Are Interim Performance Evaluations Optimal when the Evaluations are Subject to Manipulation

The paper considers, in a principal-agent framework, whether or not providing interim performance evaluations is more efficient when agents can manipulate the very performance reports by which they are evaluated. Providing interim performance evaluation introduces possible dependence of subsequent strategies on interim reports. Such dependence, while alleviating previous-period incentive compatibility constraints, also increases the (expected) later-period performance manipulation costs. The first effect provides a risk-sharing benefit whereas the second effect increases the expected compensation cost to the agent (in utilities). Correspondingly, providing interim performance evaluation is better when the manager is sufficiently risk-averse so that the risk-reduction effect dominates.