Stealth Compensation Via Retirement Benefits

This Article analyzes an important form of stealth compensation provided to managers of public companies. We show how boards have been able to camouflage large amounts of executive compensation through the use of retirement benefits and payments. Our study illustrates the significant role that camouflage and stealth compensation play in the design of compensation arrangements. It also highlights the importance of having information about compensation arrangements not only publicly available but also communicated in a way that is transparent and accessible to outsiders. To improve the transparency of executives' retirement payments and benefits, we propose several changes in current disclosure requirements. Among other things, firms should be required to report to investors each year the dollar value of all the retirement benefits to which their executives become entitled. For example, firms should disclose to investors the annual buildup in the actuarial value of executives' retirement plans, as well as the tax savings reaped by executives at the company's expense through the use of deferred compensation arrangements. Firms should also disclose to investors each year the present value of all the retirement benefits their top executives have accumulated.


I. INTRODUCTION
This Article focuses on an important form of "stealth compensation" provided to managers of public companies. We show how designers of compensation arrangements for these managers have been able to camouflage large amount of executive compensation through the use of retirement benefits and payments. Our study highlights the significant role that camouflage and stealth compensation play in the design of compensation arrangements. Our study also highlights the importance of ensuring that information about compensation arrangements not only be placed in the public domain but also be communicated in a way that is transparent and accessible to outsiders.
We begin by discussing the critical role of outrage costs and camouflage in the setting of executive compensation. Managers have considerable influence over their pay and use their influence to extract pay that is both higher and less performance sensitive than arm's-length bargaining with the board would produce. The difference between what managers' power enables them to receive and what they would receive under arm's-length bargaining constitutes "rents." Managers' ability to extract rents, however, is hardly unlimited. When a board approves a compensation arrangement that favors managers at the expense of shareholders, executives and directors will bear certain economic and social costs. The magnitude of these costs will depend on how the arrangement is perceived by outsiders whose views matter to the directors and executives. An arrangement that is perceived as outrageous might reduce shareholders' willingness to support incumbents in a proxy contest or takeover bid, might lead to shareholder pressure on managers and directors, and might embarrass directors and managers or harm their reputations. The more outrage a compensation arrangement is expected to generate, the more reluctant directors will be to approve it and the more hesitant managers will be to propose it in the first place.
The critical role of outsiders' perception of executives' compensation, and the significance of outrage costs, explain the importance of "camouflage." The desire to minimize outrage gives designers of compensation arrangements a strong incentive to try to obscure and justify-or, more generally, to camouflage-both the level and performance-insensitivity of executive compensation. Camouflage thus allows executives to reap benefits at the expense of shareholders. More importantly, attempts to camouflage can lead to the adoption of inefficient compensation structures that harm managers' incentives and in turn company performance, imposing even greater costs on shareholders.
We discuss elsewhere how various forms of non-retirement compensation, including bonuses, stock option plans, and executive loans, have been designed For an electronic copy of this paper, please visit: http://ssrn.com/abstract=583861 2 with an eye to camouflaging rents and minimizing outrage. 1 In this Article, we examine how retirement arrangements have often been designed in a way that serves this goal. As disclosure requirements for executive salaries, bonuses, and long-term compensation have become stricter, firms have increasingly turned to post-retirement payments and benefits as ways to compensate managers. Post-retirement value has been provided to executives through four main channels: retirement pensions, deferred compensation, post-retirement perks, and guaranteed consulting fees. As we will explain, these methods enable firms to provide a substantial amount of performance-insensitive value in a less transparent form than, say, salary. Firms have used these channels to make less transparent both the total amount of compensation received by managers and the extent to which pay is decoupled from managers' own performance.
Before describing outrage costs and camouflage in more detail and discussing each of the four channels, we should note two attributes they all share. First, these arrangements differ substantially from those that firms elect to provide to other employees. Although firms often provide pensions and deferred compensation to lower-level employees, they do so only to the extent that these arrangements receive a tax subsidy. This pattern suggests that, absent such a subsidy, pensions and deferred compensation are generally not efficient. Yet most of the arrangements provided to executives do not enjoy similar tax advantages. Furthermore, consistent with economists' belief that in-kind benefits are inefficient, firms do not generally provide retired employees with coverage for specified consumption expenses. Such benefits are, however, given to high-level executives. And although firms occasionally use retired employees as consultants when the need arises, they generally do not guarantee lifetime consulting fees to any employees other than executives.
The second shared attribute of these various retirement payments is that they all make it possible to obscure large amounts of performance-decoupled compensation. As we shall see, firms do not have to disclose the value transferred to executives through these channels in the same way that other forms of compensation-such as salary, bonuses, and stock options-must be disclosed. Retirement payments hence offer what might be called "stealth compensation." 2 Indeed, the dollar figures used by the media in reporting compensation levels, and by financial economists in their studies, usually do not include the large value provided to executives through retirement benefits. 2. We borrow the term "stealth compensation" from Robert Monks, who used it to refer to executives' stock option compensation because that form of payment is not expensed on the firm's income statement. ROBERT  The remainder of this Article proceeds as follows. Part II describes the importance of outrage costs and camouflage in the setting of executive compensation. Part III discusses the widespread use of supplemental executive retirement plans (SERPs). It explains how SERPs differ from the pension benefits provided to regular employees and how they can be used to camouflage a significant amount of performance-insensitive compensation to executives. Part IV discusses the deferred compensation arrangements offered to managers. It describes how these plans differ from the 401(k) plans offered to other employees and how these plans, like SERPs, are used to provide a significant amount of performance-decoupled pay to executives in a way that is largely hidden from view. Part V considers the use of post-retirement perks and consulting contracts. Part VI explains why increased transparency of retirement arrangements would be beneficial to shareholders and proposes several changes to the disclosure requirements for retirement payments and benefits. Part VII concludes.

II. OUTRAGE COSTS AND CAMOUFLAGE
This Part explores the role and significance of outrage costs and camouflage in the setting of executive compensation. Section II.A explains why outsiders' perceptions and the possibility of outrage are of concern to boards when they fashion pay packages for managers. Section II.B describes the key role of camouflage in the design of compensation arrangements. Section II.C provides empirical evidence on the effect of outrage and camouflage on managerial pay.

A. The Importance of Outsiders' Perceptions
As we discuss and document in great detail elsewhere, 3 top executives have considerable influence on their own pay arrangements. Although directors are supposed to negotiate with executives at arm's-length, they have both financial and non-financial incentives to provide managers with pay arrangements that favor managers at the expense of shareholders. A variety of psychological and social factors acting on the directors reinforce these incentives to serve managers' interests. Moreover, neither shareholder pressure nor market forces have been able to effectively constrain managerial influence over pay.
Managers have used their power to extract pay that is both higher and less performance-sensitive than arm's-length bargaining with the board would produce. The difference between what managers' influence enables them to receive, and what they would receive under arm's-length bargaining, is called "rents." The rents captured by managers come in both the form of higher pay and reduced pressure to generate value for shareholders.
However, managers' ability to extract these rents is hardly unlimited. When a board approves a compensation arrangement that favors managers at the expense of shareholders, executives and directors may bear certain economic and social costs. Although market forces, the need for board approval, and social sanctions do not altogether prevent deviations from arm's-length contracting, they do, as we explain below, place some constraints on managers' ability to obtain favorable compensation packages, and the tightness of these constraints depends on outsiders' perceptions of these pay arrangements.
In the face of these constraints, how far firms will go in favoring managers will depend not only on how much contemplated arrangements will actually favor executives, but also on how these arrangements will be perceived by outsiders. Whether directors and managers are deterred from adopting a given compensation arrangement depends on the extent to which it will be viewed by relevant outsiders as unjustified or even abusive or egregious. We have broadly referred to negative reactions by outsiders as "outrage," even though some of them may amount to criticism not reaching the level of outrage, and to the costs that such reactions impose on managers and directors as "outrage costs." 4 The more widespread and strong these negative reactions are-that is, the greater the outrage-the larger the costs to directors and managers. When the potential outrage costs are large enough, they will deter the adoption of arrangements that managers would otherwise favor. Arrangements that are deterred in this way can be regarded as ones that violate the "outrage constraint." Why should perceptions-and, in particular, outrage-matter? To begin with, the extent to which markets penalize managers and directors for the adoption of particular arrangements depends on how these arrangements are perceived. Consider the market for corporate control. This market may penalize the adoption of arrangements by increasing the vulnerability of managers and directors to a control contest. Such a penalty is likely to be significant only if the firm adopts compensation arrangements that appear sufficiently outrageous. Institutional investors may view such arrangements as a strong signal that the executives or directors are relatively insensitive to shareholder interests. These investors may become less likely to support the incumbents should a hostile takeover or a proxy fight occur. In this manner, through the operation of the market for corporate control, outrage over compensation can impose a penalty on managers and directors.
Consider also the labor market and the reputation of managers and directors in this market. Reputational damage might have an adverse effect on the future career prospects of managers and directors. It might also affect their current business dealings with others outside the firm. Indeed, some outside directors join boards partly for the prestige and connections that the posts provide, and gaining a bad reputation could eliminate these benefits and impose costs instead. Reputational losses to managers and directors will likely be significant, however, only if their firms adopt compensation arrangements that generate sufficiently negative reactions-that is, sufficient outrage. An arrangement that fails to serve shareholders would be unlikely to impose such costs as long as it falls within the range of what is perceived as conventional and legitimate.
Indeed, we believe that arrangements that are perceived as abusive or outrageous impose on executives greater costs than an analysis based solely on the above market incentives suggests. That is, we believe that constraints on rent extraction are somewhat tighter than suggested by an analysis of the (limited) market penalties that outrageous compensation arrangements involve. As we have explained elsewhere, 5 directors are affected not only by "narrow" interests of a homo economicus, but also by various social and psychological factors (such as collegiality, loyalty, and so forth) that pull them in the direction of favoring executives. Similarly, there are social and psychological factors that increase the costs that managers and directors incur from adopting arrangements that are viewed by outsiders as sufficiently outrageous.
Managers and directors likely care about the extent to which relevant social and professional groups view them with approval and esteem. Directors are also likely to prefer to avoid criticism or ridicule from the social or professional groups whose opinions they value-even if such criticism or ridicule does not involve any economic losses for them. 6 As a result, even if the economic incentives provided by the markets for corporate control and managerial labor would be insufficient to deter managers from seeking certain outrageous compensation, fear of embarrassment or criticism could discourage managers from doing so. When former General Electric CEO Jack Welch made headlines by giving up much of the retirement perks to which he was contractually entitled-including the free use of a corporate jet and a New York apartmenthe was undoubtedly seeking to protect the approval and esteem he had earlier enjoyed at the expense of his narrow economic interests. 7 Clearly, for outrage to impose significant costs, it must be sufficiently widespread among a relevant group of observers. It is not enough for a small group of researchers or arbitrageurs to identify a compensation scheme as egregiously bad for shareholders. For executives or directors to be adversely affected in a material way by market penalties or social costs, the outrage must be shared by those outsiders whose views matter most to them: the institutional investor community, the business media, and social and professional groups to 5 For an electronic copy of this paper, please visit: http://ssrn.com/abstract=583861 6 which directors and managers belong.

B. Camouflage
The main costs to directors and managers of adopting compensation arrangements that favor managers, then, depend mainly not on how costly the arrangements actually are to shareholders, but on how costly the arrangements are perceived to be by important outsiders. Perceptions matter. This brings us to another concept that is critical for understanding the compensation landscape: camouflage.
Because perceptions are so important, the designers of compensation plans can limit outside criticism and outrage by dressing, packaging, or hiding-in short, camouflaging-rent extraction. The more reasonable and defensible a package appears, the more rents managers can enjoy without facing significant outrage. Accordingly, under the managerial power approach, managers will prefer compensation practices that obscure the total amount of compensation, that appear to be more performance based than they actually are, and that package pay in ways that make it easier to justify and defend.
The greater the ability of plan designers to engage in camouflage, the more they can be expected do so. Before 1992, the SEC required firms to report executive compensation to the public but allowed them to do so in the format of their choosing. Not surprisingly, firms took full advantage of their discretion to obscure the amount and form of their pay. An SEC official describes the pre-1992 state of affairs as follows: The information [in the executive compensation section] was wholly unintelligible . . . . [T]he typical compensation disclosure ran ten to fourteen pages. Depending on the company's attitude toward disclosure, you might get reference to a $3,500,081 pay package spelled out rather than in numbers. That gives you an idea of the nature of the disclosures: it was legalistic, turgid, and opaque; the numbers were buried somewhere in the fourteen pages. Someone once gave a series of institutional investor analysts a proxy statement and asked them to compute the compensation received by the executives covered in the proxy statement. No two analysts came up with the same number. The numbers that were calculated varied widely. 8 In 1992, the SEC tightened its disclosure rules by providing standards for how information about executive pay must be presented. The standardized compensation tables that firms now must use have made camouflage more difficult. As we describe elsewhere, however, the 1992 disclosure requirements have hardly brought an end to firms' ability to camouflage the amount and form of executive pay. 9 One might reasonably ask how, if rent extraction is camouflaged, any observer (including this Article's authors) can determine that executives are enjoying rents. In theory, rent extraction could be camouflaged so well that it becomes absolutely undetectable. In fact, camouflage is successful as long as the rent extraction is not apparent to those outside observers whose outrage would be particularly costly for directors and managers, even if other observers are aware that the executives are enjoying large rents. Thus, the notion of camouflage is consistent with the possibility that an outsider might identify the hidden rents of a compensation arrangement. Such a conclusion would simply reflect the observer's judgment, not yet widely shared, that the compensation program is distorted in favor of managers. In time, of course, such conclusions might become widely accepted, in which case the rent extraction will no longer be camouflaged. But a given form of rent extraction might continue to be camouflaged long after it has been recognized by some observers.

C. Outrage and Camouflage at Work
Some critics of our earlier work argued that the idea of outrage costs, and the related idea of camouflage are not empirically testable. 10 This, however, is not the case. There is evidence that directors and executives are indeed influenced-in compensation and other types of decisions-by strong outside criticism and outrage. And there is evidence that they engage in camouflage.
To begin with, there is evidence that shareholder precatory resolutions that criticize managers' high compensation have an impact. Although such resolutions are nonbinding and generally fail to pass anyway, their appearance may shine a critical light on problematic aspects of the firm's executive compensation policies and make them less opaque. Indeed, a study by Randall Thomas and Kenneth Martin examined the effect of pay-related precatory resolutions during the mid-1990s and found that they had a moderating influence on subsequent compensation decisions. 11 The study found that during the two-year period following the passage of shareholder resolutions criticizing executive pay in particular firms, total compensation (adjusted for industry) in those firms declined by a statistically significant average of $2.7 million. In a subsequent study, the researchers also found that higher negative votes on management-sponsored proposals to ratify an option plan slowed the increase in CEO compensation in subsequent years. 12 Another study, by Alexander Dyck and Luigi Zingales, documents the effects of media scrutiny on corporate decisions in general. The authors found 10 that such attention leads firms to adopt more environmentally friendly policies, for example. As for issues of corporate governance, they also found that media attention reduces the amount of value that controlling shareholders siphon off. 13 A well-known example of how outside criticism affects governance decisions involves the campaign of shareholder activist Robert Monks against the directors of Sears. During the late 1980s and early 1990s, Monks urged the Sears board to adopt various proposals to improve the firm's dismal performance. In April 1992, having been repeatedly ignored by the board, Monks took out an advertisement in the Wall Street Journal titled "The Nonperforming Assets of Sears" and identified the directors by name. The presumably embarrassed directors then adopted many of Monks's proposals, generating an abnormal stock price return (the change in stock price adjusted for overall stock market movements) of almost 10% when the changes were announced. 14 Another example is the California State Pension Fund for Public Employees' (CalPERS) practice of identifying poorly run companies. For some years, CalPERS put poorly performing firms on what it called its "focus list" and suggested various ways to improve their corporate governance practices, such as making compensation and nominating committees fully independent. In many cases, firms placed on the list implemented some of the requested changes. Then, in 1991, after several CEOs told CalPERS that being less antagonistic would be even more effective, CalPERS decided to adopt a "kinder, gentler" approach that did not involve public shaming. 15 Absent the threat of adverse publicity, however, firms approached by CalPERS were actually much less cooperative. The then-CEO of CalPERS, Dale Hanson, said at the time, "It has shown us that a number of companies won't move unless they have to deal with the problem because it's in the public eye." In 1992, CalPERS reinstated its policy of publicly shaming uncooperative firms. 16 In fact, CalPERS' policy of shaming has had a measurable effect on targeted corporations. Yi-Lin Wu found that firms put on CalPERS' poor governance focus list were subsequently more likely to reduce the number of inside directors on their boards. These firms were also more likely to experience CEO turnover. 17 Shaming also appears to have adversely affected the careers of inside directors that left the targeted firms' boards. They were much less likely than inside directors departing nontargeted firms to land other board positions. 18 As this study makes clear, negative publicity-or outragedoes impose costs.
Finally, and perhaps most importantly, there is substantial evidence of camouflage activities. A testable implication of the camouflage idea is that when compensation arrangements deviate from arm's-length bargains, they should do so in a way that makes the amount of pay or the insensitivity of pay to performance less visible. This prediction is borne out by actual compensation practices. As we have shown elsewhere, many common nonretirement compensation practices-such as company loans and the structure of conventional options-provide camouflage benefits. 19 And as we will explain below, the four channels through which executives are paid after retirement also serve to obscure a significant amount of compensation.

III. RETIREMENT PENSIONS
Many employees are covered by pension plans that provide payments to workers after retirement. At first glance, it seems only natural for firms to provide such benefits to their executives. A closer look, however, raises serious questions about whether the extensive use of executive pensions as a form of compensation reflects arm's-length bargaining. Part III.A describes the difference between executive retirement pensions and the retirement benefits offered to ordinary workers. Part III.B explains how the executive pensions are used to camouflage a substantial amount of performance-decoupled executive pay.

A. Differences from Regular Pensions
Most of the pension plans used for employees are designed to be "qualified" for favorable tax treatment. The firm gets a current deduction for contributing funds to a qualified plan for employees-the same deduction it would have received had it paid the amount of the contribution to workers in the form of salary. Workers, however, do not pay income taxes on the pension money until they retire and begin receiving payouts from the plan. In the meantime, the funds invested by the firm grow tax-free. Neither the firm nor the employees must pay any taxes while the plan's investments increase in value. Thus, the plans provide a tax benefit to employees at no cost to the firm. 20 18. Id. 19. See BEBCHUK AND FRIED, PAY WITHOUT PERFORMANCE, supra note 1, at chs. 9-14. 20. To illustrate how the tax subsidy provided to a qualified plan operates, consider the following examples involving a hypothetical firm and employee. Assume that both face a 40% tax rate on all of their income, including capital gains. Also assume that both are able to earn, between the pre-retirement period and retirement period, a pretax return of 100% on their investments.
Given the opportunity, boards might well prefer to offer executives qualified retirement plans. A qualified pension plan, however, can use only about $200,000 of annual compensation as the basis for determining benefits under the plan. For example, a plan that promises to pay all retirees, annually, 50% of the compensation earned during their last year of service cannot pay a retired executive more than $100,000 annually, even if the executive earned $1 million of compensation during that final year. As a result, firms cannot use qualified plans to provide executives with pensions that are similar in size to their annual compensation. For this reason, most firms also provide executives with nonqualified "supplemental" executive retirement plans (SERPs). 21 SERPs differ from typical qualified pension plans in two critical ways. First, they do not receive the favorable tax treatment enjoyed by qualified plans; no investment income goes untaxed under a SERP. The company pays taxes on the income it must generate in order to pay the executive in retirement. If the money had been distributed as salary, on the other hand, the executive who invested the money for retirement would have had to pay taxes on any income generated. The effect of the SERP, therefore, is to shift some of the executive's tax burden to the firm. 22 If the employee and the firm are subject to the same tax rate and are able to earn the same pretax rate of return on their investments, a SERP cannot reduce the total amount of taxes paid by the parties. For every dollar the employee's tax burden is reduced, the firm's tax burden is increased by one dollar. Unlike a qualified plan, the SERP would not reduce the parties' total tax burden. 23 Example 1: The employee invests for retirement outside a qualified plan. Suppose the firm pays the employee $100 in the pre-retirement period. The firm deducts $100 from its taxable income, reducing its tax liability by $40. The employee pays $40 in taxes, takes the after-tax income of $60, and invests it. The $60 grows to $120 by the retirement period-a gain of $60. This $60 gain triggers a tax liability of $24 (40% of $60), leaving the employee with $96 ($60 + $36) when the employee retires. Example 2: The firm invests for the employee's retirement under a qualified plan. Now suppose the firm contributes the $100 to a qualified pension plan in the pre-retirement period. The firm again deducts $100 from its taxable income, reducing its tax liability by $40. The $100 grows to $200 by the time of the employee's retirement-a gain of $100. The $200 is distributed to the employee, who pays a tax of $80 (40% of $200), leaving the employee with $120, or $24 more than in Example 1, in which the employee had received $100 from the firm in the pre-retirement period and saved for retirement. The gain to the employee does not come at the expense of the employer: in both examples, the employer incurs an aftertax cost of $60 in the pre-retirement period.
21. Clark Consulting reports that approximately 70 percent of responding firms use SERPs. 22. A firm can shelter from taxation the investment income on funds set aside for financing executive pensions by investing these funds in life insurance policies on the lives of its executives and other employees. However, because of the fees that must be paid to the insurance company, this taxsheltering mechanism involves significant costs, which are borne by the company rather than the executive. If, on the other hand, the executive received the funds to begin with, the executive would also be able to shelter the investment returns from taxation by purchasing a variable annuity, at no cost to the company.
23. To illustrate the effect of a SERP on the tax burdens of the parties, consider the following example and explanation, which builds on the examples provided in note 20. Assume again that both the In reality, of course, the situation is more complicated. 24 In many cases, the total tax liability faced by the parties will be affected by whether the executive or the firm saves for the executive's retirement. Even if the firm and the executive are able to earn the same return on their investments, they may face different tax rates. Suppose, for example, that an executive investing personal funds for retirement in the stock market is paying a low long-term capital-gains tax rate of 15%, while the firm pays taxes on the income generated for the executive's retirement at a corporate tax rate of 35%. In such a case, using SERPs would be tax-inefficient and would increase the total amount of taxes paid by the two parties. On the other hand, if the firm had no taxable earnings and was not expected to pay taxes for a considerable amount of time, the reverse might be true: shifting retirement savings from the executive to the firm might be tax-efficient. 25 Similarly, even if the firm and the executive face the same tax rate, the investment returns available to the firm may be higher than those available to the executive. For example, firms having difficulty raising capital may enjoy a higher expected rate of return on new investments than the market generally. (This is unlikely to be the case for companies with easy access to capital, as firm and the executive face a 40% tax rate on all of their income, including capital gains. And assume that both are able to earn, between the pre-retirement and retirement periods, a pretax return of 100% on their investments. Example 3: The firm invests for the executive's retirement under a nonqualified plan. Suppose a firm seeks to use a SERP to give an executive the same retirement payment that it gives the employee in example 2 using a qualified plan. As in the case of the employee, the firm sets aside $100 to fund the executive's pension, which grows to $200 by the time the executive retires. The $200 is distributed to the executive, who, like the employee, pays a 40% tax on the retirement distribution-a tax of $80. This leaves the executive, like the employee in example 2, with $120, $24 more than the employee in example 1 made. Now consider the effect of the SERP on the firm. In examples 1 and 2, discussed in note 2, the firm reduces its tax liability by $40 in the pre-retirement period when it pays the worker $100 or contributes $100 to the worker's qualified pension plan. In example 3, the firm reduces its tax liability by $80 in the retirement period when it pays the executive $200. However, the firm must add to its taxable income in the retirement period the $100 gain on the funds it previously invested for the executive's retirement, and this increases the firm's tax liability in the retirement period by $40. The net effect of the $200 payment to the executive and the $100 gain is to reduce the firm's tax liability by $40 during the retirement period. Had the firm reduced its tax liability by $40 in the pre-retirement period, rather than during the retirement period, it could have invested the $40 and earned a pretax return of $40 (100%) by the retirement period. That $40 would also have been taxed at 40%, leaving the firm with $64. But by reducing its tax liability in the retirement period, the firm has only an extra $40, $24 less. The firm is thus worse off than in example 2, in which it received the same $40 reduction in its tax liability in the pre-retirement period. The $24 gain to the executive from the use of a nonqualified plan designed to put the executive in the same position as an employee under a qualified retirement plan comes at the expense of the firm.
24. For an explanation of the tax effects of using arrangements such as SERPs to defer compensation under various scenarios, see MYRON S. SCHOLES ET AL., TAXES AND BUSINESS STRATEGY: A PLANNING APPROACH 181-185 (Prentice Hall, 2d ed., 2002).
25. The tax efficiency of a SERP will also be affected by expected changes in the firm's (or the executive's) tax rate change over time. For example, if the firm is losing money and thus unable to get a current tax benefit by deducting executive compensation in the current period, but is expected to be subject to a higher tax rate in the future, deferring an executive's compensation will be tax efficient, all else being equal. such companies are unlikely to have unutilized investments with returns much higher than the market.) If the firm has better investment opportunities, having it invest for the executive's retirement will be efficient for both parties, even if their tax rates are identical.
However, there is no reason to believe that, absent a tax subsidy, it is generally efficient to have the firm save for the executive. On the contrary, there are good reasons to think that it is inefficient for many firms to save for their executives' retirement, given individuals' low long-term capital-gains tax rate. It is telling that firms providing SERPs to executives do not offer nonqualified retirement plans to other employees. Consider the case where it is efficient for a firm to provide a SERP to its executives because the firm has better investment opportunities than they do. In such a case, it should also be efficient for the firm to provide nonqualified retirement to its nonexecutive employees who supplement their qualified pensions with personal retirement savings. However, firms rarely, if ever, do so. This fact suggests that, absent the tax subsidy provided to qualified plans, using nonqualified retirement benefits is commonly not an efficient way to compensate employees. Yet in 2002, more than 70% of firms provided nonqualified SERPs to their executives. 26 The second important difference between executive SERPs and qualified pension plans for nonexecutive employees concerns the risk borne by the firm and by the participant. Qualified pension plans offered to new lower-level employees are usually based on a defined contribution. The firm commits to contribute a specified amount each year. The value available to an employee upon retirement depends on the performance of the plan's investments. The risk of poor investment performance falls entirely on the employee.
In contrast, SERPs offered to executives are defined-benefit plans, which guarantee fixed payments to the executive for life. All of the CEOs in the S&P ExecuComp database have defined-benefit plans. 27 These plans shift the risk of investment performance entirely to the firm and its shareholders. No matter how poorly the firm and its investments perform, the executive is guaranteed a specified lifelong stream of payments.
Given that arm's-length negotiations with most employees lead to definedcontribution arrangements, why should arm's-length bargaining with executives yield such a different result? If anything, there are reasons to believe that defined-benefit plans should be more valuable to regular employees-and thus offer a more efficient form of compensation-than they are to executives. Unlike most executives, ordinary employees are unlikely to accumulate substantial wealth over their lifetimes. They are likely to be more dependent on their pensions to meet their financial needs in retirement and therefore less able 26 to bear the investment risks associated with defined-contribution plans. In contrast, executives faced with defined-contribution plans could easily insure themselves against poor investment performance by using some of their already high salaries and option-based compensation to buy fixed annuities that would provide them with guaranteed payments. If only one of the two groups were to receive defined-benefit plans, arm's-length contracting would predict that group to be nonexecutive employees, not executives.

B. Camouflage Benefits
Although the efficiency benefits of providing executives with definedbenefit SERPs are far from clear, such plans do considerably reduce the visibility of a substantial amount of performance-insensitive compensation.
SERP payments are usually based on years of service and pre-retirement cash compensation. The higher the executive's salary and the longer the period of employment, the higher the payout. SERP payments-like salary-are therefore largely decoupled from the executive's own performance. Many firms have also credited executives with years they did not actually serve, ratcheting up the final payout under the plan's formula. 28 In their annual public filings, firms must publish a summary compensation table indicating the dollar value of different forms of compensation received by the current CEO and the four other highest-paid executives of the firm. The numbers in these tables are the most visible indicators of executive compensation in public firms. They are easily accessible to the media and others reading the public filings. Indeed, the standard databases of executive compensation, which are used by both financial economists and compensation consultants, are based on these numbers.
If an executive's pensions were structured as a defined contribution plan, the firm's annual contributions to the executive's account would be reported in the summary compensation table. An important camouflage benefit of SERPs is that the annual increase in the present value of an executive's defined benefit plan-due to pay raises and the addition of another year of service-is largely hidden from view: firms are not required to include this increase in value in the compensation tables. A person examining the compensation tables would not see the steady buildup in value of an executive's SERP.
Furthermore, and importantly, disclosure requirements require firms to include in their summary compensation tables only amounts paid to their current executives. Because the executives are no longer employed by the firm when the pension payments begin, the payments need not be included in the published tables. Thus, the value of an executive's defined-benefit SERP never appears in the place where the media and researchers collect most of their information about executive compensation. In addition, because the value of an executive's pension payouts is obscured, the performance insensitivity of such payments also gets little notice.
Consider a situation in which a CEO serves a company for ten years and then receives annually, for life, a payment that equals a large fraction of the salary earned during the last year of service. In such a case, the total value of the pension payments may in the end exceed the total value of the salary received during the CEO's actual tenure. Unlike the salary amounts, however, the value of the pension payments will never appear in the firm's published compensation tables.
For example, when IBM CEO Louis Gerstner retired after about nine years of service, he was entitled to a $1,140,000 annual pension beginning at age sixty. 29 The actuarial value of this annuity was of a similar order of magnitude as the approximately $18 million in salary he received during his nine years as CEO. IBM, however, was not required to include the pension in the summary compensation table or even place a dollar value on it. 30 Not surprisingly, SERP plans are designed and marketed specifically as ways to increase compensation "off the radar screen of shareholders." 31 Indeed, according to media reports, some directors have voted to adopt SERPs only after being reassured that the amounts involved do not have to be reported to the public. 32 To be sure, although neither the increase in value of the SERP plan before retirement nor the amount of payments after retirement appears in the compensation tables, the existence of SERPs, and the formulas under which payouts are made, must be disclosed in the firm's SEC filings. 33  precision the value-and thus the cost to the company-of these future payments. 34 As noted above, firms are not required to supply, and usually do not provide, any estimate of the dollar value of a particular executive's definedbenefit pension plan. The lack of easy access to the monetary values of these substantial benefits presumably explains their absence from the standard databases used for research on executive compensation.
Indeed, it is often difficult even to figure out the total SERP liability of a firm with respect to its executives as a group. A firm must report only one figure: the sum of the liabilities associated with all of its employee pension plans that are "unfunded" or "underfunded" (that is, plans for which the firm does not have assets set aside to cover the plans' liabilities fully). 35 The Financial Accounting Standards Board (FASB) does not require that liabilities associated with SERPs be itemized separately. 36 Thus, firms can simply report one number that represents all the liabilities associated with underfunded qualified plans and unfunded SERPs.
Although they are not required to do so, some firms do report the total obligations arising under SERPs. These figures can be staggering. In 2000, for example, GE reported a $1.13 billion pension liability for all of its executives. 37 Unfortunately, GE did not report what portion of this amount was due specifically to its CEO and other top executives. Most companies do not even break down pension liabilities into separate categories for executives and other employees.
It is worth noting at least one way in which executives' plans may not be as advantageous to their beneficiaries as the plans of lower-level employees. Firms using qualified plans are required, as a condition for favorable tax treatment, to set aside assets to ensure that they can pay their liabilities under the plans. Given that executives' SERP plans would not qualify for the favorable tax treatment even if they were so funded, firms do not bother funding SERP plans. Executives' retirement benefits are thus at greater risk of nonpayment than the benefits of ordinary workers-and Congress is considering legislation that would make it difficult for firms to shelter executives from this risk. 38 In the past, however, firms facing financial problems have often purchased insurance policies that guaranteed payment of executive retirement benefits, transferred the money to a designated trust, or taken other steps to guarantee the benefits against insolvency. 39 Delta Airlines, for example, set up an executive-protecting arrangement shortly after September 11, 2001, when the solvency of the airline industry appeared to be in danger. 40 Although putting the money beyond the reach of the firm's creditors triggers a tax liability for the executive, firms often "gross up" the payment to cover part or all of that liability. 41 It was reported in 1991 that approximately fifty major companies had set up fully guaranteed executive pension plans. 42 This practice may have been much more widespread; many firms, fearing criticism that they are insulating managers from the effects of their own failures, have failed to announce the existence of such guarantees. 43

IV. DEFERRED COMPENSATION
Deferred compensation is a second technique used to transfer large amounts of mostly performance-insensitive value to executives without attracting much shareholder attention. Many firms offer programs that permit executives, or sometimes even require them, to defer receipt of compensation until some future date. In the meantime, the deferred compensation "builds" according to a formula devised by the firm. Executives do not pay taxes on the original compensation or on the accumulated increase until they receive payment, which often occurs after they leave the company. At that time, the firm takes a tax deduction for the amount paid. Most large companies have plans of this kind. 44 Deferred compensation plans can take different forms. Some firms require that managers receiving salary in excess of $1 million, which would otherwise be nondeductible under Section 162(m) of the Internal Revenue Code, defer the excess. Other firms have purely elective plans. Some arrangements permit deferral of salary only, while others also allow deferral of long-term incentive compensation and gains from the exercise of stock options or from the sale of restricted stock. Companies frequently provide matching contributions, with 39 the amounts varying from firm to firm. At some companies, contributions are awarded at the board's discretion. At others, they are determined by formulas. 45 Plans also differ in how the deferred compensation is "invested," that is, how the amount owed to the executive at the end of the deferral period is determined. Many companies provide a guaranteed rate of return (or a guaranteed minimum rate) on the funds. 46 Firms have often granted extra benefits to executives by providing rates of return that are higher than the market rate. For example, in 2001, at a time when one-year Treasury bills offered returns of 3.39% to 4.63%, both GE and Enron guaranteed executives a 12% rate of return. Other firms have offered a market return plus a premium. For example, Lucent has offered the return on the ten-year Treasury bill plus 5%. 47 Congress is now considering legislation aimed at preventing firms from providing executives with above-market returns in their deferred-compensation plans. Although the adoption of such legislation would eliminate this particular benefit to managers, deferred compensation plans would still provide executives with significant other financial and camouflage advantages as we discuss below. Part IV.A identifies the differences between executive deferred compensation arrangements and the 401(k) plans offered to other employees. Part IV.B describes the camouflage benefits of executive deferred compensation arrangements.

A. Differences from 401(k) Plans
Deferred-compensation arrangements appear analogous to the familiar 401(k) plans used by many employees. But, just as SERPs differ from the qualified retirement plans offered to lower-level employees, there are some important differences between executives' deferred compensation and 401(k) plans.
To begin with, the 401(k) plans give workers an opportunity to put money in designated investment instruments; whatever the investments, employees get the same pretax returns they would receive by investing in similar instruments outside the 401(k) plan. In contrast, executives' deferred-compensation arrangements often provide higher returns than those available in the market.
In addition, 401(k) plans are given a tax subsidy, while executive deferredcompensation plans are not. Under a 401(k) plan, a fraction of the employee's salary is placed in a tax-deferred account. The firm may also make a separate contribution to the account. As in a qualified retirement arrangement, the funds are invested and grow tax-free. Neither the firm nor the employee pays taxes on 45 the income and capital gain generated in the account. Employees do not pay taxes on the contributions or the increase until they withdraw the funds. The employer, on the other hand, gets a deduction for both its contribution and the employee's contribution to the 401(k) plan. By placing current compensation in a 401(k) account, the employee gains the benefit of tax deferral without the employer's loss of a tax deduction. 48 Firms could provide deferred compensation to executives through 401(k) plans. However, there are limits on how much money can be contributed annually to a 401(k) account. For the tax year 2004, employees covered by such a plan ordinarily cannot defer more than $13,000 of compensation. 49 In order to provide executives with amounts exceeding this limit, firms implement deferred-compensation arrangements outside the tax-advantaged framework of 401(k) plans. Executives' deferred compensation is therefore not based solely, or even primarily, on 401(k) plans.
Rather than contribute a portion of the executive's compensation to an account where the investment grows tax-free, the firm simply withholds part of the executive's pay and credits the executive each year with a prespecified return on the money, allowing it to "grow" over time. The withheld compensation, along with the appreciation credited to it by the firm, is paid to the executive at a later date.
The company pays taxes on the income it must generate in order to pay the executive the promised buildup of the deferred compensation. If, on the other hand, the deferred compensation had been distributed when it was originally owed the executive, the executive would have invested the money and paid taxes on any income or capital gains subsequently generated. Thus, as in the case of a SERP, the effect of executive deferred compensation is to shift some of the executive's tax burden to the firm. 50 48. To illustrate how the tax subsidy provided to a 401(k) operates, consider the following examples involving a hypothetical firm and employee. As in the SERP examples found in note 20 (examples 1 and 2), assume that both the firm and the employee face a 40% tax rate on all of their income. Assume also that both are able to earn, between the pre-retirement and retirement periods, a pretax return of 100% on their investments. Example 4: The employee saves outside the 401(k) plan. Suppose the firm pays the employee $100 in the pre-retirement period. The firm deducts $100 from its taxable income, reducing its tax liability by $40. The employee pays $40 in taxes, and invests the aftertax income of $60 in an ordinary, nonqualified investment account. By the retirement period, the $60 grows to $120-a gain of $60. The employee pays a tax of $24 on the gain (40% of $60), leading to an aftertax gain of $36. The employee is thus able to withdraw a total of $96 ($60 + $36). Example 5: The employee saves under a 401(k) plan. Now suppose that the employee contributes $100 of compensation income to a 401(k) account. The firm again deducts $100 from its taxable income, reducing its tax liability by $40. The $100 grows to $200 by the time the employee withdraws the funds from the 401(k) account. The employee pays a tax of $80 (40% of $200), leaving the employee with $120-$24 more than in example 4, where the employee received $100 from the firm in the preretirement period and saved the money outside the 401(k) plan. The $24 gain to the employee does not come at the expense of the employer. In both examples, the employer pays the employee $100 in the pre-retirement period, thereby reducing its taxable income by $100 and its tax liability by $40.

A company can shelter from taxation investment income on funds set aside for financing
If the employee and the firm are subject to the same tax rate and are able to earn the same pretax rate of return on their investments, executive deferred compensation, like a SERP, cannot reduce the parties' joint tax burden. While every dollar of deferred compensation lowers the executive's taxes, it boosts the firm's taxes by one dollar. Like a SERP, and unlike qualified 401(k) and retirement plans, deferred-compensation plans for executives provide no taxefficiency benefit when the firm and the executive share the same tax rate and investment opportunities. 51 As in the case of SERPs, of course, there will be many cases in which deferred compensation outside 401(k) plans can increase or reduce the total amount of value available to the executive and the firm. 52 The firm and the executive may face different tax rates. Even if the firm and the executive face the same tax rate, the investment returns available to the firm may be higher than those available to the executive (although, as we noted in our discussion of SERPs, this is unlikely to be the case for companies with easy access to capital). However, there is no reason to believe that, absent the tax subsidy provided by qualified plans, there is generally a benefit to the parties when the firm defers the executive's compensation. In many cases, the tax burden on the firm is greater than the tax benefit to the executive, increasing the total tax that executive pensions by investing these funds in insurance policies on the lives of its executives and other employees, but this will impose other costs on the firm. See SCHOLES ET AL., supra note 24, at 181-185. 51. To illustrate the effect of executive deferred-compensation arrangements on the tax burdens of the parties, consider the following example and explanation, which refer to examples 4 and 5 provided in note 48. Example 6: The firm offers the executive deferred compensation outside a 401(k) plan. Assume, as in examples 4 and 5, that both the firm and the executive face a 40% tax rate on all of their income, including capital gains. And assume that both are able to earn, between the pre-retirement and retirement periods, a pretax return of 100% on their investments. Suppose the firm seeks to use deferred compensation to give an executive the same (100%) return that the firm provides the employee in example 5 using a 401(k) plan. As in the case of the employee, the firm sets aside $100, which grows to $200 by the time the executive withdraws the deferred compensation and the buildup credited to the designated amount of deferred compensation. The $200 is distributed to the executive. Like the employee, the executive pays 40% tax on the retirement distribution-a tax of $80. This leaves the executive, like the employee in example 5, with $120, or $24 more than the employee saving on his own ended up with in example 4. Now, let us consider the effect of the executive's deferred compensation arrangement on the firm. In examples 4 and 5, the firm reduces its tax liability by $40 in the pre-retirement period when it pays the worker $100 or contributes $100 to the worker's qualified pension plan. In example 6, the firm reduces its tax liability by $80 in the retirement period when it pays the executive $200. However, the firm must add to its taxable income in the retirement period the $100 generated to boost the executive's withdrawal payout from $100 to $200-which in turn increases the firm's tax liability by $40. The net effect of the $100 gain and the $200 payment to the executive is to reduce the firm's tax liability by $40 during the retirement period. The firm is thus worse off than in example 2, where it received the same reduction in its tax liability in the pre-retirement period. Had the firm reduced its tax liability by $40 in the earlier period, it could have invested the $40 and earned a pretax return of $40 (100%) by the retirement period. The $40 would have been taxed at 40%, leaving the firm with $64. By reducing its tax liability in the retirement period, the firm has only an extra $40, or $24 less. Thus, the $24 gain to the executive from the use of a deferred-compensation arrangement designed to put the executive in the same position as an employee under a qualified 401(k) comes at the expense of the firm.
52. For an explanation of the tax effects of deferred compensation under various scenarios, see SCHOLES ET AL., supra note 24, at 181-85. the two parties pay to the government.
Consider, for example, the case in which an executive of a profitable company is promised a return that is linked to a stock index. If the executive invests the money in shares of a stock index fund, the gains will be taxed at the long-term federal capital-gains rate, which in the highest bracket is 15% (as of 2004). 53 If, instead, the firm invests the money-in those shares, other investments, or its own business-the gains could be taxed at the marginal corporate rate of 35%. 54 Thus, it is puzzling that over 90% of firms offer deferred-compensation programs to their executives. 55 As in the case of SERPs, there are good reasons to think that, in many firms, such programs are not an efficient form of compensation. It is curious that firms offering nonqualified deferredcompensation arrangements to executives do not offer such nonqualified plans to other employees. After all, if nonqualified deferred compensation is an efficient form of compensation for the executives of certain firms-say, because the firms have better investment opportunities than the executivesnonqualified deferred compensation should also be an efficient form of compensation for the nonexecutive employees of these firms. But firms rarely, if ever, provide nonexecutive employees with the option of nonqualified deferred-compensation arrangements in addition to their 401(k) plans. This pattern suggests that, in most cases, offering nonqualified deferred compensation to an executive does not increase the joint wealth of the executive and the firm.

B. Camouflage Benefits
While it is far from clear that deferred-compensation arrangements provide efficiency benefits, their camouflage value is substantial. The compensation being deferred must be reported in the summary compensation table in the year in which it would otherwise have been received. However, the substantial benefits that have been conferred by the deferred-compensation plan-the taxfree (and sometimes above-market) buildup over time-are not evident to outsiders.
Even assuming that the nominal rate of return used by a deferred-53. I.R.C. § 1. 54. I.R.C. § 11. As in the case of SERPs, a firm can reduce the tax cost of deferred compensation by using company-owned life insurance. Under this strategy, the firm uses after-tax dollars to buy insurance on the lives of its executives and other employees. Part of the premium is invested, increasing the "cash value" of the policy. The policy is then cashed out when funds are needed to pay deferred compensation. The tax savings come from life insurance policies' capacity to shelter from taxes the buildup of the cash value. However, because the insurance company charges fees, the use of a life insurance policy to avoid taxes gives rise to transaction costs. A 1996 study found that 70% of the 1,000 largest firms did not use insurance for funding deferred compensation, which suggests that these costs can be quite high. compensation arrangement is no higher than the market rate, the effective interest rate earned by executives is higher than it appears because of the substantial tax benefits. Executives must pay taxes on investment income earned outside deferred-compensation arrangements, but investing within such plans provides them-at the expense of the firm-with a tax-free buildup. Thus, as long as the rate of return in deferred-compensation arrangements is above the executive's after-tax rate of return, the executive makes substantial gains that do not show up in the compensation tables. The New York Times reported, for example, that CEO Roberto Goizueta of Coca-Cola was able to defer taxes on $1 billion of compensation and investment gains over a seventeen-year period. 56 Coca-Cola picked up the tab, paying taxes on the earnings needed to cover the returns credited to Goizueta's deferredcompensation account. 57 Furthermore, while 401(k) plans offer lower-level workers returns equivalent to those available in the bond or stock markets, many deferredcompensation arrangements have provided executives with substantially higher returns. These executives have thus received investment income that was not only tax-free for them (at the expense of the firm) but also above-market. The benefits from these above-market returns have also been hidden to a significant extent.
The SEC requires firms to include in the summary compensation table the above-market interest earned that year by each executive on deferred compensation. In the case of a guaranteed interest rate, "above-market" interest is defined as returns in excess of 120% of the applicable federal rate (AFR) used by the IRS at the time the guaranteed interest rate is set, multiplied by the amount of deferred compensation. By exploiting the SEC's definition of "above-market rate," firms have sometimes been able to provide their executives with rates of return that are higher than those they could get on their own without including this benefit in the compensation tables.
The threshold used by firms for "market" long-term rates of return is especially generous because boards can reset interest rates whenever doing so benefits executives. If market interest rates and the AFR rise so that the current guaranteed rate is not especially attractive, the firm can simply adopt a new, higher, guaranteed rate. As long as the reset rate is lower than 120% of the new, higher AFR, the additional interest accruals need not be reported in the summary compensation table. If, however, market interest rates and the AFR fall, the firm can continue to pay at the old guaranteed rate, which is now above market. And because the AFR used for the disclosure threshold is that prevailing when the guaranteed interest rate was initially set, no matter how low market rates drop, the above-market interest paid to the executive never 56. Tax Deferred Pay for Executives, N.Y. TIMES, Oct. 18, 1996, at A36. 57. According to Coca-Cola's annual reports to shareholders, it paid taxes on its income in every year of Goizueta's tenure except in 1992. appears in the compensation table.
Finally, even benefits that have come from rates of return exceeding the SEC's threshold are unlikely to be fully reflected in the compensation tables. The reporting requirement ends when the executive retires, but the executive often has the option to continue enjoying the above-market rates after retirement. Such a stream of post-retirement benefits-which could be quite substantial in value-would never appear in the firm's publicly filed compensation tables.
As in the case of SERPs, deferred-compensation plans could expose executives to the risk of firm bankruptcy. While 401(k) plans must be backed by their assets, which cannot be seized by the firm's creditors, deferredcompensation arrangements are simply a promise by the firm to pay compensation in the future. The executives owed this compensation are unsecured creditors who may not be paid in full if the firm becomes insolvent. As in the case of SERPs, Congress is considering legislation that would make it difficult for firms to shield executives from this possibility. To date, however, firms have often taken steps to insulate executives from insolvency risk. Many firms have used "security devices," such as trusts, to ensure that funds will be available to the executives. In addition, firms have usually permitted executives to withdraw deferred compensation at any time-such as when inside information suggests that a firm is about to fail. Shortly before Enron filed for bankruptcy, for example, its executives withdrew millions of dollars of deferred compensation.
For executives and their friends on the board, SERPs and deferred compensation have been very useful. They have provided a means for channeling large amounts of performance-insensitive compensation in a way that, under current disclosure regulations, has not been highly visible to outsiders. As one compensation analyst pointed out: "The disclosure of the myriad executive compensation plans-pension, supplemental executive retirement plans, deferred compensation, split-dollar life insurance-is not adequate in answering a fundamental question: what is the projected value of these plans to the executive upon his retirement?" 58

V. POST-RETIREMENT PERKS AND CONSULTING CONTRACTS
We now turn to consider the use of post-retirement perks and consulting contracts to convey a significant amount of performance-decoupled value to executives in a way that is not transparent to shareholders. Part V.A describes some of the perks provided to executives and explains why they are unlikely to result from arm's-length bargaining between the parties. Part V.B examines the use of post-retirement consulting agreements.

A. Perks
Many compensation contracts promise executives a substantial stream of perks after retirement. For example, many executives receive a certain number of hours of corporate aircraft use annually for themselves, and sometimes for their families and guests as well. Some executives have even received unlimited lifetime use of corporate aircraft. 59 Other perks that often follow the executive into retirement include chauffeured cars, personal assistants, financial planning, home-security systems, club memberships, sports tickets, office space, secretarial help, and cell phone service. 60 Outgoing IBM CEO Louis Gerstner, for example, was given access to apartments, planes, cars, home-security services, and financial planning. Terrence Murray, former CEO of FleetBoston, received 150 hours of company aircraft use, a chauffeured car, an office, office assistants, financial planning, and a home-security system.
Another common benefit is giving contributions to charities designated by the retiring executive. FleetBoston gave retiring CEO Murray the ability to direct $3.5 million of the firm's charitable contributions to Murray's favorite institutions. 61 In addition, Ford promised retiring CEO Jacques Nasser to endow a scholarship in his name at the educational institution of his choice (in addition to providing Nasser a new car each year, financial-planning assistance, an office, and an assistant). 62 Most of these perks cost the company more than may be apparent at first glance. Consider retiree use of corporate jets, now a common perk. Although the marginal cost of allowing a retired executive to use the company jet may appear limited, 63 it can run quite high. Consider the use of a company plane for a flight from New York to California and then back several days later. Because the New York-based aircraft and flight crew will return to the East Coast after dropping the retired executive off, the actual charge to the company is two round trips: a total of eight takeoffs and landings and approximately twenty hours of flying time, most likely costing-for fuel, maintenance, landing fees, extra pilot and crew fees, incidentals, and depreciation (an aircraft's operating life is reduced for every hour it flies and, more important, for every takeoff and landing)-at least $50,000. 64 Henry R. Silverman, CEO of Cendant, was promised lifetime use of the corporate aircraft or, if the plane was in use, an equivalent chartered plane at a direct cost of thousands of dollars per hour. 65 Firms usually do not provide post-retirement perks to nonexecutive employees. There is good economic logic to avoiding such in-kind compensation. Promising a retiring employee $10,000 a year for certain travel expenses is less efficient than providing $10,000 in cash. The reason is straightforward. If the retiree views travel as the best way to spend $10,000, the cash and the travel coverage will have identical utility. However, cash is superior if there are any possible circumstances in which the retiree would prefer spending some or all of the money on goods or services other than travel, because the retiree will receive greater utility at the same cost to the firm.
A retiree's needs and preferences are likely to change over time. Thus, economic logic suggests that if in-kind retirement benefits are provided, they should not be provided for long periods. Yet such long-term, in-kind benefits are often provided to retired CEOs: for example, Louis Gerstner of IBM received use of a plane, cars, offices, and financial planning services for ten years. 66 Although post-retirement perks are unlikely to be an efficient form of compensation, they offer an effective means of camouflaging compensation. The value of post-retirement perks is not reported when they are agreed to, and the firm incurs costs only after the executive has left, at which point any value provided is no longer included in the salient summary compensation table. Post-retirement perks thus offer yet another way of providing additional value to executives without ever having to include the benefits in compensation tables or even place a dollar value on them.
Firms in the past have sometimes grudgingly provided vague descriptions about post-retirement perks, but these descriptions did not generally allow shareholders to form a good picture of the scope and value of these benefits. For example, former General Electric CEO Jack Welch received approximately $2.5 million in benefits in his first year of retirement, including access to GE aircraft for unlimited personal use and for business travel, exclusive use of a furnished New York City apartment, and unrestricted access to a chauffeured limousine (among other things). 67 However, GE's proxy statements revealed only that in retirement Welch would be entitled to "continued lifetime access to Company facilities and services comparable to those that are currently made available to him by the Company." 68 Thus, GE not only failed to put a dollar value on the perks, but did not bother even to describe them.

B. Consulting Contracts
Like perks, consulting contracts provide substantial value to retired executives. They usually offer the retiring CEO an annual fee for "being available" to advise the new CEO for a specified amount of time per year. Approximately 25% of CEOs negotiate a post-retirement "consulting" relationship with their old firm. 69 For example, AOL Time Warner is paying retired CEO Gerald M. Levin $1 million a year to serve as an adviser for up to five days a month. 70 In 2000, retiring Carter-Wallace CEO Henry Hoyt was promised annual payments of $831,000 for a similar monthly obligation. 71 Verizon co-CEO Charles Lee negotiated a $6 million consulting contract for the first two years of his retirement. Delta Airlines CEO Ronald Allen's 1997 retirement package provided him with a seven-year, $3.5 million consulting deal under which, according to Delta's public filings, he was "required to perform his consulting services at such times, and in such places, and for such periods as will result in the least inconvenience to him." 72 Allen or his heirs will be entitled to the annual fee of $500,000 even if he is totally disabled or dies. 73 These consulting arrangements provide flat, guaranteed fees for the retired executive's "being available" rather than payment for work actually done, and for a good reason: companies generally make little use of the availability for which they pay generously. For better or worse, new CEOs are usually not inclined to seek advice from their predecessors. 74 Allen, for example, reportedly "rarely talks" with the new Delta chief executive, Leo Mullin. Even compensation consultants acknowledge that retired executives add little if any value to the firm under these arrangements. According to Frank Glassner, CEO of Compensation Design Group, most of these consulting contracts are merely a way of increasing the severance payment to the departing executive. According to another executive compensation expert, Alan Johnson, "Most former CEOs are doing very little for what they're getting paid . . . . Usually, the demands [from new management] are miniscule." Like post-retirement perks, the consulting payments to retired executives never find their way into the summary compensation tables because they are provided when the executives are no longer officers. However, in contrast to 68 post-retirement benefits, these contracts enable boards to provide retired executives with cash rather than in-kind benefits. 75 Retirement consulting fees are essentially a cash severance payment, turned over in installments, disguised as compensation for post-retirement work.
If these fees are just a form of cash severance, what is the advantage of packaging them as consulting agreements? Besides ensuring that the payments are kept out of the compensation tables, dressing them up as consulting fees obscures their nature as severance payments that essentially increase the total compensation received by executives for their pre-retirement work. Some observers might believe that the outgoing CEO will in fact provide valuable advice to new management, and therefore view the payments as legitimate consideration for post-retirement services. Needless to say, these consulting agreements do not tie the retired executive's pay to any personal contribution to shareholder value either before or after retirement.

A. The Critical Role of Transparency
We turn in this Section to discussing the policy implications of our study. Our study highlights the importance of making compensation arrangements in general, and compensation via retirement benefits in particular, transparent to public investors. Although we argue elsewhere for reforms that would increase shareholder power, 76 shareholders do already have some power. This power is in part why the outrage constraint matters. The greater outsiders' understanding of compensation arrangements, the tighter the outrage constraint. Improving the transparency of compensation arrangements is therefore desirable.
Financial economists have paid insufficient attention to transparency because they often focus on the role of disclosure in getting information incorporated into market pricing. It is widely believed that information can be reflected in stock prices as long as it is known and fully understood by even a limited number of market professionals.
In the case of executive compensation, there is already significant disclosure. As we have discussed, SEC regulations require detailed disclosure of the compensation of a company's CEO and of the four most highly compensated executives other than the CEO. In our view, however, it is important to recognize the difference between disclosure and transparency, and it is transparency that should receive more attention. The main aim of requiring disclosure of executive compensation is not to enable accurate pricing of the firm's securities. Rather, this disclosure is primarily intended to provide some check on arrangements that are too favorable to executives. This goal is not well served by disseminating information in a way that makes the information understandable to a small number of market professionals but opaque to others.
The ability of plan designers to favor managers depends on how compensation arrangements are perceived by a wide group of investors and other outsiders. Because of market forces and social dynamics, managers and directors are concerned about disapproval (threatened or actual) from institutional investors or other reference groups, such as the business press or popular media. We have seen that compensation designers often seek to make the amount of pay, or the extent to which pay is decoupled from performance, less transparent. For disclosure to constrain compensation effectively, the disclosed information must reach more than just a select group of market professionals and arbitrageurs. Raw facts buried in a mountain of technical disclosure probably will not suffice. The salience of disclosure and degree of transparency are important.

B. Putting Retirement Benefits on the Radar Screen
Public officials and governance reformers, therefore, should work to ensure that compensation arrangements are and remain transparent. Having shown that pensions, deferred compensation, and post-retirement perks and consulting contracts have been used to camouflage a significant amount of performancedecoupled compensation, we now put forward several proposals designed to put these forms of compensation on investors' radar screens.
As we explain in more detail below, firms should be required to report each year the (present) dollar value of the future benefits to which executives become entitled during the year in connection with SERPs, deferred compensation, and post-retirement perks and consulting arrangements. These monetary values should be included in the firm's annual summary compensation table. The SEC should also require firms to provide a separate table disclosing the present value of all the retirement benefits-SERP payments, deferred compensation balances, and post-retirement perks and consulting arrangements-to which the CEO and other high ranking executives would be entitled were they to separate at year end (under each of the various termination scenarios contemplated in their employment contracts). Before proceeding, it is worth noting that SERPs, deferred compensation, and postretirement perks and consulting contracts are not necessarily the only forms of stealth compensation. Firms have provided-and will continue to look for ways to provide-other forms of stealth compensation. Thus, while our proposals will substantially reduce the ability of firms to provide stealth compensation, they will not eliminate it completely.

Disclosing the Full Details of Executive Compensation Contracts.
Firms should be required to disclose in their annual proxy filing the full text of the most recent versions of the employment contracts that the firm has entered into with high-level executives. Firms often file such agreements with forms other than the annual proxy statement that are not as closely followed. Requiring firms to attach executive compensation contracts to their most important SEC filing, which contains the summary compensation table and other key details about compensation,, will make executive compensation agreements more accessible and salient.
However, requiring firms to disclose in an accessible and salient way the text of executive compensation contracts is far from enough. While experts who take the time to do so may be able to calculate the monetary value of the various forms of stealth compensation provided in these contracts, these values are not as transparent to investors as the dollar values reported for many other forms of executive compensation. To make the information contained in these compensation contracts more transparent and accessible, firms should be required to attach monetary values to each of the types of compensation we have discussed. That is, the SEC should require firms to report compensation via retirement payments, deferred compensation, and post-retirement perks and consulting contracts in the same manner as other forms of compensation: by putting a dollar value on these amounts and including them in the compensation tables accompanying firms' proxy filings. We explain below, for each type of retirement benefit we have discussed, how this can be accomplished.
SERPs. As we explained, SERP payouts are generally based on years of service and historic compensation levels. The actuarial value of SERPs therefore usually increases each year. The annual buildup in value of an executive's SERP is not, under current rules, reported in firms' annual summary compensation tables. Nor are the payments themselves, which are made after the executive retires.
We propose requiring firms to add a column in the annual summary compensation table for SERPs. This column would indicate, for each of the firm's highest-paid executives, the amount by which the actuarial value of the executive's SERPs increases each year. Firms should also provide an accompanying explanation of the reasons for the increases in actuarial value reported in the compensation tables.
Deferred Compensation. Recall that firms currently must report in the publicly filed summary compensation tables only "above-market" returns credited to executives' deferred compensation accounts. As we explained, however, firms have been able to exploit the SEC's definition of "abovemarket" to provide above-market returns without disclosing them in the compensation tables. Moreover, firms have not been required to report the tax benefit to the executive-which comes at the expense of the firm-provided by deferred compensation arrangements.
We propose adding disclosure requirements that would provide outsiders with a clear and full picture of the gains to executives from deferred compensation programs. Firms should be required to disclose each year the value of an executive's deferred compensation account at the beginning of the year, the amount of earnings credited to this account by the end of the year, and the basis on which these earnings were determined. This information will enable outsiders to decide for themselves whether and to what extent executives enjoy above-market returns.
Firms should also report to investors the tax cost to the firm of covering the earnings. For example, if a firm in the 33% tax bracket credits $600,000 to an executive's deferred compensation account, the firm should disclose not only the $600,000 credited to the account, but also the $300,000 in taxes that the firm must pay on the $900,000 in pre-tax income needed to cover the $600,000 earnings credited to the executive's account.
Finally, it is necessary to ensure that summary compensation tables include the full monetary value executives derive from deferred compensation. We propose that compensation tables include a column reporting this monetary value. In calculating this value, firms should be required to regard as "abovemarket" any returns exceeding the AFR-not only returns exceeding 120% of this rate. For this purpose, firms should be required to use the AFR for shortterm loans in effect at the beginning of the firm's fiscal year. Furthermore, and perhaps more importantly, the monetary value included in the compensation tables should include any gains to the executive that come from having returns accumulate tax-free at the company's expense.
Post-retirement Perks and Consulting Contracts. Post-retirement perks and consulting contracts never appear in the summary compensation tables because the value is provided to the executive after she retires. This is the case even when such arrangements are promised in the executive's contract long before the executive retires-either when she is initially hired or later when she negotiates a new contract.
The SEC should modify the compensation tables to include a column for post-retirement perks and consulting contracts. Firms should be required to report, in this column, the actuarial present value of post-retirement perks and consulting contracts promised to executives. Critically, these amounts should be disclosed as soon as the executive is promised these payments-not when the payments are made and the executive is already out the door. Thus, for example, if an incoming CEO negotiates a package that includes the use of a corporate jet for ten years after he retires, the firm should place a dollar amount on the value of this arrangement and include that amount in the entry for non-SERP retirement payments.
Disclosing the Retirement Kitty. The proposals we have just offered would require firms to report the compensation executives receive each year in the form of buildup in pension value, earnings, and tax benefits associated with deferred compensation, and promises of post-retirement perks and consulting contracts. In addition, firms should be required to put all this information together to provide a complete picture of how much the executive would receive from the firm if she were to separate at the end of the year, under each of the termination scenarios described in her employment contract.
In particular, the SEC should require firms to publish a table showing the actuarial value of each high ranking executive's SERP at the end of the year, the balances in any deferred compensation accounts, and the value of any postretirement perks and consulting contract to which the executive would be entitled if she were to resign, retire, or otherwise separate from the firm at year's end. This picture will better enable shareholders to form a judgment as to whether it is necessary for the firm to continue spending money to ensure that the executive has a comfortable retirement. The table will also enable shareholders to easily see whether executives will enjoy a "soft landing" even if they are pushed out for failure.
We should emphasize that a retirement-payout table is not a substitute for our proposals to expand the annual summary compensation tables to include the yearly monetary value associated with SERP buildup, deferred compensation programs, and promises of post-retirement perks and consulting contracts. These measures are necessary to make transparent to investors how much executives have been paid during the past year, as well as to help investors estimate the relationship between annual pay and the executives' own performance. In contrast, the retirement-payout table is necessary to make transparent to investors the extent, if any, to which it would be necessary for subsequent pay packages to provide executives with additional retirement benefits.
Transparency and Improved Compensation Arrangements. The measures above would provide shareholders with a more accurate picture of total executive compensation. They will thereby reduce the total amount of compensation executives could get below the radar screen. These measures could thus help constrain total compensation levels. More importantly, these measures also would reduce the distortions that arise when companies choose particular forms of compensation for their camouflage value rather than for their efficiency. Improving transparency in this area could thus substantially increase shareholder value.
Our analysis has focused on the most important forms of stealth compensation via retirement benefits that have been used by firms to date. Of course, designers of compensation plans may find and use new ways to make compensation, or its insensitivity to performance, more opaque. As new practices (and new means of camouflage) develop, disclosure arrangements should be updated to ensure transparency. Regulators adopting the measures we propose should continue to monitor compensation arrangements and to refine disclosure requirements as new ways of making pay less salient are developed.

VII. CONCLUSION
This Article has explained how retirement benefits and payments have been used to camouflage the payment of large amounts of performance-insensitive compensation to executives of public companies. Our study has highlighted the significant role that camouflage and stealth compensation play in the design of compensation arrangements, as well as the importance of ensuring that information about compensation arrangements be communicated in a way that is transparent and accessible to outsiders.
We have also proposed several measures aimed at putting retirement benefits on investors' radar screens. Among other things, firms should make transparent, on a timely basis, both the monetary benefits to executives and the cost to firms of the various arrangements we have described. In particular, the summary compensation table firms must publicly file should include entries that indicate: (1) the amount by which the actuarial value of executives' SERP plans increases each year; (2) the monetary value executives derive that year from the firm's deferred compensation plans, including the value of a tax-free buildup at the firm's expense; and (3) the present value of post-retirement perks and consulting contracts promised to the executives during the year. Firms should also be required to include a separate table showing how much of each form of retirement compensation a high ranking executive would receive if she were to separate at the end of the year under each of the termination scenarios contemplated by her contract. By making it more difficult to camouflage pay through retirement benefits, the proposed requirements would contribute to improving compensation arrangements.