Estimating delinquency migration and the probability of default from aggregate data

Abstract

Defaulting on a mortgage represents the ultimate consequence of past decisions to delay payment. While many modeling approaches are available to estimate the probability of default, most if not all require account‐level data. Further, past research has not attempted to estimate the probability that a current loan will transition among delinquency states prior to default. In this paper, we present an econometric approach that makes use of publicly available aggregate data for estimating the probability of delinquency and the probability of default. The results suggest the approach may have merit for monitoring bank performance as well as usefulness for banks’ risk management efforts.