Full Implementation and Belief Restrictions

We introduce a framework to study the problem of full implementation under arbitrary restrictions on agentsbeliefs, which we call -Implementation. We show how this framework delivers insights from both a theoretical and an applied perspective. In particular, rst we provide a characterization of the properties of the mechanisms that achieve it, and relate it to several notions in the existing literature on full implementation. Then we show how, in environments with transferable utility, using minimal information about agents beliefs, full implementation can be achieved through the design of simple direct mechanisms. The mechanisms that we use are determined by a payment scheme, that maps agents reports into transfers. Such mechanisms have a clear economic interpretation, thereby bridging a gap between the literature on partial and full implementation. These results also show how the methodology of the recent literature on robust mechanism design can be extended to address more applied problems of mechanism design, overcoming important limitations of the traditional approach to the full implementation problem.