The Media and Intercollegiate Sports

In Chapter 2 we talked about the Prisoners’ Dilemma and the incentive for cartel members to cheat. Here we apply the concept to scheduling broadcasts of college football bowl games and similar events.

As we observed earlier, networks sell the sports product in an oligopoly market. In your study of microeconomics you learned that oligopoly markets consist of a few firms whose pricing and output decisions are interdependent. There is often an opportunity for these firms to increase profits by colluding in their pricing or division of the market. For the networks broadcasting college sports (ABC/ESPN, CBS/CSTV, FOX, and NBC), this often comes in the form of scheduling contests. Those that fail to coordinate schedules and instead compete head-to-head may be sacrificing potential profits.

Traditionally New Year’s Day has been crowded with college football bowl games, many competing for the same viewers. As one can imagine, those airing the games earned fewer profits than they could have if they were the only networks showing a game at any particular time. In recent years, however, broadcasts of the major bowl games (BCS games primarily) have been staggered to reduce the overlap. Game theory helps us understand and represent the media providers’ incentive to cooperate rather than engage in a bowl game arms race.

Suppose that ABC and FOX are each scheduled to broadcast a major bowl game on New Year’s Day. Assume that no other networks are broadcasting games, and that there are two good time slots available in which to broadcast bowl games, but one is better than the other (we’ll call these “best” and “second best”). Figure 7.1 below represents the payoff matrix faced by ABC and FOX given the available choices of the best and second best time slots.
Figure 7.1 The Payoff Matrix for a Bowl Scheduling Game (amounts in millions)

		FOX
		Best	Second Best
ABC	Best	FOX: $10 ABC: $10	$8 $16
	Second best	$16 $8	$5 $5

The payoff matrix tells us that if they both choose the “best” time slot, each will receive a profit of $10 million. Likewise, if they go head-to-head in the “second best” slot, each will profit $5 million. Alternatively, if they stagger the bowls such that one is broadcasting at the best time and the other at the second best time, the network with the best time earns $16 million, and the other network earns only $8 million.

If they play the game only once, and assuming they both employ a traditional strategy of maximizing the minimum gain (maximin strategy), the networks will both choose the “best” time (this is also a dominant strategy for both). However, if we look at the totals for each cell of the payoff matrix, we see that the combined profit is greater ($24 million v. $20 million) if they stagger the bowl offerings. The outcome of the game is a form of prisoners’ dilemma in that both can be made better off if a different arrangement is made. In this case, ABC and FOX have an incentive to cooperate rather than compete directly, if they can agree on some form of profit-sharing.

As we saw earlier, ABC and FOX both have extended contracts to broadcast major bowl games; in such cases there is an opportunity for a repeated game between the players. Assuming that the payoffs in Fig. 7.1 remain constant over time, ABC and FOX might agree to take the second best slot in alternating years. If they form and stick to this agreement, both will earn $24 million every two years versus the $20 million each would receive if their media providers’ dilemma extended over two years.

Suppose they form the agreement described above, with ABC taking the second best slot in Year 1 of the agreement, but in Year 2 FOX defects and also broadcasts its game in the best slot. By cheating on the agreement, FOX gains $2 million in Year 2 ($10 v. $8 million), but incurs the wrath of ABC. If the game continues, ABC is unlikely to trust FOX again, and we would expect the outcome to revert to the media providers’ dilemma for as long as the game is played. While FOX gained $2 million in the short run by cheating, it will be worse off beginning in Year 3 (they will only earn $10 rather than $16 million) than if it had honored the agreement.

If a game such as this is to be repeated into the foreseeable future, both players will be better off sticking to the agreement. But what if both networks have contracts with definite end dates and low probability of renewal? Suppose ABC and FOX both have four year contracts to broadcast their respective bowl games. They enter the agreement described above, with ABC agreeing to take the second best slot in Years 1 and 3, and FOX in Years 2 and 4. When Year 4 rolls around, what incentive does FOX have to stick to the agreement? Unless ABC has some other mechanism with which to punish FOX for defection, FOX is $2 million better off by cheating, especially if it thinks it can “kiss and make up” with ABC should such collusive opportunities present themselves in the future.

The benefits of programming cooperation can be even greater if staggering agreements also include non-sports broadcasts. Bowl games competing for the New Year’s Day audience are all vying for the same set of viewers. If a media provider not showing a bowl game can offer up a holiday favorite such as The Sound of Music at the same time as the Orange Bowl, for example, both can appeal to large audiences (and the accompanying profits) with less risk of losing viewers to the competition.

Just as college sports generate revenue for the media, the media helps college sports programs generate revenue for themselves. Colleges and the NCAA have become dependent on this revenue to maintain or increase current levels of spending on their sports programs. The benefits of media coverage, however, extend beyond the direct infusion of revenue. Colleges also benefit, or at least believe they do, from the exposure they receive during media-covered events.

Media revenue is the major source of revenue for the NCAA and its member institutions. The NCAA Men’s Basketball Tournament alone contributes 90 percent of the NCAA’s annual income, much of which is generated by media coverage (Baade & Matheson, 2004, p. 112). As we saw in Figure 1.2, the majority of this income is ultimately distributed to member institutions, and as we saw in Table 6.1, Division I-A universities rely directly on media revenue to support approximately seven percent of their budgets. This revenue supports expenditures on coaches, equipment, and transportation. Schools with greater revenue have more to spend on recruiting efforts, including facilities upgrades. Better recruiting attracts higher quality athletes, which often leads to more competitive success. Program success enhances the media attention, encouraging alumni and boosters to “support a winner” by increasing their giving. Despite the logic of this system, as we saw in the last chapter the evidence does not support the universities’ belief that greater spending systematically improves winning or the overall financial position of universities.²

Even if the broadcast revenue distributed to member schools only fuels the arms race, some of the money retained by the NCAA is used for the benefit of student-athletes. In 1990-91, after CBS and the NCAA signed a $1 billion broadcast agreement for March Madness, the NCAA set up three funds to support student athletes. The first fund expanded the Catastrophic-Injury Insurance Program to cover all NCAA athletes. The second established the Special Assistance Fund that is used to help student-athletes facing emergency expenses for things like education, medical care, travel. The third fund, the Academic Enhancement Fund, was created to distribute money to member institutions to enhance academic programs for NCAA student-athletes (Copeland, 2006).

Media coverage of bowl games is critical to generating the millions of dollars paid to the NCAA institutions and their conferences. Corporate sponsors of bowl games reap millions of dollars in benefits from having their logos appear during broadcasts. Research by Image Impact, a sponsorship measurement firm, estimated that Frito-Lay (the producer of Tostitos) received “an extra $30 million worth of exposure during the Fiesta Bowl and other BCS game broadcasts” (Goetzi, 2006). It is no wonder that companies like Frito-Lay are willing to pay millions of dollars to sponsor the bowls that generate the millions of dollars distributed to bowl participants.

Payouts to schools participating in bowl games for the 2006-07 season ranged from $325,000 to $17 million, with all but two at the $750,000 mark or above. Nineteen of the thirty-two bowl games paid at least $1 million, and the five BCS games all paid $17 million to the participating teams (“Bowl Championship Series,” n.d.). A portion of the BCS payouts stays with the teams that compete; the remainder is shared with the teams’ respective conferences. The six major conferences (ACC, SEC, Big 10, Big 12, Big East, and Pac 10) are all guaranteed the largest share of BCS money. Of the $96,160,000 in BCS revenue paid out in 2006, over $89 million went to the six major conferences, with the remaining money (just under $7 million) going to the other Division I-A and I-AA conferences (“Bowl Championship Series,” n.d.). A full listing of BCS payouts appears in Chapter 2. What is clear from the payout information is that the major conferences effectively use their power within the NCAA (a “cartel within the cartel”) to reap “winner-take-all” returns from bowl participation.

The true picture is clouded by the fact that the listed payouts don’t necessarily match what schools receive. In 2006-07, for example, the $17 million per team payouts for BCS games applied only to teams from the six major conferences. Notre Dame received only $4.5 million and Boise State only $9 million. Who gets the best deal? It isn’t apparent from the payout numbers; one must look at how those payouts are distributed. Though Notre Dame only received $4.5 million, as an independent it did not have to share its bowl revenue. Boise State, on the other hand, had to share its $9 million with five non-major conferences (Conference USA, WAC, Mountain West, Sun Belt, and Mid-American), and itself kept only about $3 million. The major conference participants in the BCS shared their $17 million payouts according to conference formulas. After expenses for traveling to bowl games are deducted, conferences such as the Atlantic Coast and Big Ten share bowl revenue equally. The Big East, on the other hand, returns a larger share to BCS-participating schools, and lesser shares to those in minor or no bowl games (O’Toole, 2006).

Payouts for other bowl games are also deceiving. In addition to revenue sharing obligations, officially listed payouts sometimes exceed what schools and their conferences actually receive. In the 2006 Texas Bowl, for example, Kansas State received the published payout of $750,000, while Rutgers received only $500,000. What makes these bowl revenue distribution figures more misleading is that they fail to account for ticket purchase requirements. Each school is required to buy a large block of tickets, and they may choose to give away some tickets to loyal supporters or simply be unable to resell them all. Bowl officials and conferences are free to negotiate payouts, so a team receiving a lesser payment may also be obligated to buy fewer tickets than the opposing school (O’Toole, 2006).

As we saw in Chapter 6, revenue generated from the college football bowl system plays a unique role in athletic department budgets. A portion of bowl revenue, the annual split between conference members, is a regular fixture in budgets. The part earned from actually participating in a bowl game in a given year is spent almost exclusively on attending the event. To understand why colleges do not use this share of bowl money to improve facilities or otherwise enhance their programs, it is important to remember that communities host these bowls to stimulate the local economy. In fact, bowl games were first started to attract tourists to the warm winter destinations where they were played – Southern California, Arizona, Texas, and Florida. The goal was to draw visitors and the dollars that come with them. Participating schools are not only required to purchase a minimum number of tickets, they are expected to spend heavily on local accommodation, food, and tourist attractions for the players, coaches, university personnel, and “friends of the program.” Schools that fail to bring freely spending fans to bowl games are less likely to be invited to future bowls (see Box 7.1). This raises the question, if schools are expected to spend all of their bowl participation revenue at the event, why bother? If bowl revenue was the only consideration, schools might not care, but colleges believe that the exposure they receive still leaves them better off, even if all of the extra bowl money is spent attending the event.

While revenue from media contracts provides substantial budgetary support, it is not the only benefit of the college sports–media relationship. Exposure benefits universities with successful teams in ways that are quite tangible, but sometimes difficult to measure. As we saw in Chapter 6, athletic success may draw the attention of high school students in the midst of their college application and selection process. To the extent it actually exists, the boost in applications from the Flutie Effect provides schools with the opportunity to enhance revenue through enrollment growth or greater selectivity.

Why would the NCAA use the media in this way? Like any cartel or monopoly, the NCAA has reason to fear the government stepping in to regulate operations. In order to quell public calls for government intervention, the NCAA uses public relations advertising to extol the virtues of college sports and its athletes. While allocating resources toward public relations advertising may not maximize short run profits for the NCAA, in the long run it may be cheaper than complying with tougher regulations or devoting additional resources to lobby those legislators willing to consider greater government oversight of college sports.

Has the media tainted the “purity” of college sports? For economists, the bigger questions are: (1) How has media involvement affected college sports; (2) What is the economic rationale behind media-driven changes; and (3) What are the economic impacts of media-driven changes? Here we look at a few changes, first focusing on football, then basketball, that have occurred in college sports that are connected with media involvement.

Scheduling of major college football and basketball games is driven heavily by the demands of television coverage. Subject to the constraint that most college football games will occur on a Saturday, times are juggled to allow networks to show multiple games in a day. In addition, there are Thursday and Friday night games broadcast every week, and occasionally games on Sunday and Monday. For every week of the 2006 college football season, ABC (including its affiliated stations ESPN, ESPN2, ESPNU, and ESPN360) broadcast at least one game each on Thursday and Friday, and multiple games each Saturday, including one weekend (Thursday to Saturday) that included seventeen regular season games. The majority of the off-Saturday games, and the staggered schedules of Saturday games, were engineered by the NCAA primarily to provide media programming.

To the dismay of traditionalists, media programming demands have also resulted in the rescheduling of games with established histories of playing on a certain date. For example, the annual football game between the University of Oregon and Oregon State University, known locally as “the Civil War,” was played on a Saturday for 79 years, often during the weekend after Thanksgiving. In 2006, in order to accommodate a FSN national telecast, the game was moved from Saturday, November 25, to the afternoon of Friday, November 24. Why did Oregon and Oregon State agree to the change? Both teams received an additional $250,000 from FSN, but as OSU’s athletic director Bob De Carolis explained, “I’m not going to say the financial part didn’t have anything to do with it, but that certainly wasn’t the driving force. It was more about getting exposure on a national basis.” (Beseda, 2006).

In response to the media providers’ dilemma described above, bowl games are now distributed across a wider spectrum of dates and times. There were 32 scheduled games for the 2006-07 bowl season, running from December 19, 2006, to January 8, 2007. Historically, on New Year’s Day the schedules for the Rose, Fiesta, Orange, and Sugar Bowls (the four non-championship BCS games) would often overlap. In 2007 there was no overlap, and the four games were spread over prime time slots on January 1^st, 2^nd, and 3^rd. The championship game was played on January 8^th. All except for the Rose Bowl (ABC) were broadcast by FOX.

Critics of building schedules around media programming demands claim that travel to and participation in these games (particularly those on Thursdays) increases absenteeism and further distracts student-athletes from their studies. Economists might support that argument on the grounds that it inhibits human capital formation, diminishing productivity growth. In measured terms, however, consumer demand for additional broadcast games, and satisfaction of that demand by networks and the NCAA, appears to have an overall positive impact on economic welfare. Despite the objections, consumers remain willing to pay for the sports product (e.g. Thursday night games)
7.8.2 Creation and expansion of the BCS

For many years, college football bowl games operated independently, focused on generating economic activity for the host community. As explained earlier, teams accepting invitations to bowl games were required to purchase a block of tickets and spend most, if not all, of their bowl payout in the local economy, often providing lavish accommodations and entertainment for players, coaches, and other university officials. Some bowls had formal arrangements with conferences that agreed to furnish participants. Since 1947, for example, the Rose Bowl had always been played between the Pacific-10 and Big 10 conference champions. Other bowls, like the Sugar and Cotton, were hosted by a particular conference champion (Southeastern and Southwestern, respectively), with the opponents not confined to a particular conference. These bowls would issue invitations to prospective opponents that schools would accept, or decline in favor of a better bowl offer. The lesser-known bowls would have a variety of arrangements, some tying themselves to the second- or third-place team of a major conference, others simply offering invitations geared at creating an appealing contest.

For the media, the old bowl system created tremendous financial uncertainty. Rights fees were negotiated and broadcast schedules were set well in advance of knowing a given bowl’s participants. Networks fortunate enough to have secured a bowl game with national championship implications or some other intriguing match-up did well. Less appealing games not only lost networks advertising revenue (from the projected lack of viewers), but dedicating three hours or more to a bowl game meant three less hours available for potentially more profitable sports or non-sports programming.

As we saw in Chapter 1, in an effort to create a more stable financial climate, in 1991 the Atlantic Coast (ACC), Big East, Big Eight (now the Big Twelve), Southeastern (SEC) and the Southwestern conferences, in conjunction with Notre Dame, formed a bowl coalition with the IBM (now Tostitos) Fiesta Bowl, the Mobil (now AT&T) Cotton Bowl, the Federal Express (FedEx) Orange Bowl, and the USF&G (then Nokia, now Allstate) Sugar Bowl. The purpose was of course to ensure that these games would attract the maximum number of viewers, as all stood to gain from the greater revenue and exposure.

The problem with the new coalition was that bowl organizers were initially reluctant to give up historic conference ties. Affiliated conference champions were still required to host their respective bowls, and it failed to correct the situation where the top ranked teams vying for a national championship might never play each other. Uncertainty of match-ups remained, leaving advertisers hesitant to spend as freely as they would for a game certain to attract a large audience.

The system was revised when the Bowl Alliance was created in 1994 between the ACC, Big East, Big 12 (the Big Eight plus four teams from the Southwest conference now merged), SEC, and Notre Dame, and the Orange, Sugar, and Fiesta Bowls. Conferences were no longer tied to a particular bowl game, so the alliance was free to create match-ups between the top-ranked teams, with the top game rotating amongst the three bowls. Each of the four conference champions plus Notre Dame (assuming a winning record) would be featured in these bowl games. Remaining slots could be filled either from within or outside the alliance.

One of the goals of the Bowl Alliance was to create a national championship game. Prior to creation of the BCS, the “mythical” national champion was determined by the sports writers’ and coaches’ polls. Each group would vote to determine national rankings, sometimes arriving at different conclusions. For example, in 1991, the University of Washington and the University of Miami both finished the season undefeated. Washington soundly defeated the University of Michigan in the Rose Bowl, while Miami trounced the University of Nebraska in the Orange Bowl. The result was a split national championship, with the writers selecting the Miami Hurricanes #1 and the coaches crowning the UW Huskies as national champions.

While the Bowl Alliance reduced the chances for a split national championship, the possibility would remain as long as the Pac-10 and Big 10 remained tied to the Rose Bowl. In 1997 they joined the Bowl Alliance and formed the Bowl Championship Series (BCS), creating a national title game that would rotate between the four bowls. The six conference champions would fill six of the eight BCS slots. Participants in the championship game are the top two teams in the BCS rankings, as determined by the Associated Press (media) poll, the USA Today/ESPN coaches’ poll, and the average of a computer ranking system. The computer rankings factor in strength of schedule (including both opponents and opponents’ opponents), losses, and quality wins (bonus points for beating a team ranked 15^th or above).

From 1997 through the 2005-06 bowl season, the BCS remained largely intact. The rating system changed over the decade, with additions and deletions of various computer rating systems. The main controversy during the time, besides the circumstances of specific years, was that the BCS allowed few opportunities for non-BCS conference teams to participate. To improve access and generate additional revenue, another BCS game was added beginning with the 2006-07 bowl season. The new fifth game is a national championship game separate from the four original bowls, but held on a rotating basis in the cities hosting the four bowls. For the 2006-07 BCS season, for example, the Rose and Fiesta Bowls were held on January 1^st, the Orange Bowl on January 2^nd, the Sugar Bowl on January 3^rd, and the BCS Title Game was held on January 8^th in Arizona (host of the Fiesta Bowl a week earlier).

Addition of a national championship game has not eliminated controversy, as we will see later in the chapter. It has increased the probability, however, that the top two teams will meet in a final game to determine a slightly less mythical national champion.