We present a prediction model to forecast corporate defaults. In a theoretical model, under incomplete information in a market with publicly traded equity, we show that our approach must outperform ratings, Altman's Z-score, and Merton's distance to default. We reconcile the statistical and structural approaches under a common framework, i.e., our approach nests Altman's and Merton's approaches as special cases. Empirically, we cannot reject the superiority of our approach.Furthermore, the numbers of observed defaults align well with the estimated probabilities. Finally, with rank transforms, we obtain cycle-adjusted forecasts that still outperform ratings.
AbstractEnding the dependence on rating agencies is a top priority for the Financial Stability Board, which coordinates the G20 financial policies. Rating agencies have been accused of contributing to the recent financial crisis by misjudging the creditworthiness of mortgage-backed securities. Their downgrading practice of sovereigns and corporations is said to amplify procyclicality and to fuel market uncertainty. However, it is one thing to criticize rating agencies, quite another to find an appropriate alternative. For the case of corporate issuer ratings, we show that there is at least one such alternative. From a statistical viewpoint, it outperforms credit ratings. From a practical viewpoint, it is straightforward to implement and depends solely on publicly available data. Therefore, we conclude that for the case of corporate credit risk, appropriate alternatives to credit ratings do exist! JEL Classification Codes: G01, G18, G24