CHAPTER 2

A DIFFERENT KIND OF MARKET

THE SUCCESSION PROCESS that resulted in Jamie Dimon’s hiring as the CEO of Bank One followed an increasingly familiar script. A company’s performance declines, and the board responds by forcing out the incumbent CEO, who is blamed for its troubles. An external search is initiated with an extraordinary emphasis on hiring a candidate with demonstrable “leadership” and “charismatic” qualities. Much less emphasis is placed on the company’s strategic situation and how appropriate the background of the candidate is in light of this. The entire search process is orchestrated to produce a corporate “savior,” to find a new CEO whom investors and the business media regard as a star. The standard profile of this savior is of an individual who has served as a CEO or president at a high-performing and well-regarded company. Only a few candidates make it through this filtering process. Using the elusive concept of leadership to distinguish among the candidates, directors come to focus their attention on a single individual. Next comes the effort of convincing this person to accept the position. To rationalize the organizational and monetary resources that go into recruiting the favored candidate, boards convince themselves that the person they have identified through the search process is, in fact, worth the effort and expense—that he (or, more rarely, she) really is a better candidate than any available inside the firm. Not only is the new executive now expected to be the solution to the company’s problems, but his very presence is interpreted by the financial markets and the business media as a vote of confidence by the directors that these problems can and will be quickly solved. Everyone’s expectations run high.

This storyline, however, often has a dubious ending from the point of view of those whose long-term interests directors are supposedly representing (i.e., shareholders). In the effort to convince the candidate to join the firm, directors yield control of the process to the demands of the recruit, who is in an unusually strong position to bargain with the board about subsequent power and compensation arrangements. If the new CEO is unable to deliver results relatively quickly, the wisdom of the selection is questioned. In some instances, boards find themselves trapped in an infinite loop of dashed expectations and CEO churn.

The extraordinary focus on the CEO as the source of a company’s problems, and the blind faith that directors show in the charismatic CEO’s powers to heal what ails the firm, introduce the first strain of irrationality into what—considering what companies believe to be at stake in the choice of a new CEO—one might expect to be a rational, carefully considered process. The widespread, firmly held belief in the overriding importance of the CEO is all the more noteworthy considering that there is no conclusive evidence linking leadership to organizational performance. In fact, most academic research that has sought to measure the impact of the CEO on firm performance confirms Warren Buffett’s observation that when good management is brought into a bad business, it is the reputation of the business that remains intact. Because so much of this research is so much at odds with the conventional wisdom, it is worth summarizing it briefly here.

Two schools of thought dominate the literature on CEOs and firm performance. The “leadership” school suggests that CEOs play a critical role in affecting firm performance. Because CEOs are at the top of the organizational hierarchy, leadership scholars argue, they can and do affect firm performance by means of the decisions that they make about the company’s mission, strategy, structure, and culture. Most of the support for this perspective comes from detailed comparative field studies and individual firm case studies that document the links between executive qualities, organizational decisions, and subsequent firm performance. While their methods have been persistently questioned by a small number of other researchers, the leadership scholars have convincingly made the case that CEOs can indeed affect firm performance in particular instances.¹

On the other side of the debate, “constraint” researchers (in fields such as, for example, population ecology, resource dependency, and new institutional theory) argue that CEOs are so constrained that they have little impact on company performance.² The most careful research on this topic strongly suggests that a variety of internal and external constraints inhibit CEOs’ ability to affect firm performance; these constraints include internal politics, previous investments in fixed assets and particular markets, organizational norms, and external forces such as competitive pressures and barriers to exit and entry.³ Glenn Carroll and Michael Hannan, who have been at the forefront of an effort to study the life-course of organizations (using samples consisting of thousands of firms from scores of industries), have found convincing evidence to support this perspective. By studying the rates of organizational foundings and dissolutions over several decades, and by keeping track of attempts at major organizational change, the authors conclude that CEOs and other top executives have no statistically significant impact on firm performance, even when they make major decisions about strategy or structure. They write:

The common, but distorted, view of organizational change as easy and beneficial likely arises from the unsound way many observers collect data on corporations. Frequently, analysts and popular management authors collect information on a set of organizations currently performing very well and then look at the evolution of their strategies and structures. This selective, retrospective view of firms often shows that successful firms went through one or several previous transformations that led to later good performance. It is tempting—and many analysts succumb—to infer from this information that, had other organizations attempted the same changes, they too would have experienced success. Unfortunately, this inference comes from considering data that are heavily biased toward the successful firms. The information available does not justify an inference of cause-and-effect with respect to the changes. Nor does it provide the basis of generalization to other firms. Such a sample cannot support dependable analyses of the consequences of change.⁴

Given the divergence between the conclusions reached by these two schools of thought on CEOs and firm performance, some scholars have recently suggested that both the leadership and the constraint researchers have been asking the wrong question. This third group of scholars contends that the appropriate question to be asking is not “Does leadership matter” but rather “When does leadership matter.” The emerging answer to this question is that the impact of a leader on a firm is highly case-sensitive, varying according to individual situations, industries, companies, and CEOs. And while the CEO does indeed affect firm performance in particular situations, even then the overall CEO effect is swamped by contextual factors such as industry and macroeconomic conditions.⁵

Yet despite the overall evidence now pointing to at best a contingent and relatively minor cause-and-effect relationship between CEOs and firm performance, the cultural context within which contemporary CEO succession is conducted gives little consideration to this contingent perspective and instead operates on the assumption that a more precise and definitive relationship exists than scholars have been able to demonstrate. What then accounts for the persistence of the popular belief? As several scholars have noted, strong social, cultural, and psychological forces lead people to believe in cause-and-effect relationships such as that between corporate leadership and corporate performance. In the United States, the cultural bias towards individualism largely discounts the influence of social, economic, and political forces in human affairs so that accounts of complicated events such as wars and economic cycles reduce the forces behind them to personifications (as when people attribute the performance of the economy to the actions of Alan Greenspan)⁶. This process of exaggerating the ability of individuals to influence immensely complex events is strongly abetted by the media, which fixate the public’s attention on the personal characteristics of leaders at the expense of serious analysis of events. More generally, in their study of what they call performance-cue effects, social psychologists have found that the construction of leader images is a process of matching leaders’ characteristics with performance outcomes: outcomes, whether positive or negative, are attributed to leaders, and then determine whether these leaders are viewed in a positive or negative light.⁷

Such social-psychological forces exert a powerful pull on all of a society’s members, elites as well as ordinary people. Their power over corporate directors becomes evident when overestimation of the CEO’s importance turns the external CEO search into a quest for a charismatic corporate savior. Recall the Bank One directors gushing about Jamie Dimon’s intelligence, and their resultant high expectations as to how analysts and the media would react to the appointment of such a star—as opposed to the negative reaction, it was presumed, with which Verne Istock’s appointment as permanent CEO would have met. All this attention to Dimon’s charisma is all the more extraordinary when one considers how little the Bank One board’s search for a new CEO seems to have been geared to evaluating candidates in light of specific, demonstrable needs of the organization. For example, while Dimon did have tremendous clout on Wall Street, he did not have much experience with retail banking or credit card operations—two of Bank One’s largest businesses, and the latter the source of many of its operational problems at the time of the search.

Furthermore, given the distinct culture of commercial and retail banking, Dimon was a somewhat odd choice. After all, he had spent virtually his entire career as a dealmaker and on the investment banking side of financial services, a world described by Roger Brunswick and Gary Hayes as focused on the “here and now” and the art of the deal.⁸ In their book Doing Deals, a study of the management of investment banks, Robert G. Eccles and Dwight B. Crane find that individuals who excel at investment banking typically have extraordinary financial acumen, impressive sales skills, indefatigable energy, keenly developed intuitive skills, and oversized egos.⁹ By contrast, commercial and retail banking are predicated on analytically and objectively calculating and assessing long-term trends and risks. Banking cultures such as those of Bank One and First Chicago value recognizable facts that can be discussed and debated beyond the influences of personality and emotions. The individuals in this industry tend to be methodical and focused on the long term, for effective coordination across their enterprises has traditionally depended on the predictable behavior of executives. This being the case, an inability or unwillingness to control one’s temper—qualities for which both Dimon and Sandy Weill had become known—would have been considered in an earlier era a prima facie disqualification for the CEO position in a commercial or retail bank (as they would have been in virtually any professionally managed organization).

Yet one other difference between commercial and retail banking, on the one hand, and investment banking, on the other, does help to explain Bank One’s choice of Dimon. In terms of organizational practices, commercial and retail banks have always paid far more attention than their investment banking counterparts to promotion from within and have traditionally invested heavily in human resource systems to support managerial recruitment and training.¹⁰ Investment banks and brokerages have largely relied on a star system for recruiting.¹¹ It is therefore not too surprising that few U.S. bankers displayed Dimon’s energy and seemingly unique persona, and even less surprising that Bank One’s board was so taken with Dimon, given the importance that it had attached to the criterion of charisma.

As the Bank One story also illustrates, however, it is not only the criteria directors use in choosing a new CEO that call into question the efficiency and overall rationality of the external CEO market. So do many other features of the search process itself. For example, while it is often assumed that CEO searches are wide and extensive, the Bank One story illustrates that the external search is actually a very closed process: in this case, all four external finalists were known, directly or indirectly, to members of the Bank One board before the search even began. Moreover, much of the information about the candidates was gathered through the contacts of two board members, John Hall and Jim Crown. While being closed in this way, the external CEO search process is at the same time surprisingly porous in being influenced by the views of outside actors such as shareholders, analysts, the press, and others who are not direct parties to the transaction. Interactions between the searching firm and the candidates, moreover, are facilitated by another set of outsiders, the executive search consultants, whose role in the process—again as illustrated by the Bank One case—appears more like that of a master of ceremonies or diplomat than that of a broker facilitating connections among atomized buyers and sellers (the latter being the commonly held view of their function).

Finally and, in some ways, most tellingly, the external CEO search often ends with an outcome that seems less than efficient from an economic point of view. Like other new CEOs hired as the result of external searches, Jamie Dimon exercised his extraordinary bargaining power to strike a highly favorable compensation scheme. While the initial stock run-up following his appointment suggested that Dimon was worth the price, several studies examining the long-term consequences of outsider CEOs on firm performance find that such initial bounces in stock prices are usually short lived. Most of the long-term studies find no effect on firm performance from hiring an outsider CEO. And in some circumstances, the arrival of an outsider results in a decline in long-term performance.¹²

In the case of Bank One, not only did the initial boost to the company’s stock price following the announcement of Dimon’s hiring prove to be very short lived, but outsiders also quickly began to question whether the board was exercising any control over its new CEO. The question was apt. CEOs have traditionally wielded enormous clout relative to the boards that are, legally as well as in theory, their bosses, and Dimon, newly installed as Bank One’s CEO, looked well set to follow in this tradition. The Bank One board invested Dimon with a great deal of power when, for instance, it returned to the practice of combining the CEO and chairman positions, which had been separated following McCoy’s departure.¹³ Then in August 2000, five months into his tenure, Dimon announced that Verne Istock and five other directors had “volunteered” to retire from the board, which would thus be shrunk from nineteen to thirteen members. Unbeknownst to shareholders, moreover, the search committee had agreed in advance to the cut during its negotiation with Dimon. Dimon had concluded that a smaller board “would be more nimble” and effective for Bank One shareholders; he probably also recognized, as do many CEOs, that a smaller board would be easier for him to control.¹⁴ Two months later Dimon also convinced the board to appoint two of his earlier acquaintances as board members: Heidi Miller, the former CFO of Citigroup, and David Novak, the CEO of the restaurant conglomerate Tricon, on whose board Dimon himself sat. Finally, in a move that was criticized by several industry observers, Dimon dismissed a number of Bank One executives and replaced them with trusted former Citigroup and Travelers colleagues.¹⁵ In all, as part of the price for obtaining the charismatic CEO that it so fervently desired, the Bank One board had handed Dimon the kind of extraordinary power and privilege that charismatic leaders, in all times and places, are accustomed to command. Whether the benefits the board thus hoped to gain for shareholders would materialize at all—let alone prove to be worth the price the directors had agreed to pay in anticipation of them—would, of course, remain an open question for an unforeseeable length of time.

How are we to account for these remarkable, ultimately disquieting features of the external CEO search: the overestimation of the CEO’s role and the fixation on charisma; the somewhat Byzantine nature of the search process itself, simultaneously closed to many presumably qualified candidates and open to the influence of many external actors; and the questionable outcomes that this process often produces? This book is an attempt to answer this very question. For the time being, any clear-eyed analysis of the external CEO search process must begin with one central fact: the external CEO labor market is not a “market” in the ordinary sense. It is not even like the market for other executive positions. Scenarios such as Bank One’s wooing and winning of Jamie Dimon unfold against the background of several conditions that make the external CEO labor market a truly different kind of market. Its uniqueness arises from a combination of three features: small numbers of buyers and sellers, high risk to participants, and concerns about legitimacy.

Small Numbers of Buyers and Sellers

Rather than being what economists call a “perfect” market, with large numbers of buyers and sellers engaged in relatively anonymous exchange, the external CEO labor market is one in which buyers and sellers—or at least those considered qualified sellers—are relatively few. The number of buyers is a function of how many (or rather how few) companies are looking for a new CEO at any given time. Indeed, most potential candidates are aware of which companies are conducting a CEO search, particularly when poor performance has led to the decision to change CEOs. (Recall that, in the case of Bank One, Dimon himself felt that he would likely be considered for the CEO position, and therefore had expected a call from the search firm.) The small number of “sellers” in the external CEO market (for reasons that will become apparent momentarily, “sellers” has to be used in a very restricted sense in this context) can be illustrated by the lists of candidates that the Bank One search committee considered at various points in its CEO search. Drawing on both the directors’ and the search consultants’ knowledge of the banking industry, the Bank One search committee and its consultants produced an initial list of the leading lights of finance, one that contained thirty names. After a week of fact checking, the search consultants reduced this first list but also added some women and minorities whose names had not originally appeared. They then gave the search committee a booklet with profiles of each of the remaining candidates. The names of almost all of these candidates were familiar to the directors since the pared-down list was a virtual dream team for banking and finance. Some of the directors even knew several of the candidates personally.

A situation such as that in which the Bank One directors were aware of the candidates, and the candidates of the vacant CEO position at Bank One, is not unusual in CEO searches by large, publicly held firms. Relatively few CEO positions open up each year in comparison to other positions. Table 2.1 presents the number of outsider CEO searches undertaken by the firms in the sample in comparison to searches for vice presidents of marketing by these same firms. Firms conduct only onesixth as many CEO searches as they do VP-marketing searches. Moreover, search firms receive no unsolicited resumes from individuals looking for CEO jobs. One search consultant cogently summarizes both consultants’ and directors’ common perception by stating that “unlike [with] other positions, the CEO market is not simply a problem of a person looking for a job, but a job looking for a person.”

Table 2.1 also quantitatively represents the most commonly held perception of directors, executive search firms, and CEO candidates about this market—that the supply of qualified candidates for the CEO position in large corporations is thin. When contrasting the search for a CEO with that for other executives, one search consultant whom I interviewed commented that “the number of people who can run a 50,000-person organization is small, and most of us know them off the top of our head.” Directors, too, lament the presumed shortage of CEO talent.

Ideally, of course, one would want lots of organizations and lots of executives looking at one another to ensure a good match over the long run between executives and open CEO positions. Most firms, however, rarely realize these conditions in their external CEO searches. Yet while describing the condition of small numbers of buyers and sellers in the external CEO market, it is important to note that this scarcity is exacerbated, if not actually created, by the participants themselves.

TABLE 2.1

Search Firm Statistics Contrasting CEO Search with VP Marketing Search

The shortage of qualified sellers is at its core a misperception largely driven by the fact that boards employ extremely limiting criteria to define the pool of eligible candidates. These criteria, which are loosely (if at all) coupled to the specific strategic challenges facing the firm, are adopted largely with the intention of producing a candidate who will be seen as legitimate by external constituents, namely, financial analysts and the business media. The application of these criteria focuses the board’s attention on a small number of candidates. As a result, the perceived shortage of qualified CEO candidates is exactly that—more social fiction than empirical reality. Bank One’s directors, for example, insisted on a candidate who would restore the credibility of the company with Wall Street and financial analysts. This translated into a candidate pool in which the final external candidates were either CEOs or presidents of major financial institutions that had performed relatively well and were recognized as being of high quality and status. (Additional support for this interpretation is found in Table 2.1, which shows that the number of candidates perceived to be qualified for the position in the initial stages of a CEO search averages only thirty. On average, only five of these candidates are formally interviewed by the directors.) For their part, most of these candidates shared Dimon’s view that there were “likely only one or two other firms I would ever be happy at.”¹⁶

How does the “thinness” of the external CEO market affect its functioning? The number of buyers and sellers in a market is both an indication of the number of alternative relationships available to actors and a critical element in their definition of the market itself. For competitive markets to exist, economists have argued, the number of buyers and sellers must be so large that the ordinary transactions of any single one of them do not appreciably affect the conditions under which other transactions are made. Conversely, the smaller the number of sellers in a given market, the greater the effect of the transactions of any one seller on the fortunes of the others. In the external CEO market, the initial, totally artificial reduction of the size of the candidate pool is generally followed by a similar market-distorting maneuver during the negotiation of compensation. Both Derek Bok and Graef Crystal have argued that the reason CEO pay continues to ratchet upward—even during periods of declining firm performance—is that it is usually anchored to a median calculated from a ranking of a small number of other CEOs.¹⁷ Because the comparison group is small and most boards have adopted a convention to pay their CEOs above the median of their comparison group, an increase in pay for one CEO will result in raising the pay of others.¹⁸ The ratcheting is further exacerbated by CEOs’ exertion of influence on the comparison group, via their chosen compensation consultants, by “eliminating certain companies and adding others.”¹⁹

Small or thin markets also reduce competition by limiting the set of other actors with whom a focal actor may contract in efforts to conduct transactions, and by making the search for these trading partners more problematic. For example, buyers and sellers in disconnected or sparse networks may not be aware of the full range of trading partners or opportunities for exchange. Even actors in the same social networks may not be aware of each other’s complementary interests if the actors constituting that network are differentiated with regard to activities or interests or beliefs about what is discussable, so that direct interchange between such actors is muted.²⁰ For example, two members of a golf club or university club may not be aware of how they could benefit each other if talking business at the club is discouraged. This suggests that two actors will be accessible to each other for exchange only if their interests become known and are tolerably consistent. This creates a problem in CEO search because both candidates and boards of directors lack the information necessary to make informed decisions. Aside from the small number of companies known to be searching for a new CEO, a potential CEO candidate will not know what opportunities exist in the market. Similarly, directors will not know if there are qualified people who would be willing to take on the position for less pay than the small number they have already identified as candidates presumably would.

Another analysis of the significance of small numbers of buyers and sellers in a market has been proposed by the sociologist Wayne Baker, who studied the Chicago commodities markets to demonstrate the implications of the numbers of buyers and sellers for a market.²¹ Baker found that in contrast to the competitive markets typically described by economists, in which no set of buyers or sellers can dominate the market, many commodities markets were specialized and esoteric, and only a handful of dealers in the world knew how to buy and sell the goods in them. Baker found that to work, these markets needed extensive coordination between buyers and sellers to create the market in the first place. That is, the formal market structure of thinly traded commodities was supplemented and occasionally subverted by an informal social structure of roles, relationships, and collective action. People would often repeatedly trade with the same people. At other times, traders would avoid taking opportunistic advantage of another trader whose financial exposure was too great. One consequence of this small number of participants is that these markets did not display the processes of equilibrium so often assumed in standard economic analysis. Instead, there were large price discrepancies and evidence that the participants’ behavior resembled that in a village store, with concerns about interpersonal relationships, reciprocity, and trust—rather than the relentless pursuit of profits—undergirding the behavior of the traders.²²

Elements such as reciprocity and trust, in turn, become important in the external CEO labor market owing to a second fundamental condition of this market: the risks faced by both buyers and sellers.

High Risk

Previous researchers have not explicitly considered the problem of identifying the potential pool of candidates in the external CEO market. Since most candidates in the CEO labor market are already employed, those seeking to employ them do not have a clear view of whether any of these individuals would be willing to leave their current positions or whether they would be interested in a particular firm that has an opening. The basic condition of a passive CEO labor market creates a high-risk situation for both firms and potential candidates.

The risk to both the candidates (sellers) and the firm (buyer) in this market is reflected in much of the behavior that we observed in considering Bank One’s CEO search. In the Bank One case, once the search committee had whittled its list of thirty candidates down to five finalists, the Russell Reynolds consultants arranged the interviews to ensure the utmost confidentiality and even secrecy. One reason for the secrecy was that, with the exception of Dimon, the external final candidates were actively employed as CEOs or presidents at other firms. The Chicago Tribune and the Chicago Sun-Times published rumors about the identities of the candidates almost daily, and if it could be confirmed that any of the finalists other than Istock and Dimon had interviewed for the position, it would not bode well for these executives at their current firms. Since all of the finalists except for Istock were from outside Chicago, the consultants were careful to schedule interviews in Chicago for days when board meetings were also being held, so as not to arouse suspicion among Bank One executives should any of them run into an out-of-town director. They also took pains to schedule the candidates’ flights to eliminate any possibility of their seeing one another at the airport. The interviews were held at Russell Reynolds’ downtown Chicago offices, where Redmond and Tribbett carefully choreographed arrivals, departures, and the use of rooms to ensure, again, that there would be no possibility of the candidates’ catching so much as a glimpse of one another.

Meanwhile, even though Jamie Dimon was not actively employed at the time of the Bank One search—and therefore not at risk himself—even his name had to be kept under wraps for the sake of Bank One. Tribbett points out that if it had become public that Dimon was being considered, and then that he was not interested in the position or had actually refused it, any CEO who was eventually appointed, no matter how good he might be, would find it difficult to succeed. “If employees and analysts think the selected CEO is the ‘second-best’ person to run the firm, he has no shot,” says Tribbett. (As confirmation of this, consider the example of AT&T’s 1997 CEO search, which will be discussed in greater detail in chapter 4. John Walter was clearly the second choice of the AT&T board when C. Michael Armstrong refused the position, and AT&T’s image was dealt a severe blow.) Moreover, not wanting to be regarded as the second choice, the remaining strong candidates often withdraw themselves from consideration in circumstances where this danger exists, thereby often forcing the search to start all over again.²³

Confidentiality also becomes an issue when obtaining information about a candidate’s skills or capabilities, since such information cannot be obtained directly from his or her current employer. As one consultant comments, “You can’t exactly go to the guy’s boss and let him know you are thinking of hiring his CEO.” Although general information about a candidate’s educational background or work history can be collected easily from public sources, the firm searching for a CEO must rely on other, more private sources to gain particular knowledge about a candidate’s capabilities, temperament, character, and skills. Economists describe this need for detailed information as linked to the problem of adverse selection, a process by which the least desirable objects from any observationally similar group enter a market in which information is poorly distributed between buyers and sellers.²⁴ Although on the surface they may appear alike, outside candidates cannot be presumed equal. Some outsiders have deceptive career patterns and are available in the market because they continue to fail upward. Legal requirements and privacy protections make it hard to find detailed information about a candidate’s history. In other words, if a board relied simply on the information produced by the conventional market process, it would know little other than what the résumé and the candidate offered. In an example of the pitfalls of such an approach, consider the CEO search at Sunbeam that resulted in the hiring of Al Dunlap, a case in which the search firm was unaware that the candidate it had placed had been dismissed from a previous job for overseeing an accounting fraud that had culminated in the firm’s bankruptcy. While I will return to this point in chapters 4, 5, and 6, for now keep in mind that much of the particular information required to minimize risks to the hiring firm is gathered not by the search firm but by the directors. To obtain such information, directors rely on their prominence and strong ties to individuals who have had direct experience with the persons under consideration.

While some degree of adverse selection is always inherent in transactions in which information about an individual’s past behavior is incomplete, the problem is aggravated by a second type of risk specific to information about present and future behavior. This second type is now commonly described as an agency problem. Whenever one individual depends on the action of another, an agency relationship arises; for example, a lawyer is an agent for his or her client, and a CEO is an agent for the shareholders. An agency problem arises when participants in an exchange have divergent objectives, and information about how they are behaving and will behave in the future is imperfect or unavailable.²⁵

In the relationships that exist between CEOs and their companies, agency costs can be quite high.²⁶ Michael Jensen and others have offered examples of cases in which self-interested decisions by CEOs at the expense of shareholders have led to multibillion-dollar losses.²⁷ Moreover, even when information comes to light about irresponsible actions on the part of a CEO, it is often difficult to remove him. A CEO cannot be dismissed as easily as other members of the organization. CEOs, in fact—despite the increase in CEO dismissals over the last decade—have several tools at their disposal to make themselves both difficult and costly to replace, including controlling the agenda of board meetings and the ability to appoint board members.²⁸ Under such risky conditions, directors would not likely participate in the exchange that external CEO search represents without a means for reducing the information uncertainty associated with adverse selection and agency relationships.²⁹

Returning again to the case of Bank One, neither Bank One’s directors nor Dimon had the information required to engage in fully informed decision-making. Let us consider the situation from Dimon’s point of view first. While the Bank One directors were somewhat forthcoming with their leading candidate regarding the problems facing the bank, they also wanted him to take the job and so tried not to make these problems look insurmountable. In any event, most had only limited understanding of the firm’s situation by virtue of being directors rather than full-time executives. (Most of the information that Bank One’s external directors relied on for assessing the firm’s condition came from the external environment and from former First Chicago executives—agents whose motivations potentially diverged from those of the directors themselves.) While Dimon, in the meantime, contacted financial analysts to better understand Bank One’s problems, he recognized that analysts rely primarily on financial indicators to analyze the company. One ironic consequence of this particular information asymmetry was that Dimon was able to negotiate a relatively high salary and guaranteed bonus as a type of insurance in the event that the problems at Bank One proved to be more severe than he had understood. Thus the apparent risk to Dimon turned out to increase the risk to the directors.³⁰

For their part, in an effort to gain specific and detailed information about Dimon, Bank One’s directors relied on trusted social connections, particularly with other directors who had known or worked with Dimon in the past. Yet even after gathering as much information of this kind as possible, the directors found it impossible to know ex ante whether they had made an intrinsically sound choice. By picking a candidate who was highly regarded (given Dimon’s performance as president of one of the most highly regarded financial institutions in the world, Citigroup), they hoped at least to be able to justify their decision to others. In its manifest concern for the opinions of others, their strategy for dealing with risk points to the third key feature of the external CEO market.

Concerns about Legitimacy

The third important feature of the external CEO labor market is the way in which it is driven by concerns about legitimacy. According to the sociologist Richard Scott, the legitimacy of an action “is determined by the amount of consensus within the relevant sector or field regarding the appropriateness of the means selected to achieve the desired ends.”³¹ Because the directors and candidates involved in external CEO search are embedded in a community of overlapping business and social relationships, they are particularly sensitive to maintaining the appearance of propriety in the conduct of the search among their peers. Moreover, because the opinions of external actors such as analysts and the business media are so important to ensuring the eventual acceptance of the candidate, that the process be perceived by these outsiders as both objective and proper is critical.

In the Bank One search, the directors faced enormous pressure from shareholders, analysts, and the business press first to fire McCoy and then to act quickly to find a successor. This successor was also going to have to be someone whose appointment would signal to this external audience that Bank One was serious about solving its problems, and that it still enjoyed sufficient prestige to achieve outside director John Hall’s stated goal of finding “the best person in the United States to lead us back to the top”—which would require mounting what outsiders would accept as a wide-ranging, objective search. (The very real link between the external CEO search process and the maintenance of a firm’s standing in the eyes of external constituents is reflected in the stock market’s response first to the news of McCoy’s departure, then to Dimon’s appointment, and eventually to the doubts that were quickly raised about the latter’s ability to perform as expected. The idea that the CEO’s performance determines a company’s fortunes may be a myth, but that does not make it any less powerful.) Finally, the search had to be pursued in the midst of more-or-less open strife between two opposing factions on the board. The highly politicized context that this conflict created made it all the more imperative that the Bank One directors and their candidates appear to have conducted themselves appropriately.

The importance of concerns about legitimacy in the external CEO search process is highlighted by one aspect of the role of the executive search firm: the part that search consultants play in overcoming the reservations both parties may feel about their own participation in the market. Although a key part of the search firm’s role, this function is easily obscured by other issues that arise between firms in the market for a CEO and CEO candidates.

The high stakes and high risks for both directors and candidates in external CEO searches greatly increase the possibility of heightened emotions among the actors: for example, the hiring firm can easily become frustrated with a candidate who seems to take too long to make a decision, or who is perceived as making extraordinary demands regarding compensation, perquisites, or employment contracts. Much as in international diplomacy, such complex, emotionally fraught negotiations usually require the participation of a third party to resolve not only substantive issues such as compensation but also human issues such as frustration or anger—factors that can easily poison a working relationship between a board and its desired candidate. One consultant, describing his job as “part recruiter, part messenger, and mostly marriage counselor,” recounts a particularly intense negotiation in which the firm had become frustrated with the “seemingly endless demands the candidate was making on compensation-related issues”:

[Steve] kept making a longer and longer list of what he thought should be covered in order to “make him whole” as a consequence of the move. Meanwhile, the board was getting pretty annoyed at looking at the detailed requests about unexercised options, initiation fees at new country clubs, and deferred compensation, etc. I saw what was happening with respect to the frustration level. . . . I stepped in and said, “Why don’t we just put all of this into a one-time signing bonus? In that way, [Steve] doesn’t have to go through accounting for every little cost he was going to incur and the board didn’t have to review and approve every one of these . . . expenses.” . . . [T]hey [quickly] came to agreement on an amount that was perceived fair by both sides. I think it was twenty percent of the first year’s cash compensation.³²

When search consultants are asked to explain why candidates and directors—often portrayed as the kinds of people who are rational, cool, and in control of their emotions—become so sensitive and easily frustrated during such negotiations, they almost always seem irritated by a question about what to them is an obvious point. In two-person interactions, they point out, directly negotiating a salary or other sensitive matters can provoke intense emotions. The use of an intermediary, however, dampens the feelings that normally arise in such negotiations and represents one party’s demands or responses to the other in a more conciliatory, objective manner.

This explanation of the search consultant’s role in facilitating negotiations between boards and CEO candidates is a plausible one, and in several other settings, such as divorces and labor-management negotiations, seems to be a principal reason why third parties are brought into tense negotiations. However, we can look at the intermediary in an external CEO search in another way by considering the hesitancy that both firms and candidates express about actively participating in this market at all. We have just seen how the involvement of an executive search firm in an external CEO search helps protect participants from risk. Yet the executive search firm’s role also helps allay concerns about how the appropriateness of their actions will be perceived both by their peers and by outsiders. For example, both directors and candidates feel constrained by norms concerning the propriety of contacts, in the course of an external CEO search, between board members and CEO candidates who belong to competing firms. In the face of such constraints, the search firm protects both parties from appearances of impropriety by eliminating the need for direct contact between the two in the early phases of a search. That directors and candidates, in turn, are generally linked to one another by personal connections creates another legitimacy issue that the executive search firm helps to resolve. In this case, the mere presence of a third party mediator lends an appearance of distance and objectivity to what otherwise might be suspected by outsiders to be an essentially social exchange.

In view of these realities, a more far-reaching interpretation of the search consultant’s role than the “marriage counselor” one would start with the supposition that the gap between buyers and sellers in a market is, in part, institutional, and is linked to the degree to which participation in a market is normatively legitimate. In legitimate markets, buyers and sellers can engage openly (and presumably with less heightened emotions) in exchanges without fear of repercussions. By contrast, in markets of questionable legitimacy, actors may be unwilling or hesitant to engage or participate, even if their interests suggest that an exchange would be mutually beneficial.³³ The role of the executive search firm thus is to help both searching firms and candidates to overcome their ambivalence about participating in a market of this particular kind.

The concern with market legitimacy on the parts of buyers and sellers, in turn, highlights an important fact about interaction, even in markets: it is shaped by a collective, communal structure. Some interactions that could take place in a market do not, in fact, do so because all interaction occurs in a context of institutional constraints, including rules and roles.³⁴ Rules impose some of these constraints: for example, the rule that organizations cannot exchange cost information with competitors. Role expectations impose constraints as well: directors, for instance, are asked to resign or to recuse themselves from decisions when their memberships on other boards present the potential for a conflict of interest. Under conditions in which a market is perceived as illegitimate, actors may be unwilling to make even initial contact or to engage in ephemeral transactions with others unless there is a means of keeping appropriate distance until a transaction can be consummated. Indeed, in the external CEO market, the role of legitimacy concerns is underscored by the numerous informal restrictions and rules that influence the manner in which interactions must take place. The enforcement of restrictions on, for example, the ways that candidates can be approached is part of this market’s distinctive character. In the external CEO market, legitimacy concerns are particularly acute because market decisions occur not only at the level of the individual organization but also at the organizational field level, in which other actors evaluate the outcome.³⁵

Under these circumstances, the best searches serve to legitimate both the search process itself and the final choice of the search committee and board so that a new CEO can have a smooth transition into the position. Missteps during the search process can leave organizational observers and constituents alike seething about the search and hostile to its outcome.³⁶ The search ends up an abysmal failure, not because the wrong candidate was selected, but because someone who may have been right for the organization is handicapped by the mishandling of the process.³⁷ The outcome of a search that, from a strategically sound perspective, may be seen as appropriate may be rejected by both internal and external constituents because of the manner in which the process was executed.

Much of the board’s concern with legitimacy in external CEO search stems, in part, from a weakening of the boundaries and secrecy that once surrounded CEO succession. Naming a new CEO is no longer considered a divine right of the CEO or even of the board. Whereas in the past CEOs or boards of directors could be allowed to choose the successor because of the perceived validity of peer review—the thinking being that people who have been in positions of leadership are best positioned to decide who can run the corporation—the internal succession process now has the aroma of a smoke-filled room. CEO succession is increasingly treated as an event in which external constituencies have both a strong interest in the outcome and a right to influence it.

These external constituencies, namely Wall Street analysts and the business media, now constitute a legitimating authority for most organizations. Ezra Zuckerman, a sociologist of markets, has documented the increased power of Wall Street analysts in the determination of an organization’s stock performance. He argues that investors, especially institutional investors, increasingly rely on the judgments of these analysts when making investment decisions.³⁸ Board members, as a result, must now pay more attention to them, too. “If you go back a few years ago,” notes George Kennedy, a director of several public firms, “I don’t recall reading the analyst reports on the boards that you served on, everybody else’s view as to how things were going. You might read the articles in the Wall Street Journal if you are on the board, obviously, but I don’t remember reading all those reports.” Other directors observe that, through clipping services, they now receive almost every newspaper or magazine article that mentions companies on whose boards they serve. The cumulative result of these changes is to focus directors’ attention away from the immediate situation of the firm and toward the externally relevant actors. Directors who control organizations now try to interpret their external environment and then make succession decisions based on their reading of those external actors whose opinions they most value—and with good reason. (Again, the example of AT&T’s 1997 CEO search—and the subsequent search the firm had to undertake nine months later—provides a vivid illustration of this point, one that I provide in chapter 4.)

A market characterized by small numbers of buyers and sellers, high risk to both, and such concerns about legitimacy as those outlined above, does indeed bear little resemblance to what is most often meant by a “market,” even though previous observers have generally described it as such. The reason for the current lack of understanding of the true nature of the external CEO labor market is that none of the lenses that have been used for observing and analyzing it are adequate for perceiving, much less comprehending, its complexity. Before looking in more detail at how this market actually works, we must explicitly consider the various frameworks that have been used to study it, and outline the perspective that will be used in examining it here.

The External CEO Market as a Socially Constructed Institution

For neoclassical economists, the external CEO labor market is a market like any other.³⁹ In their view, markets are constituted by large numbers of individuals acting primarily in pursuit of individual self-interest and engaged in relatively anonymous, friction-free exchange. Neoclassical economics also argues that social relationships and institutions do not matter. Society stands apart from the economic transaction. Parties other than the immediate buyers and sellers do not alter the choices and subsequent actions of economic actors, since perfect information and stable preferences are assumed. The story of the Bank One search obviously belies this perspective in many ways.

The other group of social scientists who have studied the external CEO labor market consists of sociologists who use a structural perspective. In contrast to neoclassical economics’ individualist framework, the structural perspective—building on the remarkable discoveries of recent sociological, economic, and historical research—views actors in a market as submerged in social relationships.⁴⁰ From a structural point of view, certain actors are connected to certain others, trust certain others, receive information from certain others, give information to certain others, are obligated to support certain others, and are dependent on exchange with certain others. This recognition points away from economic calculations and toward patterns of social relationship, reputation, information flow, constraints, opportunities, and the determining roles of community and power. It also points toward the fact that individuals, in pursuing and safeguarding their own interests, often act as much, if not more so, to safeguard their social assets and relationships. To take one example of how much more accurately a structural perspective accounts for the character of the external CEO market than the conventional economics perspective does: although external CEO searches are widely considered to be broad ranging in the way that the neoclassical economic view of the market would suggest, they are in fact (as the Bank One case illustrates) confined within the relatively narrow bounds of inter-firm networks. To take another example: while neoclassical economists’ accounts of the external CEO labor market have presumed this market to be “institution free,” external institutions such as executive search firms, interlocking directorates, and (as in the Bank One case) investors, analysts, the press, and other groups of actors, play an important role in controlling access to jobs and facilitating mobility into the CEO position.

The structural perspective on the external CEO labor market represents a considerable advance over the traditional economic analysis in that it recognizes the centrality of social networks and relationships to the search and selection process.This perspective, however, has its own limitations. And what neither view—the mainstream economists’ nor the structural perspective used by many sociologists—takes into account is the way that the market is perceived by the actors involved, and the culturally conditioned beliefs bound up with their perceptions. Put another way, the structural perspective is contextually sparse because economic action is restricted to the constraints of networks of relations. Considerations such as societal institutions, the preconditions for market exchanges, and how social practices define appropriate and inappropriate forms of market behavior are simply set aside.

To see how social institutions and beliefs exert their influence on the external CEO market, consider several details of the Bank One story once again. In the Bank One search, the business press and financial analysts both influenced and refereed the search process. Negative press, true or not, played a key role in the removal of John McCoy. Similarly, the ability of financial analysts to drive investors to buy or sell Bank One’s stock had a decisive influence on the company’s directors. It was essentially social beliefs that led the Bank One directors to limit the candidate pool to individuals who were currently CEOs or chairmen at prestigious financial institutions. Moreover, the selection of Dimon was in part determined by the Bank One directors’ desire to appease cultural institutions, such as the press, and was seen as a way to signal to competitors that, in the words of one director, “Bank One is back.”

To say that actors’ perceptions of the market and of the structural position of actors in it are important constitutive dimensions of that market is also to assert, as I do, that the external CEO labor market is a socially constructed institution. Sociologists will understand my use of the term social construction, which comes out of the sociology of knowledge, in a way that those outside the discipline may not.⁴¹ To call the external CEO labor market a socially constructed institution is not to say that it is an institution in the everyday sense of that word. Rather, the CEO labor market is an institution in the way that sociologists think of the term: that is, a pattern of practices, relationships, and obligations that are so taken for granted that they assume the status of rules governing both thought and action.

Socially constructed institutions play a pivotal role in the economic functioning of any society. The ideas of the sociologists Mark Granovetter and Viviana Zelizer have been central in developing this concept. In one case study, Granovetter and his colleagues examine the origins of the American electrical utility industry.⁴² Through an examination of the state of technology and the economic costs of producing electricity, the researchers suggest that in the 1880s, when the industry was just taking shape, it was by no means clear that it would be organized in its present, taken-for-granted form of investor-owned utility companies generating power in central stations for large areas. Instead, there were two other possibilities: generation of power at a household or neighborhood level, and public ownership of utility firms. What determined the investor-owned model (which until the 2001 California electricity shortage had been taken for granted as the most efficient model for producing electricity) were a variety of non-market forces that are typically ignored in explaining the economic trajectories of particular industries or institutional arrangements, but which converged and interlocked in powerful ways to forge the industry structure that exists today. Granovetter and his colleagues find that trade associations, director interlocks, decisions by equipment manufacturers about who to sell equipment to, and the ability to mobilize social, political, and financial resources, all played a pivotal role in creating an industry that is now dominated by a few large holding companies using standardized methods of generation and organizational structure, and protected by government agencies.

While Granovetter’s research has emphasized that networks of relations constrain and facilitate economic action, Viviana Zelizer has emphasized the role of cultural factors in creating and sustaining markets. In Morals and Markets, Zelizer solves one of the great puzzles of nineteenth-century economic history, the sudden popularity of life insurance in the United States after decades of failure to convince the public of its importance.⁴³ Based on a careful examination of historical documents related to the industry, Zelizer finds that public resistance to life insurance was largely the result of a value system that condemned putting a monetary value on human life. Many early insurance firms discovered that the juxtaposition of concepts such as death and money conjured up unpleasant images and seemed to many people to violate the boundary between the sacred and the profane; indeed, the major religious institutions in the United States officially denounced life insurance as sacrilege, arguing that the very idea was incompatible with the Christian values of charity and compassion. Not until insurers mastered the delicate problems of fixing a price on an individual life and learning how to market financial provisions for mortality did the public’s viewpoint evolve and life insurance become a legitimate and acceptable financial planning product.

Turning again to the external CEO market, one sees that many actors both affect and are affected by that market as an institution. Some of these actors—employees and shareholders, stock market analysts, the business press—are indirectly involved in influencing the process. Others—boards of directors, executive search firms, and CEO candidates themselves—play a more direct role. Their perceptions of the CEO labor market, including their understanding of the rules that determine who can be considered as buyers and sellers, and of the norms defining appropriate behavior for participants, are essential to understanding how the process of CEO succession takes place.

To call the external CEO market a social construction is not only to call attention to the roles and perceptions of these actors but also to say that, while this market appears to be an external and objective institution—to have, in other words, what some scholars call facticity—this is merely an illusion, since the process of social construction consists precisely in making an institution look like an objective and external aspect of the environment.⁴⁴ One of the important lessons of the sociology of knowledge is that very little in society had to be the way it is. As the sociologists Walter Powell and Paul Dimaggio have found, institutions that have been socially constructed in particular ways probably could have been constructed in other ways as well.⁴⁵ Part of the process of social construction is to camouflage this fact, since society becomes more stable when people accept its institutions as simply given and share a common explanation for events.⁴⁶

This type of argument is not usually applied to economic institutions, but there is no reason why it should not be. According to conventional economics, economic institutions, including markets, are simply matter-of-fact solutions to problems (for example, the need to minimize costs or to maximize certain kinds of efficiency) that require them to exist or to exist in the particular forms that they have come to assume. Scholars coming from this perspective have treated economic institutions as if they had been constituted by the laws of physics or chemistry.⁴⁷ They conspicuously neglect the fact that the entrenching of an institution is, at its roots, a social process, highly dependent on a complex interlocking of factors, including the specific and shared personal understandings of the human agents involved in the process, the social connections among them, organizational conditions, and historical context.⁴⁸ The Panglossian treatment of economic institutions as inevitable and maximally efficient has, to date, hampered the progress of social science and the development of its ability to offer convincing explanations for many important economic phenomena.⁴⁹

For example, during the heyday of the multi-business firm, or conglomerate, in the 1960s and 1970s, several economists maintained that this was the ideal form of organization. They argued that the efficiency of a company lay in the way it was managed, not what businesses it was in. The idea that the conglomerate form was the most efficient gained significant following and was used both to explain and to justify the existence of conglomerates. Some economists argued that conglomerates could use their internally generated cash flow to efficiently allocate resources to exit old businesses and enter new ones.⁵⁰ Econometric models of the cost of capital were developed to justify the conglomerate strategy and show that conglomerates had a lower cost of capital, and thereby a competitive advantage over small companies seeking funding.⁵¹ Then in the 1980s, as inefficient conglomerates became the target of hostile takeovers, economists attempted to explain the takeover movement by arguing for the inefficiency of conglomerates.⁵² New mathematical models were constructed to show that conglomerates inefficiently allocated resources. Core competencies and focus became the mantras of the economic interpretation of business strategy and contemporary economic strategy research. The rise and fall of the conglomerate actually makes it clear that both the multi-business firm and its more focused, streamlined successor were socially constructed. At the same time, economists’ contradictory analyses of the conglomerate show how it is possible—not to say necessary—to ascribe efficiency to any particular form of organization once one assumes, a priori, that economic institutions are organized as they are for the purpose of maximizing efficiency.

The social constructionist perspective is particularly useful for studying the CEO labor market because of a similarly evasive maneuver that economists have used to explain the tendency of large, publicly held corporations in recent years to turn from the internal to the external market when replacing CEOs. During the many years in which new CEOs were selected from inside the firm, economists saw the structure and rules of the internal CEO labor market as having a self-evident rationale.⁵³ The argument went something like this: CEOs were chosen from inside the firm because this was the most efficient way to select them. CEOs require know-how and skills specific to a firm to run it most effectively. Moreover, when new employees join a firm, it is not clear in advance who will acquire the requisite know-how and be the best person to run the firm. Firms therefore use internal labor markets—which utilize training and promotions—both to impart firm-specific know-how and to learn who is likely to be most effective in the CEO’s job. Today, confronted by a more ambiguous situation in which almost one-third of all the CEOs selected by large, publicly traded corporations can be classified as outsiders, students of the CEO labor market often dismiss this change as peripheral, a mere epiphenomenon.⁵⁴ A social constructionist perspective, by contrast, allows us to see that the external CEO labor market has its own social structure and way of being that cannot be explained with reference to economic criteria such as efficiency.

This is not to suggest that an institution like the external CEO labor market is completely independent of economic forces. Likewise, I am not suggesting that such economic institutions, once formed, are immune to economic factors. What I do argue is that economic institutions can be understood only within the context of broader social structures. Let me give a concrete example to illustrate this theoretical point.

One of the three critical actors in any external CEO search is the executive search firm. Many economists, and even many sociologists, treat the existence of intermediary institutions such as executive search firms as responses to an economic need. The gist of the economic argument is that when corporate directors need to find a new CEO, they are faced with a “make or buy” decision.⁵⁵ That is, they can either undertake the search by themselves or farm it out to a third party. Because executive search firms specialize in search, it is argued, they are likely to have cost or scale efficiencies that the directors do not enjoy. Directors therefore compare the transaction cost of conducting the search by themselves versus the cost of hiring an outsider. The mainstream sociological argument, in turn, suggests that because they know of candidates who are unknown to the firm, executive search firms have an information advantage over the directors for which they can and do extract economic “rent.”⁵⁶

One problem with both the economic and the existing sociological explanations of the role of executive search firms is that neither explanation meshes with the facts of how external CEO searches actually unfold. For one thing, boards of directors hire search firms even when they have no intention of looking outside for a new CEO. For another, directors seriously consider only candidates who are known to them from the start, before they have even hired a search firm. On further consideration, the conventional perspectives of both economists and sociologists on executive search firms can be seen to sidestep several important questions. Where did executive search firms come from? Why didn’t firms go outside for CEOs in the past? And why would a board hire an executive search firm even when it intended to promote an inside candidate, or when it already knew the identities of the outside candidates whom it would consider? Most social scientists who study the CEO labor market have evaded such questions because they require attention to concrete actors, such as directors and CEO candidates, who are embedded in networks of relationships, expectations about behavior, and uncertainty regarding their proper course of action. Such considerations lie very far outside traditional ways of thinking about the CEO labor market.

The CEO Labor Market as a “Closed” Market

The preceding discussion underscores the fact that key differences between the external CEO labor market and more conventional markets are so great that the former cannot be understood using standard economic concepts and market imagery. The analytical lens of social construction was introduced because the social organization and the culture of the main participants in the external CEO market affect both the organization and inner workings of this market and its outcomes so directly and profoundly that only an approach that explicitly considers the relationship between economy and society is capable of yielding insights and generalizations of importance. One idea that captures particularly clearly how both the structure of relationships and the cultural meaning of economic action are linked to the external CEO labor market is the concept of the closed market.

The core idea of closed markets can be found in the sociologist Max Weber’s discussion of social closure presented in Economy and Society.⁵⁷ Weber speaks at a general level of closed and open relationships. Here the terms “open” and “closed” will be applied to a market; no change in meaning is implied. In speaking of open and closed markets I refer to the ease or difficulty of access to the basic market relationship, which, in the CEO labor market, means meaningful access to the position of CEO.⁵⁸ Open markets are those that have the properties assumed in neoclassical economic theory, and the mechanisms that allocate people to positions in such markets are those described by the standard economic theory.⁵⁹ In open markets, a very large number of transactions occur simultaneously and independently of one another. They establish equilibrium wage rates, and no one is prohibited from working at some specific wage. No single transaction will affect the market as a whole. Employers can rely on active competition among potential workers to minimize labor costs, and workers can rely on open competition among employers to ensure they get the market wage. While they cannot influence that market wage, they can increase their earnings by working more or by supplying a different and higher quality of work.

In contrast to an open market, a closed market exists whenever the distribution of opportunities is restricted to a narrow set of eligible groups or individuals. In Weber’s discussions, closure most often took the form of the singling out of certain social or physical attributes as the justificatory basis of exclusion. Weber provided numerous examples of how groups use caste systems, professional societies, and political institutions to improve their fortunes by restricting to a limited circle access to rewards and privileges. To do this, Weber argued, groups single out certain social or physical attributes that they themselves possess and then define these as the criteria of eligibility. In practice, almost any characteristic can be used to this end, provided that it can serve as a reliable mechanism for identifying and excluding outsiders. Exclusionary social closure is thus action by a dominant group designed to secure for itself certain resources and advantages at the expense of other groups. Where the excluded themselves manage to close off to other groups access to what remains of the rewards, the social system becomes more highly differentiated and the number of strata or sub-strata multiplies. An example of this is the Indian caste system, which is based on a highly codified system of social classes that defines in detail what occupations can and cannot be pursued by each of the castes.⁶⁰ Apartheid is a more crude example of closure.

As we shall see, one of the defining features of the external CEO labor market is the way that it restricts access to the CEO position to those who fit certain socially defined criteria. While hardly as dramatic or morally repugnant an example of social closure as many others found in human societies in all times and places, the external CEO labor market, properly understood, offers a stark refutation of much of the received wisdom that surrounds it in contemporary America. For besides contradicting orthodox economic and sociological accounts of this market, the closed nature of the external CEO search process flies in the face of today’s near-religious faith in markets as a mechanism for achieving such socially desirable goals as equal opportunity based on open competition, or advancement on the basis of merit.

How have we come to this pass? Although, as I argue, the structures and beliefs constraining the major actors in the external CEO market are what constitute the market as an institution, these actors have not created these structures and beliefs themselves. Rather, they have received them as part of a social and cultural system that has developed over time in a series of steps that can be traced. Before turning to the roles of these actors in the external CEO market, we need to consider the origins of one of the key beliefs coloring their perceptions of this market: the very idea of the charismatic CEO.