“Stockholder and public disapproval of executive pay levels is growing and could lead to government intervention unless directors and managers confront the executive pay issue squarely and come to grips with it,” wrote David Kraus in a 1980 Harvard Business Review article. At that time, an average chief executive officer (CEO) earned 42 times more than an average worker, and the ratio had been fairly stable for decades. By 2005 the ratio increased to 411 (see Figure 1), and there was public outcry about greed and excess. For example, in 2006 AT&T CEO Edward E. Whitacre Jr. earned $60.7 million in total compensation, according to SEC filings—the equivalent of the yearly wages of 4,990 minimum-wage earners or 2,056 average workers. Of course, such calculations are often confounded by the rather confusing structure of executive pay, as with Apple’s Steve Jobs, whose compensation in 2006 was $1 if computed by the SEC method, whereas Forbes estimated it at $647 million.
Figure 1: Average CEO to Average Worker Pay Ratio
Source: Institute for Policy Studies / United for a Fair Economy
The top to average pay gap has been growing both within most countries and between them, within and between industries, between executives and workers, and among executives themselves. Still, the question remains: what has led to this widening gap since the early 1980s? While the most popular explanation may be the undue influence of greedy CEOs on corrupt boards, it cannot be the whole story. The trend, dubbed “star effect,” also surfaces in other areas, such as the earnings of recording artists and baseball players.
A model developed by Bård Harstad of the Kellogg School of Management explains recent trends in both executive pay and organizational design. Market size is a critical factor. According to Harstad, “the classical example of the star effect comes from the entertainment industry.” To illustrate, he uses the analogy of every village having its own best singer, forming a closed market. Even in a closed market, better performers attract more people, but their earnings are limited by what people from that village are willing to pay. But if CDs are introduced into this scenario, people from all over the land can listen to every singer and pick the best one; that performer will consequently earn much more. However, his closest second may well find himself worse off, as even the people from his own village might prefer to listen to the “star.”
Just as a star performer’s popularity determines his or her earnings, managers’ reputations are pivotal factors in the design of their job contracts and the amount they are compensated. But what determines a reputation’s value, and how is it created? As Henry Ford once said, “You cannot build a reputation on what you are going to do,” and, accordingly, companies judge potential new managers on what is known about their past performance. However, not all jobs are equal for a manager desiring to establish a stellar reputation—visibility of individual effort is of utmost importance, and it is determined by a company’s organizational structure.
In addition, as Bård Harstad writes: “While the organizational design determines the allocation of blame and fame within the firm, the value of a good reputation depends on the market structure.” By combining a model of the market with a model of organizational design, he shows how changes in the former drive changes in the latter.
In Harstad’s model, a manager’s talent determines the cost of production under her supervision: the better managers are able to produce more at a lower marginal cost. Moreover, when competition gets tougher and the market gets “thicker” (e.g., the number of potential customers increases), the most cost-effective firm is able to capture larger market share. In this situation, the best manager brings value over and above that of simple efficiency as she is able to not just reduce costs, but also to outcompete others. Companies are therefore willing to pay much larger salaries to superior managers.
As the competition for market share becomes tougher, managers’ salaries will tend to be even larger. This provides an additional incentive for other executives to gain good reputations as well. Particularly in the case of young managers with unknown talent, the desire to be noticed and to have a chance at gaining an excellent reputation dominates the aspiration for a high initial salary. They are therefore willing to accept a lower wage when offered an opportunity to work in a position with high exposure. That way, their successes will be noticed and they can expect to gain a good reputation. Consequently, firms whose organizational structure allows such exposure can afford to pay their managers less and still be attractive employers.
What type of organizational decisions could help here? One factor is whether management tasks are centralized in one person or distributed among many. Additionally, a firm can be transparent, with everyone’s performance visible to everybody, or non-transparent, in which only aggregate performance is discernible. Imagine, for example, a group of high school students making a Web site where they all write reviews of their favorite computer games. One of them, Jenny, writes particularly well. At first their audience is small, consisting mostly of their friends, and Jenny might not even bother to sign her reviews. But once the site attracts more visitors, Jenny would likely prefer to have her name known. The Web site could then make her participation even more visible by placing a link to “New Reviews by Jenny” on the home page, thus enhancing transparency.
In his model, Harstad analyzes decentralization in transparent and non-transparent firms by building a model in which there are two managers dealing with two different tasks. In non-transparent firms, the subtasks performed by different managers cannot be differentiated, and they both share the reputation from the success or failure of the entire firm. This makes it less likely for one of them to achieve a “superstar” reputation. As competition increases and the market grows thicker, an extremely good reputation becomes much more valuable. These firms then find it profitable to centralize control in one manager, as she is able to take credit for the success of the whole company and is therefore willing to accept a lower salary for such an opportunity. In transparent firms, where the performance of individual managers is observable, the contrary is true: namely, in thicker, more competitive markets, transparent firms will decentralize control, as they are able to offer opportunities to create potentially stellar reputations for several young managers at the same time.
A company’s divisional structure is also affected by these factors. In a multidivisional firm (“M-form”), responsibility is allocated by product, so that one manager is responsible for all tasks associated with one product. In a unitary firm (“U-form”), responsibility is allocated by function, so that one manager is responsible for one specific task for all products. A manager has a better chance of building an excellent reputation in M-form, as its design reveals more about individual contributions. If a particular product’s overall performance improves and profit increases, in an M-form the manager responsible for that good or service will appropriate all of the rewards, while in a U-form the multiple individuals involved makes it difficult to identify individual achievements. Thus, as before, if competition increases and the market thickens, firms will find it profitable to switch to M-form. “Outsourcing is a common way of doing such a transition: when a dedicated company is in charge of a specific task, their contribution is clearly visible,” says Harstad.
Returning to our Web site example, let’s suppose that it expands into two new game genres—flying and cooking simulators. For that, it has two new reviewers—Joe and Mary. It is possible to put, say, Joe in charge of describing controls and Mary in charge of evaluating graphics in each review. Alternatively, Mary could review flying and Joe could review cooking simulators. In the second option, the quality of each writer is more discernible, and the chances of discovering a new Jenny are high, making this option more attractive for Joe, Mary, and the Web site.
Harstad’s model also clarifies how the turnover of managers is affected by all of the previously mentioned dynamics. Firms have a trade-off between keeping a good, seasoned manager and hiring a potential young “superstar.” For the latter, the value of exposure is high, and he may accept a lower salary compared to the former. When the market becomes tougher, excellent reputations are pivotal. In this scenario, moderately good, experienced managers are replaced with younger ones whose talent is yet unknown. The possibility of discovering a “superstar” among the young and inexperienced thus overrides the “benefits of keeping a fairly good manager,” according to Harstad. The tougher the competition is in an industry, the larger the turnover of managers.
Finally, Harstad discusses these model predictions in light of empirical evidence. He notes that in recent decades “consumers’ transport costs and search costs have declined, and the variety of products has increased. The number of consumers and their budgets have increased, and the size of a typical market is larger than before.” In accordance with the model, this should increase the premium for excellent (relative to moderate) talent and the variance of pay between managers. This variance should be higher for older managers, who have had a chance to earn a reputation. In addition, the difference between salaries of older (endowed with reputation) and younger (willing to forego higher wages for a shot at “star” status) managers should increase. These trends are both present in actual data (see Figure 2) and cannot be easily explained by other theories. In addition, increased manager turnover, decentralization, outsourcing, and transition to M-form are all features of modern organizational dynamics.
Figure 2: Annual Compensation for Executives Older and Younger Than 50 Years of Age
Source: Standard & Poor’s ExecuComp. Annual averages for all executives, including value of options and restricted stock grants.
Moreover, the model sheds light on the puzzling positive correlation between firm size and wages. On one hand, larger firms are better able to exploit abilities of better managers and are therefore going to attract the best—and the most expensive—talent. On the other hand, individual contributions of workers in smaller firms are more visible, and they are therefore willing to work for lower salaries to get their talents recognized. Indeed, we know that the size-wage correlation has increased over time and is larger in the United States, which is a thicker, more competitive market than Europe.
By showing how reputation building creates a link between market structure and organizational design, Harstad’s model provides an explanation for some of the recent organizational changes. It also helps us understand increasing executive wages and their variation, but the model is ultimately “just a small step towards a deeper understanding of these issues.” Other potential applications of his theory include the study of inter-firm organizational dynamics, such as umbrella branding, outsourcing, and other practices, in relation to differences between companies and marketplace dynamics.
AFL-CIO (2007). Paywatch (visit site, accessed November 26, 2007).
Anderson, Sarah, John Cavanagh, Chuck Collins, and Eric Benjamin (2006). “Executive Excess 2006. Defense and Oil Executives Cash in on Conflict. 13th Annual CEO Compensation Survey.” United for a Fair Economy (download PDF, accessed January 4, 2008).
Kraus, David (1980). “Executive Pay: Ripe for Reform?” Harvard Business Review, 58(5): 36-48