Working Paper: CEPR ID: DP4262
Authors: Michael Raith
Abstract: This Paper examines optimal incentives and performance measurement in a setting where an agent has specific knowledge about the consequences of their actions for the principal. I study incentive contracts in which the agent?s compensation can be based on both ?input? measures closely related to the agent?s actions, and ?output? measures closely related to the principal?s pay-off. I argue that when the agent has specific knowledge (i.e. private information that is difficult to communicate) about how their actions contribute to the principal?s pay-off, output-based pay encourages the agent to use their knowledge while input-based pay does not. I show within a two-task agency model that (partially) output-based compensation is optimal even when the agent?s actions on each task can be measured perfectly. Comparative statics results show how the optimal choice of performance measures and incentives depends on the agent?s knowledge, environmental risk, technological uncertainty, and job complexity. The theory leads to several novel predictions, as well as new explanations for existing empirical findings.
Keywords: Distortion; Incentives; Input vs Output-based Pay; Multitask Agency Model; Performance Measurement; Risk; Specific Knowledge
JEL Codes: D82; J33; M52
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| output-based pay (J33) | agent utilization of specific knowledge (D83) |
| input-based pay (J33) | agent utilization of specific knowledge (D83) |
| agent's private information about productivity (L85) | principal's choice of compensation structure (M52) |
| agent's actions can be perfectly measured (C91) | principal prefers input-based pay (J33) |
| output-based measures (C67) | compensation contracts (J33) |