Screening Peers Softly: Inferring the Quality of Small Borrowers

This paper examines the performance of new online lending markets that rely on non-expert individuals to screen their peers’ creditworthiness. We find that these peer lenders predict an individual’s likelihood of defaulting on a loan with 45% greater accuracy than the borrower’s exact credit score (unobserved by the lenders, who only see a credit category). Moreover, peer lenders achieve 87% of the predictive power of an econometrician who observes all standard financial information about borrowers. Screening through soft or nonstandard information is relatively more important when evaluating lower-quality borrowers. Our results highlight how aggregating over the views of peers and leveraging nonstandard information can enhance lending efficiency. JEL codes: D53, D8, G21, L81