Semi-nonparametric estimation of secret reserve prices in auctions

Abstract

This letter proposes an estimation method for the distribution of unobservable secret seller reserve prices in auctions. Within the symmetric affiliated private values model, I construct a density estimator using the semi-nonparametric quasi maximum likelihood approach developed by Gallant and Nychka (1987). I then use Monte Carlo experiments to demonstrate its performance in relatively small samples. A discussion on applying this method to structurally similar identification problems is provided. • This letter offers an approach to identify the distribution of secret reserve prices. • The model allows for affiliated private values among buyers and sellers. • Input data include the auction’s paying bid and whether the item sold at said bid. • A semi-nonparametric estimator identifies the distribution of secret reserve prices. • This method works well in relatively small samples ( ∼ 100 ).