Reducing costly information acquisition in auctions

Most research on auctions assumes that potential bidders have private information about their willingness to pay for the item being auctioned, and that they use this information strategically when formulating their bids. In reality, bidders often have to go through a costly information-gathering process in order to learn their valuation for the item being auctioned. Recent attempts at modelling this phenomena has brought to light complex strategic behavior arising from information-gathering, and has shown that traditional approaches to auction and mechanism design are not able to overcome it. In this paper, we show that if the auction designer has some information about the agents' information-gathering processes, then it is possible to create an auction where, in equilibrium, agents have incentive to only gather information on their own valuation problems and to reveal the results truthfully to the auctioneer. Additionally, simulation results show that, from a system-level perspective, the overall cost of information acquisition is substantially lower in this new auction when it is compared to a classic auction mechanism.