WPS6289 Policy Research Working Paper 6289 How Does Deposit Insurance Affect Bank Risk? Evidence from the Recent Crisis Deniz Anginer Asli Demirguc-Kunt Min Zhu The World Bank Development Research Group Finance and Private Sector Development Team December 2012 Policy Research Working Paper 6289 Abstract Deposit insurance is widely offered in a number of the “moral hazard effect� of deposit insurance dominates countries as part of a financial system safety net to in good times while the “stabilization effect� of deposit promote stability. An unintended consequence of deposit insurance dominates in turbulent times. Nevertheless, insurance is the reduction in the incentive of depositors the overall effect of deposit insurance over the full sample to monitor banks, which leads to excessive risk-taking. remains negative since the destabilizing effect during This paper examines the relation between deposit normal times is greater in magnitude compared with the insurance and bank risk and systemic fragility in the years stabilizing effect during global turbulence. In addition, leading to and during the recent financial crisis. It finds the analysis finds that good bank supervision can alleviate that generous financial safety nets increase bank risk and the unintended consequences of deposit insurance on systemic fragility in the years leading up to the global bank systemic risk during good times, suggesting that financial crisis. However, during the crisis, bank risk is fostering the appropriate incentive framework is very lower and systemic stability is greater in countries with important for ensuring systemic stability. deposit insurance coverage. The findings suggest that This paper is a product of the Finance and Private Sector Development Team, Development Research Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at danginer@worldbank.org, ademirguckunt@worldbank.org, and mzhu1@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team How Does Deposit Insurance Affect Bank Risk? Evidence from the Recent Crisis Deniz Anginer, Asli Demirguc-Kunt, Min Zhu 1 JEL Classifications: G21, G28 Keywords: Bank risk, systemic risk, deposit insurance. 1 Anginer, World Bank, danginer@worldbank.org; Demirguc-Kunt, World Bank, ademirguckunt@worldbank.org; Zhu, World Bank, mzhu1@worldbank.org. This paper’s findings, interpretations and conclusions are entirely those of the authors and do not necessarily represent the views of the World Bank, their Executive Directors, or the countries they represent. 1. Introduction In response to the global financial crisis, a number of countries substantially increased the coverage of their financial safety nets in order to restore confidence and to avert potential contagious runs on their banking sectors. This has rekindled the debate on the impact of deposit insurance on banking sector stability and performance. While previous studies suggest that deposit insurance exacerbates moral hazard problems in bank lending and is associated with higher likelihood of banking crisis (Demirguc-Kunt and Detragiache 2002), to the best of our knowledge, there is no study that examines the impact of deposit insurance on bank risk and stability during a period of global financial instability. This is an important gap in our knowledge since economic theories suggest that deposit insurance brings both benefits and costs that are likely to vary with economic conditions. Hence while deposit insurance can increase moral hazard and make financial systems more vulnerable to crises during good times, it can also enhance depositor confidence and reduce the likelihood of contagious bank runs during turbulent periods such as the recent global financial crisis. The net effect of deposit insurance on bank risk and stability, therefore, depends on whether the benefits of deposit insurance can outweigh its costs. In this paper, we take advantage of the global financial crisis and study whether deposit insurance schemes have a stabilizing effect during financially turbulent periods. Specifically, using a sample of 4,109 publicly traded banks in 96 countries, we examine the impact of deposit insurance on bank risk and systemic stability separately for the crisis period from 2007 to 2009, as well as the three years from 2004 to 2006 leading up to the global financial crisis. We use z- score and stock return volatility to measure standalone risk of an individual bank, and the conditional value at risk measure (CoVar) of Adrian and Brunnermeier (2010) to measure the risk posed by an individual bank to the banking system as a whole. We find that generous financial safety nets increase bank risk and reduce systemic stability in non-crisis years. However, bank risk is lower and systemic stability is greater during the global financial crisis in countries with deposit insurance coverage. Nevertheless, the overall effect of deposit insurance over the full sample we study remains negative since the destabilizing effect during normal times is greater in magnitude compared to the stabilizing effect during global turbulence. Consistent with the prior literature, we also find that good supervision can enhance the positive effects of deposit insurance during turbulent periods and dampen the negative effects due to moral hazard 2 during normal times. Our findings offer new insights into the effect of deposit insurance on banking sector stability and have important policy implications for the design of banking section regulation and supervision. Starting with Merton (1977), a number of theoretical papers have studied the relationship between deposit insurance and banking sector stability. 2 Deposit insurance protects the interests of unsophisticated depositors and helps prevent bank runs which can improve social welfare. This positive stabilization effect of deposit insurance is, naturally, more important during economic downturns when contagious bank runs are more likely to occur. Consistent with this view, Gropp and Vesala (2004) show that the adoption of deposit insurance is associated with lower bank risk in the European Union. Chernykh and Cole (2011) also document that the adoption of deposit insurance in Russia is associated with better financial intermediation. Karels and McClatchey (1999) find stabilization effects from the adoption of deposit insurance for US credit unions. There is also considerable consensus in the literature that deposit insurance exacerbates moral hazard problems in the banking sector by incentivizing banks to take on excessive risk. Depositors can limit bank risk taking by charging higher interest rates. When deposits are insured, however, bank depositors lack incentives to monitor (Demirguc-Kunt and Huizinga 2004 and Ioannidou and Penas 2010). The lack of market discipline leads to excessive risk taking culminating in banking crises. Demirguc-Kunt and Detragiache (2002), Demirguc-Kunt and Kane (2002) and Barth, Caprio and Levine (2004) find supportive evidence for this view. Moreover, critics of deposit insurance argue that the government may lack the resources, information or incentives to correctly assess bank risk and charge deposit insurance premiums accordingly. Any risk based premium charged may be deemed “unfair� leading to distortions and inefficiencies in the banking sector. A number of papers have emphasized how design features of deposit insurance schemes and the larger institutional environment may affect the relationship between deposit insurance and banking risk and bank system fragility. In poor institutional settings, generous design features tend to destabilize the banking system and to undermine market discipline. Hovakimian, Kane and Laeven (2003) and Laeven (2002) show that weak institutional environments 2 See Morrison and White (2011) for recent development in the theoretical literature and references therein. The focus of the theoretical literature seems to be the design of the optimal deposit insurance scheme. 3 undermine deposit-insurance design. Cull, Senbet and Sorge (2004) produce evidence that in weak institutional environments deposit insurance reduces financial development. Features of a country’s private and public contracting environments have been shown to be important in deposit-insurance adoption and design (Demirgüç-Kunt and Kane, 2002). In this paper we test if the impact of deposit insurance is different during normal times and economic downturns. It is possible that the positive stabilization effects of deposit insurance can dominate the negative moral hazard effects during economic downturns. During downturns banks may face tightened funding and limited investment opportunities, leaving little room for excessive risk taking. Deposit insurance can then help enhance depositor confidence and prevent systemic bank runs during the crisis, leading to lower risk and greater systemic stability. During non-stress periods, however, the reverse could be true since there would be plenty of investment opportunities and little need for preventing bank runs. Alternatively, deposit insurance could also be a destabilizing (or stabilizing) influence throughout the whole sample period. For a sample of 4,109 banks in 96 countries over the time period 2004-2009, we find supportive evidence that the influence of deposit insurance on bank risk is different during normal periods versus global systemic downturns. Specially, we show that existing deposit insurance coverage is associated with lower systemic stability and higher bank risk in the 2004- 2006 pre-crisis period. However, the relationship between deposit insurance and bank risk and systemic stability is reversed in the 2007-2009 crisis period. Nevertheless, the overall effect of deposit insurance over the full sample we study remains negative since the destabilizing effect during normal times is greater in magnitude compared to the stabilizing effect during global turbulence. Our paper is related to Demirguc-Kunt and Detragiache (2002), which studies the link between deposit insurance and the occurrence of banking crisis. Unlike Demirguc-Kunt and Detragiache (2002), who focus on the likelihood of crisis of a particular country, we study the impact of deposit insurance on bank risk and system stability during a global systemic shock and compare it to normal times. Our paper is also related to Fahlenbrach, Prilmeier, and Stulz (2011) and Beltratti and Stulz (2012), who investigate the determinants of bank performance during the recent crisis. While their focus is on bank characteristics that affect performance, we focus on the effect of deposit insurance on bank risk and stability. We also add to the literature that has shown that features of a country’s private and public contracting environment are important in 4 deposit-insurance adoption, design and performance (see for instance, Demirgüç-Kunt and Kane 2002). We show that good supervision can affect the state-varying benefits and costs of deposit insurance. Strong supervision and regulation can enhance the stabilization effects during crisis periods while dampening the negative effects associated with moral hazard during normal times. While there is widespread agreement in the academic literature that deposit insurance affects bank risk through two channels, stabilization and moral hazard, our paper is the first to document that the net impact of deposit insurance varies with the relative importance of these two effects. Our results indicate that deposit insurance indeed enhanced depositor confidence and had positive stabilization effects during the recent global financial crisis. Our findings also emphasize the role of regulation and supervision in the effectiveness of deposit insurance schemes by maximizing the benefits of deposit insurance while minimizing costs associated with incentives and moral hazard. The rest of the paper is organized as follows. Section 2 describes the construction of the sample and variables. Section 3 presents the empirical results and discusses the implications. Section 4 concludes. 2. Data 2.1. Sample and Bank Level Variables Our sample consists of all publicly traded banks covered by the Bankscope database. We use stock market information from Compustat for international banks and stock market information from CRSP for U.S. banks. The Bankscope database reports detailed balance sheet and income statement information for both public and private banks and covers over 90% of the total banking assets in a given country. The Compustat database provides daily stock price information for both active and delisted companies, accounting for 98% of the global stock market capitalization. CRSP is the standard source for stock price information of U.S. companies. Our final sample consists of 4,109 banks in 96 countries over the time period 2004- 2009. Bank level variables are constructed from the Bankscope database. We follow Laeven and Levine (2009) and compute a traditional measure of bank risk, which is calculated as average bank return on assets (net income divided by total assets) plus bank equity to assets ratio, scaled by the standard deviation of return on assets over a five-year rolling window. Higher 5 z-score indicates lower bank risk. We use the natural logarithm of z-score in our regressions because the distribution of z-score is highly skewed. We also compute an additional measure of bank risk using market prices. In particular, we use bank stock return volatility to measure bank risk. Bank stock return volatility is calculated as the standard deviation of bank daily stock returns in a fiscal year. Higher stock return volatility indicates higher bank risk In addition, for each bank, each year, we calculate bank size (natural logarithm of total assets), leverage (liabilities divided by total assets), provisions (loan loss provisions divided by total assets), reliance on deposits for funding (deposits divided by total assets), and profitability (net income divided by total assets) as control variables. We winsorize all financial variables at the 1st and 99th percentile level of their distributions to reduce the influence of outliers and potential data errors. 2.2. Systemic Stability Measure The global financial crisis has led to a re-examination of risk assessment practices and regulation of the financial system, with a renewed interest in systemic fragility and macro- prudential regulation. This requires a focus not on the risk of individual financial institutions, but on an individual bank’s contribution to the risk of the financial system as a whole. Hence, there is a growing consensus that from a regulatory perspective of ensuring systemic stability, the correlation in the risk taking behavior of banks is much more relevant than the absolute level of risk taking in any individual institution. In addition to the bank level standalone measures of risk we also compute a measure of each bank’s contribution to the system as a whole. Our measure is based on the conditional value at risk measure (CoVar) of Adrian and Brunnermeier (2010). 3 As this approach requires high frequency data, we focus on publicly traded banks and use market prices to compute measures of bank default risk. First, we use the Merton (1974) contingent claim framework to measure standalone bank default risk. We then use the methodology in Adrian and Brunnermeier (2010) to measure co-dependence of bank default risk in a given country using quantile regressions. 3 Huang, Zhou, and Zhou (2009), Chan-Lau and Gravelle (2005), Avesani et al. (2006), and Elsinger and Lehar (2008) also use a portfolio credit risk approach to compute the contribution of an individual bank to the risk of a portfolio of banks. 6 Merton’s (1974) framework treats the equity value of a company as a call option on the company’s assets. The probability of default is computed using the “distance-to-default� measure, which is the difference between the asset value of the firm and the face value of its debt, scaled by the standard deviation of the firm’s asset value. The Merton (1974) distance-to- default measure has been shown to be good predictor of defaults outperforming accounting- based models (Campbell, Hilscher and Szilagyi (2008), Hillegeist, Keating, Cram, and Lundstedt (2004), and Bharath and Shumway (2008)). Although the Merton distance-to-default measure is more commonly used in bankruptcy prediction in the corporate sector, Merton (1977) points out the applicability of the contingent claims approach to pricing deposit insurance in the banking context. Bongini, Laeven and Majnoni (2002), Bartram, Brown and Hund (2007), Hovakimian, Kane, and Laeven (2012), among others, have used the Merton model to measure default risk of commercial banks. We follow Campbell, Hilscher and Szilagyi (2008) and Hillegeist, Keating, Cram, and Lundstedt (2004) to calculate Merton’s distance-to-default measure. Specifically, the market equity value of a company is modeled as a call option on the company’s assets: 𝑉𝐸 = 𝑉𝐴 𝑒 −𝑑𝑇 �(𝑑1 ) − 𝑋𝑒 −𝑟𝑇 �(𝑑2 ) + �1 − 𝑒 −𝑑𝑇 �𝑉𝐴 2 𝑉𝐴 𝑠𝐴 (1) log � 𝑋 � + �𝑟 − 𝑑 + 2 � 𝑇 𝑑1 = ; 𝑑2 = 𝑑1 − 𝑠𝐴 √𝑇 𝑠𝐴 √𝑇 In equation (1), VE is the market value of a bank. VA is the value of the bank’s assets. X is the face value of debt maturing at time T. r is the risk-free rate and d is the dividend rate expressed in terms of VA. sA is the volatility of the value of assets, which is related to equity volatility through the following equation: 𝑉𝐴 𝑒 −𝑑𝑇 �(𝑑1 )𝑠𝐴 𝑠𝐸 = (2) 𝑉𝐸 We simultaneously solve the above two equations to find the values of VA and sA. We use the market value of equity for VE and total liabilities to proxy for the face value of debt X. Since the accounting information is on an annual basis, we linearly interpolate the values for all dates 7 over the period, using beginning and end of year values for accounting items. The interpolation method has the advantage of producing a smooth implied asset value process and avoids jumps in the implied default probabilities at year end. sE is the standard deviation of daily equity returns over the past year. In calculating standard deviation, we require the bank to have at least 90 non-missing daily returns over the previous twelve months. T equals one year. r is the one year US treasury yield, which we take to be the risk free rate. We use the Newton method to simultaneously solve the two equations above. For starting values for the unknown variables, we use VA = VE + X and sA = sEVE/(VE+X). We winsorize sE and VE/(VE+X) at the 5th and 95th percentile levels to reduce the influence of outliers. After we determine asset values VA, we follow Campbell, Hilscher and Szilagyi (2008) and assign asset return m to be equal to the equity premium (6%). Merton’s distance-to-default (dd) is finally computed as: 2 𝑉𝐴 𝑠𝐴 𝑙𝑜𝑔 � 𝑋 � + �𝑚 − 𝑑 − 2 � 𝑇 (3) 𝑑𝑑 = 𝑠𝐴 √𝑇 The default probability is the normal transform of the distance-to-default measure and is defined as PD = F (–dd), where F is the cumulative distribution function of a standard normal distribution. We calculate a bank systemic risk measure to examine the correlation in the risk taking behavior of banks. Following Adrian and Brunnermeier (2010), we compute a conditional value at risk measure (CoVar) for each of the banks in our sample using quantile regressions. Quantile regressions estimate the functional relationship among variables at different quantiles (Koenker and Hallock (2001)) and allows for a more accurate estimation of the credit risk co-dependence during stress periods by taking into account nonlinear relationships when there is a large negative shock. As in Adrian and Brunnermeier (2010), we estimate a time series CoVar measure using a number of state variables. We run the following quantile regressions over the sample period: ∆𝐵𝑎𝑛𝑘𝐷𝐷𝑖,𝑡 =�𝑖 + 𝛾𝑖 𝑀𝑡−1 + 𝜀𝑖,𝑡 ∆𝑆𝑦𝑠𝑡𝑒𝑚𝐷𝐷𝑡 =�𝑠𝑦𝑠𝑡𝑒𝑚|𝑖 + 𝛽𝑠𝑦𝑠𝑡𝑒𝑚|𝑖 ∆𝐵𝑎𝑛𝑘𝐷𝐷𝑖,𝑡 + 𝛾𝑠𝑦𝑠𝑡𝑒𝑚|𝑖 𝑀𝑡−1 + 𝜀𝑠𝑦𝑠𝑡𝑒𝑚|𝑖,𝑡 (4) 8 In equation (4), ∆𝐵𝑎𝑛𝑘𝐷𝐷𝑖 ,𝑡 is the change in Merton distance-to-default variable for bank i in week t. ∆SystemDDt is similarly the change in the value-weighted Merton distance-to- default variable for all banks in a given country. 𝑀𝑡−1 are lagged state variables and include change in the term spread (TERM), change in the default spread (DEF), CBOE implied volatility index (VIX), S&P 500 return (SPRET) and the change in the 3 month t-bill rate (RATE). The ΔCoVar variable is then computed as the change in the VaR of the system when the institution is at the qth percentile (or when the institution is in distress) minus the VaR of the system when the institution is at the 50% percentile: 𝑞 ̂ 𝑞 ∆𝐶𝑜𝑣𝑎𝑟𝑆𝑦𝑠𝑡𝑒𝑚𝑡 = 𝛽 � 𝑞 � 50% 𝑠𝑦𝑠𝑡𝑒𝑚|𝑖 �∆𝐵𝑎𝑛𝑘𝐷𝐷𝚤,𝑡 − ∆𝐵𝑎𝑛𝑘𝐷𝐷𝚤,𝑡 � (5) We compute the ΔCoVar measure for each bank in our sample for the pre-crisis and post- crisis time periods separately, and higher ΔCoVar indicates lower systemic risk. 2.3. Country Level Variables Country level variables are collected from a number of sources. We obtain a deposit insurance variable from Barth, Caprio, and Levine (2008). Specifically, deposit insurance is equal to 1 if a country has explicit deposit insurance and depositors were fully compensated the last time a bank failed and 0 otherwise. We obtain a second deposit insurance variable from Demirguc-Kunt, Kane, and Laeven (2008). Full coverage is equal to 1 if a country offers full coverage and 0 otherwise. Since deposit insurance schemes are often changed during times of crises and we are interested in the impact of deposit insurance on bank risk during the recent pre- crisis and crisis periods, we take the deposit insurance coverage in existence in 2003 as our independent variable and use bank-year observations from 2004 to 2009 in our study to mitigate the reserve causality problem. Although fixing the deposit insurance measure at year 2003 leads to measurement error in the deposit insurance variable and may bias us against finding any significant results, it alleviates the concern that some unobservable country characteristics drive both changes in deposit insurance and bank risk. 4 4 It might be argued that the deposit insurance scheme in 2003 is positively correlated with government interventions in the recent crisis, which seems to favor our hypothesis that deposit insurance is associated with lower bank risk in the crisis. However, there is no evidence for such a claim and in contrast previous studies have suggested that lack 9 We also control for a number of country level variables to further deal with the potential omitted variable problem. Specifically, since both bank performance and deposit insurance can be affected by economic conditions in a country, we obtain economic development measures from the World Bank’s World Development Indictor (WDI) database. We use the natural logarithm of GDP per capita to measure the economic development of a country, the variance of GDP growth rate to measure economic stability, the natural logarithm of population to measure country size, and imports plus exports of goods and service divided GDP to measure global integration (Karolyi, Lee, and Van Dijk 2011). We also use stock market capitalization divided by GDP and private credit divided by GDP from the Financial Structure Dataset (Beck, Demirguc-Kunt, and Levine (2010)) to control for differences in financial development and structure. We define a crisis dummy to be equal to 1 for years 2007 to 2009 and 0 for years 2004 to 2006. A full list of variables and definitions are provided in Appendix I. 3. Empirical Methodology and Results 3.1. Summary Statistics Table 1 provides the summary statistics of variables used in this study. An average bank in the sample has a logzscore of 3.50, stock return volatility of 0.03 and systemic risk of -0.07. In addition, the average bank has log total asset value of 7.66, and leverage ratio of 0.90. These numbers are comparable to those in previous studies such as Anginer, Demirguc-Kunt, and Zhu (2012). Table 2 presents the sub-sample comparisons of means of our main variables of interest. In Panel A of Table 2, we partition our sample by whether a country offers deposit insurance and whether the country is in a crisis period. The results suggests that in the non-crisis period, banks in countries without deposit insurance have higher z-score, lower return volatility, and slightly lower systemic risk. Entering into crisis years substantially increases bank risk, stock return volatility and systemic risk. More importantly, compared to countries with deposit insurance coverage, countries without such coverage experienced larger increases in bank risk, stock return volatility and systemic risk. In Panel B of Table 2, we partition our sample by whether a country of deposit insurance may indicate high potential to the offering of generous financial safety nets (Gropp and Vesala 2004 and Demirguc-Kunt, Karacaovali, and Laeven 2005). 10 offers full coverage and whether the country is in a crisis year. The results are similar to those in Panel A of Table 2. Overall, the univariate results in Table 2 are consistent with our conjecture that the effect of deposit insurance on bank risk is time varying. We then conduct formal tests of our hypothesis using multivariate regressions. 3.2. Bank Risk We first examine whether the impact of deposit insurance on bank risk varies during crisis and pre-crisis years. While deposit insurance may lead to moral hazard and excessive bank risk taking (e.g., Demirguc-Kunt and Kane 2002 and Demirguc-Kunt and Huizinga 2004), it can also prevent bank runs (e.g., Merton 1977 and Matutes and Vives 1996) and thus ensure investor confidence and systemic stability during economic downturns. We use ordinary least squares (OLS) to estimate the following regression specification: logzscoreijt = β0 + Ω×bank and country controlsijt-1 + β1×deposit insuranceij2003×crisisijt + (7) β2×deposit insuranceij2003×noncrisisijt + β3×crisisijt + εijt Our dependent variable is bank i’s risk (in country j in year t), logzscoreijt, which is equal to bank average bank return on assets (net income divided by total assets) plus bank equity to assets ratio, scaled by the standard deviation of return on assets over a five-year rolling window. Our main explanatory variables of interest are the two interaction terms: one between deposit insurance coverage and crisis year dummy, and the other between deposit insurance coverage and non-crisis year dummy. This specification allows deposit insurance to have different effects on bank risk in crisis and non-crisis periods. We also control for bank characteristics that can affect risk. Our bank level control variables include log value of bank assets, leverage, provisions to net interest income ratio, reliance on deposits for funding, and profitability. We also include country level control variables such as natural logarithm of GDP per capita, variance of GDP growth rate, natural logarithm of population, imports plus exports of goods and service divided GDP, stock market capitalization divided by GDP, private credit divided by GDP, and crisis year dummy. All explanatory variables are lagged by one year to alleviate any reverse causality problem. 11 Table 3 presents the coefficient estimates. Column (1) of Table 3 shows that deposit insurance has a positive and statistically significant effect on bank zscore in crisis years and a negative and statistically significant effect on bank zscore in pre-crisis years. This suggests that deposit insurance reduces bank risk during the recent crisis and increases bank risk in the years leading up to the crisis. However, the average effect of deposit insurance over the full sample period is still negative, since the destabilizing effect during normal times is greater in magnitude compared to the stabilizing effect during the global financial crisis. In terms of control variables, we find that banks with higher leverage ratios have higher risk and banks that rely more on deposits have lower risk. Higher loan loss provisions seem to increase bank risk while greater profitability tends to be associated with lower risk. For country level variables, we find that bank risk is positively correlated with variance of GDP growth and is higher in crisis years, and bank risk is negatively correlated with private credit offered by financial institutions. Last, bank risk is substantially higher in crisis years. In addition, we consider an alternative regression specification that allows us to estimate an incremental effect of deposit insurance on bank risk in crisis years directly. 5 Specifically, we estimate the following OLS regression: logzscoreijt = β0 + Ω×bank and country controlsijt-1 + β1×deposit insuranceij2003×crisisijt + β2×deposit insuranceij2003 + β3×crisisijt + εijt (8) Column (2) of Table 3 presents the results. Our dependent variable is still bank i’s risk (in country j in year t) logzscoreijt. Our main explanatory variables of interest are deposit insurance and the interaction term between deposit insurance coverage and crisis year dummy. The control variables are similar to those we used in Column (1). We find that the main effect of deposit insurance is negatively and statistically significant at the 1% level. This indicates that, on average, deposit insurance exacerbates moral hazard and leads to excess risk taking in non-crisis years. The interaction term between deposit insurance and crisis year is positive and statistically significant at the 1% level. Moreover, the economic magnitude of the incremental effect is 5 Note that this specification is identical to the previous one except that it offers the magnitude and significance of the incremental effect of deposit insurance directly. 12 greater than that of the main effect, suggesting that deposit insurance enhances investor confidence and reduces bank risk during a crisis. As a robustness check, we also use full coverage dummy to replace the deposit insurance coverage dummy as a measure of deposit insurance coverage. The regression results are presented in Columns (3) and (4) of Table 3, and they are consistent with those in Columns (1) and (2) of Table 3. Overall, the evidence in Table 3 suggests that deposit insurance increases bank risk in pre-crisis years and decreases bank risk in crisis years, with an average negative effect for the entire sample period. In addition, Table 3 shows that the full coverage dummy has a slightly higher time varying impact on bank risk, consistent with previous studies which suggest that full coverage may further exacerbate the moral hazard problem (Demirguc-Kunt and Detragiache 2002). 3.3. Bank Stock Return Volatility Next, we examine the impact of deposit insurance on bank stock return volatility. Our dependent variable is bank i’s volatility (in country j in year t), which is calculated as the standard deviation of daily stock returns in a fiscal year. The regression specifications and control variables are the same as those used in Section 3.2 and the regression results are reported in Table 4. We find that, in Column (1) of Table 4, deposit insurance dummy is associated with lower stock return volatility in crisis years but is correlated with higher stock return volatility in pre-crisis years. Column (2) of Table 4 indicates that, deposit insurance increases bank stock return volatility in non-crisis years but the positive relationship between deposit insurance and stock return volatility is completely reversed during a crisis. In terms of control variables, we find that banks with larger size, higher reliance on deposits as funding, lower loan loss provisions, and higher profitability seem to have lower stock return volatility. For country level variables, we find that countries with higher GDP per capita, lower variance of GDP growth, smaller population, and lower private credits have banks with lower stock return volatility. In addition, stock return volatility is substantially higher in crisis years. In Columns (3) and (4) of Table 4, we replace the deposit insurance dummy with a full coverage dummy. The results are consistent with those in Columns (1) and (2) of Table 4 and are 13 again consistent with our conjecture that deposit insurance increases bank risk in normal years but reduces bank risk in crisis years. 3.4. Bank Systemic Risk From a regulatory perspective, there is a growing consensus that the correlation in the risk taking behavior of banks is much more relevant than the absolute level of risk taking in any individual institution. We thus examine the relationship between deposit insurance and bank systemic risk. Acharya (2009) suggests that if there is an implicit guarantee provided by the State to cover losses stemming from a systemic crisis, banks will have incentives to take on correlated risks. Guaranteed banks will not have incentives to diversify their operations, since the guarantee takes effect only if other banks fail as well. However, Wagster (2007) finds that the adoption of deposit insurance in Canada seems to reduce banking sector systemic risk. In our test, our dependent variable is bank i’s systemic risk (in country j in year t), whose calculation is described in Section 2.2. The regression specifications and control variables are the same as those used in Section 3.2. Table 5 presents the regression results. The results in Column (1) of Table 5 indicates that deposit insurance dummy is associated with lower bank systemic risk in crisis years but higher bank systemic risk in non-crisis years, and the overall effect of deposit insurance over the entire sample period is negative. Column (2) of Table 5 shows that deposit insurance increases bank systemic risk in non-crisis years but the positive effect of deposit insurance on bank systemic risk is reversed in crisis years. In terms of control variables, we find that banks with larger size, lower leverage, higher reliance on deposits as funding, and lower loan loss provisions are associated with higher systemic risk. For country level variables, we find that countries with higher GDP per capita, smaller population, and lower international trade have banks with higher systemic risk. Not surprisingly, bank systemic risk is also higher in crisis years. In Columns (3) and (4) of Table 5, we replace the deposit insurance dummy with a full coverage dummy in our regression analysis of systemic risk. The results are consistent with those in Columns (3) and (4) of Table 5, and are again supportive of our hypothesis. 3.5. Bank Supervision, Deposit Insurance, and Systemic Risk 14 We are also interested in the impact of the regulatory and institutional framework on the deposit insurance and systemic risk relationship. The adverse consequence of deposit insurance may well depend on the institutional environment and can potentially be mitigated through bank regulation. For instance, better bank supervision may limit the extent to which banks can engage in correlated risk taking activities in the presence of deposit insurance. We consider a bank supervisory quality index, which measures whether the supervisory authorities have the power and the authority to take specific preventive and corrective actions such as replacing the management team. This variable comes from the banking surveys conducted by Barth, Caprio, and Levine (2008). The surveys were conducted in the years 1999, 2002, and 2005. Because country level regulations change slowly over time, we use the previously available survey data until a new survey becomes available. Our dependent variable is bank i’s systemic risk, whose calculation is described in Section 2.3. The regression specifications and control variables are the same as those used in Section 3.2. Because the full coverage dummy is expected to lead to the most egregious moral hazard problem during good times, we examine whether bank supervision can mitigate such an unintended consequence of full deposit insurance coverage. Table 6 reports the coefficient estimates. We first partition our sample by the bank supervisory quality index. Column (1) of Table 6 shows the regression results for the subsample where bank supervisory quality is greater than or equal to sample median, and Column (2) of Table 6 presents that the regressions coefficients for the subsample where bank supervisory quality is lower than sample median. We find that, in Column (1) of Table 6, full coverage dummy is associated with lower bank systemic risk during global crisis years and has no adverse impact on bank systemic risk in non-crisis years, when bank supervision is strong. In contrast, Column (2) of Table 6 indicates that, full coverage increases bank systemic risk in all years when bank supervision is weak. Overall, the results in Table 6 suggest that better bank supervision may be able to mitigate the adverse consequences of deposit insurance while weak bank supervision exacerbates the negative impact of deposit insurance on bank stability. 4. Conclusion As part of the International Monetary Fund’s best practice recommendations to developing countries, deposit insurance has received increased attention from both academic 15 researchers and public policy makers. While deposit insurance is aimed at ensuring depositor confidence and to prevent bank runs, it comes with an unintended consequence of encouraging banks to take on excessive risk. In this paper, we study the relation between deposit insurance and bank risk and system fragility during the global financial crisis and the period preceding it. We show that generous financial safety nets increase bank risk and systemic fragility in the years leading up to the crisis. However, both standalone bank risk and systemic risk are lower during the global financial crisis in countries with deposit insurance coverage. Our results suggest that deposit insurance offered significant stabilization effects during the recent banking crisis. Nevertheless, the overall impact of deposit insurance on bank risk over the full sample period remains negative, since the stabilization effect during the crisis period tends to be smaller than the destabilizing effect of the insurance in the period leading up to the crisis. We also find that good bank supervision can alleviate the adverse consequences of deposit insurance on systemic risk in good times. Our results stress the importance of the underlying regulatory and institutional framework and lend support to the view that fostering the appropriate incentive framework is very important for ensuring systemic stability. 16 References Acharya, V. 2009. A theory of systemic risk and design of prudential bank regulation. Journal of Financial Stability, 5(3), 224-255. Adrian, T., and M. K. Brunnermeier. 2010. CoVaR. Working Paper, Princeton University. Anginer, D., A. Demirguc-Kunt, and M. Zhu. 2012. How does bank competition affect systemic stability? World Bank Policy Research Working Paper No. 5981. Avesani, R., A. G. Pascual, and J. Li. 2006. A new risk indicator and stress testing tool: A multifactor Nth-to-default CDS basket. IMF Working Paper. Barth, J. R., G. Caprio Jr, and R. Levine. 2004. Bank regulation and supervision: what works best? Journal of Financial Intermediation 13 (2):205-248. Barth, J., G. Caprio, and R. Levine. 2008. Bank regulation and supervision database. The World Bank. Bartram, S., B. Gregory, and J. Hund. 2007. Estimating systemic risk in the international financial system. Journal of Financial Economics 86(3), 835-869. Beck, T., A. Demirguc-Kunt, and R. Levine. 2010. Financial institutions and markets across countries and over time: The updated financial development and structure database. World Bank Economic Review 24 (1):77-92. Beltratti , A., and R. M. Stulz. 2011. The credit crisis around the globe: Why did some banks perform better? Journal of Financial Economics, forthcoming. Bharath, S. T., and T. Shumway. 2008. Forecasting default with the Merton distance to default model. Review of Financial Studies 21(3), 1339-1369. Bongini, P., L. Laeven, and G. Majnoni. 2002. How good is the market at assessing bank fragility? A horse race between different indicators. Journal of Banking and Finance (26)5: 1011-1028. Campbell, J.Y., J. Hilscher, and J. Szilagyi. 2008. In search of distress risk. Journal of Finance 63, 2899-2939. Chan-Lau, J.A., and T. Gravelle. 2005. The END: A new indicator of financial and nonfinancial corporate sector vulnerability. IMF. Chernykh, L., and R. A. Cole. 2011. Does deposit insurance improve financial intermediation? Evidence from the Russian experiment. Journal of Banking and Finance 35 (2):388-402. 17 Cull, R., M. Sorge, and L. Senbet. 2004. Deposit insurance and bank intermediation in the long run. BIS Working Paper #156. Demirguc-Kunt, A., and E. Detragiache. 2002. Does deposit insurance increase banking system stability? An empirical investigation. Journal of Monetary Economics 49 (7):1373-1406. Demirguc-Kunt, A., and H. Huizinga. 2004. Market discipline and deposit insurance. Journal of Monetary Economics 51(2), 375-399. Demirguc-Kunt, A., and E. Kane. 2002. Deposit insurance around the globe: Where does it work? Journal of Economic Perspectives 16(2), 175-195. Demirguc-Kunt, A., E.J. Kane, and L. Laeven. 2008. Determinants of deposit-insurance adoption and design. Journal of Financial Intermediation 17(3), 407-438. Demirguc-Kunt, A., B. Karacaovali, and L. Laeven, 2005. Deposit insurance around the world: A comprehensive database. World Bank Policy Research Working Paper No. 3628. Elsinger, H., A. Lehar, and M. Summer. 2005. Using market information for banking system risk assessment. Working Paper, Austrian National Bank. Fahlenbrach, R., R. Prilmeier, and R. M. Stulz. 2011. This time is the same: Using bank performance in 1998 to explain bank performance during the recent financial crisis. NBER Working Paper 17038. Gropp, R. and J. Vesala, 2004. Deposit insurance, moral hazard and market monitoring. European Central Bank Working Paper No. 302. Hillegeist, S., E. Keating, D. Cram, and K. Lundstedt. 2004. Assessing the probability of bankruptcy. Review of Accounting Studies 9, 5-34. Hovakimian, A., E. J. Kane, and L. Laeven. 2012. Variation in systemic risk at US banks during 1974-2010. NBER Working Paper 18043. Huang, X., H. Zhou, and H. Zhu. 2009. A framework for assessing the systemic risk of major financial institutions. Journal of Banking and Finance 33, 2036-2049. Ioannidou, V. P., and M. F. Penas. 2010. Deposit insurance and bank risk-taking: Evidence from internal loan ratings. Journal of Financial Intermediation 19 (1):95-115. Karels, G. V., and C. A. McClatchey. 1999. Deposit insurance and risk-taking behavior in the credit union industry. Journal of Banking and Finance 23 (1):105-134. Karolyi, G.A., K.-H. Lee, and M.A. Van Dijk. 2011. Understanding commonality in liquidity around the world. Journal of Financial Economics, forthcoming. 18 Koenker, R. and K. Hallock. 2001. Quantile regression. Journal of Economic Perspectives 15, 143-156. Laeven, L., and R. Levine. 2009. Bank governance, regulation and risk taking. Journal of Financial Economics 93 (2):259-275. Merton, R. C.. 1974. On the pricing of corporate debt: the risk structure of interest rates. Journal of Finance 29, 449–470. Merton, R.C.. 1977. On the pricing of contingent claims and the Modigliani-Miller theorem. Journal of Financial Economics 15, No. 2, 241-250. Morrison, A. D., and L. White. 2011. Deposit insurance and subsidized recapitalizations. Journal of Banking and Finance 35 (12):3400-3416. Wagster, J. D. 2007. Wealth and Risk Effects of Adopting Deposit Insurance in Canada: Evidence of Risk Shifting by Banks and Trust Companies. Journal of Money, Credit and Banking 39 (7):1651-1681. 19 Table 1: Summary statistics The sample consists of 4,109 banks in 96 countries over the time period 2004-2009. Definitions of variables are in Appendix I. Variables N P25 Mean Median P75 STD Log (zscore) 14,664 2.854 3.500 3.598 4.256 1.076 Volatility 5,473 0.016 0.028 0.022 0.034 0.019 ΔCoVar 4,389 -0.119 -0.069 -0.042 -0.008 0.079 Year 14,664 2,005 2,007 2,007 2,008 2 Log (Total Assets) 14,664 6.231 7.655 7.155 8.901 1.952 Leverage 14,664 0.889 0.899 0.913 0.933 0.074 Deposits 14,664 0.669 0.725 0.781 0.846 0.184 Provisions 14,664 0.024 0.124 0.064 0.155 0.199 ROA 14,664 0.005 0.010 0.009 0.014 0.014 Log (GDP Per Capita) 14,664 10.003 9.931 10.491 10.554 1.067 GDP Growth Volatility 14,664 0.852 1.345 1.256 1.587 0.951 Log (Population) 14,664 17.907 18.466 19.487 19.505 1.500 Trade / GDP 14,664 25.220 53.536 28.849 60.506 51.215 Stock Market Cap / GDP 14,664 0.831 1.146 1.322 1.397 0.581 Private Credit / GDP 14,664 0.984 1.433 1.733 1.884 0.620 Crisis Period Dummy 14,664 0.000 0.504 1.000 1.000 0.500 Deposit Insurance Dummy 14,634 0.000 0.324 0.000 1.000 0.468 Full Coverage Dummy 13,648 0.000 0.023 0.000 0.000 0.151 Supervisory Quality 14,632 11.000 11.872 13.000 13.000 2.041 20 Table 2: Sub-sample comparison of variable means The sample consists of 4,109 banks in 96 countries over the time period 2004-2009. Definitions of variables are in Appendix I. Panel A Crisis Period Dummy Deposit Insurance Dummy 0 1 Diff in mean p-value 0 Log (zscore) 3.9580 3.0472 0.9107 0.0000 Volatility 0.0186 0.0202 -0.0016 0.0009 ΔCoVar -0.0106 -0.1118 0.1013 0.0000 1 Log (zscore) 3.3349 3.2774 0.0574 0.0312 Volatility 0.0412 0.0276 0.0135 0.0000 ΔCoVar -0.0993 -0.1012 0.0018 0.6234 Panel B Crisis Period Dummy Full Coverage Dummy 0 1 Diff in mean p-value 0 Log (zscore) 3.7611 2.4710 1.2901 0.0000 Volatility 0.0184 0.0295 -0.0111 0.0000 ΔCoVar -0.0290 -0.0523 0.0233 0.0001 1 Log (zscore) 3.3148 3.1170 0.1978 0.0218 Volatility 0.0369 0.0303 0.0067 0.0020 ΔCoVar -0.1026 -0.0468 -0.0559 0.0000 21 Table 3: Deposit insurance, crisis, and bank risk The sample consists of 4,109 banks in 96 countries over the time period 2004-2009. The dependent variable is the log value of bank zscore. Standard errors are reported in parentheses below their coefficient estimates and adjusted for both heteroskedasticity and within correlation clustered at the bank level. *** (**) (*) indicates significance at 1% (5%) (10%) two tailed level, respectively. Variables (1) (2) (3) (4) Log (Total Assets) 0.002 0.002 0.002 0.002 (0.009) (0.009) (0.009) (0.009) Leverage -0.762*** -0.762*** -0.512* -0.512* (0.273) (0.273) (0.304) (0.304) Deposits 0.694*** 0.694*** 0.707*** 0.707*** (0.091) (0.091) (0.097) (0.097) Provisions -1.320*** -1.320*** -1.483*** -1.483*** (0.067) (0.067) (0.072) (0.072) ROA 2.888** 2.888** 4.889*** 4.889*** (1.255) (1.255) (1.408) (1.408) Log (GDP Per Capita) 0.030 0.030 -0.007 -0.007 (0.026) (0.026) (0.030) (0.030) GDP Growth Volatility -0.152*** -0.152*** -0.144*** -0.144*** (0.016) (0.016) (0.018) (0.018) Log (Population) -0.009 -0.009 -0.025 -0.025 (0.016) (0.016) (0.024) (0.024) Trade / GDP 0.000 0.000 -0.002 -0.002 (0.000) (0.000) (0.001) (0.001) Stock Market Cap / GDP 0.004 0.004 0.041 0.041 (0.036) (0.036) (0.049) (0.049) Private Credit / GDP 0.226*** 0.226*** 0.262*** 0.262*** (0.052) (0.052) (0.057) (0.057) Crisis Period Dummy -0.555*** -0.555*** -0.365*** -0.365*** (0.024) (0.024) (0.023) (0.023) Deposit Insurance Dummy × Crisis Period Dummy 0.221*** 0.693*** (0.042) (0.044) Deposit Insurance Dummy × Pre-Crisis Period Dummy -0.472*** (0.049) Deposit Insurance Dummy -0.472*** (0.049) Full Coverage Dummy × Crisis Period Dummy 0.245** 0.960*** (0.103) (0.136) Full Coverage Dummy × Pre-Crisis Period Dummy -0.716*** (0.150) Full Coverage Dummy -0.716*** (0.150) Constant 3.837*** 3.837*** 4.138*** 4.138*** (0.456) (0.456) (0.665) (0.665) Observations 14,634 14,634 13,648 13,648 R-squared 0.227 0.227 0.221 0.221 22 Table 4: Deposit insurance, crisis, and bank stock return volatility The sample consists of 4,109 banks in 96 countries over the time period 2004-2009. The dependent variable is bank stock return volatility. Standard errors are reported in parentheses below their coefficient estimates and adjusted for both heteroskedasticity and within correlation clustered at the bank level. *** (**) (*) indicates significance at 1% (5%) (10%) two tailed level, respectively. Variables (1) (2) (3) (4) Log (Total Assets) -0.001*** -0.001*** -0.001*** -0.001*** (0.000) (0.000) (0.000) (0.000) Leverage -0.001 -0.001 -0.002 -0.002 (0.005) (0.005) (0.006) (0.006) Deposits -0.008*** -0.008*** -0.007*** -0.007*** (0.002) (0.002) (0.002) (0.002) Provisions 0.021*** 0.021*** 0.025*** 0.025*** (0.002) (0.002) (0.002) (0.002) ROA -0.146*** -0.146*** -0.199*** -0.199*** (0.028) (0.028) (0.032) (0.032) Log (GDP Per Capita) -0.002*** -0.002*** -0.002*** -0.002*** (0.000) (0.000) (0.000) (0.000) GDP Growth Volatility 0.004*** 0.004*** 0.003*** 0.003*** (0.000) (0.000) (0.000) (0.000) Log (Population) 0.001*** 0.001*** 0.001*** 0.001*** (0.000) (0.000) (0.000) (0.000) Trade / GDP -0.000 -0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) Stock Market Cap / GDP -0.000 -0.000 -0.002** -0.002** (0.001) (0.001) (0.001) (0.001) Private Credit / GDP 0.007*** 0.007*** 0.009*** 0.009*** (0.001) (0.001) (0.001) (0.001) Crisis Period Dummy 0.020*** 0.020*** 0.018*** 0.018*** (0.001) (0.001) (0.001) (0.001) Deposit Insurance Dummy × Crisis Period Dummy -0.004*** -0.011*** (0.001) (0.001) Deposit Insurance Dummy × Pre-Crisis Period Dummy 0.006*** (0.001) Deposit Insurance Dummy 0.006*** (0.001) Full Coverage Dummy × Crisis Period Dummy -0.005*** -0.016*** (0.002) (0.002) Full Coverage Dummy × Pre-Crisis Period Dummy 0.012*** (0.003) Full Coverage Dummy 0.012*** (0.003) Constant 0.008 0.008 0.014 0.014 (0.008) (0.008) (0.013) (0.013) Observations 5,469 5,469 5,090 5,090 R-squared 0.347 0.347 0.368 0.368 23 Table 5: Deposit insurance, crisis, and bank systemic risk The sample consists of 4,109 banks in 96 countries over the time period 2004-2009. The dependent variable is bank systemic risk, ΔCoVar. Standard errors are reported in parentheses below their coefficient estimates and adjusted for both heteroskedasticity and within correlation clustered at the bank level. *** (**) (*) indicates significance at 1% (5%) (10%) two tailed level, respectively. Variables (1) (2) (3) (4) Log (Total Assets) -0.016*** -0.016*** -0.017*** -0.017*** (0.001) (0.001) (0.001) (0.001) Leverage 0.061*** 0.061*** 0.067** 0.067** (0.023) (0.023) (0.027) (0.027) Deposits -0.041*** -0.041*** -0.046*** -0.046*** (0.011) (0.011) (0.011) (0.011) Provisions 0.020*** 0.020*** 0.012* 0.012* (0.006) (0.006) (0.007) (0.007) ROA 0.095 0.095 0.136 0.136 (0.116) (0.116) (0.132) (0.132) Log (GDP Per Capita) -0.005* -0.005* 0.001 0.001 (0.003) (0.003) (0.004) (0.004) GDP Growth Volatility 0.003 0.003 0.014*** 0.014*** (0.002) (0.002) (0.002) (0.002) Log (Population) 0.010*** 0.010*** 0.017*** 0.017*** (0.002) (0.002) (0.004) (0.004) Trade / GDP 0.000*** 0.000*** 0.001*** 0.001*** (0.000) (0.000) (0.000) (0.000) Stock Market Cap / GDP 0.007 0.007 0.017** 0.017** (0.005) (0.005) (0.007) (0.007) Private Credit / GDP -0.002 -0.002 -0.014* -0.014* (0.009) (0.009) (0.008) (0.008) Crisis Period Dummy -0.077*** -0.077*** -0.067*** -0.067*** (0.003) (0.003) (0.003) (0.003) Deposit Insurance Dummy × Crisis Period Dummy 0.032*** 0.082*** (0.007) (0.007) Deposit Insurance Dummy × Pre-Crisis Period Dummy -0.051*** (0.008) Deposit Insurance Dummy -0.051*** (0.008) Full Coverage Dummy × Crisis Period Dummy 0.035*** 0.096*** (0.009) (0.009) Full Coverage Dummy × Pre-Crisis Period Dummy -0.061*** (0.013) Full Coverage Dummy -0.061*** (0.013) Constant -0.077 -0.077 -0.293*** -0.293*** (0.057) (0.057) (0.106) (0.106) Observations 4,389 4,389 4,036 4,036 R-squared 0.376 0.376 0.376 0.376 24 Table 6: Deposit insurance, crisis, bank supervision, and systemic risk The sample consists of 4,109 banks in 96 countries over the time period 2004-2009. The dependent variable is bank systemic risk, ΔCoVar. Standard errors are reported in parentheses below their coefficient estimates and adjusted for both heteroskedasticity and within correlation clustered at the bank level. *** (**) (*) indicates significance at 1% (5%) (10%) two tailed level, respectively. (1) (2) Supervisory Quality Supervisory Quality Variables >= Sample Median < Sample Median Log (Total Assets) -0.013*** -0.017*** (0.001) (0.003) Leverage 0.068* -0.001 (0.039) (0.043) Deposits -0.027*** 0.008 (0.010) (0.020) Provisions 0.022*** 0.039*** (0.008) (0.011) ROA -0.265 0.334** (0.220) (0.147) Log (GDP Per Capita) -0.053*** 0.016*** (0.012) (0.005) GDP Growth Volatility 0.016*** 0.012*** (0.003) (0.003) Log (Population) -0.023** 0.020*** (0.011) (0.004) Trade / GDP -0.002*** 0.001*** (0.001) (0.000) Stock Market Cap / GDP 0.043*** -0.012 (0.015) (0.008) Private Credit / GDP 0.068*** -0.054*** (0.026) (0.014) Crisis Period Dummy -0.099*** 0.006 (0.004) (0.006) Supervisory Quality Full Coverage Dummy × Crisis Period Dummy 0.114*** -0.060*** (0.015) (0.015) Full Coverage Dummy × Pre-Crisis Period Dummy -0.006 -0.042** (0.020) (0.017) Constant 0.886*** -0.500*** (0.299) (0.123) Observations 2,819 1,217 R-squared 0.490 0.288 25 Appendix I: Variable definitions Variables Definitions Dependent variables Log value of zscore, where bank zscore is calculated as the sum of ROA and equity ratio (ratio of book equity to total assets), averaged over the past five years, divided by the Log(z-score) standard deviation of ROA over the past five years. We require a bank to have at least four years of data in the past five years. Stock return volatility, which is the standard deviation of Volatility fiscal year daily stock return. Measure of bank systemic risk, which is the change in the VaR of the system when the institution is at the 5% percentile ΔCoVar minus the VaR of the system when the institution is at the 50% percentile. Bank level control variables Log (Total Assets) Log value of total assets in millions of US dollars. Leverage Total liabilities divided by total assets. Provisions Loan loss provisions divided by net interest income. Deposits Total deposits divided by total assets. ROA Net income divided by total assets. Country level variables Deposit insurance A variable indicates whether a country has explicit deposit insurance (Yes=1/No=0) and whether depositors were fully Deposit Insurance Dummy compensated the last time a bank failed (Yes=1/No=0). The variable is equal to 1 if both are true in 2003 and 0 otherwise. A variable that equals 1 if a country offers full insurance Full Coverage Dummy coverage in 2003 and 0 otherwise. Country level control variables A dummy variable that equals 1 for years 2007 to 2009 and 0 Crisis Period Dummy for years 2004 through 2006. A dummy variable that equals 0 for years 2007 to 2009 and 1 Pre-Crisis Period Dummy for years 2004 through 2006. A variable that ranges from zero to fourteen, with fourteen indicating the highest power of the supervisory authorities. For each of the following fourteen questions, a value of 1 is added to the index if the answer is yes: 1. Does the supervisory agency have the right to meet with external auditors to discuss their report without the approval of the bank? 2. Are auditors required by law to communicate directly to the supervisory agency any presumed involvement Supervisory Quality of bank directors or senior managers in illicit activities, fraud, or insider abuse? 3. Can supervisors take legal action against external auditors for negligence? 4. Can the supervisory authority force a bank to change its internal organizational structure? 5. Are off-balance sheet items disclosed to supervisors? 6. Can the supervisory agency order the bank's directors or management to constitute provisions to cover actual or potential losses? 7. Can the supervisory agency suspend the directors' decision to distribute: a) dividends? b) 26 bonuses? c) management fees? 8. Can the supervisory agency legally declare-such that this declaration supersedes the rights of bank shareholders-that a bank is insolvent? 9. Does the banking Law give authority to the supervisory agency to intervene that is, suspend some or all ownership rights-a problem bank? 10. Regarding bank restructuring and reorganization, can the supervisory agency or any other government agency do the following: a) supersede shareholder rights? b) remove and replace management? c) remove and replace directors? Log value of GDP per capital in nominal constant US 2000 Log (GDP Per Capita) dollars. GDP Growth Volatility Variance of GDP growth for the previous five years. Log (Population) Log value of population in millions. Imports plus exports of goods and services as a percentage of Trade / GDP GDP. Stock Market Cap / GDP Stock market capitalization divided by GDP. Private credit by deposit money banks and other financial Private Credit / GDP institutions to GDP. 27