The Effects of Biasing Performance Measurement Systems on Incentives and Retention Decisions
研究了委托人如何选择绩效测量系统的偏差（严格、中性或宽松）来同时服务于激励和留任决策，发现严格报告能缓解道德风险，而加入留任评估后反而可能强化对严格性的需求。
ABSTRACT We examine a principal–agent setting in which the principal uses a performance measurement system for multiple purposes—to provide incentives and for retention decisions. The principal chooses the nature and extent of bias in the system, which determines whether the performance report is stringent, neutral, or lenient relative to the unobservable actual outcome. We show that when the report is used only for incentive purposes (an incentive role), stringency alleviates moral hazard. On the other hand, when the principal's objective is to minimize the cost from incorrect retention and firing decisions (a fit evaluation role), there is a demand for leniency. Surprisingly, however, we show that adding a fit evaluation role to a system serving an incentive role can accentuate the demand for stringency because stronger incentives can also indirectly improve retention decisions.