This paper shows that in online car auctions, resellers are better at appraising the value of the cars they are bidding on than are consumers. Using a unique data set of online car auctions, I show that differences in bidding behavior between resellers and consumers can be explained by heterogeneity in the accuracy of bidders’ private signals and heterogeneity in the dispersion of private value components.
I use the asymmetric ascending auction model of Hong and Shum (2003) to quantify the differences between resellers and consumers, finding that the dispersion of reseller value signals is roughly half that of consumers and simulate three different counterfactual scenarios - one in which consumers are provided with more information, one in which consumers are subsidized and one in which consumers are allowed into all-reseller auctions.
Finally, I argue that the asymmetry in signal precision stems not from asymmetric information regarding the technical characteristics of a car but rather from uncertainty about the car’s resale value.
