Can machine learning models capture correlations in corporate distresses?

Accurate probability-of-distress models are central to regulators, firms, and individuals who need to evaluate the default risk of a loan portfolio. A number of papers document that recent machine learning models outperform traditional corporate distress models in terms of accurately ranking firms by their riskiness. However, it remains unanswered whether advanced machine learning models can capture correlation in distresses, which traditional distress models struggle to do. We implement a regularly top-performing machine learning model and find that prediction accuracy of individual distress probabilities improves while there is almost no difference in the predicted aggregate distress rate relative to traditional distress models. Thus, our findings suggest that complex machine learning models do not eliminate the need for a latent variable that captures correlations in distresses. Instead, we propose a frailty model, which allows for correlations in distresses, augmented with regression splines. This model demonstrates competitive performance in terms of ranking firms by their riskiness, while providing accurate risk measures.