Success-Driven Skill Inferences and Financial Crises

This paper develops a theory of financial crises that explains why crises should be expected to follow periods of sustained banking profitability. When there is some probability that outcomes are purely exogenous and some probability that they are driven by the talent/skill of the agents managing the assets that are associated with these outcomes, periods of sustained profitability for banks cause all agents - banks as well as those who monitor and fund banks - to revise upward their estimates ofthe skill of bankers to manage risks. All agents consequently become sanguine about the consequences of bank risk-taking, and banks invest in increasingly riskier assets. Further, as the number of banks entering this market grows, the liquidity of highly risky assets improves, making it more attractive for others to enter the market. The economy thus ends up with a large number of financial institutions investing in very risky assets that are traded in highly liquid markets. Subsequently, if it is revealed in some period that loan repayment probabilities are exogenous rather than skill dependent, investors rush to withdraw funds, market liquidity dries up, and a crisis ensues. The model also explains why the economy may not recover even after the friction that precipitated the crisis has dissipated. Empirical predictions and policy implications of the analysis are discussed. The analysis suggests that the majority of regulations put in place in response to the crises of the last few decades, including the recent financial crisis, may do little to prevent future crises.