Multitask Contract Theory

Abstract

We construct and analyze a dynamic principal-agent model with moral hazard, where the agent engages in multiple tasks, and the principal incentivizes effort by observing various related performance measures. In our model, cost interactions in the agent's production function can lead to effort distortions, and the composition and statistical properties of the principal's observation technology make certain combinations of tasks easier to enforce than others. In order to study the optimal effort, we prove results about the statistical properties of the information structure, the costs of enforcing various effort allocations, and the tradeoff between the costs of compensation and risk. We analyze the cumulative effects of effort distortions and information structures on the principal's pro t by decomposing profit into an incentive-separable space. This paper provides a general, unified way to frame and analyze static and dynamic multitasking, allowing deeper insights into why multitasking phenomena occur. Our results shed light on a broad range of multitasking applications, including performance indexation, executive compensation, window dressing, and working in teams.