Mechanism design and intentions

Abstract

We introduce intention-based social preferences into a mechanism design framework with independent private values and quasilinear payoffs. For the case where the designer has no information about the intensity of social preferences, we provide conditions under which mechanisms which have been designed under the assumption that agents are selfish can still be implemented. For the case where precise information about social preferences is available, we show that any tension between efficiency, incentive-compatibility, and voluntary participation may disappear. Impossibility results such as the one by Myerson and Satterthwaite (1983) are then turned into possibility results. We also provide a systematic account of the welfare implications of kindness sensations.