Insurance and microinsurance of agricultural risks :

public-private partnerships and innovative business

models
Pierre Casal Ribeiro

To cite this version:

Pierre Casal Ribeiro. Insurance and microinsurance of agricultural risks : public-private partnerships
and innovative business models. Business administration. Université de Nanterre - Paris X, 2021.
English. �NNT : 2021PA100149�. �tel-03817171�

HAL Id: tel-03817171

[Link]
Submitted on 17 Oct 2022

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 E D 396 : Ecole Doctorale
Economie, organisations, société
EA4429 – Centre d’Etudes et de
Recherches sur les
Organisations et la Stratégie

Membre de l’université Paris Lumières

Pierre Casal Ribeiro

Insurance and microinsurance of
agricultural risks
Public-private partnerships and innovative business models
Assurance et microassurance des risques agricoles
Partenariats public-privé et modèles économiques innovants

Thèse présentée et soutenue publiquement le 08/12/2021

en vue de l’obtention du doctorat de Sciences de gestion de l’Université Paris
Nanterre
sous la direction de M. Didier Folus (Université Paris Nanterre)

Jury :

Rapporteur : M. Philippe Dessertine Professeur des Universités, Université

Paris I - Panthéon Sorbonne

Rapporteure : Mme Isabelle Guérin Directrice de Recherche, Institut de

Recherche pour le Développement

Directeur de thèse : M. Didier Folus Professeur des Universités, Université

Paris Nanterre

Membre du jury M. Marek Hudon Professeur des Universités, Université

Libre de Bruxelles

Membre du jury : M. Bruno Lepoivre Directeur adjoint Marchés agriculture et

professionnels, Pacifica
Résumé

L’observation des marchés de l'assurance agricole dans le monde révèle qu'à l'exception des
polices d'assurance contre les risques nominaux, l'assurance agricole ne s'est développée à
grande échelle qu’en présence d’un appui public, prenant généralement la forme de subventions
à la prime. Ce constat est vrai dans les économies développées, émergentes et en
développement. Par ailleurs, la situation des petits agriculteurs pose des défis supplémentaires
pour étendre la couverture d'assurance à un grand nombre d'agriculteurs. Plusieurs décideurs
politiques dans le monde ont décidé de soutenir activement l'assurance agricole afin d'améliorer
la protection des agriculteurs. Cette évolution pose la question du modèle économique de
l'assurance agricole, et des modalités des partenariats public-privé. Cette thèse passe en revue
les motivations et les modalités de l'intervention publique sur les marchés de l'assurance
agricole. Les questions autour du modèle économique et des partenariats public-privé de
l'assurance agricole sont ensuite analysées autour de deux études de cas et d'une enquête par
sondage. Le travail de recherche évalue la possibilité de remplacer un programme public
d'assistance en cas de catastrophe par un produit d'assurance commercial, en prenant le cas de
l'assurance des prairies en France. Il analyse également les produits d'assurance récolte et bétail
proposés par une compagnie de microassurance au Kenya, afin de comprendre comment les
produits et les processus sont conçus pour cibler ce segment spécifique d'agriculteurs. Enfin, il
étudie l’opportunité pour des institutions de microfinance de proposer de la microassurance à
leurs clients, sur la base d'une enquête menée auprès de 36 institutions en Afrique et en Asie.
Le travail de recherche démontre qu'une assurance indicielle peut être une alternative viable à
un fonds public de réponse aux catastrophes pour protéger les agriculteurs. Il montre également
que s'attaquer au segment des petits agriculteurs nécessite un changement d'approche, au-delà
de la simple réduction des primes et des sommes assurées. Il invite également les décideurs
politiques à envisager d'autres leviers d'action que les seules subventions aux primes. Pour les
institutions financières, on constate que le bien-fondé économique de distribuer des polices
d’assurance à leurs clients n’est pas évident. Ce travail contribue à la littérature sur les risques
agricoles et l'assurance, en analysant les enjeux auxquels sont confrontés les opérateurs privés,
tant du côté de l'offre que de la demande. Il ne se limite pas au point de vue des assureurs et des
réassureurs, et inclut des acteurs clés tels que les courtiers d'assurance, les institutions
financières et les acteurs de la chaîne de valeur agricole. Il offre également des perspectives et
des leçons pratiques aux professionnels de l'assurance agricole et aux décideurs politiques.

Mots-clés : assurance agricole, assurance indicielle, partenariats public-privé, microassurance,

modèle économique, petits producteurs

3
Executive summary

A review of agricultural insurance markets around the world reveals that, except for simple
name-peril policies, agricultural insurance has developed at a large scale only in the presence
of government support, usually taking the form of premium subsidies. This observation holds
true in developed, emerging and developing economies. The situation of smallholder farmers
though, brings additional challenges to extending insurance coverage to a large number of
farmers. Several policy-makers around the world have decided to actively support agricultural
insurance in order to improve the protection of farmers. This situation raises the question of the
business model of agricultural insurance, and the modalities of public-private partnerships. This
thesis review public intervention in agricultural insurance markets’ rationale and modalities.
The questions around the business model and public-private partnerships of agricultural
insurance are then analyzed around two case studies and one survey. The research evaluates the
possibility to replace a public disaster assistance program by a commercial insurance product,
taking the case of grassland insurance in France. It analyses the crop and livestock insurance
products offered by a microinsurance company in Kenya, in order to understand how products
and processes are designed to target this specific segment of farmers. Finally, it investigates the
business case for microfinance institutions to offer microinsurance to their clients, based on a
survey of 36 institutions in Africa and Asia. The research demonstrates that an index-based
insurance can be a viable alternative to a public disaster fund to protect farmers. It also shows
that addressing the smallholder farmers segment requires a shift in the approach, beyond simply
scaling down policies. It also invites policy-makers to consider other levers of action than just
premium subsidies. For financial institutions, it is found that the business case of distributing
insurance is not straightforward. This work contributes to the literature regarding agricultural
risks and insurance by analyzing the challenges faced by private operators, on both supply and
demand. It is not limited to the perspective of insurers and reinsurers, and includes key
stakeholders such as insurance brokers, financial institutions and agricultural value chain actors.
It offers practical insights and lessons for agricultural insurance professionals and policy-
makers.
Keywords: agricultural insurance, index insurance, public-private partnerships,
microinsurance, business model, smallholder farmers

4
Remerciements

Je tiens à remercier mon directeur de thèse, Didier Folus, pour ses conseils et ses
encouragements. Merci d’avoir accepté de superviser le travail d’un « vieux » doctorant au
parcours plus pratique qu’académique.

Je remercie vivement Philippe Dessertine et Isabelle Guérin qui ont accepté d’être rapporteurs
de cette thèse, ainsi que Marek Hudon et Bruno Lepoivre pour leur participation au jury.

Merci à Pacifica, en particulier à Jean-Michel Geeraert et Bruno Lepoivre, pour avoir fait le
choix de soutenir la recherche en entreprise et de m’avoir permis de réaliser ma thèse dans le
cadre d’un contrat CIFRE. Je salue mes collègues de la DMAP qui m’ont accueilli avec plein
de bienveillance et qui m’ont tant appris sur le monde agricole. Je remercie Constance Collin,
pour tous ses bons conseils. Merci à Jean-Luc Perron, Eric Campos et Philippe Guichandut de
la Fondation Grameen Crédit Agricole, qui m’ont donné l’opportunité de m’impliquer sur des
projets passionnants, et permis de faire le lien entre terrain et recherche. Je garde un attachement
très fort à la Fondation, à sa mission et aux équipes qui la font vivre.
Merci à l’équipe d’ACRE pour sa disponibilité et sa franchise. Mon séjour au Kenya restera un
très beau souvenir.

Je remercie les amis qui m’ont soutenu, ont relu certains travaux ou m’ont hébergé pendant mes
séjours d’écriture. Je pense en particulier à Jérémy, Wally, Nico, Alessandra et Anna. Rocket,
ta ténacité et ta passion pour la recherche ont été une source d’inspiration au moment de me
lancer dans le travail doctoral. Je remercie mes amis, qui ont entendu parler pendant des années
de ma fameuse thèse comme d’une chose mystérieuse et lointaine, mais en m’encourageant
néanmoins. Impossible de citer tout le monde, mais j’embrasse collectivement les Pulpis, la
bande du Château d’Eau, les copains d’Amplepuis et le groupe des vacances d’été en Espagne
et en Italie.
Merci à Bastien, sans qui je n’aurais jamais réussi à mener cette aventure jusqu’au bout. Tu
m’as soutenu et aidé comme personne.
Merci à Flavien, Anne-Sophie, Nattie, Timéo, Florian, Dorian et Ana Paula.

Enfin, un merci du fond du cœur à mes parents. J’étais certes un gamin qui aimait l’école, mais
rien n’aurait été possible sans le soutien inconditionnel que vous m’avez apporté dans la
poursuite de mes études, sans vos sacrifices, sans vos encouragements, malgré vos propres
doutes et craintes. Votre fierté me touche d’autant plus que je sais tout ce que je vous dois.

5
Table of contents

List of acronyms ..................................................................................................................... 11

List of tables ............................................................................................................................ 12

List of figures .......................................................................................................................... 13

INTRODUCTION .................................................................................................................. 15

CHAPTER 1 ........................................................................................................................... 23

The business model of agricultural insurance: context and problem statement .............. 23

1. Agricultural insurance markets and environment ......................................................... 24

1.1. Risk management in agriculture ............................................................................ 24

1.2. The agricultural insurance landscape ..................................................................... 26

2. The business model of agricultural insurance in discussion ......................................... 30

2.1. Conditions of insurability ...................................................................................... 30

2.2. Insuring the production of smallholder farmers ..................................................... 31

2.3. Can agricultural insurance exist without subsidies? Viability of insurance programs
33

3. Index insurance, a new paradigm for agricultural insurance......................................... 34

3.1. Principles of index insurance ................................................................................. 34

3.2. Benefits .................................................................................................................. 37

3.3. Limitations ............................................................................................................. 37

4. Reinsurance in agriculture ............................................................................................. 39

4.1. Reinsurers as market enablers ................................................................................ 39

4.2. The example of the Spanish scheme ...................................................................... 39

4.3. Interviews with reinsurers ...................................................................................... 43

5. Problem statement ......................................................................................................... 47

5.1. Research hypotheses .............................................................................................. 47

5.2. Data and case studies ............................................................................................. 49

6
 6. Conclusion ..................................................................................................................... 50

CHAPTER 2 ........................................................................................................................... 55

From subsidies to public-private partnerships: public interventions in the agricultural

insurance market .................................................................................................................... 55

1. Context ............................................................................................................................. 56

2. Literature review methodology ........................................................................................ 57

2.1. Research objectives ................................................................................................... 57

2.2. Data ........................................................................................................................... 57

3. Literature review: findings ............................................................................................... 59

3.1. Why should governments subsidize agricultural insurance? .................................... 59

3.2. Are subsidies efficient? ............................................................................................. 65

3.3. Beyond subsidies: other instruments for public intervention .................................... 69

4. Discussion ........................................................................................................................ 75

4.1. Main results ............................................................................................................... 75

4.2. Contribution of the literature review ......................................................................... 76

4.3. Strengths and limits ................................................................................................... 77

4.4. Perspectives ............................................................................................................... 78

5. Conclusion: the agricultural insurance policy mix ........................................................... 78

CHAPTER 3 ........................................................................................................................... 87

From crisis management to risk management in agriculture: the case of grassland

insurance in France ................................................................................................................ 87

1. Introduction ................................................................................................................... 88

2. Policy framework .......................................................................................................... 89

2.1. The World Trade Organization Agreement on Agriculture ................................... 89

2.2. The European Union and the Common Agricultural Policy .................................. 90

2.3. The French framework ........................................................................................... 92

3. Data ............................................................................................................................... 93

3.1. Disaster fund data .................................................................................................. 93

7
 3.2. Index data ............................................................................................................... 94

4. Method .......................................................................................................................... 95

4.1 Private insurance system ........................................................................................ 96

4.2 Public disaster relief system................................................................................... 97

5. Results ........................................................................................................................... 98

5.1 Comparison between insurance products and disaster fund, at the same level of
guarantees ......................................................................................................................... 98

5.2 Modification of insurance product parameters .................................................... 100

5.3 Cost for the State and for the clients .................................................................... 102

6. Discussion ................................................................................................................... 104

6.1 Discussion on the model’s assumptions............................................................... 104

6.2 Evaluating the alternative from the State’s perspective ....................................... 105

6.3 Client value .......................................................................................................... 108

7. Conclusion ................................................................................................................... 110

CHAPTER 4 ......................................................................................................................... 115

Rethinking agricultural insurance for the BoP – a Kenyan case study .......................... 115

1. Introduction ................................................................................................................. 116

2. Literature review ......................................................................................................... 117

2.1. Risk management and microinsurance demand of smallholder farmers ............. 117

2.2. Creating inclusive markets ................................................................................... 120

2.3. Unmet needs are not equivalent of untapped markets ......................................... 122

3. The practitioners’ perspective: assessing client value in microinsurance ................... 123

3.1. Why and how assessing client value .................................................................... 123

3.2. The PACE tool ..................................................................................................... 123

4. Case study: methodology and data .............................................................................. 125

4.1. Research objectives .............................................................................................. 125

4.2. Method ................................................................................................................. 125

8
 4.3. Data collection ..................................................................................................... 126

4.4. Data analysis ........................................................................................................ 128

5. Results ......................................................................................................................... 128

5.1. Description of ACRE ........................................................................................... 128

5.2. 4As Framework analysis ...................................................................................... 133

5.3. PACE tool results ................................................................................................. 137

6. Discussion ................................................................................................................... 141

6.1. Confronting two approaches: the 4As framework and the PACE tool ................ 141

6.2. Limits of the case study ....................................................................................... 143

6.3. Perspectives .......................................................................................................... 144

7. Conclusion ................................................................................................................... 146

CHAPTER 5 ......................................................................................................................... 153

Microfinance institutions and insurance: where is the business case? ............................ 153

1. Introduction: from microcredit to microinsurance ...................................................... 154

2. Offering microinsurance, a non-straightforward path for MFIs ................................. 155

2.1. MFIs’ rationale to include insurance in their service offering ............................. 155

2.2. Organizational set-ups for MFIs .......................................................................... 158

2.3. Challenges faced by MFIs willing to distribute microinsurance ......................... 160

3. The case of agriculture ................................................................................................ 162

3.1. Farmers, underserved among the underserved ..................................................... 162

3.2. Insurance, the silver bullet to unlock agricultural credit markets? ...................... 163

3.3. Organizational options to link credit and insurance in agriculture ...................... 164

3.4. Conclusion ........................................................................................................... 168

4. Evidence from the field ............................................................................................... 168

4.1. Method and sample .............................................................................................. 168

4.2. Results .................................................................................................................. 170

5. Discussion ................................................................................................................... 175

9
 5.1. Sample.................................................................................................................. 175

5.2. Defining the business case of microinsurance for MFIs ...................................... 175

5.3. Sketching a progression path for MFIs ................................................................ 176

6. Conclusion ................................................................................................................... 178

CONCLUSION ..................................................................................................................... 185

BIBLIOGRAPHY ................................................................................................................ 189

APPENDICES ...................................................................................................................... 201

Appendix 1 - Client value assessment approaches ............................................................. 201

Appendix 2 - PACE scoring criteria .................................................................................. 202

Appendix 3 - Online questionnaire filled by MFIs ............................................................ 204

10
LIST OF ACRONYMS

ACRE Agricultural Climate Risk Enterprise Ltd (Kenya)

ANRT Association Nationale de la Recherche et de la Technologie
AoA Agreement on Agriculture
BIP Base Insurance Product (Mongolia)
BoP Base of the Pyramid
CAP Common Agricultural Policy (European Union)
CAT bond Catastrophe-linked bonds
CERISE Comité d’Echanges, de Réflexion et d’Information sur les Systèmes d’Epargne-crédit
CGAP Consultative Group to Assist the Poor
CIFRE Conventions Industrielles de Formations par la Recherhe
CRM Corporate Risk Management
DRP Disaster Response Product (Mongolia)
ECA Europe and Central Asia
EU European Union
FNGRA Fonds National de Gestion des Risques Agricoles (France)
FPI Forage Production Index (France)
GCAF Grameen Crédit Agricole Foundation
IBLI Index-Based Livestock Insurance (Mongolia)
LIIP Livestock Insurance Indemnity Pool (Mongolia)
MENA Middle East and North Africa
MFI Microfinance Institution
MPCI Multi-Peril Crop Insurance
NAIS National Agricultural Insurance Scheme (India)
PAR Portfolio at Risk
RCT Randomized Control Trial
RPG Replanting Guarantee (Kenya)
SSA Sub Saharan Africa
SSEA South and South East Asia
TCIP Turkish Catastrophe Insurance Pool
WTO World Trade Organization

11
 LIST OF TABLES

Table 1 - Agricultural insurance schemes in developing countries.......................................... 30

Table 2 – Producer loss ratios in seven countries .................................................................... 33
Table 3 - Stakeholders' expectations towards a fair insurance index ....................................... 73
Table 4 - Disbursed indemnities over the period by the FNGRA ............................................ 93
Table 5 - Descriptive table of FPI values ................................................................................. 94
Table 6 - Disbursed indemnities over the period for product 1 and the disaster fund. K = 800
.................................................................................................................................................. 98
Table 7 - Disbursed indemnities over the period for product 1, product 2, and the disaster fund.
K = 800 ................................................................................................................................... 101
Table 8 - Total cumulated indemnities in million euros received by the farmers depending on
the sum insured....................................................................................................................... 102
Table 9 - Yearly premium in million euros to cover 10.5 million hectares of pasture .......... 103
Table 10 - Premium cost for the State and for the farmer, depending on the coverage and the
loading factor.......................................................................................................................... 104
Table 11 - Papers assessing the correlation between demand and a series of variables ........ 119
Table 12 – BoP paradigm evolution ....................................................................................... 121
Table 13 – List of interviews.................................................................................................. 127
Table 14 – Summary of products’ features ............................................................................ 132
Table 15 – Summary of the 4As framework analysis ............................................................ 136
Table 16 - PACE tool score of the ACRE products ............................................................... 137
Table 17 - Explanation for PACE scores ............................................................................... 138
Table 18 - PACE tool sub-dimensions correspondence with the 4As framework ................. 141
Table 19 - Descriptive data of the MFIs’ sample, per region................................................. 168
Table 20 - Descriptive data of the MFIs’ sample, per portfolio size...................................... 169
Table 21 - Heterogeneity of the sample ................................................................................. 170
Table 22 – Answers to the question "Do you offer any insurance product(s) to your clients? NB:
This includes mandatory credit life policies offered to borrowers." ...................................... 171
Table 23 - Answers to the question "Do you have any current or future plans to expand your
insurance offering?" ............................................................................................................... 173
Table 24 - Answers to the question "Do you have any current or future plans to start offering
insurance to your clients?" ..................................................................................................... 173

12
LIST OF FIGURES

Figure 1 - Estimated agricultural insurance premiums worldwide in 2005 and 2011 ............. 29
Figure 2 - Estimated geographic distribution of agricultural insurance premiums in 2011 ..... 29
Figure 3 – The Spanish Public-Private Partnership ................................................................. 40
Figure 4 – Agroseguro premiums............................................................................................. 42
Figure 5 – Agroseguro technical results ................................................................................... 42
Figure 6 : Flow chart of literature search and selection ........................................................... 59
Figure 7 – Agricultural risk layering ........................................................................................ 71
Figure 8 – The agricultural insurance policy mix .................................................................... 79
Figure 9 - Yearly distribution of FPI values ............................................................................ 95
Figure 10 - Comparison of disbursed indemnities over the period between product 1 and the
disaster fund. K = 800 .............................................................................................................. 99
Figure 11 - Comparison of disbursed indemnities over the period between product 1, product 2
and the disaster fund. K = 800€.............................................................................................. 101
Figure 12 - PACE added value analysis framework .............................................................. 124
Figure 13 – ACRE’s role in the value chain .......................................................................... 130
Figure 14 – Hybrid cover payout structure ............................................................................ 132
Figure 15 – PACE results ....................................................................................................... 138
Figure 16 - Overlap between the 4As framework and the PACE tool ................................... 143
Figure 17 - Answers to the question: " What are your main reasons for offering insurance to
your clients? Please select up to 3 answers." ......................................................................... 171
Figure 18 - Answers to the question "Based on what you know about your clients, what would
you say are the risks they would be most interested in covering through insurance? Please select
up to 3 answers." .................................................................................................................... 174
Figure 19 - Answers to the question "In which areas do you think your organization needs more
support? Please select up to 3 answers." ................................................................................ 175
Figure 20 - Client value approaches ....................................................................................... 201
Figure 21 - PACE scoring criteria and sub-dimension weights ............................................. 202

13
14
INTRODUCTION

15
Introduction

Agriculture represents 28% of the world’s total employment (International Labour

Organization, 2018), and remains a source of livelihood for an even larger part of the humanity.
Millions of people’s income and livelihoods are vulnerable to natural hazards like droughts,
floods or pests. Agricultural insurance can play a role in helping farmers, herders and their
communities to mitigate the negative financial impact of such adverse events.

Farmers are exposed to several types of risks. Agricultural risks are traditionally classified into
four categories (Antón, 2009): production risks, market risks, financial risks and legal and
institutional risks. Insurance can only address production risks. Those risks have an impact on
the quantity or the quality of the farm output. Typically, an insurance policy protects farmers
against weather hazards, natural catastrophes, pests or diseases affecting yields or livestock. In
an insurance contract, an insurer shall pay specific pre-defined compensation when financial
damages are caused by predefined cost-generating events, in exchange for upfront payments of
a premium by the policyholder (Dror & Piesse, 2014). Microinsurance is not different from
traditional insurance, and it relies on the same insurance principles. The term however
emphasizes its specific target, low-income households. It is commonly accepted in the literature
and by insurance practitioners, that providing insurance to low-income people is not simply the
provision of insurance at a “micro” scale (Churchill & Matul, 2012). Microinsurance requires
a more fundamental reengineering of products and processes, in order to address properly the
needs and constraints of poor clients. This research on agricultural insurance will address
microinsurance specific challenges, as 90% of the 570 million farms in the world are run by an
individual or a family (FAO, 2014). The vast majority of world farms are small or very small:
farms of less than one hectare account for 72% of all farms, and farms between 1 and 2 hectares
account for 12% of all farms.

Governments and development organizations’ interest in agricultural insurance has grown

steadily over the last two decades. Indeed, agricultural insurance stands at the crossroads of
several of current global issues: food security, climate change adaptation, rural development,
economic, social and financial inclusion. However, most of the world’s farmers are not covered
by an agricultural insurance policy. Those who are usually live in countries where the
government is actively supporting access to agricultural insurance. Indeed, under standard
market conditions, there is often no market for comprehensive agricultural insurance policies.
Agricultural risks are often spatially correlated, meaning that insurers are not able to mutualize

16
 Introduction

those risks efficiently within their portfolios (Miranda & Glauber, 1997). Reinsurance only
offers a partial solution and remains expensive. Therefore, public intervention appears essential
to support the development of agricultural insurance markets and extend protection against
natural risks to a large number of farmers and herders.

However, innovative approaches of agricultural insurance are being tested and sometimes
scaled-up. The extension of mobile network coverage and a range of technological innovations
in satellite imagery, automated weather stations, drones or big data allow the development of
new insurance policies and extend the limits of insurability. Innovation also happen at the
organizational level, resulting in new distribution, enrolment, claim management or payment
processes.

This research aims at analyzing these disruptions, and how such changes affect both the
business model and the public policies of agricultural insurance. The concept of “business
model” has been used extensively in the academic literature, especially since the mid-1990s,
but scholars do not agree on a single definition (Zott et al., 2011). However, patterns emerge
among the various definitions used in academic work. In general, the concept of business model
is used to analyze how firms do business, adopting a systemic level, holistic approach.
Therefore, a business model analysis goes beyond the firm, and encompasses its partners and
networks. In this work, we will also mobilize the concept of business model to study revenues
generated and costs borne by a firm, with an emphasis on the long-term viability of agricultural
insurance. As agricultural insurance schemes are often struggling to attain financial viability
under regular market conditions, public actors play a key role in most agricultural insurance
programs around the world. Therefore, analyzing the business model leads logically to study
public-private partnerships. Following Linder (1999), we adopt a very broad definition of the
public-private partnerships, to include all “cooperative ventures between the state and private
business”. All those partnerships are built with the aim to foster the development of agricultural
insurance markets, but vary in nature, governance, risk-sharing and cost-sharing agreements.
The thesis reviews the instruments available to policy makers to support the development of
the agricultural insurance sector, taking into account that insurance is one among a range of risk
management tools. It also examines the potential to renew the business model of agricultural
insurance in order to make it affordable and accessible to a larger number of farmers.

The research is based on two case studies and a survey in both developed and developing
economies. Though farmers may experience different kind of challenges, innovation is
happening in both types of economies with lots of potential for cross learning and exchanges
17
 Introduction

of best practices. The thesis investigates the case of grassland insurance in France and an
agricultural insurance scheme aimed at smallholder farmers in Kenya. It also includes a survey
of potential distributors of agricultural insurance in Africa and Asia. The research makes use of
different methodologies: financial modelling, interview-based qualitative study and
quantitative survey. This variety in both case studies and methods makes possible to take-out
different natures of results.

An important methodological feature of this work lies in the fact that the research took place
within the framework of a CIFRE contract (CIFRE standing for Conventions Industrielles de
Formation par la Recherche, or Industrial Research Training Agreements in French). A CIFRE
contract is an agreement signed between a company, a research laboratory and a PhD student.
The company hires the PhD student for a period of 3 years to conduct a research work under
the supervision of an external laboratory, this work leading to the preparation of a PhD
dissertation. During those 3 years, the hosting company receives a yearly grant from the
Association Nationale de la Recherche et de la Technologie (ANRT), a public agency under the
supervision of the French Ministry of Higher Education and Research. The ANRT gathers
public and private actors involved in research and development. The agency is responsible to
administer CIFRE agreements on behalf of the Ministry and validates the doctoral projects
before the signature of the CIFRE agreement. The existence of the CIFRE scheme can be
considered a success of public and private collaboration, producing research that is valuable for
both academics and companies (Canet & Grassy, 2006). Indeed, CIFRE researchers often play
the role of « mediators » between these two worlds (Levy, 2005), allowing knowledge to flow
from the company to the laboratory, and vice versa.

In our case, the researcher worked with the Crédit Agricole group. The Crédit Agricole is a
French cooperative financial institution, composed by a network of local banks, 39 regional
banks and a holding company named Crédit Agricole SA. Crédit Agricole SA itself holds many
subsidiaries in the finance and insurance sectors, and has activities in France and abroad. As
the name suggests, the Crédit Agricole was founded at the end of the 19th century to fund the
agricultural sector. It later diversified its activities during the 20th century but remains the main
lender to the French agricultural sector. The researcher was hired by Pacifica, the non-life
insurance company of the Crédit Agricole group. He collaborated with the Agricultural market
team. This team develops and manages insurance products designed for the farming sector, to
protect farmers’ productions and assets: hail and crop insurance, truck/tractor insurance,
livestock insurance, professional liability insurance. These products are then distributed

18
 Introduction

through the Credit Agricole banking network. The researcher also worked with the Grameen
Crédit Agricole Foundation (GCAF). Created in 2008, at the joint initiative of Crédit Agricole’s
Directors and Professor Yunus, founder of the Grameen Bank in Bangladesh, GCAF’s mission
is to fund and support microfinance institutions, businesses and projects that promote inclusive
finance and the development of rural economies around the world. It offers debt, equity,
guarantees and technical assistance services, and is active in developing economies. The setting
of the CIFRE agreement has been extremely beneficial for the research, as the PhD student had
the possibility to work with very different contexts (French agriculture vs developing countries’
agriculture). Both Pacifica and GCAF provided access to their data and contacts with their
partners, and gave the practitioners’ perspective on the research conducted through regular
feedback.

This thesis contributes to the scientific literature and the debate around agricultural insurance
in different ways. One characteristic of this work is that it relies heavily on real cases. The case
studies in chapters 3 and 4 for example describe and review specific features of agricultural
insurance products available in the market. As noted by Sandmark et al. (2013), many
publications on agricultural insurance schemes lack details on the financial aspects or the
product design. This thesis aims at taking into account this remark by providing a
comprehensive information on agricultural insurance contracts, such as premiums, deductibles,
sums insured, risks covered and exclusions. Chapter 3 also proposes calculations on the
financial viability of an insurance scheme, even if it relies on assumptions and modelizations
rather than on an actual insurer’s financial data.

This research’s contribution also comes from the confrontation of scientific literature with field
reality. Chapter 4 digs into the line of work initiated by C.K. Prahalad in 2002 around marketing
for the low-income households and the concept of the “bottom of the pyramid” (Prahalad,
Hammond, et al., 2002), as well as the research on demand for microinsurance products. The
literature’s analytical framework is then mobilized together with an existing industry’s
management tool to review three crop insurance policies and one livestock insurance policy
commercialized in Kenya. Chapter 5 examines the existing research to list possible motivations
of financial institutions to distribute insurance for their clients, before surveying a panel of 36
microfinance institutions. This back and forth movement between theory and practice allows to
enrich analytical frameworks and ground the research into operational realities.

Finally, the thesis offers practical insights and lessons for agricultural insurance professionals
and policy-makers. Though it is not comprehensive of all agricultural insurance public-private
19
 Introduction

partnerships around the world, it offers a broad overview of agricultural insurance public
policies and public intervention mechanisms to support the development of agricultural
insurance markets. The thesis also analyzes the challenges faced by private operators, on both
supply and demand. It is not limited to the perspective of insurers and reinsurers, and includes
key stakeholders such as insurance brokers, financial institutions and agricultural value chain
actors. Indeed, the author is utterly convinced that agricultural insurance is not a topic that can
be addressed only by the insurance industry, or the farming sector, but rather requires the
involvement of a broader range of stakeholders.

The thesis adopts the following structure. Chapter 1 describes the agricultural insurance context
and the challenges it faces. It also introduces the research hypotheses forming the base of this
work. Chapter 2 examines the literature to question the rationale for public intervention in favor
of agricultural insurance. It then reviews the instruments available and questions their
respective advantages and shortcomings, as well as complementarities between them. Chapter
3 evaluates the possibility to replace a public disaster assistance program by a commercial
insurance product, taking the case of grassland insurance in France. Chapter 4 analyzes the crop
and livestock insurance products offered by a microinsurance company in Kenya, in order to
understand how products and processes are designed to target this specific segment of farmers.
Finally, Chapter 5 investigates the business case for microfinance institutions to offer
microinsurance to their clients, based on a survey of 36 institutions in Africa and Asia.

20
 Introduction

References

Antón, J. (2009). Risk management in agriculture : A holistic conceptual framework. Managing

Risk in Agriculture A Holistic Approach, Paris, OECD.

Canet, E., & Grassy, G. (2006). Optimizing French scientific and economic performance : The
Cifre system of public-private partnership in doctoral research and Servier’s contribution.
Medecine sciences: M/S, 22(6‑7), 664‑668.

Churchill, C., & Matul, M. (2012). Protecting the poor : A microinsurance compendium:
volume II. Geneva. International Labour Organization.
[Link]

Dror, D. M., & Piesse, D. (2014). What is microinsurance? In A practical guide to impact
assessment in microinsurance (Microinsurance Network, p. 23‑39). Chapter 2.

FAO, T. (2014). The state of food and agriculture : Innovation in family farming. Rome: FAO.

International Labour Organization. (2018). World Employment and Social Outlook—Trends

2019.

Levy, R. (2005). Les doctorants CIFRE: médiateurs entre laboratoires de recherche

universitaires et entreprises. Revue d’économie industrielle, 111(1), 79‑96.

Linder, S. H. (1999). Coming to terms with the public-private partnership : A grammar of

multiple meanings. American behavioral scientist, 43(1), 35‑51.

Miranda, M. J., & Glauber, J. W. (1997). Systemic risk, reinsurance, and the failure of crop
insurance markets. American Journal of Agricultural Economics, 79(1), 206‑215.

Prahalad, C. K. (2004). The fortune at the bottom of the pyramid: Eradicating poverty through
profits. New Jersey: Wharton School Publishing.

Sandmark, T., Debar, J.-C., & Tatin-Jaleran, C. (2013). The emergence and development of
agriculture microinsurance. Luxembourg: Microinsurance Network.

Zott, C., Amit, R., & Massa, L. (2011). The business model : Recent developments and future
research. Journal of management, 37(4), 1019‑1042.

21
22
 CHAPTER 1

THE BUSINESS MODEL OF AGRICULTURAL INSURANCE:

CONTEXT AND PROBLEM STATEMENT

23
The business model of agricultural insurance: context and problem
statement

Risk is inherent to the agricultural activity. Farmers face a variety of risks, related to production,
market trends or other events. Unforeseen events affect any business, but agriculture’s strong
dependence on nature puts farm businesses in a situation of constant instability. These risks
affect farmers’ income, which is uncertain and unpredictable. The existence of risk has a strong
influence on farming practices and investments. In particular, agricultural risk can hinder
economic development, limit growth opportunities and exacerbate poverty and vulnerability in
rural areas. Many of these risks can be managed though, and farmers implement different risk
management strategies. Farmers’ risk management decisions will depend on the financial,
social and human capital available, as well as on individual farmers’ risk-aversion. Agricultural
insurance is one instrument within a portfolio of risk management tools available to farmers.
Therefore, agricultural insurance should not be analyzed in isolation, but rather be replaced
within the broader framework of agricultural risk management.

This chapter is organized as follows. Section 1 sets the context, describing risk management
strategies and agricultural insurance markets. Section 2 presents the challenges faced by
agricultural insurance, and questions the viability of its business model. Section 3 introduces
index insurance, and how it is disrupting agricultural insurance. Section 4 focuses on the role
of reinsurance in agricultural insurance. Finally, Section 5 exposes the three research
hypotheses and the case studies that make the backbone of this thesis.

1. Agricultural insurance markets and environment

[Link] management in agriculture

1.1.1. Risks in agriculture

The term risk can be used with different meanings: it can be the chance of a bad outcome, the
variability of outcomes or the uncertainty of outcomes (Hardaker, 2000). We will adopt the
third interpretation, and define risk as the “exposure to uncertain unfavorable economic
consequences” (Hardaker, Huirne, Anderson, & Lien, 2004).

Risks originate from different sources. Antón (2009) identifies four types of risks:

24
 Chapter 1 - The business model of agricultural insurance: context and problem statement

(i) Production risk: weather hazards, natural catastrophes, pests and diseases affect the
quantity and the quality of the farm output
(ii) Market risk: Price volatility of farm inputs, outputs and assets affect the farmers’
disposable income
(iii) Financial risk: Changes in interest rates, access to credit, farm assets’ value or non-
farm income affect the household’s wealth and income
(iv) Legal/institutional risk: Changes in the regulation modify the business environment
for farmers

Another categorization of risks can be made according to the degree of covariance of losses,
between idiosyncratic and systemic risk. Idiosyncratic risk affects a limited number of farms.
It has a narrow scale, and it is strongly related with farming practices. Idiosyncratic risk events
are usually frequent events that will affect a predictable number of farmers each year. Hail or
frost damaging crops, or a non-contagious disease affecting livestock fall into this category. On
the other hand, systemic risk affects an entire region or country. Losses among farmers in the
affected area are strongly correlated. Systemic risk events are infrequent, but they can
destabilize the region’s economy. Natural disasters like severe droughts, floods and hurricanes,
or epizootic diseases are systemic risk events. Financial risks are also typically systemic risks.

Though our analysis is centered on farming households, agricultural risks do not affect only
farmers. Rural communities, local and national governments, financial institutions, private
insurers and humanitarian organizations also implement risk management strategies to deal
with risk.

1.1.2. Risk management instruments

Farmers can use and combine several instruments to manage risk. These instruments can follow
three different strategies (Holzmann & Jørgensen, 2001):

(i) Prevention strategies, to reduce the probability of an adverse event occurring

(ii) Mitigation strategies, to reduce the potential consequence of an adverse event
(iii) Coping strategies, to relieve the impact of the adverse event once it has occurred

Prevention and mitigation strategies focus on income smoothing, while coping strategies focus
on consumption smoothing (Antón, 2009). In other words, prevention and mitigation strategies
address the risk ex-ante, risk coping strategies address the risk ex-post.

25
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Prevention instruments relate to farming practices and technological choices: vaccination,

drought-resistant seeds, pesticides intend to reduce the volatility of the farm output. Some of
these technological choices, like irrigation, can reduce the risk exposure while increasing the
average production. In other cases, there is a trade-off between risk and profitability. Farmers
can choose to plant crops that are less volatile, but also less profitable. Some studies estimate
that average agricultural incomes could be 10 to 20 percent higher in the absence of risk (P.
Hazell et al., 2010b). Mitigation instruments include both on-farm technical tools and market
mechanisms: production diversification, crop sharing agreements, futures and options, contract
farming, off-farm work. Agricultural insurance also falls into this category. It mitigates the
impact of production loss (in quantity and/or quality) through a claim payment. Finally, farming
households mobilize coping instruments to recover after the occurrence of an adverse event:
savings, loans (from a formal or informal lender), sale of assets, off-farm income, family and
community solidarity, humanitarian and government relief. Risk-coping mechanisms can be
difficult to access. For example, a bank might be unwilling to disburse a loan to a farmer in
difficulty. Some risk-coping mechanisms are also less efficient for covariant risks than for
idiosyncratic risks. When a drought affects an entire region, neighbors and family are unable to
provide relief as they are coping themselves with the consequence of this event. Selling
livestock is also a very common risk-coping instrument in developing economies, but poorly
efficient in the case of a widespread crisis as every household tries to sell at the same time,
pushing prices down. Finally, some risk-coping strategies can have long-lasting effects on
farmers’ livelihoods. If farmers sell productive assets to maintain a minimal level of
consumption, it threatens their capacity to generate income in the future.

[Link] agricultural insurance landscape

Agricultural insurance is a market-based risk mitigation instrument, which reduces farmers’

risk exposure to production risk. An insurance contract protects the policyholder against one or
several pre-defined risks, against the payment of a premium. It may not be available in all
countries, for all types of production risks, for all crops and livestock.

1.2.1. Products

At a global level, crop insurance represents an estimated 90 percent of premiums collected, and
livestock insurance represents the 10 percent remaining (Mahul & Stutley, 2010). Farmers also

26
 Chapter 1 - The business model of agricultural insurance: context and problem statement

purchase insurance to be covered against personal risks (health, death) and to protect their
property and assets. These insurance products are outside the scope of this research.

[Link]. Crop insurance products

For crops, named-peril insurance is the most widely available type of insurance. Named-peril
insurance protects crops against a specific risk, usually hail. It is possible to add other pre-
defined perils, like frost, or floods; such products are known as combined insurance in Europe.
Claims are calculated by estimating the percentage of damage to the crop, multiplied by a pre-
agreed value (production costs or a pre-agreed value of the crop revenue).

Multi-peril crop insurance (MPCI) guarantees an insured yield against multiple hazards that can
cause yield losses. This insured yield is calculated as a percentage of the historical average
yield. If the actual yield falls below the insured yield, the insurer pays a claim to the
policyholder to cover the loss. Claims are calculated by the difference between the insured and
the actual yield, multiplied by a pre-agreed value of the crop value.

Named-peril insurance and MPCI are both indemnity-based contracts. Loss adjustment
traditionally implies the visit of an expert assessing two things: (i) whether the loss is due to an
insured peril and (ii) the magnitude of the loss. The expert loss assessment will determine
whether a claim payment is due or not, and its amount.

Revenue insurance combines MPCI with hedging techniques (futures) against price volatility.
It is very common in the United States. Revenue insurance guarantees a minimum revenue for
farmers who are protected against yield losses and crop price fluctuations.

Besides traditional indemnity-based crop insurance products, index-insurance is a new

insurance approach. Index-insurance does not pay claim against actual losses. Losses are
estimated thanks to an index, a proxy using third party data (see Section 3 below for more
details). Index-insurance is available in a growing number of countries, and is particularly
interesting to cover smallholder farmers, for which indemnity-based insurance would not be
feasible due to the high costs of loss adjustment compared with the sum insured.

[Link]. Livestock insurance products

Livestock insurance contracts protect against the risks of animal accident or mortality. They
sometimes include a cover against diseases, but not preventable diseases. Livestock insurance
contracts may require that farmers implement some prevention measures, like vaccination or

27
 Chapter 1 - The business model of agricultural insurance: context and problem statement

veterinary controls of insured animals. Animal mortality depends a lot on farmer’s behavior, so
livestock insurance is often plagued with adverse selection and fraud.

Index-based livestock insurance is available in a few countries. It protects herders against

animal mortality, and estimates losses thanks to average animal morality per locality
(Mongolia) or average vegetation available for animal feed (Kenya, Ethiopia).

1.2.2. Agricultural insurance markets around the world

The first agricultural insurance products emerged more than 200 years ago in Europe. Hail was
the first peril insured and remains until today the most widely available cover. The Roosevelt
administration introduced the first MPCI contract in the United States in the 1930s. Between
the 1950s and 1980s, governments of Asia and Latin America provided public agricultural
insurance, often in linkage with national agricultural credit programs. These programs were
often terminated in the 1980s and 1990s, due to their poor results and a paradigm shift towards
more liberalism in public policies. In Europe, some governments started subsidizing MPCI in
the 1980s, with an acceleration of this trend in the 1990s and 2000s. Globally, the trend in the
last two decades has been to provide agricultural insurance through the private sector, generally
with support from the national governments.

Agricultural insurance is growing rapidly around the world. Global premiums tripled between
2005 and 2011, going from $ 8 billion to $ 23.5 billion (Kalra & Xing, 2013). This is equivalent
to an annual 20% growth rate. The share of emerging markets in total premiums went from
13.4% to 22% over the same period. In these markets, agricultural insurance premiums more
than quadrupled, corresponding to a 30% annual growth rate. China and India were driving the
emerging markets growth. These two countries account for 62% of premiums in emerging
markets in 2011.

28
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Figure 1 - Estimated agricultural insurance premiums worldwide in 2005 and 2011

25

Premiums in USD billion

20

15

10

0
2005 2011

Emerging markets Advanced markets

Source: Swiss Re (2013)

Despite the rapid growth in emerging markets, USA & Canada remain the largest agricultural
insurance market, with 55% of premiums collected. Europe comes third with 18% of total
premiums. The shares of Latin America (4%) and Africa (0.5%) remain particularly low,
comparatively to their weight in food production and world population.

Figure 2 - Estimated geographic distribution of agricultural insurance premiums in 2011

US & Canada: 55%

Europe: 18%

Asia: 22%

Latin America: 4.0%

Australia & New Zealand: 0.8%

Africa: 0.5%

Source: Swiss Re (2013)

Unfortunately, to the author’s knowledge, there was no study published that aimed at estimating
global agricultural insurance premiums, since Swiss Re’s publication in 2013, which was based
on 2011 data. Since 2011, it is reasonable to think that agricultural insurance markets have most
likely grown. Indeed, many developing countries have expanded their agricultural insurance
programs or launched new pilots. Hess, Hazell & Kuhn (Hess et al., 2016) estimate that there
are 198 million policyholders in developing countries, with 37 schemes operating. Most of these
policyholders are in China (80%) or India (16%).

29
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Table 1 - Agricultural insurance schemes in developing countries

Number of Weighted average

policyholders (in Number of subsidy
millions) schemes (estimation)
Africa 0,65 18 37%
India 32,2 4 64%
China 160 - 77%
Rest of Asia 1,1 7 64%
Latin America 3,3 8 91%

Source: Hess et al. (2016)

2. The business model of agricultural insurance in discussion

2.1. Conditions of insurability

Not all perils are insurable. Skees, Barnett & Hartell (2005) and Shaik et. al (2006) identify five
ideal conditions of insurability.

(i) Determinable and measurable loss: it must be possible to determine whether or

not a loss occurred, if it was caused by an insured peril, and to measure the
magnitude of the loss
(ii) Accidental and unintentional loss: an insurance should pay claims only for random
events over which the policyholder has little or no control. However, the
policyholder’s behavior can increase the probability and/or the magnitude of loss.
Monitoring policyholders’ behavior can be costly or even impossible, which opens
the door to moral hazard situations. Fraud is an extreme case of moral hazard. Even
in the absence of unethical or illegal actions, moral hazard can be the rational
economic decision of a policyholder to reduce risk prevention and mitigation
measures when being insured.
(iii) Sufficient data is available for risk classification: Insurers must be able to
accurately estimate the level of risks of clients and set the premium accordingly.
Otherwise, adverse selection may occur, especially if policyholders have better
information on their risk exposure than the insurer. Those who benefit from low
premiums relatively to their real risk will be more likely to purchase insurance, while
those who are charged premiums too high will drop out.
(iv) Calculable expected frequency and magnitude of loss: Insurers calculate
premium rates based on the frequency and the severity of expected losses. Insurers

30
 Chapter 1 - The business model of agricultural insurance: context and problem statement

need reliable data on historical losses, as well as information on future possible

distribution of losses.
(v) Large number of independent exposure units: Risk pooling is more efficient
when losses are uncorrelated or not highly correlated. Thanks to the law of large
numbers, the variability of losses of the aggregate portfolio decreases as the number
of policies increases.

In practice, these five conditions are rarely met altogether. Insurers often address the violation
of one or several criteria by increasing premium rates. However, if premiums are too high, there
might be no demand. Shaik et al. (2006) add a sixth insurability condition, stating that (vi)
premiums must be economically feasible. For perils that deviate significantly from the five
ideal conditions, premiums may exceed the clients’ willingness to pay, and therefore there
would be no market.

Regarding crop and livestock insurance, the first criterion is often met, though it requires
significant expertise from the insurer. The second criterion is more challenging. Crop and
livestock losses are related to external and non-controllable shocks, like weather hazards or
animal accidents. However, farming practices do affect the frequency and severity of losses, so
insurers must implement monitoring systems and incentives to comply with the second
criterion. Data availability is in many cases a challenge, which makes the third and the fourth
criteria hard to meet. Historical yield and animal mortality data may be insufficient or
unreliable, especially in developing countries.

If the first four criteria can be met over time, through the collection and improvement of data
and monitoring systems, the fifth criterion poses a long-lasting challenge. The very nature of
agricultural risk contradicts the condition of independent losses. Indeed, crop losses are often
highly spatially correlated, which makes risk pooling inefficient (Miranda & Glauber, 1997).
This should preclude the formation of an agricultural insurance market. However, such markets
exist, but usually with some sort of public support.

2.2. Insuring the production of smallholder farmers

Insuring farm production is challenging, as the five insurability conditions are not met for
agricultural risks. Targeting smallholder farmers comes with additional constraints. The
number of smallholder farms is estimated between 400 and 500 million, and the number of
people living in these households lies between 1.5 and 2.5 billion people (Christen & Anderson,

31
 Chapter 1 - The business model of agricultural insurance: context and problem statement

2013). Therefore, smallholder farmers’ specific constraints have to be taken into account in
order to extend insurance coverage to a large number of farmers.

In traditional insurance theory, risk-averse individuals are always willing to buy insurance at
the actuarially fair price in order to maximize their utility, and they can accept to pay a premium
above the expected loss, in order to reduce income volatility (Brau et al., 2011). In theory, the
greater the risk-aversion, the greater the risk premium an individual is willing to pay.
Individuals will also prefer to pay a risk premium to protect themselves for large infrequent
losses to preserve their wealth. Finally, the lower the initial level of wealth, the more clients are
willing to buy insurance. Therefore, demand for insurance should be very high among
smallholder farmers as their initial wealth is low and potential losses represent a significant
percentage of their wealth.

The reality is quite the opposite: willingness to pay for insurance is particularly low among low-
income households in general. Indeed, low-income clients do not know insurance contracts and
have limited confidence in insurance companies (Brau et al., 2011). From the supply side, data
availability is often deficient in low-income environments, which makes risk-classification and
premium calculation challenging (Biener & Eling, 2012; Brau et al., 2011). To make things
worse, risk pools are too small, which limits both the potential of risk diversification and the
possibilities to generate internal data over time (Biener, 2013). Insurance for low-income clients
also faces the issue of high transaction costs in comparison with the premiums and sums
insured. This is particularly true for loss adjustment of yield losses. The cost to send an expert
evaluate losses for a smallholder farmer might exceed the yearly premium paid by the farmer
or even the potential claim. Traditional loss adjustment methods bear high transaction costs that
are not compatible with the economic feasibility of premiums for smallholder farmers.
Agricultural insurance for smallholder farmers cannot operate under the same business and
operational model as in other markets. In particular, operational costs need to be drastically
reduced to make insurance affordable for clients and viable for insurers.

Moreover, smallholder farmers often live in remote areas and lack access to basic services taken
as granted for many other citizens: transport, education, information, markets, or finance.
Insuring smallholder farmers not only requires to lower costs; the distribution, awareness-
raising and marketing mix need to address these clients’ specific needs and constraints.

32
 Chapter 1 - The business model of agricultural insurance: context and problem statement

2.3. Can agricultural insurance exist without subsidies? Viability of insurance programs

In a 1992 article, Hazell analyses the viability of seven public agricultural insurance programs
in the 1970’s and 1980’s (Hazell, 1992). He finds that these insurance programs’ financial
performance is disastrous, and that actuarial principles are often not applied. To be financially
viable, an insurer must collect more premiums than it spends annually on indemnities and
administration costs. Let us denote P the premiums collected, I the indemnities paid to
policyholders and A the administration costs borne by the insurer. For an insurer to be
financially solvent, the “Hazell ratio” Z must comply with the following condition:

+
= <1

All the programs analyzed by Hazell fail to meet this condition. They are only able to run
because they receive public subsidies. It is possible to disaggregate the Hazell ratio into two
components: the producer loss ratio I/P and the administration cost ratio A/P. Among the seven
insurance programs studied, the producer loss ratio ranges between 0.99 in Japan to 5.11 in
India. When I/P is above 1, the commercial premium is set below the actuarially fair premium.
In such cases, farmers receive more in indemnities than they pay in premium. From the farmers’
perspective, purchasing insurance is a very profitable investment: every dollar spent in
insurance premium in India is equivalent to a 5.11-dollar claim payment on average.

Administration costs can also undermine the viability of an insurance program. In the sample
of Hazell, the A/P ratio is comprised between 0.28 and 3.57. When A/P is above 1, the
premiums collected do not even cover the administration costs of the insurance program.

Hazell et al. (Hazell, Sberro-Kessler, & Varangis, 2017) ran the same calculations a few years
later using data from Mahul & Stutley (2010). Except in Japan, all producers’ loss ratios
decreased significantly, indicating a better viability of these insurance programs. However, only
one program (Mexico) had an I/P ratio below 1, meaning the insurance scheme may be able to
operate without subsidy (the level of administration costs A/P remains to be analyzed).

Table 2 – Producer loss ratios in seven countries

Country Producer Loss Ratio I/P

Brazil 1975-81 4.29 2004-2007 1.19
Costa Rica 1970-89 2.26 2003-2007 1.75
India 1985-89 5.11 2003-2007 3.36

33
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Japan 1985-89 0.99 2003-2005 1.84

Mexico 1980-89 3.18 2003-2007 0.72
Philippines 1981-89 3.94 2003-2007 1.42
US 1980-89 1.87 2003-2007 1.70
Source Hazell et al. (2017)

Skees, Hazell & Miranda (1999) propose several reasons for the insolvency of agricultural
insurance programs. The main reason identified is that agricultural risks fail to meet insurability
conditions, so an agricultural insurance scheme cannot operate without public support. The
authors also point other explanations for failure, such as the mandate of public insurers to work
with smallholder farmers, which increases administration costs, the fact that the government
may systematically cover losses, which gives no incentives to public insurers to monitor
fraudulent claims, and the use of insurance for political motives, like when excessive claims
are paid prior to elections. Crop insurance programs can also be too specialized on certain crops,
regions or types of farmers, which limits the size of the risk pool and the effect of
diversification.

3. Index insurance, a new paradigm for agricultural insurance

For many years, agricultural insurance has been available only to large-scale commercial
farmers, and only for those living in countries with pro-insurance policies. However,
technological innovations are disrupting traditional agricultural insurance markets. Relying on
alternative sources of data for risk classification, premium calculation and loss adjustment,
index insurance is profoundly modifying the business model of agricultural insurance. Index
insurance extends the scope of insurability to new risks and new types of farmers.

3.1. Principles of index insurance

In traditional crop and livestock insurance, insurers pay claims when farmers suffer from an
actual loss. An external event, like hail or a drought negatively affected yields, or animals died
because of a non-preventable disease. In some cases, assessing production loss can be
unpractical, costly or simply impossible. Index insurance overcomes this difficulty by relying
on a variable, a proxy that estimates accurately the production loss. Unlike traditional
agricultural insurance, index insurance payments are not triggered by the observation of a
production loss, but by the level of an index, which is itself correlated with the level of
production. When the index value goes below a pre-defined threshold (i.e. the insured level of

34
 Chapter 1 - The business model of agricultural insurance: context and problem statement

production), insurers pay indemnities to policyholders. Policyholders do not need to file a claim,
the index automatically calculates losses, and claims are paid accordingly.

3.1.1. Crop indices

Crop insurance indices fall into three categories (Collin, 2018).

Area-yield indices

Like in traditional crop insurance, experts evaluate yield losses through actual observation in
the field, except that instead of assessing yields of each individual farmer filing a claim, experts
measure the production of a representative sample of farmers. The index value of a certain area
corresponds to the average yield measured in this area.

Compared to other index technologies, area-yield indices are easier to understand for all
stakeholders, as they are similar to traditional insurance techniques, applied on a sample.
Furthermore, area-yield indices allow accounting for all perils that can affect yields, including
pests. However, an area-yield index requires good historical yield data, which is seldom
available. Furthermore, conducting crop-cutting experiments to evaluate average yields in each
area is costly and the results can be available several months after the season, which further
delays indemnities’ payment.

Weather indices

Weather indices rely on a network of ground weather stations and captors, measuring different
meteorological variables: rainfall, air and soil humidity, wind speed. Rainfall indices are the
most common meteorological indices, as in many contexts, water-availability is the main factor
affecting yields. The index will estimate yields by comparing rainfall with the water needs of
the crop during each growth phase of the plant.

Weather indices are relatively easy to understand. Farmers know how weather affects yields
and they can observe meteorological variables themselves. Nevertheless, in many countries
there is no sufficient historical meteorological data to build an index. Reliability of current data
being collected can also be an issue, though automated weather stations tend to increase the
timeliness and accuracy of data. Finally, the weather station network is often not dense enough
to allow an accurate yield assessment at the local level.

Remote-sensed indices

35
 Chapter 1 - The business model of agricultural insurance: context and problem statement

These indices are based on data provided by satellites. Different techniques using radar captors
or satellite imagery are used to evaluate rainfall or vegetation growth.

Remote-sensed data is available for the whole planet, with a good level of accuracy. In the case
of medium-resolution images, there is more than 30 years of data. However, remote-sensing
indices are hard to explain to the different stakeholders. They also require significant initial
investments and expertise to be set-up, though in the long run they can be easier and cheaper to
administer than area-yield or weather indices.

3.1.2. Livestock indices

Though less common than for crops, livestock index insurance also exists. We can identify two
types of indices.

Mortality indices

In this technique, animal losses are estimated by measuring average animal mortality per area.
In its construct, it is quite similar to an area-yield index, except that it relies on animal censuses
instead of crop cutting experiments. Such an approach is being implemented in Mongolia
(Mahul & Skees, 2007). The Government has been conducting an annual census of adult
animals, and recorded animal mortality per district (“soum”) for more than 50 years. Such data
include all causes of animal losses. Mortality indices are not common. The quality and
availability of animal mortality data in Mongolia is rather exceptional. The data collection
process is strictly controlled to ensure its accuracy. Animal losses per species are available at
the local level since 1971. These favorable conditions made possible the creation of the
livestock index insurance program.

Pasture availability indices

As animal mortality data is scarce, an alternative variable that insurers can use is the availability
of pasture. In arid and semi-arid areas, nomad pastoralists face the risk of drought, which results
in the lack of pasture and possibly livestock’s starvation. An index measuring the availability
of pasture can predict accurately animal losses. In Kenya, an index-based livestock insurance
scheme is running, using satellite data to measure the quality of pastureland every 10-16 days
(Index-based livestock insurance as an innovative tool against drought loss: Good practices
and impact analysis from northern Kenya., 2018). This data forms the input to build a statistical
model of animal mortality per region. When losses exceed a determined threshold, herders
receive an insurance payout. Pasture availability indices are like satellite-based crop indices,

36
 Chapter 1 - The business model of agricultural insurance: context and problem statement

with the additional component of using the estimation of vegetation growth to predict animal
losses.

3.2. Benefits

The benefits of index insurance techniques compared to traditional indemnity-based insurance

have been extensively studied (Barnett & Mahul, 2007; Collin, 2018; Hazell et al., 2010).

Index insurance overcomes the issues of adverse selection and moral hazard, which are intrinsic
to most insurance contracts. Index insurance does not operate at the farm level, but at a larger
scale. Individual farmers’ risk profiles do not really matter at this scale, eliminating the problem
of adverse selection. An individual farmer’s production will not influence significantly the
value of the index, limiting moral hazard issues. All the policyholders in a certain area have the
same contract, pay the same premium and receive the same indemnity per insured unit.
Therefore, insured farmers have the same incentives to obtain a good level of production than
a non-insured farmer. The insurance contract covers only the risk that is beyond the control of
farmers, evacuating farm-specific practices, which are difficult to monitor.

Another set of advantages lies in the administration of the insurance scheme. If the initial
upfront investments to create an index insurance contract can be high, there are large savings
opportunities in the running of the program. In particular, index insurance does not require
assessing individual losses, which opens the door to insuring smallholder farmers who would
otherwise not be insurable.

Finally, index insurance increases transparency in the system. Losses are not measured based
on an individual expertise, but rather on third-party data provided by a reliable source.
Policyholders may have more trust in this third-party data provider than in the insurance
company loss adjuster. Reinsurers and financial markets may also be more willing to take risks
if they trust the indices.

3.3. Limitations

Index insurance relies on data. This data must be accurate, available at a sufficiently small scale
and over a sufficient long time. Those conditions are not always met, especially in developing
countries. Statistical methods or the combination of several data sources can compensate only

37
 Chapter 1 - The business model of agricultural insurance: context and problem statement

partially the absence of data. Over time, the index insurance scheme will generate its own data,
which should improve the quality of predictions.

The most important shortfall met by index insurance is “basis risk”. Basis risk happens when
the index does not reflect the real losses suffered by farmers. Therefore, individual farmers
experiencing a loss may not receive the compensation they should get. On the other hand, they
may benefit from an insurance payout without suffering any loss. Such situations undermine
the confidence in the insurance scheme, and place farmers at risk.

Sandmark et. al (2013) identify three types of basis risk:

- Spatial: when the granularity of the index is not sufficient. For example, two villages
located in the same area (same data point in the index) may experience different losses
- Temporal: there can be a lag between the event and its impact on the crop
- Crop-specific: the correlation between the index and the actual yield can be low. The
index does not properly identify the factors provoking losses.

Research is still needed to improve the quality of indices and reduce basis risk. Increasing data
availability is also a way to improve the quality of indexes. It is possible to reduce spatial basis
risk by increasing the granularity of an index, setting up more weather stations for example, or
increasing the resolution of satellite images. Another approach is to modify the contract design
with a double trigger (Carter, 2011). The first trigger is based on measured yields at the
village/community level. However, at such a small scale, farmers can collude and influence the
index measurements. A second trigger at a larger scale validates the reality of widespread
losses. The first trigger is designed to improve the correlation between losses and payments
while the second trigger controls the absence of moral hazard. Finally, a third approach to
mitigate basis risk is to insure groups or larger entities rather than individual farmers.
Sometimes called “mesoinsurance”, this approach aims at protecting entities that are exposed
to aggregate yield losses risk, rather than idiosyncratic risks. Such entities can keep claim
payments, or they can use the payouts to compensate individual farmers according to their
individual losses, as these entities might be in a good position to assess actual losses.

Despite its limitations, index insurance bears the potential to extend risk coverage to farmers
who were previously excluded from insurance. It also has a different cost structure from
traditional indemnity-based insurance, and benefits from economies of scale. Index insurance
can therefore disrupt the business model of agricultural insurance.

38
 Chapter 1 - The business model of agricultural insurance: context and problem statement

4. Reinsurance in agriculture

4.1. Reinsurers as market enablers

As seen in 2.1, agricultural risks are difficult to insure because of their covariant nature.
Reinsurance can relax this constraint, by allowing insurers to transfer part of the risk.
Reinsurance companies hold large portfolios, spread across several types of businesses and
geographies. They are therefore in a better position than most national insurance companies to
efficiently diversify risk. Moreover, reinsurance companies have built very large reserves which
allows them to cope with large-scale loss events. By absorbing large losses, reinsurance can
make an insurance company’s results smoother and more predictable. This normally means that
the insurer’s exposure is reduced, and therefor limits the need to raise capital. With reinsurance,
insurers can issue policies with larger limits than they would otherwise be able to.

Insurers and reinsurers sign a contract, called a reinsurance treaty, which defines the modalities
of the risk-sharing agreement. Two types of risk-sharing agreements are the most common:

- proportional or “quota-share” reinsurance: under a quota share agreement, a fixed percentage

of each insurance policy is transferred from the insurer to the reinsurer; usually, the ceding
company seeks a quota share agreement to reduce the need for capital.

- non-proportional or “stop-loss” reinsurance: stop-loss reinsurance covers losses after a

threshold of claims has been exceeded; stop-loss reinsurance allows insurance to underwrite
catastrophic risks.

Besides the risk-transfer instruments, reinsurance companies also bring insurers technical and
actuarial expertise. In some markets, local insurers may lack the experience or the capacity to
properly rate premiums and assess probable maximum losses. Reinsurers can support their
clients, and bring their know-how acquired in various markets. The largest reinsurance
companies are private, but some countries have also set-up a national public reinsurance
company when private reinsurance capacity was not available or not satisfying.

4.2. The example of the Spanish scheme

Spain offers an interesting example of public-private collaboration to provide insurance to a

large number of farmers. The Spanish agricultural insurance system was created in 1978. It
covers all productions (agriculture, livestock, aquaculture, forests), against all natural risks. The

39
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Spanish system is an original public-private partnership based on the voluntary participation of

private insurance companies in a co-insurance pool (Agroseguro, 2015). This pool is managed
by Agroseguro, a company owned by the insurance companies. Agroseguro does not bear the
risk itself, it only manages the co-insurance pool. Policies are grouped into 3 categories: A
(experimental), B (viable) and C (cost of destroying dead animals). The participation of each
company in each category is adjusted each year, according to its share in the sale of the
corresponding insurance products.

The system leaves an important place to the State. Premium rates are determined by the co-
insurance pool and then validated by the State. The State also validates the characteristics of
the contracts, and determines as part of an annual plan the cultivation areas on the basis of
which state subsidies are paid. Competition between insurance companies only affects the costs
and services of marketing their products. The purchase of insurance is necessary to obtain
certain exceptional production aid or disaster aid schemes (direct compensatory payments,
interest subsidies and loan guarantees). In addition, renewing insurance policies allows for
higher policy discounts and subsidies.

Figure 3 – The Spanish Public-Private Partnership

ENESA
Min
Agri

DGSFP
Min
Eco
Agroseguro

Producers
organizatio
ns

Private
CCS
reinsurers

40
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Private insurers
- Carry the risk
- Market insurance products
- Compulsory participation in the Agroseguro co-insurance pool
- Agroseguro shareholders

Agroseguro
- Issues policies
- Directly receives state and regional subsidies (producers pay premiums net of subsidies)
- Manages claims reports and assessments
- Compensates policy-holders on behalf of insurers
- Performs statistical and actuarial studies
- Manages reinsurance with the CCS and the private reinsurers on behalf of insurance
companies

CCS: Consorcio de Compensación de Seguros

- Public company supervised by the Ministry of the Economy
- Compulsory reinsurer of the co-insurance pool, according to a percentage defined annually by
the Ministry of the Economy
- Provides stop-loss reinsurance

Private reinsurers
- Each insurance company decides how much of the risk it wishes to transfer to the private
reinsurers
- Reinsurers provide stop-loss reinsurance

ENESA: Entidad Estatal de Seguros Agrarios

- Public body dependent on the Ministry of Agriculture
- Defines the government strategy (annual agricultural insurance plan)
- Manages the grant budget
- Defines the subscription dates, the insurable yields and the premium rates
- Conducts viability studies for each product

DGSFP: Dirección General de Seguros y Fondos de Pensiones

- Administrative body dependent on the Ministry of the Economy
- Supervises insurance contracts
- Defines the claims expertise standards
- Collaborates with ENESA on the strategy of the co-insurance pool and the criteria for the
distribution of grants

Producers’ organizations
- Represent the sector within ENESA
- Allow farmers to take out group insurance policies

41
 Chapter 1 - The business model of agricultural insurance: context and problem statement

The insurance take-up varies depending on the production, from less than 3% for grasslands, to
nearly 90% for fruit production (Agroseguro, 2017). The amount of premiums increased almost
continuously until 2008, before starting to decline, then recovered from 2015. The amount of
subsidies followed the same trend.

Figure 4 – Agroseguro premiums

Agroseguro insurance premiums, in million euros

800,0 70%
700,0 60%

Premium subsidy rate

600,0
Premiums in M€

50%
500,0
40%
400,0
30%
300,0
20%
200,0
100,0 10%

0,0 0%

Titre de l'axe

Subsidies Paid by the farmer % premium subsidies

Figure 5 – Agroseguro technical results

Agroseguro technical results (in M€)

900,0 120%
Premium and Losses in M€

800,0
100%
700,0
600,0 80%
Loss ratio

500,0
60%
400,0
300,0 40%
200,0
20%
100,0
0,0 0%
199719981999200020012002200320042005200620072008200920102011201220132014
Titre de l'axe

Premiums Losses L/P

42
 Chapter 1 - The business model of agricultural insurance: context and problem statement

4.3. Interviews with reinsurers

4.3.1. Context

The largest reinsurance companies operate on a global scale, which allows them to improve risk
diversification across different business lines, and different countries. This unique position in
the insurance value chain gives reinsurance professionals a very broad and global perspective
on agricultural insurance markets. Therefore, it appears relevant and insightful to interview a
small sample of agricultural underwriters working in reinsurance. Those interviews were
conducted at the end of the research work, in order to fine tune research questions and get
feedback on the results of the thesis.

4.3.2. Objectives

Reinsurance professionals were interviewed on three topics that are transversal in this research
work: (i) market failures in the agricultural insurance markets and the conditions of insurability;
(ii) the role of governments in supporting agricultural insurance markets; and (iii) innovations
and other drivers that will likely affect future market development.

4.3.3. Methodology and sample

7 people were interviewed, during 5 meetings with 5 different companies. Four of the
companies interviewed are among the largest reinsurance companies in the world, and one
company is a national public reinsurer. All the professionals interviewed are either agricultural
underwriters, or working in the research and development department of their company. All are
experts in agricultural and/or catastrophic risks. The people in the sample were identified thanks
to Pacifica.

The interviews were all conducted remotely, through video calls, using the Zoom or Microsoft
Teams apps. Interviews lasted between 30 minutes and one hour. One interview was conducted
in English, the other four were conducted in French. Interviews took place in May and June
2021.

An interview guide was prepared prior to the interviews, with open questions around the three
themes presented above. After the first interview, the interview guide was refined. The
interview guide and the results of the first interview were shared with the thesis supervisor and
a Pacifica representative, in order to improve the questionnaire.

43
 Chapter 1 - The business model of agricultural insurance: context and problem statement

4.3.4. Results

Market failures in agricultural insurance

The first topic discussed is the topic of agricultural insurance market failures. All the
professionals interviewed agreed that, only name peril insurance (usually hail insurance) can
exist without direct public intervention. Experience shows that multi-peril crop insurance has
grown only in countries where premium subsidies are available. Two interviews raised the issue
of agricultural prices and margins. Agriculture tends to be a low-margin activity, it is therefore
not conceivable for most farmers, including in developed economies, to pay the full price of
insurance and dedicate several points of margin for a risk management instrument. The market
has focused on insuring crops, livestock insurance representing a very small fraction of the
business. Indeed, livestock mortality risk proves to be more complex to insure, one of the
reasons being that farming practices have a very large impact on losses and are difficult to
control. Spain is a notable exception, with livestock insurance weighting close to a third of
agricultural insurance premiums. Overall, market development is always driven by policy.
Market development for multi-peril crop insurance always follows a voluntarist policy, which
most often takes the form of premium subsidies.

Role of government

All the reinsurers interviewed underlined the importance of assessing the public policy and
support mechanisms in place to evaluate the market potential for agricultural insurance.
However, this analysis of the public policy does not translate into models, or even a
standardized evaluation grid of the public-private partnerships. Such a model would be used to
integrate the public sector into the reinsurance price calculation, and the amount of capital to
mobilize. One professional interviewed mentioned an attempt to build such a model internally
around the year 2010, but the company abandoned the project. Another interview pointed the
fact that some market research companies are developing an offer in this topic, providing data
and markets analytics to support reinsurers’ decision making, but it remains very nascent and
not central in their analysis.

On the contrary, market assessment appears to be very ad hoc, with lots of factors to take into
account as there are a lot of different set-ups among countries. There is a great variety of public-
private partnership models, no country replicating exactly what can be observed in another
country. Indeed, governments can activate and mix a variety of policy instruments (see chapter
2). Another factor limiting the standardization of market assessments is the fact that farming
44
 Chapter 1 - The business model of agricultural insurance: context and problem statement

practices can also be radically different between countries, especially between developed and
developing economies. Access to credit, inputs and technology are very heterogeneous,
resulting in average yields and yield volatilities that vary a lot from one country to the other.
The reinsurer’s assessment also includes an analysis of the players operating in the market (local
insurers, brokers, aggregators), the history of the local insurance market and the quality and
availability of data.

Three interviews insisted on the importance of the relationship with the local insurer. Reinsurers
assess the quality of operations’ execution, in particular for loss adjustment. The go/no-go
decision is not always purely technical. A reinsurance company may agree to make a deal even
if the deal does not satisfy its usual requirements, in order to maintain a good relationship with
a client. The reinsurance sector is a business with few clients (insurance companies), so
reinsurers have to take into account the total business they have with a client when evaluating
the level of risk they accept to take. Reinsurers invest on the medium and long term. They don’t
invest resources if they expect to withdraw within five years.

The expectations of reinsurers towards the public sector can be grouped into three categories:

- visibility: reinsurers expect a stable regulatory environment. The reinsurers interviewed

insisted that they invest for the long term and can’t do so if the market is unpredictable. Changes
in policy or delays in paying premium subsidies affect negatively market growth. Consistency
is also a key point: two reinsurers mentioned the fact that government payments to farmers after
severe production losses undermine market development, even in the presence of premium
subsidies. Three reinsurers also indicated that technical expertise within government often leads
to more consistent and predictable policies, and a better dialogue between the public and the
private sectors.

- data and transparency: when there is no insurance history, reinsurance companies can only
rely on secondary data, often public data, to estimate historical yields and potential losses.
Governments have a role to play in the provision of public goods and improving transparency.
The cases of Spain and the United States were cited in all the interviews, both because they are
among the largest agricultural insurance schemes, but also because the government agencies
publish a lot of information available to the public.

- risk-sharing: the role of governments as risk-takers was not unanimously appreciated.

Unsurprisingly, the public reinsurance company insisted the most on this aspect. The
professional interviewed defended the idea that the private sector only has a limited capacity to

45
 Chapter 1 - The business model of agricultural insurance: context and problem statement

take on catastrophic risk. He also supported the idea that public reinsurance rates are stable over
the years, unlike reinsurance rates which tend to be volatile. Two private reinsurers also
recognized that the public sector can play a role in covering the most extreme layer of risk.
However, for two other interviewees, risk-taking capacity is not the most limiting factor at this
stage of market development. One professional even rejected the idea that public reinsurance
could make insurance more affordable, as price is first driven by the technical rate.

We can see that there is a consensus around the idea that public policy must be predictable, and
that governments have role to play to increase transparency and supply data. The role of the
public sector as a risk-taker itself is more controversial, though the examples of Spain and the
United States are often cited as examples of success.

In the end, reinsurance companies scrutinize the 3 “Ps” of PPPs: they assess the public policy,
the track record of the private players, and the quality of the partnerships established between
the public and the private sector.

Perspectives of development

All the professionals interviewed watch closely the innovations taking place in the agricultural
insurance sector. Their appreciation of a new approach like index insurance may vary: some
see it as a “fad”, others have more confidence in its potential to open new markets. As one
interviewee said, “more than 90% of our agricultural business is traditional reinsurance”, so it
is still difficult to say what innovation will actually disrupt the market.

In terms of geographies, nowadays the market is concentrated on a few countries that have put
in place favorable policies: the United States, some European countries, and a few large
emerging economies like China and India. However, some reinsurers are willing to explore new
markets. One reinsurance company in particular has a very proactive approach to work in
countries where they consider there is potential for growth, even if the market conditions are
not optimal currently. That company explained that they adopt a “sandbox approach”. When a
country with significant market potential takes a pro-insurance policy, the company will try to
get into the market, even if the policy is far from being perfect. The company will usually limit
its exposure, observe the market, accumulate historical data and learn. Another company
described something similar: when starting from scratch, it is usually easier to start with hail
insurance, and then build from this base. When asked about the potential to address smallholder
farmers in developing economies, one interviewee appears to be voluntarist in serving such
markets and clearly sees an opportunity, two expressed caution, as they don’t see yet any viable

46
 Chapter 1 - The business model of agricultural insurance: context and problem statement

business model to work with this segment, and another indicated that the attempts in these
directions have been so far motivated by Corporate Social Responsibility considerations more
than by commercial imperatives, and are therefore limited in size.

4.3.5. Discussion

Despite the small sample size of the people interviewed, it is possible to draw preliminary
conclusions. There is a consensus around the fact that multi-peril crop insurance cannot exist
without direct public support. Reinsurers also rely on governments to improve market
transparency and expect them to implement consistent and predictable policies. The question
whether governments should bear part of the risk themselves is debated and there is no
consensus of what this role should exactly be. Another area of discussion lies in the perspectives
for the market. Is index insurance likely to disrupt agricultural insurance practices? Are
smallholder farmers a viable business opportunity for the future? As the research intends to
work on the business model and the public-private partnerships in agricultural insurance, the
interviews allowed to clarify where there is a consensus among practitioners, and what issues
are still in debate.

5. Problem statement

From the observations made in the four previous paragraphs, it is possible to conclude that
agricultural insurance cannot develop beyond a nascent stage on a purely market-based
functioning. The nature of agricultural risks prevents the development of an efficient
agricultural insurance market without public intervention, though some recent innovations bear
the potential to improve agricultural insurance affordability.

5.1. Research hypotheses

The current thesis will review the rationale and modalities of public intervention in agricultural
insurance markets. Afterwards, it will test three research hypotheses addressing issues related
to the agricultural insurance business model and public policy choices. Those hypotheses
examine challenges that are affecting the development of large, efficient and inclusive
agricultural insurance markets.

Policy-makers have various instruments available to reduce farmers’ vulnerability and income
volatility. In some cases, those instruments can undermine each other. It is the case between
disaster assistance funds, providing relief to farmers suffering production losses, which

47
 Chapter 1 - The business model of agricultural insurance: context and problem statement

compete with agricultural insurance. The first hypothesis will analyze whether supporting
agricultural insurance through the provision of premium subsidies can be a substitute to a
disaster fund (H1).

Smallholder farmers living in developing economies usually have no access to agricultural

insurance markets. There is simply no offer, or the insurance contracts are not properly designed
to address their specific constraints. Working with smallholder farmers in low-income
environments requires more changes than just offering the same services at a lower-scale or a
lower price. It is essential to reengineer products and processes in order to satisfy the needs of
this particular segment. This is the subject of the second hypothesis (H2).

Financial services providers can play a role in the distribution of insurance policies. Credit and
insurance complement each other and there are efficiency gains in offering both services
together. In low-income environments, microfinance institutions (MFIs) have established
trusted relationships with the targeted clients. However, MFIs will only get involved in the
distribution of insurance if there is a business case. This is what the third research hypothesis
will analyze (H3).

The three research hypotheses formulated are listed and detailed below.

(H1) For governments willing to reduce farmers’ income volatility, agricultural insurance
can be a substitute to disaster assistance programs.

Adopting the policy-makers’ perspective, the thesis will investigate whether an agricultural
insurance product can meet its objective to reduce farmers’ income volatility. Then it will
analyze the implications of switching from one instrument to the other, considering the impact
on public expenditures as well as qualitative criteria.

(H2) Offering agricultural insurance to smallholder farmers requires a specific design of

product features and business operations.

This hypothesis implies that it is not possible to do “business-as-usual’ at a lower scale when
working with smallholder farmers. To test this hypothesis, the thesis mobilizes different
theoretical frameworks to assess the client value of agricultural insurance. These theoretical
frameworks are used to analyze different dimensions of the client experience, such as the risk
coverage, the distribution model, the premium collection, the enrolment and claim processes.

(H3) There is a business case for microfinance institutions to offer insurance to their
clients
48
 Chapter 1 - The business model of agricultural insurance: context and problem statement

In order to test this hypothesis, the research will examine the reasons that could support MFIs’
involvement in the insurance business. It will also review the risks and challenges it poses to
MFIs, in order to see under which conditions the business case holds.

5.2. Data and case studies

Each research hypothesis is going to be tested in two case studies and one survey.

(C1): Drought risk for grasslands in France

The research hypothesis H1 is tested on the case of grassland insurance in France. A disaster
fund is currently running in France to indemnify cattle breeders when grasslands produce
animal feed below a certain threshold. A commercial index insurance product was launched in
the market in 2016 to protect farmers against the very same risk. The analysis will use two types
of data to compare the two risk management instruments. First, it presents the disaster fund
payouts related to grassland production losses over the period 2003-2015. This data was
obtained from the French Ministry of Agriculture. Second, the research uses the forage
production index that serves as a reference for the commercial insurance product. The company
Airbus Defense and Space produces this index, and the index was made available by Pacifica,
the insurer commercializing the grasslands insurance product.

(C2): Agricultural insurance for smallholder farmers in Kenya

The second research hypothesis is tested on a case study of the Agricultural Climate Risk
Enterprise Ltd (ACRE) in Kenya. The company offers three crop insurance products and one
livestock insurance product. A field research was conducted, with interviews of different
stakeholders and an analysis of the company’s documentation. This research allowed collecting
descriptive data of the insurance contracts features, as well as of the business processes and
operations.

(S3) Survey of 36 MFIs in Africa and Asia

The third research hypothesis (H3) is tested thanks to an online survey directed towards MFIs
in the portfolio of the Grameen Crédit Agricole Foundation, a social investor. An invitation to
fill the survey was sent to all the 58 MFIs who were in the Foundation’s portfolio in September
2017, and 36 filled it. The answers provided by the respondents was completed by financial and
operational data on these MFIs collected and analyzed by the Grameen Crédit Agricole

49
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Foundation itself. This survey allows understanding better how MFIs perceive insurance, what
are the expected benefits and challenges of distributing insurance policies to their clients.

6. Conclusion

Agricultural risks are ill suited for the development of insurance products, mostly due to their
covariant nature. The development of agricultural insurance markets around the world has been
uneven, and somehow correlated with the level of government support. Smallholder farmers in
developing countries remain in their vast majority excluded from agricultural insurance
markets. However, technological innovations and new approaches such as index insurance are
pushing the boundaries of insurability. The three research hypotheses will be tested in different
types of economies and study some of the main challenges hindering the development of
agricultural insurance markets. They should also provide insights for both insurance
professionals and policy-makers on opportunities and strategies to relieve these barriers and
extend coverage to a larger number of farmers.

50
 Chapter 1 - The business model of agricultural insurance: context and problem statement

References

Agroseguro. (2015). El sistema español de seguros agrarios en cifras.

[Link]
groseguro__2015-05-18__baja.pdf

Agroseguro. (2017). El seguro agrario en cifras 2016.

[Link]
URO_AGRARIO_EN_CIFRAS_2016.pdf

Antón, J. (2009). Risk management in agriculture—A holistic conceptual framework

(TAD/CA/APM/WP(2008)22/FINAL). OECD.

Barnett, B. J., & Mahul, O. (2007). Weather index insurance for agriculture and rural areas in
lower-income countries. American Journal of Agricultural Economics, 89(5), 1241‑1247.

Biener, C. (2013). Pricing in microinsurance markets. World Development, 41, 132‑144.

Biener, C., & Eling, M. (2012). Insurability in microinsurance markets : An analysis of

problems and potential solutions. The Geneva Papers on Risk and Insurance-Issues and
Practice, 37(1), 77‑107.

Brau, J. C., Merrill, C., & Staking, K. B. (2011). Insurance theory and challenges facing the
development of microinsurance markets. Journal of Developmental entrepreneurship, 16(04),
411‑440.

Carter, M. R. (2011). Innovations for managing basis risk under index insurance for small farm
agriculture. Ferdi Policy Brief, 41.

Christen, R. P., & Anderson, J. (2013). Segmentation of smallholder households : Meeting the
range of financial needs in agricultural families. Focus Note, 85.

Collin, C. (2018). Towards a working crop insurance market : An integrated strategy of

systemic risk management. Paris 10.

Hardaker, J. B. (2000). Some issues in dealing with risk in agriculture.

Hardaker, J. B., Huirne, R. B. M., Anderson, J. R., & Lien, G. (2004). coping with risk in
agriculture. Oxfordshire. CABI publishing.

Hazell, P., Anderson, J., Balzer, N., Hastrup Clemmensen, A., Hess, U., & Rispoli, F. (2010a).
L’assurance basée sur un indice climatique : Potentiel d’expansion et de durabilité pour
l’agriculture et les moyens de subsistance en milieu rural. Fonds international de
développement agricole et le Programme alimentaire mondial.

51
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Hazell, P., Anderson, J., Balzer, N., Hastrup Clemmensen, A., Hess, U., & Rispoli, F. (2010b).
The potential for scale and sustainability in weather index insurance for agriculture and rural
livelihoods. World Food Programme (WFP).

Hazell, P. B. (1992). The appropriate role of agricultural insurance in developing countries.

Journal of International Development, 4(6), 567‑581.

Hazell, P., Sberro-Kessler, R., & Varangis, P. (2017). When and how should agricultural
insurance be subsidized ? Issues and good practices. World Bank.

Hess, U., Hazell, P., & Kuhn, S. (2016). Innovations and emerging trends in agricultural
insurance. Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH.

Holzmann, R., & Jørgensen, S. (2001). Social risk management : A new conceptual framework
for social protection, and beyond. International Tax and Public Finance, 8(4), 529‑556.

Index-based livestock insurance as an innovative tool against drought loss : Good practices
and impact analysis from northern Kenya. (2018). CTA. [Link]

Kalra, A., & Xing, L. (2013). Partnering for food security in emerging markets (No 1/2013;
Sigma). Swiss Reinsurance Company.

Mahul, O., & Skees, J. R. (2007). Managing agricultural risk at the country level : The case of
index-based livestock insurance in Mongolia. World Bank Policy Research Working Paper,
4325.

Mahul, O., & Stutley, C. J. (2010). Government Support to Agricultural Insurance : Challenges
and Options for developping countries -. World Bank.
[Link]
nment+Support+to+Agricultural+Insurance&ots=mPxoweHjmn&sig=awwd6G0iPOxqRaTJ
ME4sYbVo6Mw#v=onepage&q=Government%20Support%20to%20Agricultural%20Insura
nce&f=false

Miranda, M. J., & Glauber, J. W. (1997). Systemic risk, reinsurance, and the failure of crop
insurance markets. American Journal of Agricultural Economics, 79(1), 206‑215.

Sandmark, T., Debar, J.-C., & Tatin-Jaleran, C. (2013). The emergence and development of
agriculture microinsurance. Luxembourg: Microinsurance Network.

Shaik, S., Barnett, B. J., Coble, K. H., Miller, J. C., & Hanson, T. (2006). Insurability conditions
and livestock disease insurance. The Economics of Livestock Disease Insurance: Concepts,
Issues and International Case Studies, 53.

Skees, J. R., Barnett, B. J., & Hartell, J. (2005). Innovations in government responses to
catastrophic risk sharing for agriculture in developing countries. workshop Innovations in
Agricultural Production Risk Management in Central America: Challenges and Opportunities
to Reach the Rural Poor, Antigua, Guatemala, 9‑12.

52
 Chapter 1 - The business model of agricultural insurance: context and problem statement

Skees, J. R., Hazell, P., & Miranda, M. (1999). New approaches to public/private crop yield
insurance. Unpublished Working Paper, The World Bank, Washington, DC.

53
54
 CHAPTER 2

FROM SUBSIDIES TO PUBLIC-PRIVATE PARTNERSHIPS: PUBLIC

INTERVENTIONS IN THE AGRICULTURAL INSURANCE MARKET

55
From subsidies to public-private partnerships: public interventions in the
agricultural insurance market

1. Context

Mahul and Stutley (2010) conducted a review of agricultural insurance programs around the
world for the World Bank. They find that in 2007, 104 countries in the world were offering
some kind of crop or livestock insurance. The authors surveyed more specifically 65
agricultural insurance programs, in high income, middle-income and low-income countries. 38
out of these 65 countries (58%) subsidize premiums. Premium subsidies are the most common
form of subsidization of agricultural insurance, but governments can also provide public
reinsurance subsidies, subsidies on administrative and operational costs, and loss adjustment
subsidies. The total global cost for governments reached around $10 billion, for a total gross
premium of $15.1 billion. Since then, the figures are likely to have risen up: emerging countries
like China and India have significantly expanded their agricultural insurance programs, while
at the same time many low and middle-income countries have launched pilots. Hazell et. al
(2017) estimate that the United States, China and India are altogether collecting $17.7 billion
per year in agricultural insurance premiums.

As governments globally are dedicating increasing amounts of public resources to support the
development of agricultural insurance markets, it is worthwhile to discuss the relevance of this
policy trend. There is a strong corpus of academic literature analyzing market failures in
agricultural insurance and the motives and means of governments to intervene to address these.
Another source of literature is made of the reports published either by international
organizations or by insurers and reinsurers. These documents provide information on the actual
state of the market globally, its trends and figures. They also give qualitative insights on the
challenges hindering the growth of the agricultural insurance market. Therefore, as a
preliminary work regarding the three research hypotheses presented in chapter 1, it appears
relevant to undertake a systematic literature review on the question of public policies in
agricultural insurance. Many of the questions raised by the research work have already been
addressed by scholars as well as practitioners. This literature review will aim at making a
selection and a synthesis of the state of the knowledge in the field, qualify these findings and
identify gaps where further research is needed.

56
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

This chapter will first present the methodology adopted for the literature review It will then
present the main findings on three main topics: it will question the rationale behind the use of
public money to subsidize agricultural insurance; it will also examine whether there is a
possibility to increase the efficiency of subsidies; finally, it will look at other forms of public
intervention that can support the development of agricultural insurance. Finally, we will discuss
these findings and put them in perspective with the doctoral work and the research hypotheses
that are going to be developed in the further chapters.

2. Literature review methodology

2.1. Research objectives

The main objective of the literature review is to determine the relevant instruments for public
policies regarding agricultural insurance.

The specific objectives of the literature review are to answer three questions: why should
governments subsidize agricultural insurance; are premium subsidies an efficient instrument;
what are the other instruments available to policy-makers to support agricultural insurance
market development.

2.2. Data

The approach adopted is a systematic literature review. It compiles documents from academic
literature as well as from the grey literature. The articles, reports and other documents included
in this literature have been found using two search engines specialized for academic work:
Google Scholar and Science Direct These databases were chosen because they cover a broad
range of journals. Even if these databases provide mostly academic sources, Google Scholar
also identifies documents from the grey literature. Grey literature was also identified by
inquiries on the general Google search engine.

The research of sources through search engines relied on the following keywords:
- “public-private partnerships” AND “agricultural insurance”
- “public policy” AND “agricultural insurance”
- “government” AND “agricultural insurance”
- “market failures” AND “agricultural insurance”
- “subsidies” AND “agricultural insurance”

57
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

The keywords were entered in English and in French, sources in the two languages have been
included in the literature review.

Reference tracing led to the identification of relevant sources, beyond those initial keywords.
The literature review also identified sources by browsing the websites of specialized
institutions: the Index Insurance Forum, managed by The Global Index Insurance Facility of
the World, and the Impact Insurance Facility hosted within the International Labour
Organization. The researcher also mobilized publications that he had identified in prior work
and prior studies. He attended several conferences and seminars (in presence or online) that
participated in enriching the corpus for the literature review. In particular, the researcher was
rapporteur at the International Microinsurance Conference in Casablanca in 2015 (Qureshi &
Reinhard, 2016). The data collection took place between 2015 and 2018, and include documents
published between 1973 and 2018.

Regarding the selection of documents, sources from the grey literature were included only when
they were published after 2010, in order to capture the most recent knowledge and reflect the
latest developments of the market. On the contrary, no time limit was set for academic sources,
as long as they were relevant. In particular, market failures affecting catastrophic and
agricultural risks markets have been thoroughly studied and analyzed from the 1970s to the
1990s, and more recent academic work often keeps referring to these fundamental sources. In
order to ensure a form of quality control, documents considered in the literature review are
articles from peer-reviewed journals, books or individual book chapters, articles and reports
published by reputable authors and/or institutions. The relevance of identified sources was
assessed at several stages, applying filters: at each step, some documents are discarded and
shortlisted documents are upgraded to the next step, for a further analysis and so on. The process
went on as follows (see below Figure 6 : Flow chart of literature search and selection):
1) Selection based on the title: 301 documents shortlisted for the next step
2) Screening based on the abstracts: 161 documents for the next step
3) Screening based on full text reading: 71 documents comply with the relevance and quality
criteria regarding the three questions asked for the literature review.

Documents were excluded along the different screening steps either because the focus of the
article was not directly addressing the research questions for the literature review, or because
they were redundant (several versions of the same article for example, or document not bringing
any new relevant idea from prior documents).

58
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

Figure 6 : Flow chart of literature search and selection

3. Literature review: findings

The systematic literature review included 71 documents (see Figure 6 : Flow chart of literature
search and selection). The documents were grouped per theme, each document addressing one,
two or three research questions. The results for each research question are presented below.

3.1. Why should governments subsidize agricultural insurance?

Arguments in favor of public intervention usually fall into two categories. On the one hand,
governments can use economic arguments like market failures, and the existence of
externalities to justify public support for agricultural insurance. The second category relates to
the pursuit of political and social goals. Premium subsidies can sometimes be seen as more
efficient or more politically acceptable than other instruments to implement public policies.

Whatever the rationale behind the use of subsidies, policies in favor of agricultural insurance
intend to increase the level of protection of farmers against natural risks, and boost the demand
for agricultural insurance.

59
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

3.1.1. To address market failures and externalities

Informational asymmetries

Any insurance contract faces the problems of adverse selection and moral hazard. Adverse
selection is the result of the imperfect correlation between the observable characteristics of
agents and their risk exposure (Mahul, 1998). An insurance company will calculate premium
rates according to the average population. This pricing leads the riskiest farmers to purchase
more insurance, increasing the exposure of the insurer’s portfolio. To maintain the equilibrium,
the insurer increases premiums, pushing out the less exposed farmers, and increasing further
the level of risk for the insurer. Eventually, premiums will be so high that no consumer is willing
to buy insurance. Making insurance compulsory can be a way to overcome adverse selection,
but it can be difficult to implement. In India, strong incentives to purchase insurance, like the
possibility to obtain a credit at a concessional rate, coupled with premium subsidies, have
increased the penetration of crop insurance and made possible the creation of a large pool of
risks. Nevertheless, there is strong evidence that the National Agricultural Insurance Scheme
was plagued with adverse selection (Mahul et al., 2012) before it was drastically reformed.

Moral hazard is also a consequence of informational asymmetries. Insured farmers can modify
their farming practices and take more risk. Ramswami (1993) shows that the introduction of
insurance can lead farmers to reduce their use of preventive measures (such as pesticides). The
presence of a deductible can mitigate this effect though (Sundt, 1999).

Adverse selection and moral hazard are related to the difficulty faced by insurers to accurately
classify risk and monitor farmers’ behavior, especially when data is scarce and not qualitative.
Informational asymmetries can result in high, and potentially prohibitive, premium rates, which
prevents the creation of a private agricultural insurance market. Index insurance can overcome
information adverse selection and moral hazard as it relies on an exogenous variable.

Systemic risks

Another market failure identified in the literature is the existence of systemic risk. Insurance
markets function well for uncorrelated risks. For highly correlated risks, like price volatility,
derivatives markets offer hedging possibilities. But as Skees and Barnett (1999) put it, natural
risks are “in-between” risks, for which traditional insurance markets nor financial markets offer
satisfying solutions.

60
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

A farmer’s individual risk can be broken down into a specific risk (sometimes referred to as
“idiosyncratic risk”) and a systematic risk (Mahul, 1998). The first one has to do with farming
practices, or localized events like hail. This first component is independent among farmers so
insurers can efficiently pool these risks. Systematic risk affects a large number of farmers at the
same time. It is the consequence of events like climatic shocks or a widespread disease. Such
events affect an entire region, and are the reasons for the high spatial correlation of yields among
farmers. In this case, risk pooling does not work efficiently. Systemic risk can jeopardize the
solvency of insurance companies. Insurers need to make large reserves in order to cope with
such risks. The cost of mobilizing capital to build such reserves often deters private insurers
and investors to engage in agricultural insurance.

For Miranda & Glauber (1997), systemic risk, and not information asymmetries, is the real
market failure for crop insurance. This finding has policy implications: in the presence of
information asymmetries, private insurance will never be viable. However, if systemic risk is
the problem, access to affordable reinsurance can enable a viable crop insurance market. Indeed,
the systematic nature of a risk is linked to the size of the pool for diversification. A drought can
be systematic at the country level, but insurable at a global scale. As they are supra-national
actors, reinsurers can mutualize risks at a much larger scale than national insurers. Reinsurers
diversify risks among different geographies and different business lines. Reinsurance protects
the solvency of national insurance companies when these face exceptional losses. However,
reinsurers bear high transaction costs that limit their efficiency (Mahul, 1998). Furthermore,
even if several authors have proposed pricing models based on the separation between systemic
and non-systemic risks (Chantarat et al., 2013; Mahul, 2001; Miranda, 1991), the current
practice remains to transfer risks indiscriminately in one bulk (Collin, 2018).

Some authors dispute the existence of a market failure. For Jaffee & Russel (2006), the public
sector should not be a substitute to the private sector, private markets are in capacity to cover
catastrophe risks. For the authors, as loss modeling improves, there is no actuarial impediment
for a purely private catastrophe insurance, as long as insurance companies have access to
efficient capital markets. For Priest (1996),the State has no competitive advantage compared to
the private sector in any of the three risk reduction functions of insurance: risk aggregation, risk
segregation and control of moral hazard. He actually argues that the existence of governmental
insurance or the potential of governmental relief is crowding-out private insurers. In the case
where governments decide to maintain insurance subsidies or publicly-funded disaster
assistance programs, Jaffee and Russel (2006) support the idea that the government should use

61
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

the market as a model for public intervention, adopt actuarial principles and encourage private
insurers to participate.

Externalities

Goods and services produce externalities when the benefits they bring to society (positive
externalities) or the costs (negative externalities) are not reflected in the market price.
Individuals do not factor these effects in their purchasing decision. In the case of insurance, the
global welfare of society increases when individuals purchase insurance (R. Vargas Hill et al.,
2014). This effect is well-known for health insurance, as insured individuals are more likely to
seek treatment rapidly, which will limit the spread of infectious diseases. In the case of
agricultural insurance, it has been demonstrated that when individuals buy insurance, others
learn about it. In China, Cai et al. (2011) show that receiving an insurance brings an
informational externality, and increases take-up in the next period, as members of the
community learn from seeing others’ experience. Another positive externality brought by
insurance lies in the fact that insured households are less burdensome for society when they
experience a shock. Households losing everything after a catastrophe pose a liability to society
(Cummins & Mahul, 2009). This is why property insurance is compulsory in some countries,
like Turkey or the United States (R. Vargas Hill et al., 2014). It also holds true for agriculture:
if farmers lose the entirety of their crop after a shock, governments will be compelled to provide
some kind of relief.

Just like households, private companies do not factor externalities in their investments
decisions. For example, an insurer investment in improving clients’ knowledge of insurance
may increase demand, but the returns of the investment will be public and benefit competitors
as well. Furthermore, the impact of these trainings on insurance demand is uncertain, with some
studies showing a positive impact (Cai et al., 2011; Ruth Vargas Hill et al., 2013), while others
show no or moderate increase in take-up (Dercon et al., 2011). Therefore, in the absence of
subsidy, there is little incentive to invest in insurance literacy, despite positive externalities.

3.1.2. To pursue political and social goals

If market failures in agricultural insurance markets are well documented, governments find
other arguments to subsidize insurance. Subsidies are an instrument available for governments
to pursue social and political goals.

Insurance as a substitute for disaster assistance programs and safety nets

62
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

When their citizens are affected by catastrophic events, governments are compelled to provide
assistance to the population. In some cases, governments dedicate a part of their yearly budget
to build reserves in disaster relief funds. These reserves require a strict discipline to be
maintained: with scarce fiscal resources and many pressing needs, the temptation to use these
funds for other purposes is great. In case of large-scale losses, the capacity of disaster funds are
rapidly exhausted and governments must rely on international humanitarian appeals to fund the
humanitarian response. It is also well-known that external shocks can push vulnerable
households into a poverty trap (Barnett et al., 2008), with effects on poverty being observable
several after the shock occurs.

Insurance reduces the need for emergency relief as affected farmers receive a payout that
compensate at least part of the loss. It reduces the volatility of farm income and prevents
farming households from falling into poverty after an external shock. From a public finance
perspective, the cost of insurance subsidies is also more stable and predictable than disaster
relief costs. It is also established that the quicker the humanitarian response is brought, the
better the recovery for the affected population and the cheaper the cost of the response is
(Dercon et al., 2005). Insurance also allows a more timely response (Hazell et al., 2010),
without the need to go through a lengthy humanitarian appeal process.

Overall, governments can use insurance as a substitute to disaster assistance programs, allowing
them to better plan public expenditures and be more cost-effective.

Insurance as a tool for the development of agriculture

Governments can decide to use premium subsidies to support the agricultural sector and food
production. In many middle and high-income countries, insurance serve to support farm income
(Hazell et al., 2017). This is the case when the annual insurance payout exceeds the
unsubsidized part of the premium.

Governments also often explicitly use insurance as a tool to increase access to unlock access to
credit for farmers (Meyer et al., 2017). Indeed, one of the reasons why financial institutions are
reluctant to lend to farmers is the existence of production risk. By allowing lenders to absorb
large covariate shocks, insurance should give the capacity to help financial institutions grow
their agricultural loan portfolio for larger amounts and a larger number of farmers. However,
there is no evidence that subsidized insurance has modified lending practices ex-ante (larger
volumes, broader outreach, cheaper rates or longer maturities), nor that it has protected lending
portfolios ex-post (Hazell et al., 2017).

63
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

There is ample evidence though that insurance has an impact on farmers’ behavior. Randomized
control trials have shown that insured farmers in Ghana increased farm expenditures, cultivated
land and use of fertilizers (D. Karlan et al., 2014). In India, farmers who were offered weather-
based index insurance adopted higher-yield seeds (Cole et al., 2017; Mobarak & Rosenzweig,
2013). Indeed, investment decisions are affected by the exposure to climate risk. Even if
climatic risk is not the one with the bigger impact on farm profitability, it is the most difficult
risk to diversify, and as a consequence the one with the more influence on consumption and
investment decisions (Rosenzweig & Binswanger, 1993). This mechanism strengthen
inequalities, as the most vulnerable and poorest farmers are the ones most likely to adopt low-
risk low-productivity strategies. A better access to risk-coping mechanisms such as insurance
is likely to modify farmers’ investments, to increase the profitability of their investments while
reducing income inequalities. Carter et al. (2016) developed a theoretical model to assess under
which conditions index insurance can boost the adoption of new agricultural technologies. They
show that index insurance is the most relevant when risk is high, highly covariant, and collateral
requirements are high. In low collateral environments, the lender carries a substantial part of
the risk. Index insurance will be effective if it is linked with the credit, as insurance reduces the
risk rationing practiced by financial institutions and increases the volume of credit available to
farmers.

Promoting inclusive insurance

Despite being particularly vulnerable, poor rural households are often excluded from
agricultural insurance. Subsidizing premiums is a common mechanism to increase the
participation of low-income households in insurance markets and promote equity in insurance
coverage (R. Vargas Hill et al., 2014).

Research has shown that when insurance is subsidized, demand increases. Several studies
attempted to estimate the price elasticity of agricultural insurance, through the randomization
of premium subsidies. Price-elasticity of agricultural insurance was estimated at 0.94 in China
(Cai et al., 2011), meaning that the demand for insurance falls by 0.94 percent when the price
increases by 1 percent. Other studies estimated price-elasticity in India (Cole et al., 2013; Ruth
Vargas Hill et al., 2013), Ghana (D. S. Karlan et al., 2012) and Ethiopia (Berhane et al., 2012)
for values comprised between 0.55 and 0.99.

One difficulty when using premium subsidies to increase insurance access for low-income
households is to make sure that the subsidies reach the intended target. Evidence shows that

64
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

insurance take-up is positively correlated with wealth (Eling et al., 2014), so it is likely that
richer households will reap more benefit from this policy. For an efficient use of public
resources, policy-makers must primarily identify precisely who the target population and design
appropriate targeting mechanisms to ensure that the subsidies meet their purpose.

3.2. Are subsidies efficient?

There are different theoretical and empirical arguments to justify the existence of subsidies for
agricultural insurance. In the absence of subsidies, market failures are precluding the formation
of private agricultural insurance markets. Many policy-makers around the world are supporting
agricultural insurance, whether it is to protect farmers’ income, reduce their vulnerability or
foster agricultural development. Even when governments recognize the relevance of
agricultural insurance, they should ask if premium subsidies are the better use of public money,
and if there are ways to increase their efficiency.

3.2.1. Defining smart subsidies

The notion of “smart subsidies” first appeared in the microfinance industry. Indeed, the
microfinance promise was based on the idea that it could lift millions of people out of poverty
while being profitable, or at least self-sufficient. Subsidies proved to be essential for many
microfinance actors though, at least in the first years of operation. The question for many donors
and governments was to define the appropriate level of subsidization.

For Jonathan Morduch (2005), “smart subsidies maximize social benefits while minimizing
distortions and mistargeting”. This definition, introduced initially for the microfinance industry,
can find many other applications, including for agricultural insurance.

R. Vargas Hill et al. (2014) propose some key principles for smart subsidies in inclusive
insurance:

(i) Clear objective: subsidies should address an identified market failure or pursue a
clear social goal
(ii) Transparent: rules of eligibility must be clearly stated.
(iii) Targeted: subsidies should reach those in need in priority. For example the
HARITA/R4 project in Ethiopia relied on the existing national safety net program
to identify beneficiaries (R. Vargas Hill et al., 2014).

65
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

(iv) Monitoring and evaluation: it is essential to define indicators of success and put in
place a proper monitoring system in order to review the efficiency of subsidies.
(v) Exit-strategy or long-term financing: once in place, subsidies are difficult to
withdraw. Subsidies must have a clear exit-strategy or long-term financing before
their introduction.
(vi) Costs-contained: the fiscal burden of subsidies must be anticipated and controlled.

Subsidies will influence farming practices, so policy-makers should examine what incentives
do premium subsidies give. By reducing the consequences of production losses, premium
subsidies may encourage the cultivation of crops in areas where they would not be viable
otherwise. Agricultural insurance programs can also have ambiguous effects on the
environment. Some may be positive, as through moral hazard effects, insured farmers tend to
reduce the use of agricultural chemicals (Babcock & Hennessy, 1996; V. H. Smith & Goodwin,
1996), resulting in less land and water pollution. However, some authors also showed that
subsidized insurance programs have increased soil erosion (V. H. Smith & Goodwin, 2003) and
encouraged the expansion of cultivation on environmentally sensitive land (Goodwin et al.,
2004). These results outline the importance of designing subsidies with the right incentives.

Carter et al. (2014) looked at the specificity of index insurance. The quality of an index
insurance contract is hard to assess for consumers. On top of that, unlike other insurance
products, index insurance payouts are not frequent, limiting the possibilities of learning from
others’ experience. Therefore, governments and donors can decide to subsidize insurance
“strategically” in order to help build the market. These subsidies pursue two goals: establish a
market large enough so that private insurers are able to recover their fixed costs and make a
profit; give time to insurers to accumulate data, and to consumers to learn about index insurance
contracts. Such subsidies should be time-bound and are not intended to become permanent.
Long-term subsidies are also relevant for catastrophic risks, given the social externalities.

3.2.2. Subsidies design

Insurance subsidies can take several forms. The survey conducted by Mahul & Stutley (2010)
gives us an overview of the practices in 2008 for 65 countries. Total premiums collected amount
to $15.1 billion, of which 44 percent are covered through premium subsidies, and additional 24
percent in administrative, operational and excess claims subsidies.

66
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

63 percent of the studied countries offer some kind of upfront premium subsidy for crop
insurance, and 35 percent of them offer universal premium subsidies. These subsidies cover a
fixed proportional amount of the insurance premiums for all farmers, crops, and regions. Certain
types of contracts (group contracts and multi-year contracts for example) are more heavily
subsidized to encourage good risk-management practices. Some countries (11 percent of the
sample) will also implement more targeted subsidies for vulnerable or marginalized farmers,
specific crops or regions. The survey provides less detail on livestock insurance, though we
know premium subsidies are available for 33 percent of the surveyed countries.

Administrative and operational subsidies are also available in some countries (16 percent of the
sample). These subsidies are paid directly to insurers and reinsurers. Finally, governments can
cover excess claims. In India for example, before reform, the National Agricultural Insurance
Scheme (NAIS) was putting a cap on insurance premiums, meaning that premiums were not
determined on actuarial principles. Subsidies were essential to cover claims, as insurance was
priced below the actuarially fair price.

3.2.3. Evaluating the efficiency of premium subsidies

Subsidies can take different forms, but premium subsidies are the preferential tool of many
governments, and the destination of most public funds to support agricultural insurance. Once
the objectives and principles of premium subsidies are clearly stated, it is the policy-makers’
responsibility to monitor these subsidies and evaluate their efficiency.

The Federal Crop Insurance Program in the United States, the largest crop insurance program
in the world, has been extensively studied. Critics point the cost of the subsidies, which have
grown constantly since the 1980s without eliminating ad-hoc payments when farmers face large
losses (Glauber, 2007). As the US Farm policy shifted from direct payments to premium
subsidies, the Federal Crop Insurance program has become a non-efficient way to transfer
revenue to farmers (Babcock & Hart, 2015; V. H. Smith & Glauber, 2012). Analyzing three
different legislative acts, Smith (2017) even sustains that the successive reforms of the Federal
Crop Insurance Program were intended to benefit farmers and the private insurance sector, as a
result of implicit and explicit coalition of farm interest groups and crop insurance interest
groups. Some authors consider that a better design could improve the efficiency of these
subsidies: Babcock and Hart (2015) defend that the cost of subsidies could be reduced by
excluding agents’ commissions; Skees (1999) proposes the introduction of subsidies in the form

67
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

of vouchers that farmers would be free to use in any form of risk management instrument they
find relevant.

From a macro-economic perspective, economists have tried to estimate how agricultural

insurance subsidies affect society’s welfare. Regös (2015) builds the model of a small,
developed economy in which agriculture represents a small part of the GDP, and in which prices
are determined by international markets. He concludes that premium subsidies do not result in
significant change of the economy’s performance, but leads to a transfer of revenue from
consumers to farmers. This transfer does not significantly affect consumers’ welfare though,
due to the small weight of agriculture in the economy. This impact is likely to be larger, though
still limited, in a developing economy. Gohin (2017) also analyzed the welfare effect of
agricultural insurance subsidies, examining the Federal Crop Insurance Program in the United
States. The originality of his work is that he builds a dynamic model, taking into account the
externalities of insurance on production and investment. He finds that withdrawing federal crop
insurance subsidies would lead to a $1.7 billion welfare loss for the American society. This
result does not entirely dismiss the critics made on the Federal Crop Insurance Program, but it
shows the importance of integrating spillover effects of premium subsidies when conducting a
cost-benefit analysis.

Outside the United States, welfare analyses of premium subsidies programs are scarce. The
Japanese (Tsujii, 1986) and Mexican (Bassoco et al., 1986) crop insurance subsidies have been
studied in the 1980s, to conclude that net social returns were extremely low relatively to their
cost. However, these were evaluation of classical Multi-Peril Crop Insurance programs. Since
then, the landscape and design of agricultural insurance around the world has changed
dramatically (Hess et al., 2016; Mahul & Stutley, 2010). The World Bank carried an ex-ante
evaluation, prior to the introduction of an insurance program in Bangladesh (Quayyum et al.,
2018). The report evaluates the fiscal burden for the government: between $6 million and $9
million per year, when about 10 percent of the area cultivated in paddy is covered. This amount
represents approximately 1 percent of the Ministry of Agriculture’s annual budget. The welfare
analysis shows a 41 percent increase in crop income if insurance unlocks access to credit and
higher-yield varieties. This type of evaluation is very insightful for decision-makers.
Unfortunately, we have not been able to find ex-post evaluation premium subsidies besides the
ones already mentioned above.

68
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

Ideally, policy-makers should not only assess whether subsidizing premiums increase society’s
welfare, they should also be able to compare subsidies with alternative investments and their
respective expected or observed social returns. A team of researchers compared weather-index
insurance premium subsidies with other types of public interventions in the Senegalese
groundnut basin (Ricome et al., 2017). Coupling a crop growth model and an economic model,
and using a weather-index insurance product existing in the market, they find that premium
subsidies increase farmers’ welfare, but only for those living in the driest areas, as for farmers
living in the wetter area, rainfall variations are not their major constraint. Their analysis also
finds that, if subsidies increase farmers’ utilities, it is much less efficient than other instruments,
like subsidizing credit, inputs, or simply transferring the amount of subsidy in cash.

There is now a large body of evidence that agricultural insurance brings benefits farmers and
society in general (Bertram-Huemmer & Kraehnert, 2015; Cole et al., 2017, 2017; Mobarak &
Rosenzweig, 2012). It also is true that no agricultural insurance programs managed to achieve
a large scale without some sort of subsidies, especially premium subsidies (Mahul & Stutley,
2010). Therefore, many governments and donors are willing to finance premium subsidies as a
way to increase agricultural insurance penetration and reap its benefits. Nevertheless, such
subsidies could be more efficient if they were better designed. Governments may also use these
resources in other ways to achieve better outcomes or higher welfare for the society. Evidence
is still lacking though to evaluate if subsidizing agricultural insurance premium is the best use
of public funds. Premium subsidies tend to be the preferential instruments for governments
willing to increase agricultural insurance coverage. However, the role of governments is not
limited to premium subsidies, and other mechanisms are available to support the development
of agricultural insurance.

3.3. Beyond subsidies: other instruments for public intervention

Many governments subsidize agricultural insurance premiums with the objective to decrease
the market price of insurance and increase take-up. Nevertheless, other mechanisms may
achieve the same goal. Moreover, price is not the only challenge limiting agricultural insurance
penetration. This finding calls for renewed public policies addressing specifically these other
challenges.

69
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

3.3.1. Risk-sharing arrangements between the public and private sectors

We saw that several authors deny the existence of a market failure for agricultural insurance,
as international reinsurance and capital markets are in capacity to absorb large losses (Jaffee
& Russell, 2006; Priest, 1996; Skees, 1999). Nevertheless, experience shows that reinsurance
is not always available for catastrophic losses. In order to facilitate the supply of agricultural
risk coverage, governments may decide to create their own insurance company and bear
agricultural risk directly, or create a reinsurance company, with the objective that private
insurers will progressively retain more and more risk.

Risk layering in a context of uncertainty

Building a viable agricultural insurance market takes time; insurers need to be confident about
their risk exposure. For Carter (M. Carter, 2013), the lack of historical data creates uncertainty
on potential losses. Insurers are uncertainty-averse, so they tend to price insurance not
according to the average expected losses, but based on worst-case scenarios. These “uncertainty
loadings” lead the private sector to charge commercial premiums 50%, 100% or even more
above the actuarially fair premium (M. Carter, 2013). In some situations, when data is too
sparse, insurers may simply refuse to price insurance contracts.

Therefore, there is a role for the public sector. Governments can be uncertainty-neutral, so a
public reinsurance fund could price index insurance according to the average scenario, not the
worst-case estimates, as it is typically the case for the private sector. In this public-private
partnership, each actor can retain a different level of risk, according to its own capacity. This
approach called “risk layering” (Antón, 2009; Carter, 2013; Mahul & Stutley, 2010; Jerry R.
Skees & Barnett, 2006) intends to improve cost-efficiency through an optimal combination of
technical and financial tools.

70
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

Figure 7 – Agricultural risk layering

Source: Mahul&Stutley (2010)

Such an approach can be more efficient from a fiscal point of view than offering premium
subsidies, as extreme losses can represent a significant weight in the premium. In order to
improve the efficiency of the public-private partnership, Carter, Long and Boucher (Carter,
Long, & Boucher, 2011) propose that private and public actors use one single index to evaluate
losses. They model an insurance contract based on individual rice yields in Ecuador. The first
layer of risk (80% to 100% of the historical average yield) is the “risk-retention” layer. Farming
households are able to manage such losses using a variety of traditional risk-coping
mechanisms. The second level of risk (50% to 80% of the historical average) is the “commercial
risk layer”. This type of event is expected to occur every five years or so. The third layer is the
“catastrophic risk layer”, in which governments bear losses below 50% of the historical average
yield. With this scheme, farmers benefit from a continuity of coverage, between private and
public risk-bearers.

Contingent credit as a complement or substitute to reinsurance

However, many governments in developing countries do not have the capacity to provide
reinsurance without depleting their resources. This barrier can be removed if governments can
access a contingent credit line. One of the most documented example is the Index-Based
Livestock Insurance (IBLI) in Mongolia (DeAngelis, 2013; Lailan, 2011; Mahul & Skees,
2007).

71
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

The example of Livestock Insurance in Mongolia

A Base Insurance Product (BIP) covers herders for livestock mortality rates comprised
between 7 and 25% or 30% (depending on animal species). Losses are calculated through a
livestock mortality index at the local “soum” level. Herders pay a commercial premium rate.
The Government of Mongolia covers losses exceeding the commercial layer through a
Disaster Response Product (DRP). In order to protect private insurers, a Livestock Insurance
Indemnity Pool (LIIP) was created. With the LIIP, herders are protected from the risk of an
insurer defaulting, and the livestock losses risk are separated from the insurers’ balance
sheets. The Government of Mongolia offers a stop-loss reinsurance contract to the LIIP in
exchange of a compulsory reinsurance premium.

In this scheme, private insurers’ liabilities are kept under control. Regarding the Government
of Mongolia’s fiscal exposure, it is two-fold: through the DRP, and through the reinsurance
agreement with the LIIP. The premiums paid by the LIIP go to a reserve fund. The
Government of Mongolia can obtain a contingent loan from the World Bank if the reserve
fund is exhausted and/or if DRP payments are due. This elaborate scheme gives clarity to
both public and private decision-makers on their own liabilities.

Catastrophe risks to habitat, just like agricultural risks, are hard to insure because of potential
extreme losses. In Turkey, earthquake insurance is compulsory for urban property owners and
contracts are pooled in the Turkish Catastrophe Insurance Pool (Linnerooth-Bayer & Mechler,
2007). Homeowners’ fees depend on risk-reduction measures. In case a major catastrophe
occurs, the TCIP benefits from reinsurance in the form of a very favorable contingent loan
provided by the World Bank.

The Mongolian and the Turkish cases illustrate how public-private partnerships can extend
beyond private insurers and national governments. International financial institutions, donors
and NGOs can also play a role to ensure the success of insurance public-private partnerships
(Linnerooth-Bayer & Mechler, 2007).

Potential of securitization of index insurance portfolios

Reinsurance is the most common form of transferring large-scale risks to international markets.
However, reinsurance markets are known to be volatile: after a major loss event, premiums tend
to increase dramatically (Skees et al., 2008). The first catastrophe-linked bonds (CAT bonds)
were issued in the 1990’s as a response to reinsurance markets difficulty to deal with highly-

72
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

correlated risks (Mahul, 2001). CAT bonds transfer the risk from reinsurance markets to
international capital markets. A CAT bond is a marketable security tied to a catastrophic event.
If the event does not occur, investors receive a return. If the event does occur, the investor loses
the interest and some-predefined portion of the capital (up to 100 percent). The CAT bond
issuer uses the capital raised with the CAT bond to pay claims to policyholders.

Index insurance portfolios can be reinsured through CAT bonds, and such securities are
attractive for both issuers and investors, under certain conditions. Collin (Collin, 2018) lists 6
criteria of a fair insurance index, in order to meet the expectations of farmers, insurers and
index-bond investors.

Table 3 - Stakeholders' expectations towards a fair insurance index

Index characteristics Farmers Insurers Investors
Quantifiability X X
Time consistency X X
Availability of historical data X X
Adequacy to real losses X X
Transparency and reliability X X X
Timely availability X X X
Source: Collin (2018)

As index-based techniques allow a precise layering of risk, multiple layers can be defined and
transferred to a variety of investors with different risk appetites and return expectations, from
traditional to socially oriented investors (Skees et al., 2008). Collin (2018) showed that index-
linked securities are attractive for investors as they are uncorrelated with capital markets and
offer therefore a potential for diversification.

3.3.2. Creating an enabling environment

The public sector can intervene in agricultural insurance markets in a less direct fashion, but
still playing an essential role. Efficient agricultural insurance markets require some
infrastructure, as well as an adapted regulatory environment.

Governments’ provision of public goods

Agricultural insurance, especially when it relies on index-based techniques, need data to

function. When it comes to data, agricultural insurance operators face two challenges:
availability and quality (Agricultural Data and Insurance, 2015). Agricultural insurance data
is expensive and non-rivalrous (it can be consumed by many at the same time at no additional

73
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

cost). Therefore, data collection tend to be a natural monopoly. It makes no sense that each
private insurer collects its own data, or sets-up its own weather stations network. Quality of
data is also critical. Index-insurance products must accurately evaluate losses, otherwise
farmers are exposed to basis risk. Poor quality products may even put farmers worse-off, and
encourage risk-taking behavior without the expected coverage. Quality of data is also key to
transfer risk to reinsurers and index-insurance bond investors.

Increasing availability and quality of data has positive spillover effects beyond insurance. Better
data allows better decision-making for policy-makers and actors involved in the agricultural
sector in general. It can also be argued that not only data, but indexes themselves could be
public goods (Sandmark et al., 2013). This would encourage transparency in the market, as
ensure that the market offers accurate products as indexes can be audited.

Beyond data, governments have a role to play by encouraging and funding research and
development. The development and expansion of mobile phone coverage, satellite imagery,
remote-sensing technology, automated-weather stations or big data are disrupting agricultural
insurance. Governments can support research and innovation in this field. Governments can
also contribute to the development of actuarial and sectorial expertise, through the offer of
higher education training, or by setting-up Technical Support Units.

The public sector also has role to play in raising awareness on agricultural insurance. Some
studies have shown that financial education can increase insurance take-up (De Bock & Gelade,
2012; Norton et al., 2014). However, private insurers have limited incentives to invest in
awareness-raising campaigns, as these investments would also benefit their competitors. As
greater financial education and insurance enrollment benefits the whole society, governments
can fund insurance awareness-raising programs.

Legal environment

Insurance value being intangible, trust in the system is essential. Insurance regulators around
the world define minimal capital requirements for insurance companies to make sure that
insurers can face their liabilities. Index insurance raises additional concerns for regulators. As
we have seen previously, basis risk poses a real threat to consumers, who might purchase an
insurance contract that does not fulfil its promise. Farmers, especially in developing and
emerging countries, may have limited knowledge of insurance. Index insurance is a new and
complex product, it is essential that they understand well the contract, its limitations and

74
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

inherent basis risk. This calls for action from the regulator to protect consumers from unfair
practices and false promises.

Following the definition given by Darby and Karni (1973), Clarke and Wren-Lewis (2013)
categorize index insurance as a “credence good”. A credence good is a good whose quality is
not observable prior to purchase, and at best only partially inferred after purchase. Indeed,
index-insurance claims are based on an index, and customers have a poor understanding of
these indexes. Secondly, opportunities to learn from others’ experience are limited, because
index-insurance provides coverage against infrequent and spatially covariant events.
Consequently, private insurers have no incentive to invest in quality indexes, trust in index-
insurance contracts is low and insurance take-up remains limited. The regulator has a role to
play to reinforce confidence in the system. It can license only the products that meet some
minimal quality requirements, or deliver a certification standard, based on commonly accepted
basis risk measurements (Carter et al., 2014).

Another legal challenge that is specific of index-insurance lies in the fact that some countries
do not allow index-insurance, because claim payments are not determined by actual losses. It
is a commonly accepted principle that policyholder must not make a profit from a loss event,
meaning that claim payments must not exceed the actual loss. Insurance regulatory frameworks
need to be adapted to the specific features of index-insurance loss adjustment techniques. For
innovative products using non-traditional techniques, putting in place “sandboxes” where
insurers can test their products at a small scale is a way to foster innovation, without losing
focus on consumer protection.

Finally, insurance regulators and supervisors must keep in mind that compliance requirements
can have a large impact on operational costs and commercial premiums (Acces to Insurance
Initiative, 2016). In environments where clients have limited access to financial products,
regulators need to balance the consumer protection imperative without creating compliance
costs that would deter private actors from serving the low-income customers segment (Brix &
McKee, 2010).

4. Discussion

4.1. Main results

The main results of the review are summarized in the conclusion of this chapter (Figure 8 – The
agricultural insurance policy mix).

75
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

The systematic review of the existing literature shows that the failures inherent to the
agricultural insurance market have for long been understood and analyzed. Whether the
existence of these failures justifies public intervention remains subject to debate among
scholars. Governments can also intervene in agricultural insurance markets to pursue political
and social goals. Agricultural insurance is then seen as an instrument to complement disaster
relief and safety nets, to modernize agriculture or to build a more inclusive economy.

The role of subsidies remains central in agricultural insurance public policies. Whether
subsidies are the most efficient and effective instrument in the policy mix has seldom been
analyzed though, despite increasing public resources being mobilized. Less frequently,
governments have experimented other instruments to support agricultural insurance market
development. In several countries, the public sector is bearing part of the risk itself, usually the
most extreme layers, through the provision of public reinsurance or contingent credit. There is
also a growing interest from financial markets to take on these extreme risks with the
development of catastrophic bonds.

Governments can also support market development by supporting an enabling environment.

The production of public goods is traditionally a government’s prerogative. In the case of
agricultural insurance, such public goods include data and infrastructure, knowledge, financial
education and awareness-raising. In addition, policy-makers need to implement adequate
regulation and supervision, balancing growth and innovation in the market with consumer
protection requirements.

4.2. Contribution of the literature review

Among the resources identified and analyzed in this work, Mahul & Stutley (2010) conducted
the most comprehensive review of existing agricultural insurance programs around the world.
This survey remains the only known example of an attempt to describe government support to
agricultural insurance at a global scale, based on quantitative data (including premium amounts,
number of producers covered and premium subsidies amounts). There are two main limits to
this work though: it relies on self-declared data reported by the insurance authorities of each
country; since it was published in 2010, there are many signs that the agricultural insurance
landscape has known important evolutions, with growth in emerging countries (Kalra & Xing,
2013) and technological innovations. This literature review takes a different and
complementary approach to the work of Mahul and Stutley. Rather than aiming at describing
in an exhaustive and quantified way government supports to agricultural insurance globally, the
76
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

review seeks to identify the means and rationale of public intervention modalities, and to link
these modalities with policy objectives.

A striking characteristic of the documents included in the review is that many articles focus
either on theoretical arguments, or take a macro perspective by providing global figures on
premium amounts and in some cases number of producers or sums covered. Even when case
studies were available, they often provided very limited details on the actual features of the
insurance policies available (risks covered, exclusions, pricing, deductibles, payout, or
distribution channels). Too often agricultural insurance is considered as a “black box”, without
entering into the details of product design. Such qualitative details would bring valuable lessons
to compare insurance contracts, understand differences in pricing, coverage or take-up in
different markets. The author would also have wished to find more, and more recent evaluations
on the viability and economic equilibrium of the public-private partnerships. Since the work of
Hazell (1992), little has been done in this direction. A notable exception is Gohin’s work (2017)
which proposes a welfare analysis of the United States Federal Crop Insurance Program. The
research also failed to identify articles addressing the issue of agricultural insurance public-
private partnerships governance. It remains a key issue though as such partnerships should not
be considered only under a technical view, but also in perspective with policy objectives. Too
often, articles and reports documenting agricultural insurance programs focus on the outreach
of such programs and the market size, with little consideration for the dimensions of inclusivity,
economic viability and level protection brought to farmers.

4.3. Strengths and limits

To the author’s knowledge, this is the first literature review addressing the question of public
interventions in agricultural insurance markets. The mobilization of both academic and grey
literature takes stock of the theoretical arguments within the scholars’ debate, with a grounding
in the practitioners’ perspective and actual state of the market. The triangulation of several
sources indicate that in some areas, a consensus has been reached, in particular on the question
of the insurability of agricultural risks. This work also is original in the extent that it
encompasses together developed, emerging and developing economies.

A limitation of this review is that it includes a lot of grey literature. This situation illustrates the
fact that scientific literature on the question public policies and public-private partnerships in
agricultural insurance is scarce, and underlines the relevance of this doctoral work to address
this knowledge gap. Some of the reports and case studies included in the review were written
77
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

by actors with a direct interest in the agricultural insurance programs they document: donors or
policy-makers. These actors are central in this field and are important knowledge producers,
but more independent research and evaluation would greatly benefit the sector. Finally, this
review is most likely non exhaustive, and other contributions to the debate may have been
ignored. However, at the end of the literature search, the review reached a form of saturation as
no new concept or theoretical argument arose with the integration of additional sources.

4.4. Perspectives

As mentioned earlier, despite the immense contribution of Mahul and Stutley’s work (2010) to
document agricultural insurance programs around the world, a lot has happened in this field
since the publication of their study. The sector, including both scholars and practitioners, would
greatly benefit from an update of this World Bank survey, in order to document the most recent
evolutions. Such a comprehensive work would have to be carried by an international
organization with sufficient legitimacy, such as the World Bank or another international
organization. For the sake of research but also in order to guide policy-makers, the debate on
agricultural insurance policies would benefit from more evaluations of these government
interventions in support of agricultural insurance. Such evaluations should not be centered only
on the issue of insurance take-up and outreach. The analysis of these programs economic
viability, and the efficiency of public spending in regard of policy objectives would inform
policy-makers on the right policy mix of instruments.

5. Conclusion: the agricultural insurance policy mix

After reviewing public policy instruments, it appears that governments can activate different
levers to support agricultural insurance to reach different policy objectives. It is common to
present price as the biggest challenge hindering the development of the agricultural insurance
market, and the response in most cases has been the provision of premium subsidies. However,
this assertion calls for two responses: (i) there are other barriers besides the price that limit the
growth of agricultural insurance and (ii) premium subsidies may not be the only, and not always
the most efficient instrument to lower agricultural insurance premiums.

The literature review identified that agricultural insurance policies can pursue four different and
complementary objectives:

(i) increase outreach to reduce the risk-exposure of farmers (and other value chain actors)

78
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

(ii) support the development of viable agricultural insurance markets

(iii) protect consumers and ensure the quality of products
(iv) promote equitable coverage for the most vulnerable and marginalized groups

For each of these objectives, different policy instruments are available and many times, several
mechanisms must be activated at the same time. Finding the right policy mix assumes a
thorough assessment of the challenges of each specific market, and of the ways to address these
constraints.

Figure 8 – The agricultural insurance policy mix

Risk financing Public goods

• Public insurance & • Data
reinsurance • Research & technical
• Coinsurance pool support
• Contingent credit • Financial education
• Securitization

Subsidies Legal environment

• Premium subsidies • Adapted compliance
• Administrative, requirements
operational and claim • Licensing & quality
payments subsidies
• Universal or targeted
Agricultural standards
• Consumer protection
insurance
policy mix

Reduce risk Support market Promote equitable

Protect consumers
exposure development coverage

Source: author

79
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

References
Access to Insurance Initiative. (2016). How agricultural insurance can improve food security—
And why regulation matters. Policy note.

Antón, J. (2009). Risk management in agriculture—A holistic conceptual framework

(TAD/CA/APM/WP(2008)22/FINAL). Paris. OECD.

Babcock, B. A., & Hart, C. E. (2015). Crop Insurance : A Good Deal for Taxpayers? Iowa Ag
Review, 12(3), 1.

Babcock, B. A., & Hennessy, D. A. (1996). Input demand under yield and revenue insurance.
American journal of agricultural economics, 78(2), 416‑427.

Barnett, B. J., Barrett, C. B., & Skees, J. R. (2008). Poverty traps and index-based risk transfer
products. World Development, 36(10), 1766‑1785.

Bassoco, L. M., Cartas, C., & Norton, R. D. (1986). Sectoral analysis of benefits of subsidized
insurance in Mexico. In Hazell, P., Pomareda, C., Valdes, A. (Eds.), Crop Insurance for
Agricultural Development:Issues and Experience. Johns Hopkins University Press, Baltimore,
pp. 126–142.
Berhane, G., Clarke, D., Dercon, S., Hill, R. V., & Taffesse, A. S. (2012). Financial innovations
for social and climate resilience : Ethiopia case study. Research report.

Bertram-Huemmer, V., & Kraehnert, K. (2015). Does index insurance help households recover
from disaster. German Institute for Economic Research (DIW) Discussion Paper, 1515.

Brix, L., & McKee, K. (2010). Consumer Protection Regulation in Low-Access Environments :
Opportunities to Promote Responsible Finance. Focus Note, 60.

Cai, J., de Janvry, A., & Sadoulet, E. (2011). Social networks and insurance take up : Evidence
from a randomized experiment in China. ILO Microinsurance Innovation Facility Research
Paper, 8.

Carter, M. (2013). Sharing the risk and the uncertainty : Public-private reinsurance partnerships
for viable agricultural insurance markets. I4 Index Insurance Innovation Initiative Brief 2013-
01 and FERDI Policy Brief B78

Carter, M., de Janvry, A., Sadoulet, E., & Sarris, A. (2014). Index-based weather insurance for
developing countries : A review of evidence and a set of propositions for up-scaling.
Development Policies working paper, 111.

Carter, M., Long, E., & Boucher, S. (2011). Public-Private Partnerships for Agricultural Risk
Management through Risk Layering. I4 Index Insurance Innovation Initiative Brief, 2011-01.

80
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

Carter, M. R., Cheng, L., & Sarris, A. (2016). Where and how index insurance can boost the
adoption of improved agricultural technologies. Journal of Development Economics, 118,
59‑71.

Chantarat, S., Mude, A. G., Barrett, C. B., & Carter, M. R. (2013). Designing index‐based
livestock insurance for managing asset risk in northern Kenya. Journal of Risk and Insurance,
80(1), 205‑237.

Clarke, D., & Wren-Lewis, L. (2013). Learning from Lemons : The Role of Government in
Index Insurance for Individuals. Paris School of Economics and FERDI Working Paper.

Cole, S., Giné, X., Tobacman, J., Townsend, R., Topalova, P., & Vickery, J. (2013). Barriers
to household risk management : Evidence from India. American economic journal. Applied
economics, 5(1), 104.

Cole, S., Giné, X., & Vickery, J. (2017). How does risk management influence production
decisions ? Evidence from a field experiment. The Review of Financial Studies, 30(6),
1935‑1970.

Collin, C. (2018). Towards a working crop insurance market : An integrated strategy of

systemic risk management. PhD thesis. Paris Nanterre University.

Cummins, J. D., & Mahul, O. (2009). Catastrophe Risk Financing Catastrophe Risk Financing
in Developing Countries (pp. 1–299). Washington, DC.

Darby, M. R., & Karni, E. (1973). Free competition and the optimal amount of fraud. The
Journal of law and economics, 16(1), 67‑88.

De Bock, O., & Gelade, W. (2012). The demand for microinsurance : A literature review. ILO
Microinsurance Innovation Facility Research Paper, 26.

DeAngelis, K. (2013). Index-based livestock insurance : The case of Mongolia. Inside Stories
on climate compatible development, Londong. CDKN.
Dercon, S., Gunning, J. W., & Zeitlin, A. (2011). The demand for insurance under limited
credibility : Evidence from Kenya. International Development Conference, DIAL.

Dercon, S., Hoddinott, J., & Woldehanna, T. (2005). Shocks and consumption in 15 Ethiopian
villages, 1999-2004. Journal of African economies, 14(4), 559.

Eling, M., Pradhan, S., & Schmit, J. T. (2014). The determinants of microinsurance demand.
The Geneva Papers on Risk and Insurance-Issues and Practice, 39(2), 224‑263.

Glauber, J. (2007). Double indemnity : Crop insurance and the failure of US Agricultural
Disaster Policy. AEI Agricultural Policy Series: The 2007 Farm Bill and Beyond. American
Enterprise Institute, Washington, DC. [Link]

81
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

Gohin, A. (2017). Macroeconomic impacts of the US farm policy : A second best assessment.
20th Global Economic Analysis Conference, Purdue, Indiana.

Goodwin, B. K., Vandeveer, M. L., & Deal, J. L. (2004). An empirical analysis of acreage
effects of participation in the federal crop insurance program. American Journal of Agricultural
Economics, 86(4), 1058‑1077.

Hazell, P. B. (1992). The appropriate role of agricultural insurance in developing countries.

Journal of International Development, 4(6), 567‑581.

Hazell, P., Anderson, J., Balzer, N., Hastrup Clemmensen, A., Hess, U., & Rispoli, F. (2010).
The potential for scale and sustainability in weather index insurance for agriculture and rural
livelihoods. Rme. World Food Programme (WFP).

Hazell, P., Sberro-Kessler, R., & Varangis, P. (2017). When and how should agricultural
insurance be subsidized ? Issues and good practices. Washington DC. World Bank.

Hess, U., Hazell, P., & Kuhn, S. (2016). Innovations and emerging trends in agricultural
insurance. Eschborn. Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH.

Hill, R. Vargas, Gajate-Garrido, G., Phily, C., & Dalal, A. (2014). Using subsidies for inclusive
insurance : Lessons from agriculture and health. Microinsurance Paper No. 29. International
Labour Organization.

Hill, Ruth Vargas, Robles, M., & Ceballos, F. (2013). Demand for weather hedges in India :
An empirical exploration of theoretical predictions (Vol. 1280). Intl Food Policy Res Inst.

Jaffee, D. M., & Russell, T. (2006). Should Governments Provide Catastrophe Insurance? The
Economists’ Voice, 3(5).

Kalra, A., & Xing, L. (2013). Partnering for food security in emerging markets (No 1/2013;
Sigma). Swiss Reinsurance Company.
Karlan, D., Osei, R., Osei-Akoto, I., & Udry, C. (2014). Agricultural decisions after relaxing
credit and risk constraints. The Quarterly Journal of Economics, 129(2), 597‑652.

Karlan, D. S., Osei, R., Osei-Akoto, I., & Udry, C. (2012). Agricultural decisions after relaxing
credit and risk constraints (Vol. 23). National Bureau of Economic Research.

Lailan, T. (2011). Exploring Options to Institutionalize the Dzud Disaster Response Product in
Mongolia. World Bank, Washington, DC.

Linnerooth-Bayer, J., & Mechler, R. (2007). Disaster safety nets for developing countries :
Extending public–private partnerships. Environmental Hazards, 7(1), 54‑61.
[Link]

Mahul, O. (1998). Vers une redéfinition du rôle de l’assurance agricole dans la gestion des
risques sur récoltes. Cahiers dEconomie et Sociologie Rurales, 49, 34‑58.
82
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

Mahul, O. (2001). Managing catastrophic risk through insurance and securitization. American
Journal of Agricultural Economics, 83(3), 656‑661.

Mahul, O., & Skees, J. R. (2007). Managing agricultural risk at the country level : The case of
index-based livestock insurance in Mongolia. World Bank Policy Research Working Paper,
4325.

Mahul, O., & Stutley, C. J. (2010). Government Support to Agricultural Insurance : Challenges
and Options for developping countries -. World Bank.
[Link]
agricultural-insurance-challenges-and-options-for-developing-countries

Mahul, O., Verma, N., & Clarke, D. J. (2012). Improving farmers’ access to agricultural
insurance in India. Washington. The World Bank.

Meyer, R., Hazell, P., & Varangis, P. (2017). Unlocking Smallholder Credit. World Bank
Group. [Link]
smallholder-credit-does-credit-linked-agricultural-insurance-work

Miranda, M. J. (1991). Area-yield crop insurance reconsidered. American Journal of

Agricultural Economics, 73(2), 233‑242.

Miranda, M. J., & Glauber, J. W. (1997). Systemic risk, reinsurance, and the failure of crop
insurance markets. American Journal of Agricultural Economics, 79(1), 206‑215.

Mobarak, A. M., & Rosenzweig, M. R. (2012). Selling formal insurance to the informally
insured.

Mobarak, A. M., & Rosenzweig, M. R. (2013). Informal risk sharing, index insurance, and risk
taking in developing countries. The American Economic Review, 103(3), 375‑380.

Morduch, J. (2005). Smart Subsidy for Sustainable Microfinance. ADB FINANCE FOR THE
POOR, A Quarterly Newspaper of the Focal Point for Microfinance, 6(4).
[Link]
&tdet3=2&referer=MTA1MTY%3D

Norton, M., Osgood, D., Madajewicz, M., Holthaus, E., Peterson, N., Diro, R., Mullally, C.,
Teh, T.-L., & Gebremichael, M. (2014). Evidence of demand for index insurance :
Experimental games and commercial transactions in Ethiopia. Journal of Development Studies,
50(5), 630‑648.

Priest, G. L. (1996). The government, the market, and the problem of catastrophic loss. Journal
of risk and Uncertainty, 12(2‑3), 219‑237.

Quayyum, S. N., Clarke, D., Lo Re, F., Sberro, R., & Stutley, C. (2018). Bangladesh :
Agricultural Insurance Solutions Appraisal Technical Report. World Bank Group.

83
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

[Link]
Insurance-Solutions-Appraisal-Technical-Report

Qureshi, Z., & Reinhard, D. (2016). Report 11th International Microinsurance Conference
2015. Munich Re Foundation.

Ramaswami, B. (1993). Supply response to agricultural insurance : Risk reduction and moral
hazard effects. American Journal of Agricultural Economics, 75(4), 914‑925.

Regős, G. (2015). Should governments subsidize agricultural insurances. Applied Economics

Quarterly, 61(1), 65‑92.

Ricome, A., Affholder, F., Gérard, F., Muller, B., Poeydebat, C., Quirion, P., & Sall, M. (2017).
Are subsidies to weather-index insurance the best use of public funds ? A bio-economic farm
model applied to the Senegalese groundnut basin. Agricultural Systems, 156, 149‑176.

Rosenzweig, M. R., & Binswanger, H. P. (1993). Wealth, weather risk and the composition and
profitability of agricultural investments. The economic journal, 56‑78.

Sandmark, T., Debar, J.-C., & Tatin-Jaleran, C. (2013). The emergence and development of
agriculture microinsurance. Luxembourg: Microinsurance Network.

Skees, J. R. (1999). Agricultural risk management or income enhancement. Regulation, 22, 35.

Skees, J. R., & Barnett, B. J. (1999). Conceptual and practical considerations for sharing
catastrophic/systemic risks. Review of Agricultural Economics, 21(2), 424‑441.

Skees, J. R., & Barnett, B. J. (2006). Enhancing microfinance using index-based risk-transfer
products. Agricultural Finance Review, 66(2), 235–250.

Skees, J. R., Barnett, B. J., & Murphy, A. G. (2008). Creating insurance markets for natural
disaster risk in lower income countries : The potential role for securitization. Agricultural
Finance Review, 68(1), 151‑167.

Smith, V. (2017). The US Federal Crop Insurance Program : A Case Study in Rent-Seeking.

Smith, V. H., & Glauber, J. W. (2012). Agricultural insurance in developed countries : Where
have we been and where are we going? Applied Economic Perspectives and Policy, pps029.

Smith, V. H., & Goodwin, B. K. (1996). Crop insurance, moral hazard, and agricultural
chemical use. American Journal of Agricultural Economics, 78(2), 428‑438.

Smith, V. H., & Goodwin, B. K. (2003). An ex post evaluation of the conservation reserve,
federal crop insurance, and other government programs : Program participation and soil
erosion. Journal of Agricultural and Resource Economics, 28(1835-2016‑148818), 201‑216.

Sundt, B. (1999). An introduction to non-life insurance mathematics 4th Edition. Karlsruhe:

University of Mannheim Press.

84
 Chapter 2 - From subsidies to public-private partnerships: public interventions in the agricultural insurance
market

Tsujii, H. (1986). An economic analysis of rice insurance in Japan. In Hazell, P. B. R.,

Pomareda, C. and Valdes, A. (eds), Crop Insurance for Agricultural Development: Issues and
Experience. Baltimore: Johns Hopkins University Press, pp. 143- 155.

85
86
 CHAPTER 3

FROM CRISIS MANAGEMENT TO RISK MANAGEMENT IN

AGRICULTURE: THE CASE OF GRASSLAND INSURANCE IN
FRANCE

87
From crisis management to risk management in agriculture: the case of
grassland insurance in France

1. Introduction

Adverse weather events expose farmers to potential economic losses. Governments willing to
help managing agricultural risk can mobilize several instruments, which are usually classified
into two categories (Antón, 2009) : ex-ante measures are implemented before the event takes
place, whereas ex-post response occurs after the event.

In France, agricultural risks have been covered since 1964 by a government-sponsored disaster
fund: the Fonds national de gestion des risques en agriculture (FNGRA). Reformed in 2010
(Loi de modernisation de l’agriculture et de la pêche, 2010), the FNGRA pursues 3 main goals:
to compensate economic losses related to vegetal or animal diseases, to financially support the
development of agricultural insurance, and to indemnify agricultural calamities. It is financed
through a tax on insurance premiums paid by the farmers, and an equal contribution taken from
the government’s budget. In case of an extreme event, the government may decide to allocate
an extra funding for the fund. The main shortcoming of this disaster fund, like any ad-hoc
disaster response, is that the State lacks visibility on its financial intervention, which depends
on the collected taxes and on the total damage. This is why the French Government asked
private insurers to create new policies and decided to subsidize premiums, while focusing the
FNGRA’s compensations on non-insurable crops (Loi de modernisation de l’agriculture et de
la pêche, 2010; Ménard, 2004). It has been the case with cereals in 2005 with the birth of the
first multi-peril crop insurance in France, and in 2015 with the launch of grassland insurance.

In Europe and in France, the preferred agricultural risk management framework is moving from
government-run programmes, providing ex-post assistance in case of an adverse, to insurance
public-private partnerships, offering ex-ante solutions (EU, 2013). Is this change interesting for
the State? Increasing amounts are invested through Public-private Partnerships (Wettenhall,
2003), and tools to assess their relevance appeared with the new public management wave.
Most often these tools reflect investment and delays considerations through the use of indicators
such as net present value and internal rate of return (Bonnafous & D’Arcier, 2013). They fit the
study of financing, building and operating infrastructures, but are not suited to the insurance
sector. Compared to the well documented sectors of healthcare (Torchia et al., 2015), transport,
telecommunications or energy, few articles developed criteria and indicators to compare and

88
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

assess different compensation systems. For a first global approach between two existing
systems, Hazell (1992) uses the cost (claims) to benefit (premium) ratio as indicator of an
insurance system soundness. More interestingly in Finland, where the public system is also to
be replaced by private insurance, a comparison between actual public intervention and
simulated private insurance subsidy is conducted based on the public system data (Liesivaara
et al., 2014).

Building on those methods, a comparison process adjusted to the French case is developed. For
the period 2003-2015, a model based on biomass production data evaluates the economic loss.
This model is used to set the premium for an insurance product guaranteeing a certain amount
of grassland production. We present the impact on the government budget for each risk
management policy: subsidizing insurance premiums or maintaining a disaster fund. We also
assess each policy choice according to several public management criteria. This research
provides an original contribution for governments and policy-makers when choosing which
agricultural risk management instruments to support.

The chapter is structured as follows. First, it gives the policy context at the global, European
and French levels. Section 3 describes the available datasets. Section 4 presents the model and
the method applied, and section 5 its results. The final part discusses the interest for the State
and the farmers to switch from an ex-post to an ex-ante system.

2. Policy framework

As we examine the risk management policy in France and the alternative between a
compensation fund and support to insurance, it is useful to have a look also at the international
and European context as it largely shapes the public support allowed and implemented in
France.

2.1. The World Trade Organization Agreement on Agriculture

The Agreement on Agriculture (AoA) of the World Trade Organization (WTO), signed in
Marrakech in 1994, classifies public aids into three categories: the green box, the blue box and
the amber box. The blue box contains aids for goods with a production limitation. The green
box contains measures that have “no, or at most minimal, trade-distorting effects or effects on
production”, and are therefore exempt from reduction commitments (Annex 2 of the AoA). The
amber box gathers all other types of public aids. Aids falling into the amber box are subject to

89
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

reduction commitments, unless they comply with the rule of “de minimis”: those are aids not
exceeding 5% of agricultural production (10% in developing countries) and are not specific to
a production.

The possibility to classify agriculture insurance public aids in the green box is often cited as
one of the main reasons why many governments decided to subsidize agricultural insurance
premiums in lieu of other support measure (J. W. Glauber, 2015). However, a more detailed
look at the AoA requirements shows that most agricultural insurance subsidies do not qualify
for the Green Box.

The AoA differentiates income insurance and income safety net programs (paragraph 7) on one
hand, crop insurance and disaster assistance (paragraph 8) on the other hand. Income insurance
and income safety net programs are classified in the Green Box if the “income loss exceeds 30
per cent of average gross income or the equivalent in net income terms (…) in the preceding
three-year period or a three-year average based on the preceding five-year period, excluding
the highest and the lowest entry. The amount of such payments shall compensate for less than
70 per cent of the producer's income loss in the year”.
For disaster relief and crop insurance, there needs to be a “formal recognition by government
authorities that a natural or like disaster ( ) with a production loss which exceeds 30 per cent
of the average of production in the preceding three-year period or a three-year average based
on the preceding five-year period, excluding the highest and the lowest entry.”

The necessity to have a formal recognition by government authorities is not compatible with
insurance practice. Moreover, the three- or five-years’ reference for historical yields contradicts
actuarial principles, as normally longer time series are preferred. It may actually lead to adverse
selection if farmers buy insurance if the three-year average was exceptionally high and they
predict a lower yield for the current year, and vice-versa if the historical reference is low, they
might be tempted not to buy insurance. In the end, most governments declare agricultural
insurance public aids in the Amber Box. However, these supports fall under the “de minimis”
rule, and therefore are exempt from reduction commitments.

2.2. The European Union and the Common Agricultural Policy

When it was introduced in 1962, the Common Agricultural Policy (CAP) main objective was
to maintain high and stable prices for agricultural production within the common market. From

90
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

1992, the CAP went through successive reforms in order to increase European Agriculture
market orientation. Price support instruments were gradually removed in favor of direct income
support for farmers. This evolution is reflected in the CAP expenditure: in 1992, market
intervention mechanisms represented over 90% of the CAP budget, against 5% at the end of
1993. Direct payments, 94% of which are decoupled from production, are currently the main
instrument of support.

The current CAP is in place for the period 2014-2020. It supports farmers through “Market
support measures and direct subsidies” (Pillar 1) and “Rural Development Programmes” (Pillar
2). Risk management instruments were transferred from Pillar 1 to Pillar 2, and are therefore
optional measures, which can be co-financed by Member States. They consist of three tools:

- Financial support for insurance premiums

- Mutual insurance funds, owned by the participants, to compensate farmers for
production losses
- An income stabilization tool, to compensate severe income losses (30% over three
years).

The European Commission regulation and guidelines define a framework for Member States
intervention:

- Adverse climatic events can be assimilated to a natural disaster if production loss is

above 30% (20% for less favoured areas) (articles 87 and 88; No 70/2001; EC regulation
No 1857/2006; guidelines 2007-2013; 2006/C 319/01). Compensation cannot exceed
80% (90% for less favoured areas) of the income loss. The public authorities must
formally acknowledge the weather event. Since 2010, the indemnity payment in case of
a climatic event has been reduced by half if the farmer does not insure at least 50% of
the production against the most frequent climatic risks.
- Financial support to insurance cannot exceed 65% of the premium (Article 70, No
73/2009; Article 37, No 1305/2013).

It appears that the definitions and thresholds adopted by the European Commission have been
defined to comply with WTO rules (Bielza et al., 2007). It also seems that insurance is the
preferred mechanism for risk management as certain aids are conditioned by the existence of
insurance. In a 2005 communication, the European Commission states that “insurance provides
an alternative to public ex post compensation payments for losses caused by natural disasters
at EU and national or regional level” (EC COM, 2005).
91
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

2.3. The French framework

For many years, the risk management policy in France was focused on a disaster fund. The
Fonds National de Garantie des Calamités Agricoles was created in 1964, replaced by the Fonds
National de Gestion des Risques Agricoles in 2010 (Loi n° 2010-874 du 27 juillet 2010 de
modernisation de l'agriculture et de la pêche). The FNGRA expenditures are classified into
three sections (Direction de l’initiative parlementaire et des délégations, 2016):
- Financing indemnities in case of animal or vegetal disease, or an environmental incident
- Financing premium subsidies
- Indemnifying agricultural calamities

Agricultural calamities are production losses or damages on productive assets resulting from
non-insurable risks. They are caused by a climatic event of exceptional magnitude (frost,
drought, flood…) and are officially recognized by a ministerial decree.

Following several parliamentary reports (Babusiaux, 2000; Ménard, 2004), the French
government decided to subsidize crop insurance premiums, and the first multi-peril crop
insurance policies were launched in the market in 2005. Insurance policies are available for
most crops and against all major risks. After a year of commercialization, 15% of the
agricultural production and 25% of surfaces (excluding grasslands) are covered by insurance
(Boyer, 2008). Commercial success was concentrated on cereals, concentrating three quarters
of premium subsidies. Cereals traditionally received a small part of the indemnities from the
disaster fund, as they are less exposed to risk. Another parliamentary report also notes that the
existence of a public disaster fund limits the development of crop insurance (Mortemousque,
2007).

As they became insurable, cereals were excluded from the FNGCA/FNGRA in 2009 and wine
in 2011. However, the government maintained other emergency aids, quite often in an
uncoordinated and inefficient way (Cour des Comptes, 2012, 2013) . It is still unclear whether
the transition from a public ex post compensation to an insurance premium subsidies regime is
beneficial for the government budget (Cour des Comptes, 2013).

In 2015, ten years after it was introduced in the French market, multi-peril crop insurance is not
widely spread among farmers. Multi-peril crop insurance covers 26.8% of planted area for
cereals, 23.2% for vine-growing, 15.4% for vegetables and 2.2% for fruits (Lidsky et al., 2017).
Lidsky et al. propose two main reasons for this low penetration: the level of premium subsidies

92
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

was not constant and perceived as unpredictable, and farmers prefer hail insurance which is
cheaper with a lower deductible.

3. Data

Two data sets are used; one is used to measure the cost of the French government’s spending
related to grassland production shocks over the period 2003-2015, and the other one to create
an insurance product covering grassland production.

3.1. Disaster fund data

Data from the disaster fund was obtained directly from the French Ministry of Agriculture. For
the period 2003 to 2015, the ministry provided a list of all indemnification decisions related to
forage losses, per year and per department (France is divided in 101 departments). The amounts
listed do not necessarily correspond to the credits effectively disbursed. Indeed, some affected
farmers do not claim the indemnities they are entitled to receive, or in some rare cases, credits
can be distributed in excess of the voted budget. As an illustration, we have the data for the
period 2009-2012, and over this period, only 77% of the voted budget was actually disbursed
to farmers.

Table 4 - Disbursed indemnities over the period by the FNGRA

Year Disaster fund

2003 390,4
2004 16,2
2005 227,4
2006 138,7
2007 6,3
2008 0
2009 67,4
2010 118,1
2011 254,2
2012 1,5
2013 0
2014 0
2015 166,5
Total 1 386,7
Source: French Ministry of Agriculture

Adding the claims of orders from one year, we obtain the total amount awarded that given year
(see Table 1). However, when important shocks occur, the total fund endowment does not cover
the total loss. In those cases, additional claims are awarded the following years. On Table 1, the

93
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

claims are summed by shock’s year rather than settlement year. In 2003 for example, year of
severe drought, 332 million euros were awarded, but an extra 58 million of euros was paid
during the subsequent years for that precise drought. In the following, we will say that 390
million euros were awarded in 2003. On the contrary, 87 million euros were awarded in 2004,
but only 16 million euros regarding damages that effectively happened in 2004. The other part
constitutes complementation for past years. In the following, we will say that 16 million euros
were awarded in 2004.

3.2. Index data

Grass may be eaten by livestock, harvested twice to trice a year, and there are no accounting
tracks of its financial value or exact yields. For all those reasons, we consider a policy based on
satellite data rather than traditional direct yield assessment. The index used is the Forage
Production Index (FPI), derived from the fCover technology. The fCover measures the
proportion of ground covered by active vegetation when the scene is observed vertically
(Roumiguié et al., 2015). This measurement is based on medium resolution satellite imagery.
For each year n, the FPI is the sum of the daily fCover values noted , with being the
date of the day between February 1 and October 31. From this value, a part of non-productive
vegetation (NPf) is subtracted, this last parameter being adjusted according to farming practices
in each forage region of France.

= ( − × )
!"

The FPI allows to measure local grassland production without taking into account individual
farmers’ practices. The index is available since 2003 at the zip-code scale. Each zip code has
one FPI by year, representing the level of biomass produced during the year (from February to
October). The higher the value, the higher the production.

Table 5 - Descriptive table of FPI values

Standard
FPI Minimum Quartile 1 Average Quartile 3 Maximum Deviation
2003 0.3 22.7 34.0 43.7 87.9 15.5
2004 12.3 37.7 50.6 63.0 104.6 15.4
2005 0.9 27.9 41.4 56.3 99.0 17.3
2006 5.3 31.4 41.5 51.7 96.1 12.2
2007 13.3 47.0 58.9 71.5 103.3 15.0

94
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

2008 8.9 42.0 51.3 59.8 97.1 11.8

2009 6.0 31.5 40.8 50.1 97.7 12.1
2010 1.8 34.0 42.6 50.4 102.3 13.0
2011 0.8 36.9 45.0 55.0 102.4 13.3
2012 2.5 33.7 46.8 61.7 97.5 17.1
2013 9.0 43.8 53.1 62.0 103.4 12.8
2014 0.0 49.0 58.7 69.2 118.9 14.4
2015 7.7 35.4 44.9 54.5 97.2 14.0

Figure 9 - Yearly distribution of FPI values

120

100

80 Q1
FPI values

min
60
median
max
40
Q3
20

-
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Table 5 shows some descriptive data about FPI. Figure 1 is a graphical presentation of the Table
2 values. The spread between minimal and maximal FPI values is explained by the difference
of production from one region to another due to local specificities (e.g. weather or soil
properties).

Table 2 and Figure 1 illustrate that 2003 is the year of lowest production for the majority of zip
codes, with an average FPI value of 34 for the year. Conversely, 2007 is the year of the best
production, with a 58.9 average FPI value, which is consistent with the FNGRA data.

4. Method

The aim is to compare the cost for the State of a public system of disaster relief for grassland
production losses versus a private scheme of index insurance. The State is involved in claim
settlement in the first case, and in premium subsidy in the second.

To compare the two systems, public and private, the State historical expenses due to FNGRA
claim settlement are considered. On the other hand, insurance premium and premium subsidy

95
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

are modelled based on the index data introduced in the previous section. First, both systems are
compared for the same level of coverage, corresponding to the coverage provided by the
FNGRA. Then the private scheme is assessed for various levels of coverage, getting along with
the market standards of agricultural insurance in France.

4.1 Private insurance system

The cost of the private scheme for the State is the subsidy amount required to support the
insurance product described below. The index being available since 2003, it is possible to
simulate the compensations that would have been awarded from 2003 to 2015.

The pure premium of the insurance policy is estimated with the expected value (Young, 2004)
of those historical claims. A loading factor accounting for administration costs is added to the
pure premium to obtain the commercial premium. The subsidy amount is easily deduced from
this result.

The following scheme based on the FPI is considered: claims are triggered whenever the FPI
falls below a certain guaranteed level, noted #$% . This level is made of two components:

- An insured level & , which is the Olympic average FPI (average value of the FPI
over the last five years, excluding the lowest and highest years)1
A loss threshold expressed in percentage and chosen by the insured, which grants
eligibility for compensation (in the following, the quantity ' (( = 1 −
-

) ** 'ℎ *ℎ ), is considered)

The guaranteed level #$% = & × ' (( .

The claim of year n is computed with:

- The Insured FPI, &

The FPI of year n,
The deductible or more conveniently α = 1 − , ,./' 0) . The deductible is the
-
-
threshold of loss before the insurance will make any payment.
- The level of compensation, 1 , is the loss beyond the deductible covered by the
insurance policy.
- The insured capital 2 = /34 '3) 5*. , 06 ℎ /'3 ∗ 5*. , 3 3

The deductible, trigger and level of compensation are expressed in proportions, with values
between 0 and 1.

1
In practice, the insured FPI is the FPI’s Olympic average. However, the FPI data is available since 2003, so
Olympic averages can be calculated only from 2008. For the 2003-2007 period, the insured FPI is the average FPI
for the period 2003-2015.

96
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

When there is no deductible, α = 1

The claim amount awarded for year n is:

1
= (α × − )× ×1 ×2 if ' (( × − >0
8 &
&
&

=0 if ' (( × & − ≤0

When the FPI of the year is above or equal to the guaranteed FPI (' (( × & − ≤
0), the insurance will make no payment ( = 0 ). For losses beyond the trigger (' (( ×
& − > 0), the insurance will make a payment to compensate losses beyond the
deductible, and according to the proportion of the level of compensation.

For example, if ) ** 'ℎ *ℎ ), = 0.3, ' (( = 0.7. If = 0.8 × & , then ' (( ×
& − ≤ 0 and there will be no claim payment.

Now if = 0.6 × &, which corresponds to a 40% loss, the insurance will make a
payment. If we consider a level of compensation LoC = 1, and a deductible of 0.3 (α = 0.7),
the claim payment will be equal to:

1
= (α × & − )× ×1 ×2
&

= (0.7 × & − 0.6 × &) × BC

×1×2
&

= 0.12

In addition, the commercial premium is:

DD / 3) 4 D .D = E( ) × 1 3, 5( 3/'

4.2 Public disaster relief system

No modeling is done to assess the claim amount, directly taken from the information given by
the ministry of agriculture, as described in the “data” part.

The FNGRA indemnifies farmers when losses exceed 30% of the guaranteed production. The
FNGRA covers 28% of the losses and no deductible is applied. Using the variables introduced
in the 4.2 section above, for the FNGRA we have the following values:

- loss threshold = 30% so trigger = 0.7

97
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

, ,./' 0) = 0, so α = 1
1 = 28
-
-

The FNGRA does not insure a value per hectare, but rather guarantees a minimum level of
forage production (3000 forage units per dairy cow unit2, which corresponds to the average
ration of forage consumed by a cow over the year).

5. Results

5.1 Comparison between insurance products and disaster fund, at the same level of
guarantees

We first compare a product, “product 1”, offering guarantees similar to the FNGRA scheme:

loss threshold = 30% so trigger = 0.7

deductible = 0, so α = 1
-
-
- LoC = 0.28

The amount of capital insured K is a major determinant of the compensation received by the farmer in
case of a shock. The FNGRA as such does not establish a level of capital insured by hectare, but a
minimum forage production.

On the other hand, a pilot project has been implemented in France from 2013 to 2015 by a
private insurer, covering grasslands against the risk of drought. For the approximately 200
farmers covered during the first three years of the project, the average capital insured per hectare
was of roughly 800€. In this study, several levels of insured capital will be considered. 800€/ha
as “close-to-practice” level, 600€/ha as low limit level and 1000€ as high-limit level.

Table 6 - Disbursed indemnities over the period for product 1 and the disaster fund. K = 800

Year Product 1 Disaster fund

2003 572,1 390,4
2004 3,3 16,2
2005 306,1 227,4
2006 58,7 138,7
2007 0,6 6,3
2008 1,2 -
2009 72,0 67,4
2010 151,3 118,1
2011 47,3 254,2
2012 94,8 1,5

2
[Link]

98
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

2013 2,3 -
2014 2,8 -
2015 86,1 166,5
Total 1398,5 1 386,7

Figure 10 - Comparison of disbursed indemnities over the period between product 1 and the
disaster fund. K = 800

700 000 000

600 000 000

500 000 000

400 000 000

300 000 000

200 000 000

100 000 000

0
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Product 1 Disaster fund

Table 6 presents the amounts disbursed over the period by the insurance product 1 and the
disaster fund. Figure 6 is a graphic presentation of Table 6 values. For both coverage
instruments, 2003 is the worst year for grassland production, with disbursements of 572 million
euros for product 1 and 390.4 million euros for the disaster fund.

Considering a capital insured of 800€ per hectare, between 2003 and 2015, the index-based
insurance product distributes 0.9% more in indemnities than the disaster fund. The index
identifies major shocks, 2003 and 2005 being the worst years in terms of forage production.
The insurance product indemnifies much more the farmers in 2010 (+28.1%) and less in 2011
(-81.4%).

With K =600, then farmers receive 24% less over the period compared with the disaster fund
(see table 4, in section 5.2). However, even in this case the index-based insurance still
indemnifies farmers more for the years 2003 and 2012.

Lastly, if we raise the capital insured per hectare to 1000€, the total amounts disbursed to farmed
over the period reaches approximately 1.75 billion €, a 26% increase compared with the public
disaster fund.

99
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

If both the FNGRA and the index insurance product seem to identify the major crisis, important
differences remain if we examine each year separately. For example, the FNGRA indicates
bigger shocks in 2006 and 2011, whereas the index measures medium losses in 2012 that the
FNGRA ignores. This discrepancy can be explained by two main factors. First, their levels of
reference are different: the index product is based on an average production, per zip code, over
the period 2003-2015, while the FNGRA establishes an absolute threshold (3000 forage
production units). Second, to claim an FNGRA indemnification, farmers must demonstrate a
minimum 13% loss of the gross product of the whole farm. This rule clearly disadvantages
diversified farms, in favour of specialized farms. Therefore, a climatic event affecting a region
where mixed farming is dominant will result in lower disbursement from the disaster fund than
a shock affecting a region characterized by specialized agriculture.

Besides these two factors, the 2011 year illustrates how the switch to index insurance can bring
significant changes in the risk coverage of farmers. The 2011 springtime was characterized by
a severe drought, which prevented farmers to make the first cut of grass. The disaster
recognition intervened in July, earlier than ever before for forage losses3. The 2011 political
context should be kept in mind as two major elections (presidential and legislatives) were to
take place in May and June 2012. After the spring drought, several rains and a good growth of
grass marked the 2011 summer. This is why the index does not measure a catastrophic year, as
the grass growth of the summer and autumn compensate the poorly performing spring. Despite
this situation, farmers received a compensation based on their situation at the beginning of July.

5.2 Modification of insurance product parameters

The features of the insurance product 1 match the level of guarantee of the national disaster
fund but are quite different from the standard in crop insurance market. Indeed, product 1 has
no deductible, but compensates losses at a very low level (28%). Therefore, we create another
insurance product named “product 2”, for which we modify these parameters to put them in
line with the standard practices:

threshold = 30% so trigger = 0.7

deductible = 30% so α = 0.7
-
-
- LoC = 1

3
[Link]

100
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

Table 7 - Disbursed indemnities over the period for product 1, product 2, and the disaster
fund. K = 800

Year Product 1 Product 2 Disaster fund

2003 572,1 794,3 390,4
2004 3,3 1,7 16,2
2005 306,1 326,7 227,4
2006 58,7 30,2 138,7
2007 0,6 0,2 6,3
2008 1,2 0,7 -
2009 72,0 48,8 67,4
2010 151,3 143,0 118,1
2011 47,3 42,1 254,2
2012 94,8 90,9 1,5
2013 2,3 1,4 -
2014 2,8 3,7 -
2015 86,1 66,0 166,5
Total 1398,5 1549,7 1 386,7

Figure 11 - Comparison of disbursed indemnities over the period between product 1, product
2 and the disaster fund. K = 800€

900 000 000

800 000 000
700 000 000
600 000 000
500 000 000
400 000 000
300 000 000
200 000 000
100 000 000
0
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Product 1 Product 2 Disaster fund

Table 7 presents the amounts disbursed over the period by the insurance product 1, insurance
product 2 and the disaster fund. Figure 7 is a graphic presentation of Table 7 values. For all
coverage instruments, 2003 is the worst year for grassland production, with disbursements of
572 million euros for product 1, 794 million euros for product 2 and 390 million euros for the
disaster fund.

For K= 800, the global amount of indemnifications reaches 1.550 billion euros over the period,
which is 12% above the sums distributed by the FNGRA, and 10.8% more than the index-
insurance product previously examined (product 1). We can observe that this product

101
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

indemnifies more on the years of big shocks (2003,2005,2010) and less the years of low or
medium intensity shocks (2004, 2006, 2009, 2011, 2015).

This is result is confirmed when calculating the volatility of disbursements for each type of
coverage. The standard deviation calculated in million euros is equal to 125.7 for the disaster
fund, to 163.5 for the product 1 and to 211.9 for the product 2. It seems that the FNGRA
distributes more regularly small amounts but is unable to cope with major disasters, when the
insurance products allocate most of their disbursements for major events.

The correlation coefficient between product 1’s disbursements and disaster fund’s
disbursements is Corr (Product 1, Disaster Fund) = 0.82

The correlation coefficient between product 2’s disbursements and disaster fund’s
disbursements is Corr (Product 2, Disaster Fund) = 0.80

These values indicate a significant positive correlation between the disbursements of the two
insurance products and the disbursements of the FNGRA.

The total amount of disbursements over the period varies according to the type of coverage
(Disaster Fund or insurance) and the capital insured (600, 800 or 1000 euros per hectare).

Table 8 - Total cumulated indemnities in million euros received by the farmers depending on
the sum insured

Disaster
Fund Product 1 Product 2
K= 600 1 049 1 162
K= 800 1 387 1 399 1 550
K = 1000 1 748 1 937

Depending on the sum insured and the parameters of the insurance product, the indemnities
received by the farmers vary from 1049 to 1937 million euros between 2003 and 2015, when
the disaster fund disburses 1387 over the same period.

5.3 Cost for the State and for the clients

We have seen that it is possible to adjust the parameters of the index-insurance product so it
can offer the same guarantees to the farmers as with the disaster fund (product 1). However, in
order to assess the cost of the private scheme for the state and for the farmer, it is important to
take into account the administration costs, through a premium loading.

102
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

In the first years of running an index-insurance, insurers tend to be quite conservative in their
pricing. For this analysis, we have considered a target ratio of claim/premium at 0.75. We also
compute the risk of model error, given the relative novelty of the product and the fact we only
have 14 years of data. This risk of model error is set at 1.1.

Therefore, we can calculate a loading factor of L = 1/0.75 × 1.1 = 1.47

Commercial Premium is then equal to Pure Premium× 1.

If we insure with product 1 the 10.5 million hectares of France, the yearly premium to insure
the whole country is 107.6 × 1.33 = 157.8 million €. For product 2, this yearly premium
amounts to 119.2 × 1.47 = 174.8 million €.

Now, if the insurer adopts a more conservative approach and seeks a ratio indemnities/premium
at 0.6, the loading applied is L = 1.83 The commercial premium charged for the whole country
rises to 197.2 and 218.5 million euro for product 1 and product 2 respectively.

The “loading factor” of the disaster fund is very difficult to assess: administrative costs,
borrowing costs on financial markets… These administrative costs are very hard to estimate as
they are spread among a variety of administrations and are not monitored separately. For
example, how do we estimate the cost of a meeting at the ministry of agriculture attended by
various experts and civil servants? Our hypothesis is that the public disaster fund and the
private insurer bear the same costs and have the same risk appetite. Therefore, we apply the
same loading factors to the FNGRA to calculate what a commercial premium would be. The
main difference is that the French farmers contribute only to the resources of the fund, and do
not support its administrative costs. In other words, they are not paying the loading factor.

Table 9 - Yearly premium in million euros to cover 10.5 million hectares of pasture

FNGRA Product 1 Product 2

Pure premium 107 108 119
Commercial Premium, loading =
157 158 175
1.47
Commercial Premium, loading =
196 197 219
1.83

In France, agricultural insurance premiums can be subsidized up to 65%. This has been the
subsidization rate between 2013 and 2015 during the pilot phase of the grassland insurance
project. Depending on which product the State is willing to promote and the loading factor

103
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

applied by the insurer the total cost for the State varies between 1033 and 142 million euros per
year.

Table 10 - Premium cost for the State and for the farmer, depending on the coverage and the
loading factor

FNGRA Product 1 Product 2

Total cost Cost per Total cost Cost per Total cost Cost per
in million hectare in in million hectare in in million hectare in
€ € € € € €
Pure premium 107 10,2 108 10,2 119 11,3
incl. cost for
54 5,1 70 6,7 77 7,4
the State
incl. cost for
54 5,1 38 3,6 42 4,0
the farmer
Commercial
Premium, loading 157 15 158 15 175 16,7
= 1.47
incl. cost for
104 9,9 103 9,8 114 10,8
the State
incl. cost for
54 5,1 55 5,3 61 5,8
the farmer
Commercial
Premium, loading 196 18,6 197 18,8 219 20,9
= 1.83
incl. cost for
142 13,6 128 12,2 142 13,6
the State
incl. cost for
54 5,1 69 6,6 77 7,3
the farmer

On average, the FNGRA has distributed 107 million euros per year over the period 2003 -2015.
This cost is borne by the State, but the FNGRA is partially financed through a tax paid directly
by farmers. Therefore, we will consider that the cost of the FNGRA disbursement is borne by
farmers at 50%

6. Discussion

6.1 Discussion on the model’s assumptions

For the purpose of comparison, we decided to assume that 100% of grasslands would be
insured, as the FNGRA covers all the farmers. However, given current crop insurance
penetration rates in the French market, this assumption appears quite unrealistic unless other
measures are implemented to support agricultural insurance and increase its penetration rate.

104
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

The cost of insurance premium subsidies will also depend on the guarantees and the subsidy
rate. We have seen that this subsidy rate fluctuated in the past, which is said to have hurt the
development of multi-peril crop insurance (Lidsky et al., 2017). On the other hand, the latest
evolutions of the European regulatory framework indicate that Member States will have the
possibility to increase the subsidy from 65 to 70%, and lower deductible from 30 to 20%
(Bachelet, 2017). Member States can use this possibility, but they may decide not to do so. If
Member States decide to increase the level of subsidy and lower deductibles, this will
automatically lead to increased spending, but how these changes will affect farmers’ take-up of
insurance is hard to predict.

Finally, to design the insurance scheme (product 1 and product 2), we used data from the period
2003-2015. During this period, the year 2003 was a year of extremely severe drought. We did
not give any special weight for this year. The key question here is: was the 2003 drought a 1-
in-100-year exceptional event, or just a 1-in-10-year event? The answer to this question will
define what is the loss expected by insurer, and the corresponding premium.

6.2 Evaluating the alternative from the State’s perspective

When it comes to evaluating public policies, a lot of different frameworks with their own
methodology exist (Perret, 2010). However, certain criteria are recurrent. In this case, we will
focus on three criteria:
- Effectiveness: does the policy reach its targets?
- Efficiency: is the policy cost-efficient?
- Sustainability: is the policy likely to have a long-term impact?
Two other dimensions are also classically assessed when performing a public policy evaluation
but will not be used in our analysis:
- Relevance: Are the policy’s objectives covering the different dimensions of the problem
it is supposed to address?
- Coherence: Are the means and objectives of the public intervention coherent among
themselves and with other public policies?
When comparing a disaster fund with an insurance approach, we do not question the relevance
of the public authorities intervening in the management of agricultural risks, but we rather
compare different tools available. In addition, the issue of coherence would involve analysing
the other forms of public interventions (subsidies, tax exemptions, market interventions,
tariffs…) which is out of the scope of this research.

105
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

In our analysis, we decided to integrate two additional criteria, transparency and fairness.
Indeed, the two public policy options available have different implications for these two
dimensions.

Effectiveness
The objective of the public intervention is to reduce the volatility of farmers’ income by
guaranteeing a minimum yield. Our dataset shows that the insurance fulfils better this objective
by matching better the payouts with the loss. First, the use of an index guarantees a fair and
transparent loss assessment at the local level, when the disaster fund relies on political decrees
taken at the department or sub-department level. Second, the insurers’ financial capacities and
access to the reinsurance markets allow a larger disbursement capacity in case of a major shock.
The 2003 drought depleted the FNGRA’s resources, leading to a very low compensation of
affected farmers in 2003, and delayed payments in 2004 and 2005. These two dimensions make
that insurance’s disbursements match better farmer’s losses (and cash-flow needs) than the
disaster fund.

Efficiency
It is difficult to compare the complete cost of the two tools as the cost of running the FNGRA
is not visible and there is no such thing as a “loading factor” as we used it to estimate a
commercial premium charged by an insurer. In the different scenarios presented in Table 4, the
annual cost for the State to subsidize insurance for all French farmers is estimated between 93
and 124 million euros annually, against a euro 53 million annual average cost for the FNGRA.
This assumes a 65% subsidy, and that 100% of the cattle breeders purchase the insurance. The
second assumption appears rather unrealistic, while the rate of subsidy can be modified
according the government’s financial capacity. It should also be noted that Product 2 offers a
higher protection to farmers. The impact on the public budget of switching from a disaster fund
to insurance therefore depends on the farmer’s participation, the subsidy rate and the level of
coverage chosen.

Sustainability
The main advantage for the State to switch from ex-post compensation to ex-ante premium
subsidies is that the costs are predictable. Insurance premiums are relatively stable over time,
and even if there are variations, they are not comparable to the volatility of payouts from the
fund. On the other hand, when relying on its own disaster fund, the State is not sure to have the

106
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

necessary resources to cope with a major shock. This was the case for the 2003 drought, for
which payments were spread over three years as this exceptional event depleted the resources
of the fund. An insurer is obliged by the prudential regulation to keep significant reserves. It
also has the possibility to get access to the reinsurance market to protect itself from extreme
events. Another risk faced by the disaster fund is that the government may choose to draw off
from the reserves of the disaster fund in order to balance its budget. This is what happened in
France with the 2016 budget4, for which 255 million euros were taken from the FNGRA
reserves. If the government had chosen to subsidize insurance, it would not have been in
capacity to reallocate the budget. The switch from one system to the other would then imply a
clarification of the rules and protection offered to farmers, and a stricter commitment from the
State.

Transparency and fairness

Though transparency and fairness are not always included in public policies evaluation, it
seemed interesting to include them as the choice of one tool or the other has implications for
these two dimensions.
An interesting feature for the State is that the forage insurance proposed here is based on an
index that is transparent and objective. In particular, the index should not be subject to political
pressure. It also operates at the zip-code level, whereas FNGRA indemnities are usually decided
at the level of the department, or a sub-area of the department. The FNGRA has been criticized
for its lack of objectivity regarding the decisions to give payouts, especially as most of payments
followed demands from professional organizations and unions (Cour des Comptes, 2012).
Finally, it has been shown that the disaster fund payments were structurally focused on certain
crops and certain regions (Boyer, 2008; Cordier et al., 2008). This means that some farmers are
subsidizing others through their contributions to the fund. Subsidized farmers are not paying
the real price for the risk they face, which may prevent needed structural evolutions in the
farming sector (Mortemousque, 2007).

An important limitation of this reasoning lies in the capacity of the State to “tie its hands”.
Indeed, Kydland and Prescott (1977) introduce the concept of “dynamic inconsistency” to
explain the fact that States might be unable to implement socially optimal policies because they
cannot convince agents that they will keep their promises. This theory finds its application in
monetary economy, where a government’s declaration to limit inflation will have no effect as

4
[Link]

107
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

agents anticipate the government may decide to change its policy in order to boost growth and
employment. Our situation may look very similar as a government may announce the
withdrawal of a disaster fund in favour of an insurance system, but break its promise in case of
a major shock because of political pressure, or to limit the scale of the crisis. Therefore, the
incentive for farmers to buy insurance is seriously lowered as they anticipate a government’s
disbursement in case of a major shock. To a certain extent, this has been the case with United
States Crop Insurance Program: even if crop insurance subsidies and take-up were increased
dramatically in the 1980’s and 1990’s, the Congress continued to pass disaster bills to
indemnify farmers (Glauber, 2007). It is also legitimate to question whether the government’s
goal is to actually implement socially optimal policies. It is well known since Buchanan and
Tullock (1962) and more generally in the Public Choice literature that political decision-making
outcomes can conflict with the general public’s interest, especially when groups within the
society are rent-seeking and successfully advocating their own interests. Farmers might be
politically strong enough to influence the risk management policy and increase the
government’s participation even when it involves maintaining inefficiencies.

6.3 Client value

The academic literature on corporate risk management (CRM) offers studies on the link
between corporate value and corporate hedging policies. Most of the studies use arguments that
are derived from the financial theory or from the agency theory, and there is no consensus about
the fact that risk-hedging creates corporate value. On the one hand, hedging activities create
corporate value because they decrease fiscal payments, reduce the cost of financial distress,
mitigate the underinvestment risk, or ease the external financing. On the other hand, hedging
activities are costly.

When considering the perfect capital market hypothesis, firm’s shareholders can manage their
risk through diversification. In such a context, enterprise risk management does not bring value
to shareholders. Theoretically, hedging systematic risk at the corporate level does not increase
the shares value (Dufey and Srinivasulu, 1983). However, a more realistic approach has to take
into account capital market imperfections, like agency costs resulting from conflicts between
stakeholders, transaction costs related to external financing, or cost related to financial distress
or tax scheme convexity. That is why, in the presence of capital market imperfections, corporate
risk management can increase firm’s value to shareholders. Aretz, Bartram and Dufey (2007)
give a clear and comprehensive presentation of the CRM rationales, including:

108
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

- the mitigation of the underinvestment problem;

- the reduction of the cost coming from the shareholder tendency to opt for riskier
investment projects, facing risk-adverse lenders;
- the reduction of the cost coming from managers who cannot simply diversify away their
idiosyncratic risk, contrary to shareholders;
- the design of managers’ incentives, in order to reduce the mismatch with the shareholder
interest;
- the cost resulting from the mismatch between the firm’s investment policy and the
firm’s financing policy;
- the reduction of the financial distress cost;
- the reduction of the corporate tax burden, resulting from fiscal schedule.

Those rationales for corporate risk management have been subject to empirical investigation:
for example, papers by Nance et al. (1993), Tufano (1996), Haushalter (2000), or Allayanis and
Weston (2001) give evidence for the different theoretical arguments applied to a variety of
industries.

The case of farmers and “agri-producers” has been studied by McKinnon (1967), Finger (2012),
using the « natural hedge » argument: the negative relationship between yields and prices is
relevant to understand the farmer risk management decision making. The opportunity cost of
pre-sold futures positions can be challenged by the use of crop insurance, because this last
would pay only when the yield drops below a trigger. Following the idea that crop insurance
and financial hedging may be partial substitutes, Walters (2015) examines the optimal trade-off
between crop insurance, financial hedging, and publicly funded government programs in the
US (disaster insurance program under the Farm Bill). The study shows that the choice of a
combination between revenue protection crop insurance and agricultural risk coverage
increases the expected revenue.

The insurance policy has several advantages not priced here, compared with the fund.

First, the insurance product is based on an index, for which the robustness was valuated during
a scientific evaluation. It is not subject to manipulation from the insurer or the State. Therefore,
we can argue that farmers may prefer this approach to an approach based on traditional
insurance (requiring human expertise of the damage) or on a public approach, for which
disbursements depend on a political decision. Farmers may also have more confidence that they
will be compensated in proportion to the size of their loss. We have already mentioned that

109
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

during exceptional events like the 2003 droughts, the resources of the funds had been exhausted
and farmers had to wait the following years to receive the full amount of the disbursements.
This fear of the disaster fund’s insolvency is strengthened when resources of the funds are
diverted for other purposes. Conversely, insurers are constrained by prudential regulations and
are therefore required to maintain sufficient reserves to be able to cope with large-scale shocks.

Another important difference between a disaster fund and an index-insurance product lies in
the delay for settlement. The public disbursement must follow an official decree of natural
catastrophe and evaluations from government experts to estimate the production loss. In case
the fund lacks sufficient resources, the payment will be spread over several years. On the
opposite, an index-based insurance does not require human expertise; it actually does not
require any human decision as soon as the index is available. Payment can be very swift and
the farmer’s treasury management is improved.

Finally, unlike a disaster fund, an insurance policy is a risk management tool for a farm
business. Its costs and payouts are predictable for the farmer. The farmer can also decide and
adjust certain features of the policy: capital insured, threshold, deductible… Indeed, farmers
have different capacities to cope with a shock, and have therefore different insurance needs. For
cattle breeders, it will largely depend on the amount of hay in stock (Mosnier, 2015). With the
disaster fund, farmers have no choice but to rely on the coverage features of the fund and
especially a level of compensation set at only 28% of the losses.

7. Conclusion

Switching from a public disaster fund to a public-private partnership leads to an increase in

public spending, as private insurers apply a loading on the pure premium, unlike the public fund
for which the administration and operation cost is invisible in the public accounts. However,
the fact this cost is invisible does not mean it does not exist. Moreover, the level of public
spending due to premium subsidies depends on farmers’ participation, loading factors applied
by insurers, but also on the subsidization rate. The article provides a framework for government
decision-makers to evaluate policy options according to budgetary considerations and a set of
qualitative criteria.

To be effective, such a move would also require a very clear commitment from the public
authorities that they stop ad-hoc disbursements following a disaster and concentrate their
support on crop insurance subsidies. This is the way taken by the Netherlands, where the

110
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

introduction of multi-peril crop insurance in 2010, assorted with a 65% subsidy, coincided with
a termination of the disaster fund. Since then, and despite adverse climatic events in 2010, 2012
and 2015, the Dutch government resisted political pressure to make any ad-hoc disbursement
by referring to the public-private insurance scheme (Van Asseldonk, et al., 2016). The Dutch
example proves that beyond public finance concerns, agricultural insurance public private
partnerships are also politically feasible.

111
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

References

Antón, J. (2009). Risk management in agriculture : A holistic conceptual framework.

Managing Risk in Agriculture A Holistic Approach, Paris, OECD.

Babusiaux, C. (2000). L’assurance récolte et la protection contre les risques en agriculture.

Rapport pour le Ministère de l’Economie des Finances et de l’Industrie et le Ministère de
l’Agriculture et de la Pêche.

Bielza, M., Stroblmair, J., Gallego, J., & Conte, C. (2007). Agricultural risk management in
Europe. Conference proceeding: 101st EAAE Seminar: Management of climate risks in
agriculture, Berlin, 05‑06.

Bonnafous, A., & D’Arcier, B. F. (2013). The conditions of efficiency of a PPP for public
finances. 13th WCTR-13th World Conference on Transportation Research.

Boyer, P. (2008). Assurer les calamités agricoles? Notes et études économiques, 30, 7‑32.

Buchanan, J. M., & Tullock, G. (1962). The calculus of consent (Vol. 3). University of
Michigan Press Ann Arbor.

Cordier, J., Erhel, A., Pindard, A., & Courleux, F. (2008). La gestion des risques en agriculture
de la théorie à la mise en oeuvre : Éléments de réflexion pour l’action publique. Notes et études
économiques, 30, 33‑71.

Cour des Comptes. (2012). Les aides d’urgence en agriculture.

[Link]
[Link]

Cour des Comptes. (2013). Référé 65742.

[Link]
[Link]

Direction de l’initiative parlementaire et des délégations. (2016). La couverture des risques en

agriculture et les assurances agricoles (No 269; Étude de législation comparée).
[Link]

EC COM. (2005). Communication from the Commission to the Council on Risk and Crisis
Management in Agriculture.

EU. (2013). Overview of CAP reform 2014-2020. Agricultural Policy Perspectives in Brief, 5.

Glauber, J. (2007). Double indemnity : Crop insurance and the failure of US Agricultural
Disaster Policy. AEI Agricultural Policy Series: The, 1‑36.

Glauber, J. W. (2015). Agricultural Insurance and the World Trade Organization. IFPRI
Discussion Paper 01473. Washington, D.C.: International Food Policy Research Institute.

112
Chapter 3 - From crisis management to risk management in agriculture: the case of grassland insurance in France

Hazell, P. B. (1992). The appropriate role of agricultural insurance in developing countries.

Journal of International Development, 4(6), 567‑581.
Kydland, F. E., & Prescott, E. C. (1977). Rules rather than discretion : The inconsistency of
optimal plans. The journal of political Economy, 473‑491.

Lidsky, V., Maudet, C., Malpel, G.-P., Gerster, F., Helfter, M., Lejeune, H., & Le Theule, F.-
G. (2017). Les outils de gestion des risques en agricultures (N° 2016-M-099 / N° 16-104).
Inspection Générale des Finances / Conseil Général de l’alimentation, de l’agriculture et des
espaces ruraux. [Link]

Liesivaara, P., Meuwissen, M., & Myyrä, S. (2014). Designing public–private crop insurance
in Finland. The Third International Agricultural Risk, Finance, and Insurance Conference.
June 22-24.2014 Zürich, Switzerland.

Loi de modernisation de l’agriculture et de la pêche, Pub. L. No. n° 2010-874 (2010).

Ménard, C. (2004). Gestions des risques climatiques en agriculture-Engager une nouvelle

dynamique. Assemblée Nationale, Paris.

Mortemousque, D. (2007). Une nouvelle étape pour la diffusion de l’assurance récolte. Sénat,
50p.

Mosnier, C. (2015). Self-insurance and multi-peril grassland crop insurance : The case of
French suckler cow farms. Agricultural Finance Review, 75(4), 533‑551.

Perret, B. (2010). L’évaluation des politique publiques. La découverte. Paris.

Roumiguié, A., Jacquin, A., Sigel, G., Poilvé, H., Lepoivre, B., & Hagolle, O. (2015).
Development of an index-based insurance product : Validation of a forage production index
derived from medium spatial resolution fCover time series. GIScience & Remote Sensing,
52(1), 94‑113.

Torchia, M., Calabrò, A., & Morner, M. (2015). Public–private partnerships in the health care
sector : A systematic review of the literature. Public Management Review, 17(2), 236‑261.

Van Asseldonk, M., Van der Meulen, H. A. B., Van der Meer, R. W., Silvis, H. J., & Berkhout,
P. (2016, octobre 4). Does subsidized insurance crowd out market-based insurance ? An
analysis of subsidized multi-peril crop insurance versus hail insurance in The Netherlands.
156th EAAE Seminar on Prospects for agricultural insurance in Europe, Wageningen, The
Netherlands.

Wettenhall, R. (2003). The rhetoric and reality of public-private partnerships. Public

Organization Review, 3(1), 77‑107.

Young, V. R. (2004). Premium principles. Encyclopedia of Actuarial Science.

113
114
 CHAPTER 4

RETHINKING AGRICULTURAL INSURANCE FOR THE BOP – A

KENYAN CASE STUDY

115
Rethinking agricultural insurance for the BoP – a Kenyan case study

1. Introduction

As seen in chapters 1 and 2, insuring farmers to protect them from natural and climatic risks
poses a certain number of challenges. The global volume of agricultural insurance premiums is
mostly found in North America and Europe (55% and 18 % the world premium volume
respectively), but the share of Asia is rapidly growing (Kalra & Xing, 2013). In 2011, Latin
America accounts for 4% of the premiums, and Sub-Saharan Africa for about 0.5% (Kalra &
Xing, 2013). To date, agricultural insurance programs in the developing world are either heavily
subsidized, or are remaining at a pilot level and have not reached scale.

Low-income customers face specific challenges to get access to insurance products. Among
them are the issues of price, liquidity, trust, financial education level and wealth (Cole et al.,
2013). The idea that low-income customers experience price and non-price barriers to access
some products or services is not new, and many of these challenges are not specific to the
insurance sector. In 2004, CK Prahalad introduced the paradigm that there is a market at the
“bottom of the pyramid” (Prahalad, 2004). The “BoP” (standing for “bottom of the pyramid”
or “base of the pyramid”) represents the mass of the world population, and its aggregate
purchasing power is important. However, they are currently under or poorly served by large
companies. Most firms are not addressing properly the specific needs and constraints of the
BoP segment, and are therefore missing a broad market. Anderson & Billou (Anderson &
Billou, 2007) propose the “4As” framework to analyze the barriers preventing BoP customers
to access needed goods and services. Firms must take into account the questions of
Affordability, Availability, Acceptability and Awareness when they target BoP customers.

This article relies on the 4As framework to analyze the strategy implemented by Agriculture
and Climate Risk Enterprise Ltd (ACRE), which operates under the brand name “ACRE
Africa”. ACRE designs crop and livestock insurance products for farmers in Kenya, Tanzania
and Rwanda, with a specific focus on smallholder farmers. The company was incorporated in
2014, and it follows the project “Kilimo Salama” that was initiated by the Syngenta Foundation
for Sustainable Agriculture in 2009. Based on this case, the article aims at understanding how
agricultural insurance needs to be redesigned to target the BoP. The case study is based on a
series of 15 interviews with ACRE staff, clients and partners. It also relies on documentation
provided by the company or the Grameen Crédit Agricole Foundation, one of its shareholders.
All ACRE’s products and operations features are analyzed through the 4As framework, in order
116
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

to assess how well the products and customer experience perform against each criterion. This
analytical framework is also confronted with a practitioner tool (PACE) that intends to assess
the value of insurance products.

Section 2 reviews the literature on microinsurance demand. This literature will then be
compared with the literature on the BoP paradigm. Section 3 presents the current state of
practice and the methodologies developed around microinsurance. Section 4 describes the
methodology and the data. The next section presents the main results, using the 4As framework
and the PACE tool. The last section confronts the two analytical grids and establishes the main
lessons that can be drawn from this case study.

2. Literature review

[Link] management and microinsurance demand of smallholder farmers

Low-income faming households deploy a broad array of methods to manage natural and climate
risks. These methods can be classified into two categories: risk reduction measures, and
mitigation and coping mechanisms (Antón, 2009; Bryla-Tressler, 2011; Müller, Ramm, &
Steinmann, 2014). Farmers can ex-ante try to reduce and prevent risks by implementing certain
farming practices, like using drought-resistant seeds or develop irrigation. They can also
diversify their crops and their sources of income by taking another job. After a shock happened,
typical mitigation and coping mechanisms include relying on savings, community and national
solidarity, selling assets or taking a loan.

These methods work quite well to manage frequent, low intensity events (Hazell et al., 2010).
They made possible for communities to survive and thrive even in disaster-prone areas. These
risk-avoidance strategies however bear an important opportunity cost: farmers often choose
low-risk, low-return farming strategies. For example, a farmer may decide not to purchase high-
yield seeds because they are perceived to be riskier. Low-income farming households are too
vulnerable to depart from non-efficient methods. Some studies estimate that average
agricultural income would be 10 to 20% higher in the absence of risk (Gautam et al., 1994;
Sakurai & Reardon, 1997).

Another limitation of traditional risk management tools is that they may not offer sufficient
protection in case of a catastrophic event. Poor households are often forced to sell productive
assets like livestock after a shock, which affects their livelihoods in the long term. It is now

117
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

established that external shocks and asset depletion can “trap” vulnerable households into
poverty (Barnett et al., 2008). Community-based mechanisms are also less effective to protect
from covariant risks like floods and droughts (Christiaensen & Sarris, 2007). Indeed, many
members of the community are affected and need support at the same time when large-scale
catastrophic events occur.

With such imperfect risk management tools and a high-risk aversion, one would expect that
demand for microinsurance products is high among low-income households (Brau et al., 2011;
Cohen et al., 2003). When reviewing the literature on demand for microinsurance De Bock &
Gelade (De Bock & Gelade, 2012) actually find quite the opposite. Take-up rates of index
insurance products reach 20 % in Ethiopia (Hill & Robles, 2011), 17% in Malawi (Giné &
Yang, 2009), 16% in India (Cole et al., 2013), between 6% and 36% in Ethiopia (Norton et al.,
2011), 6% in India (Gaurav et al., 2011). In many of these randomized control trials (RCTs),
insurance was subsidized. Why is insurance demand so low?

Farmers are typically risk-averse: they will prefer to exchange a certain sum (the insurance
premium) on a regular basis to guarantee a certain amount of production outcome (Patt et al.,
2009). A utility-maximizing farmer is therefore willing to pay more than the actuarially fair
price of the policy. Theory suggests than the willingness to pay for insurance will depend on
the farmer’s risk aversion, which means that setting the premium between the actuarially fair
price and the farmer’s willingness to pay would suffice to sell the insurance policy (De Bock &
Gelade, 2012; Patt et al., 2009). Two Empirical studies (Cole et al., 2013; Giné et al., 2008)
have shown results contradicting this theory. Demand for crop insurance (in both cases, index
insurance) is particularly low among the most risk-averse farmers. Karlan and Morduch (Karlan
& Morduch, 2010) advance that uncertainty is about the product itself, and that farmers doubt
that it will deliver its promise. For Binswanger-Mkhize (Binswanger-Mkhize, 2012), farmers
are too cash and credit-constrained to transform their potential demand into an actual purchase.
Clarke (Clarke, 2011) takes a different approach and develops a “theory of rational demand for
index insurance”. Index insurance bears basis risk: the producer’s loss is imperfectly correlated
with the index insurance payout. When basis risk happens, a farmer can end in a worse situation
than if he/she had not purchased insurance: he paid the premium but he/she is not compensated
for his/her loss. “The low observed demand for weather derivatives sold to poor farmers may
be consistent with objective financial advice, rather than being a result of poor understanding,
an unwillingness to experiment, or credit constraints.” (Clarke, 2011).

118
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Overall, a variety of factors affects demand for microinsurance. Eling, Pradhan and Schmit
(Eling et al., 2014) established a list of 12 characteristics determining demand for
microinsurance and reviewed 41 papers to assess the effect of each of these factors.

Table 11 - Papers assessing the correlation between demand and a series of variables

Number of papers finding a correlation

Variable Positive Negative Non-
correlation correlation significant
Economic Factors
1. Price of insurance (including transaction
costs) 10 1
2. Wealth (access to credit/ liquidity) and
income 4 1 5
Social and cultural factors
3. Risk aversion 1 4
4. Non-performance and basis risk 3
5. Trust and peer effects 12
6. Religion/fatalism 1
7. Financial literacy 11 6
Structural Factors
8. Informal Risk Sharing 1 2
9. Quality of service 5
10. Risk exposure 1 2
Personal and demographic factors
11. Age 3 1 1
12. Gender (female is positive) 2 2 1
Source: adapted from Eling, Pradhan and Schmit (2014)

Price, risk aversion and basis risk appear to negatively correlate with demand, except in one
paper finding that risk aversion correlates positively with demand (Ito & Kono, 2010).
Depending on the studies, wealth, financial literacy and risk exposure are positively correlated
with demand, or non-significant. The importance of informal risk sharing, age and gender are
not clear, different empirical studies found contradictory results. Only one paper examined
religion and found a positive correlation between religion and insurance demand (Cole et al.,
2011). In a large number of studies, trust and peer effects, and quality of service are found to
be positively correlated with demand.

119
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

2.2. Creating inclusive markets

In “The fortune at the Bottom of the Pyramid” (Prahalad, 2004), CK Prahalad states it is
possible to “eradicate poverty through profits”. The fundamental assumption of Prahalad is that
the poor are customers who have their needs improperly met. The World Resources Institute
and the International Finance Corporation (World Resources Institute & International Finance
Corporation, 2007) estimate that the bottom of the pyramid represents four billion people. The
definition of the BoP is broader than the population below the poverty line. These 4 billion
people earn individually less than 3000$ of income annually (in local purchasing power), but
their aggregated purchasing power represents $5 trillion dollars. Despite their heterogeneity in
cultures, habits, geographical locations, all BoP consumers share common characteristics. Basic
services taken as granted in developed countries are often non-existent or inadequate for the
BoP. BoP consumers often lack proper access to water and sanitation services, electricity,
transportation services, basic health care, and have limited or no access to formal financial
services. BoP markets are inefficient markets. This inefficiency negatively affects customers
who pay a high price for inadequate products and services, and it is not uncommon for low-
income customers to pay a higher price for some products and services than more affluent
households. When comparing the prices between Dharavi, a shantytown of Mumbai, and
Warden Road, an upper-class community in Mumbai suburb, Prahalad and Hammond (Prahalad
et al., 2002) conclude that the poor live in a high-cost economy. The kilo of rice is 20% more
expensive in Dharavi than in Warden Road. For diarrhea medication or a cubic meter of
municipal-grade water, the “poverty premium” is respectively 20 and 37 times the price of the
upper-class neighborhood. This paradox is, among other factors, the consequence of the high
cost to reach the poor, the “last mile” problem (Counts, 2011; Vachani & Smith, 2008). The
lack of infrastructure, especially for transport and telecommunication, creates a situation where
it is more expensive to serve poor customers than the rich ones. This problem is particularly
acute for rural areas where populations are even harder to reach and geographically dispersed.

Hence, reaching BoP customers requires solving the “4As challenge”(Anderson & Billou,
2007).

Defining the 4As

Availability – the extent to which customers are able to readily acquire and use a product or
service. Distribution channels in bottom of the economic pyramid (BOP) markets can be
fragmented or non-existent and the task of simply getting products to people can be a major

120
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

hurdle to overcome. Companies need to explore alternative methods of delivering their products
and services to even the most isolated BOP communities.

Affordability – the degree to which a firm’s goods or services are affordable to BOP
consumers. Many low-income consumers in developing countries survive on daily wages,
meaning that cash-flow can be a significant problem. Companies need to be able to deliver
offerings at a price point that enables consumption by even the poorest consumers.

Acceptability – the extent to which consumers and others in the value chain are willing to
consume, distribute or sell a product or service. In BOP markets, there is often a need to offer
products and services that are adapted to the unique needs of both customers and distributors.
Companies might need to respond to specific national or regional cultural or socioeconomic
aspects, or to address the unique requirements of local business practices.

Awareness – the degree to which customers are aware of a product or service. With many BOP
customers largely inaccessible to conventional advertising media, building awareness can be a
significant challenge for companies wishing to serve low-income consumers in the developing
world. To overcome these constraints companies must explore alternative communication
channel.

Reproduced from Anderson & Billou, 2007

Most companies do not have the capacity to tackle these challenges on their own. This is why
most BoP ventures are built around partnerships, between companies, MFIs and non-
governmental organizations (NGOs). These partnerships between NGOs and for-profits do not
come without costs and adjustments (Webb et al., 2010). Partners involved in a BoP venture
may pursue different objectives and conflicts may weaken the alliance. Moreover, producers’
cooperatives, local businesses, churches, public services or governments, even if not formally
integrated in the project, can be powerful enablers in the communities (UNDP, 2008).

The BoP “paradigm” evolved from the initial views of Prahalad to take more into account the
importance of partnership creation. Tweaking some products’ features is not sufficient to
address the needs of the BoP. This new approach is often referred to as “BoP 2.0” (Erik Simanis
& Hart, 2008).

Table 12 – BoP paradigm evolution

BoP 1.0 BoP 2.0

BoP as consumers/producers BoP as business partner

121
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Deep listening Deep dialogue

Reduce price points Expand imagination
Redesign packaging, extend distribution Marry capabilities, build shared commitment
Arm’s length relationship mediated by NGOs Direct, personal relationships facilitated by
NGOs
“Selling to the Poor” “Business co-venturing”
Reproduced from Simanis & Hart, 2008

Despite enormous potential, addressing the “BoP market” presents many challenges. A market
exists only if there are buyers willing to pay and sellers able to cover their costs, including the
cost of capital (Garrette & Karnani, 2010; Karnani, 2005). “If the supply and demand curves
do not intersect, there will be no market, even if there is an unmet need”(Garrette & Karnani,
2010). Actually, many companies fail to address the BoP segment because they fail to adopt a
“market creation” strategy, as opposed to “market entry” (Simanis, 2011). Entrepreneurs
entering a new market face “information deficiency”. They can address this information gap
through market research or establishing partnerships with businesses already active in this
market. The situation is different when there is no existing market: there are no competitors, no
customers to observe. Therefore, creating a new market carries a huge upside potential but
involves much more complexity, requires higher amount of investment with a longer payback.
Another common mistake of many BoP ventures is that their strategy could be summarized as
follows: “low price, low margin, high volume” (Simanis, 2012). In markets that do not exist
yet, assuming high penetration rates to balance the business model is simply unrealistic. BoP
ventures need to find margin-boosting solutions if they want to be sustainable.

[Link] needs are not equivalent of untapped markets

Both the literature on microinsurance demand and on BoP markets come to the same
conclusion: the existence of a need does not necessarily mean that there is a high demand for
products and services addressing this need. Low-income households face various constraints
(level of income, liquidity, remoteness) and information gaps that exclude them from insurance
services. Therefore, microinsurance products must be designed to address the needs and
constraints of these customers in order to meet their market.

122
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

3. The practitioners’ perspective: assessing client value in microinsurance

3.1. Why and how assessing client value

The question of microinsurance clients receiving good value against their premiums is central
for the microinsurance sector. Insurance providers and distributors aim at delivering good
products to attract and retain clients. Governments and donors want to evaluate if
microinsurance has a positive impact on the lives of the poor. In this chapter, we follow Matul,
Tatin-Jaleran & Kelly (Matul et al., 2012) to adopt a client-focused definition of the client value,
mixing the development and marketing perspectives. “Client value is about reducing
vulnerability due to improved risk management practices that then contributes to improved
well-being”, and this definition assumes that “valuable products are a means to accomplish
clients’ goals and satisfy their needs” (Matul et al., 2012)

The issues of “client value”, “social performance” or “impact” gained a lot of attention in the
early 2010’s and several methodologies were developed in order to track and measure client
value (see Appendix 1). These approaches vary in the level of complexity involved, the sources
of data and rationale. A market study will be more relevant at the product development stage to
understand the needs and preferences of the target population while a client satisfaction study
while focus on past and existing clients with an objective to improve an existing product. In
this research, we decided to use the PACE tool developed by the Microinsurance Innovation
Facility5 hosted at the International Labour Organization. Compared to other client value
assessment approaches, the PACE tool is relatively simple to implement. It relies on secondary
data, like employees and clients’ interviews, which were also useful for other aspects of this
research.

3.2. The PACE tool

The purpose of the PACE tool is to help “practitioners develop a better value proposition for
clients.” (Matul & Kelly, 2012). It aims at improving the client value rather than proving it, and
is not a substitute for demand or impact studies. The PACE tool is structured around the four
dimensions of Product, Access, Cost and Experience.

5
This facility is now called Impact Insurance Facility

123
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Figure 12 - PACE added value analysis framework

Reproduced from PACE technical guide v1.0 2012

The underlying assumption behind this analytical framework is that microinsurance products
need to be appropriate, accessible, affordable, responsive and simple to fulfill their purpose. A
defining feature of the PACE methodology is that it looks at both the product characteristics
and the related processes. Indeed, a very good product hard to access or not responsive to claims
will deliver little value to the target population. In many microinsurance schemes, the room for
improvement lies in the distribution or customer experience dimensions.

In its design, the PACE tool aims at comparing insurance products or risk management
mechanisms addressing the same need. However, in the context of our study, alternatives were
scarce, especially in the formal insurance market.

Each of the four dimension is divided into several sub-dimensions, corresponding to detailed
criteria (see Appendix 2). Each sub-dimension is weighted relatively to its own importance, and
receives a score between one (poor performance, ineffective and not appropriate to client needs)
and five (strong performance, effective in almost all situations, with limited or no room for
improvement). It can be argued that in some contexts, certain criteria are more important than

124
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

others, and therefore the weighting can be modified to give more importance to certain aspects.
However, in this analysis we decided to keep the standard weights.

4. Case study: methodology and data

4.1. Research objectives

The primary objective of this case study is to evaluate the adequacy of ACRE’s offer with the
needs and constraints of their target clients, smallholder farmers.

The case study pursues three specific objectives: documenting ACRE’s products and processes
from a client’s perspective to have a detailed description of the service brought to clients;
assessing the relevance of ACRE’s offer under the lens of the BoP literature, using the 4As
framework; comparing this analysis from the scholar’s perspective with a practitioner’s tool,
the PACE tool.

4.2. Method

The case study is a qualitative transversal study based on semi-directed interviews.

Studies based on a single case can be useful to test a one or several theories in order to confirm
the model, question it or complete it (Yin, 2017). In this research, the case study allows to
compare two approaches, the 4As framework and the PACE tool. Single-case studies also give
the opportunity to provide a detailed description of the content. As noted by Sandmark et al.
(2013), agricultural insurance research too often gives insufficient details on the actual
insurance contracts, such as risks covered and exclusions, premium rates, deductibles and
thresholds. It is often said that single case studies only have idiosyncratic value, and that it is
impossible to draw more general conclusions (Thietart, 2014). However, if statistic
generalizations are indeed not feasible, it is possible though to have at least a partial theoretical
generalization (Hlady-Rispal, 2015).

Case studies also raise the question of the positioning of the researcher towards the case studied.
Whereas positivist authors affirm the necessary exteriority of the researcher and the erasing of
their own subjectivity (Friedberg, 1988), for ethno-methodologists neutrality is a myth and
researchers necessarily have transformational influence on the object they study (Plane, 1996).
In this research, the author had a dual and somehow ambiguous position vis-à-vis the people
interviewed and the company. Indeed, the author was at the same time pursuing his research for

125
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

the completion of his PhD while also working for the Grameen Crédit Agricole Foundation,
one of ACRE’s shareholders. It was clearly stated from the beginning that the interviews would
serve primarily the author’s research, but of course, there was no Chinese wall in place within
the Foundation and the author never claimed that there would be confidentiality towards his
employer.

[Link] collection

4.3.1. Interviews

This study is based on data gathered during 2 weeks of fieldwork in April 2017 in Kenya. The
author led 15 interviews with the objective to understand the operations and strategy of ACRE.
8 ACRE staff were interview at the headquarters level in Nairobi. Additionally, 3 ACRE
employees based in the field were interviewed in the Meru region. The author also met with
two farmers who are ACRE’s clients (in Meru and Thika) and with two intermediaries (an
agrovet shop manager in Thika and a dairy cooperative in Meru).

An open, semi-structured protocol guided these interviews, focusing on ACRE’s operational

processes and products, and how these features may affect the client’s experience. Moreover,
the interviews at HQ level also explored the company’s strategy and its evolution over time.

All interviews were conducted individually, except for the two Business Development Officers
who were interviewed together. The grid for the interviews was established prior to the
fieldwork and was adapted to the profile of the interviewee. Indeed, the interviewees hold
different positions: employee, client or partner (agrovet shop manager, dairy cooperative
manager) of ACRE. Moreover, even within employees, each person had their own area of
expertise and perspective on the company’s products and operations. The grid also evolved
incrementally during the fieldwork as the understanding of ACRE’s functioning increased and
new questions arose through the fieldwork process.

The selection of people interviewed was made in collaboration with ACRE. The objective was
to have a diversity of profiles and perspectives. All the staff at HQ with management
responsibilities has been interviewed. It was assumed that HQ staff would bring a more strategic
vision, in their respective areas of responsibilities. Regarding the field staff, their focus is in
sales and promotion activities, and they interact more regularly with the end clients or the
aggregators. They were chosen because they hold different positions and different levels of

126
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

seniority. The sites for the field visit were selected based on the variety of agricultural activities
and taking into account practical factors (distance mainly).

All interviews were conducted in English, which is not the mother tongue of neither the author
nor the interviewees, but is the working language within ACRE. Translation was sometimes
needed for the interviews with farmers, and ACRE staff brought their support in the matter.
During the interviews in the Meru and Thika regions, an ACRE’s employee always
accompanied the author.

Table 13 – List of interviews

Location Category Interviewees

Nairobi – ACRE HQ ACRE HQ staff - Business development Officers (2)
- CEO
- Business analyst
- Chief Actuary
- Agriclimate data analyst
- HR manager
- Platform manager
Thika Intermediary - Agrovet shop manager
Thika Client - Farmer (replanting guarantee)
Meru ACRE field staff - Account Associate
- Portfolio Manager
- Field promoter
Meru Intermediary - Dairy Cooperative Manager
Meru Client - Farmer (replanting guarantee)

4.3.2. Other sources of data

On top of the interviews, the author had the opportunity to follow the Portfolio manager and
the Account Associate for half a day, during their meetings with clients in the Meru area. This
gave the opportunity for the researcher to observe the sales process in action.

The author also had access to other types of documents, made available before, during and after
the fieldwork. These sources notably include products information sheets, organization charts,
board meetings material and some company’s internal documentation on processes. Such
documents were provided either directly by ACRE, found on the company’s website, or
provided by the Grameen Crédit Agricole Foundation, in its capacity of ACRE’s shareholder
and board member. Additionally, during the field work, the researcher had the possibility to
witness a demonstration of the company’s internal software, which helped understand how
127
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

ACRE stores, organizes and manages the data it uses to run its operation and create insurance
products.

4.4. Data analysis

All interviews were recorded with a smartphone, except one due to technical reasons. The
author took manual notes during the interviews, and all interview recordings were transcribed,
analyzed and summarized after the fieldwork.

The research ensured that a triangulation of different sources was implement to attain the
research objectives, whether these sources are different interviews with different people, or the
analysis of documentation. The relative importance of the different topics was influenced by
how frequently these topics would be raised during the interviews, by the diversity of the
interviewees who would raise these topics, and by their legitimacy according to their area of
expertise to address the issue.

In the end, the data analysis process was guided by the two frameworks that had been previously
identified: an academic framework, the 4As framework, and a practitioners’ framework, the
PACE tool. These two analytical frameworks were described extensively in the sections 2.2
and 3.2 of this chapter.

5. Results

5.1. Description of ACRE

All the fieldwork was conducted in Kenya. ACRE also proposes its products in Rwanda and
Tanzania, and most findings from this research would probably be relevant for these two
countries, but the case study will focus on ACRE’s work in Kenya.

5.1.1. Presentation of the company

ACRE follows the project Kilimo Salama initiated by the Syngenta Foundation for Sustainable
Agriculture in 2009. It develops and offers affordable insurance products to African farmers
(smallholder to large-scale commercial farmers). The Kilimo Salama project transformed into
a company in 2014, and in 2017, it operates in three countries: Kenya, Rwanda and Tanzania.

ACRE is not an insurance company. It is rather a service provider for the agricultural insurance
market. ACRE main services are:

128
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

- product development: ACRE designs the insurance products, defining guarantees, calculating
risks and premiums

- data collection: ACRE collects and analyzes different natures of data to design the insurance
products: ground data obtained from weather stations, satellite data, and yield data

- brokerage: ACRE finds insurance and reinsurance capacity for its products

- marketing and distribution: ACRE markets the products that are sold under ACRE Africa
brand name and is in charge to manage the relationships with the aggregators who are
distributing the products

- financial education: ACRE implements communication campaigns to educate farmers about

agricultural insurance

- portfolio monitoring: ACRE tracks losses so insurers can issue payments to affected farmers

5.1.2. Partnerships and role in the value chain

ACRE serves as an intermediary between the supply and demand, and as a market maker and
product developer. Depending on the country, it operates as an insurance agent or an insurance
surveyor.

For each insurance product it designs, ACRE finds insurance and reinsurance capacity. ACRE
established business relationships with several local insurers and international reinsurers. If
ACRE is not carrying risk itself, ACRE does make the risk analysis and actuarial calculations.
ACRE and the insurers and reinsurers use this analysis and the indexes developed by ACRE to
establish the level of premiums. In case of a loss event, the insurers and reinsurers pay the claim.
For some products, losses are not always assessed by an index (see paragraph 3.2.3 below). In
this case, the insurance company is in charge of the loss adjustment.

Regarding distribution, ACRE does not market its products to individual farmers, which would
be too costly and ineffective (Mukherjee et al., 2017). ACRE works through aggregators, i.e.
organizations that are directly dealing with large numbers of farmers. These aggregators can be
financial institutions like banks or MFIs, producers’ organizations, NGOs or farm input
suppliers. Working through aggregators is a way to get access to a large number of potential
clients. When the aggregator has established a long-term and trustful relationship with farmers,
ACRE products will benefit from this positive image. Partnering with aggregators also makes
it simpler to bundle insurance with other products and services. Insurance is usually bundled

129
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

with farm inputs, credit, extension services or commercialization. Bundling improves the value
proposition of insurance, and makes it more attractive to farmers (Mukherjee et al., 2017).
Aggregators participate in the marketing effort towards clients. Payments go through these
aggregators, as they usually collect the premiums and receive claim payments.

Figure 13 – ACRE’s role in the value chain

Source: author

ACRE is what Bernhardt, Steinmann and McCord call a “microinsurance intermediary”

(Bernhardt et al., 2012). In traditional insurance markets, intermediaries serve as matchmakers
between demand and an existing supply. In microinsurance markets, intermediaries have a more
significant role than the matchmaking function, they are markets makers. From a client
perspective, their value lies in building a formal market that caters their needs, bringing new
products and establishing new delivery channels. On the insurer’s side, microinsurance
intermediaries provide valuable information on a potential market.

5.1.3. Products

For the research, we will analyze four agricultural insurance products offered by ACRE.

The replanting guarantee

130
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

The replanting guarantee (RPG) is ACRE’s flagship product. It is an insurance bundled with
seeds. The RPG covers the risk of poor rainfall during the plant’s germination phase. Upon
opening the bag of seeds, the farmer will find a scratch card, similar in design to phone credit
scratch cards. This card contains a unique code and instructions. Farmers are asked to send the
code by SMS to a platform, which activates the cover for the next 14 days. Weather will be
monitored at the farmer’s location and in the event of insufficient rainfall, farmers directly
receive on their mobile phone a compensation equals to the price of the bag of seeds.

The product is entirely free for the farmer: the insurance is already included in the price of the
bag of seeds at no additional cost. The premium is paid by the seed manufacturer. Farmers do
not have to fill any form; all transactions are made directly through the mobile platform.

Top-up option

This product is an option to complement the replanting guarantee. It gives the possibility for
the farmer to extend the coverage period to the whole season, after the germination phase when
the free insurance expires. ACRE was piloting this product during the field trip. Farmers who
had registered to the RPG were informed by SMS about the possibility to extend their free
coverage. Farmers need to call a hotline where a salesperson will explain the guarantees of the
product and enroll clients. The farmer pays the top-up option premium. Like the RPG, all
transactions are operated by mobile phone.

Hybrid cover

The hybrid cover is a crop insurance combining index insurance and more traditional Multi-
Peril Crop Insurance (MPCI) yield guarantee. This means some risks are monitored with an
index (drought or excessive rainfall), while other risks are monitored through traditional, in-
the-field loss adjustment methods (pests, diseases, hail, frost, fire). This combination of
methods allows benefiting from the advantages of index insurance for the most common,
rainfall-related risks, while also protecting farmers against a broader range of risks that are
difficult to track with an index. The hybrid cover is available for many different crops (maize,
wheat, potatoes, sorghum…). It bears a 15% deductible and guarantees losses up to 65% of the
guaranteed yield. Farmers can cover the cost of production or the value of the production, at a
pre-agreed market value of the harvested crop.

131
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Figure 14 – Hybrid cover payout structure

0-15 No payout (below deductible)

15-65 Payout zone

65-100 No payout (above yield guarantee)

Source: reproduced from ACRE

The hybrid cover is marketed through financial institutions, NGOs and producers’
organizations.

Livestock insurance

The livestock insurance covers dairy cows between 1 and 8 years old against some causes of
death, like non-preventable diseases and accidents. A veterinary doctor examines each cow
before enrollment, to assess its value and if there is any pre-existing condition. A set of vaccines
are mandatory. After the animal’s death, the veterinary doctor will determine the cause of death
to validate it is covered by the insurance policy. The sum insured is 80% of the animal resale
value. Premium rate is at 3.5% of the sum insured if the sum insured is below 1 million Kenyan
shillings (KES), and 2.5% from 1 million KES and above. There is also a 1000 KES
(approximately 10 US dollars) vaccination fee per animal. The livestock insurance is marketed
mostly through dairy cooperatives.

Summary of products’ features

Table 14 – Summary of products’ features

RPG Top-up option Hybrid cover Livestock insurance

Crop/Animal Maize, wheat, Dairy cows between

Maize Maize
Covered potatoes, sorghum 1 and 8 years old
Value of the seeds Value of the seeds Cost of production or
Sum insured Animal resale value
(450 KES per bag) (450 KES per bag) harvest value

132
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Drought, excessive
rainfall, frost Accidental causes of
Risks covered Drought Drought damage, fire, hail, death and non-
unavoidable pests or preventable diseases
diseases

100% of the sum 100% of the sum 15 to 65% of the sum 80% of the sum
Payout
insured insured insured insured
2,5%-3,5%
Premium Free for the farmer 5-6% Depending depending on the
total sum insured
Included in the bag of NGOs, producers' Dairy cooperatives
Distribution
seeds and mobile Mobile platform organizations and and veterinary
Channels
platform lending institutions doctors

5.2. 4As Framework analysis

In this section, we will analyze each of the four insurance products previously described, using
the 4As of Anderson & Billou as a framework. We will use their definition of each criterion.

5.2.1. Replanting guarantee

The RPG performs extremely well on the affordability criterion, as the premium is completely
free for the farmer, the insurance being already included in the price of the bag of seeds. This
bag of seeds is not more expensive than other brands and the interviews revealed it is the more
popular brand and is generally perceived as a quality product. It also does not involve high
transaction costs, as farmers can register directly with their mobile phone. The only cost borne
by farmers is the cost of one SMS. No other cost is incurred during the coverage period or in
case of a payment as the payment is made directly on the mobile money account of the farmers.

Bundling the insurance with the seeds makes the insurance readily available for the farmer.
Seeds are sold in “agro-vet” shops. Agro-vet shops can be found in many rural towns in Kenya.
They are the one-stop shop for farmers, where they can buy inputs, equipment, or services like
soil analysis and consulting.

In terms of acceptability, the RPG has mixed results. The enrollment process is designed to be
light and efficient for the farmers. It is very similar to buying phone credit, so farmers are
comfortable with it. The claim and payment processes are automated, making it effortless for
the farmer. Nevertheless, the sum insured (the value of the seeds) is very low compared to the
farmer’s risk exposure and coverage is limited to the germination phase.

133
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Awareness is also a challenge. Agrovet shop managers received training about the replanting
guarantee, but they do not have time to explain individually to each client the insurance’s
benefits. ACRE’s promotional posters are displayed inside the shop. ACRE hires seasonal rural
promoters to organize awareness raising events with farmers’ groups. These promoters are
farmers themselves and they promote the seeds and the RPG within their communities. Farmers
also receive information about the insurance with the scratch card placed in the bag. Once they
register, they receive informational SMS, even when there is no payment. The interviews
revealed that many farmers are not aware of the insurance. When they find the scratch card,
many will discard it or they will think it is an authenticity certificate. The design of the scratch
card evolved in order to increase enrollment, but the take-up rate remains low (unfortunately,
the author did not get exact figures, but he was told only a fraction the scratch cards distributed
with the seeds are activated).

5.2.2. Top-up option

The top-up option is an extension of the RPG. Only farmers who activated the RPG will have
the possibility to “top-up” and extend the coverage to the whole season. The top-up option
requires an upfront payment after the germination phase. All the process of delivering
information and enrolling clients is made by phone (voice or SMS), which limits transaction
costs for the farmer. The top-up product performs well in terms of availability. The farmer can
take-up insurance from home and without filling any form, as long there is mobile phone
network in the area. In terms of acceptability, it is an interesting complement to the RPG as it
covers the relevant period for the farmer, the whole planting and growing season. However, the
sum insured is the same as the RPG and only a fraction of the production costs are covered.
Awareness-raising is done by the rural promoters and through SMS sent all along the season.
When enrolling, farmers talk to a salesperson on the phone, so they receive adequate
information about the product features.

5.2.3. Hybrid cover

The hybrid cover is more expensive than a pure index-insurance product. This is because more
risks are covered (risks that cannot be monitored with an index) and because of loss adjustment
costs. However, the hybrid cover is about 1% cheaper than conventional MPCI available in the
Kenyan market. Furthermore, no MPCI is available for smallholder farmers. Liquidity to pay
the premium upfront is not an issue as insurance is bundled with credit, or there are credit
possibilities through the cooperatives.

134
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

The hybrid cover is sold through aggregators that are dealing with farmers on a regular basis.
These aggregators are responsible for the enrolment and premium payment. Therefore,
availability does not appear as an issue for the product.

Regarding acceptability, the hybrid cover is an interesting example of how an insurance product
was reengineered to meet farmers’ needs. ACRE’s crop insurance was initially a pure index-
insurance product based on weather and satellite data. However, when a disease devastated
maize harvest in Kenya in 2013, the index-insurance product proved unable to protect farmers
effectively. This is why ACRE decided to modify its crop insurance product, combining the
index with conventional loss adjustment methods. The hybrid cover protects farmer against a
broad range of risks. It is also customizable, as the farmer can choose to insure the value of
inputs, the credit or the harvest value (depending on the aggregator). An issue may arise as
losses are covered until 65%, so in the event of an extreme event, the farmer will be partially
covered. This feature of the product illustrates the trade-off that can exist between affordability
(not covering the most extreme losses makes the product much cheaper) and acceptability
(capping losses may not adequately protect the farmers).

Finally, the awareness raising activities are made in partnership with the aggregators. Therefore,
distribution channels must be trained in order to be in capacity to sell and explain the product.
ACRE does not have a direct view on the sales process. The hybrid product is complex,
requiring technical expertise that is often lacking with aggregators. However, ACRE benefits
from the reputation of the distribution channels who are endorsing ACRE’s products.

5.2.4. Livestock insurance

Livestock insurance is available to protect dairy cows. The premium is between 2.5 and 3.5%
of the value of the cow, and there is an additional vaccination fee. The average sum insured is
around 100 000KES, so the premium for a farmer owning less than 10 cows would be 3500
KES per cow. This can represent a large amount for a small farmer, especially as the insurance
is not always linked with credit.

Enrollment goes through the dairy cooperatives or directly the veterinary doctors. In each case,
the veterinary doctor is involved to assess the health and value of the animal. The whole process
is paperless, veterinary doctors use an app to fill in the information and the animal value is
automatically calculated. Farmers do not need to visit an office or bring documentation; the
veterinary doctor will come directly to the farm for the enrolment, so insurance is easily
available.
135
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

The livestock insurance product protects farmers against the risks they have no control over,
like non-preventable diseases, or accidents like snakebites. In case the animal is sick and the
veterinary recommends slaughtering the animal, farmers receive the insurance compensation.
The insurance covers the main risk (death), but not the risk of disease, even if it affects the milk
production. What is protected here is not the revenue, but the investment. A cow represents a
large sum for a farming household, it is often their most valuable asset and a form of saving for
rural families. Farmers can sell the animal when there is a need for liquidity. The livestock
insurance addresses well the needs of the farmers to secure their assets. The insurance comes
with risk-reduction measures (vaccination), which increases the value of the service.

Regarding awareness, the involvement of a person with a legitimacy (the veterinary doctor)
increases the confidence of farmers in the product. However, there is a long list of risks covered
and exclusions, which can make the product difficult to understand.

5.2.5. Summary of the 4As framework analysis

Table 15 – Summary of the 4As framework analysis

RPG Top-up option Hybrid cover Livestock insurance

+: bundled with +: bundled with +: sold through +: sold through dairy

seeds seeds NGOs, producers’ cooperatives and
Availability
organizations and veterinary doctors
lending institutions
+: product is free +: low transaction +/-: cheaper than -: high upfront cost
+: almost zero costs MPCI but more +: low transaction
transaction costs -: upfront payment expensive than pure costs
needed index
Affordability
+: bundled with
credit
+: low transaction
costs
+: easy to enrol +: easy to enrol +: covers a broad +: covers the main
-: low level of -: sum insured is low range of risks risk
Acceptability protection (coverage +: comes with risk
period and sum reduction measures
insured)

+: presence of rural +: presence of rural +/-: depends on the +: trust in veterinary

promoters promoters capacity of the doctors
+: information +: information given aggregators -: complex product
Awareness directly on the by SMS and call +: trusted
scratch card centre aggregators
-: no involvement of -: complex product
agrovets

136
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

-: low understanding
of the product

[Link] tool results

Many of the findings from the 4As framework analysis can be used to fill the PACE tool. Some
additional sub-dimensions need to be further explored like the 1.3 “Eligibility criteria”, 2.1
“Choice and enrolment” and 4.4 “Customer care”.

The detailed scoring on each sub-dimension is given in the table below.

Table 16 - PACE tool score of the ACRE products

Top- Hybrid
Product name Weights RPG Livestock
up cover
1. PRODUCT 2,2 2,5 3,55 3,9
1.1 Coverage, service quality, exclusions and waiting periods 0,35 2,5 3,5 4 3,5
1.2 Sum insured in relation to cost of risk 0,35 1,5 1,5 4 4
1.3 Eligibility criteria 0,15 4 4 4 3,5
1.4 Value-added services 0,15 1 1 1 5
2. ACCESS 3,9 4 3,33 3,7
2.1 Choice and enrolment 0,35 4 4 3,5 3,5
2.2 Information and understanding 0,35 3 4 3 4
2.3 Premium payment method 0,15 5 3,5 3 3
2.4 Proximity 0,15 4,5 4,5 4 4
3. COST 5 4,7 3,15 2,7
3.1 Premium in relation to benefit 0,35 5 5 3 3
3.2 Premium in relation to client income 0,35 5 5 3 3
3.3 Other costs 0,15 5 4,5 4 2
3.4 Cost structure and controls 0,15 5 3,5 3 2
4. EXPERIENCE 4,6 4,6 3,85 3,5
4.1 Claim processing procedures 0,35 5 5 4 3
4.2 Claim processing time and/or quality of service provided 0,35 5 5 4 4
4.3 Policy administration and tangibility 0,15 3 3 3,5 3,5
4.4 Customer care 0,15 4 4 3,5 3,5

We summarized the results of the PACE tool in the chart below.

137
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Figure 15 – PACE results

product RPG

Top-up

Hybrid cover

Livestock

experience access

cost

The explanations for the score of each sub-dimension are given in the table below.

Table 17 - Explanation for PACE scores

Product name RPG Top-up Hybrid cover Livestock

1. PRODUCT
Covers against the risk
Covers against the risk
of drought (main risk).
of drought (main risk)
Limited in duration Covers a large variety
and for the whole
1.1 Coverage, (only germination of risks through a Covers the risk of death
season. No possibility
service quality, phase). No possibility combination of but not the risk of
to customize the
exclusions and to customize the methods. Product disease. Exclusions for
coverage (sum insured
waiting periods coverage (sum insured offered before the preventable risks.
or risks covered). No
or risks covered). No planting season.
exclusion. Immediate
exclusion. Immediate
enrolment.
enrolment.
Possibility to choose
the sum insured (cost
Very low sum of Very low sum of
1.2 Sum insured in of inputs or harvest
insured (cost of seeds) insured (cost of seeds) 80% of the animal
relation to cost of value). Does not cover
in relation to the risk in relation to the risk resale value
risk the most extreme
(harvest value). (harvest value).
events (above 65%
loss).
No client screening or
No client screening or No client screening or exclusion. Enrolment Sold through agregators
exclusion. Requires a exclusion. Requires a through agregators, so so clients need to be
1.3 Eligibility mobile phone and mobile phone and farmer needs to be a members of a dairy
criteria access to network. access to network. client/member of a cooperative (or a
Partnership with the Partnership with the financial institution or wealthy individual
main operator main operator producers' farmer).
organization.
Farmer receives
1.4 Value-added
None None None insurance training and
services
vaccination of cows.
2. ACCESS

138
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Enrolment is
Enrolment is voluntary. Voluntary or
voluntary. No Voluntary or
No possibility to mandatory (depends on
possibility to mandatory (depends on
customize the the agregator). The
customize the the agregator). The
coverage. Option to opt agregator makes the
coverage. Option to agregator makes the
out irrelevant (product decisions in the name
opt out irrelevant decisions in the name
is for one season). of the clients, except
(product is free and of the clients, except
Simple enrolment for large individual
short term). Simple for large individual
2.1 Choice and process through a farmers who can
enrolment process with farmers who can
enrolment hotline. No customize their own
just one SMS. No customize their own
documentation needed. product. Reduced
documentation needed. product. Reduced
Provides enrolment at paperwork thanks to
Provides enrolment at paperwork thanks to the
the end of germination the partnerships with
the planting time. partnerships with
phase. No reminder agregators. Insurance
Information sent by agregators and the app.
sent to renew coverage sold at the relevant
SMS about the Insurance sold
for the following moment (before
possibility to extend throughout the year.
season. planting).
coverage period
The scratch card
provides the essential A well-designed
information about the product info sheet
A well-designed
benefits and All clients receive summarizes the
product info sheet
limitations of the information via a necessary information.
summarizes the
product. Information hotline. Information Product is complex
necessary information.
2.2 Information about the season about the season (combines index and
Financial education
and understanding outcome sent by SMS. outcome sent by SMS. traditional insurance
goes through veterinary
Limited understanding The contact via a methods). Need to
doctors and
of the product features hotline allows clients to educate the distribution
producers'organizations
by the clients. Field ask questions. channels, who will
(trusted partners).
promoters educate the educate the farmers
clients about themselves
insurance.
Upfront payment
needed. Premium
financing options
Payment via mobile Upfront payment
depend on the
money account in one needed. Premium
agregator, but is often
2.3 Premium Free (paid by the seed instalment. No financing options
possible to deduct the
payment method manufacturer) premium financing depend on the
premium from the
option offered, no agregator, but it is often
payments made by the
subsidy. linked to a credit.
dairy cooperative to the
farmer for the milk
collection.
Insurance bundled
Insurance bundled with
with the seeds, which Enrolment through Enrolment through
the seeds, which are
are sold in agrovet agregators who have agregators who have
sold in agrovet shops in
2.4 Proximity shops in rural towns. already established already established
rural towns. Enrolment
Enrolment via SMS, at close relationships with close relationships with
via phone, at the farm,
the farm, no need to the farmers. the farmers.
no need to travel.
travel.
3. COST
Ratio risk coverage =
Ratio risk coverage = 0 Ratio risk coverage = 8 Premium level is
3.1 Premium in 1000
Ratio to all benefits = Ratio to all benefits = confidential, depends
relation to benefit Ratio to all benefits =
0 11 on the crop insured.
875
3.2 Premium in Premium level is Premium to insure one
Less than 0.1% of
relation to client Product is free confidential, depends cow = 2.4% of the
monthly income.
income6 on the crop insured. annual average income
Almost zero Almost zero
Vaccination fees (1000
transaction cost transaction cost No need to travel for
KES). No need to travel
3.3 Other costs involved for the client, involved for the client, enrolment or claims.
for enrolment or
except the cost of one except the cost of one No additional fee.
claims.
SMS. phone call.

6
Estimated monthly income: 12 190KES. The author used the 2016 GNI per capita in local currency, from the
World Bank.
[Link] accessed on 28 June 2018.

139
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

High level of human

interactions (veterinary
Presence of a loss doctor needed for
Product in pilot phase,
adjuster needed for enrolment and claim
so the registration
some perils, but a good settlement), which will
process is costly
3.4 Cost structure mix with index likely maintain the
Product is free (requires a call center).
and controls techniques to minimize premium at a high
Possibility to go fully
labour costs. Requires level. App to support
by SMS in the next
large scale to take costs and ease the enrolment
season
down. process. Requires large
scale to take costs
down.
4. EXPERIENCE
For perils covered by
the index, no need to
No need to file a
4.1 Claim No need to file a claim, file a claim. But for the Needs human expertise
claim, automatic
processing automatic index-based other perils, the farmer to validate the cause of
index-based claim
procedures claim settlement needs to call the death of the animal.
settlement
company to have the
visit of a loss adjuster.
Claim is settled within Claim is settled within
4.2 Claim 7 days after the end of 7 days after the end of Claim is usually paid Claim is usually paid
processing time coverage period. Claim coverage period. Claim directly to the agregator directly to the agregator
and/or quality of paid directly on the paid directly on the (to write-off a loan for (to write-off a loan for
service provided mobile money account mobile money account example). example).
of the policy-holder. of the policy-holder.
Dematerialized policy Dematerialized policy
issuance (SMS) right issuance (SMS) right
4.3 Policy
after registration. after registration. Each client has a paper Each client has a paper
administration and
Basic information Basic information policy. policy.
tangibility
given (scratch card and given (scratch card and
SMS). SMS).
The entry point for The entry point for
Call centre. Network Call centre. Network of complaints and complaints and
4.4 Customer care
of field promoters. field promoters. feedback is the feedback is the
aggregator aggregator

Any scoring based on qualitative criteria has its part of subjectivity, which means that another
person could have given different scores. Nevertheless, the results obtained with the PACE tool
are in line with the analysis based on the 4As framework. The RPG and the top-up products
perform very well on the Cost dimension, and to a lesser extent Access and Experience
dimensions. On the other hand, they perform poorly on the Product dimension. Indeed, the RPG
and top-up are free or very cheap; they have very easy enrollment procedures and smooth claim
processes. Nevertheless, they only cover a small fraction of the risk exposure of the farmer.

On the other hand, the Livestock and Hybrid cover products score much better on the Product
dimension as they offer a more comprehensive cover to farmers. The Access and Experience
dimensions receive good scores, as the relation with the aggregators eases the Enrolment and
Claims processes, even if they are not as smooth as the total mobile experience offered by the
RPG and the Top-up products. Finally, these products are more expensive, and this can be a
significant barrier to the inclusivity of these products.

140
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

To summarize, the RPG and the top-up products propose a low-cost/low-coverage value
proposition, whereas the Hybrid cover and the Livestock products propose a higher-cost/high-
coverage value proposition. All products take advantage of existing distribution channels
(mobile operators, aggregators) to smooth enrolment and claims processes. Whether there is a
possibility to “graduate” from the low-cost product to the higher-coverage products is a
question to explore and for which there is no evidence so far.

6. Discussion

[Link] two approaches: the 4As framework and the PACE tool

The PACE tool was designed by practitioners with the explicit purpose to improve the value
proposition for microinsurance products in the market. On the other side, the 4As framework is
a theoretical analytical framework built by academics. The 4As framework is not intended to
rate any product or service. It is useful to understand and analyze marketing strategies deployed
to reach the BoP segment. Nevertheless, it is interesting to see how these two approaches are
somehow overlapping, and some dimensions of the PACE tool can be seen as a practical
application of the 4As framework to microinsurance.

Table 18 - PACE tool sub-dimensions correspondence with the 4As framework

PACE tool 4As framework

1. PRODUCT
1.1 Coverage, service quality, exclusions and waiting periods Acceptability
1.2 Sum insured in relation to cost of risk Acceptability
1.3 Eligibility criteria Not addressed in the 4As framework
1.4 Value-added services Acceptability
2. ACCESS
2.1 Choice and enrolment Availability (partially)
2.2 Information and understanding Awareness
2.3 Premium payment method Affordability
2.4 Proximity Availability
3. COST
3.1 Premium in relation to benefit Affordability
3.2 Premium in relation to client income Affordability
3.3 Other costs Affordability
3.4 Cost structure and controls Affordability
4. EXPERIENCE
4.1 Claim processing procedures Availability/Acceptability
4.2 Claim processing time and/or quality of service provided Acceptability
4.3 Policy administration and tangibility Acceptability
4.4 Customer care Acceptability (partially)
Source: author

141
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

The focus of each approach is however different. The PACE tool is a client-centered approach.
This explains why some sub-dimensions, like the 1.3 eligibility criteria sub-dimension, are
typically looking at the social performance of the product, whether it is inclusive or not. The
notion of “social performance” is well established in the microfinance sector since the early
2000s (Lapenu et al., 2004). The social performance concept includes both the results on clients,
and the management processes implemented by a lender (Bauwin, 2019). The French NGO
CERISE developed a tool to monitor and measure this social performance: the social
performance index. Currently in its fourth version, this tool (SPI4) evaluates the social
performance of a financial institution under six dimensions, in line with the Universal Standards
for Social Performance Management defined by the Social Performance Task Force (Wardle,
2017). These six standards are: define and monitor social goals; ensure board, management and
employee commitment to social goals; design products, service and delivery channels that meet
clients’ needs and preferences; treat clients responsibly; treat employees responsibly; balance
financial and social performance.

The methodology developed in assessing the social performance of microfinance has been
partially transferred to the PACE tool. The dimension 3 (design products, service and delivery
channels that meet clients’ needs and preferences) is present in both the PACE tool and the 4As
framework. However, unlike the 4As framework, the PACE tool addresses the standards 1 and
4 of social performance (define and monitor social goals; treat clients responsibly). For
example, the 2.1 Choice and enrolment sub-dimension of the PACE tool will assess if enrolment
is voluntary, or the option to opt-out. Such questions are irrelevant to the 4As framework, which
will focus only on the simplicity of the process (assuming simplicity will boost sales). The 4As
framework is business-focused, it aims at selling products and services for the BoP segment.
The issues of client protection and social performance are central in the PACE tool, but they
are present in the 4As framework only if they have a business implication. In other words, the
4As framework aims at analyzing marketing strategies to improve their effectiveness, whereas
the PACE tool aims at improving the value proposition of microinsurance products. The
marketing and the development aspects are equally important in the PACE tool whereas the

142
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

4As framework focuses on the marketing aspect, the development aspect being secondary and
serving the marketing objective.
Figure 16 - Overlap between the 4As framework and the PACE tool

4As PACE tool

framework
• Client value
• Marketing

Source: author

It is fair to conclude that the PACE tool is more comprehensive and more accurate than the 4As
framework to review a microinsurance scheme. It should not come as a surprise, as the 4As
framework is applicable to any type of product or service designed for the BoP, whereas the
PACE tool is an industry specific analytical grid. Chronologically, the PACE tool was created
after the emergence of the BoP literature. The publication of “The fortune at the Bottom of the
Pyramid” in 2004 received both praise and criticism, but it is clear that it has had a great
influence on development practitioners and policy-makers, as well as multination corporations
(Perrot, 2011). Anderson & Billou introduced the 4As framework in 2007 as analytical
framework to pursue the work initiated by Prahalad. The PACE tool was presented in 2011.
Though it does not explicitly refer to the 4As framework or even to the concept of Base of the
Pyramid, we can consider the PACE tool as the practical, microinsurance-specific outcome of
the research around BoP markets. The PACE tool integrates a multidimensional perspective
just like the 4As framework, but in addition to the marketing considerations, it also adds the
criteria of social performance and inclusivity.

[Link] of the case study

A first limit of the case study lies in the selection of interviewees, which may lead to bias.
Indeed, for practical reasons, the selection process for both employees and external stakeholders
of ACRE heavily relied on one ACRE staff. The researcher being based outside the country of
the field research, the first contact with the sample, the scheduling of interviewees and the field
logistics were delegated. This is not really an issue for ACRE’s employees, as the researcher

143
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

could interview all HQ staff and representatives from all functions in the field. However, when
it comes to external stakeholders, it is likely that ACRE contacted clients and partners who had
positive views towards the company. The number of people interviewed outside the company
is also relatively small. This study is no way intending to measure client or partners satisfaction,
so there is no necessity to have a representative sample. However, some issues or topics could
have been identified with a larger and more diverse sample of external stakeholders. In order to
mitigate this issue, the research placed emphasis on triangulating sources and multiplying
different perspectives in order to check the validity and exhaustiveness of the information
collected.

Other limits of the study are classical with qualitative interview-based studies. The first one is
the desirability bias: interviewees tend to say what they think the interviewer expects from
them. In this situation, the researcher was also identified as a shareholder’s representative which
may further reinforce this bias. To balance this risk, each interviewee was reminded at the
beginning of each meeting that the purpose of the interview was to inform research and not to
make a report to the shareholder. It was also clearly stated to ACRE’s management, the
interviewees and the researcher’s own employer that all interviews would be strictly
confidential, and that interviewees would never be quoted with their names or function, so that
no information could be linked to a specific person in particular. A second typical limit in
interview-based data collection is the memory bias. People interviewed in the sample may
simply forget some information or details during the interview. Nevertheless, the desk review
of company documentation, the multiplication of interviews and the possibility for interviewees
to contact the researcher after the meeting mitigate this memory bias.

6.3. Perspectives

6.3.1. Viability strategies for agricultural microinsurance

There are usually trade-offs among the four As, or among the four dimensions of the PACE
tool. When deciding on a marketing strategy, microinsurance practitioners have to make a
choice: do they provide an affordable product with a low level of protection, or do they increase
the level of protection (increasing the sum insured, covering various risks and reducing basis
risk) even it means that fewer clients will purchase their product? The hybrid cover is a good
example of how a company navigates between these constraints and adapts the design of its
product to move from one priority (affordability) to the other (protection).

144
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

However, formulating the problem as an issue of trade-off between price and protection is
oversimplifying. Even a free product like the RPG can struggle to meet its market if clients do
not perceive the value. Non-price challenges are equally important. In particular, the literature
has demonstrated the importance of trust and peer effects within the target market, and how
they can determine the uptake of insurance products. Innovative distribution channels, like
aggregators or mobile distribution, not only reduce costs for insurance providers, they also
address the issues of acceptability, availability and awareness. The same can be said about
bundling. Packaging insurance with other services and products not only reduces distribution
costs, it can also increase the perceived value of insurance and improve the overall client
experience, like when insurance unlocks the access to agricultural credit. It is even possible to
envision a situation where other actors involved in the value chain like farm input providers,
off-takers, agricultural lenders or farmers’ organizations, see their own interest in having
farmers insured. Mukherjee et al. (2017) identify several reasons why bundling can be an
interesting opportunity for insurance companies, farmers, and agricultural value chain actors.

For insurance companies:

- Increase outreach and market penetration without the need to build their own
distribution channel in rural areas
- Utilize the partner’s goodwill to convince farmers to try insurance
- Reduce cost of distribution, client education and premium collection
- Reduced anti-selection and fraud (especially for mandatory products)

For farmers :

- Access to insurance in a cost-effective way

- Easier access to credit, farm inputs, and market
- Access to a variety of services in a one-stop shop
- Ease of payment of premiums

For agricultural value chain actors:

- Reduce the risk for their own business

- Boost sales of their existing products by adding insurance as an additional benefit
- Increase loyalty
- Obtain additional revenues in the form of commissions

145
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Not only value chain actors can distribute or package insurance with their own services, they
can also be willing to share the cost of insurance premiums, rather than expecting commissions.
In India, Pepsico buys the production of insured potato farmers at a better price (Mukherjee et
al., 2017). This, and a guaranteed market output for their production make a strong value
proposition for farmers. The alignment of interests between the different players is the key to
the viability of the partnership and of the business case of agricultural insurance.

6.3.2. Perspectives for further research

This case study is based on a comprehensive description of the company’s products and
processes. It exploits such data with two analytical grids, the 4As framework and the PACE
tool, to evaluate the adequacy of ACRE’s offer with the needs and constraints of their target
clients, smallholder farmers. This company-centered approach could be complemented with
more client data. There are many methodological options to collect client data: focus groups,
satisfaction survey, socio-economic survey… If quantitative data is collected, it would also be
interesting to pair such data with data from the client data base. This would allow to segment
clients according to several variables: type of insurance policy, crop cultivated or livestock,
geographic area, client who received a payout vs client who have not, household social or
economic characteristic…

Another dimension that could be further investigated is the economic equilibrium of each type
of insurance policy. The focus of the study was placed on the adequacy of the products and
processes with clients’ needs, but any company needs to balance these considerations with
operational efficiency and economic profitability objectives. Quite often, there is a trade-off
between the client value, the premium rate and the company’s profitability. When designing
insurance contracts, companies need to consider these three dimensions and may have to make
concessions on one dimension to preserve one or the two other dimensions. Having the full
picture with all three dimensions would inform the decision-making process when practitioners
design insurance policies for smallholder farmers.

7. Conclusion

This case study of Kenya illustrates the variety of strategies that are available to provide
agricultural insurance to the BoP, both in terms of products than in the distribution. When Eling,
Pradhan and Schmit (2014) review the literature to identify factors affecting demand for
microinsurance, they consider two types of variables: client-related variables (wealth, risk-

146
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

aversion, trust, religion, financial literacy, access to informal risk sharing mechanisms, risk
exposure, age, gender), and insurance products variables (price, non-performance and basis-
risk, quality of service). Unsurprisingly, price and basis risk correlate negatively with demand,
while quality of service correlates positively with demand. Defining quality of service as a
variable raises two questions though. Is it possible to give a unique score to quality of service
when this aspect is fundamentally multidimensional? How customers can accurately evaluate
quality of service before purchasing insurance of before filing a claim? The same study offers
insights to answer the second question, as it evidences that financial literacy is positively
correlated with demand. It suggests that clients who are more able to understand an insurance
policy are also more likely to purchase it. However, the study also shows that risk aversion is
negatively correlated with uptake. For many low-income households, insurance is a “black box”
and they are not willing to spend their limited resources to try a new product which value
remains to be demonstrated.

The 4As framework and the PACE tools are useful in analyzing the multidimensional setting
that determine the product uptake and its client value. These two analytical approaches also
contribute to assess the viability of the agricultural insurance business model in integrating the
different business partners involved. The findings from this study bring nuance to the academic
consensus that price is negatively correlated with demand for microinsurance. Indeed, in the
case analyzed, an insurance policy available for free is struggling to meet its market. This study
confirms the hypothesis brought by Prahalad (2004), Anderson & Billou (2007) and Simanis &
Hart (2008), stating that price is not the only barrier preventing low-income households to have
access to certain products and services. Low-income households actually face a broad range of
constraints beyond income and cash availability. This realization has led to a number of new
marketing approaches, aiming at designing the whole client experience around the specific
demand of BoP customers, rather than simply trying to reduce costs and marginally adapting
higher-end products. In the insurance sector, it resulted in the elaboration and testing of various
initiatives aiming at assessing client value of insurance policies, including the PACE tool. Just
like the 4As framework, the PACE tool is multi-dimensional and client-centric. Unlike
Anderson & Billou’s framework though, it integrates the criterion of social performance.
Indeed, these tools come from different perspectives. The 4As framework takes place in the
BoP marketing literature stream, where academics as well as practitioners are looking for ways
for companies to expand into BoP markets and seize the opportunity to address this untapped
segment. Even if it aims at improving understanding of BoP clients, it is fundamentally

147
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

business-focused and oriented towards the growth of sales. The PACE tool was developed a
few years later by development practitioners. It builds on the multidimensional approach of the
4As frameworks, but adds a social performance and development agenda. It pursues an
objective of inclusivity (insurance products must be available to the largest number of
households) and checks that products effectively reduce risk exposure for policyholders.

Agricultural insurance as a stand-alone product has often met low demand, especially in low-
income markets. This observation has led policy-makers around the world to subsidize
insurance premiums, in order to increase insurance penetration. Indeed, the largest agricultural
insurance markets are also the ones where insurance is the most heavily subsidized.
Nevertheless, the agricultural insurance business case could also be strengthened by assessing
insurance not a separate product, but as component of a larger value proposition, for farmers
but also for other agricultural value chain actors and agricultural lenders. Involving such players
requires first to understand where their interests lie, what their own constraints are, in order to
build sustainable partnerships. The next chapter addresses this question of aligned interests and
constraints for microfinance institutions, as they can play a key role in making insurance
available for a large number of households who only have imperfect risk management solutions.

148
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

References

Anderson, J., & Billou, N. (2007). Serving the world’s poor : Innovation at the base of the
economic pyramid. Journal of Business Strategy, 28(2), 14‑21.

Antón, J. (2009). Risk management in agriculture : A holistic conceptual framework. Managing

Risk in Agriculture A Holistic Approach, Paris, OECD.

Barnett, B. J., Barrett, C. B., & Skees, J. R. (2008). Poverty traps and index-based risk transfer
products. World Development, 36(10), 1766‑1785.

Bauwin, M. (2019). La Gestion de la Performance Sociale en microfinance : Pratiques,

résultats et enjeux. ADA. [Link]
[Link]

Bernhardt, A., Steinmann, R., & McCord, M. J. (2012). Microinsurance intermediaries.

Protecting the Poor: A Microinsurance Compendium, 2, 503‑525.

Binswanger-Mkhize, H. P. (2012). Is there too much hype about index-based agricultural

insurance? Journal of Development Studies, 48(2), 187‑200.

Brau, J. C., Merrill, C., & Staking, K. B. (2011). Insurance theory and challenges facing the
development of microinsurance markets. Journal of Developmental entrepreneurship, 16(04),
411‑440.

Bryla-Tressler, D. (2011). Weather index insurance for agriculture : Guidance for development
practitioners. Agriculture and Rural Development Discussion Paper, 50.

Christiaensen, L. J., & Sarris, A. (2007). Rural Household Vulnerability and Insurance against
Commodity Risks : Evidence from the United Republic of Tanzania (Vol. 10). Food &
Agriculture Org.

Clarke, D. J. (2011). A theory of rational demand for index insurance. Department of

Economics, University of Oxford.

Cohen, M., McCord, M. J., & Sebstad, J. (2003). Reducing vulnerability : Demand for and
supply of microinsurance in East Africa. Nairobi, Microsave.

Cole, S., Giné, X., Tobacman, J., Townsend, R., Topalova, P., & Vickery, J. (2013). Barriers
to household risk management : Evidence from India. American economic journal. Applied
economics, 5(1), 104.

Cole, S., Sampson, T., & Zia, B. (2011). Prices or knowledge ? What drives demand for
financial services in emerging markets? The Journal of Finance, 66(6), 1933‑1967.

149
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Counts, A. (2011). Towards Reinventing Microfinance Through Solving the “Last Mile
Problem” : Bringing Clean Energy Solutions and Actionable Information to the Poor. Grameen
Foundation.
[Link]
aft%20-%20A_%20Counts(1).pdf

De Bock, O., & Gelade, W. (2012). The demand for microinsurance : A literature review.
Microinsurance Innovation Facility, Research Paper, 26.

Eling, M., Pradhan, S., & Schmit, J. T. (2014). The determinants of microinsurance demand.
The Geneva Papers on Risk and Insurance-Issues and Practice, 39(2), 224‑263.

Friedberg, E. (1988). L’analyse sociologique des organisations. L’Harmattan.

Garrette, B., & Karnani, A. (2010). Challenges in Marketing Socially Useful Goods to the Poor.
California Management Review, 52(4), 29‑47.

Gaurav, S., Cole, S., & Tobacman, J. (2011). Marketing complex financial products in emerging
markets : Evidence from rainfall insurance in India. Journal of marketing research, 48(SPL),
S150‑S162.

Gautam, M., Hazell, P. B. R., & Alderman, H. (1994). Rural demand for drought insurance
(Vol. 1383). World Bank Publications.

Giné, X., Townsend, R., & Vickery, J. (2008). Patterns of rainfall insurance participation in
rural India. The World Bank Economic Review, 22(3), 539‑566.

Giné, X., & Yang, D. (2009). Insurance, credit, and technology adoption : Field experimental
evidencefrom Malawi. Journal of Development Economics, 89(1), 1‑11.
[Link]

Hazell, P., Anderson, J., Balzer, N., Hastrup Clemmensen, A., Hess, U., & Rispoli, F. (2010).
The potential for scale and sustainability in weather index insurance for agriculture and rural
livelihoods. World Food Programme (WFP). Rome.

Hill, R. V., & Robles, M. (2011). Flexible insurance for heterogeneous farmers : Results from
a small scale pilot in Ethiopia. IFPRI discussion papers, 1092.

Hlady-Rispal, M. (2015). Une stratégie de recherche en gestion-L’étude de cas. Revue française

de gestion, 41(253), 251‑266.

Ito, S., & Kono, H. (2010). Why is the take‐up of microinsurance so low ? Evidence from a
health insurance scheme in India. The Developing Economies, 48(1), 74‑101.

Kalra, A., & Xing, L. (2013). Partnering for food security in emerging markets (No 1/2013;
Sigma). Swiss Reinsurance Company.

150
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Karlan, D., & Morduch, J. (2010). Access to finance. In Handbook of development economics
(Vol. 5, p. 4703‑4784). Elsevier.

Karnani, A. (2005). Misfortune at the Bottom of the Pyramid. Greener Management

International, 51, 99.

Lapenu, C., Zeller, M., Greely, M., Chao-Béroff, R., & Verhagen, K. (2004). Performances
sociales : Une raison d’être des institutions de microfinance et pourtant encore peu mesurées.
Quelques pistes. Mondes en développement, 2, 51‑68.

Matul, M., & Kelly, E. (2012). How to conduct a PACE client value assessment.
Microinsurance Innovation Facility, International Labour Organization.
[Link]

Matul, M., Tatin-Jaleran, C., & Kelly, E. (2012). Improving Client Value : Insights from India,
Kenya and the Phillipines. Protecting the Poor: A Microinsurance Compendium, 2, 300‑330.

Mukherjee, P., Pandey, M., & Prashad, P. (2017). Bundling to make insurance work (No 47;
Microinsurance Paper). International Labour Organization.

Müller, S., Ramm, G., & Steinmann, R. (2014). Agriculture, Microinsurance and Rural
Development. Microinsurance Network Thematic Paper, Luxembourg.

Norton, M., Holthaus, E., Madajewicz, M., Osgood, D. E., Peterson, N., Gebremichael, M.,
Mullally, C., & Teh, T. (2011). Investigating demand for weather index insurance :
Experimental evidence from Ethiopia. Annual Meeting Agricultural and Applied Economics
Association, Pennsylvania.

Patt, A., Peterson, N., Carter, M., Velez, M., Hess, U., & Suarez, P. (2009). Making index
insurance attractive to farmers. Mitigation and Adaptation Strategies for Global Change, 14(8),
737‑753.

Perrot, F. (2011). Les entreprises multinationales au bas de la pyramide économique : Un cadre

d’analyse des stratégies.

Plane, J.-M. (1996). Pour une approche ethnométhodologique de la petite entreprise. Actes du
Congrès international francophone de la PME, 1106‑1118.

Prahalad, C. K. (2004). The fortune at the bottom of the pyramid : Eradicating poverty through
profits. Wharton School Publishing.

Prahalad, C. K., Hammond, A., & others. (2002). Serving the world’s poor, profitably. Harvard
Business Review, 80(9), 48‑59.

Sakurai, T., & Reardon, T. (1997). Potential demand for drought insurance in Burkina Faso and
its determinants. American Journal of Agricultural Economics, 79(4), 1193‑1207.

151
 Chapter 4 - Rethinking agricultural insurance for the BoP – a Kenyan case study

Sandmark, T., Debar, J.-C., & Tatin-Jaleran, C. (2013). The emergence and development of
agriculture microinsurance. Luxembourg: Microinsurance Network.

Simanis, E. (2011). Needs needs everywhere, but not a BoP market to tap. In Next generation
business strategies for the base of the pyramid : New approached for building mutual value (p.
103‑128). FT Press.

Simanis, E. (2012). Reality check at the bottom of the pyramid. Harvard Business Review.

Simanis, E., & Hart, S. (2008). The base of the pyramid protocol : Toward next generation BoP
strategy. Cornell University, 2.

Thietart, R.-A. (2014). Méthodes de recherche en management-4ème édition. Dunod.

Vachani, S., & Smith, N. C. (2008). Socially responsible distribution : Distribution strategies
for reaching the bottom of the pyramid. California Management Review, 50(2), 52‑84.

Wardle, L. (2017). The Universal Standards for Social Performance Management

Implementation Guide. The Social Performance Task Force.
[Link]

Webb, J. W., Kistruck, G. M., Ireland, R. D., & Ketchen Jr, D. J. (2010). The entrepreneurship
process in base of the pyramid markets : The case of multinational enterprise/nongovernment
organization alliances. Entrepreneurship Theory and Practice, 34(3), 555‑581.

World Resources Institute, & International Finance Corporation. (2007). The next 4 billion :
Market size and business strategy at the base of the pyramid.

Yin, R. K. (2017). Case Study Research Design and Methods . (6th edition). Sage Publications.

152
 CHAPTER 5

MICROFINANCE INSTITUTIONS AND INSURANCE: WHERE IS THE

BUSINESS CASE?

153
Microfinance institutions and insurance: where is the business case?

1. Introduction: from microcredit to microinsurance

When the term “microcredit” was coined, it could give the impression that microcredit was just
traditional credit at a smaller scale. However, it appeared from the start that in order to serve
properly financially excluded clients, doing business as usual was not a viable option. The group
lending methodology, coupled with joint liability, as found in the Grameen 1.0 model
(Armendáriz & Morduch, 2010) is a good illustration. The size of the transaction was not the
only, and probably not the most important, barrier preventing low-income households to access
credit. Lenders were worried that in the absence of collateral, poor borrowers would not pay
back their loans. The joint liability was an innovative response to the specific issues that arise
when evaluating the creditworthiness of a low-income borrower. Acknowledging that low-
income households need a variety of financial services beyond credit, the term “microfinance”
started to replace the word “microcredit”. For the Consultative Group to Assist the Poor,
“microfinance means building financial systems that serve the poor” (CGAP, 2004). In this
context, the idea to provide insurance to the poor made its way. Churchill and Matul (2012)
define microinsurance as “the protection of low-income people against specific perils in
exchange for regular premium payments proportionate to the likelihood and cost of the risk
involved”. This definition could apply to traditional insurance, except for its specific target: the
poor. In order to reach this target, microinsurance providers need to adapt their processes and
operations, and rethink the distribution model. In the same way as microcredit is not only a
small credit, microinsurance is more than smaller amounts insured for smaller premiums.

Poor households do not only have to deal with the fact that their income is low. They are
affected by what Collins et al. call the “triple whammy” (Collins et al., 2009). Low-income
households have modest income, their cash-flows are irregular and unpredictable, and the
financial products available to them do not properly address the households’ needs. This
situation makes poor households particularly vulnerable to external shocks. Poverty and
vulnerability reinforce each other: poverty erodes resilience mechanisms, while vulnerability
increases the likelihood of poor households falling into a “poverty trap” from which they cannot
escape (Barnett et al., 2008) because their assets are depleted.

Microfinance institutions serve the same type of clients that microinsurance aims at reaching.
We have also seen that the needs for microinsurance services are important for this segment of
population. Therefore, Microfinance Institutions appear to be ideal distribution channels for

154
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

microinsurance products. Nevertheless, evidence shows that only a fraction of MFIs offers
microinsurance products, and when they do, they offer only a limited range of products. In
particular, even when MFIs serve smallholder farmers, agricultural insurance remains scarce.
Despite the need of risk management solutions of their clients, many MFIs do not propose
microinsurance. How to explain this paradox?

This chapter will first examine theoretical arguments in favour of the distribution of
microinsurance products by MFIs. It will then review the different organizational arrangements
available for MFIs willing to offer insurance to their clients, and the challenges they face. The
specific case of agricultural credit and insurance will be further analysed. Finally, we will see
how these elements drawn from the literature compare with the results of a survey answered by
36 MFIs, and discuss the conditions under which MFIs can successfully offer insurance to their
clients.

2. Offering microinsurance, a non-straightforward path for MFIs

[Link]’ rationale to include insurance in their service offering

Mitigating credit risk

External shocks like death, sickness or natural catastrophes, affect low-income households: they
reduce their incomes, deplete their savings and assets, and in some cases may undermine their
long-term livelihoods. For example, if the breadwinner in a family gets sick, he/she will not be
able to work and bring money home. The family will draw on savings in order to pay for medical
expenses or simply to maintain consumption and day-to-day necessities. In some cases, the
household will sell productive assets, like livestock, in order to get cash rapidly. This will affect
the future income of the household. If a member of this household has a loan with an MFI, the
probability that he/she will not be able to repay on time increases significantly. The literature
shows that external shocks play a significant role in credit defaults (Bouquet et al., 2007;
Schicks, 2010). An adverse shock causing a drop of income can push a borrower into over-
indebtedness.

Therefore, insurance can be a useful tool for MFIs to control credit risk. To do so, MFIs would
first need to examine the reasons for default, and in particular the shocks that are more likely to
deteriorate the quality of their loan portfolio.

155
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Different risks affect differently the portfolio quality of an institution. Death risk has a low
probability (borrowers are unlikely to die during the duration of the loan) and can be easily
diversified (it is very unlikely that many borrowers die at the same time). Nevertheless, in case
the risk event happens (a borrower dies), the probability of default is extremely high: the MFI
will be unable to get its money back, unless the client has a current or savings account within
the institution with sufficient funds. Furthermore, if some clients in the portfolio have an
outstanding loan much larger than average loan size, the death of one borrower may
significantly deteriorate the portfolio quality of the institution. On the other hand, for an MFI
with a large agricultural portfolio, a drought can result in a rapid increase of the portfolio at
risk. In this case, the external shock exposes the MFI to the risk of bankruptcy.

Developing the business

Depending on the organizational arrangement (see section 2.2 below), an MFI distributing
insurance products will obtain an additional income, usually in the form of a commission on
each policy sold or through underwriting gains if the insurance bears the risk itself. Compulsory
products linked with the credit are an easy way to reach scale quickly. In many institutions,
borrowers are required to take-up a compulsory life insurance policy that covers the loan
amount. This type of product is usually quite simple, standardized and cheap, in the sense that
the premium represents a small part of the interest and is therefore almost “invisible” for the
client. Credit life insurance is nowadays the second most common microinsurance product in
Africa and in the Latin America and Caribbean region (Microinsurance Network, 2016;
Microinsurance Network, 2018) both in terms of lives covered and in terms of premium
amounts.

For products based on voluntary take-up, the MFI will need to invest, in the form of marketing
campaigns, or time spent by loan officers explaining the products to the clients. Low-income
clients often have a low financial literacy and a limited experience with insurance. Therefore,
the effort and investment required to convince them to take insurance can be significantly more
important than for more well-off clients. In this case, the business development opportunity
needs to be assessed: products based on voluntary take-up often charge higher premiums,
meaning the commission received by the MFI will also be higher, but they are harder to sell.

However, leveraging on the existing relationship established by MFIs with their clients, cross-
selling opportunities exist. The new microinsurance products will benefit from the MFI’s
reputation and trust. Therefore, the acquisition and transaction costs to sell insurance policies

156
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

should remain significantly lower than setting up a specialized distribution channel of insurance
sales agents.

Selling insurance policies can also be the way for an MFI to attract new clients, especially those
who are not interested by a loan. Selling to non-clients present however a higher risk of adverse
selection(Roth et al., 2005).

Fulfilling the social mission

For MFIs having a strong social focus in their mission, linking insurance and credit can be a
way to reach their social objectives.

We have seen that low-income households are particularly vulnerable, and therefore would
benefit greatly from having access to adapted insurance products. Microinsurance improves the
resilience of poor households by protecting their income, their assets and their livelihoods.
Impact studies have also shown that even in the absence of a risk event, insurance can play a
positive role.

Rademarcher et al (2010) reviewed the impact of health microinsurance, and found that in 13
out of 15 studies, access to health microinsurance led to an increase in healthcare utilization. In
another literature review, De Bock and Ontiveros (2013) find that for most health
microinsurance schemes studied, it is possible to observe an increase in healthcare services
utilization and a reduction in out-of-the pocket expenses. They also find that health and
agricultural insurance reduce the recourse to expensive coping mechanisms, like selling
productive assets. Insurance affects positively the livelihoods of the insured in the long term.
Bertram-Huemmer and Kraenert (2015) led empirical research in Mongolia, to determine if
index-based livestock insurance helps herders’ households recover more rapidly. They find that
after the 2009-2010 winter disaster that struck Mongolia, and caused the worst livestock losses
ever recorded, households who had bought insurance and received a payment had a significantly
larger herd size in 2011, 2012 and 2013 than those who had not purchased insurance. This case
shows that a shock can have a long-lasting impact on affected households, and that the
difference between indemnified and non-insured households can be observed even several years
after the event. The J-PAL, CEGA and ATAI (2016) compiled several randomized control trials
to measure impact of index insurance. The review shows that insurance modifies farmers’
behaviour, even in the absence of a shock and of a payment. Insured farmers adopted riskier,
but more profitable production strategies. In Ghana (Karlan, Osei, Osei-Akoto, & Udry, 2014),
farmers were randomly offered free rainfall index-insurance, cash grants, both cash and

157
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

insurance, or nothing. Farmers receiving insurance increased total farming expenditure,

fertilizer use, and land cultivated. They also increased the land dedicated to maize and decreased
the production of drought resistant fruit crops like mango. Farmers receiving cash also increased
their use of fertilizer, but with less dramatic impact on other production decisions. This result
suggests that risk, not cash, is the main obstacle preventing farmers from changing their
behaviour. In Andra Pradesh and Tamil Nadu in India (Cole et al., 2017; Mobarak &
Rosenzweig, 2012), farmers benefitting from weather index insurance shifted from drought-
resistant varieties to higher-yield seeds. Similar effects were observed for area-yield insurance
products. In China, expanding access to insurance led to 22 percent increase in the production
of tobacco, the insured crop. In Mali, cooperatives that were offered crop insurance saw a 0.39
ha increase in planted land, a 15 percent increase over the comparison group.

Conclusion

From the MFIs’ point of view, there are both financial and social arguments in favour of the
provision of insurance services to their clients. Insurance can improve the quality of their
portfolio by reducing the probability of default due to an uninsured shock event. It is also a
source of additional income and a way to attract potential new clients. Insurance also has the
capacity to improve the wellbeing and resilience of its clients, allowing MFIs to pursue further
their social mission with the provision of an additional service.

[Link] set-ups for MFIs

Churchill and Roth (2006) identify four types of institutional arrangements available to MFIs
willing to offer insurance to their clients.

The partner-agent model

McCord (2006) defined this model, which is based on a partnership between an insurance
company and an MFI. In this model, MFIs play the role of an agent for the insurers. For MFIs,
this model is probably the simpler, cheaper and faster way to start offering insurance products.
By collaborating with an insurance company, MFIs limit their own risk exposure, increase their
revenues through commissions and improve the protection for its clients and the MFI itself.
From the insurer’s point of view, MFIs represent an immediate access to a large number of low-
income clients without having to deal with each client individually. This institutional
arrangement is also positive for the clients who benefit from insurance services which access
was initially limited. Collectively, they represent a large number of clients, which means that
158
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

the MFI can co-create with the insurer a product that suits their needs, and at a better rate. They
can rely on the MFI to defend their interest and enforce the guarantee.

The broker model

An MFI (or a group of MFIs) can decide to create an insurance broker. This model is a
refinement of the partner-agent model. The advantage of creating a dedicated brokerage
company is to develop insurance expertise, in order to obtain better deals with insurers. As the
broker is not tied to any specific insurer, it has the capacity to work with different companies
in order to serve better the MFI and its clients.

Self-insurance

MFIs can decide to self-insure for various reasons. They may not be satisfied with the offer of
existing insurers; they may consider that insurance is a very profitable business and want to
keep the underwriting gains for themselves or they refuse to have their clients pay for the
insurer’s operating costs. Self-insurance can be a tempting option for basic credit life insurance,
as the risks are usually not covariant and the main beneficiary is the insurance company itself.
However, self-insurance must be considered very cautiously. Even for simple products, MFIs
need to have a sufficiently large pool to diversify the risk. MFIs often lack the skills and assets
to manage efficiently insurance products. With no access to the reinsurance market, covariant
risks expose MFIs to catastrophic risks. So can poorly designed policies, or improper risk
evaluation. CARD in the Philippines and TYM in Vietnam are two MFIs that almost went
bankrupt after venturing into insurance. These two institutions had successfully piloted life
insurance products before offering more covers to their clients, without having properly
assessed the risks associated (Churchill & Roth, 2006).

Creating an insurance company

Rather than self-insuring, an MFI or group of MFIs may choose to create its own insurance
company. This model can follow the broker model, once the brokerage company has
accumulated sufficient expertise in underwriting and claim management. Compared to self-
insurance, creating an insurance company bears several advantages:

- Clear separation of the insurance and credit risks

- Accumulation of insurance expertise
- Access to reinsurance markets
- Possibility to collaborate with different distribution channels.

159
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

This was the path taken by the Réseau des Caisses Populaires in Burkina Faso. This cooperative
network initially self-insured the risk of death of their borrowers. Then it created its own life
insurance company, which offers a greater variety of products: CIF-VIE Burkina. CIF-VIE
Burkina even started to distribute its life insurance policies through other MFIs and financial
institutions, reaching an even greater number of low-income clients.

[Link] faced by MFIs willing to distribute microinsurance

Demand challenges

As we have seen in chapter 4, demand for agricultural microinsurance products from

smallholder farmers is low (see chapter 4, section 2.1). We can extend these results for all
microinsurance products, and all low-income households in general. Indeed, they all face the
same barriers: low financial literacy, especially with insurance, limited resources, imperfect
adequacy between the insurance coverage and the real risk exposure, alternative risk
management mechanisms.

Distributing insurance products to their clients requires an important investment in sales and
marketing from MFIs. Such investments can be split between the MFI and the insurance
company, as both parties are interested in achieving a high volume of sales. Nevertheless, given
the current low demand of insurance (De Bock & Gelade, 2012), the return on investment is
uncertain. In particular, MFIs may be worried to dedicate loan officers’ time and attention in
non-strategic products, instead of focusing on their core, profitable, lending business. Loan
officers themselves are likely to be reluctant to sell microinsurance products if their variable
remuneration remains based on the size and quality of their loan portfolio. Selling insurance
through MFIs requires implementing specific incentives for loan officers.

Operational issues

An MFI selling microinsurance to its clients will have to modify its operational processes. First,
as we have seen, insurance products are complex, which means that the staff will need training.
Loan officers must be able to understand the features of the insurance products (risks and sum
covered, exclusions, premium calculation…). They will also need to be able to explain these to
the clients and give a convincing sales pitch. A case study of a rainfall index insurance product
in India points that loan officers tend to offer always the same type of insurance, and not adapt
to the needs of the clientele (Giné et al., 2010). Besides the salesforce, MFIs may want to
acquire insurance expertise themselves. Even in the most common case in which they bear no
160
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

risk, MFIs need some insurance knowledge to select the most adequate products that fit their
clients’ needs, but also minimize the risk of default.

The MFI must also put in place processes for operations like enrolment, premium collection,
policy issuance, claim management and payment. New administrative tasks will have to be
performed, new forms issued and the MIS may need an update. All these adjustments can result
in operational inefficiencies. An important aspect often overlooked is the resources needed to
manage the partnership with the insurer and other stakeholders (Rendek, 2012). Before the
introduction of the products, the partners must define together the right offer and the way they
manage processes and how responsibilities and costs are shared. Once the partnership is
operational and insurance policies are being sold, efficient reporting and communication
channels are required to ensure a smooth implementation.

Risks

In the institutional arrangements involving a partnership with an insurance company, MFIs need
to be careful on the selection of their partner. Any misconduct or failure from the partner may
affect the MFI’s own reputation and the trustful relationship it has built with its clients. In
particular, if the insurance company fails to deliver its promise to indemnify policyholders in
case of a risk event, clients will lose faith in the product, and possibly distrust the MFI that sold
the product. For clients, the only entity they dealt with is the MFI, so they expect the MFI to
keep its commitment regarding the insurance they bought. A bankruptcy or fraudulent intents
to deny payouts would have devastating consequences on the MFI’s reputation. Moreover, even
if the insurer is solvent and honours the contract, the clients may feel that they are being cheated.
Exclusions or overcomplicated administrative processes to file a claim can lead clients to
believe that the insurer (and by extension the MFI) is not reliable. It is well established that
microinsurance clients give a high importance to the promptness of claim settlement. Any
administrative hurdle that unnecessarily delays payment should therefore be avoided.

In the case of an MFI bearing the risk itself, or when an MFI creates an insurance company on
its own, large losses result in large liabilities and potentially bankruptcy. In particular, MFIs
usually do not have in-house actuarial skills. Therefore, they may underestimate these risks and
be unable to pay claims.

Another challenge is the issue of basis risk for index-insurance. If a farmer suffers a shock but
the index fails to recognize the loss or underestimates it, two problems arise. The client is not
compensated for the loss; he/she will likely not renew the policy and express discontent about

161
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

the product and the MFI. Furthermore, the absence of payout puts the client in difficulty to pay
back his/her loan, exposing the MFI to a risk of default.

Finally, there is the risk of mismatch between the insurance product and the client risk exposure,
if the insurance does not protect the client against the relevant risk, or if the level of protection
is not sufficient. To continue with the example of a farmer, a farmer may buy a drought
insurance cover but be exposed to many other perils. In addition, the policy may bear a very
high deductible whereas the client has very limited resources to cope with a shock.

Conclusion

Broadening the range of services offered to their clients does not come without risks and costs
associated. MFIs willing to offer insurance to their clients must evaluate if the business case is
there. A study of a microinsurance scheme implemented by the Indian MFI ASA (Roth et al.,
2005) shows that even when loading the premium handed over to the insurer by a factor of
three, the profitability of the product is marginal. Cross-selling opportunities may not be as easy
to catch as initially thought, considering the willingness to pay for insurance of low-income
households.

3. The case of agriculture

[Link], underserved among the underserved

If MFIs target segments of the population that are traditionally under or poorly served by
financial institutions, they keep ignoring a large part of this population: farmers. Farmers,
especially smallholder farmers, are perceived by many MFIs to be more difficult to serve
because they bear specific risks and costs. Clarke and Dercon (2009) classified those challenges
into four categories: information asymmetries, transaction costs, enforcement constraints, and
ambiguity aversion. Information asymmetries have the well-known effect of restricting the
provision of credit and insurance because of moral hazard and adverse selection behaviours.
Transaction costs are high because financial services providers must deal with many small
transactions, and geographically dispersed clients. These administrative and operational costs
can represent a large proportion relatively to the size of the transaction (Rojas & Rojas, 1997).
Enforcement problems arise when borrowers can be tempted to strategically default, and
lenders have difficulty to distinguish between lack of capacity and lack of willingness to repay.
The use of collateral to induce repayment in this context is often costly and ineffective, if not

162
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

limited by regulation. Finally, ambiguity aversion refers to the preference of MFIs to serve
clients with whom they are familiar, rather than learning the complexities and specific risks of
the agricultural sector (Binswanger & Rosenzweig, 1986). Meyer et al. (2017) add a fifth
barrier, which is the systemic and covariant nature of agricultural risks, like price and
production risks. In case these risks materialize, they will affect a large number of borrowers at
the same time. Unlike larger lenders, MFIs have less possibility to diversify their risk spatially
(Barnett et al., 2008). Hence, using a longitudinal dataset of agricultural MFIs in Mexico, Peru
and Ecuador, Abrego & Guizar (2017) find that rainfall shocks have a significant negative
impact on portfolio quality. In the absence of efficient mechanisms to manage those risks, MFIs
are likely to ration credit in order to limit their exposure.

Some of these challenges exist for all MFIs clients, but they are exacerbated in the case of
farmers. Röttger (2015) shows that traditional risk and cost management mechanisms
commonly put in place by MFIs are not adapted for clients engaged in agriculture. For example,
microcredits demand quick and frequent repayments whereas farmers obtain revenues only
once they sell their crop, and would therefore prefer a bullet repayment at the end of the cycle.
Maurer (2014) denies that lending to agriculture is significantly different and riskier than
lending to other micro and small enterprises. Indeed, data showing a significantly higher risk
of agricultural loans is scarce ( Meyer, 2011). Nevertheless, only a fraction of the smallholder
farmers’ demand of financial services is currently satisfied. A 2016 report (Goldman et al.,
2016) estimates that value chain actors, informal and informal financial institutions together
provide 50 billion USD in credit to smallholder farmers in Africa, Latin America and South and
South East Asia (excluding China), when the potential demand amounts to 200 billion dollars.

[Link], the silver bullet to unlock agricultural credit markets?

Production risk pressures farmers and financial services providers, limiting the provision of
credit. In the presence of risk, farmers can be reluctant to apply for a loan because they fear
they will not be able to repay, especially when the lending institution requires a highly valued
collateral (Boucher et al., 2008). Instead, farmers prefer adopting low-risk, low-yield
production strategies. Even when they do apply for a loan, farmers may not obtain it because
lenders choose to ration agricultural credit to limit their exposure to this production risk. In the
presence of risk, agricultural credit markets are sub-optimal because bankable projects are not
financed. Insurance, if it adequately protects the farmer and the lender against production risk,
has the potential to unlock the credit supply and demand.

163
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Farmers will be encouraged to buy insurance and borrow if they anticipate that the new
investment will generate an extra income sufficient to cover the premium price. If the project
fails due to an insured event, farmers will not be pushed into over-indebtedness as the claim
payment can be used to repay the lender. One of the attractive features of credit-linked insurance
is that the premium payment can be embedded within the loan, overcoming cash-flow
constraints that farmers may face at the moment of buying insurance (Hess et al., 2016).
Financial services providers can use insurance as a form of collateral. In particular, the
insurance contract may specify that the lender is the first claimant of the payment, preventing
borrowers to strategically default when there is a payment from the insurance. Carter et al.
(2016) show with a theoretical model that the value of credit-linked insurance is the greatest in
environments where farmers have limited collateral to offer. However, insurance is not efficient
in all environments, especially when risk is not the primary barrier or when the nature of risk
makes insurance relatively ineffective to cover the risk. In favorable conditions though,
insurance is likely to boost credit markets and the adoption of improved technology, generating
a positive outcome for the farming household. The question of whether insurance adequately
covers against production risk has to do with the issue of basis risk in the case of index
insurance. Indeed, as Clarke has shown (2011), an insured farmer can be worse-off in the case
he/she faces a production loss without receiving a payment from the insurance, and having paid
the premium.

[Link] options to link credit and insurance in agriculture

Insurance can remove some of the barriers preventing farmers to get access to credit, and the
provision of credit appears as an efficient way to distribute agricultural insurance policies. The
question of how to link credit and insurance, through which organizational arrangements and
partnerships between Financial Services Providers (and in particular MFIs) and the insurance
sector must be addressed. As index insurance is the current preferred approach to insure
smallholder farmers against natural perils (Hazell et al., 2010; Hess et al., 2016), we will focus
on this particular type of agricultural insurance. Skees and Barnett (2006) as well as Meyer et
al. (2017) propose similar but not exactly identical classifications of credit and insurance
linkage mechanisms. We synthesize these two classifications and identify five different
modalities to link credit and insurance.

1) Insurance as a stand-alone product

164
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Farmers buy insurance as a separate product from their loan, and enrolment is on a voluntary
basis. Farmers can choose to insure a sum equivalent to the value of their loan, but also have
the possibility to increase this amount, if they want to cover the value of their production.
Farmers pay the premium up front. In this case, the MFI plays the role of a sales agent, insurance
being just an additional service offered to its clients. MFIs collect premiums on behalf of the
insurers and are remunerated through sales commissions.

2) Insurance embedded within credit

In this setting, insurance is automatically included in each agricultural loan. Insurance may be
compulsory to be eligible to a loan, or it may replace other forms of collateral. Farmers can still
have the option to cover more than the amount of their loan. The premium is added to the
amount borrowed, so farmers do not have to make an upfront payment.

Making insurance compulsory brings several advantages: it makes possible to reach rapidly a
critical scale, it reduces administrative and operations costs, it ensures that the whole loan
portfolio is covered and it avoids adverse selection issues. The main drawback is that insurance
increases the cost of credit. Farmers may not perceive the value of the insurance and see it
merely as an additional cost. Farmers may choose to borrow from other sources, or be excluded
from obtaining a loan because it simply is not affordable anymore. In markets where interest
rates are low, the insurance premium will weigh more significantly in the total cost of the
financing. On the contrary, when interest rates are high, the cost of insurance is less visible.
Compulsory insurance also raises the question of the insurance client value: Does the insurance
policy effectively protects the farmer against natural perils? What is the level of the basis risk?
Is the insurance price “fair”?

When insurance is embedded within credit, the question of who receives the payment must also
be raised. Usually, MFIs prefer to be the first claim recipient, in order to use the insurance
payment to write-off the loan. Any residual amount would be transferred to the client. In some
countries, regulation prevents such mechanisms and requires that the client, who paid the
premium, is the policyholder. When farmers receive directly the payment, they still have the
possibility to strategically default and use the cash received to smooth consumption. Insuring
more than the loan amount is a way to ensure that the farmers’ consumption needs will be met
in the case of a production loss, but this option increases the cost of insurance.

3) Portfolio cover

165
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Portfolio (or “meso”) insurance has gained a lot of interest in the last years. Indeed, options 1)
and 2) require managing multiple individual policies, which brings inefficiencies and costs.
When take-up is on a voluntary basis, reaching scale has proved to be an issue ( De Bock &
Gelade, 2012). They also leave farmers exposed to basis risk. The portfolio cover approach
overcomes these difficulties. With this setting, the MFI purchases index insurance for itself, to
cover its own portfolio. Indeed, as natural perils and the resulting production losses faced by
their borrowers affect portfolio quality, MFIs are willing to purchase insurance for themselves.
The link between individual loans and insurance no longer exists. Insurance is rather used as an
internal risk management tool to protect the MFI from defaults due to large and systemic risks.
Large and systemic risks put MFIs in a difficult situation on the liquidity front. Clients are likely
to withdraw their deposits to cope with the crisis when at the same time borrowers’ payment
defaults increase.

One difficulty of meso-insurance is how to pay the insurance premium. As the borrowers do
not directly benefit from the insurance cover, it is not possible to charge them the price of the
premium. The only way to pay for the premium is to “hide” the cost of insurance inside the
interest rate, which raises the question of competitiveness for the MFI. It may even be
impossible if interest rates are capped by the regulation. Meso-cover are likely to be cheaper
than individual covers though. First, there are a risk diversification and an economy of scale
effects resulting from the creation of a large risk pool. Second, the level of protection needed
by an MFI (which depends on the sum insured and the deductible) is likely smaller than the
sum of individual policies. Indeed, we can assume that MFIs are more resilient to production
losses than smallholder farmers, as they have more efficient coping mechanisms (like drawing
on their provisions or accessing external capital). Furthermore, if production losses directly
reduce farmers’ revenues, the effect is less direct for MFIs because farmers who suffer
production losses will not always default on their loans. They may use other sources of income,
sell assets or use their savings to maintain their creditworthiness towards the MFI. These
elements combined suggest that the premium for a meso-cover insurance will likely be smaller
than the sum of individual insurance policies. As Skees and Barnett (2006) put it, when
considering buying a portfolio cover, MFIs must compare the cost of the premium with the cost
of capital in a time of crisis.

Another limitation of the meso-cover is that the benefits do not trickle down directly to farmers.
In the case of a shock leading to an insurance payment, the MFI is protected but the individual
loans are not written-off and farmers will keep struggling to maintain a good repayment record

166
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

with the MFI. It is possible to imagine a situation in which an MFI recovers quite well from a
large natural risk, but in which many of its clients are over-indebted and unable to apply for a
new loan. Whether the MFI decides to use the claim payment to forgive loans is at its discretion
and asks a number of questions (who gets to have their loan forgiven, what about the risk to set
a precedent of loan forgiveness and the impact it may have on future portfolio quality…).

4) Hybrid meso and individual cover

This option takes elements from the options 2 and 3. Like in the meso-cover, the MFI purchases
one index-insurance policy for its entire loan portfolio. However, the insurance proceeds are
used to written-off individual loans, like in option 2. Skees and Barnett (2006) imagine a system
in which the MFI assesses the actual losses suffered by the borrowers to write-off part or the
entirety of the loan. In this system, the MFI decides on the allocation of the insurance payment
according to its own evaluation of the production losses in the field. The main benefit of this
model is that it mimics an indemnity-base insurance product as it compensates clients according
to the actual individual losses. It avoids basis risk for individual farmers, while taking advantage
of index-insurance markets.

This approach poses questions though. When there is no regulatory impediment, does the MFI
have the capacity and legitimacy to assess production losses? In addition, if the MFI is willing
and capable to perform these loss adjustments, the process will likely be time and resource
consuming. One of the main benefits of index insurance is that losses are automatically
calculated, which saves on loss adjustment costs and allows for prompt claim settlements. Skees
and Barnett (2006) also identify another difficulty with this option. As the insurance payment
is distributed to farmers in proportion to their individual losses, we are facing then a zero-sum
game in which each farmer has interest to declare a loss hoping to receive part of the insurance
proceeds. In the authors’ words, “any payment made to one borrower reduces the amount
available for other borrowers”.

5) Index-insurance as a form of reinsurance

In markets where traditional indemnity-based agricultural insurance is the norm, but local
insurance capacity is limited, index-insurance can be used as a form of reinsurance. Like in
options 1) and 2), farmers purchase directly or with their loan an indemnity-based insurance.
MFIs serve as a distribution channel, or in case 2) may be the policyholder. The problem is that
in many countries, local insurance companies are not willing to take agricultural risks, due to
the systemic nature of these risks (Hazell et al., 2010; Sandmark et al., 2013). Local insurance
companies can choose to buy index-insurance as a form of reinsurance if this risk-transfer
167
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

mechanism is more cost-effective. In option 5, index insurance is a tool that facilitates the
development of local agricultural insurance capacity through the transfer to international
financial markets of large and catastrophic risks.

3.4. Conclusion

Distributing insurance through MFIs makes sense for many insurance business lines, like life
and health insurance. In the case of agriculture, credit and insurance reinforce each other:
insurance allows controlling default risk of agricultural loan portfolios; it can also increase the
value proposition of credit by facilitating the adoption of new technologies and profitable
productive strategies. Different credit and insurance linkage mechanisms are available, each
with their own strengths and weaknesses. Quite often, there is a trade-off between offering a
high level of protection to farmers and reaching a large scale at a low price.

4. Evidence from the field

[Link] and sample

The results presented here are drawn from an online survey. It was sent to 58 MFIs with which
the Grameen Crédit Agricole Foundation had a business relationship. This means that all the
MFIs surveyed had an on-going loan, guarantee, or equity investment from GCAF, or had
hosted a due diligence mission in order to put in place such a transaction. The 58 MFIs were
contacted by email in September 2017. The primary purpose of the survey was not to benefit
this research, but rather to help GCAF define the best approach to support MFIs willing to
engage in insurance, by understanding their interest and their challenges. However, lessons can
be learnt from this survey about the attitude and perception of MFIs towards microinsurance.
GCAF posted the survey on the [Link] website in English and in French.
Each question appears on the screen after the other, so respondents do not have the possibility
to anticipate on the next question. 36 out of the 58 MFIs (62%) who had received the invitation
to fill the survey answered it.

Table 19 - Descriptive data of the MFIs’ sample, per region

Average Average
Average Average
Number of number of Average number of
Region loan balance portfolio size in
MFIs active PAR30 clients per loan
in EUR EUR
borrowers officer
ECA 3 18 464 827 12 013 813 2.6% 147

168
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

MENA 3 61 933 1 759 59 417 626 3.6% 284

SSEA 10 112 661 792 97 314 359 1.7% 308
SSA 20 26 302 327 7 719 087 6.2% 531

Total 36 52 607 617 37 272 768 4.4% 416

Table 20 - Descriptive data of the MFIs’ sample, per portfolio size

Average Average
Average Average
Number of number of Average number of
Region loan balance portfolio size in
MFIs active PAR30 clients per loan
in EUR EUR
borrowers officer
Tier 1 4 173 467 2 017 246 582 448 2.0% 221
Tier 2 14 57 855 677 19 492 314 2.9% 343
Tier 3 18 21 667 260 4 588 748 6.2% 517

Total 36 52 607 617 37 272 768 4,4% 416

The majority of the respondents are from Sub Saharan Africa (SSA), followed by South and
South East Asia (SSEA), the Middle East and North Africa (MENA) and Europe and Central
Asia (ECA). In the ECA region, the average loan balance per borrower is 827 euros, and the
PAR30 (portfolio at risk, with arrears of 30 days or more) represents 2.6% of the total
outstanding loan portfolio.

Half of the MFIs in the sample are Tier 3 MFIs (portfolio size below 10 million US dollars), 14
are Tier 2 (portfolio comprised between 10 and 100 million US dollars) and four are Tier 1
(portfolio above 100 million US dollars). On average, Tier 1 MFIs have a better portfolio
quality, as measured by the PAR30 (2%), they disburse bigger loans (average loan balance is
2,107 euros) and their loan officers follow a smaller number of clients (221). On the other hand
of the spectrum, Tier 3 MFIs have a lower quality of portfolio (PAR30 = 6.2%), they offer
smaller loans (average loan balance = 260 euros) and show a larger number of clients per loan
officer (517).

The MFIs in the sample are not a heterogeneous group. Some MFIs only lend to individuals
when others rely on group lending, which partly explains the difference in clients per loan
officer. The type of client they target and the local purchasing power has an impact on the
average loan balance. Some are small institutions with a regional footprint whereas others have
a national presence.

169
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Table 21 - Heterogeneity of the sample

Number of Average loan
Portfolio size in Number of clients
active balance in PAR30
EUR per loan officer
borrowers EUR
Min 3 109 79 1 005 179 0.1% 90
Median 27 330 301 9 924 171 3.6% 328
Average 52 607 617 37 272 768 4.4% 416
Max 326 892 3 844 440 187 482 13.8% 1 947
Std deviation 66 785 800 86 003 178 3.4% 364

The smaller MFI in terms of active borrowers has 3,109 active borrowers, when the larger has
326,892 active borrowers. The average loan balance varies between 79 and 3,844 euros across
the portfolio. The size of the loan portfolio goes from 1 million euros for the smallest MFI, to
440 million euros for the largest. Portfolio quality is also extremely heterogeneous, with a
PAR30 varying between 0.1% and 13.8%. The number of clients per loan officer is comprised
between 90 and 1947, reflecting both differences in productivity and credit methodologies.

GCAF explicitly targets MFIs that have a strong presence in the rural areas and dedicate a
significant part of their portfolio to lending to the agricultural sector. 13 MFIs from the sample
are part of a special facility designed to provide financing and technical assistance to rural MFIs
in Sub Saharan Africa.

We also checked if the MFIs from the sample had received a rating. We checked public ratings
given by MF-Rating, MicroRate and M-Crill, the three agencies specialized in MFIs rating.
However, only six MFIs in the sample had a recent rating (i.e. between 2016 and 2018) which
does not allow drawing conclusions.

The survey is made of closed (yes/no or multiple choices) and open questions. As expected, the
closed questions have a higher response rate than the open questions (participants had the
possibility to skip questions). Open questions however provide rich and qualitative insights on
issues like challenges limiting MFIs from entering the microinsurance market. The
questionnaire is available in Appendix 3 - Online questionnaire filled by MFIs

[Link]

Rationale and challenges to offer microinsurance

170
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

69% of the surveyed MFIs currently offer some insurance to their clients, through different
organizational arrangements and partnerships. This rate is higher in SSA (80%) than in the other
regions: 67% in ECA and MENA, and only 50% in South East Asia.

Table 22 – Answers to the question "Do you offer any insurance product(s) to your clients?
NB: This includes mandatory credit life policies offered to borrowers."
Do you offer any insurance ECA MENA SEA SSA Total
product(s) to your clients?

No 1 1 5 4 11
Yes 2 2 5 16 25

Total général 3 3 10 20 36

For the MFIs that do not offer insurance, regulation is the most often cited barrier (5 out of 11),
either because it prevents them to do so (“not allowed by central bank”), because regulation
imposes an insurance agent license to offer insurance products, or because MFIs find the
legislation not adapted (“the legislation is not developed for microinsurance in [country]”). The
other reasons given are the lack of in-house expertise, the premiums that they consider not
affordable, or in one case the absence of microinsurance offer in the market. One MFI stated
microinsurance was not a strategic issue.

For MFIs offering insurance to their clients, the two most cited reasons to do so are the reduction
of credit risk (72% of the MFIs) and the social mission (64%).

Figure 17 - Answers to the question: " What are your main reasons for offering insurance to
your clients? Please select up to 3 answers."

20
18
Number of MFIs

16
14
12
10
8
6
4
2
0
Reduce credit risk Reinforce social Obtain additional Answer clients' Other Reach new clients
mission income from demand
commissions and
fees

15 out of the 25 MFIs indicated that insurance is mandatory, and 12 make life insurance
mandatory in order to protect the client and the institution against the risk of death.

171
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Business development reasons (obtaining additional income or reaching new clients) appears
less important. For the five MFIs who stated that they offered microinsurance to answer their
clients’ demand, four out of five also cited the social mission as a reason, but only two cited the
additional income they could get from it and none cited reaching new clients. It is risky to draw
conclusions from such a small sample, but for the MFIs surveyed, answering clients’ demand
appears to be more about improving client’s wellbeing than increasing revenues or the client
base.

Only ten MFIs declared to offer voluntary insurance products. For half of them, the insurance
has no impact on their lending methodology. One MFI stated that insurance lowers the interest
rate, another MFI that it loosens the requirements to obtain a credit and another one that
insurance is accepted as a collateral. One MFI mentioned that insurance affected positively the
institution by strengthening the social mission and attracting more clients. On the other hand,
one MFI related difficulties with a livestock insurance product, which in the end harmed the
development of their livestock loan product because clients perceived it as too expensive and
complex.

20 out of the 25 MFIs offering microinsurance gave specific details about the difficulties they
faced when introducing insurance products. Ten MFIs advanced reasons around the fact that
clients do not know or have experience with insurance products. Financial illiteracy and lack
of trust towards the products are common among the clients: “a lot of clients don't believe to
insurance system, because they didn't experience in past.”; “uncertainity (sic) of whether the
insurance firm will pay in case of death”; “lack of understanding”; “weak insurance culture due
to the social context of [country]7”. Five MFIs mentioned the price as one of the main obstacles:
“clients are reluctant because they find the insurance premiums too high8”; “value for money
challenge”; “additional cost to the borrower thereby making loans expensive”. Other reasons
mentioned have to do with the extra burden for loan officers, the commissions that are too low
and the difficulty to negotiate with insurance companies. For these MFIs, the business case of
microinsurance appears to be a challenge. One MFI cited religious beliefs to be an obstacle,
and another cited regulation and staff training. Three MFIs stated they faced none or no major
difficulty.

Opportunities for development

7
Original quote in French
8
Ibid

172
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

For the 25 MFIs that already offer microinsurance products, insurance is an area for
development. 36% are already investing in the expansion of their insurance offer and 32%
intend to do so in the short term (in 1 to 3 years).

Table 23 - Answers to the question "Do you have any current or future plans to expand your
insurance offering?"
Do you have any current or future plans to Number of
expand your insurance offering? MFIs
No 3
Yes, we are already investing in it 9
Yes, in the short term (1-3 years) 8
Yes, in the long term (more than 3 years) 1
No answer 4

Total 25

For the 11 MFIs that do not have an existing insurance offer, most of them intend to introduce
microinsurance products to their clients. From the total sample, only four MFIs (11%) do not
offer insurance and do not plan to do so in the future. Their reasons advanced are regulatory
obstacles for three of them, and the absence of microinsurance product in the market for one
MFI.

Table 24 - Answers to the question "Do you have any current or future plans to start offering
insurance to your clients?"
Do you have any current or future plans to start Number of
offering insurance to your clients? MFIs
No 4
Yes, we are already investing in it 1
Yes, in the short term (1-3 years) 4
Yes, in the long term (more than 3 years) 2

Total 11

When asked about the risks that their clients would like to cover by insurance, the most
commonly cited risks are death (69% of the total sample), crop loss (47%) and health problems
(43%). 4 MFIs (11%) did not answer this question.

The high importance given to crop, but also livestock loss risk (22%) is most likely affected by
the investment policy of GCAF, as this investors targets MFIs offering services to the rural and
agricultural sectors.

173
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Figure 18 - Answers to the question "Based on what you know about your clients, what would
you say are the risks they would be most interested in covering through insurance? Please
select up to 3 answers."

30

25
Number of MFIs

20

15

10

0
Death Crop loss Health Accidents Livestock Natural Property loss Other
problems loss disasters

Regarding the support needed to introduce or expand the microinsurance offer, MFIs ask for
assistance in the product design (67% of the sample), customer education and IT systems and
technology (47% each). The high importance placed on customer education is consistent with
the difficulties expressed by the MFIs regarding the lack of financial literacy and insurance
culture of their clients. The question of technology did not pop up before in the survey but
appears to be an important topic as almost half of the sample declared to need support in this
field. 5 MFIs (14%) did not provide any answer to this question.

174
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Figure 19 - Answers to the question "In which areas do you think your organization needs
more support? Please select up to 3 answers."

30

25

20
Number of MFIs

15

10

0
Product design Customer IT systems and Claims Improving Sales Risk
education technology management client value management

5. Discussion

[Link]

The MFIs surveyed here are far from being representative of the global microfinance industry.
Instead, the sample illustrates the investment policy of GCAF. There is no MFI operating in the
Latin America and Caribbean region, or from high-income countries of Europe or North
America. Second, in terms of MFI portfolio size, Tier 2 and Tier 3 MFIs are over-represented.
Tier 1 MFIs serve a large number of clients in the world and concentrate the larger part of
microfinance portfolios, but they represent only 11% of the sample of the survey. In addition,
GCAF tends to favor MFIs operating in rural areas and serving women, whereas the
microfinance sector is mostly developed in urban areas. On the other hand, the sample also
excludes MFIs that do not meet GCAF investment criteria because they are too small, or their
performance is too weak (profitability, portfolio quality, operational self-sufficiency…). These
MFIs are at the lower end of the Tier 3 segment (sometimes referred as “Tier 4”). Rather than
representing the microfinance sector as a whole, the results of the survey are biased toward Tier
2 and large Tier 3 MFIs operating in Sub Saharan Africa and South and South East Asia.

[Link] the business case of microinsurance for MFIs

As development practitioners and policy-makers identified the possibility to protect the poor
through the provision of “micro” insurance products, it seemed natural to use the same
175
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

distribution channels used for credit and savings. MFIs would simply offer insurance as an
additional service to their clients. Experience proved that selling insurance to the poor through
MFIs was not easy. Challenges exist on both the demand and supply side. Especially, from the
MFI perspective, the profitability of distributing insurance products is limited, which is
consistent with what Roth et al (2005) found in a case study in India. The demand for such
products by their clients is scarce, and the distribution of microinsurance involves high risks
and operational costs. The MFIs survey found that they were in their majority already offering
some kind of insurance to their clients, or planning to do so in the near future. Regulation or
the absence of an insurance partner locally prevents some to venture in insurance, but almost
all of them expressed some interest to offer this kind of service to their clients. Churchill and
Roth (2006) identified four commercial objectives to drive MFIs into microinsurance:
improving loyalty, increasing profitability through cross-selling, diversifying revenue inflows,
and reaching new markets. For the MFIs in our survey, the motives however appear to be not
so much about obtaining more revenues. Rather, MFIs seem to view microinsurance as a
complementary service that strengthens their core lending business, improves risk management,
and which is in line with their social mission.

In the absence of tangible demand for microinsurance products (De Bock & Gelade, 2012),
MFIs are unlikely to engage in the distribution of a broad range of insurance products to their
clients. Instead, they tend to prioritize on products that are easy to sell (or can be even made
compulsory like credit life insurance) and that are the most likely to improve the quality of their
portfolio. Understanding the personal risks of their clients, and how those affect portfolio
quality, is key to select which products to offer to their clients.

In the case of agriculture, many MFIs limit their lending activity because they fear large and
systemic risks. Insurance can provide a solution to control for production risks and make
possible the growth of agricultural loan portfolios. Nevertheless, agricultural insurance is a
complex and expensive product and even when credit and insurance are bundled, take-up
remains a challenge.

[Link] a progression path for MFIs

The provision of credit-linked life insurance is becoming standard practice in the microfinance
sector. However, it did not result in a widespread accumulation of insurance expertise. When
asked in which area they would like to receive support, 67% of the MFIs in the sample cite

176
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

product design. Even for simple products like credit-life insurance, improvements can be made
in the credit design in order to improve the protection of clients, by extending coverage when
there is no outstanding loan, including other family members or increasing the sum insured to
cover not only the loan but also provide a capital to the family.

Building on existing insurance products, MFIs can work to improve them and gradually extend
their insurance offering. This paper proposes a 4-step progression path:

1) Start with cheap and compulsory products

This is where most MFIs are: their loans include a mandatory life insurance product that covers
the outstanding amount of the loan in case the borrower dies. Usually, these products are cheap
(premium rate is typically below 1% of the sum insured) and is therefore “invisible” for the
client who may ignore that he/she is insured.

2) Improve compulsory products and/or propose optional additional coverages

Credit-life insurance has two virtues for an MFI: it protects the MFI against the risk of
borrower’s death and it does not need to be sold. However, from the client’s point of view, the
protection offered is limited. Indeed, in case of a borrower’s death, the family would most likely
not reimburse the loan anyway, so the insurance does not make much of a difference. Additional
guarantees can be proposed without increasing substantially the cost of the insurance. The
insurance can cover the funeral’s expenses, and an additional capital can be paid to the
policyholder’s family. Coverage can also be extended beyond the duration of the loan, in order
to avoid protection gaps. Another way worth of investigating consists of linking insurance with
savings. By doing so, MFIs make available to a larger number of clients, not only their
borrowers. It also avoids issues of protection gaps for the periods between two loans.

3) Introduce new insurance business lines

Once death insurance products are well understood by both the MFI and the clients, it is possible
to introduce other insurance business lines. At this stage, it is important for MFIs to identify
which types of risks are the most likely to provoke payment defaults and which products are
demanded by the clients. Health and accidents insurance are viewed by MFIs as the most
pressing needs (ranked 3rd and 4th respectively).

Introducing new insurance business lines will most likely require a significant investment in
marketing and client education. Indeed, in stage 1 and 2, insurance products were cheap and
often compulsory. At stage 3, the value proposition must be clear to the client to provoke the
purchasing decision.

177
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Insurance products can also be linked to savings product. A savings account can be used to save
for the premium payment for example. Capitalization insurance products can also complement
existing savings options for the clients.

4) Exploit synergies between credit and insurance in agriculture

Insurance has the capacity to help MFIs control the risk and grow their agricultural loan
portfolio. Different linkages mechanisms are available to MFIs. Despite a lot of interest from
international donors, non-compulsory agricultural insurance schemes in the world have failed
to reach scale and sustainability (Hess et al., 2016) and most of them remained at a pilot stage.
Challenges for MFIs willing to develop combined services to smallholder farmers are many.
Nevertheless, the potential benefits in terms of adoption of new technologies, improved risk
management, better income-generating opportunities indicate that it is a path worth exploring.

6. Conclusion

Results from an online survey with 36 MFIs from Asia and Africa put in perspective the
business case of microinsurance. Traditionally, development practitioners and donors sought to
leverage the presence of MFIs within low-income communities to provide them insurance
services. This paper brings the MFIs perspective in the discussion. Indeed, the delivery of
insurance products through MFIs does not come without challenges. MFIs tend to prioritize
insurance products that strengthen their core lending business, rather than multiplying products
and offering a variety of risk coverages to their clients. Agricultural insurance is probably the
most complex product to sell for MFIs. Despite technological innovations, it remains often very
expensive. The complexity of such products not only hurts sales, it also means that MFIs need
to develop significant insurance technical expertise to integrate agricultural insurance in their
offer. More fundamentally, few MFIs have built sustainable business models in agricultural
credit, which seems to be a prerequisite prior the launch of crop or livestock insurance. This
fuels a vicious circle, in which MFIs do not lend to smallholder farmers because of the absence
of insurance, and insurance is not developed because MFIs have no agricultural lending
experience.

If the MFIs in the sample identify a business opportunity, they also express the need for support
in technical areas (product design, technology, claims management) but also for customer
education. MFIs will be able to reap the benefits of microinsurance only if they manage to enroll
a significant portion of their client portfolio. To do so, this paper proposes a progression path
for MFIs, starting with simple and compulsory products, before moving to more complex and

178
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

expensive coverages. Graduating from compulsory to voluntary products implies a significant

investment for MFIs: clients’ financial education, marketing, training, update of the
management information system… Inclusive insurance policies must address these challenges,
in order to integrate financial institutions within the insurance ecosystem (see Figure 13 –
ACRE’s role in the value chain for an example of the insurance partnerships in agriculture).
Typically, financial education is a public good so private actors will be less inclined to make
the necessary investments for market development as their competitors would also benefit from
it, and the return on investment very uncertain. Moreover, insurance and banking regulations
need to be consistent, as we have seen that low interest rate make insurance premiums appear
more expensive in comparison, and therefore harder to sell. Both regulators and private actors
(insurers, lenders) need to adopt a systemic perspective rather than thinking in silos. Market
development will rely on cross-sector partnerships and comprehensive financial inclusion
policies balancing credit, savings and insurance goals.

179
References

Abrego, A., & Guizar, I. (2017). Resilience of Agricultural Microfinance Institutions to Rainfall
Shocks. In 2017 Annual Meeting, July 30-August 1, Chicago, Illinois (No. 258031).
Agricultural and Applied Economics Association

Armendáriz, B., & Morduch, J. (2010). The economics of microfinance. Cambridge, MA. MIT
press.

Barnett, B. J., Barrett, C. B., & Skees, J. R. (2008). Poverty traps and index-based risk transfer
products. World Development, 36(10), 1766‑1785.

Bertram-Huemmer, V., & Kraehnert, K. (2015). Does index insurance help households recover
from disaster. German Institute for Economic Research (DIW) Discussion Paper, 1515.

Binswanger, H. P., & Rosenzweig, M. R. (1986). Behavioural and material determinants of

production relations in agriculture. The Journal of Development Studies, 22(3), 503‑539.

Boucher, S. R., Carter, M. R., & Guirkinger, C. (2008). Risk rationing and wealth effects in
credit markets : Theory and implications for agricultural development. American Journal of
Agricultural Economics, 90(2), 409‑423.

Bouquet, E., Wampfler, B., Ralison, É., & Roesch, M. (2007). Trajectoires de crédit et
vulnérabilité des ménages ruraux : Le cas des Cecam de Madagascar. Autrepart, 4, 157‑172.

Carter, M. R., Cheng, L., & Sarris, A. (2016). Where and how index insurance can boost the
adoption of improved agricultural technologies. Journal of Development Economics, 118,
59‑71.

CGAP. (2004). Key principles of microfinance.

[Link]
[Link]

Churchill, C., & Matul, M. (2012). Protecting the poor : A microinsurance compendium:
volume II. Geneva. International Labour Organization.
[Link]

Churchill, C., & Roth, J. (2006). Microinsurance : Opportunities and pitfalls for microfinance
institutions. In Churchill, C. (Ed.). Protecting the poor: A microinsurance compendium.
Geneva. International Labour Organization.

Clarke, D. J. (2011). A theory of rational demand for index insurance. Department of

Economics, University of Oxford.

Cole, S., Giné, X., & Vickery, J. (2017). How does risk management influence production
decisions ? Evidence from a field experiment. The Review of Financial Studies, 30(6),
1935‑1970.

180
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Collins, D., Morduch, J., Rutherford, S., & Ruthven, O. (2009). Portfolios of the Poor : How
the World’s Poor Live on $2 a Day. Princeton University Press.

De Bock, Ombeline, & Gelade, W. (2012). The demand for microinsurance : A literature
review. ILO Microinsurance Innovation Facility Research Paper, 26. Geneva. International
Labour Organization.

De Bock, Ombelline, & Ontiveros, D. (2013). Literature review on the impact of

microinsurance. Microinsurance Innovation Facility, Research Paper, 35. Geneva.
International Labour Organization.

Dercon, S., & Clarke, D. (2009). Insurance, Credit and Safety Nets for the Poor in a World of
Risk. DESA Working Paper 81, United Nation, Department of Economics and Social Affairs.

Giné, X., Menand, L., Townsend, R., & Vickery, J. (2010). Microinsurance : A case study of
the Indian rainfall index insurance market. In Handbook of the Indian Economy, edited by
Chetan Ghate, 167–94. New York: Oxford University Press.
Goldman, L., Tsan, M., Dogandjieva, R., Colina, C., Daga, S., & Woolworth, V. (2016).
Inflection Point : Unlocking growth in the era of farmer finance. Initiative for Smallholder
Finance and the Rural and Agricultural Finance Learning Lab.

Hazell, P., Anderson, J., Balzer, N., Hastrup Clemmensen, A., Hess, U., & Rispoli, F. (2010).
The potential for scale and sustainability in weather index insurance for agriculture and rural
livelihoods. World Food Programme (WFP).

Hess, U., Hazell, P., & Kuhn, S. (2016). Innovations and emerging trends in agricultural
insurance. Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH.

J-PAL, CEGA, & ATAI. (2016). “Make It Rain.” Policy Bulletin. Cambridge, MA: Abdul Latif
Jameel Poverty Action Lab, Center for Effective Global Action, and Agricultural Technology
Adoption Initiative.

Karlan, D., Osei, R., Osei-Akoto, I., & Udry, C. (2014). Agricultural decisions after relaxing
credit and risk constraints. The Quarterly Journal of Economics, 129(2), 597‑652.

Karnani, A. (2005). Misfortune at the Bottom of the Pyramid. Greener Management

International, 51, 99.

Kydland, F. E., & Prescott, E. C. (1977). Rules rather than discretion : The inconsistency of
optimal plans. The journal of political Economy, 473‑491.

Lailan, T. (2011). Exploring Options to Institutionalize the Dzud Disaster Response Product in
Mongolia. World Bank, Washington, DC

181
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Lapenu, C., Zeller, M., Greely, M., Chao-Béroff, R., & Verhagen, K. (2004). Performances
sociales : Une raison d’être des institutions de microfinance et pourtant encore peu mesurées.
Quelques pistes. Mondes en développement, 2, 51–68.

Levy, R. (2005). Les doctorants CIFRE: médiateurs entre laboratoires de recherche

universitaires et entreprises. Revue d’économie industrielle, 111(1), 79‑96.

Lidsky, V., Maudet, C., Malpel, G.-P., Gerster, F., Helfter, M., Lejeune, H., & Le Theule, F.-
G. (2017). Les outils de gestion des risques en agricultures (N° 2016-M-099 / N° 16-104).
Inspection Générale des Finances / Conseil Général de l’alimentation, de l’agriculture et des
espaces ruraux. [Link]

Liesivaara, P., Meuwissen, M., & Myyrä, S. (2014). Designing public–private crop insurance
in Finland. The Third International Agricultural Risk, Finance, and Insurance Conference.
June 22-24.2014 Zürich, Switzerland.

Linder, S. H. (1999). Coming to terms with the public-private partnership : A grammar of

multiple meanings. American behavioral scientist, 43(1), 35‑51.

Linnerooth-Bayer, J., & Mechler, R. (2007). Disaster safety nets for developing countries :
Extending public–private partnerships. Environmental Hazards, 7(1), 54‑61.
[Link]

Loi de modernisation de l’agriculture et de la pêche, Pub. L. No. n° 2010-874 (2010).

Mahul, O. (1998). Vers une redéfinition du rôle de l’assurance agricole dans la gestion des
risques sur récoltes. Cahiers dEconomie et Sociologie Rurales, 49, 34‑58.

Mahul, O. (2001). Managing catastrophic risk through insurance and securitization. American
Journal of Agricultural Economics, 83(3), 656‑661.

Mahul, O., & Skees, J. R. (2007). Managing agricultural risk at the country level : The case of
index-based livestock insurance in Mongolia. World Bank Policy Research Working Paper,
4325.

Mahul, O., & Stutley, C. J. (2010). Government Support to Agricultural Insurance : Challenges
and Options for developping countries -. World Bank.
[Link]
agricultural-insurance-challenges-and-options-for-developing-countries

Mahul, O., Verma, N., & Clarke, D. J. (2012). Improving farmers’ access to agricultural
insurance in India. Washington. The World Bank.

Matul, M., Tatin-Jaleran, C., & Kelly, E. (2012). Improving Client Value : Insights from India,
Kenya and the Phillipines. Protecting the Poor: A Microinsurance Compendium, 2, 300‑330.

182
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Matul, Michal, & Kelly, E. (2012). How to conduct a PACE client value assessment.
Microinsurance Innovation Facility, International Labour Organization.
[Link]

Maurer, K. (2014). Where is the risk ? Is agricultural banking really more difficult than other
sectors? In Finance for Food (p. 139‑165). Springer.

McCord, M. J. (2006). The partner-agent model : Challenges and opportunities. In Protecting

the poor. A microinsurance compendium (p. 357). ILO Publications Geneva.

Ménard, C. (2004). Gestions des risques climatiques en agriculture-Engager une nouvelle

dynamique. Assemblée Nationale, Paris.

Meyer, R., Hazell, P., & Varangis, P. (2017). Unlocking Smallholder Credit. Washington DC.
World Bank Group.
[Link]
credit-does-credit-linked-agricultural-insurance-work

Meyer, R. L. (2011). A Review of Subsidies and Investments to Support Agricultural Credit

Markets in Developing Countries. CABFIN.
[Link]

Microinsurance Network. (2016). The Landscape of Microinsurance in Africa 2015.

Luxembourg.

Microinsurance Network. (2018). The landscape of microinsurance in Latin America and the
Caribbean 2017. Luxembourg.

Mobarak, A. M., & Rosenzweig, M. R. (2012). Selling formal insurance to the informally
insured. SSRN Scholarly Paper ID 2009528. Rochester, NY: Social Science Research Network.
[Link]

Radermacher, R., Ashok, S., Zabel, K., & Dror, I. (2010). What do we know about the impact
of microinsurance? Microinsurance Academy, Presentation, 3ie Seminar Series, New Delhi.

Rendek, K. (2012). Managing microinsurance partnerships. Microinsurance Paper; no. 15.

Geneva: International Labour Organization.

Rojas, M., & Rojas, L. A. (1997). Transaction costs in Mexico’s preferential credit.
Development Policy Review, 15(1), 23‑46.

Roth, J., Churchill, C., Ramm, G., & Namerta, G. (2005). Microinsurance and microfinance
institutions : Evidence from India. CGAP Working Group on Microinsurance, Good and Bad
Practices.

Röttger, D. (2015). Agricultural finance for smallholder farmers : Rethinking traditional

microfinance risk and cost management approaches (Vol. 11). Columbia University Press.

183
 Chapter 5 - Microfinance institutions and insurance: where is the business case?

Sandmark, T., Debar, J.-C., & Tatin-Jaleran, C. (2013). The emergence and development of
agriculture microinsurance. Luxembourg: Microinsurance Network.

Schicks, J. (2010). Microfinance Over-Indebtedness : Understanding its drivers and challenging

the common myths. Bruxelles: Centre Emilee Bergheim, Solvay School of Business, CEB
Working Paper, 10/048.

Skees, J. R., & Barnett, B. J. (2006). Enhancing microfinance using index-based risk-transfer
products. Agricultural Finance Review, 66(2), 235–250.

184
CONCLUSION

185
Conclusion

This work analyzed the challenges hindering the growth of agricultural insurance, the
disruptions that are affecting its business model and the public policies that can be implemented
to support its development. The thesis is based on the assumption that agricultural insurance
markets cannot develop beyond a niche segment without changing deeply its business model,
and with no public support. Therefore, the research is structured around three research
hypotheses:

(H1) For governments willing to reduce farmers’ income volatility, agricultural insurance can
be a substitute to disaster assistance programs.

(H2) Offering agricultural insurance to smallholder farmers requires a specific design of

product features and business operations.

(H3) There is a business case for microfinance institutions to offer insurance to their clients

The research placed a strong emphasis on smallholder farmers living in developing countries,
as they are the most likely to be excluded from agricultural insurance markets, while at the same
time they are the more vulnerable to external shocks. The findings in this thesis are of interest
for corporate and public decision-makers involved in agricultural insurance, and call for
renewed partnerships between the public and the private sectors.

Chapter 2 is a reminder that government interventions in favor of agricultural insurance go

beyond premium subsidies. Policy-makers can activate a variety of instruments, depending on
the most pressing challenges in a specific market. An enabling legal and regulatory
environment, as well as the availability of public goods such as data or financial education are
pre-requisites for the introduction of agricultural insurance contracts. Subsidies and risk-sharing
arrangements between public and private actors can further support market development. Too
often, the debate around agricultural insurance is centered on premium subsidies and whether
this is an optimal use of public resources. Phrasing the debate this way leaves out of the scope
the fact that an agricultural insurance policy may pursue other goals than just “developing”
agricultural insurance markets. Policy-makers are also interested in improving farmers’
resilience to production shocks, in extending access to risk mitigation mechanisms to the most
vulnerable farmers and ensure consumers protection. Therefore, policy-makers will use other
instruments than simple premium subsidies to pursue these various goals. The definition of

186
 Conclusion

political objectives as well as the growth stage of the local market will determine the right
agricultural insurance policy mix.

The research hypothesis (H1) was addressed in chapter 3, and is based on a case study (C1) of
drought risk insurance for grasslands in France. In this case, we demonstrate that an index-
based insurance can be a viable alternative to a public disaster fund to compensate farmers for
this risk. The level of protection of farmers will be enhanced while maintaining a similar level
of public spending, the insurance premium subsidy replacing the public contribution to the
disaster fund. Chapter 3 illustrates the importance of consistency within this policy mix. In
particular, disaster relief programs and agricultural insurance should not compete and
undermine each other. Both instruments can coexist if they handle different risk layers or types
of farmers. In this case, they should adopt a common base for loss measurement in order to
avoid coverage gaps.

In Chapter 4, the second research hypothesis (H2) is tested in an agricultural insurance scheme
in Kenya (C2), specifically targeting smallholder farmers. The chapter shows that doing
business-as-usual is not an option when targeting the Base of the Pyramid segment. In order to
reach its market, ACRE Africa has brought substantial innovations to its products and its
delivery channels. Chapter 4 outlines the importance to examine agricultural insurance under
several dimensions and adopt a client-centric perspective. Price is not the only barrier limiting
farmers’ access to agricultural insurance. In accordance with the findings of chapters 2 and 3,
it invites policy-makers to consider other levers of action than just premium subsidies. The case
study of the Kenyan agricultural insurance program also shows the importance of considering
not only the product characteristics (price and coverage), but also the processes surrounding it
and the whole client experience in general. It also stresses the importance to evaluate client
value in order to improve products and extend coverage to a larger number of clients. Though
the case study focuses on smallholder farmers, the methodological approach and lessons are
applicable to a larger number of farmers.

Finally, the chapter 5 addresses the third research hypothesis (H3) thanks to a survey of 36
microfinance institutions operating in Africa and Asia (S3). It examines the roles of financial
institutions, for their role as potential distributors of insurance, but also under the perspective
of synergies between credit and insurance. The survey illustrates the necessity to make the proof
of the business case of insurance to lending institutions. The “bancassurance” model, offering
a one-stop-shop for financial and insurance services, is popular in France, but few other
countries have exploited the synergies between banking and insurance at that level. Chapter 5
187
 Conclusion

shows that for the MFIs in the sample, the provision of insurance services is primarily seen as
a way to strengthen their core lending activity. Furthermore, the delivery of other products
beyond credit can seem appealing at first as MFIs already have the client base, but the
broadening of the offer involves significant investments and is associated with risks, that may
endanger their core business. This might explain why so few MFIs in the world are distributing
insurance products.

This research provides elements to conceive an agricultural insurance development strategy. It

also bears its own limits. In chapter 3, the grasslands production index and the disaster fund
data were only available for 13 years, which is a short period of time for this kind of insurance
products. It includes several bad years, and one exceptional drought (in 2003) which is a good
thing when trying to estimate the probable maximum loss. It remains a small sample though
and each year will likely improve the relevance of this analysis. The survey in chapter 5 faces
the same kind of limit, with only 36 MFIs. It is a very qualitative sample though, as survey
respondents were CEOs or top managers of their institutions. Still, this sample does not claim
to be representative of the microfinance industry. Due to limitations in time and resources, it
was not possible to conduct additional, qualitative interviews with these institutions. It is clearly
an interesting path to explore, in order to assess the barriers and reservations of microfinance
institution towards insurance.

This thesis would benefit from additional research taking into account the expectations of other
actors in the agricultural value chains. Chapter 4 already mentioned the potential involvement
of farm input suppliers. Other stakeholders such as off-takers and the food industry in general
can also contribute to extend coverage to a larger number of farmers and support agricultural
insurance markets development. More research is also needed to further explore
complementarities between public and private actors, especially to design new risk-sharing
agreements, and test new approaches of “risk layering”, to see more practically what levels of
risks are bearable by private actors and when public intervention is actually required. The issue
of governance of public-private partnerships is also a question that is worth investigating
further, in order to not only involve the insurers and governments, but all stakeholders who
have an interest in implementing better agricultural risk management, farmers being the most
important of them.

188
BIBLIOGRAPHY

Abrego, A., & Guizar, I. (2017). Resilience of Agricultural Microfinance Institutions to Rainfall
Shocks. In 2017 Annual Meeting, July 30-August 1, Chicago, Illinois (No. 258031).
Agricultural and Applied Economics Association

Access to Insurance Initiative. (2016). How agricultural insurance can improve food security—
And why regulation matters. [Policy note].

Agroseguro. (2015). El sistema español de seguros agrarios en cifras.

[Link]
groseguro__2015-05-18__baja.pdf

Agroseguro. (2017). El seguro agrario en cifras 2016.

[Link]
URO_AGRARIO_EN_CIFRAS_2016.pdf

Anderson, J., & Billou, N. (2007). Serving the world’s poor : Innovation at the base of the
economic pyramid. Journal of Business Strategy, 28(2), 14‑21.

Antón, J. (2009). Risk management in agriculture : A holistic conceptual framework. Managing

Risk in Agriculture A Holistic Approach, Paris, OECD.

Armendáriz, B., & Morduch, J. (2010). The economics of microfinance. Cambridge, MA. MIT
press.

Babcock, B. A., & Hart, C. E. (2015). Crop Insurance : A Good Deal for Taxpayers? Iowa Ag
Review, 12(3), 1.

Babcock, B. A., & Hennessy, D. A. (1996). Input demand under yield and revenue insurance.
American journal of agricultural economics, 78(2), 416‑427.

Babusiaux, C. (2000). L’assurance récolte et la protection contre les risques en agriculture.

Rapport pour le Ministère de l’Economie des Finances et de l’Industrie et le Ministère de
l’Agriculture et de la Pêche.

Barnett, B. J., Barrett, C. B., & Skees, J. R. (2008). Poverty traps and index-based risk transfer
products. World Development, 36(10), 1766‑1785.

Barnett, B. J., & Mahul, O. (2007). Weather index insurance for agriculture and rural areas in
lower-income countries. American Journal of Agricultural Economics, 89(5), 1241‑1247.

Bassoco, L. M., Cartas, C., & Norton, R. D. (1986). Sectoral analysis of benefits of subsidized
insurance in Mexico. In Hazell, P., Pomareda, C., Valdes, A. (Eds.), Crop Insurance for
Agricultural Development:Issues and Experience. Johns Hopkins University Press, Baltimore,
pp. 126–142.

189
 Bibliography

Bauwin, M. (2019). La Gestion de la Performance Sociale en microfinance : Pratiques,

résultats et enjeux. Luxembourg. ADA.
[Link]
[Link]

Berhane, G., Clarke, D., Dercon, S., Hill, R. V., & Taffesse, A. S. (2012). Financial innovations
for social and climate resilience : Ethiopia case study. Research report.

Bernhardt, A., Steinmann, R., & McCord, M. J. (2012). Microinsurance intermediaries.

Protecting the Poor: A Microinsurance Compendium, 2, 503‑525.

Bertram-Huemmer, V., & Kraehnert, K. (2015). Does index insurance help households recover
from disaster. German Institute for Economic Research (DIW) Discussion Paper, 1515.

Bielza, M., Stroblmair, J., Gallego, J., & Conte, C. (2007). Agricultural risk management in
Europe. Conference proceeding: 101st EAAE Seminar: Management of climate risks in
agriculture, Berlin, 05‑06.

Biener, C. (2013). Pricing in microinsurance markets. World Development, 41, 132‑144.

Biener, C., & Eling, M. (2012). Insurability in microinsurance markets : An analysis of

problems and potential solutions. The Geneva Papers on Risk and Insurance-Issues and
Practice, 37(1), 77‑107.

Binswanger, H. P., & Rosenzweig, M. R. (1986). Behavioural and material determinants of

production relations in agriculture. The Journal of Development Studies, 22(3), 503‑539.

Binswanger-Mkhize, H. P. (2012). Is there too much hype about index-based agricultural

insurance? Journal of Development Studies, 48(2), 187‑200.

Bonnafous, A., & D’Arcier, B. F. (2013). The conditions of efficiency of a PPP for public
finances. 13th WCTR-13th World Conference on Transportation Research.

Boucher, S. R., Carter, M. R., & Guirkinger, C. (2008). Risk rationing and wealth effects in
credit markets : Theory and implications for agricultural development. American Journal of
Agricultural Economics, 90(2), 409‑423.

Bouquet, E., Wampfler, B., Ralison, É., & Roesch, M. (2007). Trajectoires de crédit et
vulnérabilité des ménages ruraux : Le cas des Cecam de Madagascar. Autrepart, 4, 157‑172.

Boyer, P. (2008). Assurer les calamités agricoles? Notes et études économiques, 30, 7‑32.

Brau, J. C., Merrill, C., & Staking, K. B. (2011). Insurance theory and challenges facing the
development of microinsurance markets. Journal of Developmental entrepreneurship, 16(04),
411‑440.

190
 Bibliography

Brix, L., & McKee, K. (2010). Consumer Protection Regulation in Low-Access Environments :
Opportunities to Promote Responsible Finance. Focus Note, 60.

Bryla-Tressler, D. (2011). Weather index insurance for agriculture : Guidance for development
practitioners. Agriculture and Rural Development Discussion Paper, 50.

Buchanan, J. M., & Tullock, G. (1962). The calculus of consent (Vol. 3). University of
Michigan Press Ann Arbor.

Cai, J., de Janvry, A., & Sadoulet, E. (2011). Social networks and insurance take up : Evidence
from a randomized experiment in China. ILO Microinsurance Innovation Facility Research
Paper, 8.

Canet, E., & Grassy, G. (2006). Optimizing French scientific and economic performance : The
Cifre system of public-private partnership in doctoral research and Servier’s contribution.
Medecine sciences: M/S, 22(6‑7), 664‑668.

Carter, M. (2013). Sharing the risk and the uncertainty : Public-private reinsurance partnerships
for viable agricultural insurance markets. I4 Index Insurance Innovation Initiative Brief 2013-
01 and FERDI Policy Brief B78

Carter, M., de Janvry, A., Sadoulet, E., & Sarris, A. (2014). Index-based weather insurance for
developing countries : A review of evidence and a set of propositions for up-scaling.
Development Policies working paper, 111.

Carter, M., Long, E., & Boucher, S. (2011). Public-Private Partnerships for Agricultural Risk
Management through Risk Layering. I4 Index Insurance Innovation Initiative Brief, 2011-01.

Carter, M. (2011). Innovations for managing basis risk under index insurance for small farm
agriculture. Ferdi Policy Brief, 41.

Carter, M., de Janvry, A., Sadoulet, E., & Sarris, A. (2014). Index-based weather insurance for
developing countries : A review of evidence and a set of propositions for up-scaling.
Development Policies working paper, 111.

Carter, M. R., Cheng, L., & Sarris, A. (2016). Where and how index insurance can boost the
adoption of improved agricultural technologies. Journal of Development Economics, 118,
59‑71.

CGAP. (2004). Key principles of microfinance.

[Link]
[Link]

Chantarat, S., Mude, A. G., Barrett, C. B., & Carter, M. R. (2013). Designing index‐based
livestock insurance for managing asset risk in northern Kenya. Journal of Risk and Insurance,
80(1), 205‑237.

191
 Bibliography

Christen, R. P., & Anderson, J. (2013). Segmentation of smallholder households : Meeting the
range of financial needs in agricultural families. Focus Note, 85.

Christiaensen, L. J., & Sarris, A. (2007). Rural Household Vulnerability and Insurance against
Commodity Risks : Evidence from the United Republic of Tanzania (Vol. 10). Rome. Food &
Agriculture Organization.

Churchill, C., & Matul, M. (2012). Protecting the poor : A microinsurance compendium:
volume II. Geneva. International Labour Organization.
[Link]

Churchill, C., & Roth, J. (2006). Microinsurance : Opportunities and pitfalls for microfinance
institutions. In Churchill, C. (Ed.). Protecting the poor: A microinsurance compendium.
Geneva. International Labour Organization.

Clarke, D. J. (2011). A theory of rational demand for index insurance. Department of

Economics, University of Oxford.

Clarke, D., & Wren-Lewis, L. (2013). Learning from Lemons : The Role of Government in
Index Insurance for Individuals. Paris School of Economics and FERDI Working Paper.
[Link]
Lewis/publication/265963136_Learning_from_Lemons_The_Role_of_Government_in_Index
_Insurance_for_Individuals/links/[Link]

Cohen, M., McCord, M. J., & Sebstad, J. (2003). Reducing vulnerability : Demand for and
supply of microinsurance in East Africa. Nairobi, Microsave.

Cole, S., Giné, X., Tobacman, J., Townsend, R., Topalova, P., & Vickery, J. (2013). Barriers
to household risk management : Evidence from India. American economic journal. Applied
economics, 5(1), 104.

Cole, S., Giné, X., & Vickery, J. (2017). How does risk management influence production
decisions ? Evidence from a field experiment. The Review of Financial Studies, 30(6),
1935‑1970.

Cole, S., Sampson, T., & Zia, B. (2011). Prices or knowledge ? What drives demand for
financial services in emerging markets? The Journal of Finance, 66(6), 1933‑1967.

Collin, C. (2018). Towards a working crop insurance market : An integrated strategy of

systemic risk management. Paris Nanterre University.

Collins, D., Morduch, J., Rutherford, S., & Ruthven, O. (2009). Portfolios of the Poor : How
the World’s Poor Live on $2 a Day. Princeton University Press.

Cordier, J., Erhel, A., Pindard, A., & Courleux, F. (2008). La gestion des risques en agriculture
de la théorie à la mise en oeuvre : Éléments de réflexion pour l’action publique. Notes et études
économiques, 30, 33‑71.

192
 Bibliography

Counts, A. (2011). Towards Reinventing Microfinance Through Solving the “Last Mile
Problem” : Bringing Clean Energy Solutions and Actionable Information to the Poor. Grameen
Foundation.
[Link]
aft%20-%20A_%20Counts(1).pdf

Cour des Comptes. (2012). Les aides d’urgence en agriculture.

[Link]
[Link]

Cour des Comptes. (2013). Référé 65742.

[Link]
[Link]

CTA. (2018). Index-based livestock insurance as an innovative tool against drought loss : Good
practices and impact analysis from northern Kenya.. [Link]

Cummins, J. D., & Mahul, O. (2009). Catastrophe Risk Financing Catastrophe Risk Financing
in Developing Countries (pp. 1–299). Washington, DC.

Darby, M. R., & Karni, E. (1973). Free competition and the optimal amount of fraud. The
Journal of law and economics, 16(1), 67‑88.

De Bock, Ombeline, & Gelade, W. (2012). The demand for microinsurance : A literature
review. ILO Microinsurance Innovation Facility Research Paper, 26. Geneva. International
Labour Organization.

De Bock, Ombelline, & Ontiveros, D. (2013). Literature review on the impact of

microinsurance. Microinsurance Innovation Facility, Research Paper, 35. Geneva.
International Labour Organization.

DeAngelis, K. (2013). Index-based livestock insurance : The case of Mongolia. Inside Stories
on climate compatible development, Londong. CDKN.
Dercon, S., & Clarke, D. (2009). Insurance, Credit and Safety Nets for the Poor in a World of
Risk. DESA Working Paper 81, United Nation, Department of Economics and Social Affairs.

Dercon, S., Gunning, J. W., & Zeitlin, A. (2011). The demand for insurance under limited
credibility : Evidence from Kenya. International Development Conference, DIAL.

Dercon, S., Hoddinott, J., & Woldehanna, T. (2005). Shocks and consumption in 15 Ethiopian
villages, 1999-2004. Journal of African economies, 14(4), 559.

Direction de l’initiative parlementaire et des délégations. (2016). La couverture des risques en

agriculture et les assurances agricoles (No 269; Étude de législation comparée).
[Link]

193
 Bibliography

Dror, D. M., & Piesse, D. (2014). What is microinsurance? In A practical guide to impact
assessment in microinsurance (Microinsurance Network, p. 23‑39). Chapter 2.

EC COM. (2005). Communication from the Commission to the Council on Risk and Crisis
Management in Agriculture.

Eling, M., Pradhan, S., & Schmit, J. T. (2014). The determinants of microinsurance demand.
The Geneva Papers on Risk and Insurance-Issues and Practice, 39(2), 224‑263.

EU. (2013). Overview of CAP reform 2014-2020. Agricultural Policy Perspectives in Brief, 5.

FAO, T. (2014). The state of food and agriculture : Innovation in family farming. Rome: FAO.

Garrette, B., & Karnani, A. (2010). Challenges in Marketing Socially Useful Goods to the Poor.
California Management Review, 52(4), 29–47.

Gaurav, S., Cole, S., & Tobacman, J. (2011). Marketing complex financial products in emerging
markets : Evidence from rainfall insurance in India. Journal of marketing research, 48(SPL),
S150‑S162.

Giné, X., Menand, L., Townsend, R., & Vickery, J. (2010). Microinsurance : A case study of
the Indian rainfall index insurance market. In Handbook of the Indian Economy, edited by
Chetan Ghate, 167–94. New York: Oxford University Press.
Giné, X., Townsend, R., & Vickery, J. (2008). Patterns of rainfall insurance participation in
rural India. The World Bank Economic Review, 22(3), 539‑566.

Giné, X., & Yang, D. (2009). Insurance, credit, and technology adoption : Field experimental
evidencefrom Malawi. Journal of Development Economics, 89(1), 1‑11.
[Link]

Glauber, J. (2007). Double indemnity : Crop insurance and the failure of US Agricultural
Disaster Policy. AEI Agricultural Policy Series: The 2007 Farm Bill and Beyond. American
Enterprise Institute, Washington, DC. [Link]
Glauber, J. W. (2015). Agricultural Insurance and the World Trade Organization. IFPRI
Discussion Paper 01473. Washington, D.C.: International Food Policy Research

Gohin, A. (2017). Macroeconomic impacts of the US farm policy : A second best assessment.
20th Global Economic Analysis Conference, Purdue, Indiana.

Goldman, L., Tsan, M., Dogandjieva, R., Colina, C., Daga, S., & Woolworth, V. (2016).
Inflection Point : Unlocking growth in the era of farmer finance. Initiative for Smallholder
Finance and the Rural and Agricultural Finance Learning Lab.

Goodwin, B. K., Vandeveer, M. L., & Deal, J. L. (2004). An empirical analysis of acreage
effects of participation in the federal crop insurance program. American Journal of Agricultural
Economics, 86(4), 1058‑1077.

194
 Bibliography

Hardaker, J. B., Huirne, R. B. M., Anderson, J. R., & Lien, G. (2004). coping with risk in
agriculture. Oxfordshire. CABI publishing.

Hardaker, J. Brian. (2000). Some issues in dealing with risk in agriculture. Working Paper
Series in Agricultural and Resource Economics. School of Economics, University of New
England, Armidale, NSW 2351, Australia.

Hazell, P., Anderson, J., Balzer, N., Hastrup Clemmensen, A., Hess, U., & Rispoli, F. (2010).
The potential for scale and sustainability in weather index insurance for agriculture and rural
livelihoods. World Food Programme (WFP).

Hazell, P. B. (1992). The appropriate role of agricultural insurance in developing countries.

Journal of International Development, 4(6), 567‑581.

Hazell, Peter, Sberro-Kessler, R., & Varangis, P. (2017). When and how should agricultural
insurance be subsidized ? Issues and good practices. Washington DC. World Bank.

Hess, U., Hazell, P., & Kuhn, S. (2016). Innovations and emerging trends in agricultural
insurance. Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH.

Hill, R. Vargas, Gajate-Garrido, G., Phily, C., & Dalal, A. (2014). Using subsidies for inclusive
insurance : Lessons from agriculture and health. Microinsurance Paper No. 29. International
Labour Organization.

Hill, R. V., & Robles, M. (2011). Flexible insurance for heterogeneous farmers : Results from
a small scale pilot in Ethiopia. IFPRI discussion papers, 1092.

Hill, Ruth Vargas, Robles, M., & Ceballos, F. (2013). Demand for weather hedges in India :
An empirical exploration of theoretical predictions (Vol. 1280). Intl Food Policy Res Inst.

Holzmann, R., & Jørgensen, S. (2001). Social risk management : A new conceptual framework
for social protection, and beyond. International Tax and Public Finance, 8(4), 529‑556.

J-PAL, CEGA, & ATAI. (2016). “Make It Rain.” Policy Bulletin. Cambridge, MA: Abdul Latif
Jameel Poverty Action Lab, Center for Effective Global Action, and Agricultural Technology
Adoption Initiative.

Karlan, D., Osei, R., Osei-Akoto, I., & Udry, C. (2014). Agricultural decisions after relaxing
credit and risk constraints. The Quarterly Journal of Economics, 129(2), 597‑652.

Maurer, K. (2014). Where is the risk ? Is agricultural banking really more difficult than other
sectors? In Finance for Food (p. 139‑165). Springer.

McCord, M. J. (2006). The partner-agent model : Challenges and opportunities. In Protecting

the poor. A microinsurance compendium (p. 357). ILO Publications Geneva.

195
 Bibliography

Meyer, R., Hazell, P., & Varangis, P. (2017). Unlocking Smallholder Credit. Washington DC.
World Bank Group.
[Link]
credit-does-credit-linked-agricultural-insurance-work

Meyer, R. L. (2011). A Review of Subsidies and Investments to Support Agricultural Credit

Markets in Developing Countries. CABFIN.
[Link]

Microinsurance Network. (2016). The Landscape of Microinsurance in Africa 2015.

Luxembourg.

Microinsurance Network. (2018). The landscape of microinsurance in Latin America and the
Caribbean 2017. Luxembourg.

Miranda, M. J. (1991). Area-yield crop insurance reconsidered. American Journal of

Agricultural Economics, 73(2), 233‑242.

Miranda, M. J., & Glauber, J. W. (1997). Systemic risk, reinsurance, and the failure of crop
insurance markets. American Journal of Agricultural Economics, 79(1), 206‑215.

Mobarak, A. M., & Rosenzweig, M. R. (2012). Selling formal insurance to the informally
insured. SSRN Scholarly Paper ID 2009528. Rochester, NY: Social Science Research Network.
[Link]

Mobarak, A. M., & Rosenzweig, M. R. (2013). Informal risk sharing, index insurance, and risk
taking in developing countries. The American Economic Review, 103(3), 375‑380.

Morduch, J. (2005). Smart Subsidy for Sustainable Microfinance. ADB FINANCE FOR THE
POOR, A Quarterly Newspaper of the Focal Point for Microfinance, 6(4).
[Link]
&tdet3=2&referer=MTA1MTY%3D

Mortemousque, D. (2007). Une nouvelle étape pour la diffusion de l’assurance récolte. Sénat,
50p.

Mosnier, C. (2015). Self-insurance and multi-peril grassland crop insurance : The case of
French suckler cow farms. Agricultural Finance Review, 75(4), 533‑551.

Mukherjee, P., Pandey, M., & Prashad, P. (2017). Bundling to make insurance work (No 47;
Microinsurance Paper). Geneva. International Labour Organization.

Müller, S., Ramm, G., & Steinmann, R. (2014). Agriculture, Microinsurance and Rural
Development. Microinsurance Network Thematic Paper, Luxembourg.

Norton, M., Holthaus, E., Madajewicz, M., Osgood, D. E., Peterson, N., Gebremichael, M.,
Mullally, C., & Teh, T. (2011). Investigating demand for weather index insurance :

196
 Bibliography

Experimental evidence from Ethiopia. Annual Meeting Agricultural and Applied Economics
Association, Pennsylvania.

Norton, M., Osgood, D., Madajewicz, M., Holthaus, E., Peterson, N., Diro, R., Mullally, C.,
Teh, T.-L., & Gebremichael, M. (2014). Evidence of demand for index insurance :
Experimental games and commercial transactions in Ethiopia. Journal of Development Studies,
50(5), 630‑648.

Patt, A., Peterson, N., Carter, M., Velez, M., Hess, U., & Suarez, P. (2009). Making index
insurance attractive to farmers. Mitigation and Adaptation Strategies for Global Change, 14(8),
737‑753.

Perret, B. (2010). L’évaluation des politique publiques. La découverte. Paris.

Perrot, F. (2011). Les entreprises multinationales au bas de la pyramide économique : Un cadre

d’analyse des stratégies. PhD thesis. Ecole Polytechnique X.
[Link]
Prahalad, C. K. (2004). The fortune at the bottom of the pyramid : Eradicating poverty through
profits. Wharton School Publishing.

Prahalad, C. K., Hammond, A., & others. (2002). Serving the world’s poor, profitably. Harvard
Business Review, 80(9), 48–59.

Priest, G. L. (1996). The government, the market, and the problem of catastrophic loss. Journal
of risk and Uncertainty, 12(2‑3), 219‑237.

Radermacher, R., Ashok, S., Zabel, K., & Dror, I. (2010). What do we know about the impact
of microinsurance? Microinsurance Academy, Presentation, 3ie Seminar Series, New Delhi.

Ramaswami, B. (1993). Supply response to agricultural insurance : Risk reduction and moral
hazard effects. American Journal of Agricultural Economics, 75(4), 914‑925.

Regős, G. (2015). Should governments subsidize agricultural insurances. Applied Economics

Quarterly, 61(1), 65‑92.

Rendek, K. (2012). Managing microinsurance partnerships. Microinsurance Paper; no. 15.

Geneva: International Labour Organization.

Ricome, A., Affholder, F., Gérard, F., Muller, B., Poeydebat, C., Quirion, P., & Sall, M. (2017).
Are subsidies to weather-index insurance the best use of public funds ? A bio-economic farm
model applied to the Senegalese groundnut basin. Agricultural Systems, 156, 149‑176.

Rojas, M., & Rojas, L. A. (1997). Transaction costs in Mexico’s preferential credit.
Development Policy Review, 15(1), 23‑46.

Rosenzweig, M. R., & Binswanger, H. P. (1993). Wealth, weather risk and the composition and
profitability of agricultural investments. The economic journal, 56‑78.

197
 Bibliography

Roth, J., Churchill, C., Ramm, G., & Namerta, G. (2005). Microinsurance and microfinance
institutions : Evidence from India. CGAP Working Group on Microinsurance, Good and Bad
Practices.

Röttger, D. (2015). Agricultural finance for smallholder farmers : Rethinking traditional

microfinance risk and cost management approaches (Vol. 11). Columbia University Press.

Roumiguié, A., Jacquin, A., Sigel, G., Poilvé, H., Lepoivre, B., & Hagolle, O. (2015).
Development of an index-based insurance product : Validation of a forage production index
derived from medium spatial resolution fCover time series. GIScience & Remote Sensing,
52(1), 94‑113.

Sakurai, T., & Reardon, T. (1997). Potential demand for drought insurance in Burkina Faso and
its determinants. American Journal of Agricultural Economics, 79(4), 1193‑1207.

Sandmark, T., Debar, J.-C., & Tatin-Jaleran, C. (2013). The emergence and development of
agriculture microinsurance. Luxembourg: Microinsurance Network.

Schicks, J. (2010). Microfinance Over-Indebtedness : Understanding its drivers and challenging

the common myths. Bruxelles: Centre Emilee Bergheim, Solvay School of Business, CEB
Working Paper, 10/048.

Shaik, S., Barnett, B. J., Coble, K. H., Miller, J. C., & Hanson, T. (2006). Insurability conditions
and livestock disease insurance. The Economics of Livestock Disease Insurance: Concepts,
Issues and International Case Studies, 53.

Simanis, E. (2011). Needs needs everywhere, but not a BoP market to tap. In Next generation
business strategies for the base of the pyramid : New approached for building mutual value (p.
103–128). FT Press.

Simanis, Eric. (2012). Reality check at the bottom of the pyramid. Harvard Business Review.
90(6), 2-6.

Simanis, Erik, & Hart, S. (2008). The base of the pyramid protocol : Toward next generation
BoP strategy. Cornell University, 2, 1-57.

Skees, J. R. (1999). Agricultural risk management or income enhancement. Regulation, 22, 35.

Skees, J. R., & Barnett, B. J. (1999). Conceptual and practical considerations for sharing
catastrophic/systemic risks. Review of Agricultural Economics, 21(2), 424‑441.

Skees, J. R., & Barnett, B. J. (2006). Enhancing microfinance using index-based risk-transfer
products. Agricultural Finance Review, 66(2), 235–250.

Skees, Jerry R., Barnett, B. J., & Hartell, J. (2005). Innovations in government responses to
catastrophic risk sharing for agriculture in developing countries. Paper presented at the IADB-
World Bank-CABEI workshop, "Innovations in Agricultural Production Risk Management in

198
 Bibliography

Central America: Challenges and Opportunities to Reach the Rural Poor," Antigua,
Guatemala, May 9–12.

Skees, J. R., Barnett, B. J., & Murphy, A. G. (2008). Creating insurance markets for natural
disaster risk in lower income countries : The potential role for securitization. Agricultural
Finance Review, 68(1), 151‑167.

Smith, V. (2017). The US Federal Crop Insurance Program : A Case Study in Rent-Seeking.

Smith, V. H., & Glauber, J. W. (2012). Agricultural insurance in developed countries : Where
have we been and where are we going? Applied Economic Perspectives and Policy, pps029.

Smith, V. H., & Goodwin, B. K. (1996). Crop insurance, moral hazard, and agricultural
chemical use. American Journal of Agricultural Economics, 78(2), 428‑438.

Smith, V. H., & Goodwin, B. K. (2003). An ex post evaluation of the conservation reserve,
federal crop insurance, and other government programs : Program participation and soil
erosion. Journal of Agricultural and Resource Economics, 28(1835-2016‑148818), 201‑216.

Sundt, B. (1999). An introduction to non-life insurance mathematics 4th Edition. Karlsruhe:

University of Mannheim Press.

Torchia, M., Calabrò, A., & Morner, M. (2015). Public–private partnerships in the health care
sector : A systematic review of the literature. Public Management Review, 17(2), 236‑261.

Tsujii, H. (1986). An economic analysis of rice insurance in Japan. In Hazell, P. B. R.,

Pomareda, C. and Valdes, A. (eds), Crop Insurance for Agricultural Development: Issues and
Experience. Baltimore: Johns Hopkins University Press, pp. 143- 155.

Vachani, S., & Smith, N. C. (2008). Socially responsible distribution : Distribution strategies
for reaching the bottom of the pyramid. California Management Review, 50(2), 52–84.

Van Asseldonk, M., Van der Meulen, H. A. B., Van der Meer, R. W., Silvis, H. J., & Berkhout,
P. (2016, octobre 4). Does subsidized insurance crowd out market-based insurance ? An
analysis of subsidized multi-peril crop insurance versus hail insurance in The Netherlands.
156th EAAE Seminar on Prospects for agricultural insurance in Europe, Wageningen, The
Netherlands.

Wardle, L. (2017). The Universal Standards for Social Performance Management

Implementation Guide. The Social Performance Task Force
[Link]

Wettenhall, R. (2003). The rhetoric and reality of public-private partnerships. Public

Organization Review, 3(1), 77‑107.

World Resources Institute, & International Finance Corporation. (2007). The next 4 billion :
Market size and business strategy at the base of the pyramid.

199
 Bibliography

World Bank Group (2015). Agricultural data and insurance

[Link]
Young, V. R. (2004). Premium principles. Encyclopedia of Actuarial Science.

Zott, C., Amit, R., & Massa, L. (2011). The business model : Recent developments and future
research. Journal of management, 37(4), 1019‑1042.

200
APPENDICES

Appendix 1 - Client value assessment approaches

Figure 20 - Client value approaches

Reproduced from Sandmark (2013)

201
 Appendices

Appendix 2 - PACE scoring criteria

Figure 21 - PACE scoring criteria and sub-dimension weights

202
 Appendices

Reproduced from PACE technical guide v1.0 2012

203
 Appendices

Appendix 3 - Online questionnaire filled by MFIs

204
 Appendices

205
 Appendices

206
 Appendices

207
 Appendices

208
 Appendices

209
 Appendices

210
 Appendices

211
 Appendices

212