Authority and screening in a principal agent model

I compare authority structures in a principal agent screening model where the agent’s type determines his payoff and his preferred activity level. The uninformed principal incurs an externality and she is able to propose a revelation mechanism with contingent transfers. Under ‘principal authority’ (the agent needs the principal’s agreement for his activity), the optimal mechanism is a standard screening contract and activities are downwards distorted. Under ‘agent authority’ (the agent can pursue the activity at his own discretion), reservation utilities become type dependent and the sign of the payoff externality shapes contractual outcomes. Specifically, (1) negative externalities (the principal prefers a smaller activity than her agent) now trigger upwards distortions; (2) a subset of agents is being ‘left alone’ under the mechanism; and (3) agent authority is always associated with larger activities than principal authority. Finally, I show that under simple symmetry assumptions, agent authority always generates a higher total surplus than principal authority.