Analyzing the Tradeoff Between Ratings Accuracy and Stability

Abstract

Credit rating agencies state that their clients prefer ratings that are stable as well as accurate. While accuracy could in some cases be increased by tracking market-based credit signals more closely, the required increase in rating volatility might be substantial. In this article, we discuss why users of the rating system may value both accuracy and stability and why a tradeoff between the objectives may be unavoidable. We illustrate the potential tradeoff by comparing the combinations of the accuracy and stability associated with the many different “rating systems” that can be derived by applying various filters to Moody's-KMV's expected default frequencies (“EDFTM's). The performance of Moody's traditional ratings—both adjusted and unadjusted for rating reviews and outlooks—is then compared to the performance of the various filtered versions of EDF-implied ratings.