Part One of a four-part, fiscal efficiencies series:
- (Part Two: Basketball Expenditures)
- (Part Three: Fiscal Trends and Measures)
- (Part Four: Revenue, Expenditure, and ROI)
Given resource constraints for programs of all sizes, this article explores the three areas of direct connections between Division I football programs and various outcomes in order to provide athletic administrators and practitioners with information on potential returns in budgeting processes and strategic planning.
1. Take the first known sport-specific examination of financial efficiencies while estimating the potential returns on (a) academic and (b) athletic outcomes.
2. Utilize a multilevel model to estimate not only the school-level impact of football expenditure, but also the interacting effects with conference alignment and BCS membership.
3. Discuss the existence of possible trade-off in terms of resource allocation strategies to inform athletic administrators as they make decisions within the current economic climate.
Overall this study revealed three key findings as it relates to both academic and athletic outcomes in Division I / FBS football programs:
1. Overall athletic expenditures have positive effect on both athletic and academic outcomes.
a. Athletic Outcome: Diminishing return on dollars spent; however, generally increases in athletic budgets will lead to increases in overall on-the-field success (as more fully defined below as season quality metric [SQM]).
b. Academic Outcome: Quadratic function exists within total athletic expenditures. Except for the higher end of athletic expenditure, increased athletic expenditures will lead to increase in annual program rate (APR); however, there exists a point where increasing athletic budgets has a negative impact on the APR scores.
2. Investment in high-quality assistant coaches appears to have the most significant returns on both academic and athletic outcomes.
a. Athletic Outcome: Additional $30,000 of total compensation to assistant coaches is estimated to produce a one-point increase in a team’s SQM—which is equivalent to a non-conference, regular-season win against an unranked opponent.
b. Academic Outcome: Additional $100,000 of total compensation to assistant coaches is estimated to yield a one-point increase in a football team’s APR.
3. Dichotomy exists on the impact of athletic student aid between athletic and academic outcomes.
a. Athletic Outcome: Increases in athletic aid expenditures negatively impact a football team’s annual SQM outcomes.
b. Academic Outcome: Increases in athletic aid do not have any positive effect on the academic outcomes as postulated.
To reach these observations, we begin with content surrounding current Division I football expenditures [table 1]. The mean score provides the average of each expense category. The standard deviation provides useful information regarding the range of the expenditure across schools. Results of the descriptive statistics show wide ranges in the amount expensed within each category. Specifically, within the total operating expenses category, 68.2% of all schools fell within a range of $6.8 and $20.8 million—one standard deviation around the mean.
Table 1 — FBS Football Expenditures—Descriptive Statistics
1. Total Compensation: All Coaches – includes all expenditures (direct or third party) paid to both head and assistant coaches along with severance packages.
2. Direct Student-Athlete Expenditures – includes athletic student aid; equipment, uniforms, and supplies; medical expenses and medical insurance; and recruiting expenses.
3. Facility and Game Expenditures – includes direct facilities, maintenance, and rental; game expenses; guarantees; indirect facility and administrative support; and team travel.
4. Athletic / Administrative Operations – includes fund-raising, marketing, and promotions; membership and dues; support and administrative compensation; and other expenses.
Given the nature of college football and the recent development in data access, this study limited its sample to football programs within the FBS / Division I-A. The exclusion of FCS was purposeful in that comparison across divisions would be difficult given the differences in athletic culture and post-season structure. The use of advanced economic and multilevel models allows for this study to ensure factors associated with both conference alignment and BCS eligibility are controlled for in the estimates. Finally, the use of the two years of data (2009 and 2010) is a function of data access and limitations to the sport-specific data. 1
This uses two distinct outcomes to measure potential returns on investment. The academic outcome is measured through the APR developed and used by the NCAA. The APR metric has far-reaching consequences within the NCAA and is a measure of student-athlete educational gains. The athletic outcome is a derived metric called the SQM, which uses four dimensions of weights in assigning game-level quality outputs. Weights were derived using a predictive linear regression function, which looked at the importance each factor (i.e., opponent quality, game type, championship/ bowl game) played in overall athletic success. Weights were then assigned at the individual game level and aggregated to ascertain an overall quality score.
Results: Potential Returns on Athletic and Academic Outcomes
Overall Athletic Expenditures
On the most simplistic level, while still controlling for conference alignment, BCS membership, and institutional academic quality, increases in football-specific athletic expenditures positively and significantly increase on-the-field (SQM) and academic (APR) outcomes. Specifically, for each additional $100,000 spent on a football team, there is an estimated increase of approximately 0.30 on the team’s APR. Additionally, for the same increase in expenditures, a football team can expect an increase of 0.60 on its SQM—slightly larger than one-half the effect of a non-conference, regular-season win against an unranked opponent. Interestingly, expenditures as they related to the APR had a quadratic function—meaning that there exists a level of spending that once crossed will yield no more positive impact. This function was not seen when analyzing the SQM outcome, although a diminishing return trend was visible.
Given that there exists an overarching impact of athletic expenditures on both academic- and athletic-related outcomes, the following two sections take a more in-depth look at the impact of both categorical and line-item expenditures.
Season Quality Metric:
In order to estimate the potential impact of various levels for expenditures, the multilevel model used an institutional fixed-effect (to control for unobserved factors associated with both the athletic department and institution) along with the Learfield’s Director Cup ranking / points as a measure of athletic department quality and “culture of winning”. As expected, total coaches’ compensation, athletic / administrative expenses, and facility / game expenses have significant impact on a team’s overall SQM. Surprisingly, direct student-athlete expenses did not have a significant impact on a team’s SQM.
Results found that assistant coaches, rather than head coaches, had the most significant impact on a team’s SQM. In particular, every additional $30,000 spent on assistant coaches’ total compensation is estimated to produce a one-point increase in a team’s SQM—which is equivalent to a non-conference, regular-season win against an unranked opponent. Severance payments had a negative impact on the SQM, with every $50,000 in additional severance payments producing an estimated decrease of one point to a team’s SQM.
The only non–coaches’ compensation variables that had a positive significant impact on athletic success were total travel expenses. The model estimated that every additional $40,000 spent on team travel is equivalent to a non-conference, regular season win against an unranked opponent when it comes to SQM. The amount expensed on student athletic aid had a negative impact on a team’s SQM, with every additional $62,500 estimated to decrease a team’s SQM by one point.
Academic Outcome – APR:
In order to estimate the potential impact of various levels for expenditures, the multilevel model used an institutional fixed-effect (to control for unobserved factors associated with both the athletic department and institution) along with the institutional admit rate and the 75th percentile SAT scores on both math and reading. These indicators are measures of academic quality and control for any variations in academic expectations or rigor. Similar to the outcomes on the SQM, assistant coaches’ compensation also had a significant impact on a team’s APR. Specifically, for each additional $100,000 of total compensation to assistant coaches, there is an estimated one-point increase in a football team’s APR score. Head coach compensation, while only slightly significant, had a negative impact, with every $100,000 in additional total compensation leading to an approximate decrease of 0.35 points on the team’s APR score.
In terms of non–compensation-related expenses, expenses on direct facility rental and maintenance had a significant negative impact on a team’s APR score, with each additional $100,000 of expenses estimated to decrease a team’s APR score by 0.10. Football teams that could garner third-party support of administrators and support staff members (non-coaches) saw large gains in the team APR score. For each additional $10,000 of additional third-party support, the model estimated an increase of 1.6 points on the football team’s APR score. Finally, expenses on student athletic aid had no impact on APR score.
While there are limitations to this study, such as the use of two years of data and limiting the sample to FBS / Division I-A schools, it appears that there are significant trade-offs between line-item expenditures and outcomes within football. Overall, it appears that investment in high-quality assistant coaches would have a positive impact on both a team’s athletic and academic success. Additionally, this study did not find the pervasive negative impact of the head coach’s salary on either measure (APR or SQM), except when examining the impact of severance payments.
The discount between the effect of expenses on athletic aid remains a topic of conversations for athletic departments and universities to have moving forward. Given the restrictions of 85 scholarships by the NCAA, athletic aid remains a large and rigid expense category with athletic departments. Finally, a football team’s ability to generate third-party support of both coaches and administrative/support staff produced more efficient results on both the SQM and APR.
The results within this study are essentially the average impacts. Given that each football program has its own nuances and culture, reallocation of the results might have differing results. However, given the large number of the football programs analyzed over two years, the results within this study provide a solid foundation for future discussions on resource allocation during a period of economic recession.
What are other leaders doing, and what are some emerging trends? We are proud to feature submissions from college athletics professionals to promote best practices, and give special thanks to Dennis A. Kramer II (714-514-6442 / email@example.com) for this article. Dennis is the Senior Policy and Educational Statistics Analyst for the Policy Division at the Georgia Department of Education and a Technical Adviser to Georgia’s Leadership Institute for School Improvement. Dennis previously served as the Policy & Research Fellow for the Knight Commission on Intercollegiate Athletics. He is currently pursuing his Ph.D. in Higher Education at the University of Georgia with a focus on the economic and organizational impact intercollegiate athletics has on colleges and universities.
Assistant coaches seem to have more impact on a team’s SQM and APR than do head coaches. Data indicate that a team’s SQM jumps a full point for every additional $30,000 spent on assistant coaches’ total compensation, while its APR increases approximately one point for each additional $100,000.
- In order to efficiently estimate the impact of the certain expenditure patterns on three outcomes, a multilevel mixed regression platform provided the most reliable estimates. Below describes (algebraically) the multilevel mixed model used to estimate the potential returns on outcomes from each category of expenditures. In line with the model presentation of Raudenbush and Bryk (2002) or Goldstein (1995), the multilevel model used within this analysis is as follows: Let yijkt denote the current year outcomes on outcome t for school k in conference j’s class within BCS i. The level 1 model is given by: yijkt = αijk + εijkt (1) where αijk is the school-specific football mean and εijkt are specific error terms for the football program with mean zero and variance σt2 and correlation between the errors for tests t and t’ equal to ρtt’. The level 2 model for football team means is given by: αijk = θIj +β’xijk + ζijk (2) where θIj is the conference mean, xijk is vector of football team-level covariates and ζijk are independent football team level errors with mean zero and variance ν 2. The level 3 model for conference means is given by: θij = πI +φ’cij + γ’zij + ξij (3) where πI is the BCS membership mean, cij are the football-specific financial indicators being tested in set s for school k, in conference j and zij is a vector of athletic or school-related indicators including admission selectivity—for academic outcomes —or overall athletic success—for athletic indicators—for the football-specific program, and ξij is a conference-level residual error term with mean zero and variance vs2. The final model is the school mean model: πI = μ + ηI (4) where μ is the overall mean and ηI is the school random effect with mean zero and variance ω2.: ↩