CHAPTER 4
BOARD GAMES:
THE ROLE OF DIRECTORS
IN CEO SEARCH
THERE ARE two ways of understanding the role of corporate directors in external CEO search. One is based on the prevalent belief that actors in the external CEO market are what neoclassical economists say that actors in markets generally are: autonomous individuals rationally pursuing their self-interest in a context with a high degree of transparency. This approach focuses on the motives and behavior of the individual director. Today, this is perhaps the most commonly used method of accounting for directors’ roles in corporate governance. It is linked to the view that directors are fully motivated to act in the interests of the firm and its shareholders only when they have a financial interest in so doing. Advocates of this perspective have argued that without financial inducements such as stock options or share grants, directors “have little incentive to extend themselves beyond relatively superficial oversight of their firms’ affairs. They mechanically fulfill their specified duties (certain approvals and audits) and watch for egregious derailments, but not much more.”1
An emphasis on director shareholding has become all the rage in recent prescriptions for improving corporate governance. Today, almost all publicly traded corporations have adopted some type of director stock plan. Despite the proliferation of such plans, however, little evidence supports the belief that they have any effect on director behavior or firm performance. In one of the largest sample studies on the subject, John Core and his colleagues found that “percentage ownership per outside director . . . has little effect on the monitoring role of the board of directors.” The researchers concluded that there is no empirical support for “recent conjectures suggesting that outside directors should be forced to hold a greater amount of the firm’s shares in order to ensure that they have a financial stake in the outcome of their monitoring.”2
The focus on financial incentives for individual directors misses the truly critical point about director behavior: a board is not simply an aggregation of individuals, but in fact constitutes a complex group. The actions of a board, therefore, cannot be viewed as an aggregate of individual directors’ behaviors. A second and, I argue, more powerful way of understanding the role of directors in external CEO searches sees their actions as growing out of the connections and relationships in which individual directors are embedded, and that reach beyond the confines of any particular boardroom.
The Social Character of the Board
Once one moves beyond the standard individualistic models for explaining the behavior of corporate directors, it becomes apparent that group- and community-oriented behavior takes the place of calculations of self-interest when directors undertake an action such as searching for a new CEO. The nature of board membership, combined with the small size of the overall director community and the embeddedness of directors within a larger social system of third parties (such as analysts and the business media), results in an ingrained mutual dependence among people who share many relationships: those with whom board members are economically involved are the same as those with whom they are socially linked through shared status, organizational affiliations, and social standing. Several scholars have noted that directors are often connected to other leaders in the business community by virtue of their common residence in elite neighborhoods and membership in social clubs, professional associations, and public-policy committees.3 Directors’ actions in external CEO search are fundamentally regulated by these massive social facts of structural relationships and shared values.
Directors, to begin with, are highly cognizant of their membership in the communities to which they belong by virtue of their service on a corporate board. Like members of other close-knit communities, they have a clear sense of their own social domain—how far it extends, who belongs in it, and who does not.4 There is a social map, a list of members, and a system of gradations, hallowed by time and sanctioned by tradition. Whereas most of us are often only vaguely aware of our membership in social communities and of our place in a particular social system, directors explicitly know which groups they belong to, what communities they serve, where they are in the status hierarchy, and what their roles are. Two groups to which directors belong are particularly relevant to explaining directors’ role in CEO search and therefore are worth describing in some detail.
The first membership group with which a director identifies him or herself is a particular corporate board. Individual boards of directors are characterized by a high degree of internal cohesion, facilitated by several structural factors, the first of which is group distinctiveness. By distinctiveness I mean how easily a group can be defined and how easily members of the group can identify one another.5 Tom Piper, a board member at companies including Marriott Corporation, describes the “groupiness” of board meetings: “It is hard to explain to a person who is not a director. It is in many ways a club. You have a group of highly respected individuals who you work with. Each board I belong to has its own norms and ways of doing things, we value this. . . . We are conscious that being a director constitutes a unique set of roles and responsibilities.” The centrality of this distinctiveness is so strong, according to Piper, that there are “many people whose identities get caught up in being a board member of one or another particular firm.” The distinctiveness of a particular board is not only expressed in the group consciousness or “we-feeling” of directors but is also apparent to outsiders (such as researchers or investors), who can easily identify a firm’s directors through a search on a corporate website or in the pages of an annual report.
The second structural quality that lends a board its cohesiveness is the small number of people who constitute the group. Today, the average board consists of thirteen people, only one less than the average board in 1980. This smallness of scale gives rise to particular interpersonal and group processes. As social psychological research on organizational groups has shown, the smallness of a group heightens the members’ awareness of one another not only as professionals but also as private individuals.6 One director describes what he sees as the advantages of a small board by saying, “There is less formality in the smaller board. We can engage in more constructive and frank discussions with the CEO. Also, it is easier to develop a working relationship and team feeling with the other board members when the number of participants is smaller.”
The third structural factor leading to high cohesion among members of a board is homogeneity. Homogeneity in a group, the sociologist C. Wright Mills noted, eases role-taking and identification both with other individuals and with the collectivity.7 Although not one director I met (including the one woman and one African-American I interviewed) made reference to this obvious fact, board members are similar along a variety of observable demographic dimensions such as gender (they are almost all males), race (they are almost always white), age (they are in their fifties and sixties), occupation (most occupy or have occupied the highest administrative positions in organizations), class (most are quite wealthy), and status position (they are affiliated with high-prestige institutions and clubs, and almost all have a biographical sketch in Marquis’s Who’s Who in America).8
The sense of internal cohesion on a corporate board created by these three structural factors is reinforced, in turn, by the existence of group norms of a kind that can be found within any face-to-face group. These norms guide both directors’ behavior and their evaluations of others’ behavior. Many board norms are oriented specifically toward promoting board cohesion. There is a strong emphasis on politeness and courtesy, and an avoidance of direct conflict and confrontation. Aside from legally required committees, boards avoid specialization of any kind. In their face-to-face interactions, board members also show a tendency to shun distinct honorary titles of any kind, including Dr., Professor, or even Mr. or Ms. Seniority on the board is rarely discussed among directors. As such conventions suggest, boardroom decorum seeks to avoid the emergence of prestige groups of any kind within the group. Moreover, board members tend to hold in low regard members who are boastful about individual accomplishments. Making a big fuss over some individual accomplishment is considered brash and rude. In its own particular way, each board is a gentleman’s club, in both the best and worst senses of the term.
Structural factors and group norms thus combine to give members of a corporate board a strong sense of group identity. Yet because many directors serve on multiple boards, they are inevitably members of a second group: the broader director community. By “community” I mean a group that exists in such a way that its members share not this or that particular interest but rather the basic conditions of a societal position and a common perspective.9 In a landmark 1984 study, Michael Useem presented overwhelming evidence for the existence of this broader director community. Combining in-depth interviews and multiple data sources, Useem found that board members shared remarkably similar attitudes and beliefs regarding their roles as directors, and that they were highly conscious of themselves as a community and of their roles in this community. The foundation upon which this community rested, according to Useem, was the overlapping ties of directors created through interlocking board memberships. (An “interlock” exists when a person affiliated with one organization sits on the board of directors of another.) Directors who sat on multiple boards, especially those of prestigious companies, enjoyed particularly high status and influence. It is through these interlocks that director values and beliefs are both shaped and diffused. It is also through these interlocks that information about board practices is diffused.10
At the time when Useem conducted his study, a director’s loyalty to his particular board and his loyalty to the larger corporate community rarely came into conflict. Useem found that those directors who would have faced the greatest potential for conflicting loyalties—those with multiple board memberships—easily navigated between their roles as board members of particular firms and as members of the inner-circle group of directors who represented broader, inter-corporate interests.11 This situation was soon to change as the era of managerial capitalism gave way to the new era of investor capitalism.
Although not recognized at the time, the absence of role conflict that Useem observed was dependent on the institutional autonomy enjoyed by many corporations during the 1970s and early 1980s, the period of his study. As documented by Alfred Chandler, the managerial corporation was organized for autonomy, relying on its own retained earnings to fund operations and giving managers full discretion in deciding how much to return to shareholders and how much to invest in new ventures.12 Corporations were institutionally complete. Directors and managers were institutionally autonomous because they contained within themselves the necessary resources and organization for dealing with all aspects of corporate existence, obviating the need for the organization to be too dependent on other actors. Nowhere was this corporate autonomy more evident than in the marginal role boards played in CEO selection.13 Except in cases of overt crisis, boards played almost no role in shaping the CEO succession process or influencing its outcome in any meaningful way. Philip Caldwell, CEO of Ford Motor Company in the early 1980s, aptly summarized the CEO succession process for the boards on which he had served during the heyday of managerial capitalism: “[T]he incumbent CEO is the designer of a selection process that ultimately produces his successor.”
The shift from managerial to investor capitalism (described in chapter 3) that began in the 1980s created an environment in which corporations lost the autonomy they had previously enjoyed. The threat of corporate takeovers, the emergence of a shareholder rights movement spearheaded by institutional investors, and the increased influence of Wall Street analysts and the business media are just some of the many factors that ended the insulation of corporate management and directors from the turbulence of market forces. Change in CEO succession and search set in abruptly. CEOs, who previously had insulated themselves from the authority of directors in choosing their successors, were increasingly challenged by board members, who themselves were being pressured by institutional investors to take greater control over the CEO succession process, particularly when firm performance was unsatisfactory.
In this environment, the demands of external actors began to penetrate the boundaries of the director community. The previously insular culture of the board increasingly became an aspect or dimension of the societal context of which it was part. Directors’ actions were still influenced by the social dynamics of the individual board, but were now increasingly oriented toward reacting to the concerns of actors outside the immediate corporate environment. One consequence of this new orientation was the adoption of a widespread ethos among the director community exalting the creation of shareholder value to the virtual exclusion of other, potentially competing, corporate objectives.
This new orientation toward shareholder interests sometimes directly opposed the loyalty that board members felt toward the particular boards on which they served. For example, taking control of the CEO succession process—which boards were now under increasing pressure from outside forces to do—was an action that directly challenged past corporate practices and the authority of the incumbent CEO.14 Outside forces gained leverage, however, from the presence on boards of outside directors, who felt their responsibilities to the larger community of directors (e.g., the responsibility to show that the corporate world as a whole was mindful of shareholder interests) more keenly than did others on the boards on which they served, and who were less invested in the maintenance of past practices. Table 4.1 summarizes the changes in board composition between 1980 and 1990, and shows that, by 1990, on average one less insider served on a corporate board than a decade before. (As noted earlier in this chapter, average board size also decreased by one member over the same period. Both changes were the result of pressure from institutional investors.) Table 4.2, which summarizes the variables that predict the likelihood of CEO dismissals, shows that in addition to poor performance, the percentage of outside directors and the number of other boards on which a firm’s directors serve are key predictors of CEO dismissal: the greater the number of outside directors and director interlocks on a particular board, the greater the likelihood that the CEO will be fired if performance falters.
Changes in Board Characteristics over Time
Maximum-Likelihood Estimation of Effects on Transition Rate toward Forced CEO Turnover, 1980–96, using a log-logistic model. For specification of model, see appendix.
Even as individual boards, in their local operation, still retained much of their old identity, integration, and attachment to board cohesion and loyalty to the CEO, the broader director community had lost the autonomy and self-sufficiency of Useem’s inner circle, and the war over corporate control now knocked on the boardroom door. As we saw in chapter 3, institutional investors—once passive custodians of their clients’ accounts who tended to automatically vote their shares as management recommended—were no longer passive. And while hostile takeovers subsequently waned in the 1990s, many institutional investors did not return to their old ways. In many cases they became the ones targeting underperforming companies through coordinated action, such as that directed at General Motors in 1992 and at Digital Equipment that same year. As a Fortune magazine headline put it, the “big owners” had begun to “roar,”15 and their roar was directed at outside directors who had the formal authority to exercise control over CEO turnover and succession. In the relatively short span of a decade—and despite having exercised almost no influence on the choice of a new CEO for the better part of a century—boards came to regard CEO succession as part of their job. Today, directors are almost unanimous in their view that choosing a new CEO is among the most important decisions made by a board.16
Emboldened to exert real control over the CEO succession process, directors now try to interpret their organization’s external environment, then make succession decisions based on their reading of those outside actors whose evaluations they most prize. This reliance on external judgments stems partly from the fact that in actually choosing a CEO (rather than simply rubber-stamping the outgoing CEO’s choice), directors have no habits or settled procedures to fall back on.17 External succession is different from internal, so boards have no existing version of the CEO succession process that readily solves the problems posed by external search. As the social psychologist Leon Festinger and others have discovered, when objective or institutionalized processes for evaluating a course of action are not available, individual actions and opinions as to the rightness of those actions are strongly influenced by the opinions and evaluations of relevant others.18 In the case of CEO succession, Wall Street analysts and the business media have become the relevant others.
This way of reading and responding to the external environment has had significant consequences for CEO succession, as is illustrated by the search that resulted in Jamie Dimon’s hiring by Bank One. For another example, consider the process that led first to the appointment in October 1996 of John Walter as AT&T’s president, COO, and CEO-designate, then to his firing by the AT&T board less than nine months later. On the day that Walter’s selection was announced, telecommunications analysts described it as “a mistake,”19 the CNN anchor commenting on the selection called Walter “a puzzling choice,” and AT&T’s market value fell by $4 billion.20 As the immediate plunge in the company’s stock price suggests, this initial judgment by Wall Street analysts and the press became a self-fulfilling prophecy. Walter was described in the press as lacking credibility among the executives who worked for him, and quickly lost the confidence of the board. After less than nine months as company president, John Walter was dismissed by AT&T’s board in July 1997. He was replaced three months later by a name-brand CEO whom many influential analysts felt should have originally been selected, C. Michael Armstrong.21 Walter’s fate convincingly demonstrates how the perceptions and judgments of opinion leaders in the external environment can and do weigh more heavily in the external succession process than do the individual judgments of the directors themselves.
The influence that external actors such as analysts and the business media exert on an external CEO search offers a particularly clear example of how directors act not as autonomous individuals but rather as members of groups that are, in turn, firmly embedded in larger social networks and relations. Yet in many other ways that may not be so readily apparent, directors are strongly influenced at every step of the external CEO search process by the social dimensions of corporate boards.
Launching the CEO Search
The importance of the social factors discussed above—the cultural forces of board cohesion, the structural roles of interlocking directorships in channeling both values and information throughout the broader director community, and the institutional context within which directors operate—can be seen not just in the increased involvement of corporate boards in the CEO succession process, or in the dramatic, highly publicized firings and hirings to which that process has increasingly given rise. These factors can also be observed at work throughout the external CEO search process, including those phases of it that are generally hidden from outsiders. Indeed, right from the start of the external CEO search as typically conducted today, we can see directors organizing and carrying out their deliberations in ways that defy what we would expect if the external CEO labor market were a “market” of the kind that most other observers have described, with individuals acting rationally to advance their own interests.
ORGANIZING THE SEARCH: THE SEARCH COMMITTEE
The process of board-led CEO succession, whether internal or external, starts with the board’s appointing a search committee, which manages the administrative tasks of succession. The chair of the search committee is usually the director responsible for the executive development and compensation committee. Among its other tasks, the search committee hires the executive search firm that now almost always assists with the search.
Whether a search committee exercises significant control over the succession process is most dependent on the recent performance of the firm—as is the decision about whether to conduct a serious external search. If the firm has performed satisfactorily, the outgoing CEO is asked to participate but not allowed to take control of the process. In such cases, the appointment usually passes to the internal heir apparent, whose identity has been known well in advance. To the extent that the board conducts a search in this situation, it is fairly superficial and merely symbolic. This was recently the situation at a Northeast bank, where the outgoing CEO, who by all accounts had performed well, had made his preference for his successor clear to the board and to financial analysts. The board went through the motions of forming a search committee and conducting a “search.” One search consultant hired to bless the heir apparent admitted that there was no serious data gathering or comparison of this candidate against others, either inside or outside. The consultant summarized the process in this way:
We had a short list, some internal and external candidates, but I can’t say it was meaningful. Everyone knew that [D] was [M]’s favored chosen successor. The search committee went through the motions and I was asked to evaluate the internal candidates. . . . I went around to talk to the other executives about [D]. I asked them to provide examples of when he had been visionary, his leadership style. As you can imagine, everything was described very positively. When I asked for negative examples and feedback, people were silent. They knew the end game. . . . The board acted upon the outgoing CEO’s recommendation and appointed his successor with little reaction from either the press or Wall Street.
In some internal successions, there is a horse race between known competing candidates. This has been General Electric’s historical succession practice (as described in longtime CEO Reginald Jones’s own words in chapter 3). In the most recent instance, GE’s current CEO, Jeffrey Immelt, was explicitly pitted against two other internal candidates. While GE’s board and its search committee made an admirable show to investors and the business media of having been actively involved in the search, the decision to elect Immelt was mostly made by the retiring CEO, Jack Welch. Not surprisingly, GE’s press release on Immelt’s appointment described the decision as having been supported unanimously by the board.22
GE, of course, has been a high-performing firm for many years. When firm performance has been poor, or no obvious successor or potential successors to an outgoing CEO have emerged, the search process takes on a quite different character.
When a CEO is thought to be performing poorly (generally because firm performance itself has slipped), or when the CEO position has become unexpectedly vacant, boards come under tremendous pressure to appoint a new CEO. A firm pays a price financially and in terms of organizational image when it remains leaderless for too long. Directors know that a firm’s executives, managers, and line employees all experience a great deal of uncertainty when the CEO position is empty. Moreover, Wall Street and the business media both exert unrelenting pressure on boards to announce a new chief executive as quickly as possible. One consequence of this pressure to expeditiously announce a new CEO is that the board’s search process is oriented toward completing the search as quickly as possible.
In such circumstances, boards tend to try to adapt to an external search many of the forms and practices developed for internal succession. Because most boards have not had extensive experience with outside succession, they fall back on more familiar processes. Internal succession thus becomes the template from which the external succession process is developed. In many cases, it offers a quick and apparently viable solution with little expense, since boards are in a rush to complete the search. This is not to say that directors are not consciously aware of this copying process. As one director says, “Once you’ve done a search, you have a sense of what one should look like. This then becomes the basis upon which you do other searches.” One consequence of this approach is that, because many large companies share directors with one another, there are few major differences among companies in terms of how the search committee is organized for an external search.
One of the first questions the search committee chair faces is what the committee’s size should be. This question is often vigorously debated by the board. Some board members argue that a large committee will better represent the interests of the board as a whole. Others argue that a small committee will be more focused and efficient. Interestingly, these debates do not drastically affect the outcome. In 75 percent of the field cases I examined, the search committee had between four and seven members. The smallest committee I found consisted of three members, and the largest of eleven.
Consistent with the taboo on creating distinctions among board members, as well as with the usual practice in internal searches, the composition of the search committee is not as vigorously debated as is its size. The most common method for determining the composition of the search committee is asking for volunteers. The individuals who volunteer for the assignment are disproportionately directors who are retired executives or directors from non-business backgrounds such as nonprofits and education—essentially, directors who have more time than do their peers to devote to the search. In the field cases I examined, on average only one non-retired director served on the search committee.
One unintended result of the uniform application of the practices used for internal searches to external search—particularly the voluntary method for composing a committee—is that boards do not always pay as careful attention to the composition of the search committee as the ramifications of the search process would suggest that they should. The composition of the search committee is critical to succession outcomes. In some cases that I studied, however, search committees were composed of members who had little familiarity with the company and its history. In these cases, most of the information committee members had about the company, its products, and its other executives, was limited to the information received in quarterly meetings. Meanwhile, the predominance of retired directors on search committees impedes the effectiveness of the external search process in a particularly significant way. Because the future CEO will want some familiarity with the board members with whom he or she will work, the search committee needs individuals who are likely to stay on the board for some time. This is unlikely to be the case when most of the search committee is made up of executives who have already retired.
Search committees also need directors with diverse functional backgrounds. In several instances, the narrow range of committee members’ backgrounds bias CEO succession decisions. Consider the case of one technology company with a strong engineering culture. Although the company’s products were considered the most technologically advanced in its market, poor marketing had slashed its market share. Yet the search committee, made up largely of directors with backgrounds in operations or research and development, recommended a CEO with a strong technical background but little marketing experience. Of course, it is human nature to gravitate toward persons like oneself. Yet it is also unsurprising that the company’s market share continued to erode after the appointment of the search committee’s choice of a new CEO.
DEFINING THE POSITION: THE SPECIFICATION SHEET
Even though the composition of the typical search committee seems to belie directors’ stated belief in the importance of the job of choosing a new CEO, board members are not utterly oblivious of the needs of the firm as a factor in a CEO search. Directors, in fact, portray the CEO succession process as a unique chance for them to assess the firm’s situation. The opportunity to search for a new CEO, in the words of one director, is “a clean slate to find the person to help meet the company’s objectives.” CEO succession, according to another director, offers one of the few real opportunities for board members to assess the current state of the organization and the future that they desire for it.
Such statements notwithstanding, there is little evidence that board members engage in any extensive discussion of a firm’s objectives during the actual search process. Even in cases in which the firm has been performing poorly, or the organization needs a new strategic direction because of a changed environment, most CEO searches do not begin with an analysis of the problems facing the organization. Instead, many search committees jump immediately into the task of writing a specification sheet.
The specification sheet is a formal document that describes the requirements for the new CEO. It is usually developed with the assistance of an executive search firm (which conducts one-on-one meetings with each of the board members to get their ideas), and is considered a critical element of the search process. Al Zeien, the former CEO of Gillette and a member of several corporate boards, emphasizes the importance of the specification sheet for the search process. “From my own experience on an outside board,” Zeien observes, “the search firm [and the search] is only as good as the spec that the board agrees upon at the outset, and the spec has got to be pretty specific because that is really the basis by which the process is going to take place.”
Search consultants also stress the importance of the specification sheet; according to one, the spec sheet “lays out the required qualifications . . . the desired qualifications . . . and forces the client to prioritize and focus.” Because of the commonly held view that the specification sheet provides the guidelines against which candidates can be both identified and evaluated, executive search firms emphasize its development as a strategic component of the search process, and their particular approach to assembling a specification sheet as an important way for them to differentiate themselves from their competitors. Preparation of the specification sheet is, as one consultant put it, “the consultative component of the process, and perhaps our most value added.”
In view of these descriptions, it comes as a surprise that search committee members rarely refer back to the specification sheet once it has been drawn up and initially circulated. Instead of functioning as a set of guidelines for the search, the specification sheet assumes importance only during the process of compiling it, when it acts as a Rorschach print on which directors project their hopes, desires, fears, interpretations, and solutions. In discussions ostensibly devoted to identifying the qualities that the search committee should seek in the next CEO, some directors privately air their grievances against the outgoing CEO and blame him for the firm’s problems. Others use the process of compiling the specification sheet to engage in self-congratulatory behavior—for example, by describing their roles in ousting the previous CEO. Almost every director also uses it as a chance to discuss his or her “theories of leadership” and ideas about the qualities of a good leader. The resulting list is usually so long that, if taken seriously, it would deter many candidates by its sheer length. In compiling this list of desired qualities, moreover, directors give little attention to their relative importance.
Figure 4.1a presents a specification sheet for a CEO search at an insurance company, and figure 4.1b presents a specification sheet for a search at a major software company. Both of these lists are entirely typical of those used in external CEO searches. Three aspects of these specification sheets immediately stand out as noteworthy.
The first striking feature of these specification sheets is their emphasis on individual characteristics. The content of both lists consists largely of a collection of personal traits rather than a set of concrete skills or a discussion of the situational context of the search. The CEO should be “aggressive.” The CEO should be able to “balance reward with risk.” The CEO should be able to “provide direction.” The CEO should be able to “work closely with the team to develop a direction.” Judging from these lists of qualities, a CEO’s ability to perform effectively in the position is determined entirely by individual attributes, with no reference to the particular challenges facing the firm.
The second noteworthy feature of the two sample specification sheets is the similarity between them. A blind read would make it difficult to know that one sheet comes from a software firm and the other from an insurance company. Part of this similarity stems from the fact that each of the characteristics listed is so vague—traits such as “proven leader” and “motivator” apparently do not need to be defined. Yet such descriptions are, indeed, in many ways irreducible to more particular requirements because their rightness appears to be grounded in nature and reason. After all, who is against leadership and motivation? The individual characteristics that search committees seek in a CEO represent society’s idealized definitions of leadership writ small. They are congruent with prevailing notions—those found in the business press, for example—of what a CEO should be like. Thus, this laundry list of characteristics sheds light on how societal beliefs permeate the CEO search process.
FIGURE 4.1a Specification Sheet for Insurance Company CEO Search.
The third remarkable feature of these specification sheets is the ambiguous and contradictory nature of many of their requirements when considered in relation to one another. For example, the board of the insurance company (fig. 4.1a) was looking for a CEO who would both “establish challenging goals” and “safely” deliver a return on equity of “between 13 and 15 percent.” This individual was expected to “concentrate on increasing . . . stockholder value,” but to make only such changes as would be “appropriate with risk.” The incoming candidate was to be at the “forefront of insurance issues,” but “accommodations [could] be made in industry knowledge for those candidates outside the industry.” Looking at either of the sample specification sheets, one can only conclude that no one person could fulfill all of the requirements on either one of them. How is it possible, for example, to be team oriented and a consensus builder while being directive? To be a risk taker without taking too many risks? To be flexible and yet stand up for one’s beliefs? This would be a no-win situation for candidates and board members alike—if anyone ended up taking the specification sheet the least bit seriously.
FIGURE 4.1b Specification Sheet for Software Company CEO Search.
How to account for these anomalies and deficiencies in the specification sheet? A knee-jerk explanation might be that the people conducting the search have not worked hard enough, or been given the appropriate incentives, to pinpoint the precise traits or skills necessary to the organization’s near-term success. Yet such an interpretation ignores the context formed by the contemporary view of the CEO. Because it is assumed that the quality of the CEO determines firm performance, while—in the absence of any demonstrable link between the two—it is difficult, if not impossible, to know ex ante what characteristics in a CEO are needed to improve performance, directors are left to guess about which criteria are likely to be associated with success. Consequently, they resort to using the difficult-to-define buzzwords that we see on specification sheets—terms such as “leadership,” “team builder,” and “integrity.”
Moreover, even if directors could more precisely define the qualities they were seeking in a new CEO, the search process would continue to exist in tension with the information problems associated with external CEO search. Selecting a CEO is an information-intensive decision. The information required for this task is also highly particular and fine grained. That is, it usually can be derived only through direct experience in working with and observing the candidate. Some examples of particular information include candidates’ dispositions, working styles, and mannerisms; how they work with others; and, most importantly, whether they are really capable of doing what they say they can do. For internal candidates, this information is more easily gathered than for external ones, since the board and the sitting CEO have better knowledge and experience of the individuals under consideration. Moreover, a detailed record of an insider candidate’s promotion history, performance reviews, and accomplishments can be found in the firm’s own personnel records.
For outside candidates, however, this kind of particular information is not so easily gathered. Because of the requirements of confidentiality inherent in CEO search, it is not possible for directors to interview peers or subordinates for particular information about the candidates. Instead, particular information needs to be ascertained and assembled through more informal means. Before directors undertake the difficult search for such information, however, they narrow the pool of candidates on the basis of the much less fine-grained information available from the start. With careful consideration to how both internal and external—but especially external—constituents will perceive the search, directors use structurally and culturally derived attributes to create a pool of candidates that these constituents will see as legitimate and defensible.
Creating the Candidate Pool: The Social Matching Process
For boards to consider the interests of outside parties when selecting a new CEO is, of course, potentially desirable. After all, both a company’s shareholders and the larger community for which the press is writing will be affected in some way by the decision. Yet the focus on external constituents does have some important unintended consequences. Most significantly, it has led to an emphasis on the “acceptability” of a candidate at the expense of considering whether an individual possesses the skills necessary for the position. As a result of this emphasis, search committees focus on a fairly narrow pool of candidates and evaluate them by means of a very conservative decision process. These features of the search process, in turn, are evident in the final choices that directors make when selecting a new CEO.
While the CEO succession process appears, at first glance, to offer limitless opportunities for invigorating an organization or introducing fundamental change, many search committees express frustration with the limited pool of candidates from which they end up choosing, often lamenting that candidates are seemingly indistinguishable from one another and even from the predecessor CEO. The reason for this outcome can be found in the process that search committees use to identify and qualify the candidate pool.
IDENTIFYING THE CANDIDATES
Following the compilation of a specification sheet, directors, often with the assistance of an executive search firm, begin putting together an initial list of candidates. Board members themselves, based on their knowledge of the industry and of the profiles of CEO talent at other firms, suggest the majority of these candidates. As Michael Useem and others have found, just as the leading internal candidates for a vacant CEO position are already board members at other companies, those considered plausible external candidates are also already board members.23 While executive search firms will sometimes add one or two names, their contributions to this phase of the process are negligible.
One director of several public firms, William Pounds (the former dean of MIT’s Sloan School of Management), provides a general description of the process for compiling the initial list of candidates:
The first hurdle for myself is the laugh test: if we actually named this guy and told the employees and shareholders that he was the new boss, what would they think? Then there is the tendency to find people who look like the job. You start by whacking down the job to a set of alternatives. So you consider things like performance . . . the company they are coming from, who they have worked with. From there you start to find people who look like the position.
Consider the entirely typical case of a search at a publishing firm. Here, the search committee came up with a list of forty names, thirty of which came from the directors themselves, who were asked by the search firm to list potential candidates. The search firm added ten names to the list, including those of four women or members of minority groups. Within a few days, with the help of the search firm, the search committee narrowed the initial list down to twenty candidates, almost all of them actively employed. The search firm then gave the search committee a booklet containing a photograph and a short profile of each of these candidates. The profiles included only general information on the candidates’ work histories and educational credentials, as well as estimates of their most recent compensation. One director described the booklet as a “Who’s Who of global publishing that included high-profile CEOs of other prominent publishing firms and promising top executives from related industries such as entertainment.”
The search firm then began to contact candidates listed in the book. Without revealing the identity of their client, the search consultants sought to identify candidates who would seriously consider moving from their existing positions. When the consultants came back with a short list of ten candidates, the board narrowed the list even further. Throughout this process, the board was guided by three criteria for winnowing the candidate pool: the current position of the candidate, the performance of the candidate’s current firm, and the stature of that firm.
These criteria form the critical matter of the CEO labor market. At one level, as we shall see, they represent apparently reasonable responses to the elementary conditions of the market outlined in chapter 2. The relevance of these criteria—prior position, firm performance, and firm status—seems so self-evident, so reasonable, that directors rarely question them or feel the need to justify using them.
At the same time, however, these criteria form the basis of a durable sorting process that results in a narrow definition of the candidate pool, and a distinct privileging of one set of candidates over another. Examined with a critical eye, the standards by which candidates are sorted turn out not to be aligned with the technical or efficiency goals of the CEO search. Rather, they are essentially extraneous considerations that cause search committees to shine a spotlight on certain candidates, thereby leaving others in the shadows. The mechanism governing this sorting process is one that can best be described, using James C. March and James G. March’s term, as “social matching.”24 Social matching is a filtering process that takes place when individual and organizational actors are confronted with choices that are difficult to select among because of limited information, and because the choices themselves cannot be reliably distinguished from one another. At a minimum, social matching is a convention. Its impact, however, is not benign. Social matching is one of the basic mechanisms that contribute to a closed CEO labor market, because directors seize upon one or more easily identifiable external characteristics of potential candidates to construct a narrowly defined candidate pool. By relying on social matching to identify the plausible candidate pool, directors are caught in an unforeseen trap of focusing on a small set of candidates and unwittingly undermining the original intent of a broad search.
Directors’ tendencies to fasten on particular characteristics are also conditioned by what the anthropologist Mary Douglas calls “habits of thought.”25 Habits of thought are organizing and classificatory principles that appear to be endowed with the character of common sense, theories that seem self-evident. In the case of external CEO search, the external environment furnishes many, if not all, of the habits of thought on which directors rely in what is, at its core, a thoroughly social process. Directors’ socially derived ideas about the criteria necessary to be included in the candidate pool then actively influence both directors’ perceptions of candidates and the composition of the pool. For example, as a result of limiting the pool to candidates from organizations that have been performing well, many firms pass over talented executives from underperforming firms—all because their directors accept the widespread idea that organizational performance is a direct result of the ability and efforts of the CEO. Other circumstances (such as market conditions or the skills and experience of a company’s senior management team) that are known to have a profound impact on corporate performance are not considered in sorting through the candidates. Thus, reliance on the performance of the firm from which the candidate is coming for evaluating a candidate’s record—although at one level a perfectly defensible response—is actually problematic.
Let us examine each of the three criteria that directors employ to sort external CEO candidates at greater length.
POSITION MATCHING: ALREADY A CEO OR PRESIDENT
Many search committees feel that a potential CEO candidate needs to have previous experience leading a large corporation. Fully 75 percent of outsider CEO appointees during the period between 1985 and 2000 were either CEOs or presidents in their previous jobs. According to one director, a candidate’s having already been in the CEO position or its equivalent “gives you a sense of how they are likely to act. It reduces uncertainty.” Directors point out that the fact that an individual is already a CEO indicates that another board has already rendered a favorable judgment about that person. From the directors’ point of view, coding a candidate on the basis of his previous position saves effort and energy. As Gigi Michelson, a longtime director at GE and other large firms, puts it, a board would not appoint someone to a CEO position if it “did not already have a judgment about the person, their personal stability, some knowledge of [their] family life and outside background, outside activity, outside of the corporation and so on.” In using position matching to sort external CEO candidates, directors effectively trick themselves into thinking that already occupying the CEO (or president) position is a reliable signal of leadership abilities.
Directors’ reliance on position matching in identifying candidates also makes it easier for the board to eventually defend its chosen candidate to the outside world. By choosing a candidate who has already had experience as a CEO or president, the board conforms to expectations concerning the traits or characteristics a legitimate candidate will need. Board members thus externalize what should be their own responsibility for choosing among available candidates.
Of course all of the expectations to which directors want candidates to conform create advantages for certain classes of candidates and disadvantages for others, which makes focusing on candidates with prior CEO experience—particularly at large companies—a potential liability. This is a problem that directors seem to be aware of at some level. While no two CEOs are the same, and each will have a different personality and style of leadership, such differences are often minor. And although directors do not make a link between their selection process and the uniformity of candidates that they observe, they do often comment on this uniformity and describe it as a problem. Several directors I interviewed complained about the shortage of “real leadership” and the “cookie-cutter” character of several of the CEO candidates whom they had considered. “A lot of the candidates are corporate survivors, not leaders” corporate director George Kennedy said of a search he had led at the industrial conglomerate Brunswick. Similarly, Henry Wendt, former CEO of SmithKline Beecham and a director on other companies’ boards, commented:
We’re in the most severe shortage of CEO talent in corporate history. Most of the people I see are not leaders, they are managers who know how to work the system and have worked it well. Times are different. We need leaders, not politicians. We need to find people who can define what the organization is about. Charismatic people who help people understand what they need to do. Executors who know how to get things done. People who integrate leadership into their everyday activities. It is really difficult to find a real leader based on my experience in searching for CEOs.
Such observations are not surprising, given the candidate pool that directors themselves typically create. As C. Wright Mills noted years ago, in an observation that still applies today, “The top executives of the big companies are not, and never have been, a miscellaneous collection of Americans; they are a quite uniform social type which has had exceptional advantages of origin and training, and they do not fit many of the stereotypes that prevail about them.”26 Whatever their origins, CEOs of large corporations settle into the executive suite after a selection process lasting years. Socialization in the executive suite and self-selection en route to it significantly reduce the variability of the types of candidates who are seriously considered for vacancies in the CEO job. Those at the top of an organization are all corporate survivors. Of course, the skills possessed by all such survivors—specifically, their ability to comprehend and navigate a complex corporate bureaucracy—were once considered desirable in a CEO, inasmuch as the CEO’s job, before the rise of investor capitalism, was thought to require understanding how complex organizations function and being able to work with the people in them. Yet directors shun such qualifications now that they are in search of the elusive personal characteristic called “leadership.”
Directors’ categorizing of candidates according to whether they have already been CEOs also highlights the fact that the social matching process that directors use is not so much one of consciously excluding candidates as it is one of creating expectations that can be fulfilled by only a small number of them. This has the consequence not only of narrowing the candidate pool but also of narrowing the directors’ own thought processes. In their social matching process, directors pick and choose among the criteria that they believe will give due weight to the concerns of external parties. Yet by selecting candidates on the basis of socially legitimated classifications, such as already having occupied a position of corporate leadership, they are at the same time establishing a pattern of thought and action that is adhered to in the rest of the selection process. That is, these classifications, though grounded in pragmatic reasoning, also straitjacket the minds of directors—the more so, precisely because they do not appear to be unreasonably narrow. The convention of choosing an individual who has already occupied the position of CEO overcomes any tendency that individual board members might have to select candidates who deviate from the externally derived orthodoxy. Board members thus substitute a social convention for deliberative thinking and decision-making.
PERFORMANCE SORTING: COMING FROM A HIGH-PERFORMING COMPANY
In almost every external CEO search I studied, candidates were categorized according to the performance of their current firms. In the minds of directors, sorting candidates along the criterion of firm performance is a reasonable means of gauging the quality of a potential CEO. In the case of Hewlett-Packard’s most recent CEO search, for example, several H-P directors described choosing outsider Carly Fiorina over several internal candidates because they associated Fiorina with the strong performance of Lucent Technologies, her previous firm. Lucent’s stock had indeed performed extraordinarily well since the company had been spun off from AT&T. Yet while the H-P directors had attributed Lucent’s success to its executives, much of this “success,” it was later revealed, was due to creative accounting and liberal financing of sales to customers. Another relevant factor that these H-P directors seemed to ignore was that investors had bid up almost every technology-related stock in 1999.27
Yet despite such considerations, directors everywhere admire CEO candidates whose firms have performed at a level above the average for their industry. In choosing among external CEO candidates, directors attach great value to observable performance metrics in making their selection decisions. Consequently, firms choosing a CEO from outside usually select candidates from firms that are performing well relative to their industry. An examination of all outsider CEO appointments in the S&P 500 companies during the period I have studied finds that at the time of their appointments, 70 percent of outsider CEOs are moving from companies that are in the top quartile of performance in their industries.
What is striking about this performance sorting is the casual assumption that CEOs directly affect organizational performance—an assumption that, as we saw in chapter 2, has found little empirical support in scholarly studies on the topic. Directors, however, rarely cast a critical eye on the supposed connection between firm performance and executive quality, and dismiss as inherently flawed any research that questions it. One director describes any such research as “inherently unscientific, since you can never run the controlled experiment of what the firm would have been like without the CEO,” and dismisses statistical techniques—as well as studies that match comparable firms to one another to gauge the impact of an individual CEO on performance—as “academic mumbo-jumbo, with no regard for reality.” It is as if when directors don’t like the empirical facts, they simply create their own narratives (admittedly a common enough human trait). In such a milieu, any theory or causal relationship that one cares to posit is justifiable. Meanwhile, there is little evidence of directors’ trying to deconstruct the performance of a candidate’s organization into its constituent parts, to determine what role the candidate may have played and what might be attributable to factors such as the environment, the quality of the company’s employees, the competitive structure of its industry, or even luck.
Even granting, for the sake of argument, the possible validity of the contention that the research showing little or no connection between firm performance and executive quality is inherently flawed, a question arises as to why directors so resist any suggestion that the widely credited relationship between the two might be spurious. It is difficult to convey to the reader how deeply rooted this belief in the dependent relationship between CEO quality and firm performance is among members of corporate boards, who hold it with virtually religious conviction. To openly question it is taboo.28 One reason it may be difficult for directors to talk objectively about the relationship between CEOs and firm performance is that it touches on fundamental, deeply held beliefs about, for example, corporate meritocracy and the reasons for their own ascensions to the top rungs of the corporate hierarchy. Directors have internalized the link between CEOs and performance to the point of sacralizing it. Every director defends the logic. Thus, directors’ existing theories about the relationship between CEO quality and firm performance pre-condition their responses to a challenge and even determine what can be counted as a reasonable question. Their ideas place firm limits on their ability to make objective judgments about the posited relationship.
Another reason it is difficult to question the belief in the relationship between CEOs and corporate performance is because it enjoys such broad external support and legitimacy—especially among constituencies such as the business media and analysts, who, as we saw in chapter 3, often reduce firm performance to the qualities of a single individual. The sociologist Robert Merton has written about the power that a social fact can have over human behavior.29 Once a complete and self-reinforcing system of opinions has been formed in a society, it offers strong resistance to anything that contradicts it.30 Thus the external institutions of the business media and analysts systematically reinforce the individual perceptions of directors. They fix and routinize dynamic thought processes, making them part of an invisible social structure. They hide their influence and rouse emotions to a heightened pitch when these perceptions are questioned.
As in the case of position matching, the belief in the link between CEO quality and firm performance is not without its substantive consequences. Without directors’ explicitly realizing it, this emphasis on the performance of the firms with which CEO candidates are linked sometimes leads to the inclusion of otherwise mediocre candidates, and to the exclusion of candidates who may actually have performed admirably under challenging circumstances. By connecting firm performance to CEO quality as a sorting mechanism for eliminating potential candidates, directors create a system for segmenting people on the basis of a measure that from an empirical standpoint is inherently flawed.
STATUS MATCHING: COMING FROM A FIRM OF SIMILAR OR HIGHER REPUTE
Prestige is particularly important in CEO searches, and search committees see the status of the firm from which a CEO candidate comes as amplifying or reducing the status of their own organizations. They also see the status of a firm as a property of individual executives who have worked for it. One director summarizes the effects of this outlook on external CEO searches:
While the search process is becoming professionalized, the way things work at this level is a lot on the basis of appearances. The process is basically driven by the board, and most boards are looking for either one or two things in a CEO—peers, which suggests executives who are similar to them, or more higher [sic] profile people who bring prestige and stature to the company simply because of who they are and their obvious accomplishments. So by and large, this is a biased set of selection criteria.
Faced with the need to defend their firms’ current positions in the corporate status hierarchy, or desiring to move up the ladder, boards recruit CEOs from organizations with status comparable to, or higher than, that of their own companies. In the external CEO searches I studied, 80 percent of firms recruited their new CEOs from such companies.31
Directors’ concern with the status of the firm has parallels with concerns about status in most areas of society. Status, as the economist Thorstein Veblen noted, can be used to signal quality, wealth, and, by inference, power.32 In sociology, Joel Podolny’s research has strongly underscored the importance of status in markets.33 Podolny argues that under conditions of uncertainty, organizations use their affiliations with other economic actors as signals of their own quality and worth. This reflected legitimacy or status may, in turn, have a number of positive economic benefits for the actor, ranging from survival to organizational growth to profitability. In the case of an organization whose future is uncertain, appointing a CEO who is from a high-status firm can produce advantages for the organization such as increased confidence in its prospects.34
One of the basic insights of research on status, however, is that a tight coupling between the status of individuals or organizations and their underlying quality may not always exist, a point illustrated by the case of auto parts manufacturer Federal Mogul. The firm’s appointment of Richard Snell, formerly CEO of Tenneco Automotive, as its own CEO in November 1996 prompted a Merrill Lynch analyst to write: “Frankly, we are surprised and somewhat relieved that the company was able to secure someone of Mr. Snell’s stature.” Federal Mogul’s stock price jumped to $25 per share from $22 as a consequence of Snell’s appointment. As he had done in his previous position at Tenneco, the new CEO then went on an acquisition spree. In the process, he turned Federal Mogul from a $2 billion company into a $6 billion company, while its stock price plummeted from $70 per share in mid-1998 to just $1 in February 2001.35
Yet despite such cautionary tales, board members have their own version of Q ratings, and hiring a high-status CEO is viewed as an indication that a company belongs to a particular class of firms. One notable expression of the American corporate ethos is the tendency to form ranks and lists. Business Week and Fortune magazines, for example, continually generate lists of the “100 Best Companies to Work For” or “America’s Most Admired Companies.” Fortune’s survey of “Most Admired Companies” is typical of how such tabulations are assembled. The magazine creates its ranking through a survey of executives, directors, and financial analysts in which respondents are asked to rate companies along the criteria of: (1) quality of management; (2) quality of products and services; (3) financial soundness; (4) long-term investment value; (5) use of corporate assets; (6) innovativeness; (7) responsibility to the community and environment; and (8) ability to attract and retain talented people. Such lists, in turn—although generated, in part, by directors themselves—then form part of directors’ view of which corporations and executives are inherently the best. Meanwhile, the notion that CEOs have a status that can be transferred from one company to another has become deeply ingrained in directors’ new view of the role of the CEO. And while CEO candidates might appear to most outsiders to be similar with respect to their prior titles, education, and background—just as they appear similar to board members, as we have seen, in terms of their “leadership” abilities—from directors’ perspective, no list of candidates can be described as flat or undifferentiated along the lines of status.
Directors’ placing of a primary value on status is a gesture that is aimed, in turn, at outside audiences, who are presumed to be capable of discerning the same status differences that the directors themselves do. In a reaction against what external observers see as the uniformity of the candidates in the firm’s internal labor market, directors seize on status considerations partly as a way of giving these outsiders a choice from what to the directors (although not necessarily to anybody else) seems a varied menu. Ironically, this decision has the effect mainly of limiting the external candidate pool. Yet relying on status considerations constrains directors significantly in other ways as well. The sheen of a candidate’s status often dazzles directors and reduces the level of questioning they will direct at him or her. In the pursuit of status, directors stop trusting their own responses to particular individuals, even feigning indifference to candidates whom they secretly admire and in whom they have reliably based confidence. In the quest for a high-status CEO, directors forego the opportunity to choose a candidate who might truly represent “new blood” and be “reinvigorating” in favor of one who, but for his celebrity, might not even be considered qualified for the job.
Directors’ Connections as a Source of Information
Social matching is a relatively crude way of sorting candidates for a position such as CEO. One reason that directors resort to it—in addition to the pressure they feel from internal and external constituents to produce a defensible choice for CEO—is the relative difficulty of obtaining more useful kinds of information about CEO candidates. The means that directors do use for acquiring such information, meanwhile, also restrict their focus to a relative handful of candidates.
Fine-grained information is critical to the outsider CEO selection process. As one director of a $2.6 billion manufacturing concern states:
As a person responsible for the fiduciary interests of the company, I can tell you that going outside the firm for a CEO is a very difficult decision. . . . While people think there is always a potential person out there who can do the job, this is hard to prove. For example, going outside for the bottom line is fine, but once you start considering issues such as corporate culture and norms, etc., you recognize that you don’t want a bull in a china shop, like that guy at Sunbeam [Al Dunlap]. Instead, making that outsider decision requires knowing as much as possible about the person from the outside as you can.
To get such fine-grained information on external candidates, directors rely largely on their connections to other directors. Since a large number of CEO candidates are themselves on boards, they are often connected to other board members. These connections, in turn, make the interlocking directorate particularly well suited for transferring specific information about potential candidates. One director describes this process of information transfer in a search in which he participated:
Logit Model for Predicting Outsider CEO Succession, 1980–l96
After we had narrowed down our list of candidates to the finalists, I began to make phone calls. While we liked [David] in particular, we were worried that he would try to change the company culture rather than adapt it. We were looking for a person who would change what was not working well in our culture and keep what was [working]. I made a phone call to a director of another board I sat on who had been his former boss and began to verify my assumptions about this individual and what he thought of this individual’s abilities. I did not reveal the name of the company, but described what we were looking for. He allayed our fears and we eventually selected this person because we were able to find out in fairly accurate ways what he was like and would be like once he got here.
The reliance on director ties for particular information about candidates is an important element facilitating the creation of a CEO labor market. The importance of having access to particular information about candidates is highlighted in table 4.3. This table shows that those boards that have a high level of connectivity to other boards through prominent directors are more likely to appoint outsiders than those boards that do not have such connections. This is true independently of whether the outgoing CEO was fired.
The irony inherent in directors’ reliance on social ties and in their use of the social matching process in external CEO searches is that even when directors have gone outside the firm to get away from the restrictiveness of the internal market, they often end up duplicating a similar structure in the external CEO selection process. Social matching, in particular, creates a contradiction between the ideal of opening the CEO labor market beyond the relatively narrow confines of the firm and the reality of external searches’ focus on a narrow set of candidates. One might say that the external succession process runs in the rut of social matching. The criteria used in the directors’ social matching process are harmful not least because they are substituted for others that might measure such truly individual characteristics as ability or industriousness.
The criteria employed in the social matching process that characterizes external CEO succession not only shed light on the closed nature of the CEO labor market but also account, in particular, for the durable homogeneity we observe in the nation’s CEO suites despite dramatic changes in the composition of the work force.36 When directors use these criteria to identify a legitimate set of candidates, they are at the same time maintaining and reproducing a social structure. Through their concern with maintaining group cohesion, their connections with other directors that channel information and values throughout the broader director community, and their attention to the external institutional context in which they operate, directors develop beliefs and social codes that focus their attention on a homogenous set of candidates whom they believe share these beliefs and codes. One of the most perceptive directors I spoke with understood this process. “Do you miss something by not having different criteria or a broader cross-section?” he asks. “Yes, you probably do. You should understand that I am not making any value judgments as to whether the system necessarily is right or wrong or the way it should be. This is the way that it is.”
The outcome to which social matching in CEO searches leads is not intentional, but it is locked into external CEO search by the social matching process itself. And because this process is now so routinized in external search—that is, it is taken for granted, rarely questioned, and close to universal—it reproduces a strikingly similar set of candidates in one search after another. Although directors play the most active role in this self-limiting, self-replicating selection process, they are not the only actors in external CEO search to play this particular part. To see how boards of directors receive critical reinforcement in keeping external CEO search focused mostly on a small group of the usual suspects, we need to examine the little understood role of the executive search firm.