Robust Bidding in First-Price Auctions: How to Bid without Knowing What Others Are Doing

Abstract

We propose how to bid in first-price auctions when a bidder knows her value but not how others will bid. To do this, we introduce a methodology for how to make choices in strategic settings without assuming common knowledge or equilibrium behavior. Accordingly, first eliminate environments that are believed not to occur and then find a robust rule that performs well in the remaining environments. We test our bids using data from laboratory experiments and the field. We find that our bids outperform those made by real bidders.

Keywords: robust decision-making, first-price auction, minimax loss, empirical loss, non-equilibrium model

JEL Classification: D44, D81, C57