Executive dismissal or retention? A study of performance,power,and survival for college basketball coaches

Abstract

This study applies upper echelons theory associated with executive dismissal and power to examine the relationships of performance and four types of executive power—structural, prestige, expert, and governance concentration—with dismissal. Using the context of National Collegiate Athletic Association college basketball, in which coaches are completely responsible for strategies and human capital acquisition and retention, a curvilinear relationship between performance and dismissal is found. Significant relationships for prestige and expert power with dismissal are also found, but the “honeymoon period” is longer than prior studies of executive and coach dismissal have shown. Surprisingly, concentrated governance is found to be negatively associated with the likelihood of dismissal.

Keywords

concentrated governance cox proportional hazard model executive dismissal power reputation

Introduction

Poor organizational performance generates strong negative opinions and emotions for many stakeholders, leading them to hold accountable the senior executives tasked with leading the organization’s strategy and operations (Lusthaus et al., 1999). Stakeholders feel a sense of personal loss and demand new leadership to stimulate change that improves outcomes. Financial and market returns are firms’ strongest performance indicators, with examples in 2019 at General Electric, Intel, and WeWork illustrating how failure to meet expected performance levels can foster stakeholder demands for chief executive officers (CEOs) to resign or be fired. Stakeholders’ personal loss is markedly different in professional and college sport since very few are owners or shareholders and most are not employed by the organization. Still, the subsequent demand for new leadership (i.e. coaches) when losses occur can be equally strong. Recent coach dismissals in professional (e.g. Ron Rivera—Carolina Panthers) and college sport (e.g. Steve Alford—University of California-Los Angeles) seem consistent with theory and prior scholarly research regarding performance and dismissal (see Holmes, 2011; Wangrow et al., 2018). Yet, in each of these instances of coach dismissal, the antecedents and context associated with dismissal may go beyond performance. For example, Rivera led his team to the Super Bowl in 2015 and won the Associated Press National Football League (NFL) Coach of the Year awards in 2013 and 2015, but the prestige power (see Finkelstein, 1992) associated with these accomplishments diminished and he was dismissed during the 2019 season. Indeed, prior research has shown that performance is just one of many factors associated with senior executive dismissals (Allen and Chadwick, 2012; Finkelstein et al., 2009).

While they assert performance is the predominant factor influencing dismissal, Fredrickson et al. (1988) theorize that the power executives hold can supersede other factors (e.g. expectations, allegiances) that impact decision-makers’ latitude to dismiss senior executives. Senior executives derive power from a wide range of sources, including control of critical resources, ownership, structural authority, reputation, and personal traits (Finkelstein, 1992). If they are unable to derive power from at least one of these sources, they are more vulnerable to dismissal. Other management scholars have found support for various sources of power influencing dismissal, including social and celebrity status relative to board members (Flickingeret al., 2016), loyalty (i.e. interdependence) of board members¹ (Boeker, 1992), and expertise derived from tenure (Kim, 1996). Research examining coach dismissals in professional sport supports power’s effect on dismissal, with scholars finding significant relationships for structural power from simultaneously holding the coach and general manager positions (Wangrow et al., 2018) and expert power derived from tenure (Allen and Chadwick, 2012; Foreman and Soebbing, 2015) with dismissal.

Despite prior examinations of power in management and sport literature (e.g. Amedu and Dulewicz, 2018; Wangrow et al., 2018), its effect on dismissal in intercollegiate sport has been minimally studied. This is surprising since intercollegiate sport provides a unique landscape in which several power-oriented factors are constant across all coaches. Unlike firms and professional sport teams, all college coaches hold the dual role of coach and general manager. Besides being accountable for game strategy and execution, a college coach is fully accountable for the team’s human capital via recruiting and retention of players and assistant coaches. Thus, like a CEO who serves as board chairman, a college coach has considerable power. Moreover, unlike firms in which CEOs have varying degrees of personal ownership (Daily and Johnson, 1997), college coaches have no financial stake in the team and therefore are unable to exert power from ownership. Consequently, intercollegiate sport provides a rich landscape to examine the four forms of power posited by Finkelstein (1992) to be relevant to top management: expert power associated with tenure, structural power in relation to a decision-making stakeholder, prestige power from reputational capital, and power associated with concentrated versus dispersed governance. The value of studying these forms of power within this context is further enhanced by the availability of data to support multiple objective performance measures, thus increasing the robustness of any findings by accounting for stakeholders’ varying definitions of success.

The relationship of performance with dismissal is first examined, as this is necessary for examining the relationship of power with dismissal (Wangrow et al., 2018). Theory and prior scholarly research associated with dismissal are reviewed and used to inform five hypotheses associated with the relationships of performance and four power dimensions with the likelihood of dismissal. Next, methods utilized in this study are detailed. As prior scholarly work suggests that competitive sport offers objective measures that are less affected by external factors (Giambatista, 2004), National Collegiate Athletic Association (NCAA) Division 1 college basketball (hereafter referred to as college basketball) coach dismissals in major conferences over 16 seasons are examined. This longer event window overcomes potential problems from missing complex strategic issues manifested over multiple years that may be missed if a shorter timeframe is used (Box-Steffensmeier and Jones, 2004). A Cox Proportional Hazards event history model is used to test the hypotheses.

Taken as a whole, several contributions are made to the literature on executive dismissal. First, a curvilinear relationship of performance with the likelihood of dismissal is found which suggests that the impact of performance improvements on the likelihood of dismissal depends on the current performance level. Second, while expert power from tenure is found to be negatively related to the likelihood of dismissal, the results suggest that expert power’s effect on the likelihood of dismissal occurs much later in the context of college basketball than it has in previously examined contexts. Indeed, the likelihood of dismissal continues to increase until coaches reach their ninth season, at which point it levels off and then declines with further tenure. Third, the results suggest that coaches with greater prestige power associated with reputational capital are less likely to be dismissed, though no relationship is found for changes in structural power from athletic director (AD) succession with the likelihood of dismissal. Lastly, a significant relationship between concentrated governance and the likelihood of dismissal is shown, but in a direction that is contrary to the prediction and prior research. This suggests that additional factors may influence how concentrated and dispersed governance affect the likelihood of dismissal.

Literature review

Performance and dismissal

There are many reasons why executives leave their position. Executives may leave because of death, illness, mandatory early retirement, or other personal reasons. They also may voluntarily leave for an executive position in another organization or they may leave due to stress and fatigue, though stress and fatigue are often associated with pressures from performance demands of powerful groups and individuals (Hambrick et al., 2005). Lastly, they may be dismissed by a more powerful individual or group who feels that performance will improve with new leadership (Shen and Cannella, 2002). In their model of CEO dismissal, Frederickson and colleagues (1988) stated that the most obvious reason for CEOs being fired is that their organizations perform poorly. Yet, they caution that prior studies (e.g. Salancik and Pfeffer, 1980) found that performance explained less than half of the variance of dismissal. Subsequent work has suggested that performance explains as little as 10% of variance associated with dismissal (Finkelstein et al., 2009).

The relationship between performance and dismissal in sport is similar to business, with winning percentage, instead of profitability, often serving as the “bottom line” measure (Fizel and D’Itri, 1999). Within the context of intercollegiate sport, Fizel and D’Itri (1999) examined college basketball coach dismissals from 1983 to 1992 and found that overall winning percentage, rather than a coach’s ability to efficiently manage talent, was the dominant predictor of involuntary (but not voluntary) turnover. Holmes (2011) examined NCAA college football coach dismissals from 1983 to 2006 and found that recent winning percentages were negatively related to the likelihood of dismissal. Interestingly, the models in Holmes’ analysis, which also include tenure, performance against rivals, infractions, and whether the coach was an alumnus, suggest that less than half of the variance was explained. Humphreys et al.’s (2016) examination of NCAA college football coach dismissals from 1980 to 2004 offers a different perspective for the performance–dismissal relationship, finding that performance versus expectations (instead of actual performance) was positively related to retention.

While it may not be possible for new executives to immediately improve performance (Schepker et al., 2017), the belief among decision-makers who initiate succession (e.g. boards, ADs) is that new senior executives can change personnel in key positions, implement new strategies, and deliver a strong message to current employees (Meindl et al., 1985; see also Berns and Klarner, 2017 for a review). Indeed, in a meta-analysis of CEO successions over a 40-year period, Schepker et al. (2017) found that short-term performance is likely to decline with executive succession and that strategic change is a prerequisite for long-term performance improvement with the new executive. In sport, Giambatista (2004) considered three management theoretical lenses of dismissal (common sense, vicious cycle, and ritual scapegoating) but argued that only dismissals consistent with common sense theory are associated with performance improvements. Common sense dismissals assume that decision-makers can distinguish poor executives from good executives and expect new executives to bring enthusiasm and fresh ideas.

Power and dismissal

The power to dismiss a senior executive rests with the organization’s governing body, with the likelihood of dismissal influenced by his or her degree of power relative to the governing body (Haynes and Hillman, 2010). In firms and many other types of organizations, board members collectively decide whether to dismiss the CEO, though more powerful (e.g. board chairman, committee chairs) or independent board members (i.e. directors who joined the board prior to the start of the CEO’s tenure) may strongly influence this decision (Fredrickson et al., 1988; Haleblian and Rajagopalan, 2006). Dismissal decisions are further influenced by structural authority available to the CEO, the presence of board members who joined the board after the CEO began his or her tenure, key stakeholders’ attributions regarding poor performance, and the expected challenges of finding a qualified successor who is acceptable to key stakeholders (Fredrickson et al., 1988; Hilger et al., 2013). Ultimately, power dynamics can vary greatly, with the most powerful individuals best positioned to set agendas and build coalitions that give precedence to their desired outcomes (Lukes, 1974; Zhao et al., 2019).

Finkelstein (1992) argues that executives gain power from four sources: structure, ownership, expertise, and prestige. While these sources of executive power apply to intercollegiate sport coaches, the nature of these power sources differs from traditional firms and organizations. First, a college coach and individuals or groups who are directly involved in dismissal decisions have no ownership stake. Second, college coaches have almost complete authority over acquisition, usage, and retention of human capital (i.e. players) and, thus, effectively serve in the dual role of coach and general manager. This becomes highly salient when comparing the context of traditional firms and organizations with sport teams, as human capital is even more essential for competitive advantage (Bailey and Helfat, 2003; Lajiliet al., 2020). Third, many alternative candidates are visible to the person (i.e. the AD) determining whether dismissal is required. Similar to how the awareness of suitable replacements can influence the likelihood of replacing a CEO (Fredrickson et al., 1988), ample availability of coaches from other teams and assistant coaches from highly successful teams reduces the power of all incumbent coaches (Foreman and Soebbing, 2015).

As Wangrow et al. (2018) discuss in their study of National Basketball Association (NBA) coach dismissals, prior studies examining dismissal have focused on determinants of dismissal (e.g. performance, missed expectations). Wangrow and colleagues examined how other power sources counterbalance ownership’s power, finding that structural power from the coach serving as general manager and prestige power from reputational capital decreased the likelihood of an NBA coach being dismissed. Yet, little attention has been given to how developing and sustaining power affects dismissal in an environment in which some traditional power sources are constant or unavailable. Hence, this study hypothesizes and analyzes the relationships of performance and Finkelstein’s (1992) four sources of power for executives with the likelihood of dismissal but utilizes constructs and relationships that are specific to the environment.

Hypotheses

Dismissal and performance

A variety of options are available for boards to assist CEOs or athletic departments to assist coaches in reversing performance, including providing additional resources, counseling, or internally and externally supporting executives’ abilities and initiatives (McDonald and Westphal, 2003; Vestal and Guidice, 2019). However, if individuals chartered with assessing executive performance believe that poor performance is indicative of poor management (Frick and Simmons, 2008), they may believe that any alternative assistance would be wasteful and delay taking the necessary steps for turnaround. In intercollegiate sport, where the coach has dominant control over the acquisition and usage of talent, this becomes even more acute and, therefore, leaves few alternatives for decision-makers beyond dismissal. Yet, for the entire range of performance, this relationship may not be linear. Stakeholders may not be satisfied with modest improvements when performance is poor, and they may not accept dismissal for any reason when performance is strong (Wangrow et al., 2018). Thus, the following relationship between team performance and the likelihood of coach dismissal is hypothesized.

Hypothesis 1: The relationship between team performance and the likelihood of coach dismissal is curvilinear.

Dismissal and structural power

As stated previously, intercollegiate sport coaches have considerable structural power since they essentially hold the dual role of coach and general manager. While the relationship of duality with structural power is well-established in the literature (Berns et al., 2021; Daily and Johnson, 1997; Haynes and Hillman, 2010; Shen and Cannella, 2002), interdependence between the executive and the board of directors has also been shown to be a source of structural power (Daily and Johnson, 1997; Finkelstein and D’Aveni, 1994). For intercollegiate sport, interdependence is defined by the relationship between the coach and the AD, since an AD who hires the coach will have a vested interest in seeing his or her selected coach succeed. Indeed, with the national visibility of college basketball, the AD’s reputation and ability to stay in his or her position may be dependent on other stakeholders’ perceptions of the team and its performance under the coach’s leadership. As the coach and the hiring AD are linked to one another, the coach retains greater power and, thus, is less likely to be dismissed while the hiring AD remains in place.

Hypothesis 2: Coaches will be more likely to be dismissed when the AD who hired the coach is no longer in the position.

Dismissal and prestige power

Scholars have previously theorized and shown that executives whom others view as prestigious are more likely to have poor performance and other shortcomings overlooked (Finkelstein, 1992; Rindova et al., 2005). Reputational capital can be developed in several ways, including strong performance outcomes recognized by all stakeholders and certification contests in which objective and qualified third parties recognize outstanding performance (Graffin and Ward, 2010; Rindova et al., 2005). For intercollegiate sport coaches, reputational capital provides additional power in a manner similar to executives. Highly prestigious coaches are likely to attract better talent because they may be perceived as more likely to win future championships and more effective in developing NBA-ready talent (Evans and Pitts, 2017). Additionally, teams with highly prestigious coaches are more likely to be considered elite. Thus, recent poor performance is less likely to affect other outcomes (e.g. attendance, television appearances). Conversely, losing a prestigious coach may signal other problems to players being recruited, coaching successor candidates, and other stakeholders—reducing perceptions of whether the team is elite and expectations of the team’s future outcomes. Current and incoming (i.e. recruited) players who are aligned with the prestigious coach may choose to play elsewhere. Additionally, since coaches with substantial reputational capital are more likely to have sizable, guaranteed salaries, universities may incur a substantial financial loss from dismissing such a coach. Overall, reputational capital can provide a college basketball coach with extensive power, since replacing such a coach with one who is viewed by stakeholders as equally capable is likely to be difficult and costly.

Hypothesis 3: Coaches with higher levels of reputational capital are less likely to be dismissed than coaches with lower levels of reputational capital.

Dismissal and expert power

As their tenure progresses, senior executives acquire a deeper understanding of their position and develop capabilities and contacts necessary to operate and succeed in their organization’s environment (Hambrick and Fukutomi, 1991). Comprehending critical contingencies associated with the environment and developing experience to address these contingencies are the foundations of expert power (Finkelstein, 1992). College basketball coaches develop expert power from experience serving as assistant coach or head coach of an NBA or college team. However, the most relevant experience to coaches developing expert power and retaining their position is time spent leading the team, since that experience develops the most organization-specific expertise. Each season further develops the coach’s knowledge associated with recruiting, talent utilization, game strategies against rivals and conference opponents, and interactions with the team’s other stakeholders.

Prior research has shown that CEOs and coaches are given a honeymoon period in which they are allowed time to develop and implement new strategies that can only be evaluated after several seasons (e.g. Allen and Chadwick, 2012; Hambrick and Fukutomi, 1991). The likelihood of dismissal is lower during this period, but scholars have shown that the likelihood increases with tenure. However, the likelihood of dismissal will decrease when the coach has substantial expert power associated with longer tenure. Notably, only a few studies have indicated when, during his or her tenure, an executive or coach is at the greatest risk of dismissal. Kim (1996) found that dismissal was most likely after a CEO’s fourth year and that the likelihood of dismissal did not substantially decline until a CEO’s 10th year. Like Kim, Giambatista (2004) found support for a similar honeymoon period for NBA coaches, which was later supported by Wangrow et al., (2018) who also found that after the eighth season the likelihood of dismissal is reduced by 50% from its peak.

College basketball coaches are also likely to also be granted a honeymoon period in which they are given time to develop and implement new strategies. While the likelihood of their dismissal will increase with tenure, it will eventually decline when they accrue expert power associated with tenure. Thus, the following is hypothesized:

Hypothesis 4: There is an inverted u-shaped relationship between coach tenure and the likelihood of dismissal.

Concentrated governance

The degree that the governance structure is concentrated or dispersed is likely to affect whether strategic action can be taken. When governance entails fewer people and levels, parties within the governance ranks (e.g. boards, shareholders) are better positioned to monitor and act due to greater ease in gaining consensus. Alternatively, scholars have suggested that when there are additional levels in the governance ranks involving more, dispersed individuals, it may be more difficult to develop a common position which limits the ability to reach agreement regarding key decisions (Haynes and Hillman, 2010). Thus, the parties responsible for governance are likely to have greater power relative to the senior executive when governance is concentrated than when it is dispersed (Grabke-Rundell and Gomez-Mejia, 2002).

All universities in the United States share some governance characteristics. Relative to this study, a coach reports to the AD who, in turn, reports to the university’s president or an executive (e.g. provost) who reports to the president. Presidents are accountable to a governance board (directors, regents, trustees). However, a publicly owned university has additional governance layers, since the governance board is appointed by and responsible to the state’s government. Thus, governance is more concentrated in privately owned universities than it is in publicly owned universities. Subsequently, coaches of teams at privately owned universities are likely to have less power than their counterparts at publicly owned universities, since the concentrated governance structure of privately owned universities allows for more expedient and unified decision-making. Based on coaches’ reduced power from concentrated governance at privately owned universities, the following relationship regarding the likelihood of their dismissal is hypothesized.

Hypothesis 5: A coach whose team represents a privately owned university will be more likely to be dismissed than a coach whose team represents a publicly owned university.

Method

To examine college basketball coach dismissals, a model that included all teams from the 10 major conferences (American Athletic, Atlantic Coast, Atlantic 10, Big East, Big 10, Big 12, Pacific 12, Missouri Valley, Mountain West, Southeastern) for each season from 2003–2004 to 2018–2019 was developed.² Each coach’s tenure was divided into annual spells that represent each season in which an individual coached a team. A spell is an at-risk period from the start of one season to the beginning of the following season in which a coach can potentially be dismissed. Splitting the data into annual spells is necessary because some covariates in the survival model vary during each coach’s tenure (Benner and Tripsas, 2012; Wangrow et al., 2018). The data spans 1768 spells (i.e. years in which the coach is at risk of being dismissed) involving 347 coaching tenures.³ One hundred and forty-five of these coaching tenures ended with dismissal. The remaining 202 coaching tenures were voluntarily discontinued (retirement, leaving to coach another NCAA Division 1 team or an NBA team), terminated due to the coach’s death, or continued into the 2019–2020 season.

There were very few mid-season replacements in the sample. As including mid-season replacements would create uneven time spells which violate an assumption in Cox models, only coaches who began a season coaching a team were included.

The majority of data used in this study comes from the commercial website, https://www.sports-reference.com/cbb/, which publishes official historical college basketball statistics and other information regarding teams, coaches, and players. Other data sources included www.coachesdatabase.com, www.NCAA.org, and university websites. Articles found through Google searches were the primary sources used to determine coach dismissal and the coach’s racial or ethnic affiliation.

Dependent variable

Similar to prior studies designed to examine power and performance (e.g. Haleblian and Rajagopalan, 2006), dismissal, instead of turnover, was chosen as the dependent variable. An event variable representing dismissal was assigned “1” for a season (i.e. spell) in which the coach was involuntarily dismissed or forced to resign. This event variable was coded “0” if the coach remained in place at the beginning of the following season or it was determined that the coach had voluntarily left the position or passed away.

It is difficult to determine the cause of an executive turnover event, since organizations are reluctant to publicly disclose reasons of dismissal (Shen and Cannella, 2002). Thus, methods that distinguish dismissals from other types of turnover are necessary for a valid measure of dismissal. Following methods from prior studies (e.g. Shen and Canella, 2002; Wangrow et al., 2018), a list of all coaching turnover events in the sample was created. Two of the authors independently reviewed media accounts and press releases associated with turnover events and assessed whether the coach was dismissed. The authors agreed on 97.4% of the turnover events (Cohen’s kappa = 0.95), suggesting excellent rater reliability (McHugh, 2012). The two authors then discussed information associated with turnover events in which they had disagreed and reached a consensus regarding whether those events were associated with dismissal.

Independent variables

Winning Percentage is used to measure team performance and is the number of games that a team won versus all games that it played during a season.

For structural power, the traditional measure of whether the person holds the dual role of chief executive and leading the group who hires and oversees the chief executive is not appropriate for the context of college basketball, since all coaches make decisions similar to those of professional sport general managers (e.g. assessing and acquiring talent). Instead, interdependence between the AD and the coach is used (see Daily and Johnson, 1997; Holmes, 2011). Thus, structural power was measured using a binary variable, AD Change, with “1” assigned if the current AD did not hire the coach.

Coach Reputational Capital is used to measure prestige power and is a count of coaches’ cumulative accomplishments (see Wangrow et al., 2018 for an example). Two accomplishments are associated with team achievements (reaching the Final Four and winning the national championship), while two accomplishments are associated with personal achievements (winning the Naismith National Coach of the Year Award and winning the conference’s coach of the year award). The result was a variable ranging from 0 to 25 with a mean of 2.66.

Governance concentration was measured using a binary variable, Private University, that was set to “1” if the team represented a privately owned university and “0” if the team represented a publicly owned university.

Control variables and alternative performance measures

An extensive set of control variables associated with coaches, athletic departments, player talent, and stakeholders’ expectations—as well as dummy variables for conference and year—are detailed in Table 1 and are included in the analysis. Since other performance indicators may influence the likelihood of dismissal, Table 1 also includes three additional performance measures (Conference Winning Percentage, Made NCAA Tournament, Tournament Wins) that were used to verify the robustness of findings.

Table 1.

Descriptions of control variables and alternative performance measures.

Variable	Description	Source
Control variables
Alumni	Binary variable set to “1” if the coach had attended the university as a student.	Web searches. https://www.sports-reference.com/cbb/.
Former division 1 player	Binary variable set to “1” if the coach had played NCAA Division 1 college basketball.	Web searches. https://www.sports-reference.com/cbb/.
Coach age	The coach’s age in years at the beginning of the season.	https://www.sports-reference.com/cbb/
Minority	Binary variable set to “1” if the coach represented a minority racial or ethnic group.	Web searches of pictures from press reports.
Football bowl subdivision	Binary variable set to “1” if the university had a team that competed in NCAA college football’s top division during the season.	https://www.sports-reference.com/cfb/
NCAA sanctions	Binary variable set to “1” if the basketball team was under sanctions for NCAA violations during the season.	Web searches. https://www.sports-reference.com/cbb/.
Basketball expenses	Spending (in millions) during the season on the university’s college basketball program.	United States Department of Education EADA database.
Preseason ranking	Team’s Associated Press ranking prior to the season. A team ranked number one was assigned 25, number two was assigned 24, and so on. Teams not ranked were assigned 0.	https://www.sports-reference.com/cbb/
Recruiting rank	Rank of team’s incoming players among all college basketball teams. A team ranked number one was assigned 240, number two was assigned 239, and so on. Teams not ranked or ranked below 240th were assigned 0.	Annual college basketball recruiting ranking from www.247sports.com
Draft value	Count variable representing the talent level for the season. Calculated from the number of current players who eventually played in the NBA and the place in which current players were selected in a future NBA player draft.	https://www.sports-reference.com/cbb/. https://www.basketball-reference.com/
Winning percentage (t−1)	Percentage of wins versus total games played for the team’s prior season.	https://www.sports-reference.com/cbb/
Conference	Dummy variables for each of the 10 conferences.	https://www.sports-reference.com/cbb/
Year	Dummy variables for each of the 16 seasons used in the sample.
Alternative performance measures
Conference winning percentage	Percentage of wins in conference games versus total conference games played for the season.	https://www.sports-reference.com/cbb/
Made NCAA tournament	Binary variable set to “1” if the team made that season’s NCAA Division 1 college basketball tournament.	https://www.sports-reference.com/cbb/
Tournament wins	Count variable indicating the number of games won during the season’s NCAA Division 1 college basketball tournament.	https://www.sports-reference.com/cbb/

Note: NCAA: National Collegiate Athletic Association; NBA: National Basketball Association.

Analysis

A Cox Proportional Hazards event history model (Cox, 1972) was used to examine antecedents of coach dismissals. In this study, a Cox model estimates the hazard rate h(t) which represents the probability that a coach is dismissed during or following a season—a distinct period (i.e. spell) in which the coach is at risk of being dismissed. A Cox model was chosen since it uses data to determine the underlying hazard function instead of assuming characteristics of the underlying hazard function (Benner and Tripsas, 2012; Blossfeld et al., 2007). Additionally, many variables change during a coach’s tenure, and thus need to assume different values over time for a single at-risk coach. By allowing for time-varying covariates, Cox models are ideally suited to handle fluctuations of variables over a coach’s tenure (Blossfeld et al., 2007; Box-Steffensmeier and Jones, 2004). Furthermore, many coaching tenures end voluntarily or continue into the 2019–2020 season, which requires analysis methods that retain right-censored observations. Cox models effectively handle right-censored data, preventing biased parameter estimates (Allison, 1984; Box-Steffensmeier and Jones, 2004).

The smoothed baseline hazard function for all coaches in the sample is shown in Figure 1. The hazard function is shown for 24 years of coaching tenure, in which a coach is at risk of being dismissed during each year. As shown in Figure 1, approximately 1.6% of coaches (hazard rate of approximately 0.016) are dismissed prior to the start of their second season. Of the coaches who remained in position for their second season, approximately 2% of the coaches are dismissed prior to the start of their third season. Indeed, Figure 1 shows that coaches are most likely to be dismissed after their ninth season, in which about 7% of remaining coaches are dismissed. Overall, the results suggest that if 100 NCAA Division 1 coaching tenures from the sample were randomly selected, about 38 of the coaches would be dismissed prior to their tenth season.

Figure 1.

Estimated baseline hazard function—all coaches.

Results

Table 2 shows descriptive statistics and correlations. Results from the event history analysis are provided in a stepwise fashion in Table 3. Exponentiated hazard coefficients are reported, with a positive (negative) coefficient indicating that the likelihood of a coach being dismissed increases (decreases) as the associated variable increases. The variable’s numerical effect on coach dismissal is determined by taking the exponential function of the coefficient (i.e. numerical effect = exp(exponentiated hazard coefficient)). For example, model 7 of Table 3 (i.e. the full model) reports a significant hazard coefficient for the independent variable, Private University, of −0.533. Since exp(−0.533) equals 0.575, the results suggest that, when other factors are held equal, the likelihood of a coach being dismissed decreases by about 42.5% (i.e. 1 − 0.575 = 0.425) when the team is associated with a privately owned university. Robust standard errors (clustered by coach) are reported to correct for potential heteroscedasticity.

Table 2.

Descriptive statistics and correlations.

	Mean	SD	Min	Max	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19
1. Involuntary dismissal	0.08	0.27	0	1
2. Years in position	6.14	5.93	1	43	0.01
3. Minority	0.23	0.42	0	1	0.10**	−0.15**
4. NCAA sanctions	0.02	0.15	0	1	−0.00	−0.01	0
5. Football bowl subdivision	0.72	0.45	0	1	0	0.07**	0.06*	−0.00
6. Alumni	0.13	0.33	0	1	−0.02	0.14**	0.02	−0.01	0
7. Former division 1 player	0.58	0.49	0	1	−0.00	0.02	0.01	0.01	−0.04	0.20**
8. Coach age	49.42	8.3	31	75	0.03	0.50**	−0.06**	−0.00	0.24**	−0.02	−0.10**
9. Basketball expenses	5.52	3.16	0.95	22.9	0.03	0.34**	−0.03	0.01	0.30**	0.01	0.09**	0.26**
10. Preseason rank	2.72	6.3	0	25	−0.10**	0.37**	−0.14**	−0.04	0.18**	0.05*	0.01	0.19**	0.41**
11. Recruiting ranking	152.54	75.56	1	240	−0.01	0.14**	0.05*	0	0.37**	0.05*	0.03	0.08**	0.43**	0.26**
12. Draft value	3.14	4.18	0	27	−0.09**	0.34**	−0.05*	0.02	0.28**	0.07**	0.04	0.21**	0.42**	0.63**	0.37**
13. Winning percentage	57.73	16.1	6.3	97.4	−0.26**	0.28**	−0.15**	−0.04	0.13**	0.05*	0.04	0.12**	0.29**	0.47**	0.17**	0.50**
14. Winning percentage (t − 1)	57.77	16.12	6.3	97.4	−0.16**	0.34**	−0.12**	−0.05*	0.14**	0.08**	0.04	0.13**	0.29**	0.48**	0.21**	0.38**	0.54**
15. Conference winning percentage	50.03	20.82	0	100	−0.25**	0.26**	−0.17**	−0.03	0.03	0.04	0.04	0.10**	0.22**	0.42**	0.10**	0.44**	0.92**	0.52**
16. Made NCAA	0.39	0.49	0	1	−0.20**	0.24**	−0.11**	−0.02	0.16**	0.05*	0.01	0.12**	0.31**	0.44**	0.21**	0.47**	0.72**	0.45**	0.69**
17. Tournament wins	0.51	1.09	0	6	−0.13**	0.26**	−0.12**	−0.02	0.14**	0.06**	0.01	0.14**	0.34**	0.56**	0.19**	0.57**	0.57**	0.38**	0.51**	0.58**
18. AD change	0.42	0.49	0	1	0.08**	0.39**	−0.00	0.01	0.13**	0.09**	−0.08**	0.23**	0.17**	0.15**	0.12**	0.12**	0.11**	0.15**	0.20**	0.07**	0.12**
19. Coach reputational capital	2.64	3.79	0	25	−0.09**	0.60**	−0.17**	0.01	0.21**	0.06*	−0.01	0.56**	0.49**	0.58**	0.24**	0.54**	0.37**	0.38**	0.34**	0.34**	0.45**	0.20**
20. Private university	0.33	0.47	0	1	−0.01	0.14**	−0.01	0.04	−0.47**	−0.07**	0.12**	0.03	−0.05*	−0.10**	−0.22**	−0.09**	−0.11**	−0.11**	−0.08**	−0.07**	−0.10**	−0.06*	−0.06*

Note: N = 1768. NCAA: National Collegiate Athletic Association; AD: athletic director; SD: standard deviation.

*p < 0.01; **p < 0.05.

Table 3.

Event history analysis (dependent variable: coach dismissed during or after season).

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
Independent variables
Winning percentage (H1)		−0.064*** (0.006)	−0.042^† (0.025)	0.042^† (0.025)	0.037 (0.025)	0.043^† (0.025)	0.037 (0.025)
Winning percentage–squared (H1)			−0.001*** (0.000)	−0.001*** (0.000)	−0.001*** (0.000)	−0.001*** (0.000)	−0.001*** (0.000)
AD change (H2)				−0.043 (0.170)			−0.082 (0.168)	−0.104 (0.168)	−0.029 (0.174)
Coach reputational capital (H3)					−0.108* (0.051)		−0.124* (0.051)	−0.123* (0.048)	−0.120* (0.047)
Private university (H5)						−0.469* (0.183)	−0.533** (0.180)	−0.589** (0.183)	−0.508** (0.194)
Control variables
Alumni	0.096 (0.232)	0.139 (0.222)	0.072 (0.221)	0.079 (0.225)	0.088 (0.217)	0.009 (0.228)	0.036 (0.228)	−0.020 (0.217)	0.027 (0.216)
Former division 1 player	−0.098 (0.164)	−0.155 (0.158)	−0.075 (0.160)	−0.082 (0.162)	−0.088 (0.164)	−0.040 (0.158)	−0.064 (0.165)	−0.093 (0.162)	−0.066 (0.166)
Coach age	0.002 (0.011)	−0.004 (0.010)	−0.000 (0.010)	−0.000 (0.010)	0.014 (0.012)	0.001 (0.011)	0.016 (0.012)	0.017 (0.012)	0.021 (0.013)
Minority	0.448** (0.172)	0.221 (0.169)	0.243 (0.166)	0.244 (0.166)	0.193 (0.171)	0.249 (0.165)	0.192 (0.171)	0.232 (0.164)	0.331^† (0.173)
Football bowl subdivision	−0.011 (0.345)	0.144 (0.342)	0.091 (0.342)	0.098 (0.343)	0.131 (0.346)	−0.193 (0.357)	−0.168 (0.364)	−0.261 (0.347)	−0.196 (0.353)
NCAA sanctions	0.028 (0.491)	−0.159 (0.439)	−0.020 (0.461)	−0.004 (0.473)	−0.173 (0.497)	−0.035 (0.440)	−0.177 (0.492)	−0.129 (0.499)	−0.116 (0.518)
Basketball expenses	0.199*** (0.037)	0.217*** (0.034)	0.212*** (0.035)	0.212*** (0.035)	0.226*** (0.035)	0.214*** (0.033)	0.230*** (0.034)	0.240*** (0.036)	0.222*** (0.035)
Preseason ranking	−0.073* (0.033)	−0.069* (0.032)	−0.041 (0.032)	−0.042 (0.032)	−0.024 (0.031)	−0.043 (0.032)	−0.023 (0.031)	−0.033 (0.030)	−0.003 (0.031)
Recruiting rank	0.001 (0.001)	0.001 (0.001)	0.001 (0.001)	0.001 (0.001)	0.001 (0.001)	0.000 (0.001)	0.001 (0.001)	0.001 (0.001)	0.001 (0.001)
Draft value	−0.063^† (0.036)	0.022 (0.029)	0.049 (0.030)	0.050 (0.030)	0.058^† (0.031)	0.053^† (0.031)	0.063^† (0.032)	0.039 (0.034)	0.034 (0.039)
Winning percentage (t−1)	−0.048*** (0.005)	−0.025*** (0.005)	−0.026*** (0.006)	−0.026*** (0.005)	−0.025*** (0.006)	−0.028*** (0.006)	−0.027*** (0.006)	−0.027*** (0.006)	−0.042*** (0.006)
Conference dummies significant at p < 0.05	0 of 9	0 of 9	1 of 9	1 of 9	1 of 9	0 of 9	1 of 9	1 of 9	1 of 9
Year dummies significant at p < 0.05	0 of 15	1 of 15	1 of 15	1 of 15	2 of 15	1 of 15	1 of 15	0 of 15	0 of 15
Alternative performance measures
Conference winning percentage								−0.005 (0.015)
Conference winning percentage–squared								−0.001** (0.000)
Made NCAA tournament									−1.414*** (0.333)
Tournament wins									−1.138* (0.457)
Log pseudolikelihood	−639.18	−598.75	−592.78	−592.75	−590.91	−590.72	−588.19	−591.82	−606.77
Likelihood ratio χ²	232.66***	335.97***	330.03***	335.31***	356.70***	342.35***	399.25***	356.35***	281.92***
Subjects (coaches)	347	347	347	347	347	347	347	347	347
dismissals	145	145	145	145	145	145	145	145	145
Observations	1768	1768	1768	1768	1768	1768	1768	1768	1768

Note: Robust standard errors in parentheses. NCAA: National Collegiate Athletic Association; AD: athletic director.

*** p < 0.001; **p < 0.01; *p < 0.05; ^†p < 0.1.

To test hypothesis 1, which argues for a curvilinear relationship between team performance and coach dismissal, the primary measure of team performance, Winning Percentage, is considered. As shown in model 2 of Table 3, the coefficient for Winning Percentage is negative and significant (p < 0.001). Model 3 indicates support for hypothesis 1, as the coefficient for Winning Percentage–squared is negative and significant (p < 0.001).

To further illustrate the nonlinear relationship between team performance and dismissal, Figure 2 shows the marginal effects of team performance on the likelihood of dismissal. A hazard score of 1.0 indicates that winning has no effect on the likelihood of being dismissed, which is associated with winning 61% of a season’s games. A coach whose team wins about 39% of its games in a season is four times more likely to be dismissed (hazard score of 4.0). Referring to the base rate of dismissal shown in Figure 1, a coach whose team wins 39% of its games after the first year of tenure has a 6.4% probability of being dismissed (0.016 × 4.0 = 0.064). However, a coach with that same level of performance after the fifth year of tenure has a 16% probability of being dismissed (0.04 × 4.0 = 0.16). Conversely, a coach whose team wins 67% of its games in a season is half as likely to be dismissed (hazard score of 0.5). Referring again to Figure 1, the probability of a coach being dismissed whose team wins 67% of its games after the fifth year of tenure is about 2% (0.04 × 0.5 = 0.02). The nonlinear marginal effects shown in Figure 2 provide further support for hypothesis 1.

Figure 2.

Hazard ratios at various winning percentages.

Hypothesis 2 argues for a positive relationship between AD Change and the likelihood of coach dismissal, but the results shown in model 4 indicate that this relationship is not significant. Hypothesis 2 is not supported. The results in model 5 provide support for hypothesis 3, as a significant and negative relationship is shown for Coach Reputational Capital with the likelihood of coach dismissal (p < 0.05).

Hypothesis 4 argues that there is an inverted u-shaped relationship between a coach’s tenure and the likelihood of dismissal. Figure 1 shows that the hazard rate has an inverted u-shaped relationship, in which the likelihood of dismissal increases each year until year 9, after which the likelihood of dismissal decreases. t-Tests that the hazard rates in years 1 and 18 are equal to the hazard rate in year 9 both reject the null hypotheses, indicating a significant curvilinear relationship between tenure and the likelihood of dismissal. Hypothesis 4 is supported.

Model 6 in Table 3 provides results from testing hypothesis 5. The relationship of Private University to the likelihood of dismissal is significant (p < .05), but the coefficient’s sign is opposite to the predicted direction. This suggests that, for the sample, concentrated governance decreases the likelihood of dismissal. Hypothesis 5 is not supported.

Robustness

As shown in model 7, the full model reported no meaningful differences to the previously reported results. Since decision-makers and other stakeholders may emphasize other performance criteria, full models with alternative performance measures to Winning Percentage were also assessed. Model 8 uses Conference Winning Percentage and Model 9 uses two measures (Made NCAA Tournament and Tournament Wins) associated with selection and advancement in the annual national college basketball tournament. While these performance measures were also significantly related to the likelihood of dismissal, their inclusion had no meaningful effects on the results from testing hypotheses 2 through 5.

Discussion

Consistent with prior research, performance was found to explain substantial variance in executive dismissal. In the context of college basketball, these findings are supported using multiple measures of team performance (overall winning percentage, conference winning percentage, and performance postseason advancement). Additionally, these findings add to those from recent studies (e.g. Holmes, 2011; Wangrow et al., 2018) that the relationship between performance and dismissal is curvilinear. As shown in Figure 2, performance improvements have the greatest impact on the likelihood of college basketball coach dismissal when the current winning percentage is about 30%.⁴ However, performance improvements have far less impact on the likelihood of dismissal when performance is very poor (Winning Percentage around 15%) or very good (Winning Percentage above 70%).

Three dimensions of power are significantly related to the likelihood of dismissal. The results provide evidence that prestige power decreases the likelihood of dismissal, as an executive’s performance legacy and cumulative recognition from individuals and groups external to the organization create a degree of prestige that may override the effects of poor performance on the likelihood of dismissal. Moreover, the governance representative (the AD in this study) may not wish to risk his or her reputation by firing an executive (the coach in this study) with a distinguished career, nor may the governance representative be willing to risk losing human capital by dismissing a prestigious executive.

There is also evidence that the likelihood of dismissal is altered as expert power is gained through tenure. As shown in Figure 1, which illustrates the relationship between tenure and the hazard rate for the full model, the hazard rate reaches an apex after 9 years and then declines in subsequent years (substantially after 12 years). While these findings are consistent with Hambrick and Fukutomi’s (1991) seasons of executive tenure, they suggest a longer honeymoon period for college basketball coaches than executives and coaches in other environments. Indeed, findings from studies examining the survival of coaches (e.g. Holmes, 2011; Humphreys et al., 2016; Wangrow et al., 2018) and CEOs (e.g. Brookman and Thistle, 2009; Kim, 1996) all showed honeymoon periods of 5 years or less. This suggests that executives in some contexts (in this case, college basketball coaches) are given greater time to develop expertise, acquire human capital, and implement their strategies. However, like all of the aforementioned studies examining survival (with the exception of Holmes, 2011), the effect of expert power is shown by an eventual and significant decline in the likelihood of dismissal.

Surprisingly, the results suggest that concentrated governance is associated with a reduced likelihood of executive dismissal. The analysis found a significant and negative relationship between Private University and the likelihood of dismissal, in which the likelihood of dismissal declined by 42.5% over the entire sample. This finding suggests that for some instances of concentrated governance, factions may develop that make it difficult to gain consensus or allow an executive to gain power with a sufficiently large subset of the governance structure. Alternatively, some forms of dispersed governance may yield substantial power to a few individuals within the governance structure, thus reducing an executive’s power. This finding is also contrary to agency theory predictions (Boeker and Goodstein, 1993) in which dispersed governance makes “free riding” (i.e. lack of monitoring leading to the executive acting in his or her own interests) more likely. Notably, the analysis does not show how power within the governance structures of privately owned universities (concentrated) and publicly owned universities (dispersed) causes this surprising result. Yet, as this finding runs contrary to prior findings that executives are constrained by concentrated governance (Grabke-Rundell and Gomez-Mejia, 2002), future research that examines power bases within governance structures and the impacts of concentrated versus dispersed governance structures on executive dismissal, performance, and managerial discretion across various contexts is encouraged.

No support was found for the hypothesized relationship of a coach being more likely to be dismissed when the AD who hired the coach is no longer in the position. While it is surprising that losing a strong internal advocate and source of structural power would not influence the likelihood of dismissal, it is possible that a new AD may wish to establish his or her power base before taking more extreme, visible steps. Future researchers are encouraged to examine the range of actions taken by ADs, as well as senior executives and board members in similar positions, early in their tenures. Such studies could include the influence of other powerful stakeholders (e.g. major donors, institutional shareholders) who may discretely mandate their preferred course of action.

This study’s findings are likely generalizable to a variety of contexts associated with human capital. At any given time, a college basketball team is comprised of 12 to 15 players, with a team likely replacing some players prior to each season. Given the smaller number of players, a coach is highly involved in processes and decisions associated with each player. Thus, this study’s findings may be more generalizable to intercollegiate athletics in which there are fewer players on a team (e.g. soccer) than other intercollegiate athletics with many players on a team (e.g. football). The findings may also be generalizable to organizations in which the senior executive is directly responsible for and personally involved in human capital acquisition, development, and retention. This includes firms that are new or have fewer employees, but also includes firms in which human capital associated with a specific function (e.g. research and development) is critical for success. Furthermore, the findings in this study may also be generalizable to leaders responsible for team development within organizations (e.g. middle-level managers), as many of the same dynamics apply.

Practical implications

The results from this study suggest several implications for executives and the governance structure of their organizations. First, the nonlinear relationship between performance and the likelihood of dismissal suggests agency considerations in which executives may take steps to preserve their position – but are not in the organization’s best interests. The results suggest that a coach whose team wins 39% of total games is four times more likely to be dismissed than the average of all coaches in the sample, but this is substantially reduced with minimal performance improvements. On the other hand, the results show that coaches who win more than 75% of their games have almost no chance of being dismissed. This highlights the importance of consistent monitoring and involvement by members of an organization’s governance structure. Executives whose organizations are performing at either end of the performance spectrum may apply strategies and use resources in a manner that maximizes their likelihood of not being dismissed in lieu of applying strategies and using resources in a manner that is best for the organization.

The significant relationship between concentrated governance and the likelihood of executive dismissal suggests other forces within a concentrated governance structure that may influence the likelihood of dismissal. As summarized earlier in this article, even concentrated governance structures like those found in privately owned universities have layers. In the case of college basketball coaches, members from other governance layers rely on the AD for information and recommendations regarding the coach. If ADs in privately owned universities are more likely to develop close personal ties, this may bias information and recommendations made to others within the governance structure. This reinforces the importance of involvement by multiple individuals from various layers of the governance structure when assessing senior executives.

Limitations

This study utilizes data from a variety of archival sources to proxy for performance and power. Reputational capital, for instance, is used to measure prestige power and is the cumulative career recognition from national tournament performance and coach of the year awards. As stated in prior studies, “future research could examine truer values of reputation…and examine how reputational capital from major accomplishments may decline over time” (Wangrow et al., 2018: 345). Survey and interview methodologies that capture these constructs, as well as associated latent behavior and conflict (see Lukes, 1974), may further clarify how power influences dismissal. In addition, the models do not capture streaks, trends, or margins of victory and defeat, which may elicit positive or negative perceptions that also affect the likelihood of dismissal. Further, the models do not include the contractual terms of coaches and ADs, including buyout clauses that the university must pay to a coach or AD when he or she is dismissed prior to contract expiration. Future studies that examine the relationship of the contractual terms for coaches and executives with dismissal, as well as other forms of executive turnover, would provide valuable new insights into factors that influence decision-making associated with executive retention and dismissal.

Conclusion

The motivation for this study was to advance knowledge regarding the influence of performance and multiple forms of power on the likelihood of executive dismissal. The context of college basketball offers a rich environment to study performance, power and executive dismissal, since coaches are responsible for all aspects of their teams’ strategies, as well as acquisition and retention of human capital. The results show that performance’s relationship with the likelihood of dismissal is curvilinear, with performance improvements having the greatest impact at performance levels below the sample’s performance average. Expert power from tenure is significantly related to the likelihood of dismissal, but the honeymoon period for college basketball coaches is several years longer than honeymoon periods found in prior studies of executive and coach dismissal. Prestige power from reputational capital was negatively related to the likelihood of dismissal, though no relationship for structural power with the likelihood of dismissal was found. Lastly, a surprisingly negative relationship between concentrated governance associated with privately owned universities and the likelihood of dismissal was found. These findings advance scholarly knowledge but also raise several new questions regarding the contextual nature of performance and power that are worthy of further inquiry.

Footnotes

Authors’ note

This project’s origin was research work by one of the authors as part of the science research program at John Jay High School in Lewisboro, NY. We commend this continuing program that fosters the development of future researchers in the STEM, medical, and social science fields.

Acknowledgment

We gratefully acknowledge Vince Barker, John Cotton, Donald J. Schepker, and two anonymous reviewers for their help and suggestions. However, any errors or omissions belong solely to the authors.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Allen

Chadwick

(2012) Performance, expectations, and managerial dismissal evidence from the National Football League. Journal of Sports Economics 13(4): 337–363.

Allison

(1984) Event History Analysis: Regression for Longitudinal Event Data. Thousand Oaks, CA: Sage.

Amedu

Dulewicz

(2018) The relationship between CEO personal power, CEO competencies, and company performance. Journal of General Management 43(4): 188–198.

Bailey

Helfat

(2003) External management succession, human capital, and firm performance: an integrative analysis. Managerial and Decision Economics 24(4): 347–369.

Benner

Tripsas

(2012) The influence of prior industry affiliation on framing in nascent industries: the evolution of digital cameras. Strategic Management Journal 33(3): 277–302.

Berns

Gupta

Schnatterly

, et al. (2021) Chief executive officer dismissal: a multidisciplinary integration and critical analysis. Group & Organization Management 46: 1–37.

Berns

Klarner

(2017) A review of the CEO succession literature and a future research program. Academy of Management Perspectives 31(2): 83–108.

Blossfeld

Golsch

Rohwer

(2007) Event History Analysis with STATA. New York, NY: Taylor and Francis.

Boeker

(1992) Power and managerial dismissal: scapegoating at the top. Administrative Science Quarterly 37(3): 400–421.

10.

Boeker

Goodstein

(1993) Performance and successor choice: the moderating effects of governance and ownership. Academy of Management Journal 36(1): 172–186.

11.

Box-Steffensmeier

Jones

(2004) Event History Modeling: A guide for Social Scientists. Cambridge: Cambridge University Press.

12.

Brookman

Thistle

(2009) CEO tenure, the risk of termination and firm value. Journal of Corporate Finance 15(3): 331–344.

13.

Cox

(1972) Regression models and life tables. Journal of the Royal Statistical Society, Series B (Methodological 34(2): 187–202.

14.

Daily

Johnson

(1997) Sources of CEO power and firm financial performance: a longitudinal assessment. Journal of Management 23(2): 97–117.

15.

Evans

Pitts

(2017) The determinants of NCAA basketball recruiting outcomes. Applied Economics Letters 24(5): 351–354.

16.

Finkelstein

(1992) Power in top management teams: dimensions, measurement, and validation. Academy of Management Journal 35(3): 505–538.

17.

Finkelstein

D’Aveni

(1994) CEO duality as a double-edged sword: how boards of directors balance entrenchment avoidance and unity of command. Academy of Management Journal 37(5): 1079–1108.

18.

Finkelstein

Hambrick

Cannella

(2009) Strategic Leadership: Theory and Research on Executives, Top Management Teams, and Boards. New York, NY: Oxford University Press.

19.

Fizel

D’Itri

(1999) Firing and hiring of managers: does efficiency matter? Journal of Management 25(4): 567–585.

20.

Flickinger

Wrage

Tuschke

, et al. (2016) How CEOs protect themselves against dismissal: a social status perspective. Strategic Management Journal 37(6): 1107–1117.

21.

Foreman

Soebbing

(2015) The role of candidate availability in CEO dismissals: an examination of the National Football League. Journal of Management 16(2): 11–25.

22.

Fredrickson

Hambrick

Baumrin

(1988) A model of CEO dismissal. Academy of Management Review 13(2): 255–270.

23.

Frick

Simmons

(2008) The impact of managerial quality on organizational performance: evidence from German soccer. Managerial and Decision Economics 29(7): 593–600.

24.

Giambatista

(2004) Jumping through hoops: a longitudinal study of leader life cycles in the NBA. Leadership Quarterly 15(5): 607–624.

25.

Grabke-Rundell

Gomez-Mejia

(2002) Power as a determinant of executive compensation. Human Resource Management Review 12(1): 3–23.

26.

Graffin

Ward

(2010) Certifications and reputation: determining the standard of desirability amidst uncertainty. Organization Science 21(2): 331–346.

27.

Haleblian

Rajagopalan

(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.

28.

Hambrick

Fukutomi

GDS

(1991) The seasons of a CEO’s tenure. Academy of Management Review 16(4): 719–742.

29.

Hambrick

Finkelstein

Mooney

(2005) Executive job demands: new insights for explaining strategic decisions and leader behaviors. Academy of Management Review 30(3): 472–491.

30.

Haynes

Hillman

(2010) The effect of board capital and CEO power on strategic change. Strategic Management Journal 31(11): 1145–1163.

31.

Hilger

Mankel

Richter

(2013) The use and effectiveness of top executive dismissal. Leadership Quarterly 24(1): 9–28.

32.

Holmes

(2011) Win or go home: why college football coaches get fired. Journal of Sports Economics 12(2): 157–178.

33.

Humphreys

Paul

Weinbach

(2016) Performance expectations and the tenure of head coaches: evidence from NCAA football. Research in Economics 70(3): 482–492.

34.

Kim

(1996) Long-term firm performance and chief executive turnover: an empirical study of the dynamics. Journal of Law, Economics, and Organization 12(2): 480–496.

35.

Lajili

Lin

LYH

Rostamkalaei

(2020) Corporate governance, human capital resources, and firm performance: exploring the missing links. Journal of General Management 45(4): 192–205.

36.

Lukes

(1974) Power: A Radical View (Vol. 1). London: Macmillan.

37.

Lusthaus

Adrien

Anderson

, et al. (1999) Enhancing Organizational Performance: A Toolbox for Self-Assessment. Montreal, QC: Universalia.

38.

McDonald

Westphal

(2003) Getting by with the advice of their friends: CEOs’ advice networks and firms’ strategic responses to poor performance. Administrative Science Quarterly 48(1): 1–32.

39.

McHugh

(2012) Interrater reliability: the kappa statistic. Biochemia Medica 22(3): 276–282.

40.

Meindl

Ehrlich

Dukerich

(1985) The romance of leadership. Administrative Science Quarterly 30(1): 78–102.

41.

Rindova

Williamson

Petkova

, et al. (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.

42.

Salancik

Pfeffer

(1980) Effects of ownership and performance on executive tenure in US corporations. Academy of Management Journal 23(4): 653–664.

43.

Schepker

Kim

Patel

, et al. (2017) CEO succession, strategic change, and post-succession performance: a meta-analysis. Leadership Quarterly 28(6): 701–720.

44.

Shen

Cannella

(2002) Power dynamics within top management and their impacts on CEO dismissal followed by inside succession. Academy of Management Journal 45(6): 1195–1206.

45.

Vestal

Guidice

(2019) The determinants and performance consequences of CEO strategic advice seeking. Journal of General Management 44(4): 232–242.

46.

Wangrow

Schepker

Barker

(2018) Power, performance, and expectations in the dismissal of NBA coaches: a survival analysis study. Sport Management Review 21(4): 333–346.

47.

Zhao

Thatcher

Jehn

(2019) Instigating, engaging in, and managing group conflict: a review of the literature addressing the critical role of the leader in group conflict. Academy of Management Annals 13(1): 112–147.