Yuhei Inoue*

                               Jose M. Plehn-Dujowich**

                                     Aubrey Kent*

                                    Steve Swanson*

            *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:



       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.


        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 Theory

         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;

Sloan, 1993).

       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.


Research Setting

       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” (, 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

and broadcasters.

       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.


Descriptive Results

       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.

Additional Analysis

       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,
        PA. 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
        (2): 179-199.
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
        (Supplement): 85-129.
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 The
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.
       Retrieved from
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.
       Retrieved from
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.
       Retrieved from
USA Today. 2009b. Texas' Mack Brown becomes highest-paid football coach. December 9.
       Retrieved from
USA Today. 2009c. Missouri shows how schools pay a price for football success. November 10.
       Retrieved from
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.

                                                    Table 1
                                  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

                                     Table 2
      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

                                                                               Table 3
                                                               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
1. Factor
                    -0.02      0.90
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**
    left b
10. Race            0.96       0.20    .15     .07     -.19*     -.01    -.15    -.06    -.31**    .04    -.08
11. Factor
    (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**
16. Program
                    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
18. Total
    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


                                                                             Table 4
                                                 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)
Contract years
                   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)
Career FBS
     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


                                     Table 5
    Results of the Full OLS Regression Model for BCS and Non-BCS Samplesa
                                                         BCS                              Non-BCS

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


To top