Strategic Selection of Risk Models and Bank Capital Regulation

Abstract

The regulatory use of banks' internal models makes capital requirements more risk-sensitive but invites regulatory arbitrage. I develop a framework to study bank regulation with strategic selection of risk models. A bank supervisor can discourage arbitrage by auditing risk models, and implements capital ratios less risk-sensitive than in the first-best to reduce auditing costs. The optimal capital ratios of a national supervisor can be different from those set by supranational authorities, in which case the supervisor optimally tolerates biased models. I discuss the empirical implications of this hidden model problem, and policy answers such as leverage ratios and more reliance on backtesting mechanisms.