Efficient Auctions With Common Values

TL;DR

This paper develops an efficient auction mechanism for allocating goods with common value structures when buyer information is multidimensional, extending previous models and identifying key assumptions for efficiency.

Contribution

It introduces a new efficient auction design for multidimensional buyer information under certain valuation functions, expanding the scope of prior models.

Findings

Constructed an efficient auction mechanism for multidimensional signals.

Identified necessary assumptions for the existence of such mechanisms.

Extended the class of valuation functions where efficiency can be achieved.

Abstract

Consider the problem of allocating goods to buyers through an auction. An auction is efficient if the resulting allocation maximizes total welfare, conditional on the information available. If buyers have private values, the Vickrey-Groves-Clarke mechanism is efficient. If buyers have common values and a buyer's information can be summarized as a one-dimensional signal, Dasgupta and Maskin present an efficient auction. We construct an efficient auction mechanism in case buyer information is multidimensional, for a restricted class of valuation functions, and we prove which of the assumptions made are necessary for the existence of an efficient mechanism.