Stress tests by an informed regulator
Abstract This article studies the disclosure of stress test results by a regulator who is privately informed about bank health when she chooses the stress scenario. I show that the regulator’s choice of the stress scenario depends on how heterogeneous health is across banks. There can be fewer runs than under transparency. However, there are more runs than when the regulator chooses the scenario before becoming informed, highlighting a time-inconsistency problem. Moreover, disclosure can become a source of informational contagion, as changes in the health of one bank affect beliefs about other banks. The model explains the empirical puzzle that a bank’s share price can fall even though it passes the stress test.