The Microeconomic Effects of Different Approaches to Bank Supervision by Ross Levine*
Total Page:16
File Type:pdf, Size:1020Kb
Working Paper No. 237 The Microeconomic Effects of Different Approaches to Bank Supervision by Ross Levine* This Draft: December 16, 2004 Stanford University John A. and Cynthia Fry Gunn Building 366 Galvez Street | Stanford, CA | 94305-6015 * 321 19th Avenue South; Minneapolis, MN 55455; [email protected]. I thank Maria Carkovic, George Clarke, Luc Laeven, and seminar participants at the University of Minnesota and Stanford’s Center for International Development for helpful comments. 1 I. Introduction Since a large body of evidence finds that cross-country differences in banking sector development influences national rates of economic growth (Levine, 1997, 2005), it is important to understand which public policies improve the operation of banks. Yet, it is only with the creation of a new database on bank supervision (Barth, et al., 2001a) that researchers have been able to begin assessing which bank supervisory policies improve the efficient operation of banks around the world. To assess which policies enhance the functioning of banks, I evaluate two broad government approaches to bank supervision. Barth, et al. (2004, 2005) argue that national approaches to bank supervision fall along a spectrum. Some countries adopt an overall approach to bank supervision that focuses on empowering official supervision of banks, while other countries place comparatively greater emphasis on strengthening private sector monitoring of banks. They show that countries do not select individual bank policies in isolation, rather individual policies reflect broad public policy approaches to bank supervision. From this perspective, it is more useful to evaluate which broad government approaches to bank supervision enhance the functioning of banks than it is to evaluate individual policies. I follow this strategy.1 The official supervision approach stresses that market failures -- information costs and contract enforcement costs -- and government policies interfere with the incentives and abilities of private agents to monitor banks effectively.2 Thus, a powerful supervisory agency that monitors and disciplines banks directly can, in theory, improve bank operations. The official 1 An alternative strategy involves examining individual policies. For example, the Barth, et al (2001a) database contains information on over 200 bank regulations and supervisory practices. Theory provides conflicting predictions about the impact of some of these specific policies on bank efficiency. 2 For references, see Barth, et al. (2005). 1 2 supervisory approach assumes that (i) there are market failures, (ii) official supervisors have the capabilities to ameliorate those market failures by directly overseeing, regulating, and disciplining banks, and (iii) official supervisors have the incentives to fix market failures and promote the formation of banks that foster national economic prosperity (Stigler, 1971, 1975). Theory also provides a logical basis for an alternative approach to bank supervision: the private monitoring approach. Arguments in favor of the private monitoring approach include both a critique of the official supervision approach and support for private monitoring. While market failures, such as information and enforcement costs, may impede private monitoring, government failures may be so large that empowering official supervisors produces socially counterproductive results (Hamilton, et al., 1788; Shleifer and Vishny, 1998; Haber et al., 2003).3 Specifically, politicians and government supervisors do not maximize social welfare; they maximize their own welfare. From this perspective, official supervisors do not necessarily have the incentives to fix market failures. Indeed, supervisors may use their positions of power to funnel credit to connected firms, or banks may capture supervisory agencies and use the agencies to enrich bank insiders. From this perspective, the most efficacious approach to bank supervision relies on using government regulation to empower private monitoring of banks through information disclosure rules and improved contract enforcement mechanisms. This is not laissez-faire. Rather, the private monitoring approach holds that bank policies should focus on reducing information and enforcement costs, not on empowering direct official oversight of banks. 3 Some policy analysts reject the notion that there are serious market failures in banking. In a world with (i) no information or transactions costs, (ii) governments that maximize social welfare, and (iii) well-defined and efficiently enforced property rights, markets will achieve efficient outcomes (Coase, 1960). If the prerequisites for this laissez-faire – invisible hand – theory hold, government supervision of banks would be at best irrelevant and potentially harmful to social welfare. Considerable research, however, suggests that there are non-trivial information and transactions costs in the banking industry. Indeed, the mere existence of banks may reflect 2 3 To assess empirically the impact of these two approaches to bank supervision on the efficient operation of banks, I use two microeconomic measures of the efficient operation of banks. First, across a broad cross-section of countries, I use bank-level data of that ratio of each bank’s overhead costs to its total assets. While imperfect, high overhead costs can signal excessive managerial perquisites and market power that contradict the notions of sound governance of banks and efficient intermediation. As a second measure of bank efficiency, I use firm-level survey data across a large number of countries regarding the degree to which firms need corrupt ties with banks to obtain loans. Again, while imperfect, the need for corrupt ties with bank officials to obtain funds may signal a socially inefficient form of allocating bank credit. Using these measures, this paper provides empirical evidence on which broad approaches to bank supervision promote the efficient functioning of banks. To measure policies that empower official supervisors, I use the Barth, et al. (2004) indicator of official power, which aggregates information on whether bank supervisors can take specific actions against bank management, bank owners, and bank auditors both in normal times and times of distress. Supervisory agencies can use these powers to improve the governance of banks as emphasized by the official empowerment view. Alternatively, the supervisory authority can use these powers to induce banks to funnel credit to favored ends as emphasized by the private empowerment view. Thus, the degree to which empowering the supervisory authority improves the functioning of banks is an empirical question. To measure policies that empower private monitoring of banks, I use the Barth, et al. (2004) indicator of private monitoring, which includes information on the degree to which bank regulations force banks to disclose accurate information to the public. The private empowerment information and transactions costs in credit markets. More importantly for the purposes of this paper, the empirical work below suggests that laissez faire is not the most successful approach to bank supervision around the world. 3 4 view holds that improved private governance of banks will boost the functioning of banks. In contrast, the official empowerment view holds that market imperfections in banking are so substantive and pervasive that strengthening information disclosure laws will not materially improve banking operations. Again, sound theories bolster both sides of the debate. By assessing empirically the impact of the official supervisory power and the private monitoring indicators on bank efficiency and the integrity of bank-firm relations, I provide evidence on which supervisory approach is associated with better functioning banks. While I have stressed differences between the official and private empowerment approaches, these supervisory approaches are not mutually exclusive. Countries could adopt regulations that force banks to disclose accurate information to the public, while also creating official supervisory agencies that directly oversee bank activities. Conceptually, however, these approaches reflect different attitudes toward the role of government in monitoring banks. Indeed, differences in the degree of emphasis on empowering private monitoring of banks may represent a particularly useful litmus test of different national attitudes toward the role of government in the economy. Empirically, if the official power and private monitoring indicators are highly correlated, then I will observe multicollinearity in the regressions. Alternatively, if it is useful to classify bank supervisory approaches in terms of official empowerment versus private empowerment, then I will not observe multicollinearity and will instead observe that the official power and private monitoring indicators yield distinct results. Methodologically, I use two distinct datasets, different econometric procedures, and different definitions of “better functioning banks.” First, I use a pooled bank-level, cross-country database and employ a generalized least squares random effects estimator to assess which supervisory strategies lower bank overhead costs. The paper uses microeconomic data to 4 5 minimize the possibility that reverse causality biases the results. Nevertheless, it is crucial to emphasize that bank overhead costs are an imprecise measure of bank efficiency, which advertises the importance of (i) controlling for other bank-specific characteristics and (ii) using alternative