Question

Suppose that the shareholders can hire a board of directors to monitor the CEO. The board of directors cannot perfectly monitor the effort level of the CEO, but hiring the board of directors increases the chance that they observe the true effort level of the CEO. The cost of hiring the board of directors to the shareholders is z. If hired, the board of directors will observe the effort level of the CEO with probability ½. Assume that the CEO can choose from two effort levels: high (e=1) and low (e=0). The cost of each unit of effort is c. Assume that the shareholders will pay a wage of w to the shareholder. However, if the board of directors observes low effort by the agent, then the shareholders will pay a wage of zero to the CEO. If the CEO chooses high effort the shareholders receive a payoff of Y and if the CEO chooses low effort the shareholders receive a payoff of zero. The shareholders first choose to hire the board of directors or not, then the CEO chooses the effort level.

a) Draw the game tree (Hint: only the shareholders and CEO should be in the game tree. The board of directors only influences the payoffs at the end of the tree).

b) If the shareholders hire the board of directors, then what condition must hold for the CEO to choose high effort? (Hint: The payoff for high must be greater than choosing low.)

c) If the shareholders do not hire the board of directors, then what condition must hold for the CEO to choose high effort? (Hint: The payoff for high must be greater than choosing low.)

d) Suppose that the CEO will choose low effort if the shareholders do not hire a board of directors, but will choose high effort if the shareholders do hire a board of directors. What condition must hold for the shareholders to hire a board of directors? (Hint: The payoff for hire given what the CEO will do must be greater than the payoff of not hiring given what the CEO will do).

Answer 1

This chapter is about top managers, executive compensation, boards, and “corporate governance”. Governance is how the company is managed at the highest, and most important, levels. Legally, a company’s board has “fiduciary” responsibilities to serve the interests of its many stakeholders. Its actions determine the company’s most important moves. The boards of companies are elected by shareholders, much as politicians are elected by voters in democratic countries (except there are typically many restrictions on how directors are elected).

Top management and governance are important for several reasons. The actions of managers and boards are clearly important for what businesses the company enters, the products it makes, where its operations are located, and how employees and customers are treated. If a company makes a terrible mistake and struggles, it is probably due to an error in judgment by top managers and poor oversight by the board.

Furthermore, the actions of the managers and board are often closely watched by employees as clues about how they are supposed to behave—or not. In 2003, for example, American Airlines executives were trying to negotiate large pay cuts among their employees, to cope with the recession in the hard-hit airline business after the September 2001World Trade Center attacks. It was revealed that at the same time, the CEO was awarding large pension and performance bonuses to many top managers. Airline employees were furious and the CEO ended up having to cancel the giveaway to the top people. [ADD DETAIL HERE]

Another reason top managers and boards are important is that data on what they do and how they are paid are widely available, due to regulations which require publicly-held firms to disclose many details of how top executives are paid. This means we have a rich source of data to test theories of compensation, and particularly to see how pay influences executive performance. These data provide the best empirical studies of basic principles of agency theory. Many of those results are described in this chapter.

?

I: Agency problems and executive behavior

?The central agency problem, discussed in chapter 2, is that a principal wants the agent do something—code-named “effort”—which the agent does not want to do, or does not want to do enough of. When the agents are top executives, the range of agency problems is larger than with lower-level workers because executives have so much more power to make decisions. Some of the sources of agency problems are similar to those involved in motivating other workers, but other problems are different for executives with more power.

1 Work hours and shirking. The most obvious kind of effort is just showing up at work and putting in time. Many companies measure this simple kind of effort by having workers punch a card in and out of a time clock when they arrive at work and leave. Even this kind of measurement can be “gamed”— for example, a worker might punch in his friend’s time card if the friend is supposed to come at a particular time and is late. Furthermore, “shirking” may happen when people are actually at work but not paying attention to customers or their work. (This is probably a bigger problem than ever before due to the popularity of the internet.)

2 Nepotism. “Nepotism” is the practice of hiring friends and relatives who are not ideally qualified for their jobs. (In fact, one could extend the definition to any hiring of unqualified workers who the hiring manager likes working with.) In a surprising number of firms, the sons or daughters of top executives (often founders who started a company) have high-level positions or go on to run the company when their parent steps down. It is not known whether the practice is more widespread in the private or public sectors. Well-governed firms and democratic states are presumably less likely to tolerate nepotism, but even then, it is not hard to find egregious examples. For example, when Alaska’s Senator Frank Murkowski was elected Governor in 2002, he had the power to appoint somebody to finish out his term to 2004. He chose his daughter, Lisa Murkowski, a lawyer and “relatively obscure two-term state legislator” (LATimes, 2004). But Murkowski the Governor paid a political price—his approval rating fell from 70% to 29%.

It is possible that carrying on the family name is economically efficient, because some special kind of trust is engendered by keeping corporate succession along family lines, or the children of top executives are able to learn the business better from a parent. [NATIONAL GEOGRAPHIC QUOTE FROM ECONOMIST] Family ties may also limit opportunism and other forces that affect the optimal boundaries of the firm (see chapter 3).

3 “Perquisites” (“perks”): Perks are special services given to executives which are valuable but aren’t counted as regular income (and typically, are not taxed as income is). Examples include extravagant offices and office furniture (or art), using corporate jets or flying first-class on airplanes, terrific tickets to sporting events, and country club memberships. In the corporate scandals of 2000 and later saw some dramatic examples of excessive perks. Tyco executive Dennis Kozlowski threw a $2.1 million birthday part for his wife which included a life-size ice statue modeled after Michaelangelo’s famous status David spouting vodka from his, uh, body (see SIDEBAR below). Many CEO’s in this era also were given multi-million dollar loans are low interest rates, with repayment of the loans cancelled (“forgiveable loans’) if the executives met certain conditions (like working for the company a certain number of years).

Perks are not necessarily agency costs. Usually they are defended as part of the total compensation package necessary to attract and retain fantastic executives. Since perks are often not taxed as regular income is, it could be that executives would rather be paid in the form of perks than in additional cash, to save on taxes. However, they are many well-managed companies in which top executives may a point of not consuming such perks (for example, flying coach class on airplanes rather than much more expensive first class).

Sometimes the perks are disguised as good business decisions. For example, in 1994 MasterCard moved its major operations out of Manhattan to nearby Greenwich, Connecticut. The move was forecasted to save $11-15 million per year, but actual savings were only $8-10 million. Relocation expenses and operating costs were about 20% higher than forecasted. Most importantly, 20% of the workforce quit rather than work in staid Greenwich rather than in bustling Manhattan. One reason the company moved was that the new CEO, Eugene Lockhardt, loved to play golf and said he wanted to be “an eight-iron shot from Greenwich”.

?The key point is that moving the headquarters *might have been* a good business decision. As with investment and acquisition that top managers seem to personally prefer, it is often difficult to tell whether a decision is good for the company or is just what the CEO wants to do (and isn’t good for the company). The ambiguity about whether decisions reflect agency problems, or are just good business, is probably one reason why these decisions get made in the first place.

4. “Empire-building”: Top executives have a lot of control over how major investments are made, including acquisitions of other companies and mergers. Executives may make investments or acquisitions which are bad for their company due to errors in judgment or a desire to just run a larger “empire”. (Top executive pay is also closely linked to the size of the firm, so enlarging their company by acquisition usually leads to an increase in an executive’s pay.) We’ll discuss this in much more detail later when we talk about mergers and restructuring.

5. Risk-taking: The risk-incentive tradeoff looms large at the executive level. Large publicly-owned corporations are typically owned by many, many shareholders. [example]. Typically, they would prefer to have the company take on large risks which are good bets (that is, have positive expected profits), since any single shareholder effectively holds a huge portfolio of such large bets. But top executives often appear to be averse to take such risks. One reason is that executives often own a lot of shares, and have a larger portion of their personal wealth tied up in their company’s shares. So a big project failure hits them harder than it hits the typical shareholder. Probably ore importantly, however, project failures create career risk for executives: A mistake may cost them a promotion or get them fired.

6. Short horizons: A big potential for agency costs is the mismatch between the time horizon of the shareholders and executives, especially when executives are near retirement. (Of course, companies may realize this and make executive pay more sensitive to short-term performance when executives are closer to retirement; see the discussion of “endgame” below). Whether executives turn down good long-term bets, which will take years to be profitable, also depends on the ability of the stock market to value these bets properly. If markets can forecast that bets will payoff, and are informationally efficient, then even current stock prices will reflect long-term prospects. So if executives are motivated to maximize the current share price, they will take good long-term bets if they have faith in the markets. So incentivizing executives to take the long view depends on how much stock they have, and how much faith they have in the markets.

II: Executive compensation

?Let’s start with some facts.

1: How much are top managers paid?

?What would you consider a lot of pay for a job that requires some training and experience? The median household income in the US is around $40,000 in US dollars. Teachers earn a little less and do an important job.

Compared to these figures, CEO pay is very high. Figure 1 [FROM HALL 2003] below shows the median (middle) CEO pay ikn the US from 1980 to 2001 (in dollars adjusted to 2001 dollars). The Figure separates out salary and bonus (light blue) and equity-based pay, which is the value of shares as well as stock options. The figure also shows the share of total pay which is “cash” (salary) and bonus, through 1991 (around 90%), and the share of total pay which is equity-based from 1992 on (rising from 32% to 66% of total pay). Median total pay in 2001 was around $7 million.

?Median pay tells us only about the middle of the distribution. Like many “positively-skewed” distributions, the highest-paid CEOs earn much more than the median. Forbes (2002) reported that out of the 500 CEOs of Fortune 500 companies (rated by market value), the top 20 CEO’s earned total compensation from $35-$706 million, with a median (among the top 20) of $88 M. Most of this compensation was from exercised share options. The bottom 20 CEOs in the Fortune 500, according to Forbes, earned $84,000 to $785,000. So keep in mind that the lowest-paid CEO’s, who still run large important companies, earned less than a million dollars a year.

2. Compared to what?

?While CEO pay is staggering to average folks, it is important in economics to compare pay to benchmarks. If you say “CEOs are paid too much” you have to ask—compared to what (or whom)? And if CEO’s are paid too much, exactly how much should they be paid?

Labor economics tells us that we should judge pay relative to the marginal revenue product (MRP) of a good CEO. Unfortunately, it is hard to judge MRP. Later we will talk about how sensitive pay is to performance—that is, if a CEO’s company has a very good year, or a stretch of good years, how much of the stock market value that they helped create do they earn?

?As we think about these numbers, keep in mind how difficult it is to figure out a CEO’s MRP. In sports or sales based on commission, we can often measure MRP very directly—by linking sports performance to team wins, and then to revenue, or linking sales to marginal revenue. For CEO’s it is much more difficult. A CEO may inherit a high-performing firm and earn more than they deserve. Similarly, a CEO may inherit a “sick” firm and earn less than they deserve.

?a. The recent past: Do CEO’s in early 2000’s earn a lot more than in previous years? Brian Hall (03, Figure 1 above) shows the huge increase in (inflation-adjusted) salary & bonus plus equity-based pay. There is no question that CEOs of American companies are earning a lot more than they did 20 years earlier.

?b. CEO earnings compared to regular workers: Figure 2 above (also Hall, 2002) plots three graphs from 1970-2002. The key lines are the ratio of CEO salary and bonus only to average annual earnings of workers (the flat blue line, rising from around 30 to 70 (scale on left side), and the ratio of average total CEO pay (including options and equity) to average annual earnings of workers (the red line, which rises from 30 to around 400). These two lines show that CEO earnings have gone up relative to average workers, but almost entirely because of the huge increase (as in Figure 1) in CEO earnings from options and equity. ?

?Figure 2 also plots the level of the Dow-Jones Industrial Average of stock prices (scale on right) in green. The Dow rises and falls in lock-step with the CEO total pay to worker pay ratio. This is a reminder that CEO pay basically just tracked the level of the Dow. Given our previous discussion, about why CEO pay should be benchmarked to industry averages, this is surprising. Nothing in agency theory says that CEO pay should simply rise and fall with stock prices—it should rise and fall with relative performance (to subtract out common business cycle shocks that most CEOs cannot be blamed for or credited with). But it does not.

c. CEO earnings in other countries. While CEO pay is increasingly dependent on stocks and options in other developed countries (typically Japan and Europe), the ratio of CEO pay to worker pay is still low in those countries. For example, the ratio is 17 in Japan and 24 in France/Germany [GET RECENT FIGURES] in the late 1990s. However, note that the portion of CEO pay which is “at-risk” (bonuses, options and stock) has risen in other countries from 1996-2001 and is slowly catching up to the US model (see Table 1 below). This is evidence that the American style of incentivizing CEOs with options and stock is catching on, albeit more slowly in other countries than in the US. Also, in other countries CEOs typically get more non-pecuniary perks, such as subsidized housing, memberships to private clubs, and so forth. However, at the sky-high dollar levels of American CEO salaries, it is hard to argue that these increased perks begin to equalize salaries, unless you belong to a golf club that charges $100,000 a round.

?Conyon and Murphy (2000) conduct a careful comparison of executive compensation in the United Kingdom (UK) and the US in 1997. They find that US executives earn about 50% more cash compensation and twice as much total compensation as UK counterparts, controlling for firm size, industry and other factors that are known to influence pay.

Why? First note that the percentage of US firms which have executives holding options rose steadily from 70% to almost 100% over the period 1980-97. The corresponding UK numbers start much lower, around 10% of firms in 1980, then shoots up to 95% in 1985 (actually surpassing the US percentage). However, in the popularity of options starts to *decline* slowly throughout the 1980s’ and 1990s, and ends up in 1997 at around 70%.

It turns out that the larger pay for US executives is mostly due to the number of options? US executives are issued, and details of precisely the options are created and priced. Most importantly, the US and UK are similar in many respects—mix of companies, tax treatment of options—which rule out possible explanations. Conyon and Murphy suggest that historical accidents and national culture might provide part of the explanation. In 1995, after some executives in privatized electric utilities earned a huge windfall from option exercise, a government report was issued (the Greenbury report) encouraging companies to replace options with long-term incentive plans (LTIPs— various programs designed to encourage direct stock purchase by executives). The government then restricted the amount of options that could be awarded to only £30,000. In contrast, in 1994 the US government restricted tax deductibility of non-performance related pay (cash) to $1,000,000, encouraging a shift from cash pay to options and LTIP’s.

Conyon and Murphy also note that:

The United States, as a society, has historically been more tolerant of income ?inequality [than the UK], especially if the inequality is driven by differences in ?effort, talent, or entrepreneurial risk taking. In this light, perhaps it is natural that ?the United States reacts to claims of excessive pay by increasing the link between ?pay and performance, thus exacerbating income inequality, while the United ?Kingdom reacts through wage compression and reducing the pay-performance ?link. (p. F667).

They also note that Americans tolerate, and even celebrate, much higher pay in other professions—law, medicine, investment banking, and especially sports and entertainment-- than in the UK. So maybe the issue is not the CEO pay differential at all, but simply whether a society thinks a small number of workers deserve enormous rewards. ?