We study the question of which social choice functions from an abstract type space to a set of outcomes are truthful, i.e., implementable by truthful mechanisms, when utilities are quasi-linear. For convex domains, our main theorem characterizes truthful social choice functions as those...

We study a notion of locally robust implementation that captures the idea that the planner may know agentsʼ beliefs well, but not perfectly. Locally robust implementation is a weaker concept than ex-post implementation, but we show that no regular allocation function is locally robust...