Numerological Heuristics and Credit Risk in P2P Lending

Abstract

Heuristics as mental shortcuts have a ubiquitous influence on decision-making. Despite a large strand of literature on various heuristics, scant research addresses the question of what the choices of heuristics can tell us about the decision-makers. Using detailed peer-to-peer (P2P) investment data with more than 7.5 million bidding records of 742 thousand loan applications from Renrendai, a leading Chinese P2P lending platform, we examine two important numerological heuristics, the round-number heuristic and the lucky-number heuristic, which can be observed in over 80% of submitted loan amounts. We find these two heuristics have different implications on funding outcomes and repayment performances. Specifically, round-number loans are in general less likely to get funded and more likely to default after being funded, whereas the opposite is observed for loans with lucky numbers. We simultaneously estimate a borrower’s choice between the two heuristics using the Bivariate Probit model and find that the borrowers’ choices of heuristics reveal their credit and cognitive attributes. We show that the choice of numerological heuristics can be an indication of inferior (superior) borrower qualities in the case of round (lucky) numbers. Overall, our paper offers a framework to extract credit risk information from the heuristics that borrowers choose.

Keywords: Credit risk; Numerological heuristics; Round-number heuristic; Lucky-number heuristic; Information asymmetry; P2P lending

JEL Classification: G20, G21, G23, G40, G41, D91