The role of conference externalities and other factors in determining the annual recruiting rankings of football bowl subdivision (FBS) teams

ABSTRACT

In this article, we analyse the impact of school, head coach and conference characteristics on a college football team’s annual recruiting ranking. Utilizing panel data collected from various sources covering 2002–2014, we find that measures of recent school success such as having winning seasons and finishing seasons ranked in the Associated Press (AP) top 25 poll have a positive impact on a team’s recruiting ranking. Similarly, schools with more successful head coaches tend to earn better recruiting classes, while schools facing bowl bans, scholarship restrictions and probation tend to earn worse recruiting classes. Various measures of conference achievement indicate that conference externalities in recruiting may indeed be positive as is often suggested; however, there is much potential for a negative externality as well.

KEYWORDS:

• Externalities
• recruiting rankings
• NCAA
• college football
• conferences

JEL CLASSIFICATION:

• L83
• Z20
• Z22
• Z29

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ BCS conferences include the Atlantic Coast Conference (ACC), American (2013 only), Big East, Big 12, Big Ten, Pacific-12 (Pac-12) and Southeastern Conference (SEC). Under the BCS system, the conference champions from these conferences were guaranteed a slot in a high-profile BCS bowl. Teams from other FBS conferences could qualify for a BCS bowl, but were not guaranteed an entrant.

² FBS conferences include the aforementioned BCS conferences as well as the Mid-American, Conference USA, Mountain West and Sun Belt conferences.

³ This article titled ‘ACC scores recruiting wins after Florida State’s national championship run’ is available at http://www.ncaa.com/news/football/article/2014-02-06/acc-scores-recruiting-wins-after-florida-states-national. Last accessed on 22 September 2014.

⁴ During the BCS era, the Orange, Rose, Fiesta and Sugar bowls were considered BCS bowl games.

⁵ The head coach variables are measured for the coach’s entire career rather than over the previous five seasons to avoid multicollinearity issues with the school variables.

⁶ Only six teams in our dataset changed conferences within 5 years of winning a BCS bowl game, and only one team changed conferences within 5 years of winning a national championship. No teams changed conferences within 5 years of having a player win the Heisman trophy. Fourteen teams changed conferences within 5 years of having a player selected in the first round of the NFL draft, and 20 teams changed conferences within 5 years of finishing in the Top 25 of the AP Poll. Thus, regardless of how the conference achievement variables are measured, we are confident that while there may be a tendency to slightly overstate or understate the magnitude of their regression coefficients, the sign and statistical significance of those coefficients would almost assuredly be the same in the absence of any conference realignment. This is particularly true for the variables CONF NAT CHAMPS, CONF BCS and CONF HEISMAN.

⁷ We also estimated our models using a standardized version of ADJUSTED RECRUITING POINTS as the dependent variable. The findings in those models are nearly identical to those reported in the article.

⁸ One potential caveat of our results is that certain teams’ recruiting rankings may be inflated. For example, if a recruit is initially rated as a one- or two-star recruit by Rivals but then receives a scholarship offer from a premier programme such as Ohio State, some suggest that Rivals may assign the recruit a higher rating. While this probably does not change the overall order of the recruit rankings drastically and, thus, is likely not a major concern for models employing RECRUITING RANKING as the dependent variable, it may be a concern for those models employing ADJUSTED RECRUITING POINTS as the dependent variable. For our part, we can only acknowledge this limitation of our study.

⁹ Theoretically, one might expect diminishing returns to winning percentage. While recruits certainly value a winning programme, there is likely some point at which each additional increase in winning percent means less and less to recruits. For example, an increase in winning percent from 60% to 70% may attract more recruits than an increase in winning percent from 70% to 80%. Furthermore, schools with extremely high winning percentages over the previous five seasons may have little playing time to offer incoming freshmen. A graph of RECRUITING RANKING versus WIN 5, available upon request, reveals a potential non-linear relationship, which is supported by the empirical results.

¹⁰ Similar to winning percentage, there are theoretical reasons to expect diminishing returns to increased revenues. For example, if having state-of-the-art facilities requires a certain amount of revenues, then any team with that amount of revenues will likely possess state-of-the-art facilities. However, teams with revenues in excess of that amount may have various additional amenities that attract recruits, but probably not to the extent of simply having facilities that could be classified as state of the art. A graph of RECRUITING RANKING versus REVENUES, available upon request, reveals a potential non-linear relationship, which is supported by the empirical results.

¹¹ Unfortunately, this variable is excluded from the fixed effects models due to multicollinearity between the variable and the school fixed effects.

¹² These 10 schools include Colorado, Mississippi State, California, Tennessee, South Carolina, Alabama, North Carolina, USC, Ohio State and Michigan.

¹³ Oklahoma was punished for recruiting violations in 2007, LSU was punished in 2011 and Oregon and Miami were both punished in 2013. These violations are not detected by the methodology presented here.