The relative performance evaluation (RPE) hypothesis states that firms benefit from comparing their own performance to that of a peer group when evaluating the CEO's performance. Although in theory firms should be employing relative performance to evaluate the CEO, indirect empirical tests in the 1980s and 1990s generally fail to support the RPE hypothesis. This paper examines RPE‐related disclosures found in the compensation committee reports provided in proxy statements to determine whether firms actually employ RPE, and to offer insight into why indirect tests generally fail to support the RPE hypothesis.
We find that firms do use RPE in determining executive compensation, thus supporting the RPE hypothesis, although RPE usage is not widespread. We also find that several key assumptions underlying prior indirect tests are misspecified for many firms, helping to explain the difficulty in detecting RPE in random samples of firms and suggesting improvements to methodologies employing indirect tests for RPE.
Our results also beg the question of why some firms use RPE while other firms do not. We find that RPE usage is positively related to greater monitoring and stakeholder concern about pay and performance, but that performance and CEO power and insulation from pressure do not explain cross‐sectional variation in RPE usage. We also examine disclosures related to peer groups and adverse performance‐related events since they indirectly relate to RPE and find that many firms filter out negative‐performance‐related events, but not positive ones. This is equivalent to using one‐sided RPE, where a firm excuses the CEO from factors that affect industry performance adversely, but credits the CEO for factors that aid industry performance.