Studies on Banking Performance in Vietnam.

intermediation approach) and the production approach, and concluded that smaller bank branches would have scale efficiency.

Resti (1997) conducted a study on the efficiency of the Italian banking system during the period 1988 – 1992. The study used both parametric (SFA) and non-parametric (DEA) methods to estimate bank efficiency. Resti found a negative and significant correlation between the efficiency index and bad debts/total loans. At the same time, the results of measuring bank efficiency were not much different between the two methods.

Bauer et al. (1998)) compare four methods of bank efficiency estimation including SFA, DFA, TFA, and DEA to measure bank efficiency of the US banking system from 1977 – 1988. The authors propose a set of unifying conditions for measuring marginal efficiency that is most useful in general analysis and other purposes. The main purpose of the paper is to compare US bank efficiency across the four different methods. Their finding is that the parametric method is generally consistent and the two methods (parametric and nonparametric) do not simultaneously support each other.

Eisenbeis et al. (1996) used SFA method to study 254 large and small banks in the US during the period 1986 – 2001. In the study, they used three inputs (labor, funds, equity) and five outputs (investment securities, real estate loans, commercial loans, consumer loans and off-balance sheet commitments). They found that there was significant inefficiency in the banking industry ranging from 10% to 20% on average and small banks were relatively less efficient than medium and large banks.

Liang et al. (2008) used the DEA method to measure the bank's operating efficiency of the Taiwanese banking system. The research team used the CCR, BCC and Malmquist index approaches to measure bank efficiency taking into account the influence of bad debt factors. At the same time, the study also used the DFA parameter method to compare and contrast the results. The results showed that after introducing the non-performing loans ratio (NPLR) factor, the efficiency score decreased. In addition, bad debt is also the efficiency score of private banks that is initially high and then immediately lower than that of public banks.

Most studies on banking efficiency only use structural approaches (parametric and non-parametric) with the main methods for measurement such as: stochastic frontier method (SFA); Thick Frontier Approach (TFA);

Distribution Free Approach (DFA); data envelopment analysis (DEA); and Hull free factor (FDH) method.

2.2.2 Studies on banking performance in Vietnam.

Dan (2004) has built a system of expenditures to evaluate the operational efficiency of commercial banks through descriptive statistics. The study uses data from commercial banks in Da Nang city during the period 1999 - 2002. In addition, the study also uses data from other commercial banks in the Central region for comparison. The research results show that when applying different statistical methods, the results of evaluating the operational efficiency of commercial banks will be different. Each method has its own advantages and disadvantages. When analyzing the efficiency of commercial banks, it is necessary to combine different methods to exploit the advantages and limit the disadvantages of each method to have a multi-dimensional perspective on the bank that needs to be analyzed.

Hung (2008) conducted a study on the performance of 32 Vietnamese commercial banks in the period 2001 - 2005 through qualitative and quantitative methods. The study applied DEA data envelopment analysis and SFA stochastic frontier analysis to evaluate the performance of Vietnamese commercial banks. The research results show that the commercial banking system needs to improve inefficient factors that negatively affect the performance of commercial banks, only then will the Vietnamese commercial banking system become more efficient and increase its competitiveness in the post-WTO period.

HT Vu & Turnell (2010) measured cost efficiency using the stochastic frontier method SFA in a Bayesian approach of the Vietnamese commercial banking system. The study aimed at a reasonable estimate in estimating marginal costs and using the Bayesian approach. The results of the study showed that the cost efficiency of Vietnamese commercial banks is very high, 87%. There are small and insignificant differences in cost efficiency between different groups of banks classified by ownership. However, during the study period, the banking industry experienced a slight decline in cost efficiency. This is explained by the increase in management costs of diversified activities, branch network expansion and upgrading of banking technology platforms.

H. Vu & Nahm (2013) conducted a study on the factors affecting the profit efficiency of Vietnamese banks in the period 2000 - 2006. The influence is due to four groups: bank specific characteristics, ownership, transitional environment, and macroeconomic conditions. Then, the study uses Tobit regression model to evaluate the factors affecting the profit efficiency of Vietnamese banks. The research results show that the profit efficiency of Vietnamese banks is enhanced by larger size, better management ability, and hindered by low quality of assets, as well as high level of capitalisation. High GDP growth rate per capita and low inflation rate have positive impacts on bank performance.

Through the empirical evaluation of research works on bank efficiency. The author found that most of the studies on bank efficiency use two methods: data envelopment DEA and random frontier SFA. There are studies such as Resti (1997), Bauer et al ,. (1998), Pelosi (2008), Hung (2008), Vu et al ,. (2010), Ngoc Nguyen et al,. (2013) that use both methods to measure efficiency in a country. When studying bank efficiency in a country, the DEA data envelopment method is used more and more popularly, especially from 2008 to present such as Liang et al ,. (2008), Staub et al ,. (2010), Ke et al ,. (2010), Yu et al ,. (2013), Replová (2014), Zimková (2014)... At the same time, most studies use non-allocation DEA models, allocation DEA models such as CCR, BCC, SBM, cost efficiency, profit efficiency.

Table 2.1: Summary of studies on measuring bank efficiency

Author/Author group

fake

Year of study rescue	Object	Method
Fecher & Pestiau	1993	11 TCTC Area OECD countries	Using SFA assessment HQKT
Pasiouras & Kosmidou	2007	16 Greek Commercial Banks	Using DEA evaluation HQKT
Resti	1997	Italian banking system	Using SFA and DEA to measure HQKT
Hung	2008	32 Vietnamese Commercial Banks	Using DEA and SFA to measure HQKT
Bright	2015	48 Vietnamese Commercial Banks	Using DEA and SFA to measure economic efficiency and cost effectiveness

Maybe you are interested!

[Source: Author's synthesis]

2.3 EVALUATION OF RESEARCH WORKS RELATED TO FACTORS AFFECTING BANKING PERFORMANCE

Hoang & Huan (2016) based on the model suggestion of Williams (2012) studied the factors affecting the performance of the Vietnamese commercial banking system from 2005 to 2011 using the stochastic frontier performance measurement method (SFA) and the tobit regression method to analyze the factors affecting the performance of banks. The results showed that performance is affected by two main groups: Subjective factors (market share, liquidity risk, foreign investor holding ratio and bank size) and Objective factors (gross domestic income and inflation). In which, the positive factors are the foreign investor holding ratio, bank size and market share.

Sang (2015) conducted a study of 48 Vietnamese commercial banks in the period of 1992 - 2013 on banking efficiency and its relationship with economic growth through three methods: financial index set analysis; parametric analysis with the stochastic frontier approach SFA; non-parametric analysis with the data envelopment approach (DEA). The study analyzed the factors affecting banking performance.

Vietnam's commercial banking system through the Tobit model. Technical efficiency and cost efficiency are used in the Tobit regression model. The research results show that: the efficiency of capital supply to the economy through the Vietnamese commercial banking system is still low; the financial capacity and scale of operations of commercial banks are not reasonable; the asset quality of the commercial banking system is not high; the quality of service provision and payment activities at Vietnamese commercial banks is low; the ability to mobilize capital is still low. Thereby, the factors affecting the efficiency of resource use of Vietnamese commercial banks are assessed by the Tobit regression method. In which, the factors of bank size, ROA, equity/total assets and outstanding debt over total assets are significant in both methods.

In their study, Sharma, Raina and singh (2012) used panel data through stochastic frontier analysis (SFA) model to measure the sources of technical efficiency of Indian banking sector. This study has shown that the main determinants of technical efficiency are fixed assets, deposits and deposits to total liabilities.

Using a non-parametric approach to measuring efficiency by focusing on total factor productivity (TFP)1 in assessing the determinants of efficiency of Central Asian banks during 2003-2006, Djahlilor and Piesse showed that most banking institutions were efficient and the inefficiency of some Central Asian banks was due to low capital adequacy ratios, poor asset quality and low profitability (J.Piesse (2007)).

Research by Berger and Mester (1997) shows that factors such as the size of the bank and the quality of the assets of the banks have a great influence on the efficiency of using resources of commercial banks. Capital structure measures the strength of the bank's capital resources expressed through the ratio of equity capital to the total assets of the bank, which is one of the important factors affecting the performance of banks.

Sufian (2009) analyzed the factors affecting the performance of Malaysian commercial banks in the period of 1995 - 1999 around the 1997 East Asian financial crisis. The study used non-parametric analysis DEA to measure performance and tobit regression analysis to evaluate the factors affecting performance.

1 TFP (Total Factor Productivity) is an indicator that measures the productivity of both "labor" and "capital" in a specific activity or for the entire economy.

performance of Malaysian commercial banks. The dependent variable of the model is the performance of commercial banks according to DEA, the independent variables include: (i) bank size measured by the natural logarithm of total deposits, (ii) the ratio of outstanding credit to total assets,

(iii) bad debt provision ratio to total assets, (iv) total non-interest income to total assets, (v) total non-interest expenses to total assets, (vi) equity ratio to total assets.

According to Williams (2012), using the SFA method to analyze factors affecting the performance of commercial banks includes: market power (Lerner Index), concentration level (in the deposit market: concr4deposit and the loan market: concr4loan), size (banksize), market share (marketshare), credit risk (credit risk), liquidity risk (liquidity risk), gross domestic product (gdp), inflation (inflation). The variables "ownership" and "listed dummy" are included as dummy variables: ownership takes the value "1" when the bank has a foreign investor's holding ratio and takes the value "0" when there is no foreign investor's capital contribution; listed banks will receive the listed dummy value "1" and unlisted banks will receive the value "0". The inclusion of dummy variables in the model helps the author to consider the positive or negative impacts of dummy variables on the performance of the banking system.

Ayadi (2013) analyzed the factors affecting the performance of commercial banks in Tunisia in the period 1996 - 2010. The study applied the non-parametric analysis method to measure the performance of commercial banks in Tunisia. The model uses the cost efficiency variable according to DEA as the dependent variable and the independent variables are: market concentration index (HHI), the deposit ratio of each bank compared to the system, the ratio of outstanding credit to total assets, the ratio of equity to total assets, the size of banks measured by the decimal logarithm of total assets and the dummy variable on the form of ownership of commercial banks. Ayadi (2013) used the regression method with panel data through the fixed effects and random effects models and then used Hausman test to test.

Alrafadi et al. (2014) measured operational efficiency as a factor affecting the operational efficiency of the banking system in Libya during the period 2004 - 2010 through a dataset of 17 Libyan commercial banks. The study applied the non-parametric DEA method to measure the efficiency of resource use of commercial banks and tobit regression.

to analyze the factors affecting the performance of Libyan commercial banks. The research model uses technical efficiency according to DEA as the dependent variable and independent variables including: customer deposits/total assets, equity over total assets, bank size as the natural logarithm of total assets, liquidity status and capital structure.

Table 2.2: Summary of studies on factors affecting bank performance

Author/Author Group

Research Methods

Dependent variable

Independent variable

Research results

Sufian (2009)

Measuring performance using DEA and Tobit regression to assess influencing factors

Performance of commercial banks

according to DEA

Bank size (SIZE); Credit weighting ratio; liquidity status, capital structure, deposit size, bad debt ratio, profit ratio

profit

Bank size (SIZE); Credit weighting ratio; liquidity status, capital structure and deposit size are significant

statistical meaning

Williams (2012)

Measuring performance using SFA, and Tobit regression to assess influencing factors

Performance of commercial banks

according to SFA

Market power, concentration, bank size, market share, credit risk, liquidity, GDP, Inflation and volatility

fake

Market power, bank size, market share, credit risk, liquidity, GDP are significant

in the model.

Adadi (2013)

Cost-effectiveness by DEA and Fixed Effects and Random Effects

Cost effectiveness according to DEA

Market concentration index (HHI), bank-to-system deposit ratio, loan-to-asset ratio, equity-to-asset ratio, bank size measured by the decimal logarithm of total assets, and ownership type dummy variables

commercial banks

Market concentration index (HHI), the ratio of each bank's deposits to the system, the ratio of outstanding credit to total assets, the ratio of equity to total assets, the size of banks measured by the decimal logarithm of total assets

statistical significance

Alrafadil et al

(2014)

Using DEA to measure efficiency and Tobit regression to analyze influencing factors

Technical efficiency of commercial banks

according to DEA

Customer deposits/total assets, equity to total assets, bank size as the natural logarithm of total assets, liquidity and capital structure

Customer deposits/total assets, equity to total assets, bank size is the natural logarithm of total assets, liquidity status and capital structure are meaningful

statistical meaning

[Source: Author's synthesis]

Comment