Screening in Alternative Credit Markets Can Individual Lenders Infer Borrower Credit-worthiness in Peer-to-Peer Lending?

The past few years have seen a large growth in online peer-to-peer credit markets. Yet the sustainability of these alternate credit markets crucially depends on the ability of lenders to infer borrowers’ credit-worthiness. We examine this ability using a methodology that takes advantage of lenders not observing a borrowers’ true credit score but only seeing an aggregate credit category in the particular market considered. We find that lenders are able to utilize information available to infer up to a third of the relevant content in a borrower’s credit score. This is economically significant and allows them to lend at a 200 basis points lower rate for borrowers with (unobserved to them) better credit scores within a category. While lenders infer the most from standard “hard” information that is traditionally used by banks for screening, they also use non-standard (subjective) information for inference. Our methodology shows, without needing to code information contained in pictures or personal descriptions a borrower posts, that lenders learn even from such “softer” information, particularly when it is likely to provide credible/costly signals regarding borrower credit worthiness. Our findings highlight that peer-to-peer markets do show a degree of efficiency and suggest they may even provide advantages over traditional lending markets by extracting information from non-traditional sources. JEL codes: D53, D8, G21, L81