Organization of Project Evaluation and Implementation under Moral Hazard

We examine the optimal organizational form of project evaluation and implementation under moral hazard. In the evaluation phase, two fallible risk-neutral agents sequentially screen projects. The approved projects are subsequently implemented in the development phase. We show that moral hazard renders the organization as a polyarchy less attractive than as a hierarchy. Furthermore, given moral hazard, task assignment becomes relevant: For identical agents, the principal always delegates implementation to the agent who works first in the evaluation phase.