Essays on late bidding in internet auctions

Abstract

OF THE DISSERTATION Essays on Late Bidding in Internet Auctions by Ying Zhang Dissertation Director: Matin K. Perry Empirical studies have documented that buyers in internet auctions tend to bid immediately before the scheduled closing time. This practice of “late bidding” is often called “sniping”. My dissertation investigates this phenomenon from both theoretical and empirical aspects. The rst chapter of my dissertation studies a second-price independent private value auction with two risk-neutral buyers who can submit a single bid either early or late. Early bids are received by the auctioneer with probability one, but late bids are received with a probability less than one. I prove that there can exist a symmetric cutoff equilibrium in which high-value buyers always bid early and low-value buyers bid late as long as their opponent has not bid early. A suf cient condition is then identi ed for the existence of the cutoff equilibrium with late bidding. Finally, I illustrate the equilibrium with examples from the beta distribution on values. The second chapter of my dissertation empirically investigates the impact of late bidding on the nal price of internet auctions. Using a dataset collected from eBay, I estimate the time after which late bidding has a differential impact on the nal price. The time is estimated as a structural break in the data using a method proposed by Bai (1997). The break point occurs approximately three minutes before the close of the auction. The empirical ndings indicate that buyers who won auctions by bidding in the last three minutes were typically more experienced than other winners and that they won the auctions for used items at lower prices.