Does a financial crisis change a bank's exposure to risk? A difference-in-differences approach

Can a major financial crisis trigger changes in a bank's risk-taking behavior? Using the 2008 Global Financial Crisis as a quasi-natural experiment and a difference-in-differences approach, I examine whether the worst crisis-hit Russian banks - the banks that have strong incentives to behavior-altering changes - can decrease their post-crisis exposure to risk. A shift in risk-taking behavior by these banks indicates the learning hypothesis. The findings are mixed. The evidence concerning credit risk is inconsistent with the learning hypothesis. On the other hand, the evidence concerning solvency risk is consistent with the learning hypothesis and corroborates evidence from the Nordic countries (Berglund and Mäkinen, 2019). As such, bank learning from a financial crisis may not depend on the institutional context and the level of development of national financial market. Several robustness checks with alternative regression specifications are provided.