Dual Ceteris Paribus Comparisons: Some Initial Thoughts on Laboratory Isolation of Equilibrium Forces in Auction Markets

Abstract

A technique is illustrated for isolation of equilibrium forces reputedly present in individual agents’ economic behavior. An example setting arises in a theory of affiliated auctions (MILGROM/ WEBER, 1982), which predicts expected revenue to be increased by seller’s announcement of public information correlated with the item’s unknown common value.

Within-subject ceteris paribus comparisons are obtained by a dual market technique. These within-subject comparisons are duplicated across experiments employing alternative institutional forms.

Two equilibrium forces, a strategic force and an evaluative force, are responsible for the predicted revenue increases due to public information. A series of first-price common-value auction experiments incorporates both forces. A series of second-price, common-value auction experiments eliminates the strategic force. Finally, a series of first-price, affiliated-private-value auction experiments eliminate the evaluative force.

A tentative conclusion: subjects’ errors in Bayesian assessments, before public information is announced, create overbids relative to theory, overwhelming the evaluative force, and reducing strength and reliability of the strategic force. With extensive experience within a particular parameter setting, some evaluative force emerges.

This research is supported by National Science Foundation grant NSF IST 84-08396. I thank, without implicating, John Kagel and Dan Levin for comments and suggestions, and they and Douglas Dyer for assistance in conducting the experiments.