CHAPTRE ONE : INTRODUCTION

1.1. BACKGROUND OF THE STUDY

The debate over the relationship between financial development and economic growth has been
active for many years. The basic question of the debate is whether financial development leads to
economic growth or financial development is driven by economic growth. Determining the
causal pattern between financial development and economic growth has important implications
for policy-makers’ decisions to adopt the appropriate growth and development polices.
On one hand Economist argues that financial development and economic growth are not causally
related; neither of the two has considerable effects on the other, and correlation between them is
the result of a historical peculiarity Graff (2016). On the other hand, others argue that financial
development and economic growth do have considerable causal relationship; one is the cause for
the other. According to demand- following hypothesis the lack of financial growth is
manifestation of the lack of demand for financial services. As the real side of the economy
develops, its demands for various new financial services materialize, and these are met rather
passively from the financial side. Economic growth enhances financial institution to change and
deepen, and as well as credit market grow Jung, (1986)

In contrast to the demand following hypothesis, supply leading hypothesis supposes the direction
of causation runs from financial development to economic growth. Financial development is the
one induces growth. The other view assumes direction of causation is bi direction financial
development to economic growth or economic growth to financial development Graff (2016).

As of the causation between financial development and economic growth, determining the
measure of financial depth or development is germane. The measure of financial depth also has
been controversial among economists. The common ways to measure financial depth are
monetary aggregates, such as M1 or M2, mainly because these aggregates are widely available.

1
According to Arfanuzzaman (2014) Money is remarkably significant for any economy for its
necessity and diverse characteristics. If money supply increases by expansionary monetary
policy of central bank, interest rate will go down. Consequently, cash flow and lending activities
will be prolonged. As a result, investment will gear up and gross output level is also expected to
increase defining a positive relationship between money and economic growth. From the
perspective of liquidity, money can be classified in two types. One is Narrow money (M1) and
another one is defined as broad money (M2). Besides, if monetization of the economy increases,
GDP is also expected to increase startlingly. One of the most important determinants of
economic growth is variation in the quantity of money. Therefore broad money will be taken as
Indicator of financial depth since it consist of narrow money in it.
More recently, credit to the private sector has been favored as alternative measure. The main
advantage of this indicator is that, by excluding credit to the public sector, it measures more
accurately the role of financial intermediaries in channeling funds to private sector khan and
Senhdji (2000).

Therefore, to advance economic growth developing countries, started to liberalize the financial
sector following the 1973 McKinnon and Shaw paradigm. This paradigm argues for the
liberalization of the financial sector believes that government intervention in the finance sector,
in particular through subsidized interest rates and (favored) credit allocation, not only distorts the
financial market but also depresses savings and leads to inefficient investment Alemayehu,
(2006).
Like many other developing countries Ethiopian financial system was under the control of
central government before the reform period of 1991. Particularly, during the socialist Derg
regime (1974 to 1991) all private banks were nationalized. The dominant banks during this
period were the two
Governments owned banks called Commercial Bank of Ethiopia (CBE) and Development Bank
of Ethiopia (DBE) Alemayehu, (2006).

The Commercial Bank of Ethiopia (CBE) was the outstanding provider of credit from period
1981 to 1990, shared 50%(percent) of total credit , and the Development bank of Ethiopia (

2
DBE) covered 40% (percent) , but only 10 % ( percent) of the credit were covered by the
construction bank of Ethiopia.
Proclamation No. 84/1994 allowed the private sector to engage in the banking and insurance
businesses. Although the proclamation restricted the financial sector only for Ethiopian
nationals, it marked the beginning of a new era in Ethiopia’s financial sector. Following this
proclamation the country witnessed a proliferation of private banking and insurance companies
(Alemayehu, 2006). Now 18 banks are working in the country in which 16 of them are private
and 2 public owned with 3,187 branches NBE, (2016).

According to data obtained from national bank of Ethiopia (NBE) the total credit disbursed by
the banking sector reached 263,901.6 million birr in 2015/16, the average annual rate of growth
of new credit disbursement by the banking sector over 2004/05 – 2015/16 period was 26 %.
Public banks cover 65% of the total credit disbursed in 2015/16. The commercial bank of
Ethiopia (CBE) is the dominant bank in the country as it covered 53% of the total credit
disbursed in 2015/16.
Private Banks all together granted 93,181.7 million birr or 35 % of the total fresh credit
disbursed during the same fiscal year. The outstanding loan in the banking sector reached at
263,901.9 million birr by the end of 2015/16 which was higher by 21% over the level in 2014/15.
Of the total outstanding credit, claims on public enterprises increased by 936% for the period
2008/2009 to 2015/16. In the same fashion claims on cooperatives and private sector surged by
307% and 386% respectively for the same period.

It is evident that both private and public credit has increased throughout the recent period in the
country but literature on the relationship and impact of financial depth and Economic growth in
Ethiopia is very scant. Therefore, this study will analyze the casual relationship and impact of
financial depth on economic growth.

3
 1.2. STATEMENT OF THE PROBLEM

The causal relationship between financial development and economic growth has been analyzed
at length in the literature. However, most literatures are contradicted in their findings and
conclusion over the question of the casual relationship between the financial depth and economic
growth.
Empirical studies like Chang and Steven (2006) concluded the causation between financial
development and economic growth is a uni- directional run from financial development to
economic growth which is supported by the theory of supply leading hypothesis that illustrate the
creation of financial institutions and markets increase the supply of financial services and thus
leads to real economic growth.

But other studies like (Roman, 2012), found that a contradictory finding over the pattern of
relationship between financial development and economic growth which concluded as the
causation run from economic growth to financial development. The study is consistent with the
theory of demand following hypothesis that states as the economy grow, demand is created in the
process. Thus, it is economic growth that creates demand for financial development Levine
(2005) cited in Caporale et al (2009).

In addition Ankilo and Tajudeen (2010) found different causation for different countries, uni
directional relations running from financial development to economic growth in some African
countries, while causality runs from economic growth to financial development in other
countries. The study also concluded bidirectional causality between financial development and
economic growth in Chad, South Africa, Kenya, Sierra Leone and Swaziland.

However it contradicts with Jung, (1986) that concluded that less developed countries have a
supply – leading causality pattern and developed countries have demand following hypothesis
pattern of causality. Even though empirical studies on the pattern of relationship between
financial depth and economic growth in Ethiopia are few, the studies under taken in Ethiopia also

4
contradict each other. The study by Roman (2012) and Dejene, (2016) are different in their
conclusion. Roman’s study illustrate that in Ethiopia financial development indicators have a
positive and significant relationship with economic growth in the long-run. However, in the
short-run the link is weak; the increase in financial development in the long-run has a
considerable effect causing an increase in economic growth.

In contrary to the Roman’s finding, Dejene (2006) concluded in opposite direction that implies
financial development is an essential economic growth driver in Ethiopian economy in the short
run. But the financial sector development did not reach the minimal level needed to support long
run economic growth.

Therefore, the contradiction among studies is unsettled and this is the main research gap in
which result a problem for decision making for policy makers. And to solve this problem this
study is enhanced using data of large sample and incorporating both financial depth indicators
broad Money and credit to private sector as the growth of domestic product(GDP) .

1.3. OBJECTIVES OF THE STUDY

1.3.1 General Objective
The general objective of this study is to examine the relationship between financial depth and
economic growth in Ethiopia.
1.3.2 Specific objectives
 To analyze weather the growth of financial depth causes economic growth
 To analyze the short and long run relationship between financial depth and economic
growth.

1.4. HYPOTHESIS OF THE STUDY

The study critically investigates the following research hypothesis regarding the relationship
between financial depth and economic growth in Ethiopia.
 The growth of financial depth expected to affect economic growth positively.
 By contrast Economic growth and financial depth expected to have bidirectional
relationship.

5
 1.5. SIGNIFICANCE OF THE STUDY

The findings of this study will help government and the monetary authorities to see the
effectiveness of monetary policy in the management of the Ethiopian economy in terms of
money demand and supply which have a positive economic growth. This research work
further serves as a guide and provides insight for future research on the topic and related field
for academia’s and policy makers who are interested on the topic.

1.6. LIMITATIONS OF THE STUDY

The availability of data may affect estimation technique due of lack of comprehensive data for
previous years. But attempt is made to solve this problem using different source of data
estimation technique.

1.7. SCOPE OF THE STUDY

The study is limited in scope with regard to the issue of examining the relation between bank
credit and growth as well as impact assessment. The study covers the period between the
years 1970-2016 G.C. The selection of the period only depends on the availability of data.

1.8. ORGANIZATION OF THE THESIS

This paper is organized into five chapters, following the introduction in chapter one, chapter
two present literature review. Chapter three discusses Reaserch methodology employed.
Chapter four presents’ empirical results, interpretation and discussion of the results and
finally chapter five provides conclusion and policy implications based on the findings.

6
 THAPTER TWO : LITRATURE

2.1. THEORETICAL LITERATURE

A finance is an arrangement of monetary institutions and markets trade in a variety of financial
instruments which are occupied in currency diffusion activities and the provision of loans and
credit facilities. Financial institutions and markets dwell in a key place in the economy as
intermediaries in channelling savings and other funds to borrowers and investors. In doing this,
one of their main roles is to settle the different requirements of savers and borrowers, thereby
facilitating a higher level of saving and investment in the economy than would otherwise be the
case (Levine, 2000) cited in Roman(2012).

Levine (1997) breaks the functions of financial instutions in to five (1) produce information
about possible investments and allocate capital (2) monitor investments and exert corporate
governance after providing finance (3) facilitate the trading, diversification and management of
risk (4) mobilize savings (5) facilitate the exchange of goods and services.

2.1.2 INDICATORS OF FINANCIAL DEVELOPMENT

The development of the financial sector can be measured using different kinds of indicators. The
most commonly used financial development indicators include:
1. Liquid Liabilities to GDP: major indicators to measure the size, relative to the economy, of
financial intermediaries. It consists of currency plus demand and interest bearing liabilities of
banks and other financial intermediaries divided by GDP. It is the broadest available indicator of
financial intermediation (Beck & Demirguc-kunt, 2009).
2. Private Credit to GDP: credit issued to the private sector by banks and other financial
intermediaries divided by GDP. It measures the activities of financial intermediaries by
channelling savings to investors. Countries with higher levels of private credit to GDP have been
shown to grow faster ( Demirguc-kunt and Levine 2008).
3. Commercial-Central bank: the ratio of commercial bank assets to the sum of commercial bank
and central bank assets.

2.1.2. THEORIES OF FINANCIAL DEVELOPMENT

There are many theoretical literatures about the relationship between financial development and
economic growth. The debate on whether the causal relationship runs from financial

7
development to economic growth on the one hand and economic growth to financial
development on the other hand is far from settled.

Demand following hypothesis which states that as the economy grow demand is created in the
process. It is economic growth that creates demand for financial development. Increasing
demand for financial services might lead to an expansion in the financial sector as the economy
grows Patrick, (1966). This hypothesis is shared both by Robinson J, 1952 and Lucas, (1988).
Thus, according to the Demand following hypothesis economic growth is a causal factor for
financial development.
According to supply-leading and demand-following hypothesis by Patrick (1966), the supply-
leading hypothesis argues a causal relationship from financial development to economic growth,
which means the creation of financial institutions and markets increase the supply of financial
services and thus leads to real economic growth. The same line of argument is followed by King
and Levine, (1993) Greenwood and Jovanic, (1990), Demirguc-Kunt and Levine, (2008)

2.1.3. Theories of Economic growth

The classical economists of the 18th and 19th centuries have developed different theories
regarding how and which factor of production will generate economic growth. The classical
economists believe that saving in the economy is a key factor, which is used for capital
formation. Growth in domestic product is an important determent of welfare and wellbeing of the
people in each country. Therefore, understanding and determining the primary factors which
affect the economic growth is a very interesting aspect of scholarly studies.
Modern growth theory identifies two specific channels through which the financial sector might
affect long-term growth: first through its impact on capital accumulation (including human as
well as physical capital) and second through the rate of technological progress Lucas,(1988),
Romer, (1991) and Helpman et al,(1991). These financial sector effects, nevertheless, occur from
the intermediation role of the financial institutions, which enable the financial sector to mobilize
savings for investment, facilitate and promote inflows of foreign capital such as foreign direct
investment (FDI) portfolio investments and bonds, and remittances. According to the classical
economists’ capital flow from low return to capital location to high return to capital locations.
The flow of the capital needs financial institutions as a channel. Thus, the financial institutions

8
optimize the allocation of capital between contending issues by ensuring that capital goes to its
most productive use.
Economic growth models in modern theories have been premised on the same assumption about
investment and saving as sources of economic growth. There are three economic growth models
including the neoclassical model which is proposed by Domar (1946) and Harrod (1939),
endogenous growth model introduced by Solow (1956), and financial repression hypothesis
modeled by McKinnon,(1973) and Shaw, (1973 )

2.1.3. HARROD- DOMAR MODEL

The Harrod-Domar Model states the rate of economic growth in an economy as dependent on the
level of saving and the capital output ratio.
The productivity of capital investment (this is known as the capital-output ratio)

g= S -δ
θ
Where g = gross rate of growth domestic product
S = saving
θ = capital output ratio
δ = depreciation of capital
If there is a high level of saving in a country, it provides funds for firms to borrow and invest.
Investment can increase the capital stock of an economy and generate economic growth through
the increase in production of goods and services. The capital output ratio measures the
productivity of the investment that takes place. If capital output ratio decreases the economy will
be more productive, so higher amounts of output is generated from fewer inputs. This again,
leads to higher economic growth. The model is mainly used in development economics. It
suggests that if developing countries want to achieve economic growth, governments need to
encourage saving, and support technological advancements to decrease the economy’s capital
output ratio. The Harrod-Domar model provides a framework for economic development and has
been an important influence to government policy formulation.

9
 2.1.5. SOLOW – MODEL
According to Solow model, exogenous technological improvement and capital accumulation
drive economic growth. Based on his analysis of the American data from 1909 to 1949, he
observed that 87.5% of growth of that period was attributable to technological change, and
12.5% to the increased use of capital. The result of the Solow growth model was that many came
to believe financial markets had only minor influence on the rate of investment in physical
capital, and the changes in investment were viewed as having only minor effects on economic
growth.

2.1.6. FINANCIAL REPRESSION HYPOTHESIS

According to McKinnon, (1973) and Shaw, (1973) Financial repression hypothesis states that a
measure by governments channel funds to them as a form of debt reduction. Financial repression
includes Explicit or indirect capping of interest rates, such as on government debt and deposit
rates ,Government ownership or control of domestic banks and financial institutions with barriers
that limit other institutions from entering the market, High requirements, Creation or
maintenance of a captive domestic market for government debt, directing credit to certain
industries.

Financial repression also takes the form of forcing banks to allocate credit to industries that are
perceived to be strategically important for industrial policy, ensures stable provision of capital
rather than leaving it to decisions of disinterested banks or to efficient securities markets.
Government directives and guidance sometimes include detailed orders and instructions on
managerial issues of financial institutions to ensure that their behavior and business is in line
with industrial policy or other government policies.
According to Beim and Charles, (2001), the key reason for the government to implement
financially repressive policies is to control fiscal resources. By having a direct control over the
financial system, the government can funnel funds to itself without going through legislative
procedures and more cheaply than it could when it resorts to market financing.

McKinnon, (1973) and Shaw, (1973) were the first to illuminate the notion of financial
repression. While theoretically an economy with an efficient financial system can achieve

10
growth and development through efficient capital allocation, McKinnon and Shaw argue that
historically, many countries, including developed ones but especially developing ones, have
restricted competition in the financial sector with government interventions and regulations.
According to their argument, a repressed financial sector discourages both saving and investment
because the rates of return are lower than what could be obtained in a competitive market. In
such a system, financial intermediaries do not function at their full capacity and fail to channel
saving into investment efficiently, thereby impeding the development of the overall economic
system. In general, the Mckinnon-Shaw model shows that financial repression reduces both the
quality and quantity of investment in the economy.

The early hypotheses of McKinnon and Shaw assumed that liberalization, which would be
associated with higher real interest rates. The underlying assumption is that saving is responsive
to interest rates. The higher saving rates would finance a higher level of investment, leading to
higher economic growth. Therefore, according to this view, we should expect to see higher
saving rates (as well as higher levels of investment and growth) following financial
liberalization.
Sala and Roubini, (1992) Developed a model that shows financial repression reduces the
productivity of capital and lowers savings, which hampers growth. The model also examined the
effects of policies of repression of the financial system in the form of taxes, restrictions and
various sorts of regulations on the rate of economic growth. They asked the question why an
optimizing government represses the financial sector in spite of the fact that it reduces economic
growth.

2.2. EMPIRICAL LITERATURE

Several empirical Studies have shown that there is a strong relationship between financial
development and economic growth.
Chang and Seteven (2005) Analyzed financial development and economic growth in Taiwan
using vector auto regressive model and found that financial development and Growth Domestic
product are co integrated, the Granger causality tests based on vector error-correction models
(VECM) suggest unidirectional causality running from financial development to economic
growth. Chang and Seteven (2005) result supports the supply-leading hypothesis for Taiwan.

11
In addition to chang and Seteven (2005), Odo et al (2016) had investigated the causality and
impact of financial development on economic growth in Nigeria and South Africa by employing
co integration, VECM and granger causality test. The result of granger causality indicates a
unidirectional causality running from financial development to economic growth in Nigeria but a
bidirectional causality from financial development to economic growth in South Africa
validating the Supply leading hypothesis of financial development by Patrick (1996), which
states that the direction of causality between financial development and economic growth
changes over the course of development. That is, at the early stage of development “the supply –
leading” is evident but as real growth occurs in the economy, it will spark demand for financial
services. This study therefore concludes that supply – leading phenomena is evident in both
Nigeria and South Africa economies.
Similar to Chang et al. (2005) findings, Jung (1986) analyzed the relationship between financial
development and economic growth at international level and found that less developed countries
(LDCs) have a supply-leading causality pattern more frequently than a demand-following
pattern. In this sense, what Patrick emphasized about the usefulness and the importance of
financial development in LDCs is borne out empirically. Thus, LDCs are characterized by the
causal direction running from financial to economic development, and DCs by the reverse causal
direction, regardless of which causality concept is employed.
In contrast to the Jung (1986), Ankilo et al (2010) investigated the long run causal relationship
between financial development and economic growth for ten sub – Saharan countries, but the
granger causality test result different causation for different countries even if the countries are
almost at the same level of Development. Granger causality test within the VECM framework
shows unidirectional relations running from financial development to economic growth in
Central African Republic, Congo Republic, Gabon, and Nigeria while causality runs from
economic growth to financial development in Zambia. However, within the same framework, the
results show bidirectional causality between financial development and economic growth in
Chad, South Africa, Kenya, Sierra Leone and Swaziland.

Whereas Roman, (2012) analyzed using the VAR and VECM approach to determine the long-
run and short-run relationship between financial development and economic growth in Ethiopia.

12
Furthermore, the granger causality test is employed to determine the direction of causality. Using
the financial development indicator PRIV, Roman (2012) found uni-directional causality from
economic growth to financial development; this implies that past economic growth rate is an
important determinant for the development of the financial system. As the economy grows the
demand for financial resources will increase and this in turn will boost the development of
financial sector. This finding is consistent with Patrick’s (1966) Demand-following hypothesis
which postulates a causal relationship from economic growth to financial development.

Rodgers et al. (2014) studied the relationship between financial development and economic
growth in Africa using data from 50 countries for the period [Link] study applied panel
regression and causality testing using credit to the private sectors to total GDP and the ratio of
broad money (M2) to total GDP as proxies of financial development. In contrast the Roman
(2012) findings, the regression concluded a positive relationship between financial development
and economic growth. Moreover, the causality test by Rodgers and others (2014) shows a bi-
directional relationship between financial development and economic growth in Africa. It is also
inconsistent with the demand following hypothesis that state the relationship between economic
growth and financial development is a uni-directional from financial development to economic
growth.
As seen above the empirical findings so far are mixed on the causal direction of financial and
Economic growth. Thus, most empirical studies are contradicting in their conclusion over the
direction of casual relationship of the relationship between financial development and economic
growth. The question of whether financial development is cause for economic growth or
economic growth is cause for financial development? Or is the casual relationship between
economic growth and financial development is bi- directional? Or no causal relationship between
the two variables is far from settled. Therefore, this study will analyze the impact and
relationship between financial depth and economic growth, and will determine kind of casual
relationship the two variables have in Ethiopia.

13
 CTHAPTER THREE: REASERCH METHEDOLOGY

3.1. METHOD OF ANALYSIS

The study will apply the Johansen Co-integration test to investigate the long run determinants of
real economic growth in Ethiopia. If co integration between under variables exist the study will
apply Vector Error correction Model (VECM), if not Vector Auto Regressive (VAR) approach
will applied to identify the relationship between financial depth and economic growth. The use
of VAR model help the study account for spurious correlation, and endogeneity bias as it is
designed for non-stationary time series and requires no endo-exogeneous division of variables
when compared to simultaneous equations. Other tests like causality through Granger causality
also applied Gujarati (2004).
In the simultaneous or structural equation models some variables are treated as endogenous and
some of as exogenous or predetermined (exogenous plus lagged endogenous). Estimation such
models, it has to be make sure that the equations in the system are identified (either exactly or
over). This identification is often achieved by assuming that some of the predetermined variables
are present only in some equations. Therefore, if there is true simultaneity among a set of
variables, they should all be treated on an equal footing; there should not be any a priori
distinction between endogenous and exogenous variables. It is in this spirit that VAR model
developed Gujarati,(2004).
Before estimating the VAR, we have to decide the maximum lag lengths, K to generate the white
noise of error terms. This can be done based on the Akaike information criteria (AIC) or
Schwarz (SIC). Since time-series variables have been widely noted to be non-stationary, the
results that are obtained from the level VAR are spurious and misleading (Mukhopadhyay and
Pradhan, 2010) cited in (Roman, 2006). Moreover, utilizing properly differenced variables in the
VAR may lead to model mis-specification if the level variables share the long run relationship or
are co integrated. In this case the VAR should be written in a VECM (Vector Error Correction
Model). The error correction mechanism (ECM) first used by Sargan and later popularized by
Engle and Granger corrects for disequilibrium. An important theorem, known as the Granger
representation theorem, states that if two variables Y and X are co integrated, then the
relationship between the two can be expressed as ECM Gujarati (2004).

14
 3.2. MODEL SPECIFICATION
The model employed is based on considerations of incorporation of essential variables explained
in the literature section of the paper and keep it straight and effective in explaining the impact of
financial depth on economic growth. Under these considerations, the following variables are
used to develop Real economic growth model: broad money Relative to GDP, domestic credit to
the private sector Relative to GDP, population growth rate and degree of openness (i.e. export
less import) Relative to GDP. Accordingly, the study will measure economic growth as the real
gross domestic product growth rate. Financial development measured by broad money (M2)
relative to (GDP) and Credit to private sector (PRIV) relative to GDP. Degree of openness is
measured (Trade) by import less export(X-M) and population growth measured by population
growth rate (POPG). Government expenditure excluded from the model because the study made
its base on neo classical growth model of 1956 known as Solow growth model. The model
assumes output is produced with the help of two factors of production, capital and labor Solow
(1956). Since the data of capital formation is not adequately found in developing countries then it
proxied by private investment which is credit to private sector.

Definition of Variables in the model and the possible relationship with real gross
domestic product growth rate.

Variables Definition Expected Relationship with RGDPG

RGDPG Real Gross Domestic Product Growth Rate -
PRIV Credit to Private Sector Expected to be positively related
POPG Population growth Rate Exoected effect of the variable is
mixed
M2 Broad Money Hypothesised to relate positively
Trade Export less import Expected to relate positively

15
 3.2.1. ECONOMETRICS MODEL
Therefore the model will be specified as
RGDPGt = F (M2t, PRIVt, POPGt, Tradet,)
RGDPGt = β1 + β2PRIVt +β3 M2t + β3Tradet + β4POPGt + μ
In order to avoid the problem of hetroscedacity problem we will take log on both sides of the
model as following.
LRGDPGt = β1 + β2 LPRIVt +β3 LM2t + β3 LTradet + β4 PoPGt+ μ
Where
LRGDPGt = the log of Real GDP growth rate at time (t)
LPRIVt = the log credit to private sector relative to GDP at time (t)
LM2t = the log of broad Money relative to GDP at time (t)
LPOPGt= the log of population growth rate at time (t)
LTradet = the log trade (export less import) at time (t)
μ= error term
3.3. ECONOMETRICS PROCEDURES

3.3.1. UNIT ROOT TEST

According to Greene (2002)there are two models which have been frequently used to
characterize non-stationary:
(1) Random walk without drift (i.e., no constant or intercept term)
Suppose ut is a white noise error term with mean 0 and variance σ2.

Yt = Yt−1 + ut……………………………………………….(1)
In this model the value of Y at time t is equal to its value at time (t-1) plus a random shock; thus
it is an AR (1) we can write
Y1 = Y0 + u1
Y2 = Y1 + u2 = Y0 + u1 + u2
Y3 = Y2 + u3 = Y0 + u1 + u2 + u3
In general, if the process started at some time 0 with a value of Y0, we have

16
 Yt = Y0 + Ʃut
Therefore
E(Yt) = E(Y0 +Ʃut)= Y0
In like fashion, it can be shown that
var (Yt) = tσ2
As the preceding expression shows, the mean of Y is equal to its initial, or starting, value, which
is constant, but as t increases, its variance increases indefinitely, thus violating a condition of
stationary. In short, the RWM without drift is a non stationary stochastic process. In practice Y0
is often set at zero, in which case E (Yt) = 0 It is easy to show that, while Yt is non stationary, its
first difference is stationary. In other words, the first differences of a random walk time series are
stationary .

(2) Random walk with drift (i.e., a constant term is present).

Yt = δ + Yt−1 + ut
Where δ is known as the drift parameter, the name drift comes from the fact that if we write the
preceding equation as it shows that Yt drifts upward or downward, depending on δ being
positive or negative.
E(Yt) = Y0 + t · δ
var (Yt) = tσ2
Then Yt = ρYt−1 + ut − 1 ≤ ρ ≤ 1
If ρ = 1, becomes a RWM (without drift). If ρ is in fact 1, we face what is known as the unit
root problem, that is, a situation of non-stationary; we already know that in this case the variance
of Yt is not stationary. The name unit root is due to the fact that ρ = 1.11 Thus the terms non
stationary, random walk, and unit root can be treated as synonymous. If, however, |ρ| ≤ 1, that is
if the absolute value of ρ is less than one, then it can be shown that the time series Yt is
stationary in the sense we have defined it. RWM with drift the mean as well as the variance
increases over time, again violating the conditions of (weak) stationary. In short, RWM, with
or without drift, is a non-stationary stochastic process. Therefore In order to test for the existence
of a unit root in time series, we use the popular tests: Dickey-Fuller (ADF) test. Dickey and
Fuller (1976) tested (Gujarati 2004).

17
 3.3.2. CO INTEGRATION ANALYSIS
A natural first step in the analysis of co integration is to establish that it is indeed a characteristic
of the data. Two broad approaches for testing for co integration have been developed. The Engle
and Granger (1987) method is based on assessing whether single-equation estimates of the
equilibrium errors appear to be stationary. The second approach, due to Johansen (1988, 1991)
and Stock and Watson (1988), is based on the VAR approach Greene, (2002).
There are two ways of testing the existence of co integration, the Engel-Granger or EG approach
and the Johansen approach.

A. The Engel-Granger Approach

Although regression analysis deals with the dependence of one variable on other variables, it
does not necessarily imply causation. In other words, the existence of a relationship between
variables does not prove causality or the direction of influence. But in regressions involving time
series data, the situation may be some what different because, if event A happens before event B,
then it is possible that A is causing B. However, it is not possible that B is causing A. In other
words, events in the past can cause events to happen today. This is roughly the idea behind the
so-called Granger causality test Gujarati (2004).

B. The Johansen Approach

The reason for the application of Johansen approach is Engle-Granger co integration test cannot
be used to test the number of co integration relationships that the Johansen test can do.

3.4 NATURE AND SOURCE OF DATA

This study does the empirical analysis by employing data sets for the period 1970-2016 for all
the variables specified in the model. This period is chosen based on the availability of data. Data
for real gross domestic product growth rate (GDPG), broad money (M2), Credit to the Private
Sector (Priv), and degree of openness (Trade) obtained from the National Bank of Ethiopia and
population growth rate (POPG) obtained from World Bank Database.

18
 CHPTER FOUR : EMPIRICLA AND DISCUSSION

This chapter analysis the causation between financial depth and economic growth using annual
data of fourty six years from 1970 – 2016 in Ethiopia, before econometric Analaysis the study
will apply describtive analysis
4.1 DESCRIPTIVE RESULT
As Table 4.1 shows kurtosis and skewness of variable are normaly distributed since the value for
kurutosis approximate to three for the lowest and six for the hight value and the skewness is
between negative one and positive on wich is normal.
Table 4.1 decriptive statitcs
Statistics GDPG Trade M2 POPG PRIV
kurtosis 2.1 3.2 2.2 4.4 2.42
skewnesss -0.60 0.98 -0.21 -0.62 0.39

Source own copmputation using stata 12

Figure 4.1 shows the trand of broad money credit to private sector and real domestic growth all
variables are trendinging together with time . therefore it is expected that the co integration
between variables will happen.
1800000

1600000

1400000

1200000
RGDP
1000000
MS
800000
PRIV
600000
year
400000

200000

0
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46

19
Figure 4.1 trend in broad money and private credit in relation to GDP
Source own calculation using micro soft exel
4.2. ECONOMETRICS RESULT
Before estimation it is necessity to employed the unit root test, to know the time series data’s are
stationary or not. After identifying the optimal lag length, the presence of the co-integrating
vectors is tested by means of the Johansen system. additional the granger causality test is
engaged to discover the direction of causality between financial depth and economic growth. .

4.2.1. UNIT ROOT TEST RESULT

This test is made using the Augmented Dickey-Fuller (ADF) unit root tests. When the ADF test
statistics is larger than the critical value in absolute terms, the null hypothesis of unit root is
rejected, and if the ADF test statistics is less than the critical value in absolute terms, we fail to
reject the null hypothesis, Therefore unit root tests conducted discovered that all variables have
unit root in their level hence they are not stationary. As a result, the variables have to be
differenced to accomplish stationary. From the test results on the first difference given in Tables
1.b, the null hypothesis has been rejected because of the fact that all variables become stationary
at their first difference. As the Unit root tests revealed that all variables used in this study are
stationary at their first difference

Table 4.2a Augmented Dickey-Fuller test for unit root at level

variables Test statistics Critical value at 5% P -value

LGDPG -1.904 -2.952 0.33
lnM2 0.186 -2.952 0.97
Lnpriv 0.765 -2.944 0.99
LnPOPG -1.824 -2.952 0.36
LnTrade -2.006 -2.950 0.28

Source : own computation using Stata 12

20
The Table 4.2a shows that all variables at level have a unit root since the absolute value of the
test statistics is greater than the critical value at five percent therefore we fail to reject the null
hypothesis that is the variables are unit root.
Table 4.2b Augmented Dickey-Fuller test for unit root at difference
variables Test statistics Critical value at 5% P -value
LGDPG -4.933 -2.952 0.000
lnM2 -10.595 -2.944 0.000
Lnpriv -35.931 -2.944 0.000
LnPOPg -5.260 -2.950 0.0000

LnTrade -4.161 -2.955 0.000

Source : own computation using Stata 12

As the Table 4.2b above shows the absolute values of the test statistics for all variables in the
first different are greater than its critical value at 5% level of significance. The result indicates
that the variables are stationary at first difference. So the null hypothesis that suggests each
variable has unit root can be rejected by the ADF test.
4.2.2. CO INTEGRATION TEST RESULT
To evaluate the long run relationship between variables we use the co integration technique, be
short of co-integration between variables suggests being of no long-run association between
them. consequently, the Johansen co-integration method is practical. prior to estimate have to
decide on the maximum Lag length but, including too many lagged terms will consume degrees
of freedom, not to mention introducing the possibility of multi co linearity. Including too few
lags will lead to specification errors. One way of deciding this question is to use a criterion like
the Akaike or Schwarz Gujarati (2004).
Table 4.3 VAR lag order selection criteria
lag LL LR DF P FPE AIC HQIC SBIC
0 -104.63 .00011 5.099 5.17 5.30
1 69.97 349.22 25 0.00 1.1e-07 -1.85 -1.40 -0.63*
e-07
2 97.57 55.19 25 0.00 1.0 * -1.98 -1.14 0.27
3 123.13 51 25 0.00 1.1e-07 -2.00 -0.79 1.2
4 152 58.4* 25 0.00 1.2e-07 -2.2* -61 2.09

21
Source : own computation using Stata 12
As seen on Table 4.3 LR and AIC are choose four lag length, FPE chooses two lag length
HQIC and SBIC are choose one lag length , therefore the study will apply four lag lengths
during estimation to avoid multi co linearity problem.
Table 4.4 Johansen co integration test
Maximum rank Parms LL Eigenvalue Trace statistic Critical value
0 80 100.73 - 103.23 68.52
1 89 129.82 0.74 45.05* 47.21
2 96 140.81 0.40 23.07 29.68
3 101 148.33 0.29 8.03 15.41
4 104 152.09 0.16 0.51 3.76
5 105 152.35 0.011 - -
Source : own computation using Stata 12

It can be seen from the table 4.4 that the Johansen tests for co integration rank test (Trace)
shows one co integrating vectors at the 5% critical value in the system. Thus based on trace
statistics result we can conclude that there exists meaningful long run relationship between the
variables under consideration

4.2.3 GRANGER CAUSALITY TEST RESULT

The study applied granger causality test between real growth domestic product growth rate
(RGDPG) and financial depth indicators PRIV (Private credit relative to GDP) and M2 (Broad
money relative to GDP). As the estimated granger causality test is reported in table (4.5). Shows,
we do reject the null hypothesis that LnRGDPG does not granger cause to LM2 and we accept
the null hypothesis that says LnM2 is not granger cause to LnRGDPG. Therefore the direction of
causation between LnRGDPG and LnM2 is uni directional running from economic growth to
financial depth . in this case the study is in line with demand following hypothesis that stats as
the economy growth it creates demand which makes the economy to growth. this justified as
because broad money grow relative to economic growth since the growth of money more than
economic growth rate it will lead [Link] the monetory authority will restrict the
growth of money to the level of demand for money.

22
And the test shows the granger causation between LnRGDPG and LnPRIV is bi directional
becuase we do reject the two null hypothesis, LnRGDPG is not granger cause to LnPRIV and
LnPRIV do not granger cause to LnRGDPG. Therefore the causation between credit to private
sector and real economic gross rate is bi directional running from economic growth to financial depth
and from financial depth and to economic growth. the result is in line with the findings of Haile and
Kassahun (2011) who employed data of Ethiopia from 1972-2010 to find the causal relationship
between financial development and economic growth. In addition it is consistent with Sime
(2016) who employed data from 1973 to 2008 and analyzed Causality between financial
development and Economic Growth in Ethiopia. Greenwood & Jovanovic, 1990 also found a bi
directional relationship financial development and economic growth. this is beause as the
financial depth deepen the supply for finance will icrease that increase the economy and in turn
as the economy increases the demand for finance will increase.

Table 4.5; - Granger causality Wald tests

Equation Excluded Chi2 df Prob>chi2
GDPG M2 35.71 4 0.00
GDPG PRV 42.60 4 0.00
M2 GDPG 2.16 4 0.7
PRIV GDPG 14 4 0.007

Source : own ccalculation using stata 12

4.2.4 DIAGNOSTIC TESTS

Diagnostics test is undertaken to become aware of model misspecification and as a direct for
model perfection. These tests are serial correlation, model stability and normality tests. The
serial correlation test done using the Lagrange multiplier (LM) test. The null-hypothesis of the
LM test that the residuals are not serially correlated is accepted at 5% level of significance (see
appendix C).
The Jarque-Bera normality test is used to see whether the regression errors are normally distributed.
The null-hypotheis that the residuals are normal is rejected in this particular study. However,
econometric theory states that the existence of non-normality does not affect and distort the

23
estimator’s BLUE and consistency property (Enders 1995). The non-normality of vector in our
model doesn’t affect the coefficients and t-values (see appendix D)

4.2.5 LONG- RUN AND SHORT – SHORT MODELS

The study identified one co integrated equation through Johansen trace statistics and the objectives of
the study is to examine the long and short run impact of financial depth on economic growth Hence,
we estimated the vector error correction model, the result in table (4.5) below is based on the
estimation of the Vector error correction model with four lag selected by the optimum lag length
selection criteria.

Table 4.6 Estimates of β Johansen normalization to LRGDPG

Variables Coefficient Z value P value
lnGDPG 1
lnM2 -3.760128 -6.31 0.000
lnPRIV -3.03903 5.66 0.000
lntrade .5763241 1.15 0.249
lnPOPG 5.382693 5.48 0.000
cons | -4.872834

Source own calculation using stata 12

The Johansen normalization equation can be written as:-

LnGDPG = 3.76LnM2 + 3.03LnPRIV – 0.58LnTrade- 5.4LnPOPG + 4.88
From the above equation it can be observed that LnM2 and LnPRIV have a positive and
significant impact on LnRGDPG in the long run .a percentage increase in LnM2 and LnPriv will
increase LnRGDPG by 3.76 % and 3.03% repectively in the long run with coeffient of
significant at 1 % . this is due to monitazation and credit to private sectors will take maximum
time to affect the economy showing the low development of credit channeling in the short run.
However, and LnPOPG (population growth rate) have a negative and statistically significant
relationship with LnRGDPG in the long-run a parentage increase in LnPOPG will reduce
RGDPG in 5.4 % this might be due to the rapide increase in popaation will be burden for means

24
production in which increase slower than the population [Link] LnTrade (import less
export % RGDP) has a negative and in significant relationship with long Run RGDPG the
unexpected sign for Trade is due to the is country experiencing a persistent trade deficit wich
reduce LnRGDPG growth rate (see Appendix E)

Table 4. 7 Vector Error Correction Models

GDPG coef std z p
cel -1.56 0.30 -5.07 0.000
M2 LD 0.788 2.95 0.27 0.79

L2D 3.41 2.88 1.18 0.23

L3D 2.53 3.07 0.82 0.41

PRIV LD 6.61 1.83 3.61 0.00
L2D 4.51 1.84 2.44 0.01

L3D -.069 0.34 -0.2 0.84

Trade LD 1.82 1.01 1.8 0.07
L1D -0.04 1.03 -0.05 0.96
L2D -1.65 1.08 -1.53 0.12
POPG L1D 4.3 1.45 2.97 0.00
L2D 0.72 1.37 0.53 0.59
L3D -1.03 1.08 -0.95 0.34

Source own calculation using stata 12

As Table 4.7 indicates the sign of co integration equation (c e1) is negative and significant at 1%
that indicates financial depth (LnM2 and LnPRIV) will influence lnRGDPG in the long run. Ce1
is called the speed of adjustment which means the disequilibrium will adjust towards long run
equilibrium at the speed of ce1 (1.56) ; Broad money (M2) and private credit (PRIV) have long
run impact on real growth domestic product growth rate (RGDPG).
In the short run Credit to the private sector (LnPRIV) is significantly affcte LnRGDPG
through period lagged values. LnTrade and LnPOPG affect LnRGDPG through one period

25
lagged value , while broad money (lnM2) doesn’t have a strong direct positive effect in the
short run. this shows
 the under development of the financial sector to affect the eonomy in the short run.
 lack adequate policy and effiecint supervision of financial institution

4.2.6 IMPULSE RESPONSE

Impulse response function is used to trace the effect of a one shock to one of the innovations on
current and future values of the endogenous variables. We can identify the positive or negative
impact of the variables and determine how long it would take for that effect to work. It is a
method of assessing the interaction among the variables in the VECM. Figure 4.2 below
illustrates the response of LnRGDG due to a shock of (generalized impulse) each explanatory
variable. In the first graph, response of lnRGDPG to lnRGDPG implies growth rate of real gross
domestic product in the future will depend on the current growth rate of gross domestic product
growth rate. Any shock will affect LnRGDPG immediately but it will increase after three periods
and this effect remains the same in the economy for a long time period and will not die out even
in the 10th quarter though it shows a fluctuation. Therefore current LnRGDPG rate will affect
future LnRGDPG significantly.

26
 order1, lnGDPG, lnGDPG order1, lnGDPG, lnM2 order1, lnGDPG, lnPOPG order1, lnGDPG, lnPRIV order1, lnGDPG, lntrade
10

-5

order1, lnM2, lnGDPG order1, lnM2, lnM2 order1, lnM2, lnPOPG order1, lnM2, lnPRIV order1, lnM2, lntrade
10

-5

order1, lnPOPG, lnGDPG order1, lnPOPG, lnM2 order1, lnPOPG, lnPOPG order1, lnPOPG, lnPRIV order1, lnPOPG, lntrade
10

-5

order1, lnPRIV, lnGDPG order1, lnPRIV, lnM2 order1, lnPRIV, lnPOPG order1, lnPRIV, lnPRIV order1, lnPRIV, lntrade
10

-5

order1, lntrade, lnGDPG order1, lntrade, lnM2 order1, lntrade, lnPOPG order1, lntrade, lnPRIV order1, lntrade, lntrade
10

-5
0 5 10 0 5 10 0 5 10 0 5 10 0 5 10

step
Graphs by irfname, impulse variable, and response variable

Sourec: own Computation using stata 12

Figure 4.2 Impulse Response Graph
Any positive shock in broad money (LM2) makes an immediate increase in lnRGDPG and this
effect gradually down but does not die out over the time period and may reflect a cyclical effect.
Immediate effect of LnPRIV shock is quite low and however after time period it begins to
increase its affect will not dying out in the 10th period of time.

27
Table 4.8 Impulse Response of LRGDP

step LnRGDPG LnM2 LnPRIV lntrade lnPOPG

0 1 0 0. 0

1 -.319174 6.68504 0.058497 .920173 -4.1303

2 .080878 6.95153 0.560085 -1.21843 -3.42477

3 .374883 3.46026 0.7360021 -1.49271 -.249865

4 .340102 3.26923 0.149898 -.119216 2.09006

5 .054597 5.29333 0.23394 -2.06546 -2.68175

6 .229134 4.53048 0.317172 -1.05759 -2.37534

7 .371073 3.34821 0.120988 .479857 -.245348

8 .07202 -.025814 -049985 -.708178 -.619868

9 .177814 2.45082 -0.7722 -1.50126 -1.81137

10 .253306 4.85816 0.160174 -.515127 -1.81842

Source : owne calculation using stata12

Table 4.8 presents the results of the IRF. In response to shock of LnRGDPG, LnRGDPG itself
decrease by -0.32 in the first year and continues to grow in the third period, in the long-run
reaching 0.25 in 10th period. A one disturbance originating from LnM2 produces a 6.7 increase
in RGDPG in the first year. Its effect continues to decline with a positive impact as the forecast
horizon is extended, then rise again and reaches 4.8 at the 10th year which shows the long run
which shows LnM2 has a permanent impact on GDPG. In other words, financial development
has a long- run impact on economic growth which is consistent with the above findings. The
impact of LnPRIV is positive in the long run it runs from 0.05 to 0.16. the result is consistent

28
with the findings of the study and those of LnPOPG and lnTrade affect RGDPG negative in the
long run. This might be due to the increase in population more than the increase in production
will affect RGDPG negatively lnRGDPG and the persistence trade deficit will reduce GDOPG
growth (see Appendix F).

4.2.7 SUMMERRY OF FINDINGS

the study found a bi-directional causality running from economic growth to financial depth and
from financial depth to economic growth for credit to private sector and uni directional
causality for Broad money running from Economic growth to financial depth. Besides the
granger causality the study found that financial depth have positive and significant impact on
Economic growth in the long [Link] the impact of financial depth in the short run is
insignificant only credit to private sector has effect on the economy in the short run through two
lagged period. Popalation growth rate has significant and negative impact on the gross Domestic
product growth rate but Trade has negative and insignificant impact on gross Domestic product
growth rate this might be due to persistence Trade deficit of the country

29
 CHAPTER FIVE: CONCLUSION AND RECOMMENDATION

5.1 CONCLUSION
The study examines the nexus between financial depth and economic growth in Ethiopia over the
1970-2016 periods. By Using Vector Error Correction Model (VECM) to determine the long-run
and short-run relationship between financial development and economic growth Furthermore, the
granger causality test is employed to find the direction of causality. The empirical result shows a
bi-directional causality from economic growth to financial depth and from financial depth to
economic growth for credit to private sector and uni directional causation running from
Economic growth to financial depth. Besides the granger causality test attempt was made to
determine the impact of financial depth on Economic growth of Ethiopian economy with a
reference to short run and long run effect which was determined using co-integration approach.
The result of the findings indicates that the two financial depth indicators (LnM2 and PRIV)
have positive impact on gross domestic product growth rate in the long run; Broad Money had
insignificant impact on gross domestic product growth rate in the short run but it had a positive
and significant impact in the long run. And credit to private sector had positive impact in the in
the long run but it affect the economy in the short run only through period lagged value. The
Impulse response result is also in line with the above findings that financial depth has a long-run
impact on economic growth.

30
 5.2 RECOMMENDATION

Reference to the findings and conclusion reached at the end of the study; the following are
recommended:
 Adequate policies and efficient supervision of all financial institution should be provided
and sustained.
 in order to stimulates economic growth Central bank of Ethiopia (CBE) should regulate
credit to private sector by reduce interest rate for productive sector of the economy such
as investment on capital goods.
 Since financial depth have a significant effect on economic growth of Ethiopia in the long
run, policy makers should focus on long-run policies interest rate and macroeconomic
stability
 Generally the finding of the study claims the government to create conducive
environment for the development of finance to sustain its contribution to economic
growth.

31
 Reference

Alemayehu Geda (2006) The Structure and Performance of Ethiopia’s Financial Sector in the
Pre- and Post Reform Period with a Special Focus on Banking. researchgate.

Akinlo, Anthony Enisan and Tajudeen Egbetude, (2010) Financial Development and Economic
Growth: The Experiance of 10 Sub-Saharan African Countries Revised. The Review of
Finance and Banking,

Arfanuzzaman,MD (2014) The Long-Run Dynamic Relationship between Broad Money Supply and
the GDP of Bangladesh: A VECM Approach Independent Researcher and Graduate Student,
Dhaka School of Economics

Beck, T., A. Demirgüç-Kunt, and V. Maksimovic (2005). Financial and Legal Constraints to
Firm Growth: Does Size Matter? Journal of Finance.

Beim D. and Charles W.C ( 2001) Emerging Financial Markets... New York: MCGraw Hill.

Caporale, Guglielmo Maria , Christophe Rault , Robert Sova and Anamaria (2009) Financial

Development and Economic Growth:Evidence from Ten New EU. Economics and
Finance Working Paper Series, Volume 09-37. Helpman E and Grossman G.M,
(1991). Innovation and Growth in the Global Economy, cambridge. MA,MIT press.

Chang, Tsangyao Caudill and Steven B ( 2006) Financial Development and Economic Growth
Applied Economics.

Dejene Mamo Bekana (2016) Does Financial Sector Development Drive Economic Growth in a
Post- Communist Economy? Evidence from Ethiopia. Economic Insights – Trends and
Challenges, Volume V(LXVIII)

Demirguc-Kunt, Asli and Levine R ( 2008) Finance, Financial Sector Policies and Long run Growth.
The World Bank Commission on Growth and Development Working Paper, p. 11

32
Enders, W. (1995). Applied Econometric Time Series. Iowa State University, John Wiley and Sons,
Inc.

Graff, Michael (2016) Financial Development and Economic Growth – A New Empirical Analysis.
JEL Classification: G20, N20, 016

Greenwood, J and B Jovanic, (1990) Financial Development Growth And the Distribution of
Income.

Journal of Political Economy, Volume 98, pp. 1076-1107.

Gujarati, D. N. (2004) Gujarati: Basic Econometrics, Fourth Edition. Fourth Edition ed. s.l.:The

McGraw−Hill Companies,.

Haile Kibret Fantaye and Kassahun Aberra (2011). Linkage between Financial market Development
and Economic Growth in Ethiopia. Paper Presented at the 9th International Conference of
EEA/EEPRI.

Jung, W. S. (1986) Financial Development and Economic Growth: International Evidence. the
university of chicago press journals, Volume 34, pp. 333-346
Kahan, Mohasin and Senhadji, Abdelhak S. (2000). Financial Development and Economic Growth
: over view. international monetory fund working paper.
King, Robert. and Levine, Ross. (1993) Finance and Growth; schumpeter Might Be Right. Quarterly
Journal of Economics, Volume 108, pp. 717-738.

Levine R. (2005), “Finance and Growth: Theory and Evidence,” Handbook of Economic

Growth, in: Aghion P. and S. Durlauf (ed.), vol 1, 865-934.

Levine R. (1997), “Financial Development and Economic Growth: Journal of Economic

LiteratureVol. XXXV (June 1997), pp. 688–726

Lucas, R., (1988) On Mechanics of Economic Development. Journal of Monetary Economics,
Volume 22, pp. 3- 42.

Mckinnon, R., (1973) money and capital in Economics Development. Washington DC: Brookings

33
 Institution..

Mukhopadhyay, B and Pradhan,R. (2010). An Investigation of the Finance-Growth Nexus: Study of

Asian Developing Countries Using Multivariate VAR Model. International Research Journal of

Finance and Economics, Issue 58

Murty Sreerama, Wondaferahu Mullugeta Demissie and Sailaja (2012) The Long-Run Impact of
Bank Credit on Economic Growth in Ethiopia: Evidence from the Johansen’s Multivariate

Cointegration approach. European Journal of Business and Management, 4(14).

NBE, National Bank of Ethiopia Quarterly Bulletin (20016) National Bank of Ethiopia Quarterly
Bulletin, Fourth Quarter 2015/16.

Odo Stephen idenyi Bigben [Link] Promise E. Agbi Charity I. Anoke ( 2016) Investigating
the Causal Relationship between Financial Development and Economic Growth in Nigeria and
South Africa.. Journal of Economics and Finance (IOSR- JEF), 7(2).

Roman Tesfaye (2012) The Link Between Financial Development and Economic Grpwth in Ethiopia

Robinson, J. (1952) The Generalization of the General Theory in; the Rate of Interest and other
Essays. London: Macmillan.

Romer, p., (1986) Increasing Returns and Long term Growth. journal of political Economy, Volume
94, pp. 1002- 1037.

Rodgers A. Musamali, Esman M. Nyamongo and Eliud D. Moyi , (2014) The Relationship between
Financial Development and Economic Growth in Africa. Research in Applied Economics, 6(2)

Sala and Roubini (1992) financaial repression and Economic growth. Journal of Development
Economics, 39(5-30).

Shaw, E., (1973) Financial Deepening in Economic Development. New York: Oxford Univesity
Press.

34
Sime, Workenh Eshatuu (2016) Causality between Financial Development and Economic Growth:
Empirical Evidence from Ethiopia: (Johansen and VECMApproaches) Ethiopian Civil service
University ,School of Policy Studies

Solow, Robert M, (1956) a Contribution to the Theory of Economic Growth. The Quarterly Journal
of Economics, Vol. 70, No. 1. , pp. 65-94.

Patrick, H., (1966). Financial Development and Economic Growth in Underdeveloped Countries.
Economic Development and Cultural change, Volume 14, pp. 174-189.

35
 List of Appendices
Appendix A; Lag Exclusion Wald Tests

Equation: All

lag chi2 df Prob > chi2

1 434.3639 36 0.000
2 152.0939 36 0.000
3 123.5316 36 0.000
4 138.9501 36 0.000

Appendix B: VECM Stability Test

Eigenvalue stability condition

Eigenvalue Modulus

1 1
1 1
1 1
1 1
-.3550071 + .8056018i .880355
-.3550071 - .8056018i .880355
.6978548 + .5245762i .87303
.6978548 - .5245762i .87303
-.1025047 + .8298251i .836132
-.1025047 - .8298251i .836132
.2731924 + .7796454i .826124
.2731924 - .7796454i .826124
-.7280695 + .3518887i .808648
-.7280695 - .3518887i .808648
.7277682 + .1044957i .735232
.7277682 - .1044957i .735232
.3953741 + .6129392i .729394
.3953741 - .6129392i .729394
-.359449 + .4686161i .590597
-.359449 - .4686161i .590597
-.5611668 .561167
-.1912726 + .2977968i .353932
-.1912726 - .2977968i .353932
.1484174 .148417

The VECM specification imposes 4 unit moduli.

36
 Roots of the companion matrix

1
.5
Imaginary

0
-.5
-1

-1 -.5 0 .5 1
Real
The VECM specification imposes 4 unit moduli

Appendix C: VEC Residual Serial Correlation LM Tests

Lagrange-multiplier test

lag chi2 df Prob > chi2

1 46.7340 36 0.10854
2 38.6784 36 0.34962
3 34.1421 36 0.55718
4 31.0010 36 0.70514

H0: no autocorrelation at lag order

Appendix D: VEC Residual Normality Tests

Jarque-Bera test

Equation chi2 df Prob > chi2

D_lnGDPG 1.483 2 0.47647

D_lnM2 0.008 2 0.99609
D_lnPRIV 0.068 2 0.96647
D_lntrade 6.922 2 0.03140
D_lnPOPG 13.290 2 0.00130
ALL 21.771 10 0.01632

37
 Appendivx E

J oha ns en no rma li zat ion r est ri ct io n i mp ose d

b eta C oe f. Std . Err . z P> |z| [9 5% Co nf. I nte rv al ]

_c e1
l nG DPG 1 . . . . .
l nM2 - 4.1 03 75 .70 98 137 - 5.7 8 0. 000 - 5. 494 95 9 - 2.7 12 54 1
l nP OPG 6 .29 67 79 1.0 98 936 5.7 3 0. 000 4. 142 90 4 8.4 50 65 4
l nP RIV -3.39 69 03 .66 54 415 5.1 0 0. 000 2. 092 66 1 4.7 01 14 4
ln tr ade . 664 31 18 .52 40 542 1.2 7 0. 205 - .3 628 15 5 1.6 91 43 9
_c ons -5 .38 40 79 . . . . .

Appendix F:
Impulse response Table

38
 Results from order1

(1) (2) (3) (4) (5) (6) (7)

step irf irf irf irf irf irf irf

0 1 0 0 0 0 0 1
1 -.319174 .009464 .039368 -.072122 -.020962 6.68504 1.13153
2 .080878 .023546 .016238 -.070985 -.048959 6.95153 1.15749
3 .374883 .010523 -.004074 -.033095 -.009317 3.46026 1.22037
4 .340102 .018991 -.0071 -.047529 -.027952 3.26923 1.2259
5 .054597 .025757 .028119 -.054463 -.017227 5.29333 1.28448
6 .229134 .020246 .016292 -.052148 -.032998 4.53048 1.23523
7 .371073 .018158 .008737 -.032337 -.015729 3.34821 1.22056
8 .07202 .021584 .024069 -.045712 -.029662 -.025814 1.28235
9 .177814 .024418 .026767 -.049964 -.033959 2.45082 1.33074
10 .253306 .020064 .019991 -.043256 -.024243 4.85816 1.31237

(8) (9) (10) (11) (12) (13) (14)

step irf irf irf irf irf irf irf

0 0 0 0 0 0 1 0
1 -.081857 -.776533 .058497 1.85196 -.070727 1.48508 .51126
2 -.56942 -1.09684 .560035 1.00392 -.129539 1.72891 .4368
3 -.459524 -.053181 .736021 -4.45085 -.045939 2.08168 -.016703
4 -.152978 .289647 .149898 -3.19239 .004283 1.94143 -.045073
5 .45363 .326339 .233494 -.252675 -.022547 1.61033 -.021581
6 .754821 .2157 .317172 -1.66684 .046511 1.57691 -.182967
7 .629736 .038316 .120985 .52902 .092495 1.62693 -.212475
8 .587575 -.036735 -.049985 -.536027 .053009 1.53001 -.060302
9 .44901 .006833 -.07722 533887 .040235 1.56422 .033124
10 .379947 .031926 .160174 -1.9739 .064147 1.71644 -.020964

(15) (16) (17) (18) (19) (20) (21)

step irf irf irf irf irf irf irf

0 0 0 0 0 1 0 0
1 -.229878 .920173 -.040747 .148479 .919872 -.023669 -4.1303
2 -.359821 -1.21843 .001423 .566183 .625592 -.075706 -3.42477
3 -.69619 -1.49271 -.02066 .471108 .693315 -.109039 -.249865
4 -.808028 -.119216 -.057779 .223234 .712387 -.130879 2.09006
5 -.669565 -2.06546 -.000499 .280337 .613943 -.148784 -2.68175
6 -.624731 -1.05759 .025559 .276192 .607911 -.141189 -2.37534
7 -.637495 .479857 -.018853 .186969 .66129 -.02703 -.245348
8 -.59054 -.708178 -.012445 .20232 .651174 -.041295 -.619868
9 -.535714 -1.50126 .007497 .272826 .655092 -.101643 -1.81137
10 -.611471 -.515127 -.010359 .260347 .696851 -.099608 -1.81842

(22) (23) (24) (25)

step irf irf irf irf

0 0 0 0 1
1 -.017351 -.241027 -.291252 .080143
2 .13031 -.491724 -.096825 .21922
3 .074722 -.472934 -.006829 .487648
4 .048555 -.521335 -.021938 .62771
5 .057769 -.344398 -.065713 .567862
6 .067406 -.363846 -.022043 .352462
7 .028972 -.377624 .147599
39. 4 9 3 7 6 6
8 .033426 -.340699 .054569 .484779
9 .05873 -.274241 -.062338 .464225
10 .045678 -.343871 -.051371 .421749
(1) irfname = order1, impulse = lnGDPG, and response = lnGDPG
(2) irfname = order1, impulse = lnGDPG, and response = lnM2
(3) irfname = order1, impulse = lnGDPG, and response = lnPRIV
(4) irfname = order1, impulse = lnGDPG, and response = lntrade
(5) irfname = order1, impulse = lnGDPG, and response = lnPOPG
(6) irfname = order1, impulse = lnM2, and response = lnGDPG
(7) irfname = order1, impulse = lnM2, and response = lnM2
(8) irfname = order1, impulse = lnM2, and response = lnPRIV
(9) irfname = order1, impulse = lnM2, and response = lntrade
(10) irfname = order1, impulse = lnPRIV, and response = lnGDPG
(11) irfname = order1, impulse = lnPRIV, and response = lnM2
(12) irfname = order1, impulse = lnPRIV, and response = lnPRIV
(13) irfname = order1, impulse = lnPRIV, and response = lntrade
(14) irfname = order1, impulse = lnPRIV, and response = lnPOPG
(15) irfname = order1, impulse = lntrade, and response = lnGDPG
(17) irfname = order1, impulse = lntrade, and response = lnM2
(18) irfname = order1, impulse = lntrade, and response = lnPRIV
(19) irfname = order1, impulse = lntrade, and response = lntrade
(20) irfname = order1, impulse = lntrade, and response = lnPOPG
(21) irfname = order1, impulse = lnPOPG, and response = lnGDPG
(22) irfname = order1, impulse = lnPOPG, and response = lnM2
(23) irfname = order1, impulse = lnPOPG, and response = lnPRIV
(24) irfname = order1, impulse = lnPOPG, and response = lntrade
(25) irfname = order1, impulse = lnPOPG, and response = lnPOPG

40
 Appendix G
Descriptive statistics

stats lnGDPG lntrade lnM2 lnPOPG lnPRIV

kurtosis 2.101572 3.20377 2.232366 4.40697 2.424606

skewness -.6066757 .9801157 -.2159422 -.6207907 .3932216

Appendix H
Lag Length selection criteria

Selection-order criteria
Sample: 1974 - 2016 Number of obs = 43

lag LL LR df p FPE AIC HQIC SBIC

0 -104.633 .000113 5.09919 5.17471 5.30398

1 69.9758 349.22 25 0.000 1.1e-07 -1.85934 -1.40622* -.630597*
2 97.5748 55.198 25 0.000 1.0e-07* -1.98022 -1.1495 .272474
3 123.13 51.11 25 0.002 1.1e-07 -2.00603 -.7977 1.27062
4 152.357 58.454* 25 0.000 1.2e-07 -2.20264* -.61671 2.09796

Endogenous: lnGDPG lnM2 lnPRIV lntrade lnPOPG

Exogenous: _cons

Appendix I Johansen Co Integration Test

Johansen tests for cointegration

Trend: constant Number of obs = 43
Sample: 1974 - 2016 Lags = 4

5%
maximum trace critical
rank parms LL eigenvalue statistic value
0 80 100.73788 . 103.2378 68.52
1 89 129.82924 0.74156 45.0550* 47.21
2 96 140.81989 0.40022 23.0737 29.68
3 101 148.33812 0.29509 8.0373 15.41
4 104 152.09807 0.16044 0.5174 3.76
5 105 152.35676 0.01196

41
 Appendix J Granger Causality Wald Test

Granger causality Wald tests

Equation Excluded chi2 df Prob > chi2

lnGDPG lnM2 35.718 4 0.000

lnGDPG lnPRIV 42.601 4 0.000
lnGDPG lnPOPG 17.762 4 0.001
lnGDPG lntrade 19.406 4 0.001
lnGDPG ALL 115.59 16 0.000

lnM2 lnGDPG 2.163 4 0.706

lnM2 lnPRIV 8.0843 4 0.089
lnM2 lnPOPG 4.0607 4 0.398
lnM2 lntrade 2.964 4 0.564
lnM2 ALL 31.275 16 0.012

lnPRIV lnGDPG 14.009 4 0.007

lnPRIV lnM2 10.288 4 0.036
lnPRIV lnPOPG 9.2877 4 0.054
lnPRIV lntrade 19.447 4 0.001
lnPRIV ALL 97.703 16 0.000

lnPOPG lnGDPG 4.3213 4 0.364

lnPOPG lnM2 12.053 4 0.017
lnPOPG lnPRIV 8.1792 4 0.085
lnPOPG lntrade 5.1731 4 0.270
lnPOPG ALL 53.643 16 0.000

lntrade lnGDPG 11.631 4 0.020

lntrade lnM2 18.041 4 0.001
lntrade lnPRIV 10.16 4 0.038
lntrade lnPOPG 2.4754 4 0.649
lntrade ALL 33.027 16 0.007

42
 Appendix K Vector Erro Correction

C oe f. Std . Err . z P> |z| [9 5% Co nf. I nte rv al ]

D_ ln GDP G
_ ce1
L1. -1 .56 80 53 .30 93 657 - 5.0 7 0. 000 - 2. 174 39 8 - .96 17 07 3

l nG DPG
LD. . 248 87 88 .22 71 037 1.1 0 0. 273 - .1 962 36 3 .69 39 93 8
L 2D. . 071 49 37 .19 79 176 0.3 6 0. 718 - .3 164 17 6 .4 59 40 5
L 3D. -. 060 96 62 .1 32 502 - 0.4 6 0. 645 - .3 206 65 4 .19 87 32 9

l nM2
LD. . 788 96 11 2.9 57 947 0.2 7 0. 790 - 5. 008 50 8 6. 58 64 3
L 2D. 3 .41 76 03 2.8 88 094 1.1 8 0. 237 - 2. 242 95 8 9.0 78 16 4
L 3D. 2 .53 12 48 3.0 77 532 0.8 2 0. 411 - 3. 500 60 5 8 .5 63 1

l nP RIV
LD. 6 .61 73 24 1.8 34 484 3.6 1 0. 000 3. 021 80 1 10. 21 28 5
L 2D. 4 .51 49 34 1.8 48 523 2.4 4 0. 015 .8 918 95 4 8.1 37 97 2
L 3D. - .06 93 66 .34 40 059 - 0.2 0 0. 840 - .7 436 05 2 .60 48 73 2

ln tr ade
LD. 1 .82 38 79 1.0 15 355 1.8 0 0. 072 - .1 661 79 2 3.8 13 93 8
L 2D. -. 047 64 05 1.0 35 942 - 0.0 5 0. 963 - 2. 078 04 9 1.9 82 76 8
L 3D. -1 .65 63 69 1.0 84 676 - 1.5 3 0. 127 - 3. 782 29 6 .46 95 57 8

l nP OPG
LD. 4 .31 00 45 1.4 53 189 2.9 7 0. 003 1. 461 84 6 7.1 58 24 4
L 2D. .72 83 77 1.3 73 825 0.5 3 0. 596 - 1. 964 27 1 3.4 21 02 5
L 3D. -1 .03 23 29 1.0 84 088 - 0.9 5 0. 341 - 3. 157 10 2 1.0 92 44 5

_c ons . 000 06 32 .48 43 184 0.0 0 1. 000 - .9 491 83 5 .94 93 09 9

43