Mechanism Design in Hidden Action and Hidden Information: Richness and Pure Groves

Abstract

We investigate general collective decision problems related to hidden action and hidden information. We assume that each agent has a wide availability of action choices at an early stage, which provides significant externality effects on the other agent’s valuations in all directions. We characterize the class of all mechanisms that solve the hidden action problem, and demonstrate equivalence properties in the ex-post term. Importantly, we find that pure Groves mechanisms, defined as the simplest form of canonical Groves mechanisms, are the only efficient mechanisms that solve such hidden action problems. We argue that the resolution of the hidden action problem automatically resolves the hidden information problem.