The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 349 The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking Murharsito 1 Abstract This paper attempts to examine the effect of ownership and global crisis to income diversification of Indonesian Banks during period of 2005 to 2012. The income diversification is classified as the taxonomy of De Young and Rice (2004), therefore income diversification is divided in to non-traditional stakeholder non-interest income and traditional and fee for service non-interest income. Apart of regress the whole bank sample, analysis is conducted to each of ownership types as well. Using pooled effect panel data,the study result suggests that ownership doesn t affects income diversification of Indonesian banks both to the non-traditional stakeholder and traditional and fee for service non-interest income. However, the direction effect of public ownership is negative in both non-traditional incomes in contrast direction of foreign ownership is positive. Then, in each types of ownership, capitalization affect significantly in positive direction to non-traditional stakeholder non-interest income. In terms of traditional and fee for service, in state banks, size has a positive and significant effect. In foreign banks credit risk affect in positive direction, but it affect oppositely in private banks. In addition profitability also affects positively to traditional and fee for service non-interest income in private banks. The effect of global crisis has different direction to each non-interest income, it encourages banks to generate traditional and fee for service non-interest income. And although it is not significant, it has negative effect to non-traditional stakeholder non-interest income. Keywords: Income diversification, ownership, global crisis, Indonesian banks JEL Classification: G01, G15, G21 1 Lecturer at Nahdlatul Ulama Islamic University (UNISNU) Jepara, email: murharsito@unisnu.ac.id.
350 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 I. INTRODUCTION Indonesian banking sector is about to be controlled by more diverse owners recently, the domination of state banks is gradually decreased, on the other hand private and foreign banks increased their role participating in Indonesian banking sector. According to the total asset data, at the end of 2002, state banks comprise 46. 44%, private banks (both foreign exchange and non-foreign exchange private banks) comprise 37. 30%, and foreign and joint venture banks comprise 11. 02% of banking sector, ten years later at the end of 2012 those composition shift significantly, state banks comprise 36. 02%, private banks (both foreign exchange and non-foreign exchange private banks) comprise 43. 19%, and foreign and joint venture banks comprise 12. 19% of banking sector (Bank Indonesia, 2004, 2012). The 1998 banking crisis that hit this sector seriously was then followed by privatization programs which invited foreign investors to participate in this sector, Sato (2005) stated that the combination of bank branches, foreign joint banks and foreign owned private banks accounting for 31 % of bank asset as a whole, up from 9% before the crisis occurred. This condition has intensified the competition among banks and forces them to maximize their efforts in order to increase their market share. Income diversification is one strategy that could be applied by banks to improve their profitability. Banks could extent their business not just in their traditional activities such as loan making but also create and develop non-traditional activities, such as service commission, trading revenue, insurance fee etc. In Indonesian banking, the role of income diversification has been realized as one of the important corporate income sources, study by Sufian and Habibullah (2010) about Indonesian banks profitability over 15 years period concluded that income diversification plays a key role and has a positive relationship with the banks profitability. Further issue that is interesting to be discussed is whether the ownership factor has the effect to the income diversification to the non-traditional sectors by Indonesian banks. As the owners of banks is more diversified among state, private and foreign owner, and each of them has significant portion in Indonesian banks market share. The ownership difference could lead to the performance difference, many studies have revealed the effect of the relation between firms ownership to its performance, the result of those researches mostly highlight the weak performance of state owned enterprises compared with the others (Hart et al., 1997; Shleifer and Vishny, 1997; Dewenter and Malatesta, 2001). Specifically, the effect of the ownership to banks performance also reveals similar result. Cornett et al (2010) concluded that state owned banks generate less profit, not well capitalized and more risky in the term of credit than private banks prior to 2001. Beck et al (2004) concluded that higher share of state owned banks make the effect of bank concentration acerbated, while foreign banks presence prevent the effect of bank concentration on credit obstacles. Bonin et al (2005) concluded that in terms of efficiency, foreign owned banks are the most efficient while state owned banks in the opposite place become the least efficient. If this theory also materializes in terms of bank achievement on non-traditional income activities, there will be a difference on the non-interest income among state, private and foreign banks. Recent study on Indian banking industry by Pennathur
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 351 et al (2012) indicate that different type of ownership will cause different impact on income diversification, further public sector banks earn the lowest fee income while foreign banks can generate higher fee income. On the other hand, the occurrence of financial crisis which hit world economies more frequent currently could also affect banks performance, particularly in generating non-interest income. Recently, there are two major crises period which has severe impact to the world s economics; they are 2008-2009 Subprime mortgage crises and 2012 Eurozone crisis, different with the Asian Financial crisis which rooted from Southeast Asia, those crises originated from advanced economies; however its impact also pronounced in this region. Related to the prior Asian Financial crisis Sufian and Habibullah (2010) concluded that this crisis negatively effect on bank s profitability in Indonesia, further the profitability of Indonesian s banks is relatively higher in the tranquil periods than in the crisis time. The effect of crisis to the bank s performance could be different according to their owner, Cetoreli and Goldberg (2011) notes that foreign and local banks play different role in transmitting the shocks of the crisis, foreign banks abroad channel the shocks by reduce their cross border lending, foreign banks affiliations also decrease its local lending, and local banks follows reducing their loan because of interbank lending decline. The specific character of crisis effect on each different ownership status of banks could also affect their non-interest income generating performance. The objectives of the study are first to analyze the ownership and the crisis impact to non-interest income, the non-interest income is classified and defined by Pennathur et al (2012) that group non-interest income into brokerage income and other non-interest income. We use different approach and based our classification on the model which is developed by De Young and Rice (2004) and also be used by De Young and Torna (2012) which categories non-interest income based on its production and risk return characteristics that could affect to insolvency and financial distress probability, further categorizing non-interest income into three kinds, namely non-traditional stakeholder activities, non-traditional fee for service and traditional fee. Non-traditional stakeholder activities are activities that require banks to hold risky asset i. e., investment banking, venture capital and proprietary trading. Non-traditional fee for service activities are activities that don t require banks to hold risky asset i. e., securities brokerage and insurance sales. And then traditional fee activities are activities permitted prior to deregulation i. e., fiduciary services and depositor services. However it s difficult to classify the data that we have to those groups, so we still insist the category of non-traditional stakeholder activities, but uniting the non-traditional fee for service and traditional fee. This uniting makes sense because they don t have substantial difference in the risk return characteristic; those non-interest incomes don t require banks to hold risky asset in the generating process. After analyze the effect of ownership to non-interest income, we also will examine what factors could influence the non-interest income generating for each type of ownership. Specifically we will examine the effect of size, profitability, credit risk, lending business and capitalization to non-interest income generating in public, foreign and private banks. Second, we will investigate the effect
352 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 of recent crisis, which is originated from advanced economies. Although the deterioration effect of this crisis is not as big as Asian Financial crisis for Indonesian economics, but Indonesia banking sector must be affected significantly by this crisis. We attempt to assess the effect of this crisis to the Indonesian banks non-interest income generating activities. To support the assessment we also analyze the effect of crisis to non-interest income generating in each type of ownership of Indonesian banks. The next section of this paper outlines the theory and related literatures.section three present the data and methodology applied on this paper, while section four discuss the result and its analysis. Section five provide conclusion and will close the presentation of this paper. II. THEORY 2.1. Selected Literature on Bank Ownership and Performance The bank ownership cannot be separated from privatization issue currently, regarding to the impact of ownership, privatization and bank performance, Bonin et al (2005) investigated effect of bank privatization in European transition countries by computing income, balance sheet characteristic and efficiency. In terms of efficiency, they found that foreign owned banks are the most efficient on the other hand state banks are the least efficient. In fee for business service, found that local banks perform better because they have local advantage than foreign banks. Next, method and timing of privatization effect to the banks performance, voucher privatization doesn t result in efficiency improvement, while early stage privatization resulted in better performance than the later one. The ownership also related with the financing obstacles that faces by lender, Beck et al (2004) found that ownership structure of banking system coincide with level of economic development, regulatory and county s characteristics affect relationship between financing obstacles and bank concentration. The relationship between financing obstacles and bank concentration dampens by the present of foreign banks along with high level of institutional development and efficient credit registry. On the other hand, greater restriction on bank activities, high intervention of government in banking system and higher share of banks in government ownership make the relationship of financing obstacles and bank concentration acerbated. The comparison performance of public and private banks also presented by Iannotta et al (2007), after control bank characteristic, country and time effect found that mutual and government owned banks produce lower profitability than private banks even though they have lower cost. Government owned banks also have poorer quality of loan and higher insolvency risk than the other types of bank on the other hand mutual banks have better quality of loan and lower asset risk than other types of bank. Further, ownership concentration doesn t affect profitability of banks; a higher concentration of ownership is resulted in better quality of loan and also asset and insolvency risk in lower level.
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 353 The more specific analysis of the relationship between ownership and income diversification is presented by Pennathur et al (2012) that examine the effect of ownership on income diversification and risk to Indian banks over the period 2001 to 2009. They found that noninterest income activities are affected by the ownership significantly. Furthermore public sector banks earn less fee income compared to private banks; on the other hand foreign banks generate higher fee income. Moreover public sector banks with higher share ownership of government tend to pursue non-interest income less intensively. On the relation with risk, fee based income reduces risk for instance default risk for public sector banks significantly. 2.2. Selected Literature on Bank s Income Diversification The relationship between size and technological advances to non-interest income was presented by De Young and Rice (2004) who found that non-interest income is generated relatively larger by large banks while well managed banks less depend on noninterest income. Moreover some technological advances such as mutual fund and cashless transaction are associated with noninterest income increases, while the other kind of technological advance such as loan securitization are linked with the noninterest income reduction. Further marginal increases of non-interest income are resulted in higher profit. The relation of risk return trade off are different in two periods, in the first part the risk-return trade off were improved by the expansion to non-interest income, but it worsened in the last part of observation period. The effect of bank s decision to either focuses or diversifies their activities to its return and risk is examined by Acharya et al (2006). Using data from 105 Italian firms over the sample period from 1993 to 1999, they found that there seems to be diseconomies of bank diversification particularly when it expands into industries that have higher degree of competition and without having prior experimentation in that area. Those diseconomies arise in the form of deteriorating credit quality of loan portfolios with a fall in bank returns. The impact of bank activity and short term funding strategies for bank risk and return is analyzed by Demirguc-Kunt and Huizinga (2010), from 1. 334 banks in 101 countries, they concluded that the diversification to the non-interest income activities increases the rate of return of asset and it could offer risk diversification at very low level. Non-deposit funding in contrast lowers the rate of return of assets, but it offers risk reduction at low level. Further, banking strategies that rely on generating non-interest income or attracting non-deposit funding are very risky. The research result from emerging countries observation also shows that non-interest activities increased bank s risk. One of them proposed by Berger et al (2010) that examine the effect of product and geographical focus and diversification strategies on 88 Chinese banks during 1996-2006; they found dis-economies of diversification in the loan, deposit, asset and geographic dimension among those banks. Oppositely, more focused banks can attain higher profit and lower cost, as well as higher profit efficiency and higher cost efficiency.
354 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 2.3. Selected Literature on Income Diversification and Crisis Whether income from non-traditional banking activities has a contribution to the failures commercial banks during financial crisis is investigated by De Young and Tokna (2013). They found that asset based non-traditional income increases the probability of bank failure especially for the banks that already suffered of financial crisis. The fee based non-traditional income has contrary effect, it reduces the probability that banks failed during crisis. This result confirmed that there is fundamental different production and characteristic of asset based non-traditional income and fee based non-traditional income. The comparison of ownership effect on bank performance during crisis is explored by Cornett et al (2010), who found that from period of 1997 to 2000 in state banks the deterioration in cash flow returns, credit quality and capital core are greater than private banks. And then prior to 2001, in countries which government involvement and political corruption in banking is greater, state owned banks will have more inferior than private ones, some indicator of this inferiority performance is less profitability, less well capitalized and greater credit risk. However this gap can be closed by state banks in terms of cash flow returns, core capital and nonperforming loan in period of 2001 to 2004 or in the post-crisis period. Relative similar study also conducted by Vallascas et al (2012) who examine the theory that states diversification will improve the resilience of banks during distress period. They found that banks which diversify their income in narrow level before the crisis experienced performance declining during financial crisis. Oppositely, broad diversification activities such as in lending and capital market activities prevent performance declining during the crisis. Finally, the Indonesian banks profitability determinants during Asian financial crisis occurred is investigated by Sufian and Habibullah (2010). Form the data span from 1990 to 2005 they found that income diversification coincide with capitalization are positively associated to bank profitability, while overhead cost and size negatively impacted. Indonesian banks seem to have been skimping their resources especially during crisis and pre-crisis period. Moreover the Asian financial crisis exerts negative impact on Indonesian banks profitability, while in the tranquil period Indonesian bank were more profitable. III. METHODOLOGY 3.1. Data The data in this paper are obtained from Bankscope, the banking database that contain thousands of bank data around the world. We choose commercial banks which are owned by state, private national and private foreign banks as sample of this research. The definition of state banks here is the bank which is owned by the national state, not included bank which is owned by regional government. The definition of foreign bank is the bank which is owned
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 355 majority by foreign shareholder, so it would be both subsidiaries of foreign bank in Indonesia (for instance ANZ Indonesia Bank and HSBC Bank), and private Indonesian banks which is owned by foreigner (for instance CIMB Niaga and Danamon Bank). We select it by see the owner of that bank in the bankscope output sheet, and set the bankscope searching tool to find Indonesian banks which minimum 50. 01 % of its shares are belong to the foreign shareholder. The final sample comprises of 50 commercial banks, and the periods of the data are from 2005 to 2012. Bankscope database provides the income statement of banks in certain format, dividing income in to interest income and noninterest income, for noninterest income there are six categories, they are net gain (losses) on trading and derivatives, net gain (losses) on other securities, net gain (losses) on assets at fair value through income statement, net insurance income, net fee and commissions and other operating income. We try to categories those incomes to be congruent with the taxonomy of non interest income which is developed by De Young and Rice (2004) which classifies non-interest income on to non-traditional stakeholder activities, non-traditional fee for service and traditional fee. Non-traditional stakeholder activities are activities that require banks to hold risky asset i. e., investment banking, venture capital and proprietary trading. Non-traditional fee for service activities are activities that don t require banks to hold risky asset i. e., securities brokerage and insurance sales. And then traditional fee activities are activities permitted prior to deregulation i. e., fiduciary services and depositor services. We match the categories from bankscope with the categorization of De Young and Rice (2004), there is no problem in defining the non-traditional stakeholder activities, in bankscope data they are net gain (losses) on trading and derivatives and net gain (losses) on other securities, those two categories have implication on bank to hold the risky asset, for the future this kind of income will be called non-traditional stakeholder. Then we continue to next categorization, the non-traditional fee for service and traditional fee activities, when we see the income data on bankscope, net insurance income and net fee and commission, one of them net insurance income could be fit with the non-traditional fee for service, but the other, net fee and commission could be on both non-traditional fee for service and traditional fee activities. To solve this problem, we unite these categories fee for service and traditional fee, we consider that apart for the difference between them, they have basic similarity, those noninterest incomes don t require banks to hold risky asset to be generating. As the implication we also unite the net insurance income with the net fee and commission in the bankscope, for the future this income will be called traditional and fee for service. 3.2. Model This paper attempts to investigate the effect of ownership and global crisis to the income diversification in Indonesian Banking. Further non-interest income will be divided on to two variables, they are non-traditional stakeholder and traditional and fee for service. To answer the research objectives, we use the model which has been developed by Pennathur et al (2012). For
356 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 the first objective to investigate the effect of ownership to income diversification, we deploy this model: (1) We employ two measures of non-interest income, first the ratio of non-traditional stakeholder income to total non-interest operating income and second traditional and fee for service income to total non-interest operating income. The ownership as independent variable measured by employ dummy variable for public bank and foreign bank as well. Control variables in this model are based mostly from Pennatur et al (2012), to capture the effect of size, we employ the log of total asset (LnTA), ROE is to capture the profitability difference, and the quality of loan is capture by Loan Loss Provision (LLP) controlled by total loan, the growth of business is captured by loan to Total Asset (L/TA) and to capture the capitalization effect, we employ capital to total asset (Cap/TA). Then we add the lagged of dependent variable (noninterest income at t-1) as the control variable, this variable will capture the effect of the last year non-interest income to this year non-interest income. The other objective is to measure the effect of global crisis to the income diversification on Indonesian Banking. To capture the effect of global crisis we set a dummy on the year when global crises occurred. Our sample dataset period is 2005-2012, so we inventory the crises years upon that period. In determining whether that year could be classified as the crisis year, we consider based on the existing literature review. First, we conclude that the year of 2008 and 2009 is a crisis year, based on the systemic banking crisis database which has been updated by Laeven and Valencia (2012), they provide the database that include all systemic banking, currency and sovereign debt crises that span from 1970 to 2011. After reviewing that database, we conclude that the year of 2008 and 2009 can be classified the crisis year, because in those year occurs systemic banking crisis in the extensive scale in many countries, for instance in 2008 there were banking crises in Austria, Belgium, Iceland, Latvia, Luxembourg, Netherland, United Kingdom and United States, In 2009 systemic banking crisis occurred in Denmark, Germany, Greece, Ireland, Mongolia and Ukraine. The other crisis year that we noted is in the year of 2012, we draw this conclusion based on Aizenman et al (2013) that stated in the year of 2012 the Euro zone sovereign debt crisis posed ad become the single biggest downside risk to the global outlook. We also see the record of world growth rate from the databank World Bank
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 357 database in our sampling period and find that the lowest growth rate record are occurred in the year 2008, 2009 and 2012. Then, we would like to investigate the determinant of non-interest income and the effect of crisis to each ownership type, so we will have panel data regression for state bank, foreign bank and private bank respectively. This specific analysis will provide the more comprehensive understanding and enrich the result of the previous analysis. Still the same with previous model, the dependent variables for this analysis are non-traditional stakeholder and traditional and fee for service. The model then written as the equation bellow: (2) IV. RESULT AND ANALYSIS 4.1. Ownership and Income Diversification The analysis begins with examining the impact of ownership to income diversification, further it would examine whether there are differences in non-interest income both in non-traditional stakeholder income and in traditional and fee for service income. Based on the descriptive statistic, the mean value of non-traditional stakeholder income to total non-interest operating income for all samples is 16. 08 percent, which is lower than the traditional and fee for service income to total non-interest operating income which is 46. 25 percent. This result reveals that traditional and fee for service non-interest income is much common non-interest income for Indonesian Banks. The complete result of descriptive statistic and correlation matrix is provided in Table 1. We continue in examining the effect of ownership on the income diversification. In estimating the effect of ownership and global crisis to income diversification for all banks we use pooled effect panel data, we choose that method with regard to two considerations; first, because the ownership ofbanks doesn t change over time (dummy of public and foreign banks will be the same every year) so it can t be estimated by fixed effect, second, random effect cannot be used either as can be seen on the value of hausman test (the p value is significant for all of models). In addition, we augment the year dummy to the model. The usage of pooled effect panel data to calculate income diversification also be used by Pennathur et al (2012). The result of the estimation is provided in Table 2.
358 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 Table 1. Descriptive Statistics and Correlation between Variables MEAN STD NS TFS LNAS SET ROE LLP LOAN ASSET LOAN CAP_TA PUBLIK FOREIG CRISIS N NS 0. 160811 0. 258980 1. 000000 TFS 0. 462475 0. 321136-0. 390203* 1. 000000 LNASSET 9. 942848 2. 231431-0. 114244* 0. 268656* 1. 000000 ROE 0. 143538 0. 342630-0. 004310 0. 154541* 0. 289772* 1. 000000 LLPLOAN 0. 012300 0. 110468 0. 008166 0. 119855* 0. 015212 0. 787833* 1. 000000 LOANASSET 0. 589038 0. 136115-0. 069596* 0. 095374* -0. 152637* -0. 079092* 0. 033341 1. 000000 CAPASSET 14. 40100 9. 973828 0. 160937* -0. 243321* -0. 358119* -0. 377536* -0. 381904* -0. 010389 1. 000000 PUBLIC 0. 120000 0. 325369-0. 068318* -0. 023052 0. 221172* -0. 025561-0. 011769-0. 035176 0. 135906* 1. 000000 FOREIGN 0. 480000 0. 500226 0. 058031-0. 036343-0. 208244* -0. 032200-0. 099907* 0. 255132* 0. 101896* -0. 504030* 1. 000000 CRISIS 0. 375000 0. 484729-0. 002526 0. 121198* 0. 008474 0. 034032 0. 065928* 0. 229213* -0. 017192-0. 044266 0. 013042 1. 000000 Note: * Significant at 5%. NS: Non-traditional Stakeholder; TFS: Traditional and Fee for Service; LNASSET: Log of Total assets; ROE = Return on Equity; LLPLOAN = loanloss provisions/total loans; LOANASSET: Loans to Total Asset; CapAsset: Capital to Total Asset; Public: Public Sector Banks: Foreign: Foreign Banks Independent variables in this model seem to be able to explain traditional and fee for service better than non-traditional stakeholder. It can be seen from the value of R 2, where its value is bigger in dependent variable of traditional and fee for service (38. 57% and 39. 72% in model 2 and model 4) than in dependent variable of non-traditional stakeholder (20. 58% and 21. 70% in model 1 and model 3). The lagged of dependent variable performs effectively capture the effect of last year non-interest income, for all models the lagged of dependent variable is significantly affect dependent variable at the level of 1 %. This variable also prevent the presence of first order serial correlation, it is proved by the value of Durbin-Watson statistic which is near two, indicating no first order serial correlation. Regarding to the possibility of multicollinearity between public and foreign ownership, we test its presence by running models one without public ownership variable and the other without foreign ownership variable interchangeably and finding that the value of R 2 doesn t change, it indicates that there is no multicollinearity between public and foreign ownership. Model 1 and model 2 in Table 2 reveal that ownership has no explanation power as the determinant of the non-traditional stakeholder non-interest income, the p value of public and foreign ownership in both non-traditional stakeholder and traditional and fee for service is not significant. However, the direction effect of public ownership is negative in both non-traditional incomes; in contrast the direction of foreign ownership is positive. Even though not significant, but the direction of public and foreign ownership to income diversification confirms previous arguments that public banks performance is poorer, and foreign banks has a better performance (Bonin et al, 2005; Iannotta et al, 2007; Pennathur et al, 2012).
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 359 For control variables, the variable that affect non-traditional stakeholder and traditional and fee for service is different. For non-traditional stakeholder, capitalization (capital to total asset) significantly affect non-traditional stakeholder with positive direction. It means that more well capitalized the bank more generating non-traditional stakeholder non-interest income. This positive effect of capitalization is in line with argument of Blum (1999) about the leverage effect of capital rules. It stated that banks will maximize the value of its equity by invest it in the profitable business although it is more risky. On the other hand profitability (ROE) has a positive and significant effect to traditional and fee for service non-interest income. From the coefficient value, reflected that one percent increase of ROE leads to 0. 13% increase in traditional and fee for service non-interest income. This result is contrary to De Young and Rice (2004) which found that the profitability affect negatively to the non-interest income. Table 3 provides the estimation result of some determinants of non-interest income for each of ownership type. We analyze this estimation use pooled effect panel data, we decide it based on two considerations, first, because the number of observation is small (the smallest number of observation is 30 and the largest is 145) so it is impossible to use fixed effect because it can raise small sample bias (Nickell, 1981), second we consider to use random effect panel data, but it cannot be done because of the value of Hausman test (p value is significant in all calculation). Table 2 Effect of Ownership and Global Crisis to Income Diversification Model 1 Model 2 Model 3 Model 4 Independent Variables Nontraditional Stakeholder Traditional and Fee for Service Nontraditional Stakeholder Traditional and Fee for Service NS (lagged) TFS (lagged) Public Foreign LnAsset ROE LLPLoan LoanAsset 0. 310934*** - -0. 041323 (0. 4565) 0. 017752 (0. 6717) 0. 002613 (0. 8141) 0. 031363 (0. 7053) 0. 058248 (0. 8171) -0. 094784 (0. 5239) - 0. 569372*** -0. 057759 (0. 2739) 0. 010146 (0. 782) 0. 008279 (0. 3832) 0. 13026** (0. 0335) -0. 140973 (0. 4659) 0. 076194 (0. 5748) 0. 304485*** - -0. 042053 (0. 4468) 0. 016964 (0. 6843) 0. 002251 (0. 8389) 0. 033038 (0. 6892) 0. 072418 (0. 773) -0. 078269 (0. 5981) - 0. 56732*** -0. 054802 (0. 2956) 0. 013963 (0. 7016) 0. 008622 (0. 3603) 0. 125124** (0. 0397) -0. 149157 (0. 437) 0. 044484 (0. 7427)
360 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 Model 1 Model 2 Model 3 Model 4 Independent Variables Nontraditional Stakeholder Traditional and Fee for Service Nontraditional Stakeholder Traditional and Fee for Service NS CapAsset (lagged) Year Crisis Constant R squared F statistic Hausman Durbin Watson statistic N 0. 310934*** 006844*** 0. 000625-0. 304485*** 006903*** 0. 000603- (0. 0058) 0. 003315 (0. 7415) - 0. 053364 (0. 7284) 0. 2058 4. 662936*** (0. 000017) 62. 99388*** 1. 985027 172 (0. 7762) -0. 009198 (0. 3134) Note: NS= Nontraditional Stakeholder; TFS= Traditional and Fee for Service; Public= Public Sector Banks Dummy; Foreign= Foreign Banks Dummy; LNASSET= Log of Total assets; ROE= Return on Equity; LLPLOAN=loan loss provisions/total loans; LOANASSET= Loans to Total Asset; Cap_ta= Capital TO Total Asset; Year= year Dummy; Crisis= Crisis Year Dummy. Absolute values of t-statistics are in parentheses, *, **, and *** indicate significance at 10%, 5%, and 1% levels respectively. - 0. 080375 (0. 5623) 0. 3857 18. 62367*** (0. 000000) 52. 926482*** 2. 296834 277 (0. 0053) 0. 006824 (0. 5066) -0. 053927 (0. 1298) 0. 053713 (0. 7256) 0. 2170 4. 462642*** (0. 000015) 72. 40831*** 1. 9821 172 (0. 7824) -0. 013804 (0. 1374) 0. 074822** (0. 0248) 0. 084898 (0. 5375) 0. 3972 17. 52763*** (0. 000000) 51. 713238*** 2. 290361 277 We then continue the estimation of each type of ownership. For public bank capitalization (capital to total asset) positively and significantly affect non-traditional stakeholder non-interest income. This is in line with the estimation of non-traditional stakeholder for all of bank s ownership categories. In terms of traditional and fee for service, in state banks, size (ln of asset) has a positive and significant effect. It means that the bigger the state bank, the greater the earnings of traditional and fee for service. This finding is in line with Hidayat et al (2012) who also state that the income diversification as the result of deregulation in Indonesia is done by big banks, because it has significant role in Indonesian banking industry. For foreign banks, we found that none of the independent variable has significant effect to non-traditional stakeholder non-interest income unless lagged of dependent variable. However, for traditional and fee for service non-interest income credit risk (loan loss provision to loan ratio) has a positive and significant effect. The reason which could explain this result is possibly banks attempt to seek another income because the main income from lending activities faces quite significant risk. This finding is different with Pennathur et al (2012) which found a positive and significant effect of credit risk effect on fee based income not in foreign but private banks. For private banks, similar to foreign banks, none of the independent variable has significant effect to non-traditional stakeholder non-interest income unless lagged of dependent variable. In terms of traditional and fee for service there are two variables which has significant effect.
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 361 First, profitability (ROE) which has positive value, this result in line with estimation of traditional and fee for service for all bank s ownership categories. From the coefficient value can be interpreted that one percent increase of ROE leads to 0. 25% increase of traditional and fee for service non-interest income. Second, credit risk (loan loss provision to loan ratio) which has negative value, and also the coefficient value is quite big, one percent increase in LLP to loan ratio will decrease traditional and fee for service 0. 54%. This result may indicate that main interest income activities and traditional and fee for service non-interest income activities run together in private banks. Another variables that related to main interest income activities like lending business (loan to total asset) also has positive value although not significant. This finding provides a more comprehensive understanding, that the role of credit risk to traditional and fee for service non-interest income in foreign and private banks is different. It has positive effect in foreign banks but affect oppositely in private banks. Table 3 Determinants of Income Diversification for Each Ownership Type Independent Variables Public Bank Foreign Bank Private Bank Panel A dependent variable is nontraditional stakeholder NS (lagged) Ln Asset ROE LLP Loan Loan Asset Cap Asset Year Constant R squared F statistic Hausman Durbin Watson statistic N -0.158753 (0.2664) -0.137273 (0.3158) 1.490357 (0.1685) 7.257302 (0.175) -1.415572 (0.2115) 0.019745** (0.0143) 0.048576 (0.3746) 1.724989 (0.3507) 0.4656 2.737696** (0.033277) - 2.31552 30 0. 54057*** -0. 002021 (0. 8815) 0. 020209 (0. 8241) -0. 550451 (0. 4016) -0. 281913 (0. 1332) -0. 000755 (0. 823) 0. 005078 (0. 6771) 0. 294897 (0. 1387) 0. 3549 6. 443788*** (0. 000005) 32. 966402*** 1. 949832 90 0. 558315*** -0. 000785 (0. 9573) 0. 248891*** (0. 0012) -0. 539426** (0. 0268) 0. 257583 (0. 2076) -0. 003174 (0. 5317) -0. 022576 (0. 1652) 0. 17513 (0. 4207) 0. 4729 10. 63859*** (0. 000000) 25. 370487*** (0. 0007) 2. 089733 91
362 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 Independent Variables Public Bank Foreign Bank Private Bank Panel A B dependent variable is traditional fee for service TFS (lagged) Ln Asset ROE LLP Loan Loan Asset Cap Asset Year Constant R squared F statistic Hausman Durbin Watson statistic N 0.135024 (0.485) 0.235416** (0.0116) -1.458078 (0.3032) -0.060938 (0.9933) 0.981914 (0.3427) 0.006033 (0.471) -0.065639 (0.1887) -2.470474 (0.0379) 0.411933 3.302293*** (0.00897) Note:NS: Non-traditional Stakeholder; TFS: Traditional and Fee for Service LNASSET: Log of Total assets; ROE: Return on Equity; LLPLOAN:loan loss provisions/total loans; LOANASSET: Loans to Total Asset; Cap_ta: Capital to Total Asset; Year: Year Dummy. Absolute values of t-statistics are in parentheses, *, **, and *** indicate significance at 10%, 5%, and 1% levels respectively. - 2.189925 41 0. 659874*** 0. 008416 (0. 4312) -0. 036124 (0. 6463) 1. 061116** (0. 0373) -0. 095103 (0. 5023) 0. 000955 (0. 6567) -0. 003138 (0. 7021) 0. 128285 (0. 3912) 0. 5155 20. 81987*** (0. 000000) 35. 93624*** 2. 066938 145 0. 495356*** (0. 0001) -0. 00678 (0. 6912) -0. 208717 (0. 3841) 0. 399327 (0. 5074) -0. 139914 (0. 5646) -0. 004627 (0. 5617) 0. 009888 (0. 5976) 0. 257495 (0. 2668) 0. 4019 4. 224526*** (0. 001221) 24. 756586*** (0. 0008) 2. 155818 52 4.2. Global Crisis and Income Diversification We then examine the effect of global crisis to income diversification, the result of panel data estimation is exhibited in Table 2, in detail model 3 provides the effect of global crisis to nontraditional stakeholder non-interest income, while model 4 provides the effect to traditional and fee for service non-interest income. There are two different results, for non-traditional stakeholder, the effect of crisis is negative but it s not significant, on the other hand, it has significant and positive effect to the traditional and fee for service. The negative effect to non-traditional stakeholder non-interest income could be caused by shocks in capital market and other financial markets that imply to the financial asset pricing decreases, As the nontraditional stakeholder is the non-interest income that comes from risky activities such as from investment banking, venture capital and proprietary trading that highly depend on the asset price in financial markets. Longstaff (2010) finds strong evidence the contagion of subprime
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 363 crisis to the other financial market, further by investigating the pricing of subprime asset backed collateralized debt obligations (CDOs), he finds that financial contagion to other markets is propagated through liquidity and risk premium channel. The positive effect of crisis to traditional and fee for service reveals that during global crisis, the pressure of crisis that disturbs banks performance is later compensated by intensify the other source of income, the traditional and fee for service non-interest income. Then we investigate the effect of global crisis for each of ownership type, the estimation result for this analysis is exhibited in Table 4. We calculate this estimation by using pooled effect panel data with similar reasons with previous session in calculating Table 3. For the effect to non-traditional stakeholder non-interest income, we find that the global crisis has negative effect on foreign and private banks but not significant. However it has positive effect on state banks although it s not significant either. The different direction effect of global crisis to private and foreign banks in one side and public bank in other side could be caused by different source of non-traditional stakeholder non-interest income. As mentioned before, this kind of non-interest income sourced from holding risky assets, perhaps assets which are held by private and foreign banks more exposed by global crisis (probably hold more international assets which its value fragile of global crisis), while state banks hold different type of asset. Lastly, the effect of global crisis to traditional and fee for service non-interest income is significantly occurred with the positive direction in foreign banks. It means that during the crisis traditional and fee for service non-interest income of foreign banks increased. This finding in line with Jeon and Miller (2005) that found a steady performance of foreign banks in Korea during Asian Financial crisis while it deteriorated domestic banks performance, one of the reasons for this is because foreign banks in Korea rely more on fee for service income than from lending interest income. Foreign banks in Indonesia seems to follow similar pattern, Hadad et al (2004) concluded that foreign banks in Indonesia were more focus on fee based income business but less active in its intermediation function. So the positive performance of foreign banks to generate traditional and fee for service during global crisis could be caused by the nature of foreign banks which more specialized in non-interest income.
364 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 Table 4 The Effect of the Crisis onthe Income Diversification across Bank's Ownership Independent Variables Public Bank Foreign Bank Private Bank Panel A dependent variable is nontraditional stakeholder NS (lagged) Ln Asset ROE LLP Loan Loan Asset Cap Asset Year Crisis Constant R squared F statistic Hausman Durbin Watson statistic N Panel B dependent variable is traditional fee for service TFS (lagged) Ln Asset ROE LLP Loan Loan Asset 0. 152591 (0. 4372) 0. 223535** (0. 0186) -1. 331828 (0. 3533) -1. 397026 (0. 8519) 0. 695519 (0. 5307) 0. 007084 (0. 4080) -0. 068284 (0. 1762) 0. 120602 (0. 4642) -2. 223967 (0. 0725) 0. 4218 2. 918594** (0. 014526) - 2. 171339 41-0. 189867 (0. 1882) -0. 199866 (0. 1711) 1. 863419* (0. 0975) 7. 139937 (0. 1770) -2. 142975* (0. 096) 0. 521513*** -0. 001633 (0. 9040) 0. 018253 (0. 8406) -0. 600388 (0. 8406) -0. 272622 (0. 1459) -0. 000728 (0. 8289) 0. 008622 (0. 4929) -0. 048180 0. 2559 0. 291845 (0. 1421) 0. 3651 5. 823259 (0. 000007) 42. 669459*** 1. 925548 90 0. 667466*** 0. 008206 (0. 4372) -0. 031376 (0. 6866) 1. 136742 (0. 0245) -0. 100515 (0. 4730) 0. 550921*** -0. 000903 (0. 9507) 0. 242838*** (0. 0015) -0. 542397** (0. 0256) 0. 206692 (0. 3197) -0. 003031 (0. 5492) -0. 025462 (0. 121) 0. 066649 (0. 2251) 0. 191181 (0. 379) 0. 4823 9. 550903*** (0. 000000) 26. 624891*** (0. 0008) 2. 045213 91 0. 48933*** (0. 0001) -0. 008257 (0. 6285) -0. 225947 (0. 3458) 0. 466943 (0. 4393) -0. 104915 (0. 6669)
The Effect of Ownership and Global Crisis to Income Diversification of Indonesian Banking 365 Independent Variables Public Bank Foreign Bank Private Bank Panel Cap Asset A dependent variable Year Crisis Constant R squared F statistic Hausman Durbin Watson statistic N 0. 020151*** (0. 0121) 0. 065436 (0. 2438) 0. 147621 (0. 2216) 2. 678154 (0. 1818) 0. 5031 2. 657678** (0. 034675) Note: Absolute values of t-statistics are in parentheses, *, **, and *** indicate significance at 10%, 5%, and 1% levels respectively. NS: Non-traditional Stakeholder; TFS: Traditional and Fee for Service LNASSET: Log of Total assets; ROE: Return on Equity; LLPLOAN:loan loss provisions/total loans; LOANASSET: Loans to Total Asset; Cap_ta: Capital to Total Asset; Year: Year Dummy; Crisis: Crisis Year Dummy. - 2. 204604 30 0. 001063 (0. 6166) -0. 007438 (0. 3742) 0. 063987** (0. 0371) 0. 120423 (0. 4152) 0. 5307 19. 22771*** (0. 000000) 34. 308043*** 2. 103576 145-0. 004433 (0. 5769) 0. 013383 (0. 4798) -0. 063442 (0. 2563) 0. 264916 (0. 2521) 0. 4198 3. 889134*** (0. 001592) 41. 666991*** 2. 135999 52 V. CONCLUSION Indonesian banking is growing rapidly and attracts diverse investors to participate in this business sector. Several parties which become significant shareholder of Indonesian banking are government, private and foreign owner. The turbulence of world economic raises financial crises which potentially harm the progress of Indonesian banking growth. This paper attempts to examine the effect of ownership and global crisis to income diversification of Indonesian Banks during period of 2005 to 2012. The income diversification is classified as the taxonomy of De Young and Rice (2004), in this paper we divide the income diversification to the non-traditional stakeholder non-interest income and traditional and fee for service non-interest income. Our result suggests that ownership doesn t affects income diversification of Indonesian banks both to the non-traditional stakeholder and traditional and fee for service non-interest income. However, there is a difference in terms of direction, the direction effect of public ownership is negative in both non-traditional incomes, and on the other hand the direction of foreign ownership is positive. Based on the direction of the ownership effect, this result support Pennathur et al (2012) that public banks do not intensively generate their non-interest income, and on the other hand foreign banks can maximize this source of income better. Then, when analyzing in each types of ownership, for public bank capitalization affect significantly in positive direction to non-traditional stakeholder non-interest income. In terms of traditional and fee for service, in state banks, size has a positive and significant effect. For foreign and private banks none of the variables affect non-traditional stakeholder non-interest income unless lagged of
366 Bulletin of Monetary, Economics and Banking, Volume 17, Number 3, January 2015 dependent variable. In Traditional and fee for service non-interest income, in foreign banks credit risk significantly affect in positive direction, but in it affect oppositely in private banks. In addition profitability also affects significantly and positively to traditional and fee for service non-interest income in private banks. The effect of global crisis has different direction to each non-interest income, for nontraditional stakeholder non-interest income it is not significant and has negative effect. However it significantly encourages banks to generate traditional and fee for service non-interest income, this finding reveals that the decline in interest based income due to the crisis condition push banks to compensate through maximize traditional and fee for service non-interest income. For each ownership type s investigation, we find no evidence that global crisis affect non-traditional stakeholder non-interest income generating for all types of ownership, public, private and foreign banks. In terms of traditional and fee for service non-interest income, global crisis has significant and positive effect to traditional and fee for service generating in foreign banks. This finding support Jeon et al (2005) that found better performance of foreign banks than domestic banks during the Asian financial crisis because they rely more on fee for service. These findings have several implications, first the taxonomy of De Young and Rice (2004) should be recognized well, because different kind of income diversification has different characteristics, so everyone should avoid generalizing income diversification. Second, the determinant factors that affect income diversification for each type of ownership are not similar. It should be realized that the effort to maximize non-interest income would be different. Third, related to the bank supervision during crisis period, regulator should concern about different effect of crisis to non-interest income generating by different types of banks, so necessary policies could be taken properly.