Determinants of bank profitability in vietnam does foreign presence matter

There have been some researches regarding the impact of the internal factors such as bank specific characteristics size, risk level, liquidity status and so on and external factors such

VIETNAM – THE NETHERLANDS PROGRAMME FOR M.A IN

Academic Supervisor

Dr DUONG NHU HUNG

TRUONG UY PHAP

Date: … December 2014

ACKNOWLEDGEMENT

I would not be able to write and finish my thesis without the help and support

of people surrounding me

Above all, I would like to express my greatest appreciation to my supervisor,

Dr Duong Nhu Hung, for his precious comments, advices and guidance, encouragement during the writing period Without his guidance, my thesis would not be completed

I would also like to offer my special thanks to Dr Truong Dang Thuy for the econometric guidance and valuable suggestions that help to develop this thesis

Besides my mentors, special thanks also to all the lecturers at the Vietnam – Netherlands Program for their expert knowledge and sharing during the coursework phrase

In addition, I really appreciate my friends and people who are always beside

me and provide supports for my thesis

Last, but not least, I am very deeply grateful to my family Without their warm encouragement and patience, I would not be possible to complete well this thesis

TRUONG UY PHAP

ABSTRACT

The present paper seeks to examine the determinants of bank profitability in Vietnam during the period 2004–2013 The empirical findings suggest that credit risk, non-interest income, operating expense and liquidity variables have a statistically significantly impact on bank profitability The empirical findings suggest that credit risk and expense preference behaviour are negatively related to banks' profitability, while non-interest income and liquidity have a positive impact During the period under study, the results suggest that inflation has a negative impact on bank profitability, while the impacts of economic growth and money supply have not significantly explained the variations in the profitability of the Vietnamese banks Besides, foreign presence from the macroeconomic view and enterprise view do not show the significant relationship with bank profitability In other words, the opening up policy to allow more foreign entry entering into domestic banking market and relaxing of foreign ownership restriction in local banks are considered unclear for policy implications

Key words: bank profitability, foreign presence, ownership, Vietnam, determinants

ABBREVIATION

TABLE OF CONTENTS

CHAPTER 1: INTRODUCTION 1

1.1 Problem statement 1

1.2 Research objective 3

1.3 Research questions 3

1.4 The scope of the study 3

1.5 Structure of the study 4

CHAPTER 2: LITERATURE REVIEWS 5

2.1 Theory of Bank profitability’s determinant 5

2.2 Foreign presence/ ownership and its impact on bank profitability 10

CHAPTER 3: RESEARCH METHODOLOGY AND DATA 16

3.1 Research methodology 16

3.1.1 Estimation panel regression model: 16

3.1.2 Test for an appropriate panel data regression choice 18

3.2 Model specification 20

3.3 Variable measurement and test hypothesis 25

3.3.1 Bank profitability definition 25

3.3.2 Internal determinants 26

3.3.3 External determinants 29

3.3.4 Foreign presence and foreign shareholding 30

3.4 Estimation strategy & Data collection 32

CHAPTER 4: RESEARCH RESULTS & DISCUSSION 35

4.1 Vietnam banking background 35

4.2 Descriptive statistic of the sample 39

4.3 Empirical results & findings discussion 45

4.3.1 Result of Test for panel regression model 45

4.3.2 Empirical results 46

4.4 Discussion of findings 48

4.4.1 Overall model 48

4.4.2 Internal determinants 49

4.4.3 External determinants 52

4.4.4 Foreign presence 53

CHAPTER 5: CONCLUSION 55

5.1 Conclusion 55

5.2 Recommendations 57

5.3 Limitations & suggestion for future studies 58

APPENDIX A: INFORMATION OF BANK OBSERVATION SAMPLE 62

APPENDIX B: SUMMARY OF SOME MAIN EMPIRICAL LITERATURE REVIEWS 71

APPENDIX C: DESCRIPTIVE SUMMARY ANALYSIS 77

APPENDIX D: PANEL REGRESSION MODEL 79

APPENDIX E: RESULTS OF BREUSCH – PAGAN LM TEST 87

APPENDIX F: RESULTS OF HAUSMAN TEST 89

APPENDIX G: RESULT OF ROBUST STANDARD ERROR (FIX HETEROSKEDASTICITY) 91

LIST OF FIGURES

Figure 2.1: Analytical framework of determinants of bank profitability 15

Figure 4.1: Trend of foreign presence in Vietnam 38

Figure 4.2: Average bank’ equity by different levels of foreign presence 42

Figure 4.3: Average bank’ total assets by different levels of foreign presence 42

Figure 4.4: Average bank’ return on asset by different levels of foreign presence 43

Figure 4.5: Average bank’ net interest margin by different levels of foreign presence 43

LIST OF TABLES

Table 3.1: Tests for choosing the appropriate panel regression model 20

Table 3.2: Definition, notation and expected signs of relevant variables 22

Table 3.3: Data collection sources 34

Table 4.1: Numbers of banking institutions in Vietnam (2004-2013) 37

Table 4.2: Performance of banking industry and its contribution to Vietnam economy 39

Table 4.3: Summary statistics for all variables in this research 40

Table 4.4: Correlation analysis among independent variables 44

Table 4.5: Results of F test and Breusch – Pagan Test 45

Table 4.6: Results of Hausman Test 45

Table 4.7: Result of Test for heteroskedasticity 46

Table 4.8: Results of the final models 47

CHAPTER 1: INTRODUCTION 1.1 Problem statement

In the world of financial liberalization during the last few decades, in terms of facilitating the critical roles of international trade, the rise of international banks has become important significantly Banks have expanded multinationally by not only establishing foreign susidiaries and branches as local incoporation formation but also by taking over estabished foreign banks As other corporations, banks have always operated with the aims of profit maximization The foreign banks operates in developing countries will be considered as the same strategy afterward However, to what extent will the foreign bank operating in particular country yield a better return than those domestic competitors, or vise-versa? This is the question leading to many reseaches conducted recently

There have been some researches regarding the impact of the internal factors such

as bank specific characteristics (size, risk level, liquidity status and so on) and external factors such as inflation, GDP from the country-specific factors on banking profitability (Pasiouras and Kosmidou, 2007) These determinants have been anlayzed not only in a series of pooled countries but also in a single country However, to integrate the effect of foreign ownership structure into this link is critically necessary especially for developing countries where they are concentrating on the progress of financial liberalization and seeking for foreign entry to enhance the domestic market As part of the privatization and financial liberalization strategy, more and more developing countries allow foreign-owed banks or greenfield bank (Bonin et al, 2005) to operate in their domestic market

As stated in their study, privatization of banking sector, particularly in developing countries is usally accompany with complex procedures and phrases These may dramatically change the whole banking sector operation in which they would affect to the value of banking, growth rate, risk exposures and defintely performance of banks Some markets such as China followed a more prudent approach where they open a certain limit of foreign investor in their domestic commercial bank so-called strategic partner program (Shen et al., 2009) Since foreign banks represent one of bank ownership type besides domestic state-owned or private-owned banks, will foreign ownership, either at greenfield bank or at domestic bank, with the role of strategic

foreign investor or at board management, affect the banking profitability besides bank specific charateristics and macroeconomies factors in a typical developing market?

The answer of Bonin et al (2005) and Claessens et.al (2001) was considered as

“YES” Generally, they found that foreign-owned banks are more cost-efficient than other banks and deem better performance in terms of profitability, in particular if they have a strategic foreign owner Indeed, if foreign-owned banks use modern technology, better expense management, and rely on the human capital of their parent banks, they should perform better than government-owned or domestic private banks in transition countries In the scope of a developing country like China, the findings of Shen et al (2009) have concluded the insignificant impact of the foreign presence at individual bank’s stake on banking profitability The reason for this could derived from the impact

on both revenue and expense that somehow mututally reduced the final influence on profitability That led to the conclusion of further research to investigate more on the actual influence on each component or activity rather than release more stake to foreign investors Neverthelese, they have observed a increasing trend in attracting more foreign penetration in Chinese banking industry was significant, especially when Chinese government considered a opening-up policy to enhance its own financial market That is happened in China banking industry, Should foreign penetration or foreign ownership at individual bank have the same impact and the same findings on an emerging market where it witness the fast changing environment recently ?

To the limited knowledge of author, there have been many reseaches examining how different bank profitability could be affected critically by the bank internal factors and external factors, especially in a particular country In Vietnam, a typical less developed country, the similar study unlikely yield the similar answers, or maybe

“YES” Moreover, a more interesting topic flagged up besides the traditional determinants mentioned above is the relationship between foreign presence and bank profitability As noticed Vietnam recent banking industry is applying the strategy from some countries such as China, Hungary when introducing foreign strategic partners and 100% foreign-owned bank type in our commercial banking market These could be somehow have impacts on domestic bank profitability, as our opinion That is the reason why these trends could be brought up for our study interests In this study, these impacts

would be identified to provide a clear understanding of this relationship of both the internal and external determinants of banking profitability in Vietnam banking industry during period of 2004 to 2011 This period invloves a significant improvement of economic conditions with really high boom after joining WTO in 2006-2007 In addition

to this research aim, the question of how to improve the financial market or banking market in developing countries like Vietnam could be deemed as crucial to investigate Should Vietnam allow more and more foreign penetration entering into banking sector

to improve the profitability and its impact, if any, and the policy implication that Vietnamese government should be considered when follow the similar opening-up policy to liberalize our own domestic banking market? Or any other factors, as a hint, for policy maker to understand and react could be considered significant to enhance the banking industry?

1.2 Research objective

The objectives of the thesis are to:

a To examine the impact of internal factors (i.e bank sepcific characteristics) and external factors (i.e macroeconomies & industry variables) on Vietnamese bank’s profitability in Vietnam

b To investigate the impact of foreign presence on bank profitability controlling others factors in different bank ownership types in Vietnam;

1.4 The scope of the study

This thesis used yearly data of most of commercial banks in Vietnam during the 10-year period from 2004–2013 It focuses on two aims to answer the research questions above First, this paper will examine the determinants for bank profitability, both internal and external factors Second, it will try to apply a variable to test the impact of foreign ownership or foreign presence on individual bank profitability The dataset is

collected from Bankscope, which is one of the most reliable financial resources for banking industry meanwhile some external macroeconomics variable may be found on Worldbank, ADB and other sources The data sources for foreign shareholding will be collected through banks’ annual reports and be verified against information from Bankscope, other reliable media

1.5 Structure of the study

The structure of this study is displayed as a five-chapter format template Beyond the first chapter - Introduction chapter - flagged up to steer the attentions, the second chapter introduced the concepts and theoretical study of banking profitability and its relevant determinants with the coverage of foreign ownership discussion Besides, some empirical researches has also been employed as the foundation of our study followed by the next couple chapters Next chapter - chapter three - the research methodology and data collection, variables explanation and model specification would be introduced gradually Chapter four will then drive us through the snapshot of Vietnamese banking industry background meanwhile the important part of this research – empirical findings

& discussion will be brought up as the main section among others Final chapter, the study will summarize all of the policy implication, limitations, and further approach for research

CHAPTER 2: LITERATURE REVIEWS

This chapter generally covers two main sections and the derived framework The first section provides definitions and theory of bank profitability and reviews theoretical backgrounds about the critical factors that could impacts on banking profitability The second section summarizes previous empirical researches that studied on foreign presence or foreign ownership and its impact on profitability Based on relevant contents, the final drawn part will recommend in this chapter the construction of conceptual framework for this study

2.1 Theory of Bank profitability’s determinant

Knowledge of factors influencing the bank profitability will be challenged for all the economists In Vietnam, this is vital for all banks management level to have a better view of the effect of those determinants to their bank profit A lot of researches related

in this study area examine the determinants of bank profitability within the scope of single country (Ábel & Siklos, 2004; Sufian,2009) or the pooled countries (Athanasoglou et al, 2008; Molyneux & Thornton, 1992; Kosmidou & Pasiouras, 2007; Demirgüç-Kunt & Huizinga, 1999) However, to what extent of the knowledge of the research author, there was only a limited number of research focusing on this field for Vietnam market rather than a variety of countries or individual developing country such

as Phillipine or Malaysia

The theory of bank profitability’s determinants came from the study of a more comprehensive set of variables that would be discussed in this study conducted by Demirgüç-Kunt & Huizinga in 1999 In this research, they have employed most of the important factors to examine their impact on bank profitability They had used bank-level data for 80 countries since 1988 and found that differences in interest margins and bank profitability derived from different types of factors such as bank characteristics, macroeconomic conditions, explicit and implicit bank taxation, deposit insurance regulation, overall financial structure, and underlying legal and institutional indicators Using the linear regression and controlling for fixed effect, in which Net interest margin (NIM) and return on asset (ROA) were the key dependent variables representing the interest margin and bank profitability, some conclusions results in the lower margin and profit derived from maintaining a larger assets to GDP and a lower market concentration

ratio Besides, Demirgüç-Kunt & Huizinga in 1999 also found that difference in bank activities, leverage, and the macroeconomic environment could account for the change

of margin and profit The method they used was fixed effect regressionl with weighted least squares, with the weight being the inverse of the number of banks for the country in

a given year to control the different number of banks in each country

The model was based on the accounting identity representing that the bank variables themselves had impacts on profitability This is the original theory that we based on to conduct a clear hyportheisis of internal factors on bank profitability Income statement in accounting principle relect how well the bank operating and yield profit or loss Most of the determined factors accounted for bank profitability are shown in this statement Basically, profit comes from the result of main revenue deducted by other sources of expense In this scope of study, accounting identity of profit before (PBT) over total asset (TA) named ‘return on assets’, and the definite term or value to measure the bank profitability, was shown as the equation below:

PBT is profit before tax and it has the detailed expanded accounting identity as the main resources for most of the determinants retrieved:

Where NIM is net interest margin – a fraction of , NII is non-interest income, OV

is overhead, and LLP is loan loss provisioning NII/TA represents the how diversifying the bank have exposure on other nonlending activities, such as investment banking and and exchange rate arbitrary or fee-based services, OV/TA illustrates how the bank manages its own non operating expense to serve its assets and operation, and LLP/TA measures the proportion of the bad debt in total asset In this circumstance, we are ignoring the effect of tax due to the biased effect from different taxation rate or any advatage from tax shield This accounting was considered such a generic idea for most

of the determinants of profitability had been taken into account in most recent studies, for example, Claessens et.al (2001) The researchers have found that most of the key bank-specific or internal determinants are significantly with bank profitability These variables can be grouped by profit and loss statement’s factor and balance sheet factor Firstly, Profit and loss statement or so-called income statement illustrate how well the

bank operates and earn profit or suffer loss That is obviously tight with profitability where the accounting equation (2.1) and (2.2) have shown above These variables that they have employed here net interest income, overhead, non interest income, loan loss provision They do not take into account the effect of NIM on profitability since the bank’s operation traditionally deals with the activities of earning the spread of lending rate and deposit rate That could result from the logic that bank is the financial intermediaries who circulate the money across the country’s business performance from macroeconomic perspective The second bank-specific determinants here are balance sheet related items where capital, liquidity, and bank size play significant role towards bank profitability These variables were usually used by many authors to test the relevant impact on bank profitability from Demirgüç-Kunt & Huizinga, 1999; Kosmidou

& Pasiouras, 2007; Shen et al 2009 etc Normally, these factors are indirectly affect the bank profitability For example, the larger bank seems to attract better customer thanks

to it reputation and branch network and may not compete with other smaller banks especially in price Hence it would yield a better performance through lower cost of fund and higher profitability These have been proved by Bonin et al in 2005 while testing the impact of size on the Return on asset and other authors

The regression analysis of Demirgüç-Kunt & Huizinga (1999) started by the following equation:

Where Yijt is the dependent variable (either ROE or ROA or sometimes ROAA)

Shortly, more researchers have applied the basic concept of internal and external determinants and employ some other variables such as cyclical output, privatization ownership (Athanasoglou et al, 2008), credit risk ((Pasiouras & Kosmidou, 2007; Sufian, 2009) and the scope of studies varied from European regions to a particular developing countries such as Malaysia, or Philiipine Actually, there has not been a detail study that could employed the complete sets of data induced by Demirgüç-Kunt & Huizinga (1999), the latter authors have used some critical variables and combined with

their specific research aims at a particular country scope Some authors have tried to combine with other variables to test the distinguished impacts on profitability between foreign and local banks (Shen et al, 2009; Claessens et.al 2001) The relationship between foreign entry or foreign ownership would then be discussed in more detail within the next sections

Besides the bank-specific factors like Non-interest income, capital etc, macroeconomic factors such as GDP growth, money supply growth and inflation rate are also important variables for researchers to evaluate the impacts on bank profitability Indirectly reflecting on profitability, the external or macroeconomic factors mentioned above played a significant role that have influence on the enviroment host where commercial banks are considered as players For example, GDP growth slows down during recessions may unfavor indirectly reducing bank returns by deteriorating credit quality & increasing default (Sufian & Chong 208) Many authors such as Demirgüç-Kunt & Huizinga (1999), Pasiouras & Kosmidou (2007); Sufian (2009) and so on have mentioned different macroeconomic controlling variables such as GDP, inflation, industry concentration, money supply etc and their impacts on bank performance Other authors have employed cyclical output, or interest rate Some impacts may be reflected significantly while others may not No matter how the outcomes of these studies concluded, banking industry could be seen as the main activity among other country’s activities recognizing a logical impact from macroeconomic factors These factors and their impacts will be discussed more detail later

Bonin et al (2005) chose a different approach that indirectly evaluated the impact

of internal factors on profitability They used translog functional form with the model of Stochastic Frontier approach to estimate the profit efficiency and cost efficiency across

11 transition countries from 1996 to 2000 with the sample size around 856 bank-year observation They also found that bank size is porsitive and no relationship between different ownership strucutre significantly impact on profit Rather than that, they have suggested the better cost efficiency of foreign bank than other types of bank It has also employed the normal linear regression with ROA as the dependent variables to compare with profit efficiency from different ownership types

Suffian & Hibibullah (2009) employed the fixed effect regression for 220 year observation from 2000-2005 in China banking market to illustrate the effect to bank profitability They have found that liquidity, credit risk, and capitalization played a significant role to improve bank profitability while cost has negative impact on it The impact on joint-stock banks seems to be manified than state-owned ones Moreover, they also examined that economic growth have positive impact on profitability while money supply growth is found to have reversed effect The shorfall of this study is that the author only can observed the quite short period when the economic conditions are considered to pace with “change” quickly

bank-Similarly approach, Athanasoglou et al (2008) applied General Method of moment (GMM) technique on a dynamic panel data to investigate the changing impact

on profitability from different group of independent variables including bank specific characters, industry and macroeconomic variables They have employed Return on asset

as a dependent variable in Greek banking market from 1985 to 2001 and found that most

of the internal determinants except for size have anticipating signs on bank profitability Besides, the ownership status of the banks is insignificant in explaining profitability even the provate banks yield higher profit As similar result, other macroeconomic and industry concentration status were insinificantly affect to the final profitability except for positive impact by business cycle

In the scope of study regarding the Vietnam banking market, Vu and Nahm in

2013 investigate the effects of some both internal and external factors to profit efficiency

of 56 Vietnam banks for the period of 2000- 2006 They used a techniques called Tobit

or two step instrumental variables to examine the efficicency score There were four types of variables employed in that study They consist of, firstly bank-specific characteristics such as size, credit risks and so on; secondly the ownership feature (State-owned Banks or Foreign-owned banks ); thirdly transitional indicators; and finally environmental factors such as GDP, inflation The instrumental variable that they used are the fraction of capital over total asset They have found the size and better management ability remaining positive impacts on profit efficiency, while lower asset quality and high capitalization would somehow drive a bad influencing power Besides

the positive impact of high growth in per-capita GDP, it was better for banks to improve their profitability by the success of government in maintain a low-inflation rate

Another study conducted by Dinh (2013) to examine the impact of determinants

on foreign bank profitability specifically She has used least square method of fixed effect on the foreign bank dataset, domestic bank dataset, combined dataset and compared each others for robustness check The size of data they applied here is over 51 commercial banks operating in Vietnam from 2000 to 2012 They have extended the study by adding multinational factors such as parent bank’s profitability to better measure the impact on foreign bank profitability in Vietnam market It was found that most of bank specific factor, macroeconomic factors significantly affect bank profit In detail, total asset and other income redemmed positive impacts while foreign banks parent profit negatively affect on foreign bank profit Foreign bank thanks to it ownership advantage seem to perform better than doemstic banks since the competition arised since opening up policy to allow 100% foreign incorporation.GDP growth showed a signifcant impact on profitability of all banks

Similar approach except for multinational factors has been employed by Sufian to test the impact on profitability of Malaysian banks during 2000 – 2004 However, the findings have some differences where economic growth negatively affected the profitability and inflation rate had positive impact to improve it Besides, negative effects of credit risk loan concentration on profit also had been found meanwhile the positive ones are suggested within the apperance of capitalization, non interest income and expense

In conclusion, bank specific characters such as the items in Balance sheet and income statement and macroeconomic conditions have been studied and suggested to have significant relationship with bank profitability These effects can be different between foreign and domestic banks and within the different circumstances, they may show different signs with profitability

2.2 Foreign presence/ ownership and its impact on bank profitability

Some authors have studied on detail ownership types to understand whether any relationship between different ownership structure and profitability According to Ábel and Siklos in 2004, they assessed the privatization of the banking sector in Hungary that

took place during the 1990s The Hungary government steadily chose a path named

“strategic foreign partner” to allow foreign participation in domestic banking industry instead of foreign owned banks or so-called Greenfield banks (Bonin et al, 2005) to enter into this transition market at the early stage of financial reformation They also have tested that this strategy was considered successful when the variable to measure profitability significantly has responded positively This was very similar to the strategy that Vietnam used to develop its banking and financial market at transition stage They concluded that strategic foreign partnership is supportive in terms of reforming the emerging industries

Bonin et al (2005) also studied the ownership structure’s impact on bank performance but applied another techniques called Stochastic Frontier analysis (SFA) to measure the efficiency on bank performance in 11 transition countries Regarding the differences in bank performance measurements, they have found that foreign banks with

a strategic owner earn an average ROA that is much more than other ownership categories This result was driven by better cost efficiency controlled by foreign banks rather than domestic competitors

Historically, the research of Demirgüç-Kunt & Huizinga (1999) have found that foreign entry’s impact on profitability are significantly positvely in developing country rather than developed countries by employing a dummy variable named “foreign” with a threshold of 50% ownership stake In a country level study, Vu & Nahm 2013 have found that “The effect of being a foreign bank in Vietnam is also positive and significant, even though its magnitude is less than half of the effect of being a state-owned bank This is in line with the general finding in the literature that Foreign banks are more efficient than domestic banks on average in transitional banking sectors”

over the 1988–1995 period The result has shown the reduction in profitability, interest income, and overall operating expenses of domestic-owned banks are associated with the foreign bank penetration They applied the extension measurement based on the model of Demirgüç-Kunt & Huizinga, 1999 The basis estimated equation is as descibed below

Where yi is the performance indicator for country i, fi is the share of bank assets

and γ are parameters to be estimated To extent the study and fix the problem of joint endogeneity, the “change” of performance impacted by foreign presence in subsequent year has employed by Claessens et al 2001:

They investigated the change of performance by different factors in which foreign presence is proxied by marcroeconomic view The foreign presence proxy here is the number of foreign bank at 50% threshold ownership stake over total banks They have found that the large share of foreign presence, the more rduction in profitability that domestic bank may encounter through enforce domestic bank to a higher level of competition in short term and invest in operation, technonoly to improve efficiency These results encouraged policy implication as evidence that foreign bank entry improves the efficiency of the overall market and it would show a positive view on continuing opening up signal in policy point of view

The study of Shen et al in 2009 focused more on the foreign penetration both in the policy-making point of view and in the internal individual bank’s view They have evaluated the way that foreign bank or foreign penetrations may have impact on local China bank performance The method that they used to analyze those impacts from different edge was least square dummy variable method (LSDV) with the similar approach as equation (2.4) To proxy the foreign bank penetration or bank presence, they employed two variables called macro foreign penetration and micro foreign penetration

While the macro foreign penetration was measured by the fraction of number of banks with foreign strategic partners occupying at 5%-25% threshold over total number of banks, the micro one focus more on the proportion of foreign’s shareholding in each individual banks This first one is very similar to variable used by Claessen et al (2001) where the second one focus at the detail proportion shareholding to demonstrate the impact of foreign strategic investor to each bank They have found the positive relationship when relaxing the limitation for foreign factor could lead to the improvement of profitability nor reduction of cost That has been seen as a proven trend where more local banks have followed recently to allow more foreign partners to be onboarded in their businesses This positive influence can be attributed to technology transfers from foreign investors or the creation of improved competitive environments However, while employing foreign pesence in microeconomic view, which was found neither profitability nor cost Some reaons for this have been mentioned where costs and revenues increase after foreign partner were participating in the local banks Some might said that the cost increased by investing in human capital, training and even the technology enhancement to provide the advanced level of competition with the short term expense increase Another thoughts were derived from the difference in culture and strategic objectives of their own local main shareholders Besides, lack of employees who are fluent in English and who are familiar with international affairs might also be a barrier for local banks to communicate and cooperate successfully with foreign partner’s representative working in their banks Those have led to the repressed impact from introducing foreign strategic partner in a bank’s operation That’s why, Chinese government considered if the cap of foreign investment proportion at 20–25 percentage was appropriate

Unite & Sullivan (2003) also follow this study approach to examine the effect of foreign entry in Philippine banking market when it has gradually reform its domestic market by liberalizing restrictions on foreign investment By employing the random effect model, they have found that in general the results are somehow similar to what Claessens et al (2001) have found Interest rate spreads and operating expense declined once more foreign penetrate into domestic market due the hypothesis of increasing competition and efficiency However, it was only found true for those domestic banks

that are affiliated to a family business group Foreign entry corresponds more generally with improvements in operating efficiencies, but a deterioration of loan portfolios The increase in the percentage ownership by foreign investors in domestic banks is shown to result in an increase in operating expenses and a decrease in noninterest income – a proxy of accounting profit where Philippine banking industry encourage banks to pursuit this approach rather than traditional activities They believed liberalization of foreign presence will have positive effects on economic growth and will enhance the ability of the economy to overcome negative external shocks

To conclude, there are different approaches to test the impact of foreign presence

to bank profitability It could be found that foreign presence represented by the number

of banks with foreign ownership’s influence over total number of banks might be considered to have some relationship with bank profitability However, the influence may be different if we employed another variable to represent foreign presence such as foreign proportion shareholding or dummy variable with specific threshold These approaches have yielded different results so that it is still a critical concern for research

Table B.1 (Appendix B) below summarizes all relevant empirical studies discussed in this study They greatly provide the audiences a better and concrete view on the theories that have been demonstrated and introduced above

Based on the literature reviews, the impacts of internal and external determinants together with foreign presence factors on bank profitability are graphed by figure 2.1

Figure 2.1: Analytical framework of determinants of bank profitability

The diagram shows generally the link among estimated variables in this study Dependent variable is Return on asset The internal key set of indicators of bank profitability is used as independent variables consist of size, capital, credit risk, non-traditional income, expense, liquidity Besides, the model also involves a series of macroeconomic explanatory variables such as Inflation, GDP growth, and money supply

ownership/shareholding” are employed here to test the significant relationship with bank profitability

Bank profitability Return on asset

Macroeconomic factors

Inflation rate (+/–)

Money supply growth (+)

Foreign presence factors

CHAPTER 3: RESEARCH METHODOLOGY AND DATA

This chapter includes four main parts The first part steer our attention to the introduction of suggested research methodologies considered to apply in this research Secondly, the model specification will be drawn out to answer for the research questions Third section provides the audience with the comprehensive introduction of variable measurement to answer the research question Finally, the estimation strategy and data collection should be discussed for this study

3.1.1 Panel regression model’s approaches:

Firstly, the common constant method, Pooled Ordinary Least Square or so-called Pooled OLS is brought to consideration generally The equation below described the general model for the common constant linear regression method:

The common constant method, so-called the pooled OLS method, could estimate

a common constant α for all cross–section data In other words, this estimation assumes

there is not any existing difference across estimated sections Besides, the error term uit

simple estimation apparently appears with some shortfall leading us to look for the application of other better model Firstly, when considering the application of this approach against others common panel regression model such as fixed or random effect model, the restrictions in constant intercept and slope coefficients achieved by the pooled OLS model seem to be flagged up According to Baltagi (2008), the first assumption of this technique is there is no serial correlation existing in observations, and the errors terms are homoscedasticity:

Then, another assumption of independent and identical distribution of error term rings a seemingly negligence of the panel data structure Finally, the same intercept applied for all banks together with the same coefficients of independent variables constantly unchanged across time These shortfalls would bring the biased estimation to measure the actual relationship of this model Therefore, the pooled OLS method appears to be inappropriate for this research Instead of going with this common approach, the study drives the audience to different estimation techniques where more studies apply in practice (Baltagi, 2008)

The second estimation approach introduced here is Random effects method

(REM) which has the general econometric idea described as follows:

The error term of this random effect model consists of two components The first

individual time-variant effect that is different unsystematically across individuals and

words, REM can only estimate the effects of time invariant variables, not to control for omitted variables This will give us the biased estimated outcomes Lastly, the error term

of REM assumes to absorb the different characteristics of each entity That definitely drives the audience to search for other techniques

Fixed effects method (FEM) is introduced here as a key technique to alter the

drawbacks of the first two approaches The following model describes the illustration of FEM:

FEM assists researchers to demonstrate the intercept difference across entity given the constant slope but it does not capture the effect of time variance across time This method is conducted through some assumptions First assumption is mentioned that the differences across sections are explicit by the intercept besides meanwhile the second one dealt with the missing correlation between predictors and error term In other words,

very important concept where at first stage of our estimation if our study assumes there

is underlying omitted variables correlated with variables in the model, this model may provide a tool to control for this omitted bias and cease the impact of time – invariant characteristics such as culture, customer base, etc from the predictors Consequently, the estimated coefficient of FEM would not provide biased circumstance This model is also called Least squares dummy variable (LSDV) because it is dealt with the generation of corresponding entity dummy variables to measure the individual effect However, the reductions of degree of freedom apparently ring disadvantages while using this approach rather than others

3.1.2 Test for an appropriate panel data regression choice

How can the researcher make his choice to select the appropriate model to run the study? That depends on the dataset and, of course, some other tests to select the most suitable panel estimation approach To make a choice between FEM and pooled OLS, according to Baltagi in 2008, the F-test is the best way to run The regression model of

(3.5)

Where RSS is the restricted residual sum of squares of pooled OLS model, and URSS is the unrestricted residual sum of squares of LSDV regression Technically stated, once the null hypothesis is rejected, we could state that at least one time specific

rather than the pooled OLS with this conclusion

Second test to illustrate the better selection between REM and pooled OLS is

called Breusch – Pagan LM test for random effect This test’s ideally examine whether

the variance of individual or time specific character is equal zero or we may say the null

(3.6) Technically results leads to the favorable selection of Random effect when the null hypothesis is rejected or vice-versa

After the first two selection tests, the dilemma where the researcher should choose one of the model to run has fall to conduct the test of Hausman Specification Test Since the differences in assumptions of correlation between time – invariant

uncorrelated circumstance appears The results of Hausman test can rely on the test of correlation strength by the following econometric test model:

The null hypothesis Ho in this model is testing the consistency and efficiency of

is large, the estimates are significant differences, thus the null hypothesis can be rejected It means that the fixed effects estimator is better

According to Park (2011), to conclude, there would be three basic tests for selecting the appropriate panel regression model F test should be conducted firstly to decide whether the fixed effects is preferred than pooled OLS Then, LM test drew the decision of selection between random effects against pooled OLS before applying the

Hausman test makes a choice between fixed effects and random effects model if we fall into the dilemma situation of the two models To better illustrate the test used, the following table is compiled for clearer understanding:

Table 3.1: Tests for choosing the appropriate panel regression model

No Fixed Effect (F

Test)

Random Effect (Breusch

(No fixed effects)

Ho is not rejected (No random effects)

Pooled OLS

(Fixed effects)

Ho is not rejected (No random effects)

Fixed effects model

(No fixed effects)

Ho is rejected (Random effects)

Random effects model

(Fixed effects)

Ho is rejected (Random effects)

Run Hausman test If Ho is rejected, selected model should be fixed effects model; otherwise, choose a Random effects model

3.2 Model specification

Basically, the approach that this study will follow the way that Sufian & Chong (2008) and Shen et.al (2009) and apply the least square method in which fixed effects model (FEM) is essential considered The standard error will then be controlled for the cross-section heteroscedasticity However, to ensure the appropriate model is prefered than random effects model, this study would consider to conduct Hausman test as a necessary step To reaffirm the relevance and correctness of choosing Fixed effect model, the author would like to compared the results with appropriate econometric measurements of Pooled OLS and REM against FEM For other sound variables such as internal and external determinants from previous studies, the author may employ them in further research if found availability and necessarily Especially, the research key element, foreign presence, will be illustrated and embedded into the regression model This approach is very similar to the study of Shen et al (2009)

The scope of our study was the representing data set of Vietnam banking industry, more than 45 banks including State-owned banks, 100% Foreign-owned, Joint venture and local joint-Stock banks during the period of 2004 to 2013 This dataset is collected as an unbalanced pooled data because of the different time of enter and exit market Basically, the data sources that would be used in this section was supposed to collect through Bankscope and bank’s annual report, others sources to collect appropriate bank-wide level as required

For the first research question to examine the impacts of internal factors and external factors on Vietnamese bank profitability, the original regression model is built

to illustrate the relationship between internal & external factors and bank profitability:

Next research hypothesis invloved the application of foreign presence variable to identify whether any impact of foreign bank’s entry exists on bank profitability Similar

to Claessen et al (2001) and Shen et al (2009), to answer this question, the equation with foreign presence variable from the macroeconomic view is employed to compare with the original equation:

by dividing the numbers of foreign-influence bank to total numbers of bank at time t In

relationship between foreign presence and profitability

Finally, to answer the question flagged up to measure the foreign factor’s impact

at the bank level on bank profitability, the proxier named Foreign shareholding is employed to run the regression model as belows:

supposed to appear as significant, this represents the potential relationship between foreign shareholding’s influence and profitability

Generally, three comprehensive extended empirical models used in this study are derived from the equation (3.8), (3.9) and (3.10) respectively as follows:

Where the definitions, notation and the expected effect of the explanatory variables for the empirical model on bank profitability is described as the following table:

Table 3.2: Definition, notation and expected signs of relevant variables

Internal determinant (bank specific factors)

high capital asset ratio is assumed

to be indicator of low leverage and

therefore lower risk

Habibullah (2009), Athanasoglou et

al (2008), Dinh (2013), Sufian &

etc

Variable Definition Notation

Expect-ed sign

Author

Credit

risk

Loan loss provisions/total asset

measures how much is the bank’s

provision provided for against it

A measure of diversification and

business mix, calculated as

non-interest income/total assets

accounting value of the total assets

of the bank in year t

Habibullah (2009), Athanasoglou et

Kosmidou (2007), etc

Liquidity Loan to customer deposit and

short term funding, represents the

liquidity level of bank’s activities

Kosmidou (2007), Shen et al (2009)

expense/total assets and provides

information on the efficiency of

expenses relative to the assets in

year t Higher ratios imply a less

efficient management

Habibullah (2009), Athanasoglou et

al (2008), Dinh (2013), Sufian &

Variable Definition Notation

Total number of bank with foreign

strategic partner influence* over

total number of operating banks

Foreign shareholding proportion to

total bank’s stake

Variable Definition Notation

3.3 Variable measurement and test hypothesis

3.3.1 Bank profitability definition

Generally, as mentioned in literatture above, proxy of bank profitability could be determined as the return of asset (ROA) or return of equity (ROE) derived by taking the bank’s net income divived by its total asset or total equity The key difference here was how the management ability to utilize its bank asset to generate revenue and provide profit to all stakeholders meanwhile the ROE only considered the purpose of shareholder’s profit maximization purpose In other words, ROA reflects the ability of a bank’s management to generate profits from the bank’s assets, although it may be biased due to offbalance-sheet activities ROE indicates the return to shareholders on their equity and equals ROA times the total assets-to-equity ratio or it was said to reflect the leverage in terms of return Average assets are being used in some study, in order to capture any differences that occurred in assets during the fiscal year (Pasiouras & Kosmidou, 2007) Even most of the literatures reviewed used ROA as the key dependent variables to serve for their studies, some authors such as Vu & Nahm (2013) have applied another aspect to measure the profitability called profitability efficiency They have used the Tobit model and two stage Instrumental variable regression which the main factors derived from set of bank specific indicators such as size, capital, etc No matter what the different and result of many methods that were used are, among those, most of the authors have employed ROE as the key variable to measure the bank profitability thanks to its in this scope, to limit the impact of tax shield, particular country-based data and other impacts on net return, profit before tax would be used as a numerator of this fraction

3.3.2 Internal determinants

a Capital

Within the set of internal determinants, Capital is the first bank-specific variable that essential expectedly affect profit Total Equity to total assets or sometimes so-called leverage level (“EQASS”) was considered as proxy to captial It measured to which extent the bank maintain its capital structure to serve the operational activities The expected sign of this relationship is mixed since the banks may concern the issue of having high or low level of capital Theoretical evidence showed that if the banks have low capitalisation ratio probably leading to an increased default risk related to meeting is obligation, while high one probably encounters results in higher costs of fund and finally deriving a poor return This reason were also found in Sufian, 2009; Demirgüç-Kunt & Huizinga, 1999 and most of the economists

b Credit Risk Management

Secondly, the bank operation would closely expose to Credit risk within main lending activities To proxy this variable, this study employed the loan-loss provisions

to total asset ratio (LLP_TA) In the sense of risk and accounting principle, the increasing exposure to credit risk representing by the amount of its loan loss provisioning provided for is generally accompanied with the reduction in profitability Hence, a negative relationship between ROA and profit is expected Actually, within a bank’s Balanced Sheet, Total loan almost covered its total asset where the leverage level was found to be significantly large Indeed some studies have recruited Loan loss provision over Total loan as the credit risk representative followed the principle of accounting treatment (Sufian and Chong, 2008) However, to be more consistent with the accounting identity, total asset is traditionally used to proxy this factor and reflect more meaningful in terms of credit risk especially in Vietnam market (Demirgüç-Kunt

& Huizinga, 1999) To improve the profitability, management definitely should control its own policies and improve the risk control process meanwhile different central banks would adopt different types of provisioning standards This approach was also used by Sufian, 2009; Athanasoglou et al., 2008; Kosimdou, 2008 However, some authors would prefer the non-performing loans over total loans to quantify banks asset quality such as Vu and Nahm 2013 Besides, to illustrate to what extent the expsoure of banks to

its credit risk potentially was, Demhguc-Kunt and Huizmga in 1999 used the proportion

of loans total assets representing the risk level as higher the ratio is, the higher exposure the bank is once the economics turn bad, it would negatively impact the profit if the default rate was too high

c Expense management

Next, one of the most important factor could yield the relationship towards banking profitability is Cost or expense management The debate in which the introduction of different variables to proxy this have been found recently In terms of expense management concept, cost to income ratio was used by Kosmidou, 2008 and Kosmidou & Pasiouras, 2007 They had emphasized that it was useful by providing information on the efficiency of the management regarding expenses relative to the revenues it generates Higher ratios imply a less efficient management style Some researchers employed Operating expenses over total assets such as Athanasoglou et al

2008 or different approach by using overhead over total assets to measure the variation

of management to maintain its quality & services, product mix as stated in Kunt and Huizmga 1999 Following the study of Sufian and Chong in 2008, non-interest expense over total asset is used to capture the management capbablity to control the banks expense over its total activities The relationship between this variable and profitability levels may be negative, as banks that are more productive and efficient aim

Demirguf-to minimise their operating costs Generally, in the scope of this study, the fraction of bank’s non interest expense or socalled overhead on total assets would be employed to illustrate this relationship Followed by other studies, the hypotheised impact of this variable to banking profiatbility would be considered negative

d Non - interest activities

Moreover, in this world of financial development, more and more products beyond traditional activities of a commercial bank have been considered significantly important Demhguc-Kunt and Huizmga in 1999 has also recognized its important on banking profitability Unite and Sullivan in 2003 also employed this variable but played

it as other role when considering it as proxy for accounting profit to measure the impact

of foreign entry To recognise that financial institutions in recent years have increasingly been generating income from “off-balance sheet” business and fee income general, the

ratio of non-interest income over total assets (“NII_TA”) is entered in the regression analysis as a proxy for non-traditional activities Non-interest income consists of commission, service charges, and fees, guarantee fees, net profit from sale of investment securities, and foreign exchange profit The ratio is also included in the regression model

as a proxy measure of bank diversification into non-traditional activities The variable is expected to exhibit positive relationship with bank profitability The more penetration of banking’s non-traditional activities is, the more income it would yield Actually, not all the bank could reflect the same level of non interest income as others We may found the proportion of this is higher where it came from the investment of fee generation and other advanced technologies’s outcomes Hence, the higher this ratio is, the higher bank profits is expected

e Liquidity Management

This is a measure of liquidity calculated as loans to customers and short term funding Higher figures denote lower liquidity Liquidity risk is always a critical concern for bankers To represent to what extent how the banks manage it liquidity management policy in terms of assets and cash flow to serve their obligations Banks may quickly fail

or temporary insolvent if the liquidity management is badly maintained The ratio of total loans to customers deposit and short term funding (“TL_DSTF”) is employed to measure the relationship between liquidity & performance This may show the how tight the illiquid assets such as loans when they were disbursed and the stable funding sources such as deposits and other short term funding Hence, the negative relationship between this ratio value and liquidity level the banks meet And in terms of risk-return paradigm, good liquidity level would yield lower return So it will represent the negative relationship with profitability

f Bank size

Last bank-specific variable is Bank size represented by the natural logarithm of Total asset as a proxy of size to capture the possible cost advantages associated with size In other word, different economies of scale cound account for different level of return The original variable derived from the absolute accounting value of bank assets introduced by Demihguc-Kunt & Huizmga (1999) In the literature, mixed relationships are found between size and profitability, while in some cases a U-shaped relationship is

observed This is also used to control for cost differences related to bank size and for the greater ability of larger bank to diversify In essence, This variables may have a positive effect on bank profitability if there are significant economies of scale On the other hand,

if increased diversification leads to higher risks, the variable may exhibit negative effects The same concept was also employed by most of the authors such as Sufian (1999); Athanasoglou et al (2008); Claessens et al (2005); and so on Hence, the expected sign of this relationship could appear at positive or negative effect

3.3.3 External determinants

In terms of external determinants, to measure the relationship between macroeconomic condition and bank profitability, GDP growth rate, money supply growth and annual change in inflation rate are used

a Gross domestic Product (GDP)

According to Sufian & Chong (2008), the gross domestic product growth (“GDP_G”) is one of the most common macroeconomic indicators for measuring an economy's total economic activity Banking industry which was considered as a extreme important to operate the whole economy Hence, GDP defintely is the key impact derived by numerous determinants related to the supply and demand for loans and deposits Sufian (2009) in his studies has determined that GDP growth slows down particularly during recessions, credit quality deteriorates, and defaults increase, thus reducing bank returns GDP per capita sometimes is employed to test for the macroeconomic condition of a country where banks are operating It reflected the similar relationship trend where it proved the positive impact on bank profitability (Vu & Nahm 2013) Therefore, we could expect the sign of this relationship to be positive

b Inflation

Another important macroeconomic condition possibly affecting both the costs and revenues of banks is the inflation rate (INF) The same macroeconomic theory suggests that inflation may have direct effects on increasing the price of labour, or any produtction input and indirect effects such as changes in interest rates and asset prices

In case of anticipated inflation, interest rates would be adjusted prudently resulting in revenues against its cost This would then affect positively to profitability (Pasiouras and Kosmidou, 2007) and vise-versa These effects accounts for the change in bank

profitability negatively Kosmidou in 2008 and even Demihguc-Kunt and Huizmga in

1999 in their study have proved inflation held a negative impact on profitability However, once applying this into individual country such as Phillipine or Malaysia, it seems that its relationship is not always significantly proven or sometimes investigated

as positive impact because of the period of dataset and its expected level falling at the anticipating inflation rate (Sufian, 2009) Even these studies mentioned above have different results, as the basic and logical circumstances, this result reconfirmed the basic concept of the macroeconomic conditions’ impacts on bank profitability prudently that lower inflation rate would encourage banks to create more profitability subsequently (Vu

& Nahm 2013)

c Money supply growth

Another important variable is money supply (“MSG”) in which we expected that changes in the money supply may lead to changes in the GDP and the price level respectively Especially in Vietnam, this is the crucial function of State bank to ensure the impact of its to economy as a whole is controllable Although the money supply is basically determined by the central bank's policy, it may also affect the behaviour of households and banks In the study of Sufian and Chong in 2008, they have found the money supply acted as a measure of market size and postively affected bank profitability Following among others, Kosmidou (2008) has used the growth of the money supply was used in his study and found that there was no significant impact on profitability In contradiction to what have been found by Kosmidou, Sufian & Habibullah (2009) has found the negative sign of this relationship In this case, the study would hypothesize the mixed sign of this relationship exists

3.3.4 Foreign presence and foreign shareholding

One of the objectives of this study focusing on the effect of foreign presence or ownership to bank profitability The study of Demirguf-Kunt and Huizmga in 1999 applied the dummy foreign ownership variable which equals 1 if at least 50 percent of the bank's stock is in foreign shareholding and equals 0 otherwise Testing for developing & developed cases had suggested that foreign banks realized high profitability in less developed countries It implied the circumstances where foreign banks had been avoided the unfavorable domestic banking regulations once they were

local incorporations and apply advanced banking techniques and management skills in that local market Foreign ownership would not be the unique variable type once considering the impact of ownership type to bank performance Shen et al, 2009 and Claessens et al, 2001 have applied the foreign penetration from the macroeconomic point of view represented by the total number of banks with significant influence of foreign factors over total number of banks They have found that in developing countries where authorities are encouraging the foreign penetration to boost up the financial liberalization and enhance their local market, foreign factors are considered to have larger impact on profitability It led to the encourangement of opening up policy to have spillover effect on the economy Therefore the foreign presence factor at the macroeconomic level would be expected to positively influence domestic bank profitability In contradiction, the research of Unite & Sullivan 2003 found other result where foreign bank entry tend to have negative impact on interest rate spreads and bank profits, especially for those domestic banks with their close affiliation with family business group because of high competition level they have brought into the domestic Phillipine banking market

At the enterprise level, Shen et al 2009 have applied the fraction of foreign bank shareholding at each individual bank to measure the influence of those foreign factors However, neither profitability nor cost has been found to be impacted by this proxier They have expained that its result is derived from the mutual exclusive effect where the foreign factor enhance both profits but also expense at the same time while investing in technology and operation Therefore, at this study, the author would like to employ two proxiers to measure the foreign presence impact First, we expect the positve influence

of number of foreign impacted bank over total number of commercial banks to bank profitability, so-called foreign presence (“FP”), is significant A bank defined as foreign- influence bank at this study is a bank with more than 10% at Vietnam market seems to have executing power within the bank’s decision making process Since Vietnamese market is capping individual foreign ownership at 20 per cent stake in a local bank, while the total non-Vietnamese ownership is limited at 30 per cent (Vietnamnews, 2013) Most of the acquitision deals from foreign ownership varies from 10% to 20% which they might assign representative to be member of Board of directors.Second, the