Economics and Sport

Marc Lavoie. Handbook of Sports Studies. Editor: Jay Coakley & Eric Dunning. 2000. Sage Publishing.

The economics of sport were given their first credentials when, 40 years ago, S. Rottenberg (1956) published his ‘trailblazing’ article on the economics of the baseball labor market, in a leading journal of economics, the Journal of Political Economy, edited by members of the Department of Economics at the University of Chicago. The department is well known for believing that the (neoclassical) principles of economics can be applied to all subjects and that free markets have the ability to achieve desirable outcomes, a somewhat ironic belief given the geographical location of the university, surrounded by run-down residential black ghettos. At the time Rottenberg first tackled the economics of professional baseball, the future Nobel Prize recipient, Gary Becker, was dealing with the economics of discrimination, the economics of marriage and the economics of the household.

From an historical viewpoint, one could thus see the development of the economics of sport as part of the growing imperialism of economics over other social sciences. This has been accompanied by the ever-increasing media attention over financial issues involving professional sport teams or athletic stars, their profits or their earnings, both in North America and in Europe. The players’ strike in baseball and owners’ lock-out in hockey also hyped up the media’s attention towards the economics of sport.

An attempt to give a broad picture of the economics of sport quickly forces one to realize that two traditions have emerged in the field. One may say that there is a North American, or Anglo-Saxon tradition, and a Continental European tradition. The Anglo-Saxon tradition generally has a micro-economic focus, centering on the individual and the firm, applying the standard supply and demand tools of micro-economics in an attempt to formally model the behavior of the various participants in the world of sport. North Americans, and their Anglo-Saxon colleagues from Britain and Australia, usually focus their attention on professional sport, more specifically men’s team sports. They use, whenever they can, the theories of statistics applied to economics, that is, econometrics and regression analysis. On the other hand, economists from Continental Europe make little use of econometrics, relying rather on descriptive statistics, with tables of numbers and the computation of various ratios. Their work is usually more of an institutionalist sort, that is, more descriptive, sometimes relying on more unorthodox economic theories. Although they are also concerned with the implications of professional teams, Continental Europeans are much concerned with the economics of amateur and recreational sport, and with the economics of the sports industry.

The present chapter will attempt to cover both of these traditions, but due to the sheer number of publications within the North American approach, a disproportionate amount of space is devoted to the Anglo-Saxon tradition. This chapter will be divided into two broad sections, both mainly devoted to professional team sports. We shall first deal with labor economics, while in the second section we shall deal with the professional club seen as a firm. The latter section will end with a theme covered by both traditions in the economics of sport: the development impact of major sporting events on a city or a region. We shall conclude with a brief assessment of the future outlook of the discipline.

Labor Economics

Salary Determination

Many, if not most, economists attracted to the economics of sport are labor economists. The main fascination with the sports industry is that there are reliable measures of output and productivity per individual, something which is usually lacking in other industries. Although it is possible to obtain huge data matrices about the characteristics of individual employees, it is usually very difficult to have adequate and direct measures of their productivity. In sport, the reliable measures of individual output are, of course, the individual performance statistics, about which sports fans are so fascinated. There used to be one drawback, however: whereas in other industries it was often possible to have the salaries of individual employees, in sport it was very difficult to obtain a sufficiently large set of salary data, and their reliability could always be questioned. The situation has changed considerably since about 1980. First, due to legal court actions, salary data were collected and made available (Fort 1992); secondly, players’ associations have gradually come to realize that salary secrecy was only to the advantage of the employer, and this has led to the media gaining access—officially or unofficially—to the salaries of all players on major league teams, with entire lists of them being published at once. In fact this explains why very little empirical work on this topic—in contrast to attendance determination—has been performed in Europe: besides stars, the salaries of professional or semi-professional players there, whatever the sport, are highly confidential.

There have been two approaches to the issue of salary determination in professional team sports. The first approach, closer to micro-economic theory per se, attempts to verify whether players are exploited by their employers, or it attempts to estimate which sort of player is being most exploited—star players or journeymen, those with or without bargaining power. This normative approach is based on the neoclassical concept of the marginal revenue product. The idea is that a worker should be paid according to the additional revenue that his work generates. The degree of exploitation is then defined as the ratio of the actual salary of the player to his marginal revenue product. This approach was first put into operation by Scully (1974a) for baseball. Three steps are involved. One must first estimate by how much an additional win increases team revenues, having taken into account other factors such as metropolitan population. Then one must estimate how improved athletic performance—in baseball, better pitching and better batting—increases the likelihood of a win. From the individual performance statistics, one can then proceed to the third step, estimating the marginal revenue products of each individual player and comparing them with the actual salaries. With this first approach, a very restrictive set of performance variables is usually chosen, often one per sort of player.

Things are different with the second approach, more positive than normative, which purports to explain how salaries are actually determined. In this second approach, salaries are explained by the location characteristics of the team, the characteristics of players and their performance of the previous year. The set of variables considered as determinants of salaries is then enlarged, despite the problems this might create because of collinearity, that is, because a performance measure may be highly correlated with another, thus making it impossible to properly identify the contribution of each. In a brilliant article, Pascal and Rapping (1972) were the first to proceed with this approach, again applied to baseball. As is to be expected, it was shown that salaries depend on various measures of performance and on experience. Despite some claims to the contrary, the more recent studies on salary determination have shown only marginal additional ingenuity, repeatedly coming down to similar findings.3 The real improvement is in the extent and reliability of the salary data set. Most measures of performance in these salary regressions rely on offensive statistics, such as slugging or on-base average for baseball, or points per game in basketball and in hockey. American football, where individual prowess is more elusive, has been less subjected to salary analysis (but see Ahlburg and Dworkin, 1991 and Kahn, 1992). There has been some reluctance to use and also some difficulty in finding statistically significant defensive statistics, but some regressions do find a role for these lesser-known performance measures: in basketball (Koch and Vander Hill, 1988) and in hockey (Lavoie and Grenier, 1992), the former with blocks and saves, and the latter with an indicator of short-handed play, and even in baseball, when considering golden gloves recipients of awards for defensive play (Johnson, 1992).

Coming back to the approach based upon marginal revenue product, it should be pointed out that salary regressions based on this approach have been used in the arbitration hearings between the Major League Baseball Players Association and the Major Leagues, when the union filed a grievance against the owners, accusing them of collusion in refusing to compete for the hiring of free agents in the mid-1980s (Zimbalist, 1992: 110). Scully (1989: 169) reports that in 1986 and 1987 ‘the free agents’ salary is a much smaller fraction of their contribution to team revenues than is the case of the non-free agents,’ thus giving statistical support to those claiming the existence of owner collusion. MacDonald and Reynolds (1994) reach the same conclusions, showing that players who were forced or chose to take the arbitration route wound up with significantly higher salaries than those who declared themselves free agents. These results are not robust, however. Zimbalist (1992) and Oorlog (1995) come to opposite conclusions, by using procedures that are broadly similar, but with different variables or hypotheses. Their regressions show that the degree of exploitation is smallest for free agents, intermediate for those who could go to arbitration, and highest for those without any salary arbitration rights, a ranking that one would tend to expect. This only shows that there are limits to what econometrics can prove. Whereas straightforward salary regressions tend to be robust—the significance and sign of most performance or experience variables will be consistent—this is not so when more complex theoretical and empirical constructions are being put to task. At least in baseball, it is relatively simple to relate individual performance to winning; in other team sports, it is not so.5

Salary Discrimination

Whereas measures of exploitation appear to be fragile, measures of salary discrimination have turned out to be highly resilient. The issue of salary discrimination is clearly a matter where the interests of sociologists and economists of sport have overlapped. Again, sport provides an extraordinary laboratory to identify and measure discrimination against minority groups, because sport provides direct or quasi-direct measures of individual labor productivity, something that cannot really be obtained in studies of discrimination pertaining to other industries, where evidence can only be indirect.

Once again there are two procedures that can be followed to estimate whether or not there is salary discrimination. In one procedure, first applied to sports by Scully (1974b), salary regressions are run separately for the majority and the minority groups (say whites and blacks), and the estimated parameters of the variables for the two regressions are compared. If, for instance, the salary increase for a one per cent increase in batting average is lower for a black baseball player than for a white one, we may conclude, as did Scully, that there is evidence of salary discrimination against blacks. But what if an extra year of experience is more highly rewarded for black than for white players, as Scully also found? Is there still salary discrimination?

Because its results avoid this sort of ambiguity, the most popular procedure is the one followed by Pascal and Rapping (1972). A single salary regression is run for all players, with a dummy variable identifying the players from the minority group. If the parameter of this variable is negative and statistically different from zero, then there is straightforward evidence of salary discrimination. Pascal and Rapping found no evidence of salary discrimination in baseball. This finding has been repeatedly rediscovered by others, in baseball (Kahn 1991: 400) and in football (Kahn, 1992). The case of basketball is more complex. The current view is that there is no evidence of widespread salary discrimination in basketball (Jenkins 1996; Hamilton 1997). However past studies, based on data of the early and mid-1980s, as reported by Kahn (1991), had shown significant salary discrepancies in favor of white athletes, for given levels of performance. This was rather surprising in a game now dominated by black athletes. But this domination is precisely the explanation that has been advanced: white spectators want to see at least some white players. It has been shown repeatedly that attendance rises when the proportion of whites on the team increases (Kahn 1991: 410; Hamilton 1997: 289). More precisely, still in basketball, Burdekin and Idson (1991: 185) note that there is a tight link between attendance and the racial composition of the team. Also, the greater the racial match between the team and its metropolitan population, the larger is average attendance.

This explanation, based on customer discrimination, is reinforced by several other pieces of evidence. Although there is no evidence of salary discrimination in baseball and football, it has been shown that in baseball the salaries of white players are positively linked to the proportion of black players on the team, whereas the salaries of black players are negatively linked to their representation on the team (Johnson, 1992: 200). For football, Kahn (1992: 307) finds that players make more money the greater the proportion of their race in the local population. Fans in the 1980s were not color-blind! Furthermore, evidence collected from the trading of the baseball cards of past players shows that fans in general—and not only those who go to watch the games—prefer to purchase cards of white rather than black players, for given performance achievements (Nardinelli and Simon, 1990; Andersen and La Croix, 1991). However, the market prices of baseball cards picturing current players show no such customer discrimination (Gabriel et al., 1995). Younger fans may be color-blind after all!

Salary discrimination is not necessarily limited to race. Language may also be an issue. While Latin baseball players have been subjected to some scrutiny, but with little result. There have been studies of salary determination in ice hockey which have shown that French-speaking Canadian players are sometimes significantly underpaid, compared to their English Canadian or American counterparts (Lavoie and Grenier, 1992). More recent evidence, based on the use of crossed variables, shows that such salary discrimination, when it occurs, arises from teams located in English Canada, and not from teams located in the United States (Longley, 1995). This seems to indicate that the observed differentials are due to discriminatory behavior rather than the linguistic costs imputed to French-speaking players.

The Distribution of Earnings

While salary discrimination can certainly modify the distribution of labor earnings between various groups of players, the overall question of the distribution of earnings generated increased attention in the 1990s than before. Before we leave the topic of discrimination, it should be pointed out that entry and exit discrimination are important means of modifying the structure of income. Although they have written several pages on the topics, economists have given less attention than sociologists have to entry discrimination and the phenomenon of stacking. Once again, the classic references are the papers by Pascal and Rapping (1972) and Scully (1974b), and Kahn’s (1991) survey article. By denying access to the major leagues, where salaries are so much higher than in the minor leagues, entry discrimination has substantial repercussions on income distribution.

Two other issues around the distribution of earnings have attracted some interest. With greater access to salary data it has been possible to compute reliable Gini coefficients, that is, measures of inequality in the distribution of earnings among professional athletes. This Gini coefficient must stand between zero and one. A perfectly equal income distribution would bring down the Gini coefficient to zero; and the higher the Gini coefficient, the more unequal the income distribution. As background information, the Gini coefficient for family income in most industrialized countries is in the 0.300 to 0.350 range, while in the United States it exceeds 0.380. While the focus of attention has been again professional team sports, Scully (1995: 74) has shown that earnings in golf were much more unequally distributed than in team sports, the Gini coefficients both for the men’s and the ladies’ tours being around 0.630. However, if one takes only the best 150 players, which one may liken to the athletes playing in the major leagues in contrast to the minor leagues, then the Gini coefficients of individual sports are in the range of those of team sports. As one would expect, the earnings of the top 150 female athletes in tennis and golf are much more skewed than those of the top male athletes in those two sports.

In team sports, baseball has the most unequal distribution (with a Gini of 0.510), followed by basketball (0.420), hockey (0.400) and football (0.370). Where comparisons are possible, namely in baseball, basketball and hockey, the inequality of earnings has risen considerably since the mid-1970s. An obvious cause of the above is the advent of free agency, that is, the possibility for star players to offer their talent to any team. This is particularly so in baseball, where there has been hardly any expansion in the number of major league franchises. By contrast, in basketball and hockey, there has been a two-fold and a fourfold increase respectively in the number of major league teams. As a consequence, talent has been diluted and the dispersion in the number of points scored per player has increased considerably since the mid-1970s, both in basketball and in hockey. The increase in the Gini coefficient of salary earnings is thus to some extent a reflection of the increase in the Gini coefficient pertaining to performance and talent.

While players’ earnings from professional sports teams have become more unequal, this does not mean that the economic situation of the benchwarmer or that of the journeyman has deteriorated. Although mean salaries are biased upward because of salaries from star players, the evolution of these salaries shows that there has been a tremendous increase in the standard of living of the average major league athlete. In baseball for instance, salaries in the second half of the 1950s were around $US 14,000, in current dollars (Quirk and Fort, 1992: 211). This represents approximately $US 82,950 (in 1999 dollars). In 1999, baseball average salaries were $US 1,700,000, more than 20 times the salaries of the 1950s. Similarly, in hockey, salaries were around $US 8,000 in the late 1950s, or about $US 47,500 in 1999 dollars. Mean salaries for the 1999/2000 season were $US 1,350,000, nearly 30 times the salaries of the 1950s. These multi-fold increases in purchasing power can be contrasted to the small increases in the real earnings of the average American. Consider for instance the production worker or non-supervisory employee in manufacturing. Average annual earnings were about $US 4,300 in 1957, in current dollars. This represents approximately $US 25,500 in 1999 dollars. The annual mean earnings of the same kind of worker in 1999 are only $30,160 (in 1999 dollars). Whereas in 1957 the ratio between the earnings of the average major league player and the average blue-collar manufacturing worker was somewhere between 2:1 and 4:1, this ratio is now 55:1 for baseball and 45:1 for ice hockey. In basketball, where the average salary is reported to exceed two and a half million dollars per year, this ratio is now over 80:1. No wonder sport(s) fans often think athletes are overpaid crybabies when they go on strike or threaten to go on strike!

Although the representative athlete from all major league sports has seen his earnings dramatically increase, it should be pointed out that these increases have not been simultaneous. The average earnings in some sports, relative to those of others, sometimes take off, only to be caught back later. What are the factors that explain these wide discrepancies? Undoubtedly, the popularity of the sport, that is, the demand for the sport—as reflected in the ticket prices that spectators are willing to disburse and in the fees that the television networks and the advertisers are willing to expend—explains the increase in sports salaries (Quirk and Fort, 1992: 219). There are, however, other important factors, such as the presence or the threat of a rival league (Ahlburg and Dworkin, 1991: 62; Lavoie and Grenier, 1992: 166), the availability of neutral salary arbitrators (MacDonald and Reynolds, 1994: 444), and the lack of constraints on the mobility of players. We shall deal with the latter factor in the next section, when we analyse the sports team as a firm.

Economics of the Firm

Profit Maximization

An obvious characteristic of sporting leagues is their cartel nature. This brings us immediately to the issue of profit maximization by the firm versus profit maximization by the cartel as a whole, where the latter may constrain individual firms ‘in the best interests of the league.’ While this distinction carries interesting issues, some of which are to be examined later, the crucial issue to be discussed here is whether profit maximization by the firm constitutes an appropriate assumption for sporting firms. Profit maximization is the standard assumption in mainstream neoclassical economics. There are, however, famous models of the firm where sales rather than profits are assumed to be the maxim, on the grounds that high sales may bring more satisfaction to the managerial staff than high profits. In dynamic terms, the question is whether firms try to maximize their rate of profit or their rate of sales growth. With respect to sporting firms, the issue is whether professional clubs try to maximize profits or winning, that is, whether profits or victories bring more satisfaction to club owners.

At first sight, this seems a moot point. One would expect winning teams to induce more spectators, and hence more revenues and more profits. Although it might be so, economists do not expect such a simple relationship to hold on. Two broad reasons can be advanced, one related to the demand side and the other to the supply side. First, after some point, an extra win may not generate many additional spectators or revenues; the revenue elasticity of winning may become weak. Secondly, it may become very costly to generate additional victories; the cost elasticity of winning may exceed unity. When a team is already loaded with all-star players, hiring an extra superstar may barely increase the probability of winning. This is an instance of the well-known law of diminishing returns. Thus, beyond some point, profits start to decrease although revenues still increase. This point is the point of profit maximization. Until this point is reached, the additional revenue (marginal revenue) generated by an extra win exceeds the additional cost (marginal cost) required to engineer this extra win.

Profit maximizers and winning maximizers behave differently on many fronts, as recalled by Cairns, Jennet and Sloane (1986: 7) in their exhaustive survey of the economics of professional team sports. While the latter will strive to win ‘at all costs,’ or as long as they do not suffer heavy financial losses, the former will be content to remain in contention. Profit maximizers will have self-imposed limits on team strength, while winning maximizers will try to dominate the league standings and reduce the uncertainty of outcome. Profit maximizers will sell or trade away good players, merely to unload heavy salaries and increase their profits, whereas winning maximizers will trade players in an attempt to improve the team. It is highly important to know what are the true objectives of sporting firms—profit or winning maximization—for all or most models evaluating the economic and competitive impact of diverse institutional rules and features of sporting leagues assume that owners attempt to maximize profits. Whether firms maximize profits or winning is thus crucial when discussing the effects of reserve clauses, free agency, rookie drafts, revenue-sharing arrangements or salary caps.

Unfortunately, it is very difficult to distinguish empirically between winning and profit-maximizing behavior. As Cairns et al. (1986: 9) recall, various authors have claimed that the predictions of their theoretical or statistical models have given support to the standard neoclassical profit-maximizing assumption. A careful examination of their claims usually shows that either alternative behaviors have not been considered, or that their results are not inconsistent with winning maximization. An attempt at empirically settling this issue showed that owners in baseball have responded more to profit than to winning incentives (Porter 1992). There have also been interviews of club officials, where their goals were being ascertained. Generally speaking, it would seem that profit-seeking behavior—perhaps in contrast to profit-maximizing as such—is a feature of North American teams, whereas winning would be a more dominant objective in European and Australian professional sport. For instance, some rich European soccer teams, such as Real Madrid, pay talented players huge sums to sit on the bench, to prevent poorer rival teams from dressing them up, thus preventing their rivals from challenging their dominance over the national championship (Bourg, 1989: 160).


Micro-economists show that for the marginal revenue of a firm to be above zero, the price elasticity of demand must be above one. At a unitary price elasticity, revenues—but not profits—are maximized. Since the additional costs related to additional sales are necessarily non-negative, profit-seeking behavior requires minimally that the price elasticity be equal to or above one. It turns out that most studies on team attendance show price elasticities below unity, and even near zero (Cairns, 1990: 9). An increase in the price of tickets would generate more revenues and hence more profits, for lower costs would be associated with the smaller number of spectators. This price inelasticity of attendance thus yields little evidence of genuine profit-maximizing behavior. There are, however, some authors who find price elasticities near unity, thus claiming the presence of profit-maximizing behavior on the grounds that there are hardly any additional costs associated with more spectators, for a given winning percentage, and hence that profit maximization coincides with revenue maximization (Ferguson et al., 1991).

The statistical analysis of attendance also yields some interesting insights vis-à-vis the interests of the cartel versus those of the team. In the previous subsection, it was mentioned that the league may want to impose restrictions on its team members, in particular to preserve a viable level of competition and to prevent the wealthier teams from dominating the league. Many different variables have been introduced to meter outcome uncertainty, but to no avail (Borland and Lye, 1992: 1058; Cairns, 1990). If competitive balance is not a determinant of attendance, then league officials running the sporting cartel have no economic grounds to impose restrictions such as reserve clauses upon teams and their players, since the financial viability of the cartel as a whole does not depend upon it.

Of course, this does not mean that the quality of the home team has no impact on attendance. All studies have shown, as one would expect, that a better home team attracts more spectators, but it should also be pointed out that good visiting teams also generally attract more spectators. Hence winning draws additional revenues, even when visiting, if there is gate-sharing. In addition, being in contention for a playoff spot draws more spectators.

One peculiar set of studies on attendance is the one done for ice hockey, linking attendance to violence. This is certainly a theme that is of interest to sociologists, and it may give some ammunition to those who argue that violence in European soccer is condoned or mandated by team owners because it is profitable (Brohm, 1993). Jones (1984) has shown that when a team with a fighting reputation is playing, it draws some 1,500 to 2,500 additional spectators. When acting on the quality side, to draw as many additional spectators the home team needs to improve its standing in the league by approximatelyten positions. A team may thus wish to pursue two strategies to increase revenues: a winning strategy or a strategy of violence. The winning strategy is a risky one: chosen players may not perform up to expectations, while if they do, their salary cost may skyrocket. The fighting strategy, by contrast, is less risky: goons fight when they are asked to, and they are much less expensive than talented stars. In addition, from the point of view of the cartel, the fighting strategy would seem to be the optimal one. Winning is a zero-sum game: all teams cannot have winning records, whereas all teams can encourage fights. There is thus some economic incentive for league officials not to ban fighting in professional hockey, despite moral pressures to do so, because fighting and violence are profitable.

Restrictions on Players’ Mobility

No subject has generated more controversy than the restrictions that sporting leagues impose on the mobility of players. Among these restrictions, the better known are: the draft of junior, college or amateur players, where the negotiating rights of a young player are assigned to one team only; the reserve clause, where a player may play with only one team, unless that team decides to trade the player; free agency with compensation, where a player may decide on his own to switch teams, but where the new host team must compensate the previous team by trading away some equivalent talent or future rookie draft picks. All leagues, whether in North America or in Western Europe, now have free agency for veteran players. What differentiates one league from another is the accessibility to salary arbitration—when free agency status is yet to be achieved—and the definition of veteran: one may become a free agent once the first professional contract has expired, that is, after say three, five or six years; or the player may need to reach a certain age, say 32 years old, as in ice hockey. In Europe in particular, a cash transfer fee may also be required.

As is well known, team owners defend the rookie draft and the reserve clause on the grounds that unrestricted player mobility would destroy competitive balance: wealthy clubs, located in areas with large populations, would attract star players with lucrative salaries and endorsement opportunities, and would thus purchase championships. The league would lose credibility, players would be perceived as mercenaries, poorer clubs in less populous areas would become consistent losers, their attendances would drop, and the league might be forced to fold or to reduce its operations. Many sport analysts and a handful of economists (such as Daly, 1992) support this appraisal.

Since Rottenberg’s (1956) ground-breaking article, economists have systematically objected to the reserve clause or other similar restrictions on players’ mobility. Rottenberg’s claim is that profit-maximizing club owners will behave in such a way that the distribution of talent will remain the same, whether restrictions are kept in place or removed. Thus, whether there are restrictions or not, big-city franchises should have winning teams while small-city franchises should have losing teams. The competitive imbalance is invariant to the restrictions designed to alter it. However, because of the constraints already noted in the section dealing with the objective of profit maximization, the differences in the revenue-generating capacity of franchises should still yield a sufficient degree of sporting competitiveness. The only real effect of restrictions on players’ mobility is to reduce the share of revenues going to these players, while increasing the share going to club owners. When there are restrictions, the player can negotiate a salary that is somewhere between his reservation wage—the salary that would induce him to abandon professional sport—and his marginal revenue product to the team that owns his rights. Under total free agency, the player can negotiate a salary which is in between the highest and the second highest marginal revenue products attributed to him by any team in the league (Quirk and Fort, 1992: ch. 6). The issue is one of income distribution: who gets the rent (the extra profits) generated by extraordinary athletic ability: club owners or athletes? This is most obvious in European soccer, where huge cash transfer payments are pocketed by the club when it still owns the rights of the player, whereas they are pocketed by the player, as a signing bonus, when the contract has expired.

While the statements in the above paragraph constitute the representative opinion of economists, there has always been a minority current of contrary opinion on some of these statements. Furthermore, the arguments related to small franchises versus big franchises have been refined. First it should be pointed out, as did, along with many others before them, Cairns et al. (1986: 33) and Quirk and Fort (1992: 279), that Rottenberg’s claim assumes a profit-maximizing behavior. If owners do try to maximize profits, small-franchise owners who have drafted players who turn out to be highly productive superstars will have an incentive to unload them to richer franchises, because the owners of these richer franchises will offer huge cash amounts, knowing that these superstars can generate large increases in revenues, as a result of the potentially larger attendance. If club owners maximize winning—and this may be not at all irrational in a league where there are playoffs or additional revenue-generating games as in European soccer cups—or if the league forbids the sale of players and unfair trades, Rottenberg’s claim does not hold anymore: restrictions will make the league more balanced.

However, even with free agency, large market teams may not necessarily dominate small market ones. This is the new view of economists on the topic, a view which has been put forth by Porter (1992) and Vrooman (1995). Again there are two sides to this claim. First, it may be that the salaries needed to be paid by teams located in big cities are much higher than those necessary to induce players to migrate to small-city franchises. The revenue advantage of large market teams may thus be wiped out by a cost disadvantage. There is as yet no empirical evidence of this supply-side effect however. By contrast, there is empirical evidence that the fans of various franchises react differently to the performance of their team. If attendance is winning-elastic, that is, ‘if fans demand a winner and express their distaste for losing by staying away from the game’ (Porter, 1992: 75), additional revenues generated by extra wins will be high, and this will induce profit-maximizing owners to search for more talented players. On the other hand, if attendance is winning-inelastic, that is, if fans are loyal, there will be little incentive for club owners to improve the team. Small market team owners with winning-thirsty fans may thus be forced to be competitive with large market teams with loyal fans.

This encouraging result for small market team fans has, however, some drawbacks. First, it should again be pointed out that if club owners do not maximize profits, these results do not hold. If club owners of large market teams maximize winning, they will deplete the ranks of small market teams, especially those that have recently achieved success with bright rising stars. Secondly, a small market team needs a core of loyal fans, otherwise there will be an incentive for the owners to move the franchise to some other, more complacent city. This brings in the issue of whether or not efforts should be made by local public officials to keep the franchise in town, an issue that will be discussed later in the chapter.

Revenue-Sharing Schemes and Salary Caps

Although the problems of small market teams seem to have been exacerbated recently, the issue is not a new one and it has generated some responses. Two broad answers to the problem have been or can be provided. The first possibility is to provide a revenue-sharing scheme that will allow small market teams to have access to the financial benefits inherent in markets with large metropolitan populations. Gate-sharing, where attendance revenues are split between the host and the visiting team, and sharing in national broadcast revenues are such revenue-sharing schemes. The second possibility is to devise some formula that will directly restrict the amounts spent on salaries. Salary caps are such a measure, but proposals for taxes on salary expenditures—positive taxes for teams with large payrolls and negative taxes for teams with small payrolls—were also proposed during the baseball 1994 strike and the ice hockey lockout of 1994-5. While, indeed, all of these schemes should improve the financial situation of small market teams, the issue is whether they would help to establish more balanced competition in the league.

On intuitive grounds, one would be tempted to answer ‘yes’ right away, and indeed Scully (1989: 80) does so. Surely, if all teams have more equal revenues, their capacity to hire good players should be equalized, and this ought to be a good thing for the sport. In fact, this is the standard argument made by league officials who have endorsed these schemes. The standard opinion of economists, however, is full of qualifications. Economists claim that under some circumstances the revenue-sharing schemes will not help to achieve a more competitive balance, and furthermore, that these schemes modify income distribution in favour of club owners, at the expense of players.

Take the gate-sharing scheme. Let us assume, as is standard, that team owners maximize profits. Their decisions to hire talented players will be based on the marginal revenue procured by the additional wins generated by the supplementary performance of these players. If the revenues from attendance are shared, however, the large market team will only get a portion of the increase in gate revenues at home, and marginal revenues when visiting will decrease whenever it wins away. Because winning is a zero-sum game within the league, the implications of gate-sharing will be identical for the small market team (Fort and Quirk, 1996). This is the revenue-sharing paradox. As long as all teams are subject to the same sharing formula, and as long as the shared revenue responds to win-loss records, revenue-sharing has no effect on competitive balance (Vrooman, 1995: 978). In other words, revenue-sharing is useless to equalize the field as long as the shared revenues are winning-elastic.

In addition, gate-sharing has unexpected negative consequences for players. Because it is winning-elastic and diminishes the marginal revenue product of each player for each team, gate-sharing induces profit-maximizing teams to reduce their salary offers to players. As a consequence, the share of revenues going to players should be reduced. Owners would pocket a larger fraction of the rent generated by the talent of athletes.

On the other hand, it can be surmised that national media revenues are revenue-inelastic from the point of view of each individual team. The distribution of lump sums from national media revenue-sharing should thus promote competitive balance and should have none of the above negative consequences on the share of income distribution of players. There is a conflict, however, with the interests of the cartel: one would presume that national media would be tempted to funnel more funds into contracts with professional leagues when these leagues are dominated by large market teams. A greater number of spectators would be happy to watch the network and its advertisers when their favourite large market team triumphs.

There is still some controversy about the impact of a salary cap. Quirk and Fort (1992: 291) argue that the salary cap helps to equalize competition, as long as the cap is enforced. They also show that the overall revenues of the league would be reduced, since large market teams, with high marginal revenues due to their large potential audience, would win less often than without the cap, thus generating less winning-elastic revenues. On the other hand, Vrooman (1995: 980) believes that salary caps and a payroll surtax allow the league to act as a cartel. This helps the league to keep down salary costs and increase the degree of exploitation of the players. In addition, the salary cap would increase, rather than decrease, the competitive inequality between teams. Some authors have noted that competition in baseball has been much more balanced between small market and large market teams than it has been in basketball, despite the salary caps of the latter (Gramlich, 1994). These opposite opinions no doubt explain that both the salary cap and the payroll surtax were highly contentious items during the labour stoppages in baseball and ice hockey in the mid-1990s. What kind of world is it, where capitalist owners want to impose rules and restrictions upon themselves whereas employed players want free markets to prevail?

The Impact of a Franchise in Professional Sport

With the advent of free agency, and with the increase in popularity of professional team sport, the salaries of professional athletes have skyrocketed. As a result, the future of several small market franchises has been questioned. Besides the league-wide schemes presented in the preceding section, this has induced the owners of these teams to pressure local public authorities, in the hope of getting public subsidies or other financial benefits.

The retention of major league teams has not been the only source of controversy between club owners, sports fans, city officials and economists. As A.T. Johnson (1993: 1) notes, the acquisition of sport franchises, be they at the minor or major league levels, has become the objective of sports entrepreneurs and local politicians in communities of all sizes. In North America, cities try to attract franchises from as high a league level as they can. In Europe, where teams are promoted and relegated according to their standings, city officials decide whether or not they will offer large subsidies that will allow their local team to purchase better players and help them to be promoted to an upper-grade league. In both continents, the renovation, construction, siting and use of public sports stadiums or arenas quickly become an issue on the political agenda.

Arguments of five sorts are usually advanced to justify the use of public money for what is basically a private endeavour. First, it is claimed that the presence of the sports franchise will directly create additional (low-paid) jobs: the ushers, the clerks at the sales booths, the parking attendants, etc. When a stadium is built or renovated, there will be a temporary increase in employment and tax revenues. Secondly, there will be indirect benefits, those that can be linked to the famous income multiplier. The public expenditures on the stadium will be injections in the economy, that will induce further spending by those who work on the construction site, or by those who provide goods and materials. These will produce further tax revenues to cover the granted subsidies. Thirdly, the franchise will be a business attraction, enticing visitors from outside, who will spend money inside the community, thus helping local business to be more profitable. Fourthly, the presence of a sports franchise will bring regional, national or international media attention and recognition to the city. This should help local firms to export their products and services, and more specifically it should help the tourist industry. Sport is thus seen as a strategic ploy for regional economic development. Finally, even if all these benefits are dubious when netted out against the costs of servicing a newer or bigger stadium or arena, there remains for the citizen the positive psychic and sociological benefits of being associated with a sports franchise and possibly a winning team. With this last claim, professional sport is said to be a public good, akin to good health.

While economist consultants hired to assess the economic impact of sports teams and their stadiums on host communities systematically find a positive impact, this is not so with academic economists. The latter are usually highly sceptical of the economic benefits associated with a new franchise, and a new or renovated stadium. The consensus among academic economists is that most of the benefits, if they exist, are intangible, and related to civic pride (Johnson, 1993: 4). The better known studies have been done by Baade and Dye (1988, 1990). Looking at major leagues with the help of regression analysis, they show that a new franchise, a new stadium or a renovated stadium, have no impact on the level of metropolitan income or retail sales. The only possible exception is that of baseball. The opinion that major league sports have little, if any, impact on local economic activity has been reinforced by a study that has shown that the baseball strike of August and September 1994 had no negative economic impact on cities with baseball franchises (Zipp, 1996).

An interesting case study is that of Gouguet and Nys (1993: 230-9). They look at two widely different cases: the city of Limoges, which has a basketball team that is highly successful at the European level; and the city of Rennes, which has a soccer team usually ranked in the bottom half of the top French soccer league, always on the verge of being relegated down to the second division. In both cases, it turns out that the sums of money injected in the local economy by each of these two professional teams were largely overturned by the sums of money leaking out of the local economy because of these professional sporting clubs. These leaks were due in large part to the players spending only a portion of their salaries within the local community. Most of these salaries, which constitute a large chunk of the clubs’ expenses, were either taxed by the French government or saved by the players. The little that was spent was either spent in Paris, for the soccer players, or in the States, for the American stars of the Limoges basketball team. There was no room for any positive multiplier effect, for the multiplicand was negative!

There are obvious lessons to be drawn for North American cities desperately seeking to obtain a franchise in major league sports. There are also obvious lessons to be drawn for cities wishing to organize major sporting events such as the Olympic Games, the Pan-American Games, or the Commonwealth Games, as Gouguet and Nys (1993) show. Indeed, tourism may even drop during such games, as tourists fear being crowded out.

In general one must beware of economic consultants shuffling multiplier effects and complex input-output models. Whatever their degree of sophistication, by their very method, these studies of economic impact cannot but predict substantial financial benefits to the organization of large sporting events or to the arrival of major league franchises. But similar results would be obtained whenever any project with similar additional expenditures is being proposed. Studies of economic impact only demonstrate that a reduction in unemployment is a profitable venture. If such is the goal being pursued, then one must decide what sort of public expenditure is more desirable. Building a new stadium with a large number of private luxury boxes, filled with corporate executives and paid for by the consumers of the products that these corporations produce, may not be the most appropriate choice.


Despite the extraordinary development of the economics of sport, it is clear that much of its focus has been on the economics of professional team sports. There are basically three reasons for this: the topic offers many instances of micro-economic conundrums that can be solved; economists thrive on numbers, which professional sport generates; the topic is popular whereas amateur sports are not (in North America). Of course, many other topics which have attracted the attention of economists could have been covered. On the micro-economic front there have been studies on the cost of practising sport (Michon and Ohl, 1989), the logic of which has inspired the so-called ‘economic hypothesis’ purporting to explain positional segregation in team sports, that is, stacking (Medoff, 1977, 1986). Others have also tried to predict the demographics of professional tennis, giving economic explanations as to why young players would dominate the sport, or as to why older players would remain longer on the tour (Galenson, 1993, 1995). A problem with these hypotheses is that they can predict just about anything. On the macro-economic front, there have been qualitative and quantitative studies on the increasing globalization of sport, both for athletes and for the sporting goods being produced and exported (Andreff, 1989; Harvey and Saint-Germain, 1995).

What is the future of the discipline? On the pedagogical front, the discipline has now reached the point where it is possible to have whole courses devoted to the economics of sport. Bruggink (1993) shows how one could build a first-year introduction course on the principles of micro-economics, with illustrations taken entirely from the economics of base-ball. Obviously, a special topics course on the economics of baseball, or the economics and business of any other professional team sport, could also easily be constructed. Estenson (1994) describes how students were asked to maximize profits while pretending to be baseball team owners in search of free agents.

What about the academic front? As reported by Johnson (1995: 505), in 1991 the keynote speaker at a conference on the economics of baseball ‘declared that everything important and interesting about the economics of professional sports had already been said, and no further work was necessary.’ Since then dozens of books and articles on the economics of baseball and the economics of sport have been published. It is true that many of the more recent articles deal with arcane or trifling issues, or rely on highly dubious hypotheses to arrive at some practical result. There is no lack of ingenuity, however, in some of the newer articles. Just as one would think that an old model cannot be improved any more, someone comes along with an innovation that gives more insights or that makes the model more robust. Where objections have been raised about previous interpretations of the data, new means to test these interpretations have been found.

What are future poles of research? There is still a great deal of uncertainty about the true impact of salary caps for competitive balance and the welfare of players regulated by such caps. There is also a great deal of uncertainty about the future of small market teams, and how their financial situation can be improved without damaging that of the players. The reverse-order rookie drafts have only been given scant attention, but they may be the only restriction still promoting competitive balance: should they be partially removed, like the reserve clause, limiting the draft to only the best prospects? Finally, leaving theory for practice, one may predict that if professional athletes dared to go on strike in various countries to increase their share of the revenues generated by professional sport, the day cannot be too far off when the so-called amateur athletes will threaten to go on strike in order to get their share of the huge revenues generated by worldwide mega-events such as the Olympic Games.