Estimating damages from bidding rings in first-price auctions

Bidding rings typically coordinate to rig auctions and keep prices low. Despite bid rigging being pervasive, measuring its damages (i.e., the revenue loss suffered by the auctioneer) is a challenge for antitrust authorities. Indeed, most of the methods to quantify damages compare outcomes of auctions affected by the collusive behavior with unaffected auctions, requiring data that is hard to obtain. We propose a model-based method to estimate damages. Its main advantages are that only information on affected auctions is required and that the underlying assumptions of the economic model are explicit, so they can be challenged and eventually modified for damage reassessment. In a Monte Carlo exercise, we show that our methodology performs well in moderate-size samples. We apply our method to data from the Ohio milk cartel and estimate damages similar to those found in previous studies, even when we discard information from non-affected markets.