Public Infrastructure Procurement: Detecting Collusion in Capped First-Priced Auctions

Abstract

Worldwide, numerous economies have been subjected to the unnecessary economic and social impacts that materialize from the collusion that has taken place when public infrastructure has been procured. In an attempt to mitigate this collusion, the Brazilian government introduced laws (e.g., Law 8666/93) that require the procurement of public goods and services to be subjected to capped first-price auctions. For public infrastructure projects, bids are limited by median market prices that are recorded in specific databases and relied on during the selection of bidders. We examine 187 capped first-price auctions with eight or more bids. We reveal that full collusion can be detected if the collective behavior of any auction’s bidders is compared to a “but for” scenario that expresses the expected behavior of a group of honest competitors bidding randomly. This comparison makes it possible to flag, using different confidence levels, whether participants in a particular auction fully colluded at the time the bids are known. The contribution of our approach is twofold. It can be used independently or in conjunction with other existing methods to detect the presence of collusive behavior in a capped first-price auction.