This paper has addressed the following questions: Do sovereign credit ratings systematically help predict currency and banking crises? If not, why not? What needs to change? What is the behavior of credit ratings following the crises? Are there important differences in the behavior of credit ratings between developed and emerging markets? The answers revealed by the analysis can be summarized as follows: As to the ability of rating changes to anticipate financial crises, the empirical tests presented here on sovereign credit ratings and financial crises suggest that sovereign credit ratings systematically fail to anticipate banking and currency crises. This result appears to be robust across alternative crises definitions, model specification, and approaches. Only for the Institutional Investor ratings is there some (weak) evidence that downgrades precede currency crises. In none of the cases are banking crises systematically preceded by downgrades. As regards the behavior of ratings after the crisis and differences between developed and emerging markets, there is evidence that sovereign credit ratings tend to be reactive--particularly when it comes to EMs. Both the probability of a downgrade and the magnitude of the downgrade are significantly higher for EMs. Taken together, these findings point to a procyclicality in the ratings. In a related paper (Calvo and Reinhart, 2000), also ask how these differences between developed and emerging markets in access to international capital markets influence the outcomes of a currency crisis--particularly as regards output. They present evidence that EMs are, indeed, very different from developed economies in several key dimensions. In EMs devaluations, or large depreciations for that matter, are contractionary, the adjustments in the current account are far more acute and abrupt. Hence, currency crises become credit crises as sovereign credit ratings often collapse following the currency collapse and access to international credit is lost. On why are sovereign ratings such poor predictors of financial distress, we conclude that generally, financial crises are difficult to forecast--witness the poor performance of international interest rate spreads and currency forecasts. Specifically, however, the results presented here offer a tentative (although partial) answer to this question. Rating agencies have tended to focus on the wrong set of fundamentals. For instance, much weight is given to debt-to-exports ratios--yet these have tended to be poor predictors of financial stress. Little weight is attached to indicators of liquidity, currency misalignments, and asset price behavior.
