A regulatory stress test to-do list: Transparency and accuracy

Abstract

Regulatory stress tests use forecasts of bank performance over hypothetical multi-year stress scenarios to set minimum regulatory capital requirements. Notwithstanding their popularity among US and European regulators, little is known about the accuracy of supervisory stress test models. Regulators keep the details of their stress test models confidential, inhibiting an independent verification of regulatory model performance. Using data from the 2008 financial crisis, this paper demonstrates how the lack of transparency and model inaccuracy have the potential to create costly misallocations of banking resources.