Luận án Mối quan hệ giữa hiệu quả sử dụng nguồn lực của các ngân hàng thương mại và tăng trưởng kinh tế tại Việt Nam

physical resources - the scale of fixed assets and the last is the financial resources - bank funds raised from clients to conduct business activities.  This study will use the parametric Stochastic Frontier Analysis approach (SFA) and the non-parametric Data Envelopment Analysis approach to measure Vietnamese commercial bank efficiency in the research phase.  Aside from measuring the bank efficiency with SFA and DEA, this study will apply the method of financial ratio analysis to evaluate commercial banks efficiency in Vietnam to compensate the defect of SFA and DEA. 1.3. ECONOMIC GROWTH 1.3.1. Theories of economic growth 1.3.2. Models of economic growth  Ricardo's model of economic growth (1817)  Keynes's model of economic growth (1936)  Harrod-Domar model of economic growth(1940)  Lewis's dual sector model of economic growth (1955)  Solow's model of economic growth (1956)  Kaldor's model of economic growth (1957)  Modern models of economic growth 1.4. THE RELATIONSHIP BETWEEN EFFICIENCY OF COMMERCIAL BANKS AND ECONOMIC GROWTH Banking system is an important part of the financial system of each country. The relationship between the development of the banking system and economic growth is also built from the relationship between the development of the financial system and economic growth with the expectation that operating efficiency in banking system will help to increase the efficiency of capital allocation and capital flows in the economy contributing to the economic growth of each country. 1.4.1. Theory of the relationship between efficiency of commercial banks and economic growth 6 According to Wachtel (2001), there are four channels showing the relationship in which the operating efficiency in banking system will promote economic growth:  The banking system plays the role of financial intermediaries to transfer funds in the economy from people who have extra money or surplus savings to those who do not have enough money.  The operating efficiency in commercial banking system and financial institutions will provide the tools, financial products more attractive and relevant to the needs of actors in the economy.  The operating efficiency in banking system will help to allocate funds for best projects through evaluation of loan requests and the administration of loans and the supervision of implementation  Finally, the operating efficiency in banking system will help the economy actors may reduce risks through oversight and disclosure of financial information of entities undercapitalized. On the contrary, economic growth is also to create or support the conditions for development of the banking system. The studies of Rachdi and Ben (2011), Liang and Reichert (2006) argues that while economic growth will spur the development of the financial system in general and the development of the banking system in particular through channels:  When there is economic growth, income increases, which will facilitate banking system to attract a lot of lower-cost capital that helps increase efficiency of resources and system development.  Economic growth has also created an opportunity for economic actors expanding production and business activities, which impacts on the expansion of the bank credit. When bank credit is expanded along with operation plan, the effective business will help bank to increase revenue thereby improving bank efficiency contributing to the development of the banking system.  When economic growth helps the banking system to expand and increase the activity and profitability in the banking sector thereby encouraging entry to banking sector. When demand to entry to banking sector increases, which creates the more competitive environment in the banking system, putting more self-renewal pressure on banks to develop banking system. 1.4.2. Overview of studies on the relationship between efficiency of commercial banks and economic growth The capital flows of the economy are made through two main channels: direct finance channel is performed on the financial market and indirect finance channel is performed by banks and financial institutions. The banks play an important role in the financial system of each country. The relationship between efficiency of commercial banks and economic growth is shown by the results of cross-country and domestic studies. 7 CHAPTER 2. RESEARCH MODELS AND METHODOLOGIES OF THE RELATIONSHIP BETWEEN EFFICIENCY OF COMMERCIAL BANKS AND ECONOMIC GROWTH The purpose of the Chapter is to introduce methods and models that dissertation uses include: (i) Methodology in the measurement of bank efficiency; (ii) Research model and methodology of analyzing the determinants of commercial bank efficiency; (iii) Research models and methodologies analyzing the relationship between efficiency of commercial banks and economic growth. 2.1. METHODS OF MEASURING COMMERCIAL BANK EFFICIENCY The recent studies on banking efficiency in the world usually use two main methods: method of financial ratio analysis and efficient frontier analysis method. 2.1.1. Method of financial ratio analysis The method of financial ratio analysis is a common method to measure bank efficiency. This method can use a large number of application indicators and the interpretation of results may be difficult. In other words, the financial indicators alone provide too little information to understand banking operation status clearly. These facts require the reader to have in-depth knowledge about banking and finance sector. 2.1.2. Method of parametric stochastic frontier analysis Parametric analysis or analysis based on parametric stochastic frontier model was introduced by Aigner et al (1977), and Meeusen and Broeck (1977). The first research model involved production function model with panel data and had the noise divided into two main components: (i) technical inefficiency; (ii) the random error term. The model of the form: Yi = xi + ε i=1,...,n, (2.7) Technical efficiency of banks are calculated with the following equation: TE = Yi exp(xi+Vi) = exp(xi+Vi-Ui) exp(xi+Vi) = exp (-Ui) (2.8) With  is output of bank i on efficiency frontier. Therefore, the technical efficiency of bank i is the ratio between the actual output Yi with estimated output on the effective frontier . Battese and Coelli (1988) estimate the technical efficiency through the following equation: 8 TE = exp (-Ui)=E exp -Ui εi = 1- ( *- * * ) 1- (- * * ) exp - * + 1 2 * 2 (2.9) The cost function estimation is formed via (Coelli et al 2005): lnCi = + i M m ni N n n YW lnln 11     + ε (2.10) Estimating the cost efficiency of the bank was formed from the following equation: CE= lnC lnC* = exp ( 0 + i M m ni N n n qw lnln 11     + Vi-Ui) exp ( 0 + i M m ni N n n qw lnln 11     + Vi) (2.11) With exp( 0 + i M m ni N n n qw lnln 11     + Vi) is the output of bank i on effective frontier. Therefore, the cost efficiency of bank i is the ratio between the actual cost lnC with estimated output on effective frontier lnC * . In summary, the stochastic frontier approach allows to identify effective technical and non- technical efficiency for each bank by splitting outliers of the model into random unobserved and the statistical error from technical inefficiency caused. However, this approach must designate the functional form and the distribution of inefficient. If we choose the incorrect functional form, the bank efficiency scores will not correct. 2.1.3. Method of nonparametric frontier analysis Unlike the method of parametric stochastic frontier analysis, method of nonparametric frontier analysis does not require to designate the functional form for production function or cost function of banks. Data envelopment analysis (DEA) was developed by Charnes et al (1978) to evaluate the efficiency of public sector with multiple inputs and outputs. Models of Charnes et all (1978), also known as the CCR model is used to measure the overall technical efficiency of banks. CCR model is built on the assumption that constant returns to scale (CRS), but in fact, banks operate at variable returns to scale (VRS). Therefore, Banker et al (1984) proposed more assumptions to measure bank efficiency with impact of scale or variable returns to scale assumption in the BCC model. CCR model and BCC model only use the amount of input and output, so it only allows to measure technical efficiency, but not allows to measure the allocation efficiency of banks. 9 When there is data on the prices of inputs, DEA method extends to measure cost efficiency or overall economic efficiency. The cost efficiency calculations are done through maximizing target revenue by orientating output or minimizing costs by orientating input. 2.1.4. Bank outputs and inputs in commercial bank efficiency study In this study, the banks are viewed as financial intermediaries, providing financial services and payment services to entities in the economy. Therefore, bank inputs are selected with 3 input variables: staff costs (I1), fixed assets (I2); customer deposits (I3); and the bank output variables include interest income (Y1); non-interest revenue (Y2) including net income from services, net income from trading activities of securities trading, investment and net income from other activities. The representative variable cost includes (W1) staff costs, the cost of the average fixed assets (W2) and the average interest expense (W3). 2.2. RESEARCH MODEL AND METHODOLOGY ANALYZING THE DETERMINANTS OF COMMERCIAL BANK EFFICIENCY Due to the characteristics of the dependent variables are censored or truncated variables , the value ranges from 0 to 1, therefore dissertation used Tobit regression model (or censored regression model - censored regression model) was introduced by Tobin (1958) where 0 and 1 denote the lower and upper limit and dependent variables receive ongoing value from 0 to 1. In accordance with the unbalanced panel data (unbalanced panel data) in this study and the dependent variable respectively technical efficiency and cost efficiency, the model was deployed: TEit= T it+ 2ETAit + T it + T it + RT it + R it + it + i (2.20) CEit= T it + 2ETAit + T it + T it + RT it + R it + it + i (2.21) Detailed description of the variables in the model (2.20) and (2.21) is presented in table 2.3. The group of independent variables includes: (i) Percentage of customer deposits per total assets measures the ratio of customer deposits in total capital structure of banks, this study expects that when the ratio of DTA increases, the Vietnam commercial banks have more resources to expand the business operations, thus increases the bank efficiency; (ii) the capital structure of the bank is measured by the ratio between equity and total assets (ETA), 10 it shows bank’s capital structure and capital strength that is expected the same sign with commercial banks efficiency; (iii) The percentage of outstanding credit to total assets (LTA) measures the scale of credit activities of commercial banks, this study expects that when the bank’s credit scale is expanded without affecting on the quality factor, the bank efficiency is improved. (iv) The ratio of loans to total customer deposits (LTD) while maintaining at appropriate levels may increase the commercial bank efficiency, but when this ratio is too high, there is pressure to increase the cost of raising fund to reduce the commercial bank efficiency; (v) The quality of bank assets is measured by the ratio of bad debt reserve per total loans (RTL), so banks with higher bad debt reserve ratios show lower asset quality; (vi) the profitability rate of the bank as measured by return on average total assets (ROA); (vii) the bank's size is measured by the natural logarithm of the total assets of banks (SIZE). This study expects that the bank’s scale is expanded for banks taking advantage of scale to help in increasing bank efficiency. To analyze the determinants of commercial bank efficiency, this study use Tobit regression model with the help of software STATA for unbalanced panel data in the period 1992 - 2013 with 48 commercial banks in the sample of study. 2.3. RESEARCH MODELS AND METHODOLOGIES ANALYZING THE RELATIONSHIP BETWEEN EFFICIENCY OF COMMERCIAL BANKS AND ECONOMIC GROWTH After the brief survey of relevant studies and business situation of the banking system in Vietnam, this study builds the ideas based on research model of Odedokun (1996) to analyze the relationship between efficiency of commercial banks and economic growth in Vietnam: Yt = f (Lt , Kt , Ft , Zt) (2.22) In which, Yt is gross national product or real GDP, Lt is the labor force of the country, Kt is capital, Ft measure the development of the financial system and ZT are other factors affecting economic growth. Based on the model of Odedokun (1996) to analyze the 11 relationship between efficiency of commercial banks and economic growth in Vietnam, this study used models: Yt = f (EFFk , Zt) (2.23) Yt is gross national product or real GDP, EFF is average efficiency of bank and ZT are other factors affecting economic growth. To deploy model (2.23), combined with variable M2 - money supply M2 to GDP ratio, the relationship between efficiency of commercial banks and economic growth in Vietnam is tested through model: GDPt = + i EFFt-i k i=1 + j M2t-j n j=1 + εt (2.24) GDP is annual percentage growth rate of total gross domestic product represents economic growth, EFF is annual percentage average of commercial banks efficiency, EFF is measured through SFA and DEA method. Money supply M2 to GDP ratio represents the development of the national financial system and εt is the noise reflected the impact of other factors to economic growth. Model (2.24) is deployed into two detailed models with independent variables respectively SFA_TE - technical efficiency under SFA and DEA_TE - technical efficiency under DEA: GDPt = + i SFA t-i k i=1 + j M2t-j n j=1 + εt (2.25) GDPt = + i DEA_TEt-i k i=1 + j M2t-j n j=1 + εt (2.26) Accordingly, EFF represents the commercial banks efficiency and M2 represents the development of the national financial system are expected to impact the same way with economic growth in Vietnam. The process of analyzing the relationship between efficiency of commercial banks and economic growth in Vietnam through the time series data is done according to the procedures detailed in Figure 2.3. CHAPTER 3. THE BUSINESS OPERATION OF BANKS AND ECONOMIC GROWTH IN VIETNAM Before measuring Vietnamese commercial banks efficiency as well as analyzing the relationship between efficiency of commercial banks and economic growth in Vietnam, dissertation analyzed overview of economic growth and banking operations in Vietnam in the period 1992 - 2013. The purposes of this chapter include: (i) Analyze overview of economic growth in Vietnam during the period 1992 - 2013; (ii) Analyze the role of the 12 business operation of banks for economic growth in Vietnam through the implementation of the basic functions of commercial banks; (iii) Evaluate the role of the business operation of banks for economic growth in Vietnam over the period 1992-2013. 3.1. OVERVIEW OF VIETNAM’S ECONOMIC GROWTH In the period 1992 – 2013, GDP growth rate of Vietnam averaged 6.95% per year in which the highest level is 9.54 % in 1995 and the lowest was 4.77% in 1999 due to the impact of the East Asian financial crises - Thailand in 1997, according to the World Bank (2014). At the chart of GDP growth rate of Vietnam in the period 1992 – 2013, economic growth of Vietnam was clearly affected by two crises in this period, the East Asian financial crises - Thailand in 1997 and the global financial crisis in 2007-2011. 3.2. THE ROLE OF BUSINESS OPERATION OF BANKS TO VIETNAM’S ECONOMIC GROWTH On 31/12/2013, the Vietnamese commercial banking system include: 5 state-owned commercial banks including commercial banks in which the state holds the controlling share, 33 joint stock banks, 4 joint venture banks and 5 wholly foreign-owned banks. In the period 1992 – 2013, the business operation of bank plays an important role for economic growth in Vietnam. During the study period, the ratio of bank credit to GDP peaked in 2010 when this rate was 124.66%. In the period 1992 – 2013, the ratio of bank credit to GDP continuously increased from 15.71% in 1992 and reached 108.15% at the end of 2013 reflected the increasingly important role of the banking system against Vietnam's economy. In the period of 1992 - 2013, with the ratio of bank credit to GDP continued to increase, the rate of annual credit growth in the period is always achieved at a high level. In 1993, credit growth rate of Vietnam had the highest growth rate reached 81.42%, followed by 2007 with a growth rate reached 49.73%. In the period 2009 – 2013, when credit growth continued to decline from 44.97% in 2009 to only 12.52% level in 2013 due to the effects from the global economic crisis as well as negative impacts on the growth speed of the economy during the same study period. 3.3. EVALUATING THE ROLE OF BUSINESS OPERATION OF BANKS TO VIETNAM’S ECONOMIC GROWTH 13 Through the process of analyzing the role of business operation of banks to Vietnam’s economic growth in the period 1992 – 2013, this study found that there are still a bit of constraints to overcome to enhance the role of business operation of banks contributing to economic growth in Vietnam. CHAPTER 4. ANALYZING THE RELATIONSHIP BETWEEN EFFICIENCY OF COMMERCIAL BANKS AND ECONOMIC GROWTH IN VIETNAM The purposes of this chapter include: (i) Measuring efficiency of Vietnamese commercial banks by using three methods: method of financial ratio analysis, method of stochastic frontier analysis (SFA), and method of data envelopment analysis (DEA); (ii) Analyzing the determinants of Vietnamese commercial bank efficiency under Tobit regression models; (iii) Analyzing the relationship between efficiency of commercial banks and economic growth in Vietnam through vector autoregressive model (VAR). 4.1. ANALYZING THE EFFICIENCY OF COMMERCIAL BANKS IN VIETNAM 4.1.1. Overview of sample research Data source of dissertation is taken from Fitch's Bank Scope database (2012) for the period 1992-2011, annual reports of 48 banks in 2012, 2013 and official data of the State Bank of Vietnam. Therefore, time study in this dissertation extends in the period 1992 - 2013. A number of banks in scope of this research are 48 banks, including state-owned commercial banks, joint stock banks, joint venture banks and wholly foreign-owned banks in Vietnam. 4.1.2. ANALYZING THE EFFICIENCY OF COMMERCIAL BANKS IN VIETNAM - FINANCIAL RATIO ANALYSIS APPROACH Considering the volatility of the underlying financial ratio in average over years, the study found that:  The Vietnam commercial banks maintain a ratio of customer deposits to total average assets during the study period was 60%. In particular, the highest average reached 81% in 1998 and the lowest average ratio reached only 47% in 2011.  The ratio of equity to total average assets of commercial banks in the period 1992 - 2013 reached 13%, the highest average rate reached 19% in 1999 and reached the lowest rate of 9% in 2004. 14  The ratio of outstanding loans to total assets - LTA and the ratio of total credit outstanding customer deposits - LTD averaging period 1992 – 2013 respectively reached 55% and 74% reflecting the important role of credit in the business operation of banks in Vietnam.  The ratio of bad debt reserve on total loans - RTL remained low average of 1% in the period 1999 - 2013. However, the ratio RTL tends to rise during 2010-2013.  The return on average total assets (ROA) and return on average total equity (ROE) remained at 1% and 11% in the period 1992-2013.  The size of the average total assets of commercial banks in the study period continued to increase from an average of 6,586,036.11 million VND in 1992 to an average increase of 118,941,104.94 million VND in 2013. The growth rate of total assets on average reached 14.77 % per year over the period 1992-2013 4.1.3. ANALYZING THE EFFICIENCY OF COMMERCIAL BANKS IN VIETNAM - PARAMETRIC STOCHASTIC FRONTIER ANALYSIS APPROACH  Estimation of technical efficiency To estimate the technical efficiency with stochastic frontier production function, this study designated the stochastic frontier production function based on study of Coelli et al (2005) in the form of simple Cobb-Douglas function, according to formula (2.7), analyzed in Section 2.1.2. If the bank has 2 revenue outputs and 3 inputs, the function estimating the technical efficiency of Coelli et al (2005) were estimated through the model: ln(Y) = 0 + 1ln(I1) + 2ln(I2) + 3 ln(I3)+ 4 (ln(I1)) 2 + 5 (ln(I2)) 2 + 6 (ln(I3)) 2 + 7 ln(I1)*ln(I2) + 8 ln(I1)*ln(I3) + 9 ln(I2)*ln(I3) + vi - ui (4.2) Output revenue Y includes interest income (Y1) and non-interest income (Y2). The bank inputs include staff costs representing human resources (I1), the scale of fixed assets representing the physical resources (I2), and customer deposits representing the financial resources (I3). Banks in this study used above mentioned inputs to create revenue output Y. ε = vi - ui is a noise in which vi is a random element and ui is non-negative random variables used to estimate technical inefficiency of commercial banks in Vietnam. To estimate the technical efficiency of Vietnamese commercial banks, this study used the computer program named FRONTIER 4.1 through OLS regression (Ordinary Least square - the square of the minimum) and MLE (Maximum Likelihood Estimation) as proposed by Coelli et al (2005). 15 Results of analysis of technical efficiency under the SFA in Appendix 1 show the NSB and VSB have the highest of average technical efficiency at 94% in 7 years and ANZ has the lowest of average technical efficiency at 47% with 3 years of research. During the study period, the average technical efficiency is only 55% in 1993 that is the lowest level of technical efficiency, and the highest average reaching 74% in 2013. The level of average technical efficiency of banks in the period 1992 - 2013 is very low at 66%. The main cause of low banks technical efficiency in the study sample is the scale input resources such as manpower, physical and financial resources which are too large when revenue outputs don’t correspond In a sample of 48 commercial banks over the period 1992 to 2013, 7 banks have average efficiency of 40% - 60% consisting of commercial banks: ANZ, INB, VID, EIB, MSB, SHI and HSB. The study results also showed that 35 commercial banks with technical efficiency analysis stochastic frontier function in about 60% - 80% and the remaining 6 banks have average technical efficiency of about 80% - 100% including commercial banks: GPB, OEB, TNB, SEA, VSB and NSB. The analysis results show that the group of wholly foreign-owned banks and joint venture banks has the technical efficiency at the lowest average among the sample banks in this study, when the output scale is not correspondent to the input scale maintained at large. Group of domestic commercial banks with small and medium-scale has the highest average technical when using minimum of inputs to optimize the outputs respectively. • Estimation of cost efficiency or overall efficiency Where the bank has 2 revenue outputs and 3 inputs, the model (4.3) to deploy into: ln(TC) = 0 + 1ln(Y1) + 2ln(Y2) + 3ln( + 4ln( 2 ) + ½* 5 (ln( + ½ * 6 (ln( 2 + 7ln(Y1)*ln(Y2) + 8ln(Y1)*ln( + 9ln(Y1)*ln( 2 + 10 ln(Y2)*ln( + 11ln(Y2)*ln( 2 + ½* 12*(ln( + ½* 13*(ln( 2 + 14 ln( *ln( 2 + vi - ui (4.4) TC is the total cost or resources bank used: human resources or staff cost (I1), physical resources or net fixed assets (I2), financial resources or customer deposits (I3). Y1, Y2 respectively bank outputs including interest income and non-interest income. W1, W2 and 16 W3 are the prices of the inputs I1, I2, I3 respectively the average staff costs, the average cost of fixed assets and the average cost of interest expense, = vi - ui is a noise in which vi is a random element and ui is non-negative random variables used to estimate technical inefficiency. The study used analysis software FRONTIER 4.1 effective margin to estimate the cost effectiveness or economic efficiency of commercial banks in the total study sample according to OLS and MLE method of Coelli et al (2005). This study used the computer program named FRONTIER 4.1 through OLS regression (Ordinary Least square - the square of the minimum) and MLE (Maximum Likelihood Estimation) as proposed by Coelli et al (2005) to estimate cost efficiency or overall efficiency of commercial banks in study sample. To estimate cost efficiency or overall efficiency of commercial banks needs full information on prices of inputs including W1 - staff costs or costs of human resources, the average cost of fixed assets or physical resources - W2 and the average interest expense or the cost of financial resources - W3. However, in the period 1992 - 1998 the total number of employees of commercial banks has not be published yet, this study could only collect data on the cost of the inputs W1, W2 and W3 in the period 1999 - 2013. Therefore, this study analyzed the cost efficiency and economic efficiency of 48 full commercial banks in Vietnam in the period 1999-2013 with 472 observations. The result of cost efficiency or economic efficiency analysis by software FRONTIER 4.1 (Coelli et al 2005) in the period 1992 to 2013 is presented in Appendix 2 showing that the average cost efficiency of banks in the study period only reach 49%. GPB had the lowest average cost efficiency at just 34% and HLB had the highest average cost efficiency up to 97%. During the study period 1999 - 2013, in 1999, it had the lowest average cost efficiency at only 27% and it had the highest average cost efficiency up to 68% in 2013. In the study sample, there are only 3 banks reaching average cost efficiency above 80% include LVP, MDB and HLB. There are 11 commercial banks out of 48 commercial banks in this sample had the overall economic efficiency at the level of 60-80% under the SFA in period of 1999- 2013. Other 31 banks had overall economic efficiency at 40% - 60% while VCB, GPB, and BID have average cost efficiency below 40%. 4.1.4. ANALYZING THE EFFICIENCY OF COMMERCIAL BANKS IN VIETNAM - NONPARAMETRIC FRONTIER ANALYSIS APPROACH This study used CCR model (Charnes et al 1978) and BCC model (Banker et al 1984), that were analyzed in Section 2.1.3 and 2.1 under software DEAP (Coelli et al 2005) to estimate 17 the commercial banks efficiency in Vietnam with the non-parametric analysis DEA approach. This study analyzed the technical efficiency, scale efficiency in the period 1992 - 2013, allocative efficiency and cost efficiency or overall economic efficiency in the period 1999-2013.  Estimation of technical efficiency The result of technical efficiency analysis under DEA period 1992 - 2013 in Appendix 3 showing that HLB, HSB, NSB, and TNB reached the highest average technical efficiency and ANZ had the lowest technical efficiency of the 48 banks in the research sample. The resuls of pure technical efficiency analysis in Appendix 2 showing that OEB, AGR, BID, HLB, HSB, NSB, TNB and VSB had the largest pure technical efficiency at 100%, while ANZ had the lowest pure technical efficiency at 69%. In the period 1992 - 2013, pure technical efficiency of Vietnam's commercial banks reached an average of 92% and the lowest reached 83% in 2004. Pure technical efficiency were maintained at a high average level indicating that Vietnam's commercial banks focused on increasing administrative capacity and the use of resources inputs in the right direction to reach a optimal pure technical efficiency. Scale efficiency of commercial banks in the research sample in Appendix 3 showing that the average scale efficiency of commercial banks in the period 1992-2013 is 91%. TNB, HLB, HSB, and NSB reached the highest scale efficiency and CTG has the lowest average scale efficiency at 74% in the period 1992-2013.  Estimation of cost efficiency or overall efficiency The analysis results in Appendix 4 showing tat the average cost efficiency of Vietnamese commercial banks under DEA period from 1999 to 2013 was 57%, with average allocative efficiency (AE) was 71% and average technical efficiency (TE) at 80%. This results demonstrate that Vietnam's commercial banks are using the resources with high costs to reduce overall economic efficiency of the bank. During the study period, Vietnam's commercial banks achieve the lowest economic efficiency in 2010 at 22% and the highest economic efficiency at 73% in 1999. ANZ is the bank with the lowest average cost efficiency at 20%, and NSB has the highest average efficiency at 20% of 48 banks in the sample research at 94%. During the study period, the cost efficiency of Vietnamese 18 commercial banks is low with higher technical efficiency but lower allocation efficiency. HSB and ANZ are 2 banks out of 48 Vietnamese commercial banks, with the lowest average cost efficiency at 33% and 36%, because HSB and ANZ use resource inputs with higher prices in comparison with the other bank in research sample and they are in the early stages of establishing the wholly foreign-owned banks in Vietnam. NSB reached the highest allocative efficiency at 94% when using the input costs such as human resources, the cost of fixed assets and the average cost of capital at lowest cost. 4.1.5. Conclusion on efficiency of commercial banks in Vietnam After analyzing and comparing the efficiency of the commercial banks in Vietnam through method of using financial indicators, method of stochastic frontier analysis (SFA) and the method of data envelopment analysis (DEA ) in the period 1992 - 2013, this study found that the efficiency of the commercial banks in Vietnam had inadequacy issues: Firstly, the scale of commercial banks in the research sample had significant growth up to 14.77% per year in the research period, but at the end of 2013 the total assets of commercial banks was only reached 118,941,104 million VND on average, which makes Vietnamese commercial banks underutilize advantages of scale that banks are going to be. Second, in the period 1992 – 2013, the average DTA ratio of commercial banks in the research sample only reached 60%, which makes a credit operation scale account for only 55% of total asset size of the banks in the research sample. Third, restrictions on efficiency of commercial banks are asset quality of commercial banks in the research sample, which tends to decrease during the study period when risk reserve ratio on total loans RTL increases in the period 1999-2013. Fourth, there are limits on the profitability of the Vietnamese commercial banks in the study period when return ratio on total assets ROA only reached 3% and return on equity ROE at 11 % on average. Fifth, the proportion of revenue from non-interest and non-credit services is still low in comparison with revenue from credit operations, which reduces efficiency of commercial banks. Sixth, the limit of technical efficiency of 48 banks in the research sample is low. Technical efficiency under parametric stochastic frontier analysis SFA was only 66% on average in the period 1992 - 2013. 19 Seventh, the limit of cost efficiency or economic overall efficiency of 48 commercial banks in Vietnam is low. 4.2. ANALYZING THE DETERMINANTS OF COMMERCIAL BANK EFFICIENCY IN VIETNAM 4.2.1. Analyzing the determinants of commercial bank efficiency in Vietnam From the model (2.20) and (2.21), this study conducted four regression models with the dependent variables respectively, including: technical efficiency under SFA (SFA_TE), technical efficiency under DEA (DEA_TE), cost efficiency under SFA (SFA_CE) and cost efficiency under DEA (DEA_CE). To synchronize the study period of the regression models, the model with the technical efficiency as dependent variables will have time to study for the period 1992 - 2013. The regression model with the dependent variable is cost efficiency which will have time to study from 1999 - 2013 due to data of the input prices to estimate CE can only be collected in the period 1999-2013. 4.2.2. Conclusion on the determinants of commercial bank efficiency in Vietnam Through the analysis of the determinants affecting efficiency of commercial banks in Vietnam through 4 tobit regression models with the dependent variable is the technical efficiency and cost efficiency under SFA and DEA. This study found that although there are some results not as expected, there are some notable problems.  DTA - customer deposits to total assets ratio is related inversely to the efficiency of Vietnam’s commercial banks in the 2 regression models that DTA makes sense.  The ratio of equity to total assets - ETA has unclear effects in 3 models that ETA makes sense.  LTA variable measuring scales of credit operations also impacts counterclockwise on 3 out of 4 regression models that LTA makes sense.  The ratio of outstanding loans to total deposits - LTD is only meaningful in the regression model with the dependent variable SFA_CE and reflects relationships in the same direction between LTD and Vietnamese commercial banks efficiency during the study period.  RTL - the ratio of bad debt reserve on total loans is related inversely to Vietnamese commercial banks efficiency through cost efficiency.  ROA variable is only meaningful in the model, when the dependent variable is cost efficiency under SFA (SFA_CE). This result reflects the same direction relationship between ROA and Vietnamese commercial banks efficiency in the study period. 20  The scale of commercial banks - SIZE is significant in all 4 models of this study, however, it has mixed effects in each models. 4.3. ANALYZING THE RELATIONSHIP BETWEEN EFFICIENCY OF COMMERCIAL BANKS AND ECONOMIC GROWTH IN VIETNAM 4.3.1. Analyzing the relationship between efficiency of commercial banks and economic growth in Vietnam To analyze the relationship between efficiency of commercial banks and economic growth in Vietnam , this study used models (2.25) and (2.26), analyzed in chapter 2 with the time series data through macro variables from the World Bank (2014), ADB (2014) and the average technical efficiency of Vietnam's commercial banks under SFA and DEA in the period 1992 - 2013. Analytical procedures are implemented according to Figure 2.3 in the order of steps: (i) tests for stationarity; (ii) testing for cointegration; (iii) selection of optimal lag length in VAR; (iv) regression with VAR; (v) The stability of a VAR; (vi) Granger- causality test; (vii) impulse–response function and variance decompositions.  Tests for stationarity  Tests for cointegration  Selection of optimal lag length in VAR  The stability of a VAR  Regression with VAR  Granger-causality test to VAR model The results of Granger-causality test showing that there are only 02 relationships in which null hypothesis H0 is rejected: Bank efficiency under SFA_TE - D (SFA_TE) has no impact on GDP growth - D (GDP); and GDP growth rate - D (GDP) has no impact on Bank efficiency under DEA_TE- D (DEA_TE). Analysis results showed that bank efficiency through SFA_TE variable – has an impact on GDP growth and economic growth D (GDP) and on bank efficiency through D (DEA_TE) variable.  Impulse–response function and variance decompositions 4.3.2. Conclusion on the relationship between efficiency of commercial banks and economic growth in Vietnam 21 Granger causality test showed that the relationship between bank efficiency of commercial banks and economic growth in Vietnam was unclear. The tested results was only meaningful at 2 relationships: Bank efficiency under SFA_TE - D (SFA_TE) has an impact on GDP growth - D (GDP); and GDP growth rate - D (GDP) has an impact on Bank efficiency under DEA_TE - D (DEA_TE). The relationship in the opposite direction, GDP growth - D (GDP) has an impact on bank efficiency under SFA_TE - D (SFA_TE) while bank efficiency under DEA_TE - D (DEA_TE) has impact on GDP growth - D (GDP), which does not make sense. 4.4. EVALUATING COMMERCIAL BANK EFFICIENCY AND THE RELATIONSHIP BETWEEN EFFICIENCY OF COMMERCIAL BANKS AND ECONOMIC GROWTH IN VIETNAM 4.4.1. EVALUATING COMMERCIAL BANK EFFICIENCY AND THE DETERMINANTS OF COMMERCIAL BANK EFFICIENCY IN VIETNAM Firstly, the technical efficiency of the commercial banks in Vietnam in the study period remained at low level. Second, the cost efficiency of the commercial banks in Vietnam was not high in the period 1999 - 2013. Third, Vietnam's commercial banks used primarily method of financial ratio analysis to measure the bank efficiency. Fourth, the quality of human resources is a factor affecting on the efficiency of commercial banks in Vietnam. 4.4.2. EVALUATING THE RELATIONSHIP BETWEEN EFFICIENCY OF COMMERCIAL BANKS AND ECONOMIC GROWTH IN VIETNAM Results from experimental research with qualitative and quantitative analysis of the relationship between efficiency of commercial banks and economic growth in Vietnam in the period 1992 - 2013 found that there are still limitations to be overcome in order to enhance the role of the banking system for economic growth in Vietnam. (i) The efficiency of capital supply to the economy via Vietnamese commercial banking system is low. (ii) The financial capacity and operating scale of commercial banks in Vietnam is unreasonable. (iii) The asset quality of Vietnamese commercial banking system is low. (iv) The service and payment quality of Vietnamese commercial banks is low. (v) The capacity to raise capital of commercial banks in Vietnam is limited. 22 CHAPTER 5. SOLUTIONS TO IMPROVE EFFICIENCY OF COMMERCIAL BANK CONTRIBUTING TO PROMOTE ECONOMIC GROWTH IN VIETNAM The purposes of this chapter include: (i) Analyzing the oriented development of Vietnamese commercial banks; (ii) Analyzing trends in resource use of Vietnam’s commercial banks; (iii) Proposing solutions and policy recommendations to improve efficiency of commercial bank contributing to promote economic growth in Vietnam. 5.1. ORIENTED DEVELOPMENT OF VIETNAMESE COMMERCIAL BANKS According to the master plan on “Development of Vietnam's banking sector till 2010 and orientations towards 2020” issued under Decision No.112/2006 / QD-TTg on 24/5/2006 by the Prime Minister with the common development orientation to commercial banks under the restructuring scheme of the banking system Vietnam 2020, multi-bank system is developed with modern trends, safety activities, efficient and diverse structure of ownership, and regulation models with diverse forms, and be able to competing on the basis of technology, advanced banking management consistent with practices, international standards on banking activities is to meet better demand for banking and financial services for the economy. 5.2. TRENDS IN RESOURCE USE OF VIETNAM’S COMMERCIAL BANKS The process of using the input resources of commercial banks in Vietnam remains focusing on traditional credit operations, but the cost increasing rate of using capital does not equal to the cost increasing rate of human resource, which proves that Vietnamese commercial banks have identified the quality of human resources as key competitive advantage in the modern banking business environment . 5.3. SOLUTIONS TO IMPROVE EFFICIENCY OF COMMERCIAL BANK CONTRIBUTING TO PROMOTE ECONOMIC GROWTH IN VIETNAM 5.3.1. SOLUTIONS TO IMPROVE TECHNICAL EFFICIENCY To improve bank efficiency, banks need to improve technical and cost efficiency. According to SFA model, Vietnam's commercial banks need to change the size of outputs - income from business activity corresponding to the used input resources to improve technical efficiency. First, Vietnam's commercial banks need to increase the scale of service provision and payment service to customers to increase the income scale from operating services and to meet the needs of customers better. 23 5.3.2. SOLUTIONS TO IMPROVE COST EFFICIENCY To improve bank efficiency, Vietnam’s commercial banks need to reduce input costs, such as the cost of financial resources, physical resources, wage costs or streamlining of personnel, and other expenses such as: management cost and advertising expenditure costs. Improving management capabilities and efficiency of IT systems is to help Vietnam’s commercial banks to enhance inputs using performance thereby improving technical efficiency and cost efficiency. 5.3.3. IMPROVED METHODS OF MEASURING COMMERCIAL BANK EFFICIENCY Through the analysis, the dissertation proposed that Vietnam commercial banks should adopt more frontier efficient analysis methods with two approaches (i) parametric stochastic frontier approach (SFA); and (ii) nonparametric frontier analysis – DEA approach to measure the efficiency of commercial banks in Vietnam. 5.3.4. SOLUTIONS TO IMPROVE THE QUALITY OF HUMAN RESOURCES OF VIETNAMESE COMMERCIAL BANKS 5.4. SOLUTIONS TO IMPROVE THE ROLE OF COMMERCIAL BANKING SYSTEM TO VIETNAM’S ECONOMIC GROWTH 5.4.1. Solutions to improve efficiency of capital supply to the economy via Vietnamese commercial banking system 5.4.2. Solutions to improve financial capacity and expand the operation scale of commercial banks in Vietnam 5.4.3. Solutions to improve the asset quality of Vietnamese commercial banking system 5.4.4. Solutions to expand and improve the service quality in banking operations of Vietnamese commercial banks 5.4.5. Solutions to improve the capacity of raising capital of commercial banks in Vietnam 5.5. POLICY RECOMMENDATIONS TO IMPROVE EFFICIENCY OF COMMERCIAL BANK CONTRIBUTING TO PROMOTE ECONOMIC GROWTH IN VIETNAM (i) Complete of the regulations on the safety management operation of the commercial banking system 24 (ii) Improve the performance of the National credit information center of Vietnam (CIC) (iii) Control the problems of linked loans (iv) Strengthen the supervision activities of the commercial banking system (v) Improve and develop the infrastructure of the commercial banking system (vi) Create a favorable macro environment for bank operations 5.6. LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH Although this study has achieved the set objectives; however, due to the limitations of the study period, research data and methodology, this study has main shortcomings such as:  Due to lack of information and data about the input cost for the period 1992 - 2013, the study doesn’t analyze total factor productivity (TFP) and cost efficiency of commercial banks in Vietnam in the study period to assess the better overview results.  The study period of 1992 - 2013 is not long enough to ensure reliability when analyzing the relationship between efficiency of commercial banks and economic growth in Vietnam through a VAR model.  Economic growth and the factors affecting economic growth are complex issues influenced by the economy; however, this study only focuses on analyzing the relationship between efficiency of commercial banks and economic growth in Vietnam over the period 1992-2013.  During the study period, Vietnam banking system underwent many changes when there are many newly established commercial banks, wave of mergers and consolidation or rename of bank, which leads to unbalanced research sample and makes difficulties to analyze the efficiency of each commercial bank. In the next period, if there is enough research data and full conditions of research, the study will develop in the following direction: (i) Expand the scale of banks in the sample research as well as the scope of the study; (ii) Develop and supplement the factors in the analysis model of the affecting factors to the efficiency of Vietnam’s commercial banks; (iii) Improve model to analyze the efficiency of commercial banks and economic growth in Vietnam by expanding the study period, supplementing more explanatory variables for economic growth, and accrediting the long-term relationship between the efficiency of commercial banks and economic growth in Vietnam. 25 LIST OF WORKS HAS PUBLISHED BY AUTHOR THAT IS RELEVANT TO THE DISSERTATION 1. Nguyen Minh Sang, Chung Thi Hoang Yen (2012), “Solutions to increase financial deepening to generate motivation for economic growth in Vietnam”, Financial and Monetary Market Review 362, pp. 30–34. 2. Nguyen Minh Sang (2012), “Analysis of bank efficiency of listed banks in Vietnam”, Banking Technology Review 79, pp. 23–29. 3. Nguyen Minh Sang, Nguyen Thi Hong Vinh (2013), “Research on the impact resource using on bank efficiency in Vietnam and Thailand”, CT-1301-1 – BUH’s scientific research. 4. Nguyen Minh Sang (2014), “Analyzing the determinants of bank efficiency in Vietnam”, Banking Review (02/2014) 4 , pp. 23–30. 5. Nguyen Minh Sang, Nguyen Thien Kim (2014), “The role of the business operation of banks to economic growth in Vietnam”, Economy and Forecast Review 7 (04/2014), pp. 9– 11. 6. Nguyen Minh Sang (2014), “The relationship between the business operation of banks and economic growth in Vietnam”, Journal of Development and Integration 27, pp. 17–26. THÔNG TIN TÓM TẮT VỀ NHỮNG KẾT LUẬN MỚI CỦA LUẬN ÁN TIẾN SĨ Tên luận án: Mối quan hệ giữa hiệu quả sử dụng nguồn lực của các ngân hàng thƣơng mại và tăng trƣởng kinh tế tại Việt Nam Chuyên ngành: Tài chính - Ngân hàng Mã số: 62 34 02 01 Nghiên cứu sinh: Nguyễn Minh Sáng Ngƣời hƣớng dẫn khoa học: PGS., TS. Lê Phan Thị Diệu Thảo Cơ sở đào tạo: Trƣờng Đại học Ngân hàng TP. HCM Những kết luận mới của luận án bao gồm: Thứ nhất, luận án đã chứng minh đƣợc vai trò quan trọng của hệ thống ngân hàng thƣơng mại Việt Nam trong giai đoạn 1992 – 2013 khi kênh lƣu chuyển vốn chủ yếu của nền kinh tế là thông qua kênh ngân hàng thƣơng mại tại thời điểm 31/12/2013 thì tỷ lệ tín dụng ngân hàng trên GDP đạt tới 108.15%. Thứ hai, luận án đã định lƣợng đƣợc hiệu quả sử dụng nguồn lực của các ngân hàng thƣơng mại Việt Nam giai đoạn 1992 – 2013 thông qua cả 3 phƣơng pháp: (i) phân tích các ch số tài chính; (ii) phân tích tham số cách tiếp cận biên ngẫu nhiên (SFA); (iii) phân tích phi tham số cách tiếp cận mô hình bao dữ liệu (DEA). Thứ ba, luận án đã phân tích các nhân tố tác động đến hiệu quả sử dụng nguồn lực của hệ thống ngân hàng thƣơng mại Việt Nam thông qua mô hình hồi quy tobit trong giai đoạn 1992 - 2013. Thứ tư, luận án đã định lƣợng đƣợc mối quan hệ giữa hiệu quả sử dụng nguồn lực của các ngân hàng thƣơng mại và tăng trƣởng kinh tế tại Việt Nam thông qua mô hình tự hồi quy véc tơ (VAR) trong giai đoạn 1992 - 2013. Cuối cùng, luận án đã đề xuất đƣợc các giải pháp gi p nâng cao hiệu quả sử dụng nguồn lực của các ngân hàng thƣơng mại từ đó góp phần th c đẩy tăng trƣởng kinh tế Việt Nam. Ngƣời hƣớng dẫn khoa học Nghiên cứu sinh PGS.,TS. Lê Phan Thị Diệu Thảo Nguyễn Minh Sáng A SUMMARY OF INFORMATION ON NEW CONCLUSION OF PHD DISSERTATION Title of PhD dissertation: The relationship between efficiency of commercial banks and economic growth in Vietnam Specialty: Banking and Finance Code: 62 34 02 01 PhD candidate: Nguyen Minh Sang Academic advisor: Assoc. Prof. PhD. Le Phan Thi Dieu Thao Training Institution: Banking University Hochiminh City Summary of new conclusion of dissertation includes: Firstly, the dissertation has demonstrated the important role of the Vietnamese commercial banking system in the period 1992 – 2013. The major channel of capital flows in the economy of Vietnam is the channel through commercial banking system. Second, the dissertation measured efficiency of the Vietnamese commercial banking system in the period 1992 – 2013 through the three methods: (i) analysis of financial ratios; (ii) the parametric Stochastic Frontier Analysis approach (SFA); (iii) the non-parametric Data Envelopment Analysis approach (DEA). Third, the dissertation used tobit regression model to analyse determinants affecting the efficiency of the Vietnamese commercial banking system in the period 1992 – 2013. Fourth, the dissertation was to quantify the relationship between efficiency of commercial banks and economic growth in Vietnam through vector autoregressive model (VAR) in the period 1992 - 2013. Finally, the dissertation proposes policy recommendations and solutions to improve the efficiency of commercial banking system for contributing to the economic growth of Vietnam. Academic advisor PhD candidate Assoc. Prof. Phd. Le Phan Thi Dieu Thao Nguyen Minh Sang