Banking Models
Regulating risk-taking

Competition has traditionally been considered a source of excessive risk taking in banking and in consequence regulation has tried to control it. Rate regulation, entry restrictions, and charter limitations of banks (including the separation of commercial and investment banking) have been used by regulators to limit competition. In Discussion Paper No. 1177, Research Fellow Carmen Matutes and Programme Director Xavier Vives study the links between competition for deposits and risk taking in the banking sector. Their main purpose is to examine whether `excessive' competition for deposits exists. To this end, they explore how rate setting behaviour depends on the risk position of banks both with and without deposit insurance, what are the derived incentives for risk taking in the asset side, and how incentives change with rate regulation. They develop a model of banking competition which allows them to disentangle the roles that limited liability, deposit insurance (both with flat and risk-based premia), and rivalry for deposits play in determining risk-taking incentives both in the asset and the liability side of the balance sheet

The authors' main findings are as follows. First, the unregulated-uninsured market yields excessive rates when the failure costs are high. Second, flat-premium deposit insurance exacerbates rate rivalry while risk-based insurance cancels limited liability and generates lower equilibrium rates. Yet, even with risk-based insurance, equilibrium rates may be too high. Third, rate regulation is a sufficient instrument to maximize welfare both in an uninsured market where depositors are informed about banks' asset risk, or in the presence of risk-based deposit insurance. Otherwise, that is with flat-premium deposit insurance or if depositors are not informed about asset risk, rate controls and investment restrictions are complementary as opposed to substitute measures.

Imperfect Competition, Risk Taking, and Regulation in Banking
Carmen Matutes and Xavier Vives

Discussion Paper No. 1177, May 1995 (FE)