THE ROLE OF EXECUTIVE CAPITAL AND THE MARKET FOR ALTERNATIVE CANDIDATES IN CEO DISMISSAL AND LABOR MARKET CONSEQUENCES FOR DISMISSED CHIEF EXECUTIVES BY Copyright 2012 Donald J. Schepker Submitted to the graduate degree program in Business and the Graduate Faculty of the University of Kansas in partial fulfillment of the requirements for the degree of Doctor of Philosophy ______________________________ Chairperson: Vince Barker Committee Members* _____________________________* Laura Poppo _____________________________* Jay Lee _____________________________* Clint Chadwick _____________________________* Todd Little Date defended: May 29, 2012 ii The Dissertation Committee for Donald J. Schepker certifies that this is the approved version of the following dissertation: THE ROLE OF EXECUTIVE CAPITAL AND THE MARKET FOR ALTERNATIVE CANDIDATES IN CEO DISMISSAL AND LABOR MARKET CONSEQUENCES FOR DISMISSED CHIEF EXECUTIVES ______________________________ Chairperson: Vince Barker Date approved:_______________________ iii ABSTRACT Donald J. Schepker Research on the dismissal of Chief Executive Officers has primarily examined how firm performance and executive power affect dismissal. However, the process used to evaluate a CEO’s capabilities is complex, as a myriad of factors affect firm performance outside of the CEO’s control and the board often has minimal interaction with the CEO. Instead, the board may be forced to examine external cues or signals that help provide information regarding the CEO’s capabilities. Analyzing 3,648 firm-year observations for likelihood of dismissal, this dissertation examines the role that CEO human and reputational capital play with regard to signaling the board regarding the CEO’s capabilities as well as the effects of the market for alternative CEO candidates on the likelihood of CEO dismissal. Findings from probit regression analysis indicate that CEOs are less likely to be dismissed when they have greater tenure, a greater base salary, a less negative reputation in the media, and when there are fewer non-CEO inside directors serving on the board. These results suggest that the board identifies some external cues when evaluating CEOs and evaluates visible internal candidates in the decision to dismiss a CEO. Building upon this line of research, the second chapter of this dissertation examines the career consequences of dismissal on the future job prospects of executives. I argue that dismissal serves as a stigma on executive careers which reduces job future prospects. However, executives may use human, reputational, and social capital to buffer themselves from the effects of stigmatization. Examining the re-employment prospects of 88 dismissed executives, results using Cox Proportional Hazards models indicate that executive job prospects at publicly traded organizations are lessened following dismissal for reasons of violation of fiduciary duty or personal conduct. Alternatively, executive re-employment is more likely when iv executives have experience with prestigious organizations, have a reputation for being a top CEO, have less negative publicity, and are located in a major city. These results suggest that while dismissal may be stigmatizing, such effects can be overcome with acquired human, reputational, and social capital. v Acknowledgements I would like to thank my advisor, Vince Barker, for his mentoring and advising throughout my time at the University of Kansas. I could not have completed this program without his invaluable advice throughout. Vince was always willing to provide counsel at any time about any subject, and it was his help and advice that helped me through the PhD process. However, what I will always remember most about Vince was that he was not only a mentor, but also a true friend. For that, I will always be grateful. I would also like to thank Laura Poppo, who has also provided me with advice and counsel throughout this process. I truly appreciate Laura’s willingness to work with me on a number of projects and her willingness to help me achieve what I have been able to achieve. Additionally, I would like to thank the other members of my dissertation committee: Jay Lee, Clint Chadwick, and Todd Little. Without your help, I could not have completed this dissertation, or the program. Your advice and friendship have meant a great deal to me throughout my time in Lawrence, and I appreciate your eagerness at all times to help me through this process. Another person who has helped significantly is our PhD program Administrative Director, Charly Edmonds. Charly provides advice, guidance, and counsel in the friendliest manner possible at all times. Like Vince, Charly has been a true friend as I have completed this program. It is thanks to people like Charly that this program has been such a blessing to me. While the faculty have helped me to achieve many things throughout, I could not have done this without the support and friendship of my fellow students. First, I would like to thank my fellow Strategic Management colleagues for their advice, friendship, support, and willingness to help in this process, including Alex Martynov, Won Oh, Dave Wangrow, Karl Kammerer, and vi Zheng Cheng. Additionally, I would like to thank the many friends I have made in the PhD program, many too numerous to name. In particular, I want to say thank you to Jake Messersmith, Shane Moser, Ze Wang, Yeong-Yeon Ji, Mike Ellis, Matt Luth, Carol Flinchbaugh, and Elizabeth Emeigh for your wonderful guidance, support, advice, and friendship. Finally, I could not have completed this program without the support of an amazing group of family and friends. To my family, especially my parents, who supported me on this crazy journey, I will never be able to repay the love you have given to me. But most importantly, I would like to thank my wonderful loving wife. Without your support at all times as I completed this process, I could not be where I am today. Thank you for everything you have done for me. I can never put into words how much your love and support mean to me. To everyone who made this happen, thank you from the bottom of my heart. 1 Table of Contents Study 1: The Role of Human Capital, Reputation, and the Market for Alternative Candidates in the Dismissal of Chief Executives......................................................... 4 Introduction........................................................................................................ 4 Past Research on CEO Dismissal...................................................................... 7 Antecedents of CEO Dismissal................................................................ 8 Reasons for CEO Dismissal..................................................................... 12 Uncertainty in Assessment of Executive Performance............................ 14 Signals of Executive Quality and CEO Dismissal............................................ 20 Human Capital and CEO Dismissal.......................................................... 20 Firm Tenure.................................................................................. 22 Prior Experience in Prestigious Firms......................................... 23 Prior CEO Experience.................................................................. 24 CEO Education Level................................................................... 26 CEO Elite Education.................................................................... 26 CEO Compensation...................................................................... 27 CEO Reputation and Executive Dismissal................................................ 29 The Managerial Labor Market and CEO Dismisal......................................... 32 Internal Candidates for CEO Replacement............................................... 33 Presence of a COO....................................................................... 33 Non-CEO Inside Directors............................................................ 34 Number of Firm Divisions / Business Lines.................................. 35 Presence of Interim CEO Candidates........................................... 36 External Candidates for CEO Replacement.............................................. 38 Industry Concentration.................................................................. 38 Prestigious Firms in an Industry................................................... 39 Board Memberships of Firm Directors......................................... 40 Firm Geographical Location......................................................... 41 Firm Performance and Executive Dismissal...................................................... 42 The Interaction of Performance and Executive Dismissal.......................... 42 The Interaction of Performance and the Managerial Labor Market........... 45 Methods................................................................................................................. 47 Sample........................................................................................................ 47 CEO Dismissals.......................................................................................... 47 Independent Variables................................................................................ 49 Executive Capital............................................................................ 49 Market for Alternative CEO Candidates........................................ 51 Firm Performance.......................................................................... 52 Control Variables........................................................................................ 53 2 Analysis...................................................................................................... 54 Results.................................................................................................................... 54 Discussion............................................................................................................... 61 Future Research........................................................................................... 67 Limitations................................................................................................... 68 Conclusion................................................................................................... 69 Study 2: What Happens to Dismissed Executives: The Role of Human, Reputational, and Social Capital in CEO Re-employment............................................................................ 70 Introduction............................................................................................................ 70 Executive Dismissal and Stigmatization............................................................... 73 The Effects of Dismissal Circumstances on Executive Re-Employment as a Top Manager.................................................................................................................. 76 Performance Related Dismissal................................................................... 77 Violations of Fiduciary Duty....................................................................... 79 Personal Conduct Violations........................................................................ 80 Degrees of Stigmatization based on Dismissal Circumstances................... 81 Executive Capital as a Buffer from Stigma’s Effects......................................... 83 Human Capital and Executive Stigma......................................................... 83 Experience with Prestigious Organizations..................................... 84 Executive Education Level............................................................... 85 Prestige of the Executive’s Education............................................. 85 Reputational Capital and Executive Stigma................................................ 87 Social Capital and Executive Stigma........................................................... 89 Number of Directorships.................................................................. 90 Number of Executive Positions........................................................ 91 Geographical Location.................................................................... 91 The Interaction of Dismissal Circumstances and Executive Capital................ 92 Performance Related Dismissal and Executive Capital............................... 93 Violations of Fiduciary Duty and Executive Capital................................... 95 Personal Conduct Violations and Executive Capital.................................... 98 Methods.................................................................................................................. 100 Sample......................................................................................................... 100 Dependent Variable..................................................................................... 102 Independent Variables................................................................................. 103 Circumstances of Dismissal............................................................. 103 Human Capital................................................................................. 104 3 Reputational Capital........................................................................ 104 Social Capital................................................................................... 105 Control Variables.......................................................................................... 106 Analysis........................................................................................................ 107 Results...................................................................................................................... 109 Sensitivity Analyses...................................................................................... 117 Discussion................................................................................................................. 122 Implications for Managers and Future Research........................................... 127 Limitations..................................................................................................... 129 Conclusion..................................................................................................... 130 References............................................................................................................................ 131 Figures.................................................................................................................................. 150 Tables.................................................................................................................................... 169 Appendix A – Sample Selection Model: Antecedents of Dismissal for Study 2............. 199 4 STUDY 1: THE ROLE OF HUMAN CAPITAL, REPUTATION, AND THE MARKET FOR ALTERNATIVE CANDIDATES IN THE DISMISSAL OF CHIEF EXECUTIVES Introduction Turnover of chief executive officers in organizations has long been studied, given the importance of top managers in firm strategic decision making (Fredrickson, Hambrick, & Baumrin, 1988). While CEO turnover may occur for a variety of reasons (e.g. retirement, resignation for personal reasons), researchers have been intrigued by executive dismissal; that is, instances where the organization chooses to replace the CEO with another individual. The decision to replace a CEO is one of the most important made by a firm’s board of directors (Huson, Parrino, & Starks, 2001). Despite this importance, relatively few researchers have attempted to model the antecedents of executive dismissal (Hatfield, Worrell, Davidson III, & Bland, 1999) and the vast majority of work on dismissal has centered around the role that performance and power play in the process (e.g., Boeker, 1992; Hatfield et al., 1999; Huson et al., 2001; Shen & Cannella, 2002). A number of prior studies have examined the question of why are CEOs dismissed. Research shows dismissal of chief executives may occur due to sociopsychological dynamics within the top management team, ritual scapegoating of executives for poor performance, or in an attempt to adapt an organization to changes in its environment (Shen & Cho, 2005). Additionally, performance alone only explains approximately 20 percent of variance in executive turnover (Finkelstein, Hambrick, & Cannella, 2009). However, more recent research has examined the opposite effect: why aren’t some CEOs fired. Consistent with these findings, research has noted that ambiguity exists with respect to how much a CEO contributes to firm performance, which makes evaluation of the CEO difficult (Lieberson & O'Connor, 1972; 5 Pfeffer, 1977). This ambiguity creates difficulty in evaluating managerial ability as organizational performance is affected by a myriad of factors, including factors at the organizational and industry level (Holmstrom, 1982). In other words, poor performance may not be the result of a poor CEO, while good performance may be attributed to factors other than the CEO. Compounding the problem, the board also has limited interaction with the CEO (Carter & Lorsch, 2004). Board members rarely, if ever, directly observe and monitor the CEO’s performance on the job. Instead, the board must evaluate the CEO’s capabilities based on the few meetings held per year and other, albeit extremely limited, interactions. This lack of interaction only contributes further to the difficulty of evaluating the CEO. This study seeks to add to the growing literature examining why boards choose not to dismiss some CEOs, even when scrutiny over firm performance may put pressure on the board to dismiss the CEO. Given the difficulty of CEO evaluation, the board may examine factors that serve as signals of an executive’s quality in addition to the firm’s performance. When capabilities are uncertain and difficult to assess, it is more likely that evaluators will seek out third party signals to help in the evaluation process (Graffin & Ward, 2010; Podolny, 2005; Rindova, Williamson, Petkova, & Sever, 2005). Such external signals may assist evaluators in corroborating their beliefs regarding capabilities. This study argues that a CEO’s human and reputational capital can provide a quality signal to the board of directors regarding the CEO’s capabilities, independent of firm performance, that reduces the likelihood of CEO dismissal. CEOs with high levels of reputational capital have been endorsed by reputable third parties as having a certain level of quality in the position of CEO. Under conditions of poor performance, these effects are expected to be further attenuated, as human and reputational capital may buffer an executive from dismissal, as such signals can help the board more positively evaluate a CEO even in the 6 face of poor performance. When a CEO has significant human and reputational capital, boards should be less likely to dismiss the CEO, even in the face of poor performance, as there is a signal that the executive may be more likely than an alternative candidate with a lesser reputation nor lesser accumulated human capital to lead the firm to success in the future. A second and critical aspect of the decision whether to dismiss the CEO is whether alternative candidates exist and are available that could achieve superior performance beyond the current CEO for the firm. Boards must assess not only the current CEO’s capabilities, but the likelihood of improving performance if a new CEO were to be hired. CEO dismissals are often extremely disruptive and costly (Wiersema, 2002). Due to the costs associated with dismissal, firms should be wary of dismissing a CEO if an executive of higher quality cannot be acquired to replace the CEO (Fredrickson et al., 1988). Alternatively, the board has a duty to dismiss even an adequate CEO if a higher quality candidate can be hired. Thus, the board should assess the quality of both internal and external candidates that are available in the labor market for CEOs as part of the decision of whether to dismiss a current CEO. Dismissing a CEO without a higher quality replacement available may only worsen firm performance due to the executive’s lack of capabilities or due to disruption associated with the CEO dismissal process (Wiersema, 2002). Thus, this study argues that the availability and the board’s perceived quality of potential replacements for the CEO position will also influence the board’s decision of whether to dismiss the CEO. This relationship is expected to be stronger under conditions of poor firm performance, as the board seeks solutions in order to achieve future aspiration levels. This study makes several important contributions to the literature on CEO dismissal. First, this study builds upon the research examining why some CEOs are not dismissed and identifies that the qualities and capabilities of the executive may alter the likelihood that a CEO 7 will be dismissed, such that executives with higher levels of human and reputational capital will be less likely to be dismissed. While past research has focused on the power executives have due to ownership or co-optation of a firm’s board, this study argues that influence can be wielded in a different, more subtle manner. Second, this study argues that a board’s situational assessment of managerial performance may be altered based on the stock of human and reputational capital that the executive has built up, which may signal greater or lesser capabilities. This is especially important when performance is poor, as these signals may lead the firm to retain a CEO even under poor performance. Third, this study examines the role that the market for alternative CEO candidates can have on a firm’s decision to dismiss a CEO. A board does not examine the CEO in a vacuum, but instead must determine whether a replacement candidate can achieve better performance than the current CEO. Taken together, these evaluations can help the board determine whether dismissing a current CEO is a move that strategically positions the organization for better future success. The arguments in this study indicate that replacing a CEO may occur regardless of performance if the firm believes an alternative candidate can achieve superior returns. Finally, prior research argues and often assumes that dismissal does not occur when performance is poor due to weak governance. However, this study builds a theoretical model that argues that the board utilizes and evaluates information that provides signals regarding executive quality outside of performance that rationally leads the board to retain a CEO, even when performance may indicate dismissal is warranted. PAST RESEARCH ON CEO DISMISSAL The decision to replace a firm’s CEO is one of the most critical decisions a board has to undertake (Huson et al., 2001), as top executive dismissals may lead to wide-spread changes in personnel and strategic direction. Furthermore, not all successions are equally interesting from a 8 theoretical perspective; dismissal itself represents the most theoretically interesting form of executive succession, as the reasons for the executive’s removal are not always clear and extend beyond just merely responding to poor performance (Fredrickson et al., 1988). Friedman and Singh (1989) noted that the removal of top managers, especially CEOs, is a unique form of turnover whose antecedents are different from those of other types of separation, including death (Worrell, Davidson III, Chandy, & Garrison, 1986) and mandatory retirement. Dismissal represents a decision not undertaken by the executive him or herself, but involve group dynamics and decision making processes undertaken by the firm’s board of directors. Forced CEO turnover has attracted growing attention in recent years as the process has been identified as a social and political process involving power struggles among executives, boards, shareholders, and other stakeholders with a vested interest in the firm. In recent years, the incidence of CEO dismissal has increased greatly (Huson et al., 2001). In a sample of firms from 1965-1974, Herman (1981) found only 20 firings in the largest 200 nonfinancial firms in the United States. However, in a study spanning more than 30 years, Huson, Parrino, and Starks (2001) find that the frequency of forced turnover increased from 10.2 percent of turnovers in the time period dating 1971-1976 to a high of 23.4 percent of turnovers over the period 1989-1994, with significant increases between 1976 and 1989 as well. These numbers indicate that chief executive dismissal has become even more increasingly common in recent years. Despite the increased focus on executive dismissal, relatively little is known about its antecedents (Hatfield et al., 1999). Antecedents of CEO Dismissal Early research on executive succession focuses primarily on executive turnover (e.g., Gamson & Scotch, 1964; Grusky, 1963); that is, any change in the top executive at an 9 organization. Turnover may occur for a variety of reasons, including retirement, CEO death or other medical complications, dismissal of the CEO, or the CEO voluntarily leaving for a new position. These studies examine the antecedents, primarily performance, that predict whether an executive would be replaced, voluntarily or involuntarily, after a period of time. Dismissal, on the other hand, refers to the forced or involuntary replacement of an executive as an action taken by the board of directors (Fredrickson et al., 1988). Thus, turnover includes all cases of dismissal, but dismissal is only a subset of all executive turnovers. Early research on executive turnover and dismissal focuses primarily on the economic performance of a firm as the key antecedent. Executive dismissal in the case of poor performance is expected in order to attempt to stimulate change in an organization (Barker, Patterson Jr, & Mueller, 2001) or to placate concerned stakeholders (Boeker, 1992; Gamson & Scotch, 1964). Overwhelming empirical evidence is consistent with the expectation that turnover will occur in poorly performing firms (Allen & Panian, 1982; Boeker, 1992; Grusky, 1963; James & Soref, 1981; Salancik & Pfeffer, 1980; Schwartz & Menon, 1985; Warner, Watts, & Wruck, 1988). While performance is an important antecedent of executive turnover and dismissal, performance alone explains only approximately 20 percent of variance in executive turnover according to one compilation of research on top executives (Finkelstein et al., 2009). Numerous examples exist where CEOs have retained their jobs despite poor performance, while other examples highlight dismissed executives even when firms are performing well (Fredrickson et al., 1988). Given this anecdotal evidence, researchers have begun examining other antecedents that may explain executive dismissal. 10 In recent years, research has evolved to examine other dynamics that may affect a CEO’s dismissal, particularly focusing on the role board monitoring, governance, and power play in making this decision. In order for the CEO to be dismissed, the executive must lack sufficient power to prevent his or her own dismissal. Power in the context of executive dismissal has been primarily examined from three different perspectives: 1) Board power, 2) executive power, 3) third-party power. The board of directors is ultimately charged with the decision of whether to remove a CEO from office (Walsh & Seward, 1990). A number of researchers examine whether boards with more power (e.g. outsider representation) are more likely to remove CEOs when firms perform poorly. Board allegiances can be critical when determining whether a CEO will be removed from office (Fredrickson et al., 1988). Under conditions of poor performance, boards dominated by outsiders are more likely to dismiss a CEO than insider-dominated boards (Boeker, 1992; Cannella & Lubatkin, 1993; Hatfield et al., 1999). As the board gains power and has fewer allegiances to the CEO, the board is more able to exert its power to remove the CEO. The CEO’s power can also play a significant role in whether the board can exert influence by removing the chief executive. Finkelstein (1992) notes that top managers can develop power through four means: 1) structural power (e.g. duality), 2) ownership power, 3) expert power, and 4) prestige power. To date, researchers have primarily focused on structural and ownership power, with some focus on expert power through examination of the effects of executive tenure on CEO dismissal. As CEOs gain structural power by holding multiple titles in an organization, they gain greater power and are able to dominate the board’s agenda (Harrison, Torres, & Kukalis, 1988; Ocasio, 1994). CEO duality is thought to hamper board independence and promote managerial entrenchment (Cannella & Lubatkin, 1993; Rechner & Dalton, 1991). 11 Consistent with expectations, previous research has found CEOs are less likely to be removed when the executive also serves as board chair (Goyal & Park, 2002). CEOs also gain additional power by accumulating ownership stakes in the organization. As CEO ownership increases, the CEO is able to exercise greater voting rights, increasing the CEO’s ability to remain entrenched in the organization and reducing the board’s ability to remove the CEO. Finally, CEOs also gain power through greater tenure in the organization. Tenure allows the CEO to develop greater knowledge that is firm-specific (Harris & Helfat, 1997), while also nominating directors whose allegiance lies more with the CEO than with the firm (Boeker, 1992). Research has found that CEOs with greater tenures and who have nominated more directors face lower probabilities of dismissal (Denis, Denis, & Sarin, 1997; Gregory-Smith, Thompson, & Wright, 2009; Harrison et al., 1988). Each of these potential sources of power increases the executive’s ability to remain entrenched in the organization and reduces the independence of the board when attempting to evaluate the CEO’s ability. While these studies on board monitoring, CEO ownership, and executive power add considerably to our understanding of under what circumstances CEOs will be dismissed, these studies have almost exclusively been performed using agency theory and its assumptions. Little research has examined other antecedents or used alternative theoretical viewpoints in order to understand decision making processes boards undertake when determining whether to dismiss the CEO. Despite this focus on agency theory, several new studies examine dismissal from an alternative context; specifically, the role that third parties play in framing a CEO’s performance. This line of research indicates that third parties also play a critical role in influencing whether dismissal will occur. First, significant owners may influence the process, such that large 12 blockholders or institutional investors may be able to reduce the power of the CEO or limit the problems related to non-independent or weak boards. Greater ownership power by such investors may enable them to discipline poorly performing CEOs (Koh, 2003). Firms with more concentrated ownership, as well as with greater holdings by blockholders have been shown to have an increased likelihood to dismiss CEOs, especially when performance is poor (Boeker, 1992; Gregory-Smith et al., 2009). Second, independent third parties can play a large role in influencing how the board assesses a chief executive’s performance. Specifically, media accounts expressing dismay over a CEO’s performance repeatedly, especially in high profile outlets, may pressure the board into making a change. Additionally, investment analysts who alter company ratings may also change how the board evaluates an executive’s performance. Recent research provides evidence that analysts, independent arbiters of firm current and future performance, may compel a board to make a change by rating firms’ stock less favorably (Wiersema & Zhang, 2011). Research on the importance of power in executive dismissal is still in its early stages, but results clearly provide evidence that as executives gain power vis-à-vis the board of directors, the probability of dismissal decreases significantly. However, if the board can maintain both power and independence from management, the board has the ability to remove the CEO from office. In cases where the board becomes co-opted, significant owners and independent third parties may be able to step in and reduce the board’s ineffectiveness and facilitate the dismissal of the CEO. Reasons for CEO Dismissal Evidence on the dismissal of key executives in organizations tells us that CEOs get fired for a variety of reasons. The most common of these is that organizations dismiss executives 13 when performance of the firm is poor. The behavioral theory of the firm (e.g., Cyert & March, 1963; March & Simon, 1958) argues that as firm performance declines and firms fail to meet their aspiration levels, firms solve the problem through a search for alternatives. One solution to the problem of poor performance is to dismiss the chief executive and replace him or her with another candidate. Thus, one of the primary drivers of dismissal is poor performance by the CEO’s firm. While performance is a strong driver of dismissal, evidence suggests CEOs are dismissed for a variety of other reasons as well. CEOs may be dismissed for personality reasons, strategic disagreements with the board of directors, or due to in-fighting with other top managers. For example, Bob Nardelli was fired after five years at Home Depot in part because his personal style and demeanor came off as arrogant and he often alienated other executives of the firm. Similarly, Carly Fiorina was ousted as CEO of Hewlett-Packard after a number of disagreements about the firm’s strategic direction occurred with the board. While the firm was not severely underperforming, Fiorina’s personal style and unwillingness to listen to the board of directors led to her ultimate dismissal. While firm performance and interpersonal dynamics may both serve as reasons for dismissal of CEOs, the firm’s board of directors also has a fiduciary duty to shareholders to consider when evaluating the CEO. As such, it is the board’s job to replace a CEO even if the firm has achieved satisfactory returns for shareholders during the CEO’s tenure if another executive is available that can increase returns to shareholders. Thus, if the board believes an available alternative candidate will achieve greater success for the organization, the board has a duty and responsibility to shareholders to dismiss the current CEO and hire the new candidate. For example, in 2003 the Detroit Pistons of the National Basketball Association dismissed head 14 coach Rick Carlisle, who had just won two straight divisional titles. Carlisle’s performance certainly could not be considered sub-par, as only 6 teams claim a divisional title each year. However, the Pistons immediately turned around and hired Hall of Fame head coach Larry Brown as Carlisle’s replacement. Brown represented a candidate with tremendous experience, bringing a resume that included championships in both the NBA as well as college basketball. The dismissal of Carlisle followed by the hiring of Brown represented a situation where the Pistons attempted to improve their franchise by hiring the best available candidate to coach their team, despite already having what appeared to be a capable executive as coach. Uncertainty in Assessment of Executive Performance While CEOs may be dismissed for a variety of reasons, poor performance is often the catalyst for the dismissal of a CEO. When firms perform poorly, the board of directors must make both a managerial and an environmental assessment as to why performance was below expectations (Walsh & Seward, 1990). Managerial assessment requires examining the effort and ability of the executive in place, while an environmental assessment requires examination of environmental effects, including competition and industry effects. According to Walsh and Seward (1990), decoupling ability and effort from managerial performance is difficult. As such, prior to removing the CEO from office, the firm should alter the mix and arrangement of incentives and compensation offered to the CEO. If performance persists as poor, the board must then replace the manager as he / she does not have the required ability to run the company. These arguments provide some logic as to why performance only explains a modest amount of variance in executive dismissal. More importantly, however, these arguments call into question the notion that the evaluation of executive performance is simplistic. Evaluation of the CEO is much more 15 complex, as there is ambiguity over how much the CEO actually contributes to firm performance (Lieberson & O'Connor, 1972; Pfeffer & Salancik, 1977). Furthermore, attributions of managerial ability are both difficult and noisy as organizational performance is affected by both managerial decisions and systematic factors that exist both at the firm and industry level (Holmstrom, 1982). Given the ambiguity noted above, both technological uncertainty and performance standard uncertainty may increase the difficulty of evaluating a manager’s capabilities (Graffin & Ward, 2010). Technological uncertainty exists when there is a loose coupling between an actor’s performance and his or her underlying capabilities, which presents a problem when evaluating a CEO’s ability based on firm performance. As noted above, this problem makes understanding and evaluating the CEO’s capabilities even more difficult. Alternatively, performance standard uncertainty exists when uncertainty surrounds the benchmarks or standards against which the CEO is to be judged (Graffin & Ward, 2010). With regards to technological uncertainty, identifying a CEO’s true capabilities becomes difficult to assess based on performance due to the factors outside the CEO’s control with regard to performance. Additionally, the board often has minimal interactions with the firm’s CEO (Carter & Lorsch, 2004). Board interactions with the CEO often occur only a few times a year when the CEO and board meet directly face to face. This limited interaction means the board does not understand the CEO’s complex thought patterns, the day to day operational decisions made by the CEO, or often even the CEO’s interpersonal style with others in the organization. In other words, the board does not directly monitor and observe the CEO’s behavior. While the evaluation of the CEO is one of the most important tasks in an organization, it is often one done with some of the least directly observable information available. This limited interaction forces the board to seek alternative cues that provide a signal of the executive’s quality. 16 There are three distinct problems which create performance standard ambiguity when assessing managerial performance. First, boards must determine what exactly constitutes poor performance worthy of dismissing a CEO. In other words, how bad does performance need to be in order to warrant a change in CEOs? Firms performing below the industry’s average in profitability may need change, but research has not yet examined how far below industry average performance needs to fall before the board starts to search for a solution to the performance problem. Alternatively, boards may have a duty to replace a CEO if the firm’s industry is performing extremely well and the firm is not (Jenter & Kanaan, forthcoming; Kaplan & Minton, 2010). These problems only further complicate the evaluation of the CEO by the board of directors. . Second, research has not yet addressed how long performance needs to be poor before change is necessary. Dismissing a CEO after one year of poor performance may be reactive and cause harm to the firm itself. Different boards may have different standards for performance evaluation, such that two years of poor performance for one board may justify CEO dismissal, while in other firms significantly longer downturns are necessary before change is warranted. Within each situation, ambiguity may exist where factors outside the CEO’s or firm’s control affect firm profitability. This ambiguity over when firms are performing poorly indicates that boards have latitude they can afford to CEOs when performance is poor. Finally, standards need to be identified against whom and what should performance be evaluated. Firms must examine performance within the context of historical performance and future aspirations for performance. For example, two firms may perform at slightly below industry average levels in a given year. Taken together, these firms may not see a need for 17 change. However, if one firm has historically underperformed and changed leadership recently, this firm may be extremely pleased with current performance (performance, while still below industry average, may be at an all time high). Alternatively, the second firm may historically outperform the industry. In this case, the firm may see slightly below average industry performance as extremely poor and believe change may be necessary. In this situation, the firm may be losing its competitive advantage and failure to adapt quickly may only further erode such an advantage. In both of these situation, merely using the industry as a referent, rather than considering historical context, may not provide enough information for the board to assess managerial performance. Overall, these arguments suggest that determining whether a manager’s ability (or lack thereof) contributed to poor firm performance is extremely difficult. Furthermore, these arguments call into question the simplistic notion that poor performance should lead to dismissal, as different boards may define “poor performance” in alternative fashions. When determining whether to make the decision to dismiss a CEO, the board must also determine whether the existing CEO has the ability to perform well going into the future. Higher quality CEOs may be decoupled from past performance or may signal the ability to turn around the firm in the future. Boards must also be wary of dismissing an executive if another executive of higher quality cannot be obtained. Given these arguments, it is important to begin to attempt to understand why boards choose to dismiss some executives and not others. Figure 1 presents the main hypotheses of this study. Regardless of firm performance, I argue that executives with higher levels of human and reputational capital are expected to signal higher levels of capabilities to the board. These two factors send quality signals to the board of directors to alter the assessment of an executive’s 18 capabilities, which should lead to a lowered likelihood of dismissal. Board members limited interaction with CEOs occurs as members only work part-time for the companies on whose boards they sit and often lack detailed knowledge necessary to understand the firm’s intricate operations. This limited interaction reduces the directly observable information that is evaluated in assessing managerial ability. Thus, I argue that given this limited interaction, boards will look to quality signals to assist in the evaluation of a CEO’s capabilities. These signals may include the CEO’s reputation as an executive, the CEO’s background, or the CEO’s previous affiliation with prestigious organizations to help assess the CEO’s capabilities. Additionally, as firms perform poorly, these quality signals will become more important, as the board will increase scrutiny on the CEO’s performance and capabilities. As scrutiny on the executive’s performance increases, the board will look to signals that may indicate whether the executive can turn the organization around going forward or if new management is needed. Given the uncertainty faced when assessing executive capabilities in the wake of poor performance, boards may turn to external cues, such as third-party quality signals (e.g. certifications) in order to reduce ambiguity and provide additional corroboration for conclusions reached (Graffin & Ward, 2010; Podolny, 2005; Rao, 1994; Rindova et al., 2005). The board can argue that the CEO has the required capabilities and experience to turn the firm around in the future, while arguing that factors beyond the CEO’s control contributed to poor firm performance. Human capital and reputation can serve as a signal of the executive’s quality, which reduces the noise surrounding the CEO’s ability and contribution to past performance. Furthermore, the availability of high quality candidates as CEO alternatives should impact the likelihood of CEO dismissal. If the board has access to high quality internal or external candidates that can improve firm performance in the future, the board has a duty to 19 dismiss the current CEO and replace him/her with the higher quality candidate. In cases where there are more viable internal and external candidates that can improve the expected quality of the firm’s CEO, the board should have an increased likelihood of dismissing the CEO. However, if few candidates are available, the board should be more reticent to dismiss a chief executive. These arguments are especially important when performance is poor. Under conditions of poor performance, a larger pool of high quality candidates, either internal or external to the firm, should increase the likelihood of CEO dismissal, as the current CEO has been unable to deliver adequate performance. If few quality candidates are available, the board may choose not to dismiss the current CEO, as the next best alternative may only disrupt firm operations without bringing future benefits. Finally, it is important to note that this study controls for characteristics relating to board monitoring and CEO power in the analysis. Firms with better board monitoring are expected to be more likely to dismiss a CEO, all else equal. This relationship is especially attenuated when performance is poor, as better monitoring ensures that agency problems are reduced. However, CEO power through duality and ownership stakes may also affect the decision to dismiss a CEO, such that CEO’s with greater power will have a lower likelihood of being dismissed. In situations where CEOs have a great amount of power, it will be more unlikely that the board can dismiss the CEO. Furthermore, this relationship will also be attenuated when performance is poor. CEOs will have a lower need to exercise power when performance is above average, but when performance is poor, power can buffer the CEO from effective monitoring by the board. --------------------------------------- Insert Figure 1 about here --------------------------------------- 20 The remainder of this study will address how human capital and executive reputation can serve as signals of a CEO’s capabilities to the board of directors, how the market for alternative available CEO candidates can affect the board’s decision to dismiss a CEO, and how performance interacts with the executive’s capital and the market for alternative CEO candidates to increase the likelihood of dismissal. Finally, an empirical analysis is performed regarding the likelihood of dismissal of CEOs over a 5 year period examining whether executive capital and the market for CEO candidates affects the likelihood of dismissal. SIGNALS OF EXECUTIVE QUALITY AND CEO DISMISSAL As noted above, the board is responsible for the decision of whether to dismiss the firm’s CEO. However, the board often has minimal interactions with the firm’s CEO (Carter & Lorsch, 2004). This limited interaction means the board does directly observe the CEO’s performance. Instead, the board must examine externally visible output signals regarding the CEO’s ability. Thus, boards often look to performance as an indicator of managerial ability. However, performance alone does not provide a strong enough signal regarding ability due to the variety of factors that influence performance. Therefore, the board is likely to look to external cues that may corroborate their beliefs regarding the executive’s ability including media reports and analyst ratings (Wiersema & Zhang, 2011). Building upon this logic, I argue that an executive’s stock of human capital and reputation as a CEO can influence how the board perceives the executive’s capabilities when evaluating the CEO, as these provide proxies for the executive’s knowledge, skills, and abilities. Human Capital and CEO Dismissal Human capital theory notes that a CEO has a collection of skills, knowledge, and experience (Becker, 1962; Buchholtz, Ribbens, & Houle, 2003); skills which may assist the 21 board in determining whether to dismiss a CEO. At the time employers decide whether to hire executives, employers are unsure of the individual’s capabilities (Hamori, 2006). Signaling theory (Spence, 1974), however, contends that employers can observe visible cues relating to the executive, including educational background, prior experience, and past achievements, all of which signal expected future performance (Rosenbaum, 1984). These observable characteristics and individual attributes assist in evaluating the capabilities of a potential executive (Spence, 1973). As human capital is accumulated, executives can better signal their own quality to the external labor market, as attributes of human capital are the strongest and most consistent predictors of managerial career advancement (Kirchmeyer, 1998). Furthermore, a number of human capital investments, including education and experience, not only represent past achievements and decisions, but also can serve as a signal of future performance (Fulmer, 2009). Given that employers only have limited interaction with most potential executives, accumulated human capital, including past performance, affiliation with prestigious organizations, and educational prestige, can all signal the executive’s quality and reduce the potential problem of adverse selection (Zajac, 1990). Managers may differ in the level of skills and ability they possess from previous experience (Bailey & Helfat, 2003), which may affect the decisions managers make when serving as CEO. Human capital may not only help reduce the problem of adverse selection, but can also assist the board in its assessment of the CEO’s ability when deciding whether to retain an existing CEO. Decision makers are vulnerable to their own biases when arriving at judgments (Kahneman, 2003; Tversky & Kahneman, 1974). As CEOs develop greater levels of human capital, the board’s decision making process is more likely to be biased towards retaining the 22 CEO. The CEO’s accumulated human capital will assist in the evaluation process, by providing cues to the board that the CEO has the ability to perform well. Thus, I argue that as CEOs develop greater levels of human capital, the board is more likely to positively evaluate a CEO’s ability, as the board has a signal that the CEO has the ability to successfully run the firm based on past achievements. Replacing the CEO in these situations may yield a lower quality CEO who performs worse than the existing CEO. Alternatively, CEOs with lower levels of human capital, which provides fewer signals of the CEO’s capabilities, will be more likely to be dismissed. This logic suggests that boards will utilize existing information based on past accomplishments by CEOs in office to determine whether to retain the CEO. As more signals are sent that indicate the CEO has the ability to be successful in the future, the more likely it is that the board will choose to retain the CEO, regardless of past performance. Thus, I argue that the CEO’s firm tenure, prior experience, education level, educational institution affiliation, and compensation will all impact the firm’s evaluation of the CEO’s capabilities and ultimately whether to dismiss the CEO. Firm Tenure. Past research has examined executive tenure from a power perspective, noting that greater tenure enables managerial power building and entrenchment (Harrison et al., 1988), primarily through the nomination of new directors to the board. However, firm tenure also enables managerial learning and develops firm-specific knowledge and skills (Harris & Helfat, 1997). As executives spend time in the organization, they gain a stronger understanding of the firm’s operations and begin to make better decisions (Hambrick & Fukutomi, 1991; Henderson, Miller, & Hambrick, 2006). Firm-specific skills include understanding the day to day operations of the firm and the firm’s internal technologies. Greater tenure also allows for building of relationships with key stakeholders (Haleblian & Rajagopalan, 2006; Hill & Phan, 23 1991), which can be critical when implementing new strategies. These firm-specific skills enable managers to develop superior knowledge vis-à-vis other potential CEO candidates, as they have a better understanding of all aspects of the firm (Castanias & Helfat, 1991). Firm specific skills are worthless to other firms, but can have great value to the focal firm (Castanias & Helfat, 1991). Dismissal of the existing CEO results in losing all firm-specific skills gained by the executive during his or her tenure. In cases of dismissal, firms must ensure that the loss of this knowledge will not seriously affect the firm in the future or must make plans in order to ensure minimal knowledge loss as well. Thus, as CEO’s develop greater tenure within the firm, either as an executive or as CEO, the firm should be less likely to dismiss the CEO, as the potential loss of knowledge will outweigh the potential gains from replacement. The board will risk losing less in firm-specific skills if a CEO with lesser tenure is dismissed. Hypothesis 1: Regardless of performance, the likelihood of CEO dismissal will decrease as tenure within the firm increases. Prior Experience in Prestigious Firms. The prior stock of knowledge and experience built by a firm’s CEO serves as a strong influence on cognitive decision making and strategic choices (Datta & Rajagopalan, 1998; Kiesler & Sproull, 1982). Joining industry leaders early in the career process can help individuals achieve greater success (Citrin & Smith, 2003). For instance, an examination of career origins of Fortune 100 CFOs showed that 89 percent began their careers with industry leaders, including McKinsey, General Electric, IBM, and Bain (O'Sullivan, 2004). Affiliation with prestigious firms increases an executive’s prominence as the association enables future employers to assume that industry leaders evaluated the executive positively (Stuart, 2000). Reputable and prestigious organizations bestow ‘career imprints’ that the market values highly in executives (Higgins, 2005). As such, employees of industry leaders 24 make more successful moves to other employers (Hamori, 2006) and are hired disproportionately to run start-up firms (Higgins & Gulati, 2006). Specifically, reputable organizations facilitate development of employee capabilities by hiring the most talented individuals, providing better mentors, and training and immersing the individuals in the company’s superior industry and product knowledge (Crane, 1965; Higgins & Gulati, 2003; Long, 1978). Given this logic, executives with experience at industry leading firms should have greater levels of capability and insight to bring to the role of CEO. This experience signals quality in the executive in terms of both a reputable third-party’s endorsement of the candidate and the candidate’s own development through involvement and understanding of a leading firm’s operations. Specifically, firms can reduce potential problems with adverse selection by hiring CEOs with experience at prestigious firms. For instance, firms hiring CEOs who have been executives at General Electric have been found to get an abnormal stock return solely from announcing the CEO’s hiring (Lehmberg, Rowe, White, & Phillips, 2009). Previous experience managing a prestigious organization sends a signal that the individual has developed the ability to successfully run a reputable organization. CEOs with experience in prestigious firms may signal their ability to perform better in the future. Dismissal of CEOs with these prestigious affiliations and development of past capabilities through managerial experience in these organizations may lose important capabilities and relationships with critical stakeholders. Hypothesis 2: Regardless of firm performance, the likelihood of CEO dismissal will be lower for CEOs with prior experience as an executive with a prestigious firm. Prior CEO Experience. In addition to experience with reputable firms, prior experience as a CEO can also provide a greater levels of skill and background knowledge when CEOs are hired. Managers acquire and perfect skills through their prior work experience (Castanias & 25 Helfat, 1991). When a manager is in charge of a business, he or she learns about managing in that firm and the competitive cycle of the industry (Bailey & Helfat, 2003). Managerial experience, particularly as a CEO, can provide additional insight for a CEO candidate to pull from when making decisions in the future. In particular, past experiences and insights serve as a cognitive foundation for future strategic actions and outcomes, regardless of the success of the prior outcomes. The CEO position is unlike any other in the organization and requires different knowledge, skills and abilities that can only be learned through involvement in the position (Porter, Lorsch, & Nohria, 2004). CEOs with previous experience as another organization’s CEO, therefore, should face a lower likelihood of dismissal for several reasons. First, these CEOs develop prior capabilities relating to effective management and understanding the rigors of the CEO position. These capabilities, all else equal, should allow for the CEO with prior CEO experience to make better decisions and more effectively run operations. Second, CEOs with prior experience as chief executive also send a signal that another organization had a belief in the CEO’s ability. When evaluating whether to dismiss a CEO, the board’s assessment may be altered by the belief that a CEO with prior CEO experience will be more likely to achieve higher levels of performance and therefore should not be dismissed in comparison to CEOs with no prior experience as chief executive. Finally, being a CEO allows access into an elite club of managers and provides greater connections (Useem, 1984). These connections may be leveraged to bring new resources to help build a competitive advantage for the focal firm. Hypothesis 3: Regardless of firm performance, the likelihood of CEO dismissal will be lower for CEOs with prior experience as a CEO with another firm. 26 CEO Education Level. A manager’s educational background influences the functional and other experiences that the manager acquires (Castanias & Helfat, 2001). Educational level has been shown to have a strong effect on compensation in past studies (Agarwal, 1981; Fisher & Govindarajan, 1992). As managers acquire greater levels of education, additional capabilities are acquired and managers are exposed to additional information and perspectives that may affect decision-making. Additionally, education may allow for executives to build strong social networks. Previous research has linked educational level with greater innovation, knowledge, skills, and openness to change (Datta & Rajagopalan, 1998; Wiersema & Bantel, 1992). CEOs with greater levels of education signal higher quality and greater ability. When boards assess managerial quality, education level serves as a signal of cognitive orientation for future decision making and performance. CEOs with greater levels of education should be expected to have a greater psychological orientation with regards to ability and future performance. Hypothesis 4: Regardless of firm performance, the likelihood of CEO dismissal will decrease as CEO educational level increases. CEO Elite Education. While educational level endows greater background knowledge and experience, education at prestigious institutions also yields additional human capital. Education from elite universities creates greater credibility and prestige than association with less visible schools (Baltzell, 1989; Clement, 1977; Domhoff, 1967). Useem and Karabel (1986) noted that promotion to top corporate positions was most influenced by executives who earned a degree from an elite university. Executives graduating from elite institutions signal to others their importance (D'Aveni, 1990). Graduation from an elite institution not only yields a perceived higher quality education, but also provides access to an extensive network of 27 successful contacts. Finally, graduation from an elite institution signals that the institution was willing to accept the CEO as a candidate and that the CEO had the ability to complete the rigorous process to receive a degree. Graduation from a prestigious institution serves as a high quality signal about an executive’s capability in that both the institution was willing to endorse the CEO with acceptance, as well as bestow the CEO with a degree once the CEO met the institution’s requirements for graduation. Thus, CEOs having graduated from an elite institution provide a higher quality signal to the board of directors when performance is assessed. A CEO’s elite education can assist in reducing uncertainty regarding a CEO’s ability. These arguments do not suggest that CEO’s without elite educations cannot successfully serve as CEO; however, these CEOs do not emit signals which help the board of directors favorably evaluate the CEO’s ability. Hypothesis 5: Regardless of firm performance, the likelihood of CEO dismissal will be lower when the CEO has graduated from an elite institution. CEO Compensation. Like dismissal, the CEO compensation process is typically viewed as the result of an economic or political process, whereby CEOs with greater economic contributions to the firm or greater power within the firm are afforded higher compensation. However, an alternative view notes that the value of specific human capital is related to the availability of that type of human capital within the labor market (Becker & Murphy, 1992). Executives earning above average pay premiums or compensation represent rents earned for unique and valuable managerial skills (Castanias & Helfat, 1991). Pay premiums paid to executives with superior managerial skills represent a belief about the executive’s ability to generate superior performance (Harris & Helfat, 1997). Superior managerial skills meet the 28 criteria for competitive advantage and allow the firm to generate above average returns (Castanias & Helfat, 1991). Given this logic, I argue that firms that firms will be less likely to dismiss executives with higher levels of compensation. These executives have been offered pay premiums in order to acquire the superior managerial skills of the executive, regardless of the firm’s performance during the CEO’s tenure. Given the board’s investment in the CEO’s ability through its willingness to pay a premium to the CEO, dismissal should be less likely for two reasons. First, the board has made an investment to acquire the CEO’s valuable managerial skills that they believe will generate superior returns in the future. Thus, the board should be less likely to cut ties with that investment in the short-term. Second, if the board decides to dismiss the CEO, their decision-making ability may be called into question if the executive was paid a premium and was unable to deliver on the investment made. Thus, the board may be hesitant to dismiss an executive until absolutely necessary in order to “save face” regarding past decisions. CEOs earning less in compensation may earn less due to fewer alternatives outside of the focal firm and the compensation level may be indicative of the CEO’s lower perceived level of ability vis-à-vis other potential CEOs. Hypothesis 6: Regardless of firm performance, the likelihood of CEO dismissal will decrease as CEO compensation increases. Overall, human capital can serve as a signal of an executive’s quality when evaluating the executive’s performance. As ambiguity around an individual’s performance contribution and the appropriate metrics for evaluating performance exists, these quality signals should play a significant role in helping reduce uncertainty regarding a CEO’s expected future performance and should act as a buffer from dismissal for CEOs with significant levels of human capital. 29 These signals provide at least some information to reduce the level of uncertainty that boards face when attempting to assess managerial performance and may at least provide some information that can alter the board’s evaluation of performance. CEO Reputation and Executive Dismissal While an executive’s human capital, developed over time, may serve as a quality signal regarding his or her ability, an executive’s reputation should also serve as a signal of ability and impact how others evaluate the executive. An executive’s reputation serves as the collective judgment of others regarding the capabilities or quality of the executive earned over time (Graffin & Ward, 2010; Rindova et al., 2005; Washington & Zajac, 2005). While some human capital variables (e.g. prior experience, graduation from an elite institution) may inform individual’s perceptions of the executive over time (e.g. reputation as an individual or as an employee), this study focuses on the executive’s reputation as a CEO; that is, the collective belief others hold regarding the CEO’s ability to manage an entire organization. One school of thought notes that reputation is based on a continued appraisal of the executive’s capabilities over time informed by the executive’s performance (Carter & Ruefli, 2006). However, an alternative view argues that reputation is based on a world of imperfect information in which proxies or signals are used to make assumptions about the intentions and future behaviors of actors (Fombrun & Shanley, 1990; Kreps & Spence, 1985; Rao, 1994). These two views, however, are not mutually exclusive. An executive’s reputation can be informed by his or her past performance. However, given the ambiguity associated with an executive’s contribution to performance, third party quality signals and related proxies may also provide relevant and important information regarding an executive’s quality or capabilities (or 30 lack of quality or capabilities in some cases). Most importantly, a CEO’s reputation may be viewed as a sort of brand that the CEO cultivates (Ranft, Zinko, Ferris, & Buckley, 2006). A CEO’s reputation can be valuable to both the firm and the executive for several reasons. First, prestigious executive can aid in providing legitimacy to the firm and provide access to other prestigious individuals (Daily & Johnson, 1997). In particular, executives with higher levels of reputational capital may open doors to new opportunities and influential stakeholders who invest energy in the organization due to their belief in the CEO. The extent to which the organization is recognized in its field can strongly influence the organization’s economic value (Rindova et al., 2005). Second, enhanced reputational capital can signal a higher level of managerial talent (Lehmberg et al., 2009). Managers with better reputations are thought of as having greater ability and endorsements from others signify their belief in the executive’s capabilities. Finally, once an individual develops a reputation or track record, the marginal effect of new information is weakened (Holmstrom, 1982). In these cases, the effects of recent performance will be lesser for long-tenured executives. Given the importance of a CEO’s reputation for both the firm and the executive, it stands that a CEO’s reputation may alter how the board assesses the CEO’s capabilities. This altering of the assessment may allow the CEO to use his or her favorable reputation with the board to buffer him or herself from dismissal. When executives develop reputational capital, others believe the CEO is credible, signaling the CEO has the ability to successfully lead the organization. In these cases, the board will need extra evidence or special circumstances to dismiss the CEO, as external parties provide a positive evaluation of the CEO’s performance. Boards with a strong efficacy belief in the CEO are more likely to allow the CEO to persevere as he or she attempts to control the external environment (Haleblian & Rajagopalan, 2006). Boards 31 can gain greater belief through looking at both past performance of the executive and external cues that signal the quality of the executive. Market participants in general form a belief in the CEO’s ability and update it over time as new information is received. CEOs with a higher estimate of their ability are more likely to be retained than CEOs who are initially assessed as lower in ability (Milbourn, 2003). These arguments note that past performance can be critical to whether the board retains an executive, but also note that how others perceive the executive can alter the board’s evaluation of performance. Specifically, third party certification contests can also help to alter the executive’s reputation and impact the board’s assessment of a CEO’s performance. Certification contests influence the board by providing independent signals of the executive’s quality (Johnson, Ellstrand, Dalton, & Dalton, 2005; Malmendier & Tate, 2008; Wade, Porac, Pollock, & Graffin, 2006). Certification contests exists as competitions where actors are ranked in a given domain based upon performance criteria accepted by key stakeholders as credible and legitimate (Wade et al., 2006). These contests are likely to be considered as one of the few neutral sources of information regarding a CEO’s contribution and ability (Wade et al., 2006). The marginal contribution of executives is difficult to assess (Holmstrom, 1982; Lieberson & O'Connor, 1972), thus stakeholders may contest quality beliefs about the CEO. In these cases, certifications can act as useful cues regarding managerial competence (Wade et al., 2006). Victories in certification contests have been found to re-assure risk averse stakeholders and induce support for the organization, increasing both reputation and odds of survival (Rao, 1994). Under conditions of uncertainty when evaluating ability and competence, certification contests and endorsements from reputable third parties provide clear signals of capabilities and play an important role in quality evaluations (Rao, 1994; Rindova et al., 2005). Not only can 32 certifications reduce uncertainty regarding the individual’s performance, certifications may also reduce performance standard uncertainty regarding whether the executive’s capabilities meet or exceed standards that are desirable (Graffin & Ward, 2010). In the end, certification contests regarding executives allow for clear and comparable comparisons of a CEO’s relative worth or standing vis-à-vis other chief executives (Elsbach & Kramer, 1996). Employing a publicly certified CEO can also yield organizational benefits by signaling that the CEO is of high quality and will add value to the company (Wade et al., 2006). Given these arguments, I argue boards will be less likely to dismiss CEOs with higher levels of reputational capital. Such CEOs have external endorsements regarding their quality and competence, reducing the ambiguity surrounding the CEO’s performance and the appropriate metrics on which to evaluate the CEO. Furthermore, removing the CEO without good reason may result in media backlash and a loss of critical stakeholder support. CEO’s who are publicly certified or with track records of good performance send signals to stakeholders regarding their ability to perform well in the future. Dismissing such CEOs may result in a loss of future economic value to the organization, as the current CEO shows evidence he or she can contribute positively to the organization. Hypothesis 7: Regardless of firm performance, the likelihood of CEO dismissal will decrease as CEO reputation increases. THE MANAGERIAL LABOR MARKET AND CEO DISMISSAL The board of directors does not operate in a vacuum when determining whether to change CEOs (Fredrickson et al., 1988). Instead, the board must consider the availability of potential alternatives who may replace the CEO. If there is a strong supply of qualified candidates at the board’s disposal, the CEO is more likely to be dismissed (Pfeffer & Moore, 1980). The board 33 has an obvious replacement if a decision exists to dismiss the current CEO. Firms must be wary of dismissing the current CEO if a qualified candidate does not exist, as future performance may not improve, or in some cases may get worse. In fact, Wiersema (2002) found that most companies perform no better after dismissing a CEO. When examining potential replacements for a dismissed CEO, firms may seek alternatives both internally and externally to find qualified candidates. Successful CEO succession requires the grooming of internal candidates or an extensive external search (Friedman & Olk, 1995; Zhang & Rajagopalan, 2004). Internal candidates may come through the ranks of senior management or the board of directors, while external candidates may be found within the industry or through contacts of the board of directors. As more qualified candidates exist in the market, the likelihood of CEO dismissal should also increase. Poorly performing firms with few internal or external options may be reticent to replace the CEO, as it may cause significant disruption in the organization without yielding much benefit (Fredrickson et al., 1988). Alternatively, firms performing even at an average level may consider dismissing the CEO if the potential exists to hire an extremely qualified CEO candidate. Internal Candidates for CEO Replacement Presence of a COO. The presence of a Chief Operating Officer (COO) provides a potential candidate for the position of CEO, as the COO is often groomed to replace the CEO when he or she retires (Vancil, 1987). Senior executives have a personal stake in the firm’s success, as the firm’s success impacts managerial prospects in the market for managerial talent (Fama, 1980a). The presence of a COO weakens the CEO’s power, as power becomes split among the CEO and COO (Worrell, Nemec, & Davidson, 1997). Furthermore, the COO as an heir apparent may become impatient under the CEO’s shadow and challenge the CEO in front of 34 the board (Levinson, 1993). The presence of a COO or president significantly increases the likelihood of selection of a new CEO internal to the firm (Zhang & Rajagopalan, 2003) and also significantly increases the likelihood of CEO dismissal when performance is poor or there is limited strategic change (Zhang, 2006). The COO provides a credible candidate to replace the CEO in case of dismissal and is often groomed directly by the CEO to become the chief general manager of the organization, has significant exposure to the board of directors already, and has had time to be evaluated by the board. Thus, a sitting COO provides a ready-made replacement for the board of directors to select if the CEO is dismissed. Hypothesis 8: Regardless of firm performance, the likelihood of CEO dismissal increases if a COO or President exists within the firm. Non-CEO Inside Directors. When evaluating potential candidates to replace a CEO in case of dismissal, directors typically have more complete information about the ability and personality of senior executives and other directors of the firm with whom the board has had prior experience (Zajac, 1990). Non-CEO inside directors allow for others to challenge the CEO and limit the CEO’s influence over the board (Shen & Cannella, 2002). Senior executives who serve on the board have the greatest contact and experience with other board members who make the decision of whether to dismiss the CEO, making them the most likely contenders to replace a dismissed CEO (Ocasio, 1994). A seat on the board gives senior executives exposure to the board of directors and allows for the building of social networks and coalitions (Vancil, 1987). Greater numbers of insider non-CEO directors has been shown to lead to a higher likelihood of CEO dismissal followed by insider succession (Shen & Cannella, 2002). Furthermore, the ownership of non-CEO insiders has also been found to directly influence the likelihood of forced 35 CEO turnover (Huson et al., 2001). These arguments suggest that as the board has greater exposure to senior management through seats on the board of directors the likelihood of CEO dismissal will increase. This exposure reduces the risk of adverse selection associated with removing the CEO and hiring an insider and increases the board’s comfort with making the decision to dismiss a CEO. Hypothesis 9: Regardless of firm performance, the likelihood of CEO dismissal increases as the number of non-CEO inside directors increases. Number of Firm Divisions / Business Lines. Previous research has also found that firm size plays a significant role in the decision to replace a CEO (Grusky, 1963; James & Soref, 1981). Furthermore, as firm size increases, the likelihood of external succession decreases significantly (Dalton & Kesner, 1983). Larger firms have a greater ability to find an internal management candidate with a full complement of skills and abilities to run the organization (Bailey & Helfat, 2003). This pool of general management talent leads to managerial replacement, as qualified candidates can be found internally (Pfeffer & Moore, 1980). In diversified firms, candidates frequently can be identified through a “horse race” when running different lines of business (Bailey & Helfat, 2003). As firms acquire additional businesses or develop new lines of business, general management must be hired or groomed from within to run each line of business. These diversified business lines can essentially run independently, with each general manager responsible for a business line serving as a “CEO,” responsible for that unit’s performance. This situation allows for the grooming of general management talent and creates a tournament style “horse race” to identify potential CEO candidates (Friedman & Olk, 1995). For instance, prior to being hired as General Electric CEO, Jeffrey Immelt ran the Medical Systems division of the corporate giant. Mr. Immelt’s experience 36 and success as head of this organization helped him naturally stand out as a successful replacement for the retiring Jack Welch. Organizations with more divisions or lines of business have a greater need for general management talent and are better able to groom potential CEO candidates. The board can directly observe the performance of each of the internal CEO candidate’s business units to potentially identify an alternative candidate to the CEO. If some business units significantly outperform others, the board may be able to identify better talent internally to replace a dismissed CEO. Hypothesis 10: Regardless of firm performance, the likelihood of CEO dismissal increases as the number of firm divisions or lines of business increases. Presence of Interim CEO Candidates. While candidates may be readily identified among senior management talent, the board may also identify other candidates among its own ranks to potentially replace a departed CEO. Recent research supports the notion that replacing CEOs is a disruptive process, with many firms resorting to an interim CEO until the board identifies a permanent successor (Brady, 2006; Hymowitz, 2006). The decision to employ an interim CEO is often made under duress, including the board losing confidence in the current CEO to the point where the CEO is dismissed (Ballinger & Marcel, 2010). In interviews with three different interim CEOs, Ballinger and Marcel (2010) note that all three indicated they occupied the office of CEO due to the lack of viable candidates. The dismissal of a CEO often leads to bypassing the normal CEO succession process due to the pressures of finding a new CEO (Wiersema, 2002; Zhang, 2008). This pressure to select a new CEO may lead to the process being completed too hastily, resulting in a loss of stakeholder confidence (Wiersema, 2002). 37 Interim CEOs are an attractive option as they allow for the board to assess the current situation and conduct a thorough search for a successor (Hymowitz, 2006). Utilizing an interim CEO also allows for immediate dismissal of a lackluster CEO, rather than performing a search while retaining a lame duck CEO (Brady, 2006). Overall, interim CEOs allow boards to alleviate the time constraints associated with a search for a new CEO and to adequately assess a larger number of external candidates (Bell, Raiffa, & Tversky, 1988; Simon, 1997). Boards of directors may identify two different types of interim CEOs among their own ranks to select from if the sitting CEO is dismissed. First, previous CEOs of the focal firm serving on the board of directors can serve as a signal that an experienced candidate is ready to replace the CEO, even if for only an interim period, if dismissal occurs (Fredrickson et al., 1988). Prior CEOs of the firm have in-depth knowledge of the organization’s operations, often have maintained social networks which will be advantageous to ensuring operations run smoothly, and have previous general management experience to run the organization until a new CEO is identified. These individuals should more successfully be able to transition into the role of interim CEO and reduce some disruption associated with CEO dismissal. For instance, Michael Dell reclaimed the CEO position of the computer company he founded following the dismissal of then CEO Kevin Rollins. Second, the board may contain directors with prior experience as a CEO at other firms. These individuals have general management experience with other organizations which can be applied to the focal firm and may represent an attractive option to the board of directors in order to reduce disruption following CEO dismissal. The availability of an experienced candidate to serve as caretaker CEO until a replacement candidate can be appropriately identified may reduce the uncertainty associated with dismissal. For example, in 2001, United Airlines hired board 38 member John Creighton, former CEO of Weyerhaeuser, as CEO and Chairman of the board to replace departing CEO James Goodwin. Similarly, in 2005, 3M hired Robert Morrison as interim CEO after having served previously as CEO of Quaker Oats. Both of these individuals represented board members with a credible reputation in a prior organization as a CEO. Organizations with credible candidates to replace a dismissed CEO, even only on an interim basis, provide the board of directors with greater leverage when determining whether to dismiss the current CEO. A lack of credible candidates within the board may increase the disruption associated with CEO dismissal and may leave the organization as a “rudderless ship” should the CEO be dismissed. Hypothesis 11: Regardless of firm performance, the likelihood of CEO dismissal will increase with the presence of a prior CEO of the focal firm on the firm’s board of directors. Hypothesis 12: Regardless of firm performance, the likelihood of CEO dismissal will increase with the presence of a prior CEO of another firm on the firm’s board of directors. External Candidates for CEO Replacement Industry Concentration. The number of externally available candidates to replace the CEO should also affect the board’s decision of whether to dismiss the current CEO. Due to their understanding of competitive dynamics and industry operations, most CEOs are hired from within the industry. Hires from within the industry can offer different opinions on markets, provide links to suppliers and customers, or offer better approaches (Hager, Driscoll, Weber, & McWilliams, 1991) and may have a shorter learning curve, minimizing the disruption associated with changes in CEOs (Davidson, Nemec, Worrell, & Lin, 2002). Industries with more firms 39 have more potential general manager candidates from which to choose a new CEO. Fredrickson, Hambrick, and Baumrin (1988) noted that the number of firms in an industry should matter when dismissing a CEO, as more firms in an industry indicates a greater pool of general management talent to choose from. Furthermore, a greater number of firms in an industry create a natural experiment from which to pick the most successful general managers in an industry. Thus, firms replacing a CEO should have a larger talent pool of candidates to choose from when industries have a greater number of firms or lower level of industry concentration, making CEO dismissal more likely. Hypothesis 13: Regardless of firm performance, the likelihood of CEO dismissal will increase as industry concentration decreases. Prestigious Firms in an Industry. The willingness to dismiss a CEO is not only impacted by the number of potential candidates for the position, but also the quality of those candidates. As noted earlier, larger firms have more management talent to choose from (Pfeffer & Moore, 1980). Firms with greater access to managerial talent in more prestigious organizations will be more likely to dismiss a CEO for several reasons. First, individuals affiliated with reputable organizations make more successful moves to other organizations (Hamori, 2006). Second, reputable organizations provide career ‘imprints’ on employees that are valued by the market, making candidates in reputable organizations more attractive (Higgins, 2005). Firms that operate in industries with more prestigious organizations will have greater access to these candidates with valuable pedigrees. Third, affiliation with these high status actors increases the organization’s legitimacy. Given that most CEOs are hired from within the same industry, firms within the same industry as prestigious firms may have greater access to managerial talent. Prestigious 40 organizations not only provide successful imprints on their employees, but also train greater numbers of general management talent. These organizations tend to be larger in size and have multiple lines of business, facilitating general management training. Previous research has identified that the market values hiring individuals from these prestigious organizations due to this level of training and experience (Lehmberg et al., 2009). Not all senior managers and business unit CEOs of prestigious organizations can go on to become CEO of that organization. Thus, many are willing to leave to become CEO at other organizations within the industry. When assessing whether to dismiss a CEO, the board should examine the pool of general management candidates that exists not only within the firm, but within other prestigious firms in the industry. As this pool increases in size, the board is likely to be more willing to dismiss a CEO. Hypothesis 14: Regardless of firm performance, the likelihood of CEO dismissal increases as the number of prestigious organizations in a firm’s industry increases. Board Memberships of Firm Directors. While boards often look to industry competitors to find replacements for CEOs, external relationships can also yield quality candidates. Directorships provide firms with important sources of information on other firm practices and procedures (Haunschild, 1993). External directorships of sitting directors at a firm are a good way for directors to identify outside CEO candidates (Khurana, 2001). Directors may use these relationships specifically to recruit new managers when necessary (Barry, Muscarella, Peavy III, & Vetsuypens, 1990). As directors have greater ties to other firms, they develop stronger relationships with managerial talent. These ties also allow the directors evaluate managerial talent. When directors need to find new candidates to replace CEOs at other firms on whose boards they serve, they may turn to candidates with whom they are familiar in order to 41 reduce the problems of adverse selection. As directors have access to a larger pool of managerial talent, they are able to reduce the uncertainty that comes with dismissing a CEO and reduce the likelihood of adverse selection. Thus, dismissing a CEO is more likely when directors have personal experience working with a greater pool of managerial talent, especially when this experience is first hand based on direct interaction. Hypothesis 15: Regardless of firm performance, the likelihood of CEO dismissal increases as the number of external directorships among board members increases. Firm Geographical Location. While director networks are often a source of interaction and experience with pools of managerial talent, firms located in large cities with greater numbers of corporate headquarters also have an advantage when conducting executive searches. Denser networks of elites in local areas tend to spread information quicker (Davis & Greve, 1997; Marquis, 2003). A relatively small group of business executives tends to dominate local boards and other local colleagues are offered board seats as they become available in local companies (Ward & Feldman, 2008). Geographic proximity is often a source of invitation to join corporate boards. One study noted that 27 percent of ties among Fortune 1000 firms were between firms headquartered in the same state, indicating that geographical proximity plays a large role in firm interlocks (Friedland & Palmer, 1994). Overall, ties are likely to be reconstituted among directors of firms with headquarters in the same location (Palmer, Friedland, & Singh, 1986). These arguments suggest that firms often find directors and executives through interlocks in regional locations. Furthermore, executives can be identified by their involvement on local boards through local colleagues. As firms operate in larger cities with denser networks, boards should have access to a larger potential pool of candidates should they choose to dismiss the CEO. In addition to having access to a larger pool of candidates, boards in larger cities may also 42 more easily entice executives from other networks to join as CEO. Taken together, these arguments suggest that firms will be more likely to dismiss the CEO when operating in larger cities or regional locations. Hypothesis 16: Regardless of firm performance, the likelihood of CEO dismissal will be greater when firms operate in larger cities. FIRM PERFORMANCE AND EXECUTIVE DISMISSAL As noted earlier, much of the research on executive turnover and dismissal has revolved around the theory that as firm performance declines, firms will replace top managers (Allen & Panian, 1982; Boeker, 1992; Grusky, 1963; James & Soref, 1981; Salancik & Pfeffer, 1980; Schwartz & Menon, 1985; Warner et al., 1988). The behavioral theory of the firm (e.g., Cyert & March, 1963; March & Simon, 1958) argues that poor performance forces the board into action to search for alternatives. Thus, poor performance leads the board to take action in order to solve the problem; namely, find recourse that leads to better firm performance. While ambiguity may exist regarding the CEO’s contribution to performance, firms performing at extremely low levels and failing to meet aspiration levels will seek alternatives, including finding new management to implement new strategies. Thus, as firms perform poorly, the board is more likely to dismiss the new CEO in an attempt to correct the problem of poor performance. Hypothesis 17: The likelihood of CEO dismissal will increase as firm performance decreases. The Interaction of Performance and Executive Capital Firms performing well often have little incentive to dismiss managers. Following poor performance, however, the board must seek a solution in order to achieve future aspiration levels. During this process, the board must determine whether the CEO contributed to poor 43 performance. Thus, the board will place increased scrutiny on the CEO in order to assess whether a change in management is needed. The key issue for the board to determine is whether the CEO was responsible for poor performance and whether the CEO can help lead the firm to better performance in the future. If the board does not believe the CEO is capable of improving performance, the board has a duty to dismiss the executive. However, given the board’s limited interaction with the CEO and the ambiguity surrounding the CEO’s contribution to firm performance, the board may not have direct evidence or knowledge of the CEO’s capabilities. Instead, the board may examine external cues to determine whether the CEO is responsible for poor performance or whether the CEO can achieve better performance for the firm in the future. Thus, as performance declines, an executive’s human and reputational capital may play a more important role in serving as a buffer from dismissal. Executives with higher levels of human capital and a better reputation as a CEO may signal to the board greater levels of knowledge, skills, and abilities that can be applied to the organization in order to produce better future firm performance. For example, a poorly performing firm may retain a CEO with a reputation for managing in large firms with superior performance, significant knowledge of the focal firm, and an elite education as these characteristics may signal to the board that performance was out of his / her control. Alternatively, a CEO with little management experience and a low external reputation may be more likely to be dismissed, as the board may feel other potential candidates could be more successful in the CEO position. Thus, I argue that human capital and an executive’s reputation as a CEO will moderate the relationship between performance and executive dismissal, such that executives with higher levels of human capital will be less likely to be dismissed under conditions of poor performance. 44 Hypothesis 18: Firm tenure will moderate the relationship between poor performance and the likelihood of CEO dismissal, such that executives with higher levels of firm tenure will be less likely to be dismissed as performance declines. Hypothesis 19: Prior experience in a prestigious firm will moderate the relationship between poor performance and the likelihood of CEO dismissal, such that executives with prior executive experience in a prestigious firm will be less likely to be dismissed as performance declines. Hypothesis 20: Prior experience as a CEO will moderate the relationship between poor performance and the likelihood of CEO dismissal, such that executives with prior experience as a CEO will be less likely to be dismissed as performance declines. Hypothesis 21: The level of education completed by a CEO will moderate the relationship between poor performance and the likelihood of CEO dismissal, such that executives with higher levels of education completed will be less likely to be dismissed as performance declines. Hypothesis 22: Elite education by a CEO will moderate the relationship between poor performance and the likelihood of CEO dismissal, such that executives graduating from an elite educational institution will be less likely to be dismissed as performance declines. Hypothesis 23: CEO compensation will moderate the relationship between poor performance and the likelihood of CEO dismissal, such that executives with higher levels compensation will be less likely to be dismissed as performance declines. Hypothesis 24: The executive’s reputation as a CEO will moderate the relationship between poor performance and the likelihood of CEO dismissal, such that executives with a greater reputation as a CEO will be less likely to be dismissed as performance declines. 45 The Interaction of Performance and the Managerial Labor Market While performance is expected to serve as a signal to the board of the CEO’s quality, the perceived quality and quantity of candidates available to the board to replace a dismissed CEO should also affect the board’s decision of whether to dismiss the CEO. While the board may seek to dismiss a CEO under conditions of poor performance, one important aspect of the decision must be whether a qualified individual exists to replace the current CEO. On the one hand, the board has the fiduciary duty to shareholders to dismiss CEOs if better candidates exist, even if performance is adequate. On the other hand, the board also has a fiduciary duty to retain the CEO of a poorly performing firm if no viable candidates exist that could replace the current CEO and improve firm performance. Thus, I argue that even if the CEO is determined to be responsible for poor firm performance, the board will only replace the CEO if there are quality candidates that are available to replace the current CEO. Hypothesis 25: The presence of a COO will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms with a COO will be more likely to dismiss a CEO as performance declines. Hypothesis 26: The number of non-CEO inside directors will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms with more non-CEO directors will be more likely to dismiss a CEO as performance declines. Hypothesis 27: The number of divisions in a firm will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms with more divisions will be more likely to dismiss a CEO as performance declines. Hypothesis 28: The presence of a firm’s former CEO on the board will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that 46 firms with the prior CEO on the board will be more likely to dismiss a CEO as performance declines. Hypothesis 29: The presence of a retired CEO of another firm on a firm’s board of directors will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms with a retired CEO as a director will be more likely to dismiss a CEO as performance declines. Hypothesis 30: The concentration of an industry will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms in a more concentrated industry will be more likely to dismiss a CEO as performance declines. Hypothesis 31: The number of prestigious firms in an industry will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms with more prestigious firms in their industry will be more likely to dismiss a CEO as performance declines. Hypothesis 32: The number of board memberships held by a firm’s directors will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms with more directorships held by its director will be more likely to dismiss a CEO as performance declines. Hypothesis 33: The location of a firm in a major city will moderate the relationship between firm performance and the likelihood of CEO dismissal, such that firms headquartered in a major city will be more likely to dismiss a CEO as performance declines. 47 METHODS Sample The initial sample to examine dismissals of CEOs was drawn from all United States, publicly traded, manufacturing firms (SIC codes from 2000 to 3999) with more than 1000 employees in existence in 2005 and yielded 783 firms. Full data to estimate the likelihood of dismissal was gathered to cover the period 2005-2010 for all firms, where complete data is available. An initial examination yielded 3670 firm-year observations over this period, accounting for firms that were acquired, went bankrupt, or went private during the period 2006- 2010. After removing observations for which full data could not be retrieved, the final sample of firm years was 3,648 covering 778 companies and 1,080 firm-CEO pairings. All firm and performance information was gathered from CompuStat, while all information relating to firm board composition was gathered from the Corporate Library database, where available, as well as individual firm proxy statements. Data on Chief Executive Officers was gathered from the Marquis’ Who’s Who On the Web, Standard & Poor’s Corporate Register of Directors & Executives, Dun & Bradstreet’s Reference Book of Corporate Management, and Forbes, as well as from other company and executive profiles where available. Data on CEO compensation and previous employment was gathered from Execucomp. Data on geographical locations has been gathered from the United States Census Bureau. Data on CEO reputation was gathered from Institutional Investor Magazine, as well as news searches conducted via Factiva. CEO Dismissals The dependent variable for this study is whether a CEO was dismissed in a given year. Identifying dismissals of CEOs represents a significant challenge in management research, as 48 firms may not often disclose why CEOs resign or depart (Denis & Denis, 1995; Fredrickson et al., 1988; Shen & Cannella, 2002). However, recent research has identified several techniques in order to better identify dismissals which more successfully help in their identification (e.g., Shen & Cannella, 2002). Given the success of these approaches at identifying dismissals (e.g., Shen & Cannella, 2002; Zhang, 2006, 2008), this study follows the approach used by Shen and Cannella (2002) relying on news reports and descriptions of the event surrounding the departure of a CEO from a sample firm. News reports surrounding all CEO turnover events were examined in order to classify each turnover event and determine whether dismissal occurred. First, successions were excluded that were the result of a CEO’s death, health reasons, acceptance of a similar position at another firm, or a merger or acquisition. Three criteria were then used to identify instances defined as dismissals. First, a CEO may be directly reported as fired or forced out; however, this is rare. Second, a CEO is reported as resigning under pressure, unexpectedly, or immediately, due to poor performance, personal reasons that are not indicated, or a desire to pursue “other interests.” Third, cases where CEOs chose early retirement, but discussion of performance problems is reported in the announcement were classified as dismissals. As noted by Shen and Cannella (2002), CEOs may have little choice but to resign when pressure surrounding their performance is high. One slight modification to Shen and Cannella’s methodology in this study is that a dismissal was also recorded if other problems were identified that were not performance related, including notification of an SEC investigation or an indication of accounting restatement. Turnover of a CEO was classified as a dismissal if any of the above three criteria are met. CEO dismissal is therefore recorded as a dichotomous variable for each firm/year/CEO pairing in the sample, with a value of 1 for each year where a CEO is coded as dismissed within that year and a 49 value of 0 for years where no CEO turnover occurs or where a CEO turnover does not meet the above criteria for dismissal. Because this process requires judgment to determine whether a dismissal occurs, two independent research assistants coded each turnover event in the sample. There were 347 turnover events which occurred in the sample time period. Of these, the coders reached independent agreement on classification in 314 of the 347 cases (Cohen’s Kappa=0.727, p<0.001), indicating significant agreement was achieved. Due to unavailability of the first research assistant when the second coder completed the coding process, the author of this study discussed any disagreements with the second coder until mutual agreement was achieved on the 34 cases were initial agreement was not achieved. The final process yielded the identification of 102 dismissals, with 96 of the 778 sample firms dismissing at least one CEO. Independent Variables Executive capital. CEO tenure is the number of years the CEO has worked for the firm in any capacity. Prior experience is operationalized in two different ways. Prestige Experience is a dichotomous variable coded as 1 if the CEO has served as an executive listed on a proxy statement of any Fortune 500 company, and 0 otherwise. Fortune 500 companies are highly visible and attract top talent. Executives at these organizations are regularly sought after to run other organizations. This variable serves as a proxy for human capital with regards to an actor’s credibility and an indicator of capabilities. CEO Experience is also a dichotomous variable coded as 1 if the CEO has previously served in the capacity of CEO for another firm, and 0 otherwise. Education level was calculated using a 7-point scale based on the highest level of education completed by the CEO, adapted from Hermann and Datta (2005) (1=high school, 50 2=some college, 3=undergraduate degree, 4=some graduate program, 5=graduate degree, 6=medical or law degree, 7=doctorate)1. Elite education is a dichotomous variable coded as a 1 if the CEO attended one of the US News and World Report college rankings top 25 institutions in 2011 for either undergraduate or graduate education and 0 otherwise2. CEO compensation is the amount of a CEO’s base salary earned in a given year. Base salary was chosen as it is not affected by firm performance and as more indicative of the firm’s belief in the value the CEO brings to the firm independent of the firm’s performance. CEO reputation is measured using two constructs. The first measure of CEO reputation is the result of a certification contest; whether the CEO was included on Institutional Investor magazine’s Best CEOs list, which began identifying top CEOs in 20033. Institutional Investor asks buy and sell-side analysts to identify the best executives from the companies they follow and ranks the top 5 CEOs by industry. Survey data reflects the opinions of more than 1,300 analysts from more than 550 firms. Institutional Investor has a circulation of more than 130,000 and its list of Best CEOs has been reported on by outlets such as CNN and CNBC. IIM Top 5 CEO, therefore, is a dichotomous variable based on whether the CEO has been indicated as a top 5 CEO in his or her industry in any year prior to the observation year by Institutional Investor. CEOs that have been named a Top 5 CEO prior to the observation year were given a value of 1, while those who have not made the list were given a value of 0. 1 An alternative conceptualization was also tested utilizing a 5-point scale (1=high school, 2=some college, 3=undergraduate degree, 4=graduate degree, 5=doctorate, law or medical degree). Results using the alternative conceptualization followed a similar pattern. 2 Alternative conceptualizations were also tested utilizing dummy variables for whether the CEO attended an Ivy League institution or not, or whether the CEO attended a small group of institutions identified as elite by Finkelstein (1994). A second test was also done using the continuous variable developed by Finkelstein, ranging from 0-3 based on the level of education and the number of elite institutions that were attended. Results for all specifications were not significantly different from those reported herein, and are available from the author upon request. 3 Johnson, Young, and Welker (1993) used Financial World’s list of best CEOs in order to examine managerial reputation; however, Financial World discontinued operations in 1998 and the list is no longer published. 51 The second measure is the negative publicity associated with the CEO over the 3 years prior to the observation year in major US newspapers. An independent coder examined all newspaper reports across six major US news papers and coded each article mentioning a CEO’s name as either positive/neutral or negative. Publicity regarding the CEO was only coded as negative if the article specifically mentions the CEO and would cast the CEO in a negative light (e.g. adversely affect how others view the CEO, not how others view the organization). The six newspapers examined were The Wall Street Journal, The New York Times, The Washington Post, USA Today, the Los Angeles Times, the Chicago Tribune, and the Houston Chronicle. These newspapers cover different regions of the United States, but also represent 7 of the top 10 newspapers in circulation in the United States. This measure represents a count of the negative articles regarding the CEO over the three years prior to the observation year. This measure is similar to Milbourn’s (2003) construct of CEO reputation with regard to press coverage. Market for Alternative CEO Candidates. COO is a dichotomous variable equal to 1 if the firm reports a COO or President in the given year, and 0 otherwise. Non-CEO Board Insiders represents a count variable equal to the number of inside managers on the board of directors who are not the CEO. Firm divisions is a count variable equal to the number of segments or business lines the company reports in a given year to the SEC based on data gathered from Compustat. Retired CEO on Board is a dichotomous variable equal to 1 in a given year if any previous CEO of the firm remains on the board of directors and 0 otherwise. Other Retired CEO on Board is a dichotomous variable equal to 1 if there is a retired CEO of another firm on the board of directors in a given year and 0 otherwise. Industry concentration is the concentration ratio of a given 2-digit SIC code calculated using the Herfindahl index, a well- known indicator of competitive intensity in industries that is designed to capture the number of 52 competitors and the distribution of market shares among competitors (Kotha & Nair, 1995; Li, Poppo, & Zhou, 2008). Fortune 500 firms in industry is a variable representing the number of Fortune 500 firms in a given two-digit SIC code in a given year and is a measure of the number of prestigious, large firms in an industry. Directors’ board memberships is a variable representing the sum of all external directorships held by the firm’s board of directors in the observation year and measures the number of ties that a focal firm’s directors maintain through board involvement to other firms in a given year. Finally, Firm’s geographic location represents a dichotomous variable equal to 1 if the firm’s corporate headquarters are located in one of the 28 largest metropolitan statistical areas in the United States and 0 otherwise in a given year. The top 28 metropolitan areas were used in order to examine all metropolitan areas with more than 2 million people in the surrounding area4. Firm Performance. Firm performance was analyzed using two measures: Return on assets and Tobin’s q. Both ROA and Tobin’s q were analyzed using the current year’s firm performance industry adjusted by a firm’s two-digit SIC code based on data provided by COMPUSTAT5 in order to test how the performance at the time the decision was made affected the likelihood of dismissal. 4 Alternative conceptualizations were utilized to examine the top 10 largest MSAs (greater than 4.5 million people), top 15 (greater than 3.5 million people), and top 20 (2 million people). The pattern of results were similar with less explained variance. 5 Alternative conceptualizations were tested examining prior year industry averaged performance, three year industry averaged performance, and three and five year trends. Additionally, data were examined at both the 2 digit and 4 digit SIC level. The results are available upon request. The models with the greatest explained variance are reported in this study. 53 Control Variables Given prior research on the antecedents of executive dismissal, a number of control variables relating to executive power, board power, and ownership characteristics are added to the model to control for alternative effects on the likelihood of dismissal. Duality is a dichotomous variable set to 1 if the CEO is also the board chair and 0 otherwise. Outside directors represents the proportion of outside directors on the board of directors. Outside directors are theoretically thought to provide more independence from management and therefore will be more likely to dismiss a poorly performing CEO (Pitcher, Chreim, & Kisfalvi, 2000). Blockholder ownership represents the proportion of shares held by large blockholders including institutions, which are expected to be able to better monitor managerial actions and lead to better firm governance. Debt to equity ratio represents a ratio of the firm’s level of debt to its level of equity. This measure controls for the leverage a firm has in a given year and may be an indication of financial problems in the organization. Firm size is the natural log of the total assets of the firm in the observation year. Larger firms may face different decisions than small firms, and some research has identified larger firms are more likely to dismiss CEOs (Boeker, 1992). CEO Gender is a dichotomous variable equal to 1 if the CEO is a male and 0 for female in order to determine whether the likelihood of dismissal is different for men versus women in publicly traded firms. CEO Ownership is a proportion of firm stock owned by the CEO in the observation year and is a measure of the CEO’s power. CEOs with significant ownership may be able to insulate themselves from dismissal due to such power (Finkelstein, 1992). Finally, Current ratio is a ratio of the firm’s current assets to current liabilities and measures a firm’s liquidity or financial solvency. Firms with less liquidity may 54 face financial hardship that is separate from measures of performance and may be more likely to dismiss CEOs, as firm survival may be threatened. Analysis Data analysis was conducted using event history analysis techniques, examining the likelihood of dismissal of a CEO in a given year. Each individual observation is a firm/year/CEO pairing, with the sample including 3648 of such pairings. While the same firm/year/CEO pairing can appear multiple times in the sample, non-independence of observations is not an issue, as each year the board makes the decision, explicitly or implicitly, whether to dismiss the CEO. Thus, each case represents an independent observation based on the variety of variables included in the models specified. The analysis was conducted using probit regression. To control for time effects, dummy variables for each year in the sample were entered into the model, which can capture unseen time effects outside the model. For example, if executive dismissal was fashionable in a given year, the dummy variables would capture this phenomenon. All data which vary over time were recorded as of the observation year in order to capture effects at the time in which the decision was made, with the exception of IIM Top 5 CEO, which was recorded based on the CEO’s inclusion on the list in any prior year. RESULTS Table 1 shows the descriptive statistics and correlations among variables except for the year dummies. The decision to dismiss a CEO is significantly correlated with duality, blockholder ownership, CEO ownership, a firm’s Tobin’s q, CEO tenure, CEO compensation, and negative publicity. In this study, 347 turnovers occurred, of which 102 CEOs were classified as dismissed (29.39%). 55 ----------------------------------------------- Insert Table 1 about Here ----------------------------------------------- Hypotheses are tested using probit regression as reported in Table 2. Model 1 includes only control variables. Model 2 tests the effects of firm performance on the decision to dismiss a CEO, while Model 3 includes all main effects for hypothesis testing. To correct for heteroskedasticity, robust standard errors are reported and used for hypothesis testing (Huber, 1967; White, 1982). ----------------------------------------------- Insert Table 2 about Here ----------------------------------------------- In examining the impact of control variables on the likelihood of dismissal, results indicate that CEOs with duality, or who also serve as board chair, are significantly less likely to be dismissed, indicating that power via duality insulates a CEO. Furthermore, a firm’s current ratio, an indicator of liquidity, is negatively and significantly related to the likelihood of dismissal, suggesting that when firms face issues dealing with solvency (e.g. bankruptcy potentially looms), dismissal is more likely. Firm size and CEO ownership are both negative and marginally significant, suggesting that larger firms are less likely to dismiss CEOs and that dismissal is less likely when CEOs gain greater ownership power. In models 2 and 3, I examine the impact of firm performance, human capital, and alternative CEO candidates on the likelihood of dismissal. Hypotheses 1 through 6 examine the relationship between executive human capital and the likelihood of dismissal. Hypothesis 1 predicts that as a CEO’s tenure in the firm, in any position, increases, dismissal is less likely due 56 to the firm-specific human capital the CEO has accumulated. Results provide support for Hypothesis 1, as greater CEO tenure in the firm is associated with a lower likelihood of dismissal (p ≤ 0.05). Hypothesis 2 argues that prestigious experience, in the form of experience as an executive at a Fortune 500 firm, will insulate a CEO from dismissal. The coefficient for prestige experience is not significant, failing to provide support for Hypothesis 2. Hypothesis 3 predicts that prior CEO experience will reduce the likelihood of dismissal; however, results fail to support Hypothesis 3, as Prior CEO experience is not significant. With regard to Hypothesis 4, a CEO’s education level is predicted to decrease the likelihood of dismissal; however, results fail to provide support for Hypothesis 4. Hypothesis 5 predicts that a CEO’s elite education will signal higher quality to the board of directors. The coefficient on elite education is significant and indicates a relationship between the prestige of a CEO’s education and likelihood of dismissal (p ≤ 0.05); however, results are in the opposite direction of that predicted. Findings from the analysis indicate that a CEO’s elite education increases the likelihood of dismissal, suggesting that elite education may alter how the board views a candidate in a negative, rather than positive, manner. Finally, Hypothesis 6 examines the relationship between CEO compensation and the likelihood of dismissal, predicting that CEOs with higher compensation are less likely to be dismissed. Hypothesis 6 is supported, as the coefficient on CEO compensation is negative and significant (p ≤ 0.01), indicating that greater base salaries paid to CEOs decrease the likelihood of dismissal. While some argue that compensation may represent a CEO’s power, compensation in this case may be indicative of greater value attributed to the CEO. As firms increase the base salaries of their CEOs, the salary is a signal of the economic value the firm believes the CEO brings to the organization. 57 Hypothesis 7 examines the relationship between a CEO’s reputational capital and his or her likelihood of dismissal. The first measure of reputation examines an externally awarded certification, whether the CEO has been named a top 5 CEO by analysts in Institutional Investor magazine. Results on IIM Top 5 CEO, however, fail to find any significant relationship6. The second measure examined the amount of negative publicity the CEO received over the three years prior to the observation year. Results show that as CEOs received more negative publicity, the likelihood of dismissal increased significantly (p ≤ 0.05). These results suggest that negative publicity sends a signal to the board about the CEO’s quality, which then increases the likelihood of dismissal. Taken together, these results suggest that there is some support for Hypothesis 7, that reputational capital impacts the likelihood of CEO dismissal. Hypotheses 8 through 16 examine proxies for the market for alternative CEO candidates on the likelihood of CEO dismissal. Hypotheses 8-12 represent proxies for internal candidates to replace the CEO and argue that dismissal is more likely when there is a COO, when there are greater numbers of non-CEO insiders on the board, when the firm has more divisions to cultivate and identify managerial talent, and when a retired CEO of either the focal firm or another firm sits on the board. Among these variables, only the coefficient on Non-CEO board insiders is positive and significant (p ≤ 0.05), suggesting that when more insiders are on the board, dismissal is more likely. These results are consistent with Shen and Cannella (2002) who note that insiders serving on the board represent viable candidates to replace a CEO, as the board’s familiarity with such candidates can reduce problems associated with adverse selection. Thus, results provide support for Hypothesis 9, while failing to support Hypotheses 8, 10, 11, and 12. 6 Alternative conceptualizations examined whether the CEO was ranked as the top CEO in the Institutional Investor awards at any point, the count of appearances prior to the observation year for a CEO in the top 5 or as top CEO, and a dichotomous variable indicating whether the CEO was named a Top 30 CEO by Barron’s magazine. Results followed similar patterns using each of the conceptualizations and are available from the author. 58 Hypotheses 13-16 represent proxies for external candidates to replace the CEO and argue that dismissal is more likely when the industry is less concentrated, when there are more Fortune 500 firms in an industry, when directors have more board memberships to identify candidates, and when the firm is located in a major city. Results, however, fail to identify any significant relationships, and thus Hypotheses 13, 14, 15, and 16 are not supported. The final main effect examined is the effect of firm performance on likelihood of dismissal. In both model 2 and 3, the coefficients on both Firm ROA (p ≤ 0.01) and Tobin’s q (p ≤ 0.001) are negative and highly significant, suggesting that as firm performance (both accounting and market based measures) increases, dismissal is less likely. These results provide support for Hypothesis 17. Hypotheses 18 through 24 examine the interaction between firm performance and executive capital, predicting that CEO human and reputational capital will buffer a CEO from dismissal as performance decreases. Tables 3a, 3b, and 3c present the results of the analyses regarding the effects of the interaction between firm ROA and both executive capital and the market for alternative CEO candidates on the likelihood of dismissal. Following Aiken and West (1991), each term included in an interaction was mean centered prior to calculating the interaction term7. Model 4 presents the results of the analysis with all interaction terms included, while Models 5 through 21 test and report each interaction individually8. Results of the analyses indicate that the interaction between firm ROA and CEO compensation (p ≤ 0.01) and the 7 An alternative process was also used, whereby a dummy variable was created to represent any observations where firm performance fell below industry average performance (ROA < 0) in a given year. This dummy variable was then interacted with each human and reputational capital variable and each market for alternative candidate variable in order to determine whether these relationships were greater in firms with below industry averaged performance. The results are available from the author upon request. 8 For the sake of parsimony the remaining variables are not reported. However, in any cases where results were changed by the inclusion of the interaction term on main effects that are not in the interaction term, these changes have been highlighted in the table’s footnotes. Full results are available from the author upon request. 59 interaction between ROA and negative publicity (p ≤ 0.001) are significant, providing support for Hypotheses 23 and 24. Results, however, fail to provide support for Hypotheses 18 through 22. ----------------------------------------------- Insert Tables 3a, 3b, and 3c about here ----------------------------------------------- Hypotheses 25 through 33 predict an interaction between firm performance and the market for alternative CEO candidates, such that as the firm’s board can identify more external and internal candidates, dismissal should be more likely as performance decreases. Results find support for Hypothesis 33, as the interaction between ROA and the firm’s location in a major city is significant (p ≤ 0.01). Hypotheses 25 through 32 are not supported. To better understand the significant interactions, Figures 3, 4, and 5 plot the relationships between ROA and CEO compensation, negative publicity, and location in a major US city. Each interaction graph examines the likelihood of dismissal as ROA moves from one standard deviation below the mean to one standard deviation above the mean. As seen in Figure 2, when CEO compensation is one standard deviation above the mean (“High CEO Compensation”), the likelihood of dismissal increases slightly as firm ROA increases. However, when CEO compensation is one standard deviation below the mean, the likelihood of dismissal increases at a greater rate as ROA increases. These results suggest that lower compensation significantly impacts the likelihood of dismissal when performance is poor. Figure 3 graphs the relationship between negative publicity and ROA. The plot of the effects illustrates that the likelihood of dismissal is highest for CEOs with low negative publicity and low ROA, but decreases as ROA 60 increases. For CEOs with higher levels of negative publicity, the likelihood of dismissal increases as ROA increases, suggesting that negative publicity limits the effects of ROA on reducing the likelihood of dismissal. Finally, as seen in Figure 4, a firm’s location in a major city affects the likelihood of dismissal, such that CEOs are more likely to be dismissed when the firm is not located in one of the 28 largest US metropolitan areas. This likelihood decreases significantly as the firm’s ROA decreases, suggesting that the firm’s location in a major metropolitan area actually buffers a CEO from dismissal when ROA is poor. These results suggest that major metropolitan areas may not serve as pools for attracting talent, but rather may serve as networks that insulate CEOs from dismissal. ----------------------------------------------- Insert Figures 2, 3, and 4 about here ----------------------------------------------- In order to test additional interactions, analyses were run interacting each main effect of executive capital and the market for alternative CEO candidates with firm market performance as measured by Tobin’s q. Tables 4a, 4b, and 4c present the results of these analyses. Significant results were identified for the interaction between Tobin’s q and prestige experience. Figure 5 graphs the significant relationship identified. As seen in the figure, prestigious experience does not significantly alter the likelihood of dismissal regardless of Tobin’s q; however, for executives without prestigious experience, the likelihood of dismissal decreases as Tobin’s q increases, suggesting that prestigious experience provides some insulation from dismissal when market performance is poor. ----------------------------------------------- Insert Table 4a, 4b, and 4c and 61 Figure 5 about here ----------------------------------------------- In sum, results provide support that both measures of human and reputational capital reduce the likelihood of dismissal. Results indicate that greater firm tenure and CEO compensation provide evidence of CEO capabilities which reduce the likelihood of dismissal. Additionally, negative publicity about a CEO impacts his or her reputation, which increases the likelihood of dismissal. With regards to the market for alternative CEO candidates, only the number of non-CEO insiders on the board significantly impacts the likelihood of dismissal, such that when more non-CEO insiders are on the board, dismissal is more likely. The influence of performance also has a significant impact, as both accounting and market measures of performance impact the likelihood of dismissal. Additionally, firm ROA impacts the relationship between CEO compensation, negative publicity, and the firm’s location in a major city on the likelihood of dismissal, while market performance impacts the relationship between dismissal and prestigious experience similarly. Importantly, dismissal is most likely when CEO compensation is below average and performance is poor, indicating that lack of human capital in the form of compensation impacts the board’s evaluation of CEO performance. DISCUSSION In recent decades, the number of CEO dismissals has continued to increase. Research on CEO dismissal has primarily examined the decision to dismiss a CEO using agency theory and the behavioral theory of the firm; that is, when performance is poor, dismissal is more likely. Additionally, prior studies note that CEOs are able to insulate themselves from dismissal when monitoring is weak or when they have greater power than the board. However, firms in recent years are also under ever greater scrutiny by the media, investors, analysts, and the general 62 public. This scrutiny often leads to individuals outside the organization questioning the firm’s board as to why some CEOs are not fired. This study attempts to examine the question and provide reasons outside of agency theory for some boards choose to not fire some CEOs when expectations might dictate dismissal should occur. This study makes several important contributions to the existing literature on CEO dismissal. First, this study builds a theoretical model predicated on signaling of executive talent to the board of directors based on the foundation that board decision making is imperfect for several reasons. Building on prior research, I argue that performance alone is a noisy signal regarding CEO talent; that is, while CEOs are responsible for the firm’s performance, a variety of other factors contribute significantly to performance. Thus, performance alone should not be the measuring stick for CEO talent or ability. Second, I argue that the limited interaction the board has with the CEO makes the evaluation of a CEO’s capabilities extremely difficult. Despite the important task with which boards face in evaluating CEOs, boards spend little time in direct interaction with the CEO and rarely observe the CEO performing his or her responsibility. This limited interaction leads the board to evaluate external, observable cues regarding the CEO’s capabilities, as board members cannot fully observe the true capabilities of the CEO. This study adds to the existing literature on dismissal by adding to prior research that argues and discusses how the board of directors looks to external cues in order to help develop their opinions regarding CEO capabilities by illustrating how both human and reputational capital may signal an executive’s capabilities for the board’s evaluation process. In particular, the notion of a CEO’s reputation has not been explored in great detail. This study is one of the first to examine how multiple measures of reputation may affect the evaluation of the CEO and his or her capabilities. Future studies should continue to examine how CEOs develop 63 reputations, how such reputations change over time, and how reputation affects both firm and board decision making. In this study, a CEO’s reputation assists the board by providing information that can corroborate board member’s evaluation of CEO ability. Prior research argues and often assumes that dismissal does not occur when performance is poor primarily due to weak governance. However, this perspective discredits the integrity and willingness to perform fiduciary duties of many board members, as the assumption exists that board members perform their duties to serve the interests of the firm’s executives rather than performing their fiduciary duty. However, this study builds a theoretical model to provide evidence that the board utilizes and evaluates information that provides signals regarding executive quality outside of performance that may rationally lead the board to retain a CEO, even when outside stakeholders may believe that dismissal is warranted. The results of this study indicate that human capital and reputational capital do affect the likelihood of dismissal, in addition to a firm’s performance. Firm tenure, an indication of the knowledge that the CEO has gained regarding the firm’s operations, decreases the likelihood of dismissal. As CEOs gain greater tenure, firm specific human capital accumulates and the firm risks losing such capital if dismissal occurs. Additionally, results indicate greater CEO compensation in the form of base salary decreases the likelihood of dismissal. A CEO’s base salary is indicative of the firm’s willingness to pay to acquire the capabilities the CEO has developed. Given that base salary is not dependent upon performance, this portion of a CEO’s compensation is indicative of the board’s belief in the talent component of the CEO. Finally, results indicate that greater negative publicity regarding the CEO increases the likelihood of dismissal. Negative publicity in major US newspapers sends a negative signal regarding the CEO’s reputation and his or her capabilities to the board of directors, which in turn alters the 64 likelihood of dismissal. Taken together, these findings provide support for the notion that human and reputational capital affect the likelihood of dismissal. This finding is important as it indicates that a CEO’s negative reputation as displayed in print media can influence the board’s evaluation process. Despite these findings, several of the arguments of this study were not supported. First, elite education was found to be significantly related to dismissal; however, CEOs with elite education are more likely to be dismissed. This may indicate that boards have different evaluation patterns and expectations regarding CEOs with elite educations. Future research should continue to examine how CEO’s educational backgrounds impact firm decision making and how others evaluate the CEO. Additionally, no support was found for the role of the “Best CEO” certification contest from Institutional Investor magazine. Prior research has identified that certification contests can impact how others evaluate the winner of the contest, such that winning awards improves an organization’s or individual’s reputations (e.g., Rao, 1994). However, results of this study do not find results for such a certification using the Institutional Investor list of Best CEOs. Such results may indicate that certifications do not impact board decision making with regard to CEO quality and that performance represents a stronger signal of executive quality. Alternatively, however, the life of the Institutional Investor award was still relatively short during the sample timeframe, having been instituted in 2003. Given that winning a Best CEO award is positively correlated with firm performance, firms may not have had to make true dismissal decisions on CEOs awarded the Best CEO award, as performance may still be high. Future research should continue to examine whether certification contests alter how individuals and boards evaluate CEOs. 65 Second, this study is the first to develop and test a model for whether the market for alternative CEO candidates affects the board’s decision to dismiss a CEO. The decision to dismiss a CEO does not occur in a vacuum. If the firm’s board does not believe a higher quality CEO can be acquired when dismissal occurs, the board is committing the firm to unnecessary cost (e.g. severance, golden parachutes) and disruption in the organization. Thus, boards should be wary of dismissing executives when capable successors have not already been identified. Furthermore, boards also have a duty to dismiss CEOs if better alternatives are available. These arguments suggest that boards must evaluate not only the capabilities of the current CEO, but also identify potential alternatives before choosing to dismiss a CEO. This study identifies a variety of candidates both internally and externally that can assist in reducing the likelihood of adverse selection and help the board gain comfort with the decision to dismiss the CEO. Internal candidates are visible to the board, have developed relationships within the firm, and understand the firm’s operations. Theoretically, external candidates discussed in this study belong to highly visible organizations or are executives with whom the board already has a relationship with, in order to be able to assess the capabilities of alternative executives. This study builds on prior research by examining and testing whether the market for alternative candidates affects the likelihood of dismissal. Results of this study find some support, though limited, for the notion that alternative candidates affect the dismissal decision making process. The number of non-CEO board insiders significantly increases the likelihood of dismissal, such that as the board becomes more familiar with more internal candidates, the likelihood of dismissal increases. This finding is consistent with Shen and Cannella (2002) who note that internal executives may challenge the CEO in a socio-political process and co-opt the CEO’s power by going to other board members directly. 66 Despite this finding, significant results were not obtained to support the other predictions regarding the market for alternative CEO candidates. Other proxies for internal candidates, including the presence of a COO, the number of business segments, and the existence of retired CEOs on the board of directors do not significantly influence the dismissal decision. Additionally, proxies utilized for external candidates did not display significant findings either, including the industry’s concentration ratio, the number of board memberships held by directors and the location of the firm’s headquarters. There are several possible explanations for the lack of significant results. First, the proxies utilized within this study may not be representative of the process used by the board of directors. In fact, boards may not truly evaluate potential alternatives until the decision to dismiss a CEO is made and an interim CEO has been installed. In these cases, boards may believe that any new CEO presents a better alternative and that change is necessary. Change may only occur when the board believes any executive allows for a better long-term future. Additionally, boards may dismiss a CEO and utilize executive search firms as the primary means to identify candidates. The prevalence of executive search firms has increased in recent years in order to assist in identifying quality candidates for firms seeking executives. Thus, boards may not be the parties responsible for identifying candidates and alternative processes, including the use of executive search firms, may need to be more fully examined in the future. Second, alternative proxies may better identify how firms evaluate external candidates. For instance, the number of directorships held by the board may not matter as much as the quality of those directorships or the overlap among the directorships (e.g. directorships in the same industry). Directors who serve on multiple boards may not believe knowledge will transfer from executives of one firm to another. Additionally, firms may not be 67 able to acquire talent from other firms in the industry for a variety of reasons, including the lack of prestige of the firm who wishes to dismiss its CEO. In totality, the results of this study provide support for the idea that executive human and reputational capital affect how the board evaluates the CEO’s performance and capabilities, while limited support exists for the notion that the board evaluates alternative candidates in the decision to dismiss a CEO. Additionally, some support exists that these processes change as firm performance increases or decreases. Specifically, the level of CEO compensation, negative publicity regarding the CEO, and the firm’s location in a major city all work in conjunction with firm ROA in order to affect the likelihood of dismissal, such that the likelihood of dismissal is different at lower levels of compensation, publicity and firm location as performance changes. Future Research Future research should continue to understand the decision making process that boards use when determining whether to dismiss or retain a CEO. In particular, greater attention should be paid to opening the “black box” behind how boards interact and make decisions regarding topics such as CEO dismissal. For instance, research should examine the group decision making dynamics that affect the decision to dismiss a CEO. The process may be driven by consensus or dominated by individual directors. Additionally, research should examine how these group decision making dynamics revolve around group members’ interpersonal relations and individual power in order to alter the decision made by the board. Second, future research should speak with corporate governance experts and board members to better understand the processes used in order to evaluate CEO capabilities and performance. Furthermore, this research should better link how the evaluation process interacts with the dismissal process. Research is needed to understand how the evaluation of a CEO 68 affects the board’s decision to retain or dismiss a CEO. For instance, CEO evaluation may lead to a negative evaluation of the CEO’s performance; however, the board may believe that such performance is unlikely to continue for a variety of reasons, including changes in the general environment or the CEO ‘learning curve.’ Research should continue to understand other factors that provide additional information that serves as inputs to the board in evaluating the CEO’s capabilities. Third, research should continue to understand and evaluate how the board identifies alternative candidates in the decision to dismiss a CEO and when such an evaluation is appropriate. For instance, some boards may believe the current CEO’s performance to be so poor that a change is needed regardless of any potential replacements. Alternatively, the market for alternative candidates may have its greatest impact when performance is only slightly poor, as such performance may provide the greatest level of uncertainty regarding the CEO’s capabilities. Additionally, future research should continue to examine other avenues in which board members and firms identify alternative CEO candidates and how these candidates may impact the decision to dismiss a CEO. Finally, future research should continue to explore the concept of CEO reputation. CEO reputation may be a construct with multiple indicators that is developed over time and changes as new information is received. Future research should examine both how a reputation is developed and updated over time, as well as the consequences of CEO reputation on CEO, firm, and board decision making. Limitations While this study builds on previous theoretical and empirical work regarding dismissal, it is not without its limitations. First, this study utilizes only US based, publicly traded 69 manufacturing firms. The results identified in this study may not generalize to non-US firms or firms outside of manufacturing industries. However, these results are still important as they identify how a series of homogeneous firms examine the decision to dismiss a CEO over a 6 year time period and, in many cases, across multiple CEOs. Second, this study utilizes only proxies for information evaluated by the board. Additionally, this study does not examine the group decision making dynamics that occur in the decision to dismiss a CEO. Without exploring these dynamics, significant results may not be identified as group decision making dynamics alter how the information discussed in this study is evaluated. Despite this limitation, the results of this study do provide evidence that the board examines the human and reputational capital of a CEO when determining whether to dismiss a CEO. Conclusion While CEO dismissal has become relatively more common in recent years, few studies have moved beyond examining the role that power, politics, and performance play in the decision to dismiss a CEO. This study builds on a growing body of literature that examines this decision and utilizes signaling theory, as well as theories on human capital, executive reputation, and the market for managerial talent in order to explain how signals of executive capabilities can alter the board’s decision making process in determining whether to dismiss a CEO. Results of this study provide evidence that an executive’s human and reputational capital can buffer an executive from the board’s decision to dismiss the CEO regardless of performance. Future research should continue to understand this phenomena in greater detail in order to better understand how such decisions are made and the information that is evaluated in this important decision making process. 70 STUDY 2: WHAT HAPPENS TO DISMISSED EXECUTIVES: THE ROLE OF HUMAN, REPUTATIONAL, AND SOCIAL CAPITAL IN CEO RE-EMPLOYMENT Introduction Dismissals are an increasing phenomenon in the corporate world, with the frequency of dismissals occurring at the Chief Executive Officer level reaching new highs each decade (Huson et al., 2001). Poor performance, conflicts with top management, strategic disagreements with the board, and personality conflicts may all lead to dismissal of top executives. Despite the increasing prevalence of executive dismissal, little research examines what happens to executives and their careers following dismissal. This study examines the likelihood that dismissed CEOs regain employment as a top manager of a publicly traded firm within five years following dismissal. Dismissal from any employment position can be a difficult time in an individual’s life. However, the dismissal of a CEO may indicate that the executive failed in his or her duty to effectively lead a firm. Failure on the part of a CEO may indicate his or her inability to lead ethically, to establish appropriate standards of conduct in organization, or merely failure to achieve the level of performance desired by the board. However, each of these explanations may affect the likelihood that other organizations are willing to employ the dismissed CEO as a top manager going forward, leading to the question as to whether dismissal may represent a stigma on the dismissed executive. According to the psychology literature on stigmatization, a stigma is ‘an attribute that is deeply discrediting’, reducing the individual from a ‘whole and usual person to a tainted, discounted person’ (Goffman, 1963, p. 3). This denigration of the individual negatively impacts his or her reputation (Goffman, 1963) and forms a basis for reduced social interaction (Carter & 71 Feld, 2004). Past research on the likelihood of executive re-employment following dismissal from failing firms has identified that relatively few executives regain employment as a top manager within five years (e.g., Cannella, Fraser, & Lee, 1995; Gilson, 1989, 1990), suggesting that dismissal, especially under conditions of leading a failing organization, reduces the likelihood that other organizations are willing to interact with executives. The process of firing a CEO may ‘single out’ one individual as warranting the stain and denigration of the organization’s failure and assign blame (Wiesenfeld, Wurthmann, & Hambrick, 2008). This process may be difficult to overcome, as other potential employing organizations examining the executive’s background and capabilities will not have the same information the CEO’s prior firm had when deciding to dismiss the CEO. Potential employing organizations have limited visible information regarding the CEO, of which a major piece is that the CEO was fired. Thus, dismissal likely serves as a form of stigmatization relating to the individual’s re-employability as a top executive. Furthermore, the level of stigma attached to the executive’s dismissal is likely to be based in large part on the circumstances surrounding the executive’s dismissal. For instance, CEOs dismissed for violating fiduciary duty may be more likely to be associated with greater levels of stigma than CEOs dismissed for below average performance. Despite the potential for stigmatization and reduced social interaction in the form of gainful re-employment as a top manager for dismissed executives, several factors may enhance the ability of a CEO to overcome the stigma associated with his / her dismissal regardless of the circumstances of dismissal. First, an executive’s human capital, the stock of observable characteristics relating to the executive’s cumulative past experience and knowledge, may send a signal that the executive represents a quality candidate for employment as a top manager that can be successful in a new organization (e.g., Spence, 1973). Executives with significant 72 accumulated human capital may send a high quality signal to the marketplace that reasons for dismissal were an aberration and that such executives can be a significant asset in a new position. Second, executives with high levels of social capital may be able to utilize resources and ties to enhance their position (Belliveau, O'Reilly, & Wade, 1996). Greater accumulated social capital may also buffer executives from stigmatization, as social capital may alter how observers react to these executives (Wiesenfeld et al., 2008). Finally, an executive’s reputation in the marketplace may also send a signal that the executive is a worthy candidate for re-employment as a top manager. Reputation serves as the collective judgment of others regarding the executive’s capabilities, and executives with greater levels of reputational capital signal that others collectively evaluate the executive’s capabilities positively (Graffin & Ward, 2010; Rindova et al., 2005; Washington & Zajac, 2005). Human capital, social capital, and executive reputation can each uniquely provide information to the marketplace about an executive’s quality or be used as a resource by executives in order to reduce the harmful effects of dismissal and eventually find new employment as a top manager. Furthermore, an executive’s stock of human, reputational, and social capital may moderate the relationship between circumstances of dismissal and likelihood of re-employment, such that executives with higher levels of capital will be able to regain employment, even if dismissed for reasons of personal conduct violations or poor performance. This study examines what factors may allow executives to overcome the harmful effects and stigmatization of dismissal to regain employment as a top manager following dismissal making several contributions to the existing literature. First, this study examines the role that circumstances of dismissal play in the likelihood of executive re-employment. Differing circumstances of dismissal may result in differing likelihoods that firms are willing to re-employ 73 a dismissed executive as a top manager. In other words, the reasons for an executive’s dismissal may determine the degree of stigmatization of the executive. Second, this study argues that human, reputational, and social capital can enhance the likelihood that an executive regains employment as a top manager within five years of dismissal, suggesting that an executive can reduce the degree of stigmatization associated with dismissal. Finally, this study notes that different levels of capital can also affect the degree of stigma associated with individual circumstances surrounding a CEO’s dismissal, such that in the face of different circumstances of dismissal, human, reputational, and social capital can provide different effects on the likelihood of future employment. The next sections will examine why dismissal serves as a stigmatizing process, how circumstances of dismissal affect likelihood of future employment, and how executives buffer themselves from the effects of stigma, prior to examining data related to executive re-employment as a top manager following dismissal. EXECUTIVE DISMISSAL AND STIGMATIZATION Stigma at the individual level is ‘an attribute that is deeply discrediting’ and which reduces the bearer ‘from a whole and usual person to a tainted, discounted one’ (Goffman, 1963, p.3). A stigma results when an individual is perceived as belonging to a category that is viewed as a basis for disassociation (Leary & Schreindorfer, 1998). Most importantly, stigma negatively impacts the image and reputation of the associated individual (Goffman, 1963). Overall, those attached to a stigma are viewed as tainted (Pozner, 2008). Stigmas thus represent an attribute that make others less apt to deal with the associated individual (Carter & Feld, 2004). Executive dismissal can serve as a stigmatizing event through a process of singling out and shunning. Wiesenfeld and colleagues (2008) noted that stigmatization of executives can occur through the process of ‘singling out’, or assigning blame to one individual who is then 74 seen as culpable for failure and warrants denigration. According to the authors, this ‘singling out’ implicates the individual and his or her qualities directly. In the case of corporate failures, the CEO is most likely to be the individual singled out (Sutton & Callahan, 1987; Wiesenfeld et al., 2008). Once the singling out process occurs, individuals may be involved in a ‘shunning’ process. This process may include pressure on economic arbiters to reject the shunned individuals and pushes the executives to ‘the dark recesses of the business world’ (Wiesenfeld et al., 2008, p.242) resulting in devaluation of the executive’s capabilities. Once the individual has been shunned, hiring the tarnished person becomes at odds with requirements for legitimacy, even if the stigma is believed to be unfair (Wiesenfeld et al., 2008). Executive dismissal leads to the ‘singling out’ of the CEO as the individual to blame for the organization’s problems, regardless of whether the blame is just. A 2003 article in the Economist quotes an executive recruiting expert noting that fired executives are nearly unemployable (Cairncross, 2003). Dismissed executives will rarely ever run another public company and may be lucky to retain existing board seats. The recruiting executive even likens attempting to gain employment as a top manager following dismissal to returning from the dead. Dismissal is expected to harm managerial reputation and make it difficult to gain another position through a process of ‘settling up’ (Amihud & Lev, 1981; Fama, 1980b). Empirical examination of executive career paths following dismissal provides evidence that dismissal reduces the willingness of other parties to interact with the dismissed executives in a business capacity. Cannella, Fraser, and Lee (1995) find that fired executives in the Texas banking industry were unlikely to be rehired and those that do receive gainful re-employment often find it in much lesser capacities. Gilson (1989) examines firms going through bankruptcy and finds that none of the managers from the bankrupt firms he studied were found in 75 management posts in other firms within 3 years of leaving the distressed firm. Similarly, an analysis of four bankrupt firms in the computer industry finds that managers suffer severe losses in reputation and self-esteem following bankruptcy filings (Sutton & Callahan, 1987). Fee and Hadlock (2004) find that almost 40 percent of executives under the age of 50 that depart organizations eventually were found in another executive capacity (e.g. listed on firm proxy statements); however, those that did find re-employment usually find it in inferior positions. In the most comprehensive examination of executive rehiring following dismissals, Ward, Sonnenfeld, and Kimberly (1995) find that 39 of 60 former CEOs in their sample do not resume a career as an executive. The authors note that dismissed executives have a negative stigma associated with them as damaged goods and appear as unsuitable to executive recruiting firms, who assist boards when examining candidates for executive positions. Taken together, these results provide significant evidence that the dismissal of a CEO provides a mark that discredits an executive’s career and reduces the likelihood that other firms will rehire the executive in a significant position. Despite these results, it is possible that an executive can regain employment as a top manager. For instance, Ward and colleagues (1995) find 21 of 60 former CEOs did regain executive re-employment. Furthermore, Cannella, Fraser, and Lee (1995) note that managers of Texas banks that were seen as unable to control the situation were likely to gain re-employment, while a limited subset of managers of failing banks also gained executive re-employment. These results provide some hope for dismissed managers that the stigma of dismissal can be overcome and the executive can gain re-employment as a top manager. While few managers may regain employment as an executive following dismissal, research has shown that some do, leading us to ask why can some regain employment when others cannot. 76 THE EFFECTS OF DISMISSAL CIRCUMSTANCES ON EXECUTIVE RE- EMPLOYMENT AS A TOP MANAGER The circumstances surrounding an executive’s dismissal may alter the degree of stigmatization associated with dismissal. Executives dismissed following poor performance may escape severe stigmatization, as attributions regarding managerial contributions to firm performance are often noisy and may not be the manager’s fault (Holmstrom, 1982). Alternatively, executives dismissed for personal conduct violations may provide greater levels of stigma, such that firms may be unlikely to associate with individuals who were dismissed for violations relating to personal conduct (Sigal, Hsu, Foodim, & Betman, 1988). The most serious potential violation may be a violation of fiduciary duty. An executive violating his/her fiduciary duty, or a violation of the trust and confidence of the firm’s shareholders, while CEO may never be able to regain position as an executive, as the violation indicates the executive did not uphold his/her legal obligations to shareholders. As noted in Figure 6, the circumstances surrounding an executive’s dismissal should directly affect the CEO’s ability to regain employment as an executive. Executives dismissed for reasons of poor performance, violations relating to personal conduct, or violations relating to fiduciary duty are less likely to gain re-employment as a top manager. It is expected, however, that violations of fiduciary duty are likely to be associated with the largest level of stigma on an executive. --------------------------------------- Insert Figure 3 about here --------------------------------------- 77 Despite the potential for stigma following dismissal, executives can rebuild their stature and buffer themselves from the effects of dismissal in three primary fashions. First, executives can point to signals of their quality and capabilities as an executive through the accumulation of human capital. Human capital can serve as a signal of the executive’s capabilities through endorsements from past actors and indication of the others’ willingness to employ or interact with the executive. Second, the executive’s reputation may signal his or her qualities based on past performance and also on the evaluation of others. When others evaluate the executive highly, the executive may be able to retain a strong reputation regardless of dismissal circumstances. Finally, executives can utilize accumulated social capital to call on favors from other parties, including friends (Westphal, 1999), and may also alter how others perceive the actor through their involvement in high status social circles (Wiesenfeld et al., 2008). Integration into the corporate elite and accumulation of social capital will alter judgments of the executive as members would judge the executive on more than the instance of one corporate failure (Wiesenfeld et al., 2008). Human, reputational, and social capital may be particularly important when the circumstances of dismissal are taken into account, suggesting that capital moderates the relationship between circumstances of dismissal and likelihood of employment as a top executive. This relationship is likely to vary depending on the circumstances of dismissal, such that an executive’s capital may provide a stronger buffer from stigmatization under conditions of poor performance than under conditions of dismissal due to a violation of fiduciary duty. Performance Related Dismissal The performance of the executive’s firm for which he or she was responsible is likely to be a meaningful indicator of the executive’s quality, such that firms that perform well should 78 indicate higher quality executives (Fee & Hadlock, 2004). Managers in superior performing firms know how to win and succeed and also increase the likelihood of a CEO or other executive finding gainful employment as a top manager in other firms. These managers have established a track record for performance which will be rewarded by the market (Fama, 1980a). Poor performance may indicate, though not necessarily, poorer quality executives. Firms may be wary of re-hiring managers whose prior firms performed poorly. For instance, evidence has shown that managers overseeing firms in financial distress or bankruptcy are much less likely to find employment in an executive capacity elsewhere (e.g., Gilson, 1989, 1990; Sutton & Callahan, 1987). Ward, Sonnenfeld, and Kimberly (1995) find that executives dismissed for reasons of poor performance were significantly less likely to regain active executive positions elsewhere. An executive’s reputation is based in part upon past performance and actions (Milbourn, 2003). While these actions may be difficult to observe and determine (e.g., Holmstrom, 1982), executives who run firms that perform poorly firms may be signaling their lack of quality and capabilities to successfully manage an organization. These executives will need greater levels of human, social, and reputational capital to signal the market of their capabilities and their ability to be successful in the future. In order for performance of the executive’s dismissing firm to matter, however, the firm must also indicate that the CEO was dismissed for performance-related reasons. Thus, I argue that performance related dismissals are likely to decrease the executive’s prospects for future re-employment more than dismissal due to reasons that are unspecified at the time of the dismissal. 79 Hypothesis 34: Executives dismissed for performance related reasons will be less likely to regain re-employment as an executive than executives not dismissed for performance related reasons. Violations of Fiduciary Duty Ethical and legal standards are often viewed with great significance. Therefore, violation of ethical and legal standards is often viewed with particular indignation and results in a large backlash against the violator (Tetlock, 2002). In the case of CEOs, violation of ethical and legal standards may lead to punishment or singling out of specific events leading to stigmatization (Wiesenfeld et al., 2008). One great source of potential outrage is the committing of unethical behavior, such as leading to earnings restatements, committing fraud, or shirking on the job (Wiesenfeld et al., 2008). In these cases, the executive has not proven his or her lack of competence, but rather has proven he or she lacks integrity. At the individual level, trust is much more difficult to restore for those perceived to lack integrity than for those who lack competence (Kim, Ferrin, Cooper, & Dirks, 2004). Thus, executives who commit legal or ethical violations are expected to be much more likely to not only lose their position as CEO, but also to be shunned by the market from gaining future employment as a top executive. For instance, coaches in college athletics who violate NCAA rules and commit ethical violations are significantly less likely to gain re-employment, and when opportunities arise they are at smaller or less prestigious universities and programs. The key argument here, however, is not just that illegal or unethical behavior occurs at an organization, but also that the executive’s dismissal is linked to it. Linking the executive to the behavior creates the stigmatization and warns other organizations not to trust the executive. These types of violations relate directly to a violation of the CEO’s fiduciary duty to properly manage shareholder value. Failing to uphold an 80 executive’s fiduciary duty exhibits the individuals’ lack of trustworthiness, as such actions represent opportunistic behavior on the part of the CEO at the expense of shareholders. The limited empirical literature on this topic provides some evidence that CEOs fired for ethical or legal violations are significantly less likely to find future employment as a top manager. For instance, Desai, Hogans, and Wilkins (2006) find that managers of firms with earnings restatements are not likely to find future jobs and those that do suffer a deterioration in job quality. Managers involved in scandal are also significantly less likely to find re- employment as a top manager, with scandals representing a far worse case than other types of forced departures examined (Fee & Hadlock, 2004). Finally, Ward, Sonnenfeld, and Kimberly (1995) also note that executives who were fired for illegal behavior were significantly less likely to find gainful re-employment than other executives. These findings suggest that executives responsible for ethical or legal violations at a prior firm are stigmatized, which leads to shunning by other organizations who choose not to associate with such executives. Therefore, I expect that violations of fiduciary duty will make it especially difficult to gain future employment as a top executive. Hypothesis 35: Executives dismissed for violations of fiduciary duty will be less likely to be re-employed as an executive than executives not dismissed for violations of fiduciary duty. Personal Conduct Violations Ethical and legal violations represent that the executive has not acted in accordance with the desires of his or her firm’s primary stakeholders in many cases. However, executives may also be dismissed for violations relating to personal conduct that may place the organization in a negative light. Violations of personal conduct often relate to moral and legal standards and are 81 also an important consideration with regards to executives, as CEOs are often the public face of an organization. Violations of moral standards are often also viewed in a very negative light and may yield punitive action against the violator (Tetlock, 2002). Violations of personal conduct do not relate directly to an organization’s operations, but may include illegal behavior on the part of the executive (e.g. arrested for behavior outside of the office) or personal conduct that violates societal norms (e.g. adultery, falsifying a resume). Previous research has noted that dismissed executives involved in both personal mismanagement or executive scandal are less likely to find re-employment in an executive capacity (Fee & Hadlock, 2004; Ward et al., 1995). The reduced likelihood of hiring an executive stigmatized by personal scandal is likely as scandals bring negative publicity about the individual, which creates tarnish or stain with which the organization may not want to be associated. Furthermore, stakeholders may pressure the organization not to be involved with an executive whose behaviors are not in line with their expectations of behavior. In these situations, firms may shy away from hiring stigmatized managers with personal character concerns in order to avoid having personal problems tarnish the firm’s reputation. Hypothesis 36: Executives dismissed for reasons of personal conduct violations will be less likely to regain re-employment as an executive than executives not dismissed for personal conduct violations. Degrees of Stigmatization based on Dismissal Circumstances While each set of circumstances leading to dismissal is expected to yield some degree of stigmatization with regards to the executive, each set of circumstances is also likely to be viewed differently when evaluated by other stakeholders. With regards to performance, attributions for performance are often noisy and include a variety of external factors that alter performance, 82 making it easier for managers to attribute performance to factors outside their control (Holmstrom, 1982; Lieberson & O'Connor, 1972). In situations where executives may not be in control, studies have shown that executives are more likely to regain employment (Cannella et al., 1995). In other words, executives may be able to deflect blame for poor performance onto other factors, such as industry conditions, and illustrate their other qualities in order to improve how others view the CEO’s ability. Executives attempting to bounce back from personal violations and violations of fiduciary duty may face a different story. Executives who violate personal standards of conduct, such as getting into legal trouble, may be able to re-gain employment; however, this process will be more difficult, as blame for past troubles cannot be shifted. These executives, however, should be more likely to regain employment than executives committing violations of fiduciary duty, as violations of fiduciary duty call into question an executive’s ability to exert a high standard of care on behalf of the shareholders and cannot be entrusted with managing shareholder value. Violations of personal conduct, while associated with the individual, merely indicate the individual made a personal mistake. However, this mistake may be overcome and the individual can atone for the violation and provide significant value to an organization. That is, through statements of apology and remorse, or through behaviors shown to indicate penance for the culpability of their actions, executives dismissed due to personal conduct violations may prove they are worth being provided a second chance. Violations of fiduciary duty, however, call into question whether an executive can ever lead a company with integrity again. In such situations, publicly traded firms are unlikely to associate with such executives. 83 Hypothesis 37: Executives dismissed for reasons of poor performance or personal conduct violations will be more likely to gain re-employment as an executive than executives dismissed for violations of fiduciary duty. EXECUTIVE CAPITAL AS A BUFFER FROM STIGMA’S EFFECTS While dismissal can clearly have negative effects on an executive’s future career prospects (e.g., Gilson, 1989, 1990; Houston & James, 1993; Ward et al., 1995), executives may also use accumulated human, reputational, and social capital in order to regain employment following dismissal. In the dark times following dismissal, leaders need to rally friends and acquaintances in order to rebuild their stature and reputation (Sonnenfeld & Ward, 2007). Human Capital and Executive Stigma Human capital, such as the executive’s education and experience, represents past actions and accomplishments, but can also signal future performance in the labor market (Fulmer, 2009; Rosenbaum, 1984). These observable characteristics can help employers identify the likelihood of success an individual may have in their organizations, as well as provide information regarding the expertise an individual may bring if hired (Spence, 1973). Human capital can help buffer individuals from the effects of stigmatization by signaling the executive’s quality, which may illustrate the expected future gains from the employment relationship. More importantly, executives with significant accumulation of human capital, including having worked for prestigious organizations or having attended elite universities, may gain entry into the network of corporate elites through this status (Westphal & Stern, 2006). This network of corporate elites may shield the executive from the effects of stigma. A variety of human capital attributes may help provide this buffer and continue to serve as signals of the executive’s quality even after dismissal that may allow organizations to employ 84 the executive as a top manager, including the executive’s knowledge in an industry, prior experience with high status actors, educational level and prestige of the executive’s educational institutions attended. Experience with Prestigious Organizations. Previous affiliation with prestigious organizations may impact a dismissed executive’s future job prospects. Working for prestigious organizations can help enhance future job prospects and allows individuals to make transitions to other organizations more successfully (Hamori, 2006). Organizations that are prestigious and reputable instill ‘career imprints’ on employees that are valued by other organizations in the market (Higgins, 2005). Affiliation with prestigious firms increases an executive’s prominence as the association enables future employers to assume that such industry leaders have positively evaluated the executive (Stuart, 2000). Specifically, reputable organizations facilitate development of employee capabilities by hiring the most talented individuals, providing better mentors, and training and immersing the individuals in the company’s superior industry and product knowledge (Crane, 1965; Higgins & Gulati, 2003; Long, 1978). This experience signals quality in the executive in terms of both a reputable third-party’s endorsement of the candidate and the candidate’s own development through involvement in and understanding of a leading firm’s operations. Experience as an executive with high status or prestigious organizations may help buffer the executive from stigmatization. This experience signals to the market that high status actors evaluated the executive positively. This evaluation indicates that the executive has knowledge, capabilities, and/or expertise that were valuable to a prestigious organization and that may be valuable to other organizations in the future. Furthermore, dismissed executives with previous experience in prestigious firms are more likely to have ties with members of the corporate elite 85 who are willing to assure others of the executive’s capabilities when attempting to gain re- employment as an executive. Hypothesis 38: Dismissed executives with executive experience in a prestigious firm are more likely to be re-employed as an executive than dismissed executives who do not have executive experience in a prestigious firm. Executive Education Level. A manager’s education influences the experiences, both functional and otherwise, that the manager acquires over time (Castanias & Helfat, 2001). Education has been shown to be a significant predictor of compensation in previous studies (e.g., Agarwal, 1981; Fisher & Govindarajan, 1992), indicating that the market values the education of employees. Executives with greater levels of education signal higher quality and greater ability, as greater levels of education can help instill new perspectives, greater knowledge, and new expertise in executives. When organizations assess managerial quality, educational level serves as a signal of cognitive orientation for future decision making and performance. Thus, I argue that individuals with greater levels of education will have a greater chance of achieving re- employment as a top manager in the future. Hypothesis 39: Dismissed executives with greater levels of education are more likely to be re-employed as an executive than dismissed executives who have lesser levels of education. Prestige of the Executive’s Education. Education at prestigious institutions also yields additional human capital which may send a strong signal to the market when evaluating dismissed executives. Education from elite universities creates greater credibility and prestige than association with less visible schools (Baltzell, 1989; Clement, 1977; Domhoff, 1967). Promotion to top corporate positions has been noted to be more influenced by executives who 86 earned a degree from an elite university (Useem & Karabel, 1986). Graduation from an elite institution may not only yield a perceived higher quality education, but also provides access to an extensive network of successful contacts. Finally, graduation from an elite institution signals that the institution was willing to accept the CEO as a candidate and that the CEO had the ability to complete the rigorous process to receive a degree. Graduation from a prestigious institution can therefore help to buffer an executive from the effects of the stigma of dismissal in three ways. First, graduation from a prestigious institution serves as a high quality signal about an executive’s capability in that the institution was willing to accept the CEO. This endorsement serves as a signal to the marketplace that the institution had belief in the executive’s abilities. Second, graduating from a prestigious institution signals that the executive has the knowledge and capabilities to complete a rigorous educational program. This signal provides evidence that the executive is willing to work hard, but also has obtained the requisite knowledge and skills to meet the institution’s degree requirements. Finally, graduation from a prestigious institution also grants access into the network of elite who have also graduated from the institution. Dismissed executives may utilize this network to buffer themselves from the effects of stigma by calling upon others for employment as an executive or by vouching for the executive’s capabilities and expertise to others following dismissal. Hypothesis 40: Dismissed executives who have graduated from an elite educational institution are more likely to be re-employed as an executive than dismissed executives who have not graduated from an elite educational institution. 87 Reputational Capital and Executive Stigma An executive’s reputation is the result of information exchanges and social influences among various actors (Rao, 1994; Rindova & Fombrun, 1999). The continued appraisal of an actor’s performance is one subset of the executive’s reputation that affects how others view the quality or capabilities of the executive (Carter & Ruefli, 2006). However, attributions of managerial ability are often difficult to make as the market has difficulty in determining how much of an organization’s performance can be attributed to the executive due to systematic risk factors at the organizational and industry level (Holmstrom, 1982). However, executive reputation can also be enhanced through proxies or signals that allow others to make assumptions about the quality of the executive (Fombrun & Shanley, 1990; Kreps & Spence, 1985; Rao, 1994). In this regard, reputation becomes a socially constructed entity that may be based on the outcome of legitimation processes (Rao, 1994). In particular, certification contests can provide a means of reducing uncertainty regarding the quality of an executive (Graffin & Ward, 2010; Rao, 1994; Rindova et al., 2005). When the marginal contribution of executives is difficult to assess, powerful stakeholders will look to certification contests to provide useful cues about the executive’s ability (Wade et al., 2006). Certification contests refer to competitions in which actors in a given domain are ranked based upon performance criteria that have been accepted previously as credible or legitimate by stakeholders (Wade et al., 2006). These contests generate information from reputable observers that evaluate and endorse certain actors in a given domain (Rao, 1994; Wade et al., 2006) and influence evaluations of an individual’s quality (Rindova et al., 2005). At their heart, certification contests allow for clear and comparable attributions of an actor’s quality or capabilities vis-à-vis others’ relative worth or standing (Elsbach & Kramer, 1996). Certifications 88 that are awarded by experts and independent arbiters confer a positive reputation on the certified individual or organization and lead to increased prestige power (Wade et al., 2006). With regard to CEOs, certifications are likely to be one of the few neutral and independent sources of information regarding a CEO’s contribution and capabilities (Wade et al., 2006). These contests can also reduce the ambiguity surrounding the executive’s contribution to potential performance of his or her previous organizations (Graffin & Ward, 2010). The effects of certification contests not only provide quality signals to an executive’s current employer, but also send a signal to the entire market that the executive is of high quality or provides substantial contributions to his or her organization. Thus, I argue that an executive’s reputation should have an impact on prospects for future employment in an executive capacity for several reasons. First, given that certification contests send a signal regarding the positive contributions of the executive, the executive should maintain credibility and reputational capital even after dismissal, as independent experts have highly valued the executive’s capabilities. Second, firms looking to hire dismissed executives can point to the certification contest’s endorsement of the executive as proof that the executive can bring valuable capabilities and expertise to the firm, even after the executive has been previously dismissed. This allows the potential hiring firm to downplay any loss in legitimacy that may be associated with dealing with a stigmatized or shunned executive. Third, information that negatively influences an executive’s reputation may reduce the likelihood of re-employment, as firms may not be willing to associate with executives who are negatively viewed by external parties. Hypothesis 41: Dismissed executives with greater positive reputational capital are more likely to be re-employed as an executive than dismissed executives with lower levels of reputational capital. 89 Social Capital and Executive Stigma An executive’s social capital may provide additional benefits beyond his or her human capital that can buffer the executive from the harmful effects of stigmatization. Social capital refers to resources available through both social networks and institutional ties that individuals can use to enhance their positions (Belliveau et al., 1996). Sociologists view social capital in terms of the benefits that actors can obtain through the social ties they develop over time (Burt, 1992; Coleman, 1990; Portes, 1998). Social capital may provide information, influence, and solidarity resources to prevent the shunning process from occurring for dismissed executives (Adler & Kwon, 2002). Specifically social capital can help reduce the harmful effects of dismissal on an executive’s future job prospects in several ways. First, social capital through personal contacts and network ties can allow for the executive to call on favors from others within his or her network (Adler & Kwon, 2002; Useem & Karabel, 1986) which can be crucial to advancement. Sonnenfeld and Ward (2007) note that our society has a norm of reciprocity where people can pay forward help they have received from others and seek opportunities to do favors for those in their network in case they need help in the future. Second, CEOs with higher levels of prestige through social capital may be seen as more competent, credible, or trustworthy (D'Aveni, 1990; Giordano, 1983). Third, executives with higher status may be excused for past poor performance as others may believe their attention was absorbed by other issues (Carson, 1980). Fourth, as executives become more ingrained in the corporate elite, other members of the elite are likely to judge them for more than one instance of failure (Wiesenfeld et al., 2008). Social capital can thus alter the way that other observers react to the dismissed executive. As executives gain greater social capital they have more individuals and network resources to call upon in times of 90 need. Finally, organizations may value executives with significant levels of social capital as they may provide access to new resources or valuable knowledge bases to which the organization did not have already have access (Brown & Duguid, 1998). Furthermore, firms may gain legitimacy through hiring executives with previous CEO experience to serve as a member of the executive team. Each of these arguments highlights the importance that significant external relationships may play in buffering an executive from the effects of dismissal on future employment prospects. Specifically, I argue that the size of the network through directorship contacts and relationships with other firms, as well as the location of the executive’s dismissing firm can all influence whether or not the executive can obtain another executive position in the future. Number of Directorships. An executive’s links to others in a social network can often be built and strengthened through direct contact via board memberships. Serving on the board of directors for another firm allows the firm’s other directors, as well as the firm’s management team, to assess the qualities and the capabilities of the executive (Khurana, 2002). Prior acquaintanceship through director ties allows for other directors to vouch for the executive’s capabilities and his or her qualities as a director (Khurana, 2002). As executives gain access to more directorships, the size of their network grows, increasing their potential social capital. These networks can also provide access to information regarding new opportunities and better career outcomes (Granovetter, 1995). Greater numbers of directorships can help executives overcome dismissal by allowing them to call in favors with individuals in their network when their career is at risk (Wiesenfeld et al., 2008). Given that the best jobs are likely to be filled through personal contacts (Granovetter, 1995), the greater the size of an executive’s network, the more likely he or she is able to use the 91 network to help find future employment in an executive capacity. Specifically, directorships provide a better way for executives to develop contacts and social capital that will be useful in the future when seeking executive re-employment. Other directors tend to either serve as directors or executives for other firms, who are influential when seeking external candidates for executive positions and vouching for the qualities and capabilities of executive candidates. Hypothesis 42: Dismissed executives with greater numbers of external directorships are more likely to be re-employed as an executive than dismissed executives with fewer numbers of external directorships. Number of Executive Positions. In addition to directorships, the prior experience of the candidate at other firms also increases the size of an executive’s network. Executives who have served in the capacity of an executive (CEO or otherwise) for more firms will have a greater number of personal contacts from those firms they can potentially call in when in times of need. These executives will have a greater number of ties to other executives who are influential in the hiring decisions at their firms (Granovetter, 1995). Furthermore, the familiarity of the individuals through ties related to executive employment may allow individuals in the executive’s social network to more personally vouch for the executive’s capabilities and reduce the potential for adverse selection at a firm choosing to hire a dismissed executive. As the size of the executive’s network increases, the executive is more likely to be able to utilize this capital in order to find alternative positions in the future. Hypothesis 43: Dismissed executives with greater numbers of executive positions in their career are more likely to be re-employed as an executive than dismissed executives with fewer numbers of executive positions. 92 Geographical Location. An executive’s physical location may also enhance the value of existing social networks and ties that have been developed or provide access to additional networks. While social networks may facilitate transactions worldwide, ideas particularly flow through denser networks in local areas (Davis & Greve, 1997; Marquis, 2003). Local boards tend to be dominated by a relatively small group of executives in a given area and local colleagues tend to be offered these seats when they become available (Ward & Feldman, 2008). Friedland and Palmer (1994) find that geographic proximity is an important predictor of invitations to join boards of directors, as 27 percent of ties among Fortune 1000 firms were from companies headquartered in the same state. This logic suggests that ties are likely to be reconstituted among directors of firms headquartered in the same area (Palmer et al., 1986). Geographical location of the executive can thus play a role in evaluating whether an executive can regain executive status following dismissal. Given the above arguments, executives whose corporations were headquartered in larger metropolitan locations should have access to a larger social network of other elite executives and influential colleagues. Dismissed executives in larger metropolitan areas more likely have had greater exposure to influential executives and economic arbiters to reduce the effects of stigmatization on the dismissed executive. Thus, geographic location should serve in helping certain executives enhance their social capital and reduce the effects of past job loss on future employment prospects. Hypothesis 44: Dismissed executives whose position as CEO was in a major city are more likely to be re-employed as an executive than dismissed executives whose position as CEO were in smaller markets. THE INTERACTION OF DISMISSAL CIRCUMSTANCES AND EXECUTIVE CAPITAL 93 While human, reputational, and social capital may send quality signals to the market about a dismissed executive’s capabilities and likelihood of future success, the circumstances surrounding an executive’s dismissal may limit the effectiveness of these signals. In some cases, the degree of stigmatization associated with the executive’s dismissal may not be overcome by the amount of capital accumulated by the executive. For instance, employers may be unwilling to hire an executive fired for embezzling money from a prior firm even if that executive had a strong reputation as a CEO, graduated from elite institutions, and had work experience a large number of prestigious institutions. In these circumstances, the signal sent by the circumstances surrounding the executive’s dismissal may be significantly stronger than the executive’s own accumulated human, reputational, and social capital. Performance Related Dismissal and Executive Capital Executives who are fired for performance related reasons will encounter significant hurdles to gaining re-employment, as poor performance indicates a lower quality executive (Fee & Hadlock, 2004). When dismissal is associated with poor firm performance, the executive is associated with poor personal performance and tied to the lack of success of his or her firing firm. Specifically, when firms face financial distress or bankruptcy, executives are much less likely to find future employment as a top manager (Gilson, 1989, 1990; Sutton & Callahan, 1987). Performance of an executive’s prior firm sends a signal separate from the executive’s capital about the executive’s ability to effectively manage a firm. However, alternative signals regarding an executive’s capabilities may help provide differing cues regarding the executive’s potential contribution to a new organization. An executive’s human, reputational, and social capital may help a new organization attribute prior poor performance to factors outside the executive’s control, such that executives with higher levels of human, reputational, and social 94 capital will be more likely to gain re-employment as a top manager in the future, even in the face of poor performance. In these situations, executives may be able to use the evaluations of others or call in favors through social capital in order to persuade organizations to hire the dismissed executive in the capacity of a top manager. Hypothesis 45: A dismissed executive’s prior experience in a prestigious organization will moderate the relationship between performance related CEO dismissal and likelihood of re-employment as an executive such that dismissed executives with experience in a prestigious firm will be more likely to gain re-employment as an executive than dismissed executives without experience in a prestigious firm. Hypothesis 46: A dismissed executive’s level of education will moderate the relationship between performance related CEO dismissal and likelihood of re-employment as an executive such that dismissed executives with greater levels of education will be more likely to gain re-employment as an executive than dismissed executives with lower levels of education. Hypothesis 47: A dismissed executive’s education at a prestigious institution will moderate the relationship between performance related CEO dismissal and likelihood of re-employment as an executive such that dismissed executives with education at a prestigious institution will be more likely to gain re-employment as an executive than dismissed executives without education at a prestigious institution. Hypothesis 48: A dismissed executive’s reputation will moderate the relationship between performance related CEO dismissal and likelihood of re-employment as an executive such that dismissed executives with a more positive reputation will be more likely to gain re-employment as an executive than dismissed executives with a less positive reputation. 95 Hypothesis 49: A dismissed executive’s number of directorships will moderate the relationship between performance related CEO dismissal and likelihood of re- employment as an executive such that dismissed executives with a greater number of directorships will be more likely to gain re-employment as an executive than dismissed executives with fewer directorships. Hypothesis 50: A dismissed executive’s number of previous executive positions will moderate the relationship between performance related CEO dismissal and likelihood of re-employment as an executive such that dismissed executives with a greater number of previous executive positions will be more likely to gain re-employment as an executive than dismissed executives with fewer previous executive positions. Hypothesis 51: A dismissed executive’s location in a major city will moderate the relationship between performance related CEO dismissal and likelihood of re- employment as an executive such that dismissed executives located in a major city will be more likely to gain re-employment as an executive than dismissed executives not located in a major city. Violations of Fiduciary Duty and Executive Capital Violations of fiduciary duty are likely to be associated with the greatest degree of stigma on an executive’s career. Such violations indicate the executive is untrustworthy, acts opportunistically, or intentionally mismanaged or mishandled shareholder money. Such executives are likely never to regain employment in an executive position, given the loss of legitimacy and reputation that would be associated with hiring such an offender. However, executives having been dismissed due to violations of fiduciary duty may be more likely to be employed by organizations under several potential circumstances. First, executives dismissed 96 for violating fiduciary duty may be able to leverage significant social capital in order to call in favors to gain re-employment. If dismissed executives are well tied into social networks, these networks may facilitate re-employment for another member of the ‘inner circle’ and shield the individual from the effects of stigmatization. Second, executives with higher levels of human capital may attempt to distance themselves from the effects of the stigma associated with the violation of fiduciary duty. In such situations, executives may argue that they were unaware certain actions were occurring (e.g. accounting misstatements, organizational fraud). Executives with higher levels of capital accumulated may argue that they are both an executive that acts with integrity, but also one that exhibits higher quality. If executives can decouple themselves from the violation of fiduciary duty, another organization may be willing to hire the executive in a top management capacity. Hypothesis 52: A dismissed executive’s prior experience in a prestigious organization will moderate the relationship between dismissal due to violations of fiduciary duty and likelihood of re-employment as an executive such that dismissed executives with experience in a prestigious firm will be more likely to gain re-employment as an executive than dismissed executives without experience in a prestigious firm. Hypothesis 53: A dismissed executive’s level of education will moderate the relationship between dismissal due to violations of fiduciary duty and likelihood of re-employment as an executive such that dismissed executives with greater levels of education will be more likely to gain re-employment as an executive than dismissed executives with lower levels of education. Hypothesis 54: A dismissed executive’s education at a prestigious institution will moderate the relationship between dismissal due to violations of fiduciary duty and 97 likelihood of re-employment as an executive such that dismissed executives with education at a prestigious institution will be more likely to gain re-employment as an executive than dismissed executives without education at a prestigious institution. Hypothesis 55: A dismissed executive’s reputation will moderate the relationship between dismissal due to violations of fiduciary duty and likelihood of re-employment as an executive such that dismissed executives who have won a certification contest will be more likely to gain re-employment as an executive than dismissed executives who have not won a certification contest. Hypothesis 56: A dismissed executive’s number of directorships will moderate the relationship between dismissal due to violations of fiduciary duty and likelihood of re- employment as an executive such that dismissed executives with a greater number of directorships will be more likely to gain re-employment as an executive than dismissed executives with fewer directorships. Hypothesis 57: A dismissed executive’s number of previous executive positions will moderate the relationship between dismissal due to violations of fiduciary duty and likelihood of re-employment as an executive such that dismissed executives with a greater number of previous executive positions will be more likely to gain re-employment as an executive than dismissed executives with fewer previous executive positions. Hypothesis 58: A dismissed executive’s location in a major city will moderate the relationship between dismissal due to violations of fiduciary duty and likelihood of re- employment as an executive such that dismissed executives located in a major city will be more likely to gain re-employment as an executive than dismissed executives not located in a major city. 98 Personal Conduct Violations and Executive Capital An executive’s accumulated capital may not only assist in finding re-employment when dismissal was related to performance, but may also enhance job prospects when dismissal occurs due to problems relating to personal conduct. While dismissal due to personal conduct problems may decrease the likelihood of re-gaining employment as a top manager, an executive’s capital may increase the likelihood other organizations are willing to take a chance and hire the executive for several reasons. First, accumulated human, reputational, and social capital can signal the executive’s quality, which allows the organization to point to the positive attributes of the executive, reducing the negative elements of the stigma. Organizations can argue that they are enhancing shareholder value by hiring a capable executive to serve in a top management position. Second, organizations may be willing to forgive an individual’s personal issues when the individual seeks forgiveness if the new hiring firm has individuals with a personal relationship with the executive. This social capital can allow for the dismissed executive’s network ties to vouch for the personal characteristics of the dismissed executive given past interpersonal interactions. Taken together, these arguments suggest that executives dismissed for reasons of personal conduct violations can have a significantly higher probability of re- employment as an executive if they have accumulated greater stakes of human, reputational, and social capital. Hypothesis 59: A dismissed executive’s prior experience in a prestigious organization will moderate the relationship between dismissal due to personal conduct violations and likelihood of re-employment as an executive such that dismissed executives with experience in a prestigious firm will be more likely to gain re-employment as an executive than dismissed executives without experience in a prestigious firm. 99 Hypothesis 60: A dismissed executive’s level of education will moderate the relationship between dismissal due to personal conduct violations and likelihood of re-employment as an executive such that dismissed executives with greater levels of education will be more likely to gain re-employment as an executive than dismissed executives with lower levels of education. Hypothesis 61: A dismissed executive’s education at a prestigious institution will moderate the relationship between dismissal due to personal conduct violations and likelihood of re-employment as an executive such that dismissed executives with education at a prestigious institution will be more likely to gain re-employment as an executive than dismissed executives without education at a prestigious institution. Hypothesis 62: A dismissed executive’s reputation will moderate the relationship between dismissal due to personal conduct violations and likelihood of re-employment as an executive such that dismissed executives with a more positive reputation will be more likely to gain re-employment as an executive than dismissed executives with a less positive reputation. Hypothesis 63: A dismissed executive’s number of directorships will moderate the relationship between dismissal due to personal conduct violations and likelihood of re- employment as an executive such that dismissed executives with a greater number of directorships will be more likely to gain re-employment as an executive than dismissed executives with fewer directorships. Hypothesis 64: A dismissed executive’s number of previous executive positions will moderate the relationship between dismissal due to personal conduct violations and likelihood of re-employment as an executive such that dismissed executives with a greater 100 number of previous executive positions will be more likely to gain re-employment as an executive than dismissed executives with fewer previous executive positions. Hypothesis 65: A dismissed executive’s location in a major city will moderate the relationship between dismissal due to personal conduct violations and likelihood of re- employment as an executive such that dismissed executives located in a major city will be more likely to gain re-employment as an executive than dismissed executives not located in a major city. METHODS Sample The sample for this study was drawn from all publicly traded firms not in industries related to banking or finance with more than 1000 employees. All CEO turnover events for each firm identified were researched for the years 2005 and 2006. This time period was selected in order to utilize data on CEO reputation, as well as examine a five year time period following dismissal to determine the likelihood of regaining executive employment. The initial sample consisted of 2,256 firm-year observations covering 1,267 firms. From this sample, 299 CEO turnover events were identified, which allowed for the determination by an independent coder of whether the CEO was dismissed or departed for other reasons. Identifying dismissals of CEOs represents a significant challenge in management research, as firms do not often disclose why CEOs resign or depart (Denis & Denis, 1995; Fredrickson et al., 1988; Shen & Cannella, 2002). However, recent research has identified several techniques in order to better identify dismissals that have more successfully helped identify when CEOs resign or are fired (e.g., Shen & Cannella, 2002). Given the success of these approaches at identifying dismissals (e.g., Shen & Cannella, 2002; Zhang, 2006, 2008), this 101 study follows the initial approach used by Shen and Cannella (2002) relying on news reports and descriptions of the event surrounding the departure of a CEO from a sample firm. News reports were examined to understand the reasons for all executive successions. Successions that were the result of a CEO’s death, health reasons, acceptance of a similar position at another firm, or a merger or acquisition were not coded as dismissals. Three criteria were then used to identify instances defined as dismissals. First, a CEO may be directly reported as fired or forced out; however, this is rare. Second, a CEO is reported as resigning under pressure, unexpectedly, or immediately, due to poor performance, personal reasons that are not indicated, or desires to pursue “other interests.” Third, cases where CEOs choose early retirement, but discussion of performance problems is reported in the announcement, were classified as dismissals. As noted by Shen and Cannella (2002), CEOs may have little choice but to resign when pressure surrounding their performance is high. Additionally, this study builds on Shen and Cannella’s criteria for dismissal by identifying any situation where problems of any kind (e.g. SEC investigation, accounting restatements) exist as a dismissal, in addition to discussion of performance problems. CEO turnovers were classified as a dismissal if any of the above criteria are met. After analyzing the sample of 299 turnover events for reasons relating to dismissal, 88 of the CEO turnovers identified were classified as dismissals. All firm information and performance information was gathered from CompuStat and firm proxy statements. Identification of CEOs each year was done through the Corporate Library database and supplemented by ExecuComp. Information on CEOs was gathered through Corporate Library, Marquis’ Who’s Who On the Web, Standard & Poor’s Corporate Register of Directors & Executives, Dun & Bradstreet’s Reference Book of Corporate Management, and Forbes, as well as from other company and executive profiles where available. Data on 102 executive directorships was gathered from Corporate Library and other biographical information and data on previous positions was gathered from ExecuComp. Data on geographical locations was gathered from the United States Census Bureau. Dependent Variable The purpose of this study is to examine the likelihood of dismissed CEOs regaining executive re-employment at a publicly traded organization. Thus the dependent variable of this study is whether the executive regained employment as an executive in a given year in a publicly traded firm. This variable is measured as Rehired as Public Executive, defined as whether or not the executive gained employment in an executive capacity at a publicly traded firm during the observation year. Each dismissed executive was tracked by examining biographical information following dismissal in order to identify the next job obtained by the CEO. Data on each executive was gathered for each year, up to five years; however, observations for a CEO were no longer gathered for any years after the year in which an executive regained employment in an executive capacity. Executives were considered re-employed in a given year if in a new position they were listed on a publicly traded firm’s proxy statement. The use of this metric is the same as previous research examining executive dismissal and re-hiring (e.g., Gilson, 1989). This measure is appropriate, as it determines whether or not the executive was hired into a meaningful, executive capacity in a new position. Dismissed executives may take lower level executive positions in order to regain employment, but the purpose of this study is to examine how strong the stigmatizing effects of dismissal are on executive careers and whether the effects of human, social, and reputational capital can limit the stigmatization that executives experience. The variable Rehired as Public Executive represents a dichotomous variable equal to 1 if an 103 executive appears on a publicly traded proxy statement as an executive in the observation year, and 0 otherwise. Independent Variables Circumstances of Dismissal. The reason why an executive is dismissed may influence perceptions of the executive in the external labor market following dismissal. Thus, three circumstances were identified that may impact perceptions of the executive. Performance dismissal is a dichotomous variable equal to 1 if news reports at the time of dismissal include discussion of poor performance or performance related problems. This variable serves as a measure of whether the executive was dismissed due to poor performance. Fiduciary violation represents a dichotomous variable set equal to 1 if news reports at the time of the executive’s dismissal indicate the executive was involved in an ethical or legal violation relating to the firm and its operations and 0 otherwise. Examples of these include fraudulent behavior, stock option backdating, or securities class action lawsuits. Personal conduct violation is a dichotomous variable set equal to 1 if news reports at the time of the executive’s dismissal indicate that the executive was dismissed for reasons relating to personal conduct that do not involve the firm’s operations. These may include personally illegal behavior or personal conduct problems (e.g. falsifying one’s resume), but do not include health reasons. These circumstances do not represent the entire wealth of reasons for why executives may be dismissed. For instance, executives may be dismissed due to conflicts of interest with the board or a disagreement over strategy. Thus, executives may be classified into no more than one of the three categories identified; however, not all executives must be categorized into one of the three categories. 104 Human Capital. Prestige experience is a dichotomous variable coded as 1 if the CEO has served as an executive listed on a proxy statement of any Fortune 500 company at any point in their career and 0 otherwise. Fortune 500 companies are highly visible and attract top talent. Executives at these organizations are regularly sought after to run other organizations. Thus, this variable serves as a proxy for human capital with regards to an actor’s credibility in the market. Education level is calculated using a 5-point scale based on the highest level of education completed by the CEO (1=high school, 2=some college, 3=undergraduate degree, 4=graduate degree, 5=doctorate). Elite education is a dichotomous variable coded as a 1 if the CEO attended an Ivy League institution or one of the US News and World Report top 25 institutions as designated in 2011 and 0 otherwise. These institutions chosen will represent those with national recognition for prestige, rather than those with middle-level or regional reputations. While these institutions chosen are different from those from prior research (e.g., Finkelstein, 1992), these institutions represent those that contain a higher level of prestige during the sample time period identified9. Reputational Capital. For this study, the CEO’s reputation is measured in two ways. The first measure is based on a certification contest; whether the CEO was included on Institutional Investor magazine’s Best CEOs list, which began identifying top CEOs in 2003. Institutional Investor asks buy and sell-side analysts to identify the best executives from the companies they follow. Survey data reflects the opinions of more than 1,300 analysts from more than 550 firms. Institutional Investor has a circulation of more than 130,000 and its list of Best CEOs has been reported on by outlets such as CNN and CNBC. Best CEO, therefore, is a dichotomous variable based on whether the CEO has been indicated as a top CEO by 9 Alternative conceptualizations were also analyzed, including using the list of institutions developed by Finkelstein (1992) with results being similar. Results are available from the author upon request. 105 Institutional Investor at any point up to the observation year. CEOs that have been named a best CEO were provided a value of 1, while those not appearing on the list at any point prior to the observation year were coded as 010. The second measure is the negative publicity associated with the CEO over the 3 years prior to the observation year in major US newspapers. An independent coder examined all newspaper reports across six major US newspapers for each dismissed CEO and coded each article mentioning a CEO’s name as either positive/neutral or negative. Publicity regarding the CEO was only coded as negative if the article specifically mentions the CEO and would cast the CEO in a negative light (e.g. adversely affect how others view the CEO, not how others view the organization). This measure represents a count of the negative articles regarding the CEO over the three years prior to the observation year. This measure is similar to Milbourn’s (2003) construct of CEO reputation with regard to press coverage, and is indicative of a CEO’s reputation as bad publicity may alter how stigmatized an executive is due to his or her portrayal in the media. CEOs with greater levels of negative publicity may be more “tainted” or have a greater “stain” on their character which may decrease their likelihood of re-employment. The six newspapers examined were The Wall Street Journal, The New York Times, The Washington Post, USA Today, the Los Angeles Times, the Chicago Tribune, and the Houston Chronicle. These newspapers cover different regions of the United States, but also represent 7 of the top 10 newspapers in circulation in the United States. Social Capital. Total directorships is a count variable based on the cumulative number of directorships the executive has occupied at any point in his or her career prior to the observation 10 A second measure of reputation was whether the CEO had appeared in the Institutional Investor top 5 in a given year prior to the observation year, as analysts rank the top 5 CEOs each year. Results with the alternative specification were nearly identical and are available from the author upon request. 106 year. These directorships represent the ties that the executive has developed across a wide variety of organizations. Prior executive employers is a count variable of all prior employers for which the executive has appeared on the company’s proxy statement as a top executive or for which the executive has served on the organization’s board of directors as an inside director. Finally, Residence in major city represents a dichotomous variable equal to 1 if the executive’s position as CEO was located in one of the 15 largest metropolitan statistical areas in the United States and 0 otherwise. The top 15 largest metropolitan areas were chosen as they represent all US cities with a population of greater than 2.5 million people11. Control Variables In order to control for the effects of likelihood of re-employment or whether a dismissed executive will re-enter the workforce, several other variables were added. SEC investigation represents a dichotomous variable equal to 1 if the CEO’s prior firm underwent an SEC investigation in the three years prior to the CEO’s dismissal or if an investigation was announced within 3 months of the CEO’s dismissal. Additionally, the investigation had to be related to the tenure of the dismissed CEO. Current age represents the CEO’s age at the time of the observation year. Older executives may not desire to re-enter the workforce or may enter it on a temporary basis as a consultant or in some other fashion. CEO’s prior compensation is the CEO’s total compensation in the year in which the executive was dismissed. Executives who make significant earnings may not seek out future employment as their need for earnings is lessened by their accumulated wealth. Alternatively, wealthy dismissed executives may be able to “purchase” future executive employment, either by buying publicly traded firms or establishing their own new enterprises with their wealth. Prior ownership represents the 11 Alternative specifications were utilized examining the top 10 and 20 largest metropolitan areas. A similar pattern of results emerged and are available from the author upon request. 107 proportion of ownership the CEO had in the firm from which he or she was dismissed in the dismissal year, in order to control for alternative wealth effects. Firm size represents the size of the firm firing the executive and is measured as the natural log of assets in the year of the executive’s dismissal. Executives from larger firms may find it easier to gain employment at smaller firms who wish to enhance their image or legitimacy through the employment of high profile executives or who wish to gain access to expertise. The CEO’s tenure may also impact his or her re-employability when searching for a new position. Long-tenured CEOs may have become obsolete for their firms or have too much firm-specific human capital, which is not portable to other organizations. Thus, CEO tenure represents the length of the dismissed CEO’s tenure in years at the dismissing firm. Finally, dismissals often occur due to poor performance, which may impact how organizations view the executive when determining whether to re- employ him or her. Thus, Prior firm’s performance measures the performance of the firm at which the executive served as CEO and is measured using a three year average of industry adjusted return on assets (ROA) at the 2-digit SIC level. Analysis The sample for this study is comprised of executives who underwent a potentially traumatizing career event: dismissal. This sample, however, does not represent a random sample of CEOs. Instead, given that each executive was dismissed from their prior firm, albeit for different reasons, it is important to control for the reasons for such dismissal. Thus, this study uses a Heckman selection model (Heckman, 1979; Karaevli, 2007, p.696; Zajac & Westphal, 1996, p. 72) in order to correct for any sample selection error that may exist regarding firm characteristics or characteristics surrounding dismissal. As a first-step, all firms with more than 1,000 employees in non-banking and finance industries were included in order to determine 108 likelihood of dismissal. Each observation was a firm-year pairing with the dependent variable representing dismissal, where observations with a dismissed CEO coded as 1, and 0 for all observations where a CEO was not dismissed. In order to test the likelihood of dismissal, variables relating to CEO power, board monitoring, and other CEO characteristics were included. The results can be found in Appendix A. The results of this model were used to calculate the Inverse Mills ratio, which was included in the second stage of the analysis to test the likelihood of re-employment. In order to test hypotheses, event history survival analysis was employed using a Cox proportional hazards model. Survival analysis allows for examining the likelihood of an event to occur over time after an initial baseline date is determined. The initial baseline date for this study was the calendar year in which the CEO was dismissed. In order to determine likelihood of re-employment, each record represents an observation calendar year following the dismissal year for each CEO, with each CEO having up to 5 years represented in the dataset. Thus, each observation represents an additional year since the dismissal occurred. Independent variables are allowed to vary over time. Such time varying variables for this study include the CEO’s age in the observation year, negative publicity regarding the CEO over the prior 3 years, and the executive’s cumulative number of directorships held. The Cox proportional hazards model allows for the calculation of an initial baseline hazard rate and the calculation of subsequent hazard rates for the likelihood of re-employment over each observation period (e.g. years since dismissal) given the data analyzed. Each executive drops out of the study at the time in which the event (e.g. re-employment) occurs; that is, once the executive is re-employed, no further observations relating to that executive are included. The final analyses for the likelihood of re- 109 employment for the 88 subjects resulted in 397 observations, as 10 of the 88 dismissed CEOs were rehired by a public company. RESULTS Table 5 shows the descriptive statistics and correlations among all variables included in the study. The rehiring of a dismissed CEO as an executive at a publicly traded company is significantly correlated with the size of the dismissing firm and prestigious experience obtained by the CEO, indicating that CEOs who have worked for large, prestigious firms are significantly more likely to gain re-employment following dismissal. ----------------------------------------------- Insert Table 5 about Here ----------------------------------------------- Tables 6 and 7 provide information relating to the career outcomes of the 88 CEOs identified in this study as dismissed. As seen in Table 6, 10 CEOs were able to overcome dismissal and regain a position as an executive at a publicly traded firm, with 8 of these assuming the role of CEO again during the sample period. All 10 dismissed CEOs who regained employment did so within the calendar year following their dismissal. Table 7 shows that on average, it took CEOs who regained the CEO role 1.5 years before they were re-appointed to the position. Additionally, 14 CEOs were hired as CEOs of private companies, positions which took on average 2.07 years to obtain following dismissal. Out of the 88 CEOs identified, 60 went on to assume a new position at some point during the five year sample window. 19 of these individuals accepted positions in other fields (e.g. private equity, venture capital) and 14 were classified as self-employed. Table 7 notes that on average executives were rehired 1.54 years 110 after their dismissal. This data suggests that executives are likely to overcome the stigmatization of dismissal shortly after it occurs to regain employment or not at all. ----------------------------------------------- Insert Tables 6 and 7 about Here ----------------------------------------------- Table 6 also identifies interesting information regarding reasons for dismissal and likelihood of re-employment. In total, 39 CEOs were classified as dismissed due to performance related reasons, 15 for violations of fiduciary duty, and 3 for personal conduct violations. No classification was identified for 31 of the 88 subjects. While five of the CEOs who regained a position as a public executive were dismissed with discussion of performance problems, none of the 18 CEOs dismissed for violations of fiduciary duty or personal conduct violations were re- employed as executives at publicly traded corporations following their dismissal. In fact, only 10 of these 18 individuals regained any employment in the five years following dismissal, compared to 50 of the 70 individuals dismissed for other reasons or performance related problems. This descriptive data suggests that it is more difficult to overcome violation of fiduciary duty or personal conduct violations than dismissal for other reasons. Table 8 presents the results of the Cox proportional hazards model with the dependent variable representing the rehiring of the CEO as a public company executive in the observation year. Model 1 presents only control variables, while Model 2 incorporates the inverse Mills ratio as calculated in the first stage of the selection model. Model 3 includes all main effects for hypothesis testing. Models 4 through 26 in Tables 8 and 9 present the results of interaction tests done in order to analyze how the circumstances surrounding a CEO’s dismissal are moderated by the executive’s capital in order to alter the likelihood of re-employment. 111 ----------------------------------------------- Insert Tables 8, 9a, 9b, and 9c about Here ----------------------------------------------- As seen in Model 1, an SEC investigation related to the CEO’s tenure at the dismissing firm significantly decreases the likelihood of obtaining an executive position at a publicly traded company. Dismissed CEOs from larger firms, however, are significantly more likely to obtain an executive position. Model 3 presents the results of the analysis for hypothesis testing. Hypothesis 34 predicts that executives dismissed for performance related reasons will be less likely to gain re-employment. Results, however, fail to provide support for this hypothesis. Hypotheses 35 and 36 predict that executives dismissed for violations of fiduciary duty and personal conduct violations will be less likely to regain employment in a publicly traded company. Both fiduciary violation (p ≤ 0.001) and personal conduct violation (p ≤ 0.001) are significant, providing strong support for both Hypotheses 35 and 36. These results suggest that it is significantly more difficult to regain employment following violations of fiduciary duty and personal conduct than for dismissal for other reasons. Hypothesis 37 argues that circumstances of dismissal relating to performance and personal conduct violations will have a lesser effect than violation of fiduciary duty on likelihood of re-employment. Results indicate that performance dismissal does not significantly alter the likelihood of re-employment, especially in comparison to violations of fiduciary duty; however, personal conduct violations and violations of fiduciary duty appear to affect the likelihood of re-employment in near equal fashions, providing only partial support for Hypothesis 37. These results are consistent with the statistics on re-employment in Table 6, which show a much greater likelihood of re-employment for executives dismissed for performance or other reasons than for those dismissed for violations of 112 fiduciary duty or personal conduct violations. Taken together, these results suggest that dismissal for violations of fiduciary duty or related to personal conduct serve as stigmatizing events in the career of executives, more so than performance related dismissal. Hypotheses 38 through 41 examine an executive’s human and reputational capital on the likelihood of dismissal. Among these, results provide significant support for both prestige experience (p ≤ 0.001) and Best CEO (p ≤ 0.01) with regard to increasing the likelihood of regaining employment as an executive in a publicly traded firm. Additionally, greater negative publicity regarding the CEO significantly reduces the likelihood of executive re-employment (p ≤ 0.05). These results provide support for Hypotheses 38 and 41, while failing to provide support for Hypotheses 39 and 40. Taken together, these results suggest that experience in prestigious organizations and earning a reputation from analysts enhances career outcomes following dismissal and reduces the effects of the stigmatization of dismissal. These results also provide support for the notion that being certified as a “Best CEO” enhances the attractiveness of a dismissed CEO in the labor market following dismissal. Alternatively, negative publicity negatively impacts a dismissed CEO’s reputation, such that greater publicity that portrays the CEO poorly and reduces the likelihood of re-employment. These last two results provide significant evidence that a CEO’s reputation affects his or her likelihood of re-employment. Hypotheses 42 through 44 examine the impact of CEO social capital on the likelihood of re-employment. Among these hypotheses, only Residence in major city significantly impacts the likelihood of re-employment (p ≤ 0.05), such that CEOs whose firms were located in larger cities have a greater chance of re-employment. This result may indicate that dismissed CEOs gain jobs when living in major cities due to a larger social network. Alternatively, however, major cities may offer a greater number of potential positions that are available for the executive to select 113 from and with whom are willing to hire the executive. In total, Hypothesis 44 is supported, while no support is obtained for Hypotheses 42 and 43. As noted above, interaction tests were performed in order to determine whether human, reputational, and social capital could buffer executives from dismissal based on the circumstances surrounding the dismissal. Following Aiken and West (1991), continuous variables were mean-centered prior to their inclusion in the interaction terms. Each interaction term was entered individually in order to test the impact each interaction has on the likelihood of dismissal. Models 4 through 11 test Hypotheses 45 through 51, which argue that an executive’s human, reputational, and social capital will reduce the effects of performance related dismissals on the likelihood of executive re-employment. Among these variables, prestige experience (p ≤ 0.001), elite education (p ≤ 0.001), Best CEO (p ≤ 0.05), negative publicity (p ≤ 0.05), prior executive employers (p ≤ 0.05), and Residence in major city (p ≤ 0.05) all impact the likelihood of re-employment in a publicly traded firm as an executive when interacted with performance related dismissal. These results provide support for Hypotheses 45, 47, 48, 49, 50, and 51. In order to better understand these relationships, Figures 7-12 provide descriptive information regarding executives appearing in these cases. ----------------------------------------------- Insert Figures 7-12 about here ----------------------------------------------- As seen in Figure 7, more executives with prestige experience were able to obtain re- employment than those without, regardless of whether the dismissal was performance related. Additionally, Figure 8 shows that more executives received re-employment when performance was the reason for dismissal if they did not have an elite education; however, elite education was 114 beneficial for more executives when dismissal was performance related. Figure 9 provides descriptive evidence that winning a Best CEO award helps more when performance was not the reason for dismissal, but did not help the one individual who was dismissed due to performance concerns. The effects of negative publicity and performance related dismissals have a different effect. Above average negative publicity appears to descriptively help more CEOs obtain re- employment regardless of whether the dismissal was performance related, as seen in Figure 10. In Figure 11, executives with more prior employers in an executive position did not receive a position when they had the most prior jobs (2) than when others had fewer jobs and dismissal was not performance related. However, having more positions did assist in regaining employment for some executives when dismissal was performance related. This may be an indication that previous employers help when dismissal occurs for performance related reasons. Finally, Figure 12 shows that a higher percentage of executives in non-major cities received re- employment when dismissal was not performance related than those in major cities, while executives fired for performance related issues were equally likely, in this sample, to gain re- employment whether they resided in a major metropolitan area or not. Hypotheses 52 through 58 predict that human, reputational, and social capital would buffer an executive from the harmful effects of dismissal due to violation of fiduciary duty. As seen in models 12 through 19, education level (p ≤ 0.001), Best CEO (p ≤ 0.001), negative publicity (p ≤ 0.001), prior executive employers (p ≤ 0.001), and residence in major city (p ≤ 0.001) all interact with violations of fiduciary duty significantly to impact likelihood of re- employment. These findings support Hypotheses 53, 55, 57, and 58. Given that no executives dismissed for violation of fiduciary duty regained employment, it is likely that fiduciary violations moderated the impact that human, reputational, and social capital have on the 115 likelihood of executive re-employment. In order to better understand these relationships, Figures 13-17 provide descriptive information regarding executives appearing in these cases. ----------------------------------------------- Insert Figures 13-17 about here ----------------------------------------------- Figure 13 provides information on the outcomes of dismissed executives who were dismissed for violation of fiduciary duty and their education level. As noted in the figure, no executives who were dismissed for fiduciary duty violations received new positions, regardless of education level, while only CEOs who had undergraduate and graduate degrees received new positions when fiduciary duties were not violated. With regards to CEO reputation and outcomes for dismissed CEOs, again no CEOs dismissed for violation of fiduciary duty were rehired; however, a higher percentage of CEOs winning a best CEO award were rehired when fiduciary duties were not violated. These results, however, must be examined with extreme caution, as only two cases were identified where CEOs who received a best CEO award were dismissed without violating fiduciary duties. Figure 15 illustrates that when fiduciary duties were not violated, a higher percentage of CEOs were rehired when there was more negative publicity about the CEO (e.g. negative publicity was above the average). These result statistics run counter to the main effects of negative publicity on the likelihood of firing. Statistics in Figure 16 show that a greater percentage of CEOs were rehired when they did not violate fiduciary duty and they had previously served in executive positions for other firms. These results provide limited support for the notion that a greater number of executive positions enhance an executive’s chance of rehiring, but only when fiduciary duties are not violated. Finally, Figure 17 statistics show that residence in a major city impacts a CEO’s chance of being 116 rehired only when fiduciary duties are not violated. In sum, Figures 13-17 provide some evidence on what may buffer executives from the stigmatization of dismissal; however, these statistics also clearly illustrate that the sample CEOs who were dismissed for violation of fiduciary duty faced severe consequences in the labor market, regardless of their level of human, reputational, and social capital; that is, violation of fiduciary duty limits the impact that executive capital have on regaining employment. Hypotheses 59 through 66 predict that human, reputational, and social capital will moderate the relationship between personal conduct violations and likelihood of re-employment, such that executives with greater levels of accumulated capital will be more likely to regain employment. Due to only three executives classified as having been dismissed for personal conduct violations, and a lack of variation in these executives’ traits, models could only be analyzed to test Hypotheses 59, 60, 63, 64, and 65. None of the tested interactions were significant at conventional levels. These results fail to provide support for Hypotheses 59 through 65. Overall, the results suggest that the circumstances of dismissal strongly impact the likelihood of executive re-employment, such that committing violations relating to fiduciary duty and personal conduct significantly reduce the likelihood of re-employment. At the same time, gaining human, reputational, and social capital can also reduce the harmful effects of dismissal and assist dismissed CEOs in regaining re-employment as an executive of a publicly traded company. The results of the interaction effects suggest that dismissal due to violations of fiduciary duty reduce the effects of executive capital on the ability to regain executive employment. Additionally, executive’s human and reputational capital, in particular, are likely 117 to assist executives who are fired for reasons other than fiduciary duty or personal conduct violations more so than executives fired for such reasons. Sensitivity Analyses In order to test the robustness of the results reported, additional analyses were conducted in order to examine the likelihood of re-employment using alternative specifications of the dependent variable. The first alternative dependent variable examined is Rehired as Public CEO, defined as whether or not the dismissed CEO appears as a publicly traded firm’s CEO in a given observation year. This variable represents a dichotomous variable equal to 1 if the dismissed executive appears as a publicly traded firm’s CEO in the observation year and 0 otherwise. Tables 10 and 11 display the results of the analyses conducted using this dependent variable. As seen in Model 3, executives dismissed for reasons of fiduciary violation (p ≤ 0.001) or personal conduct violation (p ≤ 0.001) are significantly less likely to regain a position as a CEO of a publicly traded firm, consistent with results reported earlier. Additionally, prestige experience significantly enhances the likelihood of an executive regaining employment as a public company CEO (p ≤ 0.05). Finally, executives with greater social capital in the form of more prior executive positions (p ≤ 0.05) and residence in a major city (p ≤ 0.01) are more likely to gain re-employment as a CEO of a publicly traded firm. These results are similar to the results reported in Table 8, with several differences. First, negative publicity is only marginally significant (p ≤ 0.10) with regards to decreasing the likelihood of regaining employment as a CEO. Additionally, the Best CEO award does not significantly impact the likelihood of employment as a public company CEO. Finally, prior executive employers significantly impacts the likelihood of regaining employment as a CEO, but not as an executive in a publicly traded firm. These results suggest that many of the same factors affect re-employment as a CEO that 118 affect a dismissed executive’s re-employment as an executive; however, some factors alter the likelihood as well. ----------------------------------------------- Insert Tables 10, 11a, 11b, and 11c about here ----------------------------------------------- Table 11 presents the results of the interaction tests using the alternative dependent variable. With regards to interactions between executive capital and performance related dismissal, results find significant interactions with prestige experience (p ≤ 0.001), elite education (p ≤ 0.001), negative publicity (p ≤ 0. 01), total directorships (p ≤ 0.01), prior executive employers (p ≤ 0.001), and residence in major city (p ≤ 0.001). Figures 18 through 23 provide statistics on career outcomes for CEOs who were and were not dismissed for performance reasons, as well as their prestige experience, education, negative publicity, directorships, prior employers, and residence in major city. In particular, it is interesting to note that regardless of performance related reasons for dismissal, more executives with prestige experience gained re-employment; however, more executives were also rehired as a CEO when their dismissal was related to performance. Additionally, more executives, and a higher percentage, were re-employed as a public CEO when dismissed for reasons not relating to performance and when they had received an elite education. Finally, executives in the sample who reside in a major city and were dismissed for performance related reasons were more likely to regain employment as a CEO than executives not in a major city or not fired for performance related reasons. ----------------------------------------------- Insert Figures 18-23 about here 119 ----------------------------------------------- With regard to violations of fiduciary duty, significant interactions were identified with education level (p ≤ 0.01), negative publicity (p ≤ 0.001), prior executive employers (p ≤ 0.001), and residence in major city (p ≤ 0.001). Figures 24 through 27 provide statistics on the outcomes for executives in these conditions. As noted earlier, no executives who were dismissed for fiduciary violations were rehired. Executives in the sample who did not violate fiduciary duties were more likely to gain re-employment as a CEO when they had greater levels of education, above average negative publicity, and resided in a major city. Finally, no interactions between personal conduct violations and executive capital reached conventional significance. ----------------------------------------------- Insert Figures 24-27 about here ----------------------------------------------- The final alternate dependent variable examined in this study is Rehired as Executive, defined as whether the dismissed executive regained employment as an executive at any firm, public or private, in the given observation year. This variable includes all executives who regained employment at publicly traded firms and appearing on proxy statements. However, all executives who were identified as gaining positions at the Vice President level or above at any private company were also included as having gained executive re-employment. This level was chosen as private companies do not have to report their top executives in the same fashion as public companies. Tables 12 and 13 present the results of the analyses. As seen in Model 3 in Table 12, the likelihood of re-employment as any executive, in a public or private firm, is significantly affected by only whether the dismissed executive had been named Best CEO (p ≤ 120 0.05). These results suggest that circumstances of dismissal and executive capital most strongly affect the re-employment of executives in publicly traded firms, rather than in private firms. ----------------------------------------------- Insert Tables 12 and 13 about here ----------------------------------------------- Additionally, Models 4 through 26 present the results of tests of interactions between circumstances of dismissal and executive capital on the likelihood of re-employment as any executive. Results regarding performance related dismissals suggest that education level (p ≤ 0.05) and elite education (p ≤ 0.05) both significantly impact the likelihood of regaining employment as an executive. Figures 28 and 29 show the career outcomes of sample CEOs by differing values of the capital variables based on the reasons for dismissal. Within the sample, Figure 28 identifies that a greater percentage of executives with graduate degrees received re- employment when dismissal was not performance related, while a greater percentage of executives regained employment when dismissal was performance related when only having received an undergraduate degree. Additionally, Figure 29 illustrates that within sample executives’ elite education helped more executives regain employment as any executive when dismissal was not performance related; however, more executives regained employment without elite education when dismissal was performance related. ----------------------------------------------- Insert Figures 28 and 29 about here ----------------------------------------------- Results regarding interactions between executive capital and fiduciary violations present a different picture regarding likelihood of re-employment as any executive. Prestige experience, 121 Best CEO, and negative publicity all significantly alter the likelihood of re-employment. Figures 30, 31, and 32 utilize sample outcomes for dismissed executives across the ranges of each of these interaction variables. Figure 30 provides evidence that for sample executives, re- employment was more likely when executives were not dismissed for violations of fiduciary duty with the most likely scenario being re-employment when dismissal was not related to violation of fiduciary duty and the executive had prestigious experience. Alternatively, for sample CEOs, more CEOs dismissed for violation of fiduciary duty were rehired as any executive when they did not have prestigious experience than when they did. ----------------------------------------------- Insert Figures 30-32 about here ----------------------------------------------- Finally, interaction tests were conducted between dismissals for personal conduct violations and executive capital, finding significant relationships with prestige experience (p ≤ 0.001), education level (p ≤ 0.001), and prior executive employers (p ≤ 0.001). Figures 33 through 35 provide outcomes for sample CEOs in each of these instances. For instance, Figure 35 provides evidence that for sample CEOs, executives were most likely to be rehired when personal conduct violations were not committed; however, for CEOs not dismissed for personal conduct violations, a greater percentage of CEOs were rehired as the number of prior executive employers increased. Alternatively, only one CEO who committed a personal conduct violation was rehired, and that executive had no prior executive employers. ----------------------------------------------- Insert Figures 33-35 about here ----------------------------------------------- 122 The results of the sensitivity analyses provide support for the initial findings regarding how circumstances of dismissal and executive capital impact the likelihood of executive re- employment following dismissal. Additionally, these analyses provide strong evidence that the circumstances of dismissal and an executive’s capital most strongly drive the impact of dismissal on re-employment when examining career outcomes of CEOs with regards to publicly traded firms. In particular, the reasons for dismissal only affect the likelihood of executive re- employment at publicly traded firms, but have a significantly less affect when examining re- employment at public or private firms, suggesting that norms of legitimacy prevent publicly traded firms from hiring executives dismissed for violations relating to fiduciary duty or personal conduct. Taken together, these results provide strong evidence that circumstances of dismissal and executive capital alter the likelihood of executive re-employment, but that such factors matter most in the labor market for executive talent at publicly traded firms. DISCUSSION CEO dismissal is an increasingly occurring phenomenon with significant consequences for the firm choosing to dismiss a CEO and the CEO him or herself. While great attention has been focused on how firms go through succession processes and choose successors following dismissal, only a small and disparate stream of research has examined what happens to CEOs after such a fall from grace. Prior studies have noted that executives are unlikely to gain re- employment following bankruptcy (Gilson, 1989) or dismissal (Ward et al., 1995), indicating that negative events in a CEO’s career can stigmatize the CEO. Such stigmatization can prevent organizations from wanting to associate with the executive in the future due to norms of legitimacy and pressure from economic arbiters. However, among the executives studied by these authors, some executives do manage to recover and regain employment in an executive 123 capacity. To date, however, research has not empirically examined how such executives are able to overcome stigmatization to regain employment with publicly traded firms. This study makes several interesting and important contributions to the literature on stigmatization of executives following harmful career events by bringing together the literature on executive stigmatization and research on executive’s career responses to dismissal. First, this study builds upon prior theoretical work (Wiesenfeld et al., 2008) regarding executive stigmatization following career failures to illustrate how dismissal can stigmatize an executive’s career. Dismissal reduces the likelihood that firms will choose to associate with executives in the future due to pressure from economic arbiters and norms of legitimacy for publicly traded firms. Further, this study builds upon this theory to predict that the circumstances surrounding an executive’s dismissal will contribute to the degree of stigmatization associated with the CEO. While prior studies find how events surrounding dismissal impact likelihood of re-employment (e.g., Fee & Hadlock, 2004; Ward et al., 1995), such studies have not developed rigorous theoretical models as to why dismissal should result in such outcomes. In particular, this study argues that CEOs dismissed for reasons relating to the violation of fiduciary duty will undergo the greatest degree of stigmatization, as violations of fiduciary duty are representative of an executive’s willingness to deceive organizational stakeholders or act opportunistically at the expense of shareholders. Furthermore, the theoretical foundations of this study provide evidence that both performance related dismissals and personal conduct violations should provide some degree of stigmatization regarding an executive following dismissal, as both reasons for dismissal provide negative information regarding the executive’s capabilities and / or character that may be difficult to overcome. 124 Second, this study extends research on overcoming career failures by examining how executive human, reputational, and social capital can assist dismissed CEOs in regaining employment as executives in publicly traded firms. In particular, building on Wiesenfeld and colleagues (2008), this study argues that greater levels of not only social capital, but also human and reputational capital should signal CEO capabilities following dismissal, such that greater levels of accumulated capital should assist dismissed CEOs in regaining employment. This study extends the research on the career outcomes of dismissed executives by going beyond examining whether executives were rehired based on why dismissal occurred to examine the characteristics of executives who regained employment. This theoretical step is important in order to understand how executives can overcome stigmatization following dismissal in order to re-enter the workforce as an executive. While the circumstances of dismissal may provide negative cues about capabilities and character, an executive’s accumulated capital may provide positive signals about the executive’s capability (e.g. human and reputational capital) and character (e.g. social capital). Third, this study tracks 88 executives dismissed in 2005 or 2006 and examines their likelihood of re-employment as an executive at a publicly traded organization each year for five years following dismissal. Consistent with prior studies, only ten of the 88 executives regain employment as executives at publicly traded firms; however, eight of the ten dismissed executives regain employment as a CEO at such firms. Results of the analyses conducted provide strong support for the notion that circumstances of dismissal stigmatize executives, as CEOs dismissed for violations of fiduciary duty or personal conduct violations are significantly less likely to be re-employed as executives in publicly traded firms than executives dismissed for other reasons or performance related concerns. In particular, none of the 18 executives in this 125 study’s dataset who were dismissed for personal conduct violations or violations of fiduciary duty regained employment as an executive in a publicly traded firm within 5 years of dismissal. These results are consistent with the concept of ‘settling up’, in that the market disciplines executives in the future who act opportunistically in their prior positions. While circumstances of dismissal negatively impact the likelihood of re-employment, an executive’s human, reputational and social capital each positively impact the likelihood of re- employment. Wiesenfeld and colleagues (2008) argue that norms of legitimacy and pressure from outside forces prevent publicly traded organizations from hiring stigmatized executives. However, results of this study provide evidence that executive capital can help organizations overcome such pressure in order to rehire dismissed executives under certain circumstances. For instance, prestigious experience, in the form of having worked as an executive at a Fortune 500 company, increases the likelihood of re-employment following dismissal. Prestigious experience serves as a signal of the executive’s capabilities, as prestigious organizations have strong training for executives and experience as an executive serves as the firm’s endorsement of the individual’s capabilities. Additionally, an executive having won an award for “Best CEO” as voted on by analysts significantly increases the likelihood of re-employment. Alternatively, executives with greater levels of negative publicity are less likely to regain employment as an executive. Both measures indicate that as a positive reputation (or less negative reputation) is developed, executives can utilize reputational capital as a signal of capabilities to others in the market. Finally, executive re-employment is more likely when the executive’s prior position as a CEO was located in one of the 15 largest metropolitan areas in the United States. An executive’s location is important as many social networks develop geographically through board interlocks and community involvement. Residence in a major city may expand a dismissed CEO’s network 126 through this community involvement, which develops greater friendships with other members of the corporate elite, who may assist the executive in finding re-employment following dismissal. Additionally, results based on the additional analyses performed indicate that circumstances of dismissal and executive capital impact the likelihood of executive re- employment more for executives who wish to regain positions in publicly traded firms. These results are significant as they indicate that the pressures facing firms with regards to hiring stigmatized executives are significantly greater on publicly traded organizations. Such constraints are likely to be less powerful on private firms who face less pressure from stakeholders and third parties with regards to personnel decisions. Finally, this study builds on the growing literature regarding executive reputation and certification contests to show how an executive’s reputation as a CEO affects his or her career outcomes following dismissal. First, results indicate that the certification of a CEO through winning the Institutional Investor magazine’s annual “Best CEO” contest can buffer a dismissed executive from the effects of stigmatization following dismissal. Institutional Investor’s contest regarding top executives surveys analysts, who closely follow each firm in an industry and may be considered experts on the executives in the firms they follow, to determine who analysts believe are the top CEOs in each industry. Thus, this “Best CEO” award represents a certification bestowed on a CEO as the top executive in an industry as awarded by a panel of industry “experts”. The results of this study indicate that when CEOs obtain such a certification they are more likely to be re-employed following dismissal. That is, externally awarded certifications signal to the market a CEO’s capabilities as determined by other expert parties and this signal reduces the stigmatizing effects of the signal provided by the executive’s dismissal. Additionally, this study finds a significant relationship between the negative publicity an 127 executive receives and his or her subsequent career outcomes, such that greater negative publicity reduces the likelihood of re-employment. These results suggest that a negative reputation can also signal the market about a CEO, such that firms choose not to associate with executives who have negative reputations. Both of these findings regarding CEO reputation advance the literature on CEO reputation, as they indicate that how external parties collectively evaluate and discuss an executive’s capabilities affect their likelihood of re-employment. Overall, this study is the first to build on the existing literature regarding career outcomes of dismissed executives and the effects of stigmatization to empirically test how executives can overcome negative career outcomes. This study provides an important contribution by showing how circumstances of dismissal negatively impact the likelihood of re-employment, while executive capital can assist in counteracting this process to increase the likelihood of re- employment. Implications for Managers and Future Research These results are not without practical implications for current executives. First, the results regarding violations of fiduciary duty and personal conduct violations have a resounding impact on executives. While CEOs likely know that opportunistically acting may cost them their executive positions, executives may be unaware of the long-term career implications of such behaviors. Executives should especially consider the long-term career consequences when engaging in opportunistic behaviors, as short-term gains may be cancelled out due to long-term career losses. Additionally, the results of this study indicate that personal conduct violations, such as falsifying resume information, also impact future career consequences of CEOs. Like with violations of fiduciary duty, executives should be acutely aware of the long-term consequences associated with committing violations relating to personal conduct. 128 Additionally, this study can assist CEOs who are dismissed in determining how best to find future re-employment. For instance, dismissed CEOs may be wise to use their reputational and human capital, in addition to leveraging social capital, in order to signal their capabilities to potential employers. The findings are important as they provide evidence of capabilities and capital that dismissed executives can leverage in order to gain re-employment. Additionally, the results when examining the likelihood of re-employment in a private or public position indicate that human, reputational, and social capital do not necessarily affect employment in the same manner that such capital affects employment at only publicly traded firms. These findings can provide evidence to managers that while dismissal may limit their ability to regain employment as an executive at a publicly traded firm; other options exist, including executive positions at private firms. Future research can build upon this study to continue to examine the consequences of executive dismissal on the career outcomes of executives. First, future research can examine additional characteristics of dismissed executives in order to determine their likelihood of re- employment. Such characteristics may include industry-specific human and social capital that is developed or individual characteristics of executives that may drive them to find new employment in order to personally overcome the distress associated with dismissal. Second, future research could examine other measures of social capital that may enhance the likelihood of re-employment. For instance, social network studies have looked at indications of capital including network centrality, betweenness, and embeddedness, which may impact an executive’s ability to regain executive employment. Using alternative measures of social capital within an executive’s firm and between other firms may generate additional insights with regards to how social capital can buffer executives from stigma. 129 A third avenue for future research would be to examine how executives perform at positions following re-employment. Executives who perform well following dismissal may be an indication that executives have learned from past mistakes and this learning allows for better success in the future. Alternatively, if re-employed executives fail to perform well in new positions, results may indicate that firms that fail to rehire dismissed executives are making wise decisions. Such results may indicate that executives have not learned from past mistakes or adapted to new environments and the capabilities and capital accumulated by executives may not help organizations succeed in the future. If such results are obtained, it may be a strong indication of why few dismissed executives gain re-employment in an executive capacity. Limitations Despite the findings of this paper, it is not without its limitations. First, the sample size is still relatively small with only 88 subjects and 397 firm-year observations. The relatively low sample size may fail to identify significant results and results in interaction effects being difficult to interpret. Additionally, only 10 subjects regain employment in publicly traded organizations. Future research should continue to track the career outcomes of dismissed executives to identify how executives are able to recover from the stigma of dismissal. A greater sample size may allow for better generalizability across firms and time and identify more significant results. Second, this study is limited to only United States based firms. The career outcomes for dismissed executives in other countries may be even more limited, especially if circumstances surrounding dismissal are due to stigmatizing events such as violation of fiduciary duty or personal conduct violations. Future research should examine how the process of ‘settling up’ varies among nations. Differing national characteristics may alter how such violations impact perceptions of executive character and capabilities. 130 Third, each executive who regained employment in a publicly traded organization did so within the calendar year following their dismissal. Results using survival analysis thus may be skewed, as the baseline hazard rate does not change across time due to the lack of executives regaining employment over time. As the sample size for future studies is increased, the baseline hazard rate may be affected over time. Conclusion In sum, this study provides evidence that dismissal serves as a stigmatizing event in the career of executives. Ten of 88 subjects identified as dismissed CEOs regained executive positions, and on average, positions were obtained within 1 calendar year of dismissal. Furthermore, executives dismissed for violations of fiduciary duty or personal conduct suffered the greatest degree of stigmatization, as they were most unlikely to regain future employment. Despite the negative impact of circumstances of dismissal, this study also illustrates how executive human, reputational, and social capital can help executives overcome the harmful effects of stigmatization to regain employment as an executive at a publicly traded organization. These results can better help executives understand how to overcome the negative effects of dismissal. 131 REFERENCES Adler, P. S. & Kwon, S. W. 2002. Social capital: Prospects for a new concept. Academy of Management Review, 27(1): 17-40. Agarwal, N. C. 1981. Determinants of executive compensation. Industrial Relations: A Journal of Economy and Society, 20(1): 36-45. Aiken, L. S. & West, S. G. 1991. Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage. Allen, M. P. & Panian, S. K. 1982. Power, performance, and succession in the large corporation. Administrative Science Quarterly, 27(4): 538-547. Amihud, Y. & Lev, B. 1981. Risk reduction as a managerial motive for conglomerate mergers. Bell Journal of Economics, 12(2): 605-617. Bailey, E. E. & Helfat, C. E. 2003. External management succession, human capital, and firm performance: An integrative analysis. Managerial and Decision Economics, 24(4): 347- 369. Ballinger, G. A. & Marcel, J. J. 2010. The use of an interim CEO during succession episodes and firm performance. Strategic Management Journal, 31(3): 262-283. Baltzell, E. D. 1989. Philadelphia gentlemen: The making of a national upper class. New Brunswick, NJ: Transaction Publishers. 132 Barker, V. L., Patterson Jr, P. W., & Mueller, G. C. 2001. Organizational causes and strategic consequences of the extent of top management team replacement during turnaround attempts. Journal of Management Studies, 38(2): 235-270. Barry, C. B., Muscarella, C. J., Peavy III, J. W., & Vetsuypens, M. R. 1990. The role of venture capital in the creation of public companies: Evidence from the going-public process. Journal of Financial Economics, 27(2): 447-471. Becker, G. S. 1962. Investment in human capital: A theoretical analysis. Journal of Political Economy, 70(5): 9-49. Becker, G. S. & Murphy, K. M. 1992. The division of labor, coordination costs, and knowledge. Quarterly Journal of Economics, 107(4): 1137-1160. Bell, D. E., Raiffa, H., & Tversky, A. 1988. Decision making: Descriptive, normative, and prescriptive interactions. Cambridge, UK: Cambridge University Press. Belliveau, M. A., O'Reilly, C. A., & Wade, J. B. 1996. Social capital at the top: Effects of social similarity and status on CEO compensation. Academy of Management Journal, 39(6): 1568-1593. Boeker, W. 1992. Power and managerial dismissal: Scapegoating at the top. Administrative Science Quarterly, 37(3): 400-421. Brady, D. 2006. The temp in the corner office. Business Week, July 3: 13. Brown, J. S. & Duguid, P. 1998. Organizing knowledge. California Management Review, 40(3): 90-111. 133 Buchholtz, A. K., Ribbens, B. A., & Houle, I. T. 2003. The role of human capital in postacquisition CEO departure. Academy of Management Journal, 46(4): 506-514. Burt, R. S. 1992. Structural holes. Boston, MA: Harvard University Press. Cairncross, F. 2003. Coming and going, The Economist. Cannella, A. A. & Lubatkin, M. 1993. Succession as a sociopolitical process: Internal impediments to outsider selection. Academy of Management Journal, 36(4): 763-793. Cannella, A. A., Fraser, D. R., & Lee, D. S. 1995. Firm failure and managerial labor markets: Evidence from Texas banking. Journal of Financial Economics, 38(2): 185-210. Carson, W. G. 1980. The institutionalization of ambiguity: Early British factory acts. In G. Geis & E. Stotland (Eds.), White collar crime: Theory and research. London: Sage. Carter, C. B. & Lorsch, J. W. 2004. Back to the drawing board: Designing corporate boards for a complex world. Boston, MA: Harvard Business School Press. Carter, S. M. & Ruefli, T. W. 2006. Intra-industry reputation dynamics under a resource-based framework: Assessing the durability factor. Corporate Reputation Review, 9(1): 3-25. Carter, W. C. & Feld, S. L. 2004. Principles relating social regard to size and density of personal networks, with applications to stigma. Social Networks, 26(4): 323-329. Castanias, R. P. & Helfat, C. E. 1991. Managerial resources and rents. Journal of Management, 17(1): 155-171. 134 Castanias, R. P. & Helfat, C. E. 2001. The managerial rents model: Theory and empirical analysis. Journal of Management, 27(6): 661-678. Citrin, J. M. & Smith, R. A. 2003. The five patterns of extraordinary careers: The guide for achieving success and satisfaction. New York, NY: Crown Business Books. Clement, W. 1977. Continental corporate power: Economic elite linkages between Canada and the United States. Toronto, ON: McClelland and Stewart. Coleman, J. 1990. Foundations of social theory. Cambridge, MA: Harvard University Press. Crane, D. 1965. Scientists at major and minor universities: A study of productivity and recognition. American Sociological Review, 30(5): 699-714. Cyert, R. M. & March, J. G. 1963. A behavioral theory of the firm. Englewood Cliffs, NJ: Prentice-Hall. D'Aveni, R. A. 1990. Top managerial prestige and organizational bankruptcy. Organization Science, 1(2): 121-142. Daily, C. M. & Johnson, J. L. 1997. Sources of CEO power and firm financial performance: A longitudinal assessment. Journal of Management, 23(2): 97-117. Dalton, D. R. & Kesner, I. F. 1983. Inside/outside succession and organizational size: The pragmatics of executive replacement. Academy of Management Journal, 26(4): 736- 742. 135 Datta, D. K. & Rajagopalan, N. 1998. Industry structure and CEO characteristics: An empirical study of succession events. Strategic Management Journal, 19(9): 833-852. Davidson, W. N., Nemec, C., Worrell, D. L., & Lin, J. 2002. Industrial origin of CEOs in outside succession: Board preference and stockholder reaction. Journal of Management and Governance, 6(4): 295-321. Davis, G. F. & Greve, H. R. 1997. Corporate elite networks and governance changes in the 1980s. American Journal of Sociology, 103(1): 1-37. Denis, D. J. & Denis, D. K. 1995. Performance changes following top management dismissals. The Journal of Finance, 50(4): 1029-1057. Denis, D. J., Denis, D. K., & Sarin, A. 1997. Ownership structure and top executive turnover. Journal of Financial Economics, 45(2): 193-221. Desai, H., Hogan, C. E., & Wilkins, M. S. 2006. The reputational penalty for aggressive accounting: Earnings restatements and management turnover. Accounting Review, 81(1): 83-112. Domhoff, G. W. 1967. Who rules america? Englewood Cliffs, NJ: Prentice-Hall. Elsbach, K. D. & Kramer, R. M. 1996. Members' responses to organizational identity threats: Encountering and countering the Business Week rankings. Administrative Science Quarterly, 41(3): 442-476. Fama, E. F. 1980a. Agency problems and the theory of the firm. Journal of Political Economy, 88(2): 288-307. 136 Fama, E. F. 1980b. Agency Problems and the Theory of the Firm. The Journal of Political Economy, 88(2): 288-307. Fee, C. E. & Hadlock, C. J. 2004. Management turnover across the corporate hierarchy. Journal of Accounting and Economics, 37(1): 3-38. Finkelstein, S. 1992. Power in top management teams: Dimensions, measurement, and validation. Academy of Management Journal, 35(3): 505-538. Finkelstein, S., Hambrick, D. C., & Cannella, A. A. 2009. Strategic leadership: Theory and research on executives, top management teams, and boards. New York, NY: Oxford University Press. Fisher, J. & Govindarajan, V. 1992. Profit center manager compensation: An examination of market, political and human capital factors. Strategic Management Journal, 13(3): 205- 217. Fombrun, C. & Shanley, M. 1990. What's in a name? Reputation building and corporate strategy. Academy of Management Journal, 33(2): 233-258. Fredrickson, J. W., Hambrick, D. C., & Baumrin, S. 1988. A Model of CEO Dismissal. Academy of Management Review, 13(2): 255-270. Friedland, R. & Palmer, D. 1994. Space, corporation, and class: Toward a grounded theory. In R. Friedland & D. Boden (Eds.), Now/here: Space, time, and modernity: 287-334. Berkeley: University of California Press. 137 Friedman, S. & Singh, H. 1989. CEO succession and stockholder reaction: The influence of organizational context and event content. Academy of Management Journal, 32(4): 718- 744. Friedman, S. D. & Olk, P. 1995. Four ways to choose a CEO: Crown heir, horse race, coup d'Etat, and comprehensive search. Human Resource Management, 34(1): 141-164. Fulmer, I. S. 2009. The elephant in the room: Labor market influences on CEO compensation. Personnel Psychology, 62(4): 659-695. Gamson, W. A. & Scotch, N. A. 1964. Scapegoating in baseball. American Journal of Sociology, 70(1): 69-72. Gilson, S. C. 1989. Management turnover and financial distress. Journal of Financial Economics, 25(2): 241-262. Gilson, S. C. 1990. Bankruptcy, boards, banks, and blockholders: Evidence on changes in corporate ownership and control when firms default. Journal of Financial Economics, 27(2): 355-387. Giordano, P. C. 1983. Sanctioning the high-status deviant: An attributional analysis. Social psychology quarterly, 46(4): 329-342. Goffman, E. 1963. Stigma: Notes on the management of spoiled identity. New York: Prentice- Hall. Goyal, V. K. & Park, C. W. 2002. Board leadership structure and CEO turnover. Journal of Corporate Finance, 8(1): 49-66. 138 Graffin, S. D. & Ward, A. J. 2010. Certifications and reputation: Determining the standard of desirability amidst uncertainty. Organization Science, 21(2): 331-346. Granovetter, M. S. 1995. Getting a job: A study of contacts and careers. Chicago, IL: University of Chicago Press. Gregory-Smith, I., Thompson, S., & Wright, P. W. 2009. Fired or retired? A competing risks analysis of chief executive turnover. The Economic Journal, 119(536): 463-481. Grusky, O. 1963. Managerial succession and organizational effectiveness. American Journal of Sociology, 69(1): 21-31. Hager, B., Driscoll, L., Weber, J., & McWilliams, G. 1991. CEO wanted: No insiders, please. Business Week, 12: 44-45. Haleblian, J. & Rajagopalan, N. 2006. A cognitive model of CEO dismissal: Understanding the influence of board perceptions, attributions and efficacy beliefs. Journal of Management Studies, 43(5): 1009-1026. Hambrick, D. C. & Fukutomi, G. D. S. 1991. The seasons of a CEO's tenure. Academy of Management Review, 16(4): 719-742. Hamori, M. 2006. Executive career advancement in career moves across employers: The role of organization-level predictors. The International Journal of Human Resource Management, 17(6): 1129-1151. Harris, D. & Helfat, C. 1997. Specificity of CEO human capital and compensation. Strategic Management Journal, 18(11): 895-920. 139 Harrison, J. R., Torres, D. L., & Kukalis, S. 1988. The changing of the guard: Turnover and structural change in the top-management positions. Administrative Science Quarterly, 33(2): 211-232. Hatfield, G., Worrell, D. L., Davidson III, W. N., & Bland, E. 1999. Turbulence at the top: Antecedents of key executive dismissal. Quarterly Journal of Business and Economics, 38(1): 3-5. Haunschild, P. R. 1993. Interorganizational imitation: The impact of interlocks on corporate acquisition activity. Administrative Science Quarterly, 38(4): 564-592. Heckman, J. J. 1979. Sample selection bias as a specification error. Econometrica: Journal of the Econometric Society, 47(1): 153-161. Henderson, A. D., Miller, D., & Hambrick, D. C. 2006. How quickly do CEOs become obsolete? Industry dynamism, CEO tenure, and company performance. Strategic Management Journal, 27(5): 447-460. Herman, E. S. 1981. Corporate control, corporate power: Cambridge University Press New York. Herrmann, P. & Datta, D. K. 2005. Relationships between top management team characteristics and international diversification: An empirical investigation. British Journal of Management, 16(1): 69-78. Higgins, M. C. & Gulati, R. 2003. Getting off to a good start: The effects of upper echelon affiliations on underwriter prestige. Organization Science, 14(3): 244-263. 140 Higgins, M. C. 2005. Career imprints: Creating leaders across an industry. San Francisco: Jossey-Bass. Higgins, M. C. & Gulati, R. 2006. Stacking the deck: The effects of top management backgrounds on investor decisions. Strategic Management Journal, 27(1): 1-25. Hill, C. W. L. & Phan, P. 1991. CEO tenure as a determinant of CEO pay. Academy of Management Journal, 34(3): 707-717. Holmstrom, B. 1982. Moral hazard in teams. Bell Journal of Economics, 13(2): 324-340. Houston, J. & James, C. 1993. Management and organizational changes in banking: A comparison of regulatory intervention with private creditor actions in nonbank firms. Carnegie-Rochester Conference Series on Public Policy, 38: 143-178. Huson, M. R., Parrino, R., & Starks, L. T. 2001. Internal monitoring mechanisms and CEO turnover: A long-term perspective. Journal of Finance, 56(6): 2265-2297. Hymowitz, C. 2006. A growing number of interim CEOs add to companies' turmoil. Wall Street Journal, December 18: B1. James, D. R. & Soref, M. 1981. Profit constraints on managerial autonomy: Managerial theory and the unmaking of the corporation president. American Sociological Review, 46(1): 1- 18. Jenter, D. & Kanaan, F. forthcoming. CEO turnover and relative performance evaluation. Journal of Finance. 141 Johnson, J. L., Ellstrand, A. E., Dalton, D. R., & Dalton, C. M. 2005. The influence of the financial press on stockholder wealth: The case of corporate governance. Strategic Management Journal, 26(5): 461-471. Kahneman, D. 2003. A perspective on judgment and choice: Mapping bounded rationality. American Psychologist, 58(9): 697-720. Kaplan, S. N. & Minton, B. 2010. How has CEO turnover changed? Ohio State University Working Paper. Karaevli, A. 2007. Performance consequences of new CEO ‘Outsiderness’: Moderating effects of pre and post succession contexts. Strategic Management Journal, 28(7): 681-706. Khurana, R. 2001. Finding the right CEO: Why boards often make poor choices. MIT Sloan Management Review, 43(1): 91-95. Khurana, R. 2002. Searching for a corporate savior: The irrational quest for charismatic CEOs. Princeton, NJ: Princeton University Press. Kiesler, S. & Sproull, L. 1982. Managerial response to changing environments: Perspectives on problem sensing from social cognition. Administrative Science Quarterly, 27(4): 548- 570. Kim, P. H., Ferrin, D. L., Cooper, C. D., & Dirks, K. T. 2004. Removing the shadow of suspicion: The effects of apology versus denial for repairing competence-versus integrity- based trust violations. Journal of Applied Psychology, 89(1): 104-118. 142 Kirchmeyer, C. 1998. Determinants of managerial career success: Evidence and explanation of male/female differences. Journal of Management, 24(6): 673-692. Koh, P. S. 2003. On the association between institutional ownership and aggressive corporate earnings management in Australia. The British Accounting Review, 35(2): 105-128. Kotha, S. & Nair, A. 1995. Strategy and environment as determinants of performance: Evidence from the Japanese machine tool industry. Strategic Management Journal, 16(7): 497- 518. Kreps, D. & Spence, A. M. 1985. Modelling the role of history in industrial organization and competition. London: MacMillan. Leary, M. R. & Schreindorfer, L. S. 1998. The stigmatization of HIV and AIDS: Rubbing salt in the wound. In V. J. Derlega & A. V. Barbee (Eds.), HIV Infection and Social Interaction. Thousand Oaks, CA: Sage. Lehmberg, D., Rowe, W. G., White, R. E., & Phillips, J. R. 2009. The GE paradox: Competitive advantage through fungible non-firm-specific investment. Journal of Management, 35(5): 1129-1153. Levinson, H. 1993. Between CEO and COO. Academy of Management Executive, 7(2): 71-81. Li, J. J., Poppo, L., & Zhou, K. Z. 2008. Do managerial ties in China always produce value? Competition, uncertainty, and domestic vs. foreign firms. Strategic Management Journal, 29(4): 383-400. 143 Lieberson, S. & O'Connor, J. F. 1972. Leadership and organizational performance: A study of large corporations. American Sociological Review, 37(2): 117-130. Long, J. S. 1978. Productivity and academic position in the scientific career. American Sociological Review, 43(6): 889-908. Malmendier, U. & Tate, G. 2008. Who makes acquisitions? CEO overconfidence and the market's reaction. Journal of Financial Economics, 89(1): 20-43. March, J. G. & Simon, H. A. 1958. Organizations. New York: Wiley. Marquis, C. 2003. The pressure of the past: Network imprinting in intercorporate communities. Administrative Science Quarterly, 48: 655-689. Milbourn, T. T. 2003. CEO reputation and stock-based compensation. Journal of Financial Economics, 68(2): 233-262. O'Sullivan, K. 2004. Career tracks: Where CFOs got their start. CFO Magazine: http://www.cfo.com/article.cfm/3329154/c_3348416? Ocasio, W. 1994. Political dynamics and the circulation of power: CEO succession in US industrial corporations, 1960-1990. Administrative Science Quarterly, 39(2): 285-312. Palmer, D., Friedland, R., & Singh, J. V. 1986. The ties that bind: Organizational and class bases of stability in a corporate interlock network. American Sociological Review, 51(6): 781- 796. Pfeffer, J. 1977. The ambiguity of leadership. Academy of Management Review, 2(1): 104-112. 144 Pfeffer, J. & Salancik, G. R. 1977. Organizational context and the characteristics and tenure of hospital administrators. Academy of Management Journal, 20(1): 74-88. Pfeffer, J. & Moore, W. L. 1980. Power in university budgeting: A replication and extension. Administrative Science Quarterly, 25(4): 637-653. Pitcher, P., Chreim, S., & Kisfalvi, V. 2000. CEO succession research: Methodological bridges over troubled waters. Strategic Management Journal, 21(6): 625-648. Podolny, J. M. 2005. Status signals: A sociological study of market competition. Princeton, NJ: Princeton University Press. Porter, M. E., Lorsch, J. W., & Nohria, N. 2004. Seven surprises for new CEOs. Harvard Business Review, 82(10): 62-75. Portes, A. 1998. Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24(1): 1-24. Pozner, J. E. 2008. Stigma and settling up: An integrated approach to the consequences of organizational misconduct for organizational elites. Journal of Business Ethics, 80(1): 141-150. Ranft, A. L., Zinko, R., Ferris, G. R., & Buckley, R. M. 2006. Marketing the image of management:: The costs and benefits of CEO reputation. Organizational Dynamics, 35(3): 279-290. 145 Rao, H. 1994. The social construction of reputation: Certification contests, legitimation, and the survival of organizations in the American automobile industry: 1895-1912. Strategic Management Journal, 15(S1): 29-44. Rechner, P. L. & Dalton, D. R. 1991. CEO duality and organizational performance: A longitudinal analysis. Strategic Management Journal, 12(2): 155-160. Rindova, V., Williamson, I. O., Petkova, A. P., & Sever, J. M. 2005. Being good or being known: An empirical examination of the dimensions, antecedents, and consequences of organizational reputation. Academy of Management Journal, 48(6): 1033-1049. Rindova, V. P. & Fombrun, C. J. 1999. Constructing competitive advantage: The role of firm– constituent interactions. Strategic Management Journal, 20(8): 691-710. Rosenbaum, J. E. 1984. Career mobility in a corporate hierarchy. Orlando, FL: Academic Press. Salancik, G. R. & Pfeffer, J. 1980. Effects of ownership and performance on executive tenure in US corporations. Academy of Management Journal, 23(4): 653-664. Schwartz, K. B. & Menon, K. 1985. Executive succession in failing firms. Academy of Management Journal, 28(3): 680-686. Shen, W. & Cannella, A. A. 2002. Power dynamics within top management and their impacts on CEO dismissal followed by inside succession. Academy of Management Journal, 45(6): 1195-1206. 146 Shen, W. & Cho, T. S. 2005. Exploring involuntary executive turnover through a managerial discretion framework. Academy of Management Review, 30(4): 843-854. Sigal, J., Hsu, L., Foodim, S., & Betman, J. 1988. Factors affecting perceptions of political candidates accused of sexual and financial misconduct. Political Psychology, 9(2): 273- 280. Simon, H. A. 1997. Administrative behavior: A study of decision-making processes in administrative organizations. New York: Simon and Schuster. Sonnenfeld, J. A. & Ward, A. 2007. Firing back: How great leaders rebound after career disasters. Boston, MA: Harvard Business School Press. Spence, A. M. 1974. Market signaling: Informational transfer in hiring and related screening processes. Boston, MA: Harvard Univ Press. Spence, M. 1973. Job market signaling. Quarterly Journal of Economics, 87(3): 355-374. Stuart, T. E. 2000. Interorganizational alliances and the performance of firms: A study of growth and innovation rates in a high-technology industry. Strategic Management Journal, 21(8): 791-811. Sutton, R. I. & Callahan, A. L. 1987. The stigma of bankruptcy: Spoiled organizational image and its management. Academy of Management Journal, 30(3): 405-436. Tetlock, P. E. 2002. Social functionalist frameworks for judgment and choice: Intuitive politicians, theologians, and prosecutors. Psychological Review, 109(3): 451-471. 147 Tversky, A. K. & Kahneman, D. 1974. Judgment under uncertainty: Heuristics and biases. Science, 185: 1124-1131. Useem, M. 1984. The inner circle: Oxford Univ. Press New York. Useem, M. & Karabel, J. 1986. Pathways to top corporate management. American Sociological Review, 51(2): 184-200. Vancil, R. 1987. Passing the baton: Managing the process of CEO succession. Boston, MA: Harvard Business School Press. Wade, J. B., Porac, J. F., Pollock, T. G., & Graffin, S. D. 2006. The burden of celebrity: The impact of CEO certification contests on CEO pay and performance. Academy of Management Journal, 49(4): 643. Walsh, J. P. & Seward, J. K. 1990. On the efficiency of internal and external corporate control mechanisms. Academy of Management Review, 15(3): 421-458. Ward, A., Sonnenfeld, J. A., & Kimberly, J. R. 1995. In search of a kingdom: Determinants of subsequent career outcomes for chief executives who are fired. Human Resource Management, 34(1): 117-139. Ward, A. J. & Feldman, D. C. 2008. Survival of the embedded: Expelling and embedding forces on members of the corporate elite. Corporate Governance: An International Review, 16(3): 239-251. Warner, J. B., Watts, R. L., & Wruck, K. H. 1988. Stock prices and top management changes. Journal of Financial Economics, 20: 461-492. 148 Washington, M. & Zajac, E. J. 2005. Status evolution and competition: Theory and evidence. Academy of Management Journal, 48(2): 282-296. Westphal, J. D. 1999. Collaboration in the boardroom: Behavioral and performance consequences of CEO-board social ties. Academy of Management Journal, 42(1): 7-24. Westphal, J. D. & Stern, I. 2006. The other pathway to the boardroom: Interpersonal influence behavior as a substitute for elite credentials and majority status in obtaining board appointments. Administrative Science Quarterly, 51(2): 169-204. Wiersema, M. 2002. Holes at the top: Why CEO firings backfire. Harvard Business Review, 80(12): 70-77. Wiersema, M. F. & Bantel, K. A. 1992. Top management team demography and corporate strategic change. Academy of Management Journal, 35(1): 91-121. Wiersema, M. F. & Zhang, Y. 2011. CEO dismissal: The role of investment analysts. Strategic Management Journal, 32(11): 1161-1182. Wiesenfeld, B. M., Wurthmann, K. A., & Hambrick, D. C. 2008. The stigmatization and devaluation of elites associated with corporate failures: A process model. Academy of Management Review, 33(1): 231-251. Worrell, D., Davidson III, W., Chandy, P., & Garrison, S. 1986. Management turnover through deaths of key executives: Effects on investor wealth. Academy of Management Journal, 29(4): 674-694. 149 Worrell, D. L., Nemec, C., & Davidson, W. N. 1997. One hat too many: Key executive plurality and shareholder wealth. Strategic Management Journal, 18(6): 499-507. Zajac, E. J. 1990. CEO selection, succession, compensation and firm performance: A theoretical integration and empirical analysis. Strategic Management Journal, 11(3): 217-230. Zajac, E. J. & Westphal, J. D. 1996. Director reputation, CEO-board power, and the dynamics of board interlocks. Administrative Science Quarterly, 41(3): 507-529. Zhang, Y. & Rajagopalan, N. 2003. Explaining new CEO origin: Firm versus industry antecedents. The Academy of Management Journal, 46(3): 327-338. Zhang, Y. & Rajagopalan, N. 2004. When the known devil is better than an unknown god: An empirical study of the antecedents and consequences of relay CEO successions. Academy of Management Journal, 47(4): 483-500. Zhang, Y. 2006. The presence of a separate COO/president and its impact on strategic change and CEO dismissal. Strategic Management Journal, 27(3): 283-300. Zhang, Y. 2008. Information asymmetry and the dismissal of newly appointed CEOs: An empirical investigation. Strategic Management Journal, 29(8): 859-872. 150 FIGURE 1 A MODEL OF EXECUTIVE CAPITAL AND ALTERNATIVE CEO CANDIDATES ON THE LIKELIHOOD OF EXECUTIVE DISMISSAL 151 FIGURE 2 PLOT OF THE INTERACTION BETWEEN FIRM ROA AND CEO COMPENSATION ON THE LIKELIHOOD OF DISMISSAL FIGURE 3 PLOT OF THE INTERACTION BETWEEN FIRM ROA AND NEGATIVE PUBLICITY ON THE LIKELIHOOD OF DISMISSAL 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Low ROA High ROA Pr o ba bi lit y o f d ism iss a l Low CEO Compensation High CEO Compensation Average CEO Compensation 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Low ROA High ROA Pr o ba bi lit y o f s u cc es s Low Negative Publicity High Negative Publicity Average Negative Publicity 152 FIGURE 4 PLOT OF THE INTERACTION BETWEEN FIRM ROA AND FIRM LOCATED IN MAJOR CITY ON THE LIKELIHOOD OF DISMISSAL FIGURE 5 PLOT OF THE INTERACTION BETWEEN TOBIN’S Q AND PRESTIGIOUS EXPERIENCE ON THE LIKELIHOOD OF DISMISSAL 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Low ROA High ROA Pr o ba bi lit y o f d ism iss a l Firm not located in major city Firm located in major city 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Low Tobin's Q High Tobin's Q Pr o ba bi lit y o f d ism iss a l Does Not Have Prestigious Experience Has Prestigious Experience 153 FIGURE 6 A MODEL FOR THE LIKELIHOOD OF EXECUTIVE RE-EMPLOYMENT AFTER DISMISSAL AS CEO 154 FIGURE 7 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND PRESTIGIOUS EXPERIENCE Performance Dismissal No Yes Prestigious Experience No 0 / 24 (0%) 1 / 17 (5.88%) Yes 4 / 25 (16.00%) 5 / 22 (22.72%) FIGURE 8 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND ELITE EDUCATION Performance Dismissal No Yes Elite Education No 1 / 33 (3.03%) 4 / 26 (15.38%) Yes 3 / 16 (18.75%) 0 / 13 (0%) 155 FIGURE 9 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND BEST CEO AWARD Performance Dismissal No Yes Best CEO No 3 / 47 (6.38%) 6 / 38 (15.78%) Yes 1 / 2 (50.00%) 0 / 1 (0%) FIGURE 10 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND NEGATIVE PUBLICITY Performance Dismissal No Yes Negative Publicity Below Average 1 / 38 (2.63%) 3 / 34 (8.82%) Above Average 3 / 11 (27.27%) 2 / 5 (40%) 156 FIGURE 11 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND NUMBER OF EXECUTIVE EMPLOYERS Executive Employers 0 1 2 Performance Dismissal No 3 / 41 (7.32%) 1 / 6 (16.67%) 0 / 2 (0%) Yes 4 / 31 (12.90%) 1 / 7 (14.29%) 1 / 1 (100%) FIGURE 12 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND RESIDENCE IN MAJOR METROPOLITAN AREA Performance Dismissal No Yes Residence in Major City No 2 / 20 (10%) 2 / 13 (15.38%) Yes 2 / 29 (6.90%) 4 / 26 (15.38%) 157 FIGURE 13 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND EDUCATION LEVEL Education Level High School Some College Undergrad Graduate Doctorate Fiduciary Violation No 0 / 2 (0%) 0 / 0 (0%) 4 / 26 (15.38%) 6 / 44 (13.64%) 0 / 1 (0%) Yes 0 / 0 (0%) 0 / 0 (0%) 0 / 7 (0%) 0 / 8 (0%) 0 / 0 (0%) FIGURE 14 STATISTICS FOR AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND BEST CEO Fiduciary Violation No Yes Best CEO No 9 / 71 (12.68%) 0 / 14 (0%) Yes 1 / 2 (50%) 0 / 1 (0%) 158 FIGURE 15 STATISTICS FOR AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND NEGATIVE PUBLICITY Fiduciary Violation No Yes Negative Publicity Below Average 5 / 62 (8.06%) 0 / 10 (0%) Above Average 5 / 11 (45.45%) 0 / 5 (0%) FIGURE 16 STATISTICS FOR AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND NUMBER OF EXECUTIVE EMPLOYERS Executive Employers 0 1 2 Fiduciary Violation No 7 / 60 (11.67%) 2 / 10 (20%) 1 / 3 (33.33%) Yes 0 / 12 (0%) 0 / 3 (0%) 0 / 0 (0%) 159 FIGURE 17 STATISTICS FOR EMPLOYMENT AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND RESIDENCE IN MAJOR CITY Fiduciary Violation No Yes Residence in Major City No 4 / 29 (13.79%) 0 / 4 (0%) Yes 6 / 44 (13.63%) 0 / 11 (0%) FIGURE 18 STATISTICS FOR EMPLOYMENT AS PUBLIC CEO FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND PRESTIGIOUS EXPERIENCE Performance Dismissal No Yes Prestigious Experience No 0 / 24 (0%) 1 / 17 (5.88%) Yes 3 / 25 (12.00%) 4 / 22 (18.18%) 160 FIGURE 19 STATISTICS FOR EMPLOYMENT AS PUBLIC CEO FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND ELITE EDUCATION Performance Dismissal No Yes Elite Education No 1 / 33 (3.03%) 2 / 16 (12.5%) Yes 5 / 26 (19.23%) 0 / 13 (0%) FIGURE 20 STATISTICS FOR EMPLOYMENT AS PUBLIC CEO FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND NEGATIVE PUBLICITY Performance Dismissal No Yes Negative Publicity Below Average 0 / 38 (0%) 3 / 11 (27.27%) Above Average 4 / 34 (11.76%) 1 / 5 (20%) 161 FIGURE 21 STATISTICS FOR AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND NUMBER OF DIRECTORSHIPS Cumulative Number of Directorships 0 1 2 3 4 5 6 Performance Dismissal No 1 / 27 (3.70%) 0 / 9 (0%) 2 / 5 (40%) 0 / 3 (0%) 0 / 2 (0%) 0 / 1 (0%) 0 / 2 (0%) Yes 2 / 12 (16.67%) 2 / 9 (22.22%) 0 / 9 (0%) 1 / 4 (25%) 0 / 1 (0%) 0 / 2 (0%) 0 / 2 (0%) FIGURE 22 STATISTICS FOR AS PUBLIC EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND NUMBER OF EXECUTIVE EMPLOYERS Executive Employers 0 1 2 Performance Dismissal No 2 / 41 (4.88%) 1 / 6 (16.67%) 0 / 2 (0%) Yes 3 / 31 (0%) 0 / 6 (0%) 1 / 2 (50%) 162 FIGURE 23 STATISTICS FOR EMPLOYMENT AS PUBLIC CEO FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND RESIDENCE IN MAJOR CITY Performance Dismissal No Yes Residence in Major City No 1 / 20 (5%) 2 / 29 (6.90%) Yes 1 / 13 (7.69%) 4 / 26 (15.38%) FIGURE 24 STATISTICS FOR EMPLOYMENT AS PUBLIC CEO FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND EDUCATION LEVEL Education Level High School Some College Undergrad Graduate Doctorate Fiduciary Violation No 0 / 2 (0%) 0 / 0 (0%) 3 / 26 (11.54%) 5 / 44 (11.36%) 0 / 1 (0%) Yes 0 / 0 (0%) 0 / 0 (0%) 0 / 7 (0%) 0 / 8 (0%) 0 / 0 (0%) 163 FIGURE 25 STATISTICS FOR AS PUBLIC CEO FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND NEGATIVE PUBLICITY Fiduciary Violation No Yes Negative Publicity Below Average 4 / 62 (6.45%) 0 / 10 (0%) Above Average 4 / 11 (36.36%) 0 / 5 (0%) FIGURE 26 STATISTICS FOR AS PUBLIC CEO FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND NUMBER OF EXECUTIVE EMPLOYERS Executive Employers 0 1 2 Fiduciary Violation No 6 / 60 (10%) 1 / 10 (10%) 1 / 3 (33.33%) Yes 0 / 12 (0%) 0 / 3 (0%) 0 / 0 (0%) 164 FIGURE 27 STATISTICS FOR EMPLOYMENT AS PUBLIC CEO FOR INTERACTION BETWEEN VIOLATION OF FIDUCIARY DUTY AND RESIDENCE IN MAJOR CITY Fiduciary Violation No Yes Residence in Major City No 2 / 29 (6.90%) 0 / 4 (0%) Yes 6 / 44 (13.63%) 0 / 11 (0%) FIGURE 28 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND EDUCATION LEVEL Education Level High School Some College Undergrad Graduate Doctorate Performance Dismissal No 0 / 1 (0%) 0 / 0 (0%) 4 / 20 (20%) 8 / 27 (29.63%) 1 / 1 (100%) Yes 0 / 1 (0%) 0 / 0 (0%) 8 / 13 (61.54%) 7 / 25 (28%) 0 / 0 (0%) 165 FIGURE 29 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN PERFORMANCE DISMISSAL AND ELITE EDUCATION Performance Dismissal No Yes Elite Education No 6 / 33 (18.18%) 11 / 26 (42.31%) Yes 7 / 16 (43.75%) 2 / 13 (15.38%) FIGURE 30 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN FIDUCIARY VIOLATION AND PRESTIGE EXPERIENCE Fiduciary Violation No Yes Prestige Experience No 8 / 33 (24.24%) 3 / 8 (37.5%) Yes 16 / 40 (40%) 1 / 7 (14.29%) 166 FIGURE 31 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN FIDUCIARY VIOLATION AND BEST CEO Fiduciary Violation No Yes Best CEO No 49 / 71 (69.01%) 4 / 14 (28.57%) Yes 2 / 2 (100%) 0 / 1 (0%) FIGURE 32 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN FIDUCIARY VIOLATION AND BEST CEO Fiduciary Violation No Yes Negative Publicity Below Average 17 / 62 (27.42%) 2 / 10 (20%) Above Average 7 / 11 (100%) 2 / 4 (50%) 167 FIGURE 33 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN PERSONAL CONDUCT VIOLATION AND PRESTIGE EXPERIENCE Personal Conduct Violation No Yes Prestigious Experience No 11 / 40 (27.5%) 0 / 1 (0%) Yes 16 / 45 (35.56%) 1 / 2 (50%) FIGURE 34 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN PERSONAL CONDUCT VIOLATION AND PRESTIGE EXPERIENCE Education Level High School Some College Undergrad Graduate Doctorate Personal Conduct Violation No 0 / 2 (0%) 0 / 0 (0%) 11 / 31 (35.48%) 15 / 51 (29.41%) 1 / 1 (100%) Yes 0 / 0 (0%) 0 / 0 (0%) 1 / 2 (50%) 0 / 1 (0%) 0 / 0 (0%) 168 FIGURE 35 STATISTICS FOR EMPLOYMENT AS PUBLIC OR PRIVATE EXECUTIVE FOR INTERACTION BETWEEN PERSONAL CONDUCT VIOLATION AND NUMBER OF EXECUTIVE POSITIONS Executive Employers 0 1 2 Personal Conduct Violation No 21 / 70 (30%) 5 / 13 (38.46%) 1 / 2 (50%) Yes 1 / 2 (50%) 0 / 0 (0%) 0 / 1 (0%) 169 TA BL E 1 M ea n s, St a n da rd de v ia tio n , a n d C o rr el a tio n s fo r St u dy 1 – Li ke lih o o d o f C EO D ism iss a l i n a G iv en Y ea r M ea n S. D . 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. D ism iss al . 03 . 17 2. D u al ity . 60 . 49 - . 08 * * * 3. O u ts id e D ire ct o rs . 84 . 08 . 00 . 06 * * * 4. Bl o ck ho ld er O w n er sh ip . 24 . 18 . 04 * - . 02 . 02 5. D eb t t o Eq u ity R at io . 03 45 . 09 - . 00 - . 02 - . 01 . 01 6. Fi rm Si ze 7. 52 1. 43 - . 03 . 15 * * * . 24 * * * - . 23 * * * . 01 7. CE O G en de r . 98 . 14 . 00 - . 00 . 01 - . 07 * * * - . 01 - . 05 * * 8. CE O O w n er sh ip 4. 10 10 . 88 - . 03 * . 11 * * * - . 29 * * * - . 17 * * * - . 00 - . 20 * * * . 04 * * 9. Cu rr en t R at io 2. 49 4. 60 - . 02 - . 02 - . 05 * * . 02 . 00 - . 15 * * * . 01 . 02 10 . Fi rm RO A - . 00 . 78 - . 02 . 03 . 01 - . 02 . 00 - . 03 . 01 - . 01 . 93 * * * 11 . To bi n ’ s Q - . 03 . 85 - . 07 * * * . 02 - . 00 - . 10 * * * . 01 . 04 . 01 - . 07 * * * . 00 . 03 12 . CE O Te n u re 16 . 37 11 . 90 - . 06 * * * . 20 * * * - . 17 * * * - . 11 * * * - . 04 * . 04 * * . 08 * * * . 23 * * * . 01 . 00 - . 00 13 . Pr es tig e Ex pe rie n ce . 15 . 35 . 01 . 06 * * . 14 * * * . 01 . 02 . 11 * * * - . 02 - . 10 * * * . 01 . 04 * . 01 - . 30 * * * 14 . Pr io r CE O Ex pe rie n ce . 21 . 41 . 03 - . 01 . 02 . 03 . 00 . 01 - . 03 . 01 . 02 . 02 - . 02 - . 33 * * * . 13 * * * 15 . Ed u ca tio n Le v el 4. 41 1. 31 . 01 . 03 . 10 * * * - . 01 - . 00 . 09 * * * - . 01 - . 06 * * . 00 . 01 . 03 - . 12 * * * . 05 * * . 05 * * 16 . El ite Ed u ca tio n . 34 . 47 . 01 . 06 * * * . 04 * - . 03 . 01 . 09 * * * - . 05 * * . 02 . 02 . 03 . 09 * * * . 02 . 02 . 03 17 . CE O Co m pe n sa tio n 78 1. 36 36 4. 01 - . 07 * * * . 22 * * * . 17 * * * - . 19 * * * . 00 . 68 * * * - . 03 - . 03 - . 14 * * * - . 02 . 02 . 08 * * * . 09 * * * . 05 * * 18 . IIM To p 5 CE O . 17 . 37 - . 02 . 16 * * * . 15 * * * - . 19 * * * . 00 . 53 * * * - . 01 - . 08 * * * - . 04 * . 01 . 20 * * * . 09 * * * - . 00 - . 01 19 . N eg at iv e Pu bl ic ity . 02 . 22 . 05 * * . 03 . 04 * - . 05 * * - . 00 . 19 * * * - . 09 * * * - . 02 - . 03 - . 00 - . 00 . 03 * . 04 * . 02 20 . CO O . 43 . 49 - . 01 . 09 * * * - . 09 * * * - . 04 * . 02 . 04 * . 00 . 03 * - . 02 - . 01 . 03 . 08 * * * - . 01 . 01 21 . N o n - CE O B o ar d In sid er s . 45 . 74 . 02 - . 08 * * * - . 84 * * * - . 07 * * * . 01 . 00 - . 04 * . 17 * * * - . 00 - . 01 . 00 . 21 * * * - . 13 * * * - . 05 * * * 22 . Fi rm D iv isi o n s 2. 55 1. 44 - . 01 . 09 * * * . 11 * * * - . 13 * * * . 00 . 40 * * * . 03 - . 09 * * * - . 04 * * . 03 - . 12 * * * . 06 * * * . 04 * * - . 01 23 . Re tir ed CE O o n Bo ar d . 27 . 44 . 02 - . 45 * * * - . 12 * * * - . 05 * * . 01 - . 05 * * . 01 - . 08 * * * . 01 - . 01 . 03 - . 07 * * * - . 07 * * * - . 02 24 . O th er Re tir ed CE O o n Bo ar d . 77 . 42 . 01 . 01 . 20 * * * . 02 . 03 . 24 * * * - . 00 - . 17 * * * - . 03 . 01 . 01 - . 11 * * * . 08 * * * . 00 25 . In du st ry Co n ce n tr at io n - 2. 64 . 58 - . 02 . 06 * * * . 03 * . 07 * * * - . 00 - . 02 - . 03 * - . 00 - . 02 . 03 . 10 * * * . 03 - . 02 . 03 26 . Fo rt u n e 50 0 Fi rm s in In du st ry 16 . 26 9. 58 - . 00 . 04 * . 05 * * - . 06 * * - . 01 . 19 * * * . 01 - . 02 - . 04 * - . 03 - . 04 * . 02 . 02 - . 02 27 . D ire ct o rs ’ B o ar d M em be rs hi ps 20 . 04 9. 17 - . 01 . 12 * * * . 28 * * * - . 16 * * * . 02 . 57 * * * - . 02 - . 16 * * * - . 08 * * * . 03 . 07 * * * - . 04 * . 14 * * * . 03 28 . Fi rm ’ s G eo gr ap hi c Lo ca tio n . 68 . 47 - . 01 - . 02 - . 02 - . 07 * * * - . 01 . 09 * * * . 02 . 01 - . 02 - . 03 . 03 * - . 11 * * * . 05 * * . 10 * * * n = 3, 64 8 Tw o - ta ile d co eff ic ie n t t es t ( N= 28 8). * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1 le ve l 170 15 16 17 18 19 20 21 22 23 24 25 26 27 16 . El ite Ed u ca tio n . 30 * * * 17 . CE O Co m pe n sa tio n . 08 * * * . 09 * * * 18 . II M To p 5 CE O . 04 * . 09 * * * . 42 * * * 19 . N eg at iv e Pu bl ic ity . 00 . 06 * * . 09 * * * . 10 * * * 20 . CO O - . 09 * * * - . 01 . 06 * * . 04 * . 01 21 . N o n - CE O B o ar d In sid er s - . 07 * * * - . 00 - . 00 - . 05 * * . 00 . 10 * * * 22 . Fi rm D iv is io n s . 05 * * . 09 * * * . 33 * * * . 15 * * * . 11 * * * - . 09 * * * . 01 23 . R et ire d CE O o n B o ar d - . 08 * * * - . 06 * * * - . 09 * * * - . 08 * * * - 0. 02 - . 02 . 18 * * * - . 06 * * * 24 . O th er R et ire d CE O o n B o ar d . 06 * * * . 04 * . 12 * * * . 12 * * * . 05 * * * . 01 - . 10 * * * . 09 * * * - . 06 * * * 25 . In du st ry Co n ce n tr at io n - . 09 * * * - . 01 . 03 - . 01 . 01 . 01 . 01 . 11 * * * - . 01 - . 06 * * 26 . Fo rt u n e 50 0 Fi rm s in In du st ry . 07 * * * . 02 . 11 * * * . 10 * * * . 07 * * * - . 08 * * * - . 04 * * - . 01 - . 03 . 11 * * * - . 56 * * * 27 . D ire ct o rs ’ B o ar d M em be rs hi ps . 09 * * * . 14 * * * . 43 * * * . 25 * * * . 12 * * * . 01 . 01 . 26 * * * - . 01 . 28 * * * . 03 . 11 * * * 28 . Fi rm ’ s G eo gr ap hi c Lo ca tio n . 05 * . 09 * * * . 09 * * * . 08 * * * . 05 * * . 04 * - . 01 . 03 - . 03 . 08 * * * - . 15 * * * . 07 * * * . 07 * * * 171 TABLE 2 Probit Analysis for Likelihood of CEO Dismissal in a Given Yeara Model 1 Model 2 Model 3 Coeff. Robust Std. Error Coeff. Robust Std. Error Coeff. Robust Std. Error Intercept -0.80 (0.70) -1.13 (0.70) -3.69* (1.49) Duality -0.40*** (0.09) -0.39*** (0.09) -0.33** (0.11) Outside Directors -0.03 (0.69) -0.03 (0.70) 2.52 (1.72) Blockholder Ownership 0.35 (0.26) 0.25 (0.28) 0.22 (0.29) Debt to Equity Ratio -0.00 (0.00) -0.00 (0.00) -0.00 (0.00) Firm Size -0.07† (0.04) -0.06 (0.04) 0.01 (0.06) CEO Gender 0.07 (0.32) 0.12 (0.32) 0.52† (0.31) CEO Ownership -0.02† (0.01) -0.02* (0.01) -0.02† (0.01) Current Ratio -0.19*** (0.05) -0.18*** (0.05) -0.19*** (0.05) H17: Firm ROA -0.83** (0.32) -0.75* (0.34) H17: Tobin’s Q -0.26*** (0.07) -0.28*** (0.07) H1: CEO Tenure -0.01* (0.01) H2: Prestige Experience -0.02 (0.13) H3: Prior CEO Experience 0.06 (0.11) H4: Education Level -0.01 (0.04) H5: Elite Education 0.21* (0.10) H6: CEO Compensation -0.00** (0.00) H7: IIM Top 5 CEO 0.25 (0.16) H7: Negative Publicity 0.31* (0.13) H8: COO 0.02 (0.10) H9: Non –CEO Board Insiders 0.34* (0.17) H10: Firm Divisions -0.04 (0.04) H11: Retired CEO on Board -0.09 (0.11) H12: Other Retired CEO on Board -0.07 (0.12) H13: Industry Concentration -0.14 (0.11) H14: Fortune 500 Firms in Industry -0.01 (0.01) H15: Directors’ Board Memberships -0.00 (0.01) H16: Firm’s Geographic Location -0.14 (0.10) -Log Pseudolikelihood 433.93 50.55*** 0.08 0.07 420.22 77.93*** 0.11 0.10 399.11 117.24*** 0.16 0.14 Wald-χ2 Nagelkerke r2 Pseudo r2 Dependent variable is whether a firm dismissed its CEO in a given year or not; n=3648 † p < 0.1, * p < 0.05, ** p < 0.01, *** p <0.001, Two-tailed coefficient tests. a. Dummy variables for each year were included in the model, but are not reported for sake of parsimony. No year dummies were significant. 172 T A BL E 3a Pr o bi t A n a ly sis o f t he L ik el ih o o d o f C E O D ism iss a l – Te st s o f I n te ra ct io n s be tw ee n Fi rm R O A , H u m a n a n d R ep u ta tio n a l C a pi ta l, a n d th e M a rk et fo r A lte rn a tiv e C a n di da te s M o de l 4 M o de l 5 M o de l 6 M o de l 7 M o de l 8 M o de l 9 M o de l 1 0 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r H 18 : R O A x CE O Te n u re 0. 03 (0. 04 ) 0. 04 (0. 05 ) H 19 : R O A x Pr es tig e Ex pe rie n ce 0. 73 (0. 70 ) 0. 40 (0. 63 ) H 20 : R O A x Pr io r CE O Ex pe rie n ce 0. 52 (0. 66 ) 0. 05 (0. 62 ) H 21 : R O A x Ed u ca tio n Le v el 0. 02 (0. 24 ) 0. 25 (0. 19 ) H 22 : R O A x El ite Ed u ca tio n - 0. 32 (0. 76 ) 1. 34 (0. 86 ) H 23 : R O A x CE O Co m pe n sa tio n 0. 00 (0. 00 ) 0. 00 * * (0. 00 ) H 24 : R O A x II M To p 5 CE O 0. 05 (1. 16 ) H 24 : R O A x N eg at iv e Pu bl ic ity 3. 08 (2. 12 ) H 25 : R O A x CO O - 0. 66 (0. 69 ) H 26 : R O A x N o n - CE O B o ar d In sid er s - 0. 47 (0. 48 ) H 27 :R O A x Fi rm D iv isi o n s 0. 19 (0. 31 ) H 28 : R O A x R et ire d CE O o n B o ar d - 0. 74 (0. 76 ) H 29 : R O A x O th er R et ire d CE O o n Bo ar d 0. 20 (0. 65 ) H 30 : R O A x In du st ry Co n ce n tr at io n 0. 61 (0. 55 ) H 31 : R O A x Fo rt u n e 50 0 Fi rm s in In du st ry - 0. 01 (0. 04 ) H 32 : R O A x D ire ct o rs ’ B o ar d M em be rs hi ps - 0. 02 (0. 05 ) H 33 : R O A x Fi rm G eo gr ap hi c Lo ca tio n 2. 41 * * (0. 77 ) - Lo g Ps eu do lik el ih o o d 38 3. 31 18 0. 09 * * * 0. 20 0. 18 39 8. 23 12 8. 07 * * * 0. 16 0. 14 39 8. 96 11 7. 26 * * * 0. 16 0. 14 39 9. 11 11 7. 23 * * * 0. 16 0. 14 39 8. 37 11 9. 67 * * * 0. 16 0. 14 39 6. 98 13 0. 93 * * * 0. 16 0. 14 39 2. 60 14 8. 94 * * * 0. 17 0. 16 W al d- χ2 N ag el ke rk e r2 Ps eu do r2 D ep en de n t v a ria bl e is w he th er a fir m di sm is se d its CE O in a gi ve n ye a r o r n o t; n = 36 48 † p < 0 . 1, * p < 0 . 05 , * * p < 0 . 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 173 TA BL E 3b Pr o bi t A n a ly sis o f t he L ik el ih o o d o f C E O D ism iss a l – Te st s o f I n te ra ct io n s be tw ee n Fi rm R O A , H u m a n a n d R ep u ta tio n a l C a pi ta l, a n d th e M a rk et fo r A lte rn a tiv e C a n di da te s M o de l 1 1 M o de l 1 2 M o de l 1 3 M o de l 1 4 M o de l 1 5 M o de l 1 6 M o de l 1 7 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r H 24 : R O A x II M To p 5 CE O 2. 96 (2. 05 ) H 24 : R O A x N eg at iv e Pu bl ic ity 6. 05 * * * (1. 51 ) H 25 : R O A x CO O - 0. 32 (0 . 64 ) H 26 : R O A x N o n - CE O B o ar d In sid er s - 0. 65 † (0. 40 ) H 27 :R O A x Fi rm D iv isi o n s 0. 38 (0. 32 ) H 28 : R O A x R et ire d CE O o n B o ar d - 1. 11 † (0. 66 ) H 29 : R O A x O th er R et ire d CE O o n B o ar d 0. 66 (0. 62 ) - Lo g Ps eu do lik el ih o o d 39 5. 23 13 0. 12 * * * 0. 17 0. 15 39 1. 63 13 6. 06 * * * 0. 18 0. 16 39 8. 94 12 2. 04 * * * 0. 16 0. 14 39 7. 65 12 2. 82 * * * 0. 16 0. 15 39 8. 05 12 6. 11 * * * 0. 16 0. 14 39 7. 43 12 3. 52 * * * 0. 16 0. 15 39 8. 60 12 2. 10 * * * 0. 16 0. 14 W al d- χ2 N ag el ke rk e r2 Ps eu do r2 D ep en de n t v a ria bl e is w he th er a fir m di sm is se d its CE O in a gi ve n ye a r o r n o t; n = 36 48 † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 174 TA BL E 3c Pr o bi t A n a ly sis o f t he Li ke lih o o d o f C EO D ism iss a l – Te st s o f I n te ra ct io n s be tw ee n Fi rm R O A , H u m a n a n d R ep u ta tio n a l C a pi ta l, a n d th e M a rk et fo r A lte rn a tiv e C a n di da te s M o de l 1 8 M o de l 1 9 M o de l 2 0 M o de l 2 1 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r H 30 : R O A x In du st ry Co n ce n tr at io n 0. 64 (0. 52 ) H 31 : R O A x Fo rt u n e 50 0 Fi rm s in In du st ry - 0. 00 (0. 04 ) H 32 : R O A x D ire ct o rs ’ B o ar d M em be rs hi ps 0. 04 (0. 06 ) H 33 : R O A x Fi rm G eo gr ap hi c Lo ca tio n 1. 88 * * (0. 67 ) - Lo g Ps eu do lik el ih o o d 39 8. 21 11 8. 71 * * * 0. 16 0. 14 39 9. 10 11 7. 97 * * * 0. 16 0. 14 39 8. 48 12 5. 35 * * * 0. 16 0. 14 39 5. 08 13 3. 90 * * * 0. 17 0. 15 W al d- χ2 N ag el ke rk e r2 Ps eu do r2 D ep en de n t v a ria bl e is w he th er a fir m di sm iss ed its CE O in a gi ve n ye a r o r n o t; n = 36 48 † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 175 TA BL E 4a Pr o bi t A n a ly sis o f t he Li ke lih o o d o f C EO D ism iss a l – T es ts o f I n te ra ct io n s be tw ee n T o bi n ’s Q, H u m a n a n d R ep u ta tio n a l C a pi ta l, a n d th e M a rk et fo r A lte rn a tiv e C a n di da te s M o de l 2 2 M o de l 2 3 M o de l 2 4 M o de l 2 5 M o de l 2 6 M o de l 2 7 Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r H 18 : To bi n ’ s Q x CE O Te n u re - 0. 01 (0. 01 ) - 0. 01 (0. 01 ) H 19 : To bi n ’ s Q x Pr es tig e Ex pe rie n ce 0. 30 * (0. 15 ) 0. 31 * (0. 14 ) H 20 : To bi n ’ s Q x Pr io r CE O Ex pe rie n ce - 0. 30 * (0. 15 ) - 0. 19 (0. 19 ) H 21 : To bi n ’ s Q x Ed u ca tio n Le v el - 0. 02 (0. 04 ) - 0. 02 (0. 05 ) H 22 : To bi n ’ s Q x El ite Ed u ca tio n - 0. 05 (0. 13 ) - 0. 04 (0. 14 ) H 23 : To bi n ’ s Q x CE O Co m pe n sa tio n 0. 00 (0. 00 ) H 24 : To bi n ’ s Q x II M To p 5 CE O - 0. 11 (0. 20 ) H 24 : To bi n ’ s Q x N eg at iv e Pu bl ic ity - 0. 33 (0. 53 ) H 25 : To bi n ’ s Q x CO O - 0. 08 (0. 13 ) H 26 : To bi n ’ s Q x N o n - CE O B o ar d In sid er s 0. 16 † (0. 09 ) H 27 : To bi n ’ s Q x Fi rm D iv isi o n s - 0. 03 (0. 06 ) H 28 : To bi n ’ s Q x R et ire d CE O o n B o ar d 0. 12 (0. 13 ) H 29 : To bi n ’ s Q x O th er R et ire d CE O o n B o ar d 0. 21 (0. 18 ) H 30 : To bi n ’ s Q x In du st ry Co n ce n tr at io n - 0. 20 (0. 15 ) H 31 : To bi n ’ s Q x Fo rt u n e 50 0 Fi rm s in In du st ry - 0. 01 (0. 01 ) H 32 : To bi n ’ s Q x D ire ct o rs ’ B o ar d M em be rs hi ps 0. 01 (0. 01 ) H 33 : To bi n ’ s Q x Fi rm G eo gr ap hi c Lo ca tio n 0. 15 (0. 14 ) - Lo g Ps eu do lik el ih o o d 39 1. 07 14 7. 08 * * * 0. 17 0. 16 39 8. 69 11 6. 69 * * * 0. 16 0. 14 39 7. 74 12 0. 20 * * * 0. 16 0. 15 39 8. 59 11 8. 22 * * * 0. 16 0. 14 39 9. 06 11 7. 74 * * * 0. 16 0. 14 39 9. 08 11 8. 17 * * * 0. 16 0. 14 W al d- χ2 N ag el ke rk e r2 Ps eu do r2 D ep en de n t v a ri a bl e is w he th er a fir m di sm is se d its CE O in a gi ve n ye a r o r n o t; n = 36 48 † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 176 TA BL E 4b Pr o bi t A n a ly sis o f t he L ik el ih o o d o f C E O D ism iss a l – Te st s o f I n te ra ct io n s be tw ee n To bi n ’s Q, H u m a n a n d R ep u ta tio n a l C a pi ta l, a n d th e M a rk et fo r A lte rn a tiv e C a n di da te s M o de l 2 8 M o de l 2 9 M o de l 3 0 M o de l 3 1 M o de l 3 2 M o de l 3 3 M o de l 3 4 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r H 23 : To bi n ’ s Q x CE O Co m pe n sa tio n 0. 00 (0. 00 ) H 24 : To bi n ’ s Q x II M To p 5 CE O 0. 00 (0. 19 ) H 24 : To bi n ’ s Q x N eg at iv e Pu bl ic ity - 0. 28 (0. 50 ) H 25 : To bi n ’ s Q x CO O - 0. 10 (0. 14 ) H 26 : To bi n ’ s Q x N o n - CE O B o ar d In sid er s 0. 13 † (0. 07 ) H 27 : To bi n ’ s Q x Fi rm D iv isi o n s - 0. 01 (0. 05 ) H 28 : To bi n ’ s Q x Re tir ed CE O o n B o ar d 0. 20 (0. 14 ) - Lo g Ps eu do lik el ih o o d 39 8. 74 12 2. 26 * * * 0. 16 0. 14 39 9. 11 11 7. 64 * * * 0. 16 0. 14 39 8. 94 12 1. 19 * * * 0. 16 0. 14 39 8. 13 11 7. 07 * * * 0. 16 0. 14 39 8. 13 12 4. 24 * * * 0. 16 0. 14 39 9. 11 11 7. 07 * * * 0. 16 0. 14 39 8. 33 12 3. 35 * * * 0. 16 0. 14 W al d- χ2 N ag el ke rk e r2 Ps eu do r2 D ep en de n t v a ri a bl e is w he th er a fir m di sm is se d its CE O in a gi ve n ye a r o r n o t; n = 36 48 † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 177 T A B L E 4c Pr o bi t A n a ly sis o f t he L ik el ih o o d o f C EO D ism iss a l – T es ts o f I n te ra ct io n s be tw ee n T o bi n ’s Q, H u m a n a n d R ep u ta tio n a l C a pi ta l, a n d th e M a rk et fo r A lte rn a tiv e C a n di da te s M o de l 3 5 M o de l 3 6 M o de l 3 7 M o de l 3 8 M o de l 3 9 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r H 29 : To bi n ’ s Q x O th er R et ire d CE O o n B o ar d 0. 32 (0. 19 ) H 30 : To bi n ’ s Q x In du st ry Co n ce n tr at io n - 0. 20 (0. 14 ) H 31 : To bi n ’ s Q x Fo rt u n e 50 0 Fi rm s in In du st ry 0. 01 (0. 01 ) H 32 : To bi n ’ s Q x D ire ct o rs ’ B o ar d M em be rs hi ps 0. 01 † (0. 01 ) H 33 : To bi n ’ s Q x Fi rm G eo gr ap hi c Lo ca tio n 0. 17 (0. 16 ) - Lo g Ps eu do lik el ih o o d 39 7. 92 11 7. 43 * * * 0. 16 0. 15 39 8. 22 12 3. 75 * * * 0. 16 0. 14 39 8. 91 11 8. 49 * * * 0. 16 0. 14 39 7. 70 11 7. 52 * * * 0. 16 0. 15 39 8. 60 11 9. 42 * * * 0. 16 0. 14 W al d- χ2 N ag el ke rk e r2 Ps eu do r2 D ep en de n t v a ria bl e is w he th er a fir m di sm iss ed its CE O in a gi ve n ye a r o r n o t; n = 36 48 † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 178 TA BL E 5 M ea n s, St a n da rd de v ia tio n , a n d C o rr el a tio n s fo r St u dy 2 – R e- em pl o ym en t o f D ism iss ed C EO s M ea n S. D . 1 2 3 4 5 6 7 8 9 10 11 1. Re hi re d as Pu bl ic Ex ec . 03 . 16 2. Re hi re d as Pu bl ic CE O . 02 . 14 . 73 * * * 3. Re hi re d as Ex ec u tiv e . 08 . 27 . 56 * * * . 37 * * * 4. SE C In v es tig at io n . 23 . 42 - . 09 - . 08 - . 00 5. Cu rr en t A ge 58 . 02 6. 71 - . 09 - . 04 - . 23 * * * - . 02 6. CE O ’ s Pr io r Co m pe n sa tio n 19 19 45 8. 15 29 21 69 4. 55 - . 02 - . 03 - . 03 . 10 * . 19 * * * 7. Pr io r O w n er sh ip . 01 . 03 - . 06 - . 05 . 04 . 10 * - . 08 - . 14 * * 8. Pr io r Fi rm ’ s Si ze 7. 33 1. 75 . 10 * . 04 . 06 . 25 * * * . 12 * . 52 * * * - . 37 * * * 9. Te n u re as CE O 6. 86 6. 33 - . 10 - . 09 - . 06 . 16 * * . 16 * * - . 01 . 47 * * * - . 09 10 . Pr io r Fi rm ’ s Pe rfo rm an ce . 01 . 10 - . 02 . 01 . 04 . 18 * * * - . 01 . 23 * * * - . 26 * * * . 36 * * * - . 12 * 11 . In v er se M ill s R at io . 75 . 07 . 03 . 04 . 01 - . 11 * - . 09 . 07 . 07 - . 06 . 05 . 04 12 . Pe rfo rm an ce D ism iss al . 43 . 50 . 07 . 09 . 09 - . 18 * * * - . 11 * . 12 * - . 04 - . 01 - . 16 * * - . 11 * - . 03 13 . Fi du ci ar y V io la tio n . 20 . 40 - . 09 - . 07 - . 04 . 75 * * * . 05 . 00 . 15 * * . 14 * * . 27 * * * . 12 * - . 15 * * 14 . Pe rs o n al Co n du ct V io la tio n . 04 . 19 - . 03 - . 03 - . 00 - . 11 * . 09 - . 03 - . 07 . 11 * - . 17 * * . 13 * - . 01 15 . Pr es tig e Ex pe rie n ce . 48 . 50 . 15 * * . 11 * . 04 - . 01 . 16 * * . 26 * * - . 30 * * * . 43 * * * - . 27 * * * . 02 - . 15 * * 16 . Ed u ca tio n Le v el 3. 57 . 65 - . 01 - . 02 . 02 . 01 . 01 . 15 * * - . 06 . 20 * * * . 07 . 17 * * . 34 * * * 17 . El ite Ed u ca tio n . 33 . 47 - . 02 - . 03 . 00 . 22 * * * - . 12 * . 01 . 16 * * . 25 * * * . 11 * . 00 - . 00 18 . B es t C EO . 03 . 16 . 07 . 09 . 11 * . 27 * * * . 09 . 29 * * * - . 07 . 20 * * * - . 02 . 16 * * . 05 19 . N eg at iv e Pu bl ic ity 1. 89 6. 02 - . 02 - . 02 - . 03 . 26 * * * - . 04 . 13 * * - . 06 . 36 * * * . 01 . 07 - . 04 20 . To ta l D ire ct o rs hi ps 1. 25 1. 59 . 07 . 02 - . 00 - . 05 . 03 . 05 - . 14 * * . 24 * * * - . 15 * * . 04 . 01 21 . Pr io r Ex ec u tiv e Em pl o ye rs 1. 09 1. 25 . 04 . 06 . 03 - . 15 * * . 22 * * * - . 07 - . 10 * - . 05 - . 33 * * * - . 01 - . 03 22 . Re sid en ce in M ajo r Ci ty . 62 . 49 . 00 . 04 . 05 . 07 - . 04 . 16 * * . 02 . 18 * * * . 06 . 10 * . 09 179 12 13 14 15 16 17 18 19 20 21 13 . Fi du ci ar y V io la tio n - . 42 * * * 14 . Pe rs o n al Co n du ct V io la tio n - . 17 * * . 10 15 . Pr es tig e Ex pe rie n ce . 05 - . 04 . 07 16 . Ed u ca tio n Le v el . 01 - . 01 - . 07 - . 10 * 17 . El ite Ed u ca tio n . 02 . 17 * * * - . 14 * * - . 04 . 43 * * * 18 . B es t C EO . 01 . 11 * - . 03 . 17 * * * . 09 . 07 19 . N eg at iv e Pu bl ic ity - . 01 . 09 . 27 * * * . 17 * * * . 01 . 13 * * - . 02 20 . To ta l D ire ct o rs hi ps . 16 * * - . 10 . 01 . 11 * . 24 * * * . 23 * * * - . 06 . 06 21 . Pr io r Ex ec u tiv e Em pl o ye rs . 14 * * - . 26 * * * . 04 . 22 * * * . 06 - . 03 . 08 . 15 * * . 10 * 22 . Re sid en ce in M ajo r Ci ty . 05 . 13 * * . 15 * * - . 03 . 09 . 11 * . 10 * . 19 * * * . 06 . 08 180 TABLE 6 CAREER OUTCOMES FOR THE SAMPLE OF DISMISSED CEOS Rehiring Position Reason for Dismissal Public Executive Public CEO Private Executive Private CEO Self- Employed Other Positiona No Position Performance Dismissal 1 (1.14%) 5 (5.68%) 1 (1.14%) 8 (9.09%) 5 (5.68%) 10 (11.36%) 9 (10.23%) Fiduciary Violation 0 (0%) 0 (0%) 1 (1.14%) 3 (3.41%) 3 (3.41%) 1 (1.14%) 7 (7.95%) Personal Conduct Violation 0 (0%) 0 (0%) 0 (0%) 1 (1.14%) 1 (1.14%) 0 (0%) 1 (1.14%) Other 1 (1.14%) 3 (3.41%) 1 (1.14%) 2 (2.27%) 5 (5.68%) 8 (9.09%) 11 (12.5%) Total 2 (2.27%) 8 (9.09%) 3 (3.41%) 14 (15.91%) 14 (15.91%) 19 (21.59%) 28 (31.82%) a. Other positions accepted include: venture capital/private equity partner, Professor, lawyer, and consultant TABLE 7 AVERAGE TIME TO NEW POSITION FOR DISMISSED CEOS New Position Average Years to Obtain Public Executive 1.00 Public CEO 1.50 Private Executive 3.00 Private CEO 2.07 Other Position 1.47 Average Years to Obtain Any Position 1.54 181 TA BL E 8 Su rv iv a l A n a ly sis fo r R eg a in in g Em pl o ym en t a s Ex ec u tiv e a t a Pu bl ic C o rp o ra tio n M o de l 1 M o de l 2 M o de l 3 M o de l 4 Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r SE C In v es tig at io n - 40 . 37 * * * (0. 63 ) - 34 . 23 * * * (0. 66 ) - 41 . 60 * * * (4. 34 ) - 19 . 64 * * * (4. 63 ) Cu rr en t A ge - 0. 07 † (0. 04 ) - 0. 07 † (0. 04 ) - 0. 10 † (0. 05 ) - 0. 09 † (0. 05 ) CE O ’ s Pr io r Co m pe n sa tio n - 0. 00 (0. 00 ) - 0. 00 (0. 00 ) - 0. 00 (0. 00 ) - 0. 00 (0. 00 ) Pr io r O w n er sh ip 8. 62 (19 . 39 ) 7. 13 (20 . 44 ) 15 . 22 (23 . 71 ) 7. 48 (31 . 05 ) Pr io r Fi rm ’ s Si z e 0. 64 * * * (0. 17 ) 0. 80 * * (0. 31 ) 0. 72 * (0. 36 ) 0. 75 * (0. 36 ) Te n u re as CE O - 0. 19 (0. 09 ) - 0. 22 * (0. 09 ) - 0. 11 (0. 11 ) - 0. 10 (0. 11 ) Pr io r Fi rm ’ s Pe rfo rm an ce - 1. 01 (3. 27 ) - 2. 05 (3. 43 ) - 2. 40 (3. 19 ) - 2. 91 (3. 81 ) In v er se M ill s R at io 14 . 65 † (8. 80 ) 3. 23 (10 . 94 ) 3. 82 (11 . 06 ) H 34 : Pe rfo rm an ce R el at ed D ism iss al 0. 25 (0. 65 ) 16 . 80 * * * (1. 65 )** * H 35 : Fi du ci ar y V io la tio n - 36 . 60 * * * (1. 68 ) - 15 . 98 * * * (1. 63 ) H 36 : Pe rs o n al Co n du ct V io la tio n - 35 . 34 * * * (1. 36 ) - 13 . 28 * * * (3. 23 ) H 38 : Pr es tig e Ex pe rie n ce 2. 35 * (1. 07 ) 18 . 69 * * * (1. 38 ) H 39 : Ed u ca tio n Le v el 0. 48 (0. 44 ) 0. 47 (0. 41 ) H 40 : El ite Ed u ca tio n - 0. 95 (0. 77 ) - 0. 81 (0. 89 ) H 41 : B es t C EO 4. 10 * * (1. 30 ) 3. 65 * * (1. 38 ) H 41 : N eg at iv e Pu bl ic ity - 0. 45 * (0. 20 ) - 0. 38 * (0. 17 ) H 42 : To ta l D ire ct o rs hi ps 0. 18 (0. 26 ) 0. 12 (0. 28 ) H 43 : Pr io r Ex ec u tiv e Em pl o ye rs 0. 27 (0. 56 ) 0. 17 (0. 60 ) H 44 : R es id en ce in M ajo r Ci ty 1. 31 * * (0. 51 ) 1. 35 * (0. 59 ) H 45 : Pe rf D ism iss al x Pr es tig e Ex p - 16 . 77 * * * (1. 96 ) - Lo g Ps eu do lik el ih o o d 37 . 11 71 06 . 96 * * * 36 . 43 39 46 . 33 * * * 31 . 45 48 28 . 44 * * * 31 . 16 14 79 . 16 * * * W al d- χ2 D ep en de n t v a ria bl e is w he th er th e di sm iss ed CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 39 7 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 182 TA BL E 9a R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g E m pl o ym en t a s Ex ec u tiv e a t a Pu bl ic C o rp o ra tio n w he n D ism iss a l w a s D u e to Pe rf o rm a n ce Pr o bl em s M o de l 5 M o de l 6 M o de l 7 M o de l 8 M o de l 9 M o de l 1 0 a M o de l 1 1 Co ef f. St an da r d Er ro r Co ef f. St an da r d Er ro r Co ef f. St an da r d Er ro r Co ef f. St an da r d Er ro r Co ef f. St an da r d Er ro r Co ef f. St an da r d Er ro r Co ef f. St an da r d Er ro r H 46 : Pe rf D ism iss al x Ed u ca tio n Le v el - 2. 79 (1. 78 ) H 47 : Pe rf D ism iss al x El ite Ed u ca tio n - 45 . 51 * * * (2. 13 ) H 48 : Pe rf D ism iss al x B es t CE O - 11 . 32 * (4 . 91 ) H 48 : Pe rf D ism iss al x N eg at iv e Pu bl ic ity - 1. 36 * (0. 68 ) H 49 : Pe rf D ism iss al x To ta l D ire ct o rs hi ps - 0. 93 † (0. 55 ) H 50 : Pe rf D ism iss al x Pr io r Ex ec u tiv e Em pl o ye rs - 3. 19 * (1. 52 ) H 51 : Pe rf D ism iss al x Re sid en ce in M ajo r Ci ty - 4. 79 * * (1 . 56 ) - Lo g Ps eu do lik el ih o o d 30 . 78 42 85 . 92 * * * 29 . 85 15 61 3. 33 * * * 31 . 45 74 89 . 45 * * * 30 . 67 43 90 . 02 * * * 30 . 97 86 45 . 50 * * * 30 . 54 64 11 . 66 * * * 30 . 09 40 48 . 10 * * * W al d- χ2 D ep en de n t v a ri a bl e is w he th er th e di sm is se d CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 39 7 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . a. El ite ed u ca tio n is po sit iv e an d sig n ifi ca n t a t t he 0. 05 le v el , w hi le el ite ed u ca tio n is sig n ifi ca n t a n d n eg at iv e at th e 0. 01 le v el . Pr io r Ex ec u tiv e Em pl o ye rs is al so sig n ifi ca n t a t t he 0. 05 le v el . 183 TA BL E 9b R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g Em pl o ym en t a s Ex ec u tiv e a t a Pu bl ic C o rp o ra tio n w he n D ism iss a l w a s D u e to V io la tio n o f F id u ci a ry D u ty M o de l 1 2 M o de l 1 3 M o de l 1 4 M o de l 1 5 M o de l 1 6 M o de l 1 7 M o de l 1 8 M o de l 1 9 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r H 52 : Fi du ci ar y V io la tio n x Pr es tig io u s Ex pe rie n ce - 6. 02 (4. 60 ) H 53 : Fi du ci ar y V io la tio n x Ed u ca tio n Le v el - 37 . 59 * * * (4. 89 ) H 54 : Fi du ci ar y V io la tio n x El ite Ed u ca tio n - 5. 88 (4. 60 ) H 55 : Fi du ci ar y V io la tio n x B es t CE O 29 . 57 * * * (4. 78 ) H 55 : Fi du ci ar y V io la tio n x N eg at iv e Pu bl ic ity 3. 36 * * * (0. 64 ) H 56 : Fi du ci ar y V io la tio n x To ta l D ire ct o rs hi ps - 2. 21 (1. 53 ) H 57 : Fi du ci ar y V io la tio n x Pr io r Ex ec u tiv e Em pl o ye rs 35 . 53 * * * (5. 57 ) H 58 : Fi du ci ar y V io la tio n x R es id en ce in M ajo r Ci ty - 40 . 03 * * * (5. 46 ) - Lo g Ps eu do lik el ih o o d 32 . 48 61 19 . 57 * * * 34 . 57 75 05 . 86 * * * 31 . 45 63 86 . 85 37 . 78 51 44 . 72 * * * 36 . 63 50 42 . 40 * * * 31 . 45 60 91 . 60 * * * 38 . 33 64 87 . 64 * * * 38 . 03 72 94 . 43 * * * W al d- χ2 D ep en de n t v a ri a bl e is w he th er th e di sm is se d CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 39 7 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 184 TA BL E 9c R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g Em pl o ym en t a s Ex ec u tiv e a t a Pu bl ic C o rp o ra tio n w he n D ism iss a l w a s D u e to a Pe rs o n a l C o n du ct V io la tio n a M o de l 2 0 M o de l 2 1 M o de l 2 4 M o de l 2 5 M o de l 2 6 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r H 59 : Pe rs Co n du ct V io la tio n x Pr es tig e Ex pe rie n ce 4. 68 (4. 11 ) H 60 : Pe rs Co n du ct V io la tio n x Ed u ca tio n Le v el - 4. 68 (4. 11 ) H 63 : Pe rs Co n du ct V io la tio n x N eg at iv e Pu bl ic ity 0. 26 (0. 21 ) H 64 : Pe rs Co n du ct V io la tio n x To ta l D ire ct o rs hi ps 2. 05 (1. 78 ) H 65 : Pe rs Co n du ct V io la tio n x Pr io r Ex ec u tiv e Em pl o ye rs 2. 59 (2. 58 ) - Lo g Ps eu do lik el ih o o d 33 . 43 65 01 . 36 * * * 31 . 45 61 85 . 58 * * * 32 . 67 60 89 . 52 * * * 36 . 64 82 09 . 26 * * * 35 . 57 58 13 . 70 * * * W al d- χ2 D ep en de n t v a ria bl e is w he th er th e di sm iss ed CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 39 7 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . a . D u e to la ck o f o bs er va tio n s a n d va ria n ce w ith re sp ec t t o o bs er ve d va ria bl es , es tim a te s for in te ra ct io n s in vo lv in g pe rs o n a l c o n du ct vi o la tio n s a n d el ite ed u ca tio n , Be st CE O , a n d Re sid en ce in M a jor Ci ty co u ld n o t b e o bt a in ed . 185 TA BL E 10 Su rv iv a l A n a ly sis fo r R eg a in in g Em pl o ym en t a s C EO a t a Pu bl ic C o rp o ra tio n M o de l 1 M o de l 2 M o de l 3 M o de l 4 Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r SE C In v es tig at io n - 38 . 89 * * * (0. 74 ) - 40 . 58 * * (0. 67 ) - 24 . 45 * * * (3. 34 ) - 6. 93 † (4. 13 ) Cu rr en t A ge - 0. 03 (0. 06 ) - 0. 05 (0. 06 ) 0. 08 (0. 09 ) 0. 06 (0. 14 ) CE O ’ s Pr io r Co m pe n sa tio n - 0. 00 (0. 00 ) - 0. 00 (0. 00 ) - 0. 00 (0. 00 ) - 0. 00 (0. 00 ) Pr io r O w n er sh ip - 22 . 29 (54 . 12 ) - 33 . 35 (53 . 91 ) - 48 . 20 (10 7. 79 ) - 16 5. 22 (25 9. 42 ) Pr io r Fi rm ’ s Si ze 0. 53 † (0. 31 ) 1. 15 † (0. 59 ) 2. 62 † (1. 34 ) 3. 44 * * (1. 30 ) Te n u re as CE O - 0. 22 † (0. 14 ) - 0. 29 * (0. 12 ) - 0. 26 † (0. 14 ) - 0. 29 † (0. 16 ) Pr io r Fi rm ’ s Pe rfo rm an ce 1. 47 (5. 69 ) 0. 24 (5. 09 ) - 12 . 76 (8. 33 ) - 17 . 61 * (7. 68 ) In v er se M ill s R at io 56 . 49 * (23 . 77 ) 12 5. 57 † (71 . 15 ) 17 3. 74 * (72 . 23 ) Pe rfo rm an ce D ism iss al 1. 54 (2. 67 ) 18 . 75 * * * (1. 94 ) Fi du ci ar y V io la tio n - 26 . 08 * * * (5. 13 ) - 17 . 82 * (7. 86 ) Pe rs o n al Co n du ct V io la tio n - 29 . 95 * * * (2. 45 ) 3. 80 (2. 45 ) Pr es tig io u s Ex pe rie n ce 2. 30 * (1. 12 ) 21 . 54 * * * (4. 45 ) Ed u ca tio n Le v el 1. 32 (1. 12 ) 1. 50 (1. 10 ) El ite Ed u ca tio n - 0. 91 (1. 51 ) - 1. 29 (1. 35 ) B es t C EO 12 . 97 (7. 94 ) 13 . 77 † (8. 16 ) N eg at iv e Pu bl ic ity - 1. 93 † (1. 06 ) - 2. 14 † (1. 13 ) To ta l D ire ct o rs hi ps - 0. 57 (0. 53 ) - 0. 50 (0. 90 ) Pr io r Ex ec u tiv e Em pl o ye rs 2. 59 * (1. 32 ) 2. 73 * (1. 38 ) R es id en ce in M ajo r Ci ty 3. 53 * * (1. 15 ) 5. 63 † (3. 30 ) Pe rf D ism iss al x Pr es t E x p - 19 . 08 * * * (3. 98 ) - Lo g Ps eu do lik el ih o o d 26 . 06 62 58 . 20 * * * 23 . 71 58 81 . 81 * * * 17 . 87 43 40 . 65 * * * 17 . 50 10 51 . 30 * * * W al d- χ2 a . D ep en de n t v a ria bl e is w he th er th e di sm iss ed CE O re ga in ed em pl o ym en t a s a CE O o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 40 8 a cr o ss 88 su bje ct s b. † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 186 T A BL E 11 a R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g Em pl o ym en t a s C EO a t a Pu bl ic C o rp o ra tio n w he n D ism iss a l w a s D u e to Pe rf o rm a n ce Pr o bl em s M o de l 5 M o de l 6 a M o de l 7 M o de l 8 b M o de l 9 M o de l 1 0 a M o de l 1 1 Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Co ef f. St d Er ro r Pe rf D ism iss al x Ed u ca tio n Le v el - 2. 64 (1. 86 ) Pe rf D ism iss al x El ite Ed u ca tio n - 10 6. 99 * * * (13 . 53 ) Pe rf D ism iss al x B es t C EO - 6. 34 (8. 15 ) Pe rf D ism iss al x N eg at iv e Pu bl ic ity - 11 . 73 * * (3. 64 ) Pe rf D ism iss al x To ta l D ire ct o rs hi ps - 31 . 11 * * (11 . 64 ) Pe rf D ism iss al x Pr io r Ex ec u tiv e Em pl o ye rs - 36 . 52 * * * (8. 38 ) Pe rf D ism iss al x R es id en ce in M ajo r Ci ty - 27 . 43 * * * (3. 28 ) - Lo g Ps eu do lik el ih o o d 17 . 59 52 57 . 69 * * * 8. 19 18 97 . 40 * * * 17 . 87 62 04 . 10 * * * 12 . 38 22 9. 15 * * * 11 . 07 16 8. 54 * * * 9. 53 29 5. 38 * * * 15 . 96 78 0. 88 * * * W al d- χ2 D ep en de n t v a ri a bl e is w he th er th e di sm is se d CE O re ga in ed em pl o ym en t a s a CE O o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 40 8 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . a. Pe rfo rm an ce di sm iss al , El ite Ed u ca tio n an d To ta l D ire ct o rs hi ps ar e al l s ig n ifi ca n t a t t he 0. 00 1 le v el w he n th e in te ra ct io n te rm is ad de d. b. Pe rfo rm an ce di sm iss al an d To ta l D ire ct o rs hi ps ar e sig n ifi ca n t a t t he 0. 00 1 le v el w he n th e in te ra ct io n te rm is ad de d. 187 TA BL E 11 b R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g Em pl o ym en t a s C EO a t a Pu bl ic C o rp o ra tio n w he n D ism iss a l w a s D u e to V io la tio n o f F id u ci a ry D u ty M o de l 1 2 M o de l 1 3 M o de l 1 4 M o de l 1 5 M o de l 1 6 M o de l 1 7 M o de l 1 8 M o de l 1 9 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Fi du ci ar y V io la tio n x Pr es tig io u s Ex pe rie n ce - 4. 61 (3. 86 ) Fi du ci ar y V io la tio n x Ed u ca tio n Le v el - 19 . 34 * * (6. 22 ) Fi du ci ar y V io la tio n x El ite Ed u ca tio n - 3. 96 (3. 86 ) Fi du ci ar y V io la tio n x B es t CE O 12 . 77 (10 . 42 ) Fi du ci ar y V io la tio n x N eg at iv e Pu bl ic ity 3. 72 * * * (0. 92 ) Fi du ci ar y V io la tio n x To ta l D ire ct o rs hi ps - 2. 22 † (1. 29 ) Fi du ci ar y V io la tio n x Pr io r Ex ec u tiv e Em pl o ye rs 22 . 81 * * * (5. 97 ) Fi du ci ar y V io la tio n x R es id en ce in M ajo r Ci ty - 30 . 87 * * * (4. 19 ) - Lo g Ps eu do lik el ih o o d 17 . 50 68 99 . 66 * * * 18 . 82 56 05 . 65 * * * 16 . 88 80 15 . 12 * * * 17 . 89 77 68 . 17 * * * 17 . 87 51 21 . 19 * * * 17 . 55 75 69 . 57 17 . 89 39 3. 04 * * * 17 . 87 77 80 . 88 * * * W al d- χ2 D ep en de n t v a ri a bl e is w he th er th e di sm is se d CE O re ga in ed em pl o ym en t a s a CE O o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 40 8 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 188 TA B LE 11 c R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g E m pl o ym en t a s C EO a t a Pu bl ic C o rp o ra tio n w he n D ism iss a l w a s D u e to a Pe rs o n a l C o n du ct V io la tio n a M o de l 2 0 M o de l 2 1 M o de l 2 4 M o de l 2 5 M o de l 2 6 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Pe rs Co n du ct V io la tio n x Pr es tig io u s Ex pe rie n ce 37 . 83 † (21 . 92 ) Pe rs Co n du ct V io la tio n x Ed u ca tio n Le v el 37 . 83 † (21 . 92 ) Pe rs Co n du ct V io la tio n x N eg at iv e Pu bl ic ity 1. 55 † (0. 94 ) Pe rs Co n du ct V io la tio n x To ta l D ire ct o rs hi ps 12 . 97 † (7. 83 ) Pe rs Co n du ct V io la tio n x Pr io r Ex ec u tiv e Em pl o ye rs 18 . 44 (11 . 75 ) - Lo g Ps eu do lik el ih o o d 17 . 90 75 50 . 15 * * * 17 . 71 61 01 . 95 * * * 17 . 05 70 44 . 62 * * * 17 . 55 81 29 . 59 * * * 17 . 44 80 90 . 43 W al d- χ2 D ep en de n t v a ria bl e is w he th er th e di sm iss ed CE O re ga in ed em pl o ym en t a s a CE O o f a pu bl ic ly tr a de d co rp o ra tio n o r n o t; n = 40 8 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . a . D u e to la ck o f o bs er va tio n s a n d va ria n ce w ith re sp ec t t o o bs er ve d va ri a bl es , es tim a te s for in te ra ct io n s in vo lv in g pe rs o n a l c o n du ct vi o la tio n s a n d el ite ed u ca tio n , Be st CE O , a n d Re sid en ce in M a jor Ci ty co u ld n o t b e o bt a in ed . 189 TA BL E 12 Su rv iv a l A n a ly sis fo r R eg a in in g Em pl o ym en t a s Ex ec u tiv e a t a Pu bl ic o r Pr iv a te C o rp o ra tio n M o de l 1 M o de l 2 M o de l 3 M o de l 4 Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r Co ef f. R o bu st St d. Er ro r SE C In v es tig at io n - 0. 32 (0. 58 ) - 0. 31 (0. 56 ) 0. 43 (0. 48 ) 0. 40 (0. 50 ) Cu rr en t A ge - 0. 08 * * (0. 03 ) - 0. 08 * * (0. 03 ) - 0. 10 * * (0. 03 ) - 0. 10 * * (0. 03 ) CE O ’ s Pr io r Co m pe n sa tio n - 0. 00 (0. 00 ) - 0. 00 (0. 00 ) - 0. 00 † (0. 00 ) - 0. 00 † (0. 00 ) Pr io r O w n er sh ip 12 . 84 † (7. 27 ) 13 . 01 † 7. 41 19 . 19 * (8. 12 ) 18 . 89 * (8. 60 ) Pr io r Fi rm ’ s Si z e 0. 29 * (0. 14 ) 0. 30 * (0. 14 ) 0. 46 * (0. 20 ) 0. 45 * (0. 20 ) Te n u re as CE O - 0. 05 (0. 04 ) - 0. 05 (0. 04 ) - 0. 04 (0. 04 ) - 0. 04 (0. 04 ) Pr io r Fi rm ’ s Pe rfo rm an ce 0. 93 (1. 72 ) 0. 95 (1. 76 ) 0. 86 (1. 79 ) 0. 90 (1. 76 ) In v er se M ill s R at io 0. 81 (2. 73 ) - 1. 06 (2. 85 ) - 1. 01 (2. 92 ) Pe rfo rm an ce D ism iss al 0. 52 (0. 46 ) 0. 40 (0. 51 ) Fi du ci ar y V io la tio n - 0. 77 (1. 04 ) - 0. 76 (1. 06 ) Pe rs o n al Co n du ct V io la tio n 0. 30 (0. 74 ) 0. 32 (0. 77 ) Pr es tig io u s Ex pe rie n ce 0. 55 (0. 52 ) 0. 43 (0. 76 ) Ed u ca tio n Le v el 0. 36 (0. 41 ) 0. 36 (0. 42 ) El ite Ed u ca tio n - 0. 54 (0. 45 ) - 0. 54 (0. 45 ) B es t C EO 1. 98 * (0. 87 ) 2. 04 * (0. 89 ) N eg at iv e Pu bl ic ity - 0. 09 (0. 06 ) - 0. 09 (0. 07 ) To ta l D ire ct o rs hi ps 0. 02 (0. 14 ) 0. 02 (0. 14 ) Pr io r Ex ec u tiv e Em pl o ye rs - 0. 09 (0. 06 ) - 0. 03 (0. 33 ) R es id en ce in M ajo r Ci ty 0. 48 (0. 46 ) 0. 49 (0. 44 ) Pe rf D ism iss al x Pr es t E x p 0. 22 (0. 72 ) - Lo g Ps eu do lik el ih o o d 11 1. 44 24 . 97 * * * 11 1. 42 24 . 61 * * 10 5. 97 36 . 35 * * 10 5. 94 37 . 79 * * W al d- χ2 D ep en de n t v a ria bl e is w he th er th e di sm iss ed CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic o r pr iv a te co rp o ra tio n o r n o t; n = 35 1 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 190 TA BL E 13 a R es u lts o f I n te ra ct io n T es ts fo r R eg a in in g Em pl o ym en t a s Ex ec u tiv e a t a Pu bl ic o r Pr iv a te C o rp o ra tio n w he n D ism iss a l w a s D u e to Pe rf o rm a n ce Pr o bl em s M o de l 5 M o de l 6 a M o de l 7 M o de l 8 b M o de l 9 M o de l 1 0 a M o de l 1 1 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Pe rf D ism iss al x Ed u ca tio n Le v el - 2. 18 * (1. 06 ) Pe rf D ism iss al x El ite Ed u ca tio n - 2. 16 * (0. 94 ) Pe rf D ism iss al x B es t C EO 3. 16 (3. 63 ) Pe rf D ism iss al x N eg at iv e Pu bl ic ity - 0. 20 (0 . 14 ) Pe rf D ism iss al x To ta l D ire ct o rs hi ps - 0. 11 (0. 22 ) Pe rf D ism iss al x Pr io r Ex ec u tiv e Em pl o ye rs 0. 98 (1. 25 ) Pe rf D ism iss al x Re sid en ce in M ajo r Ci ty 0. 91 (1. 06 ) - Lo g Ps eu do lik el ih o o d 10 3. 44 31 . 21 * 10 4. 00 60 . 41 * * * 10 5. 62 76 . 71 * * * 10 5. 27 40 . 18 * * 10 5. 90 36 . 09 * 10 5. 59 54 . 50 * * * 10 5. 59 34 . 90 * W al d- χ2 D ep en de n t v a ri a bl e is w he th er th e di sm is se d CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic o r pr iv a te co rp o ra tio n o r n o t; n = 35 1 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 191 TA BL E 13 b R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g Em pl o ym en t a s C EO a t a Pu bl ic o r Pr iv a te C o rp o ra tio n w he n D ism iss a l w a s D u e to V io la tio n o f F id u ci a ry D u ty M o de l 1 2 M o de l 1 3 M o de l 1 4 M o de l 1 5 M o de l 1 6 M o de l 1 7 M o de l 1 8 M o de l 1 9 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Fi du ci ar y V io la tio n x Pr es tig io u s Ex pe rie n ce - 36 . 87 * * * (1. 19 ) Fi du ci ar y V io la tio n x Ed u ca tio n Le v el 0. 24 (1. 74 ) Fi du ci ar y V io la tio n x El ite Ed u ca tio n 0. 64 (1. 53 ) Fi du ci ar y V io la tio n x B es t CE O - 40 . 82 * * * (1. 42 ) Fi du ci ar y V io la tio n x N eg at iv e Pu bl ic ity 0. 34 * * (0. 12 ) Fi du ci ar y V io la tio n x To ta l D ire ct o rs hi ps - 0. 72 (0. 67 ) Fi du ci ar y V io la tio n x Pr io r Ex ec u tiv e Em pl o ye rs - 23 . 46 (69 . 71 ) Fi du ci ar y V io la tio n x R es id en ce in M ajo r Ci ty - 1. 72 (1. 74 ) - Lo g Ps eu do lik el ih o o d 10 2. 42 31 95 . 88 * * * 10 5. 96 47 . 15 * * * 10 5. 87 40 . 65 * * 10 4. 82 16 20 . 66 * * * 10 3. 83 37 . 15 * 10 5. 47 70 . 38 * * * 10 4. 60 62 . 83 * * * 10 5. 39 37 . 72 * * W al d- χ2 D ep en de n t v a ri a bl e is w he th er th e di sm is se d CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic o r pr iv a te co rp o ra tio n o r n o t; n = 35 1 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . 192 TA BL E 13 c R es u lts o f I n te ra ct io n Te st s fo r R eg a in in g Em pl o ym en t a s Ex ec u tiv e a t a Pu bl ic o r Pr iv a te C o rp o ra tio n w he n D ism iss a l w a s D u e to a Pe rs o n a l C o n du ct V io la tio n a M o de l 2 0 M o de l 2 1 M o de l 2 4 M o de l 2 5 M o de l 2 6 Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Co ef f. St an da rd Er ro r Pe rs Co n du ct V io la tio n x Pr es tig io u s Ex pe rie n ce 18 . 56 * * * (2. 77 ) Pe rs Co n du ct V io la tio n x Ed u ca tio n Le v el - 18 . 56 * * * (1. 71 ) Pe rs Co n du ct V io la tio n x N eg at iv e Pu bl ic ity - 0. 18 (0. 11 ) Pe rs Co n du ct V io la tio n x To ta l D ire ct o rs hi ps 0. 13 (0. 62 ) Pe rs Co n du ct V io la tio n x Pr io r Ex ec u tiv e Em pl o ye rs - 9. 58 * * * (0. 73 ) - Lo g Ps eu do lik el ih o o d 10 5. 69 86 . 80 * * * 10 5. 69 19 6. 47 * * * 10 5. 62 57 . 95 * * * 10 5. 97 37 . 37 * 10 5. 88 12 31 . 77 * * * W al d- χ2 D ep en de n t v a ria bl e is w he th er th e di sm iss ed CE O re ga in ed em pl o ym en t a s a n ex ec u tiv e o f a pu bl ic o r pr iv a te co rp o ra tio n o r n o t; n = 35 1 a cr o ss 88 su bje ct s † p < 0. 1, * p < 0. 05 , * * p < 0. 01 , * * * p < 0. 00 1, Tw o - ta ile d co eff ic ie n t t es ts . a . D u e to la ck o f o bs er va tio n s a n d va ria n ce w ith re sp ec t t o o bs er ve d va ria bl es , es tim a te s for in te ra ct io n s in vo lv in g pe rs o n a l c o n du ct vi o la tio n s a n d el ite ed u ca tio n , Be st CE O , a n d Re sid en ce in M a jor Ci ty co u ld n o t b e o bt a in ed . 193 APPENDIX A Sample Selection Model: Antecedents of Dismissal for Study 2 Sample Coef. Std. Err. Intercept -2.70* (1.27) CEO Tenure -0.00 (0.02) CEO Age 0.00 (0.02) CEO Compensation -0.00 0.00 CEO Ownership -5.62 (4.78) CEO Duality -0.44 (0.27) Independent director proportion 0.73 (0.82) Institutional Holdings -0.07 (0.18) Block Ownership 0.06 (0.81) Firm ROA -5.20*** (0.94) Firm Tobin’s Q -0.26 (0.17) -2log Likelihood -292.59 Chi-square 54.31*** (d.f. = 10) Pseudo R2 0.08 Note: a. Dependent variable is whether a firm fired their CEO or not in a given year. b. * p < 0.05, ** p < 0.01, *** p <0.001, two-tailed coefficient tests (N= 1,916).