(Luận văn thạc sĩ) financial development and economic growth in vietnam a quantitative assessment

The thesis is among the very first attempts to examine the relationship between growth and financial development in Vietnam using recent developed time series economic techniques such as

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

FINANCIAL DEVELOPMENT AND

ECONOMIC GROWTH IN VIETNAM:

A QUANTITATIVE ASSESSMENT

BY

LÊ TẤN BỬU DUY

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

HO CHI MINH CITY, April 2014

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

FINANCIAL DEVELOPMENT AND

ECONOMIC GROWTH IN VIETNAM:

A QUANTITATIVE ASSESSMENT

A thesis submitted in partial fulfilment of the requirements for the degree of

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

By

LÊ TẤN BỬU DUY

Academic Supervisor:

DƯƠNG NHƯ HÙNG

ACKNOWLEDGEMENT

I want to send special thanks to my academic supervisor, Dr Duong Nhu Hung, for his patience as well as his continuous supports and valuable comments during my thesis writing process

I am also thankful to all of the lecturers in VNP for their wonderful lessons, to all administrators in the VNP Office for their supports, and especially to Mr Nguyen Dinh Quy for being so helpful every time I come to the VNP Library

Finally, I want to show my gratitude to my family and my friends for their endless encouragements to me so that I could finish the program as well as the thesis

to GDP, ratio of credits to GDP, and ratio of credits provided to the private sector to GDP Secondly, Toda and Yamamoto (1995) causality test was employed to determine the causal relationship between economic growth and financial development under a multivariate VAR framework Evidence of causality running from financial development

to economic growth is found for all indicators of financial development These results provide strong support to the supply-leading hypothesis on the relationship between economic growth and financial development for the case of Vietnam

ACRONYMS AND ABBREVIATIONS

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION 8

1.1 PROBLEM STATEMENT 8

1.2 RESEARCH OBJECTIVES AND QUESTIONS 10

1.3 RESEARCH METHODOLOGY 11

1.4 STRUCTURE OF THE THESIS 11

CHAPTER 2: LITERATURE REVIEW 12

2.1 KEY CONCEPTS 12

Economic Growth 12

Financial Development 12

2.2 THEORETICAL REVIEW 12

2.2.1 Theoretical Hypotheses 12

2.2.2 Financial Repression Theory 14

2.2.3 Financial Liberalization Theory 15

2.2.4 Financial Structure Theory 17

Bank-based Theory 17

Market-based Theory 19

Other Views in Financial Structure Theory 21

2.2.5 Endogenous Growth Theory 22

2.2.6 Summary of Theoretical Predictions 25

2.3 EMPIRICAL REVIEW 26

2.3.1 Traditional Cross-Country Analysis 26

2.3.2 Panel Cointegration and Causality Analysis 27

2.3.3 Time series analysis 29

VECM Approach 29

VAR Approach 32

2.3.4 Previous empirical studies about Vietnam 35

2.3.5 Summary of findings from previous empirical studies 35

2.4 CONCEPTUAL FRAMEWORK 26

CHAPTER 3: RESEARCH METHODOLOGY AND DATA COLLECTION 38

3.1 ANALYTICAL FRAMEWORK 38

3.2 MODEL SPECIFICATION AND DATA COLLECTION 39

Economic Growth (EG) 39

Financial Development (FD) 40

Trade (EX, IM) 41

Inflation (INF) 43

Real Interest Rate (RIR) 44

3.3 ECONOMETRIC TECHNIQUES 45

3.3.1 Seasonal Adjustment 45

3.3.2 Unit Root Test 47

3.3.3 ARDL Cointegration Test 48

3.3.4 Toda and Yamamoto Causality Test 50

CHAPTER 4: OVERVIEW OF ECONOMIC GROWTH AND FINANCIAL DEVELOPMENT IN VIETNAM 53

4.1 ECONOMIC GROWTH 53

4.1.1 Overview 53

4.1.2 Results 53

Structural Changes 53

Global Integration 55

4.1.3 Challenges 56

Investment-driven Economic Growth 56

Inefficient SOEs and Public Investments 57

Macroeconomic Instability 58

4.2 FINANCIAL DEVELOPMENT 58

4.2.1 Overview 58

4.2.2 Results 59

Banking Market 59

Stock Market 62

4.3 A FIRST LOOK ON THE RELATIONSHIP BETWEEN ECONOMIC GROWTH AND FINANCIAL DEVELOPMENT 64

CHAPTER 5: EMPIRICAL RESULTS 65

5.1 Seasonal Adjustments 65

5.2 Preliminary Analysis 66

5.3 Unit Root Test 69

5.4 ARDL Cointegration Test 70

5.5 Toda and Yamamoto Causality Test 72

CHAPTER 6: CONCLUSIONS 77

6.1 Summary and Conclusions 77

6.2 Limitations 78 6.3 Directions for Further Research 79

REFERENCE 80

LIST OF TABLES

Table 2.1: Summary of theoretical predictions 25

Table 2.2: Findings from previous empirical studies 36

Table 3.1: Critical values for ADF test 48

Table 3.2: Critical values for bound F-test 50

Table 4.1: Proportion of output from different economic sectors 54

Table 4.2: Industrial output share by ownership 54

Table 4.3: : Investment share by ownership 55

Table 4.4a: Structure of goods exports by kinds of economic sector 56

Table 4.4b: Structure of goods imports by kinds of economic sector 56

Table 4.5: Growth accounting of Vietnam from 1990 to 2007 56

Table 5.1: Descriptive statistics 68

Table 5.2: Correlation coefficients 69

Table 5.3: Results of ADF test 70

Table 5.4: AIC, SBC, F-statistic, t-statistic, and LM for different lag-lengths 71

Table 5.5: Optimal lag-length p chosen by AIC, SC, and LR test 73

Table 5.6: Results of VAR residual autocorrelation LM test 74

Table 5.7: Toda and Yamamoto causality test results 75

LIST OF FIGURES Figure 4.1: Openness, export, import, and net export (as percents of GDP) of Vietnam 55

Figure 4.2: The incremental capital-output ratio (ICOR) of Vietnam in 1991-2009 57

Figure 4.3: Number of banks in Vietnam from 2006 to 2010 60

Figure 4.4a: Deposit market share according to type of banks in 2005-2010 60

Figure 4.4b: Credit market share according to type of banks in 2005-2010 60

Figure 4.5: Ratio of banking credits and deposits to GDP 61

Figure 4.6: GDP growth and credit growth of Vietnam from 2000 to 2012 64

Figure 5.1: Real GDP and nominal GDP from 1999Q1 to 2012Q3 65

Figure 5.2: Seasonal adjusted real GDP and nominal GDP from 1999Q1 to 2012Q3 66

Figure 5.3: Graphical presentations of all variables from 1999 to 2012 66

CHAPTER 1: INTRODUCTION

The chapter introduces the necessary of a quantitative analysis of the relationship between financial development and economic growth of Vietnam Then it will discuss about the objectives of the thesis, the questions to be answered in the thesis, and the econometric techniques used to answer those questions The chapter ends by providing the structure of the thesis

1.1 PROBLEM STATEMENT

The relationship between financial development and economic growth was first examined in

a research by Schumpeter in 1911 (as cited in King and Levine, 1993a) However, it only receives great attention from economists since the seminal studies of McKinnon (1973), Shaw (1973), and King and Levine (1993b) Despite of the high volume in both theoretical and empirical research, the direction in the relationship is still not conclusive yet The research results on the relationship could be classified into five categories, each supporting one of the followings hypotheses:

i Supply-leading hypothesis (Patrick, 1966): financial development increases the efficiency in the allocation of economic resources and thus could promote economic growth eventually

ii Demand-following hypothesis (Patrick, 1966): financial development might passively follow economic growth as higher level of income will lead to higher demand for financial services

iii Bi-directional causality hypothesis: financial development causes economic growth and conversely the development of the economy also induces the development of the financial sector

iv Independent hypothesis: the role of financial development in economic growth process is overstated; there is no relationship between financial development and economic growth

v Stage of development hypothesis (Patrick, 1966): in the early stage of economic growth, the supply-leading hypothesis takes the lead Then, the demand-following hypothesis dominates

Most of the initial empirical studies on the relationship between growth and financial development are cross-country based Such studies are usually criticized for their robustness, especially on the econometric methods being employed Arestis et al (2001), Abu-Bader and Abu-Qarn (2008a) argued that these cross-country studies could not explore the dynamics in the relationship between growth and finance to answer questions related to causality When the growth regressions in these studies return a significant coefficient for financial development, it might just because there is an association between economic growth and development, and thus this result could not be interpreted as the evidence of causality Moreover, cross-country studies could only provide the broad view on the relationship between finance and growth Results from such studies might be correct in general but not necessary be true in a particular country due to differences in economic or political characteristics Thus cross-country studies might not be very useful for policy making purposes at specific countries For those purposes, time series based studies are needed

However, there are a few limitations associated with the current time series based studies on the relationship between economic growth and financial development Firstly, although country-specific time series studies are helpful for policy making, they are usually criticized about the model specification being used Luintel and Khan (1999), and Odhiambo (2008) argued that many of the previous time series studies employ bi-variate or tri-variate models thus they are vulnerable to omitted variable issues and might return biased causal relationship between financial development and economic growth Secondly, financial development is a complex concept thus there is no single variable could reflect accurately the whole level of financial development of a country However, many empirical studies only use one or two indicators of financial development and hence might subject to the question about robustness

of the results In short, results from previous time series based studies are subjected to the questions regarding to misspecification or omitted variables problem, and the question related

to robustness as usually only a few financial measures are employed in a single study This thesis attempts to address those limitations by providing a time series analysis based on a high dimension multivariate model with several different indicators of financial development being used

For the case of Vietnam, as shown in Nguyen (2011), most of the previous studies on the growth-financial development nexus are qualitative-based and only a few recent papers such

as Anwar and Nguyen (2009), and Nguyen (2011) employ quantitative analysis in studying

the link between them However, these later studies employ traditional panel data analysis method and thus subject to the criticisms by Abu-Bader and Abu-Qarn (2008a), Lee and Chang (2009) as this method is not appropriate to examine the causal relationship Briefly, there has been no time series analysis so far on studying the growth-finance relationship in Vietnam The thesis is among the very first attempts to examine the relationship between growth and financial development in Vietnam using recent developed time series economic techniques such as Toda and Yamamoto causality test (Toda & Yamamoto, 1995), and autoregressive distributed lag (ARDL) cointegration test (Pesaran, Shin & Smith, 2001)

The examination of the relationship, especially the causality link, between economic growth and financial development in Vietnam is important especially in the current context Vietnam has experienced a long time of high growth rate but it is now facing serious problems of macro-instability On April 2012, Vietnam’s government postulates a strategy to stabilize the economy as well as achieve sustainable economic growth The strategy aims to reconstruct three major fields which are the financial sector, public investments, and state-owned enterprises Given the limitations of economic resources, appropriate priorities should be assigned to these three projects Within this context, the analysis of the causal relationship between financial development and economic growth in Vietnam become more necessary The reason is that different causal results might return different implications for policy making According to Calderon and Liu (2003), if the evidence supports the supply-leading hypothesis, policy to reconstruct the financial sector should have higher priority Conversely,

in the demand-following case, the government should focus and give higher priority for other policies to improve economic growth Thus, the analysis of the causal relationship between economic growth and financial development could be useful for considerations and constructions of future growth enhancing policies for Vietnam

The general objective of the research is to examine the relationship between financial development and economic growth in Vietnam and then propose some suggestions for policy making In details, the research will focus on the following goals:

1 To determine the long run relationship among economic growth, financial development and other macroeconomic variables in Vietnam

2 To determine the causal relationship between economic growth and financial development in Vietnam

In other words, the research will attempt to answer to the following research questions:

1 Is there any long-run relationship among economic growth, financial development and other macroeconomic variables in Vietnam from 1999 to 2012?

2 What is the causality relationship between economic growth and financial development in this period?

1.3 RESEARCH METHODOLOGY

To answer the first research question about the long run relationship among economic growth, financial development and other macroeconomic variables in Vietnam the thesis will employ the ARDL bound testing procedure In seeking for the answer to the second research question related to the causality relationship between economic growth and financial development, this thesis will employ the Toda and Yamamoto (1995) causality test

1.4 STRUCTURE OF THE THESIS

This thesis is organized into six chapters Chapter I is introduction Chapter II is literature review Chapter III describes the research methodology Chapter IV explores economic growth and financial development of Vietnam in recent years Chapter V shows the empirical results of the thesis And finally Chapter VI includes the conclusion, limitations of the thesis, and direction for further research

CHAPTER 2: LITERATURE REVIEW

The chapter begins by providing definition to the main concepts of economic growth and financial development After that, theoretical hypothesis and models on the links between economic growth and financial development are reviewed Finally, results from previous empirical studies on the relationship are discussed

2.1 KEY CONCEPTS

Economic Growth

Perkins et al (2006) defines economic growth as the increase in income per capita and the increase in production of goods and services in a country In other words, a country has gained economic growth if it has produced more goods and services and also the average income of the people has increased

Financial Development

According to Levine (2005), as the financial system provides various functions, its development should include various conditions reflecting how great these functions are provided Specifically, financial development consists of improvements in facilitating the transactions of goods and services, better attracting of savings and allocation of capital, cost reduction in gathering and processing information about potential investments, enhancements

in corporate governance and overseeing investment projects, improvements in risk management and risk diversification Calderon and Liu (2003) have a more simple definition

of financial development According to them, financial development is referred to as the enhancements in the “quantity, quality, and efficiency of financial intermediary services” (Calderon and Liu, 2003)

2.2 THEORETICAL REVIEW

2.2.1 Theoretical Hypotheses

Schumpeter (as cited in King and Levine, 1993a) is among the earliest economists who advocate the hypothesis that financial development is important to economic growth He (as cited in Ang, 2008) argued that firms need access to credits to finance their employment of new production facilities and techniques Then banks through their intermediary activities could identify and finance the most productive firms and hence promote economic growth at the aggregate level Patrick (1966) named this hypothesis as “supply-leading” because in this

case, the supply of financial intermediate services has led to economic growth According to Patrick (1966), within growing industries, the demand for credit would be high and thus financial intermediate activities could enhance economic growth by attracting savings from none or slow growing industries and channeling these savings to high growing industries

In contrast to the hypothesis above, Robinson (as cited in King and Levine, 1993a) supports the reverse hypothesis that financial development simply follows the growth of an economy Patrick (1966) also termed this hypothesis as “demand-following” since economic growth generates higher demand for financial services When the economy of a country grows, firms will need more capital to expand their productions and hence the demand for financial intermediate services will increase This increasing in demand for financial services will in turn induce the expansion of the financial sector Thus in this case, financial development passively follows economic growth

Besides these two hypotheses, Patrick (1966) also proposes the “stage of development” hypothesis which argued that the relationship between growth and financial development will change according to the stage of development of a country Specifically, the “supply-leading” phenomenon could take the important role during the early development stage by providing finances to productive investments And when the industries which are financed through the

“supply-leading” process have achieved sustainable growth, the “demand-following” phenomenon takes the leading role Patrick (1966) noted that this chain of “supply-leading” and “demand-following” might not only happen at the country level but also at the industry level This means while some industry has achieved high growth rate and is dominant by

“demand-following”, another industry might still be at the “supply-leading” stage

Another pattern in the relationship between financial development and economic growth is bi-directional causality Lewis (as cited in Patrick, 1966) supports this hypothesis and believes that economic growth would lead to the expansion and development of the financial sector and as the financial system becomes more effective it could becomes the engine for further economic growth Depart from the above four hypotheses in which there’s always a causality link between growth and finance though with different direction, there exists the last hypothesis which views financial development and economic growth as unrelated matters Lucas (1988) argued that economists had overestimated the role of financial development to economic growth

2.2.2 Financial Repression Theory

In 1950s and 1960s, governments at developing countries intervened intensively in their financial markets The governments as these countries viewed their financial systems as underdeveloped and could not support the industrialization and economic growth processes thus their interventions are needed This motive is derived from the financial repression theory which was originated from the studies of Keynes (1936) and Tobin (1965)

As cited in Fry (1988), Keynes uses the liquidity preference theory and the liquidity trap to explain how financial repression is needed to induce economic growth In the model of Keynes, consumers allocate their wealth in two kinds of asset which are money and productive capital like government bond of which the value moves conversely with market interest rate The allocation decision of consumers will depend on their expectation about the market interest rate, e.g if they expect the market rate to fall which means that the bond value

is expected to go up, they will prefer buying and keeping bonds Based on the liquidity preference theory, Keynes introduces the liquidity trap phenomenon in which at a given point

of time, consumers consider some specific level of interest rate as “normal” When the interest rate falls below this normal level, consumers will expect it to rise Conversely, when the interest rate rises above the normal level, it is expected to fall back

In Keynes’ model, when the monetary policy increases money supply, consumers will realize that they are keeping more money than needed and thus shifting their wealth into government bonds This reaction would increase bond prices while lowering the interest rate The same phenomenon will occur continuously if money supply keeps increasing However, this will stop when the interest rate fall to a specific level called liquidity trap interest rate The reason

is that at this level, the interest rate has fallen so much below the “normal” interest rate thus consumers will expect the market rate to rise and prefer to keep the increase in money supply rather than investing in productive capital like government bonds Keynes had proposed an alternative to this adjustment process If the government represses the financial system by imposing a ceiling interest rate, and as the same time maintains fixed price level so that future expectations about price level will be stable, the expansion monetary policy would be benefit

to economic growth With the introduction of a ceiling interest rate and a fixed price level, the expansion monetary policy could reduce interest rate to stimulate investments and thus promote economic growth

Another model which also advocates financial repression is represented in Tobin (1965) In this model, consumers will devote their income for either consumption or savings purpose Regarding the savings part, consumers will partly invest in physical capital and hold the rest

as money balances Thus physical capital and money balances are the only two assets in the model and they are substitutes of each other In the monetary growth model of Tobin (1965),

by lowering the yield on holding money balances, the return from investments into capital stock will become more attractive comparing with the return from holding money and thus the consumers will allocate a higher fraction of their wealth into productive capital assets This reaction of the consumers will increase the capital to labor ratio which reflects a higher productivity of labor The transition of the capital to labor ratio from low to high will accelerate economic growth Thus the policy implications behind Tobin’s model is that government should lower the yield from holding money to prevent the flowing of savings to money balances which is unproductive This could be accomplished by setting ceilings on interest rate, or imposing tax on holding money balances

2.2.3 Financial Liberalization Theory

In 1970s, the financial repression view was challenged by the seminal research of McKinnon (1973) and Shaw (1973) These authors had established the theoretical framework for the financial liberalization theory This new theory criticized the financial repression view that government’s interventions in the financial systems at developing countries such as ceilings

on interest rate, high reserve requirements, and credit directing had brought negative effects

to savings, lead to inadequate and ineffective investments, and as a result impede economic growth The implications from financial liberalization theory is that the government at developing countries should remove distortions in the financial system, let the market determines the real interest rate itself, and allow the financial institutions to allocate credits to the most productive investment projects According to the McKinnon (1973) and Shaw (1973), such liberalized financial system would be beneficial to economic growth

McKinnon (1973) argued that in developing countries, the financial systems are inefficient and thus investments are usually self-finance, i.e investors must accumulate enough funds in form of money balances before investments could take place In the model of McKinnon, money and capital are complements rather than substitutes as in financial repression theory The complementing roles of money and capital are expressed in McKinnon’s model in two ways Firstly, he believed that the demand for money react positively real return of capital

due to the desire of economic agents in making investments The reason is that when real return of capital increases, agents would like to invest more and due to the self-finance nature

of investment, the demand for money will arise accordingly Secondly, McKinnon (1973) argued that investment depend positively on real interest rate of money balances This is because when real interest rate of money balances increases, the accumulation of money balances will occur and thus more funds are available for investments The implication behind the complementing roles of money and capital in McKinnon’s model is that government in developing countries should remove the repressions in the financial system such as interest rate ceilings, high reserve requirements, or credit directing The withdrawal

of these distortions would lead to a higher real interest rate of money balances which will speed up capital accumulation process and boost economic growth

An alternative theoretical framework for financial liberalization could be found in Shaw (1973)’s model In this model, money is debt of the financial intermediary system and an increase of money stock in the economy reflects a higher level of financial intermediation activities between savers and borrowers Shaw (1973) argued that the liberalization of the financial system would lead to the expansion of the financial intermediary system This development in the financial system will in turn promote economic growth through the savings effect and investment effect respectively

According to the savings effect, the propensity to save of the people tends to increase in line with the higher level of money stock in the economy Shaw explained that this effect may be

a result of the income effect discussed earlier as the increase in savings might be simply a result of the increase in income The increase in savings could be simply a direct result of liberalization since savings might be depressed earlier by government’s intervention and now

as the financial system is liberalized, savings increases to its normal level Moreover, the removal of financial repressions would increase real interest rate of money balances to reflect the scarcity of savings, and this high real interest rate will in turn attract more savings Another explanation by Shaw (1973) for the increase in savings is that with real money growth, entrepreneurs tend to save more to accumulate equity in order to apply for additional loans

The investments effect of real monetary growth is explained in the model of Shaw (1973) as follows He argued that the financial market at developing countries is fragmented and the

development of the financial intermediary system reflected by real money growth would unify and widen the financial market Thus the pooling of savings would be more efficient while the level of investments would increase since investors could now access to a larger market and the higher liquidity provided by the unified market will encourage large and long-term investments Moreover, the average return from investments will increase since the financial system is now allowing for more discriminating choice among investments Finally, the investments will be taken more efficiently since the development of the financial intermediary system will reduce risks through diversification, lower costs to investors through economies of scale

2.2.4 Financial Structure Theory

Although the main debate in financial structure theory is which type of financial structure is superior for promoting economic growth, the theoretical studies in this theory also provide some useful insights on how a financial system could promote economic growth A financial structure could be thought as the foundation of the financial system as it indicates how financial activities are arranged within the economy (Stulz, 2000) Competing views in financial structure theory maintain their own views on how a specific type of financial structure is beneficial to economic growth Currently there exist four distinct views in the debate which are bank-based, market-based, financial services, and finally law and finance

Bank-based Theory

The bank-based theory often stresses the important role of banks in facilitating economic growth For example, in the model developed by Diamond and Dybvig (1983), banks could promote economic growth by reducing the liquidity risks associated with long-term and high-return investment projects According to the authors, consumers invest part of their savings into long-run, high-return but illiquid projects When some consumers receive shocks, they will need to access their savings earlier and this could lead to the liquidation of those projects Banks could eliminate this liquidity risk by providing demand deposits which are short-term and liquid assets that consumers could exchange into cash whenever they want The demand deposits provided by banks act as the insurance against liquidity risk since these could satisfy the early consumption need of the consumers who receive shocks Thus the presence of banks and their demand deposits could enhance economic growth by preventing the liquidation of long-run and high-return investment projects when consumers receive shocks

A bank-based system could also be beneficial to economic growth through its positive effect

on the capital allocating process For instance, in the financial intermediation model of Diamond (1984), banks introduced cost advantage when mobilizing capital from savers to borrowers The information cost is minimized as bank act as a delegated agent in monitoring loans to individual firms Without banks, there would be “duplication of effort” if every savers tend to monitor their loans directly or there would be a free-rider issue if none of the savers monitor borrowers Diamond (1984) argued that although banks received deposits from savers, their utilization of funds through intermediary activities need not to be monitored like in the case of normal borrowers because banks could diversify their lending portfolio to reduce the risks associated with their loans Thus the delegated monitoring role and risk diversification ability of banks should enhance efficiency in capital allocation

While Diamond (1984) explores the advantage of banks in solving asymmetric information problem occurs after providing loans, Boyd and Prescott (1986) stresses the important of banks in resolving asymmetric information issue which arises before signing loan contracts or investment have taken place In the later model, financial intermediaries are coalition of agents and they are identical with banks since their activities include receiving deposits from savers, providing loans, and making investments According to Boyd and Prescott (1986), these financial intermediaries could prevent the adverse selection issue and thereby improve the capital allocation process since they evaluate and produce information about possible projects before making investment decisions Thus financial intermediaries have undertaken the costly process of searching and evaluating of possible investments for individual investors In the model of Boyd and Prescott (1986), the presence of competing financial intermediaries could support the Pareto efficiency allocation of capital in the economy

Another channel through which a bank-based system could accelerate economic growth is its positive effect on investments Banks encourage investments through their financial supports

to new firms’ establishments According to Stulz (2000), stage-financing, i.e additional finances will be provided when the investment project develops positively, is the optimal mechanism to finance new entrepreneurs This is because the entrepreneur would have better information about his own project than any potential investors and this asymmetric information problem creates a risk for any investors intending to provide finance for the project The solution is that the investors might provide initial funds so that the entrepreneur might start the project at small scale then if they discover that the project is profitable they

would provide additional finances Stulz (2000) argued that securities market is not suited to this funding mechanism due to its nature while banks could allow this by providing loans to the entrepreneur and provide renewal or expansion of the contract if the project develops positively and profitably

well-Advocates of bank-based view not only emphasize the positive role of banks but also stress the adverse effects of a market-based system on economic growth Stiglitz (1985) emphasized that the asymmetric information problem occurs in the market-based system would bring negative impact to corporate governance A downgrade in corporate governance means that the firms in the economy are not doing well which in turn implied that the capital mobilized into production is not utilized efficiently and thus impede economic growth In a market-based system, firms are owned by many shareholders who legally have the right to control and replace the management team if the firm is not performing well However Stiglitz (1985) argued that in reality, to make such decision stockholder must spend resources in collecting information to evaluate the performance of the firm and the management team This is a costly process and since assuring good governance of the firm is similar to a public good to shareholders, there will be inadequate resources and efforts dedicated to ensure it

Allen and Gale (1999) developed a model in which securities market could promote economic growth by facilitating the formation of new industries and new technology changes Their argument is that when new products or innovative technologies are introduced, there is “diversity of opinion” among investors regarding the efficiency of these products or technologies In the model of Allen and Gale (1999), securities market is superior

to financial intermediaries like banks in the case where there is “diversity of opinion” among investors In such situation, investors in the securities market will individually base on their own beliefs to gather information and make judgement about the effectiveness of the new product or technology Finally, those who are optimistic about the new product or technology will provide the necessary finances In contrast, Allen and Gale (1999) had shown in their model that the ability of financial intermediaries like banks to provide finances to innovative projects is limited when there is “diversity of opinion” among investors The reason is that in their model financial intermediaries are group of investors in which one of them will pointed

as the manager and when there is “diversity of opinion” among investors, the disagreement

on which information to be collected to evaluate the effectiveness of innovative projects arise Thus even if the manager of the financial intermediary is optimistic about the project, the disagreement from other investors within the intermediary would prevent the decision to provide finances to be executed

Besides their support for a market-based system, the market-based view also stresses the short-comings associated with a bank-based system According to Rajan and Zingales (2001), although lending and investment activities in a bank-based financial system are not always inefficient, there are many problems associated with this system Firstly, the poor price signal

in a bank-based system might lead to misallocation of funds This is because banks possess some monopoly power to the borrowing firms and this monopoly power allows banks to form

a long-run relationship with these firms by providing them “below-market rate” in the run or in the firms’ hard time and then charging them “above-market rate” in the long-run or when the firms are doing well This deviation from market rate could lead to inefficient allocation of funds since in this case, the cost of undertaking investment projects by borrowing firms only depend on their loan contracts with banks For instance, with long-term access to banks’ credit, firms might continuously making investments even when they are having problems with their cash flows

short-Secondly, Rajan and Zingales (2001) shown that banks may cooperate with borrowing firms

to secure returns from their loans and these actions would prevent the optimal resource allocation in the economy For instance, when there is catastrophic shocks to the economy, banks might not let their borrowing firm becoming bankrupt easily; instead they would provide additional finances to save those firms These distortions of banks might against the natural structural adjustment of the economy in which real financial distressed firms get bankrupt while the healthier and more efficient firms survive Thirdly, a bank-based system might not be able to accelerate economic growth since it is only well-suited for financing traditional industries where physical capital are used intensively Banks could easily provide loans to firms in these industries since the borrowing firms could provide collateral by their tangible assets Rajan and Zingales (2001) argued that a bank-based system might not be able

to support new and innovative industries since the monopoly power over their borrowing firms allows banks to extract profits from these firms this could impede the effort of these firms in creating innovative products This is not the case with traditional industries where the banks could not extract much profit because these industries are already well-understood

Other Views in Financial Structure Theory

The two later theories about financial structure get rid of the bank-based and market-based debate and focus on other factors in explaining growth Specifically, the financial services view originated from Merton and Bodie (1995) and Levine (1997) puts less attention on whether the financial system is bank-based or market-based Instead, it reconciles both views

by arguing that it is not the type of the financial system but rather the financial services provided by the system is crucial to economic growth Besides this view, La Porta et al (2000) advocate the law and finance view in which the role of the legal environment in creating a growth promoting financial system is stressed

The functional services view proposed by Merton and Bodie (1995) is based on two arguments First, the form of financial institutions fluctuates more than the financial services provided by them over time Second, the financial institution forms depend on the financial functions Within this view, six functions provided by financial institutions include facilitating exchange, pooling of economic resources, mobilizing those resources between savers and borrowers, managing risks, providing price signals to assist decision-making in the economy, resolving the incentive problems Merton and Bodie (1995) then argued that “the functional perspective views financial innovations as driving the financial system toward the

goal of greater economic efficiency” Thus for the goal of economic growth, the type of the financial system is not important than how well are the financial functions provided by the system Levine (1997) also proposed the functional approach when studying the relationship between growth and finance He argued that the functions provided by the financial system including facilitating exchange, pooling and mobilizing capital, risk management, corporate governance are the same across countries However, what matters in the relationship between economic growth and finance is the quality of providing those financial functions

Similarly, the law and finance theory proposed by La Porta et al (2000) also get rid of the debate between bank-based and market-based theories and argued that the legal system is crucial in creating a growth-enhancing financial system La Porta et al (2000) argued that when creditors and minor stockholders provide finances to firms, the yields on their investments will be affected by expropriation activities initiated by the controlling shareholders Those expropriation activities include selling firm’s assets or products to other firms owned by these controlling shareholders with low prices, putting their relatives into the firm’s management team, paying higher than usual for the management board If the legal system does not provide enough protection for outside investors such as creditors and minor stockholders, there would be high probability that these expropriation actions by the controlling stockholders will take place In turn, this would not only lower the performance of the firm but also discourage the incentive of making investment by outside investors Thus, the implication of the law and finance view by La Porta et al (2000) is that a strong legal environment and enforcement mechanism in protecting outside investors would open up more opportunities for firms to attract investments from external sources and increase the efficiency of using capital from those sources through corporate governance enhancement

2.2.5 Endogenous Growth Theory

Endogenous growth theory is originated from the seminal studies of Romer (1986) and Lucas (1988) In original studies advocating this economic school of thought, the role of financial development is often omitted or not discussed directly Within the framework of this theory, the role of financial system in inducing economic growth only get into attention after new endogenous growth models were introduced in 1990s These new models have incorporated financial activities in explaining economic growth The notably models in this strand are introduced in Greenwood and Jovanovic (1990), Bencivenga and Smith (1991), King and Levine (1993b), and Pagano (1993)

In the endogenous growth model developed by Greenwood and Jovanovic (1990), there is a bidirectional causality relationship between economic growth and financial development They argued that the establishment of financial institutions is costly thus in initial development stage of the economy, there is merely no institutions When the economy has reached its intermediate development stage, large financial institutions appear Then these institutions induce further economic growth through various financial activities According to Greenwood and Jovanovic (1990), the development of financial institutions could affect economic growth through two channels First, financial institutions could pool savings and allocate to investment projects which provide higher returns Without financial institution, when individual investors face two types of investment projects, i.e liquid projects with low returns and illiquid projects with high returns, they might choose the latter as they would not like to tie their funds in a longer period of time and expose to more risks Second, financial institutions could diversify their portfolios more efficient than individual investors thus the investments of individual investors through financial institutions would be less risky In summary, the development of the financial system could affect economic growth as it could allocate capital to higher return investments in a less risky manner

Bencivenga and Smith (1991) formulated an endogenous growth model in which financial intermediaries facilitating economic growth by changing the behaviour of agents in allocating their savings and by preventing unnecessary and early liquidation of investment projects In this model, economic agents have randomly liquidity needs and to satisfy those needs in an unfavourable time, all agents will have to allocate their savings not only in productive investments but also in liquid assets which are usually unprofitable Bencivenga and Smith (1991) pointed out that the presence of financial intermediaries in the economy could eliminate such inefficient capital allocation The reason is that agents now do not have to invest part of their savings in unproductive assets anymore; instead they could deposit it to financial intermediaries In turn, those intermediaries will keep part of the deposits as reserve

to satisfy the randomly liquidity need of agents and turn the rest into productive capital by intermediary activities Thus at the aggregate level, financial intermediaries have promoted economic growth by turning unproductive assets into productive capital Bencivenga and Smith (1991) also suggested that financial intermediaries could induce growth by altering the nature of investments made by economic agents Without financial intermediaries, agents invest using their own funds and thus those investment projects might be liquidated early if the agents face liquidity shocks Conversely, if the investments are financed through

intermediary activities, such costly liquidation could not occur Bencivenga and Smith (1999) argued that financial intermediaries could provide this function because they receive deposits from a large number of agents and according to the law of large number, the demand of agents to withdraw from these deposits is anticipated Thus financial intermediaries only need

to keep some amount of reserves to prevent the liquidation of premature investment projects while still be able to satisfy the liquidity needs of agents

While other endogenous growth models focus on the accumulation of physical capital channel when studying the relationship between financial development and economic, the model established by King and Levine (1993b) suggested that financial development promote growth mainly through the productivity growth channel In their model, financial activities induce productivity growth through four schemes First, financial institutions have comparative advantage than individual agents in evaluating potential firms to identify the most innovative and productive ones Second, after being able to sort those potential firms according to their productivity, financial institutions could rely on their large scale to attract funds from savers and allocate these funds to the most promising firms Third, there is always higher risk associated with high productivity firms comparing with firms which use the old production technique This fact might put a barrier to individual investors in providing finances to innovative activities Financial institutions could eliminate this barrier and mobilize funds of individual investors to high productive firms through risk diversification Fourth, financial institutions could attract more capital to innovative and productive firms by making the information regarding the value and profits of those firms available on the market Thus, the development of the financial system would stimulate productivity growth

in the economy and through this channel economic growth will accelerate

Pagano (1993) in his AK growth model had shown that financial development could affect economic growth through three explicit channels First, during the process of mobilizing savings to investments, financial institutions will charge some fee on providing their financial services As the financial system develops and becomes more efficient, this fee will get smaller so that the proportion of savings to be channelled into investments will be higher Second, as financial institutions could obtain information about investment projects more efficient than individual investors, they could identify investment projects which return higher yields Additionally, by providing risk sharing, financial institutions could allow the funds of investors to flow into those high return but also riskier investment projects Thus,

financial development could affect growth by providing finances to highly productive investment projects Thirdly, another way in which financial development could affect economic growth is through its effect on the saving rate However, Pagano (1993) argued that this effect is uncertain Although the development of the financial system could offer better risk-sharing and lead to higher productivity investments, this could alter the saving behaviour

of consumers since in a better risk-sharing environment, consumers may save less

2.2.6 Summary of Theoretical Predictions

Table 2.1 provides the summary of predictions from previous theoretical studies on the link between economic growth and financial development

Table 2.1: Summary of theoretical predictions

2.3 EMPIRICAL REVIEW

The empirical studies on the relationship between financial development and economic growth could be classified into three large categories based on their analysis approaches: traditional cross-country analysis, panel cointegration and causality analysis, and time series analysis

2.3.1 Traditional Cross-Country Analysis

King and Levine (1993b) is one of most influential papers which use traditional panel regression analysis to examine the link between growth and finance across countries The study analyzed annual data of 80 countries covering from 1960 to 1989 by running two batches of regressions with each contains a regression on growth and three regressions on sources of growth, namely physical capital accumulation rate, ratio of investment to GDP, and efficiency enhancement For the first set of regressions, the study used the average values

of all variables in the period 1960-1989 and run the regressions on pooled data The results from these regressions showed that there is a positive association between growth and all 4 indicators of financial development Additionally, the same relationship exists between financial development indicators and the three sources of growth such as physical capital growth rate, domestic investments, and efficiency For the second wave of regressions, the study takes average value of 1960s, 1970s, and 1980s for all variables so that there are three observations for each country Also, the indicators for financial development will be replaced

in the regressions by their initial value at 1960, 1970, and 1980 respectively to check if those initial finance development levels are related with economic growth rate of latter periods or not Results from this second batch of regressions showed that the initial level of financial development have some prediction power over the economic growth of latter periods

Beck et al (2000) assess the relationship between financial intermediation and economic growth, and determine the effect of financial intermediation to sources of growth across countries from 1960 to 1995 The study employs two estimating approaches First, it took average data in the period 1960-1995 for 63 countries and run a pooled cross-country regression to determine the impact of financial development on growth and sources of growth like productivity growth, rate of capital accumulation, and savings rate Second, the study divided the period 1960-1995 into 5-year sub-periods and then used Generalized Method of Moments (GMM) panel estimating technique to explore the relationship between finance and

growth and sources of growth This technique has comparative advantages over the traditional cross-country regression in avoiding biases brought from the endogeneity of explanatory variables, and missing of country-specific effects in the model Results from both estimating methods showed that there is a positive relationship between financial development and economic growth and productivity growth These results are robust regardless of the financial development indicator being used in the estimation However, less robust support is found for the relationship between financial intermediation and capital accumulation rate Specifically, results from the pure cross-country regression vary with different indicators of financial development Similarly, the link between financial development and savings rate is also weak Both pure cross-country regression and GMM panel regression return ambiguous results as different proxy for financial development is used

in the regression

There are two disadvantages associated with previous empirical studies which employ the cross-country analysis approach Firstly, they do not take into account the time series properties of the data Secondly, results from these studies could not shed light on the causality relationship between economic growth and financial development across countries,

as already argued in Arestis et al (2001) and Abu-Bader and Abu-Qarn (2008a)

2.3.2 Panel Cointegration and Causality Analysis

Traditional cross-country studies are usually criticized for the old and inappropriate method

in answering questions related to causality between growth and finance With the development of new panel cointegration and causality analysis techniques recently, a few papers have made an attempt of employing those new methods on studying the causality relationship between economic growth and financial development in a more satisfactory manner

Based on data of 109 developed and developing countries from 1960 to 1964, Calderon and Liu (2003) use the Geweke decomposition test to determine the causal relationship between financial development and economic growth, and between financial development with sources of growth such as growth of capital per capita, and productivity growth in those countries Besides the whole sample, the study also divided the countries into two other subsamples which are 87 developing countries and 22 developed countries and then for each sample they run Geweke decomposition tests on two panels which are a panel of five-year

periods and a panel of ten-year periods By comparing the results between the 5-year panel and the 10-year panel, they observed that the supply-leading relationship dominates at the developing countries in two panels and become stronger in the 10-year panel while for developed countries, the demand-following relationship is more crucial in the 5-year panel while the supply-leading relationship dominates in the 10-year panel Thus as the interval becomes longer the effect on growth of financial development is stronger To verify the

“stage of development” hypothesis, the study examined the subsample of low and middle income countries in two periods: the early stage of development period 1960-1979 with four 5-year observations and the developed period 1980-1994 with three 5-year observations They split the two periods at 1979 because many countries began financial reforms around this year and they hypothesized that this is a split of the early development stage and the developed stage of these countries Results from the Geweke decomposition tests shown that

at low and middle countries, the supply-leading relationship dominates at early development stage but in the developed stage its importance decreases and the simultaneous causal relationship become more relevant Thus the study provides little support for the “stage of development” hypothesis and the study explained that these countries did not become fully developed in the 1980-1994 period and hence the demand-following relationship could not become dominant On the relationship between financial development and sources of growth, Calderon and Liu (2003) observed that although financial development induces both physical capital accumulation and productivity growth, the latter channel is stronger

Christopoulos and Tsionas (2004) explored finance-growth nexus in 10 developing countries from 1970 to 2000 using both time series analysis method and new panel data analysis techniques The study followed Luintel and Khan (1999) on choosing the ratio of total bank deposits to GDP as the proxy for financial development Besides economic growth and financial development, the study also includes two more variables in its model which are investments share in GDP and inflation rate Both Johansen cointegration test at individual countries and panel cointegration test revealed that there exists a long-run relationship among growth, financial development, and the other two variables in the multivariate model By normalizing the cointegration equation with economic growth as the dependent variable, financial development is found to have significant positive impact on growth for the panel cointegration equation while for individual countries the same results are observed except the case of Mexico and Ecuador Short-run causality tests based on vector error correction models at individual countries showed that there is no short-run effect running from financial

development to economic growth except Dominican Republic Additionally, panel causality test based on the panel vector error correction model also confirm that no short-run causality running from financial development to growth is detected Based on the long-run and short-run tests, Christopoulos and Tsionas (2004) concluded that policy makers should be aware that financial policies might not induce economic growth instantly and they should expect such effect in the long-run instead

While studies using these new panel analysis techniques could exploit the time series properties of data and could answer the question related to causality in a more satisfactory manner, they might not be useful for policy making purpose The reason is that although results from these studies might be correct in general, they might be different in individual countries due to differences in financial development level, structure of the financial system, and other macroeconomic characteristics According to the stage of development hypothesis

of Patrick (1966), those differences could lead to different pattern of the relationship between economic growth and financial development across countries

2.3.3 Time series analysis

As suggested by many empirical studies in the field such as Demetriades and Hussein (1996), due to differences in economic environment, political and other characteristics, the relationship between growth and financial development tends to be country-specific and should not be generalized across countries In that context, the time series based studies might provide a more accurate picture about the relationship at each country Within the time series literature on the relationship, there are two distinct analysis approaches: vector error correction model (VECM) approach and vector autoregression (VAR) approach

VECM Approach

Perhaps this is the most frequently used approach in the time series analysis literature on the relationship between economic growth and financial development This approach usually start by employing some kind of unit root or stationarity tests such as Augmented Dickey-Fuller (ADF) test (Dickey and Fuller, 1979) to determine the stationarity properties of the variables After that, cointegration test using the procedures of Engle and Granger (1987) or Johansen and Juselius (1990) to determine the long run relationship among the variables Finally, Granger causality test based on the VECM framework will be applied to determine both the short-run and long-run causality between the variables

Various studies have employed the approach summarized earlier to discover the causality relationship between financial development and economic growth Demetriades and Hussein (1996) followed the VECM approach to examine the causality relationship between economic growth and financial development using annual data of 16 countries The results from the study shown that there are two countries, Spain and Pakistan, in which there is no cointegration relationship However, Demetriades and Hussein (1996) argued that such results do not imply that there is no long run relationship between growth and financial development but the relationship might exist in non-linear form or the employed proxy could not reflect the true financial development level of that country Additionally, there is little evidence which supports the view that finance causes growth or the reverse; instead more support is given to the view that there is bi-directional relationship between growth and financial development Interestingly, countries which possess this two-way relationship, such

as Korea and Thailand, are mostly successful with their financial reforms Thus the empirical results provide evidence that financial reforms might not only contribute to economic growth but also to the financial deepening process In general, as the causality relationship differs among countries, Demetriades and Hussein (1996) concluded that the results are very country specific and thus they criticized cross-country studies for examining the countries homogeneously

Thangavelu and Ang (2004) investigated the causality relationship between economic growth and financial development in Australia using quarterly data from 1960 to 1999 The study used both bank-based and market-based measures to represent financial development Specifically, the ratio of bank deposit to GDP and ratio of bank credits provided to the private sector to GDP are used as indicators for intermediary market while the ratio of equity turnover volume to GDP is adopted to represent equity market The study uses central bank lending rate and interbank lending rate to represent households’ propensity to save which is the only controlling variable in the studied model Results from the VECM version of Granger causality test against their tri-variate model shown that causality is running from economic growth to the development of intermediary market while for the case of equity market, the reverse causality running from its development to economic growth is found Thangavelu and Ang (2004) argued that these results are due to the facts that since 1970s and 1980s, Australian Federal Reserve Bank had administered the financial sector in a market-oriented manner which led to an intensive development of the stock market

Chang and Caudill (2005) verified both the supply leading and demand following hypotheses about the relationship between growth and finance for the case of Taiwan based on annual data from 1962 to 1998 The country had achieved dramatically economic growth since the implementation of financial liberalization policy in 1980s They employ the Granger causality test based on the VECM framework against the multivariate model including real GDP per capita, exports, imports, and ratio of M2 to GDP to verify whether the evolvement

of the financial system had promoted this extraordinary economic growth progress or not The study found a one way causality relationship running from financial development to economic growth which supports the supply leading hypothesis of Patrick (1966)

Ang and McKibbin (2007) assess the growth-finance nexus in Malaysia from 1960 to 2001 using Granger causality test based on the VECM approach Differing from previous empirical studies, Ang and McKibbin (2007) construct an index representing financial development using the logarithm values of the ratio of liquid liabilities M3 to GDP, the ratio of commercial banks’ assets to total assets of both commercial banks and central bank, and finally the ratio of total credit to the private sector to GDP The results are tested for sensitivity by replacing this financial development index with the three original proxies for financial development Also, to get rid of the omitted variables problem, the study also incorporates real interest rate and an index for financial liberalization into the model besides economic growth and the financial development index Additionally, to take into account the macroeconomic shocks, the study also includes 5 dummy variables representing these shocks into the model The study found no short-run causality relationship between growth and finance in all 4 models using 4 different proxies for financial development Both weak exogeneity test and strong exogeneity test show that growth induces financial development in the long run, but there is no evidence supporting the reverse direction

Abu-Bader and Abu-Qarn (2008a) employed the causality test based on the VECM approach

to study the relationship between economic growth and financial development from 1960 to

2001 using annual data As financial development is a complex concept, the study uses several indicators to capture different aspects of the financial sector These measures include the ratio of broad money measure M2 to GDP, ratio of M2 minus currency to GDP, ratio of bank credits to the private sector to GDP, and ratio of credits not provided to financial sector

to total credits to capture many aspects of the financial system Their model also includes the ratio of investment to GDP to form a tri-variate model Results from causality test provide

strong evidence supporting the two way relationship between economic growth and financial development as this pattern is found on all four indicators of financial development Also, investment is found to be the indirect channel through which financial development promote economic growth

Odhiambo (2008) uses annual data of Kenya from 1969 to 2005 to feed into their tri-variate model which includes real GDP per capita, savings, and ratio of broad money M2 to GDP to explore the relationship between economic growth and financial development in Kenya The study also follows many of the previous empirical studies by employing the VECM based causality test for their analysis However, result from the study does not provide support for the supply leading hypothesis as usually seen in previous research Instead, causality is found running from economic growth to savings and also from savings to financial development Based on the result, Odhiambo (2008) argued that any statement supporting the growth promoting role of financial development should be provided with caution

VAR Approach

In this approach, the long run cointegration relationship among the variables is not required

as in the VECM approach Instead, after the stationarity properties of the variables have been determined by stationarity and unit root tests, the multivariate vector autoregressive (VAR) model containing all variables will be formulated After that, traditional Granger causality test or various types of causality tests such as the tests developed by Hsiao (1982), Toda and Phillips (1993), Toda and Yamamoto (1995) will be employed to derive causality relationships among the variables

Luintel and Khan (1999) was one of the first time series studies which use multivariate VAR model in examining the relationship between growth and financial development Luintel and Khan (1999) then employed a multivariate VAR model containing 4 variables which are growth, financial development represented by the ratio of total bank deposits to GDP of previous period, investment, and finally real interest rate They used ADF test (Dickey and Fuller, 1981) and KPSS test (Kwiatkowski et al, 1992) to addressing the stationarity problem, then they used weakly exogenous test and Toda and Phillips (1993) causality tests to determine the long run causality between finance and growth The interesting result from Luintel and Khan (1999) is that in all 10 countries in the sample, there exist a two way

causality relationship between growth and financial development The study thus provided strong evidences to support the “bi-directional causality hypothesis”

Shan, Morris, and Sun (2001) employed a multivariate VAR model and applied the causality test developed by Toda and Yamamoto (1995) to determine the causal relationship between economic growth and financial development for nine OECD countries and China in the period 1960s to 1998 The study is distinct from previous time series researches in the literature by adding into the model many controlling variables like total factor productivity, ratio of trade to GDP, ratio of capital expenditure to GDP, and most importantly the stock market price index to represent stock market development Shan, Morris, and Sun (2001) run one causality test for the whole period 1960s to 1998 and another causality test for the sub-period mid-1980s to 1998 to check whether the structural economic changes in highly developed countries in mid-1980s might alter the results or not When using the whole period 1960s to 1998, no causality relationship between finance and growth is detected except for

US and Italy where growth is found to have Granger cause on financial development When using the sub-period from mid-1980s to 1988, bi-directional causality relationship is found in

5 countries while for the rest 5 countries, there exists causality running from growth to financial development in 3 countries Thus the study provides no support to the “supply-leading” hypothesis while some support are provided for the bi-directional causality” hypothesis One interesting result from the study is that there exists a bi-directional causality relationship between the proxy for stock market development and economic growth in 9 OECD countries but not China This result suggests that the development of the stock market

is an important contribution factor to economic growth at highly developed countries

Abu-Bader and Abu-Qarn (2008b) examined the growth and financial development nexus in six Middle Eastern and North African (MENA) countries from 1960 to 2004 The study employed four proxies for financial development which are commonly used in the literature

To avoid omitted variables problem, the study incorporated investment share in GDP, and ratio of government expenditure to GDP into their multivariate model Results from Toda and Yamamoto causality test provide robust evidence supporting the supply leading role of financial development in promoting economic growth as this phenomenon is found at five per six countries Wolde-Rufael (2009) also employed Toda and Yamamoto causality test to determine the causality relationship between growth and financial development for Kenya using annual data from 1966 to 2005 Additionally, the study uses nearly the same four

financial indicators to represent financial development as in Abu-Bader and Abu-Qarn (2008b) However, for controlling purposes, the multivariate model of the study chooses exports and imports Results from Toda and Yamamoto causality test show that two-way causality exists between growth and financial development which support both the supply-leading and demand-following hypotheses

Observed that many developing countries in Sub-Saharan Africa (SSA) had implemented various packages of financial reforms in 1980s, Gries et al (2009) attempted to verify the effect of these structuring changes to the economic growth process by both Hsiao-Granger causality test and VECM based causality test on annual data of those countries ranging from

1960 to 2004 As the theoretical linkage between financial development and trade has been supported by a number of studies, trade openness is also added into multivariate model as the third variable To represent financial development, the study follows the approach of Ang and McKibbin (2007) on using principal component analysis to construct an index for financial development from common financial measures including ratio of commercial bank assets to total assets of commercial bank and central bank, ratio of liquid liabilities to GDP, and ratio

of private credit provided by deposit money banks to GDP The empirical results are also checked for robustness by re-run all econometric tests using liquid liabilities to GDP in replacing of the constructed index of financial development While previous studies on developing countries found significant support for the supply-leading hypothesis (finance causes growth), the study finds no clear pattern of the relationship between growth and financial development at SSA countries, and the causality linkage is not strong across these countries Gries et al (2009) argued that the deviation of the study’s empirical results might

be due to the low development level of the financial systems at SSA countries comparing with other developing countries as the relationship between growth and finance seems to depend on the financial development level

In general, time series based studies, either using the VECM or VAR approach, are subjected

to a few limitations Firstly, many of them use only two or three different indicators of financial development This could generate biased results since financial development is a complex concept Secondly, except for the study of Shan, Morris, and Sun (2001), other studies following this approach only employ three (most commonly) or four variables in their multivariate model and thus might be subjected to the omitted of variables issues While the study of Shan, Morris, and Sun (2001) focus on the case of nine OECD countries and China,

those limitations generate the need for a time series based analysis at a developing countries like Vietnam which use several different indicators of financial development as well as employ a multivariate model which includes larger number of variables

2.3.4 Previous empirical studies about Vietnam

Before closing out the discussion about empirical studies in this field, we should review previous researches about Vietnam In the current literature on the relationship between finance and growth in Vietnam, most of the studies are qualitative based (Nguyen, 2011) Only recent papers by Anwar and Nguyen (2009), and Nguyen (2011) employ quantitative analysis methods Anwar and Nguyen (2009) examined the finance and growth nexus in Vietnam at provincial level from 1997 to 2006 based on the endogenous growth framework Their panel analysis using the generalized method of moments (GMM) estimator showed that financial development causes economic growth in Vietnam The robustness of their findings

is confirmed after different measurements of financial development were used in the growth regression Nguyen (2011) examines the relation between financial development and economic growth in Vietnam at provincial level in nearly the same period, from 1997 to

2004 The paper also determines the channels through which financial development could affect economic growth It follows the approach of endogenous growth literature in employing an AK type growth model Result of generalized least squares (GLS) panel analysis in the paper show that financial development has significant and positive impact on economic growth, efficiency in mobilizing savings, growth in investment, information technology, and the growth in productivity

Again, those previous empirical studies for the case of Vietnam also subjected to the question

of methodology in determining the long run and causality relationship between economic growth and financial development As argued in Abu-Bader and Abu-Qarn (2008a) and Lee and Chang (2009), the traditional panel analysis technique is not appropriate to examine the long run cointegration or causality relationship

2.3.5 Summary of findings from previous empirical studies

To have a broad comparison among various empirical studies which were using different methods against data from different countries, Table 2.2 below summarizes their findings

Table 2.2: Findings from previous empirical studies

King and

Levine (1993b)

Annual data,

80 countries, 1960-1989

Panel data analysis - Positive association between economic

growth and financial development

Granger causality test based on VECM

- Support the bi-directional causality hypothesis, especially at countries which were previously successful with financial reforms

Luintel and

Khan (1999)

Annual data,

10 countries, 1951-1995

Granger causality test based on VAR

- Strong support to the bi-directional causality hypothesis

Beck et al

(2000)

Annual data,

63 countries, 1960-1995

Panel data analysis

- Positve relationship between economic growth and financial development

- Financial development has strong positive relationship with productivity growth but week association with capital accumulation and savings rate

Shan, Morris,

and Sun (2001)

Annual data, China and 9 OECD countries, 1960-1998

VAR, Toda and Yamamoto causality test

- Little support for the bi-directional causality hypothesis (found on 5/10 countries)

Calderon and

Liu (2003)

Annual data,

109 countries, 1960-1994

Geweke decomposition test

- For developed countries: the following hypothesis dominates

demand For developing countries: support the supply-leading hypothesis

- Little support for the stage of development hypothesis

Panel cointegration test and causality analysis

- Long-run relationship exists among growth, financial development and other macro-variables

- No short-run causality from financial development to growth found

Table 2.2: Findings from previous empirical studies (continued)

Thangavelu and

Ang (2004)

Quarterly data, Australia, 1960-1999

Granger causality test based on VECM

- Between growth and equity market development: support supply-leading hypothesis

- Between growth and banking market development: support demand-following hypothesis

Chang and

Caudill (2005)

Annual data, Taiwan,

1962 to 1998

Granger causality test based on VECM - Support the supply-leading hypothesis

Ang and

McKibbin

(2007)

Annual data, Malaysia, 1960-2001

Granger causality test based on VECM

- Little support for the supply-leading hypothesis: long-run causality between growth and financial development is found, but there's none short-run causality

Abu-Bader and

Abu-Qarn

(2008)

Annual data, Egypt, 1960-2001

Granger causality test based on VECM

- Support the bi-directional causality hypothesis

Odhiambo

(2008)

Annual data, Kenya, 1969-2005

Granger causality test based on VECM

- Little support for the demand-following hypothesis

Abu-Bader and

Abu-Qarn

(2008b)

Annual data, 6 MENA countries, 1960-2004

VAR, Toda and Yamamoto causality test

- Support the supply-leading hypothesis

Anwar and

Nguyen (2009)

Provincial data, Vietnam, 1997-2006

Panel data analysis - Financial development has significant

positive impact to economic growth

Gries et al

(2009)

Annual data, 16 SSA countries, 1961-2004

VAR, Hsiao-Granger causality test

- No causality pattern found, might be due to the low development level of the financial systems at those countries

Wolde-Rufael

(2009)

Annual data, Kenya, 1966-2005

VAR, Toda and Yamamoto causality test

- Support the bi-directional causality hypothesis

Nguyen (2011)

Provincial data, Vietnam, 1997-2004

Panel data analysis - Significant and positive effect of

financial development to growth

2.4 CONCEPTUAL FRAMEWORK

The theories on the relationship between financial development and economic growth is summarized and demonstrated by the conceptual framework below According to the framework, financial development could promote economic growth via two explicit channels Firstly, the capital accumulation process is accelerated with the present of financial institutions and their provision of financial services within the economy Secondly, the efficiency associated with the utilization of capital is also improved Conversely, economic growth also imposes effects on financial development by generating the demand for additional financial services as well as accumulating the necessary resources for the establishments of new financial institutions

Financial

Accelerate capital accumulation

Enhance efficiency of capital utilization

Accumulate necessary resources for establishments of financial institutions

Generate additional demand for financial services