[Link to paper] 

Abstract: Stress tests disclose valuable information to dissuade stakeholders from running on a bank. We study the role of frequency as a critical policy lever in the design of stress tests. We develop a dynamic bank run game and show that conducting more tests makes it easier for stress tests to be credible. A test is credible when a bank never fails when it passes any single test, even in the worst equilibrium. We find the optimal credible stress test and provide key policy implications on how evolving market conditions affect the optimal frequency. We also show how complementary regulatory measures can enhance the effectiveness of the stress test framework.

Presentation (by one of the authors): 7th Bristol Workshop on Banking and Financial Intermediation, Bank of England Financial Stability Seminar, Chinese Finance Research Conference (CFRC), Econometric Society meetings, Fudan Workshop on Economic Dynamics, Stony Brook Game Theory meetings,

Kelley School of Business, and SWUFE