Mortgage Default: Classification Trees Analysis

We introduce the powerful, flexible and efficient nonparametric Classification and Regression Trees (CART) algorithm to the analysis of mortgage default data while conducting the first academic study of mortgage default in Israel. CART’s strengths in dealing with large data sets, high dimensionality, mixed data types, missing data, different relationships between variables in different parts of the measurement space, and outliers, is particularly appropriate for our data set. Moreover, CART is intuitive and easy to interpret. We find that borrowers’ features rather than mortgage contracts features are the strongest predictors of default if accepting “bad” borrowers is more costly than rejecting “good” ones. If these costs are equal, mortgage features are used as well. The higher (lower) the ratio of misclassification costs of bad risks versus good ones, the lower (higher) are the resulting misclassification rates of bad risks and the higher (lower) are the misclassification rates of good ones. This is consistent with real world stylized facts of rejection of good risks in attempt to avoid bad ones. JEL Codes: C12, D12, G21, R29