DETERMINANTS OF COLLEGE FOOTBALL COACHES’
Jose M. Plehn-Dujowich**
*School of Tourism and Hospitality Management, Temple University
**Fox School of Business, Temple University
Last revised: January 17, 2010
Corresponding author: Jose M. Plehn-Dujowich, Fox School of Business, Temple University,
451 Alter Hall, 1801 Liacouras Walk, Philadelphia, PA 19122, Tel: 215-204-8139, Fax:
215-204-5587, Email: email@example.com.
DETERMINANTS OF COLLEGE FOOTBALL COACHES’ COMPENSATION
Agency theory posits that executive compensation should be aligned with performance
measures because of the moral hazard and adverse selection problems (Fama, 1980; Jensen &
Meckling, 1976). However, the literature on CEO compensation has yet to provide a conclusive
result into the pay-performance relationship partly due to the unavailability of precise and
sensitive performance measures (Banker & Datar, 1989). To address this issue, we investigated
the effect of performance on compensation in the context of college football coaches where
appropriate performance measures were available. Building on previous CEO compensation
research, we further tested if other major determinants of managerial compensation, including
size, job complexity, managerial ability, market competition, and alma mater status, could
effectively explain the compensation of elite college football coaches.
Results indicated that performance did not have a significant effect on college football
coaches’ compensation. In contrast, other determinants collectively accounted for a large
variation in the compensation of these coaches. Although the insignificant effect of performance
found in this study is in conflict with the agency theory’s prediction, it indeed confirms the past
literature suggesting that performance does not play a major role in explaining executive
compensation (e.g., Jensen & Murphy, 1990a). Consequently, our finding provides further
evidence that performance measures, even though satisfying the two conditions (precision and
sensitivity) suggested by Banker and Datar (1989), do not significantly predict managerial
compensation, and hence demonstrates that current compensation practices fail to take into
account the alignment of principals’ interests with those of agents.
Agency theory; Compensation/Incentives; Sport Industry.
DETERMINANTS OF COLLEGE FOOTBALL COACHES’ COMPENSATION
Agency theory provides two perspectives that explain why executive compensation
should be aligned with performance measures. The first perspective is concerned with the moral
hazard problem, which arises because of the unobservable nature of managerial effort
(Holmstrom, 1979; Holmstrom & Milgrom, 1987; Jensen & Meckling, 1976). Given that
performance measures offer informative signals of effort, the principal is thought to design the
agent’s compensation to be contingent on such measures in order to better align the agent’s
interests with those of the principal, and hence elicit the desired level of managerial effort. The
second perspective addresses the adverse selection problem, which arises because managerial
ability is unobservable (Darrough & Melumad, 1995; Harris & Raviv, 1978; Rothschild &
Stiglitz, 1976; Salop & Salop, 1976; Spence, 1973; Wilson, 1977). This perspective assures that
the principals design the agent’s compensation to be contingent on performance measures to
allow them to screen across agents that have heterogeneous ability, and thereby ensure that the
agent has the incentive to truthfully reveal his ability.
Although both perspectives predict that compensation should increase with performance
measures, the CEO compensation literature has yet to find a conclusive result in the
pay-performance relationship. One line of research provides extensive support for the agency
theory’s prediction (Bushman, Indjejikian, & Smith, 1995, 1996; Healy, 1985; Ittner, Larcker, &
Rajan, 1997; Lambert & Larcker, 1987; Sloan, 1993). However, other studies have found the
weak effects of performance measures on CEOs’ pay (Jensen & Murphy, 1990a; Tosi, Werner,
Katz, & Gomez-Meija, 2000). In a seminal study, Jensen and Murphy (1990a) indicated that a
$1,000 change in shareholder value was associated with just over a $3 change in CEOs’ total
wealth, which led them to conclude that “the compensation of top executives is virtually
independent of performance” (Jensen & Murphy, 1990b: 138). A meta-analytic review of the
previous CEO compensation research by Tosi et al. (2000) also found that performance measures
on average accounted for less than 5 percent of the variance in CEOs’ compensation.
This inconclusive result can be partly attributed to the unavailability of appropriate
performance measures (Hengartner, 2006). According to Banker and Datar (1989), performance
evaluation measures must exhibit high sensitivity and precision in order to predict managerial
compensation; sensitivity refers to the extent to which the change in the level of managerial
efforts and abilities leads to the change in the expected performance, and precision is concerned
with the lack of noise in these measures. However, empirical evidence suggests that commonly
used performance measures, such as return on assets (ROA) and market values, may not satisfy
these two conditions (Bertrand & Mullainathan, 2001; Gabaix & Landier, 2008). With regard to
the insensitivity of market values, Gabaix and Landier (2008) found that if the 250th largest firm
in the S&P 500 replaced its CEO with the CEO in the largest firm (arguably the best CEO), the
250 firm would enjoy only a .016% increase in its market value. Bertrand and Mullainathan
(2001) also casted doubt on the precision of ROA by demonstrating that a firm’s ROA was
greatly determined by proxies for luck, such as changes in exchange rate and average industry
performance, and these luck measures influenced the level of CEO pay to a great extent.
Considering these issues of performance measures, a need clearly exists for testing the
pay-performance relationship in a context where more appropriate measures of performance are
available. In this regard, it is our contention that the compensation of elite sport coaches provides
a desirable context for testing agency theory. The sport context offers a unique opportunity for
addressing research questions in the area of labor market research since organizational goals and
performance of competitive sport organizations can be clearly defined in terms of wins (Kahn,
2000; Frick & Simmons, 2008). In addition, detailed compensation data and precise statistics of
individual and organizational performance are typically available to the public in sport, which
allows researchers to investigate how closely performance is aligned with pay (Bloom, 1999;
Kahn, 2000). Furthermore, performance measures in sport are shown to exhibit high sensitivity
to managerial efforts and abilities (Fizel & D’ltri, 1999; Kahn, 1993). For example, Kahn (1993)
demonstrated that hiring coaches with higher past performance and more experience
significantly improved teams’ organizational performance measured as winning percentage in
the context of a professional hockey league. Research by Fizel and D’ltri (1999) also showed that
the past efficiency of a newly hired college head basketball coach had a significant effect on the
winning percentage of his new team.
Along with benefits associated with clear performance measures, the compensation of
elite sport coaches is an appropriate research setting for investigating the pay-performance
relationship and identifying additional determinants of managerial compensation in the
following aspects. First, coaches of elite sport teams are commonly seen as the equivalent of
corporate executives in terms of assumed leadership roles and behaviors (Kellett, 1999). Recent
data further indicates that the compensation of these coaches is also approaching that of CEOs.
For example, head college football coaches at National Collegiate Athletic Association (NCAA)
Football Bowl Subdivision (FBS) institutions received on average $1.36 million in 2009 (USA
Today, 2009a).Thus, compensation practices and evaluation criteria used in CEO compensation
are assumed to be applicable to sport coaches’ compensation. Second, despite the availability of
observable performance measures, there is an apparent lack of the connection between pay and
performance in this context as some coaches enjoy substantial compensation without having
distinctive on-field performance (USA Today, 2006). This issue also parallels the insensitivity
of performance to CEOs’ pay identified by the CEO compensation literature (Edmans, Gabaix,
& Landier, forthcoming; Gabaix & Landier, 2008; Jensen & Murphy, 1990a; Tosi et al., 2000).
Therefore, the examination of the pay-performance relationship in the context of sport coaches
provides a further insight into whether managerial compensation is contingent on performance in
This paper is intended to address two purposes. First, we aim to test agency theory by
using precise and sensitive performance measures that are available in the context of sport.
Second, we seek to extend the existing body of the executive compensation literature by
investigating the extent to which identified determinants of CEO compensation could predict the
compensation of elite sport coaches.
BACKGROUND AND HYPOTHESES
Agency problems arise from the separation of ownership and management in modern
corporations (Fama, 1980; Jensen & Meckling, 1976). Agents choose the actions that maximize
their own interest, despite the fact that agents work on behalf of principals. Agency theory
addresses the problems of both moral hazard and adverse selection that arise from information
asymmetries between agents and principals. Agency theory posits that incentive contracts can be
designed to align the interest of managers and owners (Eisenhardt, 1989; Jensen & Zimmerman,
According to moral hazard theory, effort-averse agents tend to engage in behavior that
sacrifices shareholders’ interests. Jensen and Meckling (1976) describe performance measures as
signals of the unobservable actions undertaken by agents. Numerous studies argue that
performance-based compensation enhances congruence in the goals of agents and principals,
motivating executives to work hard so as to improve firm value (Banker & Datar, 1989;
Bushman & Indjejikian, 1993; Datar, Kulp, & Lambert, 2001; Feltham & Xie, 1994; Holmstrom,
1979). Holmstrom (1979) developed a moral hazard model in which incentive contracts using
performance measures align the interests of principals and agents. Banker and Datar (1989)
examine the relative weights that should be placed on noisy signals of the outcome of interest to
the principal. They find that a signal should be assigned relatively more weight if it is more
precise or sensitive. In multiple-action models of moral hazard, Feltham and Xie (1994) and
Datar et al. (2001) extend the results in Banker and Datar (1989) by examining the agent’s
allocation of effort across multiple actions, so as to determine how this allocation process
impacts the relative weights on performance measures. Overall, the implications stemming from
moral hazard theory are that pay should increase with better performance (Larcker, 1983;
Murphy, 1985; Sloan 1993). Accordingly, there is extensive evidence that executives are
rewarded on the basis of different performance measures, such as accounting and market
measures (e.g., Bushman et al., 1996; Healy, 1985; Ittner et al., 1997; Lambert & Larcker, 1987;
Another stream of research on executive compensation focuses on adverse selection
problems that arise from the premise that the agent’s ability is unknown to the principal. Highly
capable candidates for a managerial position need to be paid more attractive compensation than
candidates with low managerial talent (Darrough & Melumad, 1995). Adverse selection theory
examines contracts that take into account different abilities of agents in a variety of settings
(Harris & Raviv, 1978; Rothschild & Stiglitz, 1976; Salop & Salop, 1976; Spence, 1973; Wilson,
1977). Managerial compensation is associated with signals that are noisy measures of an
individual’s ability to manage an organization; such signals include education, experience, and
background (Spence, 1973). Rose and Shepard (1997) find that executives are paid more in firms
that are heavily diversified because of matching between high-ability CEOs and firms that are
difficult to manage. Henderson and Fredrickson (1996) indicate that executive compensation is
positively related to information-processing ability because the ability to deal with large amounts
of diverse information tends to be rare, but is critical to organizational performance.
To summarize, a principal may assign positive weight to noisy signals of the outcome for
two reasons. First, there is the moral hazard problem. If the signals are sensitive or precise, then
they enable the principal to better estimate the effort exerted by the agent; thus, when the signals
are assigned positive weights, they encourage the agent to exert higher effort. Second, there is
the adverse selection problem. In assigning positive weights to the performance measures, the
principal provides the agent with the incentive to reveal the truth about his hidden ability; in
other words, it enables the revelation mechanism.
Determinants of Coach Compensation
We investigate the effects of performance and other presumed determinants on managerial
compensation using the sample data of college head football coaches at National Collegiate
Athletic Association (NCAA) Football Bowl Subdivision (FBS) institutions. This research
context is chosen because there is growing attention over the rapid increase in the compensation
of these coaches. In 2007, the average salary of FBS head football coaches exceeded $1 million
for the first time in history (USA Today, 2007a). In just two years, this value went up to $1.36
million with at least 25 head coaches making over $2 million (USA Today, 2009a). In 2010,
Mack Brown, head football coach at the University of Texas, will start a new contract that
guarantees him an annual salary of at least $5.1 million, becoming the first college football coach
to be paid over $5 million (USA Today, 2009b). The academic community has shown great
concern over the rapid increase in the compensation of these coaches. For example, the results of
the Knight Commission on Intercollegiate Athletics’ survey indicate that over 85% of college
presidents believe that college football coaches’ pay are “excessive” (USA Today, 2009a).
However, athletic administrators respond to the criticism over the increase in coaches’
compensation by arguing that coaches are rewarded because of their on-field success and that
universities have to pay more for successful head coaches to keep their programs competitive
(USA Today, 2009c). Nevertheless, some evidence suggests that these coaches are not
necessarily paid based on their performance. At the University of Iowa, for example, head coach
Kirk Ferentz received a guaranteed salary of $3 million in 2007, regardless of a mediocre 6-6
regular season record for the same year. His 55% winning percentage during the previous six
years with the team was also not extraordinary in relation to his significantly high salary.
Examples such as this lead to a central question about what factors actually contribute to high
compensation of college football coaches.
We aim to address this question by proposing that the compensation of NCAA head
football coaches is a function of size (Gabaix & Landier, 2008), job complexity (Rosen, 1981),
managerial ability (Agarwal, 1981), market competition (Karuna, 2007), and performance
(Banker & Hwang, 2008), consistent with the findings of the previous CEO pay research. Indeed,
existing research on coach compensation indicated that these determinants of CEO compensation
could effectively explain the level of elite sport coaches’ salary (Frick & Simmons, 2008;
Humphreys, 2000; Kahn, 2006). For example, Frick and Simmons (2008) found that managerial
ability measured by the coach’s experience had a significant effect on the compensation of head
coaches in the German premier soccer league. Kahn (2006) also showed that managerial ability
and performance collectively explained over 70 percent of the variation in the annual salary of
National Basketball Association (NBA) head coaches. In a more comprehensive study,
Humphreys (2000) documented that the base salary of NCAA head basketball coaches increased
with size (the annual program revenue), job complexity (total enrollment of the university), past
performance (the coach’s career winning percentage), and the level of competition (membership
in the top division). However, these studies primarily focused on narrow aspects of coach
compensation, such as gender (Humphreys, 2000), race (Kahn, 2006), and managerial quality
(Frick & Simmons, 2008), and thereby failed to provide comprehensive theoretical background
for the identified determinants. To address this limitation, we develop the conceptual framework
for each proposed determinant of NCAA head football coaches’ compensation as follows.
Size and Coach Compensation
A substantial literature has demonstrated that organizational size explains a large
proportion of the variance in CEO compensation (e.g., Finkelstein & Hambrick, 1989; Gabaix &
Landier, 2008; Rosen, 1982; Tervio, 2008; Tosi et al., 2000). For example, Tosi et al. (2000)
performed a meta-analysis of the existing empirical CEO compensation research and found that
over the 40% of the variance in the level of CEO pay was determined by the size of firm.
Moreover, a recent study by Gabaix and Landier (2008) showed that a CEO’s pay proportionally
increased with both individual firm size and the average size of firms in the economy.
The literature identifies at least two perspectives that explain the positive effect of size on
CEO compensation (Agarwal, 1981; Gomez-Meija & Wiseman, 1997; Hengartner, 2006). The
first perspective posits that organizational size is an indicator of an organization’s ability to pay
(Agarwal, 1981). That is, greater size allows firms to pay a higher level of compensation to their
CEOs (Gabaix & Landier, 2008). Alternatively, firm size can be seen as a manifestation of job
complexity since larger firms tend to have more complex and diverse structures that are difficult
to manage than smaller firms (Rosen, 1982). Based on this explanation, the largest firms are
assumed to provide the highest salaries to their CEOs in order to “assign the most talented
persons to positions of greatest power and influence” (Rosen, 1982: 321). Although these
perspectives build on different rationales, both confirm that there is a positive association
between firm size and CEO pay. Consistent with this, Humphreys (2000) found that
organizational size measured by the annual revenue of the basketball program had a significant
effect on the base salary of NCAA head basketball coaches. We propose the following
Hypothesis 1: The compensation of the NCAA FBS head football coach is positively
associated with the size of the football program.
Job Complexity and Coach Compensation
Job complexity refers to “the nature and magnitude of the responsibility vested in the
job” (Agarwal, 1981: 38). It is proposed that high job complexity will require more capable
CEOs, leading to a greater level of CEO compensation (Agarwal, 1981; Hengartner, 2006; Rosen,
1982). As discussed earlier, one indicator of job complexity is size (Rosen, 1982). Additionally,
the literature has identified other indicators of complexity, such as internationalization (Sander &
Carpenter, 1998), diversification (Finkelstein & Hambrick, 1989), and market uncertainly
(Finkelstein & Boyd, 1998).
Among these indicators of job complexity identified in the literature, a politicized
environment appears to affect the job complexity of NCAA football coaches. A politicized
environment is defined as an environment where an individual is exposed to high scrutiny and
interest from major stakeholders (Hengartner, 2006). A head football coach of a higher
politicized program may face a more complex and demanding job since he must deal with high
attention from the public and the media. Consequently, coaches in highly politicized
environments are more likely to be paid for their additional efforts. We propose that a highly
politicized environment can be reflected in student body size. Previous research suggested that
U.S. intercollegiate athletic programs are influenced by several major external stakeholders, such
as general students, alumni, and faculty (Putler & Wolfe, 1999; Wolfe & Putler, 2002).
Presumably, an athletic program with a larger student body size has a larger number of these
external stakeholders, and tends to put its head coach in a more highly politicized environment.
We propose the following hypothesis:
Hypothesis 2: The compensation of the NCAA FBS head football coach is positively
associated with the enrollment of the university.
In addition, the academic quality of home institutions can affect the complexity of a
college football coach’s job. There is growing public scrutiny over the low academic standards
of student athletes, especially in major athletic programs, since these programs tend to put a
greater emphasis on field successes than the academic success of their players. To address this
issue, the NCAA has implemented a comprehensive academic reform policy, including the
introduction of the Academic Progress Rate (APR), a metric to evaluate the academic success of
student athletes (NCAA, 2005). Based on this policy, all NCAA member institutions must report
the APR score of each athletic team, and would receive sanctions, such as the loss of athletic
scholarships, if they fail to meet a minimum requirement (NCAA, 2005). Consequently, a
college head coach is required to ensure the high academic success of his players while at the
same time achieving a high winning percentage. Arguably, the difficulty of maintaining high
academic successes of student-athletes would increase when a head coach assumes a job at an
institution with a high academic reputation. High academic standards may also constrain head
coaches in terms of recruiting talented players because prospective student-athletes must meet
high admission requirements to be qualified to join football programs. Given this increased job
complexity concerning the academic performance of student athletes, a head football coach at an
academically recognized institution would likely receive a higher level of compensation. The
following hypothesis is proposed:
Hypothesis 3: The compensation of the NCAA FBS head football coach is positively
associated with the academic quality of the university.
Competition and Coach Compensation
Market competition refers to “the extent to which firms attempt to win business from
their rivals” (Karuna, 2006: 277). The literature has revealed that a level of market competition
influences CEO compensation (e.g., Cunat & Guadalupe, 2005, 2009; Hubbard & Palia, 1995;
Karuna, 2007). A firm in a highly competitive market is more likely to offer a greater level of
compensation to its CEO because she must possess a greater ability to address the wider range of
opportunities and strategic choices to compete with other firms (Finkelstein & Boyd, 1998;
Hubbard & Palia, 1995). Market competition has been conventionally operationalized as industry
concentration because less concentrated industries imply the existence of more rival companies
(DeFond & Park, 1999; Finkelstein & Boyd, 1998). Alternatively, Karuna (2007) proposed that
market competition would consist of multiple dimensions, and measured it using three
indicators: product substitutability (i.e., the degree to which substitutes are available for a given
product), market size (i.e., the level of demand for a given product), and entry costs (i.e., the
initial costs for entering an industry).
In the context of college football, however, the level of competition can be clearly
measured by a single indicator, membership in a competitive conference. In particular, among
the 11 individual conferences constituting the NCAA FBS division, six conferences are
collectively called “BCS conferences”, and are considered more competitive conferences.
Therefore, membership in BCS conferences would indicate that each program needs a more
capable head coach to stay competitive within the conference, resulting in greater pay. Thus, the
following hypothesis is proposed:
Hypothesis 4: Head coaches in BCS conferences receive greater compensation than
those in non-BCS conferences.
Human Capital and Coach Compensation
Based on human capital theory (Becker, 1964), Agarwal (1981) proposed that the amount
of human capital that a CEO possesses, such as educational level and work experience, may
indicate the CEO’s ability to perform her job, thereby influencing the level of compensation. In a
similar vein, Spence (1973) asserts that principals use human capital measures as indicators of an
agent’s unobserved ability to select a more capable agent. Consistent with this proposition,
Agarwal’s (1981) research found that work experience measured by the number of working
years had a significant positive effect on CEO compensation. Subsequent studies further
supported the positive effect of human capital (e.g., Banker, Plehn-Dujowich, & Xian, 2009;
Finkelstein & Hambrick, 1989; Fisher & Govindarajan, 1992). For example, Finkelstein and
Hambrick (1989) demonstrated that CEOs with general management experience received greater
amounts of bonuses than those without such experience. By examining the compensation of
profit center managers (PCM), Fisher and Govindarajan (1992) showed that a PCM’s
compensation was positively associated with three measures of human capital: job tenure, firm
tenure, and age. A recent study by Banker et al. (2009) found a positive relationship between
human capital variables and the compensation of university presidents. In a more relevant study,
Frick and Simmons (2008) identified the positive effect of the coach’s experience on his
compensation in the context of the German premier soccer league. Collectively, it can be
proposed that the greater amount of human capital a college football coach possesses, the greater
level of compensation he would likely receive. Our next hypothesis is:
Hypothesis 5: The compensation of the NCAA FBS head football coach is positively
associated with his human capital.
Past Performance and Coach Compensation
Agency theory posits that compensation should be contingent on performance measures
to align the interest of agents and principals (Eisenhardt, 1989; Jensen & Zimmerman, 1985).
This notion is supported by a substantial empirical literature that investigated the
pay-performance link in CEO compensation (Bushman et al., 1996; Healy, 1985; Ittner et al.,
1997; Lambert & Larcker, 1987; Sloan, 1993). Furthermore, Banker and his colleagues (2008,
2009) provided evidence that performance measures could effectively predict compensation of
non-CEOs. For example, Banker et al. (2009) demonstrated that a university president received a
higher level of compensation when she had high performance in the previous institution (Banker
et al., 2009). In a different research context, Banker and Hwang (2008) found that the price of an
e-service provider’s service was determined in part by her past performance. Consistent with
these findings, Humphreys (2000) showed that NCAA basketball coaches with higher career
winning percentages tended to receive greater base salaries than those with lower winning
percentages. The following hypothesis is proposed:
Hypothesis 6: The compensation of the NCAA FBS head football coach is positively
associated with his past performance.
Alma Mater Status and Coach Compensation
Finally, we propose that college football head coaches may accept a discount in their
compensation when they work for their alma mater institutions. This proposition builds upon the
two psychological theories: social identity theory and stewardship theory. First, according to
social identity theory, individuals have a tendency to classify themselves into a wide range of
social categories, such as gender, age cohort, and organizational membership (Ashforth & Mael,
1989). This theory further proposes that when individuals develop a high level of identification
with a given social group, they are likely to engage in behaviors that support the group (Ashforth
& Mael, 1989; Mael & Ashforth, 1992). In line with this perspective, Mael and Ashforth (1992)
showed that alumni of a university could develop a high level of organizational identification
with their alma mater institution, and that those with high identification tended to produce
pro-organizational behavior, such as generous financial contributions to the university. Based on
this finding, it is likely that a head football coach who serves for his alma mater program has a
high level of identification with the program; thereby, he would likely to accept lower salary to
support the institution’s financial status, given most programs operate with deficits.
Second, contrary to agency theory that assumes managers tend to maximize their own
interest, stewardship theory proposes that “managers are not motivated by individual goals, but
rather are stewards whose motives are aligned with the objectives of their principals” (Davis,
Schoorman, & Donaldson, 1997: 21). Based on this theory, an agent is assumed to perform her
job based on intrinsic motives, such as affiliation, rather than extrinsic motives, such as
incentives (Davis et al., 1997). Accordingly, the agent’s utility is maximized when she satisfies
these intrinsic motives, and monetary rewards exceeding “an income to survive” may not greatly
affect her utility (Davis et al., 1997). According to Davis et al (1997), individuals’ behavior can
be defined in terms of stewardship theory when they have high identification with the
organization. Therefore, given the high organizational identification that one may develop with
her alma mater institution, a head coach could be satisfied with receiving a lower level of salary
if he works for his alma mater program. The following hypothesis is proposed:
Hypothesis 7: The NCAA FBS head football coach is more likely to accept lower
compensation when he works for his alma mater institution.
The NCAA is the governing body of more than 1,000 university and college athletic
programs in the United States and Canada. While it is a voluntary association, the NCAA is the
regulating entity for major collegiate sports in the United States. The main components of the
NCAA’s purpose include equitable governance of competition and integration of intercollegiate
athletics into the higher education system (NCAA, 2009a). With stark contrasts existing between
member institutions, the NCAA is segmented into three different divisions. Division I (D-I),
Division II (D-II), and Division III (D-III) represent the differing levels of competition with D-I
being the highest. Each member institution self-determines their division and must then meet the
appropriate divisional criteria (NCAA, 2009b). For example, D-I and D-II may both offer
athletic scholarships to their athletes, while D-III may not (NCAA, 2009b). Further segmentation
exists within the Division I category. D-I institutions that offer football are classified as either
Football Bowl Subdivision (FBS) or Football Championship Subdivision (FCS) (NCAA, 2009c).
Until 2006, the FBS was formerly known as Division I-A and FCS as Division I-AA. FBS
institutions are eligible to compete in post season bowl games that can be financially lucrative.
The FCS schools are eligible to compete in the NCAA Division I Football Championship, while
the FBS is the only NCAA sport without a traditional tournament format to determine a
The current research focused on the compensation of FBS head coaches because of
significant differences between the two D-I subdivisions. One of the primary differences is the
number of athletic scholarships that may be granted (FBS 85; FCS 63). FBS institutions must
also meet a minimum average attendance figure of 15,000, while there is no minimum
requirement in the FCS (NCAA, 2009b). Many other financial disparities exist between the two
subdivisions including television and bowl payouts. The FBS consistently has television
packages bringing in millions of dollars per institution and additional large income injections
from bowl revenue sharing. In contrast, FCS conferences have inconsistent coverage and often
receive no money from the conference television agreement. Vast discrepancies also exist with
regard to the compensation of coaches. For example, the University of Montana, one of the most
successful FCS programs, paid an annual compensation of $144,500 for its head coach Bobby
Hauck in 2009 (Missoulian, 2009), while FBS coaches received an average salary of $1.36
million in the same year (USA Today, 2009a).
Of the 11 individual conferences of the FBS, six conferences receive automatic bids for
their conference champion to the Bowl Championship Series (BCS) and are therefore referred to
as “BCS conferences.” According to the BCS official website, the BCS is “a five-game
arrangement for postseason college football that is designed to match the two top-rated teams in a
national championship game and to create exciting and competitive matchups among eight other
highly regarded teams in four other games” (BCSFootball.org, 2009: 1). Due to the lack of a true
play-off system in the NCAA’s top football division, the BCS “championship” serves as the
current substitute for a traditional tournament format. In 2009, there were an additional 29
non-BCS bowl games that took place in the college football post-season (Football Bowl
Association, 2009). With the absence of a traditional tournament format, the opportunity to go to a
bowl game has represented a welcomed culminating experience for competing teams, as well as an
opportunity for institutions to garner additional revenues. The success of football programs are
also determined by several ranking systems. The Associated Press (AP) Poll, for example, has
been in existence since 1936, longer than any other poll in college football history. AP derives
their poll by compiling the top 25 rankings (named AP Top 25) from 65 designated sportswriters
The lead decision maker in the hiring and compensation of FBS football coaches is
normally either the Director of the athletic department (Athletic Director) or President of the
university. These two positions sometimes represent competing forces with politics playing a
significant role. A common public perception is that the Athletic Director simply makes the
decision. While this is indeed possible, with so much riding on these decisions it is often a more
complicated process. In 1997, NCAA legislation restructured the governance of the NCAA and
firmly placed the presidents in control of the organization. With this clear understanding of
presidential authority over athletics, it appears that Presidents must technically be involved in the
hiring of coaches and setting their compensation level. Standard protocol for most institutional
hiring and compensation decisions is for the president and board of trustees to ultimately sign off
on employment contracts.
However, there are situations where the Athletic Director is still the leader in the
decision-making of the hiring and compensation of coaches. These situations are similar to hiring
and compensation practices in other academic departments throughout the university in which
the President does not actively participate, but still has the final approval. Additionally, a high
level of political power through previous career success or longer tenure may allow an Athletic
Director to make the final decision on the hiring and compensation process. In such situations,
the Athletic Director clearly has the major influence in the process, with the President’s technical
approval being merely a formality.
Data and Sample
To test our hypotheses, we examined compensation data of head football coaches at
NCAA FBS institutions in 2006 and 2007. FBS consists of 11 different conferences, each of
which has 8 to 13 schools, and three independent schools that do not belong to any particular
conferences. The study period of 2006 and 2007 was chosen due to the availability of
compensation data in the USA Today database. There were 119 FBS institutions in 2006 and 120
in 2007, resulting in a total of 239 university-year observations during this study period. From
this initial pool, we selected our study sample using the following criteria. First, coaches who
were newly hired by the current program were excluded to control for the possible effect of
coach turnover on the level of compensation. Second, we restricted our analysis to those who
served as FBS head coach for both of the previous two years in order to take into account
previous performance effects over the two year period. Third, head coaches of independent
schools were excluded since we used membership in particular conferences (i.e., BCS
conferences) as the indicator of competition. Finally, we did not include coaches at private
institutions because compensation data were not available for most of these institutions.
Consequently, our final dataset included 151 university-year observations.
Coach compensation. We collected the compensation data of FBS head coached from
the USA Today’s online database in 2006 and 2007. This database listed three types of
compensation data: salary, other income, and maximum bonus (USA Today, 2007b). Salary
includes regular payment directly from the university, such as base salary, deferred payment, and
annuity payment. Other income refers to incomes from other agreements that are not related to
salary, such as media deals and shoes and/or apparel contracts. Maximum bonus refers to the
greatest amount of additional payment that the coach can receive if his team meets prescribed
goals related to on-field performance and other criteria (e.g., academic performance of student
athletes). Consistent with the previous CEO compensation literature (e.g., Core, Holthausen, &
Larcker, 1999; Finkelstein & Boyd, 1998), we obtained the total compensation value of each
coach by summing these three compensation data and entered it as the dependent variable.
Size. Size was measured as the total revenue generated by the football program in the
previous year (2004 and 2005). Data on the revenue of each football program was collected from
the Equity in Athletics database provided by the U.S. Department of Education.
Enrollment. As discussed earlier, we expected that large enrollment size would represent
a proxy for a high politicized environment, resulting in high complexity of the head coach’s job.
Therefore, we measured enrollment as the total number of full-time undergraduate students in
2006 and 2007 from the Integrated Post Secondary Education Data System (IPEDS).
Academic quality. Academic quality of the university was measured by constructing the
factor, Factor (Academic Quality), with four indicators: SAT scores, average professor salary,
Carnegie classification, and U.S. News rank. The description for each indicator is provided
First, SAT scores were used to capture the quality of students (Banker et al., 2009).
Specifically, we obtained the sum of the average of the 25th and 75th percentiles for math scores
and the average of verbal scores among incoming students in 2007, and used it as our first
measure of the academic quality of the university. The second indicator was the average salary of
professors of all ranks in the institution. We assumed that high salary is likely to attract and
retain professors with high reputation, leading to the high academic quality of the institution
(Gomez-Meija & Balkin, 1992). Data on the average professor salary in 2006 and 2007 was
collected from American Association of University Professors (AAUP) Faculty Salary Survey,
available at the Chronicle of Higher Education website. Furthermore, we used two additional
indicators to represent the overall reputation and quality of universities. First, the Carnegie
Classification of Institutions of Higher Education classifies U.S. universities based on the degree
level offered and the level of research activities undertaken. Specifically, universities listed in the
highest classification, “Research Universities – Very High Research Activity” (RU/ VH),
represent institutions that provide a wide range of doctoral degrees and actively engage in
research activities. Therefore, we created a dummy variable that had the value of 1 if the
university was classified in the RU/ VH classification and 0 if otherwise as a proxy for university
quality. Second, we used U.S. News National Universities Rankings to capture the overall
reputation of the university. We coded 1 for universities ranked in the top tiers (Tier 1 and 2),
and 0 for those ranked in the lower tiers (Tier 3 and 4).
Competition. To measure the level of competition for a given football program, we
created a dummy variable that had the value of 1 for football programs that belong to BCS
conferences and had the value of 0 for those that belong to non-BCS conferences.
Human capital. The literature has suggested that human capital increases with the
amount of experience that a person has in relation to her job (Agarwal, 1981). The coach’s
human capital was measured with five experience-related variables: age, current tenure, years as
FBS head coach, NFL head coach experience, and NFL player experience. Age was measured as
chronological age of the head coach at the beginning of the season. Current tenure was measured
as the number of years for which the coach has served as the current program’s head coach.
Years as a FBS head coach was operationalized as the number of years for which the coach has
served as head coach for any FBS programs. Following Banker et al. (2009), we constructed the
factor, Coach (Experience), using these three indicators, to capture the general working
experience of the coach.
Furthermore, the following two dummy variables were included to represent the coach’s
experience in relation to the National Football League (NFL), the top professional football
league in the U.S. The first variable was NFL head coach experience that had the value of 1 for
the coach with NFL head coach experience. The second dummy, NFL player experience, had 1
for coaches that served as player for NFL teams. We assumed that these two variables would
also have a positive effect on the coach’s compensation. Information related to these experience
variables were collected from various online sources, such as the university’s official athletic
Past performance. To measure on-field performance of the coach for the previous two
seasons, we constructed a factor, Factor (Performance), for each season with four variables: total
wins, conference wins, AP Top 25 rank, and bowl game participation. Total wins was measured
as the number of regular season wins that the team achieved for a given season, while conference
wins was measured as the number of conference wins. AP Top 25 rank was entered as a dummy
variable with the value of 1 if the team ranked in the AP Top 25 at the end of the season, and 0
for otherwise. Bowl game participation was measured with a dummy variable that had 1 for
teams that were eligible for either BCS or non-BCS bowl games in the post season and 0 for
otherwise. Along with these two performance factors, we included the career FBS winning
percentage of the head coach to represent his long-term performance.
Alma mater. Alma mater status of the coach was represented by a dummy code that had
the value of 1 if the coach served as the head coach in his alma mater institution and the value of
0 for otherwise.
Other control variables. Consistent with Kahn (2006), the possible effect of the coach’s
race on compensation was controlled by including a dummy variable that had 1 for white
coaches and 0 for non-white coaches. In order to take into account the effects of the coach’s
contract characteristics, we entered the number of contract years left and a dummy code
representing a new contract (1 for new contract; 0 for otherwise). Campus location was included
to control for the location effect to compensation (1 if the campus was located in either an urban
or suburban area; 0 for otherwise). Given that successful programs are more likely to offer higher
compensation, we further included the conference winning percentage of the program over the
past 10 years. Finally, a year dummy was included to control for differences in compensation by
year (1 for the 2007 season; 0 for the 2006 season).
We used the following multiple regression model to test our hypotheses:
Total compensationi,t = β0 + β1Campus locationi + β2Program success(t-10) – (t-1) + β3Race
dummyi + β4New contract dummyi,t + β5Contract years lefti,t +
β6Year2007i + β7Size i,t-1 + β8Enrollmenti,t + β9F(Academic quality)i,t
+ β10BCS conference dummyi + β11F(Experience)i,t + β12NFL head
coach experiencei + β13NFL player experiencei + β14(t
-2)F(Performance)i,t-2 + β15(t-1)F(Performance)i,t-1 +β16Career FBS
winning % of the head coachi,t + β17Alma mater dummyi + εi,
Where the subscript i refers to the head coach and t refers to the year; Total compensation
is measured as the natural logarithm of the coach’s annual total compensation value; Campus
location has 1 if the university is located at either an urban or suburban area and 0 if otherwise;
Program success is the natural logarithm of the conference winning percentage of the football
program over the past 10 years; Race dummy has 1 for white coaches and 0 for non-white
coaches; New contract dummy has 1 for coaches with new or amended contracts and 0 for
otherwise; Contract years left is measured as the natural logarithm of the number of years left
on the coach’s contract; Year 2007 has 1 for year 2007 and 0 for year 2006; Size is measured as
the natural logarithm of the total revenue generated by the football program in the previous
season; Enrollment is the natural logarithm of the total number of full-time undergraduate
students at the beginning of the school year; F(Academic quality) is the factor that has high
loadings on SAT scores, average professor salary, Carnegie classification, and U.S. News rank;
BCS conference dummy has 1 for football programs that belong to BCS conferences and 0 for
otherwise; F(Experience) is the factor that has high loadings on age, current tenure, and years as
FBS head coach; NFL head coach experience is measured as a dummy variable that has 1 for
coaches that served as head coach for any NFL teams and 0 for otherwise; NFL player
experience is measured as a dummy variable that has 1 for coaches who played for NFL teams
and 0 for otherwise; (t-2)F(Performance) is the factor that has high loadings on total wins,
conference wins, AP Top 25 rank, and bowl game participation two seasons ago; (t –
1)F(Performance) is the factor that has high loadings on total wins, conference wins, AP Top 25
rank, and bowl game participation in the previous season; Career FBS winning % of the head
coach is the natural logarithm of the career winning percentage as a FBS head coach; Alma
mater dummy has 1 for coaches who served as a head coach for their alma mater institutions and
0 for otherwise.
Table 1 illustrates the descriptive statistics of selected variables. On average, the sample
FBS football coaches had maximum annual pay of $1,369,118 in 2006 and 2007 with 4.2 years
left on their contracts. The average age of the coaches was 53.9, and they served as head coach
of their current programs for an average of 7.3 years and for any FBS football programs for 9.9
years. Regarding the performance related variables, these coaches on average won 6 to 7 games
for the regular season and about 4 games within their respective conferences in each of the
previous two seasons, and had the average career FBS winning percentage of 55 percent. As for
institutional characteristics, the universities included in the current dataset had an average of
1,136 SAT scores for their incoming freshman, paid on average $80,491 for their faculty, and
had the average undergraduate enrollment of 19,207. Finally, with regard to program
characteristics, the sample FBS football programs had the average revenue of $18,340,194 in the
previous year, and won an average of 52 percent within their conferences over the past 10 years.
Insert Table 1 about here
Table 2 presents the results of exploratory factor analysis. Using the Kaiser's criterion and
the Scree Test, four factors are derived. First, Factor (Academic Quality) consists of four
variables: SAT scores, Carnegie classification, U.S. news ranks, and professor salary. This factor
is a proxy for the academic quality and complexity of the university. Second, Factor
(Experience) is formed by three variables: age, current tenure, and years as FBS head coach. This
factor is a proxy for the general experience accumulated by the football coach. Furthermore, (t-1)
Factor (Performance) represents the coach’s on-field performance in the previous season,
whereas (t-2) Factor (Performance) represents his performance two seasons ago. Both factors
are formed by four variables: total wins, conference wins, AP top 25 ranks, and bowl game
participation, for respective season.
Insert Table 2 about here
Table 3 shows the correlations of the variables included in the regression analyses. The
results indicate that all three performance measures have significant positive correlations with
total compensation (r(t-2)Factor(Performance) = .28; r(t-1)Factor(Performance) = .40; rcareer FBS winning percentage
= .51), consistent with the prediction of agency theory. Furthermore, in lined with the
hypothesized relationships, total compensation is positively correlated with Factor (Academic
Quality) (r = .56), enrollment (r = .51), size (r = .84), and BCS (r = .77). In contrast, the results
do not support the hypothesized positive correlation between total compensation and experience
measures, and the negative correlation between total compensation and alma mater. Regarding
control variables, contract years left (r = .51), program success (r = .32), and campus location (r
= .22) are found to have significant positive correlations with total compensation.
Insert Table 3 about here
Testing of Hypotheses
Table 4 presents the results of the regression models. In Column 1, total compensation is
regressed on control variables. Columns 2 – 8 present the results of partial models which each
hypothesized determinant is separately entered into the regression with the control variables. In
Column 9, all independent variables are included in the analysis. This full model yields an
adjusted R-squares value of .78, indicating that these independent variables collectively explain a
substantial amount of the variation in the compensation of FBS football coaches.
Insert Table 4 about here
Regarding the effect of each determinant, size has a significant positive effect on total
compensation when included with control variables (Column 2). Furthermore, Column 9
demonstrates that the significant positive effect of size holds after entering the other independent
variables (β = .29, t = 3.62, p < .01). This finding supports Hypothesis 1, confirming the findings
of previous studies suggesting that executive compensation is an increasing function of
organizational size (e.g., Gabaix & Landier, 2008). Column 3 shows that enrollment has a
significant positive effect when entered into the model with the control variables. In addition, in
the full model (Column 9), enrollment still significantly explains the variance in coach
compensation (β = .27, t = 2.09, p < .05). Thus, Hypothesis 2 is supported. In contrast,
although academic quality is found to have a significant effect in Column 3, the results of
Column 9 show that this significant effect disappears after the inclusion of the other independent
variables (β = .01, t = .16, p > .10), resulting in the rejection of Hypothesis 3. As for the
competition effect, both Column 4 and Column 9 indicate that coaches in BCS conferences are
more likely to receive a higher level of compensation than those in non-BCS conferences (β
= .44, t = 2.59 p < .05). Therefore, Hypothesis 4 is retained.
In Hypothesis 5, we proposed that coaches’ compensation was an increasing function of
their human capital. In line with this hypothesis, the coefficient of Factor (Experience) provides
a significant positive result in Column 5 and 9 (β = .11, t = 2.49, p < .05), supporting that
coaches with more general work experience tended to receive higher pay. However, two NFL
experience-related variables do not show significant results in both the partial and full models.
This finding indicates that compensation of coaches is not affected by these specific aspects of
experience. Thus, Hypothesis 5 is partially supported. Furthermore, while the positive effect of
only career FBS winning percentage is found in Column 7, none of the three performance
measures provide significant results in Column 9, which leads to the rejection of Hypothesis 6.
Consequently, the results do not support agency theory that predicts the strong relationship
between pay and performance. Finally, alma mater has a marginally negative effect on coach
compensation in Column 8, but do not exhibit significant effects in the full model (β = -.29, t =
-1.11, p > .10). Therefore, we do not find evidence to support Hypothesis 7.
The correlation analysis in Table 3 indicates that BCS and size are highly correlated (r
= .85). Thus, there is a concern about whether these two variables may essentially represent the
same construct. To test this speculation, we separately ran the full regression model with BCS
and non-BCS samples, and examined if the significant effect of size would hold for each sample.
The results of this analysis are presented in Table 5. The BCS model indicates that size has a
significant positive effect on compensation of BCS coaches (β = .19, t = 2.03, p < .05).
Furthermore, the non-BCS model reveals that the compensation of non-BCS coaches
significantly increases with size (β = .44, t = 2.67, p < .05). Thus, the results indicate that size
and membership in BCS conferences independently influence coach compensation.
Insert Table 5 about here
Using appropriate measures of performance available in the context of sport, this study
tested the proposition derived from agency theory that there should be a close link between pay
and performance to align the interests of agents and those of principals. We further aimed to
investigate how effectively the identified determinants of CEO compensation could explain the
compensation of college head football coaches. With regard to the pay-performance link, our
results are unsupportive of agency theory; none of the three performance measures included in
the analysis significantly predicted the level of college coaches’ compensation. While this
finding is in conflict with agency theory, it indeed confirms the past literature suggesting that
performance does not play a major role in predicting executive compensation (e.g., Frick &
Simmons, 2008; Jensen & Murphy, 1990a; Tosi et al., 2000). As discussed earlier, Tosi et al.
(2000) found that performance measures explained less than 5 percent of the variation in
compensation of CEOs. Jensen and Murphy’s (1990a) study revealed the insensitivity of CEOs’
performance measured by shareholder wealth to their total compensation. In the sport context,
Frick and Simmons (2008) identified the insignificant effects of performance measures on
German soccer coaches’ compensation. Consequently, our finding provides further evidence that
performance measures, even though they satisfy the two conditions (precision and sensitivity)
suggested by Banker and Datar (1989), do not significantly predict managerial compensation,
and suggests that current compensation practices fail to take into account the alignment of
principals’ interests with those of agents.
Concerning other determinants of college football coaches’ compensation, the results
mostly supported that major determinants of CEOs’ pay could effectively account for a large
variation in the compensation of these coaches. First, size is found to have a significant positive
effect on coaches’ total compensation. A further examination of the regression model in Column
2 in Table 4 suggests that size solely accounts for a .40 increment in adjusted R-squares from the
baseline model (Column 1). Interestingly, this finding confirms the results of the meta-analysis
by Tosi et al. (2000) indicating that firm size on average explained about 40 percent of the
variance in CEO compensation. Second, we find the positive effect of enrollment on coach
compensation. This result supports the view that large enrollment may reflect a highly politicized
environment in which the coach is exposed to high public scrutiny (Hengartner, 2006), and thus
indicates that coaches’ salaries are determined in part by the complexity of their jobs.
Third, it is found that the level of competition significantly influences the compensation
of college football coaches. Specifically, coaches in BCS conferences are more likely to receive
higher compensation than those in non-BCS conferences, holding the effects of other
determinants constant. This finding corresponds with Humphreys’ (2000) study demonstrating
that the compensation of NCAA head basketball coaches was positively affected by membership
in the most competitive division. Thus, Humphreys (2000) and our study collectively confirm the
notion that executives in competitive markets would likely receive greater pay because they must
have a greater capability to win competition over other competitive firms (teams) in the markets
(Finkelstein & Boyd, 1998; Hubbard & Palia, 1995). Fourth, in line with human capital theory
(Becker, 1964), we find a positive effect of experience on coach compensation. This finding
suggests that a coach with higher age, more FBS head coach experience, and longer tenure tends
to receive a higher level of compensation, which supports that human capital is used to evaluate
how well the coach would perform his job (Agarwal, 1981; Fisher & Govindarajan, 1992;
Spence, 1973). Alternatively, greater experience may indicate that the coach has more power
over the governance of the athletic program (Finkelstein & Hambrick, 1989). According to
Finkelstein and Hambrick (1989), CEOs with longer tenure are likely to have more influence
over their boards of directors, and are often capable of “effectively dictating what their own pay
will be” (p.124). Building on this speculation, the coach with longer tenure and more experience
may have an ability to influence athletic administrators, such that he could receive higher
In contrast, we did not find a significant effect of academic quality. While this finding is
inconsistent with our hypothesis, it can be explained from the politicized environment
perspective. According to Hengartner (2006), political pressures sometime constrain the level of
CEO pay, which are typically manifested in the media that criticizes the high payment of top
corporate executives. In the current study context, such high political pressures are likely to
occur at academically distinguished universities because their faculty members tend to be critical
of large expenditures of athletic programs; UC Berkeley faculty, for example, recently voted
against the university’s financial support for its athletic department (San Francisco Chronicle,
2009). Given this perspective, although coaches at universities with high academic reputation
could face greater job complexity to maintain high academic standards of their players, these
institutions may not reward their coaches with greater pay because of their faculty’s high
scrutiny over athletic expenditures.
Finally, the results do not indicate the negative effect of alma mater on coach
compensation. Thus, we fail to find evidence to support social identity and stewardship theories
suggesting that individuals tend to categorize themselves into social groups, such as their alma
mater institutions, and engage in behavior that supports these in-groups, including financial
sacrifices (Ashforth & Mael, 1989; Mael & Ashforth, 1991). Alternatively, this insignificant
result may be attributable to the small number of coaches who served for their alma mater
institutions. The frequency analysis shows that the current data set contains only 28 (out of 151)
coaches with the alma mater status. Consequently, considering that the correlation and regression
coefficients of alma mater consistently show negative signs, we would find a significant negative
effect of this variable if the dataset contained more observations.
Limitations and Directions for Future Research
While this study contributes to the literature by testing agency theory and identifying the
major determinants of college football head coaches, it has some limitations. First, the current
dataset includes compensation and other related data only over the two year period. Although
this study period is chosen because of the availability of compensation data of college football
coaches, the use of the short-term observations does not allow us to examine the long-term
performance-pay relationship and to further identify the possible time lag effect of performance
measures on compensation. Second, our results are based on the data of one sector of the sport
industry. Given that compensation and performance data of other sectors of the sport industry
such as professional sport leagues are readily available, future research should investigate the
pay-performance link and the effects of other determinants of executive compensation using
different sport samples. Finally, this study does not take into account the detailed contract
structures of college football coaches. The USA Today’s database lists the contracts of the
majority of the coaches at the public institutions, and these contracts provide detailed
information regarding incentives, contract terms, and breach of contract. Therefore, further
investigation over the effects of incentives and other conditions on coaches’ performance and the
pay-performance relationship should provide a more comprehensive insight into the determinants
of managerial compensation.
Agarwal, N. 1981. Determinants of executive compensation. Industrial Relations, 20 (1): 36-45.
Ashforth, B.E., & Mael, F. 1989. Social identity theory and the organization. Academy of
Management Review, 14: 20-39.
Banker, R.D., & Datar, S. 1989. Sensitivity, precision, and linear aggregation of signals for
performance evaluation. Journal of Accounting Research, 27: 21-39.
Banker, R. D., & Hwang, I. 2008. Importance of measures of past performance: Empirical
evidence on quality of e-service providers. Contemporary Academic Research, 25 (2):
307 – 337.
Banker, R.D., Plehn-Dujowich, J., & Xian, C. 2009. The competing roles of managerial effort
and ability in executive compensation: A principal-agent theory and evidence on the
compensation of university presidents. Working paper. Temple University, Philadelphia,
BCSFootball.org. 2009. BCS Background. Retrieved from
Becker, G. 1964. Human capital. New York: National Bureau of Economic Research.
Bertrand, M., & Mullainathan, S. 2001. Are CEOs rewarded for luck? The ones without
principals are. Quarterly Journal of Economics, 116: 901-932.
Bloom, M. 1999. The performance effects of pay dispersion on individuals and organizations.
Academy of Management Journal, 42 (1): 25-40.
Bushman, R., & Indjejikian, R. 1993. Accounting income, stock price, and managerial
compensation. Journal of Accounting and Economics, 16 (1-3): 3-23.
Bushman, R., Indjejikian, R., & Smith, A. 1995. Aggregate performance measures in business
unit manager compensation: The role of intra-firm interdependencies. Journal of
Accounting Research, Supplement: 101-128.
Bushman, R., Indjejikian, R., & Smith, A. 1996. CEO compensation: The role of individual
performance evaluation. Journal of Accounting and Economics, 21: 161-193.
Core, J.E., Holthausen, R.W., & Larcker, D.F. 1999. Corporate governance, chief executive
officer compensation, and firm performance. Journal of Financial Economics, 51:
Cunat, V., & Guadalupe, M. 2005. How does product market competition shape incentive
contracts. Journal of the European Economic Association, 3 (5): 1058 – 1082.
Cunat, V., & Guadalupe, M. 2009. Executive compensation and competition in the banking and
financial sectors. Journal of Banking and Finance, 33: 495-504.
Darrough, M., & Melumad, N. 1995. Divisional versus company-wide focus: The trade-off
between allocation of managerial attention and screening of talent. Journal of
Accounting Research, 33 (Supplement): 65-94.
Datar, S., Kulp, S., & Lambert, R. 2001. Balancing performance measures. Journal of
Accounting Research, 39 (1): 75-92.
Davis, J.H., Schoorman, F.D., & Donaldson, L. 1997. Toward a stewardship theory. Academy of
Management Review, 22: 20-47.
DeFond, M.L., & Park, C.W. 1999. The effect of competition on CEO turnover. Journal of
Accounting and Economics, 27: 35 – 56.
Edmans, A., Gabaix, X., & Landier, A. forthcoming. A multiplicative model of optimal CEO
incentives in market equilibrium. Review of Financial Studies.
Eisenhardt, K. M. 1989. Agency theory - An assessment and review. Academy of Management
Review, 14 (1): 57-74.
Fama, E. 1980. Agency problems and the theory of the firm. Journal of Political Economy, 88:
Feltham, G., & Xie, J. 1994. Performance measure congruity and diversity in multi-task
principal/agent relations. The Accounting Review, 69 (3): 429-453.
Finkelstein, S., & Boyd, B. 1998. How much does the CEO matter? The role of managerial
discretion in the setting of CEO compensation. Academy of Management Journal, 41
Finkelstein, F., & Hambrick, D.D. 1989. Chief executive compensation: A study of the
intersection of markets and political processes. Strategic Management Journal, 10:
Fisher, J., & Govindarajan, V. 1992. Profit center manager compensation: An examination of
market, political and human capital factors. Strategic Management Journal, 13:
Fizel, J.L., & D’ltri, M.P. 1999. Firing and hiring of managers: Does efficiency matter? Journal
of Management, 25, 567-585.
Football Bowl Association. 2009. 2009-2010 Bowl Schedule. Retrieved from
Frick, B., & Simmons, R. 2008. The impact of managerial quality on organizational
performance: Evidence from German soccer. Managerial and Decision Economics, 29
(7): 593 – 600.
Gabaix, X. & Landier, A. 2008. Why has CEO pay increased so much? The Quarterly Journal
of Economics, 123 (1), 49 – 100.
Gomez-Meija, L., & Balkin, D.B. 1992. Determinants of faculty pay: An agency theory
perspective. Academy of Management Journal, 35 (5): 921 – 955.
Gomez-Mejia, L., & Wiseman, R. 1997. Reframing executive compensation: An assessment and
outlook. Journal of Management, 23 (3): 291-474.
Harris, M., & Raviv, A. 1978. Some results on incentive contracts with applications to education
and employment, health insurance, and law enforcement. American Economic Review,
68 (1): 20-30.
Healy, P. 1985. The effect of bonus schemes on accounting decisions. Journal of Accounting
and Economics, 7: 85-107.
Henderson, A. D., & Fredrickson, J.W. 1996. Information-processing demands as a determinant
of CEO compensation. The Academy of Management Journal, 39(3): 575-606.
Hengartner, L. 2006. Explaining executive pay: The roles of managerial power and complexity.
Unpublished doctoral dissertation. University of St. Gallen.
Holmstrom, B. 1979. Moral hazard and observability. Bell Journal of Economics, 10: 74-91.
Holmstrom, B., & Milgrom, P. 1987. Aggregation and linearity in the provision of intertemporal
incentives. Econometrica, 55(March): 303-328.
Hubbard, R.G., & Palia, D. 1995. Executive pay and performance: Evidence from the U.S.
banking industry. Journal of Financial Economics, 39 (1): 105 – 130.
Humphreys, B. R. 2000. Equal pay on the hardwood: The earnings gap between male and female
NCAA division I basketball coaches. Journal of Sports Economics, 1 (3): 299 – 307.
Ittner, C., Larcker, D., & Rajan, M. 1997. The choice of performance measures in annual bonus
contracts. The Accounting Review, 72: 231-255.
Jensen, M.C., & Meckling, W. 1976. Theory of the firm: Managerial behavior, agency costs and
ownership structure. Journal of Financial Economics, 3: 305-360.
Jensen, M.C., & Murphy, K.J. 1990a. Performance pay and top-management incentives. Journal
of Political Economy, 98 (2): 225 - 264
Jensen, M.C., & Murphy, K.J. 1990b. It’s not how much you pay, but how. Harvard Business
Review, May-June: 138-153.
Jensen, M., & Zimmerman, J. 1985. Management compensation and the managerial labor market.
Journal of Accounting and Economics, 7: 3-9.
Kahn, L. 1993. Managerial quality, team success and individual player performance in National
Hockey League. Industrial and Labor Relations Review, 46: 531-547.
Kahn, L. 2000. The sports business as a labor market laboratory. Journal of Economics
Perspectives, 15: 75-94.
Kahn, L.M. 2006. Race, performance, pay, and retention among national basketball association
head coaches. Journal of Sports Economics, 7 (2): 119 – 149.
Karuna, C. 2007. Industry product market competition and managerial incentives. Journal of
Accounting and Economics, 43: 275 – 297.
Kellett, P. 1999. Organizational leadership: Lessons from professional coaches. Sport
Management Review, 2, 150-171.
Lambert, R., & Larcker, D. 1987. An analysis of the use of accounting and market measures of
performance in executive compensation contract. Journal of Accounting Research, 25
Larcker, D. F. 1983. The association between performance plan adoption and corporate capital
investment. Journal of Accounting and Economics, 5: 3-30.
Mael, F., & Ashforth, B.E. 1992. Alumni and their alma mater: A partial test of the reformulated
model of organizational identification. Journal of Organizational Behavior, 13: 103 –
Missoulian. 2009. February 14. UM signs football coach Hauck to one-year contract. Retrieved
Murphy, K. J. 1985. Corporate performance and management remuneration: an empirical
analysis. Journal of Accounting and Economics, 7: 11-42.
NCAA. 2005. APR 101: Implementation of penalty structure triggers new terminology,
consequences, questions. Retrieved from
NCAA. 2009a. Our Mission. Retrieved from
NCAA. 2009b. What's the difference between Divisions I, II and III? Retrieved from
NCAA. 2009c. Composition & Sport Sponsorship of the NCAA. Retrieved from
Putler, D.S., & Wolfe, R.A. 1999. Perceptions of Intercollegiate Athletic Programs: Priorities
and tradeoffs. Sociology of Sport Journal, 16: 301 – 325.
Rose, N., & Shepard, A. 1997. Firm diversification and CEO compensation: Managerial ability
or executive entrenchment? The RAND Journal of Economics, 28 (3): 489-514
Rosen, S. 1982. Authority, control, and the distribution of earnings. The Bell Journal of
Economics, 13 (2): 311 – 323.
Rothschild, M., & Stiglitz, J. 1976. Equilibrium in competitive insurance markets: An essay on
the economics of imperfect information. Quarterly Journal of Economics, 90 (4):
Salop, J., & Salop, S. 1976. Self-selection and turnover in the labor market. Quarterly Journal
of Economics, 90 (4): 619-627.
Sander, W.M.G., & Carpenter, M.A. 1998. Internalization and firm governance: The role of CEO
compensation, top team composition, and board structure. Academy of Management
Journal, 41 (2): 158 – 178.
San Francisco Chronicle. 2009. UC Berkeley faculty endorse cut in athletic aid. November 6.
Sloan, R. 1993. Accounting earnings and top executive compensation. Journal of Accounting
and Economics, 16(1-3): 55-100.
Spence, M. 1973. Job market signaling. Quarterly Journal of Economics, 87: 355-374.
Tervio, M. 2008. The difference that CEOs make: An assignment model approach. American
Economic Review, 98 (3), 642 – 668.
Tosi, H.L. Werner, S., Katz, J.P., & Gomez-Meija, L.R. 2000. How much does performance
matter? A meta-analysis of CEO pay studies. Journal of Management, 26 (2): 301 –
USA Today. 2006. Contracts for college coaches cover more than salaries. November 16.
USA Today. 2007a. College football coaches calling lucrative plays. December 6. Retrieved
USA Today. 2007b. Compensation for Div. I-A college football coaches. Retrieved from
USA Today. 2009a. College football coaches see salaries rise in down economy. November 9.
USA Today. 2009b. Texas' Mack Brown becomes highest-paid football coach. December 9.
USA Today. 2009c. Missouri shows how schools pay a price for football success. November 10.
Wilson, C. 1977. A model of insurance markets with incomplete information. Journal of
Economic Theory, 16 (2): 167-207.
Wolfe, R.A., & Putler, D.S. 2002. How tight are the ties that bind stakeholder groups?
Organization Science, 13 (1): 64 – 80.
Descriptive Statistics of Selected Variables
Variables N Mean s.d. Min Max
Age 151 53.89 8.11 40 81
Current tenure 151 7.32 5.89 2 42
Past years as FBS head coach 151 9.91 7.36 3 42
(t – 2) Total win 151 6.44 2.62 0 13
(t – 2) Conference win 150 4.25 1.93 0 9
(t – 1) Total win 151 6.76 2.68 1 13
(t – 1) Conference win 151 4.25 1.87 0 8
Career FBS winning percentage 151 .55 .15 .17 .97
Contract years left 144 4.23 1.82 0 10
Maximum total pay 143 1,369,118 936,823 130,000 4,365,000
SAT scores 142 1136.04 99.24 920 1,330
Professor salary 143 80,491 9,704 58,300 104,600
Enrollment 149 19,207 7,459 4,461 36,835
(t – 1) Revenue of the football program 148 18,340,194 16,450,000 740,749 63,798,068
Program winning percentage 151 .52 .15 .16 .88
Factor Analysis for Academic Quality, Experience, and Past Performance
Academic (t-2) (t-1)
Variables Quality Experience Performance Performance
SAT Score .90 -.00 .15 .15
Carnegie Classification .75 .17 .08 -.00
U.S. News Rank .80 .11 .11 .15
Professor Salary .78 .04 .05 .13
Age .10 .76 -.09 -.06
Current Tenure .04 .65 .25 .24
Years as a FBS Head Coach .13 .90 .15 .16
(t -2 )Wins .15 .05 .97 .17
(t -2) Conference Wins -.00 .11 .88 .16
(t -2) AP Top 25 Rank .20 .03 .52 .41
(t-2) Bowl Participation .13 .10 .71 .18
(t -1 )Wins .13 .14 .15 .94
(t-1) Conference Wins -.06 .13 .13 .85
(t -1) AP Top 25 Rank .18 -.03 .25 .58
(t -1) Bowl Participation .19 .10 .17 .75
Cronbach alpha .72 .69 .87 .87
Number of observations 135 151 150 151
Descriptive Statistics and Correlationsa
Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
2. NFL head coach 0.10 0.30 .04
3. NFL player 0.11 0.32 -.10 .14
4. (t -2) Factor
0.02 1.02 -.04 -.03 -.11
5. (t -1) Factor
0.01 0.99 -.01 .00 -.03 -.03
6. Career FBS
0.44 0.10 .27** .05 -.14 .45** .48**
winning % b
7. Alma mater 0.20 0.40 -.01 -.03 .08 .08 .13 -.04
8. New contract 0.30 0.46 -.10 .09 .22* -.08 .19* -.01 -.14
9. Contract years
1.63 0.36 -.22* .02 .09 .10 .35** .17 -.05 .41**
10. Race 0.96 0.20 .15 .07 -.19* -.01 -.15 -.06 -.31** .04 -.08
(Academic 0.00 0.94 .02 .17 .04 .06 -.02 .23* -.11 -.10 .16 -.05
12. Enrollmentb 9.82 0.39 .11 -.25** .07 .07 .11 .31** -.16 .01 .21* -.13 .50**
13. Location 0.83 0.38 .13 .15 .16 .06 .16 .09 .17 .11 .26** -.09 .05 .22*
14. Size 16.33 1.08 .07 .04 -.02 .31** .40** .54** -.07 -.02 .39** -.13 .64** .48** .08
15. BCS 0.64 0.48 -.00 .08 -.10 .21* .25** .38** -.06 -.08 .35** -.16 .71** .49** .06 .85**
0.42 0.10 .19* .03 -.09 .44** .42** .64** .22* -.10 .01 -.08 .04 .10 .16 .35** .15
17. Year 2007 0.50 0.50 .05 .00 -.05 -.12 .04 -.00 .00 -.11 -.10 -.04 -.06 -.10 -.02 -.02 -.14 .01
Compensation 13.92 0.81 .13 .10 .02 .28** .40** .51** -.10 .06 .51** -.10 .56** .51** .22* .84** .77** .32** -.00
n = 122
Natural Logarithms are used
* p < .05; ** p < .01
Results of OLS Regression Models Predicting Total Compensationa, b
Variables 1 2 3 4 5 6 7 8 9
Intercept 10.97*** (0.48) 3.32*** (0.59) 3.21** (1.32) 11.25*** (0.41) 11.05*** (0.31) 11.02*** (0.51) 10.85*** (0.59) 11.02*** (0.48) 5.06*** (1.72)
Campus location 0.22 (0.16) 0.30*** (0.10) 0.13 (0.14) 0.09 (0.13) 0.23** (0.10) -0.01 (0.17) 0.16 (0.15) 0.26 (0.16) 0.08 (0.11)
Program success 1.82*** (0.59) 0.15 (0.38) 1.61*** (0.53) 2.28*** (0.49) 1.51*** (0.38) 2.02*** (0.60) -0.10 (0.79) 2.02*** (0.59) 0.50 (0.52)
Race -0.17 (0.29) 0.13 (0.18) 0.02 (0.26) -0.02 (0.24) 0.19 (0.19) -0.20 (0.31) -0.05 (0.28) -0.29 (0.29) 0.10 (0.20)
New contract -0.17 (0.14) 0.03 (0.09) -0.12 (0.13) -0.08 (0.12) 0.07 (0.09) -0.30** (0.14) -0.26* (0.13) -0.20 (0.14) -0.06 (0.09)
1.31*** (0.17) 0.47*** (0.12) 1.08*** (0.15) 1.01*** (0.15) 0.61*** (0.12) 1.42*** (0.19) 1.04*** (0.19) 1.31*** (0.17) 0.54*** (0.13)
Year 2007 0.13 (0.12) 0.09 (0.07) 0.16 (0.10) 0.10 (0.10) 0.21*** (0.08) 0.05 (0.12) 0.05 (0.11) 0.13 (0.12) 0.12* (0.07)
Size 0.57*** (0.04) 0.29*** (0.08)
Enrollment 0.83*** (0.13) 0.27** (0.13)
(Academic 0.41*** (0.05) 0.01 (0.06)
BCS 1.18*** (0.09) 0.44** (0.17)
0.19*** (0.07) 0.11** (0.05)
NFL head coach 0.24 (0.20) 0.21 (0.14)
NFL Player 0.06 (0.19) 0.14 (0.12)
(t – 2)
0.05 (0.07) 0.06 (0.05)
(t – 1)
0.07 (0.08) 0.07 (0.05)
winning 3.14*** (0.84) -0.26 (0.63)
Alma mater -0.29* (0.16) -0.11 (0.10)
N 141 138 139 124 141 124 124 141 122
Adjusted R2 .37 .77 .51 .58 .74 .38 .44 .39 .78
Standard errors are shown in parentheses
Each model has different numbers of observations due to missing data
* p ≤.10; ** p ≤.05; *** p ≤.01
Results of the Full OLS Regression Model for BCS and Non-BCS Samplesa
Intercept 7.38*** (2.05) 3.87 (3.69)
Campus location -0.05 (0.12) 0.43 (0.26)
Program success 0.57 (0.66) -0.01 (1.07)
Race 0.08 (0.19)
New contract 0.02 (0.11) -0.20 (0.17)
Contract years left 0.27 (0.18) 0.44 (0.26)
Year 2007 0.16* (0.08) 0.04 (0.15)
Size 0.19** (0.10) 0.44** (0.17)
Enrollment 0.32* (0.17) 0.15 (0.24)
Factor (Academic quality) -0.03 (0.07) 0.10 (0.15)
Factor (Experience) 0.06 (0.06) 0.08 (0.10)
NFL head coach 0.25 (0.15) -0.11 (0.36)
NFL Player -0.18 (0.16) 0.35 (0.25)
(t – 2) Performance 0.11 (0.07) 0.02 (0.10)
(t – 1) Performance 0.12* (0.06) 0.00 (0.13)
Career FBS winning percentage -0.72 (0.80) 0.69 (1.45)
Alma mater -0.03 (0.13) -0.16 (0.22)
N 78 44
Adjusted R2 .40 .53
Standard errors are shown in parentheses
Non-BCS sample includes only white coaches
* p ≤.10; ** p ≤.05; *** p ≤.01