Identification and Estimation of Affiliated Private Values Auctions with Unobserved Heterogeneity

I consider a sealed-bid first-price auction model that allows for affiliation in the private values and auction level unobserved heterogeneity. Both affiliation and unobserved heterogeneity manifest in bids that are correlated... [ view full abstract ]

I consider a sealed-bid first-price auction model that allows for affiliation in the private values and auction level unobserved heterogeneity. Both affiliation and unobserved heterogeneity manifest in bids that are correlated and it is hard to disentangle these two sources apart from bid data alone. I show that the model is nonparametrically identified once we allow for endogenous participation and leverage entry data combining ideas from the control function and measurement error literatures. I propose a nonparametric estimator and show how one can test for different information structures: IPV with and without unobserved heterogeneity and APV with and without unobserved heterogeneity. I then take the model to data from highway construction procurement auctions from California. I find evidence of both unobserved heterogeneity and affiliation.