C H A P T E R 27
What to Do—and Not Do—About Executive Pay
NEARLY EVERY DAY, some commentator or another is weighing in on the issue of executive pay. There seem to be two complaints. First, executive pay, and particularly chief executive pay, is increasingly excessive. Think about it: in 2005, the average pay of CEOs of large companies was $10.5 million. Furthermore, CEO pay is an ever larger multiple of the pay of front-line employees; by one estimate, CEO pay was 36 times what an average worker earned in 1976, 131 times in 1993, and had ballooned to 369 times as much by 2005.1 This ratio is much higher than that in other industrialized countries such as Japan, Spain, France, or even the United Kingdom. Moreover, the ratio of chief executive pay to the pay of even other executives has doubled since the 1960s.2 The second complaint is that executive pay is not tied to “performance”—typically total return to shareholders as measured by changes in the stock price—so that executives can get rich even if their firms aren’t doing that well and other shareholders aren’t getting great returns.
Many people have suggested reforms, some of which have already been implemented. These reforms typically are of three types. The first entails changing the presumed “power structure” on the board of directors, such as having a nonexecutive chairperson of the board (something that was once rare but is now increasingly common), having compensation committees comprised solely of independent directors, and changing voting rules so that outside shareholders have more influence over the process by which directors are nominated and elected. The second approach focuses on changing the tax rules to constrain executive pay. A number of years ago the tax rules were amended to make pay over $1 million not deductible unless that pay was tied to performance. This change had virtually no effect, as all compensation committee statements in proxies lay out how the CEO pay determination process is related to performance. So, it is unlikely that there are very many—if any—companies subject to this particular reform.
The third type of reform, and the one that seems to get the most attention and advocacy, argues for more disclosure about the components of executive pay and how pay is set. First, companies were required to disclose various forms of incentive compensation such as stock options and bonuses, and compensation committees were required to provide details of their pay-setting philosophy and process. Now there are going to be even more stringent disclosure requirements—for example, details of deferred compensation such as retirement benefits and perquisites such as use of the corporate jet and company-paid apartments, both for active and retired executives.
Having participated on the compensation committees of both private and public company boards, and having read the relatively large empirical literature on executive pay, I am struck by how disconnected from any sort of reality both the commentary and the suggestions for reform are.
It is relatively easy to dispose of the suggestions for reform. There is virtually no evidence they will work. That doesn’t mean they are bad ideas per se, but just that they won’t accomplish what their proponents intend.
Take, for instance, the proposal for more disclosure. In 1993, shareholder activist Ralph Whitworth closed down the advocacy group United Shareholders of America because, with SEC-mandated disclosure of many elements of compensation, he thought the problems with executive pay were over.3 Companies complied with the mandates. Proxy statements expanded to include not just base pay but bonuses as well, and then expanded further to provide information on stock option grants, including the average stock price of the grant and what percentage of all options granted were given to the CEO. Did that disclosure curtail the size of grants? Not that I can see. Did disclosing the disproportional percentage of stock options obtained by the CEO cause a more equal distribution of option grants? I don’t think so. If anything, all this public attention to the various components of pay may have set off an arms race, as executives were interested in not being worse off than their peers and the compensation of peers was ever more visible in the published disclosures. Not wanting to be worse off is the inevitable result of the social comparison process through which people assess how well they are being treated.
The change in the composition of compensation committees and the separation of the role of board chairperson from CEO, implemented several years ago, also did nothing that I can see to halt the rise in executive pay or to tie pay more tightly to performance. Comprehensive summaries of extensive empirical research find no evidence that board structure matters.4 The basic idea underlying the reforms was to get more “independence” into the decision-making process. But psychological independence is a very different thing from legal independence or independence as a matter of organizational affiliation. Many board members have been brought onto the board by the CEO, and even if there is an independent corporate governance committee, it is highly unlikely that a board member will be added without the CEO’s acquiescence. If board members work together, including with the CEO, it is unlikely that they are going to be dispassionate participants in the pay-setting process. Instead, their perceptions of what the CEO should be paid will invariably be affected by their relationship, either positive or negative, with that person. Social influence and norms of reciprocity matter, as research on compensation by Charles O’Reilly has demonstrated.5 If the various reforms have not worked in the past, there is no reason to believe they are going to work in the future.
Executive pay is invariably set through a process of benchmarking—comparing (loosely controlling for things like company size and industry) a focal company’s pay with the pay of other companies. And setting pay through a process of benchmarking actually makes a lot of sense. Yale economist Bengt Holmstrom, who serves on the board of a closely held family company, has noted that because compensation is a sensitive matter, the board (and the CEO) want to avoid contentious, arm’s-length bargaining that might poison the working relationship.6 Using outside consultants and salary surveys helps in this regard.
But as Graef Crystal pointed out more than 15 years ago, this is a process that is going to inevitably produce a rise in average salaries over time.7 That’s because of the “above-average effect”—the tendency for all people to believe they are better than average with respect to judgments about all sorts of things ranging from income to negotiating ability to intelligence. Since CEOs are going to think they are above average, they are going to want to be paid at least the average for comparable CEOs. And, by the way, compensation committees and other board members aren’t going to want to believe that their CEO is below average, because if that were the case, then they wouldn’t be doing their job of hiring and firing executives effectively so the board would then be below average. Since, by definition, half of the people are paid below the median, the desire for everyone to be paid at least the average leads to a continual upward pressure on wage rates.
So why has CEO pay increased so rapidly and increasingly diverged from the pay of lower-level executives and rank-and-file employees? I believe there are a set of interrelated reasons. The first is a complex dynamic: CEOs are getting fired at a higher rate, which means that the average tenure of a CEO is declining and the reputational risk of being a CEO—being fired is not good for one’s image—has increased. Booz Allen Hamilton reported that between 1995 and 2004, CEO turnover increased some 300 percent, and that dismissals of CEOs increased not just in North America but in Europe and Asia as well.8 In 2005, some 15 percent of CEOs left office, and were as likely to leave prematurely as to retire normally.9 If CEOs believe they are increasingly at risk for dismissal and that their tenures are going to be shorter, they are going to demand higher pay as a compensation in return. However, once companies pay their CEOs more money, they are going to have higher expectations for performance and less patience with underperformance, other things being equal. So, higher pay will drive more rapid dismissal, and more rapid dismissal will, in turn, drive demands for higher pay.
Second, it is far from clear that pay in public companies is actually too high, given the prevailing beliefs about the importance of CEOs to company success. One recent survey found that more than half of all board members expect both cash and stock compensation to continue to rise, and the Business Round-table argues that compensation has simply kept pace with increases in company value.10 That does not mean that CEOs are as crucial as conventional wisdom makes out, and in fact there are lots of reasons to think that leaders’ impact is overestimated compared with forces affecting organizational performance such as the general economy and particular industry conditions.11 But it does mean that if an organization and its board believe that the CEO is the factor that will make or break a company, they’ll be willing to pay quite a bit to get the “right” person in the job. Harvard Business School professor Rakesh Khurana has referred to this behavior as “searching for a corporate savior,” arguing that the belief in the potency of the CEO has led to a free-agent market for executive talent and an associated rise in CEO salaries.12
The belief in the potency and importance of the CEO has been driven, in part, by the business press, which has made CEOs into almost rock-star figures. Years ago, I doubt if many people could have named the CEOs of even large and prominent companies, but today, people like Apple’s Steve Jobs, for instance, have almost iconic status. More business biographies and autobiographies are being written, because more people are increasingly interested in knowing about CEOs’ lives and philosophies. And this celebrity status will then be reflected in the salaries paid to those individuals.
Yet other data suggest that CEO compensation is not necessarily unrealistic. One perceived problem—the disconnect between pay and performance—is easily explained. Pay is supposed to motivate behavior. Company performance goals are typically set on an annual basis, and compensation committees set their numbers based on these goals. But if it becomes clear about halfway through the year that, for whatever reason, the targets are not going to be met, arguments are made that if the executive team is to be motivated to do the best they can under the circumstances, the targets need to be changed. Readjusting the targets may motivate the team but invariably leads to this disconnect between pay and company performance.
And we would do well to look at the dramatic increase in the number of companies going private in leveraged buyout transactions. One reason for this trend, given by both organizational leaders and people in the private equity industry, is that the senior executive team can earn a bigger payday in private rather than public companies. If firms controlled by private equity companies, which have to be seen as the most activist of owners, actually pay CEOs of successful enterprises more than they would earn if the company were public, that suggests that CEO pay may not actually be too high. This is Bengt Holmstrom’s implicit conclusion as well, when he noted that there was not much difference in the pay-setting process in a closely held, private company—where there is no issue of dispersed shareholders with inadequate power over the board and the pay determination process—than that observed in the typical public company.13 So, my conclusion is that the supposed “problems” with executive pay may not actually be problems at all.
But if executive pay is going to be fundamentally changed, then two things will need to be altered. First, probably less disclosure, not more, is desirable. One might reasonably ask why the pay of lower level people has not also ratcheted up through the same social comparison process that occurs for CEO pay? I think one answer is that pay comparisons are more difficult to make for people farther down in the hierarchy. That’s because pay secrecy is a common policy in many companies, and some even threaten employees with being fired if they discuss their pay with colleagues. With secret pay and prohibitions on discussions of comparative pay, the social comparison dynamic is short-circuited.
Second, as long as people believe that corporate performance is the result of what the CEO does, the CEO-as-celebrity culture will persist and boards will tend to overpay the CEO. So, ending the CEO pay conundrum will require overcoming the “fundamental attribution error,” the tendency to overattribute causality to the characteristics and actions of individuals as compared to the situations in which they are acting. Since this is such a fundamental psychological principle, I wouldn’t hold my breath.