Research Paper: Topic Selection/Annotated Bibliography Assignment

4 Sport and gambling David Forrest

The economics of sport and the economics of gambling are distinct subdisciplines; but they overlap because the activities they analyse are so often bound up with each other. Indeed it may fairly be said that some professional sports would not be viable at all but for the betting markets that underpin them. Horse racing in many jurisdictions, and jai alai in certain states of America, are examples of sports where the bulk of revenue has always been extracted from the wagering markets that run alongside them. Thus, in early twentieth-century America, horse racing had to shut down altogether in states that enacted a prohibition on bookmaking at the track (Munting, 1996) while jai alai today is played only in a few states and only where betting is legal and the principal reason for attendance at the event (Skiena, 2001).

Of course, most sports are not as dependent on betting as these examples. Nevertheless, the existence of high-volume betting markets on sports such as American football, soccer and cricket matters, or should matter, to sports economists. In this review, three broad reasons for sports economists to think seriously about betting on sport are considered.

First, sports leagues and promoters generate a positive externality for bookmakers by providing extra and popular vehicles for betting. Given ownership of copyright over fixtures, sports may be able to charge for the provision of this externality, much as they extract payment from broadcasters when they provide them with opportunities for extra and popular television programmes. Many of the same issues arise as in the case of broad- casting, for example, the betting sector may be willing to pay more to a sports league if changes in game design are made to accommodate its interests.

Second, while benefiting sports by stimulating interest and providing an additional source of revenue, gambling may pose a threat by introducing incentives for corruption. Match fixing associated with betting is a threat from a commercial, as well as a moral, perspective because stadium and television demand will fall if events on the field lose credibility and sponsorship income will be threatened if firms fear being tainted by asso- ciation with cheating. Economic analysis can offer insights on how incentives to cheat for betting gain can be minimised and it can be used to illuminate public policy issues such as whether corruption will be more severe if sports betting takes place as a legal or illegal activity.

A third reason for thinking about gambling is that the existence of wagering markets can potentially serve the interests not only of sports but also the work of sports econo- mists. If odds can be assumed ‘efficient’ in the sense of providing unbiased probabilistic forecasts of sporting outcomes, they can be exploited, for example, for constructing ex ante measures of match-level outcome uncertainty. Studies of movements in odds could be similarly informative, for example by generating market-based assessments of the importance to a team of particular star players who become unavailable for particu- lar games.

40

C o p y r i g h t 2 0 0 6 . E d w a r d E l g a r P u b l i s h i n g .

A l l r i g h t s r e s e r v e d . M a y n o t b e r e p r o d u c e d i n a n y f o r m w i t h o u t p e r m i s s i o n f r o m t h e p u b l i s h e r , e x c e p t f a i r u s e s p e r m i t t e d u n d e r U . S . o r a p p l i c a b l e c o p y r i g h t l a w .

EBSCO Publishing : eBook Collection (EBSCOhost) – printed on 10/29/2024 9:08 AM via LIBERTY UNIVERSITY AN: 182123 ; Wladimir Andreff, Stefan Szymanski.; Handbook on the Economics of Sport Account: liberty.main.ehost

Potential Benefits to Sport from Gambling Gambling on sports events has, of course, a history going back centuries, indeed millen- nia. For example, Sauer (1998) quotes a contemporary description of crowds en route to the Circus Maximus in Ancient Rome as ‘already in a fury of anxiety about their bets’; and historians such as Munting (1996) document a close association between gambling and sports such as cricket and baseball in eighteenth- and nineteenth-century Britain and America. But for much of the twentieth century, demand for betting, and gambling gen- erally, was suppressed in many countries by prohibition. Until recently, for instance, betting on team sports was not possible legally in Hong Kong or in any American state other than Nevada, though horse racing flourished in many areas thanks to the tolerance of on-track betting.

The world, however, now appears to be in a phase of secularly increasing legal toler- ation of all manner of gaming activities. In America, the first state lottery game did not appear until 1964 and only Nevada permitted casinos prior to 1976. But the ‘lotto mania’ of the 1980s and 1990s, and the growth of Indian and riverboat casinos in the 1990s, gave a large majority of Americans access to these gambling media by the turn of the century. This liberalisation was bad news for those sports – horse and dog racing and jai alai – that had enjoyed protection because their product was bundled with the only legal betting opportunities available to the public. Thoroughbred racing declined in most states. Jai alai folded altogether in Connecticut after the opening of the world’s biggest casino. Recovery was achieved only where racetracks were permitted to turn themselves into ‘racinos’, with casino-style gaming machines to complement the sports offer. The same policy saved jai alai in Rhode Island. The spread of lotteries and casinos to many jurisdictions around the world had similar consequences for horse racing, and sometimes the same policy response, as in the United States.

Sports betting (a term commonly defined to exclude horse and dog wagering) has, on the other hand, flourished in most of the world in the new liberal environment for gam- bling, further boosted by new opportunities to place internet bets with offshore book- makers. Several jurisdictions (such as Singapore and Hong Kong) have introduced soccer wagering to curb the flow of bettors’ funds to offshore centres (or to the illegal sector). In the United States, however, internet betting remains illegal and active attempts are made to enforce the prohibition by barring credit card companies from settling payments to betting operators in countries, such as Costa Rica, that serve the American market. That the offshore sector remains intact – indeed Costa Rica had the highest sports betting turnover of any jurisdiction in the world in 2003 (Global Betting and Gaming Consultants, 2005) – is testimony both to the difficulty of controlling use of the internet and to the fact that Americans share the worldwide growth in enthusiasm for betting on sports events despite its illegality in nearly all states.

In 1998 the total staked (in legal markets, but excluding lottery-type games based on sports results) on sports worldwide is estimated to have been approximately US$36 billion. By 2003 this had grown by over half again, to just in excess of US$56 billion (Global Betting and Gaming Consultants, 2005). In sports betting, between 5 and 10 per cent of stakes is typically retained by the bookmaker, so that net turnover will have been at least US$6 billion.

This large industry is, of course, dependent on major leagues in America and Europe providing the high-profile sports product on which clients wish to bet. These sports

Sport and gambling 41

EBSCOhost – printed on 10/29/2024 9:08 AM via LIBERTY UNIVERSITY. All use subject to https://www.ebsco.com/terms-of-use

leagues, and particularly the English Premier Football League whose games are televised throughout Asia, are therefore conferring substantial benefit on another industry, betting, that produces a complementary good. Their ability to capture any profits generated depends on whether they have legal and de facto control of relevant property rights, in this case fixture lists used by the gambling industry. Potential revenue from control of such rights is large. Football industry sources suggest that Far Eastern betting turnover (legal and illegal) amounts to US$1 billion per weekend in season, with half of this staked on English Premier League games. Even a small levy on these bets would generate very significant revenue for the League.

The extension of World Trade Organization agreements to the area of intellectual prop- erty rights has strengthened the legal position of sports leagues; and indeed fees for the use of fixture lists have been negotiated successfully with, for example, the football-based lottery in China. But immediate prospects for significant copyright revenue are poor because betting is usually channelled through either an indigenous illegal sector or inter- national operators whose clients are footloose and very sensitive to charges.

In the longer term it appears plausible that bookmaking will be regularised in the important Asian markets. Singapore Pools made the first legal sports book in the region when it introduced betting on local soccer in 1998. The Hong Kong Jockey Club was per- mitted to offer football betting on English games from 2003. Both moves were in response to a substantial leakage of expenditure to the illegal sector and to overseas bookmakers (including, in the case of Hong Kong, to casino-based operations in Macau). Both met with some success. Increasing internet penetration is likely to raise pressure on other coun- tries in the region to provide for legal betting that keeps some of the turnover at home. The English Premier League has invested heavily in promoting fan interest in the region, with the payoff to date mainly in the form of broadcasting and merchandise income, but it would be well placed, in the event of the gradual establishment of legal betting, and pro- viding that countries adhere to commitments on intellectual property rights, to exploit gambling as a further source of revenue. Competition would be likely to come from the Premier leagues of Germany and Italy which are also popular betting vehicles (European football’s other major league, that of Spain, is less popular because matches start at an unfriendly time for viewers and bettors in Asia).

In the home markets of European leagues, it is already feasible to collect copyright fees. In England, football claims a percentage of expenditure on the lottery-style football pools and a turnover-related fee for the use of fixture lists by licensed betting offices. But, notwithstanding that ‘football betting is the fastest-growing form of gambling in the UK’ (Mintel, 2001), fees are constrained by the increasing inability of bookmakers to pass on increased costs in the face of international competition. Indeed, international competi- tion forced a substantial reform of, and reduction in, betting taxation in 2001. In other words, rents in bookmaking have been dissipated and there is less for the football indus- try (as well as less for the government) to capture.

In this context, a feasible route for football to take is to create new betting products over which football or its franchisees would have monopoly power. Interactive television offers such opportunities. In Britain it is now possible and legal (subject to licence) for viewers of a football game to be offered betting opportunities as the game progresses, for example on whether a penalty just awarded will be converted. Selling such instant bets to a large audience is possible only for the betting firm which is in partnership with the broadcaster

42 Handbook on the economics of sport

EBSCOhost – printed on 10/29/2024 9:08 AM via LIBERTY UNIVERSITY. All use subject to https://www.ebsco.com/terms-of-use

with rights to the game. Any monopoly profits that such bets offer will be factored into bids for broadcast rights and therefore captured by the League. Similarly, bets could be offered on large screens in the stadium during the game, with supporters responding by text messaging. Again, the value of franchise rights to such an operation would reflect monopoly power. The popularity of such new styles of bet remains to be demonstrated but may depend on the extent to which the League is willing to be flexible in game design. For example, could the game be paused to facilitate betting on the penalty kick or could one even alter rules to create more penalties in a match? Similar issues have of course arisen as sport has become increasingly dependent on television revenue. It is relevant to note that the American sports that have been most ready to adapt game design to the needs of broadcasters have achieved the highest rates of growth in broadcasting rights income (Szymanski, 2003).

Possible Costs to Sport from Gambling The principal source of concern here is that the prospect of betting gain, or bribery of players or officials by bettors or bookmakers, will corrupt a sport and cause match fixing. As with any illegal activity, it is hard to assess just how common the phenomenon is. A number of authors, such as Crafts (1985) for British horse racing and Gandar et al. (1998) for American basketball have demonstrated that movements in odds during the betting period are good predictors of event outcomes. This is consistent with corruption since the weight of money of those involved in any fix would drive the odds inwards from opening values compiled by bookmakers on the basis of fundamentals such as team form and quality. However, such movements could also result from the existence of any type of insider information (such as knowledge of player injuries) or indeed are consistent with a rational expectations perspective whereby the collective view of large numbers of bettors is liable to be superior to that of any individual, no matter how skilled an odds-setter he or she might be. This body of work cannot therefore be interpreted as evidence of wide- spread corruption though movements of odds might be admissible as a further indicator of the need for inquiry where unusual events are observed on the field.

More convincing evidence of widespread corruption in the particular case of US college sports is provided by Wolfers (2002). For college basketball, the betting market was not odds based. Rather, bets on all teams were at the same odds and related to whether or not a team would beat ‘the spread’ announced for its game by the bookmaker. The starting point for the analysis is that a team’s performance relative to the spread is a fore- cast error and the distribution of such errors should be approximately normal. In fact, across a sample of close to 30 000 matches, there was a statistically significant deviation from normality such that, in games whose outcome was relatively close to the spread, heavy favourites were much more likely to fail to beat, than to beat, the spread. This could arise because teams winning comfortably relaxed their effort; but it could also be explained by some teams actively choosing to win by less than they could. Athletes on such teams were on the winning side on the court and at the same time profited if they had bet against their own team in the spread market. This practice is termed ‘point shaving’ and has been alleged in several betting scandals at US colleges. Wolfers’s data raise the suspicion that about 10 per cent of teams engaged in point shaving. Concern over such corruption has now led to Las Vegas sports books withdrawing from most betting on amateur sport, though wagers remain possible in the illegal market.

Sport and gambling 43

EBSCOhost – printed on 10/29/2024 9:08 AM via LIBERTY UNIVERSITY. All use subject to https://www.ebsco.com/terms-of-use

Professional sport has also generated fixing scandals throughout its history. Apart from perennial accusations concerning the integrity of horse racing, the most notorious affairs include the attempt to fix baseball’s World Series in 1919 and the arrangement of match results by players (paid by bookmakers from the illegal sector) in the top level of inter- national cricket in the 1990s.

The contemporary cricket scandal (in a sport also notorious for corruption nearly 200 years ago) was used by Preston and Szymanski (2003) to illustrate a model designed to show the circumstances in which players will accept bribes from illegal bookmakers. Corruption was demonstrated to be more likely where there was (a) a large underground betting market, (b) a low detection rate for cheating, (c) low player wages and (d) low prize money or prestige for winning individual contests. All these conditions were present in international one-day cricket, according to Preston and Szymanski.

Forrest and Simmons (2003) employed a different model of corruption that neverthe- less yielded broadly similar conclusions. Adapting the general economic theory of crim- inal behaviour proposed by Ehrlich (1996), they represented the risk-neutral athlete as comparing the gain in wealth from a successful and undetected fix with the expected cost of cheating. The latter comprised the expected value of financial loss (equal to the prob- ability of detection multiplied by the financial penalty for fixing, which is likely to include loss of income while suspended) and the disutility of cheating itself (moral discomfort and loss of sporting glory through underachievement). Conclusions included that athletes are more likely to agree to a fix, (a) the lower the chance of detection, (b) the lower the player wage level and (c) the less the loss of sporting glory when they deliberately under- perform. These propositions, very similar to those of Preston and Szymanski, are in one sense obvious; but, as often in economic theory, a model yields a taxonomy of policy approaches around which debate may be organised.

First, consider the probability of detection. This will differ across sports for technical reasons and because sports governing bodies will invest in screening to an extent com- mensurate with how important (financially) the result of the contest is: contests where the marginal gain from winning is high are likely to be screened more rigorously, so chances of detection will be higher in rich, high-profile sports. However, the probability of detec- tion will vary also according to how well policed the betting market is and this is the basis for the argument that it is in the public interest for betting to take place in a legal rather than an illegal setting.

The generally prohibitive stance of American law has spawned an illegal betting indus- try with turnover a hundred times larger than that of the legal sports books in Las Vegas (Strumpf, 2003). Advocates of liberalisation argue that if this betting took place within a legal industry, available nationwide, betting irregularities would be more readily drawn to the attention of the authorities, thus deterring fixes. There is some force to this argument. Point shaving cases on occasion came to light because they aroused the suspicion of Las Vegas sports books. If an investigation is called for, legal bookmakers maintain records that are likely to reveal whether insiders have been betting: Nevada state law requires all bets over $3000 to be recorded and identification obtained for any bettor staking more than $10 000 in 24 hours. On the other hand, well-organised betting rings are likely to be able to distance the placing of a bet from the identity of the athletes arranging the fix. This is not so in illegal markets. Strumpf examined the activities of a range of illegal oper- ators in New Jersey. Each bookmaker focused on a small geographical area and served

44 Handbook on the economics of sport

EBSCOhost – printed on 10/29/2024 9:08 AM via LIBERTY UNIVERSITY. All use subject to https://www.ebsco.com/terms-of-use

clients with whom he had a long-term relationship. This is the most viable form of organ- isation in an industry where either buyer or seller may build up substantial liability for payment to the other but where these debts cannot be enforced in the courts. As a result, illegal bookmaking firms know their clients well, know who their contacts are and can recognise atypical betting behaviour at the micro level. Consequently, illegal bookmaking is not likely to be an easy environment in which to place substantial bets associated with a fix. Offshore internet betting firms are a greater danger to the extent that betting is more impersonal there, which makes it unlikely that suspicion will be raised because of the source of the wager. In the absence of a legal and only lightly taxed domestic betting industry, internet substitutes are likely to develop and will be unpoliced outlets for bets associated with match fixing.

The second determinant of the proportion of players who would agree to a fix at a given level of financial gain is the level of player wages. In most sports, the penalty for a player proven to have fixed a match is a lengthy, or lifetime, suspension. Plainly the result- ing loss of earnings will differ markedly over time and across sports and casual evidence suggests that this is a principal driver of how much betting-related corruption occurs in different periods and different settings. It is striking that major league US sports and top European soccer leagues have been almost devoid of betting scandal in the era of free agency and high player wages, in contrast to their earlier history; that contemporary pro- fessional cricket, where players have not won the same share of revenue as in other major team sports (Preston and Szymanski, 2000), remains notorious for its corruption; and that US college sports, where athletes are paid nothing at all, has generated the widest concern of all.

Cricket merits further discussion since it is unique among the big international team sports in that there is firm evidence of recent and very widespread corruption at the very top of the game. While not played in as many countries as soccer, cricket never- theless commands a large international following (one day’s play in an India–Pakistan match in 2004 had a worldwide television audience estimated at 600 million). Why then are its players so poorly paid relative to comparable sports? One factor is that, unlike soccer and US sports, the bulk of the game’s revenue is generated by matches and series between national teams rather than by domestic competitions (which attract only weak public interest). As a result, national teams now play nearly constantly around the world, with established players seldom appearing in domestic league matches. Players entering this elite circuit are hired into a monopsonistic labour market because, with few exceptions, an individual player is qualified to represent only one country. This arrangement works to prevent players capturing the same proportion of broadcasting and other revenue as in sports with more competitive labour markets. Cricket govern- ing bodies are thereby able to maintain control of economic rents which are then typ- ically channelled into subsidising otherwise unviable domestic leagues. Major reform of this structure is likely to be necessary before rewards to players become commensurate with the audiences that top-level international cricket can command. The risk of losing wages of that order of magnitude would be likely to deter most players from associa- tion with bookmakers.

Even in the highest-paid sports, however, some support from criminal law enforcement is likely to be needed to eliminate corruption. Players near the end of their career and ref- erees, who do not command the same rewards as players, may risk little financially by

Sport and gambling 45

EBSCOhost – printed on 10/29/2024 9:08 AM via LIBERTY UNIVERSITY. All use subject to https://www.ebsco.com/terms-of-use

trying to manipulate match outcomes to secure betting gain. High probability of detec- tion is an effective deterrent to criminal malpractice in all fields and would be more likely in this case if dedicated police units were charged with enforcement (DCMS, 2001). Again, enforcement will be difficult if there is a large and liquid offshore sector through which bets can be channelled.

The final proposition in the Forrest–Simmons model is that the amount of corruption will be greater where underperformance on the field does not carry much loss of utility for the player. This appears to have been true in the cricket case: Preston and Szymanski (2003) made the point that the number of one-day internationals became so large that the result of any one became unimportant to players, thus fostering a willingness to match fix. Again the idea is consistent with the apparently high incidence of point shaving in US college sports. Because handicap betting was offered, groups of players on a strong team could conspire to score fewer points than they could have, win bets as a result, and still enjoy being on the winning side. Handicap betting appears therefore to carry more risk of corruption than odds-based betting systems.

Some other styles of betting have also been criticised as encouraging corruption. ‘Proposition bets’ relate not to the result of a game but to some aspect, for example the identity of the first player to score a touchdown, or whether a particular batsman will reach 50, or how many players are sent off in a soccer game. The controversy over this style of bet arises because some propositions are much easier to manipulate than the match result. Index betting firms in the UK had to withdraw a popular proposition bet relating to the time at which the (soccer) ball was first put out of play because it was alleged that players regularly kicked the ball into touch just after the game started simply to win a bet. This was not only easy to fix but was acceptable to players since it was vir- tually irrelevant to which team would win the match. The withdrawal of this particular proposition bet perhaps indicates that the betting industry will be self-regulating in what is permitted. Las Vegas sports books have low ceilings on the size of wager in the case of proposition bets.

A fast-growing subsector of the gambling industry is that of the betting exchange. Gaming companies gain their profit by (internet) matching of traders who indicate the odds at which they would be willing to bet either that an event will happen or that it will not. This enables anyone to play the role of a bookmaker by bidding as a seller (agreeing to pay another party if the named event occurs and to collect from that party if the event fails to occur). The rapid growth of betting exchanges has been of much concern to the horse-racing industry because corrupt jockeys can now effectively wager that the horse they control will lose. It was always possible to lose a race in the hope of betting gain but the gain would follow only if a bet was placed on the correct rival horse. Corruption there- fore now carries higher expected gain.

Established bookmakers argue for the banning of betting exchanges (to which they have lost significant market share) because they corrupt the sport. The betting exchanges counter that a ban would drive the exchange concept offshore and deprive the authorities access to betting records that may be (indeed already have been) used to uncover corrup- tion. The debate is familiar and symmetric with issues discussed in respect of other financial products than betting. If the regulator prohibits, or imposes heavy taxes, on a particular product, the money may move offshore where the authorities will find it hard to protect the public interest and the interests of investors.

46 Handbook on the economics of sport

EBSCOhost – printed