Race, Market Size, Segregation and Subsequent Opportunities for Former NFL Head Coaches

1.1 CEO Compensation

Due to high levels of compensation and generous exit packages, CEOs generally do not reenter the executive labor market (Ward, Sonnenfeld, and Kimberly 1995). While this phenomenon has not been examined in the NFL due to the lack of consistent and accessible NFL coach salary data, it can be seen following many football coach dismissals anecdotally. Publicly available information from college football highlights the phenomenon of head football coaches receiving large payouts prior to the typical retirement age. For example, Gus Malzahn was owed $21.5 million after his dismissal from Auburn University at age 55 and Kevin Sumlin received $7.3 million when he left the University of Arizona at age 56 (Kirk 2020).

1.2 Causes of Departures

Ward, Sonnenfeld, and Kimberly (1995) described several causes that could result in CEO departures, including performance issues, improper behavior, disagreements, and personality clashes. They reasoned that CEOs who departed firms due to the aforementioned reasons might be identified as damaged goods. CEOs that have been given that label are often viewed as unhirable for future CEO positions. Similarly, NFL head coaches are evaluated and experience adverse career outcomes following instances of poor performance and improper behavior (Foreman, Soebbing, and Seifried 2019). Disagreements and personality clashes can also influence career outcomes for NFL head coaches, especially with conflicts likely to occur due to cultural misunderstandings in the increasingly diverse NFL (Nelson and Quick 2017).

1.3 Demographics and Career Outcomes

Common demographics used for examining NFL head coaches are age and race (e.g. Salaga and Juravich 2020; Solow, Solow, and Walker 2011). Ward, Sonnenfeld, and Kimberly (1995) stated that age is an important factor in former CEOs securing subsequent CEO opportunities. Older CEOs are less likely than younger former CEOs to obtain other active executive positions. However, older CEOs were more likely than younger CEOs to be offered advisory role positions. Also of importance, Ward, Sonnenfeld, and Kimberly (1995) shared that older CEOs are less likely to take on fresh or new challenges and that organizations might be hesitant about hiring an older CEO to replace a departing one. This type of move would likely lead to another quick succession due to the incoming CEOs age.

While Ward, Sonnenfeld, and Kimberly (1995) did not explicitly discuss racial factors in subsequent CEO opportunities, they commented on the existence of an ‘inner circle,’ stating, “The idea is that there is a fairly well defined and relatively closed network or community of CEOs and directors who serve on interlocking boards and who look after their own by helping to find other directorships or positions for fellow elites ousted by an organization” (p. 126). Relatedly, the NFL is known to operate in a relatively closed network of mostly White men who operate as NFL owners, general managers, and head coaches (Collins 2007; Foreman et al. 2018; Foreman Bendickson, and Cowden 2021; TIDES 2020). Therefore, cognitive biases and access limitations to the network or ‘inner circle’ of NFL decision makers may prevent underrepresented coaches from securing head coach opportunities.

1.4 Racial Disparities in NFL Coach Career Opportunities

Brooks et al. (2019) noted that Black head coaches tend to be hired at older ages, have more relevant playing experience, and do not receive equivalent subsequent opportunities. Regarding age, they reported that newly hired White head coaches were 48.4 years old, compared to 51.2 for coaches of color – a difference of almost three years. Notably, among the 14 head coaches under the age of 40 who were hired over the past decade, only two (14.3 %) were coaches of color.

Regarding relevant coaching experience, Brooks et al. (2019) stated that former NFL offensive coordinators were most frequently hired as NFL head coaches. Other common past coaching positions included NFL defensive coordinator and NFL head coach of an alternative club. Almost two-fifths of head coaches hired since 2009 left offensive coordinator positions and a little less than one-third of head coaches hired left other head coaching roles. Interestingly, even though offensive coordinators and current NFL head coaches represent the two largest pools from which new head coach hires are found, half (i.e. 6 of 12) of the newly hired coaches of color occupied the defensive coordinator position prior to becoming a head coach.

Regarding subsequent NFL head coach opportunities, Brooks et al. (2019) presented descriptive statistics showing that Black former head coaches experienced more difficulty obtaining subsequent NFL head coach positions, compared to White former head coaches. Specifically, they stated:

The next coaching positions for outgoing head coaches of color were less varied compared to outgoing white head coaches. Outgoing head coaches of color most often became NFL defensive coordinators (36 %). Outgoing white head coaches most often became NFL defensive coordinators (22 %), NFL offensive coordinators (21 %), or head coaches of other NFL clubs (14 %). Exiting white head coaches were hired for other NFL head coaching positions at twice the rate of coaches of color (14.3 % vs. 7.1 %). They also went on to offensive coordinator positions in the NFL at nearly three times the rate of coaches of color (20.6 % vs. 7.1 %) (Brooks et al. 2019, p. 24).

Those results clearly demonstrate Black coaches struggle to return to the head coaching pipeline positions. However, these descriptive statistics did not control for the quantity or quality of coaching experiences. Building upon the descriptive statistics presented by Brooks et al. (2019), Foreman, Bendickson, and Cowden (2021) found that Black coaches were about 38 % more likely to secure subsequent head coach opportunities in the NFL than their non-Black counterparts. Given these mixed results, we present the following hypothesis:

1.5 Publicity and Market Size

While boards of directors are responsible for appointing CEOs in corporate settings, in the NFL, head coaches are appointed by team owners or general managers. However, personnel decisions made by both boards of directors and NFL team owners and general managers are susceptible to biases and may be based on limited information sets (Collins 2007; Fredrickson, Hambrick, and Baumrin 1988; Foreman, Bendickson, and Cowden 2021). In addition to individual cognitive biases that plague personnel decisions, team owners and general managers are often swayed by public opinion, whether that be (a) to appease stakeholders (e.g. fans and sponsors) or (b) because information sets are often include external evaluations (e.g. from the public, media, and analysts). For example, amid mounting, or even pending, pressure from the public, reluctant team decision-makers have changed their racist mascot names and logos (Ospina 2024), suspended their own players (Foreman, Soebbing, and Seifried 2019), or even hired or fired their head coaches (Hambrick, Bass, and Schaeperkoetter 2014; Flores, Forrest, and de Dios Tena 2012). Similarly, team decision-makers have been known to make suboptimal decisions based on limited information sets that have been largely influenced by the public, such as fans, media and social media, analysts (Bursik 2012; Foreman, Bendickson, and Cowden 2021; Romer 2006). Therefore, public perceptions of NFL head coach candidates can affect team owners’ and general managers’ hiring decisions, whether based on characteristics of the candidates themselves (e.g. race) or the public perceptions in the environments where the coaches previously coached.

The publicity surrounding an ousted CEO’s exit can help or harm their prospects of obtaining another executive role. As noted by Ward, Sonnenfeld, and Kimberly (1995), “For CEOs ousted for negative reasons, the publicity would only reinforce those negative perceptions … For those CEOs ousted for neutral exit reasons, however, the media exposure would help keep them visible to executive search firms” (p. 126). Therefore, for candidates who may be perceived more negatively by the public, whether due to race, performance, or other issues, more publicity could hinder future employment opportunities, whereas candidates perceived more neutrally, for the same reasons of race, performance, etc., may benefit from enhanced media exposure, aiding them in securing future employment.

Similar statements have been made about the importance of retaining media exposure to remain visible in the NFL (Collins 2007), especially since the NFL is a relatively closed network of predominantly White decision-makers. Therefore, we believe that the NFL team’s media market size has been a neglected factor for analysis in discussing the current underrepresentation of Black coaches. Where most studies have analyzed the underrepresentation of Black coaches as a labor market problem (Bozeman and Fay 2013; Braddock, Smith, and Dawkins 2012; Madden 2004), emphasis should be placed on the impact the media market plays in influencing the hiring of Black coaches.

Cunningham and Bopp (2010) discussed how media and press releases for new college football head coaching hires helped to perpetuate the belief that Whites were better at game planning and understanding the strategy of football than Blacks, and Blacks were better recruiters and more relatable for players. Highlighting the coaching ability of Whites, while downplaying it for Blacks, they argued helps to reinforce harmful racial ideologies about Black men in leadership. Considering the role the media plays in covering the hiring of Black coaches in the NFL, combined with an understanding that larger markets are likely to have more news outlets and more widespread coverage of coaches within that market that may have more substantive and negative effects on Black coaches, we offer the following hypotheses:

1.6 Market Segregation

Just as perceptions about potential head coaches may arise based on the size of a market that a coach most recently worked in, another market characteristic that may affect Black head coach appointments is the degree to which the most recent market was segregated. For head coaches who most recently coached in more segregated markets, perceptions from that media market may be biased based on the degree of segregation from that market. Segregation within the metropolitan may be detrimental to an underrepresented candidate because of bias and discrimination that is inherent in the structure of the city, especially as it may result in less familiarity with people of diverse backgrounds. However, more segregated cities (e.g. Chicago, Detroit, New York) may be more likely to neglect people of diverse backgrounds in comparison to less segregated markets (e.g. Minneapolis, Seattle).

Moreover, in the case of high-profile jobs like CEO or NFL head coach, underrepresented leaders may benefit from less attention (Dwivedi, Misangyi, and Joshi 2021; Smith, Chown, and Gaughan 2021). Furthermore, the structure of the media within the city, especially with the prevalence of social media, may reflect the segregated structure of the city itself, resulting in more diverse media outlets leading the coverage of Black coaches and portraying them more positively. Given these competing possibilities for majority and underrepresented candidates for head coach positions in the NFL, we hypothesize:

2.1 Independent Variables

To test the hypotheses, we use three independent variables and two interaction variables. We use a dichotomous race variable that is indicative of whether a former head coach is Black (Black). Observations associated with Black coaches are coded with the value of 1 and observations associated with non-Black coaches are coded with the value of 0. Classifications of former NFL head coaches as Black or not are consistent with reports of NFL head coach races found in the NFL Racial and Gender Report Cards (TIDES 2020). The Black variable is specifically used within the present study to examine Hypothesis 1, that Black former NFL head coaches do not secure future head coach openings at the same speed as their non-Black former head coach peers.

We use two variables to examine Hypotheses 2a and 2b, that market size reduces the time it takes for former NFL head coaches to secure subsequent head coaching positions (Hypothesis 2a) and that Black former NFL head coaches wait longer to secure subsequent NFL head coaching positions relative to non-Black former NFL head coaches, after having coached in a larger market. The two variables used to examine Hypotheses 2a and 2b are team market size (MarketSize) and an interaction variable between Black and MarketSize (Black × MarketSize). To measure the size of each market, we use the population of the metropolitan statistical area where the team is located (George 2007). To account for annual fluctuations in populations, MarketSize is standardized by year, so that the mean MarketSize for each year is 0 with a standard deviation of 1.

Similarly, we use two variables to examine Hypotheses 3a and 3b, that former NFL head coaches secure subsequent head coaching positions at a different speed based on the degree of segregation in the city they coached in (Hypothesis 3a), and that these differences are also based on race (Hypothesis 3b). The two variables used to examine Hypothesis 3 are the degree of segregation in the metropolitan area (Segregation) and an interaction variable between Black and Segregation (Black × Segregation). To measure the amount of segregation in the metropolitan area, we use the divergence score developed by the University of California’s Othering & Belonging Institute (see Othering & Belonging Institute 2022).

2.2 Control Variables

A former NFL head coach’s previous experience is accounted for in several ways. First, seven variables are used to measure the quantity of a former NFL head coach’s experience. The amount of previously held head coach positions a former NFL head coach had (NFLHCJobs), the average number of NFL seasons spent in NFL head coaching positions (AvgHCJobTenure), the number of NFL seasons spent as an NFL head coach (NFLHCExp), the number of NFL seasons spent as an NFL assistant head coach or offensive/defensive coordinator (NFLAHCExp), the number of NFL seasons spent as an NFL position coach (NFLPCExp), the number of NFL seasons spent as an NFL assistant coach at lower levels than position coaches (e.g. quality control coach or assistant position coach; NFLAPCExp), and the number of seasons spent as a football coach in other professional or college football leagues (NonNFLExp) are included in the model. Similarly, the positions last held by former NFL head coaches are also accounted for in the model using dichotomous variables coded with the value of 1 if the position was held in the previous season and 0 otherwise. These positions are NFL head coach (NFLHC), NFL assistant head coach or offensive/defensive coordinator (NFLAHC), NFL position coach (NFLPC), or head coach outside the NFL (NonNFLHC). In addition to the quantity of coaching experiences and the type of recent coaching experiences, we also include measures of overall career performance and recent performance as an NFL head coach. Former NFL head coaches’ career winning percentages as NFL head coaches (CarNFLWin%) and the percent of seasons a former NFL head coach won the conference championship and went to the Super Bowl (i.e. the number of conference championship wins divided by the number of NFL seasons as an NFL head coach; ConfChamp%) represent overall career performance as an NFL head coach. Recent performance as an NFL head coach is measured by the former NFL head coach’s winning percentage in their last season as an NFL head coach (LastNFLWin%). Lastly, the age of the former NFL head coach is included in the model (Age).

The proportional hazard model takes the following form:

where i indexes individual coaches and t indexes time in NFL seasons. The vector X _it represents the vector of independent variables MarketSize and Segregation. The vector Z _it represents the vector of control variables.

2.3 Estimation Issues

Proportional hazard model was estimated using various distributions, including exponential, generalized gamma, Gompertz, loglogistic, lognormal, and Weibull survival distributions. After examining the Akaike’s information criterion (AIC) and Bayesian information criterion (BIC) yielded by each of these survival distributions, the Weibull distribution appeared to fit the data the best (see Frick and Wallbrecht 2012 for more information on survival models). Correlation coefficients among the independent variables did not exceed 0.8, as demonstrated in Table 2, indicating that multicollinearity is not likely a concern in this study (Hair et al. 2010). Robust standard errors were used to alleviate heteroskedasticity concerns.

4.1 Market Size

Fredrickson, Hambrick, and Baumrin (1988) posited that larger groups are more likely to develop factions, and when those factions disagree, adverse decisions regarding the career advancements of organizational leaders can suffer. Therefore, when there are disagreements about a former NFL head coach’s ability to lead an organization, and some stakeholders may believe that Black individuals are ineffective leaders (see Rosette, Leonardelli, and Phillips 2008), Black former NFL head coaches may be less likely to secure subsequent NFL head coaching positions. Indeed, the results of the present study support this notion, with Black former NFL head coaches not experiencing the career advancement benefits of coaching in a larger NFL market that are experienced by their non-Black counterparts. Rather, the results of the present study indicate that the career advancement opportunities, in terms of securing subsequent NFL head coaching opportunities, are diminished for Black former NFL head coaches whose most recent coaching position was in a larger media market. Descriptive statistical analyses outside of sport also provide additional support for our findings.

In a Brookings Institution report, Perry et al. (2022) evidenced that the highest representation of black-owned businesses was found amongst smaller metropolitan areas. For example, Black business owners and entrepreneurs in metropolitan areas such as Oxnard, California and Fayetteville, North Carolina had much higher levels of peer representation than the much larger regions of New York City and Atlanta.

4.2 Segregation

More segregated metropolitan areas are more likely to produce future NFL head coaches who previously served as NFL head coaches. This relationship holds especially true for Black former NFL head coaches. Given the benefit for both Black and non-Black former NFL head coaches (and especially for Black former NFL head coaches), it is possible that the structure of the media within the city, especially with the prevalence of social media, may reflect the segregated structure of the city itself, resulting in more diverse media outlets leading the coverage of Black coaches and portraying them more positively.

While the segregation literature does focus on its pejorative uses and effects, there is additional evidence of “positive” outcomes associated with segregation. Peach (1996) argues that while “good” and “bad” segregation are interrelated, they are distinct concepts, which has led to misunderstandings as to the relationship between social processes and urban spatial patterns. Gil and Marion (2018) evidenced a similar phenomenon in a historical study of small businesses, where there was a positive relationship between residential segregation and market concentration of African American movie theaters. Our study provides additional evidence to Gil and Marion’s (2018) proposal of positive outcomes (e.g. improved consumption in horizontally differentiated goods) associated with higher levels of residential segregation.

4.3 Past Coaching Experience

While both Black and non-Black former NFL head coaches are more likely to secure subsequent NFL head coach positions immediately following a previous stint as an NFL head coach, substantial differences exist between Black and non-Black former NFL head coaches whose most recent position was not an NFL head coach. Specifically, Black former NFL head coaches whose most recent position was not a head coach position are more likely to secure subsequent NFL head coach positions than non-Black former NFL head coaches who also were not most recently in NFL head coach positions. There are many possible reasons why Black former NFL head coaches experience a higher likelihood of securing subsequent NFL head coach positions, relative to non-Black former NFL head coaches, when not already holding an NFL head coach position in the preceding season.

Though Black and non-Black former NFL head coaches appear to experience a relatively equal likelihood of securing subsequent NFL head coach positions immediately after holding head coach positions, and that likelihood is substantially higher than former NFL head coaches who assumed lower level coaching positions or took time away from the NFL, there is a significant racial disparity among these non-NFL head coaches when seeking subsequent NFL head coach opportunities. This disparity may be due to perceptions of former NFL coaches who went from NFL head coaches to lesser positions (or no positions), as they are more likely to be viewed as damaged goods (Ward, Sonnenfeld, and Kimberly 1995). However, among these coaches, Black former NFL coaches who took lower-level coaching positions or time away from the NFL may be perceived differently than non-Black former NFL head coaches. For example, they may have performed poorly under already-adverse circumstances (e.g. poor management, limited resources) after an initial NFL head coach opportunity that could be viewed as a glass cliff was taken. NFL team owners and general managers may view their teams as having better management or more resources, which could allow former NFL head coaches to be more successful. If minorities are more likely to obtain positions in which they are expected to fail (Cook and Glass 2014), Black former NFL head coaches may be less likely to be viewed as damaged goods, and more likely to secure subsequent NFL head coach positions after taking lower-level coaching positions.

Additionally, more previous NFL head coach positions and longer average tenures in those previous NFL head coach positions increases the likelihood of securing subsequent NFL head coach positions. However, more years of experience in general, and the age of head coaches, does not help former NFL head coaches secure subsequent head coach opportunities, indicating that head coach experiences are valuable, when varied. Given that Black NFL coaches tend to have more substantial playing experiences (Brooks et al. 2019), they may enter the coaching profession later, making them older coaches who are less likely to secure head coach positions because of their old age. Thus, age discrimination may be causing racial disparities in NFL head coach opportunities.

4.4 Past Performance

While many non-performance related factors influence the likelihood that former NFL head coaches will secure subsequent NFL head coach positions, performance related factors also play a role. Though no statistically significant evidence was found regarding the relationship between past Super Bowl appearances and securing subsequent NFL head coach positions, the overall career winning percentage and most recent career winning percentage as an NFL head coach were significant predictors of securing subsequent NFL head coach positions. However, these two measures of NFL head coach performance have substantially different effects. As expected, former NFL head coaches with higher career winning percentages as a head coach are more likely to secure subsequent NFL head coach positions.

In contrast, a former NFL head coach’s season winning percentage in the most recent season as an NFL head coach has a negative relationship with securing a subsequent NFL head coach position. Former NFL head coaches with lower winning percentages in their final season as a head coach for a team may experience a degree of forgiveness for that final poor performance that likely played a role in their dismissal. The forgivable final season as the head coach of an NFL team may be excusable for decision makers for several reasons such as: (a) a successful coach’s message may have gotten old with a team and it may be time to move on, (b) decision makers may view that final season of poor performance as an anomaly, or (c) resources may have been unavailable to the coach due to player suspensions (Foreman, Soebbing, and Seifried 2019) or injuries.

5.1 Limitations

While several meaningful statistically significant results were identified in this study, and rationales were provided regarding why these phenomena may be occurring, we cannot be certain as to why these relationships exist. While causal relationships are easier to identify using survival analysis due to the temporal nature of the analysis (Blossfeld, Golsch, and Rohwer 2009), the underlying reasons of why these relationships occur remains uncertain and beyond the purview of the statistical analysis used in the present study. Additionally, exogenous factors may be affecting the analysis. Nevertheless, through a rigorous statistical analysis using numerous robustness checks, we are confident that the results in the study are valid, and the examination of subsequent NFL head coaching opportunities is meaningful.

5.2 Future Research Opportunities

Future research can delve deeper into these issues using qualitative or quantitative analyses. Future research can seek to understand why and how decisions are made when filling NFL head coach vacancies. A variety of qualitative tools could be used to examine this issue including observations, content analyses, interviews, focus groups, or ethnographies. From a quantitative perspective, future research can examine unanswered questions in this research such as whether there is a significant disadvantage for minority coaches in the coach labor market due to sustaining longer professional playing careers. The advent of legalized betting on American professional sport has led to greater data availability of NFL market odds. This would allow future work to assess a head coach’s success based on predicted gambling expectations. Another opportunity for qualitative or quantitative research could be to examine why, or under what conditions, is poor performance in a final season as a head coach desirable for decision makers filling head coach vacancies.