The influence of ownership concentration and the Basel III capital requirements on bank risk taking in the European Union

The influence of ownership concentration and

the Basel III capital requirements on bank risk

taking in the European Union

Student number: 2054701 Contact: v.s.skambraks@student.rug.nl

University of Groningen Faculty of Economics and Business MSc International Financial Management

Uppsala University Department of Business Studies

MSc Business and Economics

Supervisor: prof. dr. C.L.M. (Niels) Hermes

January 2015

ABSTRACT

This study researches the relations among ownership concentration, the Basel III capital standards, and bank risk taking. In this, the central theme are the different interest of owners and managers regarding risk taking. The results indicate that the relation between ownership and risk taking is nonlinear. Furthermore, they show that the effects of capital requirements on risk taking differs with the degree of ownership concentration. Altogether, the Basel III capital standards seem to be an effective tool in decrease risk taking. However, the regulator should take into account that the influence of these standards varies with the degree of ownership concentration.

2

1. Introduction

This paper studies the influence of the Basel III capital requirements on the risk taking behavior of banks from the European Union. Furthermore, it aims to determine if the degree of ownership concentration of these banks is a critical factor in how government regulations, like the Basel III capital standards, effect bank risk taking behavior. This allows to draw conclusions on whether ownership concentration is a critical factor that influences the effectiveness of Basel III in reducing risk for individual banks. Basel III is implemented in 2013. However, the main capital standards already existed in the years before that, although some had lower minimums. Therefore, it is possible to research the effects of these standards in the years before 2013. This research stretches the years 2006 to 2013.

Understanding the risk taking behavior of banks is highly important. Some banks are more risky than others. Prominent factors in determining the degree of risk taking are the efforts of diversified shareholders, who advocate for high risk taking resulting from their limited liability (Galai and Masulis, 1976; Esty, 1988). The financial crisis of 2008 showed us that bank risk taking behavior has a significant impact on the welfare of the public. Risk taking by banks can affect economic fragility, business cycle fluctuations, and economic growth (Bernanke, 1983; Calomiris and Mason, 1997, 2003a, b; Keeley, 1990). Accordingly, policy makers attempt to reduce the risk taking behavior by imposing restrictions and regulations on the banking sector. Recent examples are new capital and stability rules, developed by the Basel Committee of Banking Supervision. These regulations, referred to as Basel III, are planned to become effective over time. The first phase was implemented in 2013, and completion is expected in 2018. However, its predecessors, Basel I and Basel II, already enforced minimum capital standards in the European Union. Therefore, the effects of these capital standards can also be researched in the years before 2013.

3 the Basel III regulations. They focus on a set of Asian banks and study how the liquidity requirements influence risk taking. Their results imply, in line with theory, that higher ownership concentration results in higher risk taking behavior. However, they do not find a significant interactive effect of ownership concentration with the Basel III liquidity standards.

This paper extends on the work by Laeven and Levine (2009), by researching bank risk taking behavior in a more recent period. The banking sector underwent major changes due to new regulations, corporate governance reforms, and the financial crisis of 2008-2009. Consequently, the results of Laeven and Levine (2009) might not be applicable to the current economic situation. Furthermore, this study extends on Chalermchatvichien et al. (2014) by focusing on European banks instead of Asian banks, and by focusing on a fundamental different part of Basel III.

This study aims to analyze: (1) whether ownership concentration is related to higher bank risk taking behavior, (2) whether the capital requirements of Basel III are able to reduce the risk taking of banks, and (3) whether the magnitude and/or the direction of the impact of the Basel III capital standards on bank risk taking depends on the degree of ownership concentration. By answering these questions, this study will contribute to the discussion on the effects of Basel III. Furthermore, it will provide the first findings on the effects of the actual introduction of the Basel III capital standards on bank risk, while taking ownership concentration into consideration. The results can help policy makers to formulate rules and regulations that are effective in reducing bank risk taking. Accordingly, these effective regulations can help prevent, or lower the impact of future financial crises’.

The paper is structured as follows. Section 2 presents the theoretical background. Section 3 elaborates on the methodology. Section 4 reports on the sample and summarizes the data. Section 5 presents the empirical results. Lastly, section 6 concludes.

2. Theory

2.1 Ownership concentration and bank risk taking

4 taking than managers without shareholdings. The owners are limited liable and can therefore benefit from increasing risk after collecting funds from bond holders and non-shareholder managers (Galai and Masulis, 1976; Esty, 1988). Managers on the other hand, will pursue a lower degree of risk taking behavior. This is particularly true when they possess bank-specific human capital or private benefits of control (Demsetz and Lehn, 1985; Kane, 1985). Banks with more powerful owners can overcome these risk averse tendencies of managers (Houston and James, 1995). Consequently, the banks that empower diversified owners, with sufficient shareholdings, will show a higher tendency towards risk taking behavior than banks with a governance structure where these diversified owners have less influence. In comparison to small shareholders, these large owners have more power to influence decision making, and are more motivated to increase the risk taking of the bank (Jensen and Meckling, 1976; John, Litov and Yeung, 2008). The empirical results of Laeven and Levine (2009) and Chalermchatvichien et al. (2014) confirm this.

Managerial ownership can bring the interest of managers more in line with that of diversified owners. If the manager is owner in the bank and is diversified himself, he can also benefit from increasing the risk (Saunders et al., 1990). Furthermore, shareholder protection laws can force the managers to act in line with the interest of the shareholders (Shleifer and Wolfenzon, 2002; John, Saunders, and Senbet, 2000; Castro, Clementi, and MacDonald, 2004). These reduce the need for a large owner to mitigate this friction. Therefore, large owners are expected to be less prominent in influencing risk taking in countries with effective shareholder protection laws. Wen risk taking is mentioned throughout this paper, it refers to risk taking by banks.

2.2 Basel III regulations and bank risk taking

5 the risk management and the governance of individual banks, they also aim to strengthen the transparency and the disclosure of banks (Basel Committee on Banking Supervision, 2010a). The debate surrounding the likely effects of Basel III on risk taking behavior, and in the end its ability to forestall an economic crisis, has been lively (Yan, Hall, and Turner, 2012). Nonetheless, there is no empirical evidence that shows that these regulations actually result in lower risk taking.

This study researches the capital standards of Basel III, and whether they contribute to the lowering of risk taking. It can be argued that these standards decrease risk because they force banks to reduce leverage, and thus to put more of their own capital at risk. However, this leads to a loss in utility for diversified shareholders. They may seek to compensated for this by selecting a more risky position (Kim and Santomereo, 1980). Besides, Blum (1999) argues that one should consider the intertemporal effects of capital regulations. Raising equity to meet the higher capital standards tomorrow can be (too) costly, causing banks to choose for the alternative of raising the risk today.

This study focusses on two of the Basel III capital standards. These are the tier 1 ratio and the total capital ratio. The tier 1 ratio is the ratio between the bank’s core equity and its risk weighted assets. The total capital ratio also is a comparison between the bank’s core equity plus supplementary capital to its risk weighted assets. The minimum requirements for these two capital standards are displayed in Table 1. The table presents the minimum requirements of Basel III, as well as those of its predecessors. Basel I covers the first two years of this study (2006 to 2008), and Basel II the period of 2008 to 2012. As can be seen from Table 1, the minimum for the total capital ratio has not change since 2006. Furthermore, this ratio will remain unchanged as the phases of Basel III progress. However, Basel III raises the minimum requirement for the tier 1 ratio. This ratio already increased from 4% to 4.5%, in the year 2013. It is planned to further increase until it reaches 6% in 2015.

Table 1: Capital requirements of the Basel accords

2006-2013> 2013 2014 2015 2016 2017 2018 2019

Tier 1 ratio 4,0% 4,5% 5,5% 6,0% 6,0% 6,0% 6,0% 6,0%

Total

capital ratio 8,0% 8,0% 8,0% 8,0% 8,0% 8,0% 8,0% 8,0%

6 Next to the minimum requirements for the tier 1 ratio and the total capital ratio, the first phase also presented minimums for the common equity capital ratio and the common equity plus the capital conversion buffer. The next phases will introduce liquidity standards and additional capital standards. Furthermore, the minimum requirements for some of these standards will increase as the phases progress (for a more comprehensive overview of these phases and the minimum requirements for these ratio’s see Appendix A).

No specific research has been done on whether the capital standards of Basel III contribute to the goal of the Basel Committee of Banking Supervision, that is to lower risk taking by banks. Furthermore, Basel III is only active since 2013. Hence, an opportunity arises to test the real effects of Basel III on risk taking for the first time.

2.3 The interaction between ownership concentration and capital standards

7 like capital standards, dependent on the power of the owners in a bank, and therfore on the degree ownership concentration.

2.4 Hypothesis development

With the premises of the theoretical section this study will explore the impact of the Basel III capital standards on the degree of risk taking by banks in the European Union. Firstly, the relation between ownership concentration and risk taking is researched. Secondly, the focus will be on relation between the capital standards of Basel III and risk taking. Lastly, it is studied if there is and interactive effect between ownership concentration and the capital regulations. This is done in order to determine if the effect of capital standards on risk taking is dependent on the degree of ownership concentration. The hypothesizes are as follows:

Hypothesis 1: Banks with higher ownership concentration will show a higher degree of risk taking behavior.

Hypothesis 2: The tier 1, and the total capital standards of Basel III, will reduce the risk taking of banks.

Hypothesis 3: The effect of capital standards on the bank’s risk taking critically depends on the ownership structure of the bank. The Basel III capital standards will increase bank risk taking when the bank has a large owner.

2.5 Recent literature

8 only becomes active in 2013. Therefore, they project the NSFR in order to research its likely effects. Chalermchatvichien et al. (2014) find that the NSFR has a diminishing effect on risk taking. However, they do not find a significant relation between ownership concentration and the magnitude of the impact of the NSFR on risk taking. There are two possible explanations for this lack of significance. Firstly, their study focusses on the hypothetical impact of the liquidity requirements of Basel III. It does not examine the real effects of the NSFR. Secondly, the study examines a sample of Asian banks. However, their results indicate that the NSFR is more effective in reducing risk taking in countries with a higher GDP per capita. Therefore, Chalermchatvichien et al. (2014) argue that the effects will be more prominent in countries with a higher degree of economic development. They reason that this increasing effect arises because financial and accounting information are more reliable in developed countries. Stricter disclosure requirements lead to more transparency, which is a better basis for enforcing regulations. The same can be argued for regulations regarding capital standards. These also depend critically on financial and accounting information. Following this reasoning, banks from the European Union would form a more suitable sample for reaching the effects of Basel III. However, one should keep in mind that these findings are derived from a research on only Asian banks, and thus might not be generalizable.

3. Methodology

3.1 The European Union as sample

The focus is on the European Union for a number of reasons. Firstly, because financial and accounting information is reliable and readily available in the member states. Secondly, because these countries have an increased comparability because a high degree of regulatory convergence. Thirdly, because the European Union has the best record in the implementation of Basel I and Basel II, and accordingly already has minimum requirements for the tier 1 ratio and the total capital ratio. This is important because it enables this study to research these capital standards over the period 2006 to 2013.

3.2 Bank risk taking

9 measures how far a bank is from insolvency. This is expressed in how many standard deviations the return on assets (ROA) can drop before depleting the bank’s equity. Hence, a higher score implies a more stable bank (Roy, 1954). The z-score is calculated by dividing the ROA plus the capital-asset ratio (CAR) by the standard deviation of the return on assets [σ (ROA)], which is calculated over the sample period. The z-score displays a highly skewed distribution. Therefore, the natural logarithm of this measure is used to create a normal distribution. When the z-score is mentioned throughout this paper, it refers to the natural logarithm of the z-score.

Beside the z-score, an alternative measure for bank risk taking is used in the analysis. This is the volatility of equity returns, also used by Saunders et al. (1990), Esty (1998), Laeven and Levine (2009), and Chalermchatvichien et al. (2014). It is measured by annualizing the volatility of the total return index from Datastream. This return index calculates the theoretical capital gains, assuming that dividends and other cash distributions are reinvested in the bank. Therefore, it allows for a better comparison between banks that issue dividends, and banks that do not issue dividends but reinvest their earnings in the bank. Using this second measure gives the possibility to confirm the results with a measure based on market data instead of accounting data.

3.3 Basel III capital standards

10

3.4 Ownership Concentration

Ownership concentration is measured, similar to Laeven and Levine (2009) and Chalermchatvichien et al. (2014), as the cash flow rights of the largest shareholder. First, it is determined if the bank has a large owner. If this is the case, the cash flow rights of the largest owner are measured, with data on direct ownership of Bureau van Dijk, available via Bankscope. Banks are classified as having a large owner if the largest shareholder has cash flow rights summing up to 10% or more (Caprio, Laeven, and Levine, 2007). When there is no owner with this amount of voting rights, the bank is classified as widely held. In this case the cash flow rights variable takes the value zero. The results are also confirmed with a threshold of 20% ownership to be qualified as a large owner. Cash flow rights are highly correlated with voting rights. However, potential profits are distributed on the basis of cash flow rights, rather than voting rights. Therefore, these provide a more direct measure of the owner’s incentive to influence the risk taking behavior. As outlined in the theoretical section, ownership concentration is expected to be positively related to bank risk taking. This is due to the fact that owners are more prone to risk taking incentives than managers.

3.5 Bank-level and country-level control variables

11 provisions. The value of the bank is measured through the use of the Tobin’s Q. This variable is calculated by dividing the market capitalization plus the book value of the liabilities by the book value of the assets. A lower bank value is expected to create more incentives to engage in risk taking behavior (Gonzalez, 2005).

To control for country-level effects, economic size of the country, and the economic development of the country are accounted for. The economic size is measure through the natural logarithm of the total GDP of each country. Economic development is calculated as the natural logarithm of the per capita income for each country. These are important because strong competition can decrease the value of incumbent banks, and consequently increases the incentives for owners and managers to engage in risky behavior (Gorton and Rose, 1995).

Information is also collected on the management structure and on managerial bank ownership. A dummy variable is used to identify if the bank’s largest owner, as measured by the direct ownership rights, has a seat on the management board. This dummy takes the value of one if the largest owner has at least one seat on the management board, and zero otherwise. Furthermore, the total percentage of cash flow rights of the management board is calculated. This is relevant because managerial ownership can bring the interest of owners and managers more in line (Saunders et al. 1990). Therefore, managerial ownership can cause higher risk taking, even when the bank is widely held. Consequently, these two variables are expected to be positively related to risk taking.

4. Sample and summary statistics

The data set is created by collecting data on all the listed banks of the EU-member states that are available on Bankscope. Furthermore, the study uses data from Datastream for the calculation of the equity volatility, and information of The World Bank on the GDP of the EU-member states. Laeven and Levine (2009) and Chalermchatvichien et al. (2014) restrict their sample to the top ten largest banks per country. This in order to decrease accounting and liquidity difference, and thus to enhance comparability. In this study however, this is less of a concern. The sample consists of banks from the European Union, which are better comparable as a result of regulatory convergence. Therefore, this restriction is not imposed on this study’s sample.

12 stretches the period 2006 till 2013 (see Appendix B for the averages of key variables for each countries banks). In an attempt to create more comprehensive findings on the effects of Basel III, an attempt was made to collect data on the common equity capital ratio of the banks. However, this information was so scarcely available that it was decided to exclude it from the data set.

The descriptive statistics are displayed in Table 2. Laeven an Levine’s (2009) sample consists of banks from all over the world. Therefore, their descriptive statistics will be used as a rough proxy for the averages of the variables worldwide. As can be seen from Table 2, the z-score has an average of 3.11 with a standard deviation of 0.42. This value is higher than that of Laeven and Levine (2009), 2.88, implying that the banks from the European Union take fewer risks. However, it is lower than 3.53, the mean in the study of Chalermchatvichien et al. (2014). This indicates that Asian banks are even more risk averse. Equity volatility averages 0.38, which is slightly lower than the average of Chalermchatvichien et al. (2014), who report a mean of 0.4. It is also lower than 0.45, the average in the research of Laeven and Levine (2009). Lower equity volatility is associated with lower risk, again implying that banks from the European Union are relatively less risky. However, Asian banks are even more risk

connecting

Table 2: Descriptive Statistics

Observations

Number of

banks / countries Minimum Maximum Mean

Std. Deviation

Ln Z 1096 171 2,12 4,03 3,11 0,42

Equity volatility 1062 168 0,00 2,57 0,38 0,25

CFR 1096 171 0,00 1,00 0,35 0,33

Loan loss provision 1082 167 -0,89 1,49 0,32 0,35

Tobin's Q 1096 171 0,23 1,04 0,91 0,07 Revenue growth 1090 171 -1,21 3,11 0,00 0,15 Economic development 1096 26 4,04 4,96 4,51 0,12 Economic size 1096 26 9,81 12,56 11,75 0,64 Size 1096 171 4,10 9,41 7,18 1,08 Liquidity 1075 166 0,01 6,47 0,25 0,37

Large owner on mgt board 1096 171 0,00 1,00 0,04 0,20

Managerial ownership 1047 162 0,00 0,22 0,01 0,04

Tier 1 ratio 1096 171 -0,06 0,69 0,12 0,06

Total capital ratio 1096 171 -0,05 0,68 0,14 0,05

This table reports summary statistics of the main regression variables. The sample includes 171 publicly listed banks, from 26 countries in the European Union. Ln Z is the natural logarithm of the z-score, which is calculated as the return on assets plus the capital asset ratio, divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns, calculated using weekly data over the sample period. CFR is the ownership concentration, measured as the cash flow rights of the largest shareholder of the bank. Loan loss

provision is the ratio of the loan loss provision to the net interest revenue. Tobin’s Q is the market value of the equity plus the book value of

13 Table 3: Correlation matrix

Variable Ln Z Equity volatility CFR Loan loss provision Tobin's Q Revenue growth Economic development Economic

size Size Liquidity

Large owner on mgt

board

Managerial

ownership Tier 1 ratio

EquityIvolatility -0,392** (4,805) 1055 CFR -0,049 0,088** (0,104) (0,004) 1096 1062 Loan loss provision -0,348** -0,029 -0,046 (0,000) (0,347) (0,134) 1076 1054 1082 Tobin's Q -0,040 -0,017 0,114** _-0,030 (0,185) (0,582) (0,000) (0,330) 1096 1062 1103 1082 Revenue growth 0,074* 0,066* 0,015 -0,006 0,008 (0,015) (0,031) (0,629) (0,841) (0,781) 1083 1062 1090 1070 1090 Economic development -0,057 0,387** -0,132** 0,044 0,117** -0,026 (0,060) (0,000) (0,000) (0,149) (0,000) (0,396) 1096 1045 1103 1082 1103 1090 Economic size 0,085** 0,126** 0,008 -0,018 0,097** 0,024 0,363** (0,005) (0,000) (0,790) (0,564) (0,001) (0,435) (0,000) 1096 1062 1103 1082 1103 1090 1103 Size 0,047 0,191** 0,060* 0,019 0,481** 0,003 0,151** 0,407** (0,122) (0,000) (0,045) (0,528) (0,000) (0,911) (0,000) (0,000) 1096 1062 1103 1082 1103 1090 1103 1103 Liquidity -0,074* _{-0,049 -0,017 0,025 -0,284}** _{-0,002 0,075}* _0,067* _-0,062* (0,016) (0,113) (0,570) (0,424) (0,000) (0,960) (0,013) (0,028) (0,043) 1068 1035 1075 1060 1075 1062 1075 1075 1075 Large owner on mgt board -0,013 0,043 0,017 0,034 -0,064* 0,014 -0,131** -0,124** -0,211** -0,034 (0,672) (0,166) (0,578) (0,268) (0,035) (0,634) (0,000) (0,000) (0,000) (0,260) 1096 1062 1103 1082 1103 1090 1103 1103 1103 1075 Managerial ownership -0,096** _{-0,085** -0,022 -0,019 -0,306}** _{-0,005 0,042 0,121}** _-0,252** _0,151** _0,014 (0,002) (0,007) (0,486) (0,535) (0,000) (0,879) (0,173) (0,000) (0,000) (0,000) (0,645) 1041 1007 1047 1026 1047 1036 1047 1047 1047 1021 1047 Tier 1 ratio 0,102** -0,089** -0,025 -0,062* -0,597** -0,039 -0,021 -0,041 -0,361** 0,213** 0,043 0,250** (0,001) (0,004) (0,399) (0,041) (0,000) (0,202) (0,486) (0,172) (0,000) (0,000) (0,157) (0,000) 1096 1062 1103 1082 1103 1090 1103 1103 1103 1075 1103 1047 Total capital ratio 0,110** -0,104** -0,051 -0,056 -0,561** -0,052 0,030 -0,025 -0,326** 0,203** 0,012 0,256** 0,934** (0,000) (0,001) (0,089) (0,065) (0,000) (0,085) (0,323) (0,401) (0,000) (0,000) (0,684) (0,000) (0,000) 1096 1062 1103 1082 1103 1090 1103 1103 1103 1075 1103 1047 1103 This table reports the correlations between the main regression variables. The sample includes 171 publicly listed banks, from 26 countries in the European Union. Ln Z is the natural logarithm of the z-score, which is calculated as the return on assets plus the capital asset ratio, divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns, calculated using weekly data over the sample period.

CFR is the ownership concentration, measured as the cash flow rights of the largest shareholder of the bank. Loan loss provision is the ratio of the loan loss provision to the net interest revenue. Tobin’s Q is the market

14 averse. The ownership concentration is with 0.35 is higher than 0.24, the average of Laeven and Levine (2009). It is only slightly above the mean of Chalermchatvichien et al. (2014), who have an average of 0.32 for Asian banks. Consequently, banks from the European Union and Asia seem to be relatively more concentrated. The loan loss provision has a mean of 0.32, whereas Laeven and Levine (2009) have an average of 0.23. Banks from the European Union seem to have a loan loss provision ratio that is higher than average. Chalermchatvichien et al. (2014) find a mean of 0.17 for Asia banks, what is even lower. This could be a result of the recent financial crisis, due to the following two reasons. First, it can be argued that banks keep higher loan loss provisions to account for an increase in the amount of bad loans. Second, the denominator of this ratio could be decreased because of lower interest revenue than before the crisis. The Tobin’s Q averages 0.91, which is lower than 1.04, the average of Chalermchatvichien et al. (2014). The average of the revenue growth is 0.00, which is lower than the means of both Laeven and Levine (2009) and Chalermchatvichien et al. (2014). Furthermore, the size averages 7.18 which is a lot lower than 16.2, the mean in the study of Laeven and Levine (2009). These three differences are all expected to be a result of the recent financial crisis. Managerial ownership varies enormously, just like in the research of Laeven and Levine (2009). Also similar to their results, the average is relatively small. This sample has an average of 1% managerial ownership of the total bank shares, were they show an average of 6% managerial ownership. In only 4% of the banks, the largest owner has a seat on the management board. However, this is not taking into account owners who have less than 10% of the voting rights.

15

5. Empirical results

First the relation between the bank’s degree of ownership concentration and risk taking is examined, with the use of ordinary least squares (OLS). Next, the bank-level, and then the country-level controls are introduced to the regression. In the second section the capital requirement variables are included to examine their effect on the bank’s risk taking. The third section studies the interaction effect between ownership concentration and capital requirements. The aim is to research if the relation between capital requirements and risk taking varies with the bank’s degree of ownership concentration. The regression framework can be expressed by the following equation:

𝑍𝑏,𝑐,𝑡= 𝛼 ∗ 𝑋𝑏,𝑐,𝑡+ 𝛽 ∗ 𝑌𝑐,𝑡+ 𝛾 ∗ 𝐶𝐹𝑏,𝑐,𝑡+ 𝛿 ∗ 𝐶𝑅𝑏,𝑐,𝑡+ 𝜔 ∗ 𝐶𝐹𝑏,𝑐,𝑡∗ 𝐶𝑅𝑏,𝑐,𝑡+ 𝜀𝑏,𝑐,𝑡 , (1)

where 𝑍𝑏,𝑐,𝑡 equals the z-score of bank b in country 𝑐 at time 𝑡, 𝑋𝑏,𝑐,𝑡 and 𝑌𝑐,𝑡 are the bank and

the country control variables, 𝐶𝐹𝑏,𝑐,𝑡 is the cash flow right of the largest shareholder of bank b

in country 𝑐 at time 𝑡, 𝐶𝑅𝑏,𝑐,𝑡 are the capital ratios of bank b in country 𝑐 at time 𝑡, 𝜀𝑏,𝑐,𝑡 is the

error term, and α, β, γ, δ, and ω are the vectors of the coefficient estimates.

5.1.1 Ownership concentration and bank risk

16 of the z-score. The results of Regression 6 show that a one standard deviation increase in cash flow rights results in an increase of equity volatility with 0.025 (= 0.076 * 0.33). The findings both imply an increase in risk, and in line with the theory outlined in the theoretical section. Owners tend to advocate for a higher degree of risk taking (Galai and Masulis, 1976; Demsetz and Lehn, 1985). Larger owners (with higher cash flow rights) possess more power and have a higher incentive to influence bank risk in comparison with smaller owners (Jensen and Meckling, 1976, John et al., 2008). Regression 7 and 8 provide mixed findings for the effects of the board and management variables on this relation. Having a large owner on the

connecting

Table 4: Bank risk and ownership concentration Ownership concentration Ownership concentration Country level Country level Bank and country level Bank and country level Board and management Board and management (1) (2) (3) (4) (5) (6) (7) (8) Dependent variable Ln (Z) Equity volatility Ln (Z) Equity volatility Ln (Z) Equity volatility Ln (Z) Equity volatility Constant 3.130*** 0.353*** 3.797*** 0.311 4.535*** 0.224 4.358*** 0.292 (0.019) (0.011) (0.469) (0.290) (0.470) (0.278) (0.465) (0.261) CFR -0.064 0.070** -0.075* 0.066*** -0.085** 0.076*** -0.065* 0.074*** (0.039) (0.039) (0.039) (0.025) (0.037) (0.022) (0.037) (0.021) Revenue growth 0.238** -0.058 0.115 -0.065 0.108 -0.040 (0.105) (0.053) (0.099) (0.047) (0.098) (0.044) Economic development -0.358*** -0.074 -0.239** -0.148** -0.201* -0.162*** (0.110) (0.068) (0.105) (0.063) (0.104) (0.059) Economic size 0.081*** 0.032** 0.059*** 0.013 0.085*** 0.018 (0.021) (0.013) (0.022) (0.013) (0.022) (0.013) Loan loss provision -0.440*** 0.302*** -0.447*** 0.273*** (0.037) (0.021) (0.037) (0.020) Tobin's Q -1.293*** 0.316* -1.374*** 0.204 (0.282) (0.167) (0.285) (0.159) Size 0.045*** 0.036*** 0.015 0.042*** (0.014) (0.008) (0.015) (0.009) Liquidity -0.112*** -0.021 -0.092*** -0.015 (0.035) (0.021) (0.034) (0.019) Large owner on mgt board 0.008 0.081** (0.060) (0.034) Managerial ownership -1.957*** -0.111 (0.380) (0.211) Adjusted R² 0.002 0.007 0.020 0.011 0.156 0.207 0.169 0.206 Observations 1095 1061 1082 1053 1041 1015 990 963 This table presents regression results regarding the effect of ownership concentration. The sample includes 171 publicly listed banks, from 26 countries in the European Union. Ln Z is the natural logarithm of the z-score, which is calculated as the return on assets plus the capital asset ratio, divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns, calculated using weekly data over the sample period. CFR is the ownership concentration, measured as the cash flow rights of the largest shareholder of the bank. Loan

loss provision is the ratio of the loan loss provision to the net interest revenue. Tobin’s Q is the market value of the equity plus the book

value of the liabilities, divided by the book value of the assets. Revenue growth is the growth in revenues over the last year. Economic

development is the natural logarithm of the GDP per capita for each country. Economic size is the natural logarithm of the total GDP of each

country. Size is the natural logarithm of the total assets. Liquidity is the ratio between of the liquid assets to the deposits and borrowings.

Large owner on mgt board takes the value one if the largest shareholder has a seat on the management board and zero otherwise. Managerial ownership is the sum of the management’s total cash flow rights. The regressions are estimated using ordinary leas squares. The standard

17 management board is insignificant with the z-score. But, when equity volatility is used this has an increasing effect on risk. The degree of managerial ownership is significant and has a negative relation with the z-score. However, the results on equity volatility are insignificant. The results regarding ownership concentration confirm the findings of Laeven and Levine (2009), who researched a sample from the late 1990’s. They also confirm the results of Chalermchatvichien et al. (2014), who study a more recent period and focus on a sample of Asian banks. Concluding, a higher degree of ownership concentration is associated with a higher degree of risk taking.

5.1.2 Equity volatility as measure of bank risk

The yearly equity volatility is calculated by using weekly data of the total return index from Datastream. As can be seen from Regressions 2, 4, 6 and 8 in Table 4, the key results are robust using equity volatility as an alternative proxy for bank risk. The coefficient CFR is positive and significant in all these regressions. This implies that a higher ownership concentration results in higher equity volatility, and thus a higher degree of risk.

5.1.3 Instrumental variables

18

5.1.4 Possible endogeneity

Endogeneity can result from reverse causality. Risk taking might influence the independent variables, like ownership concentration, and not vice versa. This study addresses this concern by lagging the independent variables by one year. This approach decreases the likelihood that that risk influences these variables. Hence, if the results remain significant, causality is much more likely to run from the independent variables to risk taking than the other way around. The results on the main variable are robust (Regressions 3 and 4 in Table C1 of Appendix C). Ownership concentration still shows a negative and significant relation with the z-sore, and a positive and significant relation with equity volatility. However, most of the other variables lose their significance. Thus, endogeneity arising from reverse causality between risk and those variables cannot be excluded.

5.1.5 Qualification as large owner

This study qualifies a bank as having a large owner if they have an owner that possesses more than 10% of the cash flow rights. The CFR variable of banks below that cutoff rate takes the value of zero. However, it could be argued that a larger ownership percentage is needed to influence the bank’s risk taking (Caprio et al., 2007). To reduce these concerns, the results are verified with an 20% cutoff rate to be qualified as having a large owner. The results hold with this stricter criterion. Ownership concentration still has an increasing and significant effect on risk taking. The other variables remain relatively unchanged (Regression 1 in Table 2C of Appendix C).

5.1.6 Nonlinearity

19 coefficients of the other regression variables remain relatively unchanged, and most of them increase in significance. Furthermore, tests are also conducted with a dummy that takes the value of one if the cash flow rights of the bank are higher than the 75th percentile. However, this does not yield significant results. Secondly, a quadratic term of cash flow rights is used to test for nonlinearity. CFR² enters the regression negative and significant, while CFR is positive and significant. Again the results imply that the relation between risk and ownership concentration is nonlinear. Concluding, concentrated ownership is negatively related to risk. However, this negative effect diminishes when ownership concentration becomes very large.

Table 5: Bank risk, capital requirements and ownership concentration

Tier 1 Tier 1 Total capital Total capital

Capital requirements

Capital requirements

(1) (2) (3) (4) (5) (6) Dependent variable Ln (Z) Equity volatility Ln (Z) Equity volatility Ln (Z) Equity volatility

Constant 4.005*** 0.214 4.041*** 0.269 4.129*** 0.177 (0.481) (0.270) (0.471) (0.266) (0.482) (0.272) CFR -0.074** 0.071*** -0.072** 0.073*** -0.070* 0.069*** (0.037) (0.021) (0.037) (0.021) (0.037) (0.021) Revenue growth 0.109 -0.038 0.108 -0.039 0.108 -0.041 (0.098) (0.044) (0.097) (0.045) (0.097) (0.045) Economic development -0.227** -0.168*** -0.252** -0.166*** -0.26*** -0.159*** (0.104) (0.059) (0.105) (0.060) (0.105) (0.060) Economic size 0.084*** 0.018 0.085*** 0.018 0.086*** 0.017 (0.022) (0.013) (0.022) (0.013) (0.022) (0.013)

Loan loss provision -0.435*** 0.275*** -0.434*** 0.274*** -0.437*** 0.275***

(0.037) (0.021) (0.037) (0.021) (0.037) (0.021) Tobin's Q -0.959*** 0.296 -0.962*** 0.233 -1.046*** 0.324* (0.320) (0.180) (0.306) (0.173) (0.322) (0.182) Size 0.018 0.043*** 0.019 0.042*** 0.020 0.042*** (0.015) (0.009) (0.015) (0.009) (0.015) (0.009) Liquidity -0.098*** -0.016 -0.100*** -0.015 -0.100*** -0.015 (0.034) (0.020) (0.034) (0.020) (0.034) (0.020) Large owner on mgt board 0.010 0.081** 0.021 0.082** 0.026 0.076** (0.060) (0.034) (0.060) (0.034) (0.060) (0.034) Managerial ownership -2.001*** -0.121 -2.061*** -0.118 -2.081*** -0.098 (0.379) (0.211) (0.378) (0.211) (0.379) (0.211) Tier 1 ratio 0.740*** 0.161 -0.514 0.546 (0.265) (0.148) (0.602) (0.342)

Total capital ratio 1.027*** 0.072 1.534** -0.467

20

5.2 The impact of capital standards

In the regressions of Table 5, the capital requirement variables are introduced as independent variables. Both the tier 1 ratio and the total capital ratio have a highly significant and strong positive relation with the z-score, and thus a negative effect on risk taking. These relations do not change significantly when ownership concentration is excluded from the regressions (Appendix C, Table 3C). An increase of the tier 1 ratio by one standard deviation (0.06) will result in an increase of the z-score by 0.044 ( = 0.06 * 0.74). A one standard deviation increase of the total capital ratio (0.05), results in an increase of 0.051 ( = 0.05 * 1.027). This study also tests the relation between risk and regulation is also with equity volatility as the dependent variable. Here, the capital requirements variables do not enter with significant coefficients. Thus, the tier 1 ratio and the total capital ratio affect the z-score of the bank, but not the bank’s equity volatility. This is in line with theory, that suggest that capital standards force banks to put more of their own capital at risk, and thereby reduces bank risk taking (Kim and Santomereo, 1988). The results show that the Basel III capital standards have a diminishing effect on the risk taking of banks. Moreover, setting a minimum for these capital standards seems to be an effective tool to decrease bank risk taking. Concluding, the capital standards of Basel III contribute to the goal of the Basel Committee which are aimed at lowering risk in the banking sector.

5.3 The 2008 financial crisis

21 Table 6: Effects of the '08 financial crisis

Crisis effects with CFR

Crisis effects with CFR

Crisis effect with total capital ratio

Crisis effect with total capital ratio

(1) (2) (3) (4)

Dependent variable Ln (Z) Equity volatility Ln (Z) Equity volatility

Constant 4.554*** 0.046 4.317*** -0.019 (0.471) (0.261) (0.477) (0.266) CFR -0.071* 0.092*** -0.090** 0.076*** (0.043) (0.024) (0.037) (0.021) CFR * Crisis -0.051 -0.053 (0.083) (0.047)

Total capital * crisis 0.526 -0.242

(0.563) (0.312) Crisis -0.014 0.197*** -0.101 0.214 (0.040) (0.022) (0.083) (0.046) Revenue growth 0.111 -0.042 0.113 -0.037 (0.099) (0.045) (0.098) (0.045) Economic development -0.246** -0.104* -0.292*** -0.115* (0.106) (0.059) (0.106) (0.060) Economic size 0.060*** 0.010 0.057*** 0.010 (0.022) (0.012) (0.022) (0.012)

Loan loss provision -0.44*** 0.302*** -0.429*** 0.305***

(0.037) (0.020) (0.037) (0.020) Tobin's Q -1.276*** 0.254 -0.910*** 0.342** (0.282) (0.157) (0.307) (0.171) Size 0.045*** 0.038*** 0.049*** 0.038*** (0.014) (0.008) (0.014) (0.008) Liquidity -0.113*** -0.018 -0.121*** -0.020 (0.035) (0.020) (0.035) (0.020)

Total capital ratio 0.792** 0.253

(0.330) (0.182) Adjusted R² 0.156 0.301 0.164 0.301 Observations 1041 1015 1041 1015

This table presents regression results regarding the effect of the ’08 financial crisis. The sample includes 171 publicly listed banks, from 26 countries in the European Union. Ln Z is the natural logarithm of the z-score, which is calculated as the return on assets plus the capital asset ratio, divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns, calculated using weekly data over the sample period. CFR is the ownership concentration, measured as the cash flow rights of the largest shareholder of the bank. Crisis takes the value 1 in the years 2008 and 2009 and zero otherwise. Loan loss provision is the ratio of the loan loss provision to the net interest revenue. Tobin’s Q is the market value of the equity plus the book value of the liabilities, divided by the book value of the assets. Revenue

growth is the growth in revenues over the last year. Economic development is the natural logarithm of the GDP per capita for each country. Economic size is the natural logarithm of the total GDP of each country. Size is the natural logarithm of the total assets. Liquidity is the ratio

between of the liquid assets to the deposits and borrowings. Total capital ratio is the bank’s total capital ratio, under Basel III rules. The regressions are estimated using ordinary leas squares. The standard error is provided in parentheses. *, ** and *** indicate significance at 10%, 5% and 1% level, respectively.

interesting finding from Regression 2 is that the coefficient of the variable Crisis is highly significant and positive. This indicates that equity volatility significantly increased in the years 2008 and 2009, as a result of the financial crisis.

22

5.4 Interaction between ownership concentration and capital requirements

As discussed before, theory suggests the impact of capital regulation on risk taking depends on the bank’s ownership structure. This section explores if this premise is extendable to the effects of the Basel III capital standards on bank risk taking. Interaction variables between the capital ratios and ownership concentration are included in the regressions in Table 7. The interaction variable between the tier 1 ratio and ownership concentration enters Regression 1 positive and significant. Regression 2 shows similar results for the interaction effect of the total capital ratio with ownership concentration. The coefficient of this interaction variable is also positive and significant. These results indicate that the stabilizing effect of the capital ratios on risk taking is boosted if the bank has a large owner. However, this is in contrast with theory that suggests that the presence of a large owner diminishes the stabilizing effect of capital regulations on risk taking.

The findings regarding the nonlinearity of ownership concentration could offer an explanation for this contradiction. Prior results indicate that very large owners are associated with less risk taking. Therefore, the inclusion of very large owners could bias the results of the interaction effect. To exclude this possible bias, the sample is split in two. The study uses sample median, 0.23, as the cutoff rate. Keeping in mind that an owner with 10% ownership is qualified as a large owner, an owner with more than 23% ownership can be qualified as a very large owner. Subsample 1 is restricted to all observations that have a ownership concentration lower than sample median. Subsample 2 contains the other half of the sample, thus banks with high ownership concentration. Regressions 3 and 4 show the results of the first subsample. The signs of the interaction variables are the opposite to those of the full sample. Regressions 3 shows that the interaction effect of tier 1 ratio with ownership concentration is negative and significant when banks with very large owners are excluded. This also holds for the total capital ratio, as can be seen from Regression 4. Its interaction effect with ownership concentration is negative and significant. The results on the interaction effects in Subsample 2 are shown in Regression 5 and 6. Here the signs of the coefficients of the interaction terms change back to negative ones, bringing the results in line with those of the full sample. This indicates that the effect on banks with smaller owners is opposite to that of banks with very large owners.

23 Table 6: Interaction effects

Full sample Subsample 1 Subsample 2

(1) (2) (3) (4) (5) (6) Dependent variable Ln (Z) Ln (Z) Ln (Z) Ln (Z) Ln (Z) Ln (Z) Constant 4.182*** 4.170*** 2.188*** 2.346*** 6.348*** 5.831*** (0.483) (0.475) (0.630) (0.622) (0.741) (0.769) Tier 1 ratio * CFR 2.019*** -5.830** 9.771*** (0.694) (2.530) (2.070) Total capital * CFR 1.531** -5.997* 5.373** (0.779) (3.262) (2.434) CFR -0.303*** -0.283** 0.647* 0.828* -1.486*** -1.100** (0.087) (0.113) (0.337) (0.478) (0.257) (0.348) Revenue growth 0.106 0.106 0.017 0.019 0.140 0.147*** (0.097) (0.097) (0.122) (0.122) (0.140) (0.142) Economic development -0.192* -0.228** -0.142 -0.155 -0.195 -0.232 (0.105) (0.105) (0.142) (0.144) (0.149) (0.153) Economic size 0.088*** 0.087*** 0.136*** 0.137*** 0.112*** 0.100*** (0.022) (0.022) (0.031) (0.031) (0.035) (0.035)

Loan loss provision -0.438*** -0.436*** -0.386*** -0.390*** -0.468*** -0.469***

(0.037) (0.037) (0.038) (0.038) (0.064) (0.065) Tobin's Q -1.221*** -1.112*** -0.013 -0.158 -3.081*** -2.679*** (0.331) (0.315) (0.345) (0.320) (0.603) (0.611) Size 0.014 0.017 -0.005 -0.006 0.044 0.059** (0.015) (0.015) (0.016) (0.016) (0.030) (0.030) Liquidity -0.105*** -0.104*** -0.080** -0.080** -0.123* -0.163** (0.034) (0.034) (0.035) (0.035) (0.063) (0.064) Large owner on mgt board 0.015 0.023 0.095 0.105 -0.088 -0.044 (0.060) (0.060) (0.066) (0.066) (0.097) (0.098) Managerial ownership -1.957*** -2.040*** -1.258*** -1.331*** -2.018*** -1.996*** (0.377) (0.378) (0.446) (0.457) (0.712) (0.724) Tier 1 ratio -0.232 1.300*** -6.403*** (0.426) (0.395) (1.557)

Total capital ratio 0.267 1.287*** -2.863

(0.483) (0.435) (1.840) Adjusted R² 0.181 0.181 0.252 0.249 0.208 0.185 Observations 990 990 492 492 498 488 Number of banks 171 171 81 81 90 90 This table presents regression results regarding interaction effects between the capital standards and ownership concentration. The full sample includes 171 publicly listed banks, from 26 countries in the European Union. Subsample 1 is restricted to banks with a lower than median ownership concentration. Subsample 2 is restricted to banks with a higher than median ownership concentration. Ln Z is the natural logarithm of the z-score, which is calculated as the return on assets plus the capital asset ratio, divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns, calculated using weekly data over the sample period. CFR is the ownership concentration, measured as the cash flow rights of the largest shareholder of the bank.. Loan loss provision is the ratio of the loan loss provision to the net interest revenue. Tobin’s Q is the market value of the equity plus the book value of the liabilities, divided by the book value of the assets. Revenue growth is the growth in revenues over the last year. Economic development is the natural logarithm of the GDP per capita for each country. Economic size is the natural logarithm of the total GDP of each country. Size is the natural logarithm of the total assets. Liquidity is the ratio between of the liquid assets to the deposits and borrowings. Tier 1 ratio is the bank’s tier 1 ratio, under Basel III rules. Total capital ratio is the bank’s total capital ratio, under Basel III rules. Large owner on mgt board takes the value one if the largest shareholder has a seat on the management board and zero otherwise. Managerial ownership is the sum of the management’s total cash flow rights. The regressions are estimated using ordinary leas squares. The standard error is provided in parentheses. *, ** and *** indicate significance at 10%, 5% and 1% level, respectively.

24 First, the positive effect of capital ratios diminishes, after which it regains its negative effect on bank risk.

6. Conclusion

This study empirically assessed the relations between ownership structure, the Basel III capital regulations, and bank risk taking, using a sample of banks from the European Union. It follows from the literature that a higher degree of ownership concentration results in more risky behavior. This arises from the fact that diversified owners have higher risk taking incentives than non-shareholding managers. Next to the incentives, large owners also have the ability to influence bank risk taking. Regarding capital requirements, theory suggests that they have a diminishing effect on risk taking. It forces owners to put more of their own wealth at risk and by doing so, it decreases the leverage ratio. Furthermore, theory predicts that the effects of capital standards on risk differ with the ownership structure. Therefore, this study researches the interactive effects of ownership concentration with the Basel III capital standards on bank risk taking.

This research offers some interesting findings. The study confirms the theoretical premise that owners have stronger incentives to increase bank risk than managers, and that they are better able to do so. The results show that a higher degree of ownership concentration is associated with a higher degree of risk taking. Therefore, Hypothesis 1 can be confirmed. However, this study adds this conclusion that the relation is nonlinear. When the degree of ownership concentration rises, the increasing effect on risk taking diminishes.

The results on the Basel III capital standards are in line with expectations. The results indicate that the tier 1 ratio, and the total capital ratio have a negative relation with bank risk taking. Accordingly, a rise in these capital ratios results in a decrease in the risk taking of banks. This confirms Hypothesis 2, and is in line with theories predicting that capital standards reduce risk by forcing banks to put more of their own capital at risk.

25 3 can be confirmed. However, in addition to this confirmation, this study finds that the positive relation diminishes with a further increase of ownership concentration. When the owner becomes very large, the relation eventually changes back to a negative one. After this point, an increase of the capital standards, again, results in a decrease of risk taking. This can be explained by theories suggesting that owners who have a very large proportion of their wealth invested in the bank, advocate for less risk taking.

The findings offer important implications. Capital standards seem to be an effective way to decrease risk. Accordingly, setting minimum capital standards, as done with the Basel III accord, contributes to the goals of the Basel Committee. Under the right conditions, this indeed lowers the risk taking behavior of banks, improving their ability to absorb shocks from financial distress. This is critical in the quest of the Basel Committee to forestall an economic crisis. However, if ownership structure is ignored, one can arrive at an incomplete or incorrect conclusion regarding this relation. It is important that regulators and policy makers are aware of these effects. Furthermore, the findings are important for bank executives. The results offer insights into how capital standard influence the risk taking incentives of the bank’s owners, depending on their degree of ownership.

This research is subject to several limitations, impacting the validity of the results and their interpretations. Firstly, the study focuses on the tier 1 ratio and the total capital ratio of banks from the European Union. These are part of a package of measures that together form the Basel III accord. Researching these two standards does not take into account the effects of the other regulations, and thereby does not offer conclusions on the full effects of Basel III on risk taking. Secondly, the implementation of the Basel III regulations is currently at its first phase. With the years, banks can adjust their risk taking in reaction to the increasingly stricter Basel III regulations. Therefore, the full effects can only be known in a number of years.

26

Appendix A: Basel III Phase in arrangements (Basel Committee on Banking Supervison, 2013)

27

Appendix B: Sample distribution per country

Table 1B: Averages of the variables by country

Country Ln Z Equity volatility CFR Loan loss provision Tobin's Q Revenue growth Economic development Economic

size Size Liquidity

Large owner on mgt board Managerial ownership Tier 1 ratio Total capital ratio Nr of banks Austria 3.40 0.25 0.40 0.31 0.92 0.00 4.61 11.59 7.30 0.21 0.00 0.00 0.10 0.14 10 Belgium 2.56 0.72 0.38 0.08 0.97 -0.01 4.58 11.68 8.59 0.19 0.00 0.00 0.13 0.15 2 Bulgaria 2.94 0.39 0.48 0.45 0.84 0.00 4.16 10.68 5.87 0.22 0.14 0.01 0.16 0.18 4 Croatia 3.42 0.37 0.42 0.23 0.87 -0.05 4.29 10.79 6.21 0.25 0.00 0.01 0.17 0.18 7 Cyprus 2.42 0.46 0.49 0.40 0.92 0.00 4.48 10.36 6.59 0.20 0.00 0.00 0.10 0.12 5 Czech Republic 3.65 0.35 0.60 0.12 0.90 0.00 4.41 11.29 7.42 0.29 0.00 0.00 0.12 0.13 1 Denmark 3.17 0.32 0.18 0.31 0.89 -0.01 4.60 11.50 6.13 0.28 0.08 0.00 0.15 0.17 26 Finland 3.16 0.23 0.17 0.09 0.95 0.00 4.58 11.40 6.77 0.09 0.00 0.11 0.10 0.15 2 France 3.34 0.38 0.54 0.18 0.93 0.00 4.54 12.42 8.48 0.41 0.00 0.00 0.13 0.14 9 Germany 3.27 0.33 0.63 0.20 0.91 -0.01 4.59 12.54 7.42 0.41 0.00 0.02 0.14 0.17 17 Greece 2.30 0.67 0.75 0.59 0.94 0.05 4.55 12.39 7.46 0.33 0.17 0.00 0.09 0.11 7 Hungary 3.05 0.42 0.09 0.23 0.90 0.00 4.31 11.12 7.10 0.10 0.00 0.00 0.15 0.16 2 Ireland 2.59 0.81 0.00 0.85 0.94 -0.06 4.63 11.36 8.20 0.12 0.00 0.00 0.11 0.13 2 Italy 3.04 0.35 0.22 0.32 0.90 0.00 4.52 12.32 7.32 0.23 0.00 0.02 0.11 0.14 23 Lithuania 2.93 0.31 0.20 0.58 0.87 0.00 4.29 10.59 5.78 0.10 1.00 0.06 0.14 0.14 1 Luxembourg 2.97 0.22 0.49 0.38 0.92 0.00 4.93 10.72 7.90 0.17 0.00 0.00 0.08 0.10 1 Malta 3.26 0.21 0.36 0.11 0.90 0.00 4.43 9.92 6.36 0.50 0.00 0.00 0.14 0.18 3 Netherlands 2.97 0.27 0.58 0.18 0.94 -0.01 4.62 11.90 7.28 0.50 0.00 0.00 0.14 0.16 3 Poland 3.43 0.36 0.53 0.18 0.89 0.05 4.29 11.67 7.08 0.15 0.03 0.00 0.13 0.14 11 Portugal 2.96 0.39 0.31 0.50 0.94 -0.02 4.40 11.36 7.65 0.11 0.00 0.00 0.09 0.11 5 Romania 3.03 0.40 0.35 0.43 0.90 0.00 4.20 11.24 6.56 0.26 0.28 0.00 0.12 0.14 3 Slovakia 3.70 0.33 0.91 0.13 0.91 0.00 4.37 10.95 7.00 0.15 0.67 0.00 0.12 0.12 2 Slovenia 2.15 0.42 0.48 0.77 0.91 0.01 4.44 10.67 6.37 0.10 0.00 0.00 0.08 0.10 2 Spain 3.18 0.42 0.16 0.45 0.93 0.00 4.50 12.15 8.34 0.13 0.00 0.00 0.09 0.11 7 Sweden 3.29 0.34 0.13 0.11 0.95 0.00 4.60 11.68 8.37 0.26 0.00 0.00 0.10 0.11 5 United Kingdom 3.19 0.39 0.19 0.35 0.90 0.00 4.55 12.40 8.04 0.28 0.00 0.05 0.13 0.17 11 Total 3.11 0.38 0.35 0.32 0.91 0.00 4.51 11.75 7.18 0.25 0.04 0.01 0.12 0.14 171 This table reports the averages of the main regression variables per country. The sample includes 171 publicly listed banks, from 26 countries in the European Union. Ln Z is the natural logarithm of the z-score, which is calculated as the return on assets plus the capital asset ratio, divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns, calculated using weekly data over the sample period.

CFR is the ownership concentration, measured as the cash flow rights of the largest shareholder of the bank. Loan loss provision is the ratio of the loan loss provision to the net interest revenue. Tobin’s Q is the market

value of the equity plus the book value of the liabilities, divided by the book value of the assets. Revenue growth is the growth in revenues over the last year. Economic development is the natural logarithm of the GDP per capita for each country. Economic size is the natural logarithm of the total GDP of each country. Size is the natural logarithm of the total assets. Liquidity is the ratio between of the liquid assets to the deposits and borrowings. Large owner on mgt board takes the value 1 if the largest shareholder has a seat on the management board and zero otherwise. Managerial ownership is the sum of the management’s total cash flow rights.

Appendix C: Additional regressions

Table 1C presents the regression results of the instrumental variable approach, and regression results with the use of lagged independent variables. Table 2C presents additional tests on ownership concentration. This includes a tests on the qualification of the large owner and two test on regarding the nonlinear relation between ownership concentration and risk taking. Table 3C presents results of regressions that test the relation of the Basel III capital standards without controlling for ownership concentration.

Table 1C: Additional regressions I Instrumental variable Instrumental variable Lagged independent variables Lagged independent variables (1) (2) (3) (4)

Dependent variable Ln (Z) Equity volatility Ln (Z) Equity volatility

Constant 3.149*** 0.363*** 3.268*** -0,095 (0.041) (0.025) (0.542) (0,322) CFR -0.097 0.040 -0.102** 0,046* (0.119) (0.070) (0.043) (0,026) Revenue growth 0.014 -0,044 (0.086) (0,051) Economic development -0.173 0,078 (0.125) (0,073) Economic size 0.041 -0,030** (0.026) (0,015)

Loan loss provision -0.143*** 0,041

(0.044) (0,026) Tobin's Q 0.125 0,016 (0.324) (0,192) Size 0.014 0,062*** (0.017) (0,010) Liquidity 0.072 -0,001 (0.046) (0,023) Adjusted R² 0.017 0.060 Observations 1082 1082 911 811

This table presents regression results of the instrumental variable approach and that of lagged variables. The sample includes 171 publicly listed banks, from 26 countries in the European Union. Ln Z is the natural logarithm of the z-score, which is calculated as the return on assets plus the capital asset ratio, divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns, calculated using weekly data over the sample period. CFR is the ownership concentration, measured as the cash flow rights of the largest shareholder of the bank. Loan loss provision is the ratio of the loan loss provision to the net interest revenue. Tobin’s Q is the market value of the equity plus the book value of the liabilities, divided by the book value of the assets. Revenue growth is the growth in revenues over the last year.

Economic development is the natural logarithm of the GDP per capita for each country. Economic size is the natural logarithm of the total

29 Table 2C: Additional regressions II

A 20% threshold for

qualification as large owner Possible nonlinearity Possible nonlinearity

(1) (2) (2) Dependent variable Ln (Z) Ln (Z) Ln (Z) Constant 4.358*** 4.193*** 4.279*** (0.465) (0.470) (0.464) CFR -0.065* 0.229* -0.174*** (0.037) (0.130) (0.052) (CFR)² -0.340** (0.144) CFR > Median 0.117*** (0.039) Revenue growth 0.108 0.101 0.108 (0.098) (0.098) (0.097) Economic development -0.201* -0.194* -0.228** (0.104) (0.104) (0.104) Economic size 0.085*** 0.095*** 0.103*** (0.022) (0.023) (0.023)

Loan loss provision -0.447*** -0.442*** -0.444***

(0.037) (0.037) (0.037) Tobin's Q -1.374*** -1.365*** -1.425*** (0.285) (0.284) (0.284) Size 0.015 0.013 0.014 (0.015) (0.015) (0.015) Liquidity -0.092*** -0.098*** -0.097*** (0.034) (0.034) (0.034)

Large owner on mgt board 0.008 0.010 -0.002

(0.060) (0.060) (0.060)

Managerial ownership -1.957*** -2.118*** -2.221***

(0.380) (0.385) (0.388) Adjusted R² 0.169 0.172 0.175 Observations 990 990 990

30 Table 3C: Additional regressions III

Tier 1 Tier 1 Total capital Total capital

Capital requirements Capital requirements (1) (2) (3) (4) (5) (6) Dependent variable Ln (Z) Equity volatility Ln (Z) Equity

volatility Ln (Z) Equity volatility

Constant 3.931*** 0.276 3.957*** 0.347 4.063*** 0.230 (0.480) (0.271) (0.470) (0.266) (0.482) (0.273) Revenue growth 0.114 -0.035 0.113 -0.036 0.112 -0.039 (0.098) (0.045) (0.097) (0.045) (0.097) (0.045) Economic development -0.198* -0.197*** -0.224** -0.195*** -0.234** -0.185*** (0.104) (0.059) (0.104) (0.059) (0.104) (0.060) Economic size 0.083*** 0.019 0.084*** 0.020 0.085*** 0.019 (0.022) (0.013) (0.022) (0.013) (0.022) (0.013)

Loan loss provision -0.431*** 0.272*** -0.430*** 0.270*** -0.433*** 0.272***

(0.037) (0.021) (0.037) (0.021) (0.037) (0.021) Tobin's Q -1.034*** 0.372* -1.024*** 0.297* -1.119*** 0.403** (0.319) (0.180) (0.305) (0.173) (0.320) (0.181) Size 0.018 0.043*** 0.019 0.042*** 0.020 0.042*** (0.015) (0.009) (0.015) (0.009) (0.015) (0.009) Liquidity -0.099*** -0.015 -0.101*** -0.015 -0.101*** -0.014 (0.034) (0.020) (0.034) (0.020) (0.034) (0.020) Large owner on mgt board 0.008 0.084** 0.019 0.085** 0.025 0.078** (0.060) (0.034) (0.060) (0.034) (0.060) (0.034) Managerial ownership -2.012*** -0.108 -2.071*** -0.105 -2.095*** -0.081 (0.379) (0.212) (0.379) (0.212) (0.380) (0.212) Tier 1 capital 0.698*** 0.209 -0.598 0.662* (0.264) (0.148) (0.601) (0.342)

Total capital ratio 0.996*** 0.104 1.588 -0.552

(0.290) (0.163) (0.662) (0.376) Adjusted R² 0.172 0.198 0.179 0.197 0.197 0.197 Observations 990 990 990 990 990 990 This table presents regression results regarding the effect of ownership concentration and capital requirements. The sample includes 171 publicly listed banks. from 26 countries in the European Union. Ln Z is the natural logarithm of the z-score. which is calculated as the return on assets plus the capital asset ratio. divided by the standard deviation of the asset returns. Equity volatility is the volatility of equity returns. calculated using weekly data over the sample period.. Loan loss provision is the ratio of the loan loss provision to the net interest revenue.

Tobin’s Q is the market value of the equity plus the book value of the liabilities. divided by the book value of the assets. Revenue growth is

the growth in revenues over the last year. Economic development is the natural logarithm of the GDP per capita for each country. Economic

size is the natural logarithm of the total GDP of each country. Size is the natural logarithm of the total assets. Liquidity is the ratio between of

31

References

Basel Committee on Banking Supervision, 2010a. Basel III: International framework for liquidity risk measurement, standards and monitoring. Bank For International Settlements, Basel, Switzerland.

Basel Committee on Banking Supervision, 2010b. An assessment of the long-term impact of stronger capital and liquidity requirements. Bank For International Settlements, Basel, Switzerland.

Bernanke, B., 1983. Nonmonetary effects of the financial crisis in the propagation of the Great Depression. American Economic Review 73, 257-276.

Blum, J., 1999. Do capital adequacy requirements reduce risks in banking? Journal of Banking & Finance 23, 755-771.

Buser, S., Chen, A., Kane, E., 1981. Federal deposit insurance, regulatory policy, and optimal bank capital. Journal of Finance 36, 51-60.

Calomiris, C., Mason, J., 1997. Contagion and bank failures during the Great Depression: the Chicago banking panic of June 1932. American Economic Review 87, 863-884.

Calomiris, C., Mason J., 2003a. Consequences of U.S. bank distress during the Depression. American Economic Review 93, 937-947.

Calomiris, C., Mason J., 2003b. Fundamentals, panics, and bank distress during the Depression. American Economic Review 93, 1615-1647.

Caprio, G., Laeven, L., Levine, R., 2007. Ownership and bank valuation. Journal of Financial Intermediation 16, 584–617.

Castro, R., Clementi, G., MacDonald, G., 2004. Investor protection, optimal incentives, and economic growth. Quarterly Journal of Economics 119, 1131-1175.

Chalermchatvichien, P., Jumreonwong, S., Jiraporn, P., 2014. Basel III, Ownership concentration, risk taking, and capital stability: evidence from Asia. Unpublished working paper. Bank of Thailand, Bangkok.

Demsetz, H., Lehn, K., 1985. The structure of corporate ownership: causes and consequences. Journal of Political Economy 93, 1155-1177.

32 Esty, B., 1998. The impact of contingent liability on commercial bank risk taking. Journal of Financial Economics 47, 189-218.

Furlong, F., Keely, M., 1989. Bank capital regulation and asset risk. Economic Review, Federal Reserve Bank of San Francisco Spring, 20–40.

Galai, D., Masulis, R., 1976. The option pricing model and the risk factor of stock. Journal of Financial Economics 3, 53-81.

Gonzalez, F., 2005. Bank regulation and risk-taking incentives: An international comparison of bank risk. Journal of Banking & Finance 29, 1153-1184.

Gorton, G., Rosen, R., 1995. Corporate control, portfolio choices, and the decline of banking. Journal of Finance 50, 1377-1420.

Houston, J., James, C., 1995. CEO Compensation and bank risk: Is compensation in banking structured to promote risk taking? Journal of Monetary Economics 36, 405-431.

Jensen, M., Meckling, W., 1976. Theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3, 305–360.

John, K., Litov, L., Yeung, B., 2008. Corporate governance and managerial risk taking: Theory and evidence. Journal of Finance 63, 167-1728.

John, K., Saunders, A., Senbet, L. W., 2000. A Theory of Bank Regulation and Management Compensation. Review of Financial Studies 13, 95-125.

Kane, E., 1985. The gathering crisis in federal deposit insurance. Cambridge, Mass: MIT Press.

Keely, M., Furlong, F., 1990. A reexamination of mean–variance analysis of bank capital regulation. Journal of Banking and Finance 14, 69–84.

Kim, D., Santomero, A., 1994. Risk in banking and capital regulation. Journal of Finance 43, 1219–1233.

Koehn, M., Santomero, A., 1980. Regulation of Bank Capital and Portfolio Risk. Journal of Finance 35, 1235-1244.

33 Merton, R., 1977. An analytical derivation of the cost of deposit insurance and loan guarantees: An application of modern option pricing theory. Journal of Banking and Finance 1, 3-11.

Saunders, A., Strock, E., Travlos, N., 1990. Ownership structure, deregulation, and bank risk taking. Journal of Finance 45, 643-654.

Shleifer, A., Vishny, R., 1986. Large shareholders and corporate control. Journal of Political Economy 94, 461–488.

Shleifer, A., Wolfenzon, D., 2002. Investor protection and equity markets. Journal of Financial Economics 66, 3-27.