenue”. In 2008, however, “tax revenue” appeared
in more articles than “tax burden” did. Then, “tax burden” surpassed “tax revenue” in
2009 and in 2010. Nevertheless, and although “tax burden” was used more frequently
than “tax revenue” overall, since 2015 “tax revenue” has been more frequent than “tax
burden”.

3.2.5 Methodology of the article

The last variable analysed is the methodology used in each article. Although an article
can be composed of a literature review and the computation of certain metrics based on a
database, for instance, the methodology that is considered is the ultimate aim of the
article, which in this example is the analysis of the results obtained from the calculations
made.

599
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The methodology can be database (DB), econometric (ECO), exploratory (EXP), inquiry
(INQ), or literature review (LRE). Table 3 and graph 6 summarise this information in
what concerns the 346 articles:

Methodology Number of articles

DBA 290
EXP 24
ECO 22
LRE 9
INQ 1
Table 3: Number of articles per methodology

3% 0%

6%
7%
DBA
EXP
ECO
LRE
INQ
84%

Graph 6: Percentage of articles per methodology

Articles which make use of a database to draw conclusions from are by far the most
frequent, accounting for 84% of the total of articles. Some articles are of econometric
type (22) and some are exploratory (24). Articles which are developed as mere literature
reviews are very rare (9), and only 1 article applied an inquiry.

Table 4 depicts the number of times a methodology was used in an article with each
keyword, which allows to understand which methodology is more commonly used for
each keyword:

600
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Methodology Tax burden Tax effort Tax revenue

DBA 145 13 139
EXP 17 1 6
ECO 9 0 13
LRE 6 1 2
INQ 0 0 1
Table 4: Methodology used by keyword used

Not surprisingly, database articles are the most commonly used for every keyword.
Exploratory articles use “tax burden” more than any other keyword. Econometric articles
are mostly used with “tax revenue”, and none with “tax effort”. The low number of
articles of literature review or inquiry type do not allow for a significant conclusion.
Nevertheless, most literature review articles used “tax burden” and the only inquiry article
used “tax revenue”.

4. METRICS TO DETERMINE TAX BURDEN AND TAX EFFORT

The weight of the public sector versus the private sector in an economy can be evaluated
by measuring the tax burden of said economy (Frank, 1959). But how is the tax burden
calculated? From what has been stated, one can conclude that the most frequently used
measure to calculate the tax burden (or the tax effort, depending on the definition attached
to each concept) is the ratio between the total tax collection and a measure of income
(GDP, personal income, among others) (Reed and Rogers, 2006; Wang, 2007;
Vasiliauskaite and Stankevicius, 2009, Andrejovská and Hudáková, 2016). However,
some authors have questioned the efficiency of such measure for appraising the tax
system and its weight in the economy of a country. Two important and highly cited
authors in the tax effort literature are Henry Frank and Richard Bird, who proposed two
indices that would allow for a more adequate comparison of tax burden/tax effort levels
between countries (or states, if the level of government allows this comparison be made).

One question can consequently be posed: why would the traditionally proposed measures
of tax burden (or tax effort) not be satisfactory for the comparison of countries’ tax
burdens? If one computes the tax-to-GDP ratio as a measure of tax burden for two
countries, and compares it between them, the conclusion will be that for one of the
countries the tax burden is higher than for the other (or even identical between them).
Assuming that the definition of a country’s tax burden is given by the mentioned ratio,

601
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

then its values should be comparable, and no concern should therefore exist. In fact, they
are comparable, as they are computed through the same method. But what does such
information reveal in practical terms? What are the taxation policies and measures that
can be employed or altered by a government taking this information into account? Is it
correct to assume that the country with the larger tax-to-GDP ratio is the country in which
taxpayers take a more significant burden for paying taxes?

Tax burden measured as the tax-to-GDP ratio measures how much of GDP has been
redistributed through public budgets. It is a macroeconomic indicator. It therefore is not
indicative of the tax burden borne by entities at individual level, and should consequently
only be interpreted as an approximate measure, taking into account its degree of
simplification (Kotlan and Machová, 2012). Furthermore, a higher ratio can simply mean
that the efficiency of a government in collecting taxes has been enhanced, and not
necessarily that the tax burden, in its semantic sense of weight or load on taxpayers, has
increased (Kotlan and Machová, 2012). One can conclude that the tax-to-GDP ratio bears
a great deal of shortcomings.

Nevertheless, even where tax burden is not explicitly referred to, the ratio between public
revenue and GDP is the variable that allows for the comparison between countries or
states. It is the case in the study of Sobarzo (2004) and Bonatti (2007), for instance.
Therefore, and considering everything that has been stated throughout about the approach
to tax burden in previous studies, the tax-to-GDP ratio plays a pivotal role in taxation
literature, and it would be careless to disregard it from further analysis.

Lastly, the literature does not regard tax burden exclusively as the mentioned ratio at
country or state level. Frequently, the concept of tax burden is applied at entreprise-level,
such as in Budrytė (2005) and Strelnik et al. (2015).

In addition, tax burden can be evaluated by type of economic activity, i.e., by whether
taxation is imposed on labour, capital or consumption (Kotlan and Machová, 2012).

In this section, Frank’s and Bird’s indices will be explained, as they explore these
questions and provide the literature with new outlooks on the approach, measurement,
and assessment of both tax burden and tax effort.

602
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Firstly, a theoretical approach will be made to the tax effort indices. Secondly, the indices
will be calculated for 35 of the 36 OECD countries6. The results will be presented
separately and then compared between the indices. In addition, we will compute the tax-
to-GDP ratio for the same countries, analyse it briefly, and compare it with the tax effort
indices of Frank (1959) and Bird (1964).

4.1 Frank’s index

“Are taxes in this state higher than in that one?” (Frank, 1959: 179). This is the question
Henry Frank presents in his 1959 article’s opening, and one to which the author presents
two frequently used measures that aim at answering it: taxes per capita and taxes as a
percentage of income. These measures are efficient in the comparison between countries’
tax collections only if there is a reasonable similarity between such countries population
or income-wise, respectively (Frank, 1959). The author proposes another measure to
compare tax burdens between countries, which we shall hereafter refer to as “Frank’s
index”, an index of tax sacrifice that synthesises the two aforementioned measures (Frank,
1959). In his 1959 article, Frank computes this index for the states of the US.

With regards to per capita taxes as a measure of tax burden, which are computed “by
dividing the total tax collection by the resident population”, Frank (1959) recognises that
it is more appropriate than the use of total tax revenue because it is weighed by the size
of the population. However, each resident has the same weight (one) in the computation
of the population, regardless of their “age, condition or position in the economy”. Each
person is considered to have the same ability to pay those taxes (Frank, 1959).

On the other hand, taxes as a per cent of income do not demonstrate how much effort was
required to produce a certain amount of income. According to the author, if in a society
it takes ten persons to produce as much as it takes fifteen persons in another, then paying
any rate of that production is a higher burden for the members of that latter, poorer society
(Frank, 1959). This concept of “equality of sacrifice” is the basis of the progressivity of
income taxes (Frank, 1959).

6
Lithuania has been excluded from our study due to lack of available data.

603
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Furthermore, “taxes do not come out of income and disappear from the economy to
remain evermore in an Independent Treasury” (Frank, 1959: 181). Taxes are paid to the
government by taxpayers and return to the income stream in the form of salaries and
purchases of goods and services. Therefore, the denominator in the taxes-to-income ratio
is created by the taxes that represent the numerator. What is unknown here is how much
of those taxes returns to that income (Frank, 1959).

In order to understand the sacrifice concerned with the payment of taxes, it is

advantageous to join the two measures together, by dividing the taxes as a per cent of
personal income by per capita personal income. An important assumption is made here:
that in two countries with similar taxes over income, the sacrifice made to pay the taxes
will be lower for a resident of the country with the higher per capita personal income.
Frank’s index (F) can be computed as follows7:

Taxes Personal Income

F= ÷ (1)
Personal Income Population

According to Frank (1959), this index of tax sacrifice is more effective for international
comparison than taxes per capita or taxes as a percentage of income because the first does
not take into account how the income varies between the countries, and the second does
not consider the effort made by different members of an economy to pay the taxes.

4.2 Bird’s index

Like Frank in 1959, Bird (1964) questions the significance of the ratio between the taxes
and national income (T/Y, where T represents the tax revenue and Y the national income
measure), as he considers that “it indicates nothing about the effort required to produce
the income used as the denominator of the ratio” (Bird, 1964: 303).

7
The indices formulas are as presented by the authors in their original articles (Frank,
1959; Bird, 1964). Some alterations were made in our calculations for mere simplification
of presentation and comparison of results. Such modifications do not affect the
conclusions.

604
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Bird (1964)’s goal was to compute what he called the “tax sacrifice”, a “modified
version” of Frank (1959)’s index, in an attempt to measure the state tax burden and allow
it to be comparable between countries. The major difference between Bird’s (hereafter
Bird’s index) and Frank’s indices is the use of disposable income rather than just the
measure of income used by Frank. The reasoning here is the avoidance of “certain absurd
results which could otherwise arise in extreme cases” (Bird, 1964: 306).

Despite recognising the flaws of his measure, Bird asserts that its results are more
meaningful than the traditionally-adopted ratio T/Y. Furthermore, he claims it is a “more
useful calculation” to measure sacrifice. Bird’s index (B) can be computed as follows:

Taxes Gross Domestic Product

B = [( × 100) ÷ ] × 100 (2)
Disposable income Population

Bird (1964) also addresses the concepts of tax capacity and tax effort. He considers that
the proposed index is not a measure of tax capacity or tax effort, but of tax sacrifice (or
tax burden), i.e., “the relative importance to the citizens of the resources given up to
government in countries at different levels of income” (Bird, 1964: 303). Tax capacity is
the capacity to provide revenue for any public expenditure, and it depends mainly on per
capita income. The author adds that the willingness and the will for public services is
crucial to calculate the “feasible” capacity. Tax effort is a measurement of the extent to
which a government actually takes advantage of said capacity. He asserts that both
measures are useful for making policy appraisals, but difficult to be quantified in a
meaningful way. He acknowledges that the latter is one of the flaws of the tax sacrifice
index he proposes, but that it is a favourable addition to the other two.

For the purpose of simplification, the formula in (1) can be written as:

T Y
B = [( × 100) ÷ ] × 100 (3)
Y-T P

In this formula, T are taxes, Y is income, Y-T is the disposable income, and P is the
population.

Bird (1964) addresses two types of weaknesses concerning his index: conceptual and
data. As aforementioned, T stands for “taxes”. In reality, however, the author computes

605
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the index using the current government revenues, which includes both tax and non-tax
revenue. The inclusion of non-tax revenue in the variable is not significant in the ranking
of countries8 (Bird, 1964).

Furthermore, Bird (1964) alludes to the fact that, although the data used should be for all
levels of government, for subordinate levels of government the figures are only
approximate in most countries.

Bird (1964) adds that the figures for both income (Y) and population (P) are subject to
biases and errors. Nevertheless, according to the author, one can assume that the level of
accuracy of such figures is positively correlated with per capita income (Y/P), which
itself incorporates all the biases, therefore approximating the biased figures to the real
ones.

Finally, Bird (1964) addresses the measure of income used. For the first ratio in the index
formula, the author uses Gross National Product (GNP), while for the second, Gross
Domestic Product (GDP) is used.

According to the OECD, GDP is “the expenditure on final goods and services minus
imports: final consumption expenditures, gross capital formation, and exports less
imports”. The difference between GDP and GNP is “the net of incomes paid to and
received from foreigners” (Bird, 1964: 307). GDP includes, according to the OECD, “the
production by the resident institutional units of the country” (hence the “domestic”
denomination). Taking this into consideration, Bird’s index formula can be rewritten as
follows:

T GDP
B = [( × 100) ÷ ] × 100 (4)
GNP-T P

For the sake of presenting his results, Bird (1964) groups the countries into three groups:
Latin America (Costa Rica, El Salvador, Guatemala, Honduras, Mexico, Panama,
Argentina, Brazil, Colombia, Ecuador and Peru), North America (Canada and USA), and
Europe (Austria, Belgium, Denmark, Finland, France, West Germany, Greece, Ireland,

8
Bird (1964: 305) does, however, recognize the “problems of comparability” that derive from this
simplification, given that in some countries the “extrabudgetary revenues” like social security funds
and exchange profits are highly relevant (e.g. Brazil).

606
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Italy, Netherlands, Norway, Spain, Sweden, Switzerland and the UK). Latin America and
Latin America constitute one region, and Europe another.

Bird’s (1964) first important conclusion from his calculations is that, in general, poorer
countries in both regions, i.e., the countries with the lower per capita GBP, “try” harder,
or, in other words, have a greater “tax sacrifice”. The conclusion is the same when
considering each region separately. In addition, the results obtained by comparing
countries using the traditional tax-to-GDP ratio and the ones obtained with Bird’s tax
sacrifice index are very different from one another. The author concludes by stating that
despite the flaws his index carries, it outdoes the simple taxes-to-national-income ratio
in that it “gives us some idea of how much a little more funneled through the government
would “hurt” in some relative sense” (Bird, 1964: 308).

4.3 Computing the indices for OECD countries

4.3.1 Formulas and variables

In this section, some modifications to the indices are presented and explained. In addition,
the variables used for the determination of the indices are listed.

[Link] Frank’s index

The formula for Frank’s index (F) is:

Taxes Personal Income

F= ÷ (5)
Personal Income Population

In order to get more easily comparable and interpretable results, the formula shall be
altered by simply multiplying it by 100 twice, similar to what Bird did with his modified
version of the index:

Taxes Personal Income

F = [( × 100) ÷ ] × 100 (6)
Personal Income Population

Due to lack of available data, in particular personal income data, our calculation of
Frank’s index uses the following formula:

607
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

T GNI
F = [( × 100) ÷ ] × 100 (7)
GNI P

Where T is the total revenue in taxes, GNI is Gross National Income, and P is the
population of the country. Because GNI/P represents per capita GNI, the final formula is:

T
F = [( × 100) ÷ GNI pc] × 100 (8)
GNI

Taxes are total tax revenues in USD9, GNI is measured in USD at current prices10, and
per capita GNI is measure in USD at current prices11.

[Link] Bird’s index

The formula for Bird’s index as presented before does not require further modifications:

T GDP
B = [( × 100) ÷ ] × 100 (9)
GNP-T P

The only remark is the fact that in the 1993 SNA12, Gross National Product (GNP) was
renamed Gross National Income (GNI), and therefore the data used will be concerning
the latter. GDP/P is GDP per capita, which shall be written as GDP pc. We will therefore
apply the following modified formula:

T
B = [( × 100) ÷ GDP pc] × 100 (10)
GNI-T

Per capita GDP is measured in USD at current prices13. The remaining variables are
measured as established for Frank’s index.

9
OECD (2018), Tax revenue (indicator).
10
From the World Bank database – World Development Indicators.
11
From the World Bank database – World Development Indicators.
12
The 1993 SNA is a set of standard rules for the measurement macroeconomic indicators.
13
From the World Bank database – World Development Indicators.

608
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4.3.2 Results and analysis

Both Frank’s index (F) and Bird’s index (B) were computed for 35 of the 36 countries
that are part of the OECD as of today. The tax effort indices were calculated for each
country in each year from 2000 until 2015. Lithuania was excluded from this analysis due
to the lack of data for tax revenue, which is needed to compute both indices.

Firstly, we shall analyse each index separately. This will be done through the
interpretation of graphs that compile the most important information about each index.
Secondly, comparisons between the two indices will be made, again interpreting some
illustrative graphs that allow for that.

By virtue of the importance of the simple tax-to-GDP ratio in the literature, an analysis
of this indicator is also included. It is compared with Frank’s and Bird’s tax effort indices.

Because a comprehensive, case-by-case discussion and analysis of the indices is outside

of the scope of this dissertation, only some particular and distinct cases will be referred
to. The author proposes such a more in-depth study for future work.

[Link] Frank’s index

The computed index values14 for Frank’s index are presented in table 9 of the annexes.
The evolution of Frank’s index for each OECD country from 2000 to 2015 is represented
in graph 7:

14
The source for most data is The World Bank, due to the specifications of some variables
as provided by the OECD, which were not considered accurate for our calculations.

609
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1.0000

0.9000

0.8000

0.7000

0.6000

0.5000

0.4000

0.3000

0.2000

0.1000

0.0000

AUS AUT BEL CAN CHE CHL CZE

DEU DNK ESP EST FIN FRA GBR
GRC HUN IRL ISL ISR ITA JPN
KOR LUX LVA MEX NLD NOR NZL
POL PRT SVK SVN SWE TUR USA

Graph 7: Evolution of Frank’s index (2000-2015) 15

The evolution of Frank’s index throughout this period is not regular. It shows peaks of
ups and downs throughout the years.

From a general standpoint, Frank’s index has decreased over time. This idea is clearer
from graph 9, which represents the evolution of the average of the index throughout the
years, and which we shall analyse further into this sub-section.

15
A country name abbreviation glossary is available in the annexes of this dissertation.

610
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Some occurrences in Frank’s index, which are illustrated by graph 7, are noteworthy.
Turkey, Hungary, Estonia, Poland, Latvia, the Slovak Republic, the Czech Republic, and
Chile were above every country with a significant difference from 2000 until 2008. From
then on, and despite the ups and downs of the index, the evolution of Frank’s index for
each country did not suffer from a great increase or decrease.

For the sake of making a more just comparison between the countries’ averages, the
average for each country’s index was calculated from 2000 to 2015, which were the years
for which information was available for all countries:

611
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

HUN
POL
TUR
EST
LVA
SVK
CZE
CHL
SVN
PRT
GRC
MEX
NZL
ESP
ISR
ITA
KOR
FRA
BEL
FIN
AUT
SWE
DEU
DNK
ISL
CAN
AUS
IRL
NLD
GBR
LUX
JPN
NOR
USA
CHE
0.0000 0.1000 0.2000 0.3000 0.4000 0.5000

Graph 8: Frank’s index: average per country (2000-2015)

In terms of averages, Chile (0.0419), the US (0.0544), and Norway (0.0619) have the
lowest indices from 2000 until 2015. Hungary is the country with the highest average
during this period, with an average index of 0.4226. Poland scores 0.3969, which makes
it the country with the second highest average, and Turkey is 3rd on the rank (0.3585).

To sum up, every country’s index decreased over time, although there is a clear growth
of the index in every country between 2014 and 2015. There is a tendency for
convergence of the countries’ tax effort indices. Graph 9, in which the average index and

612
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

its standard deviation for each year is represented, is a modest attempt to demonstrate this
overall decrease of Frank’s index from 2000 until 2015:

0.3000
0.2500
0.2000
0.1500
0.1000
0.0500
0.0000

Average SD

Graph 9: Frank’s index: average and standard deviation per year (2000-2015)
The average index does not, evidently, inform us of how each country’s index behaved
through time. Although the average is not a perfect measure, it does provide us with an
overall picture of the indices’ trends.

Graph 9 depicts the overall decline of the average of Frank’s index, although, again, there
is an increase between 2014 and 2015. From this we can conclude that overall, as
calculated by Frank’s formula, the tax effort in the OECD has been declining.

One final conclusion is that the standard deviation of the index for each year also
decreased, despite the slight increase between 2014 and 2015. Therefore, there was fiscal
convergence amongst the OECD countries.

[Link] Bird’s index

The computed index values are presented in table 11 of the annexes. There is an overall
tendency for the reduction of the tax effort as measured by Bird’ index:

613
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1.6000

1.4000

1.2000

1.0000

0.8000

0.6000

0.4000

0.2000

0.0000

AUS AUT BEL CAN CHE CHL CZE

DEU DNK ESP EST FIN FRA GBR
GRC HUN IRL ISL ISR ITA JPN
KOR LUX LVA MEX NLD NOR NZL
POL PRT SVK SVN SWE TUR USA

Graph 10: Evolution of Bird’s index (2000-2015)

Similar to what was done with Frank’s index, the average of Bird’s index was computed
for each country:

614
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

HUN
POL
EST
TUR
LVA
CZE
SVK
CHL
SVN
PRT
GRC
ITA
FRA
BEL
NZL
ESP
ISR
FIN
AUT
SWE
MEX
DNK
KOR
DEU
ISL
CAN
NLD
AUS
GBR
IRL
LUX
NOR
JPN
USA
CHE
0.0000 0.1000 0.2000 0.3000 0.4000 0.5000 0.6000 0.7000 0.8000
Graph 11: Bird’s index: average per country (2000-2015)

Similar to what was concluded for Frank’s index, for the period from 2000 until 2015,
Hungary is the country with the highest average of Bird’s index (0.6667), and Poland is
2nd again, with an average index of 0.5842. Interestingly, it is now Estonia who is placed
3rd on the rank (0.4928).

Chile and the US are the countries with the lowest average index (0.0584 and 0.0738
respectively), alike Frank’s index average for the same period. Japan is the country with
the third lowest average (0.09384).

615
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The average for Bird’s index throughout the years was also computed, and it is
represented in graph 12:

0.4500
0.4000
0.3500
0.3000
0.2500
0.2000
0.1500
0.1000
0.0500
0.0000

Average SD

Graph 12: Bird’s index: Average and standard deviation per year

The author reiterates the limitations associated with the average. The aim is simply to
provide the literature with a general idea of how Bird’ index behaved throughout the
period from 2000 until 2015. A thorough analysis of each country’s tax index evolution
is not the objective of this dissertation, but is certainly proposed for future work.

Graph 12 depicts the overall decline of the average of Bird’s index. Alike what was
asserted with regards to Frank’s index, Bird’s index increased between 2014 and 2015.
As calculated by Bird’s formula, the tax effort in the OECD has been declining from an
overall perspective.

From graph 12, it is also clear that there was fiscal convergence amongst the OECD
countries, as overall the standard deviation of the index decreased throughout the years.

[Link] Comparing the two indices

This sub-section is dedicated to a short comparison between the two indices and the
results obtained. This is more easily done by comparing the averages of the two indices
in each year, which are represented in graph 13:

616
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0.4500
0.4000
0.3500
0.3000
0.2500
0.2000
0.1500
0.1000
0.0500
0.0000

B F

Graph 13: Averages of index F and index B from 2000 to 2015

The first obvious conclusion is that Bird’s index is always higher than Frank’s index. This
is because of the way each formula is established: while Frank’s index is calculated using
income, Bird’s index gives emphasis to the disposable income, which is smaller than the
income measure. Therefore, the first part of Bird’s formula is always higher than Frank’s.

A second relevant point is that the movement of the averages is almost parallel. This is
not surprising, given that the way that the formulas are built is very similar. Bird’s index
formula is a modification of Frank’s.

4.3 The tax-to-GDP ratio

The third measure to be analysed is the traditional tax as a percentage of income measure.
Particularly, the tax-to-GDP ratio, due to how frequently used it is in literature. The
results for each country in each year are presented in graph 14:

617
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0.6000

0.5000

0.4000

0.3000

0.2000

0.1000

0.0000

AUS AUT BEL CAN CHE CHL

CZE DEU DNK ESP EST FIN
FRA GBR GRC HUN IRL ISL
ISR ITA JPN KOR LUX LVA
MEX NLD NOR NZL POL PRT
SVK SVN SWE TUR USA

Graph 14: Evolution of the tax-to-GDP ratio (2000-2015)

Clearly, the evolution of the tax-to-GDP ratio is different from Frank’s and Bird’s
indices’: overall, the tax-to-GDP ratio is growing. This is corroborated by graph 16, which
will be analysed after a short analysis of the average ratios for each country, presented in
graph 15:

618
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

DNK
SWE
BEL
FRA
FIN
ITA
AUT
NOR
HUN
LUX
NLD
ISL
SVN
DEU
GRC
ESP
GBR
CZE
CAN
ISR
PRT
POL
NZL
EST
SVK
AUS
JPN
IRL
CHE
LVA
USA
TUR
KOR
CHL
MEX
0.0000 0.1000 0.2000 0.3000 0.4000 0.5000

Graph 15: Average tax-to-GDP ratio per country (2000-2015)

Denmark (0.4444), Sweden (0.4303), and Belgium (0.4191) score the highest tax-to-GDP
ratios. In contrast, Mexico (0.1280), Chile (0.1880), and South Korea (0.2230) have the
lowest ratios.

Finally, the overall evolution of the tax-to-GDP ratio is of increase. There is a clear peak
of the average in 2008, from 0.2885 in 2006 to 0.3725:

619
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

0.4500
0.4000
0.3500
0.3000
0.2500
0.2000
0.1500
0.1000
0.0500
0.0000

Average SD

Graph 16: Tax-to-GDP ratio: average and standard deviation per year

The ratio decreased from 2008 until 2010 (0.2961), and has been increasing since,
although not constantly, as there were periods of peaks and lows throughout.

Unlike what has been observed for Frank’s and Bird’s indices, fiscal convergence as
measured by the standard deviation of the tax-to-GDP ratio cannot be concluded.
Although there was convergence between 2014 and 2015, the standard deviation proves
to be somewhat constant, without a clear growth or decrease movement.

4.4 The World Tax Index (WTI)

The last part of section 4 is to briefly address an alternative metric for the measurement
of tax burden - the WTI - which would be more efficient than the traditional tax-to-GDP
ratio, and which was proposed by Kotlán and Machová (2012). The authors call “tax
quota” to this ratio in order to distinguish it from the concept of tax burden as a more
general term, which can be evaluated employing several measures.

In summary, the WTI incorporates several sub-indices which regard distinct groups of
taxes, according to the OECD classification: CIT (Corporate Income Tax), PIT (Personal
Income Tax), VAT (Value Added Tax), PRO (Individual Property Taxes), and OTC
(Other Taxes on Consumption). These 5 sub-indices are then divided into sub-
components (Kotlán and Machová, 2012). Apart from taxes which are particular to certain

620
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

economies, whose inclusion would impede inter-country comparison, the WTI gathers all
OECD country taxes.

One feature which positively distinguishes the WTI from the simple tax-to-GDP ratio is
that its scope is expanded to other aspects which can affect the tax burden, namely tax
progression, administrative difficulties of the taxpayer’s tax collection, tax credits, and
incentives (Kotlán and Machová, 2012; Machová and Kotlán, 2013).

Furthermore, the WTI combines hard data with soft data, which is derived from the
Qualified Expert Opinion (QEO), obtained through OECD countries’ tax specialists’
answers to a comprehensive questionnaire survey. These tax specialists scored (as a
percentage) the sub-components within each sub-index according to how each
contributed to the tax burden in their country. The average percentage for each sub-
component in each country is assigned to that sub-component in that country. These
specialists also scored the sub-indices, although now through the pair-wise comparisons
method (Kotlán and Machová, 2012). Machová and Kotlán (2013) performed some
modifications to the computation of the WTI. The authors redefined the structures of three
of the five sub-indices. The list of sub-indices and sub-components according to Machová
and Kotlán (2013) is presented in Figure 1 of the annexes.

The formula to compute the WTI is

5
QEOk (11)
WTI = ∑ (Sk )
100
k=1

where Sk represents the value of the k-th sub-index, and QEOk is the weight determined
by the result of the QEO for the k-th sub-index. A higher WTI means a higher tax burden.

Kotlán and Machová (2012) compare the WTI with the “tax quota”, i.e., the tax-to-GDP
ratio for OECD countries. On one hand, they conclude that the ratio can be a satisfying
indicator of tax burden for personal income tax (which includes social security
contributions), and VAT. On the other hand, however, its quality can be questionable in
the case of corporate taxes.

621
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5. DETERMINANTS OF TAX BURDEN AND TAX EFFORT

“What affects revenues (measured as the ratio tax revenues to GDP) has been the subject
of a long debate.” (Gupta, 2007). In this section, a literature review is performed in order
to gather the previously used determinants and the rational underlying their application
on previous studies.

Morrissey et al. (2016) mention that the typical approach in the literature is “to model the
revenue to GDP ratio as determined by variables chosen to proxy for the tax base and
structure of the economy”. Kiser and Karceski (2017) focused their study on the
“structural and institutional determinants of tax revenue”. Amongst such indicators, the
most frequent for the study of tax revenue performance in developing countries are
“agriculture and industry value added as a percentage of GDP”, “openness to international
trade”, and “GDP per capita” (Morrissey et al., 2016: 1690). To these, the authors add
others mentioned in the literature, namely “aid, demographic features such as
urbanisation, or indicators of governance and institutions” (Morrissey et al., 2016: 1690).

As Streimikiene et al. (2018) point out, previous literature on the determinants of tax
revenue distinguish between the effects in developed and developing countries.

Gupta (2007) establishes an important and clear distinction that groups the determinants
of revenue. The author distinguishes between structural, institutional and policy variables.
This grouping is used in the list below with a merely structural purpose. An important
point here is that some of the variables included in these groups were not considered by
the author himself, and are therefore inserted by the author of this dissertation where
believed to be appropriate.

Other authors, such as Mahdavi (2008), have empirically tested the determinants of tax
revenue according to revenue composition.

622
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.1 Structural variables

5.1.1 Level of development

Per capita GDP is a proxy for the overall economic development of a country (Gupta,
2007; Bird et al., 2008; Morissey et al., 2016). The level of development is expected to
be positively related to the level of tax collection of a country for two main reasons. On
the one hand, a higher level of development increases the demand for public expenditure
(Bird et al., 2008; Fenochietto and Pessino, 2010). On the other hand, it is related to a
higher capacity of payment and, therefore, tax collection (Bird et al., 2008). Lastly,
countries of higher income, i.e. with a higher per capita GDP, are likely to have a better
tax administration (Agbeyegbe et al., 2006). Per capita GDP is the most commonly found
determinant of tax revenue in literature. Fenochietto and Pessino (2010) empirically
prove the positive and significant relationship between the two variables.

5.1.2 Sectoral composition of output

Certain economic sectors are easier to tax than others (Gupta, 2007). In general, the
agricultural sector is one that is difficult to tax (Agbeyegbe et al., 2006; Gupta, 2007).

There are two reasons for this. One the one hand, for political reasons, in some countries
agricultural products are free of consumption taxes, i.e., VAT. On the other hand, the
agricultural sector is fairly difficult to supervise, especially if dominated by small
producers (Fenochietto and Pessino, 2010). “A large industrial sector is easier to monitor
and tax, and a larger share of manufacturing in GDP captures economic development and
a larger formal (taxable) sector” (Morissey et al., 2016: 1961). The ease (or difficulty)
of tax collection can be represented by the value added of the agriculture sector as percent
of GDP (Fenochietto and Pessino, 2010). The amount of collected tax will be lower the
larger the share of agriculture in GDP and the smaller the share of industry or
manufacturing.

623
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.1.3 Degree of trade and financial openness of an economy

Trade taxes are fairly easy to collect (Morissey et al., 2016). The sum of exports and
imports as a percentage of GDP is commonly used as the variable that reflects the degree
of openness of an economy (Gupta, 2007; Fenochietto and Pessino, 2010; Morissey et
al., 2016). A greater volume of exports and imports will result in higher revenue
(Morissey et al., 2016).

In their study of tax performance, Morissey et al. (2016) include the “shares of
agricultural, mining, manufacturing and fuel exports to GDP along with imports to GDP
rather than a single combined measure of trade to GDP.”

Agbeyegbe et al. (2006) test the impact of trade liberalisation on tax revenue in Sub-
Saharan countries with two measures: one is the share of external trade in GDP, and the
other is the collected tariff, i.e., the ratio between import duties and the value of imports.

5.1.4 External indebtedness

External indebtedness is another variable that can affect revenue levels (Gupta, 2007).
The reasoning is as follows: in order to service their debt, countries may choose to
decrease their import levels, hence lowering the import taxes; in alternative, they can
increase tariffs over imports (or other taxes) aiming at generating a primary budget
surplus which will finance the debt.

5.1.5 Foreign aid

Gupta (2007) considers foreign aid as a possible determinant of revenue collection levels.
The variable used by the author is the ratio between foreign aid and GDP.

624
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.1.6 Income distribution

Fenochietto and Pessino (2010) assert that an improved distribution of a country’s income
should ease both revenue collection and voluntary compliance of taxpayers. These
authors use the GINI coefficient to represent this determinant. The GINI coefficient is a
measure of the deviation of the distribution of income amongst individuals from the equal
distribution. However, the authors’ study actually finds a negative relationship between
the two variables.

5.1.7 Inflation

Fenochietto and Pessino (2010) choose the percentage change of consumption price index
(CPI) to evaluate inflation. According to these authors, in general, countries whose
resources are attained from printing money have negative efficiency in tax collection,
hence the expected negative sign of this variable. The authors’s study empirically
corroborates such negative relationship between tax revenue as a percentage of GDP and
inflation as measured by the CPI.

5.1.8 Foreign direct investment

The relationship between tax revenue and foreign direct investment has contradicting
findings in the literature – either a negative one or no relationship (Kiburi et. al, 2017).

5.2 Institutional variables

Gupta (2007) mentions “corruption, law and order, government stability, political
stability and economic stability“ as possible determinants of tax revenue. Of several
“inefficiencies” Fenochietto and Pessino (2010) refer to that can impede countries from
reaching their tax capacity, corruption is the one used in the authors’ analysis. To
represent this determinant of tax revenue, they use the corruption perception index.
Corruption happens when someone misuses or abuses public office in order to obtain

625
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

personal benefit (Arif and Rawat, 2018: 2). Fenochietto and Pessino (2010) find a
negative relationship between tax burden and corruption. In their study, Brasoveanu and
Brasoveanu (2009) empirically test the relationship between corruption overall and tax
burden. Their findings are not conclusive.

In their study, Bird et al. (2008) include a determinant of tax revenue to which they call
“voice”. They consider that the willingness of the taxpayers to contribute is raised if they
see their interests accurately represented, because they feel they have a “meaningful
voice” that influences the state.

Castro and Camarillo (2014) assess the impact of other institutional factors on tax
revenue. They consider “political rights”, which measure essentially the level of
democracy of a country, and “civil liberties”, which reflect “freedom of expression,
assembly and thought, and legal security”. They conclude that only “civil liberties” are
significant.

5.3 Political variables

Gupta (2007) assesses how the specific sources of tax revenue impact the countries’ total
revenues over GDP. They find that countries that count more on the taxation of goods
and services have a decreased revenue performance. The rationale is that these taxes are
generally regressive, which means they are not discriminated according to income levels,
i.e., they are the same for every individual. Subsequently, in such countries, the inequality
in the distribution of income can be aggravated and decrease the tax base, which can result
in a decline of the revenue. In contrast, Gupta (2007) finds that a higher dependence of a
country on progressive taxes (taxes that are based on a taxpayer’s ability to pay them)
like those over income, profits and capital gains can increase revenue.

In addition, Gupta (2007) considers corporate and individual tax rates, and average tariffs.
However, he concludes they do not have a significant impact on revenue performance.

626
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.4 Demographic/social variables

5.4.1 Level of education

“More educated people can understand better how and why it is necessary to pay taxes.”
(Fenochietto and Pessino, 2010: 73). Therefore, compliance will be higher in countries
with a higher level of education. The variable commonly used as a proxy for the level of
education is labour force with secondary education as a percentage of total labour force.
However, as suggested by Fenochietto and Pessino (2010), not only is this variable not
available for all countries, but also secondary education is different amongst countries.
These authors therefore use the total public expenditure on education percent of GDP as
the variable that represents the level of education. Castro and Camarillo (2014)’s analysis
also includes a variable to evaluate the effect of education on countries’ tax revenue levels
– gross tertiary school enrolment.

5.4.2 Population

Bird et al. (2008) consider that the rate of population growth could also exert an influence
over tax effort. The rational here is that countries that have a faster growing population
may have a decreased capacity of capturing new taxpayers. Castro and Camarillo (2014)
included life expectancy and child mortality rate as in their study.

5.5 Other variables

Morissey et al. (2016) assess the effect of the exposure to exogenous shocks on revenue,
particularly “exchange rate pressure, terms of trade and intensity of natural catastrophes”.
For lower income countries, they evaluate the impact of the countries’ “natural resource
endowments and political regime type”.

627
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

6. CONCLUSION

The purposes of this dissertation were to provide the literature with a compilation and an
explanation of fundamental taxation-related concepts, to perform a systematic literature
review on the investigation on taxation, to study the evolution of both the tax burden and
the tax effort in the OECD, and to address some of the determinants of tax revenue used
in previous literature.

One of the main conclusions is that of the importance of taxation for the financing of the
government’s expenses. It is clear that there is some misconception in what concerns the
concepts of tax burden and tax effort, which are often used interchangeably and rarely
have an explicit definition. This suggests there should be a more unanimous construction
of these concepts in the literature.

A satisfactory level of tax burden or tax effort is crucial for the decision-making process
of tax authorities. Therefore, adequate metrics must be applied in order to measure the
level of taxation. The ratio between total tax revenue and GDP is the most frequently used
metric for the measurement of the tax burden of a country and for inter-country
comparison due to its simplicity. However, it is not efficient for making tax policy
decisions, nor does it indicate who or what takes the burden. The indices proposed by
Frank (1959) and Bird (1964) represent an advance in what concerns the measurement of
the level of taxation. The metrics were computed gathering data from both the World
Bank and the OECD databases. Lithuania was excluded from the study due to the lack of
tax revenue information. Furthermore, the analysis of the evolution of the level of taxation
as measured by the tax burden or the tax effort would be extensive were it to be made for
each country separately. Averages were computed in order to perform this analysis, and
therefore the conclusions are only valid taking into consideration the limitations of
interpretation of the average. Overall, both Frank’s and Bird’s indices decreased from
2000 and 2015, despite an upward movement in the last years. In addition, fiscal
convergence amongst the OECD countries was observed for both indices. In contrast, the
tax-to-GDP ratio depicts a tendency to growth, and fiscal convergence is not verified.
Since these conclusions are conflicting, it is suggested that more adequate metrics for the
measurement of the level of taxation are developed.

628
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

BIBLIOGRAPHY

Agbeyegbe, T. D., Stotsky, J. WoldeMariam, A. 2006. Trade liberalization, exchange rate

changes, and tax revenue in Sub-Saharan Africa. Journal of Asian Economics, 17(2):
261-284.

Andrejovská, A., & Hudáková, M.. 2016. Classiﬁcation of EU countries in the context of
corporate income tax. Acta Universitatis Agriculturae et Silviculturae Mendelianae
Brunensis,64(5): 1699–1708.

Arif, I., & Rawat, A. S. 2018. Corruption, governance, and tax revenue: Evidence from
EAGLE countries. Journal of Transnational Management, 23(1): 1-15.

Bingyang, L., & Qingwang, G. 2012. Why China's tax revenue is likely to maintain its
rapid growth: An explanation within the framework of tax capacity and tax effort. Social
Sciences in China, 33 (1):108-126.

Bird, R. M. 1964. A note on "tax sacrifice" comparisons. National Tax Journal, 17(3):
303-308.

Bird, R. M., Martinez-Vasquez, J., & Torgler, B. 2008. Tax effort in developing countries
and high income countries: The impact of corruption, voice and accountability. Economic
Analysis and Policy, 38(1): 55-71.

Bonatti, L. 2007. Fiscal transfers and distributive conflict in a simple endogenous growth
model with unemployment. German Economic Review, 8 (1): 41–63

Brasoveanu, I. V., & Brasoveanu, L. O. 2009. Correlation between corruption and tax
revenues in EU 27. Economic Computation & Economic Cybernetics Studies &
Research, 43(4): 133-142.

Buehn, A., Dell’Anno, R., & Schneider, F. 2018. Exploring the dark side of tax policy:
An analysis of the interactions between fiscal illusion and the shadow economy.
Empirical Economics.

Castro, G. A., & Camarillo, D. B. 2014. Determinants of tax revenue in OECD countries
over the period 2001-2011. Contaduría y Administración, 59 (3): 35-59

629
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Cronin, J., DeFilippes, P, & Lin, E. Y. 2012. Effects of adjusting distribution tables for
family size. National Tax Journal, 65 (4): 739-758.

Dubauskas, G. 2016. The management of public finance literacy for sustainable

economical environment. Journal of Security and Sustainability Issues, 5(3): 403-409.

Fedyshyn, N. I. 2013. Tax burden: Essence, types and factors of influence. Actual
Problems of Economics, 150 (12): 208-212.

Frank, H. J. 1959. Measuring state tax burdens. National Tax Journal, 12 (2): 179-185.

Gupta, A. S. 2007. Determinants of tax revenue efforts in developing countries.

Working paper no. 7/184. International Monetary Fund.

Gur, N. 2014. Taxation and democracy: An instrumental variable approach. Applied

Economics Letters, 21 (11): 763–766.

Kiburi, W. W., Mirie, M. W., Okiro, K. O., & Ruigu, G. M. 2017. The relationship
between tax burden and foreign direct investment inflows: a review of empirical
literature. European Journal of Accounting, Auditing and Finance Research, 5 (5): 67-
77.

Kim, K., & Lim, S. 2018. Analysis of state general sales tax policy: A complementary
approach to the effects of causes and the causes of effects. Social Science Journal, 55
(2): 128-138.

Kiser, E., & Karceski, S. M. 2017. Political economy of taxation. Annual Review of
Political Science, 20 (1): 75-92.

Koch, S. F., Schoeman, N. J., & Toder, J. J. Economic growth and the structure of taxes
in South Africa: 1960 – 2002. South African Journal of Economics, 73 (2): 190-210.

Kotlán, I., & Machová, Z. 2012. World Tax Index: Methodology and data. DANUBE:
Law, Economics and Social Issues Review, 3 (2): 19-33.

630
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Lago-Peñas, I., & Lago-Peñas, S. 2010. The determinants of tax morale in comparative
perspective: Evidence from European countries. European Journal of Political
Economy, 26 (4): 441-453.

Le, T. M., Moreno-Dodson, B., & Bayraktar, N. 2012. Tax capacity and tax effort:
Extended cross-country analysis from 1994 to 2009. Working paper no. 6252,
International Trade and Investment Unit, Investment Climate Department, The World
Bank.

Macek, R. 2018. Labour taxation and its impact on economic growth – Complex
analysis. DANUBE: Law and Economics Review, 9 (1): 49-61.

Machová, Z., & Kotlán, I. 2013. World Tax Index: New methodology for OECD
countries, 2000–2012. DANUBE: Law, Economics and Social Issues Review, 4(2): 165-
179.

Mahdavi, S. 2008. The level and composition of tax revenue in developing countries:
Evidence from unbalanced panel data. International Review of Economics and Finance,
17 (4): 607–617.

Morrisey, O., Haldenwang, C., Schiller, A., Ivanyna, M., & Bordon, I. 2016. Tax Revenue
Performance and Vulnerability in Developing Countries. The Journal of Development
Studies, 52(12): 1689-1703.

Nisha, R. 2018. Technical Analysis of Tax Revenue and Non- Tax Revenue of India.
International Journal of Pure and Applied Mathematics, 199 (12): 14957-14970.

OECD. 2017. Revenue Statistics: 1965-2016. Paris: OECD Publishing.

Pessino, C., & Fenochietto, R. 2010. Determining countries’ tax effort. Hacienda Pública
Española, 195(4): 65-87.

Rabiei, M., & Balagetabi, F. E. 2013. Estimation of tax capacity and effort and oil
revenue. Research Journal of Applied Sciences, Engineering and Technology, 6(10):
1799-1804

631
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Reed, W. R., & Rogers, C. L. 2006. Tax burden and the mismeasurement of state tax
policy. Public Finance Review, 34 (4): 404-426.

Schneider, F. 2005. Shadow economies around the world: What do we really know?
European Journal of Political Economy, 21 (3): 598-642.

Silva, S. T. 2015. Direito fiscal: teoria geral. Coimbra: Imprensa da Universidade de

Coimbra.

Sobarzo, H. 2004. Tax effort and tax potential of state governments in Mexico:
A representative tax system. Working paper no. 315, Kellogg Institute for International
Studies, University of Notre Dame, Notre Dame, Indiana.

Streimikiene, D., Ahmed, R. R., Vveinhardt, J., Ghauri, S. P., & Zahid, S. 2018.
Forecasting tax revenues using time series techniques – a case of Pakistan. Economic
Research-Ekonomska Istraživanja, 31 (1): 722-754.

Strelnik E. U., Usanova D. S., & Khairullin I. G. 2015. Problematic aspects of corporate
tax risk empirical analysis. Asian Social Science, 11 (11): 374-378.

Švaljek, S. 2005. The 2000 tax reform in Croatia: Causes and consequences. Ekonomski
Pregled, 56(12): 1217-1236.

Thornton, J. 2014. Does foreign aid reduce tax revenue? Further evidence. Applied
Economics, 46 (4): 359-373.

Vallés-Giménez, J., & Zárate-Marco, A. 2017. Tax effort of local governments and its
determinants: The Spanish case. Zou, H. (Eds.), Annals of Economics and Finance,
18(2): 323-348.

Vasiliauskaite, A., & Stankevicius, E. 2009. Tax burden management and GDP growth:
Case of EU countries. Economics and Management, 14: 202-209.

Wang, D. H. 2007. Convergence tests on tax burden and economic growth among China,
Taiwan and the OECD countries. Physica A: Statistical Mechanics and its Applications.
380: 278-286.

632
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Winer, S. L., & Hettich, W. 2008. Taxation. In Rowley, C. K., & Schneider, F. G. (Eds.),
Readings in Public Choice and Constitutional Political Economy: 391-422. New York:
Springer.

Xing, W., & Zhang, Q. 2018. The effects of vertical and horizontal incentives on local
tax efforts: Evidence from China. Applied Economics, 50(11): 1222-1237.

633
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ANNEXES

Abbreviation Country
AUS Australia
AUT Austria
BEL Belgium
CAN Canada
CHE Switzerland
CHL Chile
CZE Czech Republic
DEU Germany
DNK Denmark
ESP Spain
EST Estonia
FIN Finland
FRA France
GBR United Kingdom
GRC Greece
HUN Hungary
IRL Ireland
ISL Iceland
ISR Israel
ITA Italy
JPN Japan
KOR South Korea
LUX Luxembourg
LVA Latvia
MEX Mexico
NLD Netherlands
NOR Norway
NZL New Zealand
POL Poland
PRT Portugal
SVK Slovak Republic
SVN Slovenia
SWE Sweden
TUR Turkey
USA United States
Table 5: List of abbreviations of countries’ names

634
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Year Number of articles

1972 1
1982 1
1989 1
1990 1
1992 1
1996 1
1997 1
1998 2
2000 5
2001 2
2002 3
2003 1
2004 3
2005 11
2006 5
2007 12
2008 9
2009 10
2010 22
2011 25
2012 24
2013 32
2014 40
2015 40
2016 45
2017 48
Table 6: Number of articles published per year

635
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country Number of authors

Argentina 2
Australia 19
Austria 4
Azerbeijan 1
Belgium 7
Brazil 21
Canada 3
Chile 3
China 62
Cote d'Ivoire 2
Croatia 1
Cyprus 1
Czech Repubic 67
Denmark 1
Estonia 6
Finland 4
France 6
Germany 44
Ghana 2
Greece 9
Hungary 2
India 12
Indonesia 5
Iran 5
Israel 3
Italy 3
Japan 10
Lithuania 11
Malaysia 8
Mexico 4
Netherlands 8
New Zealand 13
Nigeria 3
Norway 5
Pakistan 8
Romania 11
Russia 50
Sierra Leone 1
Slovak Republic 9
Slovenia 2
South Africa 13

636
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

South Korea 10
Spain 46
Sweden 3
Switzerland 16
Taiwan 7
Tanzania 2
Turkey 8
UAE 1
Uganda 8
UK 21
Ukraine 22
US 94
Venezuela 2
Table 7: Number of authors per country of affiliation

637
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Year Tax burden Tax effort Tax revenue

1972 1
1982 1
1989 1
1990 1
1992 1
1996 1
1997 1
1998 2
2000 3 1 1
2001 1 1
2002 2 1
2003 1
2004 2 1
2005 9 2
2006 4 1
2007 8 4
2008 3 6
2009 7 3
2010 15 1 6
2011 10 1 15
2012 12 2 10
2013 20 2 11
2014 21 2 18
2015 18 2 21
2016 17 30
2017 18 2 29
Total 177 15 161
Table 8: Number of articles with each keyword per year

638
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2000 2001 2002 2003 2004 2005 2006

AUS 0.1465 0.1618 0.1728 0.1622 0.1170 0.1044 0.0970
AUT 0.1752 0.1840 0.1639 0.1324 0.1143 0.1072 0.0990
BEL 0.1786 0.1814 0.1687 0.1354 0.1184 0.1140 0.1077
CAN 0.1548 0.1529 0.1446 0.1225 0.1072 0.0944 0.0839
CHE 0.0638 0.0644 0.0629 0.0491 0.0439 0.0411 0.0397
CHL 0.3962 0.4433 0.4629 0.4466 0.3650 0.3266 0.3004
CZE 0.5624 0.5228 0.4486 0.3775 0.3272 0.2844 0.2549
DEU 0.1549 0.1504 0.1400 0.1162 0.0978 0.0958 0.0915
DNK 0.1588 0.1536 0.1401 0.1143 0.0989 0.0968 0.0872
ESP 0.2282 0.2191 0.1999 0.1567 0.1395 0.1364 0.1307
EST 0.8248 0.7443 0.6469 0.4817 0.3902 0.3138 0.2699
FIN 0.1914 0.1736 0.1608 0.1308 0.1096 0.1071 0.1007
FRA 0.1847 0.1830 0.1703 0.1383 0.1211 0.1184 0.1134
GBR 0.1179 0.1180 0.1042 0.0897 0.0787 0.0756 0.0736
GRC 0.2733 0.2514 0.2289 0.1684 0.1396 0.1379 0.1270
HUN 0.9206 0.7903 0.6242 0.4877 0.4038 0.3673 0.3575
IRL 0.1543 0.1391 0.1194 0.0913 0.0816 0.0770 0.0738
ISL 0.1203 0.1259 0.1039 0.0908 0.0825 0.0768 0.0824
ISR 0.1887 0.1863 0.1986 0.1893 0.1778 0.1669 0.1581
ITA 0.2041 0.1989 0.1810 0.1479 0.1267 0.1220 0.1206
JPN 0.0652 0.0738 0.0749 0.0686 0.0644 0.0678 0.0727
KOR 0.1823 0.1963 0.1731 0.1611 0.1386 0.1228 0.1136
LUX 0.0960 0.0956 0.0983 0.0857 0.0573 0.0548 0.0598
LVA 0.6927 0.6264 0.5558 0.4945 0.4321 0.3754 0.3119
MEX 0.1936 0.1966 0.1993 0.2075 0.1754 0.1565 0.1453
NLD 0.1392 0.1374 0.1244 0.0983 0.0860 0.0861 0.0794
NOR 0.1128 0.1090 0.0978 0.0822 0.0734 0.0624 0.0577
NZL 0.2824 0.2586 0.2079 0.1619 0.1511 0.1500 0.1551
POL 0.7394 0.6641 0.6397 0.5820 0.5103 0.4254 0.3877
PRT 0.2806 0.2739 0.2516 0.2040 0.1723 0.1691 0.1687
SVK 0.6291 0.5728 0.5075 0.4149 0.3210 0.2832 0.2372
SVN 0.3055 0.3167 0.3001 0.2485 0.2209 0.2112 0.1944
SWE 0.1673 0.1734 0.1523 0.1186 0.1061 0.1050 0.0941
TUR 0.5637 0.8597 0.6794 0.5530 0.3969 0.3239 0.2994
USA 0.0754 0.0706 0.0639 0.0606 0.0578 0.0574 0.0551
Average 0.2836 0.2791 0.2505 0.2106 0.1773 0.1604 0.1486
SD 0.2278 0.2195 0.1854 0.1552 0.1260 0.1047 0.0926
Table 9: Frank’s index

639
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2007 2008 2009 2010 2011 2012 2013

AUS 0.0915 0.0589 0.0704 0.0599 0.0500 0.0436 0.0412
AUT 0.0867 0.0786 0.0855 0.0860 0.0795 0.0859 0.0836
BEL 0.0937 0.0851 0.0942 0.0903 0.0887 0.0941 0.0937
CAN 0.0753 0.0692 0.0825 0.0682 0.0616 0.0617 0.0614
CHE 0.0407 0.0418 0.0373 0.0317 0.0301 0.0308 0.0301
CHL 0.2737 0.2329 0.1958 0.1766 0.1630 0.1527 0.1362
CZE 0.2135 0.1692 0.1902 0.1928 0.1795 0.1944 0.1936
DEU 0.0810 0.0759 0.0827 0.0806 0.0726 0.0789 0.0755
DNK 0.0786 0.0681 0.0758 0.0747 0.0697 0.0744 0.0704
ESP 0.1167 0.0951 0.0953 0.1045 0.1014 0.1145 0.1146
EST 0.2175 0.1932 0.2522 0.2533 0.2001 0.1974 0.1737
FIN 0.0854 0.0767 0.0847 0.0859 0.0820 0.0891 0.0873
FRA 0.0978 0.0889 0.0958 0.0990 0.0942 0.1056 0.1034
GBR 0.0650 0.0694 0.0826 0.0820 0.0794 0.0793 0.0777
GRC 0.1140 0.1027 0.1078 0.1240 0.1372 0.1576 0.1624
HUN 0.3277 0.2864 0.3316 0.3165 0.2854 0.3261 0.2951
IRL 0.0662 0.0619 0.0754 0.0796 0.0780 0.0838 0.0753
ISL 0.0630 0.0990 0.1173 0.1137 0.0976 0.0967 0.0767
ISR 0.1379 0.1127 0.1132 0.1045 0.0943 0.0972 0.0883
ITA 0.1105 0.1044 0.1139 0.1172 0.1099 0.1265 0.1249
JPN 0.0735 0.0647 0.0609 0.0565 0.0543 0.0550 0.0673
KOR 0.1081 0.1208 0.1307 0.1057 0.0992 0.0997 0.0925
LUX 0.0452 0.0467 0.0802 0.0681 0.0660 0.0744 0.0755
LVA 0.2128 0.1737 0.1987 0.2418 0.2022 0.2083 0.1896
MEX 0.1341 0.1339 0.1651 0.1461 0.1341 0.1341 0.1372
NLD 0.0691 0.0663 0.0680 0.0705 0.0646 0.0699 0.0689
NOR 0.0498 0.0431 0.0509 0.0469 0.0411 0.0402 0.0379
NZL 0.1222 0.1219 0.1182 0.1007 0.0878 0.0878 0.0799
POL 0.3274 0.2532 0.2867 0.2665 0.2450 0.2604 0.2465
PRT 0.1490 0.1382 0.1397 0.1445 0.1449 0.1623 0.1619
SVK 0.1932 0.1614 0.1772 0.1770 0.1695 0.1706 0.1693
SVN 0.1618 0.1384 0.1510 0.1604 0.1484 0.1664 0.1602
SWE 0.0792 0.0734 0.0907 0.0784 0.0683 0.0710 0.0681
TUR 0.2426 0.2181 0.2667 0.2371 0.2329 0.2166 0.2057
USA 0.0548 0.0526 0.0485 0.0476 0.0463 0.0443 0.0465
Average 0.1274 0.1136 0.1262 0.1225 0.1131 0.1186 0.1135
SD 0.0765 0.0619 0.0706 0.0693 0.0619 0.0661 0.0611
Table 9 (cont.): Frank’s index

640
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2014 2015

AUS 0.0463 0.0475
AUT 0.0824 0.0989
BEL 0.0938 0.1115
CAN 0.0636 0.0758
CHE 0.0306 0.0317
CHL 0.1418 0.1583
CZE 0.1928 0.2165
DEU 0.0736 0.0865
DNK 0.0721 0.0815
ESP 0.1146 0.1312
EST 0.1733 0.2051
FIN 0.0862 0.1017
FRA 0.1026 0.1201
GBR 0.0712 0.0761
GRC 0.1620 0.1990
HUN 0.2945 0.3474
IRL 0.0708 0.0598
ISL 0.0761 0.0744
ISR 0.0842 0.0893
ITA 0.1228 0.1444
JPN 0.0747 0.0825
KOR 0.0879 0.0923
LUX 0.0749 0.0830
LVA 0.1837 0.2150
MEX 0.1380 0.1790
NLD 0.0723 0.0845
NOR 0.0375 0.0474
NZL 0.0786 0.0926
POL 0.2390 0.2767
PRT 0.1606 0.1895
SVK 0.1714 0.2072
SVN 0.1523 0.1850
SWE 0.0690 0.0827
TUR 0.2063 0.2333
USA 0.0450 0.0442
Average 0.1128 0.1301
SD 0.0603 0.0726
Table 9 (cont.): Frank’s index

641
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country Average
AUS 0.0919
AUT 0.1089
BEL 0.1156
CAN 0.0925
CHE 0.0419
CHL 0.2733
CZE 0.2825
DEU 0.0971
DNK 0.0947
ESP 0.1374
EST 0.3461
FIN 0.1096
FRA 0.1210
GBR 0.0838
GRC 0.1621
HUN 0.4226
IRL 0.0867
ISL 0.0936
ISR 0.1367
ITA 0.1360
JPN 0.0673
KOR 0.1266
LUX 0.0726
LVA 0.3322
MEX 0.1610
NLD 0.0865
NOR 0.0619
NZL 0.1410
POL 0.3969
PRT 0.1819
SVK 0.2852
SVN 0.2013
SWE 0.0998
TUR 0.3585
USA 0.0544
Table 10: Frank’s index: average per country

642
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2000 2001 2002 2003 2004 2005 2006

AUS 0.2056 0.2266 0.2536 0.2499 0.1737 0.1525 0.1404
AUT 0.3030 0.3271 0.2858 0.2299 0.1970 0.1815 0.1664
BEL 0.3183 0.3225 0.3008 0.2394 0.2092 0.2010 0.1895
CAN 0.2368 0.2294 0.2136 0.1801 0.1579 0.1389 0.1239
CHE 0.0916 0.0903 0.0882 0.0699 0.0621 0.0589 0.0564
CHL 0.4749 0.5339 0.5552 0.5245 0.4229 0.3857 0.3519
CZE 0.8228 0.7634 0.6612 0.5644 0.4902 0.4245 0.3751
DEU 0.2423 0.2307 0.2122 0.1770 0.1484 0.1455 0.1408
DNK 0.2982 0.2836 0.2564 0.2100 0.1848 0.1863 0.1633
ESP 0.3411 0.3246 0.2980 0.2335 0.2108 0.2088 0.2026
EST 1.1727 1.0409 0.9165 0.6768 0.5522 0.4381 0.3777
FIN 0.3529 0.3055 0.2842 0.2269 0.1889 0.1851 0.1745
FRA 0.3274 0.3224 0.2967 0.2405 0.2111 0.2085 0.2008
GBR 0.1758 0.1757 0.1532 0.1315 0.1173 0.1134 0.1101
GRC 0.4116 0.3713 0.3381 0.2435 0.1979 0.2006 0.1804
HUN 1.4703 1.2422 0.9765 0.7628 0.6273 0.5653 0.5507
IRL 0.2079 0.1787 0.1485 0.1163 0.1060 0.1009 0.0999
ISL 0.1863 0.1899 0.1590 0.1411 0.1286 0.1258 0.1365
ISR 0.2818 0.2801 0.2956 0.2789 0.2630 0.2507 0.2400
ITA 0.3432 0.3330 0.3002 0.2465 0.2088 0.2006 0.2032
JPN 0.0888 0.1004 0.1007 0.0920 0.0871 0.0933 0.1014
KOR 0.2309 0.2498 0.2212 0.2080 0.1774 0.1577 0.1483
LUX 0.1460 0.1475 0.1491 0.1295 0.0874 0.0858 0.0868
LVA 0.9008 0.8101 0.7231 0.6584 0.5787 0.5165 0.4306
MEX 0.2095 0.2136 0.2186 0.2313 0.1946 0.1730 0.1608
NLD 0.2232 0.2141 0.1926 0.1524 0.1334 0.1327 0.1257
NOR 0.1935 0.1884 0.1697 0.1411 0.1275 0.1091 0.1009
NZL 0.4185 0.3725 0.2961 0.2328 0.2236 0.2297 0.2355
POL 1.1002 0.9877 0.9534 0.8578 0.7378 0.6295 0.5781
PRT 0.4028 0.3916 0.3628 0.2952 0.2447 0.2422 0.2415
SVK 0.9455 0.8514 0.7543 0.5997 0.4589 0.4058 0.3297
SVN 0.4456 0.4763 0.4615 0.3899 0.3510 0.3395 0.3108
SWE 0.3279 0.3259 0.2784 0.2184 0.1955 0.1970 0.1751
TUR 0.7306 1.1350 0.8815 0.7304 0.5125 0.4197 0.3887
USA 0.1058 0.0979 0.0858 0.0806 0.0771 0.0780 0.0762
Average 0.4210 0.4095 0.3669 0.3075 0.2584 0.2366 0.2193
SD 0.3306 0.3075 0.2584 0.2137 0.1727 0.1449 0.1285
Table 11: Bird’s index

643
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2007 2008 2009 2010 2011 2012 2013

AUS 0.1360 0.0785 0.0958 0.0821 0.0686 0.0592 0.0553
AUT 0.1458 0.1342 0.1449 0.1460 0.1352 0.1475 0.1458
BEL 0.1643 0.1507 0.1635 0.1587 0.1564 0.1696 0.1711
CAN 0.1109 0.0999 0.1211 0.0978 0.0882 0.0889 0.0885
CHE 0.0551 0.0543 0.0517 0.0446 0.0414 0.0425 0.0417
CHL 0.3266 0.2795 0.2242 0.2079 0.1975 0.1873 0.1644
CZE 0.3147 0.2461 0.2707 0.2746 0.2594 0.2848 0.2856
DEU 0.1253 0.1178 0.1307 0.1251 0.1142 0.1253 0.1206
DNK 0.1468 0.1235 0.1380 0.1357 0.1271 0.1372 0.1309
ESP 0.1816 0.1380 0.1342 0.1507 0.1460 0.1684 0.1709
EST 0.3051 0.2736 0.3822 0.3697 0.2845 0.2828 0.2509
FIN 0.1462 0.1304 0.1440 0.1458 0.1417 0.1557 0.1548
FRA 0.1711 0.1547 0.1644 0.1719 0.1667 0.1900 0.1890
GBR 0.0975 0.1025 0.1203 0.1219 0.1196 0.1173 0.1141
GRC 0.1634 0.1462 0.1535 0.1805 0.2041 0.2451 0.2515
HUN 0.5286 0.4625 0.5345 0.4948 0.4384 0.5212 0.4701
IRL 0.0883 0.0797 0.0935 0.0983 0.0957 0.1033 0.0958
ISL 0.1014 0.1406 0.1566 0.1576 0.1402 0.1426 0.1182
ISR 0.2095 0.1641 0.1593 0.1495 0.1355 0.1370 0.1263
ITA 0.1895 0.1784 0.1966 0.2015 0.1891 0.2255 0.2232
JPN 0.1033 0.0899 0.0837 0.0782 0.0763 0.0779 0.0968
KOR 0.1434 0.1603 0.1713 0.1382 0.1314 0.1335 0.1228
LUX 0.0671 0.0692 0.1246 0.1020 0.0983 0.1156 0.1181
LVA 0.2908 0.2383 0.2875 0.3389 0.2797 0.2897 0.2651
MEX 0.1525 0.1537 0.1900 0.1689 0.1537 0.1522 0.1556
NLD 0.1084 0.1032 0.1052 0.1108 0.1015 0.1103 0.1092
NOR 0.0860 0.0733 0.0868 0.0809 0.0710 0.0689 0.0632
NZL 0.1799 0.1750 0.1666 0.1412 0.1236 0.1268 0.1143
POL 0.4940 0.3810 0.4106 0.3818 0.3533 0.3774 0.3564
PRT 0.2149 0.1981 0.1953 0.2037 0.2121 0.2348 0.2440
SVK 0.2682 0.2248 0.2481 0.2427 0.2324 0.2353 0.2414
SVN 0.2551 0.2152 0.2354 0.2532 0.2328 0.2623 0.2505
SWE 0.1449 0.1321 0.1631 0.1392 0.1196 0.1245 0.1199
TUR 0.3134 0.2819 0.3460 0.3139 0.3127 0.2874 0.2739
USA 0.0751 0.0710 0.0632 0.0626 0.0616 0.0595 0.0634
Average 0.1887 0.1663 0.1845 0.1792 0.1660 0.1768 0.1704
SD 0.1090 0.0884 0.1023 0.0991 0.0876 0.0980 0.0908
Table 11 (cont.): Bird’s index

644
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2014 2015

AUS 0.0627 0.0632
AUT 0.1439 0.1735
BEL 0.1710 0.2019
CAN 0.0917 0.1106
CHE 0.0419 0.0444
CHL 0.1719 0.1941
CZE 0.2787 0.3144
DEU 0.1174 0.1386
DNK 0.1403 0.1514
ESP 0.1725 0.1976
EST 0.2542 0.3062
FIN 0.1538 0.1818
FRA 0.1879 0.2194
GBR 0.1035 0.1111
GRC 0.2529 0.3129
HUN 0.4661 0.5554
IRL 0.0907 0.0669
ISL 0.1223 0.1157
ISR 0.1218 0.1294
ITA 0.2171 0.2535
JPN 0.1098 0.1218
KOR 0.1167 0.1236
LUX 0.1149 0.1245
LVA 0.2580 0.3024
MEX 0.1578 0.2091
NLD 0.1156 0.1346
NOR 0.0620 0.0781
NZL 0.1134 0.1339
POL 0.3451 0.4023
PRT 0.2423 0.2857
SVK 0.2471 0.3029
SVN 0.2377 0.2868
SWE 0.1208 0.1459
TUR 0.2719 0.3091
USA 0.0618 0.0609
Average 0.1696 0.1961
SD 0.0888 0.1088
Table 11 (cont.): Bird’s index

645
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country Average
AUS 0.1315
AUT 0.1880
BEL 0.2055
CAN 0.1361
CHE 0.0584
CHL 0.3252
CZE 0.4144
DEU 0.1507
DNK 0.1758
ESP 0.2050
EST 0.4928
FIN 0.1920
FRA 0.2139
GBR 0.1241
GRC 0.2408
HUN 0.6667
IRL 0.1106
ISL 0.1414
ISR 0.2014
ITA 0.2319
JPN 0.0938
KOR 0.1646
LUX 0.1104
LVA 0.4480
MEX 0.1809
NLD 0.1358
NOR 0.1063
NZL 0.2052
POL 0.5842
PRT 0.2632
SVK 0.4118
SVN 0.3127
SWE 0.1830
TUR 0.4693
USA 0.0738
Table 12: Bird’s index: average per country

646
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2000 2001 2002 2003 2004 2005 2006

AUS 0.3285 0.2854 0.2777 0.2735 0.2966 0.3052 0.2825
AUT 0.4218 0.4062 0.3486 0.3692 0.3998 0.3854 0.3495
BEL 0.4356 0.3996 0.3545 0.3714 0.4140 0.4076 0.3733
CAN 0.3568 0.3324 0.2818 0.2864 0.2872 0.2913 0.2947
CHE 0.2677 0.2482 0.2345 0.2397 0.2549 0.2509 0.2359
CHL 0.2064 0.1939 0.1760 0.1437 0.1540 0.1647 0.1960
CZE 0.2958 0.2683 0.2746 0.2864 0.3033 0.3020 0.2796
DEU 0.3624 0.3287 0.2857 0.3079 0.3338 0.3226 0.3012
DNK 0.4667 0.4237 0.3724 0.3961 0.4412 0.4489 0.4117
ESP 0.3161 0.2917 0.2587 0.2811 0.3154 0.3214 0.3073
EST 0.2831 0.2591 0.2323 0.2521 0.2687 0.2473 0.2331
FIN 0.4452 0.4001 0.3540 0.3689 0.4028 0.3973 0.3576
FRA 0.4281 0.3952 0.3440 0.3676 0.4082 0.4063 0.3772
GBR 0.3347 0.2997 0.2745 0.2665 0.3079 0.3063 0.2887
GRC 0.3188 0.2834 0.2460 0.2598 0.2906 0.2828 0.2604
HUN 0.3384 0.2989 0.2965 0.3054 0.3410 0.3583 0.3012
IRL 0.2815 0.2442 0.2127 0.2363 0.2653 0.2679 0.2652
ISL 0.3969 0.3039 0.2783 0.2937 0.3009 0.3893 0.3253
ISR 0.3533 0.3750 0.3242 0.3123 0.3170 0.3120 0.2951
ITA 0.3989 0.3701 0.3212 0.3501 0.3820 0.3731 0.3580
JPN 0.2936 0.2694 0.2304 0.2275 0.2550 0.2765 0.2722
KOR 0.2261 0.1907 0.1964 0.2021 0.1870 0.2001 0.2123
LUX 0.3692 0.3376 0.2988 0.3197 0.3396 0.3325 0.2979
LVA 0.2202 0.1969 0.1873 0.2055 0.2235 0.2192 0.1989
MEX 0.1124 0.1248 0.1391 0.1234 0.1079 0.1070 0.1118
NLD 0.3604 0.3302 0.2892 0.3100 0.3374 0.3299 0.3151
NOR 0.4126 0.3749 0.3618 0.3607 0.3632 0.3808 0.3688
NZL 0.3348 0.2617 0.2389 0.2733 0.3098 0.3743 0.2925
POL 0.2972 0.3152 0.3016 0.2771 0.2665 0.2930 0.2700
PRT 0.3024 0.2793 0.2546 0.2735 0.2893 0.2912 0.2722
SVK 0.3190 0.2863 0.2466 0.2659 0.2878 0.2775 0.2398
SVN 0.3057 0.2959 0.2786 0.3141 0.3541 0.3482 0.3099
SWE 0.5305 0.4253 0.3607 0.3947 0.4481 0.4317 0.3961
TUR 0.3220 0.2141 0.1830 0.1939 0.1882 0.2124 0.1925
USA 0.2731 0.2628 0.2379 0.2288 0.2308 0.2451 0.2552
Average 0.3347 0.3021 0.2729 0.2839 0.3049 0.3103 0.2885
SD 0.0800 0.0711 0.0582 0.0651 0.0772 0.0756 0.0631
Table 13: Tax-to-GDP ratio

647
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2007 2008 2009 2010 2011 2012 2013

AUS 0.2765 0.3050 0.2268 0.2359 0.2599 0.2725 0.2862
AUT 0.3663 0.4442 0.4185 0.3723 0.4331 0.3978 0.4149
BEL 0.3887 0.4620 0.4246 0.3911 0.4568 0.4224 0.4425
CAN 0.3080 0.3536 0.2760 0.2803 0.3028 0.3094 0.3195
CHE 0.2251 0.2691 0.2504 0.2199 0.2818 0.2588 0.2596
CHL 0.2196 0.2226 0.1372 0.1698 0.1995 0.2049 0.2122
CZE 0.2747 0.3823 0.3210 0.2959 0.3661 0.3342 0.3436
DEU 0.3199 0.3880 0.3609 0.3183 0.3786 0.3438 0.3545
DNK 0.4197 0.4923 0.4487 0.4191 0.4720 0.4335 0.4465
ESP 0.3290 0.3495 0.3111 0.3003 0.3473 0.3164 0.3276
EST 0.2876 0.3862 0.3521 0.2797 0.3169 0.2904 0.3031
FIN 0.3737 0.4640 0.4155 0.3693 0.4482 0.4060 0.4319
FRA 0.3868 0.4576 0.4212 0.3883 0.4605 0.4229 0.4453
GBR 0.3517 0.3925 0.3067 0.3021 0.3286 0.3173 0.2930
GRC 0.2803 0.3321 0.3377 0.3331 0.3942 0.3638 0.3587
HUN 0.3487 0.4763 0.3890 0.3469 0.3994 0.3633 0.3661
IRL 0.2983 0.3305 0.2908 0.2509 0.2889 0.2591 0.2612
ISL 0.4700 0.4772 0.3083 0.3000 0.3533 0.3220 0.3199
ISR 0.2833 0.3331 0.2653 0.2751 0.3146 0.2645 0.2923
ITA 0.3843 0.4556 0.4326 0.3908 0.4603 0.4272 0.4361
JPN 0.2465 0.2581 0.2395 0.2452 0.2744 0.3392 0.3093
KOR 0.2776 0.2741 0.1961 0.2130 0.2377 0.2322 0.2248
LUX 0.3288 0.3977 0.3688 0.3316 0.3923 0.3526 0.3562
LVA 0.2442 0.3765 0.3066 0.2364 0.2780 0.2640 0.2751
MEX 0.1184 0.1606 0.1100 0.1188 0.1292 0.1221 0.1318
NLD 0.3231 0.3965 0.3626 0.3377 0.3867 0.3446 0.3599
NOR 0.3652 0.4943 0.3711 0.3606 0.4109 0.4043 0.4179
NZL 0.3531 0.3605 0.2537 0.2666 0.2939 0.2976 0.2998
POL 0.2785 0.4143 0.2865 0.2848 0.3367 0.3066 0.3072
PRT 0.2917 0.3404 0.3060 0.2959 0.3660 0.3043 0.3354
SVK 0.2513 0.3270 0.2869 0.2559 0.3009 0.2689 0.2949
SVN 0.3209 0.4018 0.3789 0.3454 0.4037 0.3546 0.3511
SWE 0.4269 0.5259 0.3879 0.3751 0.4404 0.4003 0.4329
TUR 0.2044 0.2750 0.1965 0.2307 0.2472 0.2299 0.2578
USA 0.2628 0.2622 0.2218 0.2266 0.2296 0.2330 0.2457
Average 0.3110 0.3725 0.3133 0.2961 0.3426 0.3196 0.3290
SD 0.0696 0.0847 0.0830 0.0671 0.0822 0.0710 0.0736
Table 13 (cont.): Tax-to-GDP ratio

648
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country 2014 2015

AUS 0.2971 0.2881
AUT 0.4940 0.4213
BEL 0.5258 0.4362
CAN 0.3598 0.3249
CHE 0.2793 0.2776
CHL 0.2093 0.1988
CZE 0.3681 0.3191
DEU 0.4239 0.3599
DNK 0.5674 0.4508
ESP 0.3875 0.3261
EST 0.3806 0.3260
FIN 0.5138 0.4279
FRA 0.5299 0.4465
GBR 0.3346 0.3511
GRC 0.4332 0.3682
HUN 0.4326 0.3775
IRL 0.2532 0.2205
ISL 0.3916 0.3032
ISR 0.3213 0.2948
ITA 0.5102 0.4249
JPN 0.3372 0.2731
KOR 0.2510 0.2459
LUX 0.4294 0.3647
LVA 0.3349 0.2845
MEX 0.1576 0.1727
NLD 0.4355 0.3645
NOR 0.5009 0.3991
NZL 0.3640 0.3051
POL 0.3653 0.3284
PRT 0.3945 0.3354
SVK 0.3591 0.3142
SVN 0.4193 0.3500
SWE 0.4908 0.4173
TUR 0.2670 0.2495
USA 0.2493 0.2552
Average 0.3820 0.3315
SD 0.0982 0.0702
Table 13 (cont.): Tax-to-GDP ratio

649
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Country Average
AUS 0.2811
AUT 0.4027
BEL 0.4191
CAN 0.3103
CHE 0.2533
CHL 0.1880
CZE 0.3134
DEU 0.3431
DNK 0.4444
ESP 0.3179
EST 0.2936
FIN 0.4110
FRA 0.4179
GBR 0.3160
GRC 0.3214
HUN 0.3587
IRL 0.2642
ISL 0.3459
ISR 0.3083
ITA 0.4047
JPN 0.2717
KOR 0.2230
LUX 0.3511
LVA 0.2532
MEX 0.1280
NLD 0.3489
NOR 0.3967
NZL 0.3050
POL 0.3080
PRT 0.3083
SVK 0.2864
SVN 0.3458
SWE 0.4303
TUR 0.2290
USA 0.2450
Table 14: Tax-to-GDP ratio: average per country

650
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Nominal Tax Rates

Corporate Income Tax Progressivity
Incentives
(CIT)
Tax Deductibility of Costs
Administration

Nominal Tax Rates

Personal Income Tax Progressivity Rate
(PIT) Personal Deductions
Social Security Countributions

Value Added Tax Standard Tax Rate

(VAT) Reduced Tax Rate and Exemptions
Registration Duty

Recurrent Taxes on Immovable Property

Recurrent Taxes on Net Wealth
Individual Property Taxes Estate, Inheritance and Gift Tax
(PRO) Taxes on Financial and Capital Transactions
Other Property Taxes

Beer
Other Taxes on Consumption Wine
(OTC) Alcohol
Tobacco
Mineral Oils

Figure 1 – WTI sub-indices and sub-components in Machová and Kotlán (2013)

651
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

652
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

THE DETERMINANTS OF UNEMPLOYMENT: A CASE OF SOUTH AFRICA

Priyanka Patel
Faculty of Economic and Management Sciences
Department of Economics
North West University, North West, South Africa
priyankabpatel01@[Link]

Ireen Choga
Faculty of Economic and Management Sciences
Department of Economics
North West University, North West, South Africa
[Link]@[Link]

ABSTRACT

The unemployment rate in South Africa is complicated by extremely serious socio-political

issues confronted by the South African government. As a result, this research investigates the
determining factors which cause unemployment in South Africa and proposes a policy
recommendation to alleviate unemployment in South Africa. The determinants of unemployment
are analysed using Johansen cointegration analysis technique from an econometric perspective
with annual time series data from 1986 to 2016. The Johansen cointegration test established that
there is a long-run relationship between unemployment and chosen variables. Diagnostic and
stability tests results show that the residuals behaved well. Impulse response outcomes were
steady with the long-run dynamic model. The outcomes of the variance decomposition showed
that most of the forecast error variance in unemployment rate is explained by government
expenditure on education, whereas limited proportion of variation was explained by population
growth and foreign direct investment in unemployment rate. The Granger causality test results
shows that variables such government expenditure on education, and foreign direct investment
provide unidirectional relationship and gross domestic product reinforce the inverse relationship
suggested by economic theory. The results that have arisen from this study confirm the
theoretical predictions and are also supported by previous researchers.

Keywords: unemployment rate, population growth, government expenditure on education,

foreign direct investment, gross domestic product, VECM, South Africa.

JEL codes: H52, E24, C22, E66

1. INTRODUCTION

Unemployment is a worldwide problem that every nation relatively struggle with. This
unemployment challenge has become so enormous in the worldwide economy that the protector

653
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of the worldwide economic framework, the World Bank and the International Monetary Fund
depends on the institutional unemployment figures as the catalogues of the socio-economic
virtuous fortune of countries (Mafiri, 2002).

South Africa faced a tragedic phase of global financial crisis after a very long period of 17 years
which made the economy to go through recession in the year of 2008 till 2009. Due to the
recession, the labour market of South Africa started to experience a plunge in production which
leads to reduce in demand for labour. The pressure of recession was felt by most of the
companies and experienced major financial problems, as a result companies end up retrenching
the individuals. The effect of retrenchment in labour distress subsequently unannounced strikes
occurred along these lines further bringing down the volumes of production in South Africa
(Steytler and Powell, 2010).

In general, unemployment significantly generates poverty, homelessness and nurtures family

disunity. According to Kingdon and Knight (2007), unemployment remains a genuine issue of
concern because it brings about overwhelming impact on economic welfare, crime, the erosion
of human capital, unhappiness and social insecurity. Unemployment also breeds some
psychological problems such as hopelessness, frustration hostility and some obvious activities
that unemployed youth go through the direction of criminal behaviour (Bakare, 2011). Moreover,
the unemployed also undergo psychological costs such as involuntary unemployment which
bring about a loss of confidence and self-esteem. The psychological disorders, divorces, suicides
and criminal activities tend to rise when the rate of unemployment increases, (Mohr, Fourie and
Associates, 2012).

Ever since the end of apartheid in 1994, the problem of unemployment remains determinedly
high in South Africa. The economy showed some improvement in 1995 and at some point it was
anticipated that unemployment would decline constantly. In spite of the fact that the South
African economy has been demonstrating adequate, positive improvement in the previous 20
years, development in employment remains generally slow. Growth in employment is not
sufficiently quick to absorb the new entrants of labour, thereby heightening the impact of this
challenge on the performance of the national economy generally (Hendriks, 2016).

By looking at all these facts, South Africa faces its highest unemployment rate. The several
attempts that have been made by various scholars in different nations to alleviate the problem of
unemployment seem to have failed because this has kept on increasing in every nation alongside
South Africa. This raises questions for all the economists and policy makers on how to, alleviate
the problem of unemployment in both developing and developed countries. Even government of
South Africa have also attempted by implementing policies to alleviate the problem of
unemployment but failed to do so.

The aims of the study are to investigate the determinants of unemployment in South Africa and
investigate the causal relationship between unemployment and its determining factors. The

654
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

objectives of the study are to analyse the impact of the determinants of unemployment on the
economy in South Africa, to utilize the granger causality test to determine the existence of a
causal relationship among the variables, empirically determine the factors that cause
unemployment in South Africa using econometric tools and make policy recommendations in
respect of unemployment. The remainder of the paper is as follows: the following section
highlight on the literature review, followed by the research method which includes data and
model specification along with methodology. The followed section deals with estimation and
analysis of results and lastly the concluding remarks.

2. LITERATURE REVIEW

This section discusses the theoretical framework on unemployment. The theories reviewed are
the classical theory, the Keynesian theory, and the monetarist theory of unemployment.

The Classical Theory of Unemployment

Classical economics principles derived from Adam Smith were developed in the 19th century.
The principal cause in the classical theory is that the economy is self-regulating. Classical
economists uphold that the economy is often accomplished in regard to attaining the natural level
of real gross domestic product; this can be obtained when the resources of the economy are fully
employed. According to this theory, the solution to unemployment is to reduce wages.

Keynesian Theory of Unemployment

Keynesian theory was introduced by British economist Keynes in his book The General Theory
of Employment, Interest, and Money issued in 1936 amid the great despair. This theory was
based on a criticism of the classical theory. In this criticism, Keynes disputed that savers and
investors have mismatched the procedures which might not guarantee that equilibrium happens
in the money market. The costs and remunerations have a tendency to be severe, which might not
contribute to equilibrium in the product and labour market. Due to this period of severe
unemployment, Keynes argued to prove the deficiencies in classical theory.

Monetarist Theory of Unemployment

According to Garrison (1984), Milton Friedman is known as the well-known monetarist and has
disagreed with the economics of John Maynard Keynes. This theory has a direct criticism of
Keynesian economics theory, not because of its implications for the way taken by
macroeconomic theory however as well because of its implications for economic policy. This
prompted the advancement of a counter unemployment theory known as the Monetarist theory of
unemployment.

1.1 Empirical Literature

655
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

This section will explores the research of national and international studies on determinants of
unemployment. A wide range of evidence derives from developing countries, whereas literature
from developed counties and South Africa is limited. Empirical literature is categorized by
nation (developed and developing), by number of years and by types of analysis used by studies
(surveys and econometric studies). Moreover, the empirical review in this study follows similar
concepts where literature review is categorized in developed and developing countries.

1.1.1 Literature from Developed Countries

Bakas and Papapetrou (2012) investigated the nature of Greek unemployment taking into
account cross-sectional dependence amongst Greek areas as well as the existence of structural
breaks. The quarterly data was used for unemployment for the 13 NUTS-II areas for the period
of 1998:Q1 to 2011:Q2 and all the data were obtained from the Hellenic Statistical Authority
(ELSTAT). The study employed recently developed as well as more powerful panel unit root
tests, for instance the Lagrange Multiplier (LM) used for to evaluate if regional unemployment
rates are subject to hysteresis behaviour. The outcomes of the study present some evidence in
favour of the hysteresis hypothesis for unemployment rate of Greece, even though this evidence
becomes weaker, if they change analysis in the panel of the 13 Greek areas. The outcomes
obtained from the empirical research specifically based on the panel root tests which account for
the way that unemployment in Greece is subject to a structural break for both mean and the slope
of the series, therefore the null hypothesis of the unit root was not rejected and indicated that the
Greek regional unemployment series are non-stationary with the existence of a structural break.
Tercek and Simmons (2014) studied the determinants of European and United States
unemployment by using annual data of every variable for France, Germany, Spain, and the
United States for the period 1990 to 2012. The aim of the study was to establish the relationship
between unemployment and a set of economic factors along with social indicators. The results
show that every one of the seven regression models was statistically significant, with a few of the
individual variables in each model. However, after running the seven regression models in this
study and realizing that there was a high level of correlation between a numbers of the variables,
this would likewise be beneficial to take out a portion of the variables from the models and run a
new regression model to limit the impact of these variables.

1.1.2 Literature from Developing Countries

The study was examined by O’Nwachukwu (2016) on the determinants of the unemployment
rate in Nigeria. The secondary time series data were utilized from the period of 1980 to 2016; all
the data was acquired from the International Monetary Fund (IMF), World Outlook Database and
the Central Bank of Nigeria (CBN) Statistical Bulletin. The examination used the Ordinary Least
Squares (OLS) technique for assessing the model and the ADF unit root test was likewise used.
The results reveal that government expenditure, inflation rate and population growth are
statistically important variables in enlightening the ups and downs in unemployment of Nigeria.
On the other hand, first lag of unemployment and actual GDP are not statistically important

656
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

variables in explaining the unemployment rate in Nigeria. This research recommends that
government needed to allocate greater quantity of money to capital expenditure in the budget,
and awarded plans should be monitored to see that they are accomplished. Lastly, the
examination likewise recommends that corrupt official that commits fraud money should be
punished as needs to and technologies that are required human labour to function ought to be
introduced.
Sahin (2016) explored the study by empirically analyzing the determinants of unemployment in
China. A time series data was utilized for the time period of 1982 to 2014 and every one of the
data were collected from the International Monetary Fund (IMF) and the World Bank.
Autoregressive Distributed Lag (ARDL) approach by Pesaran et al (2001) was employed to test
the determinants of unemployment in China. The result of the long-run estimation found a
negative and significant correlation among gross domestic product and an unemployment rate in
China. A positive and insignificant relationship was found among inflation rate and foreign
direct investment with unemployment rate. The outcome of the short-run shows that gross
domestic product; inflation and foreign direct investment have a negative and insignificant
relationship with unemployment rate. Furthermore, the result of the error correction model was
negative and statistically significant. The result of CUSUM and CUSUMSQ also shows that the
model was structurally steady within 5 percent critical bounds.

1.1.3 Literature from South Africa

Malakwane (2012) examined the study in view of the social and economic impact of
unemployment in South Africa. The study utilized both quantitative data that was gathered
information by the scholar through numerous sources and used existing data through different
sources for instance libraries, the internet, and observations. According to the information
received in this study suggest that there be existent relationships among unemployment, and
different factors, for example, crime, health, education as well as skills. Policies have been
introduced to alleviate the problem of unemployment and poverty such as Growth, Employment
and Redistribution (GEAR), Reconstruction and Development Programme (RDP) and Black
Economic Empowerment (BEE). The study recommends that more radical approach and policy
intervention is vital in order to address these challenges more efficiently. Features such as clear
rural development approach, education system should be reform, providing a beneficial business
environment for the private as well as for small, medium, and micro-sized enterprises, and lastly
government should make an enhancement in employment creation in order to alleviate the
problem of crucial skills shortages.

An examination was researched by Viljoen and Dunga (2014) on the determining factors which
influenced the female unemployment in a South African township, a case of Bophelong
Township. It employed a logistic regression model to examine the potential determining factor of
female unemployment. The study was based on a household survey by making use of
questionnaires and a sample of 300 households was surveyed but 295 questionnaires were

657
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

completed in July 2013. The outcomes acquired from the logistic regression model indicates that
household size, age, marital status, access to social grants and poverty status were found to be
important factors of either of one are utilized or not. Even though there are occupations in that
area, the quality is such that household can still be beneath the normal poverty line. Therefore,
further investigations are needed to be done and this would have policy allegations through
improving the amount of works as well as nature of occupation in the Bophelong Township.

The reviewed studies have made valuable contribution to the analysis and understanding the
unemployment phenomenon and coming up with various policy solutions to fight the problem in
South Africa. However, the studies done by developed countries and South Africa have
neglected the impact of government expenditure, foreign direct investment, economic growth,
and population growth on the unemployment rate yet these factors have a significant impact on
the resulting unemployment rate in a country especially in developing countries such as South
Africa. Government expenditure as an effect on the unemployment rate is supported by the
Keynesian Theory which suggested the increase of government spending in order to reduce
unemployment.

Moreover, most of the reviewed studies done by South Africa only focusing on factors such
crime, age, marital status and many more. Yet the studies fail to add the other factors that cause
unemployment in developed countries and South Africa. This study therefore incorporates these
aspects that have been neglected by other scholars in order to gain insight into the phenomenon
of unemployment on a broader spectrum and more specifically in South Africa.

3. DATA AND METHODOLOGY

The research used yearly time series information from the period of 1986 – 2016, with sample of
30 observations intended for estimating a regression model. All the data for the variables are
available and acquired from the International Monetary Fund (IMF) Bank, the World Bank, and
South African Reserve Bank (SARB). The study uses secondary data for analysis and all the
empirical estimations in this study are carried out using the time series econometric package
Eviews 9.

The model is adopted from Arslan and Zaman (2014). The intention of this research is to
investigate factors that cause employment rate in South Africa. In this examination, the
dependent variable is unemployment rate as influenced by the factors by other independent
factors that are population growth, government expenditure on education, foreign direct
investment, and gross domestic product. The study has modified this model by adding another
variable which is government expenditure on education. The study added this variable because
government spends a higher portion of its budget on education, but still the rate of
unemployment in the country keeps increasing with each passing year. Abdullah, Harun, and Jali
(2017), Government expenditure on education is also a way out to tackle the economic issues,
for example, unemployment, poverty, and economic recession. It is on the grounds that the

658
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

government expenditure on education could increase the number of skilled laboures in economic
sectors. Therefore, the modified model of the study is:

UNRt   0  1 PGt   2 GEEDU t   3 FDI t   4 GDPt   t .......... .......... .......... .......... .......... ...(1)
To get rid of the impact of outliners, every one of the variables is changed into the logarithm
form because logarithmic values reduce the tendency of fluctuation over time. Therefore, the
regression is calculated from the formula:

LUNR   0  1 LPG   2 LGEEDU   3 LFDI   4 LGDP  .......... .......... .......... .......... .......( 2)

A time series econometric technique called Vector Error Correction Model (VECM) is used in
this study for to examine the relationship between unemployment in South Africa and its
independent variables. The study uses VECM technique rather than Vector Autoregression
(VAR) model since VAR models are misspecified when series are cointegrated (Mah, 2012).
Before analyzing the determinants of unemployment, the Augmented Dickey-Fuller and the
Phillips-Perron unit root tests are utilized for examining the stationarity properties in regard to
the variables. Augmented Dickey-Fuller (ADF) test was developed from adaptation of the
Dickey-Fuller test for bigger arrangements of time series models. If the number is found to be
negative in the test, then the Augmented Dickey-Fuller (ADF) statistic is utilized. The Phillips-
Perron test is used as a non-parametric correction for managing each relationship within the error
terms and t-statistics. Augmented Dickey-Fuller (ADF) unit root test is based on following
equation (Chemma and Atta, 2014).
p
X t  yX t 1    i X t i   t .......... .......... .......... .......... .......... .......... .......... .......... .......... ......(3)
i 1

The Phillip - Perron (PP) unit root test is based on the model:

X t     t  X t 1   t .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... ....( 4)

Cointegration test is the econometric method that makes use of testing the correlation among
time series variables that are not stationary. At least two series are cointegrated once variables
are moving all together at similar wavelength (Gujarati and Porter, 2010). To perform the
cointegration test, we have to choose an appropriate lag for our variables in order to get credible
outcomes. This can be done by using unrestricted VAR.
The examination employs the Johansen cointegrations approach. According to Tetteh (2015), the
Johansen approach to cointegration test involves two test statistics: first test is the maximum
eigenvalue test (λmax) and the null hypothesis states that there is (r + 1) cointegrating vectors,
where the alternative hypothesis states there is (r) cointegrating vectors. The second test is trace
test, utilized to test the hypotheses that there are at most (r) cointegrating vectors. The trace and
Max eigenvalue statistics is determined by:

659
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

n
trace  T  In (1   r 1 )......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......( 5)
i  r 1

max (r , r  1)  TIn (1   r 1 )......... .......... .......... .......... .......... .......... .......... .......... .......... .......... ..(6)

Where, T denotes the size of the sample and  denotes the ith largest eigenvalue of matrix Π. The
vector autoregressive (VAR) model is a common structure that is utilized to depict the dynamic
interrelationship between stationary variables. This model presents restricted VAR designed
model aimed at utilizing non-stationary variables at level form, particularly recognized as
cointegrated. This one might be tested for cointegration by utilizing an expected VAR equation.
The vector error correction mechanism permits the series of long-term elements for complying
with equilibrium limitations, whereas a short-run element has flexible dynamic specification.
This essentially implies that an extent of disequilibrium as of one period will be rectified within
the following time frame (Mah, 2012).
The diagnostic and stability test should be performed with the goal that, the model as a final
point can be selected as a best fit in the sense that every one of the estimated coefficients have
the right signs, they are statistically significant based on the t-stats and f-tests, and furthermore
the value of R-square is rationally high (Gujarati, 2004: pp. 516). In this regard the study uses
normality test, heteroscedasticity test, Autocorrelation test, and AR roots graph. The stability test
decides whether the model of the study is steady, so as to be utilized for economic decisions and
also forecasts. Diagnostic and stability tests should be done on the short run equation.
Additionally to the VECM model, the impulse response analysis and variance decomposition are
further analysed in the short-run dynamics of the models. The utilization of these econometrics
methods allows the scholars to look at the impact of shocks on explained variables and the
relative importance of each shock within the whole framework (Sibanda, 2012). Brooks (2008)
observes that impulse responses plot the reactivity of the independent variables within the VAR
model to influence alternate factors.
Brooks (2002) observes that the variance decomposition analysis gives minutely differed
methods for testing the VAR framework processes. This analysis gives the division of the
dynamics in the explained factor caused by internal shocks that could be attributed to various
factors. A shock towards a variable directly influences the variable itself. That shock is
additionally passed on to other factors by the VAR framework elements. Variance
decomposition decides the amount of the error variance represented by advancements to each of
the explanatory variables. As a matter of fact, it has been observed that advancements of the
own-series accounted the most for the error variance of the series within the autoregressive
procedure.
The Granger Causality test is the last techniques used in this study. This test is done to determine
if one variable is causally related to another as a result allowing determining the capacity of one
variable to cause the other variable. In the bivariate case, the standard Granger causality test

660
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

amounts to testing whether past values of Y together with past values of X explains the current
change in X better than the past values of X alone will do. Failure to reject this null hypothesis
leads to the conclusion that Y Granger causes X. This process is repeated interchanging the two
variables (Y and X). Granger causality measures precedence and information content but does
not indicate the causality in the common use of the term. If a cointegration vector exists there is
causality amongst the variables at least in one direction (Sarbapriya, 2012).

4. ESTIMATION AND ANALYSIS OF RESULTS

This section provides estimated results through interpretation of findings and analyzing the data
obtained from steps presented in the previous section. All the results in this chapter are rounded
off to the nearest 3 decimal places. Descriptive statistics of variables used in this study and is
summarized in Table 1 below. The residuals from all of the variables in levels form do satisfy the
normality test. It is observed that the null hypothesis of residuals for LOGUNR, LOGPG,
LOGEEDU, LOGFDI and LOGGDP cannot be rejected at 5 percent of significance as indicated
by the high p-value of the Jaque-Bera statistics; therefore, we accept the null hypothesis that the
residuals from these variables are normally distributed. This indicate that in testing for stationary
of such variables, structural break and outliers will not had to be accounted for using
conventional unit root test.

Table 1: Descriptive statistics of variables used in the study at level form

LOGUNR LOGPG LOGGEEDU LOGFDI LOGGDP
Mean 3.112 0.499 2.964 2.940 7.652
Median 3.148 0.426 2.965 2.953 7.605
Maximum 3.325 0.917 3.091 3.142 8.030
Minimum 2.776 -0.081 2.868 2.719 7.315
Std. Dev. 0.157 0.259 0.060 0.107 0.248
Skewness -0.831 0.078 0.196 -0.151 0.216
Kurtosis 2.619 2.178 2.058 2.303 1.516

Jarque-Bera 3.758 0.904 1.344 0.746 3.086

Probability 0.153 0.636 0.511 0.689 0.214

Sum 96.477 15.485 91.878 91.136 237.223

Sum Sq. Dev. 0.739 2.010 0.106 0.341 1.848

Observations 31 31 31 31 31

Before analyzing, the Augmented Dickey-Fuller and Phillips-Perron unit root tests are utilized
for examining the stationarity properties in regard to the variables. The null hypothesis of both
tests indicates the variable have a unit root, while on the other hand the series will not have a unit

661
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

root stated by alternative hypothesis. The methodology section underlines the specified
estimation equation utilized when carrying out the unit root tests. In order to reduce the
variability of the series, all the series were transformed into logarithms (Khumalo, 2014). The
unit root tests performed are presented in Table 2 and 3 for ADF and PP tests respectively. From
these Tables, it is clear that all the variables are non-stationary at level form. Therefore the study
fails to reject the null hypothesis. Consequently, the variables should be tested for stationarity at
1st difference which is integrated of order 1 or I (1), in order to get strong results for every one
of the variables are stationary at all levels. The variables turn out to be stationary at 1% for both
tests as shown by results obtained in Table 3; this indicates its mean and variance are steady after
some time. Subsequently, we can proceed to cointegration test after the stationary test.

Table 2: ADF and PP Test Results at Level Form

ADF TEST PP TEST

Variables Model Specification ADF T-Value ADF P-Value PP T-Value PP P-Value
Intercept -2.295 0.180 -2.286 0.183
LUNR Trend and Intercept -2.919 0.171 -2.673 0.254
None 0.898 0.897 2.036 0.988
Intercept -1.483 0.528 -1.494 0.523
LPG Trend and Intercept -1.194 0.894 -1.255 0.880
None -1.171 0.215 -1.183 0.211
Intercept -2.279 0.185 -2.259 0.191
LGEEDU Trend and Intercept -2.203 0.471 -2.007 0.574
None 0.190 0.734 0.516 0.821
Intercept -2.406 0.149 -2.406 0.149
LFDI Trend and Intercept -2.662 0.258 -2.654 0.261
None -0.252 0.587 -0.304 0.568
Intercept -0.312 0.911 0.158 0.965
LGDP Trend and Intercept -2.347 0.397 -1.593 0.772
None 2.060 0.989 5.104 1.000
***statistically significant at 1%
**statistically significant at 5%
*statistically significant at 10%

Table 3: ADF and PP Test Results at First Difference

ADF TEST PP TEST

Variables Model specification ADF T-Value ADF P-Value PP T-Value PP P-Value
Intercept -4.803 0.001*** -9.350 0.000***
D(LUNR) Trend and Intercept -4.964 0.002*** -10.835 0.000***
None -6.125 0.000*** -6.215 0.000***
Intercept -6.220 0.000*** -6.157 0.000***
D(LPG) Trend and Intercept -6.272 0.000*** -6.209 0.000***
None -6.274 0.000*** -6.204 0.000***
Intercept -5.382 0.000*** -6.689 0.000***
D(LGEEDU) Trend and Intercept -5.339 0.001*** -8.611 0.000***
None -5.474 0.000*** -6.835 0.000***
Intercept -5.766 0.000*** -6.341 0.000***
D(LFDI) Trend and Intercept -5.588 0.001*** -6.045 0.000***

662
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

None -5.869 0.000* -6.473 0.000*

Intercept -2.783 0.073* -2.795 0.071*
D(LGDP) Trend and Intercept -2.694 0.246 -2.705 0.242
None -1.756 0.075* -1.660 0.091*
***statistically significant at 1%
**statistically significant at 5%
*statistically significant at 10%

Table 4 presents the confirmation of the lag lengths that are carefully chosen by various
information criteria and it indicates the information criteria approach produced agreeable results
as well as a decision to adopt lag 1. As a result, the Johansen cointegration test was conducted by
using lag 1 for the VAR. the selection of this lag is based on the lag order identified by LR:
Sequential modified LR test statistic (each test at 5% level), Final Prediction Error (FPE), Akaike
Information Criterion (AIC), Schwarz Information Criterion (SC), and Hannan-Quinn Information
Criterion (HQ).

Table 4: Lag Length Selection Criteria Results

Lag LogL LR FPE AIC SC HQ14 Conclusion

0 127.817 NA 1.44e-10 -8.470 -8.234 -8.396 Not good
1 264.438 216.709* 6.75e-14 -16.168 -14.754* -15.725* Good
2 292.001 34.216 6.69e-14* -16.3449* -13.752 -15.533 Not Good
*indicates lag order selected by the criteria

The trace test indicates two cointegrating vectors while Max-Eigen value does not identify any
cointegrating equation at the 5 percent significant level as shown in Table 5. The null hypothesis
of trace test states that the number of cointegrating equation is more prominent than the number
of variables included. Supposing that the test statistic value is smaller than the critical values of
the trace test and then the study fails to reject the null hypothesis. The null hypothesis of the
maximum eigenvalue test states that the number of cointegrating equation is (r) in contrast to the
number of alternative hypothesis of cointegrating equations in addition to one which is (r + 1). In
the event that the test statistic is smaller than the maximum eigenvalue test critical value then we
cannot reject the null hypothesis (Adrino, 2012). Lütkepohl et al. (2001) argues that the trace test
is stronger than the maximum eigenvalue test; therefore this research endorses the outcome of
the trace test and concludes there are two cointegrating equations. Therefore, we reject the null
hypothesis of no cointegrating equation and accept the alternative hypothesis of at most two
cointegrating equations, since the p-value is more than percent significant level. It implies that a
long-run relationship does exist among the factors.

Table 5: Cointegration Rank Test (Trace) and (Maximum Eigenvalue) Results

Hypothesized Eigenvalue Trace Statistics 0.05 Prob.** Conclusion

No. Of CE(s) Critical Value
None* 0.640 82.230 69.819 0.004 Reject Ho
At most 1* 0.597 52.637 47.856 0.017 Reject Ho

663
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

At most 2 0.443 26.279 29.797 0.121 Accept Ho

At most 3 0.264 9.317 15.494 0.337 Accept Ho
At most 4 0.015 0.438 3.841 0.508 Accept Ho
Trace test indicates 2 cointegrating eqn(s) at the 0.05 level
Hypothesized Eigenvalue Max-Eigen 0.05 Prob.** Conclusion
No. of CE(s) Statistic Critical Value
None 0.640 29.593 33.877 0.149 Accept Ho
At most 1 0.597 26.358 27.584 0.071 Accept Ho
At most 2 0.443 16.962 21.132 0.174 Accept Ho
At most 3 0.264 8.879 14.265 0.296 Accept Ho
At most 4 0.015 0.438 3.841 0.508 Accept Ho
Max-eigenvalue test indicates 0 cointegrating eqn(s) at the 0.05 level
The summary of VECM results are presented in Table 6 and 7 showing the long run and short
run equation respectively. If trace and Max-Eigen found more than one cointegrating equation, it
is not sensible to take the unrestricted estimated equation in β specifically as significant long-run
estimating parameters. It is vital to impose restrictions on the components of β in an endeavor to
acquire the essential relationship among the variables (Eita, 2007). Subsequently there are two
cointegrating vectors the VECM is visualized along these lines in the form of matrixes in
equation 7:

 11  12   LUNR t 1 
    LPG 
 21 22   t 1 
 31  32    11  21  31  41  51  61   LGEEDU t 1 
 t 1    t 1    .......... .......... .......... .......... ...( 7)
 41  42    12  22  32  42  52  62   LFDI t 1 
 51  52   LGDP 
t 1
   
 61  62  C t 

The long-run restriction is done within this investigation. Testing for the long-run factor helps to
find which variables ought to be included into the model and which ones should not be included
in the model (Eita, 2007). Four long-run restrictions were imposed on two cointegrating vectors
as appeared in Equation 8:
 11  12   LUNR t 1 
    LPG 
 21 22   t 1 
 31  32  1 0  31  41  51  61   LGEEDU t 1 
  t 1    t 1    .......... .......... .......... .......... (8)
0 0  0 1  32  42  52  62   LFDI t 1 
 51  52   LGDP 
t 1
   
 61  62  C t 

Where,  1 and  1 signify the parameters for long-run and short-run of the VECM respectively.
The study imposed restrictions in both the short and the long-run model accordingly, the
unemployment rate equation and population growth model specified. Summary of the restricted
long-run parameters outcomes are presented in Table 6 beneath.
Table 6: Results of Long-Run Cointegration Equation 1 and 2

664
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

EQUATION 1 EQUATION 2
Variable Coefficient Standard T-Statistic Coefficient Standard T-Statistic
Error Error
Constant 0.299 - - -10.015 - -
LUNR(-1) 1.000 - - 0.000 - -
LPG(-1) 0.000 - - 1.000 - -
LGEEDU(-1) -0.890 0.322 -2.762 3.132 1.400 2.237
LFDI(-1) 1.006 0.207 4.868 -1.380 0.898 -1.537
LGDP(-1) -0.487 0.065 -7.467 0.560 0.284 1.975

The standard system is applied in interpreting the result of VECM long-run model. A negative
coefficient is interpreted as a positive coefficient and the other way around. Therefore, the entire
coefficients in the model were multiplied by -1, (Lawana, 2016). For that reason, this study also
applied the same procedure. The effect of long run effect from the independent variables on
unemployment (UNR) as presented by Table 5 is shown by utilizing equation 9 and 10:
UNR  0.299122 0.890017GEEDU  1.00553FDI  0.487284GDP.......... .......... .......... .....( 9)
PG  10.01540  3.132391GEEDU  1.38001FDI  0.559982GDP.......... .......... .......... .......... .(10)

Eita (2007) deduces that the results of the first cointegrating vector is the most important than the
second cointegrating vector is not important. Therefore, this study focuses on first equation or
cointegrating vector which is the unemployment rate equation. That is because the focal point of
the study is on the unemployment rate.
Equation 9 shows that each one of the independent variables is statistically significant in
enlightening the dependent variable (unemployment) as the value of absolute t-value is above 2.
In the first cointegrating long-run equation, zero restriction is imposed on population growth
(PG) as it has now turned out to be the explained variable of the second equation. A one percent
increase in government expenditure on education (GEEDU) would result in a rise in
unemployment (UNR) by 89 percent. As per economic theory, this finding suggests that
government expenditure would reduce unemployment in an economy and the outcomes of the
model do not appear to be in line with this economic theory. Foreign direct investment (FDI) has
a negative long run relationship with unemployment (UNR). A one percent rise in FDI would
results in a drop in unemployment (UNR) by 100.6 percent. A one percent increase in GDP
would results increase in unemployment by 48.73 percent. The results correspond to economic
theory, which implies that an increase in GDP would reduce unemployment and a decrease in
GDP would increase unemployment. Equation 10, on the other hand shows restrictions imposed
on unemployment rate (UNR) in the second equation since it is an endogenous variable in the
first equation. The imposed restrictions indicate the actual variable does not play a role in the
determination of the explained variable.
According to the results obtained for cointegrating equation 1 in Table 6, the estimation of
equation is theoretically correct as the sign of the vector error correction model of D (LOGUNR)
is negative and reported as -1.125. This indicates the speed of adjustment is about 112.52

665
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

percent. The result proves the existence of a problem in the long-run equilibrium correlation
among the explanatory variables within South Africa therefore the VECM is well determined.
This implies that there is a deviation from equilibrium; just 112.52 percent are rectified in one
year while the variable moves to reestablishing equilibrium. This speed of adjustment is slightly
higher compared to those from previous studies on South Africa. The value of adjustment is a
relatively high-speed of adjustment to long-run equilibrium. This implies that short-run shocks or
disturbances in the unemployment rate would quickly move the economy towards the long-run
equilibrium. The speed of adjustment value is statistically meaningful and a value of t-stats is
negative at -3.908.

On the other hand the results obtained for cointegrating equation 2 in Table 6 shows, the
coefficient of D (LOGUNR) is negative and stated as -0.240. This indicates the speed of
adjustment is around 23.99 percent. This also suggests that if there is a deviation from
equilibrium, just 23.99 percent will be rectified in one year as the variable moves to
reestablishing equilibrium. The value of the speed of adjustment was found as statistically
meaningful with a negative value of t-stats about -3.599. However, study only focusing on
equation 1 not equation 2, as we stated above in the discussion.

Table 7: Results of Short-Run Cointegration Equation 1 and 2 (VECM)

COINTEGRATING EQUATION (1) COINTEGRATING EQUATION (2)

Variables Coefficient Standard Error T-Stats Coefficient Standard Error T-Stats
D(LOGUNR) -1.125 0.288 -3.908 -0.240 0.067 -3.599
D(LOGPG) 1.092 0.520 2.102 0.039 0.120 -0.322
D(LOGGEEDU) -0.074 0.275 -0.269 -0.049 0.064 -0.776
D(LOGFDI) 0.000 0.000 [NA] 0.000 0.000 [NA]
D(LOGGDP) 0.046 0.089 0.520 0.013 0.021 0.643
LR test for binding restrictions (rank = 2):
Chi – Square (2) 5.779104
Probability 0.055601

The diagnostic and stability tests are responsible for checking whether a model is correctly
specified along with the goodness off fitness. Table 8 and Figure 1 present the summary of
diagnostic and stability test. The interpretation of the tests is considered on the following criteria
when probability value is less than (p<0.05), we reject the null hypothesis and when the
probability value is more than (p>0.05), we accept the null hypothesis. This Table shows that the
estimated model has no serial correlation and heteroscedasticity as well as residuals are normally
distributed. Figure 1 depicts that all the modulus of the roots lies in the unit circle; therefore we
can determine that this model for South Africa is steady and the results obtained would be
reasonable.
Table 8: Summary Diagnostic Results

666
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Test Null Hypothesis T-Statistic P-Value Conclusion

Jarque-Bera There is normal 13.764 0.184 The residuals are normally
distribution. distributed.
White (Chi-Sq.) No conditional 244.608 0.051 There is no heteroscedasticity.
heteroscedasticity
Breusch-Godfrey No serial correlation 26.249 0.394 There is no serial correlation.
AR Roots Graphs Stable model n The model is stable.
a
r 1
1 1

Figure 1: Stability Test Results

Inverse Roots of AR Characteristic Polynomial
1.5

1.0

0.5

0.0

-0.5

-1.0

-1.5
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5

The next step is the impulse response analysis. Sims (1980), established that a shock to the
variable does not just impact the variable itself; the shock leads to all other independent variables
through the lag or dynamic structure of the Vector Autoregression (VAR). The impulse response
results are from VECM. As indicated in Figure 2 underneath, the response of LUNR to LUNR
has a decreasing positive impact on itself up to the 10 periods. This implies that a shock to
LUNR in the economy of South Africa brings about a decrease in LUNR.

Figure 2: Impulse Response Results

Response to Cholesky One S.D. Innovations Response to Cholesky One S.D. Innovations
Response of LOGUNR to LOGUNR Response of LOGUNR to LOGPG
.08 .08

.04 .04

.00 .00

-.04 -.04

-.08 -.08
2 4 6 8 10 2 4 6 8 10

667
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Response to Cholesky One S.D. InnovationsResponse to Cholesky One S.D. Innovations

Response of LOGUNR to LOGGEEDU Response of LOGUNR to LOGFDI
.08 .08

.04 .04

.00 .00

-.04 -.04

-.08 -.08
2 4 6 8 10 2 4 6 8 10

Response to Cholesky One S.D. Innovations

Response of LOGUNR to LOGGDP
.08

.04

.00

The response of -.04 LUNR towards LPG and

LUNR toward LFDI has a negatively increasing
-.08
impact over similar 2 4 6 8 10 time periods. This implies
shocks to LPG and LFDI would bring about LUNR to decrease. The outcomes of impulse
response shows a negative relationship amongst LUNR and LFDI, linking with the Johansen
cointegration results of long-run equation 1 that is discussed above. The response of LUNR to
LGDP was equally increasingly negative between second and third period then negatively
increased up to the 10th period. This implies shock to LGDP results in a decrease in LUNR, an
outcome that tallies with the Johansen cointegrating results of long-run cointegrating equation 2
that show above. Therefore, we conclude that LGDP and LFDI will have an impact on LUNR in
the long-run. .

The response of LUNR to LGEEDU has an increasing positive effect during the first two
periods. After these periods, decrease until fourth period and again it turns out to be gradually
positive until 10th period. This indicates that a shock to LGEEDU would make LUNR to rise.
The results of impulse response of a positive relationship among LUNR and LGEEDU links with
the outcomes of Johansen cointegration long-run equation 1, that is discussed above. Therefore,
we could conclude that LGEEDU will have an impact on LUNR in the long-run.

Table 9: Variance Decomposition of LUNR Results

Period S.E. LUNR LPG LGEEDU LFDI LGDP

1 0.060 100.000 0.000 0.000 0.000 0.000
2 0.111 34.919 22.952 34.367 6.562 1.110
3 0.151 22.204 32.342 33.003 8.367 4.084
4 0.176 18.935 33.195 32.887 9.576 5.407
5 0.199 16.886 33.576 33.644 10.100 5.794
6 0.224 15.205 33.401 34.962 10.298 6.134
7 0.248 14.103 32.412 36.146 10.174 6.593
8 0.272 13.332 32.412 37.225 9.938 7.093
9 0.296 12.748 31.905 38.099 9.670 7.578
10 0.318 12.288 31.484 38.773 9.433 8.023

668
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The variance decomposition analysis as presented in Table 9, a time period of 10 years with the
intention of finding out the impacts of the variables and their influence on the LUNR for a
comparatively longer period of time. The forecast error variance of LUNR is clarified through its
own innovations at 12.29% but mostly explained by 38.77 percent of LGEEDU, towards the end
of 10 years. According to Brooks (2002:342), all the variation in the dependent variable in the
first year is clarified by its own innovations or shocks. From Table 8 above, the results show that
from the second period up to the tenth period, LGEEDU is the variable that brings about the
maximum variation in LUNR.

The last step is Granger causality test and presented in Table 10 below. According to Gujarati
and Porter, (2009), Granger causality test is a statistical hypothesis test for determining whether
one series is useful in forecasting another. For instance, a time series X is said to Granger-cause
Y. Using T-tests and F-tests, the values of X can provide information about the future values of
Y. The results obtained from Granger Causality test indicate that causality does not occur
between LOGUNR and LOGPG in both directions. Variables such LOGGEEDU, and LFDI
provide unidirectional relationship and LOGGDP reinforce the inverse relationship suggested by
economic theory. Therefore, in order to reduce and possibly eradicate the unemployment level,
means to draw up and implement policies that will create an enabling environment for economic
growth that is required.

Table 10: Pairwise Granger Causality Test with Lags: 2

Hypothesis Obs F-Statistic Probability

LOGPG does not Granger cause LOGUNR 29 1.77981 0.1902

0.31463 0.7330
LOGUNR does not Granger cause LOGPG
LOGGEEDU does not Granger cause LOGUNR 29 11.0800 0.0004
0.27224 0.7640
LOGUNR does not Granger cause LOGGEEDU
LOGFDI does not Granger cause LOGUNR 29 2.64046 0.0919
0.00160 0.9984
LOGUNR does not Granger cause LOGFDI
LOGGDP does not Granger cause LOGUNR 29 1.09788 0.3498

LOGUNR does not Granger cause LOGGDP 3.60370 0.0428

α = 0.10 Decision rule == reject Ho if p-value<O .1 0

5. CONCLUSION

The intention of this research was to investigate the determining factors of unemployment in
South Africa, from the period of 1986 to 2016 utilizing time series annual data. The ADF and PP
tests of stationarity were used to identify the presence of unit root between the variables. Each
variable was non-stationary at level form; however after being differenced once, all the variables

669
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

become stationary. The Johansen Cointegration, VECM, variance decomposition and impulse
response techniques were employed for this investigation.

After the stationarity test was done, cointegration tests are needed to find out if there is any long-
run relationship between unemployment rate and other independent variables. The trace test
found two cointegrating vectors, whereas Max-Eigenvalue test found zero cointegrating vectors.
Since the study found more than one cointegrating vector, restriction on components was applied
on the long and short-run model. There are two cointegrating equations but only one equation is
considered in which the unemployment rate (UNR) model and a vector error correction model
was estimated for unemployment rate (UNR). The model was tested for diagnostic and stability
checks. The results of long-run model equation 1 showed that FDI negatively impacted on UNR.
GEEDU and GDP positively impacted on UNR. All the independent variables were statistically
significant explaining UNR. The vector error correction model indicates the estimation of
equation is theoretically correct as the sign of the vector error correction model of D (LOGUNR)
is negative and reported as -1.125 (112.25 percent). This speed of adjustment is slightly higher
compared to those from previous studies on South Africa. The value of adjustment is a relatively
high-speed of adjustment to long-run equilibrium. This implies that short-run shocks or
disturbances in the unemployment rate would quickly move the economy towards the long-run
equilibrium. The speed of adjustment value is statistically meaningful and a value of t-stats is
negative at -3.908.

All the diagnostic and stability tests reveal that residuals behaved well. Since all the tests results
are correct, therefore it could be concluded that our estimated model for South Africa is a good
predictor and may be utilized for further analysis. The outcomes of the Impulse response were
found to be reliable with the long-run dynamic model. Shocks on UNR, and GEEDU positively
impacted on UNR. Shocks on PG, FDI, and GDP generated a negative reaction on UNR. The
result of variance decomposition demonstrates that GEEDU is the most vital component in
explaining UNR in South Africa at the end of the period of this study. This likewise indicates,
shock toward the variables accounting for less variation in UNR in the short-run but on the long-
run, the shock toward GEEDU and PG add to great variation of the forecast error. Shock to FDI
and GDP add less towards the fluctuation in UNR, both in the short-run and the long-run. The
results obtained from Granger Causality test indicate that causality does not occur between
LOGUNR and LOGPG in both directions. Variables such LOGGEEDU, and LFDI provide
unidirectional relationship and LOGGDP reinforce the inverse relationship suggested by
economic theory

The policy recommendation is that government should provide grants and bursaries to students at
all levels (primary, secondary, and tertiary) for particular subjects, in order to encourage them to
perform better. Government should invest more in education and training to help reduce
structural unemployment. The government should adopt an expansionary fiscal policy that
spends money correctly towards measures set at reducing the level of unemployment.
Government should improve the skills levels as this accelerates growth, and promotes an

670
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

increase in the production level in the economy. South Africa could begin implementing policies
that encourage foreign direct investment, particularly alleviating macroeconomic factors and
promoting the appraisals of organisations. Government must likewise implement such policies to
entice and lure foreign investors to come to South Africa and invest. By doing this, it would
promote sustained economic growth which would generate more job opportunities within the
country. This study focused on the determinants of unemployment, an area considered the most
debatable and most concerning topic for researchers as well as for policy makers. However, this
study is not a conclusive one; therefore, there is a dire need for further research.

REFERENCES

Abdullah, M.B., Harun, M. and Jali, M.R.M., 2017. Employment Generated by Government Spending on
Education. International Journal of Academic Research in Business and Social Sciences, 7(2), pp.738-
742.

Adrino, M., 2012. The Effect of Foreign Direct Investment on Economic Growth: Evidence From South
Africa. University of Fort Hare.

Arslan, M. and Zaman, R. 2014. Unemployment and Its Determinants: A Study of Pakistan Economy
(1999-2010). Journal of Economics and Sustainable Development, Vol. 5, No. 13.

Bakare, A.S. 2011. The determinants of urban unemployment crisis in Nigeria: An Econometric analysis.
Journal of Emerging Trends in Economics and Management Sciences, vol. 2(3), pp.184-192.

Bakas, D. and Papapetrou, E (2012, june). Unemployment in Greece: Evidence from Greek regions.
(ISSN 1109-6691). Economic Research Department – Special Studies Division, Greece.

Brooks, C. 2002. Introductory Econometrics for Finance. 1st Edition. Cambridge University Press.

Brooks, C. 2008. Introductory econometrics for finance. 2nd Edition. Cambridge: Cambridge University
Press.

Cheema, A.R. and Atta, A. 2014. Economic Determinants of Unemployment in Pakistan: Co-integration
Analysis. International Journal of Business and Social Science. Vol 5, No.3, pp.209-221.

Eita, J.H. 2007. Estimating the equilibrium real exchange rate and misalignment for Namibia. (Doctoral
dissertation, University of Pretoria).

671
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Garrison, C.B. 1984. Friedman versus Keynes on the Theory of Employment. Journal of Post Keynesian
Economics, 7 (1): 114-127.

Gujarati, D. 2004. Basic Econometrics. India, McGraw-Hill Education

Gujarati, D. and Porter, D. 2010. Essentials of Econometrics, Fourth Edition (New York: McGraw-Hill).

Gujarati, D.N. and Porter, D.C. 2009. Basic Econometrics. Mc Graw-Hill International Edition.

Hendriks, J.F. 2016. Critical evaluation of possible policy options to reduce unemployment in South
Africa. (Masters thesis, University of Western Cape).

Khumalo, Z.Z. 2014. The Determinants of Unemployment in Swaziland: An Econometric Perspective and
Investigative Analysis (Doctoral dissertation, North-West University).

Kingdon, G. and Knight, J., 2007. Unemployment in South Africa, 1995–2003: causes, problems and
policies. Journal of African Economies, 16(5), pp.813-848.

Lawana, N. 2016. The impact of foreign direct investment on labour productivity of the automotive sector
in South Africa. (Doctoral dissertation, University of Fort Hare).

Lütkepohl, H. Saikkonen, P. and Trenkler, C. 2001. “Maximum eigenvalue versus trace tests for the
cointegrating rank of a VAR process”, The Econometrics Journal, pp.287-310.

Mafiri, M. 2002. Socio economic impact of unemployment in South Africa. MCom Dissertation. Pretoria:
University of Pretoria.

Mah, G. 2012. An econometric analysis of the eurozone sovereign debt crisis: the case of Greece
(Doctoral dissertation, University of North West).

Malakwane, C.T. 2012. Economic and social effects of unemployment in South Africa: Prospects for the
Future (Doctoral dissertation, Tshwane University of Technology).

Mohr, P., Fourie, L. & Associates. 2012. Economics for South African Students. 4th ed. Pretoria: Van
Schaik Publishers.

O’Nwachukwu, C.I. 2016. Determinants of the Rate of Unemployment in Nigeria. International Journal
of Information Research and Review, Vol. 04, Issue, 01, pp.3593-3595.

Sahin, D. 2016. Determinants of unemployment: Empirical analysis for China. The journal of academic
social science, 22(4), pp. 50-58.

Sarbapriya, R. 2012. Impact of Foreign Direct Investment on Economic Growth in India: A Co-
Integration Analysis. India: University of Calcutta.

Sibanda, B. 2012. Exchange Rates and Economic Growth in Emerging Economies: The Case of South
Africa (Doctoral dissertation, University of Fort Hare).

672
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Sims, C., 1980. Macroeconomics and reality., Econometrica, 48, 1–48. Mathematical Reviews
(MathSciNet): MR559536 Digital Object Identifier: doi, 10(00031305.1980), p.10482711.

Steytler, N. and Powell, D., 2010. The impact of the global financial crisis on decentralized government
in South Africa. L'Europe en Formation, (4), pp.149-172.

Tercek, D. and Simmons, W. 2014. Determinants of European and United States Unemployment. Senior
Honors Projects, pp.1-46.

Tetteh, C. 2015. Exports and Economic Growth: The Case of Ghana (Doctoral dissertation, University Of
Ghana).

Viljoen, D.J. and Dunga, S.H., 2014. Determining the Factors that influence female unemployment in a
South African township. International Journal of Social Sciences and Humanity Studies, 5(1), pp.63-72.

673
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The Moderating Role of Corporate Governance on the Effect of

Covid-19 Pandemic on the Saudi Corporate Profitability
Rawia Reda Obaid‫٭‬, Raed Reda Obaid‫٭٭‬
‫٭‬Associate Professor, Accounting Department, Faculty of Economic and
Administration, King Abdulaziz University, responding author; robead@[Link]
‫٭٭‬Assistant Professor, Manufacturing Engineering Department, Faculty of Engineering,
King Abdulaziz University

Abstract
This paper aims to study the influence of corporate governance in
moderating the impact of the covid-19 pandemic on the Saudi corporate
profitability. We hypothesize that; covid-19 pandemic negatively
influences Saudi companies‟ profitability, and corporate governance
moderates the influence of covid-19 on Saudi companies‟ profitability.
We analyzed a sample of 65 annual reports of companies listed in the
Saudi capital market (Tadawul) that covers the vast majority of firms for
the first quarters of 2019/2020 respectively. Results confirmed research
hypotheses as corporate governance minimized the influence of covid-19
consequences on Saudi companies‟ profitability.
1. Introduction
The corporate governance concept is a set of procedures that revolves
around achieving several internal and external goals in any organization,
when well utilized, governance always leads to increasing organizations'
value and profitability (Balios and Zaroulea, 2020; Sulaiman et al., 2018;
Cengiz, 2016). Internally, profitability would increase because applying
a well-structured corporate governance aids in managing and controlling
the organization in a legitimate, effective, and efficient way (Scherer and
Voegtlen, 2020; Zohrah and Hamidah, 2017). While externally,
implementing corporate governance results in the increase of
profitability as it guarantees a fair return on investment for outside
investors (Bhagat and Bolton, 2019).
In a survey of corporate governance implementation, Milosevic (2015)
found that corporate governance includes legal arrangements that
connect between large investors and a system that shields investors'
rights. The study results revealed that the corporate governance system
applied in the United States, Germany and japan, has more in common
than assumed. The benefits and importance of implementing corporate
governance could be recognized generally and locally. Generally, it
embraces an efficient and effective scheme that helps organizations in

674
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

reaching their objectives, monitoring and rewarding executive actions

and performance. Further, it aids in selecting the board committee, which
protects the shareholders‟ interests by maximizing their wealth and
regulating managers' behavior. Finally, corporate governance alleviates
the conflict of interest between shareholders, management, and board of
directors (Murhadi, 2009; Minow and Monks, 2004). Locally, the
Council of the capital market authority (CMA) in Saudi Arabia has
announced the modified edition of the corporate governance regulations
in 2019 that was issued earlier and is, by law, mandatory to be
implemented by Saudi companies that trade their shares in the Saudi
capital market1.
The effect of corporate governance on increasing firms' profitability has
attracted a wide spread of researchers around the globe to investigate
either the direct or the moderating influence of corporate governance on
firm‟s performance. On the direct level, For example, Balios and
Zaroulea (2020) examined the effect of corporate governance and
internal audit on firms‟ profitability in four countries (Portugal, Italy,
Greece, and Spain), using a sample of 74 firms during the 2011-2016
period. They found a positive impact of increasing the board size on
profitability. In Asia, Anwar et al. (2019) examined the direct
relationship between corporate governance practices and firm
profitability for the agricultural sector companies of Asian countries (20
countries) for the period of 2008 to 2017. The results indicated that
board independence, ownership concentration, and audit committee
independence, had a positive impact on firms‟ profitability.
Locally, in Saudi Arabia, Gerged and Agwili. (2020) analyzed the direct
impact of corporate governance mechanisms on the firm market and
accounting value. They examined a sample of 60 annual reports of
financial and non-financial companies listed in Saudi capital market
from 2012 to 2016. The results suggested that, averagely, better-
governed firms tended to attain better market value, but not necessarily
better profitability.
On the moderating level, Iryadi et al. (2019), for example, analyzed the
effect of tax planning and asset use on profitability moderated by
corporate governance in 75 listed manufacturing companies in Indonesia

1
[Link]

675
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

in the period 2014-2018. The results revealed that the use of assets
moderated by good corporate governance has an effect on profitability.
Whereas, Lu et al. (2015) from a sample of 225 family firms,
investigated the differential effects of internationalization on growth and
profitability in China. They found that internationalization had a positive
effect on growth yet, a negative effect on profitability. They argued that
the positive moderating effect of corporate governance, have highlighted
the need for appropriate corporate governance to support the
implementation of internationalization. While, in Saudi Arabia, Hamdan
et al. (2017) examined the moderating effect of corporate governance on
the relationship between intellectual capital efficiency and financial,
operational and market performance. They used data of 171 firms listed
in the Saudi capital market in the period 2012-2014. The results
indicated that corporate governance played a positive moderating role on
the interaction between intellectual capital elements and financial,
operational and market performance.
Yet, the covid-19 pandemic is considered the greatest challenge the
world is facing since World War Two. The disease is caused by the most
recently discovered coronavirus, the pandemic is much more than a
health crisis, and it has the potential to create devastating social,
economic and political disasters that would leave deep scars to any
nation it touched. Every day, people are losing jobs and income, with
no way of knowing when normality will return 2.
Since the disease started in December 2019, up to now, several
research papers have examined the effect of the pandemic on different
aspects of the national, international economy and private sectors (see
Table 1). For example, Ding et al. (2020) examined the extent to which
companies have immunity to the covid-19 pandemic in international
companies in Europe. Using data during the first quarter of 2020, they
evaluated the link between the stock price reactions to covid-19 and five
corporate characteristics (financial conditions, international supply chain
and customer locations, CSR activities, corporate governance, and
ownership structures). They found that companies with special
characteristics had more immunity to the covid-19 pandemic, as the
share prices did not fall as fast as similar to companies with less

2
Covid-19 pandemic. 2020. United nation development program.

676
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

immunity. Those characteristics were; solider financial conditions (high

liquidity, less dept., more profitability), less exposure to covid-19
through their international supply chains and customer locations,
stronger CSR policies and activities, embedded executives performance,
and finally, firms owned by non-financial corporations.
In the same line of research, Aifuwa et al. (2020) examined the influence
of the coronavirus pandemic on firms' performance in Nigeria. Using an
online survey questionnaire, three hundred and eighty-four business staff
respondents, they examined the influence of covid-19 pandemic (social
distance, boarder closure, lock down, non-essential travel avoidance,
usage of Hydrochloroquin to treat covid-19) on the financial
performance (costs, resource spectrum, economic recession, company's
turnover) and the non-financial performance (work from home,
monitoring, workforce reduction, business image). The study results
showed a decrease in the financial and non-financial performance in the
private companies due to the effect of the coronavirus.
A third example of examining the effect of coronavirus on companies'
performance is the study by Secinaro et al. (2020). They investigated the
influence of 2003 SARS epidemic on the financial performance of
SMEs, to highlight how the coronavirus would impact on financial
statements of such companies. By using financial data from the balance
sheets of three publicly listed European SMEs operating in the tourism
sector in 2002 and 2003, they examined the influence of SARS epidemic
on the financial performance (turnover, total assets, capital expenditure,
total equity and liabilities, net profit, earnings before interest, tax,
depreciation and amortization and cash flow from operating activities).
The authors found that the impact was negligible, because of the short
period of time that SARS lasted (six months) they suggested that the
effect of covid-19 would be modest if the virus remained for a short
period.

677
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1: Covid-19 effect on private sector studies (alphabetically)

No. Author and Study aim Study results
year
1 ACCA, 2020 Analyze the challenges and Highlighted some of the practical challenges
considerations for the impact of covid- and key considerations for auditors by referring
19 on audit and assurance to the relevant international auditing standards
in light of the covid-19
2 Aifuwa et al., Effect of coronavirus pandemic on A decrease in the financial and non-financial
2020 firms performance in Nigeria performance in the private companies
3 Avdiu and Effect of covid-19 (social distancing) Negative distributional impacts on lower-
Nayyar, 2020 is likely to affect face to face jobs paid and female workers.
4 Baur and Impact of the coronavirus pandemic on Large differences across firms and sectors
Trench, 2020 Australian companies‟ share prices affected by the virus.
5 Broadstock et the role of environmental, social and The role of environmental, social and
al., 2020 governance performance during governance performance is important during
the covid-19 global pandemic crisis.
6 Ding et al., The corporate immunity to the covid- Companies with specific characteristics have
2020 19 pandemic more immunity to the covid-19 pandemic
7 Egyptian The covid-19 influence on the A decrease in the informal economy and the
Center for Egyptian economy GDP
Economic
Studies, 2020
8 Khoo and The economic impact of covid-19 Workers' sense of insecurity due to the risks
Lantus, 2020 exposed to when operating
9 Coibion et al., The effect of covid-19 on Increase in the unemployment rate and
2020 employment rate and unemployment employees who lost their jobs, not willing to
find alternative jobs.
10 Secinaro et Examine the accounting impact of The accounting impact on European SMEs
al., 2020 coronavirus on small and medium would be immense if the virus persisted for
firms in Europe many months.
11 Spurk and Influence of covid-19 on flexible Increase in flexible and part time employment
Straub, 2020 employment relationships and careers in the health sector versus other sectors

In Saudi Arabia, before the occurrence of the first Covid-19 infected

case in March 2020, the government has announced and continued to
take several precautionary procedures (Table 2) to control and
minimize the spread of the virus.

678
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Private sector precautious procedures in Saudi Arabia3

Date Cautious measure
th
February 27 Suspending abroad Umrah
March 14 th Suspending international flights
March 14 th Closure of gymnastic and leisure centers
th
March 15 Closure of shopping malls and all shopping activities
except supermarkets and pharmacies
March 18 th Suspending private sector employees to attend in offices
and imply online working

Therefore, with the implementation of the procedures above, and

following the line of research (for example, Aifuwa et al., 2020; Baur
and Trench, 2020; Ding et al., 2020) that examined the economic and
social impacts of the covid-19 pandemic. Taking into account the
importance of the corporate governance regulations announced and
adjusted by the CMA authority in Saudi Arabia in 2019, which is
consistent with the positive effect of corporate governance on
profitability (for example, Balios and Zaroulea, 2020; Gerged and
Agwili, 2020; Sulaiman et al., 2018; Cengiz, 2016; Lu et al., 2015). It is
important to examine to what extent the Saudi companies that trade
their shares in the Saudi capital market are influenced by the covid-19
pandemic. Consequently, this paper aims to investigate the influence
of corporate governance on the profitability of Saudi companies with the
presence of covid-19 (see Fig. 1). This will be achieved through the
following two objectives
a) Examine whether the profitability of the Saudi companies is
affected by the consequences of covid-19
b) Examine whether corporate governance moderates the
influence of covid-19 pandemic the profitability of Saudi
companies

3
[Link]

679
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Corporate
Governance

Covid-19 Saudi
Pandemic Corporate
Consequences Profitability

Fig. 1: The conceptual framework underlying the current study

2. Methodology
2.1 Population and sample
The targeted population in the current study are all Saudi companies
listed in Tadawul (Saudi capital market). The list of companies covers a
wide range of sectors for example energy, materials, capital goods,
commercial and professional services, transportation, consumer durables
and apparel, consumer services, media and entertainment, retailing, food
and staples retailing, food and beverages, healthcare equipment and
services, pharma and biotech, telecommunication services, software and
services, and utilities4.

The total number of companies listed in Tadawul website is 200

companies operating in the capital market (required to implement
corporate governance regulations). As the population of the current study
was targeted to include all Saudi companies that are presumed to
implement corporate governance, it enables such variable to being
measured, as a whole, in connection with the profitability within the
chosen companies during the covid-19 pandemic. However, we excluded
the following from the research population:
a) Companies operating in sectors that are beyond the scope of the
current research; thirty-two companies in the financial sector
including banks (eleven), investment funds (seventeen), financial
firms (four) and thirty-two companies are operating in the
insurance sector.

4
[Link]

680
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

b) Companies that their financial statement did not include needed

information for the current research; recently joined the capital
market and their 2019 financial reports covered six months
(three), did not upload their 1st quarter 2020 reports (five), or
temporarily or permanently suspended firms (four).
Considering the above and in agreement with the current research aim
and objectives, the sample of the study is 65 Saudi companies (52% of
research population) randomly selected and represented all industrial,
merchandising and service sectors trading in the Saudi capital market.
2.2 Data collection instrument
The list of companies in Saudi Arabia were obtained from Tadawul
([Link]), which is a continuously updated website where all the
Saudi Stock Exchange‟s share price information, quarter and annual
financial reports, dividend news, bonus/rights issue news for all the listed
companies were provided. From the list of companies in Tadawul the
website acquired, a validation and double confirmation, this is made to
ensure the reliability of the information.
Further, a survey questionnaire is designed, a 5-point Likert scale,
ranging from 1 (totally disagree) to 5 (totally agree), to measure the
independent variable (covid-19) effect and collect data for the subject
matter under investigation (Aifuwa et al., 2018). We divided the
questionnaire into four sections; the first included general information
from respondents, and the following three divisions represented, in
agreement with the current research aim, covid-19 measuring items. A
copy of the questionnaire is electronically sent to one representative of
the companies respectively; the suitable sample size for the current study
reached 65 managers of companies who should have the capability
within the firm, in addition to their appropriate knowledge and expertise
in both the operation and direction of the companies.
2.3 Study variables and hypotheses
To fulfill the aim of the current study, the proposed framework shows
specific items related to the moderating role of corporate governance
on the impact of covid-19 on Saudi companies' profitability that can
be demonstrated through the results of the Saudi companies trading in
the Saudi capital market and make recommendations that can be
utilized through the empirical study. A brief description of the measure
and source of measurement of the study variables (Table 3) is provided,

681
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

followed by a detailed explanation of research variables, followed by

specifying research hypotheses (Fig. 2).
Table 3: Measure and source of study variables
Independent variable (covid-19)
Name Measure Source
Increase labor costs 5-point Likert scale ranged from Aifuwa et al. (2020)
Increase material costs totally disagree (1) to totally
Decrease sales agree (5)
Moderator (corporate governance)
Name Measure Source
Board size Between 3 and 11 members The corporate
Duality Majority non executives governance
Independence Minimum 2 or 1/3 of members, regulations in Saudi
whichever is greater Arabia (2019)5
Dependent variable (profitability)
Name Measure Source
ROE Net income/shareholders‟ equity Gerged and Agwili
NPM Net income/total income (2019)
ROA Net income/total assets

2.3.1 Covid-19
The covid-19 disease imposes a threat to human beings, it causes illness
ranging from the common cold to more severe diseases, and some of the
common symptoms include fever, cough, and shortness of breath, which
can sometimes develop into pneumonia or cause severe complications in
persons with immunodeficiency6.
The covid-19 pandemic is a non-numeric variable and its effect on
different aspects of the social, economy, and companies is not equal; it is
an ordinal variable7. Consequently, social science authors measured the
covid-19 variable using different methods. For example, some have
compared either the income of different sectors before and after the

5
[Link]
6
[Link]
7
Introduction to SAS. UCLA: Statistical Consulting Group. From
[Link]
features-of-sas/ (accessed August 22, 2016).

682
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

pandemic existed (Egyptian Center for Economic Studies, 2020), or

specific items listed in the balance sheet for two years (Secinaro,
2020). Others have calculated the covid-19 by computing the number
of confirmed affected cases by the virus (Ding et al., 2020). Finally,
some have measured the covid-19 variable through a 5-Likert scale
questionnaire (Aifuwa et al. (2020); Coibion et al., 2020).
In the current research, we measured the covid-19 by designing a 5-
Likert scale questionnaire (ranged from totally disagree to totally agree).
We pulled three items from the accounting literature to measure the
covid-19 impact, on the ground that they are essential elements of
determining the net income (Bhimani et al., 2019). The three items are
labor costs, material costs, and sales, further, we linked each of these
items to four precautious procedures announced by the Saudi
government such as suspending travel abroad and the total lock down.
2.3.2 Corporate governance
In the literature, the moderating influence of corporate governance was
measured through several indicators. For example, some have examined
the effect of corporate governance through the board size, board
independence, family ownership, foreign and institutional ownership
with controlling effects of size, market capital and liquidity position
(Busru and Shanmugasundaram, 2017). They examined a sample of 255
companies listed in the Indian capital market for the period of 2008-
2015. The results indicated that only family ownership moderated the
relationship between investing in R&D and profitability. Corporate
governance practices among other researchers included board structure,
ownership structure, and transparency and disclosure with controlling
effects of firm size, firm growth and risk (Iqbal and Javed, 2017). They
analyzed 173 manufacturing firms during the period of [Link]
results indicated that corporate governance moderated significantly and
positively the interaction between capital structure and financial
performance. Others have examined the moderating role of CEO duality
on the impact of geographic diversification on firm performance in the
US lodging industry (Song and Kang, 2019). Their sample included 258
firms for the period of 1990-2015. The results indicated a positive and
significant moderating role of CEO duality on the geographic
diversification-firm performance.

683
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In the current research, we measure corporate governance by three

indicators; board size, board independency, and duality role of board
committee members. We measured the application of each item
according to the corporate governance regulations (Article 16 and 17;
p.17) in Saudi Arabia8:
“The following shall be taken into consideration when composing the
Board: 1) the number of its members shall be suitable for the size and
nature of the Company's activities… (The Company’s bylaws shall
specify the number of the Board members; such number shall not be less
than three and not more than eleven). 2) The majority of the Board
members shall be of Non-Executive Directors. 3) The number of
Independent Directors shall not be less than two members or one-third
of the Board members, whichever is greater”.

2.3.3 Profitability
Accounting and financial textbooks suggested several ratios to measure
firm‟s profitability such as, profit margin, return on assets, return on
shareholders‟ equity, and payout ratio (Kimmel et al., 2018). In the
current research, we evaluate profitability by three ratios; return on
shareholders‟ equity (ROE), net profit margin (NPM), and return on
assets (ROA).
In the literature, the extent to which profitability was affected by
corporate governance was measured through either one ratio or more
than one ratio at a time. Examples of studies that measured profitability
through more than one ratio are the study by Buallay et al. (2017). They
measured the influence of corporate governance on 171 companies‟
performance of listed companies in Saudi capital market for the period
from 2012 to 2014. Companies‟ performance was measured using ROA,
ROE and Tobin's Q. They found that ownership and the size of the board
of directors significantly influences firm's performance. Another
example is the Cengiz (2016) study, where five indicators; earning per
share, return on assets, market book value, net profit margin, and return
on equity measured profitability in 44 companies listed in the Borsa
Istanbul in 2012. The results revealed that companies that were in the
corporate governance index had significant higher profitability than the
ones that were not indexed. Examples for studies that measured

8
[Link]

684
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

profitability through one ratio is the study by Sulaiman et al. (2018). The
study investigated the impact of corporate governance on profitability
(return on assets) of listed companies in Nigeria for the period 2010-
2017. The results indicated that board size, women on board, board
meetings and firm size positively influences the return on assets.
Following the same line of research, Zorah and Hamidah (2017)
examined the effect of corporate governance on increasing firm
performance (return on assets) in Indonesian companies 2005-2014. The
results revealed that audit committee meeting and audit quality have
positively influenced profitability.

Based on the above, and in agreement with research scope, aim and
objectives, the relationship between variables of the current research can
be indicated through the following hypotheses (see fig. 2).
h1: covid-19 pandemic negatively influences Saudi companies’
profitability
h2: corporate governance moderates the influence of covid-19 on Saudi
companies’ profitability

Independent Moderator Dependent

Variable Effect Variable

Covid-19 Corporate Saudi

pandemic Governance corporate
consequences (h2) profitability
Increase labor
ROE
costs

Increase (h1)
material costs NPM

Decrease
sales ROA

Control variable: size of Firm

Fig. 2: research hypotheses and variables

685
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3. Results
3.1 Reliability Analysis
We measure the reliability through the Cronbach‟s Alpha test, the value
of Alfa ranges between zero and one, if there is no stability in the data,
the value of this parameter equal to zero. On the contrary, if there is
complete stability in the data, the value of this parameter equal to one,
and if this measure has equal or increased 0.60, then it is possible to rely
on the results of the study.
Table 4: The reliability measures on the sections of the questionnaire
Alfa Honesty
Section Description # of questions )*(
Reliability Coefficient Coefficient
B Increase labor cost 4 0.923 0.961
C Increase material cost 4 0.678 0.823
D Decrease sales 5 0.909 0.953
(*) Honesty coefficient is the square root of reliability coefficient
Results in Table 4, confirms reliability of questionnaire; the value of
Alfa is ranged between 0.678 and 0.909.
3.2 Internal consistency
Internal consistency is calculated by computing the Spearman
Correlation Coefficient between each of the survey phrases and the
general mean of the dimension to which that phrase belongs (Table 5).
Table (5): Correlation between questionnaire phrases and mean of dimension
Phrases Increase labor cost
Safety requirements reduced number of workers at workplace .915**
Lock-down hours reduced job efficiency .893**
Suspending travel abroad minimized number of workers .833**
Not being able to dismiss workers, increased labor costs .902**
Increase material
Phrases
cost
Safety requirements reduced units produced .623**
Lock-down hours increased inventory costs .731**
Suspending travel abroad limited abroad material purchase .769**
Usage of more expensive options to buy material .677**
Phrases Decrease sales
Safety requirements reduced number of clients being served .880**
Lock-down hours minimized number of customers/ shoppers .879**
Suspending travel abroad reduced abroad trade .859**
Customers‟ shortage in online buying knowledge reduced sales .928**
**. Correlation is significant at the 0.01 level (2-tailed).

686
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Results of Spearman Correlation Coefficient test (Table 5) confirm

internal consistency between each of the survey phrases and the general
mean of the dimension to which that phrase belongs for all three items.

3.3 Descriptive Statistics

3.3.1 Independent variable
To measure the consequence of covid-19 pandemic in terms of the
increase in labor cost, increase in material costs, and decrease in sales,
according to the responses, we conducted descriptive statistics (mean,
relative mean, Standard deviation, coefficient of variation as follows.

Table 6: results of descriptive statistics

Phrases Mean Relative Std. C.V(2) Agreement Order
mean (1) Deviation (3)

B Increase labor cost (Overall

3.72 74.4 1.04 28.0 Agree
mean)
B1 Safety requirements reduced
3.91 78.2 1.26 32.2 Agree 1
number of workers at workplace
B2 Lock-down hours reduced job
3.69 73.8 0.97 26.3 Agree 2
efficiency
B3 Suspending travel abroad
3.66 73.2 0.97 26.5 Agree 3
minimized number of workers
B4 Not being able to dismiss
3.60 72.0 1.36 37.8 Agree 4
workers, increased labor costs
C Increase material cost (Overall
3.88 77.6 0.83 21.4 Agree
mean)
C1 Safety requirements reduced
3.86 77.2 0.98 25.4 Agree 3
units produced
C2 Lock-down hours increased
3.95 79.0 1.15 29.1 Agree 1
inventory costs
C3 Suspending travel abroad limited
3.88 77.6 1.26 32.5 Agree 2
abroad material purchase
C4 Usage of more expensive
3.85 77.0 1.24 32.2 Agree 4
options to buy material

687
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Phrases Mean Relative Std. C.V(2) Agreement Order

mean (1) Deviation (3)

D Decrease sales (Overall mean) 3.83 76.6 1.09 28.5 Agree

D1 Safety requirements reduced
3.92 78.4 1.28 32.7 Agree 2
number of clients being served
D2 Lock-down hours minimized
3.78 75.6 1.21 32.0 Agree 3
number of customers/ shoppers
D3 Suspending travel abroad
3.92 78.4 1.14 29.1 Agree 1
reduced abroad trade
D4 Customers‟ shortage in online
3.69 73.8 1.29 35.0 Agree 4
buying knowledge reduced sales
(1) Relative mean = mean / 5 * 100.
(2) C.V (Coefficient of variation) = Standard deviation / mean * 100.
(3) Agreement: we weighted the responses of the sample according to the
Likert scale ordinal quintile as follows
Totally disagree Disagree Neutral Agree Totally agree
1 2 3 4 5
We distributed the range (5-1=4) into five levels, every level=4/5=0.8,
then the weighted mean degree of agreement ranges as follows
Totally disagree Agree Neutral Agree Totally agree
1-1.79 1.80-2.59 2.60-3.39 3.40-4.19 4.20-5

Results shown in Table 6 indicates that 72% of responses agree that

labor costs increased (mean = 3.72, coefficient of variation = 28%),
while 78.6% of responses agree that material costs increased (mean =
3.88, coefficient of variation = 21.4%) and finally 71.5% of responses
agree that sales have decreased (mean = 3.83, coefficient of variation =
28.5%). Accordingly, it can be confirmed that consequences of covid-19
pandemic emerged through the increase in labor and material costs and
in the decrease of sales in Saudi companies.

3.3.2 Control variable

The control variable in the current study is the size of the company,
which is measured by the natural log of total assets (Table 7).

Table (7) descriptive statistics for company size

Std.
Variables Range Minimum Maximum Mean
Deviation

Company size 8.26 18.64 26.91 21.88 1.60

688
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The results (Table 7) reveal that the mean of company size which is
measured by the natural log of total assets is 21.88 i.e. the mean of total
assets is SR3,179,532,467, while the standard deviation is 1.6, the
natural log reinforces homogeneity and reduces any conflict, since the
minimum value is 18.64 and the maximum value is 26.91.
3.3.3 Dependent variables
Results of descriptive statistics show to what extent there is variation in
profitability indicators in research sample (Table 8).
Table 8: Descriptive statistics for corporate profitability
Variables Range Minimum Maximum Mean Std. Deviation
ROA 17.89% -5.50% 12.40% 0.97% 2.13%
ROE 43.18% -27.21% 15.97% 1.27% 5.19%
NPM 683.14% -402.88% 280.26% 4.59% 66.04%
Results show that there is variation in research sample regarding ROA
(mean= 0.97%, standard deviation= 2.13%), ROE (mean= 1.27%,
standard deviation= 5.19%) and NPM (mean= 4.59%, standard
deviation= 66.04%).
4. Hypotheses testing results
h1: covid-19 pandemic negatively influences Saudi companies’
profitability.
To test this hypothesis; we first conducted correlation analysis (Table 9),
followed by stepwise analysis (Table 10) as follows.
Table 9: Person correlation matrix between research variables
labor cost

Decrease
Increase

Increase
material

C Size
NPM
ROA

ROE
sales
cost

Variables

Increase labor cost 1

Increase material cost .456** 1
Decrease sales .632** .532** 1
ROA -.691-** -.423-** -.616-** 1
ROE -.466-** -.245-* -.494-** .840** 1
** *
NPM -.325- -.308- -.294-* .630** .479** 1
C Size .080 -.212- -.057- .088 .166 .221 1
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).

689
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Results of correlation analysis (Table 9) show that there is a significant

negative relationship between "Increase labor cost" and both "ROA",
"ROE", "NPM" since the person correlation coefficient is -0.691, -
0.466, -0.325 respectively at 0.01 level. It further show a significant
negative relationship between "Increase material cost" and both "ROA",
"ROE", "NPM" since the person correlation coefficient is -0.423, -
0.245, -0.308 respectively at 0.01 and 0.05 levels. Finally, the results
show a significant negative relationship between "Decrease sales" and
both "ROA", "ROE", "Net profit margin" since the person correlation
coefficient is -0.616, -0.494, -0.294 respectively at 0.01 and 0.05 levels.

Table 10: Results of stepwise regression

Independent Dependent Variables
Parameters
Variables ROA ROE NPM
Coefficient 0.071 0.121 -1.390
Constant T test 9.387 5.350 -1.301
Sig, 0.000 0.000 0.198
Coefficient -0.010 -0.013 -0.219
T test -4.480 -1.848 -2.964
Increase labor cost Sig, 0.000 0.069 0.004
T test
Sig,
Coefficient -0.006 -0.016
Decrease sales T test -2.665 -2.396
Sig, 0.010 0.020
Coefficient 0.103
Company Size T test 2.140
Sig, 0.036
F test: 35.184 12.280 6.216
Sig.: 0.000 0.000 0.003
R square adjusted: 0.516 0.261 0.140

Results of stepwise regression analysis (Table 10) show that the return
on assets is significantly and negatively influenced by (F= 35.184, P-
value=0.000 < 0.01) the increase of labor costs (β1=-0.010) and the
decrease of sales (β2=-0.006), this can be expressed statistically, after
estimating the parameters of the regression model, as the following
equation
ROA=0.071 - 0.010*Increase labor cost – 0.006*Decrease sales + ε

690
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Further, t-test results confirms (T=-4.480., P-Value<0.000 < 0.01) the

"Increase labor cost" variable and (T=-2.665, P-Value = 0.010 <0.01)
the "Decrease sales" variable at 0.01 level. Finally, the value of R2 equals
0.516, which means that 51.6% of the changes at "ROA" are explained
by the "Increase labor cost" and the "Decrease sales".

Regarding the return on shareholders‟ equity, the results of stepwise

regression analysis (Table 10) show that the return on assets is
significantly and negatively influenced by (F= 12.280, P-value=0.000 <
0.01) the increase of labor costs (β1=-0.013) and the decrease of sales
(β2=-0.016). This can be expressed statistically, after estimating the
parameters of the regression model, as the following equation
ROE=0.121 - 0.013*Increase labor cost – 0.016*Decrease sales + ε
Further, t-test results confirms (T=-1.848, P-Value<0.069 < 0.10) the
"Increase labor cost" variable and ((T=-2.396, P-Value = 0.020 <0.05)
the "Decrease sales" variable at 0.01 level. Finally, the value of R2 equals
0.261, which means that 26.1% of the changes at "ROE" are explained
by the "Increase labor cost" and the "Decrease sales".

Finally, regarding the net profit margin, the results of stepwise regression
analysis (Table 10) show that the return on assets is significantly (F=
6.216, P-value=0.003 < 0.01) and negatively influenced by the increase
of labor costs (β1=-0.219). While “NPM” is positively influenced by the
size of the firm (β2=-0.103), this can be expressed statistically, after
estimating the parameters of the regression model, as the following
equation
NPM=-1.390 - 0.219*Increase labor cost + 0.103*Company size + ε
Further, t-test results confirms (T=-2.964, P-Value<0.004 < 0.01) the
"Increase labor cost" variable and (T=-2.140, P-Value = 0.036 <0.05)
the "company size" variable at 0.01 level. Finally, the value of R2 equals
0.140, which means that 14.0% of the changes at "NPM" are explained
by the "Increase labor cost" and the "company size".

We accept h1: "covid-19 pandemic negatively influences Saudi

companies’ profitability".

691
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

h2: corporate governance moderates the negative influence of convid-19

on Saudi companies’ profitability.

To test this hypothesis; we first conducted correlation analysis (Table

11), followed by stepwise analysis (Table 12) as follows.

Table 11: Person correlation matrix between research variables

Increase labor

Decrease sales
material cost

Net Profit
Increase

Margin

C Size
ROA

ROE
cost
Corporate Governance

Increase labor cost 1

Non Corporate

Increase material cost .362* 1

Governance

Decrease sales .702 .633 1

ROA -.792-** -.504-** -.667-** 1
**
ROE -.712- -.319-* -.608-** .849** 1
**
Net Profit Margin -.562- -.466-** -.514-** .921** .689** 1
C Size .162 -.254- .028 .041 .089 .125 1
Increase labor cost 1
Increase material cost .629** 1
Governance
Corporate

Decrease sales .495* .309 1

ROA -.504-* -.244- -.508-** 1
ROE -.200- -.155- -.391- .887** 1
Net Profit Margin -.148- -.186- -.119- .412* .328 1
C Size -.029- -.176- -.196- .196 .269 .337 1
**. Correlation is significant at the 0.01 level (2-tailed).
*. Correlation is significant at the 0.05 level (2-tailed).

Results of correlation analysis (Table 11) show that there is a

significantly negative relationship between "Increase labor cost" and
"ROA", "ROE", "NPM" since the correlation coefficients are -0.792, -
0.712, -0.562 at 0.01 level respectively in the condition of "Non-
corporate governance". While in the condition of "Corporate
governance”, there is only significant negative relationship between
"Increase labor costs" and "ROA" since the correlation coefficient is -
0.504 at 0.05 level. Further, the results reveal that there is a significantly
negative relation between "Increase material cost" and "ROA", "ROE",
"Net profit margin" since the correlation coefficients are -0.504, -0.319,
-0.416 at 0.01 level respectively in the condition of "Non-corporate
governance" while in the condition of "Corporate governance" there is

692
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

no relationship between "Increase material cost" and corporate

profitability. Finally, the results show that there is a significant negative
relationship between "Increase labor cost" and "ROA", "ROE", "Net
profit margin" since the correlation coefficients are -0.667, -0.608, -
0.514 at 0.01 level respectively in the condition of "Non-corporate
governance". While in the condition of "Corporate governance" there is
only significant negative relationship between "Increase labor cost" and
"ROA" since the correlation coefficient is -0.508 at 0.05 level.

Table 12: Results of stepwise regression

Dependent Variables
Independent ROA ROE Net profit Margin
Parameters
Variables Non- Non- Non-
Govern. Govern. Govern.
govern. govern. govern.
Coefficient 0.092 0.046 -0.007 0.101 1.524 -3.623
Constant T test 8.849 3.195 -0.092 2.027 5.116 -1.777
Sig, 0.000 0.004 0.927 0.054 0.000 0.089
Coefficient -0.015 -0.031 -0.201
Increase T test -7.054 -6.673 -3.300
labor cost
Sig, 0.000 0.000 0.002
Coefficient -0.007 -0.163
Increase T test -2.506 -2.200
material cost
Sig, 0.017 0.034
Coefficient -0.010
Decrease T test -2.832
sales
Sig, 0.009
Coefficient 0.006 0.158
Company
T test 1.881 1.718
Size
Sig, 0.068 0.099
F test: 39.636 8.018 22.595 4.149 12.080 2.953
Sig.: 0.000 0.009 0.000 0.053 0.000 0.099
R square adjusted: 0.665 0.226 0.525 0.116 0.362 0.075

693
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Results of stepwise regression analysis (Table 12) show that the return
on assets is significantly and negatively influenced by "Increase labor
cost", "Increase material cost" in the condition of "Non-corporate
governance", while it is only significantly negative affect by "Decrease
sales" in the condition of "Corporate governance". This can be expressed
statistically, after estimating the parameters of the regression model, as
the following equation
Non-corporate governance
ROA = 0.092 -0.015*Increase labor cost – 0.007*Increase material*cost + ε
Explanatory power is 66.5%
Corporate governance
ROA = 0.046 -0.010*Decrease sales + ε
Explanatory power is 22.6%

Regarding the return on shareholders' equity, the results of stepwise

regression analysis (Table 12) show that the "ROE" is significantly and
negatively influenced by the "Increase labor cost" in the condition of
"Non- corporate governance", while it is only significantly and
negatively affected by " Decrease sales " in the condition of "Corporate
governance". This can be expressed statistically, after estimating the
parameters of the regression model, as the following equation
Non-corporate governance
ROE = -0.007 -0.031*Increase labor cost + ε
Explanatory power is 52.5%
Corporate governance
ROE = 0.101 -0.006*Decrease sales + ε
Explanatory power is 11.6%.

Finally, the results of the stepwise regression analysis (Table 12) show
that the "Net profit margin" is significantly and negatively affected by
"Increase labor cost", in the condition of "Non-corporate governance",
while it is significantly and negatively affected by the "Company size" in
the condition of "Corporate governance". This can be expressed
statistically, after estimating the parameters of the regression model, as
the following equation
Non-corporate governance
NPM = 1.524-0.201*Increase labor cost -0.163*Increase material cost+ ε
Explanatory power is 36.2%
Corporate governance
NPM = -3.623 +0.1*Company size + ε

694
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Explanation power 7.5%.

From the above analysis, we conclude that covid-19 negatively

influenced Saudi companies' profitability in the condition of "Non-
corporate governance", while this influence is weakened in condition of
"Corporate governance". Therefore, corporate governance reduces the
negative influence of convid-19 on Saudi companies‟ profitability.

We accept h2: “corporate governance moderates the negative influence

of convid-19 on Saudi companies’ profitability”.

5. Conclusion
This study aimed to examine the moderated effect of corporate
governance on the influence of covid-19 on Saudi corporate profitability.
We analyzed 65 companies‟ financial statement listed in the Saudi
capital market (Tadawul). The results attained drives to the conclusion
that even in times of crises (covid-19) implementing corporate
governance reinforces firms‟ performance. These results are closely
linked to the results presented by Iryadi et al. (2019), and Hamdan et al.
(2017).

We found that under the “Non-corporate governance” condition, all three

profitability ratios (ROA, ROE, and NPM) are negatively and
significantly affected by the covid-19 consequences (increase in labor
costs, increase in material costs, and decrease in sales), which led us to
accept that (h1)"covid-19 pandemic negatively influences Saudi
companies’ profitability". While under the "Corporate governance"
condition, the effect of covid-19 consequences is minimized and
moderated; there is only significant negative influence of the “Decrease
of sales” on two profitability ratios (ROA and ROE), and only a
significant negative influence of the control variable (company size) on
the profitability ratio (NPM). This made us accept (h2) “corporate
governance moderates the negative influence of convid-19 on Saudi
companies’ profitability”.

The results of the current study can be generalized to all Saudi

companies listed in the capital market taking into consideration that
results obtained are limited to only testing the independent variable
(covid-19 consequences) and the moderating effect of corporate

695
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

governance on profitability. For example, the Saudi Arabian Monetary

Agency (SAMA) has announced a three actions program to support the
private sector and promote financial stability especially for the SME‟s
sector9. In light of attained results, we recommend that corporate
governance can be improved by the ongoing efforts of the Council of the
Capital Market Authority (CMA) in Saudi Arabia to enhance the
efficiency and effectiveness of in times of crises such as the covid-19
pandemic. Further, we recommend that companies must ensure the
effectiveness of control systems, and develop strategic plans in times of
disasters and reduce the risks of loses. Such points can be considered
fruitful area for future research.

References
ACCA. (2020). The impact of Covid-19 on Audit and Assurance – challenges
and considerations.

Aifuwa, H., Musa, S., and Aifuwa, S. (2020). Coronavirus pandemic outbreak
and firms performance in Nigeria. Management and Human Resource Research
Journal. Vol.9, No. 4, pp. 15-25.

Anwar, Z., Khan, M., and Danish, R. (2019). Corporate governance and cost of
equity: evidence from Asian countries. Journal of Political Studies,
Vol. 26, No. 1, pp. 207-230.

Avdiu, B., and Nayyar, G. (2020). When face-to-face interactions become an

occupational hazard- jobs in the time of covid-19. Policy Research Working
Paper. World Bank Group. May, pp. 1-11

Balios, D., and Zaroulea, T. (2020). Corporate governance, internal audit and
profitability: “evidence from P.I.G.S. countries. Journal of Accounting and
Auditing: Research & Practice. Vol. 2020, pp. 1-17

Baur, D., and Trench, A. (2020). Covid-19 infection of Australian companies.

Available online at [Link]

Bhagat S., and Bolton B. (2019). Corporate governance and firm performance:
the sequel. J. Corp. Finance. Vol. 58, pp.142–168.

9
Sky news, March 14th, 2020

696
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Bhimani, A., Horengren, C., Datar, S., and Rajan, M. (2019). Management and
Cost Accounting. Pearson; 7th ed.

Broadstock, D., Chan, K., Cheng, L., and Wang, X. (2020). The role of ESG
performance during times of financial crisis: evidence from covid-19 in China
(June 15, 2020). Available online at [Link]

Buallay, A., Hamdan, A., and Zureigat, Q. (2017). Corporate governance and
firm performance: evidence from Saudi Arabia. Australasian Accounting,
Business and Finance Journal. Vol. 11, no. 1, pp. 78-98.

Busru, S., and Shanmugasundaram, G. (2017). Effects of innovation investment

on profitability and moderating role of corporate governance: empirical study
of Indian listed firms. Indian Journal of Corporate Governance. Vol. 10, no. 2,
pp. 97–117.

Cengiz, H. (2016). Corporate governance and firm profitability: evidence from

Turkey. International Journal of Trade, Economics and Finance. Vol.7, no.6,
pp. 238-241

Coibion, O., Gorodnichenko, Yuriy., and Weber, M. (2020). Labor markets

during the covid-19 crisis: a preliminary view, CESifo Working Paper, No.
8238, Center for Economic Studies and ifo Institute (CESifo), Munich

Ding, W., Levine, R., Lin, c., and Xie, W. (2020). Corporate immunity to the
covid-19 pandemic. NBER Working Paper No. 27055. National Bureau of
Economic Research, pp. 1-52

Egyption Center for Economic Studies. (2020). Sector analysis of the

implications of covid - 19 on the Egyptian economy.

Gerged, A., and Agwili, A. (2020). „How corporate governance affect firm
value and profitability? evidence from Saudi financial and non-financial listed
firms‟. International Journal of Business Governance and Ethics. Vol. 14, no. 2,
pp. 144–165.

Hamdan, A., Buallay, A., and Alareeni, B. (2017). The moderating role of
corporate governance on the relationship between intellectual capital efficiency
and firm‟s performance: evidence from Saudi Arabia. Int. J. Learning and
Intellectual Capital. Vol. 14, No. 4, pp.295–318.

Iqbal, M., and Javed, F. (2017). The moderating role of corporate governance
on the relationship between capital structure and financial performance:

697
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

evidence from manufacturing sector of Pakistan. International Journal of

Research in Business and Social Science. Vol. 6, no. 1, pp.89-105

Iriyadi, I., Tartilla, N., and Gusdiani, R. (2019). The effect of tax planning and
use of assets on profitability with good corporate governance as a moderating
variable. Advances in Economics, Business and Management Researc. Vol.
143, pp. 220-227

Khoo, E., and Lantos, J. (2020). Lessons learned from the COVID-19
pandemic. Acta paediatrica (Oslo, Norway : 1992), 109(7), 1323–1325.

Kimmel, P., Weygandt, J., and Kieso, D. (2018). Financial Accounting: Tools
for Business Decision Making. Wiley; 9th ed.

Lu, J., Liang, X., Shan, M., and Liang, X. (2015). Internationalization and
performance of Chinese family firms: the moderating role of corporate
governance. Management and Organization Review. Vol. 11, no. 4, pp. 645-
678.

Milosevic, D. (2015). A survey of corporate governance. The Journal of

Finance. October, pp. 737-741.

Minnow, N., and Monks, R. (2004). Corporate Governance. Black well; 3rd ed.

Murhadi, W. (2009). Good corporate governance and earnings management

practices: an Indonesian cases. Published in Proceedings ICBMR, pp.100-120

Scherer, A., and Voegtlin, C. (2020). Corporate governance for responsible

innovation: approaches to corporate governance and their implications for
sustainable development. Academy of management perspectives. Vol. 34, no.
2, pp. 1-68

Secinaro, S., Calandra, D., and Biancone, P. (2020). Reflection on coronavirus

accounting impact on small and medium sized enterprises (SMEs) in Europe.
International Journal of Business and Management. Vol. 15, No. 7, pp. 48-56

Spurk, D., and Straub, C. (2020). Flexible employment relationships and

careers in times of the covid-19 pandemic. Elsevier Public Health Emergency
Collection. Journal of Vocational Behavior. 119

Sulaiman, A., Mijinyawa, U., and Khadijah, K. (2018). Corporate governance

mechanisms and profitability of listed companies in Nigeria. Journal of
Accounting and Management. Vol. 1, no. 2, pp.172-178.

698
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Song, H., and Kang, K. (2019). The moderating effect of CEO duality on the
relationship between geographic diversification and firm performance in the US
lodging industry. International Journal of Contemporary Hospitality
Management. Vol. 31 No. 3, pp. 1488-1504.

Zohrah, N., and Hamidah. (2017). The Role of Corporate Governance in Firm
Performance. SHS Web of Conferences 34, pp. 1-6

699
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Institutional Determinants of Emerging Market Returns, Flows, and Risk

Premiums

Ralph Sonenshine
Corresponding Author: American University, Department of Economics, 4400 Massachusetts Avenue, NW, Washington,
DC 20016. E-mail: Sonenshi@[Link]; Telephone: (202) 885-3733.

Bradley Erickson
Additional author: Bradley Erickson
E-mail: Be9669a@[Link]

700
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Institutional Determinants of Emerging Market Returns, Flows, and Risk

Premiums

Abstract

It is well established that investors incorporate a market risk premium when allocating capital

between assets that are imperfect substitutes. Market risk premiums include economic risk,

political instability, debt burden, etc. Institutional factors such as levels of corruption,

transparency, and democracy may also influence investor allocation decisions. This paper

examines the effects of these three institutional factors on overall market performance among

emerging market economies. We find that institutional quality impacts emerging market returns

and risk premiums, but it does so predominantly in the countries and sectors where the corruption,

transparency, and democracy levels are below average. Our industry findings are mixed with the

primary government owned/managed sectors being negatively impacted by a deterioration in the

corruption index, but positively impacted by a decline in the democracy index.

JEL Codes: G11, G15, G18

Key words: Returns, corruption, transparency, institutions, democracy, risk premium

701
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

I. Introduction

Globalization increases the movement of goods and assets around the world. As barriers to

capital mobility come down, assets naturally move to markets that generate the highest risk-

adjusted return. Thus, given the limited supply of capital, the equity and bond markets in emerging

markets have grown substantially providing greater liquidity and access to capital for businesses.

In addition, emerging markets grow as investors seek to diversify risk, assuming investors can

properly evaluate emerging market equity or bond risks.

Like with any investment, interest in emerging markets is a function of the risk-adjusted return

on investment, the volatility of returns, and the liquidity of the investment. The latter two can be

viewed as a function of the stability of the markets. For equities, return on investment depends on

idiosyncratic business risk as well as market risk, which is often higher in emerging versus

developed markets. Market risk in emerging markets has historically been higher due to

macroeconomic concerns (e.g., inflation, currency risk, political risk, changes in the business

cycle).

Market stability also impacts the equity or market risk premium. Market volatility is a

function, in part, of the quality of prevailing institutions, which include the levels of transparency,

corruption, and prevalence of democratic institutions. Alquist (2006) in his study of the causal

impact of economic policy on investor portfolio allocation argues that portfolio investors rely

primarily on macroeconomic factors e.g. inflation and GDP growth, relegating institutional factors

to a lower tier of consideration. However, a growing body of research1 suggests that institutional

1
For example, the Frazier Institutes’ Economic Freedom Report (2019) cites, “studies have found that countries
with

702
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

factors, such as democracy, corruption, and transparency, influence the return on investment, even

for short term portfolio (equity) investments (Basinger et al. 2008; Donadelli et al. 2014; Lehkonen

and Heimonen 2015; and Gelos and Wei 2005). Transparency impacts the equity risk premium by

affecting the level of uncertainty in the investment decision. Corruption levels can also have a

negative impact on returns and the risk premium due to higher transactions costs and/or other

economic inefficiencies inherent in corrupt economies. Finally, the level of democracy may

impact emerging market equity returns as democratization can increase the size, efficiency, and

transparency of capital markets.

Measuring the quality of institutions uniformly is a difficult task. This paper leverages key

indices from the Frazier Institute, Transparency International, and the Economist Intelligence Unit

(EIU) relating to corruption, transparency, and democracy to assess the impact that changes in

these indicators have on emerging market equity returns, the market risk premium, and emerging

market returns by sector. The hypothesis is that emerging market equity returns and flows will be

negatively impacted by the growth in bribery and the decline in transparency and democracy. We

also expect the market risk premium to be higher or adversely impacted by deterioration in these

three indicators. In addition, we hypothesize that government-owned or operated industries

(consumer services, healthcare, oil and gas, and utility sectors) will be negatively affected by a

deterioration in the bribery, transparency, and democracy indexes. We find that the change in

institutional quality does indeed impact emerging market returns, flows, and market risk

premiums, but it does so predominantly in the sectors where the bribery, transparency, and

institutions and policies more consistent with economic freedom have higher investment rates.”

703
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

democracy levels are below average. Our industry findings are mixed with the primary government

owned/managed sectors being negatively impacted by deterioration in the bribery index, but

surprisingly being positively impacted by a decline in the democracy index.

The layout of the paper is as follows. Section II discusses the relevant literature leading to the

testable hypotheses. Section III covers the data and empirical methodology. Section IV presents

the regression results, and Section V provides some concluding remarks.

II. Literature Review

The traditional approach to investment (Markowitz, 1959) assumes that investors are

primarily evaluating risk-adjusted return when making equity allocation decisions. Furthermore,

economists have traditionally argued that competition among rational investors will lead to an

equilibrium in which stock market prices and, therefore, investor returns equal the discounted

value of expected cash flows (Baker and Wurgler, 2006). A critical part of this argument, however,

is that the assets are perfectly substitutable. In cases where assets are imperfect substitutes, a risk

premium needs to be assessed to enable investors to determine the best risk adjusted location and

sector in which to invest.

Equity Premium and Emerging Markets

Boyer et. al. (2017) note that emerging markets present investors with additional risks, which

are aggregated into country risk. Boyer et. al. (2017) suggest risk premiums encompass problems

relating to liquidity and transparency of financial markets, less developed corporate and legal

frameworks, economic uncertainty, political instability, concerns over expropriation, and

704
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

problems repatriating earnings.

Embedded in the risk premium calculation are also institutional quality factors. In this paper

we estimate the impact of institutional quality regarding transparency, corruption, and democracy

on the emerging market risk premium as well as market returns, equity flows, and market returns

by sector.

Corruption and Foreign Investment

Corruption sways economic growth (Mauro 1995), foreign direct investment (Wei 2000; Wei

and Wu 2002), and investment in education, and health care (Gupta, Davoodi, and Tiongson 2001).

The definition of corruption is as varied as its impact. For our inquiry, we follow the definition of

“public corruption” by Rose-Ackerman (1975) and Rodriguez et. al. (2006), among others who

define corruption as the misuse of public office for private gain.

Political economists often frame corruption as a “grabbing hand” or a “grease on a wheel”

phenomenon. The “grabbing hand” effect refers to the potential for corruption to increase

transaction costs for foreign investors (Bardhan 1997). Investors who bribe politicians for

contracts, bureaucrats for permits, and police for protection increase their overall costs and

decrease the return on their investment. Another potential transaction cost is the risk that investors

ruin to their reputation should they get caught bribing foreign officials (Zhao, Kim, and Du 2003).

As such, corruption can increase costs and reduce the return on investment, but the benefits of

corruption can outweigh its costs.

705
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The “greasing the wheel” effect suggests the benefits of bribery outweigh the costs for

investors, especially in countries with weak legal systems and regulatory regimes (Bardhan 1997).

For example, Houston (2007) observes that corruption increases economic growth in countries

with weak legal systems. Similarly, Swaleheen and Stansel (2007) observe corruption increasing

economic growth, but in countries with high levels of economic freedom.

Whether supporting the “grabbing hand” tor “greasing the wheel” theory, most research on

the economic impact of corruption examines corruption’s effect on the macroeconomy rather than

portfolio investment. For example, Quazi (2014) finds support for the “grabbing hand” theory, but

observes the negative impact of corruption on foreign direct investment. Dutt and Traca (2010)

find that corruption strains trade the majority of the time, but bolsters trade in countries with

exceptionally high tariffs. Diamonte et al. (1996) did study the effect of corruption on equity

investing and found higher average market equity returns in more corrupt countries than in less

corrupt countries. They also find political risks (i.e. corruption) to impact stock market returns

more in emerging markets than in developed markets. Additionally, Lee and Ng (2009) note

corruption erodes future cash flows and atrophies shareholder value over time. To parse out

opposing findings at the national-level, more recent research examines the impact of corruption by

industry at the firm-level.

Corruption at the Firm-Level by Industry

Doh et al. (2003) describe two ways to view corruption at the firm level: 1) the uncertainty

inherent in corrupt governments, and 2) the frequency of transactions in corrupt governments. Both

of these perspectives incur direct costs (bribes or “grease”) and indirect costs (losses in firm

706
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

productivity, macroeconomic instability, etc.). Direct costs are relatively straightforward to

measure, but rarely observed. Bribers are a clandestine bunch. The paucity of direct cost

measurements of corruption force researchers to analyze indirect costs.

Researchers studying the indirect cost of corruption typically focus on a particular country

or region. For example, Athanasouli et al. (2012) find that bribery and corruption lead to smaller

firm size and smaller firm growth in Greece. In Latin American, Gaviria (2002) notes corruption

reduces a firm’s sales and growth, implying corruption enervates a firm’s competitiveness.

Sahakyan and Stiegert (2012) find Armenian firms that are younger, relatively larger, and lack

significant competition perceive corruption more favorably than their older, smaller, more

competitive counterparts. McArthur and Teal (2002) and Faruq et al. (2013) observe firms in

Africa that bribe public officials are less productive than firms that refrain from bribery and.

Additionally, they find firms in corrupt African countries are less productive than firms in

corruption-free African countries.

Donadelli and Persha (2014) engage in one of the few studies that disaggregate by industry

to analyze corruption’s impact on portfolio investment. They examine equity risk premia across

10 industries in 19 emerging markets. Their measure of country governance2 is negatively

impacted industry-level returns in the consumer goods, basic materials, financial, and industrial

goods sectors. In a similar study Donadelli et al. (2014) analyze the impact of corruption

specifically on industry average equity returns. They find that corruption hampers average equity

returns in the consumer service, financial, oil & gas, and technology industries. This study follows

2
This measure was taken from the World Bank’s Worldwide Governance Indicators which includes a control for corruption.

707
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Donadelli et. al. (2014) but diverges in that we analyze the impact of corruption along with

transparency and democracy on industry-level equity returns in addition to market risk premiums

and equity flows. Moreover, we segment the countries by high and low levels of institutional

quality, to compare the effects in these groups.

Transparency

Similar to the impact of corruption on equity returns, studies find that transparency,

unsurprisingly, influences equity returns and flows (Gelos and Wei 2005 and Lang et al. 2001).

Gelos and Wei (2005) examine two types of transparency: government transparency or “the

timeliness and frequency of macroeconomic data availability and transparency in the conduct of

macroeconomic policies” and corporate transparency referring to the clarity and “availability of

financial and other business information” (pg. 2,988). They find that each type of transparency

positively affects investment flows. In other words, more transparent countries attract more

investment. Addtionally, foreign investors are more likely to pull out their funds in less transparent

countries during crises than more transparent countries. Lang et al.’s (2011) results mimic Gelos

and Wei’s. Lang et al. (2011) find that foreign investment into a country increases with

transparency.

The Democratic Advantage

There is a large amount of research covering the economic advantages and disadvantages

of democracies. Most notably, Schultz and Weingast (1996; 1998; 2003) suggest that democracies

tend to honor their debt obligations. Countries honoring debt obligations encourage greater

inflows

708
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

of capital, greater access to loans, and lower interest rates. Saiegh (2005), however, challenges the

notion of a democratic advantage. His analysis of developing countries suggests that democracies

reschedule their debts more often than authoritarian regimes, which suggests democracies honor

debt obligations less than authoritarian regimes. Additionally, both democracies and non-

democracies pay similar interest rates.

Few studies, however, have examined the impact of varying levels of democracy on equity

returns. Biglaiser et al. (2008) investigate the relationship between political regimes of developing

countries and credit ratings and portfolio investment. They argue that democracy matters greatly

for “lower-end” developing nations, but not as much for developed economies. Biglaiser et al.

(2008) find that higher interest rates prevail in developing countries, which can lead to higher risk

premia. As such, portfolio investors tend to shelter their large investments in highly rated debt,

often in countries with strong democratic institutions. Lehkonen and Heimonen (2015) find similar

results, except they observe a parabolic relationship between democracy and political risk,

meaning political risk is low in pure authoritarian regimes and low in pure democracies. When

political risk is higher, they observe higher equity returns. Their results suggest transitioning

democracies (i.e., from authoritarian to democratic) experience the highest equity returns.

In summary, there is a fair amount of research relating to the impact of institutional quality

on investment. The contribution of this paper is to estimate the impact of all three institutional

concerns (corruption, transparency, and democracy) on market and industry returns, equity flows,

and risk premium, and to drill down into differences in performance between high and low levels

of institutional quality.

709
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

III. Data and Methodology

IIIa. Data description

The data set consists of 18 years of equity returns and flows from 26 emerging market

countries, which are tracked by Morgan Stanley’s emerging markets index. These countries

include 10 high income, 11 upper middle income, and five lower middle income countries, per the

World Bank classification.3 These countries are commonly classified as emerging markets in that

they are moving toward developed market status.4 These countries typically have rapid but volatile

growth, low per capita income relative to developed countries, and developing regulatory

frameworks, . See the Appendix (Table 7) for a list of countries in the data set as well as their

classification by year into high and low for the three key institutional categories of study,

corruption, transparency, and democracy.

Data for the corruption measurement was obtained from the Frazier Economic Freedom

Report under Extra Payments, Bribes, and Payments. Data for the democracy rankings were

gathered from the Economist Intelligence Unit (EIU) voice and accountability index, which is a

weighted average of the EIU rankings of four subcategories: vested interests, accountability of

public officials, human rights, and freedom of association. 5 The transparency index was taken

from the Transparency International Corruption Perceptions Index (CPI), published annually

by Transparency International. The CPI ranks countries by their perceived levels of public sector

3
See [Link]
4
See [Link] for a definition.
5
See [Link].

710
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

openness as determined by expert assessments and opinions. 6

Emerging market returns were gathered from the Refinitive (formerly Thomson Reuthers)

DataStream database covering total annual equity returns and industry equity returns for the 26

countries from 2000 through 2017. Net equity flows were gathered from the world bank data set

covering portfolio equity net inflows. Market risk premiums, defined as the additional return that's

required on an index or portfolio of investments above the given risk-free rate, were gathered from

Fernando, et. al. (2016) annual survey7 of finance / economics professors, equity analysts, and

company managers.8 The data was only available for the years 2010 through 2017. Finally,

following other work (Basinger et al. 2008; Donadelli et al. 2014; Lehkonen and Heimonen 2015;

among others) we include macro data (real economic growth, the real interest rate, and the current

account balance) and financial data (foreign ownership levels and the highest marginal tax rate).

This data was obtained from the World Bank indicators.

IIIb. Methodology

To assess the impact of institutional indicators on the annual equity returns, flows, and

market risk premium, we developed the following reduced form regression equations.

𝑙𝑛𝑅𝑒𝑡𝑢𝑟𝑛𝑠 𝑖𝑗 = 𝛽0 + 𝛽1 𝐸𝑀𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑖𝑗 + 𝛽2 𝑙𝑛𝐼𝑛𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑜𝑛𝑎𝑙𝑖𝑗 + 𝛽3 𝑡𝑀𝑎𝑐𝑟𝑜𝑖𝑗 + 𝛽4 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙𝑖𝑗 + 𝜀, 𝑓𝑒 (1)

6
See Transparency Index 2011. The index covers the year 2000 through 2017, excluding 2001, when data was not available.
7
This survey was published each year in the Journal of International Business Research and Marketing.
8
Respondents typically referred to internal estimates as well as published information from Damodaran, Duff & Phelps,
Ibbotson/Morningstar, DMS, Graham-Harvey, Bloomberg, Analysts, Exeperience, Own judgement, Grabowski, Pratt’s and
Grabowski, Brealy & Myers, and Siegel (Fernandez et a., 2016).

711
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

𝑙𝑛𝐹𝑙𝑜𝑤𝑠 𝑖𝑗 = 𝛽0 + 𝛽1 𝐸𝑀𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑖𝑗 + 𝛽2 𝑙𝑛𝐼𝑛𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑜𝑛𝑎𝑙𝑖𝑗 + 𝛽3 𝑡𝑀𝑎𝑐𝑟𝑜𝑖𝑗 + 𝛽4 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙𝑖𝑗 + 𝜀, 𝑓𝑒 (2)

𝑙𝑛𝐸𝑞. 𝑅𝑖𝑠𝑘𝑃𝑟𝑒𝑚 𝑖𝑗

= 𝛽0 + 𝛽1 𝐸𝑀𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑖𝑗 + 𝛽2 𝑙𝑛𝐼𝑛𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑜𝑛𝑎𝑙𝑖𝑗 + 𝛽3 𝑡𝑀𝑎𝑐𝑟𝑜𝑖𝑗 + 𝛽4 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙𝑖𝑗 + 𝜀, 𝑓𝑒 (3)

𝑙𝑛𝐼𝑛𝑑. 𝑟𝑒𝑡𝑢𝑟𝑛 𝑖𝑗 = 𝛽0 + 𝛽1 𝐸𝑀𝑟𝑒𝑡𝑢𝑟𝑛𝑠𝑖𝑗 + 𝛽2 𝑙𝑛𝐼𝑛𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑜𝑛𝑎𝑙𝑖𝑗 + 𝛽3 𝑡𝑀𝑎𝑐𝑟𝑜𝑖𝑗 + 𝛽4 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙𝑖𝑗 + 𝜀, 𝑓𝑒 (4)

A fixed effects model was run for each regression with i serving as the country ID and j as

the year. 9 The primary control variable is emerging market returns (EMreturns), which refers to
10
the MSCI emerging market index returns. Ln(Institutional) refers to the three institutional

variables, Corruption Perceptions Index (CPI), the Bribery index, and the Democracy Index, all

logged.11 In addition, we include a fourth institutional index, the Frazier Institute Regulation index,

as a control variable. Macro refers to real GDP growth logged, the real interest rate, and the

country’s current account balance. We expect equity returns and flows to be positively impacted

by increasing real GDP growth, a depreciating real exchange rate, and an improved current account

balance. Financial refers to foreign ownership of banks and the highest marginal tax rate. We

anticipate increasing foreign ownership to have a positive impact on equity returns and flows as

higher foreign ownership signals lower risk. In contrast, we presume increasing marginal tax rate

at the highest level would have a negative impact on equity returns and flows. The regressions in

both equations were run for the total sample and restricted samples by high (above the median)

9
We ran a Hausman test on both regressions and found the fixed effects model to be the appropriate model
versus a random effects model.
10
The MSCI Emerging Markets Index includes large and mid cap companies in 26 emerging markets
countries. As of 2019, the index covered approximately 85% of the market capitalization in each country. See
[Link]
11
See Table 3A in the appendix for the institutional variable wording and definitions.

712
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and low (below the median) measurements for the three institutional indices of interest, corruption,

transparency, and democracy.

A second regression covered the total sample and the six restricted samples with the response

variable being equity flows. Similarly, we used the same regression format to assess the impact

of the institutional factors on the market risk premium in the third regression.

Finally, a fourth regression analysis covered the second research question relating to the

impact of the institutional quality on emerging market industry returns for each country. As such,

the response variable in equation (4) is the annual equity return in each country in basic materials,

consumer goods, consumer services, financial services, healthcare, industrial, oil and gas,

technology, and utilities sectors. The hypothesis is that the impact of the institutional variables on

industry returns will vary based on the level of government ownership/management between the

industries.

The summary data on returns by industry segmented by high and low institutional (corruption,

democracy, and transparency) index is shown in Figure 1. See Table 7 in the appendix for a list

of the countries that fall into the above and below average segments for the democracy,

transparency, and bribery/corruption indexes.

713
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1. Industry Emerging Market Returns by High vs. Low Institution Index
45.00

40.00

35.00
Mean Percnet Returns

30.00

25.00

20.00

15.00

10.00

5.00

-
Basic Consumer Consumer Financial Industrial Oil and Gas Utilities Health CareTechnology Total
Goods Goods Services Returns
High Corruption Low Corruption High Democracy Low Democracy High Transparency Low Transparency

From the chart we see on the right-hand side that the total returns are the highest in the low

democracy and transparency countries. We see this result in each of the sectors except for health

care, where the returns were the highest in the high transparency sector. The average annual total

returns along with the average annual returns in the basic, financial, consumer service and

industrial industries passed the t-test, suggesting significant differences in the means between the

low and high transparency and democracy groups. 12

With regard to countries that are above or below the median for corruption or bribery, there

is no clear pattern, as the returns in roughly half of the sectors (basic goods, financial, oil and gas,

12
The healthcare and oil and gas sectors did not pass the t-test in either segmentation, while the technology passed only in the
transparency breakout and the utility industry passed only in the democracy breakout.

714
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and health care) are higher in the above median countries for bribery/corruption, while the opposite

is the case in the consumer goods, consumer services, industrial, utilities, and technology sectors.

Only the healthcare sector, however, passed the t-test suggesting the mean average annual returns

are different in healthcare between countries in the high versus the low bribery/corruption groups.

Next, we examine in Figure 2 differences in the volatility of average annual industry returns

between low and high institutional quality segments.

715
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 2. The Volatility of Average Annual Industry Returns by Institutional Group

High Corruption Low Corruption

10 20 30 40

High Democracy Low Democracy

10 20 30 40

High Transparency Low Transparency

10 20 30 40

20 40 60 80 20 40 60 80
Standard Deviation
Basic Goods Consumer Goods Consumer Services Financial Industrial
Oil and Gas Utilities Health Care Technology Total Returns

From the figure we see the returns and standard deviations by sector are generally higher in

the low transparency and democracy sectors than in the high transparency and democracy sectors.

This result is particularly the case in the democracy breakout with the standard deviations in the

716
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

high democracy countries at or below 40 percent, but greater than 40 percent in the low democracy

countries. The primary exception is healthcare where the average annual return and standard

deviation appear higher in the high transparency and democracy countries. For the

bribery/corruption segmentation, no clear pattern emerges.

Summary data for the other two response variables, market risk premium and equity flows,

are presented below.

Table 1: Average Annual Risk Premium and Equity Flows by High and Low Institutional Index

High Low High Low

High Low Tran Tran Dem Dem
Cor- Cor- sp- sp- o- o-
Aver rupti uptio aren aren crac crac
Variable age on n cy cy y y
(Std) (Std) (Std) (Std) (Std) (Std) (Std)

7.8% 9.1% 6.3% 7.3% 8.6% 7.7% 7.8%

Market Risk Premium (2.3 (3.0 (0.8 (1.9 (2.5 (2.6 (1.4
%) %) %) %) %) %) %)

$2,850 $8,82 $3,97 $3,03 $2,62 $3,42 $2,71

Net Equity Flows (8,800) 0 0 0 0 0 0
(in millions of $) (5,80 (13,00 (9,10 (8,90 (9,49 (9,09
0) 0) 0) 0) 0) 0)

We see from this table that the market risk premium is fairly similar between countries

segmented by high and low democracy. In contrast, the countries that scored high (good) on

corruption tended to have high average market risk premiums relative to the low corruption (below

the median) countries. Also, we see the countries that scored below the median on transparency

had a higher market risk premium than the countries that scored high on transparency.

Regarding net equity flows we see far higher net equity flows among country scoring high on

717
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

corruption (good) versus the below median countries. This result may be because the countries

scoring high on corruption are likely larger and higher income/capita countries versus the countries

scoring below median. We see slight differences in the average net equity flows among the high

versus low transparency and democracy groups.

Table 2 shows the mean level for each of the macro and financial variables used in the

regression analysis.

Table 2: Segmentation of Macro and Financial Variables by Institutional Index

Macro/ High High High Low High Low

Financial Average Cor- Cor- Trans- Trans- Demo- Demo-
Variables ruption ruption parency parency cracy cracy
Real GDP
3.2% 3.4% 3.2% 2.8% 3.7% 2.8% 4.0%
Growth Rates
Max. Marginal
30.5% 29.9% 30.0% 27.5% 31.9% 30.6% 29.9%
Tax Rates
Current Account
Balance (in 2.1 4.4 0.6 4.7 0.6 1.6 2.7
billions of USD)
Real Interest Rate 5.7% 6.8% 4.9% 5.7% 5.7% 5.1% 5.9%

Percent Foreign
6.8% 7.4% 6.2% 7.5% 6.2% 7.0% 6.4%
Ownership

We see from the chart that real GDP growth rates are higher in the low versus high

transparency and democracy countries. The higher country growth rates in the low institutional

segments may in part explain the higher returns found in figure 1. We will explore this finding

more in the regression results. We see the high transparency group has the lowest maximum

marginal tax rate; otherwise, we find little difference in the top marginal tax rates between the

groups. We see large differences in the annual current account balances between the high and

low

718
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

corruption and transparency segments. This difference may be due to the size of the countries in

the respective segments. In contrast, the real interest rate appears to be similar between groups

except for the high corruption group, where we see a higher real interest rate prevails than in the

low corruption group. Finally, we see foreign ownership percentages are higher in each of the

high versus low institutional segments, suggesting foreign investors may view these countries to

have lower risk or that these countries have fewer restrictions on foreign ownership. We will

explore how these explanatory variables along with the institutional indices impact the response

variables in the next section.

IV. Results

Table 3 shows the results from the regression analysis of emerging market returns segmented

by high and low democracy, transparency, and corruption index.13 Only 286 of the potential 468

observations were covered due to data availability. We see from the first row that the emerging

market returns are significant as expected in all of the regressions. We also find in the total sample,

column (1), that the coefficient for the democracy index logged is negative and significant

suggesting that countries that scored higher in the democracy index had lower total returns. The

coefficient for the democracy index was not significant in any of the other restricted regressions,

except for the low transparency group. Though this finding is only significant at the 0.10 level, it

is surprising and contradicts the parabolic relationship between political risk and market returns

Lehkonen and Heimonen (2015) observe. The difference may be that we include other institutional

13
A correlation matrix, shown in the Appendix, indicated moderate correlation (0.68) between the log of the bribery and
transparency index and minimal correlation between other indexes.

719
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

indexes. Alternatively, the disparity could be attributed to the differences in the indexes used.14

When the sample is segmented into above and below median institution scores, we find that

the coefficient for the transparency index logged is positive and significant in the low or below

average sample for corruption, transparency, and democracy. The coefficients were not significant

in the above median samples for these institutional categories. This finding indicates that

improved transparency has a positive effect on emerging market returns in the countries who score

low on corruption, democracy, and transparency, but not on the countries that have a high score.

In fact, none of the variables outside of emerging market returns had a significant impact on equity

returns in the countries with above average institutional indexes. This result indicates that country

equity returns are primarily influenced by overall emerging market returns when a country’s

institutional quality (principally transparency level) is increasing.

From column (1) we also see that coefficients for the current account and tax rate were

positive and significant. The rationale behind this result likely relates to the improving strength

of the economy as the current account might improve as the country becomes more competitive,

causing equity returns to increase. Similarly, increasing tax rates may occur in a strengthening

economy, enabling policy makers to improve the country’s fiscal position, which could positively

impact market returns. Also, the coefficient for economic growth is positive and weakly

significant, providing limited evidence that emerging market returns were positively impacted by

a growing economy as measured by real GDP growth.

14
We used the EIU democracy index, while Lehkonen and Heimonen (2015) use the Polity index of Polity IV and the
democratic accountability index from the Political Risk Service, published in ICRG

720
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Turning to Table 4, equity flows, we find that the coefficient for the corruption index has a

negative, significant impact on equity flows in the total sample, suggesting that as corruption

scores improve, net equity flows into emerging market countries decrease. This result is quite

surprising, and we do not find the coefficient for the corruption index to be significant in any of

the restricted regressions. This finding suggests that equity investors are attracted to countries

where the level of corruption or bribery is increasing. This result may be due, as Wei and Shleifer

(2000) point out, to increasing FDI into China, a country with higher lower scores on corruption.

We also find the coefficients for the other three institutional variables to be insignificant when

regressed against equity flows. In fact, only the coefficient for economic growth was significant

in its impact on equity flows in the total sample. In addition, we find the coefficients for the

democracy and transparency variables to be positive and significant in the low democracy and

transparency segments, suggesting that improving democracy in countries with relatively weak

democracy and transparency measurements will have a positive impact on equity flows. We also

see the coefficient for growth to be positive and significant in the low transparency and corruption

countries, while the coefficient for the percent foreign ownership is positive and significant in the

low democracy countries. These findings again suggest that countries with relative weak

institutions relating to corruption, democracy, and transparency will attract equity flows with

improving macro conditions and greater levels of foreign ownership. We did not observe this

result for the countries with higher than median institutional scores in columns 2, 4, and 6. In

contrast, equity flows in these countries are influenced to a greater degree by emerging market

returns and perhaps other unobserved factors, since the R-squared is lower in the high versus low

institutional sectors.

721
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In Table 5, we analyze the impact of institutional quality on emerging market equity returns

by industry. Our hypothesis is that institutional quality will have a significant impact on industry

returns in those sectors that are owned or operated by the government. What we find is that in

addition to the impact of the overall emerging market index, which is significant in all regressions,

the returns in the consumer services sector are negatively impacted by improvements in the

corruption level. This result suggests that returns on consumer services, whose sub-segments

include retail and wholesale services, leisure and hospitality services, health and social services,

and education, are higher when there is a greater level of corruption (lower score). Similarly, in

the technology, oil and gas, and utility sectors, we find a negative, significant coefficient for the

democracy index logged, suggesting higher returns in these industries are associated with declining

democracy index ratings. We also see the coefficient for the corruption index is negative and

significant in the utility industry, indicating equity returns are higher in the utility industry in

countries that have worsening corruption levels, perhaps as corruption helps utilities wield market

power.

In Table 5, we analyze the effect of the institutional indices on the required equity premium

in the emerging market countries. In column one we see the coefficient for the corruption index

is negative and significant while the coefficient for the transparency index is positive and

significant. These findings suggest that the higher (better) the corruption index the lower the

required market risk premium. This result appears to be driven by the lower corruption group

(column 3) where we also see a negative coefficient for the corruption coefficient and a positive

coefficient for the transparency index coefficient. We did not, however, find the corruption index

to be significant when regressed against the log of market returns. In contrast, we find the

722
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

transparency index to have a positive impact on both market return and the equity premium, with

the result driven by the below median countries for corruption with respect to the equity premium

and all three institutional indices (corruption, democracy, and transparency) when regressed

against market return.

V. Conclusions

This paper examines the effect that varying levels of institutional quality relating to

corruption, transparency, and democracy have on annual total equity returns, industry returns,

equity flows, and the market risk premium in emerging market countries over an 18 year period.

We find that an increasing level of transparency has a positive, significant impact on average

annual market returns in each of the below median group of countries relative to corruption,

transparency, and democracy. This result suggests that for emerging market countries where

corruption, transparency, or democracy are poor, policies that improve transparency will have a

positive impact on market returns. We also find that for countries in the below median group for

corruption, the equity premium required will decline as these countries improve their corruption

levels.

In addition, we see changes in these institutional indices impact several emerging markets

industry returns. Specifically, we see higher democracy scores result in lower returns in the utility

and oil and technology industries, while higher scores for corruption result in lower returns for the

consumer services and utility industries. The commonality between these industries is that they

have a high level of government ownership or management, particularly in emerging market

countries. Finally, we see higher corruption scores result in lower net equity flows, a result that

723
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

may be driven by FDI into China. Less surprising is the finding that net equity flows are positively

impacted by improved democracy indices in countries with below median democracy and

transparency scores.

In summary, varying corruption and transparency do seem to impact market returns, flows,

and market risk with the impact found primarily in emerging market countries that are ranked in

the bottom half in institutional quality. We found democracy indicators to have minimal impact

on market returns, but to have a positive impact on equity flows. The policy implication is fairly

clear. Investors should incorporate changing levels of transparency and corruption when investing

in emerging market countries where transparency and corruption is a problem. However, there

are certainly vested interests (e.g. industry interests) that will want the status quo to continue given

the benefits they receive from the institutions in place.

This paper only covers the impact of institutional quality in emerging markets on equity

investing. It is left to other research to study the effect of varying institutional quality on the debt

markets.

References

Ahlquist, John S. 2006. “Economic Policy, Institutions, and Capital Flows: Portfolio and Direct
Investment Flows in Developing Countries.” International Studies Quarterly 50 (3): 681–704.
Al-Sadig, Ali. 2009. “The Effects of Corruption on FDI Inflows.” Cato Journal 29 (2): 267–94.

724
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Athanasouli, Daphne, Antoine Goujard, and Pantelis Sklias. 2012. “Corruption and Firm Performance:
Evidence from Greek Firms.” International Journal of Economic Sciences and Applied Research
5 (2): 43–67.
Bardhan, Pranab. 1997. “Corruption and Development: A Review of Issues.” Journal of Economic
Literature 35 (3): 1320–46.
Biglaiser, Glen, Brian Hicks, and Caitlin Huggins. 2008. “Sovereign Bond Ratings and the Democratic
Advantage: Portfolio Investment in the Developing World.” Comparative Political Studies 41 (8):
1092–1116.
Boyer, Benoit, Ralph Lim, and Bridget Lyons. 2017. “Estimating the Cost of Equity in Emerging
Markets: A Case Study.” American Journal of Management 17 (2).
Diamonte, Robin L., John M. Liew, and Ross L. Stevens. 1996. “Political Risk in Emerging and
Developed Markets.” Financial Analysts Journal 52 (3): 71–76.
Doh, Jonathan P., Peter Rodriguez, Klaus Uhlenbruck, Jamie Collins, and Lorraine Eden. 2003.
“Coping with Corruption in Foreign Markets.” Academy of Management Perspectives 17 (3): 114–
27.
Donadelli, Michael, Marco Fasan, and Barbara Sveva Magnanelli. 2014. “The Agency Problem,
Financial Performance and Corruption: Country, Industry and Firm Level Perspectives.” European
Management Review 11 (3–4): 259–72.
Donadelli, Michael, and Lauren Persha. 2014. “Understanding Emerging Market Equity Risk Premia:
Industries, Governance and Macroeconomic Policy Uncertainty.” Research in International
Business and Finance 30 (January): 284–309.
Dreher, Axel, and Lars HR Siemers. 2005. “The Intriguing Nexus between Corruption and Capital
Account Restrictions.” Swiss Institute for Business Cycle Research (KOF) Working Paper No.
113.
Durham, J. Benson. 2004. “Absorptive Capacity and the Effects of Foreign Direct Investment and
Equity Foreign Portfolio Investment on Economic Growth.” European Economic Review 48 (2):
285–306.
Dutt, Pushan, and Daniel Traca. 2010. “Corruption and Bilateral Trade Flows: Extortion or Evasion?”
The Review of Economics and Statistics 92 (4): 843–60.
Egger, Peter, and Hannes Winner. 2006. “How Corruption Influences Foreign Direct Investment: A
Panel Data Study.” Economic Development and Cultural Change 54 (2): 459–86.
Fernandez, P., Alberto, O., Acin, I. 2016. Market Risk Premium Used in 71 Countries in 2016: A
Survey with 6,932 Answers. Journal of International Business Research and Marketing 2 (6): 23-
31.
Faruq, Hasan, Michael Webb, and David Yi. 2013. “Corruption, Bureaucracy and Firm Productivity in
Africa.” Review of Development Economics 17 (1): 117–29.

725
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Frazier Economic Freedom Report, [Link]

the-world-2019-annual-report
Gaviria, Alejandro. 2002. “Assessing the Effects of Corruption and Crime on Firm Performance:
Evidence from Latin America.” Emerging Markets Review 3 (3): 245–68.
Gupta, Sanjeev, Hamid Davoodi, and Erwin Tiongson. 2001. “Corruption and the Provision of Health
Care and Education Services.” In The Political Economy of Corruption, 123–53. Routledge.
Gelos, R. Gaston, and Shang-Jin Wei. 2005. “Transparency and International Portfolio Holdings.” The
Journal of Finance 60 (6): 2987–3020.
Gwartney, James, Robert Lawson, and Joshua Hall. 2019. Economic Freedom of the World: 2019
Annual Report.
Hooper, Vince, and Suk-Joong Kim. 2007. “The Determinants of Capital Inflows: Does Opacity of
Recipient Country Explain the Flows?” Economic Systems 31 (1): 35–48.
Houston, Douglas A. 2007. “Can Corruption Ever Improve an Economy?” Cato Journal 27: 325.
Javorcik, Beata S., and Shang-Jin Wei. 2009. “Corruption and Cross-Border Investment in Emerging
Markets: Firm-Level Evidence.” Journal of International Money and Finance 28 (4): 605–24.
Lambsdorff, Johann Graf. 2003. “How Corruption Affects Persistent Capital Flows.” Economics of
Governance 4 (3): 229–43.
Lang, Mark, Karl V. Lins, and Mark Maffett. 2012. “Transparency, Liquidity, and Valuation:
International Evidence on When Transparency Matters Most.” Journal of Accounting Research 50
(3): 729–74.
Lee, Charles MC, and David Ng. 2009. “Corruption and International Valuation: Does Virtue Pay?”
The Journal of Investing 18 (4): 23–41.
Lehkonen, Heikki, and Kari Heimonen. 2015. “Democracy, Political Risks and Stock Market
Performance.” Journal of International Money and Finance 59 (December): 77–99.
Mathur, Aparna, and Kartikeya Singh. 2013. “Foreign Direct Investment, Corruption and Democracy.”
Applied Economics 45 (8): 991–1002.
Mauro, Paolo. 1995. “Corruption and Growth.” The Quarterly Journal of Economics 110 (3): 681–712.
McArthur, John, and Francis Teal. 2002. “Corruption and Firm Performance in Africa.” University of
Oxford, Department of Economics, Economics Series Working Paper: WPS/ 2002-10.
Papaioannou, Elias. 2009. “What Drives International Financial Flows? Politics, Institutions and Other
Determinants.” Journal of Development Economics 88 (2): 269–81.
Quazi, Rahim M. 2014. “Corruption and Foreign Direct Investment in East Asia and South Asia: An
Econometric Study.” International Journal of Economics and Financial Issues 4 (2): 231–42.
Rodriguez, Peter, Donald Siegel, Amy Hillman, and Lorraine Eden. 2006. “Three Lenses on the
Multinational Enterprise: Politics, Corruption, and Corporate Social Responsibility.” Journal of
International Business Studies 37 (6): 733–46.

726
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Rose-Ackerman, Susan. 1975. “The Economics of Corruption.” Journal of Public Economics 4 (2):
187–203.
Sahakyan, Narek, and Kyle W. Stiegert. 2012. “Corruption and Firm Performance.” Eastern European
Economics 50 (6): 5–27.
Saiegh, Sebastian M. 2005. “Do Countries Have a ‘Democratic Advantage’? Political Institutions,
Multilateral Agencies, and Sovereign Borrowing.” Comparative Political Studies 38 (4): 366–87.
Schultz, Kenneth A., and Barry R. Weingast. 1996. The Democratic Advantage: The Institutional
Sources of State Power in International Competition. Essays in Public Policy 67. Stanford, CA:
Hoover Institution on War, Revolution, and Peace, Stanford University.
———. 1998. “Limited Governments, Powerful States.” In Strategic Politicians, Institutions, and
Foreign Policy, R. M. Siverson, 15–50. Ann Arbor: University of Michigan Press.
———. 2003. “The Democratic Advantage: Institutional Foundations of Financial Power in
International Competition.” International Organization 57 (1): 3–42.
Straub, Stephane. 2008. “Opportunism, Corruption and the Multinational Firm’s Mode of Entry.”
Journal of International Economics 74 (2): 245–63.
Swaleheen, Mushfiq us, and Dean Stansel. 2007. “Economic Freedom, Corruption, and Growth.” Cato
Journal 27 (3): 18–25.
Wei, Shang-Jin. 2000. “How Taxing Is Corruption on International Investors?” Review of Economics
and Statistics 82 (1): 1–11.
Wei, Shang-Jin, and Andrei Shleifer. 2000. “Local Corruption and Global Capital Flows.” Brookings
Papers on Economic Activity, No. 2: 303–46.
Wei, Shang-Jin, and Yi Wu. 2002. “Negative Alchemy? Corruption, Composition of Capital Flows,
and Currency Crises.” In Preventing Currency Crises in Emerging Markets, 461–506. University
of Chicago Press.
Zhao, John Hongxin, Seung H. Kim, and Jianjun Du. 2003. “The Impact of Corruption and
Transparency on Foreign Direct Investment: An Empirical Analysis.” Management International
Review 43 (1): 41.

727
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 3: Emerging Market Equity Returns Segmented into high and Low Institutional Indices

(1) (2) (3) (4) (5) (6) (7)

High Low High Low High Low
VARIABLES Total Corruption Corrption Demo- Demo- Trans- Trans-
Returns Total Total cracy cracy parency parency
Returns Returns Total Total Total Total
Returns Returns Returns Returns

Emerging Market 0.557* 0.552* 0.546* 0.594* 0.487* 0.543* 0.557***

Returns (0.0464) (0.0670) (0.0715) (0.0531) (0.0762) (0.0600) (0.0678)
Bribes (log) 7.027 -1.767 1.578 -3.141 3.331 3.699 -6.727
(9.370) (26.48) (15.07) (12.67) (15.60) (15.75) (12.90)
Democracy (log) -18.68** -15.61 -22.73 -21.83 -4.326 -14.89 -25.31*
(9.068) (13.26) (17.08) (28.67) (13.12) (12.84) (13.71)
Regulation (log) -13.44 -0.615 -26.46 -19.91 -88.74* -11.27 -9.109
(15.31) (23.39) (27.91) (13.95) (45.80) (14.47) (32.19)
Transparency (log) 15.46 -4.360 65.60*** -3.055 86.71*** -10.86 75.55***
(14.41) (24.00) (24.34) (18.66) (31.89) (21.12) (23.32)
Percent Foreign 17.40 40.28 5.704 2.817 18.21 7.441 21.38
Ownerhsip(log) (14.69) (25.17) (20.65) (21.29) (21.11) (25.20) (19.39)

Growth (log) 3.644* 4.364 2.767 2.825 5.863* 4.436* 4.466

(1.936) (2.971) (2.816) (2.252) (3.234) (2.638) (2.724)
Tax Rate (log) 28.16*** 28.43 33.81* 14.37 111.0*** 7.288 68.27***
(10.40) (19.47) (19.10) (10.38) (27.79) (10.31) (26.03)
Current Account. 1.408*** 0.291 3.444*** 0.148 3.117*** 0.0344 4.035***
(0.466) (0.621) (1.073) (0.543) (1.019) (0.505) (0.960)
Real Interest 0.343 0.729 -0.152 0.173 0.843 0.0579 0.621
(0.366) (0.553) (0.613) (0.476) (0.665) (0.499) (0.544)
Constant -146.1*** -172.5* -164.9** -11.81 -358.3*** -9.468 -349.1***
(48.08) (100.2) (78.59) (55.44) (114.1) (57.51) (93.77)
Observations 286 142 144 156 130 135 151
R-squared 0.480 0.468 0.466 0.549 0.524 0.517 0.549
Number of id 20 19 16 17 16 15 13
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

728
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 4: Emerging Market Equity Flows Segmented into high and Low Institutional Indices

(1) (2) (3) (4) (5) (6) (7)

High Low High Low High Low
VARIABLES Total Corruption Corrption Democracy Democracy Trans- Trans-
Equity Flow Equity Flow Equity Flow Equity Flow Equity Flow parency parency
Equity Flow Equity Flow

Emerging Market 94.43* 96.74* 108.9* 125.4* 74.34* 96.48* 107.8***

Returns (17.64) (31.57) (20.69) (28.51) (19.15) (33.26) (19.90)
Bribes (log) -9,660*** -15,891 -551.1 -10,037 -347.7 -8,933 -5,609
(3,560) (12,549) (4,360) (6,791) (3,923) (8,692) (3,788)
Democracy (log) 269.4 -6,476 4,723 -17,416 8,938*** -15,969** 7,615*
(3,495) (6,486) (4,941) (15,006) (3,333) (7,434) (4,026)
Regulation (log) 5,490 21,864* -3,970 4,787 21,974* 4,723 2,987
(5,793) (11,082) (8,074) (7,585) (11,515) (7,996) (9,450)
Transparency (log) 1,320 -8,710 -71.17 3,616 -9,878 3,049 832.4
(5,483) (11,368) (7,042) (10,094) (8,006) (11,727) (6,844)
Percent Foreign 5,303 11,859 4,631 -8,844 10,594* -13,102 5,178
Ownership (log) (5,580) (11,853) (5,974) (11,538) (5,361) (13,934) (5,693)
Growth (log) 1,517** 564.0 2,233*** 1,209 1,163 789.9 1,804**
(736.4) (1,411) (814.8) (1,210) (823.6) (1,466) (799.6)
Tax Rate (log) 2,621 -7,238 6,540 4,976 5,980 2,877 15,277**
(3,942) (9,199) (5,525) (5,569) (6,843) (5,693) (7,641)
Current Account 48.59 106.6 -385.5 240.6 -192.6 90.11 -72.91
(176.5) (293.7) (310.4) (290.3) (245.4) (279.3) (281.9)
Real Interest 18.06 245.5 18.90 -164.8 248.6 128.7 -44.99
(139.3) (261.9) (177.4) (257.8) (168.3) (278.7) (159.8)
Constant -14,279 34.58 -18,700 -3,069 -60,177** 12,250 -53,239*
(18,280) (47,332) (22,739) (30,410) (28,733) (31,875) (27,526)

Observations 285 141 144 141 145 134 151

R-squared 0.136 0.139 0.268 0.122 0.308 0.121 0.281
Number of id 20 19 16 16 17 15 13
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

729
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 5: Emerging Market Equity Returns by Industry

(1) (3) (4) (5) (6) (7) (8) (9)

VARIABLES Basic Consumer Financial Health- Tech- Gas Industrial Utility
Goods Services care nology

Emerging Market 0.815*** 0.533*** 0.696*** 0.543*** 0.491*** 0.505*** 0.656*** 0.653***
Returns (0.0955) (0.110) (0.113) (0.0968) (0.134) (0.0960) (0.0927) (0.0936)
Bribes (log) 18.89 -40.01* 24.71 -16.68 -49.16 14.19 -5.096 -52.13**
(19.53) (22.94) (22.69) (21.30) (35.29) (22.46) (18.87) (20.39)
Democracy (log) -2.883 -30.11 3.675 -8.189 -82.89** -50.90*** -3.548 -36.25**
(18.86) (22.54) (22.89) (20.33) (35.71) (18.77) (18.43) (18.32)
Regulation (log) 3.565 9.931 -2.423 15.10 -67.52 15.40 9.506 -54.77*
(32.37) (35.66) (35.03) (33.12) (42.33) (34.78) (29.81) (30.71)
Transparency (log) 2.776 1.524 19.97 26.17 15.70 23.81 25.40 14.82
(29.13) (36.98) (34.13) (30.40) (48.81) (33.03) (29.15) (31.08)
Percent Foreign 35.89 31.19 2.176 -0.115 21.30 -0.597 43.23 106.2***
Ownership (log) (29.47) (38.05) (34.85) (32.09) (47.77) (32.42) (29.01) (30.64)
Growth (log) 5.894 8.946** 4.898 7.667* 2.466 1.773 2.344 2.251
(3.870) (4.524) (4.648) (4.004) (5.910) (4.164) (3.904) (4.190)
Tax Rate (log) -18.84 114.2*** 29.98 7.618 33.79 39.82* 22.90 38.31*
(31.11) (31.46) (25.62) (23.87) (32.13) (23.72) (21.46) (20.61)
Current Account 1.089 2.734** 3.095*** 0.234 -0.991 2.321** 1.502 1.111
(1.010) (1.142) (1.153) (0.918) (1.361) (0.893) (0.950) (0.920)
Real Interest 0.175 -0.0106 0.308 0.218 1.540 -0.358 0.666 0.481
(0.740) (0.876) (0.935) (0.833) (1.326) (0.706) (0.743) (0.731)
Constant -49.77 -422.0*** -166.7 -62.01 2.176 -247.5** -205.4** -182.0*
(142.2) (144.7) (121.0) (108.3) (151.3) (101.6) (101.6) (96.20)

Observations 253 261 264 203 137 224 272 248

R-squared 0.326 0.228 0.242 0.185 0.164 0.259 0.237 0.272
18 19 19 15 13 16 19 19
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.10

730
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 6: Emerging Market Risk Premium Segmented into High and Low Institutional Indices
(1) (2) (3) (4) (5) (6) (7)
VARIABLES Total High Low High Low High Low
Returns Corrup- Corrup- Demo- Demo- Trans- Trans-
tion tion cracy cracy parency parency

Emerging Market -0.0001 0.001 -0.0009 -0.0002 0.0007 0.0005 -0.001

Returns (0.0006) (0.001) (0.0008) (0.0009) (0.001) (0.0009) (0.001)
Bribes (log) -0.245** 0.261 -0.315** -0.190 0.0684 -0.121 -0.315
(0.107) (0.604) (0.127) (0.149) (0.248) (0.201) (0.190)
Democracy (log) -0.200 -0.492 -0.269 -0.515 -0.254 -0.194 -0.396
(0.217) (0.775) (0.327) (0.727) (0.231) (0.265) (0.407)
Regulation (log) 0.083 -0.546 0.300 0.0509 -0.633 0.060 0.065
(0.168) (0.338) (0.267) (0.204) (0.456) (0.189) (0.415)
Transparency (log) 0.319** 0.201 0.438** 0.239 0.142 0.187 0.417
(0.147) (0.228) (0.216) (0.197) (0.434) (0.192) (0.418)
Percent Foreign 0.182 0.900 0.239 0.230 0.185 0.0230 0.235
Ownership (log) (0.195) (0.848) (0.227) (0.239) (0.488) (0.312) (0.305)
Growth (log) -0.005 0.029 -0.002 -0.012 -0.042 0.017 -0.006
(0.017) (0.049) (0.021) (0.024) (0.030) (0.029) (0.026)
Tax Rate (log) -0.181 1.105 -0.302 -0.216 -1.230 -0.212 0.0290
(0.157) (1.905) (0.186) (0.185) (1.348) (0.166) (0.618)
Current Account -0.002 -0.015 -0.001 0.005 -0.035** -0.011 0.003
(0.007) (0.013) (0.012) (0.010) (0.016) (0.009) (0.019)
Real Interest 0.009** 0.005 0.008 0.013** -0.009 -0.005 0.01
(0.004) (0.012) (0.005) (0.005) (0.007) (0.009) (0.007)

Constant 1.858* -3.702 1.704 1.816 6.656 2.273* 1.034

(0.588) (7.782) (0.682) (0.814) (4.831) (0.701) (2.149)

Observations 114 34 80 75 39 57 57
R-squared 0.202 0.306 0.279 0.208 0.444 0.159 0.304
Number of ids 20 8 16 14 9 12 12
Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1

731
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Appendix

Table 7: Institutional Segmentation of Emerging Market Countries by years

High Low High Low High Low

Countries Corrution Corruption Transparency Transparency Democracy Democracy
2000, 2002,
Argentina 2003, 2004
2005-2017
- All year 2004-2017 2000-2003

2000, 2004,
Brazil 2000-2005 2006-2017
2013-2017
2006-2012 2006-2017 2000-2005

Chile All years _ All years _ All years _

2000-2008,
China 2015-2017
2009-2014 2016, 2017 2000-2015 - All years

Colombia 2000-2006 2007-2017 - All year 2015- 2017 2000-2014

Czech 2000,2001,2
2002-2007 All years _ All years _
Republic 008-2017
2000-2008,
Greece 2000-2006 2007-2017
2013-2017
2009-2012 All years _

Hong Kong All years _ All years _ All years _

Hungary 2000-2007 2008-2017 All years _ All years _

2004,2005, 2000-2003,
India 2016-2017 2006-2014
2016, 2017 2000-2015 2006-2017 2000-2005

Korea 2000-2009 2010-2017 All years _ All years _

Malaysia All years _ All years _ - All year

2000-2005,
Mexico 2000-2006 2007-2017 - All year 2006-2014
2015-2017

Pakistan NA NA NA NA NA NA

Peru 2000-2006 2007-2017 2000-2002 2003-2017 - All year

Phillippines - All year - All year 2014-2017 2000-2013

732
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2000,2002,20 2001,2003- 2000-2002,

Poland 2003-2006 All years _
14-2013 2017 2007-2017

Qatar All years _ All years _ - All years

Singapore All years _ All years _ All years _

South Africa 2014-2017 2000-2013 All years _ All years _

Sri Lanka 2002 2003-2017 All years _ 2015-2017 2000-2014

Taiwan All years _ All years _ All years _

Thailand 2000-2007 2008-2017 2000-2007 2008-2017 2000-2001 2002-2017

2000,2002,2
2001,2004,20 003, 2005-
Turkey 2000-2006 2007-2017 - All years
08-2013 2007,2014-
2017
UAE All years _ All years _ - All years

733
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 8: Independent Variable Correlation Matrix

Foreign Real
Demo- Regula Trans- Owner- Tax Interest Current
cracy -tion parency Bribes ship Rate Growth Rate Account
Demo-
1.0000
cracy
Regula-
0.1179 1.0000
tion
Trans-
0.5072 0.0881 1.0000
parency
Bribes 0.1610 0.0470 0.6815 1.0000

Foreign
Owner- 0.4070 0.0601 0.6094 0.6342 1.0000
ship
Tax Rate -0.1647 0.0331 -0.3244 -0.0940 -0.2044 1.0000

Growth -0.1848 -0.0264 -0.0157 0.1642 0.1174 0.0751 1.0000

Real
Interest -0.0107 -0.1767 -0.0807 0.0407 0.0206 0.0446 -0.0112 1.0000
Rate
Current
-0.0322 0.0798 0.5344 0.4954 0.2980 -0.3770 0.1524 -0.1628 1.0000
Account

734
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 9: Instittutional Variable Definitions

Variable Source Description

“Voice and accountability captures perceptions of the extent to
which a country's citizens are able to participate in selecting their
Democracy Economic Intelligence Unit (EIU) government, as well as freedom of expression, freedom of
association, and a free media.”

“Ranksks 180 countries and territories by their perceived levels of

The Transparency International public sector corruption, according to experts and business
Transparency Corruption Perceptions Index (CPI) people.”

“This sub-component is based on the Global Competitiveness

Report questions: (1) “In your industry, how commonly would you
estimate that firms make undocumented extra payments or bribes
connected with the following: A – Import and export permits; B –
Connection to public utilities (e.g., telephone or electricity); C –
Annual tax payments; D – Awarding of public contracts
(investment projects); E – Getting favourable judicial decisions.
Common (= 1), Never occur (= 7)”. (2) “Do illegal payments
Bribes Frazier Economic Freedom Report
aimed at influencing government policies, laws or regulations
have an impact on companies in your country? 1 = Yes, significant
negative impact, 7 = No, no impact at all”. (3) “To what extent do
government officials in your country show favouritism to well-
connected firms and individuals when deciding upon policies and
contracts? 1 = Always show favouritism, 7 = Never show
favouritism”. The wording of the questions has varied slightly over
the years.”

735
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Investigation of the Effects of Non-Macroeconomic Indicators on the Exports of Goods and Services in the

OECD Member Countries

Associate Professor Dr. Sevgi SEZER

University of Balıkesir

Department of Banking and Finance

Turkey

Abstract

In the literature dealing with the factors that influence the export performances of countries, studies based on micro-

level and macro-level variables mostly encountered. This study, however, aims to explicate the effects of non-economic

variables such as trade freedom index, R&D expenditures, number of patent entitlements, and the number of newly

established companies on exports of goods and services using the related data obtained from 36 OECD member

countries over the period 2010 - 2018 via a dynamic panel data analysis method. The obtained findings of the study

reveal that trade freedom has a negative impact on the percentage share of exports of goods and services in the GDP,

whereas both R&D expenditures and the number of newly established companies have positive effects on the

percentage share of exports of goods and services in the GDP, and the number of patent entitlements has no impact at

all.

Keywords: Export Performance, Trade Freedom Index, R&D, Patent Entitlements, OECD.

1. Introduction

Along with globalization, the world countries tend to export their domestic products to benefit from international

market opportunities to achieve a reliable market position and to maintain a sustainable and highly competitive market

domain. The most important factor that determines and affects the competitiveness of countries and in the economic

life is the export performance and value-added created by the enterprises of that country in both domestic and foreign

markets. Although export performance is a widely studied subject, there is no clarity on its definition. Conceptually,

export performance is generally used in terms of efficiency, competence and interest in exporting (Thorelli and Tesar,

1990; Katsikeas, Piercy and Ioannidis, 1996; Voerman, Wedel and Zwart, 1999; Sousa, 2004; Cadogan, Kuivalainen

and Sundqvist, 2009). Therefore, the measurement of export performance generally includes export intensity,

perceived profitability, level of satisfaction, and continuing export activity (Wedel and Zwart, 1999; McGeehan, 1968;

Reid, 1983; Stöttinger and Schlegelmilch, 1998). Wedel and Zwart, 1999 states that the conceptual definition of export

performance should point to export and performance, which are two terms of the concept, separately. Export, in

general, is defined as the marketing of goods produced in domestic markets toward foreign markets as a stage of

736
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

internationalization. In this way, exports are seen as the decisions and activities of internationally active firms regarding

international and marketing (Çavuşgil and Nevin, 1981). There have been many studies conducted on the determinants

of export performance (Miesenbock, 1988; Katsikeas, Piercy and Ioannidis, 1996; Leonidou, Katsikeos and Samiee,

2002). Some of these studies concentrate on environmental factors as the determinants of export performance

(Diamontopoulos and Inglis, 1988; Beamish 1985; Mc Dougall and Oviatt 1996), whereas others focus on intra-

industry factors (Ansari, Aafaqi, and Jayasingam, 2000). Theoretical and empirical studies indicate that both

environmental factors and firm resources have certain influences on export performance (Lages, 1999). For example,

Katsikeas, Piercy, and Ioannidis, 1996 divide the factors that determine export performance into three groups as

environmental, organizational, and strategic factors. On the other hand, it is seen in other studies that factors affecting

export performance are usually classified under two groups as external environmental factors and internal

environmental factors (Lages, 1999; Thorelli and Tesar, 1990). In accordance with this distinction, the overall work

conducted later focuses on either environmental factors or intra-industry factors as determinants of export performance.

There are many factors that influence international trade transactions. In this context, the effects of macroeconomic as

well as non-macroeconomic factors on exports can be mentioned. Pioneering studies have mostly dealt with the effects

of macroeconomic factors. Nonetheless, today, there are findings claiming that various economic factors have effects

on exports.

The objective of this study is to examine whether or not trade freedom index, R&D expenditures, the number of

companies established, and the number of patent entitlements received affect the exports of goods and services in the

OECD member countries. With the help of data obtained from 36 OECD member countries between 2010-2017, the

GMM analysis is performed to examine the subject. Accordingly, this study consists of five parts. Following the

introduction in the first part, the pertinent literature is reviewed in the second part. In the third part, the data and

methodology of the analyses are introduced. Obtained findings of the performed analyses are presented in the fourth

part. Consequently, brief conclusive policy recommendations are made in the last part.

2. Literature Review

In the literature, studies conducted on macroeconomic factors are mostly emphasized. However, unlike the current

literature, this study considers the research studies in which the subject was explicated with non-macroeconomic

factors.

Upon analyzing the relationship between non-macroeconomic factors and export performance, it is seen that the most

widely used variables appear to be R&D expenditures and patent entitlements as well as newly established companies.

Studies in this literature are the studies conducted within a similar framework.

737
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Hulst, Mulder, and Soete (1991) investigated the association between international trade and technology and of five

OECD countries, namely Germany, Sweden, Netherlands, Japan, France, in three different years (1979, 1983 and

1987) with cross-sectional regression method. The study tested whether or not a country’s or industry’s strong

technology capability has an impact on the foreign trade of that country or industry. The study found that creating a

technological advantage for Germany, Sweden, the Netherlands, and partly for Japan provides a competitive advantage

in foreign trade. However, for France, this result could not be confirmed.

Amable and Verspagen (1995) examined 18 industries in five industrialized OECD member countries over the period

1970-1991 and found that patents, namely commercial innovations, had positive effects on international trade

performance. The study tried to determine the importance of price and non-price factors in determining international

competitiveness. A dynamic model of the export market stocks adapted from Magnier and Toujas-Bernatte (1992) was

applied to a dataset of OECD-export market shares, wages, investment, and a patent for a sample.

Verspagen and Wakelin (1997) analyzed the data of nine OECD member countries regarding their mutual trade

between 1970-1978 and 1980-1988 using a dynamic panel data method. The results of the study revealed that R&D

intensity had a significant and positive effect on exports.

Narula and Wakelin (1998) examined the association between both export performance and the ratio of foreign direct

investments and variables that affect innovation with the 4-year data (1975, 1979, 1984, 1989) on 41 developed and

developing countries. As a result of the study, the positive effect of the patent variable on export performance in

developed countries was determined. In developing countries, it was concluded that the patent variable had a negative

effect on export performance.

Wakalin (1998) analyzed reciprocal international trade flows in nine OECD member countries and 22 manufacturing

industry sectors using the OLS regression method over the period 1988-1992. According to the results of the study, an

affirmative correlation was found between total international trade of countries and export performances related to

some sectors, whereas the R&D expenditures were determined to have an adverse impact on the countries’ exporting

activities.

Montobbio and Rampa (2005) investigated the association between innovation levels and export performance in nine

developing countries during the 1985-1998 period using the structural decomposition method. According to the results

of the study, technological activities increased export performance. It was concluded that China, Malaysia, Singapore,

and Thailand, together with their high levels of innovation, have significantly increased their importance in world

exports.

738
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Uzay, Demir, and Yıldırım (2012) tested the relationship between the exports of manufacturing sectors and R&D

expenditure in Turkey over the period 1995-2005. It was found that R&D expenditures had a significant but rather

delayed effect on exports.

Tekin and Hancıoğlu (2017) utilized the data of the developing countries included in the Global Innovation Index

between 201 -2015 using the panel data analysis method. According to the results of the research, it was concluded

that innovation had a positive effect on export performance in developing countries. Furthermore, the study stated that

especially R&D activities had an important role in export performance of developing countries.

3. Data and Methodology

The sample of this study is constituted by 36 OECD member countries (Australia, Austria, Belgium, Canada, Chile,

Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Japan,

Korea, Rep., Latvia, Lithuania, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Slovak

Republic, Slovenia, Spain, Sweden, Switzerland, Turkey, the United Kingdom, and the United States). The study

period is determined as between the years 2010-2017 in annual frequency depending on data constraints. The data used

in their analysis are obtained from various sources. The ratio of the exports of goods and services to the gross domestic

product (EXPGDP) data included in the study as a dependent variable is obtained from the World Bank and OECD

national accounts databases. The related data include the total price of the product, freight, insurance, and transfer in

the export of goods and services, and indicate the proportional weight of the amount of exports in the GDP. Trade

freedom refers to the basic right of all individuals to control their labor and property. Individuals are free to work,

produce, consume and invest the way they wish to do so in economically free societies where governments allow labor,

capital, and goods to move freely and avoid coercion or restriction of freedom beyond what is necessary to protect and

maintain freedom. States determine the institutional and political framework within which individuals, businesses, and

governments operate. “Trade Freedom Index” has been reported by the Heritage Foundation since 1995. The relevant

index consists of property rights, the rule of law, government integrity, tax burden, government spending, financial

health, business freedom, labor freedom, monetary freedom, commercial freedom, investment freedom, and

commercial freedom items. The Trade Freedom Index documents the affirmative association between economic

freedom and various positive social and economic objectives. Trade freedom ideals are strongly associated with

healthier societies, cleaner environments, more prosperity per capita, human development, democracy, and eradication

of poverty. Countries with a high degree of trade freedom tend to be more prone to improvement as individuals benefit

more from their ability to innovate and develop when they are released, with heavy government regulations and

taxation. The free market system promotes the most efficient allocation of resources and creates a dynamic

739
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

environment that maximizes available options for production and consumption. The Trade Freedom (TF) included in

the study as a dependent variable is part of the economic freedom index, which has been calculated by the Wall Street

Journal and the Heritage Foundation. The index value ranges from 0 to 100, 40 - 49 being not free, 50 - 59 being mostly

not free, 60 - 69 being partially free, 70 - 79 being mostly free, and 80 - 100 being free. The trade freedom included in

the economic freedom index has been calculated by taking the average customs tariff rate weighted with commercial

goods and non-tariff barriers into consideration. The ratio of research and development (R&D) expenditures to the

GDP (RDGDP), as another independent variable used in the analysis, is obtained from the World Development

Indicators database of the World Bank. The relevant data indicate the proportional share of the budgets reserved for

research and development activities in the countries’ domestic incomes. Similarly, another data obtained from the

World Development Indicators database of the World Bank involve the number of newly established companies (NBR)

and indicate the number of companies established in status with limited liability. The last variable used in the analysis

is the number of patent entitlements (TP). The data obtained from the World Bank, World Intellectual Rights

Organization database for each member country included in the sample show the annual patent applications made by

citizens and non-citizens in each country separately. Since the variables used in the analyses are on different scales,

the logarithms of EXPGDP, TF, NBR, TP variables are taken. However, since the RDGDP variable can take values

below 1, no logarithmic transformation is applied for it. Descriptive statistics of the variables included in the analysis

can be seen in Table 1.

Table 1 Descriptive Statistics

EXPGDP TF RDGDP NBR TP

Mean 1.653800 1.935800 1.914900 4.430800 3.4834

Maximum 2.344700 1.954200 4.553200 5.821900 5.7831

Minimum 1.074400 1.850000 0.329400 3.207900 1.3979

Std. Dev. 0.245200 0.016600 1.011600 0.525900 0.9982

# of Obs. 288 288 269 267 275

Upon examining Table 1, it can be seen that the highest standard deviation is in the ratio of research and development

expenditures to GDP (RDGDP). The highest average is in the NBR series, indicating the number of newly established

companies. Also, a striking fact in Table 1 is that the numbers of observations for each variable differ on a serial basis.

It is either partially or completely impossible to access some of the data of the variables included in the analyses.

Therefore, the analyses are performed with an unbalanced panel. The analyses of the study are performed via the GMM

(Generalized Moments Method), also known as the Dynamic panel, which was first proposed by Arellano and Bond

(1991). Later on, this method was improved by Blundell and Bond (1998), and the system GMM methodology was

740
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

proposed in the related study. Neither methodology has an advantage over each other. In this study, the difference

GMM approach developed by Arellano and Bond (1991) is used in a way that two-step estimators are computed.

4. Findings

Correlations between the independent variables included in the analyses can be examined in Table 2. The findings in

Table 2 reveal that trade freedom and all other independent variables have negative correlations. Besides, the highest

correlation coefficient in the Table is 0.57, which falls between the number of newly established companies and the

total number of patent entitlements.

Table 2 Correlation Matrix

TF RDGDP NBR TP

TF 1

RDGDP -0.2057 1

NBR -0.2046 -0.0300 1

TP -0.5186 0.4374 0.5718 1

The Two-Step Arellano-Bond Difference GMM results are presented in Table 3. According to Table 3, the Wald test

results exhibiting the consistency of the GMM estimators indicate the overall significance of the variables used in the

analyses, the Sargan test results reveal the validity of the over-determination constraints, and the AR2 test results assert

that there is no second-order autocorrelation in the model. Under these conditions, the estimate can be determined to

be valid and coefficient estimators can be interpreted.

Table 3 Arellano-Bond Difference GMM Results

Coefficient Std. Dev. Probability

Constant 1.9225 0.1022 0.000

Lagged (MHGDP) 0.3398 0.0190 0.000
TF -0.5015 0.0418 0.000
RDGDP 0.0105 0.0035 0.003
NBR 0.0298 0.0065 0.000
TP 0.0042 0.0061 0.487
Wald Chi2 0.000
Sargan 26.3967 0.153
AR2 -0.8390 0.401

The estimator coefficients in Table 3 indicate that trade freedom has a statistically significant and adverse effect on the

proportional share of exports of goods and services that take place in the GDP. The fact that countries with low trade

freedom included in the sample tend to use their foreign exchange rate advantage and increase their exports is thought

741
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

to be the underlying reason. Upon evaluating the OECD member countries in the sample regardless of being developed

or developing, it is seen the relevant data are heterogeneously distributed. In this case, trade freedom levels of

developed countries with more stable exports are rather higher than those of developing countries. The trade freedom

levels of the developing countries are relatively low, whereas their exports tend to increase faster due to the exchange

rate advantages. Subsequently, the exports of goods and services proportionally included in the GDP of developed

countries are more likely to fluctuate, and the analysis results in support of this finding are evaluated. Another result

obtained involves the fact that the percentage share of research and development expenditures in the GDP has a

significant and positive effect on the percentage share of exports of goods and services in the GDP. Accordingly, the

investments made by countries in research and development activities cause their exports to increase and the incurred

costs of research and development activities can be justified in this way. Similarly, it can be said that exports are

positively and significantly affected as long as the number of newly established companies increases according to

Table 3. This situation can be interpreted as referring to the fact that the newly established companies perform direct

export activities or they indirectly provide the firms engaged in export activities with goods and services, thereby

causing the rise in the country’s exports. Undoubtedly, the most surprising result obtained from the study is that the

number of patent entitlements does not have any significant impact on the share of export of goods and services in the

GDP. The number of patent entitlements provides the advantage of protecting intellectual and industrial rights.

However, the increase in the number of patent entitlements does not necessarily indicate that it has made a statistically

significant contribution to the national economy through the export of products subject to patents by changing the

current structure of the country’s exports. In other words, the share of patented products in total exports is considered

to be relatively low. Nevertheless, in order to support this idea, no data has been obtained indicating how much of the

exported products would consist of patented products.

5. Conclusions and Discussion

It can be said that there are many factors that affect export performance. In the current literature, these factors are

mostly considered as macroeconomic factors. Whether or not non-macroeconomic factors have impacts on the exports

has been the subject of merely a limited number of studies. In this study, the export performances of OECD member

countries are analyzed over the period 2010-2017 and it is tried to determine the effects of trade freedom index, research

and development expenditures, the number of newly established companies and the total number of patent entitlements

each year on the exports via the GMM analysis. Accordingly, it is determined that the trade freedom factor negatively

affects the exports of goods and services. This result is basically due to the heterogeneous structure of the data obtained

from OECD member countries in the sample. Trade freedoms of developing countries are rather low and their export

742
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

volatility is high. This situation may account for an unexpected result in this analysis. On the other hand, research and

development expenditures positively affect the share of exports in the GDP. The mentioned positive effect of research

and development activities is consistent with the results obtained in Verspagen and Wakelin (1997), Uzay, Demir and

Yıldırım (2012) and Tekin and Hancıoğlu (2017). However, the number of newly established companies positively

affects exports due to the fact that newly established businesses export directly and/or contribute to the activities of

exporting enterprises. No significant effect of the number of patent entitlements on the share of exports in the GDP is

detected. In the study of Amable and Verspagen (1995), there are findings indicating that the number of patent

entitlements positively affects export performance. Nonetheless, the study sample consists of only five developed

OECD member countries. In the study of Narula and Wakelin (1998), it is concluded that the number of patent

entitlements positively affects the exports in developed countries, whereas negatively in developing countries. The fact

that the number of patent entitlements does not have a significant effect on exports can be explained by the relatively

low share of patented products in the exported products. However, it is not possible to obtain the data supporting this

idea.

Upon evaluating the results obtained from the study in terms of economic decision-makers, it can be said that incentive

policies that increase the number of new businesses to be established and measures to boost research and development

activities would contribute to export performance. However, it would not be rational to suggest decision-makers about

increasing the number of patent entitlements since the findings on this subject in the current literature differ according

to the developments of the countries. At this point, some shortcomings of the study can be mentioned. For instance,

the sample group has a heterogeneous structure. Moreover, the study period examines only a limited process due to

the data constraint. In the studies to be carried out in the future, performing separate analyses in distinct country groups

with homogeneous features, such as developed and developing countries, can foster more striking results on the subject.

References

Amable, B., & Verspagen, B. (1995). The Role of Technology in Market Shares Dynamics. Applied Economics, 27,

197-204.

Ansari, M., Aafaqi, R., & Jayasingam, S. (2000). Entrepreneurial success, gender and leadership behavior/Mahfooz A.

Ansari, Rehana Aafaqi and Sharmila Jayasingam. Journal of International Business, Economics and

Entrepreneurship (JIBE), 8(2), 33-46.

Arellano, M., & Bond, S. (1991). Some Tests of Specification for Panel Data: Monte Carlo Evidence and An

Application to Employment Equations. The Review of Economic Studies, 58(2), 277-297.

743
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Beamish, P. W. (1985). The Characteristics of Joint Ventures in Developed And Developing-Countries. Columbia

Journal of World Business, 20(3), 13-19.

Blundell, R., & Bond, S. (1998). Initial Conditions and Moment Restrictions in Dynamic Panel Data Models. Journal

of Econometrics, 87(1), 115-143.

Cadogan, J. W., Kuivalainen, O., & Sundqvist, S. (2009). Export market-oriented behavior and export performance:

quadratic and moderating effects under differing degrees of market dynamism and

internationalization. Journal of international Marketing, 17(4), 71-89.

Çavuşgil, S. T., Nevin, J. (1981), “Internal Determinants of Export Marketing Behaviour: An Empirical Investigation”,

Journal of Marketing Research, February, (114-119).

Diamantopoulos, A., & Inglis, K. (1988). Identifying Differences between High‐and Low‐İnvolvement

Exporters. International Marketing Review.

Hulst, N. V., Mulder, R., & Soete, L. (1991). Exports and Technology in Manufacturing Industry. Weltwirtschaftliches

Archiv, 127 (2), 246-264.

Katsikeas, C. S., Piercy, N. F., & Ioannidis, C. (1996). Determinants of Export Performance in a European

Context. European Journal of Marketing.

Lages, L. F. (1999). Marketing lessons from Portuguese wine exporters: the development and application of a

conceptual framework. Journal of Wine Research, 10(2), 123-132.

Leonidou, L.C., C. S. Katsikeas ve S. Samiee, (2002), “Marketing Strategy Determinants of Export Performance: a

Meta-Analysis”. Journal of Business Research, 55 (1), ss. 51-67.

Magnier A. y J. Toujas-Bernate (1994). “Technology and Trade: Empirical Evidence for The Mayor Five Industrialized

Countries”, Weltwirtschaftiches Archiv, 131, 494-520.

McDougall, P. P., & Oviatt, B. M. (1996). New Venture Internationalization, Strategic Change, and Performance: A

Follow-Up Study. Journal of Business Venturing, 11(1), 23-40.

McGeehan, J. M. (1968). Competitiveness: a survey of recent literature. The Economic Journal, 78(310), 243-262.

Miesenbock, K. J. (1988). Small Businesses And Exporting: A Literature Review. International Small Business

Journal, 6(2), 42-61.

Montobbio, F., & Rampa, F. (2005). The Impact of Technology and Structural Change on Export Performance in Nine

Developing Countries. World Development, 33(4), 527-547.

Narula, R., & Wakelin, K. (1998).Technological Competitiveness, Trade, and Foreign Direct Investment. Structural

Change and Economic Dynamics, 9, 373-387.

Reid, S. (1983). Firm internationalization, transaction costs and strategic choice. International marketing review.

744
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Sousa, C. M. (2004). Export Performance Measurement: An Evaluation of the Empirical Research in the

Literature. Academy of Marketing Science Review, 2004, 1.

Stöttinger, B., & Schlegelmilch, B. B. (1998). Explaining export development through psychic distance: enlightening

or elusive?. International Marketing Review.

Tekin, E., & Hancıoğlu, Y. (2018). İnovasyon Belirleyicilerinin İhracat Performansına Etkisi Üzerine Bir

Araştırma. Uluslararası Yönetim İktisat ve İşletme Dergisi, 14(4), 897-917.

Thorelli, H. B., & Tesar, G. (1990). Entrepreneurship in international marketing: A continuing research

challenge (No. 37). Indiana Center for Global Business.

Uzay, N., Demir, M., & Yıldırım, E. (2012). İhracat Performansı Açısından Teknolojik Yeniliğin Önemi: Türkiye

İmalat Sanayi Örneği. Doğuş Üniversitesi Dergisi, 13(1), 147-160.

Verspagen, B., & Wakelin, K. (1997). Trade and Technology from a Schumpeterian Perspective. International Review

of Applied Economics, 11(2), 181-194.

Voerman, L. J., Wedel, M., & Zwart, P. S. (1999). The dynamics of exporting SMEs’ information

behaviour. International Council for Small Business, Washington, DC.

745
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Exploring the Dimensions Forming the HEIs Image for the Current International
Students: The Context of Malaysia

Shamima Raihan Manzoor1, Faculty of Management, Multimedia University Malaysia

Chinnasamy Malarvizhi, Faculty of Management, Multimedia University Malaysia
Junainah Mahdee, Faculty of Management, Multimedia University Malaysia

Abstract

In recent decades, Malaysia has witnessed a notable influx of international students in its HEIs
with strong support from the Malaysian government to build Malaysia as the educational centre
of the region. This is in line with the country’s aspiration to be a global education hub by the
year 20202. Among these rapidly developing countries, Malaysia has emerged as a top study
destination for students all around the world. Hence the investigation of image dimensions
from the perspectives of international students is a burning issue which requires further
research, especially in the context of Asian countries. As Malaysia is one of the emerging
contenders as an education hub with a sound number of international students, the country
deserves to be the context of research in this study. Since the past studies are mostly done in
the Western context, a gap still exists regarding the dimensions forming the university image
in Malaysia, especially among the international students.

The study involved 200 international postgraduate students studying five different universities
consisting of both private and public universities in Malaysia. A seven-point Likert scale for
15 items included on the questionnaire related to the academic institutional image. Principal
component analysis using varimax rotation was used for questionnaire validation and
categorisation of resulting factors. The factor analysis result showed three-dimensional
abstraction of the image construct. The three dimensions explored in the study are expected to
capture a broader spectrum in understanding the real essence of the university image due to the
nature of multiplicity of this study respondents.

This instrument would be considered applicable as a tool for image determination to be adopted
by the university prior and/or post-marketing campaigns performed by the university. From a
practical standpoint, the outcome of this study will provide ideas for the universities managers
in concentrating on all the factors instead of investing plenty of resources and efforts just on
one dimension. Finally, the main implication of this research is that universities should focus
on these appeals, since it seems that the quality of education is taken for granted by students
and, as such, may not be a differentiating factor.

Keywords: Higher Education Institutions Image, International Students, Malaysia

1
Corresponding author, Email: shamimaraihanmanzoor@[Link]
2
Vision 2020

746
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1. Introduction

The growth of international student mobility is now a well-documented facet of the

internationalisation of higher education (HE). According to OECD (2014), 4.5 million students
pursued higher education outside their home countries in 2012, and a number predicted to rise
to 7.2 million by 2025 (Bohm et al. 2002). The dominant flow of international students
continues to be from the Global South to the Global North (Stein and de Andreotti 2016), with
economically developed and Anglophone countries in North America, Europe and Oceania
comprising the centre of the global HE system (Chen and Barnett 2000). However, this picture
is changing as former ‘exporters’ of students, particularly countries such as China, Singapore
and Hong Kong, pursue strategies to internationalise their higher education systems (Bhandari
and Blumenthal 2011) and become regional hubs of higher education (Knight 2011). Although
research studies about such ‘non-centre’ contexts are increasing, they are still limited in number
and scope, and more is required to enable the scholarly field to reflect better the contemporary
realities of international student mobility.

Organisational image today is a primary concern for managers and directors of both public and
private organisations (Vigoda-Gadot, Vinarski-Peretz, & Ben-Zion, 2003). Members’
perceptions of how others view their organisation play an important role in how they respond
to the demands made by the organisation (Dutton, Dukerich, & Harquail, 1994). Theories
suggest that instead of focusing on the actual quality of products or services provided by an
organisation, individuals often tend to make choices or decisions based on a series of complex
perceptions or attitudes which shape an overall image of products and services offered by that
organisation (Vigoda-Gadot et al., 2003). Organisational image is essential because it allows
members of an organisation to compare their own perceptions of the organisation with the
assessments of external stakeholders (Dutton et al., 1994). A favourable organisational image
can not only affect attitudes and behaviour of organisation members and enhance affective
commitment, citizenship behaviour, motivation, and positive opinions (Mishra et al., 2012),
but also may serve as a tool for obtaining a competitive advantage for organisations. It also
improves the organisation’s capabilities in employing job applicants and enhances their
satisfaction and loyalty (Duarte et al., 2010).

The expansion of demand experienced during the second half of the 20th century gave rise to
an increase in supply in terms of reach and variety (Maringe & Gibbs, 2009) which was boosted
by the effects of other phenomena such as globalisation and the decrease in public financing.
Globalisation has favoured a growing internationalisation of HE, considerably increasing the
numbers of international students (Altbach, Reisberg, & Rumbley, 2009) and interest from
universities in capturing them. All of the foregoing has stepped up international competition
and rivalry between HEIs to attract home-based and overseas students, resources and
prestigious teaching staff, leading many universities to perceive a need to build a solid
favourable image and reputation among their stakeholders, these being factors of
differentiation that influence their affective responses and behaviour vis-à-vis the institution
(Alves & Raposo, 2007; Belanger, Mount, & Wilson, 2002; Drydakis, 2015; Helgesen &
Nesset, 2007; Kheiry, Rad, & Asgari, 2012; Nguyen & LeBlanc, 2001). Consequently,

747
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

universities have begun to assign more resources to bolster their image (Curtis, Abratt, &
Minor, 2009).

Higher education institutions image is also an essential concern for the higher education
institutions in Malaysia. However, Limited research has been conducted in Malaysia, the
context for this study, which aims to be recognised as an international hub of higher education
excellence by 2020 (Ministry of Higher Education 2011). To achieve the country’s goal to
become the regional education hub, higher education institutions in Malaysia need to bring in
several changes to maintain their service excellence in the global educational platforms
(Manzoor, Malarvizhi & Mahdee, 2019). The government’s Malaysian Internationalisation
Policy document states that international student mobility is the main indicator of the
internationalisation of Malaysian HE (Ministry of Higher Education 2011). Hence the study
intends to explore the dimensions of university image for the international students studying in
Malaysia.

Although research interest in the study of university image is steadily growing (Sung & Yang
2008; Aghaz et al. 2015), those who have examined the subject in detail consider that attention
in the field of academic research has been scarce (Duarte et al., 2010; Wilkins & Huisman
2015) and that more empirical research is needed (Aghaz et al. 2015). While the idea that
university image is a multidimensional concept subject to the influence of a variety of aspects
is shared by academics (Arpan, Raney, & Zivnuska, 2003; Luque & Del Barrio 2008; Aghaz
et al. 2015), there is no consensus as to the dimensions that comprise it (Lafuente-Ruiz-de-
Sabando et al. 2018).

Hence this study objective is to explore the dimensions of university image for the current
international students in the context of Malaysia. The paper starts with the background of the
study, and the literature review, methodology, results, discussions and implications are
presented in the subsequent sections.

2. Literature Review

Image is configured, identifying the essential aspects that affect its formation. Ressler and
Abratt (2009), meanwhile, are of the view that universities should at least study the perceptions
held about them by prospective students, current students, graduates and employers or
companies. The growth of publications focused on the concept of HE image during the past
decade reveals a growing interest in the topic among academics (Sung and Yang 2008; Curtis
et al. 2009; Aghaz et al. 2015). According to Lafuente-Ruiz-de-Sabando, Forcada &
Zorrrrillallalla (2019), the concept of overall university image was defined as the evaluations
that the different stakeholders make of a university through perceptions, beliefs, ideas, and
impressions that, as the consequence of a consistent expression over time, they have
accumulated about it. This definition is also aligned with the meaning generally attributed to
the concept of corporate image by researchers in the marketing domain.

748
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Various studies indicate students as one of the HEI’s most influential audiences (Guilbault,
2016; Wagner Mainardes, Alves,& Raposo, 2012), and so it is essential to understand how its
image is formed in order to draw up strategies and improve university management (Kazoleas,
Kim, & Moffitt, 2001; Paniagua & García, 2015; Theus, 1993).

Various studies indicate students as one of the HEI’s most important audiences (Guilbault,
2016; Wagner Mainardes, Alves,& Raposo, 2012), and so it is essential to understand how its
image is formed in order to draw up strategies and improve university management (Kazoleas,
Kim, & Moffitt, 2001; Theus, 1993). The dimensions analysed in this study is based on the
university image model developed by Purificación Alcaide-Pulido, Helena Alves & Belén
Gutiérrez-Villar (2017). As these researchers used a systematic review of variables studied in
the past literature related to university image is expected to provide more meaningful insights
through analysing their dimensions in the context of international students in Malaysia.

Purificación Alcaide-Pulido, Helena Alves & Belén Gutiérrez-Villar (2017) university image
model consisted of four dimensions such as ‘external communication and values’, ‘national
and international awareness’, ‘economic value’ and ‘university facilities’. Internal and external
communication (Hatch & Schultz 2001) reduce the gap in communication consistency
associated with future crisis (Dowling, 1993). As students can be influenced by multiple
‘significant others’, institutions must communicate regularly and effectively with each of these
stakeholders including parents, schools, employers, regulatory bodies and the media, a
comprehensive and integrated communications strategy is suggested (Wilkins & Huisman,
2015).

Creating and building brand awareness, reaching consumers’ minds, and encouraging them to
develop a preference for the brand, are important steps in ensuring a successful brand (Keller,
2003). Some of the variables are also found to be significant in earlier studies, such as Ivy
(2001) when referring to lecturers’ reputation and the variable of the impact of research, in
Kazoleas et al. (2001) where the fact of being a local institution emerges as well as productive
and quality research, and in the study by Beerli Palacio et al. (2002) obtaining the fact of the
university having a good or bad reputation.

The study outcome of Azoury et al. (2014) shows that one of the cognitive factors of image
that exercises the greatest positive influence is the “reputation and age”. Their findings show
that, when a university is prestigious and has an illustrious history, campus life is dynamic and
the students can quickly get jobs.

Landrum et al. (1999) also have used one of the variables as the value of university using the
latent construct underlying the perception in terms of tuition costs while formulating the
university image model in their study. Similarly, economic value is referred to in the study by
Ivy (2001) when mentioning the fact of having lower fees than the competition; and in the
study by Zaghloul et al. (2010), where the cost of education is found to be significant.

University image consists of a subjective viewpoint of students about the quality of the
programs; and the social and physical environment of the university (Gafoor & Ashraf, 2012).
It has been noted in some studies that the university environment, whether physically or

749
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

psychologically, influences the students’ perception of university image. The last significant
construct is the one relating to Facilities, formed of the variables of classrooms being well
equipped technologically, the number of students per class being appropriate, subjects
generally having a right balance of theory/practice, and the campus being of proper size.

3. Research Method

As such, the higher education sector in Malaysia is chosen for this study. HEIs in Malaysia are
of two main types: public and private. So, data is collected from three top public universities
and two leading private universities in Malaysia, having a higher number of enrolled
international postgraduate students using a self-administered questionnaire. Asal Aghaz, Amin
Hashemi & Maryam S. Sharifi Atashgah (2015) have stressed on the importance of
postgraduate students in one of their studies related to university image due to the higher
experience level of these group of students. Hence, this study has used international
postgraduate students studying in Malaysia as a source of data collection. Response options are
provided on a seven-point Likert scale (from 1= strongly disagree to 7=strongly agree). The
final survey has yielded quantitative data, while the pilot study yielded some rich qualitative
data that have enabled a degree of triangulation to corroborate the findings and conclusions.
This study has used the scale developed by Alcaide-Pulido et al., 2017 in order to measure the
four dimensions of the university image model. Five items were used to measure external
communication and values’; ‘National international awareness’ was measured by four items;
two items were used to measure ‘economic value’ and ‘university facilities’ was measured by
four items. The questionnaire comprises two sections with a total of 19 items. Fifteen items
were used to measure the image dimensions, and four questions were focused on demographic
and academic aspects of the respondents. A total of 210 responses were obtained, of which 200
were complete and usable. The response rate was 60%.
With the purpose of reducing eventual questionnaire errors, a pre-test of the questionnaire was
conducted with the sampled universities. The Cronbach’s alphas for ‘external communication
and values’ is .850, ‘national and international awareness’ is 0.850, ‘economic value’ is .748
and ‘university facilities’ is .799, indicating acceptable reliability.

4. Results and Discussions

Based on the outcomes of the survey, 79.0% of the respondents were male, and 21.0%
respondents were female international postgraduate students (see table 1). 60.5% of the
respondents were studying in the Ph.D. programme; 37.0% were doing Masters and rest were
studying in other postgraduate programmes. Based on the nationality of the respondents were
from Bangladesh (22.0%), Indonesia (4.5%), Yemen (8.0%), Nigeria (11.5%), Pakistan
(9.0%), India (7.0%), Others (38.0%). Majority of the respondents (43.0%) were staying more
than three years in Malaysia.

750
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table1: Characteristics of Respondents

Percentage (%)
Gender:
Male 79.0
Female 21.0
Programme of Study:
37.0
Masters Students
60.5
PhD Students
2.5
Students in other postgraduate programmes
Nationality:
22.0
Bangladesh
4.5
Indonesia
8.0
Yemen
11.5
Nigeria
9.0
Pakistan
7.0
India
38.0
Others
Stay Duration:
< 1 Year 18.0
1-2 Years 16.5
2-3 years 22.5
More than 3 Years 43.0

Table 2: Means and Standard Deviations of Variables

Descriptive Statistics
Mean Std. Deviation
University Facilities:
Classrooms are technologically well equipped. 5.55 1.074
The number of students per class is appropriate. 5.50 1.137
Subjects generally have a good theoretical-practical balance. 5.31 1.051
The campus is of an appropriate size. 5.74 1.169
National and international awareness:
It is known in the country. 5.90 1.103
It is known internationally. 5.68 1.153
It has a good academic reputation and prestige. 5.75 1.014
It is well placed in university rankings. 5.52 1.186
Economic value:
It offers good value for money. 5.18 1.267
The price is right. 5.05 1.268
External communication and value:
It has a good website. 5.59 1.113
It is present in social networks. 5.56 1.033
It is committed to society. 5.59 1.018

751
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

It conveys ethical values and social responsibility. 5.67 .990

It conveys the values of environmental sustainability. 5.69 .985

Table 2 shows the mean and standard deviation of variables. According to this table, the mean
values for the majority of the variables associated with the factor ‘national and international
awareness’ are higher compared to the variables related to other factors such as external
communication and values, economic value and university facilities.

Exploratory factor analysis has been conducted with the overall 15 items of the university
image model to reveal the number of the dimensionality of ‘image’ in the context of
international students studying in Malaysia. Table 3 shows an exploratory factor analysis using
principal components with Varimax rotation conducted (using IBM SPSS version 23) to
determine the underlying components of 15 items presenting the university image instrument
for the international postgraduate students studying in Malaysia.
The Kaiser–Meyer–Olkin test has produced a value of .898, far higher than the cut-off point of
.70, thus indicating that the sample size of 200 is adequate. The Bartlett test of sphericity (p =
.000) indicates that the data have a high enough degree of correlation between at least some
variables adequate to make it suitable for exploratory factor analysis. Using the criteria
eigenvalue > 0.70 (deemed acceptable by Jolliffe, 1986) and factor loading > .4, three factors
have been extracted, which account for 63.472% of the total variance.

Table3: Results of Factor analysis (SPSS Output)

Rotated Component Matrixa

Component
Factor 1 Factor 2 Factor 3
Classrooms are technologically well equipped. .735
The number of students per class is appropriate. .690
Subjects generally have a good theoretical-practical
.437 .418
balance.
The campus is of an appropriate size. .720
It is known in the country. .816
It is known internationally. .697 .427
It has a good academic reputation and prestige. .637 .506
It is well placed in university rankings. .524 .490
It offers good value for money. .825
The price is right. .841
It has a good website. .656
It is present in social networks. .708
It is committed to society. .799
It conveys ethical values and social responsibility. .753
It conveys the values of environmental sustainability. .726

752
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Extraction Method: Principal Component Analysis.

Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 5 iterations.

The findings indicate that there are three factors forming the image for the international
students studying in Malaysia’s HEIs. The first factor has the greatest influence on the
university image of international postgraduate students. This factor comprises 8 items, such as
Classrooms are technologically well equipped; The number of students per class is appropriate;
It has good academic reputation and prestige and so on which explains 25.747% of the variance.
This factor is named as “Institutions national-international awareness and facilities”. The
second factor consisted of 5 items and accounted for 25.324% of the variance. This factor is
named as “external communication and values”. Lastly, the third factor comprised of two
variables that explain 12.401% of the variance. This factor is designated as “economic value”.

5. Research Implications

Our study contributed to both the theoretical and practical field of research. This study has
explored three dimensions such as “institutions national-international awareness and facilities”,
“external communication and values” and “economic value” forming the university image that
differs from the classification of ‘image’ identified by the past researchers. Besides, the sample
used for this study is the current international postgraduate students. These students group not
only have more experience but also add diverse inputs due to their multiplicity.

Maintaining and enhancing desirable organisational image require an emphasis on all the
factors associated with the three dimensions such as “institutions national-international
awareness and facilities”, “external communication and values” and “economic value”
contributing to university image. The variables related to ‘national-international awareness’
and ‘university facilities’ are grouped together as one of the new factors in this study, forming
the university image. This shows that international students prefer to evaluate university
reputation and facilities as combined matters rather than separate entity while evaluating the
university image.

The three dimensions forming the university image in this study will help the higher education
marketing experts to revise their marketing and communication strategies for international
students attraction and retention through satisfying their needs and expectations. It will not
only motivate the postgraduate students to willingly apply for PhD programs or even
membership of faculties in the same university but also influence the other undergraduate
students and prospective students to enrol in the particular university. Theories also support the
idea that members of an organisation with favourable image are more likely to stay in that
organisation (Rashid et al., 2012).

6. Conclusion and Recommendations

753
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Due to the increased competition for international students, the higher education institutions
should seriously consider developing and maintaining a positive and distinctive image in order
to achieve a competitive advantage. This study reveals three dimensions that form the
university image for the international students studying in Malaysia’s HEIs. The first factor
was named as ‘institutions national-international awareness and facilities. The HEIs in
Malaysia should concentrate on developing reputation and prestige along with creating world-
class education infrastructure to enhance their image. These results are in agreement with those
of Ivy (2001), Kazoleas et al. (2001), Arpan et al. (2003). Based on the second factor that is
‘external communication and values’, the universities should strengthen their external
communication aspects such as developing interactive websites in constituting the university
image. The second dimension also emerges as significant in the results of Ivy (2001), Kazoleas
et al. (2001). Lastly is Economic Value, with two variables: an excellent quality-price
relationship and the right price.

Similarly, economic value is referred to in the study by Ivy (2001) when mentioning the fact
of having lower fees than the competition. Furthermore, universities cannot be excellent at
everything, and they cannot address the needs of every customer group, so they should identify
specific aspects around which they can position themselves (Maringe and Gibbs 2009).
Positioning involves specifying and communicating the desired organisational image so that
the students in the target segment understand what the university stands for in relation to other
institutions in the market. Establishing a distinctive image seems to be key because students
generally judge universities on their reputations and not on the actual quality of their teaching
or research (Marginson 2006). The outcomes of this study can help the universities to revise
their marketing and communication strategies to retain international students, especially in the
competitive global education market. Besides universities can assess their performances
related to the variables of university image dimensions revealed in this study and precisely
focus on the areas that need development or more allocations of marketing budget to uphold
their image among the international students.

This study is also beyond limitations. The main limitation of this study is considered to be the
fact of using a sample of international students studying in one country such as Malaysia.
However, the future studies can focus on cross-cultural studies by incorporating international
students studying in various countries in Southeast Asia and beyond to analyse the image
dimensions in order to enhance the generalisability of the current research findings.

Acknowledgement: The authors would like to thank the Research Management Centre (RMC)
of Multimedia University for their support in funding this research (PD20190667) administered
by the Faculty of Management.

754
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References

Altbach, P. G., Reisberg, L., & Rumbley, L. E. (2009). Trends in global higher education:
Tracking an academic revolution. Paris: UNESCO.

Arpan, L. M., Raney, A. A., & Zivnuska, S. (2003). A cognitive approach to understanding
university image. Corporate Communications: An International Journal, 8(2), 97–113Jolliffe,
I.T. 1986. Principal component analysis. New York: Springer.

Asal Aghaz, Amin Hashemi & Maryam S. Sharifi Atashgah. (2015). Factors contributing to
university image: the postgraduate students’ points of view. Journal of Marketing for Higher
Education, 25(1), 104-126, DOI: 10.1080/08841241.2015.1031314

Belanger, C., Mount, J., & Wilson, M. (2002). Institutional image and retention. Tertiary
Education and Management, 8(3), 217–230

Beerli, A., Díaz Meneses, G., & Pérez Pérez, P. (2002). The configuration of the university
image and its relationship with the satisfaction of students. Journal of Educational
Administration, 40(5), 486–505.

Bhandari, R., & Blumenthal, P. (2011). Global student mobility and the twenty-first century
silk road: national trends and new directions. In R. Bhandari & P. Blumenthal (Eds.),
International students and global mobility in higher education: national trends and new
directions, 1–23, New York: Palgrave Macmillan.

Bohm, A., Davis, D., Meares, D., & Pearce, D. (2002). Global student mobility 2025: forecasts
ofthe global demand for international higher education. Sydney: IDP Education Australia.

Chen, T. M., & Barnett, G. A. (2000). Research on international student flows from a macro
perspective: a network analysis of 1985, 1989 and 1995. Higher Education, 39(4), 435–453.

Curtis, T., Abratt, R., & Minor, W. (2009). Corporate brand management in higher education:
The case of ERAU. Journal of Product & Brand Management, 18(6), 404–413.

Dowling, G. R. (1993). Developing your company image into a corporate asset. Long Range
Planning, 26, 101–109.

Drydakis, N. (2015). Economics applicants in the UK labour market: University reputation
and employment outcomes. International Journal of Manpower, 36(3), 296–333.

Duarte, P. O., Alves, H. B., & Raposo, M. B. (2010). Understanding university image: A
structural equation model approach. International Review on Public and Nonprofit Marketing,
7(1), 21–36.

755
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Dutton, J. E., Dukerich, J. M., & Harquail, C. V. (1994). Organisational image and member
identification. Administrative Science Quarterly, 39(2), 239–263

Gafoor, K. A., & Ashraf, P. M. (2012). Contextual influences on sources of academic self-
efficacy: A validation with secondary school students of Kerala. Asia Pacific Education
Review, 13, 607–616.

Guilbault, M. (2016). Students as customers in higher education: Reframing the debate.

Journal of Marketing for Higher Education, 26(2), 132–142.

Helgesen, Ø., & Nesset, E. (2007). Images, satisfaction and antecedents: Drivers of student
loyalty? A case study of a Norwegian university college. Corporate Reputation Review,
10(1), 38–59.

Ivy, J. (2001). Higher education institution image: A correspondence analysis approach.

International Journal of Educational Management, 15(6), 276–282.

Kazoleas, D., Kim, Y., & Moffitt, M. A. (2001). Institutional image: A case study. Corporate
Communications: An International Journal, 6(4), 205–216.

Kheiry, B., Rad, B. M., & Asgari, O. (2012). University intellectual image impact on
satisfaction and loyalty of students (Tehran selected universities). African Journal of Business
Management, 6(37), 10205–10211

Knight, J. (2011). Regional education hubs: mobility for the knowledge economy. In R.
Bhandari & P. Blumenthal (Eds.), International students and global mobility in higher
education: national trends and new directions, 211–230, New York: Palgrave Macmillan.

Lafuente-Ruiz-de-Sabando, A., Zorrilla, P., & Forcada, J. (2018). A review of higher

education image and reputation literature: Knowledge gaps and a research agenda. European
Research on Management and Business Economics, 24(1), 8–16.
[Link]

Lafuente-Ruiz-de-Sabando, A., Forcada, J., & Zorrrrillallalla, P. (2019). The university

image: a model of overall image and stakeholder perspectives. Cuadernos de Gestión, 19(1),
63–86. [Link]

Marginson, S. 2006. Dynamics of national and global competition in higher education. Higher
Education, 52(1), 1–39.

Maringe, F., & Gibbs, P. (2009). Marketing higher education: Theory and practice. UK:
McGraw-Hill.

Ministry of Higher Education. (2011). Internationalisation policy for higher education

Malaysia 2011. Kuala Lumpur: Ministry of Higher Education.

756
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Mishra, S. K., Bhatnagar, D., D’Cruz, P., & Noronha, E. (2012). Linkage between perceived
external prestige and emotional labor: Mediation effect of organisational identification among
pharmaceutical representatives in India. Journal of World Business, 47, 204–212.

Nguyen, N., & LeBlanc, G. (2001). Image and reputation of higher education institutions in
students’ retention decisions. International Journal of Educational Management, 15(6), 303–
311.

OECD. (2014). Education at a glance in 2014: highlights. Paris: OECD Publishing.

Phair, J. T. (1992). 1992 education report card: Educational institutions, newly vulnerable,
confront issues of economy and public trust. Public Relations Journal, 48,22–40. Retrieved
from http:// [Link]/openview/baebfe8c00e5efb83b1e69afaed4aa52/1?pq-
origsite = gscholar&cbl=34602

Purificación Alcaide-Pulido, Helena Alves & Belén Gutiérrez-Villar. (2017). Development of

a model to analyse HEI image: a case based on a private and a public university, Journal of
Marketing for Higher Education, 27:2, 162-187, DOI: 10.1080/08841241.2017.1388330

Rashid, M., Spreckelmeyer, K., & Angrisano, N. J. (2012). Green buildings, environmental
awareness, and organisational image. Journal of Corporate Real Estate, 14(1), 21–49.

Manzoor, S.R., Malarvizhi, C., & Mahdee, J. B. M. (2019). Investigating value Co-creation
behaviour among international postgraduate students in Malaysia’s HEIs. Indian Journal of
Public Health Research and Development , 10 (6), 1367–1372. [Link]
5506.2019.01488.8

Sung, M., & Yang, S. U. (2008). Toward the model of university image: The influence of brand
personality, external prestige, and reputation. Journal of Public Relations Research, 20(4),
357–376.

Theus, K. T. (1993). Academic reputations: The process of formation and decay. Public
Relations Review, 19(3), 277–291.

Vigoda-Gadot, E., Vinarski-Peretz, H., & Ben-Zion, E. (2003). Politics and image in the organ-
izational landscape: An empirical examination among public sector employees, Journal of
Managerial Psychology, 18(8), 764–787.

Wagner Mainardes, E., Alves, H., & Raposo, M. (2012). A model for stakeholder classification
and stakeholder relationships. Management Decision, 50(10), 1861–1879.

757
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

INNOVATION CAPABILITY AND LOGISTICS SERVICE QUALITY IN

IMPROVING THE PERFORMANCE OF MALAYSIAN 3PL SERVICE PROVIDERS

Siti Nur ‘Atikah Zulkiffli1, Maisarah Sebadak2, Siti Falindah Padlee3 & Juhaizi Mohd
Yusof4
1, 2, 3 & 4
Faculty of Business, Economics and Social Development, Universiti Malaysia
Terengganu, 21030 Kuala Nerus, Malaysia

Authors’ Email: atikahzulkiffli@[Link].my1, juhaizi@[Link].my4,

mysarah16102@gmail.com4
Corresponding author’s e-mail: [Link]@[Link].my3

Abstract
The Malaysian Government Transformation Programme states that transportation is one
of the national priority areas. It plays a significant role in the country’s well-being and has
contributed much to the development and improvement of the economic, social, political,
and cultural conditions of the nation. The role and contribution of transportation is
important in strategic business and it is also one of the key focal points in the Eleventh
Malaysia Plan for 2016 to 2020. Moving forward, while the prospects for Malaysia’s
logistics industry are positive, there is still much room for improvement in terms of
business performance. Most enterprises that sell or produce goods require facilitating
agencies such as third-party logistics (3PL) service providers to assist them in transporting
their products and services from one destination to another until they reach the end
customers. Therefore, this study attempts to identify the relationship between innovation
capabilities, logistics service quality and business performance among 3PL service
providers in Malaysia. A simple random sampling was used and a quantitative research
design was adopted to address two research questions to 152 3PL service providers in
Malaysia. The study employs the multiple regression analysis to analyse the data. The
results show positive relationship between innovation capabilities and logistics service
quality towards business performance. This indicates that all two hypotheses are accepted.
Such findings will benefit the policy makers, academicians, and transportation industry to
resolve current problems pertaining to the topic and enhance the competitiveness among
3PL service providers in Malaysia.
Keywords: Resource-Based View, Capability, Innovation, Service Quality,
Performance, Malaysia

758
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1 Introduction

Managing logistics process can be a complex and difficult task. The increase in
globalization also complicates the logistics systems. Yet, as logistics functions become
more complicated and the gap between demanding firms to achieve and the level of their
internal capabilities continues to grow, the rational for facilitating agencies such as third -
party logistics (3PL) service providers has increased (Sohail & Sohal, 2003). It seems that
the role of 3PL service providers is vital in assisting any business to transport their products
to the end customers.
Globally, multinational distribution and transportation companies have never stopped
looking for better solution for logistics (Yong, 2001). In Malaysia, the intense competition
in logistics has driven many 3PL service providers to the verge of dilemma. Evidence by
existing studies have reported that innovativeness is one of the prominent issues that halts
the enhancement of productivity and competitiveness in the logistics industry (Lawson &
Samson, 2001; Birkinshaw, 2011; Racela, 2014; Saunila, 2016). Among the challenges
faced by 3PL companies as they serve their customers across the globe include the
difficulty to find the right candidates, lack of good reporting systems, extreme emphasis
on rules and regulations, lack of collaboration, unable to meet the specific logistical needs
of the shippers, and lack of infrastructures (Flatworld, 2017). Damiri (2009) states that the
high logistics cost in Malaysia is caused by poor transport infrastructure, underdeveloped
transport logistics services, as well as the slow and costly bureaucratic procedures in
dealing with both exported and imported goods.
From the local perspective, the majority of 3PL service providers in Malaysia are still
offering traditional services such as transportation, warehousing, and customs clearance
(Umar, 2004). The warehouse industry in Malaysia is dominated by medium and small-
sized operators and most warehouses are lacking in specialised services such as cold
storage facilities, pick and pack facilities, and pre-retail services (Umar, 2004; Malaysia
Productivity Corporation, 2013). Besides, there is also limited governance on the local
warehouse segment. This leads to its non-strategic location where most warehouses in
Peninsular Malaysia are situated far from central locations.
The Ministry of Science, Technology, and Innovation (MOSTI, 2010) reports that
Malaysia was ranked 24th in Innovation Competitiveness and 26th in the Global
Competitiveness Index among 140 countries in the World Economic Forum (WEF) from
2010 to 2011. However, there are several innovation issues that affect the business

759
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

performance of local 3PL service providers particularly on the aspect of postal and courier,
outdated technology, low level of research and development (R&D), limited innovation
capacity, lack of support from the government and government-linked companies (GLC),
and limited financial capabilities (Noordin et al., 2013). Besides that, Malaysian 3PL
service providers also express their frustration with the traffic bottlenecks and
inefficiencies that impede the progress of the logistics industry due to the loss in
productivity hours (STAR, 2018). These issues cause further loss in the business margin
performance.
Furthermore, studies about the relationship between innovation capabilities, logistics
service quality, and business performance with 3PL service providers are scarce in the
context Malaysia. Most existing studies on 3PL service providers were conducted outside
the Malaysian context (Yang et al., 2009; Malgorzata & Gajewska, 2014; Sohail, 2016;
Rajapathirana & Hui, 2017; Chang et al., 2019) with more focus being placed on the
manufacturing industry (Sobhani & Hamid, 2008; Bustinza et al., 2010; Halim et al., 2012)
and food industry (Rafiq, 2007: Karia & Asaari, 2016). This creates the need for more
studies to be done on the context of 3PL service providers in Malaysia.
Hence, this study aims to investigate how Malaysian 3PL service providers can
develop their innovation capability and logistics service quality to enhance their
performance and lead to competitive advantage. This forms an important contribution of
the study.

2 Literature review

Scholars from different fields in the social sciences provided meaningful explanations of
the theories and drivers of innovation capabilities, logistics service quality and business
performance, as discussed in the section 2.1 and section 2.2.

2.1 Resource-based view

The Resource-Based View (RBV) theory defines how organizations achieve and maintain
competitive advantages by exploiting and utilizing their own strategic resources and
various capabilities (Wernerfelt, 1984). It defines firm resources as the assets, capabilities,
organizational processes, firm attributes, information, and knowledge controlled by a firm
(Barney, 2001). The theory proposes that a firm’s resources cannot be duplicated by other
competitors if those resources are rare, valuable, and/or cannot be imitated, and have no

760
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

substitute for the product itself. Such resources can lead to the creation of competitive
capabilities.
In recent years, logistics service providers have grown via a mixture of organizational
expansions, mergers, acquisitions, and alliances. These activities are among the ways to
acquire resources to achieve growth. According to Yang et al. (2009), RBV has been
employed in logistics-related research to assess the contributions made by logistics
activities on firm performance. Thus, the theory provides a key theoretical foundation for
logistics service providers to achieve competitive advantage.
Existing literature have provided empirical evidence on the connection between
innovation capabilities, logistics service quality, and business performance. In order to
further advocate the link, this study conceptualizes the relationship using the RBV theory.
Wong and Wong (2011) state that when firms have unique abilities to exploit and creates
resources, they will be able to improve their performance and competitiveness.
Ombaka et al. (2015) also argue that upgrading firms’ resources and capabilities will
maintain its competitiveness and growth in a changing organization environment. It seems
that innovation capabilities can be viewed as a resource and the competitive advantage of
an organization relies on the power of innovation capabilities development.

2.2 Innovation capability, logistics service quality and business performance

Nowadays, innovation represents competitive advantages that are supported by strong
mainstream capabilities in quality, efficiency, speed, and flexibility. Innovation can help
firms to play a dominant role in shaping the future of their industries where high
performing innovators will be able to maintain a giant juggling act of capabilities and
consistently bring new and high quality products to the market faster, more frequent, and
at a lower cost than other competitors (Lawson & Samson, 2001).
Scholars indicate that innovation capability is the most important dynamic that enables
firms to achieve high level of competitiveness (Saunila, 2016). Innovation capability refers
to a firm’s capacity to engage in innovation in terms of introduction of new processes,
products, or ideas in the organisation (Hult et al., 2004).
Also, logistics has always been a supporting function for production and consumption.
Previously, most traditional courier service companies only considered functions as part
of their cost. However, since the 1990s, this perception started to change in the marketing
principle as part of the effort to determine the capacity of courier service and subsequently
lead to greater customer satisfaction and loyalty.

761
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The majority of existing studies agree that innovation may contribute positive
influence towards business performance (Yamin et al., 1997, Hult et al., 2004; Rahab,
2012; Yıldız et al., 2014; Nadarajah, 2015). According to Oke (2007), innovation related
to radical or incremental has the potential of offering interesting contributions to firms'
performance. It acts as an important determinant for the business performance in spite of
the market fluctuating in which the firm conducts (Hult et al., 2004). Innovation process
can also be viewed as effective drivers to enhance the innovation and trade performance
of the organization (Lendel & Varmus, 2014). This is further supported by several
empirical evidences that report on the positive influence of innovation capability to
positively improve firms’ financial performance particularly in areas such as sales growth,
profitability, and market share (Hult et al., 2004; Keskin, 2006; Panayides, 2006; Oke et
al., 2007; Yang et al 2009).
Previous studies stated that logistics service quality shall increase clients’ satisfaction
with the companies’ ability to solve problems, keep accurate records, deliver services on
time, and communicate effectively, which are viewed as among the necessary
improvements in a business performance (Leuthesser & Kohli, 1995). Tan et al. (1998)
indicate that the creation of service quality contributes to positive impact on firms‟
profitability. In addition, delivering high quality logistics service also has an influence on
the performance of an organization in terms of timeliness, flexibility, accuracy, responsive
and problem solving. This is further supported by several other empirical evidences (Lai,
2004; Panayides, 2006, 2007; Lu & Yang, 2007) that report significant positive impact by
logistics service quality on the overall performance of a firm.

3 Research hypotheses and theoretical framework

Figure 1 illustrates the theoretical framework that serves the objective of the study.

Innovation
Capability H1
Business
Performance
Logistics Service
Quality H2

Figure 1: Theoretical Framework of the Study

762
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

By referring to the literature review (section 2.2), therefore, the hypotheses of the
study are:

Hypothesis 1, H1: Innovation capability positively affects to the business

performance of 3PL service providers in Malaysia, and
Hypothesis 2: Logistics Service Quality positively affects to the business
performance of 3PL service providers in Malaysia.

4 Research methodology

A total of 43 items were generated from the literature for inclusion in the questionnaire
that was used to collect the empirical data. The questionnaire employed a five-point Likert
scale ranging from 1 = strongly disagree to 5 = strongly agree in order to measure the
innovation capability constructs and the logistics service quality construct. It also
contained a five-point Likert scale ranging from 1 = very poor to 5 = excellent performance
to measure the business performance construct. The questionnaire was designed by
adapting those used in previous studies. The questionnaire was assessed by experts who
confirmed its validity.
For the study sample, a total of 285 3PL service providers were randomly selected
from the Federation of Malaysian Freight Forwarders (FMFF) 2016/2017 directory. A
simple random sampling was conducted in the study. A copy of the questionnaire was sent
to the selected 3PL service providers together with a pre-paid return envelope and a cover
letter addressed to the 3PL service providers’ director or manager. The initial letter was
followed by two reminder letters.
A total of 152 usable questionnaires were received from the participating 3PL service
providers. This represented a 53.33 percent response rate.

5 Data analysis and discussion

5.1 Demographic analysis

Tables 1 exhibits the demographics profiles of the 152 samples of third-party logistics
(3PL) service providers in Malaysia. The analysis indicates that the majority 3PL in
Malaysia (97 service providers or 63.8 percent) use more than one mode of transportations,
including multi-modal transport such as air, land, rail, and sea. Whereas, 44 service

763
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

providers (28.9 percent) resort to the use of land transportation and the least used of
transportation mode is sea freight with 11 service providers (7.2 percent). Generally, 124
or 81.6 percent of 3PL service providers in Malaysia operate more than one type of
business activities including warehousing and storage, custom clearance, transportation
services, order picking and packaging, overseas distribution, pick-up and delivery,
inventory control, freight consolidation, bill payment, and return food handling.
Meanwhile, 11 service providers (7.2 percent) operates transportation services, seven
service providers (4.6 percent) are involved with warehousing and storage, and eight
service providers (5.3 percent) operating other types of 3PL services.

Table 1: Demographic Profile of 3PL Service Providers in Malaysia

Business Profile Features Frequency Percent
Mode of Transport
Land Transport 44 28.9
Sea Freight 11 7.2
Use Multi-Mode of Transports 91 63.8
Total 152 100.0

Type of Core Business Activities

Warehousing and Storage 7 4.6
Custom Clearance 1 .7
Transportation Services 11 7.2
Order Picking / Packaging 1 .7
Others 8 5.3
Use Multi-type of Core Business 124
Activities 81.6
Total 152
100.0

5.2 Reliability analysis

Scale reliability was evaluated using the Cronbach’s alpha. The scale reliability was
conducted using the Cronbach’s Alpha coefficient for internal consistency via SPSS in
order to ensure that the research instrument is comprehendible by the respondents. Table
2 shows that the Cronbach’s Alpha value for construct innovation is 0.957, logistics
service quality with 0.916, and business performance with 0.961. Hair et al. (2010) posit
that Cronbach’s Alpha value of 0.70 is acceptable in exploratory research, which suggests
the positive reliability of these constructs.

764
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Reliability Analysis of the Constructs

Construct Cronbach’s Number of
Alpha Items
Innovation Capabilities 0.939 15
Logistics Service Quality 0.902 4
Business Performance 0.950 7

5.3 Multiple regression analysis

A multiple linear regression was used to predict the innovation capability and logistics
service quality to predict the business performance of 3PL service providers in Malaysia.
Preliminary analyses were conducted to ensure no violation of the assumptions of
normality, linearity, multicollinearity and homoscedasticity.

Table 3: Model Summary

Model Summaryb
Adjusted R Std. Error of the
Model R R Square Square Estimate
a
1 .820 .672 .668 .37539
a. Predictors: (Constant), Innovation Capability, Logistics Service Quality
b. Dependent Variable: Mean_BP

Table 3 shows that the value of R (.820) indicates a good level of prediction. Also, the
coefficient of determination indicates that value of .672 (R2= .672) for innovation
capability and logistics service quality explain 67.2 percent of the variability of business
performance, hours per week. And 32.8 percent (100 percent - 67.2 percent) of the
variation is caused by factors other than the predictors included in this model.

Table 4: ANOVA
ANOVAa
Sum of
Model Squares df Mean Square F Sig.
1 Regression 43.003 2 21.502 152.584
Residual 20.997 149 .141 .000b
Total 64.000 151
a. Dependent Variable: Business Performance
b. Predictors: (Constant), Innovation Capability, Logistics Service Quality

765
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In this study, ANOVA is measure to test the statistical significance of the model.
Therefore, Table 4 shows that the model is statistically significantly predicted the business
performance of 3PL service providers in Malaysia, as F (2, 149) = 152.584, p (.000) < .05,
R2 = .672.

Table 5: Coefficients
Coefficientsa
Standar
Unstandardi dized 95.0%
zed Coeffici Confidence Collinearity
Coefficients ents Interval for B Correlations Statistics
Std. Lower Upper Zero- Toleran
Model B Error Beta t Sig. Bound Bound order Partial Part ce VIF
1 (Constant) -
.239 -.675 .501 -.634 .311
.162
Innovation
.508 .085 .407 5.976 .000 .340 .676 .752 .440 .280 .474 2.108
Capability
Logistics
Service .526 .075 .475 6.972 .000 .377 .675 .770 .496 .327 .474 2.108
Quality
a. Dependent Variable: Business Performance

Table 5 shows the Coefficients result to assess the Statistical significance of the
independent variables. Both variables - innovation capability (p=.000, p<.05) and logistics
service quality (p=.000, p<.05) added statistically significant to the prediction. The highest
contributing predictor is Logistics Service Quality (.475) and the next is innovation
capability (.407) to explain the business performance. Table 5 also shows that the
Multicollinearity problem does not exist in the model as VIF for all variables is <10 (or
Tolerance > .1).
Meanwhile, the results show that Part Coefficients of Determination are (.280)2 and
(.327)2 for the predictors of innovation capability and logistics service quality,
respectively. These unique contributions of the predictors when added up, approximates
18.53 percent (7.84 + 10.69) of the variation in the outcome variable. And this percentage
of variance in the response variable is different from the R2 value (67.2 percent) in the
model. Meaning that 48.67 percent (67.2-18.53=4 5.9) overlapping predictive business
performance was done by the predictors. This proves the combination of the variables had
been quite good.
Therefore, it can be concluded that the relationship for both innovation capability and
logistics service quality towards business performance of 3PL service providers in
Malaysia are statistically significant (p<.05), and therefore, all hypotheses are accepted.

766
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The finding from Hypothesis 1 indicates that innovation capabilities have positive
influence towards the business performance of 3PL service providers in Malaysia. This
shows that advancement in innovation capabilities will subsequently enhance their
business performance. This finding is consistent with Panayides (2006), Oke et al. (2007),
Yang et al. (2009), and Zawawi et al. (2017). Hence, the increase in innovation capabilities
can be associated with the increase in business performance (Zawawi et al., 2017).
As discussed, the RBV theory underlines that a firms innovativeness leads to its firm
performance as empirically proven by various previous studies (Wong & Karia, 2010;
Zawawi et al., 2017) From the perspective of the RBV theory, turbulence of market and
competitive intensity are two main factors in the market which may influence the need for
firm to be innovative. Innovation can lead firms to achieve competitive advantage at a
certain extent, thereby improving its performance (Damanpour, 1991). In addition, it is
generally conceded that innovation contributes to firm and business performance
regardless of any fluctuated market (Hult et al., 2004). This justifies the relevance of the
RBV theory to the finding of Hypotheses 1. The results provide a significant impact to the
majority of 3PL service providers in Malaysia by highlighting the importance of
innovation capabilities in enhancing the performance of their business.
On the other hand, the finding from Hypothesis 2 indicates that logistics service
quality does have positive influence towards the business performance of 3PL service
providers in Malaysia. This shows that the improvement of logistics service quality will
lead towards better business performance. This finding is consistent with the result by
Panayides (2006, 2007), Yang et al. (2009), and Lyu et al. (2019).
Therefore, the result in Hypothesis 2 will give great impact to the Malaysian 3PL
service providers as the improvement in logistics service quality will lead to better
business performance. 3PL service providers can become competitive by achieving cost
leadership or service differentiation. According to Lai et al. (2006), logistics service
providers must achieve service variety, quality advantage, and cost advantage to achieve
their competitive advantage. As the finding analysis supports the hypothesis, therefore; the
study affirms that logistics service quality can be an influence to the performance of
business firms‟ in a developing country like Malaysia.

767
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

6 Conclusion and Recommendations

This study provides empirical evidence to support the theoretical model that links the
innovation capabilities, logistics service quality, and business performance of Malaysian
third-party logistics (3PL) service providers. In addition, the Resource Based View (RBV)
theory that underlies this study was also discussed in detail.
The research methodology was developed by using the quantitative method. The
target population of this study consists of 1161 third party logistics (3PL) service providers
as listed in the Federation of Malaysian Freight Forwarders (FMFF) 2016/2017. The two
research objectives developed for this study were achieved by using multiple regression
analysis. The contributions of the study cover several aspects including the literature,
industry or practitioner, society (clients of 3PL service providers), future academicians,
and policy makers. The findings of the study are hoped to resolve several issues pertaining
to the local and global market of the 3PL service providers in Malaysia. In addition, this
study also provides empirical supports for facilitating agencies to be more innovative in
improving their logistics service quality and subsequently lead towards a higher level of
business performance.
In addition, future studies may include other potential variables for a more
comprehensive analysis. This can include relationship orientation which might play an
important role in determining the influence of innovation capability and logistics service
quality towards the business performance 3PL service providers. It is also recommended
for future studies to use other relevant and appropriate theories besides the RBV theory to
describe the comprehensive literature. It is also recommended for future studies to expand
the research by focusing on different sectors or geographical areas. This is important in
identifying the innovation capability of logistics companies across various areas and how
this may affect the quality of their logistics services for them to be competitive.

Acknowledgment

This research is funded by a grant from the Fundamental Research Grant Scheme (FRGS)
(FRGS/1/2016/SS01/UMT/02/10) awarded by the Ministry of Higher Education,
Malaysia.

768
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References

Barney, J. B. (2001). The resource-based view of the firm. Journal of Management,

27, 625–641.
Birkinshaw, J. & Bouquet, C. & Barsoux, J.L. (2011). The 5 Myths of Innovation. MIT
Sloan Management Review. 52. 43-50.
Bustinza, O. F., Arias-Aranda, D., & Gutierrez-Gutierrez, L. (2010). Outsourcing,
competitive capabilities and performance: An empirical study in service firms.
International Journal of Production Economics, 126(2), 276–288.
Chang, H. H., Wong, K. H., & Chiu, W. S. (2019). The effects of business systems
leveraging on supply chain performance: process innovation and uncertainty as
moderators. Information and Management, (1), 2–47
Damanpour, F. (1991). Organizational Innovation: A Meta-Analysis of Effects of
Determinants and Moderators. The Academy of Management Journal, 34(3), 555–
590.
Damiri, M. H. (2009). Issues and challenges of logistics in Malaysia: A perspective.
United Nations Economic and Social Commission for Asia and Pacific
(UNESCAP), (8), 1–11.
Flatworld, F. (2017). Challenges and Trends in 3PL and Distribution. Retrieved November
25, 2017, from [Link]
[Link].
Hair, J. F. (2010). Multivariate Data Analysis: Pearson College Division
Halim, H. A., Ahmad, N. H., & Ramayah, T. (2012). Probing into the issues of outsourcing
among SMEs in Malaysia. Paper presented at ICSSBE 2012 International Conference
on Statistics in Science, Business, and Engineering International Conference, 10–12
September, Langkawi, Kedah, Malaysia.
Hult, G. T. M., Hurley, R. F., & Gary A. Knight. (2004). Innovativeness: Its antecedents
and impact on business performance. Industrial Marketing
Management, 33(1), 429–438.
Karia, N., & Asaari, M. H. A. H. (2016). Halal business and sustainability: strategies,
resources and capabilities of halal third-party logistics (3PLs). Progress in
Industrial Ecology: An International Journal, 10 (2/3), 286.

769
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Keskin, H. (2006). Market orientation, learning orientation, and innovation capabilities in

SMEs: An extended model. European Journal of Innovation
Management, 9(4), 396–417.
Lai, F., Zhao, X., & Wang, Q. (2006). The impact of information technology on the
competitive advantage of logistics firms in China. Industrial Management & Data
Systems, 106(9), 1249–1271.
Lai, H., & Zhu, S. C. (2004). The determinants of bilateral trade. Canadian Journal
of Economics, 37(2), 459–483.
Lawson, B., & Samson, D. (2001). Developing innovation capability in organisations: a
dynamic capabilities approach. International Journal of
Innovation Management, 5(3), 377–400.
Lendel, V., & Varmus, M. (2014). Evaluation of the Innovative Business Performance.
Procedia - Social and Behavioral Sciences, 129, 504–511.
Leuthesser, L. & Kohli, A.K. (1995). Relational behaviour in business markets:
Implications for relationship management. Journal of Business Research, 34(2), 221-
233.
Lu, C-S. & Yang, C-C. (2010) Logistics service capabilities and firm performance of
international distribution center operators. The Service Industries Journal, 30(2), 281-
298.
Lyu, G., Chen, L. & Huo, B. (2019). Logistics resources, capabilities and operational
performance: A contingency and configuration approach. Industrial Management &
Data Systems, 119(2), 230-250.
Malaysia Productivity Corporation (2003). Reducing unnecessary regulatory burdens on
business: Warehousing service. Draft Report, Malaysia Productivity Corporation
(MPC).
Malgorzata, L.K. & Gajewska, T. (2014). Customer satisfaction with the quality of the
logistic services. Scientific Journal of Logistics – Log Forum, 10(1), 13-19.
MOSTI (2010). Malaysian Science & Technology Indicators 2010. Malaysian Science
and Technology Information Centre, Ministry of Science and Technology.
Nadarajah, G. (2015). Factors influencing third party logistics performance in Malaysia:
The role of trust as a mediator. International Journal of Supply Chain Management,
4(4), 108–114.

770
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Noordin, A., Hasnan, N. & Osman, N. H. (2013). Six main innovation issues: A case of
service innovation of postal and courier services in Malaysia. Journal
of Chemical Information and Modelling, 53(9), 1689–1699.
Oke, A. (2007). Innovation types and innovation management practices in service
companies. International Journal of Operations and Production Management,
27(6), 564–587.
Ombaka, B., Machuki, V. N., & Mahasi, J. (2015). Organizational resources, external
environment, innovation and firm performance: A critical review of literature. DBA
Africa Management Review, 5(1), 60–74.
Panayides, P. (2006). Enhancing innovation capability through relationship management
and implications for performance. European Journal of Innovation
Management, 9(4), 466–483.
Panayides, P. (2007). The impact of organizational learning on relationship orientation,
logistics service effectiveness and performance. Industrial
Marketing Management, 36(1), 68–80.
Racela, O. C. (2014). Customer Orientation, Innovation Competencies, and Firm
Performance: A Proposed Conceptual Model. Procedia - Social and Behavioral
Sciences, 148, 16–23.
Rafiq, M. (2007). Meauring customers‟ perceptions of logistics service quality of 3PL
service providers. Journal of Business Logistics, 28(2), 159–175.
Rahab. (2012). Innovativeness model of small and medium enterprises based on market
orientation and learning orientation: Testing moderating effect of business operation
mode. Procedia Economics and Finance, 4, 97–109.
Rajapathirana, R. P. J., & Hui, Y. (2017). Relationship between innovation capability,
innovation type, and firm performance. Journal of Innovation &
Knowledge, 3(1), 44–55.
Saunila, M. (2016). Performance measurement approach for innovation capability in
SMEs Minna. International Journal of Productivity and Performance
Management, 65(2), 162–176.
Sobhani, H. & Hamid, A. (2008). Comparative analysis of the factor productivity in Iran's
large manufacturing. Tahghighat-e-Eghtesadi, 43(82), 87-119.
Sohail, M. S. (2016). Benchmarking usage of third-party logistics: A comparison of
practices between firms in Malaysia and Saudi Arabia. World Review of Intermodal
Transportation Research, 1(1), 69-81.

771
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Sohail, M., & Sohal, A. (2003). The use of third-party logistics services: A Malaysian
perspective. Technovation, 23(5), 401–408.
Star (2018). Tackling Malaysia’s Logistics Challenges. (2018, Jul 02). The Star, p.1.
[Link]
logistics-challenges/
Tan, K.C., Kannan, V.R. & Handfield, R.B. (1998). Supply chain management: Supplier
performance and firm performance. International Journal of Purchasing & Materials
Management, 34(3), 2-9.
Umar, S. (2004). The Development of Third-Party Logistics in Malaysia: An Overview.
Unpublished Master of Science dissertation. Malaysia University of Science and
Technology.
Wernerfelt, B. (1984). A Resource based view of the firm. Strategic Management
Journal, 5(2), 171–180.
Wong, C. Y., & Karia, N. (2010). Explaining the competitive advantage of logistics service
providers: a resource-based view approach. International Journal of
Production Economics, 128(1), 51–67
Wong, W. P., & Wong, K. Y. (2011). Supply chain management, knowledge management
capability, and their linkages towards firm performance. Business
Process Management Journal, 17(6), 940–964.
Yamin, S., Mavondo, F., Gunasekaran, A., & Sarros, J. C. (1997). A study of competitive
strategy, organisational innovation and organisational performance among Australian
manufacturing companies. International Journal of Production Economics, 52(1-2),
161–172.
Yang, C. C., Marlow, P. B., & Lu, C-S. (2009). Assessing resources, logistics service
capabilities, innovation capabilities and the performance of container shipping
services in Taiwan. International Journal of Production Economics,
122(1), 4–20.
Yildiz, S. M. (2012). Service quality evaluation in the school of physical education and
sports: An empirical investigation of students’ perceptions. Total Quality
Management & Business Excellence, 25(1-2), 80–94.
Yong, G. (2001). Structuring logistics activities in multinational companies (MNCs).
Master Thesis, 1-82, Göteborg University, Sweden.
Zawawi N.F.M., Wahab, S. A., Mamun, A. A, Ahmad, G.B, & Fazal, S. A. (2017).
Logistics capability, information technology, and innovation capability of logistics

772
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

service providers: empirical evidence from east coast Malaysia. International Review
of Management and Marketing, 7(1), 326–336.

773
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

CREDIT DRIVEN HOUSEHOLD DEMAND CHANNEL AND

REAL ECONOMIC ACTIVITY WITH MACROPRUDENTIAL

POLICY

1 ∗
Toure Talnan Aboulaye , NAKAMURA Tamotsu

July 15, 2020

Abstract
The recent nancial crisis of 2008, made it clear that non-Walrasian
features of credit markets have important macroeconomic consequences,
and highlighted the question about how to combine both monetary and
prudential policies. In this paper, we construct a Bayesian vector au-
toregressive dynamic stochastic general equilibrium model (BVAR-DSGE)
that integrates borrowing and using it to analyze the eect of macropru-
dential policy and the way monetary and prudential policies can be con-
ducted as a coherent whole. First, we nd that consumption appears to
be procyclical, while credit to households appears to be countercyclical.
Moreover, consumption and credit to households appear both much more
volatile than output. We also nd that nancial stress index can eec-
tively monitor nancial stability even in projection, while changes in the
policy interest rate cannot. When nancial stress index curtails credit
supply to households, a contraction ensues, and the policy interest rate
is cut. With this chain of causality, prudential policy is procyclical and
monetary policy is countercyclical. Therefore, in our model, one policy is
contractionary and the other expansionary in order to manage households'
credit expansion and deal with the business cycle.
Keywords: credit to households, macroprudential policy, nancial stress index,

nonlinear estimation of DSGE models, bayesian vector autoregressive DSGE

model, over-borrowing, nancial stability, business cycle.

1 PhD student at Kobe University, Graduate School of Economics,

tatoure7@[Link]/171e706e@[Link]. ∗ Supervisor, Economist, Professor of

Economics at Kobe University, Graduate School of Economics, nakamura@[Link].

774
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1 Introduction
The nancial crisis of 2008-2009 made it clear that non-Walrasian features of

credit markets have important macroeconomic consequences. Disruptions in

credit markets can cause large swings in economic activity, and credit-market

imperfections can have large eects on how other shocks aect the macroe-

conomy. Expansions in credit supply, operating primarily through household

demand, called the credit-driven household demand channel, have been an im-

portant driver of business cycles. While this channel helps explain the recent

global recession, it also describes economic cycles in many countries (Mian and

Su, 2018).

For example, over the past 40 years, the boom-bust business cycle gener-

ated by a rise in debt is unique to household debt; increases in rm debt or

government debt do not produce the same pattern (Mian et al., 2017b). Alter-

natively, it could boost aggregate demand by enabling households to increase

consumption (Mian and Su, 2018). Furthermore, Mian et al. (2017b), show

that periods of rising household debt are associated with a rise in the consump-

tion to GDP ratio, an increase in imports of consumption goods and no change

in the investment to GDP ratio. Moreover, in advanced economies, a rise in

household debt generates a consumption boom-bust cycle that is signicantly

more severe than the real GDP boom-bust cycle (IMF, 2017). Therefore, house-

hold debt appears to be crucial in generating these cycles; for example, a rise

in the consumption to GDP ratio by itself does not predict subsequently lower

growth. But a rise in consumption to GDP ratios concurrent with a large rise in

household debt does predict lower growth (Mian et al., 2017b). Recent studies

also show that borrowers' household who experience debt payment reductions

have a lower probability of default and use additional funds to increase spending

on durables (Di Maggio et al., 2017; Dobbie and Goldsmith-Pinkham, 2015).

In addition, the prominence of household debt is also found in emerging

economies. For instance, Bahadir and Gumus (2016), focus on Argentina,

Brasilia, Chile, Korea, Mexico, South Africa, Thailand, and Turkey; and they

show that household debt to GDP ratios in almost all of these countries have

risen substantially since the early 1990s. In contrast, business credit to GDP ra-

tios have been relatively stable. They also show signicant comovement between

household credit and real economic outcomes such as output, consumption, and

investment. Increases in household credit are also associated with substantial

real exchange rate appreciations. In contrast, changes in business credit have

weaker correlations with other real economic outcomes. Further, in a sample of

36 countries with data back to 1970, Mian et al. (2017a), nd that household

debt booms are associated with a rise in the non-tradable to tradable employ-

775
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ment ratio, a rise in the non-tradable to tradable output ratio, and a rise in

the non-tradable price to tradable price ratio. In contrast, a rise in rm debt

is uncorrelated with these outcomes. Mian et al. (2017a) also shows that the

employment patterns are more supportive of credit supply expansion operating

through household demand than an expansion in productive capacity by busi-

nesses. This pattern suggests that the household demand channel is dominant.

However, several empirical studies underline that the credit-driven house-

hold demand channel depresses economic activity level. Indeed, by examining

the real economic consequences of a sudden increase in household debt burdens

by exploiting spatial variation in exposure to household foreign currency debt

during Hungary's late 2008 currency crisis, Vener and Gyongyosi (2017), show

that the revaluation of debt burdens leads to higher default rates and a collapse

in spending. These responses translate into a worse local recession and depressed

house prices. A 10-point increase in debt-to-income raises the unemployment

rate by 0.6 percentage points, driven by employment losses at non-exporting

rms. Consistent with demand externalities of debt nancing, regional foreign

currency debt has negative spillovers on nearby borrowers without foreign cur-

rency debt. Moreover, several studies nd that expansions in household debt

predict more severe recessions (Jordà, Schularick, and Taylor, 2016).

Similarly, debt can also lower labor supply through a debt overhang eect

(Donaldson, Piacentino, and Thakor, 2016). Before that Dynan (2012), analyzes

whether household debt overhang constrained consumption in the U.S. during

the Great Recession, but notes that debt is strongly correlated with regional

housing booms and busts, which Mian and Su (2014a), show have strong eects

on local consumption and employment. Individual-level data also shows that

those taking on the most debt during the expansion phase of the credit cycle cut

spending the most during the ensuing economic downturn (for evidence from

the United Kingdom, see Bunn and Rostom (2015); for Denmark, Andersen et

al. (2014); for a sample of European households, IMF, 2017). This channel

from high household leverage to a fall in demand was rst articulated as the

debt deation hypothesis by Irving Fisher, who pointed out that an economic

slowdown would raise the real burden of debt, which would further slow the

economy through reduced aggregate demand (Fisher, 1933).

Further, in international data, the IMF (2017) study nds a substantial drop

in consumption in the aftermath of household debt expansions. Many countries

nd themselves at the zero-lower bound on nominal interest rates in the after-

math of large expansions in household debt. As illustrated by Hall (2011), and

Eggertsson and Krugman (2012), an economy that hits the zero-lower bound

during the period in which leveraged households cut demand is plagued with

a real interest rate that is too high. As a result, less leveraged households

776
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

do not boost spending suciently to oset the decline in demand coming from

leveraged households. This friction is aggravated by the fact that consumption

of less leveraged households may in general be less sensitive to credit conditions

and interest rates (Su, 2015; Agarwal et al., 2017b). Households that in normal

times have the highest sensitivity of consumption to interest rates and credit

availability nd themselves either unwilling or unable to borrow in the midst of

the downturn that follows credit booms (Mian and Su, 2018).

Furthermore, price rigidities play an important role. For example, the nega-

tive eect of household debt expansion on subsequent growth is larger in coun-

tries with less exible exchange rate regimes (Mian et al, 2017b; IMF, 2017).

In addition, the eect of a change in household debt on subsequent growth is

non-linear: a decline in household debt does not predict subsequently stronger

growth, but a large increase in household debt predicts subsequently weaker

growth (Mian et al., 2017b).

In addition to nominal rigidities, the rise in debt may further depress con-

sumption in the presence of nancial constraints. The rise in debt may increase

defaults and foreclosures, leading to re sales that depress local house prices. A

decline in house prices can tighten collateral constraints, further lowering con-

sumption (Kiyotaki and Moore, 1997). A worse recession also depresses house

prices, creating a two-way feedback between the demand and re-sale channels.

Finally, real rigidities, such as frictions that inhibit a reallocation of employ-

ment towards exporting rms, strengthen the negative eects of debt on output

(Huo and Ríos-Rull, 2016). More generally, recent research suggests that any

shock that leads to a large rise in unemployment in the short-term may have

large and persistent eects on the labor force and large spillovers onto local

economic activity (Yagan, 2017; Acemoglu and Restrepo, 2017). If a large drop

in household demand generates a substantial rise in unemployment, we should

expect the consequences to be large and long-lived. In others approaches, Gupta

(2016) isolates exogenous variation in foreclosures using shocks to interest rates

resulting from details in adjustable rate mortgage contracts. He nds that a

foreclosure leads to further foreclosures and lower house prices in the surround-

ing area. Furthermore, a foreclosure leads to diculty in renancing mortgages

into lower rates for those living close to the foreclosed property, as banks tend

to use the depressed foreclosure price as a comparison. Jordà et al. (2016) use

their disaggregated bank credit data set to estimate the relationship between

bank credit and subsequent nancial crises in 17 advanced economies since 1870.

They nd that since World War II, elevated mortgage credit to GDP ratios pre-

dict nancial crises to the same degree as non-mortgage credit to GDP ratios.

Thus, the aforementioned studies imply that the credit-driven household

demand channel generates a boom-bust cycle in real economic activity. On

777
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the one hand, the rise in household debt can boost the economic activity. On

the other hand, the rise in household debt and house prices heightens the risk

of a nancial crisis (Jordà et al., 2016; Krishnamurthy and Muir, 2017), as

demonstrated by the recent Great Recession. In addition, household debt is a

key asset held by banks, and so a rise in household defaults will directly aect

the banking sector. As mentioned above, Jordà et al. (2016) show that a rise in

mortgage credit to GDP ratios predicts banking crises. Additionally, they show

that recessions associated with high mortgage debt growth and a banking crisis

are the most severe.

Therefore, we ask about the following main question: Should regulators im-

pose macroprudential limits on household debt? Several studies have dealt with

this question, while some implement an optimal capital requirement (Ignazio

and Ester, 2013; Van den Heuvel, 2008; Collard et al., 2017; Begenau, 2020),

others focus on collateral constraint (Kiyotaki and Moore, Justiniano et al.,

2015, and Guerrieri and Iacoviello, 2017). Our main goal in this paper is to

study the eects on economic activity of macroprudential limits on household

debt. To do so, we propose a modied standard monetary DSGE model along

the lines of Kiyotaki and Moore (1997), Justiniano et al (2015), and Guerrieri

and Iacoviello (2017). As in these papers, housing serves the dual role of durable

good and collateral for borrowing. Unlike both elds of research, we incorporate

nancial stress index into our model that we assume is eective to monitoring

the stability of the nancial system and deal with the business cycle. In addition,

our macroprudential policy acts as a shock following an autoregressive process.

We estimate a Bayesian vector autoregressive dynamic stochastic general equi-

librium model by using Junior Maih's perturbation type. The remainder of the

paper is as follows: Section 2 describes the model with macroprudential policy.

Section 3 indicates the estimation of the model. Section 4 shows the results.

Section 5 presents conclusion and policy implication.

2 A DSGE model with macroprudential policy

This section presents a conventional DSGE model with nominal wage and Calvo-

type price rigidities. There are ve types of agents: households, producers,

a monetary authority using a Taylor rule, the government and a prudential

authority using the nancial stress index as a regulatory instrument.

2.1 Households
There is a continuum of measure 1 of agents in each of the two groups (patient

and impatient). The economic size of each group is measured by its wage share,

778
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

which is assumed to be constant. The only dierence between these agents is

that patient's discount factor (βp ) is higher than impatient one (βi ). We assume

that a representative patient household maximizes the expected utility:

∞
X 1
E0 βpt dt [log(cpt − γcpt−1 ) + Φt log(hpt ) − np1+θ ], (1)
t=0
1+θ t

The variables with the symbols (i, p) refer respectively to impatient and pa-

tient households. The terms ct , ht and nt are consumption, housing and hours

worked. The term Φ captures shocks to housing preferences. An increase in

Φshifts preferences away from consumption and leisure towards housing and,

ultimately, housing prices. The term dt captures shocks to intertemporal prefer-

ences. The parameter γ measures habit formation in consumption. The shock

processes follow an AR(1) representation with n.i.i.d normal innovations as fol-
lows:

log(dt ) = ρd log(dt−1 ) + σd ηtd , (2)

log(Φt ) = (1 − ρΦ )log(Φ̄) + ρΦ log(Φt−1 ) + σΦ ηtΦ . (3)

Where ηtd and ηtΦ are n.i.i.d processes with variance respectively σd2 and
2
σΦ , and

Φ̄is the steady state value.

Patient households maximize utility subject to a budget constraint that in

real terms:

(1 + Rt−1 )
cpt + qt hpt + bt + it = wtp npt + qt hpt−1 + bt−1 + rt kt + Θt − τt (4)
πt

Investment and capital are linked by the dynamic neoclassical capital accu-

mulation equation with adjustment costs as follows:

kt+1 = (1 − δ)kt + (1 − AC)it , (5)

Where the adjustment costs is dened by:

Ψ it
AC = ( − 1)2 , (6)
2 it−1

and the capital depreciate at the rate δ . We assume that there is no adjustment
cost in the deterministic steady-state (AC = 0). The ow of expenses includes
current consumption ct , housing services ht (priced, in units of consumption,

at qt ), and loans to impatient households bt . Resources are composed of wage

earnings wt nt , housings on last period, gross interest income on last period

(1+Rt−1 )
loans
πt bt−1 , where πt is the gross ination rate in the price index for
consumption, patient households own the production sector from which they

779
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

receive nominal prots for an amount, Θt , and τt is a lump-sum tax paid by

households. Patient households choose (ct ,ht ,bt ,nt ,it ,kt ), t ≥ 0 to maximize

utility subject to (5) and (6), while the calvo type price is dened as:

Pt
πt = (7)
Pt−1

In which resources spent for consumption, housing and reimbursement of

past borrowing have to be nanced with the income and new borrowing, housing

of past period and labor union fees to be paid.

Given our description of the households problems, the lagrangian function

associated with them is:

∞ ∞
X 1 X
Lp = E0 βpt {dt [log(cpt −γcpt−1 )+Φlog(hpt )− np1+θ
t − λt [cpt +qt hpt +it −wtp npt −qt hpt−1
t=0
1+θ t=0

∞
(1 + Rt−1 ) X Ψ it
− bt−1 −rt kt −Θ+τt ]− λt Qkt [kt+1 −(1−δ)kt −(1− ( −1)2 )it ]}.
πt t=0
2 it−1

Lp stands for patient lagrangian function and λt is the marginal utility on

income.

We assume that impatient households do not save, do not accumulate capi-

tal and do not own nal goods rms, they only work and borrow from patient.

Moreover, at the equilibrium, impatient consumption equal to patient consump-

2
tion following Cúrdia and Woodford (2016). In addition to them, we assume

that impatient consumption is equal to that of patient, which is exactly the

consumption from their labor income, and the share of borrowing likely to be

aected to consumption. Further, we suppose that the share of borrowing to

consumption, and the share of borrowing to house prices is unity. Therefore,

impatient real consumption at the equilibrium mathematically expresses as fol-

lows:
qt
cit = cpt + 1 − . (8)
bt
Where cit is the impatient's consumption which is a function of patients con-

sumption, house prices and borrowing at the steady state.

2 Cúrdia and Woodford (2016) prove that as their model is equivalent to the basic New
Keynesian model, in this case there is a single interest rate; then the marginal utility of
income of both households must equal in equilibrium, implying that the common marginal
utility of income of all households satises the usual Euler equation and then the consumption
of savers and borrowers must equal in equilibrium.

780
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2.2 Producers
There are several stages of production in the economy. Intermediate goods rms

produce dierentiated goods and sell to nal goods producers. The nal good

may be used for consumption, investment, and the monitoring of rms.

2.2.1 Intermediate goods producers

There is a continuum of monopolistically competitive rms of measure 1 denoted
by j producing intermediate goods. They rent capital and labor from two types

of households to produce the intermediate goods yt (j). They solve:

maxyt − wt npt − wti nit − rtk kt , (9)

where wt and wti are respectively patient and impatient real wage and rtk capital
price and operate the following Cobb Douglas production function:

yt (j) = at ktα (j)[(npt (j))1−υ (nit (j))υ ]1−α , (10)

Where 0 < α < 1, kt (j) is capital rented by rm j , npt (j) is the labor input
i
from patient while nt (j) is the labor input from impatient households, and

at denotes the total factor productivity (TFP) shock and follows an exogenous

AR(1) process:
log(at ) = ρa log(at−1 ) + σa ηta . (11)

Where ηta is a n.i.i.d with variance σa2 . We assume that rms set their prices

facing a Calvo-type price rigidity.

2.2.2 Final goods producers

Final good rms (owned by patient households) are perfectly competitive com-

bine the intermediate goods yt (j) to form the nal good yt respectively at the

price Pt and Pt (j). The production function is as follows:

1
ω−1 ω
yt = ( yt (j) ω dj) ω−1 , (12)

Where ω > 1. Prot maximization leads to the demand for good j,

Pt (j) −ω
yt (j) = ( ) yt , (13)
Pt

781
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and free entry leads to the price index:

1
1
Pt = ( Pt (j)1−ω dj) 1−ω . (14)

2.3 Nominal rigidities and monetary policy

There are Calvo-style price rigidities and wage rigidities in the nal good sector.

The Calvo pricing assumption, while appearing rather ad-hoc and unrealistic is

used so frequently because it facilitates aggregation.

Each period, rms have a xed probability 1− of getting to update their

price. Since there are an innite (i.e. continuum) number of rms, there will

be exactly the fraction 1− rms who get to update in any period, and exactly
1 1
the fraction rms who do not. We have : Pt = ( Pt (j)1−ω dj) 1−ω equivalent
1 0
to Pt1−ω = 0
Pt (j)1−ω dj .
The term on the right hand side is both the sum and the average of Pt (j)1−ω over
j. There is total mass 1 of rms, so the sum and the average are the same.

Let's aggregate the production function. By combining equation (12) and

(15), and taking integration, we get:

1 1
kt Pt (j) −ω
at ( )α Nt (j)dj = ( ) yt dj,
Nt Pt
0 0

1 1
kt Pt (j) −ω
at ( )α Nt (j)dj = yt ( ) dj,
Nt Pt
0 0
1
Where Nt = [(npt (j))1−υ (nit (j))υ ], and
0
Nt (j)dj = Nt . After simplication,

we get the following aggregate production function:

at ktα Nt1−α at ktα [(npt (j))1−υ nit (j)υ ]

yt = = (15)
vt vt
1
Where vt = 0
( PPt (j)
t
)−ω , vt essentially measures a distortion introduced by

dispersion in relative prices. As we have seen, there is already a distortion asso-

ciated with the monopoly power of rms (so that there is less capital and labor

in steady state than there would be without the monopoly power). This shows

that there is an additional distortion associated with relative price uctuations

owing to price stickiness. In a exible price model, all rms would choose the

same price. Hence, vt = 1 at all times.

With sticky prices, relative prices will uctuate, and vt 6= 1in general. Nev-

ertheless, to a rst order approximation (about a zero ination steady state),

782
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

this term disappears.

It turns out, however, that we can again appeal to Calvo pricing to write vt
in terms of only gross ination rate and gross reset price ination as follows:

1 1
Pt (j) −ω
( ) = Ptω Pt (j)−ω ,
Pt
0 0

1
1− 1
Pt (j) −ω
= Pt#−ω + Pt−1 (j)−ω ,
0 0 1−

1 1
Pt (j) −ω
= (1 − )Pt#−ω + Pt (j)−ω ,
0 1−

Back to the expression of vt , we have:

1 1
−ω P#
vt = Ptω Pt (j) = (1 − )( t )−ω + Ptω Pt−1 (j)−ω ,
Pt
0 1−

1
P# Pt−1 −ω
vt = (1 − )( t )−ω ( ) + Ptω Pt−1 (j)−ω
Pt−1 Pt
1−

1
−ω ω Pt−1 (j) −ω
vt = (1 − )π #−ω πtω + Pt−1 Pt ( ) dj,
Pt−1
1−

Thus, the combination results in a backward Phillips curve that can be

written as:

vt = (1 − )πt#−ω πtω + πtω vt−1 (16)

Where π# is the steady state reset price ination. Similarly, dene gross

reset ination as:

Pt#
πt# = = 1 + πt# .
Pt−1
Divide both sides of the equation for the aggregate price level by Pt1−ω to

get:

1
πt = [(1 − )πt#1−ω + ] 1−ω (17)

By following the canonical three-equation new Keynesian model of Clarida

et al. (2000), we assume the central bank follows a modied forward-looking

interest-rate rule, adjusting the interest rate in response to changes in expected

783
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

future ination and output. Thus, ination and output should be interpreted

as deviation from their steady-state or trend values as follows:

πt+1 yt+1
Rt = απ + αy + et , απ > 0, αy ≥ 0 (18)
π yt
and et denotes an autoregressive monetary policy shock as follows:

log(et ) = ρr log(et−1 ) + σr ηtR . (19)

2.4 Prudential policy

A prudential policy that is sucient to lead to nancial stability is one that

rules out households' over-borrowing, and any losses banks may incur, so that

households internalize the eect of their actions on others during the credit con-

traction. This subsection rst gives some details of what causes over-borrowing,

and characterizes the prudential policy that rules out over-borrowing.

One such reason for over-borrowing is the presence of aggregate demand

externalities (for example, Eggertsson and Krugman, 2012; Farhi and Werning,

2015; Huo and Ros-Rull, 2016; Korinek and Simsek, 2016; Schmitt-Grohe and

Uribe, 2016; Guerrieri and Lorenzoni, 2017). In these models, there is a friction

such as nominal wage rigidity or a lower bound on the real interest rate that

prevents the economy from adjusting when credit contracts and there is a drop

in demand from leveraged households. Thus, households do not internalize the

eect of their future decline in demand on the income of other households, and

they therefore rationally take on more debt than is socially optimal.

Another reason for over-borrowing is the presence of pecuniary externalities

due to re sales as discussed in Shleifer and Vishny (1992), Kiyotaki and Moore

(1997), Caballero and Krishnamurthy (2001), Lorenzoni (2008), Bianchi (2011),

Davila (2015), and others. Suppose that an asset, such as a house, is used as

collateral for borrowing. If households borrow in the present, they will tend

to drive up the price of the asset. After a negative shock, households will be

forced to delever by re-selling the collateral which reduces the price of collateral

and hence tightens the borrowing constraint. In this way, the collateral price

channel adds to the aggregate demand externality. In both cases, households

may rationally decide to take on more debt during an expansion than is socially

optimal because they do not internalize the eect of their actions on others

during the credit contraction.

Moreover, a reason from heterogeneous beliefs is that high levels of bank

credit also seem to be associated with a predictable crash in equity prices for

banks (Baron and Xiong, 2017), and banks that expand credit most rapidly have

predictably worse returns in the subsequent years (Fahlenbrach et al., 2017).

784
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Another approach, relying on behavioral biases can be viewed as part of a

process that leads to credit supply expansions. For example, perhaps lenders

begin lending to lower credit quality borrowers because they mistakenly believe

that the probability of default for such borrowers is lower. Or perhaps mortgage

credit spreads fall because lenders become more optimistic about house price

growth, as in Kaplan et al. (2017).

Thus, most of these cases can heighten the risks of nancial crises. There-

fore, to deal with such issues, regulators impose a macroprudential limits on

households debt.

Moreover, the recent public debate over nancial regulatory reform has been

heated and involved. Some have argued that we need an agency charged with

monitoring nancial market developments that are deemed to pose a risk to the

entire nancial system or to rms deemed systemically important. Presumably,

one of the requirements of such a systemic risk regulator would be to measure

nancial market stress.

There are many ways to measure nancial market stress. One is to look at

an interest rate spread designed to measure default risk, such as the dierence

between yields on a risky asset (e.g., corporate bonds) and a risk-free asset

(e.g., U.S. Treasury securities). However, nancial stress can also arise in other

dimensions. One type of risk prominent in the recent nancial crisis was the

inability of many nancial institutions to secure funding to nance their short-

term liabilities, such as repurchase agreements. This type of risk is known as

liquidity risk.

To overcome a potential problem of focusing solely on one indicator at the

expense of others, some economists have combined several indicators designed to

measure nancial market stress into one summary variable, like an index number

called nancial stress index (f si). The nancial stress index is constructed by

the Federal Reserve Bank of St. Louis (ST LF SI ); it is based on 18 weekly

data series. There are seven interest rate series, six yield spreads, and ve other

nancial series (credit, volatility, spread, leverage, etc.) that begin in late 1993.

Each of these variables captures some aspect of nancial stress. Accordingly,

as the level of nancial stress in the economy changes, the data series are likely

to move together. Because each variable was standardized, the coecient of

a variable represents the inuence of a 1 standard deviation change in that

variable on the nancial stress index.

As one contribution to this work, we add nancial stress index as our pruden-

tial policy instrument to the lending constraint along the line of Guerrieri and
3
Iacoviello (2017). In our model, the macroprudential policy indicator (f si) acts

as a macroprudential policy schock with an autoregressive AR(1) representation

as follows.

785
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

bt−1
bt = ν + (1 − ν)mqt hit + f sit , (20)
πt

f sit = ρf si f sit−1 + σf si ηtf si . (21)

Where ν measures the degree of inertia in the borrowing limit as in Guerri-

eri and Iacoviello (2017), and m is the maximum loan-to-value ratio. Gertler,

Kiyotaki and Prestipino (2016) show that for example, nancial crises are in-

herently nonlinear events, often featuring a simultaneous sudden collapse in

asset prices and rise in credit spreads. Similarly, Kiyotaki and Moore (1997)

underline that in addition to nominal rigidities, the rise in households debt may

increase default, leading to re sale that can depress house prices, tighten the

nancial constraint and further lowering consumption. Therefore, a number of

papers have captured this nonlinearity by allowing for the possibility that the

balance sheet constraints do not always bind (Justiniano et al., 2015; Guerrieri

and Iacoviello, 2017). Incorporating into the borrowing constraint as in equa-

tion (Eq.20), we allow nancial stress index for curtailing households debt to

prevent over-borrowing and contain nancial crises that can happen from credit
4
supply expansions. Financial stress index also solves for nonlinear events when

nancial constraints bind as it imposes a limit to credit expansions as shown in

section 4.
The average value of the index, which begins in late 1993, is designed to be

zero. Thus, zero is viewed as representing normal nancial market conditions.

Values below zero suggest below-average nancial market stress, while values

above zero suggest above-average nancial market stress. However, it is impor-

tant to be cautious when interpreting the sign of a coecient. For instance, the

de-meaned, standardized data series can (and will) take on negative values, so

a negative coecient does not imply that the variable always contributes to a

decline in nancial stress. A negative coecient multiplied by a negative data

value will result in a positive contribution to nancial stress. In other words,

although the eective federal funds rate cannot dip below zero, the de-meaned,

standardized eective funds rate data series can take on a negative value when a

federal funds rate observation is less than the federal funds rate sample average.

Thus, since the federal funds rate variable has a negative coecient, federal

funds rate data points below the sample average contribute to an increase in

the FSI, while observations greater than the sample average contribute to a

decrease in the nancial stress index.

3 According to Guerrieri and Iacoviello (2017), an interpretation of this borrowing constraint

is that, with multi-period debt contracts, the borrowing constraint on housing is reset only for
households that acquire new housing goods or choose to renance. Of course, in the face of
home equity line of credits, adjustments of the borrowing constraint may also reect lenders'

786
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2.5 Government
In our model, the government consumes some private output as set in the next

equation:

Gt = gt yt , (22)

Where gt denotes exogenous government expenditures which follows an AR(1)

process:

log(gt ) = (1 − ρg )logḡ + ρg log(gt−1 ) + σg ηtg . (23)

Where ηtg is a n.i.i.d with variance σg2 and ḡ is the steady state value.

2.6 Market clearing conditions

The market clearing condition for goods is:

yt = ct + υit + Gt (24)

Where ct denotes aggragate consumption and is given by:

ct = (1 − υ)cpt + υcit (25)

Where it is the amount of per-capita investment undertaken by the savers,

who are the only households accumulating capital. This constraint is obtained

by aggregating the budget constraints of borrowers and savers with that of

the Government, using the zero prot conditions of the competitive rms, the

denition of prots for the intermediate rms.

Equilibrium in the housing market is given by:

hpt + hit = 1 (26)

2.7 Equilibrium
A denition of equilibrium in this economy is standard and the symmetric equi-

librium policy functions are determined by the following equations:

perceived changes in the collateral value. Justiniano et al. (2015) , who study the determinants
of household leveraging and deleveraging in a calibrated dynamic general equilibrium model,
adopt an analogous specication.
4 Financial crises are then periods where the constraints bind, causing an abrupt contraction
in economic activity (Gertler, Kiyotaki and Prestipino, 2016).

787
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

•The rst order conditions of the patient household.

dt (cpt − γcpt−1 )−1 − γβp Et dt+1 (cpt+1 − γcpt )−1 = λt ,

λ1t+1
λt = βp (1 + Rt )Et { },
πt+1

Φdt
+ βp Et [λt+1 qt+1 ] = λt qt ,
hpt
Labor supply:

dt (npt )θ = λt wtp ,

The Euler equations for investing in physical capital are:

λt Qkt = βp Et λt+1 [(1 − δ)Qkt+1 + rt+1

k
],

Ψ it it it λt+1 k it+1 it+1 2

1 = Qkt {[1− ( −1)2 ]−Ψ( −1)( )}+βp Et Q Ψ( −1)( ) .
2 it−1 it−1 it−1 λt t+1 it it

Where Qkt is the Tobin's Q on capital.

•The rst-order conditions of the impatient household.

Labor supply:

dt (nit )θ = λ2t wti

At each point in time producers employ the stocks of labor from both house-

hold and capital, pay them their marginal products, and sell the resulting out-

put. Hence, from the Cobb-Douglas production function, we have:

yt
rtk = (1 − α) ,
kt

yt
wtp = α(1 − υ) ,
npt

yt
wti = αυ .
nit

2.8 Stationary equilibrium

Since we have growth in this model induced by technological change, most of

the variables are growing in average.

788
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

To solve the model, we need to make the variables stationary.

First, we work on the rst order conditions of the patient household.

at−1 −1 at+1
dt (cpt /at − γcpt−1 /at−1 ) − γβp Et dt+1 (cpt+1 /at+1 − γcpt /at )−1 = λt at ,
at at

λt+1 at+1 at
λt at = βp (1 + Rt )Et { },
πt+1 at+1

Φdt
+ βp Et [λt+1 at+1 qt+1 /at+1 ] = λt qt at ,
hpt /at

dt npθ p
t at = λt wt at ,

λt+1 at+1
Qkt = βp Et [(1 − δ)Qkt+1 + rt+1
k
],
λt at

Ψ it /at at it /at at it /at at

1 = Qkt {[1 − ( − 1)2 ] − Ψ( − 1)( )}
2 it−1 /at−1 at−1 it−1 /at−1 at−1 it−1 /at−1 at−1

λt+1 at+1 at k it+1 /at+1 at+1 it+1 /at+1 at+1 2

+βp Et Q Ψ( − 1)( ) ,
λt at at+1 t+1 it /at at it /at at

Ψ it /at at
AC = ( − 1)2 .
2 it−1 /at−1 at−1
Second, we work on the rst-order conditions of the impatient household.

dt niθ i
t at = λt wt at .

From the production function.

yt /at
rtk = (1 − α) ,
kt /at
By substituting wtp and wti in the optimality conditions of labor supply re-

spectively, we draw hour work of each household npt and nit as follows:

α(1 − υ)yt /at λt at 1+θ

1
npt = [ ] , (27)
dt

αυyt /at λt at
nit = [ ]. (28)
dt

789
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Then, central bank follows its Taylor rule:

πt+1 yt+1 /at+1 at+1

Rt = απ + αy + et , απ > 0, αy ≥ 0 (29)
π yt /at at
Besides, (Song, 2010) underlines that the house price depends on the hous-

ing demand shock, transaction costs, loan-to-values, the ratio of consumption

to house ξ, the elasticity of intratemporal substitution κ and the average of dis-

count factors across unconstrained households and constrained households. In

addition, he stipulate that house prices increase when house loan-to-value ratio

m increases and when the transaction costs decreases, ceteris paribus. When

the transaction costs are equal to zero, house prices are determined by the dis-

count factor, consumption share to houses and κ. Finally, by using minimum

distance estimation with macro data from 1970Q1 to 2008Q4, Song estimated

κ to be 0.592 with standard error of 0.053, thus providing strong support of

preferences over housing and consumption for nonseparability and complemen-

tarity. Hence, at the steady state in absence of transaction costs in our work,

house prices satisfy the following equation:

Φ̄ 1
qt = (ξ) κ . (30)
1 − βi − (βp − βi )m
We compute the ratio of consumption to house ξ using our database.

3 Estimation of the model

F igures1 − 4 show a look at the data that motivates our analysis. The gures

indicate that the credit-driven households demand channel positively inuence

the economic activity by boosting consumption and therefore the real gross

domestic product during the periods before 2008 and after 2009. However, a

disruption in credit markets can cause large swings in economic activity, and

credit-market imperfections can have large eects on how other shocks aect the

macroeconomy as shown in gures for the period between 2008-2009. During

that period, gures elucidate that household credits were about 98.6% of GDP

and the default in payement led to a decline in consumption, as a result the real

output collapsed as well for U.S. economy. As a signal to the accelerated increase

in credit, nancial stress index shows a sharp rise about 5.307% weekly or 55.71%

quarterly in october 2008 reporting a very high credit expansion between 2008-

2009 in order to track stress and vulnerabilities of nancial system.

We rst use Bayesian estimation methods to size the structure parameters

of the model so that its steady state matches some key statistics.

790
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1: Total Credit to Households and Non-Prot Institutions Serving House-

holds, Adjusted for Breaks, for United States, Percentage of GDP, Quarterly,
Not Seasonally Adjusted. Code:QUSHAM770A. The data sample runs from
1993Q1 through 2018Q4 for U.S. national economy.
4.6

4.55

4.5
Logarithm-of-Credit-to-Households

4.45

4.4

4.35

4.3

Credit-to-Households
4.25

4.2

4.15

4.1
1993Q1 1996Q2 1999Q3 2002Q4 2006Q1 2009Q2 2012Q3 2015Q4
Time-in-Quarter

Figure 2: Real Personal Consumption Expenditures, Billions of Chained 2012

Dollars, Quarterly, Seasonally Adjusted Annual Rate. Code: PCECC 96. The
data sample runs from 1993Q1 through 2018Q4 for U.S. national economy.
9.5

9.4

9.3
Logarithm-of-consumption

9.2

9.1

8.9

consumption

8.8

8.7
1993Q1 1996Q2 1999Q3 2002Q4 2006Q1 2009Q2 2012Q3 2015Q4
Time-in-Quarter

791
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 3: Real gross domestic product is the ination adjusted value of the goods
and services produced by labor and property located in the United States. For
more information see the Guide to the National Income and Product Accounts
of the United States (NIPA). For more information, please visit the Bureau of
Economic Analysis. Code GDPC1. Source: U.S. Bureau of Economic Analysis.
Units: Billions of Chained 2012 Dollars, Seasonally Adjusted Annual Rate.
Frequency: Quarterly.
9.9

9.8

9.7

9.6
Logarithm-of-GDP

9.5

9.4

real gross domestic product

9.3

9.2

9.1
1993Q1 1996Q2 1999Q3 2002Q4 2006Q1 2009Q2 2012Q3 2015Q4
Time-in-Quarter

Figure 4: St. Louis Fed Financial Stress Index, Index, Weekly, Not Seasonally
Adjusted. Code: STLFSI. The average value of the index, which begins in
late 1993, is designed to be zero. Thus, zero is viewed as representing normal
nancial market conditions. Values below zero suggest below-average nancial
market stress, while values above zero suggest above-average nancial market
stress.
60

30
Index

-10

financial stress index

-20
1994Q1 1997Q2 2000Q3 2003Q4 2007Q1 2010Q2 2013Q3 2016Q4
Time-in-Quarter

N ote :we convert the weekly data in quarter before plotting the nancial stress
index series.

792
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

3.1 Calibration
The coecients of the monetary policy rule are assigned the values απ = 2
and αy = 0.25 as in (Woodford, 2003, chapter4; Cúrdia and Woodford, 2016).

Therefore, we choose a value of 0.999 for the savers'discount factor (βp ) close to
that in Justiniano et al. (2015), to obtain an annualized steady state nominal

3.24%, close to the average Federal Funds Rate. For the borrow-
interest rate of

ers'discount factor (βi ), we pick a value of

0.990 akin to Justiniano et al. (2015).
We also pick a Frisch elasticity of labor supply (θ ) equal to 1, typical in the Real

Business Cycle literature (Hansen, 1985). We set the share of borrowers equal

to 61% (see Justiniano et al., 2015). The capital share α = 0.3 and the deprecia-
tion rate δ = 0.025. The weight on housing in the utility function Φ̄is set at 0.05
close to that calculated by Guerrieri and Iacoviello (2017) which is 0.04. We

keep the maximum loan-to-value ratio m simulated by Guerrieri and Iacoviello

(2017) to set it at 0.9. We set the Central Bank's ination target (π ) equal to the

average gross rate of ination (1.005, or 2% per year), which gives a gross reset

price ination to1.0266. = 0.8, this says that rms change their prices once
every 5 quarters on average. ω = 6, implying steady state markups of around

20 percent. We set the elasticity of intratemporal substitution κ = 0.592, as

estimated by Song (2010). Finally, we estimate the ratio consumption-to-house

prices ζ = 1.6216 using data. We use a Bayesian Vector Autoregressive methods

to estimate the priors distributions of all other parameters with quarterly data

over 1993Q1-2018Q4, except the macroprudential policy series for which we t

an AR(1) process using weekly data over 1993-2018. All data are related to

U.S. national economy.

3.2 Model ltering

Prior to estimation, Kalman lter is used to smooth, update and lter the data.

Indeed, in a state-space representation, we need a transition equation, which in

our case is given by the solution of the system and also a measurement equation

connecting the data to the variables of the model. Thus, the transition and

measurement equations can be respectively expressed as follows:

ςt = Tt (ςt−1 , ηt ), ηt ∼ N (0, Ht ).

yt = Zt (ςt , εt ), εt ∼ N (0, I).

Given those we can compute

p(ςt | ςt−1 ), p(yt | ςt ), p(yt | ςt−1 ).

Purpose of ltering: compute conditional densities.

793
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1: Estimated and calibrated parameters values.

Calibrated parameters Nomenclature Value

βp Discount factor, patient agents 0.999

βi Discount factor, impatient agents 0.990
αp Capital share in production 0.3
δ Capital depreciation rate 0.025
υ Share of borrowers 0.61
ω Price markups 6.0
1− Share of rms who update their price 0.2
κ Elasticity of intratemporal substitution 0.592
π Steady-state gross ination rate 1.005
π# Steady-state gross reset price ination 1.0266
θ Labor disutility 1
ζ Consumption-house prices ratio 1.6829
m Loan-to-value ratio 0.9

Estimated parameters Priors [mean, std]

απ Ination response, Taylor rule gamma [2, 0.1]

αy Output response, Taylor rule beta [0.25, 0.1]
ρh AR(1) housing shock beta [0.75, 0.01]
ρr AR(1) monetary shock beta [0.5, 0.1]
ρg AR(1) government spending shock beta [0.8, 0.1]
ρa AR(1) technology schock beta [0.75, 0.1]
ρf si AR(1) macroprudential shock beta [0.75, 0.1]
ν inertia borrowing constraint beta [0.75, 0.01]
γ habit formation in consumption beta [0.70, 0.1]
Ψ investment adjustment cost gamma [5, 2]
ρd intertemporal shock beta [0.75, 0.1]
σh Std. housing demand shock invgamma [0.01, 1]
σr Std. interest rate shock invgamma [0.01, 1]
σg Std. government spending shock invgamma [0.01, 1]
σa Std. technology shock invgamma [0.01, 1]
σf si Std. macroprudential shock invgamma [0.01, 1]
σd Std. intertemporal shock invgamma [0.01, 1]
N ote: The table summarizes calibrated parameters, priors estimates of the

parameters for the full model.

794
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 5: Kalman ltering.

10
filtered
updated
smoothed
8

-2

-4
1993Q1 1996Q2 1999Q3 2002Q4 2006Q1 2009Q2 2012Q3 2015Q4

N ote : we computed the kalman lter for seven series: output growth (YGR),
consumption (CONS), credit to households (BOR), leverage (LEV), house prices
(HOUSE), ination (INFL), and interest rate (INT). We do not include nancial
stress index (FSI) series as it does not appear as a measurement equation in our
model.

p(ςt | y1:t−1 ), p(ςt | y1:t ).

Thereby update our knowledge of the system each time a new observation

yt is brought in.

Bayesian ltering such as Kalman lter provides a unied recursive approach

for these types of problems. Kalman showed two important things:

First, a wide class of problems could be encapsulated in a simple linear state-

space model. Second, due to the Markovian structure of the model, the calcula-

tions needed for practical application of the model could be set up in a recursive

form, which is convenient for computer implementation. Thus, Kalman lter is

used for constructing the likelihood function which is useful for Bayesian estima-

tion and model comparison. Hence, the gure 5 presents the result of Kalman

ltering for seven observed variables.

3.3 Model solution and estimation

The model is solved nonlinearly around the balance growth path by using Junior

Maih's perturbation type. The solution method links the rst-order approxi-

mation of the model around the same point under each regime. The rst-order

perturbation solution can be expressed as the whole system in terms of the state

vector.

As the model is stationary (subsection 2.8), hence at the steady state, xt =

795
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

>(z̄) = x, where x is the approximation point dened in (appendixA).

 
xt
zt+1 ≡  σ  = J(zt ) + U zt ,
 

σηt+1
 
xt−1
where zt =  σ is the state variables vector, σ the perturbation term and
 

ηt
ηt the shock.

 
>(zt ) " #
0(nx +1)×nz
J(zt ) ≡  mσ  , U ≡
 
ηt+1 mσ
0
Expanding v    
xt+1 >(J(zt ) + U zt )
 xt   >(zt )
   

   
v ≡ xt−1  = 
   mx zt ,

 ηt   mη0 zt
   

σ mσ zt

Hence v = v(z, U ).
The steady state implications are as follows:

xt = >(z̄) = x,

   
>(z̄) x
 x  x
   
   
v=
 x  = x ,
  
 0  0
   

0 0

Hence at the steady state, Et f (xt+1 , xt , xt−1 , ηt ) = 0,becomes:

f (x, x, x, 0, 0) = 0.

The rst-order expansion involves:

Dierentiating

Et f (v(z, U )) = 0

with respect to z
Et [fv ]α [vz ]α
j = 0.

796
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[fv ]iα denotes the derivative of the ith row of f with respect to the αth row
α th
of v . Similarly, [vz ]j denotes the derivatives of the α row of v with respect to
th i α i α
P
the j row of z . So we have that [fv ]α [vz ]j implies α [fv ]α [vz ]j .
The tensor above is unfolded into

Et fv vz = 0,

where vz = a0z + a1z U.

 
>z Jz
 >z 
 

a0z ≡ 
 
 mx  ,

 mη0 
 

mσ
 
>z
0
 

a1z ≡ 
 
 0 ,

0
 

0
and  
>z
Jz = mσ  ,
 

Evz = a0z .

By partitioning the rst-order expansion, we get:

fv = [f+ f0 f− fη fσ ],

>z = [>x >σ >η0 ],

hence,

0 = f+ (>x >z + >σ mσ + >z EU ) + f0 >z + f− mx + fη mη0 + fσ mσ .

The quadratic matrix polynomial can be written as:

f+ >x >x + f0 >x + f− = 0.

The solution of this problem can be obtained using various procedures and

797
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

requires the computation of derivatives of the structural model (symbolic, au-

tomatic or numerical, etc.). In our work, we computed symbolic derivatives.

The impact of perturbation term leads to the following equation:

f+ [>x >σ + >σ ] + f0 >σ + fσ = 0.

If we dene:

Aσ ≡ f+ >x + f0 ,

Aσ >σ + f+ >σ + fσ = 0.

If fσ = 0 then >σ = 0 and we have certainty equivalence meaning that all

residuals should be zero after solving the model.

The impact of contemporaneous shocks gives:

f+ >x >η0 + f0 >η0 + fη = 0.

Finally, the Taylor's rst-order approximation of the solution implies:

>(zt ) ' x̄ + >z (zt − z̄).

4 Results
The estimation is based on observations for eight series: real gross domestic

product, real household consumption, price ination (CPI), total household

credit, leverage, real house prices, eective federal funds rate and nancial stress

index. The observations span the period from 1993Q1 to 2018Q4. Note that

we convert the weekly data in quarter before plotting the gure 4 (Appendix B

describes the data in detail). The estimation period has a particular importance

as far as it presents three key-subperiods. First, period of relative stability of the

1990s with a deregulated nancial system; second, the period of large swings in

debt and house prices observed in the 2000s, which are likely to represent large

deviations from the stationary state. Finally, the post-crisis of the 2010s marked

by a relative stability again and a deregulated credit market. The model features

5 shocks: total factor productivity shock, macroprudential policy, monetary

policy, intertemporal preferences and housing preferences.

4.1 Business cycle components

The detrending procedure used is the method employed by Hodrick and Prescott
5
(1997). Perhaps the most popular way of isolating the cyclical component of

798
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 6: HP detrended gross domestic product and Time series plots of the
cyclical component of output along with the cyclical component of households'
debt and consumption.
60 80
Hodrick-Prescott Detrended GDP GDP
Credit-to-Households
60
40

20
20

0
0

-20
-20

-40
-40

-60

-60
-80

-80 -100
1993Q1 1996Q2 1999Q3 2002Q4 2006Q1 2009Q2 2012Q3 2015Q4 1993Q1 1996Q2 1999Q3 2002Q4 2006Q1 2009Q2 2012Q3 2015Q4
80
GDP
60 Consumption

-20

-40

-60

-80

-100

-120
1993Q1 1996Q2 1999Q3 2002Q4 2006Q1 2009Q2 2012Q3 2015Q4

a series is by using the HP lter. For quarterly data, it is common to use

a smoothing parameter of 1600. A smoothing parameter of innity gives rise

to a linear trend and the cycle being dened as deviations about that linear

trend. Below is a plot of the HP detrended component of output and time series

plots of the cyclical component of output along with the cyclical component of

households' debt and consumption.

The graphs in gure 6 underline relevant business cycle ndings for the se-

lected observations from the data. First, consumption appears to be procyclical

in the sense that its cyclical component appears to be positively correlated with

the cyclical component of output. Conversely, credit to households appears rela-

tively countercyclical as its cyclical component looks to be negatively correlated

with the cyclical component of output. Secondly, both consumption and credit

to households appear much more volatile than output. These results are con-

sistent with existing literature that deals with credit-driven household demand

channel and economic activity. For example, evidence from IMF (2017) shows

that in advanced economies, a rise in household debt generates a consumption

boom-bust cycle that is signicantly more severe than the real GDP boom-bust

799
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Selected Business Cycle Moments.

Variables Standard deviation (a) Correlation (b) Autocorrelation (c)
Output 1.00(0.18) 1.00 0.93
Credit-to-households 1.80(0.33) -0.64 0.97
Consumption 1.60(0.30) 1.00 0.93
N ote : The U.S. time series used are the Real gross domestic product, total
Credit to Households and Non-Prot Institutions Serving Households, and real
Personal Consumption Expenditures. All series are seasonally adjusted, logged
and detrended using Hodrick and Prescott lter. All series are described in
details in AppendixB . The standard deviations and correlations with output are
sample means of statistics computed for each of 100 simulations. The numbers in
parentheses are sample standard deviations of these statistics. Before computing
any statistics each simulated time series was logged and detrended using the
same procedure used for the U.S. time series.

cycle. Further, as housing prices rise, household borrowing also rises fueling

a debtdriven consumption boom. As housing prices decline, households are

forced to borrow less and the deleveraging pushes the economic contraction into

overdrive (Guerrieri and Iacoviello, 2017).

The table 2 formalizes some of relevant observations in our work. It shows

the relative volatility with output (a), the correlation with output (b),and the

autocorrelation coecient (c) of credit to households and consumption. The ta-

ble conrms our above ndings in gure (6). Consumption and credit to house-

holds are more volatile than output. Consumption is procyclical and perfectly

positively correlated with output, while credit to households is countercyclical

and negatively correlated with output. All of the series are fairly persistent in

the sense of having autocorrelation coecients close to 1. Therefore, a rise in

households' debt subsequently predicts output growth operating through con-

sumption. Thus, the table also shows that a rise in household debt generates

a consumption boom-bust cycle that is signicantly more severe than the real

GDP boom-bust cycle.

4.2 Impulse response functions to a macroprudential policy shock

We now turn to the impulse-response functions, focusing on nancial stress in-

dex, and expressing the responses of some unobserved and observed (obs) vari-

ables such as credit to households, leverage, patient, impatient, and aggregate

5 This method involves choosing smoothed values {s }T for the series {x }T which solve the
t t 1
P 1 −1
following problem: min{(1/T ) Tt=1 (xt −st )2 +(λ/T ) Tt=2 [(st+1 −st )−(st −st−1 )]2 },where
P
λ > 0 is the penalty on variation, where variation is measured by the average squared second
dierence. A larger value of λ implies that the resulting {st } series is smoother. Following
Prescott (1997), I choose λ = 1600. Deviations from the smooth series are formed by taking
dt = xt − st .

800
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 7: Impulse responses functions to a macroprudential policy shock.

Impulse responses to a macroprudential policy shock
0.01 0.02 0.2

0.005 0.01 0.1

0
0 0
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -4 10 -5
0.015 2 6

0.01 4
1
0.005 2

0 0 0
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -4 10 -4 10 -3
1.5 2

1 10
1
5
0.5

0
0 0
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -6 10 -6
0 0
0.02

-2 -2
0.01

-4 -4
0
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

consumption, output, interest rate and ination since our specication stands

for a state space representation. Figure 7 reports the responses to a macropru-

dential policy shock (a 1 standard deviation innovation to ηtf si ) governed by the

process f sit in Eq.(21). To the policy shock, credit to households increases in
the model as well as in the data. At its peak, credit to households rises about

1.6% above its baseline, while its measurement variable (BOR) rises about 21%
above its baseline. When nancial stress index reaches its steady state (nancial

stress index is zero) at periods twelve and half (13), credit to households (B) is

0.5% above its baseline. These ndings suggest that the rise in credit to house-

holds likely to ensure the nancial system stability uctuates between 0.5% and

1.6%, or increases at about 21% of GDP quarterly with a 1.2% of leverage. Fur-

ther, macroprudential policy lowers and raises lending rate thereafter to keep

ination at low and then zero; output, impatient and patient consumption, and

aggregate consumption rise. Moreover, patient consumption is lower than that

of impatient. This nding is consistent with economic literature dealing with

households debt and its interaction with economic growth.

For example, in international data, the IMF (2017) study nds a substan-

tial drop in consumption in the aftermath of household debt expansions. Many

countries nd themselves at the zero-lower bound on nominal interest rates in

the aftermath of large expansions in household debt. As illustrated by Hall

(2011), and Eggertsson and Krugman (2012), an economy that hits the zero-

lower bound during the period in which leveraged households cut demand is

plagued with a real interest rate that is too high. As a result, less leveraged

households do not boost spending suciently to oset the decline in demand

801
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

coming from leveraged households. This friction is aggravated by the fact that

consumption of less leveraged households may in general be less sensitive to

credit conditions and interest rates (Su, 2015; Agarwal et al., 2017b). House-

holds that in normal times have the highest sensitivity of consumption to interest

rates and credit availability nd themselves either unwilling or unable to borrow

in the midst of the downturn that follows credit booms (Mian and Su, 2018).

4.3 Impulse responses to an horizon shock

We now focus on the eectiveness of macroprudential policy in projection. In

addition, regulators impose macroprudential policy on households debt in order

to monitor nancial stability in 4 years (horizon shock). Figure 8 indicates the

responses to a macroprudential policy shock (a 1 standard deviation innovation

to ηtf si ) governed by the process f sit in Eq.(21),and expressing the responses of
credit to households, leverage, patient, impatient, and aggregate consumption,

output, interest rate and ination. To the 4 years horizon shock, we nd similar
results as in gure 7,an increase of credit to households (B) respectively at about
1.6% above its baseline, and 0.5% as nancial stress index is at steady state. By

contrast, credit to households as a measurement variable (BOR) rises about 20%

above its baseline. Therefore, the results of horizon shock conrm our ndings

as in gure 7,stipulating that to maintain nancial stability, the rise in credit to

households should uctuate between 0.5% and 1.6%, or between 20% of GDP

and 21% of GDP quarterly with an increase of 1.2% of leverage. Further, in

horizon shock, macroprudential policy response is a small increase in lending rate

to keep ination at low, and thereafter at zero; output, impatient and patient

consumption, and aggregate consumption rise. Again, patient consumption is

lower than that of impatient as in the previous case.

4.4 Impulse responses to a monetary policy shock

Figure 9 shows the responses to a monetary policy shock (a 1 standard devi-
R
ation innovation to ηt ) governed by the processet in Eq.(19). We express the
responses of credit to households, leverage, patient, impatient, and aggregate

consumption, output, interest rate and ination. Since a monetary shock does

not aect nancial stability in our model, prudential policy does not respond to

this shock. Moreover, the responses to the monetary shock is set in two stages.

In the rst stage, the monetary response is a sharp increase in the lending rate

to moderate the expansion of borrowing and keep ination at low. This dis-

courages consumption (patient, impatient and aggregate), and output. In the

second stage, the monetary response is a steady decrease in lending rate to rise

the expansion of credit supply, as a result, ination increases and thereafter

802
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 8: Impulse responses functions to a macroprudential policy for a horizon

shock.
Impulse responses
-3
to a macroprudential policy shock
10
0.01 20 0.2

0.005 10 0.1

0
0 0
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -3 10 -4 10 -5
2 6
10
4
1
5
2

0 0 0
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -4 10 -4 10 -3
1.5 2

1 10
1
5
0.5

0
0 0
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36
-6
10 10 -6
5
0.02 2

0 0
0.01
-2
-5
0 -4
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

N ote :The blue line indicates the impulse responses functions of stochastic simu-
lations as in gure 7, while the red line stands for the impulse responses functions
for four years horizon. We do not the impulse responses function of monetary
policy shock as we would like to simulate the eectiveness of macroprudential
policy by doing so.

stays at zero. Therefore, consumption (patient, impatient and aggregate), and

output rise.

Thus, our model highlights important policy implications: as in some re-

search focusing on capital requirement to manage risk-taking incentives, nan-

cial stress index can eectively monitor nancial stability, while changes in the

policy interest rate cannot. When nancial stress index curtails credit supply to

households, a contraction ensues, and the policy interest rate is cut. With this

chain of causality, prudential policy is procyclical and monetary policy is coun-

tercyclical. This nding has been proved by a large number of relevant works of

how monetary and prudential policies may be substitutes for each other or move

to oset each other's eects aftermath of the recent nancial crisis of 2007-2009

(Yellen, 2010; Canuto, 2011; Macklem, 2011; Cecchetti and Kohler, 2012; Wolf,

2012; Ignazio and Ester, 2013; Collard et al., 2017).Therefore, in our model, one

policy is contractionary and the other expansionary in order to manage credit

risk with the smallest possible adverse eects on consumption.

803
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 9: Impulse responses functions to a monetary policy shock.

Impulse responses
-6
to a monetary policy shock-5
10 10
1 0
0
-2
-2
0 -4
-4
-6
-6
-1
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -5 10 -7 10 -8
0 0 0

-0.5 -0.5 -2

-1 -1 -4

-1.5 -1.5 -6
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -7 10 -7 10 -6
0 0 0

-0.5
-0.5 -5

-1
-1 -10
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

10 -6 10 -9
0 0.01 2

1
-10 0.005
0

-20 0 -1
0 6 12 18 24 30 36 0 6 12 18 24 30 36 0 6 12 18 24 30 36

5 Conclusion
Financial stability matter remained challenging since the latter nancial crisis

whose one of the main causes has been a credit risk. A main concern raises

questions about how to combine monetary and prudential policies. This paper

moves a step forward by constructing a macro-model that integrates borrowing

and using it to analyze the eect of macroprudential policy and the way mon-

etary and prudential policies could be conducted as a coherent whole. We nd

that consumption appears to be procyclical, while credit to households appears

to be countercyclical. Moreover, consumption and credit to households appear

both much more volatile than output. We also nd that nancial stress index

can eectively monitor nancial stability even in projection, while changes in

the policy interest rate cannot. When nancial stress index curtails credit sup-

ply to households, a contraction ensues, and the policy interest rate is cut. With

this chain of causality, prudential policy is procyclical and monetary policy is

countercyclical. Therefore, in our model, one policy is contractionary and the

other expansionary in order to manage households' credit expansion and deal

with the business cycle. Finally, we do not include credit to businesses as it is

beyond the scope of this work. We think that incorporating this series into our

model could reveal important ndings. However, we leave this for future study.

References

[1] Acemoglu, Daron and Pascual Restrepo, 2017. Robots and jobs: Evidence

from US labor markets.

804
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[2] Agarwal, Sumit, Gene Amromin, Itzhak Ben-David, Souphala Chomsiseng-

phet, Tomasz Piskorski, and Amit Seru, 2017. Policy Intervention in Debt

Renegotiation: Evidence from the Home Aordable Modication Program.

Journal of Political Economy, 125 (3).

[3] Andersen, Asger L., Charlotte Duus, and Thais L. Jensen, 2014. Household

Debt and Consumption During the Financial Crisis: Evidence from Danish

Micro Data. Working Paper.

[4] Bahadir, Berrak and Inci Gumus, 2016. Credit decomposition and business

cycles in emerging market economies. Journal of International Economics,

103, 250-262.

[5] Baron, Matthew and Wei Xiong, 2017. Credit expansion and neglected

crash risk. Quarterly Journal of Economics, pp. 713-64.

[6] Begenau, Juliane, 2020. Capital Requirements, Risk Choice, and Liquidity

Provision in a Business Cycle Model. Journal of Financial Economics 136

(2020) 355378.

[7] Bianchi, Javier, 2011. Overborrowing and Systemic Externalities in the

Business Cycle. American Economic Review, 101 (7), 3400-3426.

[8] Bunn, Philip and May Rostom, 2015. Household Debt and Spending in

the United Kingdom. Bank of England Working Paper.

[9] Caballero, Ricardo J. and Arvind Krishnamurthy, 2001. International and

domestic collateral constraints in a model of emerging market crises. Jour-

nal of Monetary Economics, 48 (3) (2), 513-548.

[10] Canuto, Otaviano, 2011. How Complementary Are Prudential Regulation

and Monetary Policy? Economic Premise. Washington, DC: Poverty Re-

duction and Economic Management (PREM) Network of the World Bank.

[11] Cecchetti, Stephen G., and Marion Kohler, 2012. When capital adequacy

and interest rate policy are substitutes (and when they are not). Bank for

International Settlements (BIS) Working Paper 379.

[12] Collard, Fabrice, Harris Dellas, Behzad Diba, and Olivier Loisel, 2017.

Optimal Monetary and Prudential Policies: Dataset. American Economic

Journal: Macroeconomics. [Link] mac.20140139.

[13] Cúrdia, Vasco, and Michael Woodford, 2016. Credit Frictions and Optimal

Monetary Policy. Journal of Monetary Economics, 84, 30-65.

805
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[14] Davila, Eduardo, 2015. Dissecting Fire Sales Externalities. Working Pa-

per.

[15] Di Maggio, Marco, Amir Kermani, Benjamin J. Keys, Tomasz Piskorski,

Rodney Ramcharan, Amit Seru, and Vincent Yao, 2017. Interest Rate

Pass-Through: Mortgage Rates, Household Consumption, and Voluntary

Deleveraging. American Economic Review.

[16] Dobbie, Will and Paul Goldsmith-Pinkham, 2015. Debtor Protections and

the Great Recession. Working Paper.

[17] Donaldson, Jason Roderick, Giorgia Piacentino, and Anjan Thakor, 2016.

Household Debt and Unemployment. Working Paper.

[18] Dynan, Karen, 2012. Is a Household Debt Overhang Holding Back Con-

sumption? Brookings Papers on Economic Activity, 300-344.

[19] Eggertsson, Gauti B. and Paul Krugman, 2012. Debt, Deleveraging, and

the Liquidity Trap: A Fisher-Minsky-Koo Approach. Quarterly Journal

of Economics, 127 (3), 1469-1513.

[20] Eric, Sims, 2011. Graduate Macro Theory II: Notes on Medium Scale

DSGE Models. University of Notre Dame, Spring 2011.

[21] Fahlenbrach, Rudiger, Robert Prilmeier, and Rene M Stulz, 2017. Why

Does Fast Loan Growth Predict Poor Performance for Banks? Review of

Financial Studies, forthcoming.

[22] Farhi, Emmanuel and Ivan Werning, 2015. A Theory of Macroprudential

Policies in the Presence of Nominal Rigidities. Working Paper.

[23] Fisher, Irving, 1933. The debt-deation theory of great depressions.

Econometrica, pp. 337-357.

[24] Gertler, M., Kiyotaki, N., Prestipino, A., 2016. Anticipated Banking Pan-

ics. Am. Econ. Rev. Pap. Proc. 106 (5), 554559.

[25] Guerrieri, Luca and Iacoviello Matteo, 2017. Collateral Constraints and

Macroeconomic Asymmetries. Journal of Monetary Economics 90, 28-49.

[26] Gupta, Arpit, 2016. Foreclosure contagion and the neighborhood spillover

eects of mortgage defaults.

[27] Hall, Robert E, 2011. The Long Slump. American Economic Review, 101,

431-469.

806
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[28] Hodrick, R. J. and Prescott, E. C., 1997. Postwar U.S. Business Cycles: An

Empirical Investigation. Journal of Money, Credit and Banking, 29(1):1.

[29] Huo, Zhen and Jose-Vctor Ros-Rull, 2016. Financial Frictions, Asset

Prices, and the Great Recession. Working Paper.

[30] Ignazio, Angeloni and Ester Faia, 2013. Capital Regulation and Monetary

Policy With Fragile Banks. Journal of Monetary Economics 60, 311-324.

[31] IMF, 2017. Household Debt and Financial Stability, in Global Financial

Stability Report, October October 2017.

[32] Jorda, Oscar, Moritz Schularick, and Alan M Taylor, 2016. The great

mortgaging: housing nance, crises and business cycles. Economic Policy,

31 (85), 107-152.

[33] Jesus, Fernandez-Villaverde and Juan, F., Rubio, 2006. A Baseline DSGE

Model. Duke University and Federal Reserve Bank of Atlanta, October.

[34] Justiniano, Alejandro, Giorgio E. Primiceri, and Andrea Tambalotti, 2015.

Household Leveraging and Deleveraging. Review of Economic Dynamics

18, 3-20.

[35] Kalman, R. E., 1960. A New Approach to Linear Filtering and Predic-

tion Problems. Transactions of the ASMEJournal of Basic Engineering,

82(Series D), 3545. 7.

[36] Kaplan, Greg, Kurt Mitman, and Giovanni L Violante, 2017. The Hous-

ing Boom and Bust: Model Meets Evidence. Technical Report, National

Bureau of Economic Research.

[37] Kevin L., Kliesen and Douglas C. Smith, 2010. Measuring Financial Mar-

kets Stress. Economic Synopses, Short Essays and Reports On The Eco-

nomic Issues of the Day,(2).

[38] Kiyotaki, Nobuhiro and John Moore, 1997. Credit Cycles. The Journal

of Political Economy, 105 (2), 211-248.

[39] Korinek, Anton and Alp Simsek, 2016. Liquidity trap and excessive lever-

age. The American Economic Review, 106 (3), 699-738.

[40] Krishnamurthy, Arvind and Tyler Muir, 2017. How credit cycles across a

nancial crisis. Working paper.

[41] Lorenzoni, Guido, 2008. Inecient Credit Booms. The Review of Economic

Studies, 75 (3), 809-833.

807
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[42] Macklem, Ti, 2011. Mitigating Systemic Risk and the Role of Central

Banks. Speech, Conférence de Montréal, Montréal, Québec, June 6, 2011.

[Link]

[43] Maih, Junior, 2018. Macroeconomic Modeling of Regime Switching, Lec-

tures Notes. Kobe University, Summer 2018.

[44] Mian, Atif and Amir Su, 2014. What Explains the 2007-2009 Drop in

Employment? Econometrica, 82 (6), 2197-2223.

[45] Mian, Atif R, Amir Su, and Emil Verner, 2017. Credit Supply and Busi-

ness Cycle Amplication: Evidence from Banking Deregulation in the

1980s.

[46] Mian, Atif R, Amir Su, and Emil Verner, 2017. Household debt and

business cycles worldwide. Forthcoming, Quarterly Journal of Economics.

[47] Mian, Atif R and Amir Su, 2018. Finance and Busi-

ness Cycles: The Credit-Driven Household Demand Chan-

nel. Available at SSRN: [Link] or

[Link]

[48] Romer, David. Advanced Macroeconomics. Fifth Edition, McGraw-Hill.

[49] Schmitt-Grohe, Stephanie and Martn Uribe, 2016. Downward Nominal

Wage Rigidity, Currency Pegs, and Involuntary Unemployment. Journal

of Political Economy.

[50] Shleifer, Andrei and Robert W. Vishny, 1992. Liquidation Values and Debt

Capacity: A Market Equilibrium Approach. The Journal of Finance, 47

(4), 1343-1366.

[51] Song, In Hon, 2010. House Prices and Consumption. MPRA Paper,

No.27481.

[52] Su, Amir, 2015. Out of many, one? Household debt, redistribution and

monetary policy during the economic slump. Andrew Crockett Memorial

Lecture, BIS.

[53] Van den Heuvel, Skander J., 2008. The welfare cost of bank capital re-

quirements. Journal of Monetary Economics 55 (2): 298320.

[54] Verner, Emil and Gyozo Gyongyosi, 2017. Household Debt Revaluation

and the Real Economy: Evidence from a Foreign Currency Debt Crisis.

808
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[55] Wolf, Martin, 2012. After the Bonre of the Verities. Financial

Times, May 1st. [Link] content/b2132c34-92aa-11e1-b6e2-

00144feab49a.

[56] Woodford, M., 2003. Interest and Prices: Foundations of a Theory of

Monetary Policy. Princeton University Press, Princeton.

[57] Yagan, Danny, 2017. Employment Hysteresis from the Great Recession.

Technical Report, National Bureau of Economic Research.

[58] Yellen, Janet L., 2010. Macroprudential Supervision and Monetary Pol-

icy in the Post-crisis World. Speech, Annual Meeting of the National

Association for Business Economics, Denver, CO, October 11, 2010.

[Link]

Appendix A.

After we have detrended the variables, the entire set of equilibrium conditions

is given by:

yt = ct + υit + Gt (31)

at ktα [(nt (j))1−υ nit (j)υ ]

yt = (32)
vt

Gt = gt yt , (33)

Pt
πt = (34)
Pt−1

1
πt = [(1 − )πt#1−ω + ] 1−ω (35)

vt = (1 − )π #−ω π ω + π ω vt−1 (36)

yt
rtk = (1 − α) (37)
kt

α(1 − υ)yt /at λt at 1+θ

1
npt = [ ] (38)
dt

αυyt /at λt at
nit = [ ] (39)
dt

809
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

at−1 −1 at+1
dt (cpt /at − γcpt−1 /at−1 ) − γβp Et dt+1 (cpt+1 /at+1 − γcpt /at )−1 = λt at
at at
(40)

λt+1 at+1 at
λt at = βp (1 + Rt )Et { }, (41)
πt+1 at+1

Φdt
+ βp Et [λt+1 at+1 qt+1 /at+1 ] = λt qt at (42)
hpt /at

λt+1 at+1
Qkt = βp Et [(1 − δ)Qkt+1 + rt+1
k
], (43)
λt at

Ψ it /at at it /at at it /at at

1 = Qkt {[1 − ( − 1)2 ] − Ψ( − 1)( )}
2 it−1 /at−1 at−1 it−1 /at−1 at−1 it−1 /at−1 at−1

λt+1 at+1 at k it+1 /at+1 at+1 it+1 /at+1 at+1 2

+βp Et Qt+1 Ψ( − 1)( ) , (44)
λt at at+1 it /at at it /at at

Ψ it /at at
AC = ( − 1)2 (45)
2 it−1 /at−1 at−1

kt+1 = (1 − δ)kt + (1 − AC)it (46)

bt−1
bt = ν + (1 − ν)mqt hit + f sit , (47)
πt

qt
cit = cpt + 1 − (48)
bt

hpt + hit = 1 (49)

πt+1 yt+1 /at+1 at+1

Rt = απ + αy + et , απ > 0, αy ≥ 0 (50)
π yt /at at

ct = (1 − υ)cpt + υcit (51)

Φ̄ 1
qt = (ξ) κ . (52)
1 − βi − (βp − βi )m

810
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Qkt = 1 (53)

log(dt ) = ρd log(dt−1 ) + σd ηtd (54)

log(Φt ) = (1 − ρΦ )log(Φ̄) + ρΦ log(Φt−1 ) + σΦ ηtΦ (55)

log(et ) = ρr log(et−1 ) + σr ηtR (56)

f sit = ρf si f sit−1 + σf si ηtf si (57)

This is a system of 27 equations in 27 variables. The 27 variables are stacked

into the vector xt as follows:

xt = (yt , ct , cit , cpt , dt , Gt , gt , it , at , kt , nit , npt , vt , πt , πt# , Rt , rtk , et , bt , f sit , hpt , hit , Φt , λt , Qkt , ACt , Pt )

Qkt = 1 implies that AC = 0 at the steady state.

Appendix B. Data and sources

This appendix treats about data and sources over 1993Q1-2018Q4.

Real gross domestic product (YGR): Real gross domestic product is

the ination adjusted value of the goods and services produced by labor and

property located in the United States. For more information see the Guide to

the National Income and Product Accounts of the United States (NIPA). For

more information, please visit the Bureau of Economic Analysis. Code GDPC1.

Source: U.S. Bureau of Economic Analysis. Units: Billions of Chained 2012

Dollars, Seasonally Adjusted Annual Rate. Frequency: Quarterly.

Real consumption (CONS): Real Personal Consumption Expenditures,

Billions of Chained 2012 Dollars, Quarterly, Seasonally Adjusted Annual Rate.

Code: PCECC 96. Source: [Link]

Real house prices (HOUSE): All-Transactions House Price Index for the
United States, Index 1980:Q1=100, Quarterly, Not Seasonally Adjusted. Code:

USST HPI. Source: [Link]

Ination (INFL): Consumer Price Index: Total All Items for the United

States, Growth Rate Previous Period, Quarterly, Not Seasonally Adjusted.

Code: CPALTT01USQ657N. Source: [Link]

Households borrowing (BOR): Total Credit to Households and Non-

811
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Prot Institutions Serving Households, Adjusted for Breaks, for United States,

Percentage of GDP, Quarterly, Not Seasonally Adjusted. Code:QUSHAM770A.

Source: [Link]

Leverage (LEV): 30-Year Fixed Rate Mortgage Average in the United

States, Percent, Weekly, Not Seasonally Adjusted. Code: MORTGAGE30US.

Source: [Link] We specify that we converted the weekly

data in quarter for leverage in this work.

Interest rate (INT): Eective Federal Funds Rate, Percent, Monthly, Not
Seasonally Adjusted. Code: FEDFUNDS. Source: [Link]

Similarly, we converted the monthly data in quarterly data in this work.

Financial stress index (FSI): St. Louis Fed Financial Stress Index, Index,

Weekly, Not Seasonally Adjusted. Code: STLFSI. Source: [Link]

812
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Financial inclusion and welfare: New

evidence from the role of government

1 ∗
Toure Talnan Aboulaye , NAKAMURA Tamotsu

June 30, 2020

Abstract
Financial inclusion has become an important public policy priority fol-
lowing the recent global nancial crisis. Global and national-level policy
makers have been embracing nancial inclusion as an important develop-
ment priority. We study whether an increase of nancial inclusion through
greater access to bank accounts aects greater welfare in Japan, India and
Ivory Coast. We use a Ramsey-Cass-Koopmans's model in competitive
economy that includes government. We nd that greater access to bank
accounts positively aects economic growth and welfare in these three
economies. However, the obtained well-being is greater in Japan, India
and Ivory Coast respectively, meaning that more a country's nancial in-
clusion is higher more the welfare is greater. Moreover, our results show
that government should play a key role in promoting nancial inclusion
through subsidy, tax incentive schemes, etc. In addition, evidence from
Japan and India indicates that greater access to banks accounts mitigates
the distortionary eect of tax made person to government on the welfare.
Finally, emerging and developing countries gain much in terms of eco-
nomic growth and welfare by getting their nancial ecosystem to be more
inclusive.
Keywords: nancial inclusion, accessibility, bank accounts, economic growth,

welfare, deterministic dynamic programming methods.

1 PhD student at Kobe University, Graduate School of Economics,

tatoure7@[Link]/171e706e@[Link]. ∗ Supervisor, Economist, Professor of
Economics at Kobe University, Graduate School of Economics, nakamura@[Link].

813
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1 Introduction

Financial inclusion has become an important public policy priority following the

recent global nancial crisis. Global and national-level policy makers have been

embracing nancial inclusion as an important development priority. The G20

made the topic one of its pillars at the 2009 Pittsburgh Summit (G20, 2009).

Thus, the World Bank Group in October 2013 postulated the global goal of

universal access to basic transaction services as an important milestone toward

full nancial inclusion-a world where everyone has access and can use the nan-

cial services, he or she needs to capture opportunities and reduce vulnerability

(World Bank, 2013b).

Indeed, lack of nancial inclusion has in recent years been identied as a

critical engine of economic growth. Sarma (2008), referring to Kempson and

Whiley (1999), distinguishes between ve factors that account for the lack of

nancial inclusion (exclusion): access exclusion due to geography and risk man-

agement of the nancial system, condition exclusion due to conditions that are

inappropriate for some people, price exclusion due to non-aordability of nan-

cial services, marketing exclusion due to the non-attractiveness of conducting

business with certain groups within society (lending risk), and self-exclusion,

due to fear of refusal or due to psychological barriers. These dierent factors

arise from either supply or demand side channels. For example, cultural and

religious factors may undermine demand for banking services. In addition, from

an economic classication perspective, bank accounts are normal, if not luxury

goods. Dealing with banks also requires trust in markets and regulatory over-

sight, which are often not in place, especially not in early stages of development

or transitional periods. Demirguc-Kunt and Klapper (2012) report seven self-

2
reported reasons for lack of nancial inclusion, which rank lack of trust sixth.

Last but not least, a strong informal sector in certain areas may also reduce de-

mand for formal banking services (Sarma and Pais, 2011). Chibba (2009) also

reports from a case study of Botswana a certain deterring fear of complications

attitude, which is likely to be observable in other places as well. Moreover, as far

as the supply side is concerned, the absence of clearly specied land and property

titles is a major obstacle for banks to oer the full range of nancial services.

Similarly, if a region has a small economic base, banking activities will not be
3
oered due to the lack of economies of scale on the side of the banks. Chibba

(2009) explicitly reports market power as a source of lack of nancial inclusion

in the case of Botswana, which is, however, a problem that can be observed in

other countries as well. He moreover laments poor governance in areas such as

monetary policy, land ownership, public sector agency government, government

procurement, and the legal and regulatory framework. Furthermore, transac-

814
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

tion costs are a signicant barrier to the take-up and use of formal nancial

services. Account opening fees and minimum balance requirements prevent the

poor from opening bank accounts (Dupas and Robinson, 2013). Indirect trans-
4
action costs such as travel time are also a barrier: the distance to the nearest

bank or mobile money agent is a key predictor of take-up of savings accounts

(Dupas et al. forthcoming) and mobile money (Jack and Suri, 2014). Therefore,

lack of nancial inclusion is costly to society and the individual.

However, a growing number of evidence suggests that nancial inclusion has

signicant benecial eects for individuals and rms. In addition, more e-

cient payment systems can help individuals by allowing better integration into

modern market economies. Not having to rely on cash, but rather using safer,

less costly and swifter means of transferring payments allows more economic

transactions across greater geographic distances. This can have a direct impact

on income earning opportunities and thus incomes of the poor (Demirgüç-Kunt

et al., 2017; Beck, 2015). In addition, gaining better access to savings and

credit services enables the poor to pull themselves out of poverty by investing

in human capital accumulation, thus reducing rural, aggregate poverty and in-

creasing welfare (Galor and Zeira, 1993; Burgess and Pande, 2005; Suri and

Jack, 2016). Similarly, gaining access to credit services allows the poor to invest

in their micro-enterprises, again gaining broader income earning opportunities

and ultimately higher incomes (Banerjee and Newman, 1993; Bruhn and Love,

2014). Interestingly, a similar argument has been made for access to savings ser-

vices; by protecting resources from intra-family claims through formal savings

accounts, micro-entrepreneurs can invest more in their businesses (Ashraf et al.,

2010; Dupas and Robinson, 2013). Then, access to ecient savings, credit and

insurance services allows the poor to smoothen consumption when hit with in-

come or expenditure shocks (Jappelli and Pagano, 1989; Bacchetta and Gerlach,

1997; Ludvigson, 1999). In turn, increased access to nancial services increases

household savings (Aportela, 1999; Brune et al., 2016; Somville and Vandewalle,

2016; Allen et al., 2016), employment (Prasad, 2010; Bruhn and Love, 2014),

income (Bruhn and Love, 2014), improves mental well-being (Karlan and Zin-

man, 2010; Angelucci et al., 2013), favours education (Flug et al., 1998), helps

making better decision (Mani et al., 2013), enhances new rm creation (Guiso et

al., 2004; Klapper et al., 2006; Banerjee et al., 2013), and ensures bank stability

(Ahamed and Mallick, 2019). Hence, better nancial inclusion can have welfare

eects that extend beyond benets in the nancial realm to the real economy.

Therefore, we ask about the following question : Does an increase of nancial

inclusion aect greater welfare? The goal of this work is to study the role of

government in the link between nancial inclusion and welfare.

At the core of our analysis is a Ramsey (1928), Cass (1965), and Koop-

815
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

mans (1965) in a competitive economy to include government. One reason for

choosing this structure is that the model avoids all market imperfections and

all issues raised by heterogeneous households and links among generations. An-

other main reason is technical namely under conditions where certain structures

of the economy (such as utility functions) don't change over time, we could use

dynamic programming methods to solve for stationary states. To this frame-

work, we add two key elements. First, nancial inclusion through access to

bank accounts increases individual's savings and aects capital accumulation.

Kempson and Whyley (1999) investigate the type of people that are excluded

from formal nance systems in Britain and point out that a necessary nan-

cial service for low-income people is a basic bank account. Therefore, owning a

bank account turns out to be a priority in access to credit, savings and insur-

ance opportunities. Second, government plays a key role in promoting nancial

inclusion either through subsidy or tax incentive schemes. We use the Newton-

Raphson's method to solve the nonlinear dynamic of the system which is based

on the Fair-Taylor (1983) algorithm. The numerical problem consists of solving

a nonlinear system of simultaneous equations in n endogenous variables in T

periods. Our model has a property of handling endogenous savings by allowing

the agents to live forever and making plans taking into account their future

consumption stream. The rest of the paper is as follows: Section 2 describes

the social planner's economy. Section 3 indicates the Ramsey-Cass-Koopmans

model in competitive economy. Section 4 shows equilibrium conditions, solution

of the model and calibration. Section 5 treats the results. Section 6 presents

conclusion and policy implication.

2
3
4

2 Demirguc-Kunt and Klapper (2012) report seven self-reported reasons for lack of nancial
inclusion, which rank lack of trust sixth. In descending order of importance the other reasons
were: (1) Not enough money, (2) Too expensive, (3) Family member already has account, (4)
Too far away, (5) Lack of necessary documentation, (7) Religious Reasons.
3 Operational costs may be a consideration when banks need to decide whether or not
to serve remote or sparsely populated areas (Claessens and Perotti, 2007; Andrianaivo and
Kpodar, 2011). Setting up a bank is often associated with high sunk and xed costs, which
can serve as a market entry barrier and lend rst movers monopolistic market power that
prevents nancial inclusion of certain strata of the society.
4 The reduction in indirect transaction costs could also improve nancial inclusion: Bachas
et al. (2018) observe that use of the account and savings are correlated with the change
in travel distance to access the account. These eects could be particularly important in the
context of a developing country with low nancial inclusion: 61 percent of the adult population
in Mexico has no bank or mobile money account, and 74 percent of those in the poorest fth of
the population have no account (Demirgüç-Kunt et al. 2015). While the relationship between
nancial activity and the change in travel distance is correlational, Bachas et al. (2018) exploit
the staggered rollout of cards over time to causally estimate how debit cards aect account
use and savings.

816
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2 The social planner's economy

Consider an economy with only one individual. At time t, an innitely lived

individual wants to maximizes a lifetime utility function of the form:

∞
X
β i u(ct+i ), (1)
i=0

where u(•) is a subutility function that does not change with time, 0 < β < 1
is the factor by which future utility is discounted, ct is the consumption of the
individual in period t. The function u(•) is assumed to be increasing, continuous,

concave, and with as many derivatives as are required.

The individual begins period t with a given amount of capital stock carried
over from the previous period, kt , will provide one unit of labor in every period,
and decides how much to consume subject to the sequence of single period

budget constraints of the form:

kt+1 = (1 − δ)kt + it , (2)

and

yt = f (kt ) ≥ ct + it . (3)

The capital stock, kt , is per worker capital by denition in a one -person

economy and by the assumption that individual provides one unit of labor in

each period. Investment, it , is the only other use for production that is not

consumed. δ is the depreciation rate.

2.1 Deterministic dynamic programming methods: prin-

ciple of optimality

An optimal policy has the property that whatever the initial state and initial

decision are, the remaining decisions must constitute an optimal policy with

regard to the state resulting from the rst decision. Therefore, the dynamic

programming method breaks this decision problem into smaller sub-problems.

Using the two budget constraints at equality, we can write consumption at

time t as:

kt+1 = f (kt ) + (1 − δ)kt − ct . (4)

In discrete time, the social planner maximization problem is as follows:

∞
X
max β i u(ct+i ), (5)
i=0

817
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

subject to

kt+i+1 = f (kt+i ) + (1 − δ)kt+i − ct+i . (6)

Setting it otherwise, we get:

∞
X
max β s−t u(cs ), (7)
s=0

subject to

ks+1 = f (ks ) + (1 − δ)ks − cs . (8)

Suppose that we are at time t on the optimal path. Since maximized sum of
discounted utilities depends solely on state at time t, let denote it as V (kt ),which

is called value function:

∞
X
V (kt ) = {max β s−t (u(f (ks ) + (1 − δ)ks − ks+1 )}. (9)
s=t

Denoting it as

∞
X
V (kt ) ≡ β s−t u(c∗s ),
s=t

V (kt+1 ) becomes:

∞
X
V (kt+1 ) = β s−(t+1) u(c∗s ),
s=t

and

∞
X
V (kt ) = u(c∗t ) + β β s−(t+1) u(c∗s ),
s=t

V (kt ) = u(c∗t ) + βJ(kt+1 ). (10)

Eq.10 is called Bellman Equation or Recursive equation.c∗t satises two-period

optimality condition, and hence optimality condition for entire program. This

is principle of optimality.

substituting ks+1 = f (ks ) + (1 − δ)ks − cs into the Bellman equation and

deriving rst order conditions:

V 0 (ks ) = u0 (cs ) − βJ 0 (ks+1 ) = 0,

Then

u0 (cs ) = βV 0 (ks+1 ). (11)

818
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Applying the Envelope theorem to Bellman equation (Eq.10), yields:

V 0 (ks ) = βV 0 (ks+1 )((1 − δ) + f 0 (ks )),

and

V 0 (ks+1 ) = βV 0 (ks+2 )((1 − δ) + f 0 (ks+1 )).

Eq.11gives:
u0 (cs+1 ) = βV 0 (ks+2 ),

Thus,

β −1 u0 (cs ) = u0 (cs+1 )((1 − δ) + f 0 (ks+1 )).

Therefore,
u0 (cs )
f 0 (ks+1 ) + (1 − δ) = . (12)
βu0 (cs+1 )
Eq.12 is the so-called Euler equation for consumption. At the stationary

state, ks+1 = ks = ks−1 = k̄ , hence the rst-order condition yields:

1
f 0 (k̄) = − 1 + δ. (13)
β
Since the parameters β and δ are known, as is the function f (•), this equation
13 can be solved for k̄ . Therefore, using the dynamic programming methods,

we nd that in a stationary state, the marginal product of capital is equal to

the net real interest rate implicit in the discount factor plus depreciation. Once

the stationary state capita stock is determined, the budget constraints can be

used to determine the stationary state values of output and consumption.

The solution to Eq.9 gives the values of the control variables as a function

of the time t state variables as follows:

kt+1 = H(kt ). (14)

5
Where the function H(kt ) is called the policy function.
5

3 The Ramsey-Cass-Koopmans model in compet-

itive economy

This section treats the long-run Ramsey-Cass-Koopmans model in a competitive

economy in which there are consumers who provide labor to the market and

5 The policy function describes how the controls behave as a function of the current state
variables, kt .

819
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

rms who hire this labor at the competitive wage. We assume that there is

a continuum of identical agents of measure 1. All individuals are the same,

therefore, we consider the representative household's problem.

3.1 Firms

Firms are owned by households. In addition, the representative rm produces

by using the Cobb-Douglas production function technology as follows:

yt = At (IF Ikt−1 )θ . (15)

Where yt , At , θ, IF I and kt are respectively per-capita output, total factor

productivity, capital share in production, index of nancial inclusion and per-

capita capital standing for both physical and banking service infrastructure

as far as building a banking service infrastructure is capital, too. Further,

we assume that nancial inclusion through access to bank accounts increases

individuals' savings, hence, investment. Therefore, greater access to nance

could positively aect savings (Aportela, 1999; Brune et al., 2016; Somville

and Vandewalle, 2016; Allen et al., 2016), investment (Ashraf, Karlan and Yin,

2010; Dupas and Robinson, 2013), enhance new rm creation (Guiso et al.,

2004; Klapper et al., 2006; Banerjee et al., 2013) and capital accumulation.

As nancial inclusion increases the amount of funds being made available and

reduces borrowing costs, capital should increase (Claessens and Perotti, 2007;

D-W. Kim et al., 2018).

Firms also hire capital from households at rate rt . As a result of prot

maximization, we get:

rt = θAt IF I(IF Ikt−1 )θ−1 , (16)

πt = yt − rt kt−1 = (1 − θ)yt . (17)

Where πt is the prot in period t of rm at the aggregate level.

3.2 Government

Policy makers increasingly recognize that nancial exclusion is a risk to political,

social, and even nancial stability, which may impede economic advancement,

and that nancial inclusion presents an opportunity to improve lives. Moreover,

account opening fees and minimum balance requirements prevent the poor from

opening bank accounts (Dupas and Robinson, 2013). Hence, government should

play a key role in nancial inclusion. Therefore, we assume that government

820
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

pays a cost to support nancial inclusion (greater acccess to bank accounts) by

determining the routing and volume of payments either through tax incentive

schemes or small subsidies that can lead to large increases in take-up (Cole,

Sampson, and Zia, 2011). In turn, households pay taxes to government whose

by assumption they (taxes) distort capital income. This process favorably shape

the nancial ecosystem to be more inclusive. As a result, the government's

budget constraint is given by:

τ rt kt−1 = Tt . (18)

Where τ is corporate tax rate and Tt stands for lump-sum transfert in period
t.

3.3 Households

An individual i ∈ [0, 1] maximizes:

∞
X
β t logcit , (19)
t=0

Subjet to

cit + it + τ rt kt−1
i
= rt kti + πti + Tt . (20)

where ct is per-capita consumption, β is the discount factor. In addition,

capital and investment are linked by the following capital accumulation equa-

tion:

kti = (1 − δ)kt−1
i
+ it , (21)

whereit is the investment and δ is the depreciation rate. Note that the house-
hold takes both πti andTt as given. First order conditions (F.O.C) are shown
in appendixA., while appendixB. indicates the aggregation rules.

4 Equilibrium conditions, solution of the model

and calibration

4.1 Equilibrium dynamics

After computing the rst-order-conditions of the lagrangian function (see appendixA.),

the equilibrium conditions at the aggregate level are set as follows:

yt = At (IF Ikt−1 )θ , (22)

821
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ct
1=β (1 + (1 − τ )θAt+1 IF I θ ktθ−1 − δ), (23)
ct+1
and

ct + kt − (1 − δ)kt−1 − yt = 0, (24)

The following is how equations are interpreted: Eq.22 is per-capita-output,

Eq.23 represents Euler equation for consumption, and Eq.24 is the resource

constraints.

There are three variables {ct , kt , yt }and three equations, so we can solve the

model.

4.2 The long-run solution of the model

The conditions for the stationary state are:

From Eq.22 to Eq.24 respectively, we get

1
− (1 − δ) = (1 − τ )θĀIF I θ k̄ θ−1 ,
β

and

δ k̄ = ĀIF I θ k̄ θ − c.

From the rst of these two conditions, we get the stationary state of capital,

¯
(1 − τ )θAβIF I θ 1−θ
1
k̄ = ( ) , (25)
1 − β(1 − δ)
The second gives the stationary state of consumption,

c̄ = ĀIF I θ k̄ θ − δ k̄. (26)

4.3 Calibration

We choose the discount factor (β) to 0.980 which is consistent with estimations of
the discount factor for models with annual data. We set θ = 0.36 and δ = 0.10 as
capital share and depreciation rate respectively. Following trading economics,

the corporate tax rate is respectively equal to τ = 30.62%,τ = 25.17% and

τ = 25% in Japan, India and Ivory Coast. Concerning nancial inclusion,

Global Findex Database of World bank Demirguc-Kunt et al. (2018) provides

the latest version of annual data on nancial inclusion measures. Thus, in our

work, we choose the index of accessibility to bank accounts (account (% Age

15+)) namely IF I = 98% in Japan, IF I = 80% in India and IF I = 41% in

822
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1: Calibrated parameters values

Calibrated parameters names Japan India Ivory Coast

Account (% Age 15+) (IF I ) 0.98 0.80 0.41

Discount factor (β ) 0.980 0.980 0.980
Capital share (θ ) 0.36 0.36 0.36
Depreciation rate (δ ) 0.10 0.10 0.10
Corporate tax rate (τ ) 0.3062 0.2517 0.25
N ote: The table summarizes calibrated parameters for the full model.

Ivory Coast. Moreover, our calibration covers three countries sample for our

policy simulation mainly: Ivory Coast, India and Japan respectively a lower

middle, upper middle and high income country. By choosing these three groups

of countries, we would like to study the degree of nancial inclusion and its eect

on economic growth and social welfare in each case of country. The following

table describes the parameters of the model.

5 Results

This section indicates the results of the model in the long-run namely the steady

state equilibrium, approximating value function and policy function.

5.1 The steady state equilibrium

The following results indicate the eects of nancial inclusion on economic

growth and so the social welfare of Ramsey-Cass-Koopmans's model in a com-

petitive economy with government in Japan, India and Ivory Coast's economy.

Indeed, gure 1 shows how consumption (c) and capital (k) evolve over time to

satisfy the representative household's intertemporal optimization condition in

the economy given by the equilibrium conditions (Eq.23 and Eq.24). The equa-

tion (Eq.24) relates the change in k to output and consumption given initial

values of c and k. Therefore, the gures (f ig.1) describe the phase diagram of

each economy, so our results show that each economy is at its steady state. Im-

mediately after the policy shock of nancial inclusion on capital accumulation,

the equilibrium requires that each of the three economies is on its saddle path

indicated by the star blue line. Hence, after the policy shock there is a jump of

economy from the initial point in red star to the equilibrium represented by the

intersection of the two blue lines.

In addition, our results show that nancial inclusion has signicant economic

outcomes in terms of promoting consumption and capital accumulation for the

823
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

three economies. Therefore, in their turn investment and output should increase

as well. Furthermore, comparing the equilibrium level from a country to another

we notice that the equilibrium obtained by including nancial inclusion in our

model is respectively greater in Japan, India and Ivory Coast's economy. Thus,

greater access to bank accounts provides the greatest welfare in Japan's economy

despite its high tax rate (table.1) followed by India and Ivory Coast. This result

shows that the high level of nacial inclusion crowds out the tax's distortion

eect in Japan and India economies. Moreover, the capital at the steady state

almost identical in Japan and India's economies is higher in both countries

relative to that in Ivory Coast's economy. This implies that the capital stock in

India's economy could faster grow as Japan's economy whether there is a greater

access to bank accounts and that policy makers encourage savings through tax

incentive schemes in India's economy. Intuitively, a permanent increase in taxes

reduces households' lifetime wealth and welfare.

In short, our ndings highlight three dierent scenarios of the eect of -

nancial inclusion on economic growth and welfare namely that in developed,

emerging and developing countries. First, results from Japan show that nan-

cial inclusion promote greater welfare in high nancial inclusion economy such

as high income countries. For example, individual's access to formal accounts

(% age 15+) is about 98% in Japan. Second, the case of India indicates that

nancial inclusion also has signicant benecial eects on well-being and ef-

forts made by policy makers of providing access to bank accounts avail more to

businesses. Furthermore, the emergence of the middle class and the nancial
6
literacy's improvement allow people a large access and use of formal nancial

services to undertake and reduce vulnerability in emerging countries such as

India. For example, the eect of providing savings accounts seems to be robust

as people shift away from storing money at home or holding it in the form of

livestock or jewelry (Demirguc-Kunt et al., 2017). Third, the results from Ivory

Coast economy as developing country highlight that the welfare obtained of ac-

cessing to bank accounts is lower despite its low tax rate compared to that in

Japan and India. This means that the rate of accessibility and use still remains

low due to a lagre informal sector, population illiteracy, an embryonic private

sector and poor institutions etc., in these countries. Therefore, nancial inclu-

sion should have a positive marginal impact on welfare in developing countries

for now, hence, it should be improved.

Our ndings comply with a growing body of evidence suggesting that provid-

ing access to bank accounts increases take-up rates of these accounts, household

savings (Aportela, 1999; Brune et al., 2016; Somville and Vandewalle, 2016;

Allen et al., 2016), employment (Prasad, 2010; Bruhn and Love, 2014), income

(Bruhn and Love, 2014), private and business expenditures (Ashraf, Karlan and

824
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1: The capital-consumption space in the competitive economy.

2 2

1.8 1.8

1.6 1.6

1.4 1.4

1.2 1.2

1 1

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10

1.8

1.6

1.4

1.2

0.8

0.6

0.4

0.2

0
0 1 2 3 4 5 6 7 8 9 10

N ote: We set the initial value of k equal to 0.1, which is determined by the
assumption that the initial capital stock is an exogenous parameter of the model
indicated by the red star. The star blue line represents the saddle path from the
initial value of capital towards the equilibrium determined by the intersection of
capital locus and consumption locus both in blue line. While the blue vertical
line is the consumption locus, the red line pins down the consumption at the
stationary state.

825
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Yin, 2010; Dupas and Robinson, 2013), improves mental well-being (Karlan and

Zinman, 2010; Angelucci et al., 2013), favours education (Flug et al., 1998),

helps making better decision (Mani et al., 2013), enhances new rm creation

(Guiso et al., 2004; Klapper et al., 2006; Banerjee et al., 2013), reduces rural

poverty and improves welfare (Burgess and Pande, 2005; Beck et al., 2007a).

Likewise, some studies nd that the impact of nancial inclusion on growth de-

pends on rms'access to credit rather than households (Beck, Buyukkarabacak,

Rioja and Valev, 2012), most notably by reducing the nancing gap faced by
7
small-and medium sized rms and industries (GPFI , 2011). Financial inclu-

sion reduces liquidity constraints and encourages investment. The distribution

of credit across rms at the sectoral level therefore has important eects on

the industrial structure, competition, or the degree of informality in the sector,

particularly in low-income countries (Beck, Demirguc-Kunt and Maksimovic,

2005). Moreover, some specic weaknesses of developing countries, such as poor

institutions (Demetriades and Hook Law, 2006), insucient nancial competi-

tion due to political deadlock (Rajan and Zingales, 2003), and high ination

(Rousseau and Wachtel, 2002), have been highlighted in the literature as damp-

ening or suppressing the nance-growth relationship.

6
7

5.2 Value function analysis

The following graphs in gure 2 indicate the optimal payo that may provide the
policy shock of accessing to bank accounts over time. Therefore, gure 2 shows
how the value function as in equation (Eq.9) is converging after 240 iterations

in three economies namely Japan, India and Ivory Coast. The green line is

the highest line calculated. It should be clear from the graphs that the steps

are gradually getting smaller as the number of iterations increases. Moreover,

gure 2 shows that while the line moves upward, remains in positive region for

Japan and India economy, it moves downward, remains in negative region for

that of Ivory Coast. Therefore, the optimal payo is greater for Japan, followed

respectively by India and Ivory Coast.

Furthermore, the results from gure 2 highlight that greater access to bank

6 Grohmann, Kluhs and Menkho (2018) show that while nancial inclusion is typically
addressed by improving the nancial infrastructure, a higher degree of nancial literacy also
has a clear benecial eect. They nd that regarding access to nance, nancial infrastruc-
ture and nancial literacy are mainly substitutes. However, regarding the use of nancial
services, the eect of higher nancial literacy strengthens the eect of more nancial depth.
Moreover, the positive impact of nancial literacy holds across income levels and several sub-
groups within countries.
7 Households' access to credit may also increase economic growth, particularly if it nances
durable goods rather than consumption (GPFI, 2013).

826
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 2: Approximating the value function.

Japan India
14 8

12 6

10
4
Value function

Value function

8
2
6
0
4
numits=48 numits=48
-2
2 numits=96 numits=96
numits=144 numits=144
0 numits=192 -4 numits=192
numits=240 numits=240
-2 -6
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Capital at time t Capital at time t
Ivory Coast
-2

-4

-6

-8
Value function

-10

-12

-14

-16
numits=48
numits=96
-18 numits=144
numits=192
-20
numits=240
-22
0 1 2 3 4 5 6
Capital at time t

N ote : numits indicates the number of iterations. Each of the lines shown in the
gure represents the results of 48 iterations, so there are a total of 5 lines shown
that come from the calculations. The value function measures the welfare over
time of the representative household's intertemporal maximization given the
state variables, for example the capital in our work. Welfare measured through
the value function is in % consumption equivalent units.

827
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

accounts has signicant benecial eects for individuals and rms, and then both
8
social and economic benets. In particular, geater access to nance increase

savings (Aportela, 1999; Ashraf et al., 2010a), female empowerment (Ashraf et

al., 2010b), productive investment (Dupas and Robinson, 2013), consumption

(Dupas and Robinson, 2013; Ashraf et al., 2010b). Consequently, ndings in

gure 2 conrm those in gure1above and prove that nancial inclusion provides
greater welfare in higher income (Japan), followed by emerging (India) and

developing (Ivory Coast) economy respectively. Thus, we may conclude that the

positive impact of nancial inclusion holds across income levels within countries.

Allen et al. (2012) study the relationship between nancial inclusion and GDP

per capita and nd that nancial inclusion increases as GDP per capita increases

at the country-level. Meaning that more a country is richer more there is a great

access and use of formal nancial and digital services to facilitate payments and

nancial transactions between economic agents.

5.3 Policy function

Figure 3 respectively gives the policy function for Japan, India and Ivory Coast's
economy after the policy shock of nancial inclusion. In addition, the policy

function for each economy, which nds the optimizing value of consumption (c)
(control variable) for each value of time t capital (k) (state variable), set by

equation (Eq.14) is generated at the same time as the value functions above

(f ig.2). Therefore, the one for our sample economies, after 240 iterations, is

shown in red line in gure 3, while the blue line stands for the 45 degree line.

We nd moreover that the policy function crosses the 45 degree line at the
stationary state value of about k = 5.60 for Japan, k = 4.9 for India and
k = 3.52 for Ivory Coast. So, Japan's economy has the highest capital at the

stationary state, followed by that of india and then Ivory Coast. Our ndings

therefore suggest that greater access to bank accounts has a positive impact on
9
capital accumulation, capital stability , hence, economic growth (D-W. Kim et

al., 2018). Moreover, the positive impact on capital accumulation holds across

income levels within countries depending on whether the country is high income,

emerging or developing.
9

8 See, for example: Wurgler (2000), Beck et al. (2000), Klapper et al. (2006), Demirgüc-
Kunt et al. (2008), Demirgüc-Kunt et al. (2015), Burgess and Pande (2005), and Banerjee
et al. (2013) in India; Bruhn and Love (2009) and Bruhn and Love (2014) in Mexico; Karlan
and Zinman (2010) in South Africa; Dupas and Robinson (2009) in Kenya.
9 D-W. Kim et al. (2018) study the relationship between nancial inclusion and economic
growth in Organization of Islamic Cooperation (OIC) countries. Based on the results of
dynamic panel estimations, they nd that nancial inclusion has a positive eect on economic
growth. The IFRs results derived from the panel VAR analysis suggest that nancial inclusion

828
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 3: The policy function after 240 iterations.

Japan India
6 6
c = H(k) c = H(k)
45 degree line 45 degree line
5 5

4 4
Consumption (c)

Consumption (c)

3 3

2 2

1 1

0 0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
Capital (k) Capital (k)
Ivory Coast
6
c = H(k)
45 degree line
5

4
Consumption (c)

0
0 1 2 3 4 5 6
Capital (k)

N ote :The policy function describes how the controls behave as a function of
the current state variables. In our model, the individuals pay a tax to the
government through a percentage tax that aects capital and receive a lump
sum transfer from the government for exactly the same amount as the tax they
paid. These two amounts cancel out in the budget constraint (Eq.20) in the
computation of the policy function.

829
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

6 Conclusion

The term nancial inclusion has received much attention since the late 1990s,

as the policy-making issue of socially excluded people and research studies about

the nancial exclusion of socially excluded people have emerged (Leyshon and

Thrift, 1993, 1994, 1995; Collard, 2010). Later, Kempson and Whyley (1999)

investigate the type of people that are excluded from formal nance systems in

Britain and point out that a necessary nancial service for low-income people

is a basic bank account. In this work, we examine whether nancial inclusion

brings about greater welfare by using Ramsey-Cass-Koopmans's model in a

competitive economy with government. Moreover, we focus on three countries's

economies namely Japan, India and Ivory Coast as we would like to analyze

the well-being level that provides greater access to bank accounts in high in-

come, emerging and developing country. Consequently, we nd that the ease

of accessibility and availability of the formal nancial services, such as bank

deposit, credits, savings, insurance, etc., for all participants in an economy has

both social and economic benets. It allows businesses to make investment so

that capital increases on the one hand, households to smooth their consumption

along their lifetime, properly spend in health, education, increases their income

and reduces poverty. We also nd that benecial eects to greater access to

banks accounts is more important in Japan, India and Ivory Coast respectively.

Therefore, the positive impact of nancial inclusion holds across income lev-

els within countries depending on whether the accessibility and use are either

higher, relatively high or lower at the country level. Moreover, the introduction

of the government in the model through the tax's implementation highlights the

important role that it plays in promoting nancial inclusion. Government sup-

ports nancial inclusion in at least three keys areas. First, government sets the

rules that regulate the environment for nancial inclusion, balancing the drive

to bring nancial services to poor households with measures to protect con-

sumers as well as the stability and integrity of the nancial system. Second, it

can promote infrastructure, either nancially or by incentivizing private sector

investments, to support the expansion of nancial services. Such infrastruc-

ture might include identity systems, connectivity, payment systems or credit

registries. Third, government can support nancial inclusion by determining

the routing and volume of payments made government-to-person (e.g., social

payments) and person-to-government (e.g. taxes) that favorably shape the -

has positive eects on the economic growth and nancial inclusion and economic growth have
mutual causalities with each other based on the panel Granger causality tests. Therefore, it
seems reasonable to conclude that nancial inclusion has positive eect on the economic growth
in OIC countries. Put another way, a high level of inclusivity in a nancial society means
that the circumstances in which most of an economy's participants are using formal nancial
systems allow them to benet from nancial services and pursue their capital stability.

830
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

nacial ecosystem to be more inclusive. Thus, from a policy perspective, at all

levels of nancial depth, government's support such as tax incentive schemes is

useful for improving nancial inclusion. Since in Japan groups that are consid-

ered nancially vulnerable include women, children and youth, the elderly and

non-Japanese residents, tax incentive schemes should increase their income to

sustain living. In both emerging (India) and developing countries (Ivory Coast)

tax incentive schemes and greater access of women to formal nance should

improve women empowerment, increase income, savings and consumption. Fur-

ther, in developing countries a lot of people undertake in the informal sector,

therefore, policy makers should organize the sector and implement long term

eective nancing policies to shift them out of informal in order to promote

greater access to bank accounts.

References

[1] Ahamed, M., M., & Mallick, S., K., 2019. Is Financial Inclusion Good for

Bank Stability? International Evidence. Journal of Economic Behavior &

Organization, 157 (2019), 403427.

[2] Allen, F., Demirguc-Kunt, A., Klapper, L., & Peria, M.S.M., 2012. The

foundations of nancial inclusion. World Bank Policy Research Working

Paper Servey (6290).

[3] Allen, F., Demirguc-Kunt, A., Klapper, L., & Peria, M.S.M., 2016. The

Foundations of Financial Inclusion: Understanding Ownership and Use of

Formal Accounts. Journal of Financial Intermediation, 27, 130.

[4] Andrianaivo, M., & Kpodar, K., 2011. ICT, nancial Inclusion, and

Growth: Evidence from African countries. International Monetary Fund,

Working Paper, 11/73.

[5] Angelucci, M., Karlan, D., & Zinman, J., 2013. Win Some Lose Some?

Evidence from a Randomized Microcredit Program Placement Experiment

by Compartamos Banco. National Bureau of Economic Research.

[6] Aportela, F., 1999. Eect of Financial Access on Savings by Low-Income

People.

[7] Ashraf, N., Karlan, D., & Yin, W., 2010. Female Empowerment: Further

Evidence from a Commitment Savings Product in the Philippines. World

Development, 38 (3), 333344.

831
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[8] Bacchetta, P. & Gerlach, S., 1997. Consumption and Credit Constraints:

International Evidence. Journal of Monetary Economics, 40, 207-238.

[Link]

[9] Bachas, P., Gertler, P., Higgins, S., & Seira., E., 2018. How Debit Cards

Enable the Poor to Save More. National Bureau of Economic Research

Working Paper 23252.

[10] Banerjee, A., V., & Newman, A., F., 1993. Occupational Choice and the

Process of Development. Journal of Political Economy, Vol. 101, No. 2

(Apr., 1993), pp. 274-298.

[11] Banerjee, A.V., Duo, E., Glennerster, R., & Kinnan, C., 2013. The Mir-

acle of Micronance? Evidence from a Randomized Evaluation. MIT Bu-

reau for Researchand Economic Analysis of Development Working Paper.

[12] Beck, T., Levine, R., & Loayza, N., 2000. Finance and the sources of

growth. Journal of Financial Economics, 58, 261300.

[13] Beck, T., Demirgüc-Kunt, A., & Maksimovic, V., 2005. Financial and

legal constraints to growth: does rm size matter? Journal of Finance,

60, 137177.

[14] Beck, T., Demirguc-Kunt, A., & Peria, M. S. M., 2007. Reaching out:

Access to and use of banking services across countries. Journal of Financial

Economics, 85(1), 234266.

[15] Beck, T., 2015. Micronance: A Critical Literature Survey. Working Pa-

per No. 2015/4. World Bank Group.

[16] Beck, T., Buyukkarabacak, B., Rioja, F. K., Valev, N. T., 2012. Who

gets the credit? And does it matter? Household vs rm lending across

countries. The BE Journal of Macroeconomics, 12(1), 146.

[17] Bruhn, M., & Love, I., 2009. The Economic Impact of Banking the Un-

banked: Evidence from Mexico. World Bank Policy Research Working

Paper No. 4981.

[18] Bruhn, M., & Love, I., 2014. The real impact of improved access to nance:

Evidence from Mexico. Journal of Finance, 69(3), 13471369.

[19] Brune, L., Giné, X., Goldberg, J., & Yang, D., 2016. Facilitating sav-

ings for agriculture: Field experimental evidence from Malawi. Economic

Development and Cultural Change, 64(2), 187220.

832
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[20] Burgess, R., & Pande, R., 2005. Do rural banks matter? Evidence from

the indian social banking experiment. American Economic Review, 95(3),

780795.

[21] Cass, D., 1965. Optimum Growth in an Aggregative Model of Capital

Accumulation. Review of Economic Studies, 32 (July): 233-240.

[22] Chauvet, L. & Jacolin, L., 2017. Financial Inclusion, Bank Concentration,

and Firm Performance. World Development Vol. 97, pp. 113, 2017.

[23] Chibba, M., 2009. Financial Inclusion, Poverty Reduction and the Mil-

lennium Development Goals. European Journal of Development Research,

21(2): 213-230 ·

[24] Claessens, S., & Perotti, E., 2007. Finance and Inequality: Channels and

Evidence. Journal of Comparative Economics, 2007, vol. 35, issue 4, 748-

773.

[25] Cole, S., Sampson, T., & Zia, B., 2011. Price or knowledge? What drives

demand for nancial services in emerging markets? Journal of Finance,

66(6), 19331967.

[26] Collard, S., 2010. Toward nancial inclusion in the UK: progress and chal-

lenges. Publ. Money Manag. 27 (1), 1320.

[27] Demetriades, P., & Hook Law, S., 2006. Finance, institutions and economic

development. International Journal of Finance & Economics, 11(3), 245

260.

[28] Demirgüc-Kunt, A., Beck, T., & Honohan, P., 2008. Finance for All?:

Policies and Pitfalls in Expanding Access. World Bank Publications, The

World Bank,Washington, DC.

[29] Demirgüc-Kunt, A., & Klapper, L., 2012. Measuring Financial Inclusion:

The Global Findex Database. World Bank Policy Research Paper, No.

6025.

[30] Demirgüc-Kunt, A., Klapper, L., Singer, D., & Van Oudheusden, P., 2015.

The Global Findex Database 2014: Measuring Financial Inclusion Around

the World. World Bank Policy Research Working Paper, No. 7255.

[31] Demirguc-Kunt, A., Klapper, L., & Singer, D., 2017. Financial inclusion

and inclusive growth A review of recent empirical evidence. World Bank

Policy Research Paper, No. 8040.

833
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[32] Demirgüç-Kunt, A., Klapper, L., Singer, D., Ansar, S., & Hess, J., 2018.

The global ndex database 2017: Measuring nancial inclusion and the

ntech revolution. Washington, DC: World Bank.

[33] Dupas, P., & Robinson, J., 2013. Savings Constraints and Microenterprise

Development: Evidence from a Field Experiment in Kenya. American

Economic Journal: Applied Economics, American Economic Association

5(1): 163-92.

[34] Dupas, P., Karlan, D., Robinson, J., & Diego, U., 2018. Banking the

Unbanked? Evidence from Three Countries. American Economic Journal:

Applied Economics, 10 (2): 257-97.

[35] Flug, K., Spilimbergo, A., & Wachtenheim, E., 1998. Investment in edu-

cation: do economic volatility and credit constraints matter? Journal of

Development Economic, 55, 465481.

[36] G20 Information Centre., 2009. G20 Leaders Statement: The Pittsburgh

Summit. [Link]

[37] Galor, O., & Zeira, J., 1993. Income Distribution and Macroeconomics.

The Review of Economic Studies, Vol. 60, Issue 1 (Jan., 1993), 35-52.

[38] George, McCandless, 2008. The ABCs of RBCs: An Introduction to

Dynamic Macroeconomic Models. Harvard University Press, Cambridge,

Massachusetts, and London, England.

[39] Guiso, L., Sapienza, P., & Zingales, L., 2004. Does local nancial develop-

ment matter? The Quarterly Journal of Economics, 119, 929969.

[40] Grohmann, A., Klühs, T., & Lukas Menkho, L., 2018. Does Financial

Literacy Improve Financial Inclusion? Cross Country Evidence. World

Development 111 (2018) 8496.

[41] GPFI (Global Partnership for Financial Inclusion), 2011. G20 Princi-

ples for Innovative Financial Inclusion. Alliance for Financial Inclusion

Bangkok Thailand.

[42] GPFI, 2013. The 20 Basic Set of Financial Inclusion Indicators. Global

Partnership for Financial Inclusion April.

[43] Fair, R., & Taylor, J., 1983. Solution and Maximum Likelihood Estimation

of Dynamic Nonlinear Rational Expectations Models. Econometrica, 1983,

vol. 51, issue 4, 1169-85.

834
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[44] Jack, W., & Suri, T., 2014. Risk Sharing and Transactions Costs: Evidence

from Kenya's Mobile Money Revolution. American Economic Review 104

(1): 183223.

[45] Jappelli, T., & Pagano, M., 1989. Consumption and Capital Market Im-

perfections: An International Comparison. American Economic Review,

1989, vol. 79, issue 5, 1088-1105.

[46] Karlan, D., & Zinman, J., 2010. Expanding Credit Access: Using Ran-

domized Supply Decisions to Estimate the Impacts. Review of Financial

Studies, 2010, vol. 23, issue 1, 433-464.

[47] Kempson, E., & Whyley, C., 1999. Kept Out or Opted Out? Understand-

ing and Combating Financial Exclusion. The Policy Press.

[48] Kim, D-W., Yu, J-S., & Kabir Hassan, M., 2018. Financial Inclusion and

Economic Growth in OIC Countries. Research in International Business

and Finance 43 (2018) 114.

[49] Klapper, L., Laeven, L., & Rajan, R., 2006. Entry Regulation as a Barrier

to Entrepreneurship. Journal of Financial Economics, 82, 591629.

[50] Koopmans, T., C., 1965. On The Concept of Optimal Economic Growth.

In The Economic Approach to Development Planning. Amsterdam: Else-

vier.

[51] Leyshon, A., & Thrift, N., 1993. The Restructuring of The U. K. Financial

Services Industry in the : A Reversal of Fortune? Journal of Rural Studies

9 (3), 223241.

[52] Leyshon, A., & Thrift, N., 1994. Access to Financial Services and Financial

Infrastructure Withdrawal: Problems and Policies. Area 26, 268275.

[53] Leyshon, A., & Thrift, N., 1995. Geographies of Financial Exclusion: Fi-

nancial Abandonment in Britain and the United States. Trans. Inst. Br.

Geogr. 20 (3), 312341.

[54] Ludvigson, S., 1999. Consumption And Credit: A Model Of Time-Varying

Liquidity Constraints. The Review of Economics and Statistics, 1999, vol.

81, issue 3, 434-447.

[55] Mian, A., 2006. Distance Constraints: The Limits of Foreign Lending in

Poor Economies. Journal of Finance, 61, 14651505.

[56] Prasad, E.S., 2010. Financial Sector Regulation and Reforms in Emerging

Markets: An Overview. National Bureau of Economic Research.

835
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[57] Rajan, R. G., & Zingales, L., 2003. The Great Reversals: The Politics of

Financial Development in The Twentieth Century. Journal of Financial

Economics, 69(1), 550.

[58] Ramsey, F., P., 1928. A Mathematical Theory of Saving. Economic Jour-

nal 38 (December): 543-559.

[59] Romer, David. Advanced Macroeconomics. Fifth Edition, McGraw-Hill.

[60] Rousseau, P. L., & Wachtel, P., 2002. Ination Thresholds and The

Finance-Growth Nexus. Journal of International Money and Finance,

21(6), 777793.

[61] Sarma, M., & Pais, J., 2011. Financial Inclusion and Development. Jour-

nal International Development, 23 (5), 613-628.

[62] Sarma, M., 2008. Index of Financial Inclusion. ICRIER Working Paper

215.

[63] Somville, V., & Vandewalle, L., 2016. Saving by Default: Evidence from a

Field Experiment in Rural India. Graduate Institute of International and

Development Studies Geneva Working Paper Series, No. HEIDWP02-2016.

[64] Suri, T., & Jack, W., 2016. The Long-Run Poverty and Gender Impacts

of Mobile Money. Science 354 (6317): 128892.

[65] Bank, World, 2013. Global Financial Development Report 2014: Financial

Inclusion. World Bank Publications.

[66] World Bank., 2013b. Universal Financial Access Is Vital to Re-

ducing Poverty, Innovation Key to Overcoming the Enormous Chal-

lenge, Says President Jim Yong Kim. Press release, 11 October.

[Link] [Link]/en/news/press-release/2013/10/11/universal-

nancial-access-vital-reducing-poverty-innovation-jim-yong-kim.

[67] Wurgler, J., 2000. Financial Markets and The Allocation of Capital. Jour-

nal of Financial Economics, 58, 187214.

Appendix A.

We set the Lagrangian:

∞
X
Li = β t {logcit − λt (cit + kti − (1 − δ)kt−1
i i
− rt kt−1 i
+ τ rt kt−1 − πti − Tt )}.
t=0

836
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

F.O.C:

∂Li 1 1
= 0 =⇒ λt = i =⇒ λt+1 = i
∂cit ct ct+1

∂Li
= 0 =⇒ λt = βλt+1 (rt+1 − τ rt+1 + 1 − δ)
∂kti

∂Li
= 0 =⇒ cit + kti − (1 − δ)kt−1
i i
− rt kt−1 i
+ τ rt kt−1 − πti − Tt = 0.
∂λt

Then, we get the following optimality conditions:

cit i(θ−1)
1=β (1 + (1 − τ )θAt+1 IF I θ kt − δ),
cit+1

cit + kti − (1 − δ)kt−1

i i
= At (IF Ikt−1 )θ .

Appendix B.

The aggregation rules are:

R1 R1
ct = 0
cit di,kti = 0
kti di,
R1 R1
yt = 0
yti di,πt = 0
πti di.

Appendix C.

Account (% age 15+): the percentage of respondents who report having an

account (by themselves or together with someone else) at a bank or another type

of nancial institution (see denition for nancial institution account) or report

personally using a mobile money service in the .past 12 months (see denition

for mobile money account).

Financial institution account (% age 15+) : the percentage of respon-

dents who report having an account (by themselves or together with someone

else) at a bank or another type of nancial institution.

Mobile money account (% age 15+) : the percentage of respondents

who report personally using a mobile money service in the past 12 months.

GDP per capita: Gross Domestic Product per capita.

837
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Trade Credit, Trade Income Elasticity and the

International Transmission of Shocks

Anna Watsony

University of Cambridge

June 2020

Abstract
The paper examines the impact of trade credit on cyclical ‡uctuations in international
trade. It provides new empirical evidence based on …rm-level UK and Irish data showing
that exporters use trade credit more actively and intensively than non-exporters. The
study introduces inter-…rm lending into an open economy general equilibrium model with
heterogeneous …rms and endogenous entry into the exports market. It demonstrates that
trade credit ampli…es the impact of macroeconomic shocks on international trade both along
the intensive and extensive margins and that it signi…cantly contributes to the high trade
income elasticity observed in the data.
JEL classi…cation: F41; E32; E51
Keywords: Trade Credit, International Trade, Business Cycle Fluctuations, Trade
Income Elasticity

I am very grateful to Sean Holly, Giancarlo Corsetti, Philip Lane, Sergejs Saksonovs, Paul Youdell, Oliver
de Groot as well as the participants of the 2018 Royal Economic Society Annual Conference, the 2018 European
Trade Study Group Annual Conference and the EEA-ESEM 2019 for their helpful comments and suggestions.
Financial support from the Economic and Social Research Council and Corpus Christi College, Cambridge, UK,
is gratefully acknowledged.
y
Faculty of Economics and Fitzwilliam College, University of Cambridge, Storey’s Way, Cambridge CB3 0DG,
United Kingdom, E-mail: ac531@[Link], Tel. +441223768470.

838
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1 Introduction
One of the prevalent features of international business cycle ‡uctuations which is di¢ cult to
reconcile with standard open economy macroeconomic models is the high volatility of imports and
exports relative to output and strong procyclicality of international trade. The average elasticity
of world trade to world income was estimated to be equal to 1.7 during the second half of the last
century (Irwin, 2002) and 3.7 in the 2000s (Freund, 2009). In contrast, a canonical international
business cycle model implies trade income elasticity equal to one. While recent research has
demonstrated that the high sensitivity of international trade to changes in output can to some
extent be explained by the composition of international trade, vertical trade integration and
inventory adjustment (Levchenko et al., 2010; Bems et al., 2010; Alessandria et al., 2010; Engel
and Wang, 2011; Bussière et al., 2013; Eaton et al.; 2016), the impact of …nancial factors on
cyclical trade ‡uctuations is not well understood. This paper investigates whether and to what
extent trade credit representing inter-…rm lending contributes to the high procyclicality of the
trade to GDP ratio.
The contribution of this paper is two-fold. Firstly, it provides new empirical evidence on
the di¤erences in the intensity of trade credit use by exporting and non-exporting …rms based
on …nancial data of 60,000 …rms in the UK and Ireland. It shows that the percentage of …rms
supplying and receiving trade credit is higher among exporters than non-exporters and that …rms
engaged in international trade extend signi…cantly more credit to their customers as a fraction
of their revenues than …rms serving only the domestic market. Secondly, the study introduces
inter-…rm lending and counterparty risk into an open-economy general equilibrium model with
heterogeneous …rms and the extensive margin of trade. It then examines how trade credit alters
the transmission of shocks in the economy and the dynamics of international trade.
Trade credit, which allows customers to delay payment until some time after delivery and is
mostly associated with the purchase of intermediate goods, is one of the most important sources
of short-term …nancing for …rms (Rajan and Zingales, 1995; Petersen and Rajan, 1997; Kohler et
al., 2000; Demirguc-Kunt and Maksimovic, 2001). Corporate surveys indicate that the majority
of …rms make the majority of their sales on credit (Wilson and Summers, 1997). Raddatz (2010)
reports that in 60 percent of countries covered by Worldscope trade credit constitutes a more
important source of short-term …nancing for …rms than bank credit.
Trade credit plays a particularly important role in facilitating international trade. It is
estimated that between 80 to 90 per cent of international transactions rely on some trade …nance
facilities (Auboin, 2009). According to the IMF/BAFT-IFSA Trade Finance Surveys, over 60 per
cent of these transactions are supported by trade credit (Asmundson et al., 2011).1 While most
1
The most common form of trade …nance, used in about 40 per cent of transactions, is open account …nancing,
in which the exporter grants credit to the importer directly and bears the non-payment risk. A further 20 per
cent of transactions are …nanced through cash-in-advance in which the exporter receives trade credit from the
importer.

839
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

empirical studies on trade credit do not distinguish between its use in domestic and international
trade, Eck et al. (2012) demonstrate, based on German survey data, that companies that export
their goods are signi…cantly more likely to extend and receive trade credit than …rms selling
their goods only on the domestic market and they also have relatively larger average shares of
transactions for which trade credit is given and received. There are a number of factors which
may contribute to the relatively more intensive use of trade credit in international transactions.
Firstly, due to the fact that international trade involves longer shipment times and additional
time for the completion of cross-border administrative procedures, it is associated with longer
time lags between production and the receipt of sales revenues, including trade credit repayments.
This leads to greater working capital requirements and it also increases the value of accounts
receivable in proportion to annual sales revenues, which is the standard measure of the intensity
of trade credit provision.2 Secondly, exporters, which are on average larger and more productive
than non-exporting …rms3 tend to have better access to external …nance and may be in a better
position to supply trade credit to their customers.4 Furthermore, Eck et al. (2012) show within a
theoretical model, for which they …nd empirical support, that the widespread use of trade credit
in international trade may be due to the inherently greater degree of uncertainty in international
than in domestic transactions. Trade credit can reduce this uncertainty and alleviate the resulting
…nancial constraints by serving as a quality signalling device.
Trade credit exposes its supplier to a counterparty risk which has been shown to be substan-
tial. Using French …rm-level data, Boissay and Gropp (2007) show that defaults on trade credit
are common – on average 19 per cent of …rms default at least once per quarter. The defaults
faced by …rms correspond on average to about 2 per cent of their receivables. The average share
of quarterly defaults in current liquid assets is as high as 44 per cent –they can therefore con-
stitute a major liquidity shock. Jacobson and Schedvin (2015) demonstrate using Swedish data
that trade credit suppliers incur considerable trade credit losses due to trade debtor failures and
that their bankruptcy risks increase with the size of the losses incurred. The authors …nd that
8.3 percent of …rms face at least one trade debtor failure in a year and that …rm failures impose
larger credit losses on the corporate sector than on the banking sector. Bradley and Rubach
(2002) and Bradley and Cowdery (2004) …nd that non-payment of trade credit by customers
is one of the most important causes of bankruptcy among US …rms. The high risk associated
with trade credit is re‡ected in the relatively high cost of this type of …nancing. The discount
rates typically o¤ered for early payment, which can serve as a proxy for trade credit price, are
2
Due to the way in which the ratio of accounts receivable to sales revenues is calculated, it re‡ects both the
fraction of goods which are sold and purchased on credit terms and the average time period within which the
credit is repaid.
3
Bernard and Jensen (1995, 1997, 1999, 2004) provide extensive empirical evidence of this phenomenon while
Melitz (2003) explains it within a theoretical framework.
4
Manova (2013) and Chaney (2016) show theoretically that …nancially constrained …rms are less likely to
export their goods and their results were con…rmed empirically by Buch et al. (2010), Bellone et al. (2010),
Minetti and Chun Zhu (2011) and Muuls (2015) for German, French, Italian and Belgian …rms, respectively.

840
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

equivalent to annual interest rates of over 40 per cent (Petersen and Rajan, 1997; Boissay, 2006).
Cuñat (2007) shows theoretically and empirically that the high interest rates on trade credit
arise from the existence of a default premium and a premium for insurance against potential
liquidity shocks, which the supplier provides for the customer. As the non-payment risk depends
on macroeconomic conditions and changes over the business cycle, inter-…rm lending can be
expected to play a role in the propagation of shocks in the economy.
While the motivation behind the widespread use of trade credit has been studied extensively
both from a theoretical and an empirical perspective,5 the macroeconomic implications of this
type of …nancing are not well understood. Until recently, research in this area focused predomi-
nantly on a closed economy and the use of trade credit in domestic transactions. Kiyotaki and
Moore (1997) showed within a partial equilibrium framework that trade credit can amplify the
e¤ects of shocks on the level of economic activity. In the presence of trade credit supply chains
a temporary shock to the liquidity of some …rms may trigger a sequence of bankruptcies and a
severe decline in output, as …rms facing a default by their customers may be forced to default on
their own suppliers, propagating the shock through the economy. This theory was subsequently
extended to a general equilibrium framework and tested empirically (Cardoso-Lecourtois, 2004;
Boissay, 2006; Boissay and Gropp, 2007; Raddatz, 2010; Jacobson and Schedvin, 2015). Con-
versely, a number of studies have investigated whether and to what extent trade credit may
help to mitigate the impact of shocks to the economy by providing an alternative to bank loans
as a source of external …nance for …rms in times of tight monetary policy and …nancial dis-
tress (Calomiris et al., 1995; Kohler et al., 2000; Atanasova and Wilson, 2003; Choi and Kim,
2005; Guariglia and Mateut, 2006; Mateut et al., 2006; Love et al., 2007; Huang et al., 2011;
Garcia-Appendini and Montoriol-Garriga, 2013). Both propositions found some empirical sup-
port. While the great trade collapse of 2008-2009 triggered by the …nancial crisis generated
considerable interest in the role of trade credit in an international context, recent studies on this
topic focus almost exclusively on the microfoundations of inter-…rm lending and the choice of
trade …nance instruments by …rms (Ahn, 2011; Schmidt-Eisenlohr, 2013; Antràs and Foley, 2015)
rather than on their consequences for international business cycle ‡uctuations. Furthermore, they
are conducted within a partial equilibrium framework. This paper, by contrast, investigates the
impact of trade credit on the transmission of shocks in an open economy and, in particular, its
e¤ects on cyclical ‡uctuations of international trade along the intensive and extensive margins.6
Moreover, the analysis is based on a general equilibrium model.
In the model developed, macroeconomic shocks a¤ect the default rate among …rms and the
fraction of trade credit which is not repaid. Changes in the riskiness of inter-…rm loans lead to
changes in their cost, which in turn alters …rms’marginal costs and prices. Due to the di¤erences
5
For an overview of the literature see Petersen and Rajan (1997) and Klapper et al. (2012).
6
Changes in trade along the intensive margin refer to changes in the volume of already traded varieties, whereas
changes in trade along the extensive margin correspond to changes in the number of varieties traded.

841
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

in the intensity of trade credit use in domestic and international transactions, macroeconomic
shocks and the accompanying changes in the counterparty risk associated with inter-…rm lend-
ing a¤ect the domestic and foreign market price of traded varieties di¤erently, which leads to
changes in the relative price of domestic and imported goods. Macroeconomic shocks as well
as the resulting changes in international relative prices a¤ect the pro…tability of exports and
induce changes in the fraction of varieties which are traded internationally. This causes further
adjustment in the relative price of and the demand for domestic and imported goods and, in
consequence, also in the volume of international trade.
This study demonstrates that due to di¤erences in the riskiness of international and domestic
transactions, cyclical changes in the cost of trade credit amplify the impact of macroeconomic
shocks on trade both along the intensive and extensive margins. The model generates trade
income elasticities ranging between 1.3 and 1.5 depending on the type of shock and shows that
inter-…rm lending signi…cantly contributes to the strong procyclicality of international trade
observed in the data.
The remainder of the paper is organised as follows. Section two presents new empirical
evidence on the use of trade credit by exporters and non-exporters in Ireland and in the UK.
The third section outlines the model developed to account for the impact of inter-…rm lending on
cyclical ‡uctuations in international trade. Section four describes the calibration of the model
parameters. In section …ve the quantitative e¤ects of trade credit for the transmission of shocks
are discussed. The last section concludes.

2 Trade credit use - empirical evidence

There are two main measures of the intensity of trade credit use: the ratio of accounts receivable
to sales revenues, which represents trade credit that …rms provide to their customers, and the
ratio of accounts payable to the cost of goods sold, which corresponds to trade credit that …rms
obtain from their suppliers.7 Trade credit constitutes a substantial component of corporate assets
and liabilities. Dass et al. (2015) show using Compustat data that in the period from 1997 to
2008 the average value of accounts receivable to sales in the US was 17.8 per cent, while the
average ratio of accounts payable to sales was 12.9 per cent. Raddatz (2010) reports that in a
sample of 43 countries the ratio of accounts payable to the cost of goods sold ranges from 7 to
33 per cent and its mean and median values are equal to 15 and 14 per cent, respectively.
While trade credit is prevalent both in domestic and international transactions, it is par-
ticularly widely used by exporting …rms. Eck et al. (2012) examined the di¤erences in the
patterns of inter-…rm lending between internationally active and non-active …rms based on Ger-
7
For a given …rm and a given year these ratios are equal to the average value of accounts receivable/accounts
payable at the end of the current and the previous year divided by the total value of sales revenues/cost of goods
sold in the current year.

842
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

man data from the 2004 Business Environment and Enterprise Performance Surveys (BEEPS).
The authors demonstrate that exporters use trade credit more actively and more intensively
than non-exporting …rms. 94 per cent of exporters supply trade credit to their customers, as
compared to 80 per cent of non-exporters. The fractions of …rms obtaining trade credit from
their trading partners are equal to 96 per cent for exporters and 92 per cent for non-exporters.
The average share of transactions for which trade credit is provided is also signi…cantly higher
for internationally active …rms and is equal to 64 per cent for exporters and 49 per cent for
non-exporters. The corresponding shares of transactions for which trade credit is received are
equal to 75 per cent among exporters and 63 per cent among non-exporting …rms.
This study provides new evidence on the discrepancies in the pattern of trade credit use
between exporters and non-exporters. Analysis of …nancial data from the 60,000 largest British
and Irish companies reveals that trade credit is a major source of short-term …nancing for …rms
and that it is considerably more widely used by companies engaged in international trade than
by …rms serving only the domestic market. The …rms covered by the analysis jointly represent
over 99 per cent of the total revenue and over 99.9 of the total employment in the corporate
sector in Ireland and in the UK.
An overwhelming majority of British and Irish …rms gives and receives trade credit. Between
2003 and 2014 on average 80 per cent of the 60,000 largest British and Irish companies recorded
positive values of ‘trade debtors’, which corresponds to trade credit granted, and 83 per cent
of these …rms had positive values of ‘trade creditors’, which represents trade credit received.
Inter-…rm lending was even more widespread in the manufacturing sector, in which on average
84 per cent of …rms provided trade credit to their customers and 83 per cent of …rms obtained
trade credit from their suppliers. The average ratio of trade debtors to sales revenues in the
period covered by the analysis was equal to 13 per cent among manufacturing …rms and 12 per
cent overall.
The fraction of UK and Irish …rms providing trade credit to their customers is considerably
higher among exporters than non-exporters. The pattern of a more intensive use of trade credit by
internationally active …rms is very pronounced both in manufacturing and in the entire population
of …rms. It is also independent of the …rms’scale of operation. Table 1 shows a range of measures
of the intensity of trade credit use by UK and Irish manufacturing …rms in the period 2003-2014,
while Table 2 provides similar statistics for …rms from all sectors of the economy. The three
measures of the intensity of trade credit provision presented are the mean and median values of
the ratio of trade debtors to turnover (TN and TD, respectively), which represents trade credit
granted as a fraction of sales revenues, and the percentage of …rms for which the value of trade
debtors is positive (TF) and which are therefore trade credit providers.8
8
Among the 15,000 largest manufacturing …rms about 66 per cent reported their national and overseas turnover
separately, which made it possible to determine their export status and include them in the sample for which the
measures of trade credit use were calculated. The fraction of …rms for which export status could be determined
was equal to 53 per cent among the 60,000 largest UK and Irish …rms.

843
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1: Trade credit use among manufacturing …rms

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Mean
Top 1,000 manufacturing …rms
TN E 12.6 12.7 13.1 12.5 13.1 12.2 11.8 12.0 12.8 12.4 12.6 13.1 12.6
N 9.7 9.5 8.1 8.3 8.3 8.0 8.0 9.3 10.3 9.5 10.0 9.9 9.1
D 2.9* 3.2* 5.0* 4.2* 4.8* 4.2* 3.8* 2.7* 2.5* 2.9* 2.6* 3.2* 3.5*
TD E 12.7 12.7 13.3 12.5 12.7 11.6 11.5 12.0 12.2 11.9 12.3 12.6 12.3
N 8.6 7.3 7.8 7.7 8.1 7.8 7.7 9.2 9.8 10.0 9.2 9.2 8.5
TF E 97.7 97.1 96.7 96.5 96.7 96.8 96.2 96.6 97.1 97.0 96.9 96.9 96.9
N 90.9 89.9 92.8 91.5 90.8 92.5 93.0 92.2 92.1 94.9 94.5 96.8 92.7
Top 5,000 manufacturing …rms
TN E 15.5 15.4 15.5 15.5 15.5 14.8 14.4 15.2 15.6 15.2 15.6 16.2 15.4
N 12.3 12.5 12.2 12.0 12.2 11.4 11.3 12.4 13.3 12.6 12.7 13.9 12.4
D 3.2* 2.9* 3.3* 3.5* 3.3* 3.4* 3.1* 2.8* 2.3* 2.6* 2.9* 2.3* 3.0*
TD E 15.4 15.4 15.6 15.7 15.6 14.6 14.4 15.2 15.2 15.0 15.3 15.8 15.3
N 11.8 12.0 11.8 12.0 11.9 11.0 10.8 11.9 12.9 12.4 12.6 14.0 12.1
TF E 98.3 97.7 97.4 97.5 97.8 97.3 97.4 97.5 98.1 98.1 98.2 98.0 97.8
N 92.8 94.0 93.3 93.2 93.4 93.0 93.3 94.0 93.7 93.4 93.6 95.6 93.6
Top 10,000 manufacturing …rms
TN E 16.6 16.4 16.5 16.6 16.5 15.6 15.4 16.5 16.7 16.1 16.6 16.6 16.3
N 14.1 14.3 14.1 13.9 14.0 13.0 13.0 14.5 15.1 14.2 14.2 13.9 14.0
D 2.5* 2.1* 2.4* 2.7* 2.5* 2.6* 2.4* 2.0* 1.6* 1.9* 2.4* 2.7* 2.3*
TD E 16.3 16.2 16.3 16.3 16.3 15.2 15.2 16.2 16.2 15.8 16.2 15.9 16.0
N 14.3 14.3 13.9 14.1 14.0 13.0 12.7 14.5 14.8 14.3 14.1 13.2 13.9
TF E 98.1 97.6 97.2 97.3 97.4 97.1 97.4 97.8 98.4 98.5 98.2 97.0 97.7
N 92.4 92.6 91.7 91.5 91.3 90.4 92.9 93.8 94.6 93.7 94.2 85.3 92.0
Source: Fame
Notation: TN and TD –the mean and the median of the trade debtors to turnover ratio
respectively; TF –the fraction of …rms for which the value of trade debtors is positive;
E –exporters; N –non-exporters;
D –the di¤erence between the values of TN for exporters and non-exporters;
* denotes that the di¤erence between the values of TN for exporters and non-exporters
is statistically signi…cant at 1 per cent signi…cance level.

844
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Trade credit use among all …rms

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Mean
Top 5,000 …rms
TN E 12.4 12.6 12.8 15.8 12.5 11.8 11.8 12.3 12.5 12.5 12.4 13.0 12.7
N 8.6 8.8 8.1 14.8 8.1 7.7 7.7 8.7 9.6 9.6 9.4 9.4 9.2
D 3.8* 3.8* 4.7* 1.0 4.4* 4.1* 4.1* 3.6* 2.9* 2.9* 3.0* 2.6* 3.5*
TD E 12.3 12.2 12.2 14.0 11.7 10.8 10.8 11.2 11.3 10.8 10.7 11.7 11.6
N 6.1 5.9 5.7 9.1 5.1 5.0 4.7 5.8 6.2 5.8 5.7 5.1 5.9
TF E 95.7 95.4 94.9 97.8 94.0 93.4 93.8 94.8 94.8 94.7 94.6 94.0 94.8
N 88.3 88.1 87.3 94.6 85.9 87.8 86.3 87.1 90.7 90.1 90.8 85.9 88.6
Top 10,000 …rms
TN E 14.2 16.4 14.0 16.4 13.7 12.9 12.7 13.3 13.9 13.6 13.6 14.0 14.1
N 9.3 13.7 9.2 13.7 8.9 8.4 8.4 9.4 10.3 10.1 10.0 10.5 10.2
D 4.9* 2.7* 4.8* 2.7* 4.8* 4.5* 4.3* 3.9* 3.6* 3.5* 3.6* 3.5* 3.9*
TD E 13.9 15.1 13.7 15.1 13.2 12.2 11.9 12.5 12.4 12.4 12.2 13.2 13.2
N 6.8 9.1 6.7 9.1 6 5.8 5.5 6.2 7.6 6.7 6.5 6.0 6.8
TF E 96.0 97.8 95.6 97.8 94.7 93.9 94.3 94.8 95.5 95.3 95.5 94.7 95.5
N 88.6 94.6 87.6 94.6 86.2 86.9 86.9 86.9 89.7 89.6 90.0 86.2 89.0
Top 60,000 …rms
TN E 16.4 16.3 16.3 16.4 16.5 15.8 15.415.9 16.2 16.1 16.2 16.1 16.1
N 11.0 11.2 11.3 11.1 11.1 10.5 10.811.5 12.2 11.7 11.7 11.3 11.3
D 5.4* 5.1* 5.0* 5.3* 5.4* 5.3* 4.6*4.4* 4.0* 4.4* 4.5* 4.8* 4.9*
TD E 15.4 15.5 15.5 15.6 15.7 14.6 14.615.2 15.2 15.1 15.3 15.7 15.3
N 6.2 6.3 6.4 6.6 6.7 6.4 7.3
7.8 8.2 7.8 7.8 6.7 7.0
TF E 93.2 92.5 92.7 92.7 92.9 93.1 94.394.0 95.0 95.3 95.3 92.9 93.7
N 76.2 76.5 78.0 78.3 79.2 80.2 84.583.8 85.9 85.8 86.5 79.2 81.2
Source: Fame
Notation: TN and TD –the mean and the median of the trade debtors to turnover ratio
respectively; TF –the fraction of …rms for which the value of trade debtors is positive;
E –exporters; N –non-exporters;
D –the di¤erence between the values of TN for exporters and non-exporters;
* denotes that the di¤erence between the values of TN for exporters and non-exporters
is statistically signi…cant at 1 per cent signi…cance level.

845
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

In the manufacturing sector the measures of trade credit use are calculated for three di¤erent
subgroups of …rms –the largest 1,000, 5,000 and 10,000 companies in terms of their revenues.9
Among the top 1,000 manufacturing …rms, 97 per cent of exporters and 93 per cent of non-
exporters extended trade credit to their customers. The average ratio of trade credit granted to
turnover was equal to 12.6 per cent among internationally active …rms and 9.1 per cent among
…rms selling their goods only on the domestic market. If we assume that trade credit is usually
granted for about two months, in line with the empirical evidence provided by Klapper et al.
(2012), then this would imply similar estimates of the intensity of trade credit use by exporters
and non-exporters to the ones obtained by Eck et al. (2012) for German …rms based on survey
data. For all the di¤erent sample sizes and all the di¤erent years covered by the analysis the
mean ratio of trade debtors to revenues was higher for exporters than for non-exporters –with
the average di¤erence ranging from 2.3 to 3.5 percentage points, depending on the number of
…rms included in the analysis. For all sample sizes and all years the di¤erences in the means
are statistically signi…cant at the 1 per cent signi…cance level. Similarly, for all sample sizes and
all years considered the fraction of exporters granting trade credit is higher than the fraction
of non-exporters acting as trade credit providers. Over the time period analysed, the average
di¤erence between these fractions lies between 4.2 percentage points for the largest 1,000 …rms
and 5.7 percentage points for the largest 10,000 …rms.
The relatively more widespread use of inter-…rm lending by exporters is not limited to man-
ufacturing …rms and can be observed among all …rms in the economy. Among the top 60,000
British and Irish …rms, representing over 99 per cent of corporate sector revenues, on average 94
per cent of exporters acted as trade credit providers, as compared to 81 per cent of non-exporters.
The average ratio of trade debtors to turnover in the period analysed was equal to 16.1 per cent
for exporters and 11.3 per cent for non-exporting …rms. For all sample sizes and for all years
covered by the analysis apart from one the di¤erences in the ratios of trade debtors to turnover
for exporters and non-exporters were positive, substantial and statistically signi…cant at the sig-
ni…cance level of 1 per cent.10 Furthermore, for all sample sizes and all years the percentage
of …rms extending trade credit is relatively higher among …rms engaged in international trade
and the di¤erences in the ratios of trade credit providers among exporters and non-exporters
range from an average of 6.2 percentage points for the largest 5,000 …rms to an average of 12.5
percentage points for the largest 60,000 …rms.
The ratio of accounts receivable to sales revenues re‡ects the extent to which the trade credit
supplier is exposed to non-payment risk on the part of its customers. The more extensive use of
trade credit by internationally active …rms leaves them more exposed to the counterparty risk
associated with such credit and more vulnerable to changes in macroeconomic conditions which
9
The revenues of the …rms included in these di¤erent subgroups represent on average 82, 96 and over 99 per
cent of total revenues generated in the manufacturing industry, respectively.
10
The revenues of the companies included in the subsamples of the top 5,000, 10,000 and 60,000 …rms examined
represent on average 87, 92 and over 99 per cent of total revenues generated by UK and Irish …rms, respectively.

846
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

a¤ect the risk of non-payment. The next section develops a model which introduces trade credit
and counterparty risk into an open-economy general equilibrium framework with heterogeneous
…rms and which is then used to analyse the e¤ects of inter-…rm lending on the international
transmission of shocks.

3 Model
The world economy consists of two symmetric countries: Home and Foreign. Each country is
populated by utility-maximising households and pro…t-maximising, monopolistically competitive
…rms. There are two types of …rm: intermediate goods producers and …nal goods producers.
Intermediate goods producers employ domestic labour in order to produce intermediate inputs
and sell them to the producers of …nal consumption goods. They decide whether to sell their
goods only in the domestic market or whether to export them as well. Final goods producers
combine intermediate inputs produced both in the domestic and in the foreign economy in order
to produce di¤erentiated …nal consumption goods, which are then sold to domestic households.
When selling their goods, the intermediate goods suppliers provide trade credit to …nal goods
producers and are thereby exposed to credit risk. In each period an endogenously determined
fraction of trade credit granted is not repaid as a result of defaults in the sector producing …nal
consumption goods. Business cycles are driven by aggregate productivity shocks and preference
shocks leading to changes in aggregate demand. It is assumed that prices are ‡exible and that
households have access to a complete set of state-contingent assets traded internationally.
All …rms in the economy are subject to idiosyncratic productivity shocks. In the sector of
intermediate goods, …rm heterogeneity with regard to productivity leads to time-varying entry
into the exports market and generates ‡uctuations in international trade along the extensive
margin. Idiosyncratic productivity shocks in the sector of …nal consumption goods result in
time-varying defaults in this sector which a¤ect the cost of trade credit and the relative prices
of domestic and foreign goods and contribute to changes in trade along the intensive margin. As
the number of traded varieties determines the share of each variety in the consumption basket
and as changes in the relative demand for domestic and foreign goods a¤ect the pro…tability of
exports, these two margins of trade interact with each other. As in the seminal paper by Melitz
(2003), …rm heterogeneity is integrated into the model in such a way that all the information
on the distribution of productivity levels across …rms which is relevant for aggregate outcomes
is completely summarised by productivity averages for di¤erent categories of …rms.

3.1 Households
In each country there is a continuum of identical, in…nitely-lived households. A representative
household supplies labour to the producers of intermediate and …nal consumption goods in the

847
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

domestic economy and consumes a basket of …nal consumption goods. Its period utility function
is additively separable in consumption, Ct , and labour, Lt , and is given by:
" #
X
1 1
Ct+k L1+'
t+k
k
Et eut (1)
k=0
1 1+'

where 2 (0; 1) is the intertemporal discount factor, > 0 is the inverse of the intertemporal
elasticity of substitution in consumption, ' 0 is the inverse of the Frisch elasticity of labour
supply, eut represents a shock to the marginal utility of consumption such that ut = u ut 1 + ut ,
u
u 2 (0; 1) and t N (0; 2u ):
Households choose their consumption and labour supply in order to maximise their expected
discounted lifetime utility subject to a sequence of budget constraints:

Pt Ct + Et [Ot;t+1 Dt+1 ] = Wt Lt + Dt + t (2)

where Pt is the aggregate price level in the Home economy at time t, Wt denotes nominal wage
and t is a lump sum component of households’income including dividends from the ownership
of …rms. Ot;t+1 is the stochastic discount factor for one-period ahead nominal payo¤s, Dt+1 is
the nominal payo¤ in period t + 1 of the portfolio held at the end of period t. Households have
unrestricted access to a complete set of internationally traded state-contingent claims.
The solution to the household decision problem gives the following intratemporal labour
supply equation:

Wt
= Ct L't e ut
(3)
Pt
which equalises real wage with the marginal rate of substitution between consumption and leisure.
The optimal consumption choice in the Home economy satis…es the following intertemporal
Euler equation:

1 Pt
Ct eut = Et Ct+1 eut+1 (4)
Ot;t+1 Pt+1
while in the Foreign economy the consumption Euler equation is given by:11

1 S t Pt
Ct eut = Et Ct+1 eut+1 (5)
Ot;t+1 St+1 Pt+1
where St is the nominal exchange rate, expressed as the price of one unit of the foreign currency
in terms of the domestic currency.
Assuming that the Home and Foreign economies are initially perfectly symmetric, the well-
known risk-sharing condition holds:
11
Throughout the paper, Foreign variables are denoted by an asterisk.

848
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Ct S t Pt
e(ut ut )
= (6)
Ct Pt
according to which the ratio of marginal utilities of consumption in the Home and Foreign
economy is equal to the ratio of relative prices of consumption in these economies.
Each household consumes all varieties of the …nal consumption goods which are produced in
the domestic economy and are indexed by j. The consumption aggregator, Ct , is de…ned as:

2 1 3 B
Z B 1
B 1

Ct = 4 Cj;t B dj 5 (7)
0

where B > 0 is the elasticity of substitution between …nal consumption goods. The utility-based
consumption price index, Pt , is therefore given by:
2 1 31 1
Z B

Pt = 4 Pj;t
1 B
dj 5 (8)
0

Given households’preferences speci…ed above, the demand for variety j of the …nal consumption
good, denoted by Cj;t , is positively related to the aggregate demand in the Home economy, Ct ,
and negatively related to the price of this variety, Pj;t , relative to the aggregate price index , Pt ,
and can be written as:

B
Pj;t
Cj;t = Ct (9)
Pt

3.2 Firms
3.2.1 Intermediate goods producers

Each economy is populated by a continuum of monopolistically competitive intermediate goods

producers, indexed by i on the unit interval. They employ domestic labour in order to produce
di¤erentiated intermediate goods using a production technology given by:

Vi;t = Zt Ai;t Li;t (10)

where Vi;t is the output produced by …rm i at time t and Li;t is the labour input used in the
production of that good. Firms are heterogeneous with regard to their productivity. Their pro-
ductivity level in a given period is determined by two components: Zt , which is common to all
intermediate goods producers and is subject to aggregate shocks such that ln Zt = z ln Zt 1 + zt
where z 2 (0; 1) and zt N (0; 2z ) as well as Ai;t , which is …rm-speci…c. The idiosyncratic pro-
ductivity draws Ai;t are independent across …rms and over time and they come from a distribution

849
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

which is time-invariant.
All the varieties of intermediate goods produced in the domestic economy are sold to the
producers of …nal consumption goods in the domestic economy. Some of them are in addition
sold abroad and the decision to export is endogenous. When selling their goods, intermediate
goods producers receive a fraction of their revenues at the beginning of the period, before delivery,
while the remaining part of the revenues is paid at the end of the period, after the goods are
delivered. The fraction of revenues which is paid in advance of delivery and is therefore safe
from counterparty risk is equal to (1 dH ) for the domestic buyers and (1 dF ) for the foreign
buyers. The parameters dH and dF can be interpreted as a measure of the trade credit provided
by the domestic producers of intermediate goods to the domestic and foreign producers of …nal
consumption goods, which is subject to non-repayment risk. It is assumed that dF > dH ,
which re‡ects the relatively more intensive usage of trade credit by exporting …rms resulting,
among other factors, from the longer time lags between the dispatch and delivery of goods in
international trade.12
Due to the fact that in each period a fraction of …nal goods producers incurs losses and defaults
on their debt, at the end of the period intermediate goods suppliers recover only a fraction of the
trade credit extended. This fraction is equal to (1 t ) for trade credit granted to domestic …nal
goods producers and (1 t ) for trade credit granted to foreign …nal goods producers, where

t and t are the Home and Foreign trade credit default rates respectively – the fractions of
the total value of trade credit received by Home and Foreign …nal goods producers which are
not repaid. The default risk, and as a result also the trade credit default rates, are endogenous
and change over time with changes in macroeconomic conditions driven by aggregate shocks. As
trade credit suppliers are assumed to hold su¢ ciently large and diversi…ed portfolios to ensure
perfect risk pooling, they behave as if they were risk neutral.

Domestic market All intermediate goods producers sell their goods in the domestic economy.
The supply of the variety i produced and sold in the Home economy is denoted by VH;i;t and is
equal to:

VH;i;t = Zt Ai;t LH;i;t (11)

Firm i chooses the amount of labour LH;i;t used in the production of intermediate input i for
the Home market and the price of the input on this market QH;i;t in order to maximise its pro…ts
subject to the demand function and the production technology. The pro…t of intermediate goods
producer i obtained from sales of goods in the Home economy is given by:
12
While in this paper the di¤erence in the trade credit provision by exporters and non-exporters is exogenously
imposed, it would arise endogenously in a richer model incorporating multi-period loans and a time-to-ship friction
in international trade.

850
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Wt
H;i;t = [(1 dH ) + dH (1 t )] QH;i;t XH;i;t XH;i;t (12)
Ai;t Zt
where XH;i;t is the total demand for the intermediate good i by Home producers of …nal con-
sumption goods and it is given by:

Z1
XH;i;t = XH;i;j;t dj (13)
0

XH;i;j;t is the demand for the intermediate variety i by …nal goods producer j, which depends
QH;i;t
on the relative price of this variety on the Home market, Q t
, the total number of varieties
of intermediate goods available for sale in the Home economy, equal to 2 Nt , as well as the
aggregate demand for intermediate goods by the Home producer j, denoted by Xj;t .13
The optimal allocation of expenditures by producer j between the di¤erent varieties of the
intermediate input available in the domestic market implies:

A
1 QH;i;t
XH;i;j;t = Xj;t (14)
2 Nt Qt
where A is the elasticity of substitution between intermediate inputs. Qt is the aggregate price
index for the intermediate goods in the Home economy, which is given by:
1
1 1 N 1 A
Qt = Q1H;t A
+ Q1F;t A
(15)
2 N 2 N
where QH;t and QF;t are the aggregate price indices for the Home and Foreign intermediate goods,
respectively, and are equal to:
0 1 11 1
Z A

QH;t = @ Q1H;i;tA A (16)

and
0 11 1
Z1 A

QF;t =@ Q1F;i;tA A (17)

The total domestic demand for the variety produced by …rm i is therefore equal to:

A
1 QH;i;t
XH;i;t = Xt (18)
2 Nt Qt
13
The number of varieties of intermediate goods available for sale in the Home economy is equal to 2 Nt ,
as 2 is the mass of intermediate goods producers in the whole world and Nt is the share of intermediate goods
producers in the Foreign economy who do not export their goods.

851
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

where:

Z1
Xt = Xj;t dj (19)
0

In equilibrium the total demand for the intermediate good i in the Home market, XH;i;t , needs
to be equal to the supply of this good on this market, VH;i;t : XH;i;t = VH;i;t .
At the beginning of each period the intermediate goods supplier sets the price for its good
and is then prepared to sell it to any …nal goods producer at this price. When setting the price,
the intermediate goods producers know the realisation of all the aggregate shocks as well as their
idiosyncratic productivity level in that period. They also know the properties of the distributions
of …rm-speci…c shocks in their sector and in the sector producing …nal consumption goods, which
are the same at home and abroad and which do not change over time. The pro…t-maximising
price for the Home market is equal to a constant mark-up over marginal cost:

A Wt
QH;i;t = gH;t (20)
A 1 Ai;t Zt
where gH;t is given by:

1
gH;t = (21)
(1 dH ) + dH (1 t)

and can be interpreted as the cost of the trade credit suppliers’ insurance against the domes-
tic buyers’ default. Equation (20) shows that endogenous changes in the trade credit default
rate among …nal goods producers, t , a¤ect the marginal cost and prices of the producers of
intermediate goods and, as a result, also the demand for these goods.

Export decision In each period the intermediate goods producers decide whether to export
their goods in addition to selling them in the domestic economy. If a Home intermediate goods
producer i decides to export its good, then the supply of this variety on the foreign market is
equal to VF;i;t :

VF;i;t = Zt Ai;t LF;i;t (22)

and it is a function of the …rm’s level of productivity and the domestic labour employed in the
production of this good for exports, denoted by LF;i;t .
In order to sell their goods abroad, the intermediate goods producers need to pay a …xed cost
of entering the exports market, equal to FA e¤ective units of domestic labour. When exporting,
they also incur an additional variable iceberg trade cost equal to a constant fraction of the
value of the goods exported. Firms can decide whether to enter the exports market in period t
after observing all the aggregate shocks in that period and also their idiosyncratic productivity

852
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

levels, Ai;t . When making their export decisions, the intermediate goods producers also know
the properties of the distributions of …rm-speci…c productivity in the sector producing …nal
consumption goods at home and abroad, Bj;t and Bj;t , which are time-invariant. Firms only
choose to export their goods if the additional pro…ts from the sale of goods abroad are greater
than or equal to the …xed export market entry cost:

Wt Wt
[(1 dF ) + dF (1 t )] St QF;i;t XF;i;t (1 + )XF;i;t FA 0 (23)
Ai;t Zt Zt
where QF;i;t is the foreign-currency price of the variety produced by …rm i and sold in the foreign
market and XF;i;t is the demand for this variety in the Foreign economy.
If a …rm decides to export its variety, its price for the foreign market (in foreign currency) is
set optimally and is equal to a constant mark-up over marginal cost:

A Wt 1
QF;i;t = (1 + )gF;t (24)
A 1 Ai;t Zt St
where gF;t is given by:

1
gF;t = (25)
(1 dF ) + dF (1 t)

Comparison of (24) with (20) reveals that due to the presence of iceberg trade costs and the
fact that international transactions are associated with greater counterparty risk than domestic
transactions, the domestic currency price of variety i on the Foreign market is higher than the
price of this variety in the Home economy and the law of one price does not hold. This, combined
with the fact that only a fraction of domestically produced varieties is exported, generates home
bias in consumption and leads to deviations from purchasing power parity (PPP).
After substituting (24) into (23), the export market entry condition can be expressed as:

1 Wt
St QF;i;t XF;i;t FA 0 (26)
A Zt
where

Z1
XF;i;t = XF;i;j;t dj (27)
0

and XF;i;j;t is the demand for the intermediate variety i by …nal goods producer j, which depends
Q
on the relative price of this variety on the foreign market, QF;i;t , the total number of varieties
t
of intermediate goods available for sale in the foreign economy, equal to 2 Nt , as well as the
aggregate demand for intermediate goods by the Foreign producer j, denoted by Xj;t and given
by:

853
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1 QF;i;t A

XF;i;j;t = Xj;t (28)

2 Nt Qt
The total foreign demand for the variety produced by …rm i is therefore equal to:

1 QF;i;t A

XF;i;t = Xt (29)
2 Nt Qt
where

Z1
Xt = Xj;t dj (30)
0

and in equilibrium it needs to be equal to the supply of this good on the foreign market: XF;i;t =
VF;i;t .
In consequence, the threshold level of productivity above which …rms export their goods,
denoted by At , is determined by the condition:

1
1 1 A gF;t A
Wt A

Xt (St Qt ) A
= FA (31)
A 2 Nt A 1 At Zt

Productivity distribution As in Ghironi and Melitz (2005) and in line with empirical evi-
dence on …rms’size distribution, it is assumed that the …rm-speci…c level of productivity, At , in
the intermediate goods sector is Pareto distributed. Its cumulative distribution function is given
Amin k
by F (A) = 1 A
, where Amin is the lower bound and k is a shape parameter, such that
k > A 1, which determines the dispersion of productivity draws across …rms. As k increases,
dispersion decreases and the …rm productivity levels are increasingly concentrated towards their
lower bound, Amin .14
Assuming that …rms’productivity is Pareto distributed, the probability that at time t it is
not pro…table for a …rm to export its goods is equal to:

kA
Amin
Nt = F (At ) = 1 (32)
At
whereas the probability of exporting is equal to:

kA
Amin
1 Nt = 1 F (At ) = (33)
At
Hence, Nt and (1 Nt ) denote the fractions of non-exporters and exporters among intermediate
goods producers in the Home economy.
14
Following Ghironi and Melitz (2005), the assumption that …rms’productivity is Pareto distributed has been
widely adopted in the literature on heterogeneous …rms and trade (for an overview see Redding, 2011).

854
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

h i 1
kA A 1
Denoting vA = kA ( A 1)
, the average productivity among all intermediate goods pro-
ducers, AA , can be de…ned as:
2 3 1
Z1 A 1

AA = 4 At A 1 dF (A)5 = vA Amin (34)

Amin

The productivity averages among exporters and non-exporters in the Home economy can be
expressed as functions of the threshold level of productivity, At , and the probability of non-
exporting, Nt . The average productivity among non-exporting intermediate goods producers,
AN;t , is given by:

2 3 1
A 1
ZAt 1
6 1 1 7 1 A 1 (1 Nt ) A 1 A 1
AN;t =4 At A
dF (A)5 = vA A At (35)
F (At ) Nt min Nt
Amin

The average productivity among exporting intermediate goods producers, AE;t , is equal to:
2 3 1
A 1
Z1
6 1 7
AE;t = 4 At A 1 dF (A)5 = v A At (36)
1 F (At )
At

3.2.2 Final goods producers

In each economy there is a continuum of …nal goods producers, indexed by j on the unit interval.
As in the case of intermediate goods producers, …nal goods producers di¤er with regard to their
productivity. At the beginning of period t, before learning their productivity levels in that
period, domestic retailers need to incur a …xed cost of production, which is equal to the cost of
FB e¤ective units of domestic labour. After the aggregate and …rm-speci…c shocks are revealed,
domestic …nal goods producers buy intermediate goods from the intermediate goods producers
and transform them into a …nal consumption good using the following production technology:

Yj;t = Bj;t Xj;t (37)

where Bj;t is …rm j’s idiosyncratic level of productivity at time t and Xj;t is the aggregate input
used in the production of Yj;t , which is equal to:

" 1 1
# A
A 1 A 1 A 1
1 A 1 N A
Xj;t = XH;j;t +
A
XF;j;t
A
(38)
2 N 2 N

XH;j;t is the aggregate input provided by the Home intermediate goods producers and XF;j;t is

855
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the aggregate input provided by the Foreign intermediate goods producers to the …nal goods
producer j:
In their production, …nal goods producers use all varieties of intermediate goods which are
available for sale in the domestic economy. XH;j;t is a CES aggregator of all intermediate varieties
produced in the Home economy:

2 1 3 A
Z A 1
A 1
XH;j;t = 4 (XH;i;j;t ) A di5 (39)
0

Similarly, XF;j;t is a CES aggregator of all intermediate varieties produced in the Foreign economy
which are exported:

2 1 3 A
Z A 1
A 1
XF;j;t = 4 (XF;i;j;t ) A di5 (40)
N

The di¤erent varieties are purchased by the …nal goods producers in proportions which min-
imise the cost of producing one unit of output Yj;t . The total demand for domestic intermediate
inputs by the …nal goods producer j is therefore given by:

A
1 QH;t
XH;j;t = Xj;t (41)
2 N Qt
The total demand for the input provided by foreign intermediate goods producers to the …nal
goods producer j is equal to:

A
1 N QF;t
XF;j;t = Xj;t (42)
2 N Qt
where Xj;t is the demand for intermediate inputs by the …nal goods producer j which is equal to
the quantity required to produce the pro…t-maximising level of output, Yj;t . QH;t and QF;t are
the aggregate price indices for the Home and Foreign intermediate goods, respectively. They are
equal to the minimum expenditure required to buy one unit of the composite intermediate good
XH;t and XF;t given the prices of the di¤erent varieties of the intermediate inputs.
The pro…ts of the …nal goods producer j are given by:

Qt Wt
j;t = Pj;t Cj;t Cj;t FB (43)
Bj;t Zt
The price of the …nal consumption good produced by …rm j at time t, Pj;t , is set optimally,
as a mark-up over marginal cost:

B Qt
Pj;t = (44)
B 1 Bj;t

856
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

where B is the elasticity of substitution between …nal consumption goods and Qt is the aggregate
price index for intermediate goods, equal to the minimal cost of a unit of the aggregate input.
The demand for the variety of …nal consumption good produced by …rm j, denoted by Cj;t ,
is equal to the supply of this variety, Yj;t , and is given by:

B
Pj;t
Yj;t = Cj;t = Ct (45)
Pt
where Ct is the aggregate demand for the …nal consumption good in the domestic economy.

Productivity distribution and non-payment risk As with intermediate goods producers,

it is assumed that in the …nal consumption goods sector the …rm-speci…c level of productivity,
Bt , is distributed Pareto with lower bound Bmin and shape parameter kB > B 1 where
Bmin k B
F (B) = 1 B
.
The average productivity among all …nal goods producers can now be de…ned as:
2 3 1
Z1 B 1

BA = 4 Bt B 1
dF (B)5 = vB Bmin (46)
Bmin

h i 11
where vB = kB k( BB 1) B .
Due to the presence of …xed costs of production and …rms’heterogeneity with regard to their
productivity, there is a fraction of …rms in each period whose revenues from sales are smaller than
the total production costs and who are therefore not able to ful…l all their …nancial obligations.
If at the end of the period revenues from sales exceed a …rm’s total cost, then the …rm repays all
trade credit due and keeps the remaining pro…ts, which are then passed on to households in the
form of dividends. If total costs exceed revenues then the …rm declares bankruptcy and receives
nothing. The trade credit suppliers receive the …rms’ revenues net of wages and the advance
payment for the intermediate goods made at the beginning of the period.
For any given realisation of aggregate shocks, there is a threshold level of …rm-speci…c pro-
ductivity B t above which …rms are able to pay all their costs of production, including the cost
of intermediate inputs and the …xed production cost. This threshold level of productivity is
determined by the condition:

1 B 1 Wt
Qt B t BA B Ct = FB (47)
B 1 Zt
For the purpose of calculating the trade credit default rate, it is useful to de…ne a thresh-
old level of …rm-speci…c productivity B t above which …rms are able to pay all their …nancial
obligations apart from trade credit. This threshold productivity level satis…es the condition:

857
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Wt B 1 B
FB = Bt Ct BA;tB Qt A Q1t A
(48)
Zt B 1
B 1 1 N 1
Bt Ct BA;tB Qt A (1 dF ) Q1F;t A
+ (1 dH ) Q1H;t A
2 N 2 N

The default rate on trade, t , which is de…ned as the fraction of the total trade credit which
is not repaid, is then equal to:

ZB t
[(dH XH;j;t QH;t + dF XF;j;t QF;t ] dF (B)
Bmin
t = Z1 (49)
[dH XH;j;t QH;t + dF XF;j;t QF;t ] dF (B)
Bmin

ZB t h i
Yj;t Pj;t (1 dH )XH;j;t QH;t (1 dF )XF;j;t QF;t FB W
Zt
t
dF (B)
Bt
Z1
[dH XH;j;t QH;t + dF XF;j;t QF;t ] dF (B)
Bmin

Given the productivity distribution among …nal goods producers, the probability that at time
t a …rm is able to meet all its …nancial obligations and earn non-negative pro…ts is equal to:

kB
Bmin
1 MH;t = 1 F (B t ) = (50)
Bt
Similarly, the probability that a …rm is able to pay for all of its costs of production apart
from the trade credit is given by:
!kB
Bmin
1 ML;t = 1 F (B t ) = (51)
Bt
In consequence, MH;t is the probability that a …rm defaults on at least a fraction of the trade
credit received and ML;t is the probability that a …rm defaults on the entire value of the trade
credit granted to it.
It is now possible to de…ne the following productivity averages. The average productivity
among non-defaulting …nal goods producers which earn non-negative pro…ts after paying all
their …nancial obligations including trade credit is equal to:

858
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2 3 1
B 1
Z1
6 1 1 7
BHH;t = 4 Bt dF (B)5 = vB B t (52)
1 F (B t )
Bt

The average productivity among defaulting …nal goods producers which default on at least a
fraction of their trade credit is given by:

2 3 1
B 1
ZB t 1
6 1 1 7 1 (1 MH;t ) 1 B 1
B B 1
B
BHL;t =4 Bt B
dF (B)5 = vB Bt (53)
F (B t ) MH;t min MH;t
Bmin

The average productivity among …nal goods producers which earn non-negative pro…ts after
paying all their …nancial obligations apart from the trade credit is equal to:
2 3 1
B 1
Z1
6 1 1 7
BLH;t = 4 Bt B dF (B)5 = vB B t (54)
1 F (B t )
Bt

The average productivity among …nal goods producers which default on the entire value of
the trade credit received is given by:

2 3 1
B 1
ZB t 1
6 1 1 7 1 B 1
(1 ML;t ) B 1 B 1
BLL;t =4 Bt B
dF (B)5 = vB Bmin Bt (55)
F (B t ) ML;t ML;t
Bmin

3.3 Aggregation
Given the assumptions concerning …rms’productivity distribution, it is now possible to express all
the aggregate variables and the equilibrium conditions of the model in terms of the productivity
averages for the di¤erent groups of intermediate and …nal goods producers which are de…ned
above.
Using the productivity averages de…ned for …rms producing intermediate goods, the aggregate
price index for these goods produced and sold in the Home economy is equal to:

A Wt
QH;t = gH;t (56)
A 1 Z t AA
Similarly, the price index for domestically produced intermediate inputs which are sold in the
Foreign market is given by:

859
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

A Wt 1
QF;t = (1 + ) gF;t (57)
A 1 Zt AE;t St
The aggregate supply of the domestically produced intermediate goods sold on the domestic
market, VH;t , can be expressed as a function of the average level of productivity in the Home
market, equal to Zt AA , and the aggregate labour supply used in the production of these goods,
LH;t :

VH;t = Zt AA LH;t (58)

where:

Z1
LH;t = (LH;i;t ) di (59)
0

In turn, the aggregate supply of the domestically produced intermediate goods which are ex-
ported, VF;t , can be expressed as a function of the average level of productivity among domestic
exporters, equal to Zt AE , and the aggregate labour supply used in the production of these goods,
LF;t :
VF;t = Zt AE;t LF;t (60)

where:

Z1
LF;t = (LF;i;t ) di (61)
0

Using the productivity averages for the producers of …nal consumption goods, the aggregate
supply of the …nal consumption good in the Home economy, Yt , can be expressed as:

Yt = BA Xt (62)

The aggregate price index for the …nal consumer goods in the Home economy is equal to:

B Qt
Pt = (63)
B 1 BA
The aggregate demand for intermediate goods in the Home economy, Xt , can be expressed
as a function of the aggregate demand for domestically produced intermediate goods, XH;t , and
the aggregate demand for the intermediate goods which are imported, XF;t :

" 1 1
# A
A 1 A 1 A 1
1 A 1 N A
Xt = XH;tA
+ XF;tA
(64)
2 N 2 N

860
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

where:

2 1 3 A
Z A 1
A
4
A 1 1 QH;t
XH;t = (XH;i;t ) A di5 = Xt (65)
2 N Qt
0

2 1 3 A
Z A 1
A
A 1 1 N QF;t
XF;t = 4 (XF;i;t ) A di5 = Xt (66)
2 N Qt
N

Using the de…nitions of average productivity levels among di¤erent categories of …nal goods
producers, the trade credit default rate can be expressed as:

B 1
MH;t BHL;t + Bt B t FB W t 1
B B
Zt Qt Ct A;t
t = h i (67)
dH 2 1N Q1H;t A + dF 12 N
N
1
Q1 A
F;t
B 1
BA;t Qt A
h i
1 A 1 A
1 1 N
(1 dH ) 2 N QH;t + (1 dF ) 2 N QF;t Qt A 1 ML;t BLL;t
B 1

h i
1 A 1 A B 1
1 1 N
dH 2 N QH;t + dF 2 N QF;t BA;t Qt A 1
B 1 B 1
B
B 1
MH;t BHL;t ML;t BLL;t
h i
dH 1
2 N
Q1H;t A + dF 1 N
2 N
B 1
Q1F;t A BA;t Qt A 1

3.4 Resource constraints

Equilibrium in the goods market requires that in each country the supply of intermediate goods
produced for the domestic and foreign market is equal to the demand for these goods in the
Home and Foreign economy: VH;t = XH;t and VF;t = XF;t . Similarly, the supply needs to be
equal to demand in the sector of …nal consumption goods: Yt = Ct .
Labour market equilibrium requires that in each country the labour employed in the produc-
tion of intermediate goods and the labour employed to cover the …xed costs of producing …nal
consumption goods and exporting is equal to the domestic labour supply:

1
Lt = LH;t + LF;t + [FB + (1 Nt ) FA ] (68)
Zt
The model is closed by normalising the Home price index of the …nal consumption goods to
one: Pt = 1 and by setting the nominal exchange rate equal to one: St = 1.15
15
A complete set of the model’s equilibrium conditions for the Home economy are presented in Appendix A.1.

861
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4 Parametrisation
The model is calibrated assuming that one period of time corresponds to one quarter. The
calibration of preference parameters determining households’utility function follows the interna-
tional business cycle literature. Both the inverse of the intertemporal elasticity of substitution in
consumption, ; and the inverse of the Frisch elasticity of labour supply, ', are set equal to 2:0.
The discount factor is set to = 0:99. The elasticity of substitution between di¤erent varieties
in both the sectors of intermediate goods and …nal consumption goods is equal to A = B = 6 to
give a mark-up of 20 per cent, as in Obstfeld and Rogo¤ (2001). Following Obstfeld and Rogo¤
(2001), Ravn and Mazzenga (2004) and Anderson and van Wincoop (2004), the iceberg trade
cost is parametrised to be equal to 20 per cent of the value of goods exported.
It is assumed that the distribution of …rm-speci…c productivity is the same in both the sector
for intermediate goods and the sector for …nal consumption goods. The parameters kA and kB
governing the dispersion of productivity levels among …rms are set equal to kA = kB = 6:25,
which is in line with empirical evidence on the distribution of …rms’ size (Axtell, 2001). The
lower bound for idiosyncratic productivity is set equal to one, Amin = 1. The …xed production
costs in the sector of …nal consumption goods, FB , are calibrated to match the default rate on
accounts payables/receivables, which Boissay and Gropp (2007) estimate to be equal to 2 per
cent, based on French …rm-level data. The …xed cost of exporting, FA , is set to match a share
of exporters among the intermediate goods producers equal to 35 per cent, which is the average
share of exporters among the top 15,000 manufacturing …rms in the UK and Ireland in the period
2003-2014, based on Fame data.
The fractions of trade credit to total sales revenues granted to domestic and foreign buyers
are set equal to dH = 0:55 and dF = 0:75 respectively, to match the empirical evidence for UK
and Irish …rms discussed in the second section of this paper. Among the top 1,000 of British
and Irish manufacturing …rms, which cover 82 per cent of total manufacturing revenues and
therefore re‡ect well the use of trade credit by …rms which determine aggregate macroeconomic
‡uctuations, the average ratios of trade debtors to revenues among exporters and non-exporters
in the period from 2003 to 2014 were equal to 12.6 and 9.1 per cent respectively. Assuming that
the average duration of trade credit is 2 months, as demonstrated by Klapper et al. (2012), non-
exporting …rms sell 55 per cent of their goods on trade credit, while …rms engaged in international
trade extend trade credit for 76 per cent of their sales.16 The values of all the model parameters
adopted in the analysis are listed in Table 3 in Appendix A.2.
16
In the same time period, the average ratios of trade debtors to revenues for exporters and non-exporters
among the top 5,000 of all …rms, which cover about 87 per cent of total revenues, were very similar and were
equal to 12.7 and 9.2 per cent respectively, which also corresponds to the fractions of goods sold on trade credit
of 76 and 55 per cent respectively

862
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5 Results
This section examines the impact of trade credit on business cycle ‡uctuations by tracing the
responses of key macroeconomic variables to demand and productivity shocks. In order to
illustrate the way in which trade credit alters the transmission of shocks in the economy, three
di¤erent versions of the model developed in Section 3 are compared: the benchmark (BN) model
in which there is no inter-…rm lending and in which the fraction of internationally traded varieties
is constant over time, the extensive margin (EM) model which extends the benchmark model
by endogenising the decision to export, and the trade credit (TC) model, which features both
inter-…rm lending and the extensive margin of trade.17 The comparison of the di¤erent versions
of the model makes it possible to examine the contributions of trade credit and of the extensive
margin of trade to changes in the volume of international trade in response to shocks.

5.0.1 Symmetric shocks

In order to understand the impact of trade credit and the associated counterparty risk on the
‡uctuations of international trade relative to output, it is useful to examine the e¤ects of macro-
economic shocks which are symmetric across countries.

Productivity shocks Figure 1 compares the e¤ects of a one per cent positive symmetric
productivity shock in the intermediate goods sector in the BN model, the EM model and the
TC model. It shows that the use of trade credit by …rms can lead to signi…cant di¤erences
in the impact of macroeconomic shocks on international trade. A favourable shock raising the
productivity level of all intermediate goods producers in the world economy leads to an increase
in output and consumption in all three models analysed. The increase in output is to some extent
dampened by a reduction in labour supply which results from an increase in real wages and the
fact that as the intertemporal elasticity of substitution in consumption is smaller than one, the
income e¤ect dominates the substitution e¤ect. In the EM and TC models in which the decision
to export is endogenous, the increase in the aggregate level of productivity in the economy reduces
the threshold idiosyncratic productivity level above which exporting is pro…table and, as a result,
the share of exporting …rms among the intermediate goods producers increases. This increase
and the resulting rise in trade along the extensive margin is greater in the presence of inter-
…rm lending, which is due to the changes in relative prices of intermediate inputs of domestic
and foreign origin following the shock and the associated changes in the relative demand for
intermediate inputs produced domestically and abroad.
17
The TC model nests both the EM and the BN models. In the EM model, the fraction of sales for which trade
credit is granted is set equal to zero: dH = dF = 0, as is the trade credit default rate: = 0. In the BN model,
in addition to the above restrictions, the fraction of non-exporting …rms is set to N = 0:65 and is constant over
time. The parameterisation of all three versions of the model used in the analysis is presented in Appendix A.2.

863
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1: Impulse responses to a 1 per cent positive symmetric productivity shock

in the sector producing intermediate goods

Output Labour supply Wage

1 0 2
% deviation

% deviation

% deviation
0.5 -0.2 1

0 -0.4 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Trade credit default rate Export profitability threshold Fraction of exporters

0 0 2
% deviation

% deviation

% deviation
-10 -0.1 1

-20 -0.2 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Average productivity - exporters Average productivity - non-exporters Imported intermediate goods price index
0 0 0.2

% deviation

% deviation
% deviation

-0.1 -0.05 0.1

-0.2 -0.1 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Domestic intermediate goods price index Relative price of imported and domestic goods International trade
0.05 0.2 2
% deviation

% deviation
% deviation

quarters
0 0.1 1

-0.05 0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Note: BN model –dashed line, EM model –dotted line, TC model –solid line

The behaviour of the relative prices of intermediate goods sold in the domestic and foreign
market di¤ers substantially in the three models analysed. In the BN model, the average price
of these goods remains constant regardless of their destination market, as the increase in the
marginal product of labour is o¤set by an equal increase in real wages. In both the EM and
the TC models the productivity shock is accompanied by an increase in the average price of
intermediate inputs sold in the foreign market relative to the average price of the intermediate
inputs sold domestically. This is due to the fact that the decline in the export pro…tability
threshold, At , and the increase in the share of exporting …rms following the rise in aggregate
productivity reduce the average level of idiosyncratic productivity among exporters, AE;t , by
more than they reduce the average …rm-speci…c productivity level among non-exporters, AN;t .
In the TC model the changes in the average prices of intermediate goods sold at home and abroad
are smaller than in the EM model due to the changes in the cost of trade credit. A favourable
productivity shock increases output and …rms’ pro…ts and reduces the default rate among the
producers of …nal consumption goods. The decline in the risk of non-repayment associated with
trade credit leads to a decrease in the marginal costs and in the prices of the producers of
intermediate inputs. As trade credit is more intensively used in international than in domestic
transactions, the fall in the marginal cost is relatively larger for goods sold in the foreign market.
As a result, in the presence of inter-…rm lending the relative price of intermediate inputs sold
abroad increases by less as compared to the price of goods sold domestically.
Changes in aggregate demand, the share of exporting …rms in the economy and the relative
prices of exported and non-exported intermediate goods which follow the productivity shock

864
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

lead to changes in the volume of international trade. The increase in trade caused by a rise in
aggregate productivity is smallest in the benchmark model (BN) and largest in the model with
trade credit (TC).
In the BN model, in which the share of exporting …rms is constant and in which relative
prices of goods sold domestically and abroad are una¤ected by the productivity shock, the only
factor in‡uencing international trade is aggregate demand. As a result the percentage increase
in trade following the productivity shock is the same as the percentage increase in output and
the trade income elasticity is equal to one.18
In the EM model, the volume of trade is not only a¤ected by changes in aggregate demand
but also by changes in the fraction of intermediate goods producers who export their goods and
the resulting changes in the composition of the consumption basket of …nal goods producers who
purchase all varieties of intermediate inputs available for sale in their economy. Following the
shock, the number of foreign varieties available increases in proportion to the number of varieties
produced domestically. As a result, the share of imported goods in the consumption basket
increases for any given relative price of domestic and imported goods –there is an endogenous
decline in the home bias of …nal goods producers. While there is an increase in the relative
average price of intermediate inputs of foreign and domestic origin, this is entirely due to the
fact that the new foreign varieties available for sale in the domestic market have a relatively
higher average price than the foreign varieties traded initially. For any variety sold in both the
domestic and foreign market, its relative price on these markets remains unchanged. As a result,
the increase in the number of imported varieties leads to an increase in the share of imported
intermediate inputs in the production of …nal consumption goods and an increase in international
trade which exceeds the rise in domestic output. The elasticity of real trade to real income in
this model is equal to 1.18.
In the TC model, the impact of the productivity shock on trade is even larger due to changes
in the relative prices of domestic and foreign varieties resulting from changes in the relative
riskiness of domestic and international transactions following the shock. In the TC model, an
increase in aggregate productivity is not only followed by an increase in the number of exported
varieties, which is even larger than in the EM model, but it is also accompanied by an increase
in the domestic price of each traded variety as compared to its price on the foreign market. This
is due to the decline in the cost of trade credit associated with international transactions relative
to domestic transactions. As a result, the increase in the relative price of imported to domestic
goods in the TC model is smaller than in the EM model and there is an even greater increase
in the demand for intermediate inputs produced abroad relative to the demand for intermediate
inputs of domestic origin. The increase in international trade in the TC model is considerably
18
The trade income elasticity is calculated as the percentage deviation of real trade volume from its steady
state divided by the percentage deviation of real world output from its steady state. It is the same for each period
following the shock.

865
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

higher than in models without inter-…rm lending –the trade income elasticity is equal to 1.47.

Demand shocks The impact of a one per cent positive symmetric demand shock on interna-
tional trade, which is illustrated in Figure 2, is very similar to that of a favourable productivity
shock. An exogenous increase in the marginal utility of consumption for any given consump-
tion level leads to an increase in aggregate demand and output in the economy. Labour supply
increases as the marginal utility of leisure declines for any level of work e¤ort. In the EM and
TC models which account for the extensive margin of trade, the increase in aggregate demand
induced by the shock raises the pro…tability of exporting and the share of exporters among
intermediate goods producers. Changes in wages are small and di¤er across the three models
considered. Pro…t maximisation on the part of intermediate goods producers requires that the
real wage is equal to the real marginal revenue product of labour. In the BM model wage rate
is constant as the marginal revenue product of labour does not change. In the EM model there
is a slight decrease in the real wage as an increase in the percentage of internationally traded
varieties leads to a decline in the average price charged by intermediate goods producers. In
the TC model the marginal revenue product of labour depends additionally on the trade credit
default rate, which declines following the shock leading to an increase in the wage rate.

Figure 2: Impulse responses to a 1 per cent positive symmetric demand shock

Output Labour supply Wage

0.4 0.4 0.05
% deviation

% deviation

0.2 0.2 0

0 0 -0.05
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Trade credit default rate Export profitability threshold Fraction of exporters

0 0 0.4
% deviation

% deviation

-2
-0.05 0.2
-4
-0.1 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Average productivity - non-exporters Imported intermediate goods price index
Average productivity - exporters 0 0.04
0
% deviation

% deviation
% deviation

-0.05 -0.02 0.02

-0.1 -0.04 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Domestic intermediate goods price index Relative price of imported and domestic goods International trade
0.02 0.1 0.5
% deviation
% deviation

% deviation

0 0.05

-0.02 0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Note: BN model –dashed line, EM model –dotted line, TC model –solid line

The demand shock causes similar changes in the relative prices of imported and domestic
varieties as well as their weights in the consumption aggregator of the …nal goods producers as a
favourable productivity shock. It results in an increase in international trade which signi…cantly
exceeds that of output. As in the case of the productivity shock, the trade income elasticities
generated by the BN, EM and TC models are equal to 1.00, 1.18 and 1.46, respectively.

866
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

5.0.2 Asymmetric shocks

The analysis in the previous section showed that trade credit ampli…es the e¤ects of macro-
economic shocks on international trade along the intensive and extensive margin even if shocks
are symmetric across countries and have no impact on international relative prices – the real
exchange rate and the terms of trade. Inter-…rm lending also alters the e¤ects of macroeconomic
shocks when they are country-speci…c. In the case of asymmetric shocks, changes in macroeco-
nomic conditions and the resulting changes in the counterparty risk associated with inter-…rm
loans a¤ect not only the relative domestic and foreign market price of traded varieties but also
the relative export price of intermediate inputs produced at home and abroad.

Productivity shocks Figure 3 shows the e¤ects of a one per cent positive productivity shock
in the sector producing intermediate goods in the Home economy. Through the risk sharing
mechanism the productivity shock leads to an increase in consumption in both countries in all
three versions of the model considered. The rise in consumption is higher in the Home than
in the Foreign economy. The e¤ect of the shock on labour supply in the Foreign economy is
unambiguously negative whereas in the Home economy it di¤ers across the models, re‡ecting
di¤erences in the relative strength of the income and substitution e¤ects.

Figure 3: Impulse responses to a 1 per cent positive productivity shock in the sector
producing intermediate goods in the Home economy

Output (H) Labour supply (H) Trade credit default rate (H)
1 0.05 0
% deviation

% deviation

0.5 0 -5

0 -0.05 -10
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Output (F) Labour supply (F) Trade credit default rate (F)
0.4 0 0
% deviation
% deviation

% deviation

0.2 -0.2 -5

0 -0.4 -10
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Fraction of exporters (H) Relative export price (H) Terms of trade
4 0.4 0.5
% deviation

% deviation

2 0.2 0

0 0 -0.5
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Fraction of exporters (F) Relative export price (F) International trade

0 0.5 1
% deviation

% deviation

-1 0 0.5

-2 -0.5 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Note: BN model –dashed line, EM model –dotted line, TC model –solid line; H –Home economy, F –Foreign
economy

The impact of the shock on relative prices also varies signi…cantly across the three models.
In the BN model, an increase in productivity in the Home economy lowers the relative price
of traded goods produced in this economy and leads to a depreciation of the terms of trade

867
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and an increase in exports. However, the gain of the market share of Home exporters in the
Foreign economy is accompanied by a loss in the market share of Foreign exporters in the Home
economy and, as a result, the increase in international trade following the shock is the same as
the increase in output in the global economy – the elasticity of world trade to world income is
equal to one. In the EM and TC models, the productivity shock leads to changes in the number
of traded varieties and movements in international trade along the extensive margin. In the
Home economy, the share of exporters increases due to both an increase in aggregate demand in
the Foreign economy and an increase in the competitiveness of Home producers on the Foreign
market. In the Foreign economy the positive impact that an increase in aggregate demand in
the Home economy exerts on exports pro…tability is more than o¤set by a decline in the relative
demand for Foreign versus Home goods and there is a decline in the number of Foreign varieties
which are exported. Due to the resulting changes in the average productivity of exporters in
both economies, the average price of Home exports increases as compared to the average price
of Home imports, leading to an appreciation of the terms of trade in the Home economy.19
However, changes in the relative number and price of domestic and foreign intermediate goods
in both countries lead to an increase in the share of imported inputs in the consumption basket
of …nal goods producers in the world economy. In consequence, the increase in trade following
the shock is about 1.13 times larger than the increase in aggregate demand. In the TC model,
the increase in the share of imported inputs in the production of …nal consumption goods in the
world economy is even larger due to the fact that the favourable productivity shock reduces the
cost of trade credit in both economies, which additionally improves the pro…tability of exporting
and stimulates international trade, generating trade income elasticity equal to 1.34.

Demand shocks A one per cent favourable demand shock in the Home economy, which is
illustrated in Figure 4, leads to an increase in consumption in the country a¤ected by the shock
and a decline in consumption abroad. Labour supply increases in both countries. In the Home
economy this is due to the exogenous increase in marginal utility of consumption for any con-
sumption level. In the Foreign economy it results from the increase in the marginal utility of
consumption associated with the fall in the level of income. In the BN model, an increase in
the relative consumption at home and abroad leads to an appreciation of the terms of trade in
the Home economy. However, as in the case of the productivity shock, the share of imported
inputs in the consumption of …nal goods producers is constant as there are no changes in the
number of traded varieties or the relative domestic and foreign market price of these varieties.
As a result, the percentage increase in world trade following the shock is once again the same
as the percentage increase in world GDP. In the EM model, the shock reduces the fraction of
19
This result is in line with the …ndings of a number of studies showing that in models accounting for the
extensive margin of trade an increase in a country’s productivity can have a positive impact on the country’s
terms of trade (Krugman, 1989; Ghironi and Melitz, 2005; Corsetti et al., 2007). It is also consistent with
empirical evidence (Corsetti et al., 2008).

868
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 4: Impulse responses to a 1 per cent positive demand shock in the Home
economy

Consumption (H) Labour supply (H) Trade credit default rate (H)
0.5 0.2 0
% deviation

% deviation

% deviation
0.1 -5

0 0
0 5 10 15 20 25 30 -10
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Consumption (F) Labour supply (F) Trade credit default rate (F)
0 0.2 4
% deviation

% deviation

% deviation
-0.1 0.1 2

-0.2 0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Fraction of exporters (H) Relative export price (H) Terms of trade
0 0.1 0.2
% deviation

% deviation

% deviation
-0.2 0 0

-0.4 -0.1 -0.2

0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Fraction of exporters (F) Relative export price (F) International trade
1 0.2 0.4

% deviation
% deviation

% deviation

0.5 0 0.2

0 -0.2 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
quarters quarters quarters

Note: BN model –dashed line, EM model –dotted line, TC model –solid line; H –Home economy, F –Foreign
economy

exporters among intermediate goods producers in the Home economy due to the decline in aggre-
gate demand abroad. However, the share of exporters among Foreign …rms increases. Due to the
resulting changes in the average productivity of exporters in both countries, the Home economy
terms of trade depreciate. The shock leads to an increase in international trade which is 1.24
times larger than that of GDP. In the TC model, the demand shock causes diverging changes in
the default rates among …rms in the Home and Foreign economies; the latter increases while the
former declines. However, as the average trade credit default rate in the world economy falls, the
overall impact of the shock on international trade is positive and greater than in the EM model,
with the trade income elasticity reaching 1.4.

6 Conclusions
This study analysed the impact of trade credit on the transmission of shocks in an open economy
and, in particular, on the dynamics of international trade. Using British and Irish …rm-level data,
it showed that an overwhelming majority of …rms makes an extensive use of this type of …nancing.
It also demonstrated that …rms engaged in international trade use trade credit more actively and
more intensively than …rms serving only the domestic market. The percentage of …rms supplying
and receiving trade credit is higher among exporters than non-exporters. Firms exporting their
goods also have a signi…cantly higher ratio of trade debtors to sales revenues. The pattern of a
more extensive trade credit use by …rms participating in international trade prevails not only in

869
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the manufacturing sector, which is responsible for a major share of international trade, but also
in the entire population of …rms in the economy. It is observed among both large and small …rms.
Given the importance of inter-…rm lending in day to day business operations, the objective of this
study was to examine whether and to what extent trade credit contributes to the high volatility
of imports and exports observed in the data and the high sensitivity of international trade to
macroeconomic shocks, which is di¢ cult to reconcile with standard international business cycle
models.
The study introduced inter-…rm lending into an open economy general equilibrium model with
heterogeneous …rms in which the decision to enter the exports market is endogenous. The analysis
showed that trade credit ampli…es the impact of supply and demand shocks on trade both along
the intensive and extensive margin. This is the case both for symmetric and country-speci…c
shocks. Trade credit in‡uences the transmission of shocks through the marginal cost channel.
Due to the fact that international transactions are associated with greater counterparty risk than
domestic transactions, the foreign market price of traded varieties is more sensitive to changes
in macroeconomic conditions than their domestic market price, which leads to changes in the
relative demand for domestic and foreign goods over the business cycle. The study demonstrated
that trade credit has considerable impact on the propagation of shocks in an open economy. The
elasticities of real trade to real income generated by the model accounting for both inter-…rm
lending and the extensive margin of trade range between 1.3-1.5, depending on the type of shock
considered. They are much higher than the elasticities implied by standard international business
cycle models and models accounting for the extensive margin of trade but not for inter-…rm loans.

870
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References
[1] Ahn, J. (2011): A Theory of Domestic and International Trade Finance, IMF Working
Paper No. 11/262.
[2] Alessandria, G., J. Kaboski, V. Midrigan (2010): The Great Trade Collapse of 2008-09: An
Inventory Adjustment?, IMF Economic Review, Vol. 58(2), pp. 254-294.
[3] Anderson, J., E. van Wincoop (2004): Trade Costs, Journal of Economic Literature, Vol.
42(3), pp. 691-751.
[4] Antràs, P., C. F. Foley (2015): Poultry in Motion: A Study of International Trade Finance
Practices, Journal of Political Economy, Vol. 123(4), pp. 853-901.
[5] Asmundson, I., T. Dorsey, A. Khachatryan, I. Niculcea, M. Saito (2011): Trade and Trade
Finance in the 2008-09 Financial Crisis, IMF Working Paper No. WP/11/16.
[6] Atanasova, C., N. Wilson (2003): Bank Borrowing Constraints and the Demand for Trade
Credit: Evidence from Panel Data, Managerial and Decision Economics, Vol. 24(6-7), pp.
503-514.
[7] Auboin, M. (2009): Restoring Trade Finance During a Period of Financial Crisis: Stock-
Taking of Recent Initiatives, WTO Sta¤ Working Paper No. ERSD-2009-16.
[8] Axtell, R. (2001): Zipf Distribution of U.S. Firm Sizes, Science, Vol. 293(5536), pp. 1818-
1820.
[9] Bellone, F., P. Musso, L. Nesta, S. Schiavo (2010): Financial Constraints and Firm Export
Behaviour, The World Economy, Vol. 33(3), pp. 347-373.
[10] Bems, R., R. Johnson, K. Yi (2010): The Role of Vertical Linkages in the Propagation of
the Global Downturn of 2008, IMF Economic Review, Vol. 58(2), pp. 295-326.
[11] Bernard, A., J. Jensen (1995): Exporters, Jobs and Wages in U.S. Manufacturing, 1976-
1987, Brookings Papers on Economic Activity, Microeconomics, Vol. 1995, pp. 67-119.
[12] Bernard, A., J. Jensen (1997): Exporters Skill-Upgrading and the Wage Gap, Journal of
International Economics, Vol. 42(1), pp 3-31.
[13] Bernard, A., J. Jensen (1999): Exceptional Exporter Performance: Cause, E¤ect or Both?,
Journal of International Economics, Vol. 47(1), pp. 1-25.
[14] Bernard, A., J. Jensen (2004): Why Some Firms Export, Review of Economics and Statistics,
Vol. 86(2), pp. 561-569.
[15] Boissay, F. (2006): Credit Chains and the Propagation of Financial Distress, ECB Working
Paper No. 573.
[16] Boissay, F., R. Gropp (2007): Trade Credit Defaults and Liquidity Provision by Firms, ECB
Working Paper No. 753.
[17] Bradley, D., M. Rubach (2002): Trade Credit and Small Business: A Cause of Business
Failure?, Technical Report, University of Central Arkansas.

871
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[18] Bradley, D., C. Cowdery (2004): Small Business: Causes of Bankruptcy, Technical Report,
University of Central Arkansas.

[19] Buch, C., I. Kesternich, A. Lipponer, M. Schnitzer (2010): Exports vs. FDI Revisited: Does
Finance Matter?, CEPR Discussion Paper No. 7839.

[20] Bussière, M., G. Callegari, F. Ghironi, G. Sestieri, N. Yamano (2013): Estimating Trade
Elasticities: Demand Composition and the Trade Collapse of 2008-09, American Economic
Journal: Macroeconomics, Vol. 5(3), pp. 118-151.

[21] Calomiris, C., C. Himmelberg, P. Wachtel (1995): Commercial Paper, Corporate Finance,
and the Business Cycle: A Microeconomic Perspective, Carnegie-Rochester Conference Se-
ries on Public Policy, Vol. 42, pp. 203-50.

[22] Cardoso-Lecourtois, M. (2004): Chain Reactions, Trade Credit and the Business Cycle,
Econometric Society 2004 North American Summer Meetings, No. 331.

[23] Chaney, T. (2016): Liquidity Constrained Exporters, Journal of Economic Dynamics and
Control, Vol. 72(C), pp. 141-154.

[24] Choi, W., Y. Kim (2005): Trade Credit and the E¤ect of Macro-Financial Shocks: Evidence
from U.S. Panel Data, Journal of Financial and Quantitative Analysis, Vol. 40(4), pp. 897-
925.

[25] Corsetti G., L. Dedola, S. Leduc (2008): International Risk Sharing and the Transmission
of Productivity Shocks, Review of Economic Studies, Vol. 75(2), pp. 443-473.

[26] Corsetti, G., P. Martin, P. Pesenti (2007): Productivity, Terms of Trade and the ‘Home
Market E¤ect’, Journal of International Economics, Vol. 73(1), pp. 99-127.

[27] Cuñat, V. (2007): Trade Credit: Suppliers as Debt Collectors and Insurance Providers,
Review of Financial Studies, Vol. 20(2), pp. 491-527.

[28] Dass, N., J. Kale, V. Nanda (2015): Trade Credit, Relationship-Speci…c Investment, and
Product-Market Power, Review of Finance, Vol. 19(5), pp. 1867-1923.

[29] Demirguc-Kunt, A., V. Maksimovic (2001): Firms as Financial Intermediaries – Evidence

from Trade Credit Data, World Bank Policy Research Working Paper No. 2696.

[30] Eaton, J., S. Kortum, B. Neimann, J. Romalis (2016): Trade and the Global Recession,
American Economic Review, Vol. 106(11), pp. 3401-38.

[31] Eck, K., M. Engemann, M. Schnitzer (2012): How Trade Credits Foster International Trade,
CEPR Discussion Papers No. 8954.

[32] Engel, C., J. Wang (2011): International Trade in Durable Goods: Understanding Volatility,
Cyclicality and Elasticities, Journal of International Economics, Vol. 83(1), pp. 37-52.

[33] Freund, C. (2009): The Trade Response to Global Downturns: Historical Evidence, World
Bank Policy Research Working Paper No. 5015.

872
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[34] Garcia-Appendini, E., J. Montoriol-Garriga (2013): Firms as Liquidity Providers: Evidence

from the 2007-2008 Financial Crisis, Journal of Financial Economics, Vol. 109(1), pp. 272-
291.

[35] Guariglia, A., S. Mateut (2006): Credit Channel, Trade Credit Channel, and Inventory
Investment: Evidence from a Panel of UK Firms, Journal of Banking and Finance, Vol.
30(10), pp. 2835-2856.

[36] Ghironi, F., M. Melitz (2005): International Trade and Macroeconomic Dynamics with
Heterogeneous Firms, Quarterly Journal of Economics, Vol. 120(3), pp. 865-915.

[37] Huang, H., X. Shib, S. Zhang (2011): Counter-cyclical Substitution between Trade Credit
and Bank Credit, Journal of Banking and Finance, Vol. 35(8), pp. 1859-1878.

[38] Irwin, D. (2002): Long-Run Trends in World Trade and Income, World Trade Review, Vol.
1(1), pp. 89-100.

[39] Jacobson, T., E. Schedvin (2015): Trade Credit and the Propagation of Corporate Failure:
An Empirical Analysis, Econometrica, Vol. 83(4), pp. 1315-1371.

[40] Kiyotaki, N., J. Moore (1997): Credit Chains, Edinburgh School of Economics Discussion
Paper No. 118.

[41] Klapper, L., L. Laeven, R. Rajan (2012): Trade Credit Contracts, Review of Financial
Studies, Vol. 25(3), pp. 838-867.

[42] Kohler, M., E. Britton, A. Yates (2000): Trade Credit and the Monetary Transmission
Mechanism, Bank of England Working Paper No. 115.

[43] Krugman, P. (1989): Di¤erences in Income and Trends in Real Exchange Rates, European
Economic Review, Vol. 33(5), pp. 1031-1054.

[44] Levchenko, A., L. Lewis, L. Tesar (2010): The Collapse of International Trade During the
2008-2009 Crisis: In Search of the Smoking Gun, IMF Economic Review, Vol. 58(2), pp.
214-253.

[45] Love, I., L. Preve, V. Sartia-Allende (2007): Trade Credit and Bank Credit: Evidence from
Recent Financial Crises, Journal of Financial Economics, Vol. 83(2), pp. 453-69.

[46] Manova, K. (2013): Credit Constraints, Heterogeneous Firms, and International Trade,
Review of Economic Studies, Vol. 80(2), pp. 711-744.

[47] Mateut, S., S. Bougheas, P. Mizen (2006): Trade Credit, Bank Lending and Monetary Policy
Transmission, European Economic Review, Vol. 50(3), pp. 603-629.

[48] Melitz, M. (2003): The Impact of Trade on Intra-Industry Reallocations and Aggregate
Industry Productivity, Econometrica, Vol. 71(6), pp. 1695-1725.

[49] Minetti, R., S. Chun Zhu (2011): Credit Constraints and Firm Export: Microeconomic
Evidence from Italy, Journal of International Economics, Vol. 83(2), pp. 109-125.

873
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[50] Muuls, M. (2015): Exporters, Importers and Credit Constraints, Journal of International
Economics, Vol. 95(2), pp. 333-343.

[51] Obstfeld, M., K. Rogo¤ (2001): The Six Major Puzzles in International Macroeconomics: Is
There a Common Cause?, NBER Chapters, in: NBER Macroeconomics Annual 2000, Vol.
15, pp. 339-412.

[52] Petersen, M., R. Rajan (1997): Trade Credit: Theory and Evidence, Review of Financial
Studies, Vol 10(3), pp. 661-691.

[53] Raddatz, C. (2010): Credit Chains and Sectoral Comovements: Does the use of Trade Credit
Amplify Sectoral Shocks?, Review of Economics and Statistics, Vol. 92(4), pp. 985-1003.

[54] Rajan, R., L. Zingales (1995): What Do We Know about Capital Structure? Some Evidence
from International Data, Journal of Finance, Vol. 50(5), pp.1421-60.

[55] Ravn, M., E. Mazzenga (2004): International Business Cycles: the Quantitative Role of
Transportation Costs, Journal of International Money and Finance, Vol. 23(4), pp. 645–
671.

[56] Redding, S. (2011): Theories of Heterogeneous Firms and Trade, Annual Review of Eco-
nomics, Vol. 3(1), pp. 77-105.

[57] Schmidt-Eisenlohr, T. (2013): Towards a Theory of Trade Finance, Journal of International

Economics, Vol. 91(1), pp. 96-112.

[58] Wilson, N., B. Summers (1997): Trade Credit Terms and the Motives for Trade Credit
Extension: Theory and Evidence, University of Bradford Management Centre Working
Paper No. 9708.

874
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

A Appendix

A.1 Flexible Price Equilibrium

This section lists the Home economy equilibrium conditions of the model.
Risk-sharing condition:

Ct S t Pt
e(ut ut )
= (69)
Ct Pt
Intratemporal labour equation:
Wt
= Ct L't e ut
(70)
Pt
Intermediate goods aggregator:
" 1 1
# A
A 1 A 1 A 1
1 A 1 N A
Xt = XH;t A
+ XF;tA
(71)
2 N 2 N
Aggregate intermediate good –domestic producers:
A
1 QH;t
XH;t = Xt (72)
2 N Qt
Aggregate intermediate good –foreign producers:
A
1 N QF;t
XF;t = Xt (73)
2 N Qt
Price index for intermediate goods produced and sold in the Home economy:
Wt A
QH;t = gH;t (74)
A 1 Z t AA
Price index for intermediate goods produced in the Home economy and sold in the Foreign
economy:

A Wt 1
QF;t = (1 + ) gF;t (75)
A 1 Zt AE;t St
Trade credit insurance cost for Home intermediate goods producers for the trade credit granted
to Home …nal goods producers:
1
gH;t = (76)
(1 dH ) + dH (1 t)
Trade credit insurance cost for Foreign intermediate goods producers for trade credit granted to
Home …nal goods producers:
1
gF;t = (77)
(1 dF ) + dF (1 t)

875
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Production function –domestic intermediate goods producers:

VH;t = Zt AA LH;t (78)

Production function –foreign intermediate goods producers:

VF;t = Zt AE;t LF;t (79)

Threshold level of productivity –exports market participation condition:
1
1 1 A gF;t A
Wt A

(1 + ) Xt (St Qt ) A
= FA (80)
A 2 Nt A 1 At Zt
Probability of exporting:
k
Amin A
Nt = 1 (81)
At
The average level of productivity among domestic intermediate goods producers:
1
kA A 1
AA = Amin (82)
kA ( A 1)
The average level of productivity among exporting intermediate goods producers:
1
kA A 1
AE;t = At (83)
kA ( A 1)
The average level of productivity among non-exporting intermediate goods producers:
1 1
kA A 1 1 A 1 (1 Nt ) A 1 A 1
AN;t = A At (84)
kA ( A 1) Nt min Nt
Production function in the sector of …nal goods producers:

Yt = BA Xt (85)
The threshold level of productivity above which …rms are able to pay all their …nancial obligations
including trade credit:
1 Wt B 1
Bt Ct Qt BA (86)
B
= FB
B 1 Zt
The threshold level of productivity above which …rms are able to pay all their …nancial obligations
apart from trade credit:

Wt B 1 B
FB = B j;t Ct BA;tB Qt A Q1 A
(87)
Zt B 1 t
B 1 1 N 1
B j;t Ct BA;tB Qt A (1 dF ) Q1F;t A
+ (1 dH ) Q1H;t A
2 N 2 N

The probability that a …rm will default on the entire value of trade credit received:

876
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

!kB
Bmin
ML;t = F (B t ) = 1 (88)
Bt
The probability that a …rm will default on at least a fraction of trade credit received:
k
Bmin B
MH;t = F (B t ) = 1 (89)
Bt
The average level of productivity among all …nal goods producers:
1
kB B 1
BA = Bmin (90)
kB ( B 1)
The average level of productivity among non-defaulting …nal goods producers which are able to
pay all their …nancial obligations:
1
kB B 1
BHH;t = Bt (91)
kB ( B 1)
The average level of productivity among …nal goods producers which default on at least a fraction
of trade credit received:
1 1
kB B 1 1 (1 MH;t ) 1 B 1
B B 1
B
BHL;t = Bt (92)
kB ( B 1) MH;t min MH;t
The average level of productivity among non-defaulting …nal goods producers which are able to
pay all their …nancial obligations apart from the trade credit:
1
kB B 1
BLH;t = Bt (93)
kB ( B 1)
The average level of productivity among …nal goods producers which default on the entire value
of trade credit received:
1 1
kB B 1 1 (1 ML;t ) B 1 B 1
BLL;t = B B 1 Bt (94)
kB ( B 1) ML;t min ML;t
Trade credit default rate among …nal goods producers:

B 1
MH;t BHL;t + Bt B t FB W t 1
B B
Zt Qt Ct A;t
t = h i (95)
dH 2 1N Q1H;t A + dF 12 N
N
Q1 A
F;t
1 B 1
BA;t Qt A
h i
1 A 1 A
1 1 N
(1 dH ) 2 N QH;t + (1 dF ) 2 N QF;t Qt A 1 ML;t BLL;t
B 1

h i
dH 2 1N Q1H;t A + dF 21 N
N
Q1F;t A BA;tB 1
Qt A 1
B 1 B 1
B
B
1
MH;t BHL;t ML;t BLL;t
h i
dH 1
2 N
Q1H;t A + dF 1 N
2 N
1 A
QF;t B 1
BA;t Qt A 1

877
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Labour market equilibrium condition:

1
[FB;t + (1 Nt ) FA;t ]
Lt = LH;t + LF;t + (96)
Zt
Resource constraint in the market for …nal consumption goods:

Yt = Ct (97)
Resource constraint in the market for intermediate goods:

VH;t = XH;t (98)

VF;t = XF;t (99)

Normalisation:
Pt = 1 (100)
St = 1 (101)
The complete model consists of all the above listed equilibrium conditions for the Home economy
and the Foreign-economy equivalents of these conditions.

878
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

A.2 Structural parameters

Table 3: Structural parameters

Parameter Description Value

TC EM BN

Inverse of the intertemporal elasticity 2.0 2.0 2.0

of substitution in consumption
' Inverse of the Frisch elasticity 2.0 2.0 2.0
of labour supply
Discount factor 0.99 0.99 0.99

A Elasticity of substitution 6.0 6.0 6.0

–intermediate goods
B Elasticity of substitution 6.0 6.0 6.0
–…nal consumption goods
Iceberg trade costs 0.2 0.2 0.2

kA Pareto distribution shape parameter 6.25 6.25 6.25

–intermediate goods producers
kB Pareto distribution shape parameter 6.25 6.25 6.25
–…nal consumption goods producers
FA Fixed cost of exporting 0.1489 0.1489 –

FB Fixed cost of production 0.0235 0.0235 0.0235

–sector of …nal consumption goods
dH Fraction of sales for which trade credit 0.55 0.0 0.0
is granted to domestic buyers
dF Fraction of sales for which trade credit 0.75 0.0 0.0
is granted to foreign buyers
a
Trade credit default rate – 0.0 0.0

Nb Share of non-exporters among – – 0.65

intermediate goods producers
z Persistence of productivity shocks 0.9 0.9 0.9
in the intermediate goods sector
u Persistence of demand shocks 0.9 0.9 0.9
a
The parameter is an endogenous variable in the TC model
b
The parameter N is an endogenous variable in the TC and EM models
TC - trade credit model, EM - extensive margin model, BN - benchmark model

879
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Value at Risk Estimation of Sector Index Futures: Evidence from Thailand Futures Exchange

Woradee Jongadsayakul

Associate Professor

Department of Economics, Faculty of Economics

Kasetsart University

Bangkok, Thailand 10900

Phone: +66-2-561-3474 ext. 315, Fax: +66-2-579-8739

fecowdj@[Link]

Author Note

The author acknowledges financial support from Department of Economics, Kasetsart

University.

880
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Abstract

Although Sector Index Futures has been traded in Thailand Futures Exchange since October

29, 2012, it has faced the liquidity problem. This study uses daily data from the period July 2,

2015 to December 27, 2019 for the Sector Index Futures to estimate Value at Risk (VaR)

using 3 methods as follows: non-parametric method using historical simulation approach,

parametric method using the model of Generalized Autoregressive Conditional

Heteroskedasticity (GARCH) Family (GARCH, TARCH, and EGARCH), and semi-

parametric method using volatility-weight historical simulation. The VaR estimation can

provide investors valuable information for examining the potential risk for their future

investment in Sector Index Futures market. Empirical results show that parametric method

using the model of EGARCH (1,1) is the best method for forecasting VaR. The results of

VaR estimation at confidence level of 95% using both non-parametric and parametric

methods also report the lowest potential loss in Commerce Index Futures investment.

Keywords: Value at Risk; Risk management; Hedging; Volatility

881
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Value at Risk Estimation of Sector Index Futures:

Evidence from Thailand Futures Exchange

Thailand Futures Exchange (TFEX), the only derivatives exchange in Thailand, was

established on May 17, 2004. At present, derivatives products include: SET50 Index Futures,

SET50 Index Options, Single Stock Futures, Precious Metal Futures, Deferred Precious

Metal, Interest Rate Futures, USD Futures, Sector Index Futures, and Rubber Futures. Sector

Index Futures, first offered on October 29, 2012, is a futures contract with sector index as an

underlying asset. Recently there are a total of 5 underlying sectors available for trading:

Energy and Utilities (ENERG), Food and Beverage (FOOD), Banking (BANK), Commerce

(COMM) and Information and Communication Technology (ICT). Although Sector Index

Futures contract serves as a tool for speculation and hedging sector-related risk, there has

been a lack of trading volume since 2013. This study therefore provides investors valuable

information for examining the potential risk for their future investment in Sector Index

Futures market by estimating Value at Risk (VaR), a measurement of risk exposure. VaR can

indicates the worst possible expected loss for an examined time horizon and a specified

confidence level. This study uses the following three methods for VaR estimation: non-

parametric method using historical simulation approach, parametric method using the model

of Generalized Autoregressive Conditional Heteroskedasticity (GARCH) Family (GARCH,

TARCH, and EGARCH), and semi-parametric method using volatility-weight historical

simulation. The purpose of this study is to test the relative performance of these three

methods in estimating and forecasting VaR in the Sector Index Futures market in Thailand.

The rest of the paper is organized as follows. The next section discusses previous

studies in this area. The sources of data and the methods adopted for calculating VaR are

described in the third section. The findings are reported in the forth section. Finally, I

conclude the paper in section 5.

882
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Literature Review

Value at Risk (VaR) was introduced by J.P. Morgan in 1994 as the method of risk

management. Previous study has examined the estimation of VaR by using three

conventional methods as follows:

(1) The non-parametric method measures a portfolio VaR without making strong

assumptions about returns distribution. It uses recent returns empirical distribution to estimate

VaR and assumes that the near future will be sufficiently similar to the recent past such that

the recent past data can be used for forecasting. Historical simulation is the most widely

implemented non-parametric approach. It was used to estimate VaR in the literature such as

Beder (1995), Hendricks (1996), Richardson, Boudoukh, and Whitelaw (1997), Raaji and

Raunig (1998), Chen, Chang, and Hsieh (2009), Chen and Chen (2013), and Abad and Benito

(2013).

(2) The parametric method measures risk by fitting probability curves to the examined

data sample and then inferring the VaR from the fitted curve. To capture the various volatility

effects observed in both the returns and prices of financial assets, the most popular models

are the GARCH family volatility models. Studies undertaken by Angelidis, Benos, and

Degiannakis (2004), So and Yu (2006), Carchano et al. (2010), Degiannakis, Floros, and

Livada (2012), Restrepo E. (2012), Abad and Benito (2013), Cera, Cera, and Lito (2013),

Wong, Chin, and Tan (2016), Smolović, Lipovina-Božović, and Vujošević (2017), Lipovina-

Božović, and Vujošević (2017), Gupta and Rajib (2018), Quang et al. (2018), and Wu (2018)

have applied many variants of the GARCH model in VaR estimation.

(3) The semi-parametric method combines the non-parametric method with the

parametric method. One of the most important semi-parametric methods is volatility-weight

historical simulation, which combines the benefit of historical simulation with volatility

883
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

models. Example of papers that reports the estimation of VaR using this approach is Hull and

White (1998).

Data and Methodology

Sector Index Futures contracts with 5 underlying sectors, including Energy and

Utilities (ENERG), Food and Beverage (FOOD), Banking (BANK), Commerce (COMM)

and Information and Communication Technology (ICT), have been offered since October 29,

2012; however, there has been a lack of trading volume since 2013. Instead of closing prices,

this research therefore uses the daily data of settlement prices for a period starting from July

2, 2015 to December 27, 2019 due to sector reclassifications in 2015. This study estimates

VaR of Sector Index Futures at 95% confidence interval using three following methods:

1. The non-parametric method using historical simulation approach.

2. The semi-parametric method using volatility-weight historical simulation.

3. The parametric method using Generalized Autoregressive Conditional

Heteroskedasticity (GARCH) family models for estimating volatility.

The following GARCH family models are estimated to characterize the volatility

clustering properties.

Model 1: Generalized Autoregressive Conditional Heteroskedasticity (GARCH) Model

The GARCH (1,1) model with constant mean can be written as follows:

Mean equation: Rt = c0 + ϵt ; ϵt|It-1 ̴ N(0, σt2)

Variance equation: σt2 = α0 + α1ϵt-12 + β1σt-12

where Rt is the futures return, α1 is the ARCH coefficient, β1 is the GARCH coefficient.

Model 2: Threshold ARCH (TARCH) Model

884
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

To capture asymmetries in terms of negative and positive shocks, the TARCH (1,1)

model is used as follows:

Mean equation: Rt = c0 + ϵt; ϵt|It-1 ̴ N(0, σt2)

Variance equation: σt2 = α0 + α1ϵt-12 + γϵt-12dt-1 + β1σt-12

where dt = 1 if ϵt < 0, and 0 otherwise. If γ > 0, the leverage effect is observed as the impulse

α1 + γ of negative shocks is larger than the impulse α1 of positive shocks.

Model 3: Exponential GARCH (EGARCH) Model

To capture asymmetries in terms of negative and positive shocks and to guarantee a

positive conditional variance, the EGARCH (1,1) model is used as follows:

Mean equation: Rt = c0 + ϵt; ϵt|It-1 ̴ N(0, σt2)

Variance equation: ln(σt2) = α0 + α1 |ϵt-1/σt-1| + γ ϵt-1/σt-1 + β1ln(σt-12)

where the negative sign of γ indicates the leverage effect.

The daily returns of the Sector Index Futures are computed as the natural logarithm of

the current day’s settlement price divided by the previous day's settlement price for a period

of 1,100 days, between July 3, 2015 to December 27, 2019. The first 1,000 days of the return

series are used to estimate the VaR models at significance level of 0.05, while the last 100

days are used to backtest VaR. Both unconditional coverage test (UC) and conditional

coverage test (CC) developed by Kupiec (1995) and Christoffersen (1998) respectively are

applied at the test significance level of 10% as recommended by Christoffersen (2012).

The unconditional coverage test, following a chi-squared distribution with one degree

of freedom, is applied for the null hypothesis, H0:   0.05 , using the following likelihood

ratio test statistic:

885
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

  N
 
LR UC  2 ln 1    0  N1  ln 1    0  N1 
N

 
where N0 = the number of days in which VaR is not violated

N1 = the number of days in which VaR is violated

 = the percentage of violation

The independence test checks, following a chi-squared distribution with one degree of

freedom, is applied for the null hypothesis of serial independence, using the following

likelihood ratio test statistic:

 
LR IND  2 ln 1   01 
N00
 01N 1   11 
01
N10
 11N 11
  ln 1    N0
N 
1

where N00 = the number of days in which VaR is not violated, following a non-violation in
VaR

N01 = the number of days in which VaR is violated, following a non-violation in VaR

N10 = the number of days in which VaR is not violated, following a VaR violation

N11 = the number of consecutive VaR violations

N01
 01 =
N00  N01

N11
 11 =
N10  N11

However, with no consecutive VaR violations (N11 = 0), the test statistic is as follows:

 
LR IND  2 ln 1   01 
N 00
 01N 01
  ln 1    N0
N  1

The conditional coverage test, following a chi-squared distribution with two degrees

of freedom, examined the joint hypothesis of unconditional and independence tests. The test

statistic can be described as LRCC = LRUC + LRIND.

886
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Results

Table 1 presents the VaR estimation at 95% confidence interval using historical

simulation approach in the Sector Index Futures market. It also reports the test statistics along

with the p-values generated on the basis of the chi-squared distribution. The historical

simulation VaR model in all Sector Index Futures except ICT Index Futures passes

unconditional coverage test, independence test, and conditional coverage test at a significance

level of 10%. However, for a poor significance level of 1%, the estimated historical

simulation VaR in ICT Index Futures is able to pass all tests. As a result, the model accuracy

is accepted. By comparison, COMM Index Futures market has a VaR value at -1.3254% and

this value is well above other Sector Index Futures. Therefore, based on the VaR measure,

COMM Index Futures is less risky than other Sector Index Futures.

Table 1. Results of the estimated historical simulation VaR and model evaluation

Types of VaR π LRUC LRIND LRCC

contract (α = 0.05) (π11) (P-value) (P-value) (P-value)
BANK Index 0.08 1.6158 1.3931 3.0089
-1.4364%
Futures (0.00) (0.2037) (0.2379) (0.2221)
ICT Index 0.10 4.1308 1.0158 5.1467
-1.8266%
Futures (0.20) (0.0421) (0.3135) (0.0763)
ENERG Index 0.05 0.0000 0.5266 0.5266
-1.7385%
Futures (0.00) (1.0000) (0.4680) (0.7685)
COMM Index 0.07 0.7530 1.0548 1.8078
-1.3254%
Futures (0.00) (0.3855) (0.3044) (0.4050)
FOOD Index 0.06 0.1984 0.7665 0.9649
-1.3582%
Futures (0.00) (0.6560) (0.3813) (0.6173)

Next, this study estimates the volatility of Sector Index Futures returns using GARCH

family. The estimation results of GARCH (1,1), TARCH (1,1), and EGARCH (1,1) are

shown in Table 2, Table 3, and Table 4 respectively, displaying the estimated coefficients and

their P-values, as well as diagnostics tests. The validity of the estimated models of GARCH

family is assessed first by employing Ljung–Box Q-test statistics up to lags 36 to check for

serial correlation in the standardized residuals. Then a Lagrange Multiplier test is employed

887
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

to examine whether the standardized squared residuals exhibit additional ARCH. The

insignificant Ljung-Box Q statistics and LM ARCH statistics imply that the residuals of the

estimated models are reasonably well behaved and adequately capture the persistence in the

variance of returns.

Table 2. Summary of GARCH (1,1) model

Types of c0 0 1 1 Q-Stat LM ARCH

contract (P-value) (P-value) (P-value) (P-value) (P-value) (P-value)
BANK Index 0.000158 3.60E-07 0.032079 0.962472 36.706 1.1398
Futures (0.5570) (0.0506) (0.0000) (0.0000) (0.436) (0.7675)
ICT Index 0.000422 1.86E-06 0.071520 0.918709 38.833 1.5924
Futures (0.1780) (0.0004) (0.0000) (0.0000) (0.343) (0.6611)
ENERG Index 0.000666 7.61E-07 0.060114 0.935683 35.990 1.2617
Futures (0.0273) (0.0361) (0.0000) (0.0000) (0.469) (0.7383)
COMM Index 0.000404 1.86E-06 0.081708 0.897010 27.157 0.0531
Futures (0.1134) (0.0001) (0.0000) (0.0000) (0.856) (0.9968)
FOOD Index 0.000112 2.08E-06 0.064339 0.908693 41.437 3.4622
Futures (0.6673) (0.0010) (0.0000) (0.0000) (0.245) (0.3257)

Table 3. Summary of TARCH (1,1) model

Types of c0 0 1  1 Q-Stat LM ARCH

contract (P-value) (P-value) (P-value) (P-value) (P-value) (P-value) (P-value)
BANK Index 4.31E-05 5.80E-07 0.013915 0.044916 0.956329 37.863 0.8215
Futures (0.8690) (0.0068) (0.1207) (0.0001) (0.0000) (0.384) (0.8443)
ICT Index 0.000269 1.98E-06 0.037161 0.052949 0.923204 39.514 1.2102
Futures (0.3960) (0.0000) (0.0000) (0.0000) (0.0000) (0.316) (0.7506)
ENERG Index 0.000444 1.50E-06 0.022827 0.070078 0.929503 35.792 1.3993
Futures (0.1479) (0.0012) (0.0482) (0.0000) (0.0000) (0.478) (0.7057)
COMM Index 0.000352 2.00E-06 0.065897 0.027647 0.896769 27.341 0.1104
Futures (0.1870) (0.0000) (0.0003) (0.1277) (0.0000) (0.850) (0.9906)
FOOD Index 3.02E-06 2.78E-06 0.036645 0.073899 0.890126 45.939 2.4608
Futures (0.9908) (0.0000) (0.0220) (0.0022) (0.0000) (0.124) (0.4824)

888
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 4. Summary of EGARCH (1,1) model

Types of c0 0 1  1 Q-Stat LM ARCH

contract (P-value) (P-value) (P-value) (P-value) (P-value) (P-value) (P-value)
BANK Index 4.24E-05 -0.126988 0.078339 -0.041464 0.993020 37.141 0.9865
Futures (0.8698) (0.0002) (0.0000) (0.0000) (0.0000) (0.416) (0.8045)
ICT Index 0.000161 -0.180646 0.126927 -0.048000 0.990188 40.887 0.6663
Futures (0.6030) (0.0000) (0.0000) (0.0000) (0.0000) (0.264) (0.8811)
ENERG Index 0.000500 -0.228519 0.116835 -0.055720 0.984792 36.509 1.3965
Futures (0.0940) (0.0000) (0.0000) (0.0000) (0.0000) (0.445) (0.7064)
COMM Index 0.000355 -0.401127 0.163194 -0.028592 0.970888 26.355 0.1872
Futures (0.1589) (0.0000) (0.0000) (0.0132) (0.0000) (0.880) (0.9796)
FOOD Index 1.46E-05 -0.504302 0.142783 -0.070778 0.958736 45.520 2.6997
Futures (0.9558) (0.0000) (0.0000) (0.0000) (0.0000) (0.133) (0.4403)

Table 2 presenting the estimation result of GARCH(1,1) model shows that the

coefficient for the previous shock (the ARCH coefficient:  1 ) and that for its lagged

conditional variance (the GARCH coefficient: 1 ) are highly statistically significant as their

P-values less than 0.001. For the estimation results of TARCH (1,1) model in Table 3, the

leverage effect is is observed as the impulse α1 + γ of negative shocks is larger than the

impulse α1 of positive shocks. The EGARCH (1,1) model in Table 4 also shows the existence

of leverage effect. All estimated coefficients in the variance equation are significant at the

level of 0.05.

After the parameters were estimated, the volatility of the daily returns of Sector Index

Futures is forecasted using GARCH family models that cater for volatility clustering. Table

5, 6 and 7 show the results of VaR modelling using GARCH (1,1), TARCH (1,1) and

EGARCH (1,1) respectively.

889
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 5. Results of VaR modelling using GARCH (1,1) and model evaluation

Types of Avg. VaR π LRUC LRIND LRCC

contract (α = 0.05) (π11) (P-value) (P-value) (P-value)
BANK Index 0.06 0.1984 0.9190 1.1174
-1.6393%
Futures (0.17) (0.6560) (0.3377) (0.5720)
ICT Index 0.09 2.7510 0.0508 2.8018
-2.0163%
Futures (0.11) (0.0972) (0.8217) (0.2464)
ENERG Index 0.05 0.0000 0.5266 0.5266
-1.5612%
Futures (0.00) (1.0000) (0.4680) (0.7685)
COMM Index 0.06 0.1984 0.7665 0.9649
-1.3780%
Futures (0.00) (0.6560) (0.3813) (0.6173)
FOOD Index 0.05 0.0000 0.5266 0.5266
-1.4001%
Futures (0.00) (1.0000) (0.4680) (0.7685)

Table 6. Results of VaR modelling using TARCH (1,1) and model evaluation

Types of Avg. VaR π LRUC LRIND LRCC

contract (α = 0.05) (π11) (P-value) (P-value) (P-value)
BANK Index 0.05 0.0000 0.5266 0.5266
-1.7729%
Futures (0.00) (1.0000) (0.4680) (0.7685)
ICT Index 0.09 2.7510 0.0508 2.8018
-2.0524%
Futures (0.11) (0.0972) (0.8217) (0.2464)
ENERG Index 0.05 0.0000 0.5266 0.5266
-1.5805%
Futures (0.00) (1.0000) (0.4680) (0.7685)
COMM Index 0.06 0.1984 0.7665 0.9649
-1.4062%
Futures (0.00) (0.6560) (0.3813) (0.6173)
FOOD Index 0.04 0.2253 0.3334 0.5588
-1.4431%
Futures (0.00) (0.6350) (0.5637) (0.7562)

Table 7. Results of VaR modelling using EGARCH (1,1) and model evaluation

Types of Avg. VaR π LRUC LRIND LRCC

contract (α = 0.05) (π11) (P-value) (P-value) (P-value)
BANK Index 0.05 0.0000 0.5266 0.5266
-1.7357%
Futures (0.00) (1.0000) (0.4680) (0.7685)
ICT Index 0.09 2.7510 0.0508 2.8018
-2.1484%
Futures (0.11) (0.0972) (0.8217) (0.2464)
ENERG Index 0.05 0.0000 0.5266 0.5266
-1.6551%
Futures (0.00) (1.0000) (0.4680) (0.7685)
COMM Index 0.05 0.0000 0.5266 0.5266
-1.4415%
Futures (0.00) (1.0000) (0.4680) (0.7685)
FOOD Index 0.04 0.2253 0.3334 0.5588
-1.4761%
Futures (0.00) (0.6350) (0.5637) (0.7562)

890
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

The results of parametric method show that the VaR estimations at 95% confidence

interval using GARCH family models in Sector Index Futures market pass easily

independence test and conditional coverage test at a significance level of 10%, as well as

unconditional coverage test at a significance level of 5%. As a result, the model accuracy is

accepted. Empirical results show that parametric method using the model of EGARCH (1,1)

is the best method for forecasting VaR since the model findings based on EGARCH (1,1) has

the lowest numbers of VaR exceptions. By comparison, COMM Index Futures market has a

lowest average VaR value so it provides a lowest risk investment compared to other Sector

Index Futures.

Combining the non-parametric method with the parametric method, this study

estimates the historical simulation VaR from the returns that has been adjusted to the current

state of the market volatility with GARCH family. The estimation results of volatility-

adjusted historical simulation VaR with GARCH(1,1), TARCH (1,1), and EGARCH (1,1)

are presented in Table 8, Table 9, and Table 10 respectively.

The results of the estimated volatility-adjusted historical simulation VaR with

GARCH family shows that the VaR estimation in FOOD Index Futures passes unconditional

coverage test, independence test, and conditional coverage test at a significance level of 10%.

However, for a significance level of 5%, the VaR estimation in ENERG Index Futures is able

to pass all tests. Whereas there is no evidence to reject the null hypothesis of serial

independence, the unconditional coverage test and conditional coverage test show undeniable

results that the estimated volatility-adjusted historical simulation VaR with GARCH family

in Sector Index Futures with BANK index, ICT index, and COMM index as underlying assets

are inaccurate. As a result, volatility-weight historical simulation approach has a poor

performance in estimating VaR compared to other methods.

891
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 8. Results of the estimated volatility-adjusted historical simulation VaR with

GARCH(1,1) and model evaluation

Types of VaR π LRUC LRIND LRCC

contract (α = 0.05) (π11) (P-value) (P-value) (P-value)
BANK Index 0.25 45.0136 0.4596 45.4732
-0.8605%
Futures (0.20) (0.0000) (0.4978) (0.0000)
ICT Index 0.13 9.5367 0.0721 9.6088
-1.6110%
Futures (0.15) (0.0020) (0.7883) (0.0082)
ENERG Index 0.09 2.7510 0.0508 2.8018
-1.1628%
Futures (0.11) (0.0972) (0.8217) (0.2464)
COMM Index 0.13 9.5367 0.4194 9.9561
-0.9523%
Futures (0.08) (0.0020) (0.5172) (0.0069)
FOOD Index 0.05 0.0000 0.5266 0.5266
-1.4766%
Futures (0.00) (1.0000) (0.4680) (0.7685)

Table 9. Results of the estimated volatility-adjusted historical simulation VaR with

TARCH(1,1) and model evaluation

Types of VaR π LRUC LRIND LRCC

contract (α = 0.05) (π11) (P-value) (P-value) (P-value)
BANK Index 0.19 24.9027 0.1628 25.0656
-0.9905%
Futures (0.16) (0.0000) (0.6866) (0.0000)
ICT Index 0.13 9.5367 0.0721 9.6088
-1.5753%
Futures (0.15) (0.0020) (0.7883) (0.0082)
ENERG Index 0.08 1.6158 0.2099 1.8257
-1.2164%
Futures (0.13) (0.2037) (0.6468) (0.4014)
COMM Index 0.14 11.7103 0.7358 12.4461
-0.9426%
Futures (0.07) (0.0006) (0.3910) (0.0020)
FOOD Index 0.04 0.2253 0.3334 0.5588
-1.5984%
Futures (0.00) (0.6350) (0.5637) (0.7562)

Table 10. Results of the estimated volatility-adjusted historical simulation VaR with
EGARCH(1,1) and model evaluation

Types of VaR π LRUC LRIND LRCC

contract (α = 0.05) (π11) (P-value) (P-value) (P-value)
BANK Index 0.19 24.9027 0.1628 25.0656
-0.9869%
Futures (0.16) (0.0000) (0.6866) (0.0000)
ICT Index 0.12 7.5402 0.2585 7.7987
-1.6590%
Futures (0.17) (0.0060) (0.6112) (0.0203)
ENERG Index 0.08 1.6158 0.2099 1.8257
-1.2515%
Futures (0.13) (0.2037) (0.6468) (0.4014)
COMM Index 0.13 9.5367 0.4194 9.9561
-1.0113%
Futures (0.08) (0.0020) (0.5172) (0.0069)
FOOD Index (0.04) 0.2253 0.3334 0.5588
-1.5982%
Futures (0.00) (0.6350) (0.5637) (0.7562)

892
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Conclusion

Thailand Futures Exchange (TFEX) started trading Sector Index on October 29, 2012.

However, there has been lack of trading volume since 2013. Instead of closing prices, this

research therefore uses the daily data of settlement prices for a period starting from July 2,

2015 to December 27, 2019 to estimate VaR of Sector Index Futures at 95% confidence

interval using three following methods: (1) Non-parametric method using historical

simulation approach (2) The parametric method using GARCH family for volatility modeling

and (3) The semi-parametric method using volatility-weight historical simulation.

The empirical results show that both non-parametric method using historical

simulation approach and parametric method using GARCH family pass unconditional

coverage test, independence test, and conditional coverage test. They provide the lowest VaR

value in COMM Index Futures. Moreover, VaR model based on EGARCH (1,1) tends to be

more accurate than others due to the lowest numbers of VaR exceptions.

The VaR estimation in this paper can provide investors valuable information for

examining the potential risk for their future investment in Sector Index Futures market. Based

on the VaR measure, COMM Index Futures is less risky than other Sector Index Futures.

However, with lack of trading volume in Sector Index Futures, liquidity risk should also be

taken into account for future investment in this market.

References

Abad, Pilar, and Benito, Sonia. (2013). A Detailed Comparison of Value at Risk Estimates.

Mathematics and Computers in Simulation, 94, 258-276. Available at

[Link]

Angelidis, Timotheos, Alexandros, Benos, and Degiannakis, Stavros. (2004). The Use of

GARCH Models in VaR Estimation. Statistical Methodology, 1(1–2), 105-128.

Available at [Link]

893
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Beder, Tanya. (1995). VAR: Seductive but Dangerous. Financial Analysts Journal, 51(5),

12-24. Available at [Link] _

Seductive_but_dangerous.

Carchano, Oscar, Rachev, Svetlozar T., Kim, Young S., Sun, Edward W., and Fabozzi, Frank

J. (2010). Forecasting VaR in Spot and Futures Equity Markets. Available at

[Link]

Equity_Markets.pdf.

Cera, Gentjan, Cera, Edmond, and Lito, Gerdi. (2013). A GARCH Model Approach to

Calculate the Value at Risk of Albanian Lek Exchange Rate. European Scientific

Journal, 9(25), 250-260. Available at [Link]

download?doi=[Link].2981&rep=rep1&type=pdf.

Chen, Cheng-Te, Chang, Hae-Ching, and Hsieh, Chin-Shan. (2009). Forecasting Value at

Risk (VAR) in the Futures Market Using Hybrid Method of Neural Networks and

GARCH Model. Available at [Link]

[Link].

Chen, Qi, and Chen, Rongda. (2013). Method of Value-at-Risk and Empirical Research for

Shanghai Stock Market. Procedia Computer Science, 17, 671-677. Available at

[Link]

Christoffersen, Peter F. (1998), Evaluating Interval Forecasts. International Economic

Review, 39(4), 841-862. Available at [Link]

. (2012). Elements of Financial Risk Management. 2nd Edition. Academic Press.

894
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Degiannakis, Stavros, Floros, Christos, and Livada, Alexandra. (2012). Evaluating Value-at-

Risk Models Before and After the Financial Crisis of 2008: International Evidence.

Managerial Finance, 38(4), 436-452. Available at

[Link]

Gupta, Apoorv, and Rajib, Prabina. (2018). Do VaR Exceptions Have Seasonality? An

Empirical Study on Indian Commodity Spot Prices. IIMB Management Review,

30(4), 369-384. Available at [Link]

Hendricks, Darryll. (1996). Evaluation of Value-at-Risk Models Using Historical Data.

Economic Policy Review, 2(1), 39-70. Available at

[Link]

Hull, John, and White, Alan. (1998). Incorporating Volatility Updating into the Historical

Simulation for Value at Risk. Journal of Risk, 1, 5-19. Available at

[Link]

Quang, Paul B., Klein, Tony, Nguyen, Nam H., and Walther, Thomas. (2018). Value-at-Risk

for South-East Asian Stock Markets: Stochastic Volatility vs. GARCH. Journal of

Risk and Financial Management, 11(2), 1-20. Available at

[Link]

Raaji, Gabriela de, and Raunig, Burkhard. (1998). A Comparison of Value at Risk

Approaches and Their Implications for Regulators. Focus on Austria, 4, 57-71.

Available at [Link]

5881d054e71e/value_at_risk_e_tcm16-[Link].

Restrepo E., María I. (2012). Estimating Portfolio Value at Risk with GARCH and

MGARCH Models. Perfil de Coyuntura Económica, 19, 77-92. Available at

895
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[Link]

42142012000100004&lng=en&tlng=en.

Richardson, Matthew P., Boudoukh, Jacob, and Whitelaw, Robert F. (1997). The Best of

Both Worlds: A Hybrid Approach to Calculating Value at Risk. Available at

[Link]

Smolović, Julija C., Lipovina-Božović, Milena, and Vujošević, Saša. (2017). GARCH

Models in Value at Risk Estimation: Empirical Evidence from the Montenegrin

Stock Exchange. Economic Research-Ekonomska Istraživanja, 30(1), 477–498.

Available at [Link]

So, Mike K.P., and Yu, Philip L.H. (2006). Empirical Analysis of GARCH Models in Value

at Risk Estimation. Journal of International Financial Markets, Institutions and

Money, 16(2), 180-197. Available at [Link]

Wong, Zhen Y., Chin, Wen C., and Tan, Siow H. (2016). Daily Value-at-Risk Modeling and

Forecast Evaluation: The Realized Volatility Approach. The Journal of Finance and

Data Science. 2(3), 171-187. Available at [Link]

Wu, Chunchou. (2018). Measuring and Comparing the Value-at-Risk Using GARCH and

CARsR Models for CSI 300 Index. Theoretical Economics Letters, 8(6), 1179-1187.

Available at [Link]

896
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

AUTHENTICITY THROUGH THE EYES OF THE HOST:

THE COMPLICATION OF PRESENTING CULTURE AS
COMMODITY
YODMANEE TEPANON
Department of Operations Management, Faculty of Business Administration
Kasetsart University
Bangkok, THAILAND
yodmanee.t@[Link]

THANWA BENJAWAN
Faculty of Social Science and Humanities,
Payap University
Chiang Mai, THAILAND
thanwa.b@[Link]

CHAWAN MALEEHOM
Faculty of Social Science and Humanities,
Payap University
Chiang Mai, THAILAND
cmaleehom@[Link]

Abstract: Experiencing local culture undeniably is one of the main tourist motivation. Hill tribe
tourism is among famous tourism activities in the northern part of Thailand and therefore is
often included in a tourist itinerary. Hill tribe communities are visited and now undergoing
through an ongoing change in displaying their culture. This study investigates the host
perception of authenticity of their local culture when presenting to tourists within the tourism
landscape. Using qualitative methodology, local key informants at Bann Mae Klang Luang hill
tribe community in Chiang Mai, Thailand, were interviewed. The findings indicate that Karens
at Bann Mae Klang Luang opened their village to tourists not only for the better income but
also to correct wrong image and perception outsiders hold about them. While attempting to
earn money by attracting tourists with their unique culture and way of life, Karens impulsively
modify their culture or unintentionally borrow tangible or intangible cultural presentation from
other cultures for that purpose. Pursuing tourism as an easy way out of poverty has not always
been the bed of roses. Their lives have changed and gone through similar processes as
happened to other villages ahead of them. This study conclusively stresses the needs to
inform or educate tourists for a better understanding of the local culture and urge them to
change their behavior. Tourists are encouraged to learn more and change their modern
lifestyle requests. This will lead to the appreciation and preservation of the traditional Karen
culture in the end.
Keywords: Authenticity, Hill tribe, Ethnic tourism, Thailand

1. Introduction
Hill tribe communities in Thailand are alike those living in other remote locations all
over the world that they impress tourists or the outsiders by their mysterious, traditional ways
of living. The unique, primitive and alienated experiences are usually expected by tourists
when visiting these locations. Considering that ethnic groups and culture have widely drawn
tourists’ interest and visits, interactions between the host and the guest could lead to
unintentional exploitation of the culture through reckless presentation and interpretation.

897
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Evidences have shown that locals’ alteration of their culture or tradition unwittingly to impress
visitors could lead to its extinction. Therefore, in this study, I aim to conduct a preliminary
investigation of the community perception of how they present they culture and whether it has
been transformed, manipulated, or changed to better fit with the tourist expectation or lifestyle.

2. Literature Review
2.1 Authenticity in Tourism
Authenticity is an eminent issue in tourism literature. Even though it has been widely
studied (MacCannell, 1973; Cohen, 1988; Heidegger, 1996; Wang, 1999; Chhabra, Healy &
Sills, 2003; Reisinger & Steiner, 2006; Zhu, 2012), the concept is still considered unsettled.
The term ‘authenticity’, coming from a Greek word, means “made by oneself” and “original”
(Zhou, Zhang, Zhang & Li, 2018). MacCannell (1976) suggested that “touristic consciousness
is motivated by its desire for authentic experience” (pp.101). It is defined as the “originals”,
related to or Most literature in tourism has looked at the authenticity from the tourist
perspectives (Zhang, S. N. et al., 2019; Engenset & Elvekrok, 2015; Meng & Choi, 2016).
Trilling (1972), one of the pioneers in researching the authenticity, proposed that the original
usage of authenticity was in the museum “where persons expert in such matters test whether
objects of art are what they appear to be or are claimed to be, and therefore worth the price
that is asked for them…, worth the admiration they are being given (pp. 93).
However, the evolution of the authenticity concept has shown that authenticity in
tourism covered more than tangible objects. It includes ‘experiences’ and the ‘objects’
themselves (Wang, 1999). According to Wang (1999), there are 3 types of authenticity in
tourism experience. The objective authenticity explains the experiences in tourism could be
created in a contrived tourism environment which, to the tourists, is original. The authentic
experienced tourists feel due to their inability to differentiate the real and not real. The
constructivism approach looks at authenticity from the perspective of constructivists. Tourists
are not searching for objective authenticity, or whether the objects or experiences are original
or real. However, objects or experiences are perceived as the symbols of authenticity. The
last type of authentic experience, the existential authenticity, is explained as Berger (1973)
suggests that it “denotes a special state of Being in which one is true to oneself, and acts as
a counterdose to the less of ‘true self’ in public roles and public spheres in modern Western
society” (Wang, 1999 pp. 358). Authenticity is normally perceived with the host is being the
object of tourists’ authentic experience (Chhabra, 2005). Later, Wang (2007) carefully
summarizes the authenticity concept found in research and concludes the 2 paths to
understand it as the ‘object’ and the ‘self’ which refer to the other (the toured) and the center
(the tourist), respectively.
Nonetheless, the meaning of authenticity is frequently given by looking or interpreting
from the tourists’ perspectives. The host perception of authenticity has not been much of the
quest for research. It does not concern tourists’ evaluation of the authenticity of the host’s
culture or lifestyle or the customed authenticity felt during the interaction (Wang, 2007). It is
about how the host see their culture as authentic as it should be or has been. Unfortunately,
often times particular performances, activities, events are particularly chosen due to their
“authentic” impression requested by tourists. As the results, local and traditional culture is
sometime twisted and modified and eventually transformed.

898
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

2.2 Ethnic Tourism and Presentation of Culture

Ethnic tourism has gained tourists’ interest in many parts of the world and is used by
governments in order to increase and attract tourism revenues. It is defined as tourism that is
“marketed to the public in terms of the ‘quaint’ customs of indigenous and often exotic peoples”
(Smith, 1977, p. 2). Ethnic tourism is supposed to endure minorities as their, particularly if not
all, cultural identities are selected, illustrated, and maintained through tourism-related activities
and system for tourist entertainment (Santos & Yan, 2008). Through the host-guest (or local-
tourist) interaction, negotiation and evaluation of their culture and tradition are analyzed and
hence displayed to best fit with tourist interest (Anderson, 1987; MacCannell, 1992). During
those interactions between the host and the guest, hosts are the best representatives of their
own culture and traditional cultural heritage. Cultural representation is shown to and
participated by tourists in different activities; passively and actively. Hosts are merely the
creator and supplier of the unique experiences that are sought by tourists. Nevertheless,
when local life-styles and tourists (with their certain expectation) are met in person and culture
and traditions are transformed as tourism commodities within the realm of pre-arranged social
scape, it creates a complex minority representation (Ryan & Aicken, 2005).

3. Methodology
The research site is located in Chom Thong district, Chiang Mai, Thailand. It is a Karen
community called Bann Mae Klang Luang (or Mae Klang Luang village). Situated on the way
to Doi Inthanon, the popular tourist destination of Chaing Mai, Bann Mae Klang Luang has
welcomed both Thai and foreign tourists every year, especially during the cold season (see
figure 1-3). The community consisting of 358 residents has been led by Poh Luang Pong To,
the head of the community, and has taken tourism as a tool to increase their annual income
since 1999. Starting from couple families seeing the opportunities to earn more income from
tourism by providing tourist accommodation, at present, more than half of the villagers involved
in tourism. The village also practices the community-based tourism (CBT) by including all
stakeholders in planning and implementing tourism activities in the area. Since then, although
the Mae Klang Luang is not the most visited hill tribe attraction when considering the number
of tourists, it has gradually gained tourist interests and visits due to its unspoiled and pristine
natural environment and the hill tribe culture.
This research employed the qualitative approach. Using an in-depth and informal
interview method, 7 key informants who were Karens residing in the Bann Mae Klang Luang
were interviewed from September 2019 to January 2020. The first time of our visits, we were
greeted by Poh Luang Pong To and his assistant, Oun. Other informants included the
restaurant owners, servers, kitchen staff and villagers who we met when exploring the village.
These people have had, in one way or another, experiences with tourists. Nonetheless, most
of the interviews took place at the main restaurant as it serves as the main gathering spot.
Each interview lasted approximately 45 – 90 minutes, formally and informally. However, the
more informal interviews also took place during every visit, some times more than 3 hours of
casual conversation. I developed the interview protocol by looking at how the host viewed
their practices and presentation within the tourism realm to be ‘authentic’. The term ‘authentic’
could be objective, constructive, or existential authenticity and how it has been converted,
intentionally or unintentionally, through different presentation modes during host-guest
interaction.

899
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 1-3: Bann Mae Klang Luang Maps

([Link]
[Link]

4. Results
After 5 trips to the village, roughly 3-5 hours each trip, 7 key informants (detailed in
table 1) were interviewed. Our informants have been involved in businesses or activities that
provide them with opportunities to meet/interact with tourists. Our questions simply were
whether they felt complied to act or change their ways of life based on the needs and
expectation of tourists and how tourism has changed their living. Eventually, their stories were
unfolded.

900
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 1: Key informant list and information

Name Who they are Age What they do

(approx.)
1 Pong To Head of the 50-55 Hold authority in deciding and
village managing tourism in Bann Mae Klang
Luang
2 Oun Assistant 40-45 Assist the head to manage tourism
Head of the activity, including homestay service,
village food and beverage, and transportation
to facilitate tourists visiting the village
3 D To Main 40-45 Prepare food and beverage service
restaurant on-premises (restaurant) or deliver
owner food and beverage to the villa
4 Nok Kitchen 30-35 Prepare food and beverage service
staff/server on-premises (restaurant) or deliver
food and beverage to the villa
5&6 O & Bud Noodle shop 30-35 Own a noodle & cook-to-order meal
owner shop
7 Joy Noodle shop 25-30 Prepare food and beverage service
staff on-premises (noodle shop)

4.1 “We want to show them how we live (and not only concerning tourism)”
The tourism story of Bann Mae Klang Luang started differently when compare to other
hill tribe villages. Unquestionably, Karens people at Mae Klang Luang have witnessed the
growth of tourism in other nearby areas, hence they are fairly aware of and expect the changes
and impacts of tourism. However, the locals revealed the main reasons why they decided to
open their village to tourists which were, of course, including drawing more income and
creating a better understanding of the hill tribe people. The Karens, being one of several hill
tribe communities in Thailand, have been blamed for the wildfire occurring now and then
especially in the northern part of Thailand. The negative perception painted by “city people”
about hill tribes is more or less the deforester. The hill tribe people are accused of destroying
the forest due to their agricultural norms and traditions. Shifting cultivation, watershed forest
destruction and opium cultivation are among activities that have frequently been seen and
portraited to and by outsiders (FAO, 2002). Nonetheless, with my own eyes, Mae Klang Luang
village shows no sign of deforestation. The area is a lush jungle with some areas used as rice
fields.
Interestingly, once when we started our interview with Poh Luang Pong To and Oun,
the idea of having tourism in the area at the beginning was not so much on generating
revenues. Frustration arises due to the depiction the outsiders have about hill tribes. City
people (mainly in Bangkok and the metropolitan areas) perceive the cause of several recent
floods are caused by deforestation (Posttoday, 2019). Having tourism activities was more on
the purpose of showing the outsiders how Karens live, the traditions, way of life and culture
that depend on and respect nature.
“There have been various media and information… from the outside.. like Ajarn (the
researcher) lives in Bangkok, right? You may think that we destroy the forest. When
it rains, the town is flooded. Yes… that must be because of the hill tribes. We have
been hearing things like that a lot” (Pong To, September 2019).

901
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

“Well, if we invite them (outsiders) to visit as tourists, they will understand that there
are many tribes… and when you say “hill tribes”, it is stereotyping. We hear the word
“hill tribes” and that, of course, includes us. Hmong, Lisu and Ahka, we are all hill tribe
groups. But we are here, we feel.. we have been condemned. So, I think ‘what should
we do?’ ‘Should we go to explain about this in Bangkok?’ That would be no chance.
‘Who is going to take us?’ This ethnic/hill tribe tourism trend actually has helped us.
People started to visit. They want to see whether or not we destroy the forest.
Sometimes they come unannounced just to check the condition of the trees and forest”
(Pong To, September 2019).

4.2 Adapt but not manipulate

From our conversation with the informants, there is a mix of feelings toward opening
the village to tourism. It was a doubtful at first by several groups of villagers. Nonetheless,
Pong To’s fierce ambition to involve in tourism has brought an understanding and a
compromise to the community. The involvement had increased when locals realized that they
could be a part of tourism, either as the service provider or supplier. The participating level is
normally depended on the returned benefit one would expect to gain from participating in
tourism.
“At first, not all seemed to care about tourism. We (Karens) are and have been
farmers. They seemed skeptical and did not know how tourism will help the village or
their lives. So, they let us (some of the tourism pioneers) try first. But I noticed that
they have been watching. Once we officially started welcoming tourists, things
changed. Those who did not show interest, later came to me and asked for
involvement in tourism. It took some time but eventually they agreed to join” (Pong to,
September, 2019).
“We are locals here. We are not very involved with tourism at the beginning but once
tourists had arrived, we saw opportunities. Selling food and drinks is easy as we do
not have to drastically change our living or ways of life.” (O, January, 2020).
However, being hill tribe people, even though Karens’ culture has long been treasured
and preserved, Pong To and Oun humbly accept that Karens do have limited and somewhat
common culture. Being farmers and having more a ‘gather – hunter’ lifestyle, most of their
culture or ways of life are in the fields, planting and harvesting. Regrettably, most tourists
expect more of an astounded, complexed experience. Therefore, they are inevitably forced
to borrow others’ performance, for example, the Northern Thai dance which utterly is not
directly related to their lives. Their original and traditional musical instrument, Te Nah (figure
4), has deferentially been used to present their culture. Despite the fact that it cannot capture
tourists’ interest that much due to most tourists visiting hill tribes prefer a more hurried and
schematic experience. The informants informed us that there were some groups that are into
Karen local life and choose to spend time exploring the culture. Nonetheless, the majority of
visitors are still looking for shallow, easy-going experiences or activities e.g. taking photos or
watching any kind of performance presented by the villagers.
“Karens, we do not have that kind of high level or complexed culture. We live in the
jungle. We respect the nature. Tourists come in with an expectation of us having
some things or performance to show them. But we have nothing like that. Our only
tangible cultural heritage that I can think of now is Te Nah. So, we have made use of
the northern Thai performance. We do not feel that they are others’ and resent it. And

902
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

tourists (seem to) like it. (laugh) but we never said that it (the performance) was ours.
Well, no one has asked anyway” (Pong To, January, 2020).
At Bann Mae Klang Luang, we noticed that there was a mix between villagers who
wore traditional Karen costume (see figure 5) regularly and those who did not. Though the
informants express their pride of waring their own costume, we observe a slight variation of
cultural attachment. Those very involved in tourism wear the Karen’s clothing more regularly
that those who are not, yet not a full costume as what are worn in their special occasions.
(From the researcher’s experience, when comparing to other villages, Mae Klang Luang
villagers seem to wear Karen clothing in the original way, without a modern style t-shirt
underneath). Still, the costume would be readily available for usage if it means to please
tourists.
“We normally do not wear Karen costume. I think modern and normal costume is
easier to wear to work. But we understand that sometimes we have to play along so
that it attracts tourists and make them happy (O, January, 2020).”
“I wear them (Karen clothes) everyday. I think they are comfortable and tourists like
seeing us in this costume. I think. I think when they come to see Karens, they should
see Karens (with the costume). So ‘why not?’ I have no problem wearing it. I actually
am proud to do that (laugh)” (Pong To, December, 2019).

Figure 5: Karen’s traditional costume

Figure 4: Te Nah [Link] (photo from [Link]
[Link]/chiangrai/article_attach/ar
ticle_fileattach_20161103104554.pdf)

Regards to the food and beverage experience, I found that the menus available to
tourists at the village contain mostly the popular and familiar Thai dishes, for example, stir fried
pork or chicken with holy basil or central-style rice porridge. Locals even serve western style
coffee and homemade delicatessen e.g. crapes and waffles. In fact, Karens do have their
own food called Kao Bur, which is rice porridge cooked with basically most ingredients
available in the kitchen (meat, vegetables, spices, herbs etc.). They eat this on top of cooked
rice. This is a very traditional and authentic way to eat Karen’s food. Although, this dish, along
with some other original Karen’s dishes, are not served or even offered to the tourists.

903
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

(researcher: why don’t you cook Kao Bur or make it available to tourists?)
“Will they like it? I do not think so. Kao Bur is not something tourists would eat. It
does not taste good for them and looks too mushy. (researcher: I would like to try.
Could you please make one for me?). Yes, I can. But I cannot assure you that you
will like it. Most tourists do not ever ask for it or, in fact, any other Karen food. Are you
sure you do not want something else? Something more suitable to your (outsider’s)
taste?” (D To, November, 2019).

5. Discussion and Conclusion

The Bann Mae Klang Luang hill tribe community is an example of a host community
attempting to play a part in the tourism industry by adapting its norms and ways of life to fit
with those of the tourists. For them, the authenticity may as well mean being true to
themselves. Activities requiring too much of manipulation of their culture as seen in other
ethnic destination are not made available only for tourists’ enjoyment. Their cultural
presentation comes naturally, does not refer to any concept or theory. This could be seen as
existential authenticity for the host in which the so-called borrowed performance seems to be
real to them. The service process which includes modifying their food or practices only to fit
with tourists’ is seen to be appropriate and acceptable. Presenting the real Karen tribe,
adjusting their lifestyle, and borrowing other culture seem to be quite natural and harmoniously
blended together with only slight effort.
However, these findings are understandable since Bann Mae Klang Luang is still
considered to be at the beginning stage of tourism development where tourism activities are
not yet fully developed. The host is not fully aware of the changes tourism could cause and,
therefore, unknowingly fail to notice their needs and allow tourist to take advantage of them.
The community and tourism practitioners/stakeholders need to be prepared for the following
stages which normally bring about tremendous changes to local living conditions. A careful
planning is required both for the host and the guest. Since most tourists hold different
expectation of what will be offered or experienced at Bann Mae Klang Luang, visitor education
before arriving at the village should be put in place via various platforms (e.g. promotional
broadcast, social media). Market segmentation is needed to carefully identify the target for
designing appropriate tourism products and services. Community agreement and regulations
are crucial to protect and rejuvenate their real culture and traditions. Otherwise, they would
possibly be vanished or transformed into a tourism contrivance or a tourism theme park similar
to those ethnic destinations that went through this path before.

904
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

References
Anderson, K., 1987. The Idea of Chinatown: The Power of Place and Institutional Practice in
the Making of Racial Category. Annals of the Association of American Geographers, 77(4),
pp. 580–598.
Berger, P. L., 1973. Sincerity and Authenticity in Modern Society. Public Interest (31), pp. 81-
90.
Chhabra, D., Healy, R. and Sills, E., 2003. Staged authenticity and heritage tourism. Annals
of Tourism Research, 30 (3), pp. 702-719.
Chhabra, D., 2005. Defining authenticity and its determinants: Toward an authenticity flow
model. Journal of Travel Research, 44 (1), pp. 64–73.
Cohen, E., 1988. Authenticity and commoditization in tourism. Annals of Tourism Research,
15, pp. 371-386.
Engenset, M. G. and Elvekrok I., 2015. Authentic concepts: effects on tourist satisfaction.
Journal of Travel Research, 54 (4), pp. 456-466.
FAO, 2002. Case Study on Education Opportunities for Hill Tribes in Northern Thailand:
Implications for Sustainable Development, [online] Available at:
<[Link]
Heidegger, M. 1996. Being and Time. Albany: State University of New York.
Hsu, P. H. and Nilep C., 2015. Authenticity in indigenous tourism: the provider’s perspective.
International Journal of Critical Indigenous Studies, 8 (2), pp. 16-28.
MacCannell, D.,1973. Staged authenticity: Arrangement of social space in tourist settings.
American Journal of Sociology, 79 (3), pp. 589 - 603.
MacCannell, D.,1992. Cannibalism Today. In Empty Meeting Grounds, D. MacCannell, ed.,
pp. 17–73. London: Routledge.
Meng, B. and Choi, K., 2016. The role of authenticity in forming slow tourists’ intentions:
developing an extended model of goal-directed behavior. Tourism Management, 57, pp. 397-
410.
Posttoday, 2019. Poll suggests deforestation causes floods and landslides (in Thai). In
Posttoday [online] Available at: <[Link]
Ryan, C. and Aicken, M., 2005. Indigenous tourism: the commodification and management of
culture. Amsterdam: Elsevier.
Reisinger, Y. and Steiner, C. J., 2006. Reconceptualizing object authenticity. Annals of
Tourism Research, 33 (1), pp. 65-86.
Santos, C., and Yan, G., 2008. Representational politics in Chinatown: The ethnic other.
Annals of Tourism Research, 35(4), 879–899.
Smith, V., 1977. Hosts and guests: The anthropology of tourism. Philadelphia: University of
Pennsylvania Press.
Trilling, L., 1972. Sincerity and Authenticity. London: Oxford University Press.
Wang, N., 1999. Rethinking authenticity in tourism experience. Annals of Tourism Research,
26 (2), pp. 349 – 370.

905
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Wang, Y., 2007. Customized authenticity begins at home. Annals of Tourism Research, 34
(3), pp. 789 – 804.
Zhang, S. N., et al., 2019. How does authenticity enhance flow experience through perceived
value and involvement: the moderating roles of innovation and cultural identity. Journal of
Travel & Tourism Marketing, 36 (6), pp. 710-728.
Zhou, Q., Zhang, J., Zhang, H. and Li, X., 2018. Is all authenticity accepted by tourists and
residents? The concept, dimensions and formation mechanism of negative authenticity.
Tourism Management, 67, pp. 59-70.
Zhou, Q., Zhang, J., Zhang, H. and Ma, J., (2015). A structural model of host authenticity.
Annals of Tourism Research. 55, pp. 28 – 45.
Zhu, Y., 2012. Performing heritage: rethinking authenticity in tourism. Annals of Tourism
Research, 39 (3), pp. 1495 – 1513.

906
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

ECONOMIC DEVELOPMENT AND NATIONALISM: IN THE

CASES OF CATALONIA AND SCOTLAND

Yusuf Omur Yilmaz†

Mardin Artuklu University

In globalized world, the number of ethnic nationalist movements in multi-

ethnic countries substantially increase. It is widely believed that economic

developments in regions where ethnic minorities live weaken these movements. The

aim of this study is to discuss whether economic developments in a multi-ethnic

country, like the UK and Spain, affect ethnic nationalist movements. For this purpose,

disparity in economic development indicators, Gini coefficients, GDP per capita,

natural resources and unemployment rates - between regions, Scotland and Catalonia

and countries, Spain and the UK and for ethnic nationalism indicators, public opinions

and the ratios of votes obtained by nationalist parties in elections, are used to discuss

this discourse in both cases. It is indicated that Catalonia and Scotland are

economically developed over the period. On the other hand, ethic nationalist

indicators considerably increase in both regions. It is concluded that the above

discourse is not supported by the economic development indicators and ethnic

movement indicators. Moreover, it is discussed whether economy is another reason

for independence-seeking in Scotland and Catalonia or not.

†
Corresponding Author. Address: University of Mardin Artuklu, Department of Economics, 47200
Mardin , Turkey, E-mail: yusufomuryilmaz@[Link]

907
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Key Words: Development, GDP per capita, Gini Coefficient, Nationalism

JEL Codes: F52, P16, P48

908
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

1. Introduction

The modern world history has several impressive and revolutionized

milestones, such as the French Revolution, the first and the second world wars. These

along with historical cases have fundamentally changed the structure of states. The

French Revolution, for example, led to the gradual destruction of the Ottoman and

British empires. The Revolution gave consciousness of being a nation to people and

made the concept of nationalism came on the world stage. This concept created

disastrous effects and dissolved states where multi-ethnic societies lived. World War

1 (WW1) demolished the two major empires and weakened all the states involved in

the war. Thus, new nation states came up and the colonizers of the states changed

hands. As one of the results of World War 2 (WW2), the other ethnic groups, had not

obtained their independencies after WW1, became independent with some

exceptions, Catalonia, Basque, Scotland and Northern Ireland.

Following WW2, regionalism accelerated across the entire world. This

concept became reality through European Economic Community, Arab League and

Common-wealth Countries. What the regionalism provides the benefitted parties in

the Regionalist Institutions is to create a common identity and action in a specific

region1. In the 1980s, globalization started to challenge acquis of regionalism for the

parties. The benefited states thought independently from their regionalist institutions

1
Ethier, W.J. (1998), The International Commercial System, Essays in International Financial System
pg.11.

909
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

and they sought to maximize their own states’ interests. Globalization and

regionalism affect not only sovereign countries, but also ethnic minorities.

Needless to say this, there have been a couple of fully homogenous states

where people shared the same ethnic background and same religious belief. Japan and

Korea might be on the top of this list. States applied the modernization apparatus,

mainly national education, to discard the obstacle of multi-ethnicity. However, not all

the states applied the policy successfully. While France was a successful example, the

UK and Spain were the opposite examples where tension stemming from both ethnic

and religious minorities is there. Central governments tried to solve the ethnic issues

through different ways, from military way to federalization. However, no states can

solve out ethnic issues completely. Mostly ethnic tensions stemmed from cultural and

political differences between politically dominated societies and minorities. The

political, cultural and historical differences have become main motivations for ethnic

groups to demand independence so far.

Regional investment/divestment and economic enrichment/ impoverishment

relative to the center have an effect on ethnic consciousness. As people get richer,

they may forget their ethnicity, and integrate into bigger cultural environment, the

center. Contrary to this rationale, people may be selfish, not want to share richness

with the bigger environment. Thus, the ethnicity feeling may flourish. On the other

hand, if the center gets richer, social dynamics, integration or disintegration, may run

conversely according to the above dynamics mentioned. What matters is that the

region or center gets relatively wealthier. In this paper, it is discussed whether changes

910
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

in economic indicators affect ethno-nationalist movements or not. For this purpose,

the cases of Catalonia and Scotland are analyzed. Moreover, economic roots behind

Scottish and Catalan nationalist movements are also discussed.

The paper is structured as follows. Section 2 provides literature review on the

relation between economic development and nationalism. Section 3 discusses

whether economic indicators affect nationalist movements in Scotland and Catalonia.

In Section 4, it is evaluated that economy is a reason for independence-seeking in

Scotland and Catalonia. Section 5 concludes the paper.

2. Literature Review

It is often argued that political, cultural and historical factors drive ethnic

movements. Although there exists a consensus on the importance of political, cultural

and historical factors, which lead to emergence and development of ethic awareness,

scholars still debate on the significance of economic factors. While some scholars say

that disparity of economic development between region where ethnic minorities live

and state push desire of ethnic movements, the others do not share the same opinion.

Hechter (1975) and Gourevitch (1979) highlight opposite views on the

relation between economic conditions and autonomy movements. Hechter (1975)

argues that disparity in the industrialization level across the regions leads to different

welfare levels across these regions. While some regions become poor, the others

become rich. He expresses that if the poor regions have distinct ethnic background

911
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

from the rest of state, the solidarity and intra-group communication between residents

in poor regions arise. Thus, autonomy movements emerge and develop in such

economic environment. On the other hand, Gourevitch (1979) argues that people

which live in the poor regions prefer benefiting from richer regions instead of

demanding autonomy.

Some writers move one step forward than nationalism and argue the effect of

economic inequality under the framework of secession concept. Roehner (2002)

discusses that validity of the economic inequality argument mentioned the above,

between ethnically-divided societies within a country, leads to separatist actions, and

he says that the separatist movements arise from the state belonging to peculiar

geography and history to the related ethnic society rather than economic inequality2.

Bookman (1992) argues that perception of economic advantages is very important for

secessionist movements. He says that the movements can be seen in both poor and

rich countries. For example, poor regions aim to depart from the central when they

think that the departure provides more welfare and richness themselves. On the other

hand, rich regions also want to depart because they think that poor regions benefit

from the budget more than their contributions to it, and consume the portions of rich

regions.

Gurr (2000) emphasizes that ethnic tensions stem from discriminatory actions

against ethnic societies, and he argues that the majority prevents minority from

2
Roehner, B.M. (2002), Separatism and Integration, A Study in Analytical History (Lanham,
Maryland:Rowman & Littlefield, ).

912
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

obtaining economic goods, conditions and positions3. Thus, people develop their

ethnic conscious on facing economically based discrimination. Crawford (1998) puts

forwards a similar argument. He argues that once cultural groups experience any

discrimination, this discrimination shapes political groups4. The discrimination may

range from economic, political to cultural factors. So, any economic disparity may

lead to ethnic movements between political groups. Wilmsen (1996) argues that

people tend to act with solidarity when they are threatened through economic, social

and political forces5. Thus, the ethnic movements can be seen in such an environment.

Fearon and Houten (2002) analyze the relation between regional GDP and regional

party success in 136 regions and 9 industrialized countries, and they find that there is

a positive relation between regional GDP and the number of the votes6.

Nagel and Olzak (1982) discuss the effect of urbanization on ethnic tension in

three aspects, urban immigration, organizational development and economic

competition7. They state that ethnic solidarity helps people move into inner-cities,

organize better ethnic networks and lead to ethnic competition for positions. Thus,

urbanization increases common sense of sharing the same ethnicity. Hunter and Dahl

3
Gurr, T. (2000), Peoples Versus States, Minorities at Risk in the New Century (Washington, D.C.
United States Institute of Peace Press).
4
Crawford,B. (1998). The Causes of Cultural Conflict: An Institutional Approach in Beverly Crawford
and Ronnie Lipschutz, eds., The Myth of ‘Ethnic Conflict’: Politics, Economics, and ‘Cultural’ Violence
pg 38 and 556
5
Wilmsen, E.N. (1996) “Premises of Power in Ethnic Politics,” in Edwin N. Wilmsen and Patrick
McAllister.
ed., The Politics of Difference, Ethnic Premises in a World of Power (Chicago: The University of
Chicago Press), pp. 4-5.
6
Fearon, J. D. and Pieter von Houten (2002) the Politicization of Cultural and Economic Difference, A
Return to the Theory of Regional Autonomy Movements, Laboratory in Comparative Ethnic
Processes: Stanford University).
7
Nagel, J. and Olzak, S. (1982) Ethnic Mobilization in New and Old States: An Extension of the
Competition Model, pg 131-132

913
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

(1962) and Wolfinger (1965) work on the American society and support Nagel and

Olzak (1982)’ statement. On the other hand, there is opposite statement on relation

between the level of urbanization and ethnic awareness. Green (2011) studies the

Turkish data set and finds that as people become urbanized, they put their ethnic

identity way and are assimilated8.

3. Political and Economic Indicators

In this section, political and economic indicators, regional election results,

regional poll results on independence for Scotland and Catalonia, GDP per capita,

unemployment rate and Gini coefficient for Scotland, Catalonia, the UK and Spain

are presented and the economic figures for each region are compared to the ones for

country to which each region belongs to analyze whether disparity the figures

between the region and state causes increase in nationalist movements in both cases

or not.

Scotland

Table 1 shows the ratios of parties’ votes taken in the regional elections in the

selected years during the period of 1999-2016. As can be seen from the table, the total

number of votes taken by independence-seeking parties increases during the period.

8
Green, E.,(2011), Endogenous Ethnicity, London: Department of International Development,
London School of Economics pp. 19

914
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

While it is just 20.9 percent in 1999, it increases to 48.4 in 2011 and remains nearly

the steady during the next five years. It can be concluded that nationalist rhetoric

gains importance over the period of seventeen years.

Table 1: the ratio of votes in favor of independence in the regional elections,

Scotland9

Year

1999 2003 2007 2011 2016

Scottish Nationalist Party (SNP) 27.3 20.9 31 44 41.7

Scottish Green Party (SGP) 3.6 6.9 4 4.4 6.6

Source: Data are retrieved from [Link]

Table 2 summarizes polls on support of Scotland independence over the period

between 1999 and 2016. As can be shown from Table 2, while 27% of the Scottish

People supported independence in 1999, this ratio increased to 46 % in 2016. In 2014,

the Scotland Independence Referendum was held and 45 % of people vote for

independence. It can be concluded that the number of independence supporters

increase during the period.

9
Scottish Nationalist Party and Green Party only support for independence in Scotland ( See
[Link] for SNP and
[Link]
content/uploads/2013/09/[Link] for SGP)

915
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 2: Support of Scotland Independence: 1999-2016

Opinion 1999 2003 2007 2011 2016

Independence 27% 26% 24% 37 % 46 %

Devolution 59% 56% 62% 55% 42 %

No Scottish Parliament 10% 13% 9% 8% 8%

Source: Data are retrieved from [Link]

[Link] pg 5

Tables 3 and 4 present economic development indicators-GDP per capita,

unemployment rate and Gini coefficients- which summarize levels of economic

development in Scotland and the UK between 1999 and 2016. Using three indicators,

economic disparity between Scotland and the UK is analyzed. GDP per capita for

both Scotland and the UK increase during the selected period. However, the one for

UK is higher than the one for Scotland over the period. For unemployment rate, in

1999, 2004 and 2016, the rate is higher in Scotland and in the rest of the selected

years, the rate is higher in the UK. It can be concluded that unemployment rate for

Scotland fluctuates over the period and it can be stated that it is not too high. For Gini

Coefficient, the figures for Scotland are better than for the UK except for the one in

2015/6. As can be concluded from the analysis that which economy is better than

another one is difficult analysis, but it can be said that both economies are in good

circumstances.

Together with the results in Tables 1 and 2, increase in the GDP per capita

Scotland, and good circumstances in both unemployment rate and Gini coefficient do

916
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

not lead to decline in nationalist movements. It can be concluded that economic well-

being does not prevent nationalism in Scotland.

Table 3: GDP per capita and Unemployment rate in Scotland and the UK

1999 2004 2007 2011 2016

GDP per capita in Scotland 15743 20180 23486 24547 28027

GBP UK 17720 22020 25275 28421 30397

Scotland 7,1 5,7 4,8 8,1 5,2

Unemployment Rate UK 6 4,8 5,4 8,3 5,1

Source: Office for National Statistics - Regional economic activity by gross domestic product, UK:
1998 to 2018. Data retrieved from
[Link]
lnutslevelregions and ILO unemployment rates for Scotland and UK
[Link]
meseries/mguk/lms and
[Link]
n/lms

Table 4: Gini coefficients in Scotland and the UK

1998/9 2003/4 2006/2007 2010/2011 2015/2016

Gini Scotland - 31 % 32 % 30 % 34 %

Coefficient UK 35.4% 34 % 34.7% 33.7 % 31.6 %

Source: Gini coefficients are retrieved from Scottish Government website and ones for the
UK are retrieved from Office for National Statistics website:
[Link]
omeandwealth/datasets/nowcastinghouseholdincomeintheuk

917
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Catalonia

Table 5 indicates the ratios of votes obtained by independence-seeking parties in

the regional elections of Catalonia in the selected. According to Table 5, the ratio of

parties, which are favor of independence, is 9.7 percent in 2006 and considerably

increases to 47.71 percent in 2017.

Table 5: the ratio of parties’ votes in favor of independence in the regional

elections, Catalonia10

Year

2006 2010 2012 2015 2017

Republican Left Party 9.7 7.2 13.85 - 21.47

Popular Unity Party - - 3.52 8.25 4.48

Together for Yes - - - 39.79 -

Together for Catalonia - - - - 21.76

Source: Regional election results for Catalonia are retrieved from

[Link]

10
In Catalonia, independence-seeking parties are as follow: Together for Catalonia movement,
Republican Left of Catalonia, Popular Unity Candidacy in 2017 regional election. (See Ortega, M.
(2017). The Political Rights of EU Citizens and the Right of Secession. In C. Closa (Ed.), Secession from
a Member State and Withdrawal from the European Union: Troubled Membership (pp. 134-152).
Cambridge: Cambridge University Press). In 2015 election, Together for Yes alliance and Popular Unity
Candidacy are pro-independence movements. (See [Link]
Catalan%20elections%20three%20possible%20scenarios%20that%20could%20shape%20Catalonias
%[Link]). Republican Left Party and Popular Unity Party are pro-independence parties (See
[Link] and [Link]

918
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Table 6 shows polls on which kind of political entity Catalan want between 2006

and 2014. While 13,9 % of Catalan people who supported independence state in 2006,

this ratio increased considerably and reached 39.1 and 37.1 in 2015 and 2017.

Moreover, in October 2017, Catalonia Independence Referendum is held. In the

referendum, 90.2 percent of people who are favor of independence in Catalonia

independence referendum. According to the figures in Table 5 and 6, the number of

independence supporters increase in Catalonia over the years.

Table 6: Polls on which kind of political entity Catalan want between

2006-2017

Federal Autonomous Spanish

Year Independent State State State Region

March

2017 37.1 21.7 28.5 4.1

2015 39.1 26.1 24 7,4

2012 29 30,8 27,8 5,2

2010 19,4 29,5 38,2 6,9

2006 13,9 33,4 38,2 8,1

Source: Data are retrieved from

[Link] pg 911

Table 7 summarizes the economic development indicators- in Catalonia and

Spain in the selected years. Over the years, GDP per capita increases and

11
In the source, data are shown on the graph. Using data on the graph, Table 6 is generated.

919
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

unemployment rate decreases for both Catalonia and Spain. Gini coefficient fluctuates

for Catalonia and Spain. It can be concluded that economic development indicators

are well above in Catalonia than in Spain. Economic well-being of Catalonia relative

to the one of Spain and the increase in nationalist movement co-move. It can be

claimed that economic welfare does not weaken nationalist movement for Catalonia.

Table 7: GDP per capita, Unemployment rate, Gini Coefficient in Catalonia

and Spain

2006 2010 2012 2015 2017

Catalonia 28.168 28040 27098 29.009 31.208

GDP per capita Spain 22.630 23040 22050 23.220 24.970

Catalonia 6,5 17,7 22,5 18,6 13,4

Unemployment rate Spain 8,45 19,86 24,79 22,06 17,22

Catalonia - - - 32,3 31,8

Gini coefficient Spain 33,5 35,2 35,4 36,2 34,7

Sources: GDP per capita in terms of Catalonia and Spain data are retrieved from
[Link]
Unemployment rate data for Catalonia is retrieved from
[Link] , Unemployment rate data for Spain is
retrieved from [Link] Gini
coefficient for Spain is retrieved from
[Link] and Gini Coefficient for Catalonia
is retrieved from [Link]

4. Economic Reasons Behind Scottish and Catalan Nationalisms

It is undeniable that political, cultural and historical reasons are behind the

increase in the nationalist movements in both Scotland and Catalonia. In this part, it

is evaluated that economic reason can be added to the reasons mentioned above or

920
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

not for both Scotland and Catalonia concerning factors, which are liberalization,

natural resource and financial imbalance. While liberalization and natural resource

factors are used for Scotland, liberalization and fiscal imbalance factors are used for

Catalonia in the analysis below.

[Link]

As widely known, liberalization mainly has two pillars: Economic and Political

ones. In this part, it is discussed whether economic liberalization and nationalist

movements co-move or not in Scotland and Catalonia, and whether economic

liberalization is one of the reasons of nationalist movements in both regions or not.

Scotland

After the Thatcher government came to power in 1979, Thatcher

fundamentally changed the structure of the economic system and followed

liberalization policies (i.e. privatization, deregulation, low taxation and weakening

unions so on). These policies weakened British welfare state. Paquin (2002) states

that the changes lead Scots to benefit less from the welfare state compared to the

previous circumstance. This caused two important consequences. First, discontent

between Scotland and England escalated and second voting behaviors of Scots

changed12. The number of votes obtained by parties which concern policies related to

12
Holitscher, M. and Roy, S. 'The Paradox of Economic Globalisation and Political Fragmentation:
Secessionist Movements in Quebec and Scotland', Global Society, Vol.13, No.3 (1999), p.251

921
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

domestic issues increase like Labor Party and Scottish National Party13. In other

words, decline in the welfare state makes people follow more Scotland-oriented

parties and policies.

Figure 1: Export Volume in Scotland in GBP between 2000 and 201814

100000

80000

60000

40000

20000

2014
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013

2015
2016
2017
2018
Scotland

Source: Data are retrieved from the Statista website:

[Link]

Along with detrimental effects of decline in the welfare state to Scotland, as

can be seen in Figure 1, export volume considerably increases in Scotland starting

from 1990 despite small fluctuation in 2010. Figure 1 indicates that it is around £ 45

billion in 2000, it increases to around £ 88 billion in 2018. It can be said that export

volume nearly becomes two-fold over the period. Concerning the figure, Scotland is

more economically developed compared to the pre-2000 period. Together with the

decline in the welfare state, the increase in export volume may lead people to think

13
Stéphane P. (2002) Globalization, European integration and the rise of neo‐nationalism in Scotland,
Nationalism and Ethnic Politics, pg 65.
14
Export volume is the sum of both export to the the rest of the UK and export to the rest of World
for Scotland.
15
In the source, data are shown in a graph. Using the data in the graph, Figure 1 is drawn.

922
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

that Scotland can economically survive after independence. This thought is supported

by increase in the number of voted obtained by independent seeking parties.

Catalonia

Spain implemented step by step trade liberalization policies from 1959 to

1986. In 1986, Spain accessed into European Union, and Spain integrated Peseta into

European Monetary System in 1989. Then, Spain accepted Euro as its currency in

199916. All these policies make Spain and Catalonia open-market economy.

Figure 2: Export Volume in Catalonia between 2000 and 2018

80000

70000

60000

50000

40000

30000

20000

10000

Catalonia

Source: Data are retrieved from

[Link]

16
Aguado, S. (2002). Spain in Globalization Process. the Conference on Spanish Investment, Miami
European Union Center and the FIU Latin American and Caribbean Center, Miami, 18-19 October
2002.

923
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

Figure 2 indicates export volume in Catalonia between 2000 and 2018. In

2000, export volume is around € 34 billion in Catalonia while it is around € 71 billion

in 2018. This volume continuously increases and makes Catalans richer during the

period. The increase in export volume coincides with the increase in the number of

votes obtained by independence-seeking parties in Catalonia.

The figures indicate that Scotland and Catalonia are economically developed

over the years. Concerning the discourse, nationalist movements should have

decreased in Scotland and Catalonia, but in fact according to the above tables, they

increase during the period. In my opinion, there could be two reasons why economic

liberalization stimuluses nationalist movements. First, since welfare state demolished

in the UK, benefits of staying with the UK decline for Scotland. Second, increase in

trade volume in Scotland and Catalonia can imply that both regions are less dependent

on the UK and Spain and they do not share welfare created by the increase in trade

with the UK and Spain. it can be claimed that benefits of staying with the UK and

Spain would not overwhelm the benefits of being an independent state. Thus, it can

be said that liberalization could be one of the reasons behind increase in nationalist

movements in Scotland and Catalonia.

4.2. Natural Resource (i.e. North Sea Oil)

The question of who owns the property rights of North Sea Oil Petroleum has

been a matter of debate between Scotland and the United Kingdom. This issue is not

only a subject of political debate, but also has an economic dimension. The table

924
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

below shows the UK's income from North Sea Oil between 2008/9 and 2018/9. Over

the years, although income has fluctuated, income is significant, except for 2015/2016

and 2016/2017 and the yearly average revenue is around £ 4,770 billion. Even if Scots

do not directly receive the revenue of the North Sea Oil, they indirectly receive a share

of this income. Despite this, the question related to the royalty right is important topic

in Scottish politics. According to the simulation conducted by Scotland Government,

if the royalty right belongs to Scotland, this ratio reduces over the selected years. it

can be argued that a significant part of this oil field, in the case of independence,

geographically is thought to remain dominated by Scotland. This income could be

one of the most important motivations for the Scots to seek independence.

Table 8: Annual North Sea Oil Revenue: 2008/9 to 2018/9 in GBP

2008/9 2009/10 2010/1 2011/2 2012/3 2013/4 2014/5 2015/6 2016/7 2017/8 2018/9

Revenue 12,456 5,991 8,406 10957 6,020 4,499 1,691 -85 36 1,297 1,237

Source: Data is retrieved from Statistia website [Link]

kingdom-uk-north-sea-revenue/ 17

Table 9: Budget Deficit in Scotland 2014-15 to 2018-19

2014/5 2015/6 2016/7 2017/8 2018/9

Scotland - Excluding North Sea -8.9% -8.7% -7.1% -6.8% -8.5%

Scotland - Including North Sea -7.3% -8.0% -6.4% -5.5% -7.0 %

(geographical share)

Source: Table is taken from Government Expenditure and Revenue in Scotland (GERS): 2018 to
2019 report [Link]

17
In the source, figures are shown in a graph. Using the figures in the graph, Table 8 is generated.

925
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

4.3. Fiscal Imbalance

Catalonia region of Spain is historically rich compared to the rest of

country. Catalans claim that although they send tax levied in Catalonia, central

government does not transfer money proportionally compared to tax revenue

collected from Catalonia. Thus, fiscal imbalance between Catalonia and Spain comes

out. Pons-i-Novell and Tromesa-i-Balcells (2005) indicate that fiscal imbalance

between Catalonia and Spain fluctuate between 4 and 10 percent of Catalonia GDP

over the period of 1990 and 2004. This level of the imbalance could be another reason

for Catalans to seek independence.

5. Conclusion

In this paper, it is discussed whether economic factors have influence

ethnic nationalism in Scotland and Catalonia or not. It can be concluded that as

Catalan and Scottish economies develop, desire for independence increases in both

Catalonia and Scotland. It implies that economic development does not reduce

nationalist movements. Moreover, it is also discussed and can be claimed that

economic development is another reason for independence-seeking. The intuitions

behind this are that people do not voluntarily share their economic wealth with the

majority of people who have a different ethnic background even if they share the same

national borders.

926
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

REFERENCES

1. Aguado, S. (2002). Spain in Globalization Process, the Conference on Spanish

Investment, Miami European Union Center and the FIU Latin American and
Caribbean Center, Miami.
2. Annual government revenues from offshore oil and gas activity on the North
Sea in the United Kingdom (UK) from 2008/09 to 2018/19,
[Link]
revenue/
3. Bookman, M. Z. (1982), The Economics of Secession, New York: St. Martin’s
Press).

4. Cetra, D., Catalan elections: three possible scenarios that could shape
Catalonia’s future, [Link]
Catalan%20elections%20three%20possible%20scenarios%20that%20could
%20shape%20Catalonias%[Link]
5. Constitution, Scottish National Party, [Link]/policy-area/constitution/
6. Crawford, B. (1998), The Causes of Cultural Conflict: An Institutional
Approach, in Beverly Crawford and Ronnie Lipschutz, (eds.), The Myth of
‘Ethnic Conflict’: Politics, Economics, and ‘Cultural’ Violence.
7. Eleccions al Parlament de Catalunya. (n.d.). Retrieved from Història
Electoral: [Link]
8. Ethier, W.J. (1998). The International Commercial System, Essays in
International Financial System pg.11.
9. Exports. By economic destination of the goods. (n.d.). Retrieved from
Statistical Institute of Catalonia:
[Link]
10. Fearon, J. D. & Pieter von Houten (2002). the Politicization of Cultural and
Economic Difference, A Return to the Theory of Regional Autonomy
Movements, Laboratory in Comparative Ethnic Processes: Stanford
University.

927
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

11. GINI index (World Bank estimate). (n.d.). Retrieved from Worldbank:
[Link]
12. Gourevitch, P.A. (1979), the Re-Emergence of “Peripheral Nationalisms”:
Some Comparative Speculations on the Spatial Distribution of Political
Leadership and Economic Growth. Comparative Studies in Society and
History, 21 pg 303–322.
13. Green, E. (2011). Endogenous Ethnicity, London: Department of
International Development, London School of Economics pg. 19.
14. Gross domestic product (GDP) per capita in Catalonia, Spain as a whole and
the Eurozone between 2004 and 2018. (2020). Retrieved from Statistia:
[Link]
eurozone/
15. Gurr, T. R. (2000) ed., Peoples Versus States, Minorities at Risk in the New
Century, Washington, D.C. United States Institute of Peace Press.
16. Hechter, M. (1975). Internal Colonialism: The Celtic Fringe in British
National Development, (pg. 1536-1966), Berkeley: University of California
Press.
17. Holitscher, M. and Roy, S. (1999), The Paradox of Economic Globalisation
and Political Fragmentation: Secessionist Movements in Quebec and
Scotland', Global Society, 13(3), pg.251.
18. Hunter, F. And Dahl, R. A. (1962). Who Governs: Democracy and Power in
an American City. Administrative Science Quarterly. 6(4), Johnson Graduate
School of Management, Cornell University.
19. Inequality of income distribution indicators . (n.d.). Retrieved from Statistical
Institute of Catalonia:
[Link]
20. LFS: ILO unemployment rate: Scotland: All: %: SA. (n.d.). Retrieved from
Office for National Statistics:
[Link]
mployment/timeseries/ycnn/lms
21. LFS: ILO unemployment rate: UK: All: Aged 16 and over: %: NSA. (n.d.).
Retrieved from Office for National Statistics:

928
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

[Link]
mentandemployeetypes/timeseries/mguk/lms
22. Nagel, J. and Olzak, S. (1982). Ethnic Mobilization in New and Old States:
An Extension of the Competition Model, pg 131-132.
23. Nowcasting household income in the UK . (n.d.). Retrieved from Office for
National Statistics:
[Link]
oldfinances/incomeandwealth/datasets/nowcastinghouseholdincomeintheuk
24. Ortega, M. M. (2017). The Political Rights of EU Citizens and the Right of
Secession. In Closa, Carlos (ed.). Secession from a Member State and
Withdrawal from the European Union: Troubled Membership, (pg. 134-152),
Cambridge University Press.
25. Paquin, S. (2002), Globalization, European integration and the rise of neo‐
nationalism in Scotland, Nationalism and Ethnic Politics, pg 65.
26. Political Context Survey. (2018). Retrieved from Generalitat de Catalunya:
[Link]
[Link]
27. Pons-i-Novell, J. and Tremosa-i-Balcells, R. (2005). Macroeconomic effects
of Catalan fiscal deficit with the Spanish state (2002–2010), Applied
Economics, 37(13), pg 1455-1463.
28. Regional gross domestic product all NUTS level regions. (n.d.). Retrieved
from Office for National Statistics:
[Link]
lgrossdomesticproductallnutslevelregions
29. Result Analysis. (n.d.). Retrieved from Scotland Parliament:
[Link]
30. Roehner, B .M. (2002), Separatism and Integration, A Study in Analytical
History, Lanham, Maryland:Rowman & Littlefield, pg. 11.
31. Scotland: total value of annual exports from 2000 to 2018 Published by Daniel
Clark, Jun 25, 2019 This statistic shows the total value of Scottish exports
from 2000 to 2018. The value of Scottish exports has grown continually over

929
32nd EBES Conference Proceedings - Volume I August 5-7, 2020

the period, except for . (n.d.). Retrieved from Statistia:

[Link]
32. Scottish Government. (n.d.). Retrieved from Government Expenditure and
Revenue in Scotland (GERS): 2018 to 2019:
[Link]
scotland-gers/

33. Scottish Independence, Scottish Green Party,

[Link]
uk/wp-content/uploads/2013/09/Scottish-Greens-Independence-Briefing-
[Link]
34. Statistia. (n.d.). Retrieved from Annual government revenues from offshore
oil and gas activity on the North Sea in the United Kingdom (UK) from
2008/09 to 2018/19: [Link]
kingdom-uk-north-sea-revenue/
35. The State of Nationalism in Scotland. (n.d.). Retrieved from Scottish Social
Attitudes: [Link]
[Link]
36. Unemployment rate. (n.d.). Retrieved from Statistical Institute of Catalonia:
[Link]
37. Unemployment, total (% of total labor force) (modeled ILO estimate) - Spain.
(n.d.). Retrieved from Worldbank:
[Link]

38. What is CUP, Popular Unity Candidacy, [Link]

39. What we are, Republican Leftist Party, [Link]
40. Wilmsen, E. N. (1996). Premises of Power in Ethnic Politics,” in Edwin N.
Wilmsen and Patrick McAllister, ed., The Politics of Difference, Ethnic
Premises in a World of Power, Chicago: The University of Chicago Press.
41. Wolfinger, R. (1965). The Development and Persistence of Ethnic
Voting. American Political Science Review, 59(4).

930

Common questions