Detecting Bidders Groups in Collusive Auctions

Francesco Decarolis (Boston University)

Abstract: This paper studies entry and bidding in procurement auctions were contracts are awarded to the bid closest to a trimmed average bid. We characterize equilibrium under competition and show that it is weak due to strong incentives for cooperation. We present statistical tests motivated by a model of coalition entry and bidding. We show that our tests perform well in a validation dataset with known cartels. We also use them to investigate cooperation in a larger dataset where cartels are suspected but not known. We detect several suspiciously cooperative groups with potentially substantial, positive effects upon auctioneers revenues.


Download the paper