Incentives and the existence of Pareto-optimal revelation mechanisms (Q804453)

This paper is concerned with conditions under which truthful revelation of information can be elicited from individuals under decentralized decision procedures (mechanisms) when information is incomplete. More specifically, it explores the connection between individual belief structures and the existence of Pareto optimal strongly incentive compatible (SIC) mechanisms [\textit{J. R. Green} and \textit{J.-J. Laffont}, ``Incentives in public decision making'' (1979; Zbl 0417.90001)] and Pareto optimal Bayesian incentive compatible (BIC) mechanisms [the first and third author, ``Incentives and incomplete information'', J. Publ. Econ. 11 (1979)]. The following main results are derived: (1) If there are only two types per individual, then a Pareto optimal SIC mechanism exists. (2) If there are only two individuals with more than three types, then there are sets of beliefs, open in the class of all beliefs such that Pareto optimal BIC mechanisms do not exist. If there are more than two individuals, then the class of beliefs for which Pareto optimal BIC mechanisms exist is open and dense in the class of all beliefs.