Pricing Mechanism in Online Credit Markets

Abstract

We compare two prominent selling mechanisms in online platforms, auctions and posted prices, in the context of Prosper.com, an online peer-to-peer lending marketplace, which switched from auctions to posted prices. We first develop a predictive model using Random Forests to approximate the pricing rule that Prosper.com uses for the posted pricing, which allows us to predict the interest rates for the borrowers who listed under the auctions, if Prosper.com were to use the posted pricing. We find that the posted pricing leads to higher interest rates than the auctions. Using the predicted interest rates and the estimates of the structural model from Kawai, Onishi and Uetake (2018), which estimate the model under the auctions, we simulate the equilibrium outcomes under the posted pricing. We find the funding probabilities are lower under the posted prices than the auctions. This can be attributed to either adverse selection due to the lack of signaling in posted prices or ex-post moral hazard due to higher interest rates in the posted prices. By simulating also the posted price mechanism under symmetric information, we find both effects. Our results imply the increase in funding probabilities observed in the data after the mechanism change mainly results from the change in the composition of borrowers and lenders.