A simple explanation of the relative performance evaluation puzzle

Abstract

We study a simple moral hazard model in which two risk-neutral owners establish incentives for their risk-averse managers to exert effort. Because the probability distributions over output realizations depend on a common aggregate shock, optimal contracts make the compensation of each manager contingent on own performance but also on a performance benchmark-the performance of the other firm. If the marginal return of effort depends on the aggregate state, optimal contracts are not monotonically decreasing in the performance benchmark. This provides a simple explanation of the Relative Performance Evaluation (RPE) Puzzle-the documented lack of a negative relationship between CEO compensation and comparative performance measures, such as industry or market performance. Our simple model can also explain one-sided RPE-the documented tendency to insulate a CEO's rewards from bad luck, but not from good luck. We clarify that our results are robust in several dimensions and we discuss other applications of our findings.JEL classification: C72; D82; M52

Illusory Correlation in the Remuneration of Chief Executive Officers: It Pays to Play Golf, and Well

“…The empirical results in Blanchard et al (1994) cast serious doubts on the empirical relevance of the models in Oyer (2004), Himmelberg and Hubbard (2000), and Celentani and Loveira (2006).…”

“…There are extensions to the basic principle-agent model, which can make pay for luck potentially optimal -see, e.g., Oyer (2004), Himmelberg and Hubbard (2000), and Celentani and Loveira (2006). The key feature of these models is that CEO marginal productivity or the value of a CEO's outside options fluctuates.…”

Due Diligence, Research Joint Ventures, and Incentives to Innovate

“…Assume the occurrence of the signal to be independent across researchers. 4 The owners of each firm i interview their researcher, who reports m i ∈ {∅, s, s},…”

“…Hence, it suffices that one researcher reveals his signal for both firms' owners to be informed. After forming an R&D joint venture, which we assume to be irreversible, and before executing the project, the owners learn the true value of s, regardless of whether the agents sent a message 4 These assumptions on the signals make it impossible for the principals to devise mechanisms that make the researchers reveal their private information without cost to the principals. See Crémer and McLean (1985) or Maskin and Riley (1985) for such mechanisms.…”

“…Competition offers a yardstick to the shareholders because the firm's performance can be measured against that of its close competitors, and it enables relative performance evaluation (Celentani and Loveira, 2006) by providing informative signals that must be considered in the contracting between principals and agents (Holmström, 1979 and1982). Competition offers insurance to risk-averse managers by reducing the impact of exogenous (demand or cost) shocks, thus increasing incentives (Rey and Tirole, 1986).…”