Working Paper: CEPR ID: DP14500
Authors: Marta Reynal-Querol; Jose G. Montalvo
Abstract: In this paper we analyze the effect of loan officers' gender on the approval of loans and, in particular, on their subsequent performance. Using detailed bank information on a sample of close to half a million loans, we show that female loan officers have, conditional on the risk score, around a 15\% lower delinquency rate than that of male officers. In addition to the original scoring of the loans, we also have the recommendation of the expert system. We find that the risk profile of applicants screened by male and female loan officers is very similar, but conditional on risk score, women follow the recommendations more often than men. Moreover, we find evidence of gender bias in terms of a mistake-punishment trade-off, which could explain, at least in part, women's higher compliance with the recommendations. Indeed, there is a double standard in terms of the consequences for breaking the rules: errors, in the form of delinquent loans as a result of not following the recommendation of the system, are forgiven more often for male than for female loan officers.
Keywords: gender; credit risk; loan officers
JEL Codes: G21; G32; J16
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Female loan officers (J16) | Lower delinquency rate (G33) |
| Female loan officers (J16) | Higher compliance with bank recommendations (G21) |
| Higher compliance with bank recommendations (G21) | Lower delinquency rate (G33) |
| Gender bias in mistake-punishment tradeoff (J16) | Higher compliance with bank recommendations (G21) |
| Male loan officers (J79) | Higher delinquency rate (G33) |
| Loan officer gender (J79) | Management of credit risk (G21) |