TCS Daily

The Compensation Conundrum

By Stephen Bainbridge - February 4, 2005 12:00 AM

Executive compensation has grown by leaps and bounds over the last two decades. By 2003, the average large firm CEO made 500 times what the average worker made. In the aggregate, during the five-year period 1998-2002, compensation paid to the top five executives at 1500 large public corporations totaled roughly $100 billion.

Yet, we live in an era in which many occupations carry such vast rewards. Lead actors routinely earn $20 million per film. The NBA's average salary is over $4 million per year. Top investment bankers can earn annual bonuses of $5 to $15 million. Unless one's objection is solely based on the size of executive compensation, perhaps for redistributionist reasons, critics of executive compensation must be able to distinguish corporate managers from these other highly paid occupations.

In their new book Pay Without Performance (Cambridge, MA: Harvard University Press, 2004. Pp. xii, 278. $24.95), law professors Lucian Bebchuk and Jesse Fried contend that actors and sports stars bargain at arms'-length with their employers, while managers essentially set their own compensation. As a result, they claim, even though managers are under a fiduciary duty to maximize shareholder wealth, executive compensation arrangements often fail to provide executives with proper incentives to do so and may even cause executive and shareholder interests to diverge. In other words, the executive compensation scandal is not the rapid growth of management pay in recent years, as too many glibly opine, but rather the failure of compensation schemes to award high pay only for top performance.

Bebchuk and Fried's analysis is premised on what they call "managerial power." They claim that "directors have been influenced by management, sympathetic to executives, insufficiently motivated to bargain over compensation, or simply ineffectual in overseeing compensation." As a result, executive pay has greatly exceeded the levels that would prevail if directors loyal to shareholder interests actually bargained with managers at arms'-length.

In a true tour de force, Bebchuk and Fried march relentlessly through one form of executive compensation after another, advancing a consistent story of how management influence taints and distorts the compensation process. Although they frequently reference theoretical models and empirical studies supporting their argument, this section was clearly written with a lay reader in mind. Bebchuk and Fried's efforts in this regard are quite successful; they have produced a highly accessible indictment of executive compensation practices.

Bebchuk and Fried's Managerial Power Model

Conventional wisdom says that the principal-agent problem is the central issue of corporate governance. Agents who shirk do not internalize all of the costs thereby created; the principal reaps part of the value of hard work by the agent, but the agent receives all of the value of shirking. Wherever a principal-agent problem is found, we thus expect to see a mixture of carrots and sticks designed to constrain shirking. The sticks include ex post sanctions, up to and including dismissal. The carrots include incentives that align the agent's interests with those of the principal.

In theory, the principal-agent problem could be ameliorated by executive compensation schemes that realign the interests of corporate managers with those of the shareholders. In practice, however, as Bebchuk and Fried show, the most common forms of executive compensation fail to remedy the principal-agent problem. Indeed, some forms exacerbate the problem.

According to Bebchuk and Fried, boards of directors -- even those nominally independent of management -- have strong incentives to acquiesce in executive compensation that pays managers rents (i.e., amounts in excess of the compensation management would receive if the board had bargained with them at arms'-length). Among these are: Directors often are chosen de facto by the CEO. Once a director is on the board, pay and other incentives give the director a strong interest in being reelected; in turn, due to the CEO's considerable influence over selection of the board slate, this gives directors an incentive to stay on the CEO's good side. Directors who work closely with top management develop feelings of loyalty and affection for those managers, as well as becoming inculcated with norms of collegiality and team spirit, which induce directors to "go along" with bloated pay packages. Finally, Bebchuk and Fried argue that those few directors who resist these incentives and seek to put shareholder interests first face a number of obstacles in both the law and practice of corporate governance.

The net effect of managerial power is that CEO pay packets are higher than would obtain under arms'-length bargaining and less sensitive to performance. As a result, compensation of CEOs and other top managers has become a not insignificant chunk of corporate earnings. Yet, Bebchuk and Fried claim, much of that pay has been insensitive to the performance of those companies.

A Cautionary Note on the Evidence

Whether one is persuaded by Pay Without Performance depends in large measure on whether one buys Bebchuk and Fried's interpretation of the evidence, which is often plausible but contestable. In fact, a forthcoming review of Bebchuk and Fried's book by John Core, Wayne Guay, and Randall Thomas argues "that in many settings where 'managerial power' exists, observed contracts anticipate and try to minimize the costs of this power, and therefore may in fact be written optimally." Accordingly, they contend, the managerial power and optimal contracting models are "complementary, and not competing, explanations."

Indeed, while Bebchuk and Fried marshal considerable evidence in support of their managerial power model, there is much competing evidence suggesting that executive compensation packages are designed to align managerial and shareholder interests. Consider, for example, the much maligned practice of management perquisites. If managerial power has widespread traction as an explanation of compensation practices, one would assume that the evidence would show no correlation between the provision of perks and shareholder interests. In fact, however, The Economist recently reported on an interesting study of executive perks finding just the opposite:

"Raghuram Rajan, the IMF's chief economist, and Julie Wulf, of the Wharton School, looked at how more than 300 big companies dished out perks to their executives in 1986-99. It turns out that neither cash-rich, low-growth firms nor firms with weak governance shower their executives with unusually generous perks. The authors did, however, find evidence to support two competing explanations.

"First, firms in the sample with more hierarchical organizations lavished more perks on their executives than firms with flatter structures. Why? Perks are a cheap way to demonstrate status. Just as the armed forces ration medals, firms ration the distribution of conspicuous symbols of corporate status.

"Second, perks are a cheap way to boost executive productivity. Firms based in places where it takes a long time to commute are more likely to give the boss a chauffeured limousine. Firms located far from large airports are likelier to lay on a corporate jet.

"In other words, executive perks seem to be set with shareholder interests in mind, which is inconsistent with the possibility that managerial power offers a unified field theory of executive compensation.

This example is not intended as a comprehensive rebuttal of the managerial power model, but rather to highlight the possibility that many executive compensation practices are at least as consistent with an arms'-length bargaining model as the managerial power model. Indeed, in an important analysis, Bengt Holmstrom and Steven Kaplan noted that while many observers complain that executive compensation packages "represent unmerited transfers of shareholder wealth to top executives with limited if any beneficial incentive effects," the evidence offers "several reasons to be skeptical of these conclusions."

Does the Managerial Power Model Still Hold True?

Pay Without Performance is based in very large part on a law review article Bebchuk and Fried published (with then coauthor David Walker) in 2002. Several versions of that paper were posted to the Social Science Research Network Electronic Library, including one as early as December 2001. In the interim, the world has changed.

In June 2002, the New York Stock Exchange's board approved new listing standards that, inter alia, significantly enlarged the role and power of independent members of listed companies' boards of directors. NASDAQ and AMEX adopted similar changes. A number of the new listing standards speak directly to the problems identified by Bebchuk and Fried.

In July 2002, President Bush signed into law the "Public Company Accounting Reform and Investor Protection Act" of 2002 (the Sarbanes-Oxley Act), which he praised for making "the most far-reaching reforms of American business practices since the time of Franklin Delano Roosevelt." Again, a number of those reforms speak directly to the sources of managerial power identified by Bebchuk and Fried.

Unfortunately for Bebchuk and Fried, they were thus caught in a time bind. Many incremental reforms that they might have proposed are now law, but we do not yet know how well those reforms will work. In response, Bebchuk and Fried adopted a uniform tactic of dismissing those reforms as inadequate.

Under the NYSE's new listing standards, for example, listed companies must create a compensation committee, comprised solely of independent directors, whose minimal duties include setting the CEO's compensation. The committee must adopt written charters specifying their roles, duties, and powers, which must at a minimum conform to the listing standard's detailed requirements. The committee is now charged with hiring compensation consultants, a task previously left to management. To address the longstanding problem of director interlocks, in which the CEO of two companies would sit on each other's board of directors' compensation committee, and presumably scratch each other's back, the new listing standards provide that a director may not be deemed independent if he is employed (or has been employed in the last five years) by a company in which an executive officer of the listed company serves as a member of the board of directors' compensation committee.

Bebchuk and Fried dismiss these developments on grounds that the new standards "merely make mandatory a practice that most public companies already have been following for some time." Surely this is too glib. In the first place, all public corporations -- not just "most" -- must now comply with these requirements. In the second, transforming what were once mere matters of good practice into affirmative legal requirements may have a salutary effect. To be sure, I am speculating here, but the larger point is that we simply don't know yet. As a practical matter, the process of implementing the new listing standards has just begun. Inevitably, it will take time to determine whether they are as toothless as Bebchuk and Fried claim.

Instructively, however, a recent study of compensation committees in IPO firms found "little support for the managerial power model." They found no evidence that including insiders or executives from other firms on the compensation committee was correlated with either higher CEO pay levels or lower incentives. Instead, noting that higher director compensation was coupled with higher CEO compensation and that the presence of large blockholders was correlated with lower CEO compensation, the study concludes that "CEO pay and incentives are set taking into account the supervisor (compensation committee)'s self-interest rather than simply assuming the supervisor will act on the owner's behalf or co-opt with CEO."

The NYSE's new standards also require that all listed corporations create a nominating and corporate governance committee comprised solely of independent directors. Again, Bebchuk and Fried dismiss this change as inadequate to limit the power of the CEO. Here, however, there is research to the contrary. Westphal and Zajac have demonstrated that as board power increases relative to the CEO -- measured by such factors as the percentage of insiders and whether the CEO also served as chairman -- newly appointed directors become more demographically similar to the board. Accordingly, enhanced powers of independent directors to control the nomination process should prove an effective constraint on CEO influence.

The new listing standards also substantially limit the CEO's ability to financially reward compliant directors. The new rules, for example, limit the amount of additional compensation directors may receive over and above their director fees. They also limit the amount of business that may be conducted between the corporation and a business associated with one of its independent directors.

Finally, the NYSE's new stock exchange listing standards significantly tighten the definition of director independence. To be sure, the evidence as to the corporate governance value of director independence is, at best, mixed. (For a summary of the evidence and the NYSE rules, see my article A Critique of the NYSE's Director Independence Listing Standards.) Yet, the new rules are only now taking effect. In addition, as Bebchuk and Fried concede, the amount of time independent directors are devoting to board work, including compensation decisions, has been increasing significantly in recent years. Once again, it may be too early to tell whether the new rules will make significant changes.

In sum, Bebchuk and Fried acknowledge that the changes just described will move the executive compensation process towards the arms'-length ideal, but contend that those changes do not eliminate the various factors that they claim give directors incentives to favor executives at the expense of shareholders. Accordingly, they claim, sweeping reforms are required. But how do we know? Indeed these changes are only now taking affect. Surely it is too soon to tell, especially since there is evidence that the changes already in place can be expected to have salubrious effects.


Should we change the law to more closely regulate executive compensation. If you believe executive compensation is just plain too high, none of the foregoing will matter. You'll want the law to drive down compensation levels. If the absolute level of executive compensation isn't what concerns you, however, the policy question depends on whether you think compensation is set at arms'-length between managers and directors acting on behalf of the shareholders or is essentially set by the managers themselves without regard to shareholder interests.

Bebchuk and Fried clearly fall into the latter category. Accordingly, Pay Without Performance closes with a slew of recommended reforms to change the law.

In this essay, by contrast, I have argued that the managerial power model was never a complete explanation of what is going on with executive compensation. In addition, I note that we've already seen important reforms that promise to significantly constrain both managers and directors. The bottom line thus is that we should wait and see what happens next before rushing forward with more changes in the law.

John Randolph famously opined that "Providence moves slowly, but the devil always hurries." Let's stay on the side of the angels.

Stephen Bainbridge is Professor of Law, UCLA School of Law. He writes two blogs: and Professor Bainbridge on Wine. A considerably longer version of this essay will be published in the Texas Law Review, Volume 83, Issue 6 (May 2005). A draft of that version is available for downloading from the Social Science Research Network.


TCS Daily Archives