Auctioning to buyers with correlated values

Abstract

In a laboratory setting, a monopolistic auctioneer sells to buyers as the level and nature of demand changes. I ask whether sellers correctly recognize the role played by correlation among buyers' values. The prices set by subjects closely match the risk-neutral benchmark predictions when demand follows the independent-private-values framework. In contrast, subjects fail to correctly account for correlation among buyers' values once the independence assumption is dropped. I offer two new models of pricing in a correlated-values environment. The model that suggests sellers ignore correlation outperforms both the benchmark and the model that suggests sellers incorrectly account for correlation.

Acknowledgements

I thank the seminar participants at North Carolina State University and Vanderbilt University as well as Andrew Daughety, Charles Knoeber, Tong Li, Thayer Morrill, Jennifer Reinganum and especially Mikhael Shor. I also thank the Household International Research Support Program for financial support and Ben Gotow for excellent programming assistance.

Notes

¹ Wang (1998) provides the second-order conditions, which are met by the distributions used in the experiment.

² In the affiliated-signals model, for :

The first term corresponds to μ − 5 ≥ r, whereas the second term corresponds to μ − 5 < r < 10. For :

³ I investigate heterogeneity in subjects' pricing decisions using regression analysis, controlling for subject performance on the pre-experiment quiz and a proxy for risk aversion derived from a Holt–Laury (Holt and Laury, 2002) risk survey as well as dummies variables for males, graduate students, economics majors and frequent eBay users. Pricing decisions are not correlated with subjects' quiz performance or risk aversion to an economically meaningful degree. Additionally, demographic characteristics have no effect on the comparative-statistic results but are correlated with price levels (e.g. graduate students set higher prices whereas frequent eBay users set lower prices).

⁴ In 2008, eBay's cited ‘consumer electronics; home and garden; clothing and accessories; and tickets’ among the categories in which its gross merchandize volume grew most rapidly (eBay Inc., 2008, p. 52). In contrast, for art, Sotheby's continues to be the industry leader in market capitalization with $616.63 million in auction-related revenues (Sotheby's, 2008, p.15).