EN RU DE
News Archive

CEO Compensation

07.12.2010

http://www.nber.org/papers/w16585

Carola Frydman, Dirk Jenter

NBER Working Paper No. 16585
Issued in December 2010
NBER Program(s):   CF   LS

This paper surveys the recent literature on CEO compensation. The rapid rise in CEO pay over the past 30 years has sparked an intense debate about the nature of the pay-setting process. Many view the high level of CEO compensation as the result of powerful managers setting their own pay. Others interpret high pay as the result of optimal contracting in a competitive market for managerial talent. We describe and discuss the empirical evidence on the evolution of CEO pay and on the relationship between pay and firm performance since the 1930s. Our review suggests that both managerial power and competitive market forces are important determinants of CEO pay, but that neither approach is fully consistent with the available evidence. We briefly discuss promising directions for future research.

Executive compensation is a complex and contentious subject. The high level of CEO pay in the United States has spurred an intense  debate about the nature of the pay-setting process and the outcomes it produces. Some argue that large executive pay packages are the result of powerful managers setting their  own pay and extracting rents from firms. Others interpret the same evidence as the result of optimal  contracting in a competitive market for managerial talent. This survey summarizes the research on CEO compensation and assesses the evidence for and against these explanations.


Our review suggests that both managerial power and competitive market forces are important determinants of CEO pay, but that neither approach alone is fully consistent with the available evidence. The evolution of CEO compensation since World War II can be broadly divided into two distinct periods. Prior to the 1970s, we observe low levels of pay, little dispersion across top managers, and moderate pay-performance  sensitivities. From the mid-1970s to the early 2000s, compensation levels grew dramatically, differences in pay across managers and firms widened, and equity incentives tied managers’ wealth closer to firm performance. None of the  existing theories offers a fully convincing explanation for the apparent regime change that occurred during the 1970s, and all theories have trouble explaining some of the cross-sectional and time-series patterns in the data.

Many of the theoretical studies we review explore how various characteristics of real-world compensation contracts can be consistent with either rent extraction or optimal contracting. Although useful, demonstrating that a given compensation feature can arise in an optimal contracting (or rent extraction)  framework provides little evidence that the feature is, in fact, used for efficiency reasons (or to extract rents). Partly as a result, there is no consensus on the relative importance of rent extraction and optimal contracting in determining the pay of the typical CEO. To help answer this question, models of CEO pay will have to produce testable predictions that differ between the two approaches.


We expect that the renewed interest in theoretical work on CEO pay and the emergence of new data will deliver both the predictions and the testing opportunities needed to resolve this debate. Promising recent contributions have examined the effects of exogenous changes in the contracting environment on CEO compensation, firm behavior, and firm performance. For example, industry deregulations have  been linked to higher CEO compensation, suggesting that increased  demand for CEO talent  raises pay levels. However, declines in CEO compensation observed after new regulations strengthen board
oversight suggest rent extraction.

...Both the level and the composition of CEO pay have changed dramatically over time. The post-World War II era can be divided into at least two distinct periods. Prior to the 1970s, we observe low levels of pay, little dispersion across top managers, and only moderate levels of equity compensation. From the mid-1970s to the end of the 1990s, all compensation components grew dramatically, and differences in pay across executives and firms widened. By far, the largest increase was in the form of stock options, which became the single largest component of CEO pay in the 1990s. Finally, average CEO pay declined from 2000 to 2008, and restricted stock grants  have replaced stock  options as the most popular form of equity compensation. It is arguably too early to judge whether the post-2001 period constitutes a third regime in CEO compensation or just a temporary anomaly caused by the technology bust of 2000-2001 and the financial crisis of 2008.

...Since Jensen & Murphy (1990a), our understanding of the link between firm performance and CEO compensation has improved substantially. The long-run evidence shows that compensation arrangements have served to  tie the wealth of managers to firm performance—and perhaps to align managers’ with shareholders’ interests—for most of the twentieth century. Much of the effect of performance on CEO wealth works through revaluations of stock and option holdings, rather than through changes in annual pay. The sensitivity of CEO wealth to performance surged in the 1990s, mostly owing to rapidly growing option portfolios. In contrast, the  typical CEO’s fractional equity ownership remains low, suggesting a continued need for direct monitoring by boards and investors.

...Our reading of the evidence suggests that both managerial power and competitive market forces are important determinants of CEO pay and that neither approach alone is fully consistent with the available evidence. On the one hand, several compensation practices, as well as specific cases of outrageous and highly publicized pay packages, indicate that (at least some) CEOs are able to extract rents from their firms. On the other hand, efficient contracting explanations are arguably more  successful at explaining differences in pay practices across firms and at explaining the evolution  of CEO pay since the 1970s. However, none of these theories provides a fully convincing explanation for the apparent regime change in CEO compensation that  occurred during the  1970s. Moreover, both approaches fail to explain the explosive growth of options in the 1990s and their recent decline in favor of restricted stock, which  may be, in part, a response to changes in accounting practices. Although a combination of the proposed  explanations may explain the changes in CEO pay in recent decades, the relative importance of the different theories remains an open question.

...The literature provides ample evidence that CEO compensation and portfolio incentives are correlated with a wide variety of corporate behaviors, from investment and financial policies to risk taking and  manipulation. Arguably, the widespread use of incentive compensation and the large cross-sectional differences in managerial contracts would make little sense if compensation had no effect on CEO behavior. However, because compensation arrangements are endogenous  and correlated with many unobservables, measuring their causal effects on behavior  and firm value is extremely difficult and remains one of the most important challenges for research on executive pay.

Both the executive compensation literature  and our understanding of  pay practices have grown tremendously in recent years. Yet, many important questions remain unanswered. For example, the causes of the apparent regime change in CEO compensation that occurred during the 1970s remain largely unknown. The relative importance of rent extraction and optimal contracting in determining pay for the  typical CEO is still to be determined, and even less is known about the causal effects  of CEO pay on behavior and firm value. Finding answers to these questions will require a combination of new theory predictions; the creative use of exogenous changes in the contracting environment; and new data from other countries, prior decades, and different  types of firms. The progress made by recent studies on all these dimensions is cause for optimism and suggests that answers may not be far off.


News archive