Private firm default probabilities via statistical learning theory and utility maximization

Xuelong Zhou, Jinggang Huang, Craig Friedman, Robert Cangemi, Sven Sandow

Abstract

ABSTRACT

We estimate real-world private firm default probabilities over a fixed time horizon. The default probabilities are conditioned on a vector of explanatory variables which include financial ratios, economic indicators and market prices. To estimate our model we apply a recently developed method from statistical learning theory. This method leads to a model that is particularly appropriate for financial market participants, who would use the model to make financial decisions. We compare our model with various benchmark models with respect to a number of performance measures. In all these tests our model outperformed the benchmark models. We discuss possible reasons for this outperformance.