Machine learning methods have achieved great success in modeling complex patterns in finance such as asset pricing and credit risk that enable them to outperform statistical models. In addition to the predictive accuracy of machine learning methods, the ability to interpret what a model has learned is crucial in the finance industry. We address this challenge by adapting interpretable machine learning to the context of corporate bond recovery rate modeling. In addition to the best performance, we show the value of interpretable machine learning by finding drivers of recovery rates and their relationship that cannot be discovered by the use of traditional machine learning methods. Our findings are financially meaningful and consistent with the findings in the existing credit risk literature.

We study the relationship between U.S. corporate bond recovery rates and macroeconomic variables used in the credit risk literature. The least absolute shrinkage and selection operator (LASSO) is used in selecting macroeconomic variables. The LASSO-selected macroeconomic variables are considered to be explanatory variables in ordinary least squares regressions, bootstrap aggregating (bagging), regression trees, boosting, LASSO, ridge regression and support vector regression techniques. We compare the out-of-sample predictive power of two types of models (LASSO-selected models with models that add principal components derived from 179 macroeconomic variables as explanatory variables). We find the recovery models with LASSO-selected macroeconomic variables outperform suggested models in the literature.