First Impressions in a Sequential Auction

Should an informed seller lead with the best or worst good in a sequential auction? Considering the sale of two stochastically equivalent goods over two periods, we show that if second period buyers can observe the rst period price, the seller has an incentive to lead with the best good so as to send a positive signal about the quality of the following good. This result holds even though the goods' values are independent because the seller's sequencing strategy endogenously generates correlation in the quality of the goods across periods. In contrast, a best for last strategy may not be as credible as the seller has an incentive to then sell his better good early. We also show that ex-ante expected pro ts from either of these strategies is higher than a babbling strategy of randomly sequencing the sale, even when the second period buyers do not observe the rst period price.We discuss implications for the choice of sequential versus simultaneous auctions, the strategic choice of auction houses, the sequential auction of items of varying expected quality, the declining price anomaly observed in auction data, and the e ects of selection bias on empirical studies of privatization auctions. Preliminary and Incomplete. Please do not quote without authors' permission. Baruch College, City University of New York, archishman [email protected] University of Pittsburgh, [email protected] Claremont McKenna College and Claremont Graduate University, rick [email protected]