Asymmetric Effects of Winning and Losing Experience in Multi-Unit Auctions

This study empirically links current behaviour with past performance in a competitive multi-unit auction setting to test extensions of the learning-by-doing hypothesis. A particular advantage of the data set is that it contains complete bidding histories from a multi-unit auction that recently underwent a design change from a uniform to a discriminatory auction format. Characteristics of this unique natural experimental setting ensures bidders were inexperienced in the discriminatory auction format and that experience at the individual bidder level can be accurately measured. Following the change in auction format, bidders are identified as adjusting their bid prices in a direction consistent with equilibrium predictions, confirming bidders’ ability to learn from experience. Furthermore, while controlling for individual effects through bidder fixed effects, regression results suggest that winning and losing experiences have asymmetric effects on adjustments individual bidders make to their bidding strategy. In particular, losing experiences prompt bidders to increase bid prices whereas winning experiences explain reductions in bid prices. These learning effects are found to taper off shortly after a year following the implementation of the discriminatory auction format. Results of this study are important for policy makers and auctioneers of real-life auctions that may be contemplating a change in auction format. This study provides some insight into the time it takes for bidders to learn optimal bidding strategies which is important from a practical perspective as suboptimal behaviour negatively impacts the allocative efficiency and revenue earned by the auction mechanism.


Subject(s):
Issue Date:
2016
Publication Type:
Conference Paper/ Presentation
PURL Identifier:
http://purl.umn.edu/236279
Total Pages:
18




 Record created 2017-04-01, last modified 2017-08-29

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)