Title: Auctioning with Strategically Reticent Bidders

Abstract: We propose and study a novel mechanism design setup where each bidder holds two kinds of private information: (1) type variable, which can be misreported; (2) information variable, which the bidder may want to conceal or partially reveal, but importantly, not to misreport. We refer to bidders with such behaviors as strategically reticent bidders. Among others, one direct motivation of our model is the ad auction in which many ad platforms today elicit from each bidder not only their private value per conversion but also their private information about Internet users (e.g., user activities on the advertiser's websites) in order to improve the platform's estimation of conversion rates.
We show that in this new setup, it is still possible to design mechanisms that are both Incentive and Information Compatible (IIC). We develop two different black-box transformations, which convert any mechanism $\mathcal{M}$ for classic bidders to a mechanism $\bar{\mathcal{M}}$ for strategically reticent bidders, based on either outcome of expectation or expectation of outcome, respectively. We identify properties of the original mechanism $\mathcal{M}$ under which the transformation leads to IIC mechanisms $\bar{\mathcal{M}}$. Interestingly, as corollaries of these results, we show that running VCG with bidders' expected values maximizes welfare, whereas the mechanism using expected outcome of Myerson's auction maximizes revenue. Finally, we study how regulation on the auctioneer's usage of information can lead to more robust mechanisms.