A main issue in the design of incentive contracts concerns the choice of performance measures. In practice, the job of a worker typically involves several different aspects or tasks that all contribute to an outcome valued by the principal; often both the outcome and the agent's actions in some tasks are observable to the principal. Thus, a basic choice the principal faces when designing incentive contracts is whether to reward the agent for output or for effort. In this paper we show that it can be optimal for the principal to incentivize an agent using only output-based performance pay, even if the principal can monitor some of the agent's actions perfectly and knows exactly which action is optimal for each task, as long as there is one action of the agent she cannot observe.
展开▼