75 Pages Posted: 23 Jun 2016 Last revised: 5 Oct 2017
Date Written: October 4, 2017
We develop a deep learning model of multi-period mortgage risk and use it to analyze an unprecedented dataset of origination and monthly performance records for over 120 million mortgages originated across the US between 1995 and 2014. Our nonparametric estimators of term structures of conditional probabilities of prepayment, foreclosure and various states of delinquency incorporate the dynamics of a large number of loan-specific as well as economic and demographic variables at national, state, county and zip-code levels. The estimators highlight the importance for mortgage risk of local economic conditions, in particular state unemployment. The behavior of mortgage risk can vary strongly depending upon the geographic region. Moreover, the relationship between factors and mortgage risk is often highly nonlinear. Higher-order interactions between multiple factors are prevalent. By incorporating these nonlinear effects, our deep learning estimators offer superior out-of-sample predictions of multi-period mortgage risk at loan- and pool-levels and enable the selection of performing mortgage investment portfolios.
Keywords: Deep Learning, Machine Learning, Mortgages, Loans, Credit Risk
Suggested Citation: Suggested Citation
Sirignano, Justin and Sadhwani, Apaar and Giesecke, Kay, Deep Learning for Mortgage Risk (October 4, 2017). Available at SSRN: https://ssrn.com/abstract=2799443