自动贷款审批是否减少了中小企业贷款中的性别偏见？

11361 Does Automatic Loan Approval ReduceGender Bias in SME Lending? Irani ArraizMiriam BruhnClaudia Ruiz-Ortega Policy Research Working Paper11361 Abstract This paper examines whether automatic credit approvalcan reduce gender bias in lending to small and medi-um-sized enterprises. The study team collaborated with abank in Peru that piloted a new screening tool to generatepsychometric credit scores. Applicants who scored abovea predefined threshold were automatically offered a loan loan offers and loan sizes of female and male applicantswithin a narrow window around the automatic approvalthreshold. The results show that female applicants belowthe threshold are less likely to take out loans and receivesmaller loan amounts than men. However, this gender biasdisappears for applicants above the threshold, suggesting This paper is a product of the Development Research Group, Development Economics. It is part of a larger effort by theWorld Bank to provide open access to its research and make a contribution to development policy discussions around theworld. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors maybe contacted at mbruhn@worldbank.org and cruizortega@worldbank.org. A verified reproducibility package for this paper The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about developmentissues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry thenames of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those Does Automatic Loan Approval Reduce Gender Bias in SME Lending? Irani Arraiz, Miriam Bruhn, and Claudia Ruiz-Ortega(IDB Invest, The World Bank) Keywords: Gender bias, SME lending, financial inclusionJEL codes: G21, G23, G41, J16, L26, O12 1. Introduction Loan applications from women-owned firms are often less likely to be approved than thosefrom men-owned firms, even after controlling for firm characteristics (Muravyev, Talavera, andSchäfer, 2009; Morazzoni and Sy, 2021; Presbitero, Rabellotti and Piras, 2014). Among firms thatdo receive loans, women-owned businesses tend to face higher interest rates (Alesina, Lotti, andMistrulli, 2013; Chen, Li, and Lai, 2017; Mascia and Rossi, 2017; Muravyev, Talavera, and Some of these observed gender gaps may be driven by omitted variables. However,evidence from audit studies that randomly vary gender in loan applications while holding othercharacteristics constant also reveals gender disparities in loan approval rates (Montoya et al., We study whether the gender gap in SME lending decreases when using an automatic loanapproval procedure that limits loan officer discretion. We collaborated with a large commercialbank in Peru that piloted a new screening tool in 2012 and 2013. SME owners who applied forworking capital loans were assessed using a psychometric tool, which generated a three-digitpsychometric credit score. Applicants who achieved a score above a predefined threshold, set by In this setting, if loan officers discriminate against women applicants, we would expectloan approval rates or loan sizes to be less favorable for women than for men below the automatic Wetest these hypotheses using a regression discontinuity design(RDD)withadministrative data from our partner bank and credit history data from Equifax Peru. About 50percent of the loan applicants in our sample are women. To assess gender bias in loan offers andloan sizes, we run OLS regressions comparing outcomes across women and men, below and above Our results show gender bias below the threshold, where loan offers and terms depend onloan officer decisions. Here, female applicants have the same probability of being offered loans asmale applicants. However, controlling for various background characteristics, female applicants Above the threshold, where loans are automatically approved and loan size depends on thepsychometric score, we find no statistically significant difference in approval rates, loan sizes, andoffer acceptance rates across men and women. Our findings thus suggest that automatic loan A key question is whether automatic approval leads to worse loan repayment since unlikealgorithms, loan officers can gather soft information on borrowers, which may be particularly Our paper contributes to the literature on automated loan decisions and discrimination. Somestudies for the U.S. have shown that loan approval processes with less human involvement canreduce taste-based discrimination against borrowers from racial minorities (Fei, 2021; Howell et Garcia, Garcia, and Rigobon (2024) point out that algorithms that are based on past decisionsrecorded in financial institutions’ datasets often consolidate existing biases against groups definedby race, sex, sexual orientation, and other attri