Optimal Performance Measures with Task Complementarity

Abstract

This study examines optimal performance measures under moral hazard where a risk neutral agent with limited liability is assigned two interdependent tasks. Task interdependency, or complementarity, is captured by assuming that when the agent provides high effort on one task, it affects the probability of success on a second task. Two performance measures are considered, disaggregate and aggregate. The relative benefit of aggregate performance measures derives from the agent's uncertainty about compensation when performing the second task. The relative benefit of disaggregate measures arises because it produces more information about the agent's actions on each task. The comparative advantage of aggregate measures is maximized for relatively independent tasks and diminishes with strong positive or negative complementarity. Endogenous task assignment is then explored. The principal can bundle the two tasks together and assign them to a single agent, maintaining the effects of task complementarity, or unbundle the tasks and assign each to a different agent, eliminating the effect of task complementarity. The inclination to avoid negative task complementarity increases the likelihood, conditional on bundling the tasks, that the principal prefers aggregate performance measures.