ANALYSIS OF BANK S PERFORMANCE AND EFFICIENCY IN INDONESIA. M. Yasser Arafat Agung D. Buchdadi Suherman

Transcription

1 ANALYSIS OF BANK S PERFORMANCE AND EFFICIENCY IN INDONESIA M. Yasser Arafat Agung D. Buchdadi Suherman Faculty of Economics, Jakarta State University Rawamangun Muka Street, East Jakarta Phone: / <aa71_2001@yahoo.com> Abstract Keywords: Bank Characteristic, ROA, ROE, NPL, Growth, Privatization This research aims to know that bank characteristics influence the performance, the efficiency level of saving deposit in banks in Indonesia. Return on Asset (ROA), Return on Equity (ROE) and net interest income are used as proxy variables in performance measurement and non performing loan (NPL) is used in both as an independent variable in performance model and as a dependent variable in efficiency measurement. Bank Characteristic variables used in this study are size, asset growth, and a dummy variable privatization. The data used are obtained by a purposive sampling that only bank which are listed during The results indicate that bank characteristic have a significant effect to performance measurement only. JEL Classification: G21 Banks INTRODUCTION A. Problem Background Package of government deregulation of the banking sector, issued in October 1988 (pakto'88) is a very controversial deregulation package in the history of the Indonesian banking. Pakto 88 makes it easy for anyone to set up new banks with only 10 billion rupiah, (Asikin, 1997) On the positive side, besides increasing the number of banks, pakto'88 spur growth in the value of banking and financial transactions. But on the negative side, competition between the owner and manager of the bank also rose and pushed to the breach of the principle of prudence, such as violation of legal lending limit, which is distributed both to the individual debtor customers and the debtor bank concerned business groups (Danusaputro, 1999). To refine pakto'88, the Indonesian government issued a package of banking policies again in March 1989, which among others includes the provisions of the health assessment results of the merger, the capital component for calculating the Capital Adequacy Ratio is more clarified, the provision of legal lending limit, and provide an opportunity for banks to invest funds in other institutions and to provide long and medium-term investment loans (Tobing, 2002). Two years later, to increase public trust in banks, the government issued a new policy which is famous for the 1991 February package. This package regarding capital adequacy ratio (CAR), restrictions on lending that is Electronic copy available at:

2 not supported by public funds (LDR), requirements for ownership and management, provision of legal lending limit and the establishment of reserves to cover risk. Slowly but sure the 1991 February package brought little influence. The banks began to consolidate meeting the capital requirements (CAR) to reach 8% by the end of 1993, the maximum LDR 110%, and lending provisions in the form of small business loans 20%. Regarding compliance issues of CAR, although some banks have been able to reach 8%, but the average CAR for many banks is below 7%. This situation by itself also raises new fears that further delay in economic activity will increasingly make it difficult for banks themselves. Table 1.1. Banking Sector Indicators (in percent) Debt to Asset Ratio Loans to Deposits Ratio Capital to Assets Ratio Cash to Deposits Ratio Ratio of Non Performing Loans to Total Loans sgn 9.2 sgn Total Bad Debt sgn 7.1 sgn sgn Net Foreign Debt to Total Liabilities Proportion of Property Sector Loans sgn sgn Return on Assets sgn sgn 0:53 1: :13 1:22 1:37 Return on Equity sgn sgn 5:19 9: :14 16:38 12:06 Operating Income sgn sgn Capital Adequacy Ratio sgn sgn :19 Statutory Reserves - Rupiah sgn sgn sgn sgn Sgn sgn 3: Currency sgn sgn sgn sgn Sgn sgn 3:28 3:09 Source: Bank Indonesia, Annual Report and IMF, International Financial Statistics With a view to strengthen the fundamentals of banking industry in Indonesia, the Bank of Indonesia has prepared a blueprint of national banking arrangements in the future as well as vision, mission and direction that will be achieved, as known as the Indonesian Banking Architecture (API). API is the basic framework of the Indonesian banking system that is comprehensive for the range of five to ten years ahead. API is needed considering the current banking challenges in the future that will be increasingly large and complex, in line with the increasingly rapid changes that occur within and outside the banking industry (Hadad, 2004). First, the capacity of bank credit growth is still low. Second, the banking structure is not optimal, causing the demand a return on equity (ROE) that is lower in state banks compared to private banks. Third, banking capabilities are still weak. Evaluation of the ability of the Indonesian banking operations showed an alarming disparity in some fields, particularly in the field of risk management, corporate governance, and core banking skills. However, in general, the operational capability of Indonesian banks is still below international best practices. This is reflected, among other things, the granting of loans is generally less heed to prudential aspects, which raised various issues that lead to the increase in NPLs. Fifth, the profitability and operational efficiency of banks are not sustainable. Viewed from profitability, Indonesian banks have shown a good performance. In industry, ROA (return on assets) Indonesian banks reached 2.4% (as of September 2003). However, profitability and operational efficiencies achieved is less Electronic copy available at:

3 sustainable in general, owing to several factors, among others, the weak structure of banking assets, in which more margins is gained from wide interest rate. Returns provision for earning assets (PPAP) is significant in supporting the Indonesian banking ROA. Without this income component, ROA of Indonesian banks will be lower than reported. Returns PPAP stemmed primarily from debt restructuring and sales of NPLs which are unlikely to be sustainable. Indonesian banking operational costs tends to be higher compared with other countries. Based on the background of the problems mentioned above researchers interested in studying about ANALYSIS OF BANK'S PERFORMANCE AND EFFICIENCY IN INDONESIA B. Problem Formulation Based on the description above, the authors formulate a research problem as follows: 1. How do Characteristics bank, company size, asset growth and privatization affect the bank performance? 2. How do Characteristics bank, company size, asset growth and privatization affect the bank efficiency? C. Research Benefit The benefits of the research are as follows: 1. The results of this research can enrich the results of previous studies in banking consolidation ever undertaken within and outside the country. 2. The results of this study will assist in providing the deliberations concerning the banking rules, including whether the Bank of Indonesia has provided protection to customers / public. 3. The results could be used as information about things that significantly affect the performance, efficiency, debt levels and the amount deposited in the banking industry.

4 THEORETICAL FRAMEWORK AND HYPOTHESIS TESTING A. Theoretical Description 1. Privatization The definition of privatization has been proposed by various authors, among others, Kay and Thompson (1975), who proposed privatization is a terminology that includes changes in the relationship between the government and the private sector. The above definition was reinforced by Beesly and Littlechild (1984) and Dunleavy (1980) in Bastian (2002) which revealed that privatization is the formation of the company or transfer of company ownership from government to private parties. While Dunleavy (1980) argued that privatization is the permanent removal of goods and services production activities undertaken by state enterprises to private companies, or in the form of non-public organizations such as NGOs. Related to the role of government in the state company, Savas (1987) gives the definition of privatization as an act of reducing the role of government or increase private participation, especially in activities concerning the ownership of assets. This definition is consistent with that proposed by Butler (1991), namely that privatization is the turn of the functions of the public sector to the private sector, either in whole or in part. Bastian (2002) asserted that the basic assumption of the management of delivery of public services to the private sector is in order to increase efficiency of resource used. Privatization is seen as the best alternative, because the market mechanism will allow the efficiency of economy. According to Makhija (2003), there are three methods of privatization, namely: 1) direct sales; 2) auctions; and 3) through a tender offer. Similarly, there are several alternatives that are available in determining the ownership structure of corporate equity in the process of privatization, among others, include: 1) direct sales to domestic purchasers; 2) direct sales to foreign buyers; 3) submitted to the shareholders equity, and 4) equity held by the government. 2. Banking Performance Cuervo and Villalonga (2000) say that the reference to company performance is the financial and operating efficiency of a company. This definition is then developed in a more operational term by Makhija (2003) using performance variables consisted of: 1) size, 2) the variation of profitability, 3) leverage, 4) market share, 5) managerial efficiency, and 6) ownership by management, community, and government. Financial ratios included in the CAMEL can be used as measure of banking performance. CAMEL consists of five criteria: capital, assets, management, earnings and liquidity (the dictionary of Banking, 1999). The following ratios contained in the CAMEL, namely: a. CAR (Capital Adequacy Ratio) CAR is the ratio that shows how much the total bank assets that contain risks involved are financed from its own capital in addition to obtaining funds from sources outside the bank. This ratio can be formulated as follows: CAR. = Bank capital x 100% Total ATMR b. Fixed Assets to Capital Ratio (ATTM). This ratio measures the ability of bank management in determining the amount of fixed assets and inventory owned by the bank concerned to capital. This ratio can be formulated as follows: ATTM = Fixed Assets and Inventory x 100% Capital

5 c. Ratio of problem productive assets (APB). This ratio indicates the ability of management to manage the bank problem productive assets to productive assets. This ratio can be formulated as follows: APB = Problem productive assets x 100% Total productive assets d. NPL (Non Performing Loan). This ratio indicates the ability of bank management in managing problem loans granted by banks. This ratio can be formulated as follows: NPL = Non performing loans x 100% Total loans e. Ratio PPAPAP (Allowance for Earning Assets Earning Assets). This ratio is defined as follows: Losses on Earning Assets = PPAP which has been formed x 100% Total earning assets f. The ratio of the fulfillment of PPAP. This ratio is defined as follows: Fulfillment of PPAP = PPAP has formed x 100% PPAP shall be created g. ROA (Return on Assets). This ratio is used to measure the ability of bank management in order to obtain the profits generated from the average total assets of the bank concerned. This ratio is defined as follows: ROA = Profit after tax x 100% Total assets h. ROE (Return on Equity). This ratio is used to measure the bank management performance in managing capital available to make a profit after tax. This ratio is defined as follows: ROE = Profit after tax x 100% Equity i. NIM (Net Interest Margin). This ratio is used to measure the ability of bank management in managing productive assets to generate net interest income. This ratio is defined as follows: NIM = Net Interest Income x 100% Productive assets j. BOPO (Ratio of Operating Expenses to Operating Income). The ratio is often called the efficiency ratio was used to measure the ability of bank management in controlling operating expenses to operating income. This ratio is defined as follows: ROA = Operating Expenses x 100% Operating Income k. LDR (Loan to Deposit Ratio). This ratio is used to assess the liquidity of a bank by dividing the total loans given by banks to third party deposits. This ratio can be formulated as follows: LDR = Total loans x 100% Total deposits

6 3. Previously studies review Havrylchyk (2005) examined the relationship of banking consolidation and efficiency in Poland. Efficiency is measured by cost efficiency, allocative efficiency and technical efficiency. The research reveals that foreign ownership influence efficiency. Subsequently it was found that foreign banks are more efficient than domestic banks in comparison with the level of 73.23% and 52.92%. This means that foreign banks be able to empower the superior technology Okazaki and Sawada (2006) revealed the relationship between banking consolidation and performance. Samples of their research are companies that have consolidated starting January 1927 until December 1932 in Japan. The number of consolidation samples is 164. The regression result shows that consolidation has a significantly positive effect on deposit growth, but significantly negative effect on change in ROA. Size, changes in branch and an urban dummy variable have significantly positive effect on almost all panel ROA regression. Laeven (2006) researched the effect of competition, diversification, and ownership on the performance of banks in some Asian countries. Bank performance is measured by operating income to total assets, and activity-adjusted operating income to total assets. Explanatory variables used are income diversity, size (total assets), total operating income, market share of deposits, deposits to liabilities, and equity to assets, growth in operating income, domestic private ownership, and foreign ownership. Laeven said that generally the process of banking consolidation is still running and the increase in foreign ownership of banks has improved bank performance and stability. 4. Hypothesis Temporary answers of this study are as follows: H1: a. bank size effects on banks performance b. Privatization effects on banks' performance H2:a. Bank size effects on efficiency b. Asset growth effects on efficiency c. Privatization effects on efficiency

7 RESEARCH RESULTS Results 1. Description of Statistics This study used data from bank financial statements listed in the Indonesia stock exchange during The year The data for 25 banks in three years i.e. 75 data is summarized in the following table: Table 2. Description of Selected Variables Data Statistics Variable N Mean Max Min Std.Dev KS Z Sig ROA ROE NICA LOAN 75 22,331, e ,547, NPL TIME 75 37,823, e ,033 57,664, SIZE GROWTH PRIVATE Source: Data processed Based on the data processing, the maximum ROA is obtained by the Bank Rakyat Indonesia in 2005 by 5% while the minimum ROA is obtained by the Bank Executives International, Tbk in 2005 by -3%. The next performance variable is ROE which has average 10, 52% with large standard deviation that is equal to 10.13%. Maximum ROE earned by Bank Rakyat Indonesia in 2005 was 37.92% while the minimum ROE earned by Bank Executives International, Tbk in 2005 amounted to %. Performance variable, net interest income to total asset has almost the same average with ROE i.e % but has a smaller standard deviation compared to ROE that is equal to 2.06%. Maximum net interest income to total asset is acquired by Bank Danamon in the year 2007 amounted to 14.99% while the minimum value obtained by Century Bank in 2005 amounted to 6.52%. NPL variable is available only 71 data because the data and the Bank the Bank NPL Executive International, Tbk in SIZE variable is obtained by logarithmic the total assets of the banks. The maximum value of SIZE Variable is obtained by Bank Mandiri Tbk in 2007 i.e. 8, 5 while the minimum value is obtained by the Bank Himpunan Saudara, Tbk in 2005 amounted to GROWTH variable has an average rating of 16.27% with a standard deviation of 15.34%. The maximum value of GROWTH is obtained by Bank Victoria International Tbk in 2007 i.e. 82%, while the minimum value is obtained by Bank Bumi Arta Tbk in 2005 amounted to % PRIVATE variable is a dummy variable indicating whether such bank is the result of the privatization process or not. It is valued 1 if the bank was privatized, and the value 0 if not. According to the data, there are four banks that are privatized namely Bank Mandiri, Tbk, Bank Negara Indonesia Tbk, Bank Bukopin, and Bank Rakyat Indonesia, Tbk. Based on data collected, LOAN and NPL variables do not have normal distribution as shown by Kolgomorov- Smirnov Z which has a significance value smaller than So these three variables are logarithmic to get the value in normal distribution. Results of normality test method Kolgomorov-Smirnov Z is presented in table 4.2.

8 Table 3. Description of Variable Data Stats on Sports Chosen After Variable N Mean Max Sun Std.Dev KS Z Sig ROA ROE NICA LOG NPL SIZE GROWTH PRIVATE Source: Data processed In addition to the normality test, stationarity test was also carried out to test if the data is stationary or not. By using e-views software data already known to be stationary as shown in Table 4 where the value of F smaller than the critical value for confidence level of 99% amounting to Discussion The Effect of Bank Characteristics on Bank Efficiency Performance measurement in this study is represented by the variables ROA, ROE, and net interest income to total asset. So the discussion in this study based on the sequence of these variables. A. Test assumptions Classic with Variable ROA as Dependent Variable See appendix... The test results show the existence of the autocorrelation phenomenon. So that the variables ROA, LOGNPL, SIZE, GROWTH use the data with first lag. Table 4.4. Indicates that the data is distributed normally which is shown by Kolgomorov Smirnov Z that have significant value above Multicollinearity test is indicated by VIF and tolerance values. All VIF values are below 10 and all the variables tolerance values are above 0.01, then we can conclude this model is free from symptoms of multicollinearity. While heteroscedasticity test is conducted using white-heteroscedasticity that is available in the program E-views. With the OBS value * R-squared and probability> 0.05, it can be indicated that the model is free from symptoms of heteroscedasticity. B. Classic assumptions test with Variable ROE as dependent variables See appendix... On autocorrelation test, the Durbin-Watson coefficient is 1.693, and then this research model is assumed free from autocorrelation phenomenon. Tests for multicollinearity indicated by VIF and tolerance values in the Table 4.5. VIF values below 10 and all the variables tolerance value above 0.01, and then we can conclude this model is free from symptoms of multicollinearity. While heteroscedasticity test is conducted using white-heteroscedasticity that is available in the program E-views. OBS value * R-squared and probability> 0.05, it can be indicated that the model is free from symptoms of heteroscedasticity. C. Classic assumptions test with variable net interest income to total asset as dependent Variable See appendix... The test results show the existence of the autocorrelation phenomenon. So that variable net interest income to total asset, LOGNPL, SIZE, GROWTH use the data with first lag. From the table 4.6. These data is distributed normally which is shown by Kolgomorov Smirnov Z values with significant value above Autocorrelation phenomenon has been

9 successfully removed considering Durbin-Watson value = in accordance with the correlation values between the encyclopedia using the formula. Tests for multicollinearity are indicated by VIF and tolerance values in Table 4.6. VIF values below 10 and all the variables tolerance value above 0.01, and then we can conclude this model is free from symptoms of multicollinearity. While heteroscedasticity test is conducted using white-heteroscedasticity that is available in the program E-views. With the value of OBS * R-squared and probability> 0.05, it can be indicated that the model is free of heteroscedasticity symptoms. D. Bank Performance Measurement Results The results of regression to the measurement of bank performance are summarized and presented in table 5: Table 5. Regression Results Bank Performance Measurement Variable ROA ROE Net interest income to total asset Constant LOGNPL (-3.814) SIZE (5.052) GROWTH 0.01 (0.08) PRIVATE (-0.714) (-4.302) (5.093) (0.999) (2.269) (0.057) (-1.403) (-1.363) (0.08) R Adj. R F N Source: Data are processed, the numbers in parentheses are t values hirung, numbers printed italic are significant value Table 5. Indicated the bank's performance using ROA and ROE with the value of F reaches 99% confidence level. NPL variable in both models has negative and significant at the 99% confidence level. Similarly, the variables SIZE have a positive effect and achieve 99% confidence level. When compared to the value of adj. R 2, it can be concluded that the bank characteristic variables in this study better in explaining the ROE because the resulting value is greater than when used ROA variable. The experimental results show significant negative NPL variable in line with research by Basu, Druck, Marston and Susmel (2004) on the measurement of bank performance in Argentina. However a PRIVATE variable that has a positive effect on performance is not observed in this study. Research by Okazaki and Sawada (2006) who found a significant positive effect of SIZE variable on bank performance is also indicated in this study. Research by Laeven (2006) who study the effect of competition, diversification, and

10 ownership on the bank performance in some Asian countries conclude a significant positive effect of the variable SIZE and GROWTH on bank performance. However, this study is only line on the SIZE variable, while the variable GROWTH found a positive effect but not statistically significant. This occurs because of the difference in which the proxy is used; this study uses asset growth, while his study used operating income growth. In this study, bank efficiency variable is represented by NPLs that shows the percentage of loans that categorized as default. Test of the classical assumption on the model is presented in the following table: See appendix... The test results show the existence of the autocorrelation phenomenon. So that variables LOGNPL, SIZE, and GROWTH use the data with first lag. Outlier test results showed that only 70 qualified data in this study. Table 4.8. Shows that these data is distributed normally with Kolgomorov Smirnov Z significant value above Tests for multicollinearity are indicated by the value of VIF and tolerance on the table 4.8. VIF values below 10 and all the variables tolerance value above 0.01, and then we can conclude this model is free from symptoms of multicollinearity. While heteroscedasticity test is conducted using whiteheteroscedasticity that is available in the program E-views. With the value of OBS * R-squared and probability> 0.05, it can be indicated that the model free from symptoms of heteroscedasticity. The results of regression on the bank efficiency measurement are summarized in table 4.9. As follows: Table 6. Measurement of Bank Efficiency Regression Results Variable Non Performing Loan (NPL) Constant SIZE (1.564) GROWTH (-1.213) PRIVATE (2.111) R Adj. R F N 70 Source: Data are processed, the numbers in parentheses are t values hirung, numbers printed italic are significant value Based Table 6. It can be concluded that the model is valid for 99% confidence level where the value of F has a significance <0.01. If we take a closer look on the table, it shows that only PRIVATE variable effects positively and significantly for 95% confidence level. While SIZE variable is positively related and GROWTH is negatively related but not statistically significant. Value adj. R 2 is small enough that the independent variables are only able to explain the change of variable NPLs amounted to 13% CONCLUSIONS AND RECOMMENDATIONS

11 A. Conclusions This study aims to examine the effect of bank characteristics on bank performance, bank efficiency, total bank loans, and total bank deposits. Based on the research and discussion on this study can be summarized as follows: Calculation of bank performance using proxy ROA and ROE can be explained by the variables SIZE and NPL of the banks. NPL variable in both models have negative effect at significant level of 99%. While the variable SIZE has a significant positive effect on the 99% confidence level. Value adj. R 2 amounted to 41.1% in the model with ROE proxy is better than a model that is proxies by ROA of 30%. Bank Characteristics selected in this study have no significant effect on bank efficiency measurement. Dummy variable PRIVATE is statistically has positive effect on the level of 95%. Value adj. R 2 generated in this study is also quite low at 13% Bank characteristic variables can be inferred to explain only variable SIZE has positive effect and significant for 99% confidence level. While PRIVATE and GROWTH variables have positive and negative effect but not statistically significant. This study concluded that bank characteristic variables are valid to explain PRIVATE and SIZE variables have positive effect and significant in this study. While the GROWTH variable has positive effect but not significant. B. Recommendations Based on the conclusions of the research results above, the researcher suggests the following: For investors in banking stocks should pay attention to the amount of NPLs held by banks considering that NPL variable is a bank efficiency proxy which effects significantly on bank performance. SIZE variable that indicates the total assets of banks should also be taken into consideration in such investments as loans and deposits are in one direction with the amount of bank assets. For subsequent studies, here are the following suggestions: a. Attempt to process the data with the panel system considering the bank characteristics can not be united for granted. However, the availability of such data is needed in this research. b. Insert other variables in performance measurement, efficiency, loans and deposits. Suppose enter a variable rate of interest on the LOAN model, or on the DEPOSITO model. c. If it is possible to extend this research not only for banks that listing only. But banks and financial institutions that are not listing on the stock market to give a comprehensive view of the Indonesian banking industry.