Optimal Delegation with a Finite Number of States

This paper contributes to the literature on optimal delegation, dating back to Holmstrom's (1984) seminal work. In contrast to models in the Holmstrom tradition, we assume that the set of states is finite. We provide a full characterization of the class of optimal delegation sets under this assumption, and show that they have a different structure from that in the continuous-state model. As the number of states tends to infinity, however, every optimal delegation set converges to that of Holmstrom (1984). We also show that, for intermediate bias, the Ally Principal fails for small changes in bias, the Uncertainty Principle may fail, and the principal prefers to appoint an amateur agent.