Search and Wholesale Price Discrimination

Abstract

Many markets for homogeneous goods feature market power and heterogeneity in the prices paid by buyers. Search costs are a common explanation for this phenomenon and are a concern as they generate inefficiencies. In this paper, I study a competitive market for homogeneous goods and, by exploiting a unique dataset, I find three facts that are opposite to what one would expect from a market with these characteristics. First, sellers enjoy market power. Second, one can find customers paying 50 or 60% more than others for the same product at the same day. Third, price differences are systematic at the buyer level, providing evidence that sellers actively practice price discrimination. Inspired by these facts and by evidence supporting search costs as the source of market power, I propose and estimate a structural search model for two purposes. First, to measure how the market power generated by search costs affects welfare and, second, to study how price discrimination may magnify or reduce the welfare effects of search costs by altering competition intensity. My results address two important issues. First, search costs imply price distortions that generate a loss in total surplus that is about two-thirds of the welfare loss when shifting from perfect competition to monopoly. That is, even for a competitive market for homogeneous goods, search costs can have a severe effect on welfare. Second, price discrimination increases total surplus by as much as six percent relative to when sellers set uniform prices. The increase in welfare can be partially explained by price discrimination increasing search incentives and, hence, intensifying competition. ∗Department of Economics, Northwestern University; e-mail: g-marshall@u.northwestern.edu. Acknowledgements: I am especially grateful to Igal Hendel for his guidance and support. I am also grateful to Aviv Nevo and Rob Porter for their help and support. I also thank Germán Bet, Laura Doval, José Esṕın, Aanchal Jain, Chris Lau, Fernando Luco, Álvaro Parra, Esteban Petruzzello, Tiago Pires, Anthony Wray, and Jaber Zarezadeh for helpful suggestions and conversations, and seminar participants at Northwestern University and Pontificia Universidad Católica de Chile. All mistakes are my own.