Title. The relation between bank ownership concentration and financial stability. Wilbert van Rossum Tilburg University

Transcription

1 Title The relation between bank ownership concentration and financial stability. Wilbert van Rossum Tilburg University Department of Finance PO Box 90153, NL 5000 LE Tilburg, The Netherlands Supervisor: Thomas Mosk European Banking Center - Tilburg University Department of Finance PO Box 90153, NL 5000 LE Tilburg, The Netherlands Telephone: , Fax: Abstract: In this thesis I study the impact of bank ownership concentration on the financial stability, and give answer to the question; does ownership concentration reduces financial instability by improving loan quality and capital adequacy of banks? As indicator for loan quality I use impaired loans and as indicators for capital adequacy I use total capital ratio and Tier 1 capital ratio. Using data for around 1800 banks from almost 80 countries, I find that a higher level of ownership concentration significantly improves the impaired loans ratio in comparison to lower levels of ownership concentration. I find no significant relation between bank ownership concentration and the capital adequacy ratios. Keywords: Ownership concentration Capital adequacy Impaired loans JEL: E44 G21 G30 May 18, 2012

2 1. Introduction The recent crises had a serious impact on banks from all over the world, many large banks collapsed or had to be saved by governments. Therefore bank riskiness and bank stability are intensively discussed in the literature but also an important issue for policy makers. Failing governance of state-owned banks might be a cause of the crisis according to some analysts, others suggest that the focus on profit and short-term thinking of private banks is a cause for the crisis (Beck, Hesse, Kick & Von Westernhagen, 2009). Therefore it may be important to look how management control can be improved such that bank riskiness will decrease. According to Beck et al. (2009) bank stability seems to be related to ownership structure and size. In this paper I will focus on bank ownership concentration and how it can reduce bank financial instability by improved management control. The research question for this study is: Does bank ownership concentration reduce bank financial instability? The indicators I use for bank financial instability are impaired loans and capital adequacy. I analyze data for more than 1500 banks from almost 80 countries averaged over This study is about how ownership concentration reduces bank financial instability by reducing the potential agency problem. Therefore ownership concentration has to improve the capital adequacy ratios and decrease the impaired loans to gross loans ratio. I find that bank ownership concentration significantly affect loan quality, because impaired loans ratios are better with a higher degree of ownership concentration. I find not enough significant results to conclude that capital adequacy is affected by ownership concentration. This study may contribute to the literature because impaired loans is an important indicator of financial stability and bank ownership concentration can reduce this ratio. The remainder of this paper is constructed as follows. Section 2 discusses the related studies about ownership concentration and bank stability and bank performance that already exist in the literature. This section is organized in two subsections; section one discusses the existing theoretical literature and section two discusses the related empirical studies. Section 3 describes the data and shows the frequency distribution of the different ownership levels. Section 4 describes the model I use. Section 5 presents the results of this study. Section 6 gives the conclusion and discusses the implications of the results of this study.

3 2. Literature review 2.1. Theory The relationship between the owners of a firm and the managers is defined as an agency relationship where the owners are the principals and the corporate managers are the agents. According to Jensen and Meckling (1976) the principals give authority for decision making on their behalf to the agents. There are also firms where owners are controlling the firm, when the owners are in the managers of the firm, in that case there is no agency relationship. In this paper I will focus on firms with a separation of ownership, because that is the case for almost all large banks in the world. Separation of ownership and control occurs when owners of a firm do not manage the firm themselves but hire outside managers to control the firm. Owners of a firm with separation of ownership and control could face potential agency problems because there could be a difference between their own interests and that of the management. Jensen and Meckling (1976) pointed this out in the agency theory where they defined the agency problem as managers that may act on their own behalf rather than that of the firm s shareholders. According to Berk and DeMarzo (2007) managers have incentives to invest in risky negative- NPV projects, what is conflicting with the interests of the shareholders because such projects would decrease the value of their shareholdings. Monitoring the management of the firm is a solution to the agency problem for the shareholders and is a measure to control the management. With monitoring the shareholders will control the management because they receive information about the performance of the management and they can replace bad performing managers on basis of this information. However monitoring brings costs with it and as we see later on in this paper that costs makes monitoring not beneficial for all shareholders. According to Jensen and Meckling (1976) the costs of monitoring are part of the agency costs, they defined agency costs as the sum of the monitoring expenditures by the principal, the bonding expenditures by the agent and the residual loss. Because management control and monitoring is important for the owners it is necessary to know who the shareholders are, what characteristics they have and how the ownership structure looks like. According to Ianotta, Giacomo and Sironi (2007) ownership structure

4 can be divided into two main dimensions. Firms may differ because of their degree of ownership concentration, firms can have a high or low concentration of ownership. If their degree of ownership concentration is the same, firms may differ by the nature of the owners. In the banking industry there are different types of owners, banks owned by the government, banks owned by private stockholders and mutual banks. The degree of ownership concentration defines whether the firm has a dispersed ownership or concentrated ownership. The firm has dispersed ownership when many shareholders own a small percentage of the shares. Figure 1 shows an example of a dispersed ownership structure. When the firm has one or more large shareholders with a large percentage of the firm s shares it has a concentrated ownership, large shareholders are also called blockholders. A higher percentage of shares owned by the large shareholder(s) will increase the degree of ownership concentration. Figure 2 shows an example of a concentrated ownership structure. Figure 1 Dispersed ownership structure. Figure 2 Concentrated ownership structure. Shareholders S 1 S 2 S 3 S 4 Etc. Shareholders Large blockholder(s) S 2 S 3 Etc. Management Management Firm Firm In the literature there are conflicting theories about the impact of ownership concentration on firm performance and what the optimal ownership structure is. Berle and Means (1933) studied the consequences of the separation of ownership from control of the modern corporation and concluded that the effective power shareholders have on the control of management is decreasing when ownership structure is more dispersed. Shleifer and Vishny (1986) studied the relation between large shareholders and corporate control and the freerider problem of small owners in a corporation. According to Shleifer and Vishny (1986) better corporate control is achieved by concentrated ownership because it improves the monitoring of management. With dispersed ownership there are many small shareholders that have little incentives to monitor the management because they do not have enough

5 shares of the firm to absorb the costs of monitoring, therefore monitoring will not be profitable for them. Large shareholders in a concentrated ownership do have incentives to monitor the management because they can bear the costs of it. When monitoring will improve the management all the shareholders will benefit from this improvement, this is the free-rider problem. Winton (1993) and Bolton and Von Thadden (1998) argued also that large shareholders act on behalf of all the shareholders by monitoring the management and face a free-rider problem with this. According to Gomes and Novaes (1999) there is also a difference in incentives between large shareholders and minority shareholders, they show that minority shareholders may benefit from the difference of interests between controlling shareholders. However, they conclude that ownership structure has to be more dispersed because minority shareholders are required for an optimal ownership structure since they reduce disagreement costs between controlling shareholders. Burkart, Gromb and Panunzi (1997) point out that dispersed ownership will commit the shareholders to control management not excessively, because that may have a negative impact on the level of noncontractible investments by managers. With ownership concentration increases the effective control by shareholders; however monitoring can reduce the effectiveness of performance-based incentive schemes. There are also studies concluding that regulation may decrease the impact of ownership concentration. Regulation in the banking sector means that banks are subjected to certain rules, principles or laws that are designed to control or govern procedure and behavior of banks. The impact of regulation on firms has been studied by Demsetz and Lehen (1985). They cast doubt on the thesis of Berle and Means (1933), because they found no significant relationship between ownership concentration and firm performance and point out that regulation leads to better disciplining of managers in industries with strong regulation. This is important for this paper because the financial sector does have a strong regulation. Elyasiani and Jia (2008) found also that ownership concentration matters less if there is strong regulation and concluded that regulation can be a substitute for management control. The relation between the financial stability of a bank and ownership structure is the importance of good monitoring, because managers have incentives to invest in risky projects

6 (Berk & DeMarzo, 2007) monitoring is necessary to keep managers in control. Banks riskiness will increase when there is less control, because managers may invest more in risky negative-npvprojects when they do not bear the full consequences of their actions (Shehzad et al., 2010). According to Shleifer and Vishny (1986) large shareholders have more incentives to influence corporate behavior than smaller owners and from this perspective Boyd and Hakenes (2008) concluded that ownership structure has an impact on the ability of owners to adjust bank risk to their own interest in response both to shifting incentives of standard risk and to incentives created by official regulations. Thus, risk-taking behavior of banks is shaped by the way ownership structure interacts with bank regulation (Laeven & Levine, 2009) Empirical There are several studies that suggest that ownership structure has impact on the monitoring and control of firm s managers and that different types of ownership structures have different consequences for the performance and stability of firms. Ianotta et al. (2007) studied the impact of alternative ownership models and different ownership levels on the performance and risk of European banks, using a sample of 181 large banks from 15 European countries averaged over They found that ownership concentration does not have a significant impact on the profitability, but a higher degree of ownership concentration is related to better loan quality, lower asset risk and lower insolvency risk. This study is more focused on the dimension of ownership concentration and uses a larger sample which contains data of banks from all over the world instead of only European banks. This paper is most related to Shehzad, De Haan & Scholtens (2010) who studied the impact of bank ownership concentration on banks riskiness for around 500 banks from more than 50 countries averaged over They tested the traditional Berle-Means thesis that banking firm performance is improved by ownership concentration against the view that bank s riskiness is not influenced by ownership concentration, where banks riskiness is measured with impaired loans and capital adequacy. They found that the impact of ownership concentration was negative on the impaired loans to gross loans ratio and has a positive impact on capital adequacy. They show also that bank riskiness is reduced by ownership concentration at low levels of shareholder protection rights and supervisory

7 control. This study uses two ratios to measure capital adequacy, total capital ratio and tier 1 capital ratio, instead of the Shehzad et al. (2010) study. They also use some regulatory control variables like; shareholder protection rights, supervisory control, activities restrictions and bank concentration. The focus in this paper is not on regulations and protection rights for shareholders but I include extra measures for the performance and stability of a bank, with return on assets and interbank ratio. Shehzad et al. (2010) uses an extra threshold of 10% for shareholders owning less than 10% of the shares, where I only use ownership levels of less than 25%, between 25% and 50% and more than 50%. 3. The data 3.1. Sample In this section I discuss the data for the independent variable and the dependent and control variables I use in this study. The sample consists of data of around 1800 of the largest banks over the world (as defined by total assets averaged over ), these banks are from almost 80 countries and the data is averaged over The summarized statistics including the means and standard deviations are presented in Table 1. Table 2 shows the correlation matrix and reports the correlation coefficients between all the variables in the model Ownership concentration The data I use for ownership concentration is from Bureau Van Dijk s Bankscope database. To measure this dependent variable I use the BvD Independence Indicator, defined by BvD as an indicator that characterize the degree of independence of a company with regard to its shareholders. Where A indicates that there is no shareholder with shareholdings more than 25%, B indicates that there is no shareholder with shareholdings more than 50% and at least one with shareholdings more than 25%, and C and D indicate a company that has at least one shareholder with shareholdings more than 50%. The data I collected for all banking companies is for the years as reported in the 2012 version of the Bankscope database. Figure 3 shows the distribution of the different levels of bank ownership concentration. Of the almost 1300 banks in this sample 820 banks at least one shareholder with shareholdings more than 50%, 120 of the almost 1300 banks have at least one shareholder with shareholdings more than 25% but no shareholders owning more than 50%,

8 almost 360 banks from the sample have no shareholders owning more than 25%. For empirical analyses I generated two indicators for ownership concentration. Ownership level 1 is a dummy that equals one when the BvD Independence Indicator is C or D and zero otherwise. Ownership level 2 is a dummy that equals one when the BvD Independence Indicator is B and zero otherwise. Figure 3 Distribution of different ownership levels Frequency <25% 25% - 50% >50% Ownership level 3.3. Impaired loans and capital adequacy The data for impaired loans and capital adequacy are also from Bureau van Dijk s Bankscope database. For impaired loans also called non-performing loans I take the impaired loans to gross loans ratio from the period This ratio is a standard proxy for bank s asset risk, it is the impaired loans divided by the gross loans of the bank. Impaired loans or nonperforming loans are loans that are unlikely to be paid back for the full amount and therefore one of the bank s greatest risks. To measure capital adequacy I use two ratios; the total capital ratio and the Tier 1 capital ratio, both are from data of the years 2008 till The total capital ratio is an important ratio to measure bank s solvency and is therefore included in the standards of the Basel Committee of the Bank for international Settlements

9 (BIS). This ratio is the total capital divided by the risk-adjusted assets, which is a requirement for banks, to have a minimal amount of capital based on a percentage of the assets weighted by risk, to cover their liabilities. The Tier 1 capital ratio is also a standardized requirement measured under the Basel rules. With this ratio only the Tier 1 capital is divided by the riskadjusted assets. Tier 1 capital is the core capital of a bank and consists of common stock and disclosed reserves Control variables For this study I use bank characteristics and one macroeconomic characteristic as control variables. The bank-level data are all from Bureau Van Dijk s Bankscope database averaged over the period I use six bank-level control variables. First, the return on assets as indicator of bank performance. This ratio is bank s net income divided by its assets. Second, the interbank ratio as indicator of bank independency and measure of bank s liquidity, defined as the money from a bank lent to other banks as a percentage of the money borrowed from other banks. Third, the cost to income ratio to measure bank efficiency. A low cost to income ratio implies better bank efficiency, because the costs are measured as a percentage of income. Fourth, bank equity is used as a proxy for bank size, it is the total capital of the bank. Fifth, loan growth gives information about growth opportunities of the bank and is measured with the growth of gross loans ratio. Sixth, a dummy variable to control whether the bank is listed or not. The last control variable is a macroeconomic one, GDP per capita for the country where the bank is located derived from data of World Bank over the period This variable measures the relative performance of a country compared to another country. It is the total output of a country divided by the number of people living in that country. 4. The model I use a model with the dependent variables impaired loans and capital adequacy. Where impaired loans is defined as the impaired loans to gross loans ratio, which is a standard proxy for bank s asset risk. Capital adequacy is measured with the total capital ratio and the tier 1 capital ratio. Both ratios are included in the international bank solvency standards of the Basel Committee of the Bank for International Settlements. The total capital ratio measures bank capitalization, banks with low capitalization are more risky and less stable

10 (Demirüç-Kunt, Detragiache & Tressel, 2006; Podpiera, 2004). And tier 1 capital ratio measures bank riskiness on basis of bank s core equity capital. The independent variable in this model is ownership concentration. The control variables are return on assets as indicator of bank profitability, cost to income ratio as indicator of bank efficiency, interbank ratio as measure of liquidity management, bank equity to measure the bank s size, loan growth as a proxy for a bank s growth opportunities, listed bank is a dummy that equals 1 and 0 for a listed bank or unlisted bank respectively, and GDP per capita of the country where the bank is located. So this is the model: Y ij = β 0 + β 1 (OC ij ) + β 2 (ROA ij ) + β 3 (Cost/income ij ) + β 4 (Interbank ij ) + β 5 (Equity ij ) + β 6 (LoanGrowth ij ) + β 7 (Listed i ) + β 8 (GDPpercapita j ) + ε ij Where Y is the independent variable, and could be the impaired loans to gross loans ratio, the total capital ratio and the Tier 1 capital ratio, of bank i in country j. OC is an indicator of bank ownership concentration, ROA is the return on assets, Cost/income is the cost to income ratio, Interbank is the interbank ratio, Equity is an indicator of bank size, LoanGrowth is the growth of gross loans rate, Listed is a dummy indicating whether the bank is listed or not, and GDP per capita is income per capita for the country where the bank is located. ε t is the disturbance term, it is appended to the model because Y can be affected by unforeseen events or factors that are not measured. Table 5 shows the expected signs for the variables used in this model. 5. Results 5.1. Regressions This section presents the simple regression results of the model, the main results are shown in Table 3. Where regression 1 relates the impaired loans only to bank-level variables. Regression 2 in addition includes the macroeconomic variable GDP per capita. Regressions 3 and 4 in the table have total capital as the dependent variable, and are otherwise fully analogous to regressions 1 and 2. Regressions 5 and 6 in the table have the other measure of capital adequacy, the Tier 1 ratio, as the dependent variable, and are otherwise fully analogous to regressions 1 and 2.

11 5.2. Results for the control variables on impaired loans In the model for the impaired loans the coefficients of the control variables have the expected signs. The coefficient of return on assets is in both regressions negative, implying that banks with higher returns on assets have lower impaired loans to gross loans ratios. The coefficient is significant at the 1% significance level. The cost to income ratio shows only in regression 1 a significant regression coefficient at the 10% level, the coefficient is negative, implying that with an increasing cost to income ratio impaired loans will decrease. Interbank ratio has a small negative impact on impaired loans, what means that banks with higher interbank ratios have lower impaired loans to gross loans ratios, indicated by a low negative coefficient. However, the coefficient is only significant at the 10% and 5% level for regression 1 and 2 respectively. The size of the bank measured by bank equity has no significant impact on impaired loans in both regressions. Banks with more growth opportunities, as measured by the growth of gross loans ratio, have lower impaired loans to gross loans ratios. However, the coefficient is also only significant at the 5% level in both regressions. In accordance with the findings of Shehzad et al. (2009) listed banks appear to have higher impaired loans to gross loans ratios. The coefficient is significant at the 1% level in both regressions. The variable GDP per capita has a negative coefficient implying that banks located in a country with a higher income per capita have lower impaired loans to gross loans ratios. The coefficient is significant at the 1% level. The model explains more than 17% of the variation in the data Results for the control variables on capital adequacy Capital adequacy is measured with two variables, therefore I use two models; one for the total capital ratio and one for the Tier 1 ratio. The model for the total capital ratio explains only 0.9% and 0.5% of the variation in the data for regression 3 and 4 respectively. Only return on assets has the expected sign and is also the only variable that is significant, at the 5% significance level, in the regression of the model that includes the macroeconomic variable. The coefficient is positive, implying that banks with higher return on assets have higher total capital ratios. Bank equity has only in regression 3 a significant coefficient, at the 10% level. The coefficient is negative, implying that larger banks have lower total capital ratios. Loan growth is also only significant in regression 3, at the 5% level. With a negative

12 coefficient it indicates that banks with more growth opportunities have lower capital adequacy. All the other control variables are not significant. The model for the Tier 1 capital ratio explains 4.9% and 6.6% of the variation in the data for regression 5 and 6 respectively. Return on assets has the expected positive sign, what means that banks with higher returns on assets have higher Tier 1 ratios. The coefficient is significant at the 1% level in both regression 5 and 6. The cost to income ratio does not have the expected sign, but shows a positive effect on the Tier 1 ratio, implying that banks with lower bank efficiency have higher Tier 1 ratios. The coefficient is significant at the 1% level in both regression 5 and 6. The coefficient of the interbank ratio is positive, this is the expected sign, what means that banks with higher interbank ratios have higher tier 1 ratios. Thus, banks that are more independent form other banks have better capital adequacy. The coefficient is significant at the 5% level in both regression 5 and 6. The control variables for bank size and growth opportunities are insignificant in both regressions. Also in accordance with the findings of Shehzad et al. (2009) listed banks have lower Tier 1 ratios, what means that listed banks appear to have worse capital adequacy. The coefficient is significant at the 1% and 5% level for regression 5 and 6 respectively. The GDP per capita variable has a positive coefficient, significant at the 1% level. Implying that banks located in a country with higher income per capita have better capital adequacy Results for ownership concentration on impaired loans and capital adequacy This study is about the difference between banks with large controlling shareholders and banks with small shareholders with no significant control. There are significant results for the different ownership levels in both regressions for the model with the dependent variable impaired loans and for the model with the dependent variable total capital ratio in the regression without the macroeconomic variable. The highest level of ownership concentration is ownership level 1, banks with at least one owner with shareholdings greater than 50%, for this variable the coefficient is in regression 1 and in regression 2. The coefficient is in both regressions significant at the 1% level. For ownership level 2, banks with at least one owner with shareholdings greater than 25% but no owners with shareholdings greater than 50%, the coefficient is in regression 1 and in regression 2. For both ownership levels the coefficients are

13 positive. The coefficients of ownership level 2 are larger than the coefficients of ownership level 1, implying that banks with ownership level 2 have higher impaired loans than banks with ownership level 1. This suggests that when ownership concentration is above 50% the impaired loans to gross loans ratio will increase less than with a lower degree of ownership concentration. This result is in line with the findings of Shehzad et al. (2010). The model with the dependent variable total capital ratio shows only significant coefficients for the different ownership levels in regression 3. The coefficient of ownership level 1 is and is significant at the 5% level. The coefficient for ownership level 2 is and is significant at the 10% level. The coefficient of ownership level 2 is lower than the coefficient of ownership 1, this means that banks with ownership level 2 have lower total capital ratios than banks with ownership level 1. This suggests that banks with an ownership concentration above 50% have better total capital ratios than banks with lower ownership levels. 6. Conclusion In this thesis I studied the impact of ownership concentration on impaired loans and capital adequacy ratios for a sample of around 1800 banks from almost 80 countries. I find that loan quality is significantly affected by ownership concentration. Banks with a higher degree of ownership concentration tend to have lower impaired loans to gross loans ratios. This is in line with the findings of Shehzad et al. (2010). The effect of ownership concentration on the capital adequacy ratio was only significant in the model without the macroeconomic variable for the total capital ratio. I find not enough significant results from this study to conclude anything about the impact of ownership concentration on capital adequacy. Therefore I can only answer the research question partly, that is that loan quality is improved by ownership concentration. Acknowledgement I am thankful to T.C. Mosk, PhD from Tilburg University for supervising this thesis. His comments, criticism and support contributed to the quality and analysis of this paper.

14 Appendix Table 1 Summary statistics on bank and macroeconomic variables. The table presents the summary statistics of variables. Impaired loans is the impaired loans to gross loans ratio, which is defined as a percentage of gross loans. Capital adequacy is one variable that is the total capital (Tier 1 and Tier 2) divided by the risk-adjusted assets. And one variable that is the Tier 1 ratio, which is defined as the Tier 1 capital divided by the risk-adjusted assets. Ownership level 1 is a dummy that is one for banks with at least one owner with shareholdings greater than 50%. Ownership level 2 is a dummy that equals one for banks that has at least one owner with shareholdings greater than 25% but no owners with shareholdings greater than 50%. Ownership level 3 is a dummy variable that equals one for banks that has no owners with shareholdings greater than 25%. Return on assets is the return on average assets for the banks in US dollars. Cost/income is the cost to income ratio, which is defined as the costs as a percentage of income. Interbank is the interbank ratio, which is defined as the money lent to other banks as a percentage of the money lent from other banks. Bank equity is the total equity of the bank in millions of US dollars. Loan growth is the growth rate of gross loans. Listed bank is a dummy variable that is one for banks that are listed. The data for these bank level variables are obtained from Bankscope. GDP per capita is in thousands of constant US dollars, this variable is obtained from World Development indicators of the World Bank. Variable Observations Mean Std. Dev. Minimum Maximum Bank variables Impaired loans 3, Total capital ratio 3, Tier 1 ratio 3, Ownership level 1 5, Ownership level 2 5, Ownership level 3 5, Return on assets 4, Cost/income 4, Interbank ratio 3, Bank equity 4, Loan growth 4, Listed Bank 5, Macroeconomic variable GDP per capita 3,

15 Table 2 Correlations between the variables. Variable The correlation matrix presents the correlation coefficients between the variables. Impaired loans is the impaired loans to gross loans ratio, which is defined as a percentage of gross loans. Total capital is the total capital ratio defined as the Tier 1 and Tier 2 capital divided by the risk-adjusted assets. Tier 1 ratio is defined as the Tier 1 capital divided by the risk-adjusted assets. OC 1 is a dummy that is one for banks with at least one owner with shareholdings greater than 50%. OC 2 is a dummy that equals one for banks that has at least one owner with shareholdings greater than 25% but no owners with shareholdings greater than 50%. Return on assets is the return on average assets for the banks in US dollars. Cost to income is the cost to income ratio, which is defined as the costs as a percentage of income. Interbank is the interbank ratio, which is defined as the money lent to other banks as a percentage of the money lent from other banks. Bank equity is the total equity of the bank in millions of US dollars. Loan growth is the growth rate of gross loans. Listed bank is a dummy variable that is one for banks that are listed. The data for these bank level variables are obtained from Bankscope. GDP per capita is in thousands of constant US dollars, this variable is obtained from World Development indicators of the World Bank. Impaired loans Impaired loans 1.00 Total capital Total capital Tier 1 ratio Tier 1 ratio OC OC 1 OC 2 Return on assets OC Return on assets Cost to income Cost to income Interbank Interbank Bank equity Bank equity Loan growth Loan growth Listed bank GDP per capita Listed bank GDP per capita

16 Table 3 The impaired loans and capital adequacy. The dependent variable in column 1 and 2 is Impaired loans, and the dependent variable in column 3-6 is Capital adequacy. Impaired loans is the impaired loans to gross loans ratio, which is defined as a percentage of gross loans. Capital adequacy is one variable that is the total capital (Tier 1 and Tier 2) divided by the risk-adjusted assets. And one variable that is the Tier 1 ratio, which is defined as the Tier 1 capital divided by the riskadjusted assets. Ownership level 1 is a dummy that is one for banks with at least one owner with shareholdings greater than 50%. Ownership level 2 is a dummy that equals one for banks that has at least one owner with shareholdings greater than 25% but no owners with shareholdings greater than 50%. Return on assets is the return on average assets for the banks in US dollars. Cost/income is the cost to income ratio, which is defined as the costs as a percentage of income. Interbank is the interbank ratio, which is defined as the money lent to other banks as a percentage of the money lent from other banks. Bank equity is the total equity of the bank in millions of US dollars. Loan growth is the growth rate of gross loans. Listed bank is a dummy variable that is one for banks that are listed. The data for these bank level variables are obtained from Bankscope. GDP per capita is in thousands of constant US dollars, this variable is obtained from World Development indicators of the World Bank. Standard errors are in parentheses. *, ** and *** denote significance at 10%, 5%, and 1%, respectively. Impaired loans Capital adequacy Total capital ratio Tier 1 ratio Variable (1) (2) (3) (4) (5) (6) Ownership level *** (0.226) 0.815*** (0.221) ** (1.514) (1.102) (0.281) (0.327) Ownership level *** (0.299) 0.902*** (0.297) * (2.018) (1.478) (0.378) (0.441) Return on assets *** (0.078) *** (0.076) (0.519) 0.966** (0.376) 0.979*** (0.096) 1.125*** (0.112) Cost/income * (0.003) (0.003) (0.022) (0.015) 0.013*** (0.004) 0.013*** (0.005) Interbank * (0.001) ** (0.000) (0.003) (0.002) 0.001** (0.001) 0.002** (0.001) Bank equity (0.004) (0.004) * (0.024) (0.019) (0.004) (0.005) Loan growth ** (0.003) ** (0.002) ** (0.018) (0.012) (0.003) (0.003) Listed bank 1.198*** (0.209) 0.653*** (0.205) (1.401) (1.028) *** (0.258) ** (0.304) GDP per capita *** (0.004) (0.021) 0.027*** (0.006) N 2,485 1,811 2,447 1,792 2,385 1,771 R²

17 Table 4 Variable definitions and sources. Dependent variables Description Data source Impaired loans Impaired or non-performing loans are loans that are Bankscope unlikely to be paid back for the full amount. The impaired loans to gross loans ratio is used to measure bank s asset risk. Capital adequacy Tier 1 capital ratio is used to measure the capital Bankscope adequacy of banks. Is the Tier 1 capital divided by the risk-adjusted assets. Total capital Ratio is the total capital (Tier 1 and Tier 2) divided by the risk-adjusted assets. Explanatory variables Ownership concentration Return on assets Interbank ratio Cost/Income Equity Loan growth Listed bank GDP per capita The amount of stock owned by individual investors or institutions. A high degree of ownership concentration implies that there is one ore more investor(s) that owns a significant amount of stock. While dispersed ownership contains that there is not a single investor that owns enough stock to control the bank. The profits of the bank divided by the assets. Ratio to measure the dependency between banks. It is the money lent to other banks as a percentage of the money borrowed form other banks. The cost/income ratio is used to measure the efficiency of the bank. Total equity of the bank. The bank equity is used to measure the size of the bank. Growth of gross loans of the bank. This ratio is used to measure the growth opportunities of the bank. Whether the bank is listed or not, listed means that the shares are listed on a stock exchange for public trading. GDP per capita is the income per person living in the country the bank is located. Bankscope Bankscope Bankscope Bankscope Bankscope Bankscope Data World Bank

18 Table 5 The expected signs for the variables The expected sign for the independent variable Ownership concentration explains the impact of the degree of ownership concentration on the dependent variables. An increase in value of the variable Ownership concentration indicates an increase in the degree of ownership concentration. Expected Sign for dependent variables Explanatory variables Impaired loans Total capital ratio Tier 1 ratio Ownership concentration Return on assets Interbank ratio Cost/income Bank equity Loan growth Listed bank +/- +/- +/- GDP per capita - + +

19 References Beck, T., Hesse, H., Kick, T. & Von Westernhagen, N. (2009). Bank ownership and stability: Evidence from Germany. Retrieved from VOX website: Berk, J., & DeMarzo, P. (2007). Corporate Finance. Boston, MA: Pearson Education. Berle, A.A., & Means, G.C. (1933). The Modern Corporation and Private Property. New York: MacMillan. Bolton, P. & Von Thadden, E.L. (1998). Blocks, Liquidity, and Corporate control. Journal of Finance, 53(1), Boyd, J.H. & Hakenes, H. (2008). Looting and Gambling in Banking Crises, mimeo, University of Minnesota. Burkart, M., Gromb, D. & Panunzi, F. (1997). Large shareholders, monitoring, and the value of the firm. Quarterly Journal of Economics, 112, Demirüç-Kunt, A., Detragiache, E. & Tressel, T. (2006). Banking on the Principles: Compliance with Basel Core Principles and Bank Soundness. Research Working Paper No Demsetz, H. & Lehn, K. (1985). The structure of corporate ownership: causes and consequences. Journal of Political Economy, 93, Elyasiani, E & Jia, J.J. (2008). Institutional ownership stability and BHC performance. Journal of Banking and Finance, 32, Gomes, A. & Novaes, W. (1999). Multiple Large Shareholders in Corporate Governance (Working Paper). The Wharton School., Philadelphia, PA. Ianotta, G., Giacomo, N., & Sironi, A. (2007). Ownership structure, risk and performance in the European banking industry. Journal of Banking and Finance, 31, Jensen, M.C. & Meckling, W.H. (1976). Theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3, Laeven, L.A. & Levine, R. (2008). Bank Governance, Regulation, and Risk Taking (NBER Working Paper No ). Podpiera, R. (2004). Does Compliance with Basel Core Principles Bring any Measurable Benefits? (IMF Working Paper No. 04/204). Shehzad, C.T., De Haan, J. & Scholtens, B. (2010). The impact of bank ownership concentration on impaired loans and capital adequacy. Journal of Banking and Finance, 34,

20 Shleifer, A. & Vishny, R.W. (1986). Large shareholders and corporate control. Journal of Political Economy, 94, Winton, A. (1993). Limitation of Liability and the Owner Structure of the Firm. Journal of Finance, 48,