International propagation of the credit crisis: lessons for bank regulation

The authors use a large sample of non‐U.S. banks to examine the origins and spread of the 2007–2009 crisis. Using both stock market and structural variables, they test whether the effects of the crisis on individual banks are better explained by crisis models or by the VaR‐type analysis of the Basel system. The latter emphasizes risk weightings for individual assets while ignoring linkages that could leave banks exposed to systemic shocks. Consistent with crisis models, the authors find that a small set of pre‐crisis measures of a bank's international linkages, leverage, and the fragility of its liability structure does a good job of discriminating between banks that suffered a large impact and those that did not. (Indeed, these measures explain almost 50% of the differences among banks' stock returns during the crisis period, and almost 40% of the changes in the variability of those returns.) The authors also provide evidence of both a direct linkage among banks' stock returns and an indirect linkage that could reflect either linkages in the real economy or common demands by investors for liquidity. The authors run a “horse race” that demonstrates that simple measures of book leverage were better predictors of bank performance than the Basel capital ratios. They find that banks with lower Basel risk weightings prior to the crisis proved, on average, to be more exposed to the crisis. The authors' explanation is that banks with lower Basel risk measures tended to operate with higher leverage and more aggressive funding strategies, which in turn exposed them to greater crisis risk (even as they conformed to the letter of the Basel system in terms of asset risk measures). Finally, the authors find no evidence that substandard governance was a separate contributing factor to crisis exposure. Banks with substantial international business that were exposed to systemic shocks had high governance scores