Essays on contracts and organizational design
This dissertation consists of three essays on contracts and organizational design. The first essay studies optimal organizational design in a principal-agent setting in which the agent must be motivated both to acquire costly information and to use the information in the principal's interest. The second deals with the foundations of incomplete contracts in the context of a hold-up problem. The third considers multidimensional screening when transfers are unavailable. In the first essay we formulate a model in which a principal simultaneously designs both a monetary incentive scheme and a delegation of decision rights for an agent. Consistent with the empirical evidence, we show that three key variables in the model are mutually complementary: the level of environmental uncertainty, the scope of optimal delegation, and the strength of optimal incentives. We then examine another important element of organization, communication, by examining a variant of the model in which the agent reports his findings to the principal, who uses them to decide which project to pursue. The result is that while an increase in the level of uncertainty still increases the strength of optimal incentives, it decreases the optimal amount of reports the agent is allowed to make. The final variant of the model generalizes the others by allowing the principal to use all three instruments, communication, delegation, and incentive pay. Our results in this case suggest that the principal should not include communication in his organizational design--delegation together with incentive pay tends to dominate communication. Overall, our model yields results that are consistent with empirical facts about organizations, unlike those of the standard models of delegation and agency. In the second essay we use the Mirrlees approach to mechanism design with renegotiation, as propounded by Segal and Whinston (2002), to reexamine recent attempts to lay foundations for incomplete contracts. The first result is that significant simplification is achieved by restricting attention to utility mappings that we call permutation invariant, a property related to the welfare neutrality of Maskin and Tirole (1999). We then provide simpler proofs and clearer interpretations of the main results of the foundations literature, namely, that environmental complexity and relationship specific investments with positive spillovers can cause the null contract to be optimal or approximately optimal. A new result is that the structure of the state space matters: the null contract need not be optimal if the state space is continuous instead of discrete. We also find that whether contingencies are describable is important for the practical implementation of a contract, but not for the achievable final utility. In the third essay we study multidimensional screening when transfers are unavailable. We find that the usual restriction to continuous decision rules is too restrictive: the only continuous rule that is implementable (incentive compatible) is a constant rule that does not depend on the reported values. Such a constant rule elicits no information. We then weaken the continuity restriction to a property we refer to as "continuous from all permutation directions". We show that within this class of decision rules, attention can be restricted to a small and relatively simple class consisting of the rules that are permutation invariant and monotone.
|Year of publication:||
|Type of publication:||Other|
Dissertations available from ProQuest