Reinforcement learning and mortgage partial prepayment behavior

Abstract

An agent can learn from previous experience to make decisions. Several important studies claim that reinforcement learning plays a key role in explaining the evolution of the individual learning process. This paper studies the likelihood of mortgage partial prepayments and the process through which mortgage borrowers learn from making partial prepayment decisions in the residential mortgage market in China. Learning dynamics are measured by studying the mortgage partial prepayment behavior of individual borrowers. A longitudinal discrete choice model of the choice of the mortgage payment is presented and estimated using a rich set of mortgage loan history data from a leading mortgage lender in China. The results indicate that path dependency and reinforcement learning arise whenever a borrower's “partial prepayment” decision depends not only on current-stage variables and his/her individual characteristics but also on the learning experience (both from himself/herself and others). Borrowers with more partial prepayment experience in previous stages have a higher probability of making the same decision in the future. Moreover, learning dynamics are not monotonic, and recent experience plays a larger role than distal experience in determining a partial prepayment decision.