3.10.1 Pay Restrictions for Financial Institutions

While the pay restrictions in the TARP legislation apply only to banks receiving government assistance, the Dodd–Frank Act goes much further by regulating pay for all financial institutions (public or private, TARP recipients and non-recipients) including broker-dealers, commercial banks, investment banks, credit unions, savings associations, domestic branches of foreign banks, and investment advisors. Specifically, Part (a) of Section 956 of the Dodd–Frank Act requires all financial institutions to identify and disclose (to their relevant regulator) any incentive-based compensation arrangements that could lead to material financial loss to the covered financial institution, or that provides an executive officer, employee, director, or principal shareholder of the covered financial institution with excessive compensation, fees, or benefits. In addition, Part (b) of Section 956 of the Dodd–Frank Act prohibits financial institutions from adopting any incentive plan that regulators determine encourages inappropriate risks by covered financial institutions, by (1) providing an executive officer, employee, director, or principal shareholder of the covered financial institution with excessive compensation, fees, or benefits; or (2) that could lead to material financial loss to the covered financial institution.

Since at least the early 1990s, there has always been a tension between shareholders (the firm’s legal owners) concerned about CEO incentives, and uninvited guests (such as politicians and labor unions) concerned about high levels of pay. After the TARP bailouts in the financial crisis, the analogous tension was between taxpayers (who wanted to be protected from excessive risks while receiving an appropriate return on their investment) and politicians who were outraged about perceived excesses in banking bonuses. Section 956(b) of the Dodd–Frank Act deliberately conflates these tensions, by explicitly defining excessive compensation as an inappropriate risk. Moreover, Section 956(a) of the Dodd–Frank Act requires banks to inform their regulators of compensation plans that provide excessive compensation, delegating to the regulators the Herculean task of defining what compensation is excessive (or, indeed, which risks are inappropriate).

The responsibility for implementing Section 956 of the Dodd–Frank Act fell jointly to seven agencies: the Securities and Exchange Commission (SEC), the Federal Reserve System, the Office of the Comptroller of the Currency, the Office of Thrift Supervision, the Federal Deposit Insurance Corporation, the National Credit Union Administration, and the Federal Housing Finance Agency. In March 2011, the seven agencies issued a joint proposal for public comment, modeled in part on Section 39 of the Federal Deposit Insurance Act. While the proposal stops short of explicitly limiting the level of executive compensation, it prohibits compensation that is unreasonable or disproportionate to the amount, nature, quality, and scope of services performed. In addition, the proposal calls for firms to identify individuals who have the ability the expose the firm to substantial risk, and demands that (for the larger institutions) such individuals have at least 50% of their bonuses deferred for at least three years; deferred amounts would be subject to forfeiture if subsequent performance deteriorates. Final rules were expected in late 2012.

3.10.2 Pay and Governance Reforms for all Publicly Traded Companies

While ostensibly focused on regulating firms in the financial services industry, the authors of the Dodd–Frank Act seized the opportunity to pass a sweeping reform of executive compensation and corporate governance imposed on all large publicly traded firms across all industries. The new rules include:

Say-on-Pay (Section 951). Shareholders will be asked to approve the company’s executive compensation practices in a non-binding vote occurring at least every three years (with an additional vote the first year and every six years thereafter to determine whether the Say-on-Pay votes will occur every one, two, or three years). In addition, companies are required to disclose, and shareholders are asked to approve (again, in a non-binding vote), any golden parachute payments in connection with mergers, tender offers, or going-private transactions.

 In January 2011 – and effective for the 2011 proxy season – the SEC adopted rules concerning shareholder approval of executive compensation and “golden parachute” compensation arrangements. Shareholders of 98.5% of the 2532 companies reporting 2011 results by July 2011 approved the pay plans; over 70% of the companies received more than 90% favorable support.¹⁴³Similarly, shareholders of 98.2% of the 1875 companies reporting 2012 results by June 2012 approved the pay plans; 72% of the companies received more than 90% favorable support.¹⁴⁴Twenty six of the 30 companies receiving less than 50% positive votes in 2011 passed in 2012, and year-over-year favorable votes increased by 14% for companies receiving between 50% and 70% favorable votes in 2011.

 Clawbacks (Section 954).

Companies must implement and report policies for recouping payments to executive based on financial statements that are subsequently restated. The rule applies to any current or former executive officer (an expansion of Sarbanes-Oxley, where only the CEO and CFO were subject to clawbacks), and applies to any payments made in the three-year period preceding the restatement (Sarbanes-Oxley only applied for the twelve months following the filing of the inaccurate statement).

 The SEC had neither adopted nor proposed rules regarding the recovery of executive compensation by August 2012. However, Equilar reports that 86% of the Fortune 100 companies issuing proxy statements in 2012 had publicly disclosed clawback arrangements; in half of the companies the clawback triggers were related to financial restatements and ethical misconduct.¹⁴⁵

 Additional Disclosures (Sections 953, 955, 972).

Companies must report the ratio of CEO compensation to the median pay for all other company employees. Companies must analyze and report the relation between realized compensation and the firms financial performance, including stock-price performance. In addition, companies must disclose its policies regarding hedging by employees to protect against reductions in company stock prices. Finally, the Dodd–Frank Act requires companies to disclose their policies and practices on why the company chooses either to separate the Chairman and CEO positions, or combine both roles.

 The SEC had neither adopted nor proposed rules regarding the disclosure of pay ratios, pay-for-performance, hedging and CEO/Chair combinations by August 2012.

 Compensation Committee Independence (Section 952).

Following Sarbanes-Oxley (2002) requirements for Audit Committees, publicly traded companies are required to have compensation committees comprised solely of outside independent directors (where independence takes into account any financial ties the outside directors might have with the firm. In addition, companies must assess the independence of compensation consultants, attorneys, accountants, and other advisors to the compensation committees.

 In June 2012, the SEC adopted final rules directing exchanges to establish listing standards guaranteeing that members of the compensation committee (or directors who oversee executive compensation matters in the absence of a committee) to be independent. While leaving the precise definition of “independence” to the exchanges, the final rule required exchanges to consider the director’s source of compensation (including consulting or advisory fees) paid by the issuer, and whether the director is affiliated with the issuer, a subsidiary of the issuer, or an affiliate of a subsidiary of the issuer.

 In addition, the new SEC rules require firms to ensure that compensation committees have authority and funding to retain compensation consultants. While neither the Act nor the June 2012 Final Rule issued by the SEC required compensation advisors to be independent, the SEC imposed a list of independence criteria that boards must consider in retaining a consultant. Finally, proxy statements issued in connection with annual shareholder meetings in 2013 and after must disclose whether the work of the consultant has raised any conflict if interest and, if so, the nature of the conflict and how the conflict is being addressed.

 Proxy Access (Section 971).

The Dodd–Frank Act authorized the SEC to issue rules allowing certain shareholders to nominate their own director candidates in the company’s annual proxy statements.

 The SEC issued its rules on Proxy Access in August 2010, but delayed implementation after lawsuits by the Business Roundtable and the US Chamber of Commerce claimed that the rules would distract management and advance special-interest agendas. In July 2011, the US Circuit Court of Appeals (Washington, DC) ruled in favor of the business groups and rejected the SEC’s rule. As of August 2012, the SEC had not announced whether it would attempt to rewrite the rule in a way that would be acceptable to the Court.

It is too early to assess the ultimate effect of Dodd–Frank on executive compensation, since many of the rules have just been implemented or are still being written. However, based on experiences with similar rules, I can speculate on the ultimate impact.

Say on Pay. In mandating Say-on-Pay, the Dodd–Frank Act follows similar rulings for non-binding shareholder votes enacted in the United Kingdom in 2002 and later in Australia, Denmark, France, Portugal, Spain, and Sweden; the Netherlands and Norway went a step further by allowing binding shareholder votes. Say-on-Pay had long been a favorite objective of Democrats in Congress, and the Say-on-Pay Bill passed the House in April 2007 by a 2:1 margin. While the companion bill introduced in the Senate by then-Senator Obama was shelved prior to a vote, Say-on-Pay was widely expected to become law following the 2008 presidential election, especially after Say-on-Pay was mandated for TARP recipients as part of the Dodd Amendments.

In spite of the support, however, there is modest evidence that Say-on-Pay results in important changes to compensation practices. In the United Kingdom (where we have the most data), there is some evidence that negative Say-on-Pay votes have led to some reductions in salary continuation periods in severance agreements and some changes in performance-based vesting conditions in equity plans, but no evidence that the votes have affected compensation levels (Ferri and Maber, 2010). In the United States, where shareholders voted on the compensation for TARP executives for the first time in early 2009, the plans were passed at all firms, with an average of 88.6% of the votes cast in favor of management. Among the TARP recipients garnering the strongest support were the Wall Street firms whose compensation systems allegedly fostered the financial crisis, including Goldman Sachs (98%), AIG (98%), JPMorgan (97%), Morgan Stanley (94%), Citigroup (84%), and Bank of America (71%).¹⁴⁶

As emphasized in this chapter, regulation inevitably produces unintended consequences. The most obvious (and most negative) unintended consequence associated with Say-on-Pay reflects the increasing influence of proxy-advisory firms (primarily Institutional Shareholder Service (ISS)). To fulfill their required fiduciary duties to vote proxies, institutional investors routinely rely on ISS and other proxy-advisory firms for recommendations on how to vote on Say-on-Pay and other proxy matters. In turn, the proxy-advisory firms rely on a limited (and controversial) set of quantitative criteria to determine whether to offer positive or negative voting recommendations.¹⁴⁷ In a broad sample of Russell 3000 firms, Larcker, McCall, and Ormazabal (2012) show: (1) the recommendations of the proxy-advisory firms do, indeed, affect voting outcomes; (2) anticipating this result, firms change their compensation policies to avoid negative recommendations; and (3) the market reaction to these changes is statistically negative.

Firms inherently face different competitive and incentive challenges, and there is neither a “one-size-fits-all” solution to these challenges, nor a limited set of quantitative criteria that can substitute for a careful and holistic assessment of compensation plans that takes into account company-specific situations and objectives. Ultimately, the benefits of adhering to the ISS criteria must be weighed against the cost associated with reduced innovation and flexibility in the provision of compensation and incentives.

Compensation Committee and Advisor Independence. The Dodd–Frank provisions on the independence of the compensation committee will have little practical effect for large companies, since the listing requirements of the NYSE and NASDAQ have required independent compensation committees since 2003, and the IRS has required independent compensation committees (for Section 162(m) purposes) since 1994. The provision related to the independence of compensation consultants, in combination with SEC disclosure rules introduced in December 2009, will encourage more committees to retain their own independent consultant in addition to the consultants engaged by management.¹⁴⁸

Clawback Provisions. The Sarbanes-Oxley experience shows that companies rarely try to recover erroneously awarded compensation from their CEO and CFO, often citing potential litigation costs and the feasibility of recouping money that has already been paid and taxed. The Dodd–Frank provision makes it more difficult for boards to shirk their responsibility to recovery erroneously awarded pay, and indeed likely subjects boards to shareholder litigation if they fail to even try.

Ratio of CEO-to-Worker Pay. The most mischievous and controversial compensation provision in Dodd–Frank is the required disclosure of the ratio of CEO pay to the median pay of all employees. The calculation costs alone can be immense for large multinational or multi-segment corporations where payroll is decentralized: to compute the median the company needs an often non-existent single compensation database with all employees worldwide. More importantly, however, is what shareholders are supposed to do with this new information, or how they should determine whether a ratio is too high or too low. Ultimately, this provision reflects a belief in Congress that CEO pay is excessive and its sole purpose is the hope that disclosing the ratio will shame boards into lowering CEO pay.

Proxy Access. Finally, potentially most important is the Proxy Access rule allowing shareholders to include their director nominees on the proxy alongside with the board’s nominees. In issuing its rule in August 2010, the SEC limited access to shareholders who have held at least 3% of the company’s stock for at least three years. One view is that Proxy Access will provide shareholders with a critical mechanism to replace poor directors with better ones. A more-cynical view—expressed by the Wall Street Journal and others—is that 3% was chosen as the sweet spot for labor unions and other politically motivated organizations who will use their leverage over the proxy statement to force companies to support political causes rather than increasing shareholder value.¹⁴⁹ In its July 2011 ruling rejecting the SEC’s rule, the US Circuit Court of Appeals (Washington, DC) issued a sharp rebuke to the SEC, saying that the SEC failed in analyzing the cost the rule imposes on companies and in supporting its claim that the rule would improve shareholder value and board performance.¹⁵⁰

5.1.1 Boards of Directors

The first line of defense against agency problems is the board of directors, elected by shareholders and responsible for monitoring, hiring, firing, and setting the compensation of the CEO and top-management team. For most of the prior century, boards were dominated by current executives and other corporate insiders. However, beginning with the shareholder movement in the 1980s (Section 3.6.2 above), firms have faced pressures for increased outsider representation on boards. By the end of the 1990s, the fraction of outside directors serving on the average board had increased to 80%, and the CEO was the sole insider in nearly half of all firms (Horstmeyer, 2011).

Conceptually, outside directors reduce agency problems by threatening errant executives with termination and by implementing incentive contracts that tie pay to value creation. The contracts that evolve from this setting will typically tie CEO pay to the creation of shareholder value, thus providing the theoretical justification for stock options, restricted stock, and other forms of equity-based compensation. Under the efficient contracting hypothesis, the contracts will be those that maximize shareholder value, while paying the CEO enough “expected” compensation or utility to get him to take the job, and recognizing that CEOs will respond predictably to the incentives provided by the contract.¹⁵⁷

However, outside directors—who often own only a trivial fraction of their firm’s common stock—are in no sense perfect agents for the shareholders who elected them. Board members are “reluctant to terminate or financially punish poor-performing CEOs because [board members] personally bear a disproportionately large share of the non-pecuniary costs [of such terminations], but receive essentially none of the pecuniary benefits” (Baker, Jensen, and Murphy, 1988, p. 614). Similarly, board members are willing to over-compensate adequately performing CEOs, since they are paying with shareholder money and not their own. As documented by Fracassi and Tate (2012), even “outside” board members often share important social ties with incumbent CEOs, especially in cases with powerful CEOs who presumably influence the director-nomination process. This agency problem between shareholders and their elected representatives forms the basis of the “managerial power hypothesis”, in which powerful CEOs are able to influence both the level and composition of their own compensation packages. However, as discussed in Section 5.2.1 below, the agency problem is perhaps even more apparent in situations not involving powerful incumbents, such as directors overpaying CEOs hired from the outside.

5.1.2 Capital Markets

As discussed in Section 3.6.1, the executive compensation practices of the 1970s provided few incentives for executives to pursue value-increasing reductions in excess capacity and disgorgements of excess cash. However, pressures to improve performance, disgorge cash, and create wealth were ultimately introduced by the capital markets. The takeovers in the 1980s—often financed with newly available high-yield debt—provided credible competition for poorly performing incumbent managers. Wealth was created by both the post-merger activities of the acquiring firms (such as firing incompetent incumbent managers) and by responses to the takeover threat (such as excess spending cash to repurchase shares). Debt created value by providing commitments that the firm would pay its cash flows to debtholders, reducing the amounts available for executives to waste.

Capital markets—in particular, shareholder activists and large-block institutional stockholders—have mitigated agency problems by pressuring companies to strengthen links between CEO wealth and company stock-price performance. Fernandes et al. (2013), for example, show that the fraction of CEO pay delivered in the form of stock or options increases with institutional ownership. Hartzell and Starks (2003) show that CEO pay-performance sensitivities increase with the concentration of institutional ownership. In an international study, Aggarwal et al. (2011) find that the performance-related CEO turnover also increases with institutional ownership.

Capital markets have also, however, contributed to agency problems by providing executives with incentives to take actions to meet or beat analyst and market earnings expectations. As discussed in Section 2.4 and shown in Figure 12, executives have incentives to beat analysts forecasts by a small amount but not by too much because the abnormal stock-price response from beating the forecast by a lot is not much higher than the response for beating it by a little. Moreover, if an executive is going to miss the forecast, the executive may as well miss it by a lot since the negative abnormal stock-price response for a large miss is not much higher than for a small miss.

More generally, as argued by Jensen and Murphy (2012) and Martin (2011), capital-market pressures teach executives to focus on the “expectations market” (in which investors bet on expectations of future performance) rather than the “real market” (in which goods and services are produced and sold, and value is created or destroyed). Focusing on the expectation market is problematic because executives inherently have access to information about future prospects that are not publicly known and incorporated into stock prices. Executives with such a focus will be tempted to take actions that increase short-run stock prices at the expense of long-run value.

Temptations to manipulate the expectations market will clearly be higher for executives holding large quantities of stock and options that can be sold or exercised before markets adjust to the “real” information. As discussed in Section 2.4, there is substantial evidence that executive option and equity holdings are indeed higher in companies that restate their earnings or are accused of accounting fraud.¹⁵⁸ There is less evidence, however, that executives actually exercise and sell large fractions of their exercisable options or sell large fractions of their unrestricted stock holdings prior to restatements or indictments. The ominous hypothesis is that executives focused on the expectations market are not following a “pump and dump” strategy (which can be controlled by imposing longer holding requirements for shares), but rather that they are legitimately confused about the difference between increases in the short-run stock price and true value creation.

5.1.3 The Political, Legal, and Regulatory System

Agency costs are mitigated by laws prohibiting embezzlement, corporate theft, and fraudulent conveyance, as well as securities rules, regulations, and listing requirements designed to protect shareholders and other financial claimants. For example, the Securities Act of 1933—which regulated new securities issues—sought to protect shareholders by mandating full disclosure of all information that a “reasonable shareholder” would require in order to make up his or her mind about the potential investment. The Securities Act of 1934—which regulated secondary trading of securities—introduced in Section 16(b) the “short swing” profit rule (discussed above in Section 3.5.4) requiring executives to return any profits realized from buying and selling (or selling and buying) shares of their company’s stock within any period of less than six months. More sweeping (at least in its interpretation) was the anti-fraud provision Section 10(b) (and the corresponding SEC 10b-5 rule), which restricts insider trading, earnings manipulation, and price fixing. More recently, Regulation FD (August 2000) requires publicly traded companies to disclose material information to all investors at the same time (rather than favoring certain investors). While there are substantive arguments for allowing trading on material nonpublic information (since new information is more quickly introduced into the market), insider-trading rules are generally believed to benefit shareholders by reducing self-dealing by unscrupulous executives.

In addition to the general Securities Acts, the government has directly regulated the composition of the board of directors. Since 1994, companies have been required to have compensation committees consisting solely of independent directors in order for any pay to be exempt from the $1 million deductibility cap. In 1999, full independence of the auditing committee was required for all NYSE-listed firms; this requirement was extended to all firms in the 2002 Sarbanes-Oxley Act. In 2003, NYSE and NASDAQ listing requirements tightened the definition of independence and mandated that boards of listed firms have a majority of outside directors; the NYSE further required full independence for the compensation and nominating committees.

Critics hoping that independence requirements would reduce levels of executive pay have been disappointed. Both the level of pay and the use of equity-based compensation increase with the fraction of outsiders on the board; Fernandes et al. (2013) show that pay levels increase with board independence even after controlling for the risk associated with higher incentives. The evidence is therefore consistent with the hypothesis that directors—paying with shareholder money and not their own—prefer better-aligned incentives but are not particularly interested in restraining pay levels. The evidence is also consistent with directors not fully understanding (or believing) the opportunity cost of equity-based compensation (see Section 5.2.3 below).

Moreover, evidence that board independence “improves” pay is elusive. Bizjak and Anderson (2003) analyze the level and structure of compensation for CEOs who sit on their companies’ compensation committees (a relatively common occurrence before the early 1990s). Most critics of CEO pay (including Bebchuk–Fried and many shareholder activists) are horrified by the idea that the CEO could be a member of his own compensation committee, and would predict that such CEOs would inflate their own pay with few constraints.¹⁵⁹ And yet, Bizjak and Anderson (2003) find that the CEOs sitting on their own compensation committees earn substantially less (and not more) than other CEOs, have significant shareholdings and are typically company founders or their family members. These CEOs sit on their compensation committees not to inflate their own salaries, but rather to influence the level and structure of pay for their subordinates. Prohibiting such CEOs from sitting on (or chairing) their compensation committees harms shareholders, and illustrates a cost of the “one-size-fits-all” nature of corporate governance regulation.

In addition to general securities laws and independence requirements, this study has chronicled the history of government intervention into executive compensation. Over the past 80 years, Congress has imposed tax policies, accounting rules, disclosure requirements, direct legislation, and other rules designed explicitly to address perceived abuses in executive compensation. With few exceptions, the regulations have been either ineffective or counterproductive, typically increasing (rather than reducing) agency problems and pay levels, and leading to a host of unintended consequences. For example, the 1984 laws introduced to reduce golden parachute payments led to a proliferation of change-in-control arrangements, employment contracts, and tax gross-ups. Similarly, a variety of rules implemented in the early 1990s are largely responsible for fueling the subsequent option explosion, and the enhanced disclosure of perquisites in the 1970s is generally credited with fueling an escalation in the breadth of benefits offered to executives.

The emerging conclusion is that the myriad attempts to regulate CEO pay have been mostly unblemished by success. Part of the problem is that regulation—even when well intended—inherently focuses on relatively narrow aspects of compensation allowing plenty of scope for costly circumvention. An apt analogy is the Dutch boy using his fingers to plug holes in a dike, only to see new leaks emerge. The only certainty with pay regulation is that new leaks will emerge in unsuspected places, and that the consequences will be both unintended and costly.

Another part of the problem—as suggested above in the context of CEOs sitting on their firm’s compensation committees—is that government regulation inevitably imposes a “one-size-fits-all” solution to a perceived problem. For example, as I emphasize in Murphy (2012), claims (unfounded or not) that the banking bonus culture created incentives to take excessive risks were relevant at most for a relatively small number of large publicly traded Wall Street security brokers and dealers (along with some large commercial banks with significant investment banking operations). And yet, the Dodd–Frank provisions designed to reduce such incentives in the future were imposed on all public and private financial institutions, including broker-dealers, commercial banks, investment banks, credit unions, savings associations, domestic branches of foreign banks, and investment advisors.

A larger part of the problem is that the regulation is often mis-intended. The regulations are inherently political and driven by political agendas, and politicians seldom embrace “creating shareholder value” as their governing objective. While the pay controversies fueling calls for regulation have touched on legitimate issues concerning executive compensation, the most vocal critics of CEO pay (such as members of labor unions, disgruntled workers and politicians) have been uninvited guests to the table who have had no real stake in the companies being managed and no real interest in creating wealth for company shareholders. Indeed, a substantial force motivating such uninvited critics is one of the least attractive aspects of human beings: jealousy and envy. Although these aspects are seldom part of the explicit discussion and debate surrounding pay, they are important and impact how and why governments intervene into pay decisions.

5.1.4 Product Markets

While competition in the product market can theoretically either reduce or increase agency problems (see Hart, 1983; Scharfstein, 1988 respectively), companies that cannot compete in the product market cannot survive. The product market, therefore, provides inevitable discipline for value-destroying managers, but only after most of the value has been destroyed. Moreover, relying on product markets to discipline managers encourages managers to view “survival” rather than value-creation as their governing objective.

5.2.1 Managerial Power

The “managerial power” approach begins with the self-interested executives envisioned by Berle and Means (1932) and Jensen and Meckling (1976) and adds a new element: the ability of these executives to influence both the level and composition of their own compensation packages, often (if not invariably) at the expense of shareholders. One of the early contributors to this view is David Yermack, who has argued that CEOs extract rents from shareholders by timing their option grants to occur just before the release of good news (Yermack, 1997), by insider trading through their family charitable foundations (Yermack, 2009), through lucrative severance and change in control provisions (Hartzell, Ofek, and Yermack, 2004; Yermack, 2006b; Dahiya and Yermack, 2008), and by consuming excessive perquisites (Yermack, 2006a).

The researchers most closely associated with the managerial power approach are Lucian Bebchuk and Jesse Fried, who have argued in a series of papers that both the level and composition of pay are determined not by competitive market forces but rather by captive board members catering to rent-seeking entrenched CEOs.¹⁶⁰ In addition, the authors argue that the CEO’s ability to extract rent is limited by outside scrutiny and criticism (the “outrage constraint”), and CEOs respond by extracting rents through difficult-to-observe or assess forms of compensation rather than through increased base salaries. They use their model to explain several common features of executive compensation plans, including the use (and misuse) of compensation consultants, the prevalence of stealth compensation (pensions, deferred pay, perquisites, and loans), gratuitous severance payments, and stock options that are uniformly granted at the money and not indexed for the market or industry.

Can managerial power explain the trends in CEO pay? There is no doubt that executives (like the rest of us) are self-interested and would prefer higher compensation to lower compensation. There is also little doubt that—while CEOs are never explicitly involved in setting their own pay (even those sitting on their own compensation committees)—CEOs have subtle ways of influencing the compensation committee and the pay-setting process.¹⁶¹ However, as emphasized by Holmstrom and Kaplan (2003) and Frydman and Jenter (2010), there is no evidence that boards have become weaker or more captive over time. Indeed, every measure of board independence has improved since the mid-1980s. As discussed in Sections 5.1.1, the fraction of outside directors serving on the average board had increased to 80% by the end of the 1990s, and the CEO was the sole insider in nearly half of all firms. Since IRS Section 162(m) in 1993 (which required independence as a prerequisite for deductibility), most compensation committees have been fully independent. The 2003 NYSE listing requirements and 2010 Dodd–Frank Section 952 are appropriately characterized as tightening the definition from “independent” to “really independent” to “really, really independent”, reflecting a mistaken belief that true independence can be measured by an objective standard applicable across all publicly traded companies without regard to the individual director. The increase in board independence during the 1990s should reduce managerial influence over pay, suggesting that the trends in CEO pay over the period were not driven by managerial power. In addition, the secular increase in disciplinarily firings of poorly performing CEOs (Kaplan and Minton, 2011; Huson, Parrino, and Starks, 2001) offers no evidence that boards are becoming systematically more passive over time.

Moreover, it is worth noting that many of the most generous and widely criticized option and severance payouts over the past two decades have been the direct result of formal employment agreements negotiated with external candidates, and not deals reached with powerful incumbents. Indeed, Murphy and Zábojník (2008) attribute the increase in executive pay to the increased prevalence of hiring CEOs from outside the firm. During the 1970s, under 15% of newly appointed CEOs were hired externally. By, the late 1990s, nearly a third of all CEO appointments came from outside of the firm, suggesting increasing competition in the managerial labor market. While the Murphy–Zábojník results (discussed in the next section) are often cited as evidence for the “efficient-contracting” approach, they are also consistent with directors systematically overpaying (and over-protecting) CEOs brought in from outside the firm.

In fact, compensation committees almost invariably pay “too much” for newly appointed CEOs, especially for those hired from outside the firm. Corporate directors seeking new CEOs from outside typically hire a professional search firm to identify qualified candidates for the position (Khurana, 2002a, 2002b). The pool of qualified candidates is narrowed through extensive research, background and reference checks, and interviews until a single individual is selected for the position. Negotiations over pay typically begin only after the favored candidate is identified and told that he or she is to be the new CEO. Indeed, many times negotiations are still on-going when the appointment is announced publicly. At this point the board is effectively locked in to the particular candidate CEO, which dramatically shifts the bargaining power to the seller (the candidate) rather than the buyer (the firm). This procedure is a reasonable way to identify top candidates when “price” is not an issue, but is clearly a recipe for systematically paying too much for managerial talent.

The tendency to pay too much and to pay it in the wrong way is exacerbated by potential CEOs who hire skilled contracting agents to negotiate on their behalf. In contrast, compensation committees rarely retain their own expert negotiators. The outcome is what one would expect in a game where there is such a clear mismatch: no matter how well intentioned, the typical compensation committee is no match against a professional negotiator, and overly generous pay packages become ubiquitous. But often the problem is worse: the incoming CEO (and his professional agent) negotiate not with the compensation committee but rather with the company’s general counsel or head of human resources, knowing they will report to the CEO when the contracting is complete.

Overpaying newly hired CEOs is an agency problem caused by directors paying the new hires with shareholder money rather than their own. It is not, however, a “managerial power” problem, since the board is not captive and these are arms’ length negotiations with a non-incumbent CEO candidate. The distinction is important because the policy prescriptions are different: the solution to overpaying new hires is to strengthen the negotiation process, while the solution to managerial power is to weaken the incumbent CEO’s influence. More importantly, the “problem” of overpaying (and over-protecting) new hires may be small compared to the costs of selecting the wrong CEO.

In any case, hiring managerial talent from either inside or outside the firm is expensive, and the price of talent increased significantly during the 1990s and early 2000s. Kaplan and Rauh (2010) and Kaplan (2008), for example, present evidence that the increased pay for top executives is comparable to pay trends for top lawyers, investment bankers, hedge-fund managers, venture capitalists, private-equity managers, and athletes. The rise in incomes for top talent in these disparate sectors—most with active and mobile labor markets—cannot plausibly be explained by managerial power. It seems unproductive to attribute gains in these other sectors to competitive market forces while inventing a different explanation for the rise in CEO pay. Indeed, the secular increase in external CEO appointments documented by Murphy and Zábojník (2008) suggests that the managerial labor market is becoming more rather than less competitive.

Can managerial power explain the growth in the use of stock options?Bebchuk, Fried, and Walker (2002) suggest that firms can “camouflage” excessive pay by substituting stock options for cash compensation, under the theory that such grants are difficult to value and are easy to hide in annual disclosures. Under disclosure rules effective before 1992, information on option grants was indeed difficult to obtain.¹⁶² However, the centerpiece of the sweeping new disclosure rules introduced in October 1992 focused on option grants, and two new tables were added to the proxy statements to describe the details of both the grant and the number and value of options held at the end of the year. Bebchuk, Fried, and Walker (2002) would predict that options grants would fall as the amount of information increased. However, option grants escalated (rather than fell) following the new rules.

Bebchuk and Grinstein (2005) attempt to provide a managerial power explanation for the 1990s increase in stock options as follows. First, they argue that the stock market boom weakened the outrage constraint, giving executives more latitude to increase their own pay. Second, they argue that increasing compensation in the form of options caused less outrage than increasing base salaries, not because of “camouflage” but because options offered the possibility of improved incentives. When the market declined in 2000–2002, the outrage constraint strengthened as investors became less forgiving of perceived managerial over-reaching, stemming the escalation in both pay and the use of stock options. Bebchuk and Grinstein (2005) use this framework to explain the correlation between CEO pay and general stock-price movements, as illustrated in Figure 18 in Section 3.7.5. Their framework would therefore also predict an increase in pay and options during the 2003–2007 bull market, and yet pay increases were modest and options were declining over this period. They could, of course, provide arguments for the existence of an “outrage constraint” for this period that would explain why pay levels moderated and options were replaced by restricted stock. This points to a basic problem with the Bebchuk and Grinstein (2005) explanation (and the managerial-power hypothesis more generally): there is no principled way to refute any trend in pay given the authors’ flexible (and unmeasurable) definition of both the “outrage constraint” and its importance.

5.2.2 Efficient Contracting

The “efficient contracting” camp maintains that the observed level and composition of compensation reflects a competitive equilibrium in the market for managerial talent, and that incentives are structured to optimize firm value. The survey article by Edmans and Gabaix (2009) considers optimal-contracting explanations for the pay practices criticized under the managerial power camp, and the survey article by Frydman and Jenter (2010) discuss how these theories can predict increases in CEO pay over time.

Unlike the “managerial power” camp, the “efficient contracting” camp is not neatly characterized by a well-defined set of authors or articles. The modern executive compensation literature paralleled the emerging agency theory literature, and the majority of CEO pay papers written since the 1980s have been explicitly or implicitly based on agency or optimal-contracting theories. Indeed, the managerial power approach largely evolved as researchers—perhaps beginning with Jensen and Murphy (1990b) and Yermack (1995)—uncovered anomalies seemingly inconsistent with optimal contracts.

Can efficient contracting explain the trends in CEO pay? Beyond optimal incentive contracts, the efficient contracting approach includes market equilibrium models of managerial productivity, matching, and sorting that predict secular increases in CEO pay. For example, Murphy and Zábojník (2008) and Frydman (2007) offer general equilibrium models attributing the increase in executive pay to the increased prevalence of hiring CEOs from outside the firm. In particular, both papers attribute the trend toward outside hiring as reflecting gradual changes in the nature of the CEO job, modeled as a shift in the relative importance of general “managerial capital” (human capital specific to CEO positions) over firm-specific capital (reflecting skills, knowledge, contacts, and experience valuable only within the organization). The shift in the relative importance of general vs. firm-specific managerial capital leads to fewer promotions, more external hires, and an increase in equilibrium average wages for CEOs relative to the wages of lower-level workers. Ultimately, while it is plausible that the increased prevalence of outside hiring will increase average wages (if nothing else, employers must always pay a premium when hiring from outside compared to promoting from within), it is less plausible that the doubling of outside hiring from the 1970s to the 2000s could lead to such a huge increase in real CEO pay over this time period.

Alternatively, Gabaix and Landier (2008) build an equilibrium model in which the marginal product of managerial ability increases with firm size (so that it is optimal to assign the most talented managers to the largest firms). As shown by Rosen (1981) and Rosen (1982), such assortative matching produces equilibrium wages that are convex in ability, such that small increases in ability can lead to large increases in wages (since the CEO is assigned to a larger firm). Gabaix and Landier (2008)’s key insight is that the wage of a CEO will depend not only on firm size, but also on the size distribution of all firms in the relevant market: as the average firm becomes larger, managerial marginal products increase and competition for scarce managerial talent will bid up compensation. In particular, they show that a shift in the size distribution of firms will lead to a proportional shift in compensation, and conclude that “the six-fold increase in CEO pay between 1980 and 2003 can be fully attributed to the six-fold increase in market capitalization of large US companies”.

Gabaix and Landier (2008)’s results are consistent with the near-perfect correlation between CEO pay and general stock-price movements observed from 1980 to 2002 (see Figure 18 in Section 3.7.5). However (and similar to the critique of Bebchuk and Grinstein (2005) above), their results are not consistent with time trends in CEO pay and the stock market since 2002. In addition, while their insights on the size distribution are potentially important, their focus on market capitalization as the size measure is problematic since it conflates size, stock-price performance, and the vagaries of the market. Few would argue, for example, that Apple was really the largest firm in the world economy in 2012 (and yet their market value in early 2012 eclipsed that of Exxon-Mobil, PetroChina, and Royal Dutch Shell). Similarly, Volkswagen was not the second-largest firm on the planet for a couple of days in late October 2008 after its stock price increased by 350% over a two-day period (before tumbling by 60% over the following week).¹⁶³ While average CEO pay may have moved roughly proportionately with average market capitalization between 1980 to 2003, it far outpaced the growth in more traditional measures of size. For example, average revenues for the 500 largest US firms ranked by revenue grew only by 50% after inflation from 1980 to 2003, while average employment for the 500 largest US employers grew only by 19%.¹⁶⁴

Can efficient contracting explain the growth in the use of stock options? The CEOs in most market-equilibrium models (including Murphy and Zábojník (2008), Frydman (2007), Gabaix and Landier (2008), and the informal model in Kaplan and Rauh (2010)) contribute only ability and not effort. Therefore, there is no role for incentives and thus no obvious reason why the increase in pay would come in the form of increased equity-based compensation (or, in particular, in stock options and why the preferred form of equity incentives would shift to restricted stock after 2002). To “explain” trends in CEO pay, it is not enough to predict increases in the level of pay, independent of dramatic changes in its composition. Indeed, as discussed above in Section 2.1.2, CEOs naturally demand a “risk premium” for accepting stock options in lieu of safer forms of compensation, and this risk premium will increase when the CEO is less diversified (i.e. when holding more shares of stock, or when the value of option portfolios increase relative to other wealth). Therefore, any increase in stock options will naturally be associated with an increase in total compensation, especially in a rising market. As shown in Figure 7 in Section 2.1.2, median risk-adjusted CEO pay actually fell from 1998 to 2001 (at least given the assumptions in the figure), even as the median unadjusted pay was exploding. In fact, the puzzle to be solved in Figure 7 is not why pay levels increased in the late 1990s (because they actually declined after adjusting for risk), but rather why risk-adjusted pay levels increased dramatically from 2002–2007, as companies replaced risky stock options with less risky restricted stock, without substantial declines in the grant-date fair-market value of pay.

Optimal-contracting theory (i.e. the subset of efficient contracting predicting that incentives are structured to optimize firm value) offers few explanations for the increase in option-based pay (i.e. increases in pay-performance sensitivities) in the 1990s. Consider, for example, the benchmark model where firm value is given by y = a + ɛ, where a is executive effort, and ɛ is (normally distributed) uncontrollable noise, ɛ≈N(0,σ²). Moreover, suppose that managerial contracts take the simple linear form w(x) = s + by, where s is a fixed salary and b is the sharing rate (or “pay-performance sensitivity”). Assuming that the executive has exponential utility, U(x) = −e^r(w-c(a)), where r is the executive’s absolute risk aversion and c(a) is the convex disutility of effort, the optimal sharing rate is given by:¹⁶⁵

Traditional contracting theory therefore suggests a finite number of factors that might explain higher incentives among CEOs in the 1990s. First, perhaps CEOs became less risk or effort averse in the 1990s, but to my knowledge there is no theory or empirical work suggesting such declines in risk- or effort-aversion parameters. Second, perhaps CEO performance became estimated with less noise in the 1990s. While potentially consistent with the increase in director independence (if taken as a proxy for board monitoring), most measures of cash-flow or shareholder-return volatility increased rather than decreased over this time period.

Alternatively, suppose that firm value is given by y = θa + ɛ, where the primary source of uncertainty is variations in θ (i.e. managerial productivity assumed to be observed by the CEO but not by directors or shareholders). Zábojník (1996) and Prendergast (2002) show that optimal pay-performance sensitivites increase with the volatility of θ (incentives are more important when the CEO has private information about his or her marginal productivity). Again, to my knowledge there is no theory or empirical work suggesting that CEO marginal productivity became more volatile during the early 1990s.

Moreover, optimal-contract theories must explain not only the increase in equity-based compensation, but why that increase came almost entirely in the form of stock options as opposed to restricted shares or other equity-based instruments. Several papers have attempted, with only limited success, to provide theoretical justification for stock options. For example, traditional principle-agent models based on constant relative risk aversion and lognormally distributed stock prices (e.g., Hall and Murphy, 2002; Dittmann and Maug, 2007), suggest that—when salaries can be adjusted—contracts with restricted shares or options granted in-the-money are generally are superior to contracts with at-the-money options.¹⁶⁶

Ultimately, the most compelling optimal-contracting explanation for the increase in equity-based compensation in the 1990s is that contracts were suboptimal before the 1990s, and got better. As explored above in Sections 3.5.6 and 3.6, year-to-year changes in executive pay in the 1970s largely reflected changes in company revenues (rather than performance), contributing to unproductive diversification, expansion and investment programs. The takeover and LBO market of the 1980s demonstrated vast potential for value creation in previously inefficient firms, leading academics, institutions, and shareholder advocates to demand that pay be more closely tied to shareholder performance. As emphasized by Holmstrom and Kaplan (2001), stock options allowed executives to share in the value created by internal restructurings that reduced excess capacity or reversed ill-advised diversification programs. The growing importance of shareholder activists and large institutional investors (Gompers and Metrick, 2001) increasingly pressured firms to tie pay more closely to stock-price performance. Stock options also became the currency of choice for high-tech start-ups, rich with ideas but (allegedly) short of cash or sources of capital. As a result, the popularity of options soared with the stock market during the 1990s, to the benefit of shareholders and executives alike. In fact, part of the increase in options during the 1990s plausibly reflects the fact that they seemed to be working: corporate boards and top managers began to associate option grants with successful company performance, especially during the high-tech and Internet boom of the late 1990s. Indeed, the increase in options coupled with the renewed focus on shareholder value creation may help explain the overall growth in stock market during this period.

This optimal-contracting explanation for stock options cannot, however, explain the magnitude of the explosion, and why it came in the form of options rather than stock. Consider, for example:

Even more difficult for the optimal-contracting camp is explaining why so many options were granted to so many employees well below the executive suite (see Figures 19 and 20). Since non-tradable stock options are an unusually inefficient method of conveying compensation (see Section 2.1.2), the incentive benefit from stock options must exceed the substantial difference between the company’s opportunity cost of granting options and the “value” of those options from the perspective of risk-averse undiversified employees. While there may be efficient contracting justifications for granting options to top-level executives and other critical employees who can directly impact company stock prices (such as R&D scientists), there is (to my knowledge) no compelling incentive theory explaining option grants for rank-and-file employees.

Existing theories of broad-based option plans focus not on incentives but on other aspects of the employment relation. Oyer (2004), for example, argues that broad-based options may help satisfy participation constraints when reservation wages are correlated with the “market” and when it is costly to adjust other terms of employee compensation. Oyer and Schaefer (2005) and Bergman and Jenter (2007) argue that it might be optimal to grant options rather than cash when employees are irrationally optimistic about company prospects. Core and Guay (2001) argue that firms grant options to lower-level employees as a substitute for cash compensation, and document a greater use of options for firms facing financing constraints.¹⁶⁷ Babenko, Lemmon, and Tserlukevich (2011) argue that financially constrained firms rely on cash inflows from employee option exercises to finance investments.

The common failing in the aforementioned theories of broad-based stock option plans is neither recognizing nor incorporating the substantial difference between the company’s cost and the employee’s value of non-tradable stock options. For example, Oyer (2004) offers no compelling argument or evidence that options are an efficient substitute for flexible employment terms (indeed, he largely ignores the efficiency cost of options, and assumes that contract adjustments are prohibitively costly)¹⁶⁸, and Core et al. (2001) and Babenko et al. (2011) implicitly hold but provide no theoretical or conceptual evidence for the implausible assumption that risk-averse undiversified employees are efficient sources of capital. Bergman and Jenter (2007) suggest that firms can reduce compensation costs by paying over-optimistic employees with (potentially overvalued) options, but provide no evidence that options are offered as a substitute for other forms of compensation. Indeed, all these models ignore the fact that most broad-based option plans were layered on top of existing compensation arrangements, and were not substitutes for cash compensation. The dominant option granters in the 1990s were not small cash-poor internet start-ups (where a compelling incentive-based rationale for broad-based options can be made), but rather large cash-rich giants such as Microsoft, Intel, Cisco, and Apple.

5.2.3 Perceived Cost

In a series of papers—admittedly garnering less traction than either the “managerial power” and “efficient contracting” approaches—I’ve suggested an alternative explanation for the growth of option-granting in the 1990s: decisions over options were made based on the “perceived cost” of options rather than on their economic (or “opportunity”) cost.¹⁶⁹ When a company grants an option to an employee, it bears an economic cost equal to what an outside investor would pay for the option. But, it bears no outlay of cash, and (prior to the 2006 changes in accounting rules) bears no accounting charge. Moreover, when the option is exercised, the company (usually) issues a new share to the executive, and receives both the exercise price and (for non-qualified stock options) a tax deduction for the spread between the stock price and the exercise price. These factors make the “perceived cost” of an option much lower than the economic cost.

From the perspective of many boards and top executives who perceive options to be nearly costless—or indeed deny that options have value when granted—the only way they can quantify the options they award is by the number of options granted. During the 1990s, the focus on the quantity rather than the cost of options was further solidified by the institutions that monitor option plans. For example (see Section 3.7.5), SEC disclosure rules in place between 1992 and 2006 required companies to report only the number of, rather than the value of, options granted in the “Summary Compensation Table”, the primary or most visible compensation table in the company’s annual proxy statement. Similarly (see Section 3.7.6), under the pre-2003 NYSE listing requirement companies must obtain shareholder approval for the total number of options available to be granted, but not for the cost of options to be granted.¹⁷⁰ In addition, advisory firms (such as Institutional Shareholder Services) often base their shareholder voting recommendations primarily on the option “overhang” (that is, the number of options granted plus options remaining to be granted as a percent of total shares outstanding), and not on the opportunity cost of the proposed plan. Therefore, boards and top executives often implicitly admitted that the number of options granted imposes a cost on the company, while at the same time denying that these options have any real dollar cost to the company.

In addition, boards and top executives understand that options, when exercised, dilute the shareholdings of current equity holders. The number of options granted is included in fully diluted shares outstanding and therefore increased grants will decrease fully diluted earnings per share. Thus the negative consequences associated with these reductions in earnings per share also vary with the number of options granted, and not with the dollar-cost of the grants, and are consistent with the observed excessive focus on the number of options awarded and outstanding and not their dollar cost to the firm.

The perceived-cost view of stock options explains why options were granted in such large quantities to large numbers of executives and employees and also explains why the grant-date opportunity cost of options rose dramatically and subsequently declined with the stock market from 1980 to 2003 as shown in Figure 18 in Section 3.7.5. If boards focused only on the number of options granted, and the number of options granted stayed constant or varied positively with stock market performance, then the cost of the annual option grants would rise and fall in proportion to the changes in stock prices.

The perceived-cost view also explains why the relation between executive pay and the S&P 500 Index shown in Figure 18 weakened beginning in 2003. As discussed in Section 3.8.4, while FAS 123R required firms to expense their options beginning in 2006, many firms began voluntarily expensing in early 2003. Expensing options brings the perceived cost of options more in line with their opportunity cost, and companies responded to the robust stock market from 2003 to 2007 by decreasing the number of options granted as stock prices increased (rather than increasing the quantity of options as happened from 1993 to 2001). Moreover, expensing brings the accounting treatment of options in line with the accounting treatment of restricted stock, explaining the shift from options towards restricted stock.

Finally, the perceived-cost view explains many prevalent features of stock options offered by the managerial-power camp as evidence for their position. For example, Bebchuk, Fried, and Walker (2002) cite the scarcity of “indexed options” (i.e., options where the exercise price adjusts over time to market- or industry-wide price indices) as evidence for the managerial power hypothesis. However, prior to the 2006 imposition of FAS 123R, indexed options were subject to an accounting charge while traditional options were not, increasing the relative perceived cost of indexed options. Similarly, Bebchuk, Fried, and Walker (2002) suggest that firms use uniform option terms (e.g., granting options “at the money”) because diverging from normal practice by granting in-the-money options would spark outrage. Under the perceived-cost view, companies grant at-the-money options to avoid the accounting expense associated with in-the-money options. Indeed, the unsavory practice of “backdating” (in which firms granted in-the-money options but retroactively set the exercise date so the options appeared to be granted at the money; see Section 3.8.2) allowed firms to convey a given level of compensation without an accounting charge using fewer options than would be required without backdating. While the apparently common practice subsequently became “criminalized”, many of the participants at the time viewed the practice as a minor accounting transgression that saved the shareholders a little dilution.

The perceived-cost view is readily acknowledged by practitioners and compensation consultants, but is usually denied or dismissed by financial economists because it implies systematic suboptimal decision-making by managers and a fixation on accounting numbers that defies economic logic. But executives often respond to accounting concerns in ways that seem irrational to economists. For example (as discussed in Section 3.7.4), the practice of repricing options following stock downturns virtually disappeared in December 1998 after an accounting expense was imposed on repriced options, illustrating how companies respond to accounting rules that have no affect on company cash flows. Similarly (as discussed in Section 3.8.4), firms accelerated the exercisability of existing options in advance of the implementation of FAS 123R in order to avoid an accounting charge for previously granted but unexercisable options; such acceleration hurts shareholders by reducing retention incentives and allowing executives to unwind their equity positions. As another example (only slightly beyond the executive compensation arena), companies systematically scaled back retiree healthcare benefits after FASB required companies to record a current accounting charge for anticipated future medical costs.¹⁷¹ The new accounting rule apparently increased the perceived cost of these benefits, putting them more in line with their actual economic cost, and as a result companies reduced benefit levels.

While the perceived-cost approach can explain why so many options were granted to so many people (because options were free, or at least cheap), it cannot explain why the explosion in grants started in earnest in the early 1990s: after all, the accounting and tax rules governing non-qualified stock options had been in place since 1972. In addition, before the May 1991 ruling that allowed stock acquired by exercising options to be immediately sold (see Sections 3.5.4 and 3.7.2), companies routinely granted Stock Appreciation Rights (typically subject to an accounting charge) rather than stock options (typically subject to no accounting charge), suggesting that the choice of equity-based incentives were not solely driven by accounting considerations.¹⁷²

More fundamentally, the problem with the perceived-cost approach is that stock options are, of course, neither free nor even cheap to grant. Indeed, non-tradable options are an unusually expensive way to convey compensation to risk-averse and undiversified employees (Hall and Murphy, 2002; Section 2.1.2 above). A tempting theory—consistent with the managerial power approach—is that executives fully understood the opportunity cost of options but duped gullible directors into believing they were free. However, this explanation is inconsistent with the fact that 95% of options were granted below the CEO level: it seems implausible that the CEO would support such as huge transfer of wealth from shareholders to employees for a modest increase in his or her own compensation.

More plausible is the idea that executives and directors simply misunderstood the nature of opportunity costs. There is ample evidence that executives routinely ignore the opportunity cost of equity capital, leading firms to excess capacity and inefficient levels of inventories, cash and working capital. Indeed, the “Economic Value Added” programs that became popular in the 1990s were explicitly designed to teach managers about the opportunity cost of capital. If executives have a hard time grasping the opportunity cost of equity, they will have an even harder time grasping the opportunity cost of a derivative on that equity, especially when told that the “cost” is not the accounting cost but rather is estimated using a seemingly arcane theoretical formula. It is worth recalling that—while the Black–Scholes methodology was twenty years old by the early 1990s and was increasingly being used in academic research on executive compensation—it had only recently gained limited traction among compensation consultants, and was not considered a useful tool in most corporate human resources departments.

5.2.4 Politics of Pay

A central theme in this study has been the futility of “explaining” CEO pay without explicit consideration of the causes and consequences of government intervention into executive compensation over the past century. The option explosion in the 1990s, which in turn caused the escalation in pay levels that spawned both the efficient contracting and managerial-power literatures, is a prime example of this futility. In Section 3.7 I discuss six factors that I believe contributed to the 1990s explosion in stock options (and hence the escalation in pay):

The first of these factors (“shareholder pressure for equity-based pay”) is consistent with the efficient contracting explanation (at least the version of the theory that contracts were suboptimal before the 1990s, and got better). However, the remaining factors reflect government intervention into the pay process, often as unintended consequences of attempts to curb perceived excesses in executive pay (and executive stock options in particular).

For example, the May 1991 SEC rules that allowed executives to sell shares immediately after exercising options was an unintended consequence of an attempt to curb excessive grants. As discussed in Section 3.7.2, corporate insiders are required to report stock purchases on SEC Form 4, but were not (before May 1991) required to report option grants. To provide more transparency for option grants, the SEC redefined the “stock purchase” as the date the option was granted rather than when it was exercised (thus triggering Form 4 disclosure of grants within 10 days of the end of the month when options were granted). As a result of this new definition, the six-month holding period required by the Securities Act started when the option was granted and not when it was exercised, allowing immediate sales upon exercise and greatly enhancing the appeal of options.

Similarly, Bill Clinton’s campaign promise to limit deductibility of executive pay covered all forms of pay, and was only later modified to exempt deductibility limitations for pay tied to productivity. After substantial debate, stock options with an exercise price at or exceeding the grant-date market price were defined as related to productivity, while options with a lower exercise price were (arbitrarily) defined as non-performance related. But, as discussed in Section 3.7.3, the intent of the Congressional sponsors of the ultimate legislation was to reduce “excessive compensation”, and not to promote the use of stock options.

However, the government faced an interesting political quandary: while it sought to curb perceived excesses in executive pay and options, it simultaneously sought to encourage firms to issue options to lower-level employees. For example, in its 1992 disclosure rules, the SEC required firms to report not only the number of options granted to each proxy-named executive, but also report that number as a percentage of options granted to all employees. The sole purpose of this requirement—similar to the Dodd–Frank requirement to report the ratio of CEO pay to the pay of the median employee—was to encourage (or “shame”) companies into spreading awards more equally across the organization.

The NYSE listing requirements—which required shareholder approval for executive option plans but not broad-based option plans—were also designed to encourage option grants to lower-level employees. As discussed in Section 3.7.6, until January 1998 it had generally assumed that “broad-based plans” excluded substantial grants to top executives, which limited their use. The “clarifications” in 1998 (revised in 1999) defined how companies could grant top-executive options without approval, so long as a sufficient percentage of either the eligible employees or options granted were below the top-executive level. As a consequence, grants to both executives and lower-level employees escalated.

Similarly, FASB’s 1995 compromise (which allowed companies to continue to grant options without an accounting expense, while recommending expensing fair market values) was driven primarily by concerns about expensing’s implications for lower-level grants (and not concerns with top executives). Countering Carl Levin’s (D-MI) Corporate Pay Responsibility Act requiring option expensing (Section 3.7.4), bilis were introduced in both the House and Senate against expensing. In May 1994, the US Senate passed (by a 88-9 vote) a non-binding “sense of Congress” resolution demanding FASB to drop its expensing proposal, claiming that expensing would affect the ability of companies to raise capital, create jobs, and attract the best employees.¹⁷³ The Senate was joined by the Clinton administration—in no means an advocate of high CEO pay—concerned that FASB’s proposal would hurt the competitiveness high-tech companies.¹⁷⁴

The political obsession for broad-based option programs continued into the early 2000s, even as the popularity of options waned due to stock-market declines and pressures towards voluntary expensing (Section 3.8.4). Advocates of broad-based plans in Congress, fearing that fair-market-value accounting for options would end of option grants to low-level employees, introduced several (ultimately shelved) bills to protect such programs, including:

These bilis, and several others, were shelved in committee and the factors that had encouraged broad-based options were reversed:

Ultimately and predictably, these changes curtailed the practice of broad-based option plans: firms that already had such plans granted fewer options, and virtually no firms without plans introduced one. Indeed, as evident from Figure 19 in Section 3.7.6 the average number of options granted by firms to all employees in the S&P 500 fell by half from 2001 to 2005 (from 2.6% of outstanding shares each year in 2001 to 1.3% in 2005).