This report analyzes loan payment deferment as a result of COVID-19 related shutdowns in the US. We focus on a portfolio of unsecured consumer loans originated by 2 different institutions. Our analysis focuses on a few key questions:
- what is the magnitude of COVID-related deferment?
- are there systematic relationships between loan attributes and payment deferment?
- how are labor market trends related to the probability of loan deferment?
- does the sensitivity to labor market shocks vary by region?
The model and results presented provide a general framework that can be applied not only to unsecured consumer loans but also more broadly to other lending sectors. While the data are still preliminary and the events they capture relatively recent, our conclusions are based on a rigorous and transparent statistical analysis and presented with confidence bounds that respect the intrinsic uncertainty of the data-generating process. The chief contribution of this paper, in terms of techniques, is the use of a mixed-model with random effects within a bayesian estimation framework which has enabled us to answer some of the questions we posed earlier and which would not have been possible using more traditional approaches.
This study should be useful to investors and policy-makers alike, allowing for data-driven estimates of potential deferment (and distress) rates on loan portfolios.