Deep Learning for Mortgage Risk
Preprint
Abstract
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 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 macroeconomic variables down to the zip-code level. The estimators uncover the highly nonlinear nature of the relationship between the variables and borrower behavior, especially prepayment. They also highlight the effects of local economic conditions on borrower behavior. State unemployment has the greatest explanatory power among all variables, offering strong evidence of the tight connection between housing finance markets and the macroeconomy. The sensitivity of a borrower to changes in unemployment strongly depends upon current unemployment. It also significantly varies across the entire borrower population, which highlights the interaction of unemployment and many other variables. These findings have important implications for mortgage-backed security investors, rating agencies, and housing finance policymakers.Keywords
All Related Versions
This publication has 48 references indexed in Scilit:
- Risk and risk management in the credit card industryJournal of Banking & Finance, 2016
- The failure of models that predict failure: Distance, incentives, and defaultsJournal of Financial Economics, 2015
- Adverse selection in mortgage securitizationJournal of Financial Economics, 2012
- The role of securitization in mortgage renegotiationJournal of Financial Economics, 2011
- Consumer credit-risk models via machine-learning algorithmsJournal of Banking & Finance, 2010
- Spillover Effects of Foreclosures on Neighborhood Property ValuesThe Journal of Real Estate Finance and Economics, 2007
- Multi-period corporate default prediction with stochastic covariatesJournal of Financial Economics, 2007
- Testing for neglected nonlinearity in time series modelsJournal of Econometrics, 1993
- Approximation capabilities of multilayer feedforward networksNeural Networks, 1991
- Multilayer feedforward networks are universal approximatorsNeural Networks, 1989