The effect of bank capital and ownership structure on bank performance: evidence in Vietnam

The findings show that long-term debt has a positive impact on ROA and ROE, while short-term debt and total debt have a negative impact on business performance measured by ROA and ROE.Re

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

THE EFFECT OF BANK CAPITAL AND

OWNERSHIP STRUCTURE ON BANK

PERFORMANCE: EVIDENCE IN VIETNAM

BY NGUYEN THI VIET ANH

HO CHI MINH CITY, DECEMBER 2015

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

THE EFFECT OF BANK CAPITAL AND OWNERSHIP STRUCTURE ON BANK PERFORMANCE: EVIDENCE IN VIETNAM

i

DECLARATION

I, Nguyen Thi Viet Anh hereby declare that this thesis is my own work under the guidance

of instructor, Dr Nguyen Thi Thuy Linh It has not yet been presented and will not be presented to any similar or other degrees

Date: 2nd December 2015

Signature

Nguyen Thi Viet Anh

ii

ACKNOWLEDGEMENTS

First and foremost, I would like to express my sincere thanks to my main supervisor, Dr Nguyen Thi Thuy Linh, for her candid comments, helpful hints and valuable scientific guidance during times of deployment, research and completion

of this thesis

I am greatly indebted to the Board of Directors, the teachers and all staff in Vietnam-Netherlands Programme for M.A in Development Economics– VNP, for their directly teaching, conveying knowledge and experiences as well as supporting

me and my class-mates necessary information in the last two years

I express my heartfelt thanks to General Director of Asia Commercial Bank - Mr

Do Minh Toan, the Board of Managers of ACB Saigon Branch – Mr Ngo Tan Long and Ms Bui Thi Anh Hoa for advice and creation of all favorable conditions for my work to attending VNP

I am grateful to Dr Pham Phu Quoc and Dr Truong Dang Thuy, who suggested research topic and gave me enthusiastic guidance in the early stages of writing the concept note and thesis research design

I greatly express my special thanks to my colleagues at The Center for Corporate Credit and Financial Institutions – Asia Commercial Bank, ACB Securities Company – ACBS, for supporting me in collecting data; colleagues at ACB - Saigon Branch, especially Corporate Department, for being always cooperative, enthusiastically supporting and encouraging me

I would like to acknowledge the enthusiastic help of fellow practitioners VNP - Class 20, especially these members, Nguyen Son Kien, Vo Tan Thanh Diep,

iii

Nguyen Le Phuong Linh, Nguyen Phuong Tram and Vu Thi Thuong, for being my close companions during the last two school years

Finally, I wish to dedicate this thesis for my family, my parents, siblings, especially

my dear husband, Nguyen Hoang Than for trust, love, sharing, help, following, encouragement and being always beside me I would also like to thank my lovely children, little daughter, Nguyen Hoang Viet Ha and little son, Nguyen Hoang Quan You are the endless motivation that helps me overcome all difficulties and move forward

iv

ABSTRACT

This study aimed at assessing the impact of capital structure and ownership structure on Vietnamese bank performance The study used secondary data of forty-nine Vietnamese banks with 387 observations in the period 2005-2014, employing Feasible General Least Square as well as Discoll-Kraay Robust for cross-sectional dependenceestimation Empirical results show that capital structure is significantly and positively related to Vietnamese bank performance Meanwhile, ownership structure impact negatively on profitability of Vietnamese banks (measuring by returns on total assets) The findings indicate that the profitability of the Private Banks is higher than the State-owned ones

Keys words:Bank performance, Capital structure, Ownership structure, Net Interest Margin, Returns on Assets, Returns on Equity

v

TABLE OF CONTENTS

DECLARATION I ACKNOWLEDGEMENTS II ABSTRACT IV LIST OF TABLES VIII LIST OF FIGURES IX LIST OF ABBREVIATIONS X

CHAPTER 1:INTRODUCTION 1

1.1 Problem Statement 1

1.2 Research Objectives 4

1.2.1 General research objective 4

1.2.2 Specific research objectives 5

1.3 Research questions 5

1.4 Significances of the study 5

1.5 Scope of the Study 6

1.6 Thesis structure 6

CHAPTER 2:LITERATURE REVIEW 7

2.1 Introduction 7

2.2 Conceptual Definitions 7

2.2.1 Capital Structure 7

2.2.2 Ownership structure 9

2.2.3 Bank performance 9

2.3 Theoretical review 11

2.3.1 Capital structure theory developed by Modigliani and Miller (MM Model) 11

2.3.2 The trade – off theory 12

2.3.3 Agency cost theory 13

2.4 Empirical review 14

2.4.1 Bank capital structure and Bank performance 14

vi

2.4.2 Ownership structure and bank performance 17

2.5 Conceptual Framework 19

CHAPTER 3:RESEARCH METHODOLOGY 20

3.1 Introduction 20

3.2 scope of Study 20

3.3 Epirical model 21

3.3.1 Bank performance measures: 21

3.3.2 The Models 23

3.4 Methodology 30

3.4.1 The Pooled OLS method 30

3.4.2 The Fixed Effects Model (FEM) 31

3.4.3 The Random Effects Model (REM) 32

3.4.4 Relevant tests to choose the most appropriate estimation method 33

3.4.5 The Feasible Generalized Least Square (FGLS) 34

3.4.6 Discoll-Kraay Robust for cross-sectional dependence – XTSCC 35

3.4.7 Problem of Endogeneity 36

CHAPTER 4:EMPIRICAL FINDINGS AND ANALYSIS 37

4.1 Introduction 37

4.2 The overview on Vietnamese banks 37

4.3 Descriptive statistics 44

4.4 Empirical results 47

4.4.1 Explanatory variables and bank performance 47

4.4.2 Control variables and bank performance 49

4.5 Robustness test - common panel data methods 50

CHAPTER 5:CONCLUSIONS AND POLICY IMPLICATIONS 56

5.1 Main findings 56

5.2 Policy implications 56

vii

5.3 Limitation of the study 58

5.4 Suggestions for further studies 58

REFERENCES 59

APPENDIX 69

viii

LIST OF TABLES

Table 2-1 – Indirect performance indicators for financial institutions 9

Table 3-1 Definitions of variables in equation 3.2 28

Table 4-1 Summary statistics for variables 45

Table 4-2 Correlation 46

Table 4-3 Vif index 46

Table 4-4 Results from feasible generalized least square (fgls) and discoll-kraay robust for cross-sectional dependence (xtscc) 47

Table 4-5 Regression result of nim model 51

Table 4-6 Regression result of roa model 52

Table 4-7 The results of f test and breusch – pagan test 53

Table 4-8 Hausman test 53

Table 4.9 Test of heteroskedasticity 54

Table 4.10 Test of autocorrelation 54

ix

LIST OF FIGURES

Figure 2-1 Components of Capital Structure 8

Figure 2-2 Static trade-off theory of capital structure 13

Figure 2-3 Conceptual Framework 19

Figure 4-1 Banks' Charter Capital updated to December 31th, 2014 38

Figure 4-2 Vietnamese banks’ total assets in the period 2005-2014 39

Figure 4-3 Vietnamese banks’ total equity capital in the period 2005-2014 40

Figure 4-4 The Return on Total Assets of the Vietnamese banks in the period 2005-2014 41 Figure 4-5 The Return on Total Equity of the Vietnamese banks in the period 2005-2014 42 Figure 4-6 The Net Interest Margins of the Vietnamese banks in the period 2005-2014 43

Figure 4-7 The average equity ratio and ROA, ROE, NIM of the Vietnamese banks in the period 2005-2014 44

x

LIST OF ABBREVIATIONS

DEA Data Envelopment Analysis technique

FEM Fixed Effects Model

Pooled OLS Pooled Ordinary Least Squares

REM Random Effects Model

ROA Returns on Total Assets

ROE Returns on Total Equity

FGLS Feasible Generalized Least Square

XTSCC Discoll-Kraay Robust for cross-sectional dependence

as the main objectives Focusing on raising equity capital standards, Basel III aims

at three major issues First, it requires an increase in the minimum common equity capital ratio from 2% to 4.5% Second, while the minimum total capital is maintained at 8%, the high-quality capital standards, i.e equity capital and Tier 1 capital standards have been increased Specifically, the minimum Tier 1 capital increases from 4% in Basel II to 6% in Basel III Besides, the assets with inherent quality issues will be gradually eliminated from Tier 1 capital and Tier 2 capital, as these investments exceed the limit of 15% of financial institutions In particular, the Basel III requirements apply additional minimum leverage ratio of 3% This is the ratio of Tier 1 capital to total assets plus off-balance sheet items Applying this ratio allows the Basel Committee on Banking Supervision to monitor changes in financial leverage ratio that the banks make in accordance with economic cycles and the relationship between capital requirements and leverage ratio

In Vietnam, after the global financial crisis, Vietnamese commercial banks were exposed to many problems of instability One of the most important factors affecting the competitiveness of Vietnamese commercial banks was the weak financial capability, especially the extremely low level of equity capital With the aim of enhancing the banking system performance and stability, on March, 1st,

2

2012, the Prime Minister issued the plan to restructure the system of banks and other credit institutions during the period 2011-2015 with Decision 254 / QĐ-TTg According to this process, the Prime Minister required the Ministry of Finance combine with the State Bank of Vietnam to approve plan to increase charter capital

of State-owned Commercial Banks From 2008 until now, Vietnamese commercial banks have actively increased their registered capital to meet the regulatory requirement of capital under the provisions of Decree No.141 / 2006 / ND-CP and Decree 10/2011 / ND-CP dated 26/01/2011 on amending and supplementing some articles of Decree 141 To achieve such requirement, all commercial banks implemented different approaches, such as selling shares to domestic or foreign shareholders, merging with other joint stock commercial bank(s)

Besides that, Decision 254/QĐ-TTg also required State-owned Corporations gradually exit their investment capital in the credit institutions The equitization of state-owned commercial banks and decreasing state-owned capital in these bank are also seen as a solution to reduce state ownership and increase the capacity of the banks based on various ownership structures (state, economic entities and individuals come from foreign countries) aiming at increasing competitiveness

There are many studies examining the impact of capital structure as well as ownership structure on bank performance

Based on agency problem (Jensen and Meckling, 1976), the agency view suggests that the banks with higher equity capital ratio have lower profitability (Berger and

Di Patti, 2006) Bank management suffers less pressure to maximize the value of the banks which maintain higher equity capital ratio This causes conflicts between management and shareholders so agency cost increases

On the other hand, many other researches prove that well-capitalized capital impacts positively on bank performance Using US banking system data in period 1984: Q.1 – 2010: Q.4, N.Berger and Bouwman (2013) suggested that equity capital enhances bank performance through three channels Higher equity capital banks

on Net Interest Margin However, their study uses data for thirty-three commercial banks in the period 2008-2011 with lower observations than my one Phuc (2014) uses the data of 217 companies listed on the Ho Chi Minh City and Ha Noi Stock Exchange in the period 2007 to 2012 to study the effects of capital structure on the performance of the company after equitisation The findings show that long-term debt has a positive impact on ROA and ROE, while short-term debt and total debt have a negative impact on business performance measured by ROA and ROE.

Research on the relationship between ownership structure and bank efficiency by using cost and profit frontier approach to examine efficiency of banks with different ownership types in German, Altunbas, Evans et al (2001) shows that public banks are less efficient than privately owned banks Study on ownership structure reform

in Nigeria in period 2004-2006, Ani, Odo et al (2012) points out that government ownership is negatively correlated with bank performance Uses data for commercial banks in 179 countries around the world in period 1995-2002 to examine the relationship between ownership and bank performance, Micco, Panizza

et al (2007) concludes that, in developing countries, the performance of owned banks is lower than their private counterparts because their profitability are lower while non- performing loans and overhead cost are higher However, Fuentes and Vergara (2007) points out the opposite conclusion that it still exist one state-

Although there have been many studies on the impact of capital structure and ownership structure on Vietnamese bank performance, they still at earlier stage and ues smaller dataset The mixed results and inconclusive literature motivates this study on the impact of capital structure and ownership structure on bank performance using Vietnamese data during the period from 2005 to 2014 Particularly, the bank performance is reviewed through two proxies: Net Interest Margin (NIM) and Returns on Assets (ROA)

1.2 RESEARCH OBJECTIVES

1.2.1 General research objective

Under the scheme to restructure the Vietnamese banking system in 2010-2015 periods, two out of four solutions given by Vietnamese Government to improve the operational efficiency of the banking system are increasing equity capital and equitization of state-owned commercial banks This plan is going into the final stage In order to investigate the relationship between capital structure, ownership structure and bank performance, I carry out this research

5

1.2.2 Specific research objectives

This thesis aims at the following two specific research objectives to achieve the above general research objective:

The first one is to evaluate the impact of capital structure on Vietnamese bank performance

The second one is to investigate the effect of ownership structure on Vietnamese bank performance

1.3 RESEARCH QUESTIONS

This study focuses on answering the following two research questions

Research question 1: Does bank capital structure impact on Vietnamese bank

performance?

Research question 2: Does bank ownership structureeffect onVietnamese bank

performance?

1.4 SIGNIFICANCES OF THE STUDY

This study aims at testing the impact of capital structure as well as ownership structure on Vietnamese bank performance Based on the results, the study offers some suggestions for the State Bank of Vietnam for its policies related to requirements in raising equity capital, as well as changing in ownership structure based on recommendations of Basel Committee on Banking Supervision and consistent with activities of commercial banks in Vietnam This study also supports Vietnam commercial banks to find an appropriate solution to build up a suitable capital structure, ownership structure to improve bank performance

6

1.5 SCOPE OF THE STUDY

The Vietnamese banking system has a total of fifty banks, including one Wholly State - Owned Commercial Bank, thirty-four Joint – Stock Commercial Banks in which there are three banks which the government owns more than 50% of the banks’ charter capital but is not the only shareholder, five Wholly Foreign Owned Banks, two Policies Bank, four Joint-Venture Banks, one Cooperative Bank and forty-nine Foreign Bank Branches updated to December 31th, 2014

This study focuses on forty-nine banks (whose names are listed in Appendix 1) for which during the period from 2005 to 2014 are available This limit is due to the fact that I cannot collect complete financial statements of Joint-Venture Banks, Policies Bank, Cooperative Bank, Wholly Foreign Owned Banks as well as the Foreign Bank Branches The selected banks also have limited data in some years during the sample period Thus, the final sample only includes 387 bank-year observations (unbalanced panel data)

1.6 THESIS STRUCTURE

This thesis consists of five chapters:

Chapter 1: Introduction

Chapter 2: Literature review

Chapter 3: Research Methodology

Chapter 4: Study findings and analysis

Chapter 5: Conclusion and policy implications

In which, Chapter 1 introduces the research objectives Chapter two presents the literature review In this chapter, some conceptual definations, previous findings of other authors related totopic are introduced Chapter three introduces the research methodology employed in this study Chapter four presents and analyzes the findings Chapter 5 concludes the thesis as well as provides some policy implications and suggests further research

2.2 CONCEPTUAL DEFINITIONS

2.2.1 Capital Structure

The capital structure is a term used to describe the origin and method of forming the capital to the business can be used for purchase of assets, tangible material means and support for production and business activities There are many different definitions of capital structure

For example, according to Chandra (2011, p.464): “Capital structure is how a firm decides to divide its cash flow into broad components, including a fixed component that is earmarked to meet the obligations toward debt capital and a residual component that the belongs to equity shareholders Alternatively, Abor and Biekpe (2005) defines that: “the capital structure of a firm is the specific mixture of debt and equity that it employs in financing the operations” Besides that, in the perspective words of Gerestenbeg “Capital Structure of a company refers to the composition or make up of its capitalization and it includes all long-term capital resources” (as cited in Paramasivan C & Subramanian T., chapter 5, p.47) Another definition of capital structure is: “The mix of a firm’s permanent long-term financing represented by debt, preferred stock, and common stock equity”,

in Paramasivan C & Subramanian T., chapter 5, p.47)

Additionally, in the research of R.H Wessel, capital structure is defined as: “The long term sources of fund employed in a business firm” (as cited in Paramasivan C

& Subramanian T., chapter 5, p.48)

Based on these definitions, the components of Capital Structure are demonstrated as following diagram:

Figure 2-1 Components of Capital Structure

Source: structure-with.html

http://articles-junction.blogspot.com/2013/10/components-of-capital-An optimal capital structure is the capital structure that maximizes the market value

of the firm, minimizing costs, maximizing earnings per share thus increased the dividend to shareholders, improving the ability for the firm to access new investment opportunities

Components of Capital Structure

2.2.2 Ownership structure

According to Gürsory & Aydogan (2002), ownership structure is understood in two ways, including ownership concentration and ownership mix In particular, ownership concentration represents the ownership rate of major shareholders who own at least 5% charter capital of the firm The shareholders holding dominant shares suffer the higher risks incurring during the operations of the firm To preserve their investment capital and ensure the expected profitability rate, they usually strongly monitor business operations of the firm as well as the decision of the executive committee Therefore, capital concentration may improve firm performance Meanwhile, ownership mix shows the type of shareholders in the firm, includes state ownership, private ownership or foreign ownership

2.2.3 Bank performance

Bikker (2010) points out some types of performance indicators mentioned in Table 2-1 for financial institution

Table 2-1 – Indirect performance indicators for financial institutions

Performance indicators Indicators represented as

1 Efficiency Cost X-efficiency

Profit X-efficiency Scale economies Scope economies

2 Costs Cost-to-income ratio

Cost margin Total costs/total income

3 Profit Return on capital

10

Return on assets Net interest margin

by Returns on Assets (ROA) (Goddard et al., 2008, as cite in Arias, Jara-Bertin, and Rodriguez (2013)) and Returns on Equity (ROE) (Al-Kayed, Zain, and Duasa (2014); Micco, Panizza, and Yanez (2007))

Net Interest Margin - NIM, calculated by the difference between interest income and interest expense divide to total profitable assets It is suggested that bank with higher NIM has better performance because higher NIM indicates higher profitability(Saunders and Schumacher (2000))

The first model concerning to NIM was mentioned in the study of Ho and Saunders (1981) At first, these authors established the formula for calculating net interest margin Accordingly, net interest margin function depends on the following variables: elasticity between demand and supply of capital in the market, the level

of risk aversion of the bank, the size of transactions and the variance of interest rates After that, Ho and Saunders built up a model of factors affecting the net interest margin in which, Net Interest Margin depend on the degree of risk aversion

of the banks, the market structure, the average transaction size and the variance of interest rates

ROA is a financial ratio used to demonstrate the profitability of a firm It shows profits generated from total assets or overall resources of a firm It is measured by the ratio of Net Income on Total Assets

11

2.3 THEORETICAL REVIEW

Some theories on capital structure

2.3.1 Capital structure theory developed by Modigliani and Miller (MM Model)

The theory of modern capital structure was given by (Modigliani & Miller, 1958) referred as MM model These authors made given some of assumption of perfect market: (1) Asymmetric information, (2) No transaction costs, (3) There is no firm income tax, (4) There is no personal income tax, (5) Lending interest rate and Borrowing interest rate are the same, (6) Opportunity accessing to funding is the same for individuals or firm, (7) No bankruptcy costs, no financial distress costs, (8) The entire profit is distributed to the owner: no reinvestment, no increase growth

Although the assumption of perfect capital markets is not true, however, there are two hypotheses need to be emphasized and they have a significant impact on the outcome of Modigliani and Miller study First, the assumption that there is no taxation: this is an important issue and one of the key advantages of the debt is tax relief for interest expenditure Second, the risk in the Modigliani and Miller theory

is calculated entirely by variability of cash flows They ignore the possibility of cash flow may stop because of insolvency This is a significant problem with this theory if other high debt

With the above assumed perfect capital market, postulate I of Modigliani and Miller said that the firm value is not affected by capital structure It was explained that the use of debt gives the owner a higher rate of return so higher income but this is offset

to the increased risks due to increased debt/equity ratio

Vg = Vu: The total value of firm using debt equal the total value of firm with no debt

In 1963, Modigliani and Miller launched a follow-up study with the removal of the assumption of firm income tax According to Modigliani and Miller, the firm

12

income tax, the use of debt will increase the value of the firm Because interest expense is reasonable expenses are deductible when calculating firm income tax, so that part of income of firm that use debt can be transferred to the investors by the equation:

Vg = Vu + T*D: value of the firm using debt equal value of firm does not use debt plus any benefits from the use of debt In particular, D is the market value of total debt, T is the firm income tax rate, T*D is benefits from the use of debt Thus the value of the firm using financial leverage will be higher than firm unlevered an amount equal the present value of the tax shield

2.3.2 The trade – off theory

The trade-off theory founded by Myers (1984)said that, besides the benefits from tax shield, the use of debt raises additional cost, typically bankruptcy cost including direct costs and indirect costs of bankruptcy caused by debt When the debt ratio rises to the point where present value of benefit from tax shield equal to present value of bankruptcy cost, the use of debt is no longer beneficial to the firm Because

of this, the firm always seeks solution to optimize its value basing on the principle

of balance to determine how much debt and equity in its capital structure The capital structure is optimal when the present value of the tax shield equals the present value of bankruptcy costs

According to the trade-off theory, the value of a firm is demonstrated as follows:

Vg = Vu + T * D - PV (bankruptcy costs)

In which:

Vg: total value of firm without debt

Vu: total value of firm using debt

T: firm income tax rate,

D: the market value of total debt

T * D is benefits from using debt

PV (bankruptcy cost): present value of bankruptcy cost

13

Optimal capital structure is the structure in which T * D = PV (bankruptcy cost)

Figure 2-2 Static trade-off theory of capital structure

Source: Myers (1984:577)

2.3.3 Agency cost theory

According to Jensen and Meckling (1976), agency cost arises because of the representatives of asymmetric information between agencies and shareholders The agencies did not make decisions aiming to maximize the profit for the shareholders; even their decision may hurt shareholder’s rights There are two major reasons leading to the asymmetric information between agencies and shareholders The first, there is difference in goals between agencies and shareholders Second, the level of risk aversion of the agencies and shareholders are different To monitor these problems, the firm should increase using of debt As the debt ratio increases, business managers will make decisions carefully not only in sponsor decisions but also in use of capital that help them manage business more effectively

14

2.4 EMPIRICAL REVIEW

2.4.1 Bank capital structure and Bank performance

Results on the association between bank capital structure and bank performance are mixed While some studies show a positive relationship between the level of bank capital and its performance (Hirschey (1999); Arias et al (2013); Valverde and Ferna´ndez (2007); J Maudos and Guevara (2004); Saunders and Schumacher (2000); J Maudos and Solís (2009); Claeys and Vennet (2008) and Ahokpossi (2013)), others provide contradictory evidence, suggesting a negative relationship(Berger and Patti (2006), Hamadi and Awdeh (2012) and Chortareasa et

al (2012)) The following section discusses both views and then develops the first hypothesis

2.4.1.1 Bank leverage is negatively related to bank performance

On the one hand, using the data of 695 commercial banks in The United States during the period from 1990 to 1995,Berger and Patti (2006) indicate that the ratio

of equity to total assets decrease 1% lead an increase of 16% in bank profit It means the banks with higher equity capital, have lower profitability Such result supports the agency cost theory, banks’ management suffers less pressure to maximize the value of the banks which maintain higher equity capital ratio This causes conflict between management and shareholders so agency cost increases In line with Berger & Patti (2006), the findings of Hamadi and Awdeh (2012) and Chortareasa et al (2012) suggest that the relationship between equity ratio and bank performance, measured as net interest margin (NIM), is negative Investigating the determinants of Lebanese banks’ NIM in period 1996-2009, Hamadi and Awdeh (2012) explains that bank with higher capital equity, i.e, higher financial autonomy ability, could be willing to pay higher deposit interest rate to mobilize more and more money so the loan volume of these banks increases as a result Besides that, well-capitalized banks are also willing to lend with lower interest rate aiming to

15

expand market share and benefit from economic scale, which then causing lower bank net interest margin

2.4.1.2 Bank leverage is positively related to bank performance

On the other hands, other recent researches indicate that the relationship between leverage and bank performance is strongly positive Some of those researches include Hirschey (1999); Ash Demirgüç-Kunt and Harry Huizinga (1999); Saunders and Schumacher (2000); Maudos and Guevara (2004); Valverde and Ferna´ndez (2007); Claeys and Vennet (2008); J Maudos and Solís (2009);Chortareasa et al (2012); N.Berger and Bouwman (2013); Ahokpossi (2013); Ameur and Mhiri (2013); Jara-Bertin, Moya, and Perales (2014) and Al-Kayed, Zain, and Duasa (2014)

Using ROA as an indicator measuring bank, Hirschey (1999) finds that debt to asset ratio is negative related to bank ROA This means equity to asset ratio is positively correlated with bank performance

The findings of Ash Demirgüç-Kunt and Harry Huizinga (1999) is consistent with Hirschey (1999) They use data of 80 countries around the world in period 1988-

1995 to investigate the determinants of commercial bank’s NIM ROA They conclude that equity to total assets lagged one period ratio is positive related to NIM and ROA According to Saunders and Schumacher (2000), because holding equity capital is relatively more costly than holding debts, banks with high capital ratio for regulatory or credit reasons are willing to cover some of the cost by generating an extra spread in the NIM over the pure spread for the interest rate risk Their findings remark that equity ratio positively impacts on NIM To offset the cost of holding high equity capital, banks often apply higher NIM Using FEM method to determine the factors affecting NIM of the banks, located in European countries including Germany, France, The United Kingdom, Italy and Spain in the period 1993-2000, Maudos and Guevara (2004) point out that the degree of risk aversion, proxied by equity to total assets ratio, is positively correlated with NIM J Maudos and Solís

16

(2009) use FEM to estimate the determinants of Mexican banks’ NIM in period 1993-2005 Their findings is consistent with Maudos and Guevara (2004) Valverde and Ferna´ndez (2007 )use GMM estimator to verify the factors affecting European banks' NIM The results of this study indicate that equity ratio is positivelyassociated with NIM The same result is found by Claeys and Vennet (2008) when they investigate the determinants of Central and Eastern European countries’ banks

It has been argued by Chortareasa et al (2012) that capitalization presented by equity ratio is related to higher NIM Using US banking system data in period 1984: Q.1 – 2010: Q.4, N.Berger and Bouwman (2013) suggested that equity capital enhances bank performance through three channels Higher equity capital banks require managers to monitor more and choose safer portfolios to invest Bank with higher equity capital can make customers, investors and other partner believe that it

is more reliable and safer so the banks can mobilize deposit money with lower interest rates as well as increase the volume of loans Therefore they can improve performance and increase market share as well As noted by Ameur and Mhiri (2013), well-capitalized banks are more profitable in comparison with others located in Tunisia Investigating factors affecting performance of Latin American banking system in the period of 1995-2010 by using Generalized Method of Moment (GMM), Arias et al (2013) use both NIM and ROA as proxies of banks performance Their conclusions indicate that capital level may improve banks performance including NIM and ROA The findings of Ahokpossi (2013) are the same when they show a positive relationship between LEV and NIM Study on determinants of Net Interest Margin in Vietnamese Commercial Banks, Thu and Huyen (2014) use data for thirty-three commercial banks in the period 2008-2011 Their findings indicate that equity to total assets ratio impacts positively on Net Interest Margin Phuc (2014) uses the data of 217 companies listed on the Ho Chi Minh City and Ha Noi Stock Exchange in the period 2007 to 2012 to study the effects of capital structure on the performance of the enterprise after equitisation His findings show that long-term debt has a positive impact on ROA and ROE,

Hypothesis 1: Capital structure impacts positively onVietnamese bank

performance

2.4.2 Ownership structure and bank performance

It is argued that, in poor countries, banks with dominant shareholders who are foreigners, perform better than others (Demirgüç-Kunt and Harry Huizinga, 1999) This result is explained as, in poor countries banks with foreign ownership type have some advantages including higher capital capacity, better banking technology

as well as better management skills in comparison with others However, in industrial countries their findings show negative relationship between foreign ownership and bank performance because of more competitive environment that make advantages of foreign-owned banks disappear Use DEA method to investigate the efficiency of banks located in Croatia in period 1995-2000, Jemric and Vujcic (2002) claims that the most efficient banks are banks owned by foreigners According to Altunbas et al (2001), among German banks in period 1989-1996, the efficiency of private banks is highest Mutual and public banks are less efficient in comparison with private owned but they do not point out any agency problem in these banks As noted by Micco, Panizza, and Yañez (2004), in developing countries, state owned banks perform lower than others However, in industrial countries, their findings indicate that ownership type does not impact on bank performance

18

Researching on Chinese banks performance in the period of 2003-2008, Wen (2010) uses both ROA and ROE to represent for bank performance His study focuses on the impacts of ownership structure including ownership concentration and ownership type on banks performance He finds out a positive relationship between state-owned type and ROE, in this case state-owned banks have better ROE His result does not give any conclusion about relationship between Ownership and ROA Otherwise, determining the effects of government ownership type on Nigerian banks performance, represented by ROA indicator, in the period of 1998-2008, the research results of Ani, Odo, and Okelue (2013) emphasizes that government ownership type is negatively and significantly correlated with ROA They suggest that the government should decrease the rate of government ownership in Nigerian banks to improve banks performance Similar to Ani et al (2013), the findings of Micco, Panizza, and Yañez (2004) points out the same result which indicates that dummy variable representing state-owned banks is negatively and strongly associated with banks performance represented by ROA, ROE and NIM indicators in developing countries It means that, in developing countries, private banks have better performance than their state-owned counterparts Research on the relationship between ownership structure and banks performance of The Middle East and North African banking system in the period of 2000-2007, Kobeissi et al (2010) highlights that private banks have better performance in comparison with other ones Son, Tu et al (2015) use data of thirty – four Vietnamese banks with 102 observations in the period 2010-2012 to investigate the impact of ownership structure on bank performance Their findings show that the percentage of private ownership impacts positively on bank profitability proxied by ROA

The equalization of State-owned Banks and requirement of the State-owned Enterprises divest capital gradually from commercial banks is being considered one

of solutions to improve the competitiveness of the banking system and the health of

Figure 2-3 Conceptual Framework

Source: Author’s own Design

CAPITAL STRCUTURE

OWNERSHIP

STRUCTURE

SIZE

CR CIO LIQ MPO

NIM/ ROA/ ROE

H 1

H 2

3.2 SCOPE OF STUDY

The Vietnamese banking system has a total of fifty banks, including one Wholly State - Owned Commercial Bank, thirty-four Joint – Stock Commercial Banks in which there are three banks which the government owns more than fifty percent of the banks’ charter capital but is not the only shareholder, five Wholly Foreign Owned Banks, two Policies Bank, four Joint-Venture Banks, two Policies Banks, one Cooperative Bank and forty-nine Foreign Bank Branches as of December 31th,

2014

Due to the limit of data on Joint-Venture Banks, Policy Banks, Wholly Owned Banks, this thesis could not include these banks in the sample Similarly, the financial statements of some banks are not fully collected during the period 10 years from 2005 to 2014 Therefore, this study uses the data of forty-nine banks (listed in Appendix 1) for which data are available during the period from 2005 to2014 These selected banks also have limited data in some years during the sample period Thus, the final sample only includes 387 bank-year observations (unbalanced panel data)

Foreign-Most of the bank-specific information is collected from financial statements and annual reports of Vietnamese commercial banks, which are publicly available on

21

their websites Further information is also collected from other websites including http://www.cafef.vn, vietstock.vn and http://www.cophieu68.vn/ Based on the banks’ financial statements, all indicators are calculated to run regression mentioned

in Chapter 4

3.3 EMPIRICAL MODEL

3.3.1 Bank performance measures:

Although NIM (Net Interest Margincalculated as the difference between interest income and interest expense divided by total average earning assets) is a common and important indicator used to represent bank performance, many empirical studies use other indicators such as return on assets (ROA) and return on equity (ROE) instead of NIM Following previous studies, for example, Chortareasa, Girardoneb, and Ventouri (2012), Claeys and Vennet (2008), J.Maudos and Guevara (2004), Hirschey (1999), Arias, Jara-Bertin, and Rodriguez (2013), Kobeissi, Nada, Sun, and Xian (2010), this study employs ROA and NIM indicators as proxies to measure bank performance Banks with higher NIM and ROA index are seen as having better performance

To assess European bank performance in the period of 2000-2008, Chortareasa, Girardoneb, and Ventouri (2012) uses NIM as one of two basic proxies Their study uses two banks specific characteristics which are determinants of NIM including equity to total assets ratio presenting capitalization and natural logarithm of banks’ total assets presenting bank size Their findings show the negative relationship between NIM and capitalization as well as bank size Using the data of 1,130 banks

in Western and Eastern European countries in the period of 1994-2001 to investigate the determinants of NIM, Claeys and Vennet (2008) inserts three bank specific factors affecting NIM including capitalization, market power as calculated

by total assets of banks divided by total assets of all banks at time t and credit risk measured by total loans divided by total assets in their estimated model To

22

investigate the determinants of NIM in the European banking system, besides capitalization, J.Maudos and Guevara (2004) adds some control variables including credit risk calculated by total loans divided by total assets and quality of management presented by cost-to-income ratio into NIM model Their results conclude a negative relationship between NIM and cost-to-income ratio and adverse relationship between NIM and credit risk Besides all bank-specific factors affecting NIM mentioned above, Hamadi and Awdeh (2012) introduces another control variable which is liquidity measured by liquid assets divided by total assets They suggest that the relationship between NIM and liquidity is negative

Hirschey (1999) uses ROA as an indicator measuring the profitability or performance of the banks in which ROA depends on bank size presented by total assets and leverage (a quotient of debt and total assets), an inverse measure of capitalization His findings show a negative relationship between ROA and leverage which means a positive relationship between ROA and capitalization Investigating factors affecting performance of Latin American banking system in the period of 1995-2010 by using Generalized Method of Moment (GMM), Arias, Jara-Bertin, and Rodriguez (2013) uses both NIM and ROA as proxies of bank performance Their conclusions indicate that capital level may improve bank performance including NIM and ROA

Research on relationship between ownership structure and performance of banks located in the Middle East and North African countries in the period of 2000-2007, Kobeissi, Nada, Sun, and Xian (2010) also uses both ROA and ROE indicators to present performance Focusing on ownership type, their findings indicate a better performance of private banks in comparison with state-owned banks

In summary, following previous studies, this thesis employs ROA and NIM as different measurements of bank performance

23

3.3.2 The Models

To examine factors affecting Vietnamese bank performance, based on previous theoretical and empirical researches mentioned above, the Vietnamese bank performance, in this study, is demonstrated through the following model:

PERF it = β 0 + β 1 LEV it + β 2 CR it + β 3 MS it + β 4 CIO it + β 5 LIQ it + β 6 SIZE +

β 7 OWNERSHIP + u it (3.1)

In which:

PERF indicates of bank performance, measures as NIM and ROA

NIM: Net Interest Margin, calculated by the difference between interest income and interest expense divide to total profitable assets

ROA: is a financial ratio used to demonstrate the profitability of a firm It shows profits generated from total assets or overall resources of a firm It is measured by the ratio of Net Income on Total Assets

Besides the two main explanatory variables including capital structure (proxied by leverage) and ownership structure, model 3.1 also takes into account other control variables such as credit risk (CR), market share (MS), cost to income ratio (CIO), liquidity risk (LIQ) and bank size (SIZE) to control for other bank characteristics

uit is the error term

LEV (LEVERAGE)

In this study, LEV is defined as the ratio between Equity and Total assets (in percentage) It represents capitalization or level of equity capital of the bank It is used to assess the financial autonomy of the banks

According to Hirschey (1999); Arias et al (2013); Valverde and Ferna´ndez (2007);

J Maudos and Guevara (2004); Saunders and Schumacher (2000); J Maudos and

24

Solís (2009); Claeys and Vennet (2008) and Ahokpossi (2013), well-capitalized banks presented by high level of equity to total assets ratio, are more profitable in comparison with others Based on these findings, the positive correlation between LEV and three bank performance indicators mentioned above, including ROAand NIM are expected

OWNERSHIP

Ownership structure mentioned in this thesis is a dummy variable representing ownership type It take value of 1 if the bank has shareholder who is government or agency of the government owns more than 50% charter capital of that bank, otherwise it equals 0

As mentioned in chapter two, Micco, Panizza, and Yañez (2004) indicate that dummy variable representing state-owned banks is negatively and strongly associated with banks performance represented by ROA, ROE and NIM indicators

in developing countries It means that, in developing countries, private banks have better performance than their state-owned counterparts The findings of Ani, Odo, and Okelue (2013); Ani et al (2013) and Kobeissi et al (2010) give the same results Son, Tu et al (2015) show that the percentage of private ownership impacts positively on bank profitability proxied by ROA.Vietnam is also a developing country so, in this study, the expected sign of impact of state-owned ownership type

on bank performance is negative

CONTROL VARIABLES

CR (CREDIT RISK)

There are many definitions of CR in different researches According to Ahokpossi (2013), CR is measured by the ratio between Loans to Deposits and Short-term funding CR is defined as the quotient of Loan Loss Provisions over Loans in study

of J.Maudos and Solís (2009) Valverde and Ferna´ndez (2007) defines CR is a one

25

lagged-value of the ratio loan default/ total loan Because of lack of information, inthis study, CR is measured by the ratio of Total loans over Total assets (in percentage) This definition is consistent with Maudos and Guevara (2004)

Valverde and Ferna´ndez (2007) investigate a negative relationship between CR and NIM They say that banks specialized in granting loans tend to charge lower NIM The study of J Maudos and Solís (2009) gives the same result Explaining the lower NIM in Mexican banking system, they says that banks specialized in granting loans may benefit from economies of scale because of low intermediate costs so they may apply lower NIM However, to find out the determinants of NIM in European Union banking industry, J Maudos and Guevara (2004) shows totally opposite result Their result demonstrates a positive relationship between CR and NIM The banks with higher loan volume are often faced with higher credit risk To offset credit risk, these banks tend to apply higher NIM Supporting this point of view are studies of Claeys and Vennet (2008); Ahokpossi (2013) and Hamadi and Awdeh (2012) We also expect CR has positive effects on NIM and ROA

MS (MARKET SHARE)

MS is calculated as Total assets of bank divide by sum of Total assets of all banks (in percentage) The findings of Claeys and Vennet (2008), which based on a sample of 1,130 banks in Western and Eastern Europe, highlight that MS may improve NIM This is interpreted that banks with higher market share, have more advantages as well as power to set prices autonomously aiming to achieve the expected profit margin That might lead them to apply higher NIM Hence, this study expects a positive impact of MPO on NIM, ROA in Vietnamese bank industry

26

CIO (COST TO INCOME RATIO)

CIO is measured by the ratio between operating cost and income is an indicator used to evaluate the quality of executive management and bank efficiency Higher CIO indicates lower management quality as well as less efficiency With the research on the Mexican banking system, J Maudos and Solís (2009) finds out a negative correlation between the CIO and NIM The same results are given by Hamadi and Awdeh (2012) after they do research on the Lebanese banking sector Based on these researches, negative relationship between CIO and NIM, ROA, is expected

LIQ (LIQUIDITY RISK)

LIQ is defined as the ratio of Liquid assets and Total assets (in percentage) The bank with higher LIQ has less liquidity risk Hamadi and Awdeh (2012) and Ahokpossi (2013)detect a negative relationship between LIQ and NIM They explain that banks with low liquid usually offer high deposit interest rates to increase the volume of deposits or have to borrow capital at high cost to improve liquidity That makes their NIM become lower Our expectation is consistent with these studies

SIZE (BANK SIZE)

This indicator is calculated by the natural logarithm of average total assets Using data of European countries in the period from 2000 to 2008 to examine the factors affecting banks performance, the findings of Chortareasa et al (2012) highlights that SIZE is negatively associated with NIM The research of Hamadi and Awdeh (2012) gives the same result Explaining this result, they say that the incomes of larger banks are mostly based on service fees compared to smaller banks Moreover, larger banks often have more advantages, and thus they are willing to pay higher interest rate for depositors with the aim to strengthen cross-selling products

Dependent variables – Bank performance

Net Interest Margin NIM Calculated by the difference between

interest income and interest expense divide to total profitable assets

Author’s calculation using banks’

financial statements

Return on Assets ROA ROA is the return on assets which is

the ratio between Net income and Average Total Assets (in percentage)

Author’s calculation using banks’

financial statements

Main explanatory variables

Leverage LEV by the ratio between Equity and Total

assets (in percentage)

Author’s calculation using banks’

Ahokpossi (2013); Ameur

& Mhiri (2013); Bertin, Moya, and Perales (2014) and Al-Kayed, Zain

Jara-& Duasa (2014)

Ownership type OWNERSHIP Dummy variable representing

ownership type It takes value 1 if the government or agencies of the government own more than 50% of equity capital of the bank, otherwise

Author’s calculation using data collected from website of The State Bank of Vietnam.

- Jemric and Vujcic (2002); Altunbas et al (2001); Micco, Panizza, and Yañez (2004)