WPS3954 Banking on the Principles: Compliance with Basel Core Principles and Bank Soundness by Asli Demirgüç-Kunt, Enrica Detragiache, and Thierry Tressel* Abstract This paper studies whether compliance with the Basel Core Principles for Effective Banking Supervision (BCPs) improves bank soundness. BCP compliance assessments provide a unique source of information about the quality of bank supervision and regulation around the world. We find a significant and positive relationship between bank soundness (measured with Moody's financial strength ratings) and compliance with principles related to information provision. Specifically, countries which require banks to report regularly and accurately their financial data to regulators and market participants have sounder banks. This relationship is robust to controlling for broad indexes of institutional quality, macroeconomic variables, sovereign ratings, as well as reverse causality. Measuring soundness through z- scores yields similar results. These findings emphasize the importance of transparency in making supervisory processes effective and strengthening market discipline. Countries aiming to upgrade banking regulation and supervision should consider giving priority to information provision over other elements of the Core Principles. Keywords: Bank Soundness, Regulation and Supervision, Basel Core Principles JEL Classification: World Bank Policy Research Working Paper 3954, June 2006 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 IMF, the World Bank, their Executive Directors, or the countries they represent. Policy Research Working Papers are available online at http://econ.worldbank.org. * The World Bank, International Monetary Fund, and International Monetary Fund, respectively. - 1 - 1. Introduction With increasing deregulation and globalization beginning in the 1980s, banking systems have become more fragile and banking crises have proliferated, causing or aggravating economic downturns and leading to significant fiscal costs (Caprio and Klingebiel, 1999). To improve crisis prevention and management, many countries are working to upgrade their bank regulation and supervision. This is a complex and difficult process, particularly in developing countries, where the required expertise may be scarce, the legal environment weak, and governance problems may lead to regulatory capture. But what exactly is good regulation and supervision? How can countries do it with limited resources? What should reforms focus on? To answer the first question, in 1997 a group of representatives of bank supervisors from advanced countries ­ the Basel Committee on Banking Supervision ­ issued the Core Principles for Effective Bank Supervision (BCPs), a document summarizing best practices in the field (Table 1).1 Most countries in the world have endorsed the BCPs and have undertaken to comply with them, making them an almost universal standard for bank regulators. Beginning in 1999, the IMF and the World Bank have conducted joint evaluations of member countries' compliance with this standard, mainly within their joint Financial Sector Assessment Program (FSAP).2 These assessments provide a unique source of information about the quality of supervision and regulation around the world. 1The countries represented in the Committee are Belgium, Canada, France, Germany, Italy, Japan, Luxembourg, the Netherlands, Spain, Sweden, Switzerland, United Kingdom and United States. The Committee consults widely with supervisors from non-member countries. 2FSAPs are a comprehensive evaluation of the financial sector and include assessments of compliance with several standards and codes. Many FSAPs are published and available on the IMF and World Bank websites. - 2 - In this paper, we rely on assessments of compliance with the BCPs to study whether better banking supervision and regulation is associated with sounder banks. In addition, we look at which elements of the regulatory framework are most closely related to soundness. The goal is to shed light on how to prioritize efforts to improve supervision. BCP compliance ratings have several advantages as measures of the quality of banking supervision and regulation. First, the BCPs are accepted as the blueprint of good supervision and regulation worldwide. Second, the assessment takes into account not only which laws and regulations are on the books, but also the extent to which they are implemented in practice. This is an important distinction. Third, separate compliance assessments are available for the 25 core principles, so that it is possible to separate out different aspects of supervision. The measure is only on a four-point scale, but to cut finer distinctions would likely be unrealistic. Because the evaluation reflects the judgment of the assessors, it inevitably contains an element of subjectivity. To limit subjectivity and ensure cross-country comparability, the Basel Committee has developed a standardized methodology. 3 In addition, the evaluations are conducted by expert supervisors from foreign countries and reviewed by internal teams at the Fund and the Bank. In any case, to the extent that measurement error is independent of bank soundness, it should not bias our results. We measure bank soundness using Moody's financial strength ratings. This is a comprehensive measure of the ability of a bank to meet its obligations to depositors and other creditors as viewed by specialized analysts. To the extent that Moody's analysts have access to both quantitative and qualitative information about banks and their operating environment, ratings should be a more accurate measure of bank soundness than indicators built using only 3The assessment methodology was published by the Basel Committee in 1999. - 3 - balance sheet variables (such as reported non-performing loans, profitability, or Z-scores).4 Another advantage of Moody's ratings is they aim at capturing bank solvency independent of the safety net, so that cross-country differences in the safety net, which are difficult to observe and measure, should not affect the results. A limitation of using Moody's ratings is that it restricts the sample to larger banks, as smaller banks are not rated. Thus, our investigation will not address the impact of the regulatory framework on smaller banks. As the latter are not likely to be of systemic importance, this limitation should be relatively minor. A more serious limitation is that a number of low-income countries have no rated banks, and thus are excluded from the sample.5 Because so far data on BCP compliance are available only at one point in time, our study cannot rely on time series variation. This limitation forces us to be very careful in controlling for other potential sources of cross-sectional variation as well as joint endogeneity. One important problem is that the BCP compliance indicator may be associated with bank soundness because it proxies for the overall quality of the institutional and macroeconomic environment rather than capture specific features of supervision. While we recognize that this concern cannot be fully laid to rest, we perform extensive robustness tests 4In support of this view, Sironi (2000) finds that credit ratings outperform balance sheet variable in predicting spreads on bank subordinated notes and debentures in Europe. Other studies have shown that changes in credit ratings cause changes in equity prices of banks in the U.S. (Schweitzer et al., 1992; Billet et al., 1998) and in Europe (Gropp and Richards, 2001), indicating that ratings agency are believed by the market to have superior information. In emerging markets, however, this does not seem to be the case (Richards and Deddouche, 1999). Rating agencies have been faulted for failing to give advance warnings of bank fragility before the East Asia crisis. Our testing strategy does not rely on the suitability of ratings as early warning signals. 5In related work we study the impact of BCP compliance on bank Z-scores, which allows us to investigate a larger set of institutions, including smaller banks from poor countries (Demirgüç-Kunt, Detragiache, Tressel, 2006). - 4 - controlling for broad indexes of institutional quality, macroeconomic variables, as well as sovereign credit ratings. A second concern is the endogeneity of supervision. The level of compliance with the BCPs is chosen by the countries themselves, and countries with sounder banks may face less opposition in enacting more rigorous supervision than countries with distressed banks.6 On the other hand, it can also be argued that countries that have experienced episodes of distress, and whose banks are still weak, may have the strongest incentives to upgrade their supervisory capacity. So in principle the endogeneity of supervision may bias the relationship with soundness in either direction. To take endogeneity into account, we resort to instrumental variables estimation. We find that an index of overall compliance with the BCPs is positively correlated with bank ratings after controlling for institutional quality, the macroeconomic environment, and bank characteristics. However, this relationship is not very robust. When we distinguish among groups of BCPs, on the other hand, we find a very robust positive relationship between compliance with information provision (BCP No. 21) and bank soundness. More specifically, countries in which banks have to report regularly and accurately their financial data to regulators and market participants have more highly rated banks. The relationship between bank soundness and transparency remains even after we instrument for transparency, suggesting it does not reflect reverse causality. 6Another source of endogeneity bias might be that assessors may unconsciously tend to give higher compliance ratings to countries whose banks are known to be financially sound. This type of perception bias is called the "Halo effect" which occurs when survey respondents respond more favorably to questions about richer countries, as explained in Glaeser, La Porta, Lopez de Silanes and Shleifer (2004). - 5 - The remainder of the paper is organized as follows: the next section briefly reviews related literature. Section 3 describes the data and the methodology. The results of the empirical estimation are in Section 4. Section 5 concludes. 2. Literature Review In their empirical studies of bank crisis determinants, Demirgüç-Kunt and Detragiache (1998, 2002) find that countries with better institutions are less likely to experience banking crises and are less affected by moral hazard due to deposit insurance. They interpret institutional variables such as rule of law and quality of bureaucracy as proxies for supervision and regulation, but do not include direct measures of this dimension, which were not available at the time. Barth, Caprio, and Levine (2001) (BCL) assembled the first extensive cross-country database on the characteristics of the supervisory and regulatory framework. The data come from a survey of bank supervisors, and measure the presence or absence of a series of regulatory features. In the first comprehensive theoretical and empirical study of alternative approaches to bank supervision, BCL (2004) find that regulatory and supervisory practices that force accurate information disclosure, empower private sector monitoring of banks, and foster incentives for private agents to exert corporate control work best to promote bank performance and stability. Specifically, in a cross-country setting they show that regulatory and supervisory regimes with these features have suffered fewer crises in the past two decades, have lower non-performing loans, and have deeper credit markets. One limitation of this study is that survey information reflects whether laws or regulations are on the books, but not to what extent they are implemented in practice. - 6 - Judging from FSAP reports, implementation is a serious concern in many countries. BCP compliance assessments allow us to better capture this important dimension, at least to the extent the assessors are able to evaluate it. Another difference between our approach and BCL is that we use bank ratings as measures of bank soundness in cross-sectional regressions, while BCL use the occurrence of systemic banking crises and non-performing loans. Because systemic crises are rare events, in BCL's analysis crises predate information on bank regulation and supervision which is available only for the end of the sample period. To the extent that bank regulation and supervision may have evolved over the sample, the relationships identified may not be accurate. Non-performing loans do not suffer from this problem, but are not usually comparable across countries, since countries have difference rules for marking loans as non-performing (Barth et al., 2006). On the other hand, an advantage of the BCL survey data is that it does not rely on the subjective views of compliance assessors. Three papers have used information on BCP compliance to study bank performance: An early paper by Sundararajan, Marston, and Basu (2001) uses a sample of 25 countries to examine the relationship between an overall index of compliance and bank soundness measured by the size of non-performing loans (NPLs) and loan spreads. They found BCP compliance not be a significant determinant of soundness. Podpiera (2004) extends the set of countries and finds that better BCP compliance lowers NPLs. Das et al. (2004) broadens the measure of bank soundness to include also compliance with standards for monetary and financial policies, and shows that better regulatory governance is associated with sounder banks, particularly in countries with better institutions. - 7 - We claim that bank ratings are a more accurate measure of bank soundness across countries than NPLs, loan spreads, interest margins, or capital adequacy. Because different countries have different reporting rules, NPLs are notoriously difficult to compare across countries, as already mentioned above. On the other hand, loan spreads or interest margins, and capitalization are affected by a variety of forces other than fragility, such as market structure, differences in risk-free interest rates and operating costs, and varying capital regulation. Another departure from existing studies is that we conduct much more extensive robustness test to separate out the effects of banking regulation from those of other country characteristics, such as institutional quality and macroeconomic performance. Finally, and perhaps most importantly in light of the results, our work is the first to explore how compliance with different components of the BCPs affects soundness. Our study is also related to the growing literature on transparency in financial markets. Patel, Balic, and Bwakira (2002) compare transparency and disclosure in emerging markets corporations. Mitton (2002) argues that transparency helps weather financial crises, as East Asian companies with auditors from major international accounting firms were found to have better stock performance after the crisis of 1997-98. Glennerster and Shin (2003) shows that countries that disclose more information to the IMF and the markets are rewarded through lower borrowing costs. More transparency also attracts more foreign portfolio investment (Gelos and Wei, 2005). Perotti and Von Thadden (2005) study how the presence of a dominant investor affects a firm decision of how much to disclose. - 8 - 3. Methodology and Data Basic Econometric Model The empirical model is a basic cross-sectional regression using bank level data. The dependent variable is the bank's rating. Moody's rates banks' financial strength on a 15-point scale, ranging from E- (least sound) to A+ (most sound). Since this is a limited dependent variable, the appropriate econometric model is an ordered probit, which we use to estimate the baseline specification. We also show that the results do not change if we transform the ratings in a numerical index and estimate an OLS model. Standard errors are clustered by country to allow for correlated residuals within each country. Specifically, the regression equations we estimate are of the form: Ratingij = + 1compliancej + 2bank characteristicsij +3institutions j + 4macroeconomiccontrolsj +ij, (1) where the subscript j denotes the country and the subscript i denotes the bank. The variable of interest is compliance with the BCPs. Assessors separately rate compliance with each principle using a four-point scale.7 We assign numerical values to these ratings, aggregate over all the principles, and standardize the sum to obtain an index that varies from zero to one. Since we are also interested in differentiating among the various dimensions of regulation and supervision, we compute compliance indexes for subgroups of principles following the grouping by chapters used by the Basel Committee (See Table 1 and Appendix 1). Compliance for each chapter is used as an alternative variable of interest. The first set of control variables includes various bank characteristics that might affect financial strength: size, measured by the logarithm of bank assets; profitability, 7The four categories are compliant, largely compliant, materially non-compliant, and non-compliant. - 9 - measured alternatively as return on assets or return on equity; and capitalization, measured by the ratio of equity to total assets. We also use the ratio of bank loans to total assets and the ratio of deposit to total liabilities to control for the type of intermediation and funding sources of each bank. Finally, we control for whether the bank is owned by the government or by foreigners and for whether the bank is not a commercial bank.8 A second group of control variables captures the overall quality of the institutions. Combining information from a variety of available indexes, Kaufman, Kraay, and Mastruzzi (2003) create various broad measures of perception of institutional quality which have been widely used in empirical studies. In our baseline specification, we use an index capturing the extent to which the rule of law is respected. This index is strongly correlated with other institutional indexes from the same source, such as lack of corruption, contract enforcement, etc., and we obtain similar results using these alternative indexes or an average of all the indexes. As alternative control for the overall quality of institutions we use GDP per capita. Bank soundness is also affected by the macroeconomic outlook, as slow output growth, high inflation, a depreciating currency, high real interest rates, and rapid credit expansion have been found to be associated with bank instability (see, for instance, Demirgüç-Kunt and Detragiache, 1998). Thus, as a third set of controls, we employ various combination of these macroeconomic variables in alternative specifications. We also use S&P's sovereign rating as a comprehensive indicator of the quality of macroeconomic policies and institutions which might affect bank stability in a country. 8The sample includes a few mutual banks, investment houses, and similar institutions. Excluding non- commercial banks does not change the results. - 10 - Sample and Data Sources BCP compliance data come from assessments carried out by the IMF and the World Bank beginning in 1999. Some of these assessments are public information and can be found on the institutions' websites. Others are kept confidential at the request of the country authorities. Moody's financial strength ratings have been compiled from Bankscope database. Bankscope reports the current rating, the last date in which the rating was revised, and the rating that prevailed before then, but no other historical rating information. Since we need to roughly match the timing of the BCP rating with that of Moody's ratings, we dropped from the sample banks that had a rating change after more than one year since the BCP compliance evaluation. Only a few banks fall in this category. Bank-level variables have also been constructed from Bankscope, while macroeconomic variables are mainly from the IMF's International Financial Statistics. Information about bank ownership is obtained from Bankscope and other miscellaneous sources. Banks are considered state-owned if the government has a controlling share or is the sole owner, and similarly for foreign-owned banks. Detailed variable definitions and sources are reported in Table 1 in the Appendix. We use BCP assessment data for 39 countries for which we have bank ratings from Moody's.9 Not surprisingly, this group includes mostly developed countries and emerging markets, while low-income countries are not represented (Table 2). Some countries have only one or two banks in the sample, but since we are not relying on intra-country variation to 9The BCP assessment data covers 67 countries, but only 39 of these countries have banks large enough to receive a rating from Moody's. - 11 - identify the relationship of interest this should not be a major limitation. The total number of banks in the baseline sample is 203. A First Look at the Data Figure 1 shows the overall compliance index by region and level of development. As expected, advanced countries have the highest compliance. Among the other countries, East Asia and Sub-Saharan Africa are closer to meeting the Basel standards, while South Asia, the Middle East and North Africa and Latin America lag behind. If we examine subgroups of chapters, the highest compliance is with principles concerning the licensing of banks and the structure of the banking market (Figure 2). At the other extreme, principles regarding the formal powers of supervisors were implemented to the least degree. Differences in compliance are not very large across different categories of principles, but averages across countries may hide larger discrepancies within each country. Table 3 presents correlations about compliance levels in different subgroups of chapters. Compliance levels are all significantly and positively correlated. The strongest correlation (over 80 percent) is between compliance with principles concerning prudential regulation and those concerning methods of ongoing supervision. Other correlations are considerably weaker, even falling below 50 percent in the case of principles regarding the formal powers of supervisors. All in all, these correlations suggest that there is enough variation in compliance across chapters to investigate the effect of various aspects of supervision separately. Turning now to bank ratings, Table 4 presents correlations between ratings and bank level variables. There is a strong positive correlation between bank size and bank soundness, - 12 - as larger, better diversified banks are seen as more stable. This correlation may be partly driven by the fact that larger banks tend to be located in more advanced countries, which have more stable economies. More profitable and better capitalized banks with a lower ratio of loans to assets also receive higher ratings. State banks are seen as more vulnerable, while the presence of large liquid assets is associated with a less favorable rating. This may reflect the existence of high compulsory liquidity requirements or lack of opportunities to lend in some unstable emerging markets. Interestingly, foreign-owned banks do not receive significantly higher ratings. Among country level variables (Table 5), bank ratings have a strong positive correlation with the country sovereign rating, the index of rule of law, and GDP per capita. Rapid exchange rate depreciation is associated with lower ratings. The correlation with the variable of interest, compliance with the BCPs, is positive, large, and significant. 4. The Results Results Using the Aggregate Index and Their Robustness Regressions of bank financial strength ratings on the overall index of BCP compliance are in Table 6. The index enters the regressions positively and significantly both by itself and when bank characteristics are controlled for. Controlling for institutional development with GDP per capita does not alter the relationship, while when we control for the sovereign rating the coefficient of compliance becomes smaller and statistical significance drops from one percent to five percent. When the quality of institutions is controlled for through the rule of law index, the coefficient of compliance becomes even smaller and significance drops to ten percent, suggesting that rule of law is an important - 13 - control variable. We take the specification including rule of law as the baseline in the rest of the paper.10 Excluding advanced country banks from the sample does not change regression results much. However, when we exclude extreme observations, the coefficient of compliance becomes quite small and is no longer significant. Significance also disappears if we estimate the model using OLS instead of ordered probit.11 As far as the control variables, larger banks are rated significantly higher and state- owned banks are considered more vulnerable. Foreign-owned banks are also rated more favorably and the coefficient is significant in the specification that includes the overall institutional quality of the country, but not if we restrict the sample to developing countries.12 These results suggest that once we control for bank size, ownership, and institutions, the additional explanatory power of bank balance sheet variables is small. We conclude that while there is a positive correlation between bank soundness and the overall index of BCP compliance, this result is sensitive to controlling for the institutional quality of the country and to the exclusion of outliers. An interesting question is whether different principles that make up the overall BCP index affect bank soundness differently. We turn to this issue next. 10Using an average of the Kaufman et al. (2003) institutional indexes rather than rule of law does not change any of the results. 11Extreme observations are identified by dropping the top one percent from the top and bottom of the distribution of each of the five bank characteristics. This results in a loss of about ten percent of the observations. 12The baseline specification excludes liquidity variables to preserve sample size. However including liquidity throughout does not change the results. - 14 - Results by Group of Principles In Table 7 we investigate the impact of different core principles as summarized by the seven chapters by including them in our baseline specification one at a time. Our results indicate that principles related to licensing and structure (Chapter 2) and information requirements (Chapter 5) are the most significantly associated with higher bank ratings at one percent significance. At the ten percent level, preconditions for effective bank supervision (Chapter 1) also enter significantly. To check if these results are robust we control for compliance with the rest of the principles. Thus we create three aggregate indexes excluding Chapters 1, 2, and 5 respectively. When we include these aggregate indexes in the relevant regressions, we see that only compliance with information requirements (Chapter 5) remains significant. As an additional robustness test, we re-run these three specifications excluding individual countries one by one. The last three columns of Table 7 report the regressions that result in the largest standard error for the compliance variable of interest. The results show that Chapter 5, the principle measuring compliance with transparency, is the only principle that remains significantly associated with bank soundness. Assessing the magnitude of the effect of Chapter 5 compliance on ratings is somewhat complicated by the fact that the econometric model is non-linear and that the compliance indicator is discrete. Nonetheless, computing marginal effects at the sample mean of the regressors, and assuming a constant elasticity around the mean, a decline in compliance from largely compliant to materially non-compliant would lower by one notch the rating of a bank rated D (from D to D-). - 15 - Robustness of the Relationship Between Bank Ratings and Compliance with Information Provision In this section we conduct additional robustness checks on the empirical relationship between information disclosure and bank soundness and investigate reverse causality. Compliance with Chapter 5 of the BCPs is positively and significantly associated with bank ratings also when we exclude from the sample advanced countries or extreme observations. In further robustness tests, we add to the baseline regression additional macroeconomic controls, which have been found in the literature to be associated with bank fragility (Demirgüç-Kunt and Detragiache, 1998 and 2005): depreciation of the nominal exchange rate, the inflation rate and its standard deviation, real credit growth, real GDP per capita growth.13 Among the macro variables, exchange rate depreciation and inflation volatility are negatively and significantly associated with bank ratings, while the level of inflation, real GDP growth, and real credit growth are not significant (results not reported) . Compliance with information requirements remains significant and positive throughout these different specifications, confirming that the impact of this principle is not sensitive to omitted macro controls. In yet another specification, we include in the regressions two measures of the frequency and timeliness of macroeconomic data release obtained from IMF's Special Data Dissemination Standards (Allum and Agça, 2001). These measures might help us control for unobservable characteristics that make a country more transparent. The frequency index is positive and significant, while the timeliness is not (result not reported). More importantly, compliance with chapter 5 remains significant and positive also in these regressions. 13For detailed variable definitions and sources see Appendix Table 1. - 16 - In the last set of regressions, we try to separate out compliance with Chapter 5 from other design features of supervision and regulation by introducing as regressors four indexes constructed using BCL's survey of bank regulators. Specifically, these indexes measure the presence in the laws and regulations of various provisions related to discipline, information disclosure, auditing requirements, and lack of banking restrictions (Barth, Caprio, Levine, 2001). Among the BCL variables, only the disclosure index has a significant impact on ratings once compliance with Chapter 5 is controlled for. Surprisingly, however, its coefficient is negative, perhaps suggesting that having many disclosure requirements but not enforcing them is detrimental to bank soundness. Endogeneity of BCP Compliance In this paper, we use individual bank level information to test the impact of compliance with the BCPs. Since an individual bank's soundness rating is unlikely to have an impact on country-level measures of supervisory quality, reverse causality should not be a serious concern. However, soundness ratings are correlated within countries, so it is possible that assessors may assign better compliance ratings to countries whose large banks are generally considered to be financially sound. Furthermore, since the level of compliance is a policy choice made by the countries themselves, countries with sounder banks may face less political opposition in adopting more rigorous regulatory and supervisory processes compared to those countries with distressed banks. On the other hand, it can also be argued that countries with weak banks coming out of a financial crisis may have the strongest incentives and political support to reform their regulations and supervision. So in principle the endogeneity of supervision may bias the relationship with soundness in either direction. - 17 - It should be emphasized that endogeneity concerns should not be overblown, particularly because we are examining only one of the chapters and using compliance with the other chapters as a control variable. Thus, the index of compliance with all but Chapter 5 should work as a comprehensive control for unobserved country characteristics or shocks that make the country more likely to comply with BCLs. It is only factors specific to compliance with information disclosure that remain to be controlled for. To deal with the potential endogeneity of compliance with Chapter 5, we use instrumental variables estimation. In estimation we apply a two stage least squares technique. We follow the established practice in the literature on institutions and use legal origin dummy variables as instruments.14 The idea is that fundamental historical country characteristics, such as legal origin, affect a country's institutions, and thus have a bearing on the particular institutions of bank regulation and supervision. On the other hand, bank soundness is unlikely to be affected by these historical characteristics, particularly after controlling for broad institutional quality and supervision itself. The 2SLS estimation results are in Table 9. First, legal origin variables are jointly highly significant in the first stage regressions. In all specifications the F-test for the joint significance of the instruments is well above 10, suggesting that there is no concern about weak instruments (Stock and Yogo, 2005). In the second stage, compliance with information provision remains positive and significant in all specification. Following Moreira (2003), we have also computed CLR confidence intervals, which should be robust to weak instruments. Also based on these confidence intervals compliance with Chapter 5 has a positive effect on 14See Beck, Demirgüç-Kunt and Levine (2003) for a discussion of alternative instruments of financial institution development. - 18 - bank ratings. Using religion as an alternative instrument gives similar results. On the other hand, other basic country characteristics, such as latitude or ethnic fractionalization, have poor predictive power for BCP compliance.15 Measuring Bank Soundness with Z-Scores As discussed in the introduction, Moody's ratings are a nice measure of bank financial strength, as they are prepared by specialists with access to variety of sources of information and whose business is to make accurate assessments. On the other hand, ratings are subjective, and it may be argued that analysts may be better disposed towards bank that provide more accurate and timely information even though these banks are not necessarily more sound. To address this concern, in this section we measure bank soundness using an alternative, objective, and commonly used indicator, namely the Z-score.16 We compute Z- scores for the banks in our sample and investigate if we continue to find evidence that compliance with information requirements is a significant correlate to bank soundness. The specification is broadly similar to the previous baseline, although we dropped equity and return on assets from the set of regressors since these variables are now used to compute the left-hand-side variable. The z-scores have a positive and statistically significant correlation with Moody's ratings, but the correlation is not very high (about 20 percent). Nonetheless, the relationship between bank soundness and BCP compliance remains 15 Barth et al. (2004) use religion and latitude as instruments for bank supervision. 16 The Z-score is defined as (average return on assets +equity/assets)/(standard deviation of the return on assets). It can be interpreted as the number of standard deviations below the mean by which returns would have to fall to wipe out bank equity (see Boyd and Runkle, 1993). In the regressions, we use as the dependent variable ln (1+ Z-score). - 19 - remarkably similar to the one uncovered using Moody's ratings (Table 10): the overall compliance index is positive but not significant, and when we examine individual chapters the only one that is robust is Chapter 5 on information provision. Additional tests using more macroeconomic control variables (exchange rate depreciation, inflation, inflation volatility, credit growth) confirm this finding (results not reported). To summarize, the positive relationship between compliance with information provision regulation is in evidence whether bank soundness is measured through Moody's ratings or through Z-scores. 5. Conclusions Strong regulation and supervision plays an essential part in ensuring a safe and sound banking system. To curb bank fragility and improve crisis management, many countries are in the process of strengthening their regulatory and supervisory systems, a complicated and costly process for many developing countries where human resources are scarce and other supporting institutions are weak. What type of regulations and supervisory practices are most effective in ensuring bank soundness? This is the question that we have addressed in this study. Using bank-level investor ratings for 39 countries, we study whether compliance with Basel Core Principles (BCPs), the standard of best practices in bank supervision, is associated with bank soundness. BCP compliance assessments, carried out under the auspices of the World Bank and IMF Financial Sector Assessment Program, provide a unique source of information about the quality of bank supervision and regulation around the world. - 20 - An important aspect of our study is the attempt to differentiate among different elements of the regulatory framework, to help prioritize reform efforts. We find a significant and positive relationship between compliance with information provision and bank soundness, which is robust to controlling broad indexes of institutional quality, macroeconomic variables, sovereign ratings, as well as reverse causality. Specifically, countries which require their banks to report regularly and accurately their financial data to regulators and market participants have more highly rated banks, as timely disclosure of high quality information strengthens monitoring by regulators and markets alike. Our results also suggest that countries aiming to upgrade banking regulation and supervision should consider giving priority to information provision over other elements of the Core Principles. Because information provision is a necessary condition for effective discipline, this policy recommendation is consistent with the approach to regulation and supervision recommended by Barth, Caprio, and Levine (2006), who stress the importance of mechanisms to empower market discipline and are skeptical of structures that assign too much power to regulators. - 21 - References Allum, Peter F., and Mehmet Agça, 2001, Economic data dissemination: What influences country performance on frequency and timeliness? IMF Working paper 01/173. Barth, J.R., Caprio Jr., G., Levine, R., 2001, The regulation and supervision of bank around the world: a new database. In: Litan, R.E., Herring, R. (Eds.), Integrating Emerging Market Countries into the Global Financial System. Brookings­Wharton Pap. Finan. Services. Brookings Institution Press, pp. 183­240. _________________, 2004, Bank regulation and Supervision: What Works Best?, Journal of Financial Intermediation, 13, 205-248. _________________, 2006, Rethinking Bank Supervision and Regulation: Until Angels Govern. Cambridge, UK: Cambridge University Press. Billet, M., J. Garfinkel and E. O'Neal, 1998, The Cost of Market Versus Regulatory Discipline in Banking, Journal of Financial Economics, 48, pp. 333-358. Boyd, John H., and David E. Runkle, 1993, Size and Performance of Banking Firms, Journal of Monetary Economics, 31, 47-67. Caprio, G., and Klingebiel, D., 1996, "Dealing with Bank Insolvencies: Cross Country Experience", Policy Research Working Paper no. 1620, The World Bank, Washington, D.C. Das, Udabair, S., Plamen Iossifov, Richard Podpiera, and Dmitry Rozkhov, 2005, Quality of Financial Policies and Financial System Stress, IMF Working Paper No. ? Demirgüç-Kunt, Asli, and Enrica Detragiache, 1998, Determinants of Banking Crises in Developed and Developing Countries, IMF Staff Papers, _________________, 2002, Does Deposit Insurance Increase Banking Sector Fragility? Journal of Monetary Economics, Journal of Monetary Economics, 49 (7), pp. 1373- 1406, November 2002. _________________, 2005, Cross-Country Empirical Studies of Systemic Bank Distress: A Survey, National Institute Economic Review, No. 192, pp. 68-83, April 2005. Gelos, Gaston, and Shang-Jin Wei, Transparency and International Portfolio Holdings, Journal of Finance, 60, 2987-3020. Glaeser, E; R. La Porta, F. Lopez de Silanes and A. Shleifer, 2004, Do institutions cause growth? Journal of Economic Growth 9(3), 271-303. Glennerster, Rachel, and Yongseok Shin, 2004, Is Transparency Good for You and Can the IMF Help?, IMF Working Paper No. 03/132. - 22 - Gropp, Reint, and Anthony Richards, 2001, Rating Agency Actions and the Pricing of Debt and Equity of European Banks: What Can We Infer about Private Sector Monitoring of Bank Soundness?, European Central Bank Working Paper No. 76. Kaufmann, D., A. Kraay, and M. Mastruzzi, 2003, Governance Matters III: Governance Indicators for 1996-2002 (Washington: World Bank). Levine, Ross, Gerard Caprio Jr., and James R.Barth, 2001. The Regulation and Supervision of Banks around the World - a New Database," Policy Research Working Paper Series 2588, The World Bank. Micco, Alejandro, Ugo Panizza, and Mónica Yañez, 2004, "Bank Ownership and Performance: Are Public Banks Different?," (unpublished; Washington: Inter-American Development Bank). Moreira, M. J., 2003, "A Conditional Likelihood Test for Structural Models," Econometrica, Vol. 71, No. 4, pp. 1027-48. Patel, Sandeep A., Amra Balic, and Liliane Bwakira, 2002, Measuring Transparency and Disclosure at Firm Level in Emerging Markets, Emerging Markets Review, 3, 325-337. Perotti, Enrico, and Ernst-Ludwig Von Thadden, 2005, Dominant Investor and Strategic Transparency, Journal of Law, Economics, and Organization, Vol.21, No. 1, 76-102. Podpiera, Richard, 2004, Does Compliance with Basel Core Principles Bring Any Measurable Benefits?, IMF Working Paper No. 04/204. Richards, Anthony, and David Deddouche, 1999, Bank Rating Changes and Bank Stock Returns--Puzzling Evidence from the Emerging Markets" IMF Working Paper No WP/99/151. Schweitzer, Robert, Samuel H. Szewczyk and Raj Varma, 1992, Bond Rating Agencies and Their Role in Bank Market Discipline, Journal of Financial Services Research 6, 249- 263. Sironi, Andrea, 2003, Testing for Market Discipline in the European Banking Industry: Evidence from Subordinated Debt Issues, Journal of Money, Credit, and Banking, 35, 3, 443-472. Stock, J.H., and M. Yogo, 2005, "Testing for Weak Instruments in IV Regressions," in Identification and Inference for Econometric Models: A Festschrift in Honor of Thomas Rothenber, (New York: Cambridge University Press). Sundararajan, V., David Marston, and Ritu Basu, 2001, Financial System Standards and Financial Stability--The Case of Basel Core Principles, IMF Working Paper 01/62. - 23 - Table 1. Basel Core Principles--Definitions Chapter 1: Preconditions for effective banking supervision Principle 1. Objectives, autonomy, powers, and resources Principle 1(1). There should be clear responsibilities and objectives set by legislations for each supervisory agency. Principle 1(2). Each supervisory agency should possess adequate resources to meet the objective set, provided on terms that do not undermine the autonomy, integrity and independence of supervisory agency. Principle 1(3). A suitable framework of banking laws, setting bank minimum standard, including provisions related to authorization of banking establishments and their supervision. Principle 1(4). The legal framework should provide power to address compliance with laws as well as safety and soundness concerns. Principle 1(5). The legal framework should provide protection of supervisors for actions taken in good faith in the course of performing supervisory duties. Principle 1(6). There should be arrangements of interagency cooperation, including with foreign supervisors, for sharing information and protecting the confidentiality of such information. Chapter 2: Licensing and Structure Principle 2. Definition of permissible activities Principle 3. Right to set licensing criteria and reject applications for establishments that do meet the standard sets Principle 4. Authority to review and reject proposals of significant ownership changes Principle 5. Authority to establish criteria for reviewing major acquisitions or investments Chapter 3: Prudential Regulations and Requirements Principle 6. Prudent and appropriate risk adjusted capital adequacy ratios must be set. Principle 7. Supervisors should evaluate banks' credit policies. Principle 8. Banks should adhere to adequate loan evaluation and loan-loss provisioning policies. Principle 9. Supervisors should set limits to restrict large exposures, and concentration in bank portfolios should be identifiable. Principle 10. Supervisors must have in place requirements to mitigate the risks associated with related lending. Principle 11. Policies must be in place to identify, monitor and control country risks, and to maintain reserves against such risks. Principle 12. Systems must be in place to accurately measure, monitor and adequately control markets risks and supervisors should have powers to impose limits or capital charge on such exposures. Principle 13. Banks must have in place a comprehensive risk management process to identify, measure, monitor and control all other material risks and, if needed, hold capital against such risks. Principle 14. Banks should have internal control and audit systems in place. Principle 15. Adequate policies, practices and procedures should be in place to promote high ethical and professional standards and prevent the bank being used by criminal elements. Chapter 4: Methods of On-Going Supervision Principle 16. An effective supervisory system should consist of on-site and off-site supervision. Principle 17. Supervisors should have regular contact with bank management. Principle 18. Supervisors must have a means of collecting, reviewing and analyzing prudential reports and statistics returns from banks on a solo and consolidated basis. - 24 - Principle 19. Supervisors must have a means of independent validation of supervisory information either through on-site examinations or use of external auditors. Principle 20. Supervisors must have the ability to supervise banking groups on a consolidated basis. Chapter 5: Information Requirements Principle 21. Each bank must maintain adequate records that enable the supervisor to obtain a true and fair view of the financial condition of the bank of the bank, and must publish on a regular basis financial statements that fairly reflects its condition. Chapter 6: Formal Powers of Supervisors Principle 22. Adequate supervisory measures must be in place to bring about corrective action when banks fail to meet prudential requirements, when there are regulatory violations, or when depositors are threatened in any other way. This should include the ability to revoke the banking license or recommend its revocation. Chapter 7: Cross-Border Banking Principle 23. Supervisors must practice global consolidated supervision over internationally active banks, adequately monitor and apply prudential norms to all aspects of the business conducted by these banks. Principle 24. Consolidated supervision should include establishing contact and information exchange with the various supervisors involved, primarily host country supervisory authorities. Principle 25. Supervisors must require the local operations of foreign banks to be conducted at the same standards as required of domestic institutions, and must have powers to share information needed by the home country supervisors of those banks. Source: Core Principles for Effective Banking Supervision, Basle Committee on Banking Supervision, Basle, Sept. 1997 - 25 - Table 2. Descriptive Statistics (by county) Sovereign GDP Av. NB. Countries Rating per Rule of Law Growth Depreciation Bank ROAA ROAE NLTA ETA LogTA of Foreign- State- capita Rating Banks owned owned BAHRAIN 0.56 10591 0.86 0.01 0.00 0.44 1.5 14.2 37.4 10.3 16.3 3 0 1 BRAZIL 0.19 4580 -0.19 0.01 0.20 0.37 2.1 24.6 27.1 8.9 15.2 10 4 2 BULGARIA 0.38 1541 -0.09 0.04 0.88 0.25 1.6 14.2 52.4 12.2 12.0 4 1 0 CANADA 1.00 22745 1.89 0.03 0.02 0.67 0.5 9.4 54.3 5.3 17.0 3 0 0 COLOMBIA 0.31 2278 -0.63 0.01 0.11 0.36 -0.2 -4.3 56.0 14.5 7.4 3 1 0 CROATIA 0.44 5023 -0.07 0.04 0.10 0.40 1.1 11.8 35.0 7.9 13.3 1 0 0 CZECH REP. 0.75 5387 0.64 0.00 0.08 0.37 1.1 14.9 38.1 7.2 11.5 2 2 0 EGYPT 0.38 1211 0.18 0.03 0.00 0.32 0.8 12.8 51.4 5.5 14.5 6 1 5 ESTONIA 0.75 3787 0.60 0.06 0.08 0.44 2.3 16.9 59.3 12.5 11.2 3 2 0 FINLAND 1.00 31222 2.04 0.03 . 0.67 0.5 10.8 59.0 4.6 16.1 3 0 0 FRANCE 1.00 29935 1.46 0.02 . 0.64 0.5 8.0 32.3 5.6 17.4 14 1 0 HONG KONG 0.75 24483 1.58 0.02 0.00 0.45 1.2 11.7 51.6 11.0 13.6 5 0 0 HUNGARY 0.75 5368 0.79 0.04 0.13 0.49 2.8 21.9 51.3 14.3 9.6 3 1 0 INDIA 0.44 465 0.13 0.04 0.09 0.33 0.9 17.7 45.9 5.2 12.7 5 0 4 INDONESIA 0.13 1010 -0.75 0.02 0.34 0.22 -1.4 28.5 29.5 9.4 6.7 5 0 1 IRELAND 1.00 27466 1.77 0.08 . 0.47 0.3 6.6 20.4 3.6 15.3 4 0 0 ISRAEL 0.69 17254 1.06 0.01 0.06 0.48 0.4 6.1 57.6 5.9 15.5 5 0 2 JAPAN 1.00 44459 1.62 0.01 0.02 0.30 -0.3 -16.8 65.0 4.6 13.0 23 0 1 KAZAKHSTAN 0.44 1552 -0.79 0.06 0.18 0.24 2.0 16.3 57.9 10.9 8.0 9 0 0 KOREA REP. OF 0.63 12962 0.78 0.04 0.08 0.29 0.4 6.0 64.3 4.8 10.1 7 0 1 KUWAIT 0.69 12392 0.94 -0.03 0.01 0.44 1.9 18.0 39.9 12.3 16.9 6 1 0 LATVIA 0.69 2620 0.24 0.07 0.02 0.33 1.2 12.6 61.7 9.6 13.8 2 0 1 MAURITIUS 0.50 4198 0.85 0.04 0.08 0.37 2.2 19.9 70.3 11.2 10.9 2 0 0 MEXICO 0.56 3720 -0.27 0.04 0.08 0.47 0.6 8.3 62.8 8.8 14.9 3 1 0 MOROCCO 0.38 1394 0.29 0.02 0.05 0.42 1.0 9.3 50.1 10.6 12.5 3 1 0 PERU 0.25 2314 -0.43 0.02 0.09 0.33 0.3 3.3 52.1 9.7 13.9 2 1 0 PHILIPPINES 0.19 1163 -0.29 0.01 0.12 0.29 1.4 25.5 49.7 11.4 11.3 6 0 2 POLAND 0.69 3643 0.57 0.04 0.08 0.28 -0.8 -18.2 54.6 8.5 13.9 7 2 1 RUSSIAN FED. 0.44 2905 -0.80 0.04 0.50 0.21 1.8 13.7 48.9 17.0 10.7 11 0 2 SLOVAKIA 0.69 4293 0.24 0.04 0.09 0.33 0.5 16.3 29.5 5.8 10.4 2 2 0 SLOVENIA 0.81 11641 0.84 0.04 0.14 0.43 -1.1 -12.3 44.6 9.9 8.8 2 1 1 SOUTH AFRICA 0.56 3913 0.26 0.00 0.14 0.53 1.0 15.3 72.9 5.2 14.5 2 1 0 SWEDEN 1.00 32523 1.94 0.03 0.01 0.73 -0.1 -1.5 43.6 4.6 15.7 1 0 0 - 26 - THAILAND 0.56 2805 0.40 0.01 0.10 0.26 -1.2 -46.0 67.0 4.5 12.6 7 1 0 TUNISIA 0.50 2472 0.31 0.03 0.05 0.29 1.3 14.1 67.1 9.3 13.9 5 1 1 TURKEY 0.19 2903 0.07 0.00 0.72 0.36 0.3 -2.8 29.8 12.3 15.5 8 1 0 UKRAINE 0.19 891 -0.72 0.04 0.26 0.20 1.5 15.9 55.1 8.6 12.1 1 0 0 UNITED ARAB EM. 0.75 18409 1.10 0.01 0.00 0.47 2.5 16.2 64.4 15.8 14.0 4 0 2 U.K. 1.00 22164 1.91 0.03 0.01 0.63 0.7 11.9 58.3 7.0 16.6 13 2 0 Average/ total 0.63 13843 0.63 0.02 0.15 0.39 0.8 6.1 50.9 8.6 13.4 205 28 27 - 27 - Table 3. Correlations between BCP chapters Preconditions Prudential Methods of Formal Overall Index effective Licensing regulations and on-going Information powers of Cross-border supervision and structure requirements supervision requirements supervisors banking Overall Index 1.00 Preconditions 0.77 1.00 effective supervision 0.00 Licensing and 0.78 0.65 1.00 structure 0.00 0.00 Prudential regulations 0.94 0.64 0.71 1.00 and requirements 0.00 0.00 0.00 Methods of on-going 0.90 0.60 0.63 0.84 1.00 supervision 0.00 0.00 0.00 0.00 Information requirements 0.77 0.61 0.58 0.67 0.77 1.00 0.00 0.00 0.00 0.00 0.00 Formal powers of 0.61 0.57 0.47 0.51 0.46 0.43 1.00 supervisors 0.00 0.00 0.00 0.00 0.00 0.00 Cross-border banking 0.83 0.71 0.64 0.69 0.69 0.64 0.48 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Note: p values in italics - 28 - Table 4. Correlations among bank characteristics Bank rating ROAA ROAE NLTA ETA logTA Foreign State owned owned Bank rating 1.00 Return on assets 0.04 1.00 0.54 Return on equity 0.16 0.77 1.00 0.02 0.00 Net loans-to-assets -0.22 -0.16 -0.23 1.00 0.00 0.02 0.00 Capitalization -0.14 0.41 0.20 -0.18 1.00 0.04 0.00 0.00 0.01 - Total assets (logs) 0.63 -0.02 0.01 -0.23 0.30 1.00 0.00 0.73 0.86 0.00 0.00 Foreign owned 0.05 -0.05 -0.02 -0.04 0.04 -0.05 1.00 0.50 0.52 0.78 0.56 0.59 0.45 - State owned -0.19 0.03 0.07 -0.02 0.07 0.01 -0.16 1.00 0.01 0.68 0.29 0.82 0.35 0.87 0.03 Note: p-values are in italics - 29 - Table 5. Correlations among country-level variables Av. GDP Percent Percent bank Country BCP Rule of per foreign state Lack rating rating Index Law capita Depreciation Growth owned owned restrictions Auditing Disclosure Discipline (BCL) (BCL) (BCL) (BCL) Av. bank rating 1.00 Country rating 0.69 1.00 0.00 BCP Index 0.58 0.52 1.00 0.00 0.00 Rule of Law 0.76 0.89 0.60 1.00 0.00 0.00 0.00 GDP per capita 0.63 0.80 0.56 0.85 1.00 0.00 0.00 0.00 0.00 Depreciation -0.43 -0.51 -0.14 -0.49 -0.35 1.00 0.01 0.00 0.40 0.00 0.03 Growth -0.05 0.20 0.19 0.00 -0.06 -0.04 1.00 0.77 0.21 0.25 0.99 0.73 0.84 Percent foreign owned -0.04 -0.02 -0.26 -0.17 -0.32 -0.04 -0.07 1.00 0.81 0.91 0.11 0.30 0.04 0.82 0.68 Percent state owned -0.21 -0.19 0.16 -0.14 -0.20 -0.13 0.03 -0.21 1.00 0.20 0.25 0.32 0.41 0.23 0.46 0.87 0.19 Lack restrictions (BCL) 0.30 0.18 0.35 0.22 0.19 0.10 0.05 -0.13 -0.19 1.00 0.06 0.28 0.03 0.17 0.23 0.54 0.76 0.43 0.25 Auditing (BCL) 0.21 0.07 0.16 0.08 0.04 -0.13 0.09 0.22 -0.10 -0.03 1.00 0.19 0.69 0.33 0.61 0.82 0.43 0.58 0.19 0.54 0.88 Disclosure (BCL) 0.33 0.39 0.42 0.48 0.35 -0.35 -0.01 0.04 -0.02 0.52 0.26 1.00 0.04 0.02 0.01 0.00 0.03 0.04 0.95 0.81 0.91 0.00 0.11 Discipline (BCL) -0.24 -0.16 -0.05 -0.15 -0.23 -0.25 0.09 0.08 0.18 0.03 0.14 0.39 1.00 0.15 0.34 0.76 0.35 0.15 0.15 0.57 0.62 0.26 0.84 0.41 0.01 Note: p-values are in italics - 30 - Table 6: Impact of overall index of compliance with BCPs on bank ratings (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Baseline LDCs Tails excluded Index of compliance with BCPs 4.927 4.259 4.363 4.125 2.626 2.805 2.824 1.647 1.448 1.855 0.686 0.14 [7.38]*** [4.94]*** [5.18]*** [4.98]*** [3.10]*** [3.41]*** [2.94]*** [2.01]** [1.66]* [1.75]* [0.69] [1.48] Foreign-owned 0.32 0.252 0.261 0.261 0.286 0.277 0.25 0.381 0.406 0.308 0.433 0.036 [1.65]* [1.30] [1.31] [1.35] [1.50] [1.43] [1.49] [2.33]** [2.48]** [1.59] [2.50]** [1.87]* State-owned -0.658 -0.721 -0.702 -0.709 -0.724 -0.677 -0.772 -0.505 -0.515 -0.521 -0.499 -0.065 [2.95]*** [2.96]*** [2.90]*** [2.71]*** [2.77]*** [2.54]** [2.63]*** [2.09]** [2.02]** [1.96]* [1.98]** [2.58]** Other banking institutions 0.296 0.389 0.392 0.377 0.293 0.284 0.338 0.073 0.14 0.069 0.123 0.022 [2.02]** [2.34]** [2.38]** [2.07]** [1.50] [1.49] [1.50] [0.36] [0.71] [0.21] [0.65] [0.91] Return on equity 0.005 0.005 0.005 0.006 0.006 0.008 0.008 0.001 0.014 0.001 [0.91] [0.87] [0.89] [1.22] [1.34] [1.55] [1.59] [0.32] [2.62]*** [1.65] Capitalization 0.011 -0.01 0.013 0.018 0.01 0.028 0.032 0.029 0.026 0.002 [1.15] [0.35] [0.53] [0.65] [0.46] [1.13] [1.52] [1.03] [1.27] [1.12] Net loans-to-assets -0.007 -0.002 -0.003 -0.001 -0.006 -0.006 -0.005 -0.006 -0.001 [1.08] [0.26] [0.50] [0.21] [1.06] [1.06] [0.74] [0.98] [0.93] Total assets 0.214 0.205 0.213 0.216 0.188 0.074 0.165 0.022 [3.81]*** [3.81]*** [3.80]*** [4.07]*** [3.14]*** [1.28] [2.82]*** [2.90]*** Return on assets 0.017 [0.32] GDP per capita -3.4E-06 [0.31] Index of sovereign rating 1.284 [1.90]* Index of Rule of Law 0.509 0.883 0.568 0.056 [1.97]** [3.28]*** [2.14]** [2.15]** Observations 260 206 206 203 203 203 203 205 203 132 180 203 Pseudo R2 or R2 0.12 0.1 0.11 0.11 0.15 0.15 0.15 0.18 0.17 0.12 0.15 0.55 Method of estimation Ordered Ordered Ordered Ordered Ordered Ordered Ordered Ordered Ordered Ordered Ordered OLS probit probit probit probit probit probit probit probit probit probit probit Robust z statistics in brackets, observations are clustered by country. * significant at 10%; ** significant at 5%; *** significant at 1%. - 31 - Table 7. Impact of individual chapters Dropping countries one by one 1/ (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) Foreign-owned 0.427 0.309 0.409 0.394 0.38 0.397 0.403 0.449 0.186 0.295 [2.65]*** [1.94]* [2.24]** [2.52]** [2.80]*** [2.63]*** [2.37]** [2.37]** [1.03] [1.8]* State-owned -0.52 -0.575 -0.443 -0.381 -0.673 -0.406 -0.447 -0.498 -0.639 -0.691 [2.07]** [2.07]** [1.82]* [1.51] [2.35]** [1.64] [1.75]* [-1.86]* [-2.16]** [-2.14]** Other banking institutions 0.105 0.234 0.099 0.083 0.252 0.093 0.124 0.077 0.296 0.340 [0.56] [1.29] [0.50] [0.41] [1.25] [0.45] [0.62] [0.4] [1.64] [1.51] Return on equity 0.008 0.008 0.013 0.01 0.007 0.01 0.009 0.013 0.002 0.002 [1.57] [1.58] [1.83]* [1.91]* [1.40] [1.99]** [1.82]* [1.73]* [0.31] [0.56] Capitalization -0.005 -0.006 -0.006 -0.007 -0.009 -0.008 -0.005 -0.004 -0.001 0.560 [1.00] [1.01] [0.98] [1.38] [1.61] [1.39] [0.95] [-0.63] [-0.14] [-0.77 ] Net loans-to-assets 0.03 0.036 0.03 0.035 0.028 0.038 0.039 0.027 0.013 0.013 [1.42] [1.60] [1.39] [1.60] [1.52] [1.70]* [1.86]* [1.26] [0.54] [0.64] Total assets 0.2 0.211 0.195 0.221 0.212 0.205 0.206 0.210 0.184 0.184 [3.47]*** [3.74]*** [3.04]*** [3.22]*** [3.59]*** [3.38]*** [3.52]*** [3.34]*** [2.65]*** [2.85]*** Index of Rule of Law 0.551 0.5 0.603 0.724 0.434 0.744 0.576 0.575 0.939 0.858 [ [2.26]** [2.03]** [2.25]** [3.68]*** [1.79]* [4.05]*** [2.28]** [2.21]** [4.05}*** 3.85]*** Index chapter 1 1.14 1.379 [1.73]* [1.07] Index chapter 2 2.538 1.491 [3.71]*** [1.39] Index chapter 3 0.568 [0.56] Index chapter 4 -0.632 [0.59] Index chapter 5 2.037 1.573 [3.17]*** [2.13]** Index chapter 6 -0.509 [0.90] Index chapter 7 0.682 [1.31] Av. chapters, excl. cha. 1 -0.399 [-0.26] Av. chapters, excl. cha. 2 0.116 [0.07] Av. chapters, excl. cha. 5 0.326 [0.33] Observations 203 203 189 203 203 203 203 186 166 175 Pseudo R2 0.17 0.18 0.17 0.17 0.19 0.17 0.17 0.1736 0.2432 0.25 Method of estimation Ordered Ordered Ordered Ordered Ordered Ordered Ordered Ordered Ordered Ordered probit probit probit probit probit probit probit probit probit probit 1/regression with the largest standard error for the variable of interest. Robust z statistics in brackets, observations are clustered by country. * significant at 10%; ** significant at 5%; *** significant at 1%. - 32 - Table 8. Impact of information requirements (Chapter 5) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) No tails LDCs OLS Compliance with ch. 5 2.108 2.128 2.089 0.187 2.324 2.884 2.326 2.098 2.407 2.053 1.89 [2.79]*** [2.86]*** [2.99]*** [2.21]** [3.42]*** [5.43]*** [3.16]*** [3.13]*** [4.55]*** [2.47]** [2.36]** Foreign owned 0.384 0.452 0.294 0.032 0.398 0.44 0.315 0.416 0.461 0.382 0.374 [2.58]*** [2.70]*** [1.44] [1.88]* [2.34]** [2.50]** [2.18]** [2.45]** [2.94]*** [2.54]** [2.39]** State owned -0.627 -0.605 -0.521 -0.073 -0.786 -0.585 -0.579 -0.527 -0.568 -0.62 -0.688 [2.10]** [2.09]** [1.64] [2.57]** [2.40]** [1.93]* [1.94]* [1.81]* [2.06]** [2.09]** [2.39]** Return on equity 0.011 0.02 -0.001 0.001 0.01 0.01 0.011 0.011 0.01 0.011 0.011 [1.61] [3.13]*** [0.13] [1.92]* [1.29] [1.34] [1.63] [1.59] [1.37] [1.62] [1.55] Net loans-to-assets -0.009 -0.009 -0.005 -0.001 -0.009 -0.007 -0.009 -0.007 -0.003 -0.008 -0.007 [1.53] [1.30] [0.86] [1.15] [1.55] [1.26] [1.59] [1.24] [0.64] [1.48] [1.26] Capitalization 0.028 0.025 0.026 0.002 0.04 0.047 0.036 0.037 0.046 0.03 0.043 [1.42] [1.45] [1.17] [0.91] [1.99]** [2.74]*** [1.73]* [1.93]* [2.76]*** [1.48] [2.20]** Total assets 0.222 0.184 0.12 0.024 0.223 0.198 0.219 0.233 0.272 0.222 0.276 [3.62]*** [3.04]*** [1.95]* [3.46]*** [3.62]*** [3.52]*** [3.62]*** [3.78]*** [4.45]*** [3.58]*** [4.00]*** Index of Rule of Law 0.509 0.597 0.782 0.053 0.375 0.373 0.613 0.561 0.827 0.517 0.507 [2.25]** [2.71]*** [3.16]*** [2.21]** [1.53] [1.71]* [2.75]*** [2.60]*** [3.78]*** [2.19]** [2.19]** BCP Compliance -0.648 -1.064 -0.686 -0.056 -0.296 -1.268 -0.706 -0.761 -1.044 -0.662 -0.68 ( excl. chapter 5) [0.58] [0.93] [0.64] [0.46] [0.31] [1.44] [0.62] [0.76] [1.12] [0.61] [0.63] Inflation -0.005 [0.65] Inflation volatility -0.008 [4.83]*** Depreciation -0.835 [2.72]*** Macro data frequency 0.347 [2.46]** Discipline (BCL) -0.882 [1.45] Disclosure (BCL) -2.882 [3.71]*** Auditing(BCL) 0.129 [0.34] Lack restrictions (BCL) -0.916 [1.86]* Observations 189 167 118 189 176 167 189 189 189 189 189 Pseudo R2 0.19 0.19 0.15 0.59 0.2 0.19 0.2 0.2 0.22 0.19 0.2 Robust z statistics in brackets, observations are clustered by country. * significant at 10%; ** significant at 5%; *** significant at 1%. - 33 - Table 9. Bank ratings and information transparency: instrumental variables regressions (2SLS) Second Stage (1) (2) (3) (4) (5) (6) Compliance with ch.5 0.623 0.496 0.715 0.536 0.689 0.715 [2.86]*** [2.49]** [3.00]*** [2.37]** [2.73]*** [2.83]*** Foreign owned 0.020 0.019 0.021 0.029 0.021 0.018 [0.83] [0.94] [0.76] [1.39] [0.73] [0.61] State owned -0.124 -0.113 -0.116 -0.096 -0.103 -0.138 [2.78]*** [3.02]*** [2.43]** [2.49]** [2.23]** [2.62]*** Return on equity 0.000 0.000 0.000 0.000 0.000 0.000 [0.21] [0.22] [0.48] [0.20] [0.25] [0.49] Net loans-to-assets -0.001 -0.001 -0.001 -0.002 -0.001 -0.001 [1.28] [1.36] [1.19] [1.49] [1.45] [1.21] Total assets 0.023 0.023 0.017 0.018 0.011 0.023 [2.89]*** [2.60]*** [2.02]** [2.56]** [1.28] [2.62]*** Capitalization -0.001 0.000 0.001 0.002 0.002 0.003 [0.16] [0.11] [0.36] [0.71] [0.72] [0.96] Index of Rule of Law -0.024 -0.044 -0.031 [0.69] [1.15] [0.84] Inflation 0.000 [0.31] Standard deviation of inflation -0.002 [3.64]*** Index of sovereign rating 0.103 0.025 [1.24] [0.31] Depreciation -0.092 -0.042 [1.24] [0.44] Observations 203 203 181 203 181 190 R-squared 0.360 0.470 0.200 0.460 0.220 0.340 Hansen J test 1.260 3.150 2.060 4.360 1.410 0.660 p value 0.740 0.370 0.560 0.230 0.700 0.880 95 percent confidence interval [0.527, [0.694, CLR 1/ [0.726, 0.747] 0.542] 0.714] [0.725, 0.745] [0.746, 0.766] [0.781, 0.804] * significant at 10%; ** significant at 5%; *** significant at Robust z statistics in brackets, observations are clustered by country 1% 1/Confidence interval robust to weak instruments (Moreira, 2003) First stage (1) (3) (4) (5) (6) (7) (excluded IVs) French legal origin -0.099 -0.171 -0.050 -0.121 -0.013 -0.063 [-1.08] [-2.08]** [-0.47] [-1.18] [-0.1] [-0.57] German legal origin -0.215 -0.154 -0.212 -0.193 -0.223 -0.192 [-2.1]** [-1.37] [-1.91]* [-1.97]** [-2.39]** [-1.63] Scandinavian legal origin 0.111 0.132 0.145 0.119 0.136 0.143 [1.32] [1.51] [1.37] [1.48] [1.59] [1.42] Socialist legal origin -0.143 -0.232 -0.107 -0.230 -0.191 -0.124 [-0.96] [-1.64]* [-0.64] [-1.69]* [-1.34] [-0.76] Partial R-squared 0.14 0.17 0.11 0.1635 0.14 0.12 F statistic (excluded IVs) 22.67 21.53 19.34 15.59 13.36 21.45 p value 0 0 0 0 0 0 - 34 - Table 10. Bank Z-Scores and BCP Compliance (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Foreign owned -0.057 -0.013 -0.103 -0.157 -0.027 -0.041 -0.072 -0.142 -0.122 -0.129 -0.095 [0.20] [0.05] [0.36] [0.49] [0.10] [0.15] [0.26] [0.48] [0.36] [0.42] [0.33] State owned 0.039 0.027 0.096 0.043 -0.023 -0.066 0.141 0.142 0.016 -0.007 -0.083 [0.19] [0.13] [0.42] [0.19] [0.12] [0.35] [0.61] [0.59] [0.07] [0.03] [0.43] Non commercial bank 0.161 0.112 0.143 0.064 0.188 0.324 0.098 0.121 0.062 0.107 0.346 [0.65] [0.45] [0.56] [0.30] [0.86] [1.45] [0.41] [0.52] [0.29] [0.56] [1.52] Total assets 0.064 0.073 0.1 0.095 0.064 0.088 0.106 0.1 0.089 0.082 0.081 [1.28] [1.94]* [2.56]** [1.34] [1.61] [2.53]** [3.24]*** [2.62]** [2.66]** [2.24]** [2.02]* Overheads / Total Assets -10.245 -9.214 -9.171 -10.36 -9.357 -9.261 -7.603 -6.454 -10.344 -9.903 -8.275 [2.20]** [1.88]* [1.85]* [1.96]* [1.92]* [1.87]* [1.38] [1.11] [2.11]** [2.05]** [1.60] Index of Rule of Law -0.143 -0.073 -0.031 -0.03 -0.275 -0.154 0.005 0.045 -0.039 -0.147 -0.14 [0.88] [0.44] [0.22] [0.20] [1.59] [1.20] [0.03] [0.31] [0.26] [0.93] [1.16] Compliance with BCPs 0.02 [1.48] Index chapter 1 1.65 0.742 [2.08]** [0.68] Index chapter 2 0.796 [0.66] Index chapter 3 0.169 [0.14] Index chapter 4 2.137 1.332 [2.63]** [1.53] Index chapter 5 2.014 1.979 [3.58]*** [3.66]*** Index chapter 6 0.108 [0.19] Index chapter 7 0.066 [0.12] Compliance, excl. cha. 1 -0.158 [0.12] Compliance, excl. cha. 4 -0.43 [0.48] Compliance, excl. cha. 5 0.192 [0.21] Observations 160 160 160 146 160 160 160 155 146 146 155 R-squared 0.16 0.18 0.13 0.14 0.19 0.23 0.12 0.12 0.14 0.15 0.23 Robust t statistics in brackets * significant at 10%; ** significant at 5%; *** significant at 1% - 35 - Appendix 1 Basle Core Principles -- Information Requirements of Banking Organizations Principle 21: Banking supervisors must be satisfied that each bank maintains adequate records drawn up in accordance with consistent accounting policies and practices that enable the supervisor to obtain a true and fair view of the financial condition of the bank and the profitability of its business, and that the bank publishes on a regular basis financial statements that fairly reflect its condition. For banking supervisors to conduct effective off-site supervision of banks and to evaluate the condition of the local banking market, they must receive financial information at regular intervals and this information must be verified periodically through on-site examinations or external audits. Banking supervisors must ensure that each bank maintains adequate accounting records drawn up in accordance with consistent accounting policies and practices that enable the supervisor to obtain a true and fair view of the financial condition of the bank and the profitability of its business. In order that the accounts portray a true and fair view, it is essential that assets are recorded at values that are realistic and consistent, taking account of current values, where relevant, and that profit reflects what, on a net basis, is likely to be received and takes into account likely transfers to loan loss reserves. It is important that banks submit information in a format that makes comparisons among banks possible although, for certain purposes, data derived from internal management information systems may also be helpful to supervisors. At a minimum, periodic reporting should include a bank's balance sheet, contingent liabilities and income statement, with supporting details and key risk exposures. Supervisors can be obstructed or misled when banks knowingly or recklessly provide false information of material importance to the supervisory process. If a bank provides information to the supervisor knowing that it is materially false or misleading, or it does so recklessly, supervisory and/or criminal action should be taken against both the individuals involved and the institution. 1. Accounting standards In order to ensure that the information submitted by banks is of a comparable nature and its meaning is clear, the supervisory agency will need to provide report instructions that clearly establish the accounting standards to be used in preparing the reports. These standards should be based on accounting principles and rules that command wide international acceptance and be aimed specifically at banking institutions. 2. Scope and frequency of reporting The supervisory agency needs to have powers to determine the scope and frequency of reporting to reflect the volatility of the business and to enable the agency to track what is happening at individual banks on both a solo and consolidated basis, as well as with the banking system as a whole. The supervisors should develop a series of informational reports for banks to prepare and submit at regular intervals. While some reports may be filed as often as monthly, others may be filed quarterly or annually. In addition, some reports may be "event generated", meaning they are filed only if a particular event occurs (e.g. investment in a new affiliate). Supervisors should be sensitive to the burden that reporting imposes. Consequently, they may determine that it is not necessary for every bank to file every report. Filing status can be based on the organizational structure of the bank, its size, and the types of activities it conducts. 3. Confirmation of the accuracy of information submitted It is the responsibility of bank management to ensure the accuracy, completeness and timeliness of prudential, financial, and other reports submitted to the supervisors. Therefore, bank management must ensure that reports are verified and that external auditors determine that the reporting systems in place are adequate and provide reliable data. External auditors should express an opinion on the annual accounts and management report supplied to shareholders and the general public. Weaknesses in bank auditing standards in a particular country may require that banking supervisors become involved in establishing clear guidelines concerning the scope and content of the audit program as well as the standards to be used. In extreme cases where supervisors cannot be satisfied with the quality of the annual accounts or regulatory reports, or with the work done by external auditors, they should have the ability to use supervisory measures to bring about timely corrective action, and they may need to reserve the right to approve the issue of accounts to the public. In assessing the nature and adequacy of work done by auditors, and the degree of reliance that can be placed on this work, supervisors will need to consider the extent to which the audit program has examined such areas as the loan portfolio, loan loss reserves, nonperforming assets (including the treatment of interest on such assets), asset valuations, trading and other securities activities, derivatives, asset - 36 - securitizations, and the adequacy of internal controls over financial reporting. Where it is competent and independent of management, internal audits can be relied upon as a source of information and may contribute usefully to the supervisors' understanding. 4. Confidentiality of supervisory information Although market participants should have access to correct and timely information, there are certain types of sensitive information27 that should be held confidential by banking supervisors. In order for a relationship of mutual trust to develop, banks need to know that such sensitive information will be held confidential by the banking supervisory agency and its appropriate counterparts at other domestic and foreign supervisory agencies. 5. Disclosure In order for market forces to work effectively, thereby fostering a stable and efficient financial system, market participants need access to correct and timely information. Disclosure, therefore, is a complement to supervision. For this reason, banks should be required to disclose to the public information regarding their activities and financial position that is comprehensive and not misleading. This information should be timely and sufficient for market participants to assess the risk inherent in any individual banking organization. - 37 - Appendix II Variable Definition Source ROAE Return over average equity Bankscope ROAA Return over average assets Bankscope NLTA Net loans, in percent of total assets Bankscope LogTA Log of total bank assets Bankscope ETA Equity in percent of total assets Bankscope LA Liquid assets / customer and short-term funding Bankscope Kaufman, Kraay and Mastruzzi Index of Rule of Law Average 1996-2002 of Index of Rule of Law (2003) Average annual depreciation of the nominal IMF, International Financial Depreciation exchange over the previous 5 years Statistics Index of sovereign rating S&P sovereign rating S&P Average annual inflation rate over the previous 5 IMF, International Financial Inflation years Statistics Average annual real credit growth over the IMF, International Financial Real credit growth previous 5 years Statistics Standard deviation of inflation over the previous IMF, International Financial Standard deviation of inflation 5 years Statistics Average annual real GDP per capita growth over IMF, International Financial Real GDP per capita growth the previous 5 years Statistics Standard deviation of growth over the previous 5 IMF, International Financial Standard deviation of growth years Statistics Index of discipline (BCL) Index of discipline Barth, Caprio and Levine (2001) Index of information disclosure (Barth, Caprio Index of disclosure (BCL) and Levine (2003)) Barth, Caprio and Levine (2001) Index auditing requirements Index of auditing requirements (Barth, Caprio and (BCL) Levine (2003)) Barth, Caprio and Levine (2001) Index lack restrictions (Caprio et Index of (lack of) overall restrictions (Barth, al.) Caprio and Levine (2003)) Barth, Caprio and Levine (2001) Index of frequency of data release - IMF Special Index Frequency SDDS Data Dissemination Standards Allum and Agca (2001) Index of timeliness of data release - IMF Special Index Timeliness SDDS Data Dissemination Standards Allum and Agca (2001) Foreign owned Dummy variable for foreign-owned banks Bankscope State owned Dummy variable for state-owned banks Bankscope English legal origin Dummy for English legal origin La Porta et al., 2002 French legal origin Dummy for French legal origin La Porta et al., 2002 German legal origin Dummy for German legal origin La Porta et al., 2002 Scandinavian legal origin Dummy for Scandinavian legal origin La Porta et al., 2002 Dummy for Socialist legal origin (La Porta et al., Socialist legal origin 2002) La Porta et al., 2002 - 38 - Figure 1. Compliance with the BCPs (by region) Mean Aggregate Index Minimum 1 Maximum 0.8 0.6 0.4 0.2 0 Transition LatinAmerica Middle East and SouthAsia Sub-saharan East Asia and Industrialized countries NorthAfrica Africa Pacific countries - 39 - Figure 2. Average Compliance with the BCPs (by Chapter) Mean Minimum Maximum 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Preconditions Licensing and Prudential Methods of on- Information Formal powers of Cross-border effective structure regulations and going supervision requirements supervisors banking supervision requirements