Learning to bid: An experimental study of bid function adjustments in auctions and fair division games

We examine learning behavior in auctions and Fair division games with independent private values under two different price rules, first and second price. Participants face these four games repeatedly and submit complete bid functions rather than single bids. This allows us to examine whether learning is influenced by the structural differences between games. We find that within the time horizon which we investigate, learning does not drive toward risk neutral equilibrium bidding and characterize some features of observed learning: Bid functions are adjusted globally rather than locally, decision time matches the sequencing structure of game types, game rules do matter, and directional learning theory offers a partial explanation for bid adjustments. The evidence supports a cognitive approach to learning.