Increasing Competition and the Winner's Curse: Evidence from Procurement

We empirically measure the effects increasing competition on equilibrium bidding in procurement auctions. In common value auctions, the winner's curse counsels more conservative bidding as the number of competitors increases. First, we estimate the structural parameters of an equilibrium bidding model and test for the importance of common value components in bidders' preferences. Second, we use these estimates to measure the effects on increasing competition on both individual bids as well as winning bids (ie. procurement costs). Methodologically, we have estimated a model which allows bidders' latent valuations for a contract to have both common and private value components. We analyze bid data from construction procurement auctions run by the New Jersey department of transportation in the years 1989-1997. Our results show that different types of contracts differ significantly in the degree that private and/or common value components are important, and these have contrasting implications on the effects of increasing competition on equilibrium bids and expected government procurement outlays. Auctions for highway work contracts are very close to a pure common value auction, while both common value and private value are important in auctions of bridge repair contracts. Furthermore, our results indicate that the winner's curse is particularly strong in these highway contract auctions. Simulated bid functions show that for the median bidder, the percentage markup increases from 50% with 2 bidders to above 70% with 10 bidders. Furthermore, winning bid simulations indicate that the average procurement cost is strictly increasing in the number of bidders as competition intensifies: for example, the median costs rise about 30%, as the number of bidders is increases from 3 to 6. These results emphasize how asymmetric information can overturn the common economic wisdom that more competition is always desirable, and have potentially important policy implications.