Efficient Sequential Auctions for Mixed Differentiation

Abstract

We study sequential auctions for allocating multiple heterogeneous objects with mixed differentiation to unit-demand agents. Objects can be divided into two groups: in one, valuations are strictly increasing in type; in the other, strictly decreasing. If within each group the objects can be ordered in decreasing variation of values, then any combination of static standard auctions implemented in this sequence achieves full efficiency. Moreover, the principal's expected revenue equals the sum of the constrained optimal revenues for the two groups. These results highlight the advantage of sequential auctions (over simultaneous auctions) in achieving efficiency while still generating reasonable revenue.

Keywords: Sequential Auction, Mixed Differentiation, Efficiency, Revenue

JEL Classification: D44, D82