Predicting serial credit rating downgrades

This paper examines the predictability and implications of sequential credit rating downgrades under the through-the-cycle rating methodology. Using 28,847 firm-year observations for North American corporates, we show that firms partway through downgrade sequences exhibit weaker fundamentals than non-downgraded peers with the same rating, implying interim ratings misstate credit risk. Predictive models of future downgrades achieve up to 75 % accuracy, with out-of-sample tests and feature-importance measures confirming robustness. A simple portfolio exercise illustrates economic significance. Our findings extend the literature on rating inertia and highlight the usefulness of credit-quality signals for anticipating rating actions.