Determinants of bank capital structure the case of vietnamnese commercial bank system

Using panel data analysis like Fixed Effects Method FEM and Random Effects Method REM, the study investigates the effect of size, growth, collateral value, profitability, dividend policy

VIETNAM- NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

DETERMINANTS OF BANK CAPITAL

STRUCTURE: THE CASE OF VIETNAMESE

COMMERCIAL BANK SYSTEM

BY PHAM TUAN ANH

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

HO CHI MINH CITY, JANUARY 2013

VIETNAM- NETHERLANDS

PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

DETERMINANTS OF BANK CAPITAL STRUCTURE: THE CASE OF VIETNAMESE

COMMERCIAL BANK SYSTEM

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

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

ACKNOWLEDGEMENT

I am grateful to all those who help me to complete the thesis I would like to thank the Vietnam-the Netherlands Programme for giving me permission to carry out the thesis after Thesis Research Design (TRD) defense and providing many precious researches through an account ofEUR library website and others

Especially, I am deeply thankful to my supervisor Dr Cao Hao Thi from Industrial Management Department of Ho Chi Minh City University of Technology for his instructions, suggestions and encouragements during the implementation of the thesis

Moreover, I would like to express my deep gratitude to my family, friends and colleagues for all their support and valuable hints in order to help me to complete the thesis

.-ABSTRACT

The study aims to identify determinants of Vietnamese banking capital structure Using panel data analysis like Fixed Effects Method (FEM) and Random Effects Method (REM), the study investigates the effect of size, growth, collateral value, profitability, dividend policy and business risk determinants to capital structure of commercial banks in Vietnam from 2007 to 20 11 The main findings of the study are that Fixed Effects Model (FEM) is the most appropriate model in explaining the banking capital structure; furthermore, size, growth, collateral and profitability variables are statistically significant and have the expected sign with the book leverage variable However, risk and dividend variables are not statistically significant

Based on reality situations, it is widely believed that the capital structure of Vietnamese banking system is not really good and does not meet the international standard at present The system does not have enough equity so as to prevent risks from the negative externalities and unexpected change of business environment Therefore, the study is going to recommend some policy implications to improve the capital structure situation by adjusting the determinants Moreover, the study also makes valuable instructions to banking managers of each commercial bank According to specific reality situation of each commercial bank, they will have right decisions to increase or decrease book leverage so as to not only maximize the own commercial bank's value but also avoid facing up bankruptcy events

Keywords: capital structure, book leverage, commercial bank, Fixed Effects Method (FEM), Random Effects Method (REM)

TABLE OF CONTENTS

Acknowledgement i

Abstract ii

Table of Contents iii

Abbreviation vi

List of Table vii

L 1St 0 fF' Igure •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• VIII

CHAPTER 1 INTRODUCTION 1

1.1 Problem Statement 1

1.2 Research Objectives 6

1.3 Research Questions 6

1.3.1 Main question 6

1.3 2 Sub questions 7

1.4 Justifications ofthe study 7

1.5 Scope of the study 7

1.6 Organization ofthe study 8

CHAPTER 2 LITERATURE REVIEW 9

2.1 Theoretical literature 9

2.1.1 Modigliani and Miller (M&M theory, 1958) 9

2.1.2 Agency theory 10

2.1.3 The pecking- order theory 11

2.1.4 The static trade-offtheory 12

2.2 Empirical literature 12

2.2.1 Foreign empirical literature 12

2.2.2 Vietnamese empirica1literature 22

2.3 Research Hypothesis 27

2.4 Conceptual Framework 27

2.5 Chapter Summary 28

CHAPTER 3 DATA AND RESEARCH METHODOLOGY 29

3 .1 Research Methodology 29

3.2 Data 35

3.3 Chapter Summary 37

CHAPTER 4 FINDINGS AND DISCUSSION 38

4.1 Descriptive Statistics 38

4.2 Panel regressions results 41

4.3 Comparison with previous studies 48

4.4 Chapter Summary 50

CHAPTER 5 CONCLUSIONS AND POLICY IMPLICATIONS 51

5 .1 Conclusions 51

5.2 Policy Implications 52

5.2.1 Policy implications for total Vietnamese banking system 52

5.2.2 Policy implications for specific Vietnamese commercial banks cases 55

5.2.2.1 Policy implications for high leverage commercial banks 55

5.2.2.2 Policy implications for low leverage commercial banks 56

5.3 Limitations and Further Studies 57

5.3.1 Limitations 57

5.3.2 Further Studies 58

References 59

Appendices 63

-ABBREVIATION

BCBS: Basel Committee on Banking Supervision BL: Book Leverage

CAR: Capital Adequacy Ratio

FEM: Fixed Effects Method

GDP: Gross Domestic Product

GMM: Generalized Method of Moments

REM: Random Effects Method

•

•

LIST OF TABLE

Table 1.1 Quantity of commercial banks in Vietnam 3

Table 2.1 Definition ofVariables 15

Table 2.2 Foreign empirical studies in researching the bank capital structure 20

Table 2.3 Empirical studies in researching the Vietnamese capital structure 25

Table 3.1 Definition ofVariables 31

Table 3.2 Classifying banks in the sample by certain asset level on December 31, 2011 36

Table 3.3 Financial items of the sample and total Vietnamese banking system on December 31, 2011 36

Table 4.1 Summary statistics of each variable 38

Table 4.2 Correlation matrix 40

Table 4.3 FEM and REM regression results 42

Table 4.4 Hausman test 43

Table 4.5 Likelihood ratio test 44

Table 4.6 Summary of research results 50

LIST OF FIGURE

Figure 1.1 Capital Adequacy Ratio (CAR) of developing countries 5

Figure 2.1 The determinants of bank capital structure 28

Figure 3.1 Research process of the study 29

Figure 4.1 Total asset in the sample from 2006 to 2011 39

Figure 4.2 Total charter capital in the sample from 2006 to 2011 47

CHAPTER 1 INTRODUCTION

1.1 Problem Statement

Based on Ross, Westerfield and Jaffe (2010), capital structure is the combination of debt and equity Each company should choose the suitable debt-equity ratio (capital structure) in order to maximize the value of its company When the capital structure problems are mentioned in any researches, it is certainly that the famous theory Modigliani and Miller (M&M theory, 1958) are applied in these researches According to M&M theory, in the perfect capital markets (case 1 of M&M proposition 1 ), choosing any capital structure do not affect the value of specific company However, if all assumptions in M&M proposition 1 are held except taxes and costs of financial distress, the value of the specific company will be affected by choosing the capital structure because of the tax shield benefits and costs of financial distress After the capital structure theory of Modigliani and Miller, many theories about capital structure are in tum released; in particular, agency theory was created by Jensen and Meckling (1976); pecking - order theory and static trade-off theory was developed by Myers (1984) These theories have indicated that capital structure affects not only the value of each enterprise but also the stability in business Therefore, there have many authors who research about the capital structure They also try to find out the independent variables that affect the enterprises capital structure; from that point, each enterprise have ability to choose the appropriate capital structure in order to maximize the value of the firm and stabilize the business operations

Capital structure of each company is usually measured by the book leverage It is the basic formulation It is calculated by debt-to-equity or debt-to-asset, with asset

is equity plus debt Commercial banks access the primary debts by attracting the money from depositors The specific characteristic of commercial bank business is

to trade in money; thus, equity of commercial banks always accounts for small proportion of total asset Moreover, in the banking industry, they also use Capital Adequacy Ratio (CAR), which is the other proxy to measure the capital structure of

each commercial bank (Basel Committee on Banking Supervision, 2006) CAR has been mentioned in Basel Accords, which are issued by Basel Committee on Banking Supervision (BCBS) Basel Accords have been built and developed in order to control the risk from the banking system and prevent the collapse of commercial banks in reality CAR is calculated by equity over asset; however, unlike book leverage, CAR emphasizes the different risk level of each asset In general, if commercial banks have low book leverage or high CAR, then they have not ability to take full advantage of tax shield benefit Conversely, if commercial banks have high book leverage or low CAR, then they have to face up the high financial distress cost such as the bankruptcy cost

Vietnam is still a developing country Since Doi Moi (economic reforms in 1986), Vietnamese economy has changed from planned-economy into socialism-oriented market economy The financial sector, especially commercial bank system still have

a rapid development both quality and quantity The quantity of Vietnamese commercial bank system increases from 9 in 1991 to 94 in 2009 This leads to the increase in competitive level of Vietnamese banking system With the high competitive level in Vietnamese commercial bank system at present, each bank has

to perform the right policy to maximize the value and reduce the business risk Among them, they need to choose the appropriate capital structure in order to carry out the purpose Table 1.1 presents quantity of commercial banks in Vietnam from

1991 to 2009

The Vietnamese banking system has a rapid development; in addition, it contributes

to economic growth and living standard improvement However, it has been exposed many weaknesses and shortcomings; from that point, they have caused the unsafe banking operations and unstable macroeconomic situation Therefore, the Government and State Bank of Vietnam have advocated restructuring the Vietnamese commercial bank system by approving the project, which is namely

"Orientations and Policies to Restructure Vietnamese Banking System in

2011-20 15" Based on the project, the State Bank of Vietnam has recognized weaknesses and shortcomings of Vietnamese banking system Among them, the financial capacity of Vietnamese banking system is still limited According to the 14112006/ND-CP decree that was signed by Vietnamese Prime Minister on November 22, 2006, each commercial bank has to meet the legal capital requirement, minimum 3.000 billion VND in the end of 2010 However, based on the information from the project, there have 3 commercial banks that do not meet

•

the requirement to 2011 yet; moreover, there are 30 commercial banks whose charter capital are smaller than 5.000 billion VND (equivalent 240 million USD) in the end of 20 11

Moreover, Vietnamese commercial bank system is collapsed easily from the negative externalities and unexpected change of business environment such as global financial crisis in 2009 and unstable macroeconomic situation of Vietnam in

2011 because the capital structure of Vietnamese banking system does not meet the international standard Based on the information from the project, the average CAR

of Vietnamese commercial bank system was 11.85% at September 30, 2011, especially the average CAR of State-Owned Commercial Banks was only 8.49% Meanwhile, the average CAR of Foreign-Owned Commercial Banks is very high (28.58%) If the Foreign-Owned Commercial Banks was excluded from the Vietnamese commercial bank system, then the real average CAR of domestic commercial banks was only 11.13% The Financial Management and Analysis of Projects, issued by Asian Development Bank (ADB) in 2005, has recommended that the minimum CAR of Asian Development Bank's Developing Member Countries (ADB DMCs) should be 12% because of loose banking regulations and supervisions, the expansion of directed lending It is realized that the CAR of many commercial banks does not meet the ADB standard, especially State-Owned Commercial Banks Besides, in reality, CAR of Vietnamese banking system is much lower than many other developing countries Other countries which are in the Southeast Asia with Vietnam such as Indonesia, Philippines, Malaysia and Thailand have the safe commercial bank system, with the CAR are very high and meet the international standard well Figure 1.1 compares the CAR of Vietnam in 2011 with the one of other developing countries

CAR

10.0%

5.0%

0.0%

Figure 1.1: Capital Adequacy Ratio (CAR) of developing countries

Source: State Bank ofVietnam-SBV

In short, the capital structure of Vietnamese banking system is not really good at present The system does not have enough equity in order to prevent risks from the negative externalities and unexpected change of business environment Therefore, the capital structure problem needs to be solved seriously The study desires to find determinants of bank capital structure in Vietnam; thus, they will be controlled logically in order to improve the Vietnamese banking capital structure

Moreover, according to the static trade-off (Myers, 1984 ), each Vietnamese commercial bank need to choose the own appropriate capital structure by considering the trade-off between the benefit of debts and cost of debts The benefit

of debts is the tax deductibility of interest payments and the cost of debts is the cost

of financial distress such as bankruptcy cost and the losses that the firm has to incur from the bankruptcy events In reality, it is difficult to choose the optimal capital structure because Vietnamese commercial banks almost have not ability to identify the cost of financial distress Therefore, managers of each Vietnamese commercial bank only have ability to identify the range of own appropriate leverage in order to

not only maximize the own commercial bank's value but also avoid facing up the bankruptcy events The study has no ambition to find the optimal capital structure

of Vietnamese commercial bank system, but merely to identify its determinants; thus, banking managers have ability to control these determinants in order to gain the own leverage that is suitable for the business of each commercial bank

According to the previous research such as Octavia, M & Brown, R (2008), Gropp,

R & Heider, F (2009), Caglayan, E & Sak, N (2010) and Chau, N.H (2012), it is widely believed that the capital structure of commercial banks are affected by size, profitability, business risk, growth, collateral value and dividend policy Therefore, the study tests whether these factors as above are really impacting on capital structure in Vietnamese commercial bank system

1.2 Research Objectives

The objectives of this study are to:

i) Determine the elements which have impacted on capital structure of Vietnamese commercial bank system and test whether each element causes capital structure of Vietnamese commercial bank system

ii) Recommend some policy implications m order to not only develop the commercial bank system but also control the risk from their business; in addition, the study desires to make valuable instructions to improve the capital structure of each Vietnamese commercial bank by controlling the determinants

1.3 Research Questions

1.3.1 Main question

The main question of the study is "What are the determinants of capital structure in Vietnamese commercial bank system?"

- - - -

-1.3.2 Sub questions

i) Do factors like size, growth, collateral value, profitability, business risk and

dividend policy have statistically significant in explaining the model that related

with capital structure of commercial banks in Vietnam?

ii) Do these factors as above have the expected sign in accordance with capital

structure theories and previous studies?

1.4 Justifications of the study

There have many empirical studies that related to the capital structure problem;

however, there have few empirical studies which research the determinants of bank

capital structure Moreover, most of the studies focus on the situation of Europe and

USA Therefore, this paper contributes to the academic study as one of

comprehensive studies about the capital structure of commercial banks in Vietnam

In addition, this study aims to update the hottest issues of the capital structure of

Vietnamese commercial bank system The study analyzes capital structure problem

in the background of restructuring banking system process of Vietnamese

Government and State Bank of Vietnam According to the useful and informative

results, authorities have ability to propose the right policies to develop the

commercial bank system and control the risk from their business; besides, managers

of each Vietnamese commercial bank have ability to choose the own appropriate

capital structure

1.5 Scope of the study

The study determines the elements which have impacted on bank capital structure in

Vietnam from 2007 to 20 11

1.6 Organization of the study

The remaining study includes 4 chapters: Chapter 2 gives a review of theoretical and empirical studies which are related to capital structure topic Chapter 3 mentions the research methodology and data collection; thus, findings and discussion are presented in Chapter 4 Finally, Chapter 5 draws conclusions, recommends some policy implications and mentions limitation; thus, some directions for further studies are also presented in the chapter

•

- - - -

-CHAPTER 2 LITERATURE REVIEW

Chapter 2 gives a review of theoretical and empirical studies which are related to capital structure topic The chapter presents 4 main parts which contain theoretical literature, empirical literature, research hypothesis and conceptual framework First, theoretical literature part introduces 4 famous theories which include Modigliani and Miller (M&M theory, 1958), agency theory (Jensen and Meckling, 1976), pecking - order theory and static trade-off theory (Myers, 1984 ) Second, empirical literature part presents cases of developed and developing countries which are related to both non-bank and bank capital structure Thus, research hypothesis and conceptual framework will be mentioned in the study based on the results of previous empirical studies

2.1 Theoretical literature

When mentioning the capital structure, there are several theories that are related to this issue Among them, the first cornerstone of capital structure theories has been carried out by Modigliani and Miller (M&M theory, 1958) After the capital structure theory of Modigliani and Miller, many theories about capital structure are

in tum released The study expresses 3 famous other capital structure theories, which include agency theory was created by Jensen and Meckling (1976); pecking-order and static trade-off theory was developed by Myers (1984)

2.1.1 Modigliani and Miller (M&M theory, 1958)

According to the M&M theory, there are 2 basic propositions: proportion 1 examines the relationship between the firm value (V) and capital structure and proportion 2 investigates the relationship between cost of capital (r5) and capital structure In each proportion, authors considered 3 specific cases In my topic, the proportion 1 ofM&M theory is researched deeply

In case 1 of the proportion 1, all assumptions hold in this model: homogeneous expectations, homogeneous business risk classes, perpetual cash flows, and

especially, capital market is perfectly; it means that there exists perfect competition, firms and investors have ability to borrow or lend at the same interest rate and they equally access to all relevant information Moreover, there have no financial distress costs and taxes Besides, Modigliani and Miller also mentioned other cases, in particular, all assumptions as above hold except the corporate income tax assumption (case 2) and all assumptions as above hold except the corporate income tax and financial distress costs assumptions (case 3 )

Based on the M&M theory, the value of unlevered firm and levered firm are the same in case 1, it means that choosing any capital structure do not affect the value

of the firm However, in case 2 and 3, the value of the firm is affected by the capital structure, in particular, the functions as below show the relationship between the value of the firm and capital structure:

Case 1: VL = Vu

Case 2: V L = Vu + T c * B

Case 3 : V L = V u + T c * B - PV (costs of financial distress)

With V L: value of the levered firm; V u: value of the unlevered firm; B: debt value; Tc: rate of corporate tax, Tc * B: present value of tax shield and PV(costs of financial distress) is the present value of financial distress cost

Especially, in case 3, the value of the firm rises with leverage and falls with costs of financial distress Therefore, each firm should choose the appropriate capital structure in order to take advantages of tax shield benefits; however, if the leverage

of the firms is too high, then they will face up the high costs of financial distress such as bankruptcy cost The possibility of bankruptcy has a negative impact on the value of the firm

2.1.2 Agency theory

Jensen and Meckling (1976) proposed that the firm may choose capital structure according to the existence of agency cost There are 2 main conflicts in choosing the

capital structure First, the conflicts between managers and shareholders will happen since managers desire to receive more benefits but doing less effort in maximizing the value of the company Therefore, managers decide to increase the level of the debts so as to carry out the high risky projects The purpose of decisions is to complete the tasks and business plan from the requirements of stockholders; however, if these projects face up losses, then stockholders will bear the consequences of losses Second, rising level of debts may arise the conflicts between shareholders and debt-holders Debt-holders always receive the fixed return The conflicts arise when the shareholders decide to perform the high risky projects If the investments earn the return that higher than the value of debts, then shareholders will receive most of the gain; otherwise, the debt-holders have to incur the losses because the maximum losses of stockholders are only the amount of their initial stockholders' investment (limited liabilities)

2.1.3 The pecking-order theory

The pecking-order theory was developed by Myers (1984) He argued that the capital structure choice of the firm can be explained from the perspective of asymmetric information and the existence of transaction costs The more level of asymmetric information, the more risk that outsider investors facing; thus, the outsider investors require more discount rate from issuing stocks Therefore, the preference of the firm is that they would like finance by internal funds first such as retained earnings, instead of with external funds If the firm must to attract the external funds, then debt is preferred to equity

The theory has explained that companies usually use less debt when they have ability to earn the high return because they prioritized to finance their projects by internal funds Besides, the theory has also appreciated the importance of financial slack When enterprises have the financial slack, they have spare cash or transfer the asset such as government bond, stock into cash easily From that point, they have ability to take advantage of investment opportunities as they appear On the

contrary, if enterprises have no financial slack, then they have to choose the external funds by borrowing debts with high interest rate or issuing the common stock that price of each stock is under the face value

2.1.4 The static trade-off theory

The static trade-off theory was also developed by Myers (1984 ) This theory suggests that each firm need to choose the appropriate capital structure by considering the trade-off between the benefit of debts and cost of debts carefully The benefit of debts is the tax deductibility of interest payments and the cost of debts is the cost of financial distress such as bankruptcy cost and the losses that the firm has to incur from the bankruptcy events However, in reality, it is difficult to choose the optimal capital structure because the firms almost have not ability to determine the cost of financial distress

In general, commercial bank is the special kind of enterprises so M&M, agency, pecking-order and static trade-off theories are absolutely applied for choosing the appropriate bank capital structure

2.2 Empirical literature

2.2.1 Foreign empirical literature

There have many empirical studies which research about the capital structure However, almost empirical studies specialize in researching non-banks capital structure Booth et al (200 1 ), Frank and Goyal (2005) argue that several independent variables which may affect the leverage of each non- financial institution are size, profitability, business risk, growth, collateral value and dividend policy

First, size has the positive relationship with the leverage Large firms normally finance their projects by using more debts They are difficult to cope with the bankruptcy because of higher diversification The positive relationship between size

-

-and leverage has been in accordance with the argument of asymmetric information (Myers and Majluf, 1984) Small enterprises usually do not announce enough information about their financial situation or extraordinary events; accordingly, potential lenders are afraid of providing credit fund to small-size enterprises On the contrary, large firms, especially listed firms, are easy to access loans from financial institutions because there have a little asymmetric information between the owners and creditors The firms have ability to make the best of economies of scale in accessing long-term loans or issuing long-term debts such as corporate bonds; moreover, they may have bargaining power to lenders

Second, firms with higher profit tend to have lower debt-equity ratio It has been suitable for the viewpoint of the pecking-order theory (Myers, 1984) High-profit companies usually finance its project by retained earning first, instead of with external fund because of the perspective of asymmetric information and the existence of transaction costs

Third, there is negative relationship between the business risk and leverage Frank and Goyal (2005) argued that firms face up high business risk when they have more volatile cash flow; thus, they cope with the higher cost of financial distress They have considered the trade-off between tax-shield benefits and bankruptcy cost More financial distress cost decreases the probability that tax-shield benefits will be used up; as a result, they should use less debt in order to reduce the volatile cash flow

Fourth, firm's growth has the negative relationship with the leverage Booth et al (200 1 ), Frank and Goyal (2005) both use market-to-book ratio as a proxy for growth A high market-to-book ratio means that firms have ability to carry out the high growth opportunities; however, these firms also face up the higher cost of bankruptcy Therefore, companies which have high market-to-book ratio tend to use less leverage Besides, Frank and Goyal (2005) also have recommended that the growth of assets and ratio between capital expenditures and assets are the other

proxies to measure the firm's growth They have argued that these variables should

be positively with the book leverage according to the viewpoint of the pecking order theory When profits do not increase correlatively with the increase in investment activities of firms, then they should use more debt in order to keep the high growth

-Fifth, if companies have many tangible assets, then they will use more debts because these assets can be used as collateral The agency theory has explained the relationship clearly Jensen and Meckling (1976) have emphasized that there always exist agency cost of debts The conflicts arise when the shareholders decide to carry out the high risky projects If the investments earn the return that lower than the value of the debts, the debt-holders have to incur the losses Therefore, debt-holders are willing to lend more money and reduce the credit risk when firms have many tangible assets that can be used as collateral Furthermore, based on the static trade-off theory (Myers, 1984), tangible assets such as land, equipments, factories have lower expected financial distress cost; on the contrary, intangible assets will lose most their value if bankruptcy event occurs

Finally, when firms have ability to pay dividend, then they will use less debts Based on the pecking-order theory (Myers and Majluf, 1984 ), dividend is considered as good signal about future prospect of firms; accordingly, they will issue more equity and less leverage

According to Mishkin (2000), capital structure of commercial banks may be affected only by the regulations of capital requirement, especially each commercial bank must reserve amount of capital to reduce the probability of failure according to the standard of Basel Accords However, several empirical studies like Octavia, M and Brown, R (2008), Gropp, R and Heider, F (2009), Caglayan, E & Sak, N (2010) found that standard determinants of non-firm capital structure which include size, profitability, business risk, growth, collateral value and dividend policy are also significant in explaining the capital structure of banks The dependent variables

-of the 3 empirical studies, which was considered as the proxy -of commercial bank capital structure, was the book leverage or market leverage It is measured by debt-to-equity or debt-to-asset, with asset is equity plus debt In banking industry, commercial banks access primary debts by attracting the money from depositors These studies used fixed effect of panel data method to identify determinants of bank capital structure Octavia, M and Brown, R (2008), Gropp, R and Heider, F (2009) have shown that book leverage has been affected by 6 main determinants, which include size, profitability, growth, collateral value, dividend policy and asset risk Table 2.1 presents ways to measure each variable in Octavia, M and Brown,

R (2008), Gropp, R and Heider, F (2009) and Caglayan, E & Sak, N (20 1 0)

Table 2.1: Definition of Variables

A DEPENDENT VARIABLES

1 Book Leverage (BL) 1-(book value of equity I book value of assets)

2 Market Leverage (ML) 1-(market value of equity I market value of assets)

B INDEPENDENT VARIABLES

2 Profitability (Prof) Profit after tax I book value of assets

3 Market-to-book ratio (MTB) Market value of assets I Book value of assets

4 Collateral (Coli) Tangible assets I book value of assets

5 Dividend (Div) Equal1 when bank pays dividend in a given year

6 Asset risk (Risk) Yearly standard deviation of daily stock price

returns * (market value of equity I market value of bank)

The general function has been expressed as follow:

BLi,t =~ỡt Ln(Sizei,t-t)+~zProfi,t-1 +~3MTBi,t-I +~4Colli,t-I +~sDivi,t +

~6ln(Riski,t-I) +uit MLi,t =~ỡtLn(Sizei,t-I)+~zProfi,t-I+~3MTBi,t-I+~4Colli,t-I +~sDivi,t +

~6ln(Riski,t-I) +uit According to the results of Gropp, R & Heider, F (2009), all coefficients are statistically significant and have the expected sign in case of developed countries First, size has the positive relationship with the leveragẹ Large commercial banks are easier to attract deposits than small ones because there have a little asymmetric information between depositors and large ones Depositors normally have confidence in the brand name of large commercial banks Besides, the large commercial banks have many branches than small ones; thus, depositors are easy to access banking services from large commercial banks

Second, commercial banks with higher profit tend to have lower debt-equity ratiọ It

is suitable for the viewpoint of pecking-order theory (Myers, 1984) High-profit commercial banks usually give the loans and develop banking services by retained earning first, instead of depending too much on the money from depositors because

of the perspective of asymmetric information and the existence of transaction cost

On the contrary, low-profit commercial banks have to choose the external funds by borrowing the debts with high deposit rates The more level of asymmetric information, the more risk that depositors facing; thus, depositors require high deposit rate from the commercial banks

Third, market-to-book ratio has the negative relationship with the leveragẹ A high market-to-book ratio means that commercial banks have ability to carry out the high growth opportunities; however, these banks also face up the higher cost of bankruptcỵ Therefore, commercial banks which have high market-to-book ratio tend to use less leveragẹ

Fourth, if commercial banks have many tangible assets, then they will use more debts because these assets can be used as collateral Depositors normally have confidence in commercial banks that have many tangible assets because they certainly will receive more money when high tangibility commercial banks go bankruptcy Furthermore, based on the static trade-off theory (Myers, 1984), tangible assets such as money, government bond, land have lower expected financial distress cost; on the contrary, intangible assets will lose most their value if bankruptcy event occurs

Fifth, when commercial banks have ability to pay dividend, then they will use less debts Based on the pecking-order theory (Myers and Majluf, 1984 ), dividend is considered as good signal about future prospect of commercial banks; accordingly, they will issue more equity and less leverage

Finally, there is negative relationship between the asset risk and leverage Commercial banks have to face up high financial distress cost when they have high asset risk They have considered the trade-off between tax-shield benefits and bankruptcy cost according to static trade-off theory (Myers, 1984) More financial distress cost decreases the probability that tax-shield benefits will be used up; as a result, they should use less debt in order to decrease the probability of happening bankruptcy events

About the case of developing countries, Octavia, M and Brown, R (2008) realized that there have 4 independent variables which include size, profitability, asset risk and dividend policy have the expected sign; on the contrary, 2 other variables have not the expected sign They have suggested that there exists overpricing situation in the capital market when commercial banks have high market-to-book ratio This leads the high level of asymmetric information between the bank owners and external investors; thus, commercial banks will issue more debts In addition, collateral value variable has not the expected sign according to the reality situation Based on Octavia, M and Brown, R (2008), there have average 43.5% assets of

developing countries banks are used as collateral, that are much higher than 26.6%

of developed countries banks Accordingly, one extra dollar of tangible asset may have lower marginal benefit in ensuring debts than the case of developed countries Moreover, Caglayan, E & Sak, N (2010) has researched the bank capital structure

in the specific case of Turkey Dividend and asset risk variables were not applied in their model because of unmeaning results They only used 4 independent variables, which include size, profitability, growth and tangibility in the model These variables are statistically significant and have the expected sign

Besides, some empirical studies such as Asarkaya, Y.& Ozcan, S.(2008) and Romdhane, M (2010) have used Capital Adequacy Ratio (CAR) as the proxy of bank capital structure According to Asarkaya, Y & Ozcan, S (2008), 7 independent variables which are risk level (risk), CAR of previous period (CAR0 ),

asset size (asset), return on equity (ROE), share of deposit in non-equity liabilities (deposit), average CAR of the sector (AVCAP) and economic growth (GDP) have affected the CAR of Turkish banking sector They have used Generalized Method

of Moments (GMM) to run the regression The desired level of CAR can be written

as follows:

CARi,t = (1-a.) CARi,t-1 +a CAR\t + ei,t (a.: speed of adjustment)

The proxy variables which include risk level (risk), asset size (size), return on equity (ROE), share of deposit in non-equity liabilities (deposit), average CAR of the sector (A VCAP) and economic growth (GDP) have been used to estimate the unobserved CAR* because the target level of CAR has been not observed

~6GDPi,t + Ui,t·

The parameter estimate of GMM is expressed by the coefficient of lagged CAR (A.1) The target level of CAR can be rewritten as following:

Based on Asarkaya, Y & Ozcan, S (2008), it is widely accepted that the results are

in accordance with the theory CAR of previous period, portfolio risk, economic

growth, average CAR of the sector, return on equity have the positive relationship

with CAR; moreover, share of deposits and asset size have the negative relationship

with CAR

Besides, Romdhane, M (20 1 0) used fixed effect and random effect of panel data

method to identify the determinants of bank's capital ratio in Tunisia He has

proved that 8 independent variables which contain asset size, risk, net margin

interest, equity cost, ratio of deposits, deposit variability, intermediation rate, ratio

of the sector have affected the bank's capital ratio in Tunisia According to the

results of Romdhane, M (20 1 0), all coefficients are statistically significant and

have the expected sign In detail, the deposit ratio, the equity cost and asset size

have negative relationship with CAR All the other variables have positive

relationship with CAR

In general, foreign empirical studies in researching the capital structure of banks are

presented briefly in Table 2.2

Table 2.2: Foreign empirical studies in researching the bank capital structure

No Authors Methodology Data Results

1 Octavia, Fixed effect of panel data From the financial All coefficients are

Mexico, Pakistan,

have the expected Dependent variables: Thailand, Turkey

sign book leverage and and Zimbabwe)

market leverage from 1996-2005 Market-to-book

ratio and collateral variables have not the expected sign

2 Gropp, R Fixed effect of panel data 200 biggest listed All coefficients are

F (2009)

6 main determinants: of 16 countries significant and have size, profitability, (US and 15 EU the expected sign growth, collateral value, members) from

dividend policy and asset 1991 to 2004 risk

Dependent variable:

book leverage and market leverage

3 Caglayan, Fixed effect of panel data 25 deposit banks All coefficients are

N (2010)

4 independent variables: period 1992-2007 significant and have size, profitability, growth the expected sign and tangibility

Dependent variable:

book leverage

No Authors Methodology Data Results

4 Romdhane Fixed effect and random Half-yearly data All coefficients are , M.(2010) effect of panel data from 18 Tunisian statistically

method commercial banks significant and have

8 independent variables: in the period the expected sign risk, net margin interest, January 2002 - The deposit ratio, equity cost, ratio of December 2008 the equity cost and

intermediation rate, ratio relationship with ofthe sector, asset size capital adequacy

capital adequacy ratio All the other

variables have positive relationship with capital adequacy ratio

;;: 5 Asarkaya, Generalized Method of Monthly data from The results are in

Y.& Moments(GMM) 20 Turkish accordance with the Ozcan, S

7 determinants: risk commercial banks theory

and asset size have the negative relationship with CAR

2.2.2 Vietnamese empirical literature

In the specific case of Vietnam, there are several empirical studies which research Vietnamese non-financial capital structure such as Nguyen,T.D.K & Ramachandran, N.(2006), Biger, N & Nam, N.V & Quyen, H.X (2007) and Dzung, N & Rainey, I.D & Gregoriou, A (2012) Similar to the empirical studies about the capital structure of developed and other developing countries cases, they have also used the basic determinants such as size, growth, profitability, business risk and tangibility in their model Among them, size, growth and profitability have really affected the capital structure of Vietnamese companies In particular, profitability has the negative relationship with the book leverage, size and growth have the positive relationship with one The determinants have supported theories strongly Conversely, the impact oftangibility and business risk on leverage are not

in accordance with theories According to 3 empirical studies as above, the negative relationship between tangibility and leverage is in accordance with the context of Vietnam Vietnamese firms with few tangible assets are difficult to access the credit funds from commercial bank system; thus, they tend to depend on short-term liabilities like trade credit Moreover, based on the results of Nguyen,T.D.K &

Ramachandran, N.(2006), business risk has the positive relationship with leverage

in Vietnamese case because the State Bank of Vietnam supervised the banking system during 1998-2001 The interest rate was controlled within a band; as a result, high risk companies had ability to access loans with interest rates that were lower than the ones when credit market was set by the market forces

Besides, the 3 empirical studies have also used the ownership variable to measure the impact on the leverage It is widely accepted that state-owned firms are easier to access the bank loans than private ones When state-owned firms have new investment projects, they have ability to access the loans from state-owned commercial banks because the Government is the same owner of theirs

When researching the Vietnamese small-medium size enterprises, Nguyen, T.D.K

& Ramachandran, N (2006) suggested that banking relationship and networking variables have positive relationship with the leverage If firms focus on establishing and maintaining the good relationship with banks, then they will receive more trust from commercial banks It is widely believed that the level of asymmetric information between the firms and banks is decreased when the firms have good banking relationship Moreover, firms with wide networking always get reliability from commercial banks Therefore, they have ability to receive trade credit or loans from commercial banks

About the banking industry, Chau, N.H (2012) researched the capital structure of Vietnamese banking system He only used 4 determinants which include size, growth (Grow), profitability (Prot) and collateral value (Coli) to determine the impact on book leverage Pooled and fixed effect panel least square methods were used in his studies The regression function is expressed as follow:

BLi,t =Po+P 1 *Ln(Sizei,t-t)+P2 *Profi,t-t+P3 *Growi,t+P4 *Colli,t-t+uit·

Based on the results of his paper, coefficients of size and profitability are statistically significant and have the expected sign In detail, size has the positive relationship and profitability has the negative relationship with the book leverage Moreover, there is negative relationship between collateral value and book leverage

It means that if commercial banks have more assets as collateral, then they will use less debt It is not in accordance with the agency theory but suitable for the situation

of developing countries, similar to Octavia, M and Brown, R (2008) and Caglayan,

E & Sak, N (2010) There have average 40% assets of Vietnamese banking system are used as collateral, similar to average 43.5% assets in developing countries and much higher than average 26.5% assets in developed countries based on the results

of Octavia, M and Brown, R (2008) Consequently, one additional currency unit of tangible asset in Vietnamese banking system have lower marginal benefit in ensuring debts than the case of developed countries

In addition, commercial banks with higher growth tend to have higher debt-equity ratio in Vietnamese case It is not in accordance with static trade-off theory according to the viewpoint of the author However, he has argued that the variable should be positively with the book leverage according to the real situation of Vietnam When profits of commercial banks do not increase correlatively with the increase in their investment activities, then they should use more debt in order to keep the high growth

In sum, empirical studies m researching the Vietnamese capital structure are presented briefly in Table 2.3

Table 2.3: Empirical studies in researching the Vietnamese capital structure

No Authors Methodology Data Results

1 Nguyen, Multiple regressions Financial Size, ownership, banking T.D.K&

8 determinants: statements of relationship and networking Ramacha-

growth, tangibility, 176 state- variables have positive ndran, N

business risk, owned and relationship with dependent (2006)

profitability, size, firm 382 private variables They have ownership, banking SMEs in supported theories strongly relationship, Vietnam Growth and profitability networking from 1998 to variables are significant but

not in accordance with theories

2 Biger, N Ordinary Least Square 3,778 Profitability, size, growth

&Nam, (OLS) observations and industry classification N.V.& 8 determinants: from are statistically significant Quyen,

collateralized assets, Vietnamese and have the expected sign H.X

profitability, tax rate, enterprises Collateralized assets and (2007)

non-debt tax shield, census 2002 - effective tax rate are not in size, growth, industry 2003,which accordance with theories and firm ownership is carried out

Non-debt tax shied is not

by General Dependent variables:

Statistics statistically significant liabilities ratio and

Office (GSO) debt ratio

No Authors Methodology Data Results

3 Dzung,N Generalized Method of 116 non- Profitability, growth,

& Rainey, Moments (GMM) financial liquidity and state

leverage and positive relationship with the long- term leverage

4 Chau, N.H Pooled and fixed effect 29 Size and profitability (2012) panel least square Vietnamese variables are statistically

4 main determinants: commercial significant and have the size, growth, banks from expected sign

profitability and 2006 to 2010 Collateral variable is

Dependent variable: important factors to explain book leverage the leverage decisions

Growth variable is not in accordance with static trade-off theory

2.3 Research Hypothesis

According to previous research such as Octavia, M & Brown, R (2008), Gropp, R

& Heider, F (2009), Caglayan, E & Sak, N (2010) and Chau, N.H (2012), it is widely believed that the expected sign of each determinant is that size, growth are positively and collateral, profitability, business risk, dividend policy are negatively with leverage Therefore, the study has 6 null hypotheses as follows:

H 1: Size has positive relationship with book leverage

H 2 : Growth has positive relationship with book leverage

H 3 : Collateral has negative relationship with book leverage

H 4: Profitability has negative relationship with book leverage

H 5: Business risk has negative relationship with book leverage

H 6 : Dividend policy has negative relationship with book leverage