Very Credible Auction

Thanawat Sornwanee

Keywords: Auction, Credibility, Data Privacy

Abstract: Online platforms routinely ask users to reveal private information to improve ``targeting'': which product to recommend, which advertisement to show, or which offer to display. At the same time, revealed data can be used for more aggressive price discrimination, leaving users worse off. This paper studies a minimal disclosure-and-pricing game in which neither side can commit: the user cannot commit to a willingness-to-pay claim and the platform cannot commit to a pre-announced pricing rule. We call mechanisms implementable under such constraints very credible. We first show a one-dimensional impossibility: with a single product and a single buyer, any very credible mechanism is outcome-equivalent to posted monopoly pricing and uninformative communication. We then show that multidimensional preferences create room for positive-sum disclosure: with multiple products but limited attention (the platform can display only one offer), users can truthfully reveal \emph{which} product they prefer without enabling stronger price discrimination. A simple two-product example yields strictly higher seller revenue and strictly higher buyer surplus relative to no disclosure.

Track: Short Paper

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 109