North American Leagues

Despite many willing entrants, today's top-tier North American professional sports leagues maintain a fixed roster of teams—a “closed corporation,” as National League promoters referred to their structure.³

Only a limited number of teams can compete, and that number is fixed until the incumbents decide to admit others. Each team enjoys a league-enforced territorial monopoly or duopoly.⁴ With three exceptions, teams in the same sport now play in different stadiums.⁵

Starting a new team is expensive in North America. Existing teams, acting through the league, create and sell new franchises for prices vastly in excess of the value of the contracts of players drafted from existing teams or any other tangible property that an entrant receives. A minor league team in the expansion city characteristically moves elsewhere or ceases operation. A rejected entrepreneur's only recourse is to organize an entire new league. Attempts have been relatively common, but few have succeeded.⁶

Existing teams have relocated with league consent to new cities as far as 2,600 miles (4,100 km) away. An abandoned city has difficulty attracting a replacement.⁷

Although long-distance moves engender acrimony, relocations of several dozen miles are uncontroversial—neither the Angels' 30-mile (50 km) move from Los Angeles to Anaheim nor the Patriots' move from Boston halfway to Rhode Island generated much comment.

League rules specify roster size. They forbid loaning players to other teams at the same level, although baseball and ice hockey teams contract with more players than their active rosters can hold, assigning the surplus to “farm” teams that play elsewhere in lower-level leagues. Though teams can trade player contracts, if a cash-strapped owner attempts to capitalize the roster by selling the contracts of a number of team stars, the league will nullify the deals.

European Leagues

Automatic promotion and relegation is practiced by nearly every European league; strong teams at one level regularly replace the weakest teams from the next higher level. For example, in England the top 20 football teams play in what we refer to throughout the article as the Premier League; three teams from the top division of the Football League annually interchange with three Premier League teams, and so on down.⁸

For illustration, any of the 138 teams presently playing in the five tiers immediately below the Premier League could in principle be Premier League teams in five years or less. Thus the industry is open to entry through superior performance, and any entrepreneur with ambitions to operate a Premier League team enjoys wide latitude to try.

Few (if any) European leagues recognize territorial rights. By North American standards, English professional football teams are tightly packed. Indeed, rival teams at each competitive level sometimes play within walking distance of each other. Measured as the crow flies, the stadiums of three major London teams—Chelsea, Fulham, and Queens Park—are separated by less than three miles (five km). Thirteen football teams playing inside London's ring road have competed in the Premier League at some time, and multiple teams in smaller cities also have reached that level.

A newly formed team simply registers to play in football's lowest tier. It may be geographically near an existing competitor; it needs no one's permission to be there. It then works its way up. Relegated teams do not relocate.⁹

Teams seldom trade player contracts, but they often sell them for cash. The maximum number of players who may play in a game is defined, but the roster from which they can be drawn varies widely across the teams. Surplus players are loaned to other teams in the same or a different tier, or assigned to a reserve team that plays locally. No farm teams play elsewhere.

North America

Baseball evolved in the Northeast from informal children's games such as rounders.¹¹

Young men formed social clubs that made arrangements for fields where members played intramurally. In 1862 William Cammeyer made his Brooklyn property available free to local clubs that agreed to let him charge spectators admission.¹² Cammeyer's innovation spread quickly, but soon the teams began to collect the revenues. When receipts surpassed operational expenses, the surplus was used to entice “crack” players away from other teams, though players of more modest talent expected no compensation.¹³ The game spread rapidly among the troops during the Civil War, and among the population at large when the troops returned home.¹⁴ Eventually club all-star “first nines” began to travel to play other clubs' all-stars, but there was no consolidated schedule or championship.

A protoleague, the National Association of Professional Base Ball Players, broke away from an earlier, staunchly amateur predecessor in 1871.¹⁵

National Association membership was unrestricted; any team could compete by paying an annual fee; clubs were instructed to arrange, on their own, best-of-five series against each of the others. The best-of-five format meant that a three-game sweep would precede two days without paying fans. Scheduling protocols meant that a western team with an eastern series to play might belatedly withdraw if no intervening opponents with open dates could be found to help defray travel costs. Alternatively, after a team began a trip, an opponent might disband before that part of the itinerary had been reached, leaving the travelers with idle time.¹⁶

The National Association failed miserably to develop an orderly process for teams to contest a championship and balance their books, so the Chicago Tribune urged restructuring. The result was the first modern professional sports league, “a new baseball association based on territorial monopolies and a restricted number of strong, financially healthy clubs.”¹⁷

The National League excluded weak teams that had imposed external costs on their fellow National Association members. To the modern observer, the league's centrally determined schedule may appear an obvious way to avoid many of the National Association's failings, but it was a major innovation at the time, one that rivals would quickly imitate. In order to facilitate extended, multigame trips that economized on travel costs, the new league selected its member clubs with an eye to geography—four in the old Northwest (Chicago, Cincinnati, Louisville, St. Louis) and four in the Northeast (Boston, Hartford, New York, Philadelphia). The major leagues then remained within the northeast quadrant of the United States for the next 82 years. Only one club could represent a city. Until 1892 the territorial monopoly that came with league membership represented a transitory right, and nonleague teams often supplanted solvent clubs between seasons.¹⁸

Of the several early attempts to circumvent National League exclusion, only the American Association survived for more than a season. Because three founding clubs had previously played in the National League, they understood the advantages of a coordinated schedule and a one-team-per-city limit. During 1882, the American Association's first season, none of its teams played in a National League city. That segregation ceased the next season when each league invaded one rival market. Over the next ten years, the two leagues competed head-to-head in eight cities. The National League characteristically won those confrontations, occasionally pirating an American Association club in the process.¹⁹

The coup-de-grâce came after the 1891 season when the eight National League teams invited four from the American Association to amalgamate with them to form a 12-team league.²⁰

The American Association's demise marked the end of city entry into the National League until the league began to transfer franchises 62 years later, and it marked the end of team entry into that league until its first expansion franchises were sold 71 years later.²¹

A new National League constitution converted transitory membership into a near permanent right by forbidding involuntary removal of teams that met specified conditions. By 1900 the league had bought out four member clubs (three of them had become virtual farm teams of other league members) and was back to eight, marking a permanent end of team (as opposed to city) exit from today's major leagues.²² Putting aside its territorial monopolies, National League club turnover before 1892 resembled the system the Football League employed until that same year. It was the closest approach North American professional sports leagues would ever make to the promotion-and-relegation system that is now commonplace throughout the rest of the world.

The only long-lasting challenge to the National League came from the minor Western League. This upstart developed quality teams in small markets, and then in 1901 moved them into three of the four cities that the National League had just abandoned, while going head-to-head in three others. Renaming itself the American League, the newcomer refused to recognize the National League's system of reserving the services of a player to the team with whom he first contracted.²³

Over the next two years the American League transferred two more franchises and was in direct competition with six National League clubs in the nation's five largest markets.²⁴ Perceiving a credible threat, the National League grudgingly accepted the junior circuit as co-equal via the Major League Agreement of 1903. The U.S. Supreme Court would soon legitimize baseball's closed cartel by granting an antitrust exemption.²⁵ The major leagues then stood pat for half a century, retaining precisely the same teams and remaining in exactly the same locations.

Europe

The evolution in England was similar. Informal games resembling football trace back millennia, but interscholastic competition between English schools sparked the first rule codification, the Cambridge Rules of 1848.²⁶

Graduating students formed clubs in their home cities. Eventually all-star elevens began to compete. As in early baseball, there was no central schedule, no championship.

In 1863 the Football Association was formed to refine the code and resolve disputes, incidentally severing soccer from rugby. Eight years later—the year that baseball's National Association was formed—the Football Association established the FA Cup, an open-entry knockout tournament that still exists.²⁷

The following year, the Football Association initiated international matches among England, Ireland, Scotland, and Wales.²⁸ Well before the Football League emerged, the FA Cup and the internationals had grown popular, influencing the continents' divergent evolution—whenever the Football League has tried to distance itself from nonleague members, the Football Association has threatened to exclude league teams from the FA Cup competition and their players from the internationals. Were they to do so, the disciplined teams would find it difficult to hire players.

During the last quarter of the century, working-class men increasingly played football and imitated the club structure. Oxbridge graduates were over-represented both in the Football Association administration and the football-playing population in and around London. Many refused to play against anyone who was paid to play. Industrial worker-players from the Midlands and the north of England, however, could ill afford trips to distant games.²⁹

In 1885, three years before the Football League formed, the Football Association reluctantly voted to allow payment for “broken time,” or wages lost while playing and traveling.³⁰ Clearly some players had been paid surreptitiously even before 1885, because a number of working-class football clubs had been suspended or otherwise punished by the FA for that infraction.³¹ Because, after 1885, some locales generated more revenue to pay broken-time compensation to semiprofessional players and even more to stars, the variance of skill levels across teams increased, and with it the number of unappealing one-sided contests. Teams began to seek parity, in other words to seek opponents of a similar caliber.

The Football League formed in 1888 when Birmingham's Aston Villa club proposed that a few top teams annually play one another home and away, and 11 others agreed.

The Football League's founders were familiar with the National League and imitated its coordinated schedule, but, after debating the matter, rejected territorial restrictions. Six Lancashire teams, three in-and-around Birmingham, and three in-between comprised the initial league.³³

The FA's rule change means that Football League clubs were most likely comprised of a mix of semiprofessionals and full professionals, though we have been unable to find evidence regarding the speed with which professionalism spread.³⁴

It is plausible that, on the best-supported teams, the most skilled players earned enough to give up outside employment. Even today a gradual transition from semi-professional to full professional teams in English football continues. At present about half of the 22 Conference teams—those in the fifth tier whose strongest members are annually promoted into the Football League—employ semiprofessional players. At its inception, the Football League was not a national league in either name or fact; it kept games near its semi-professional players' main jobs.

As had happened in America after the National League provided its seminal example, the advantages of the Football League's coordinated schedule quickly became apparent to other English clubs. The Football Alliance, a rival league with a similar geography, sprang into existence the following year.³⁵

From the beginning, the Football League required annual reapplication by the four lowest-ranking clubs.³⁶ In its third season the league replaced Stoke with Sunderland, an even more northerly team that it pirated from the Football Alliance. Two years later the league initiated its now characteristic divisional structure upon absorbing the rest of the Football Alliance.³⁷ Annually interchanging a few teams between divisions maintained parity more systematically, reapplication being required only of the lower division's bottom teams. Thus in 1892, the precise year that the National League abandoned what Eckard termed its “quasi team-promotion system,” firmly and finally closing its cartel, the Football League began crystallizing its once similar structure into the comprehensively open system one finds today.

In 1893 Arsenal became the Football League's first London team, though it languished in the second division until 1904. The league only moved south with alacrity in 1920 when the upper division of still another rival, the Southern League, which had formed in 1894, was absorbed as a third division.³⁸

The following year the league added 20 small northern teams. From 12 teams to 86, the Football League's average expansion rate between 1888 and 1921 was well over two teams per season. Over a comparable period following National League formation, the major leagues expanded from eight teams to 16, less than a quarter of a team per season.

The Inadequacy of a Cultural Explanation

The difference in league structures is sometimes said to reflect a difference between European “sportsmen” and American “profit-maximizers.” Certainly the founders of the Football League did not manifest as much commercial zeal as the National League's founders.³⁹

That is explainable, however, by the different economic environments the teams faced. Innovators do not dictate to their economy; an economy selects viable innovations from among the possibilities that are offered by its array of innovators. In the daunting environment in which baseball operated, teams faced substantially higher costs, which forced them to find ways to generate commensurate revenues. The environment in which football operated permitted a goal of winning championships to be primary, but football could not have survived as its costs increased without recognizing its financial constraint. In either environment, financial considerations were (and remain) crucial.

Empirically, history reveals that the first teams on both shores originated from informal amateur recreations, sportsmen's teams run for and by the players. Prior to league formation on both continents, when spectators betrayed a willingness to pay, a few teams began to use some receipts to entice superior players to join. The strongest teams absorbed more player time during the season and paid them accordingly. For a variety of reasons, players' salaries and travel costs were higher in the United States, so teams were forced to focus on finance at an earlier date than was necessary in England. This focus meant professional management grew in importance more rapidly in the United States.⁴⁰

Methodologically, distinguishing a financial goal from a financial constraint would seem to rely on an unobservable phenomenon—entrepreneurial motivation—which would render the distinction untestable and, as will be shown, unnecessary.

Compactness

Uncoordinated scheduling proved easier in densely populated England than in sparsely populated America, because teams in the United Kingdom had many more nearby opponents among which to choose. Even top-flight English clubs could ordinarily arrange games against comparably strong opponents at locations that permitted round-trip travel within a day, requiring little disruption of the work schedules of the fans as well as those players who continued to hold outside employment.

A similar environment existed for the lesser clubs in urban America, but the best teams discovered that they could either travel short distances to play generally one-sided games against inferior opponents before few fans, or they could travel long distances to meet comparable teams before large crowds composed almost entirely of home-team fans. Long-distance travel disrupted players' work at their full-time jobs, so U.S. baseball's top teams became fully professional much earlier than any English football club. Once a team had to travel so far that its players could not hold outside jobs during the season, the marginal cost of an increase in travel distance diminished. Though baseball's total travel costs were much higher, their marginal travel costs with respect to distance were lower; thus, they became less sensitive to distance than were U.K. football teams.

Old England occupies about three-quarters as much area as New England, as shown in Figure 1, where the countries are drawn to the same scale. According to the 1890 U.S. and 1891 U.K. censuses, however, Old England contained roughly half as many people as the entire United States. Even today England and Wales hold more than a fifth of the U.S. population in an area the size of Michigan. The Football League covers only a middling area among major European soccer leagues, barely half that of Italy, but even the league in France, Europe's most extensive, covers less than half the area of the northeast quadrant of the United States, where the National League began play.⁴¹

Due to their greater population densities, European nations were blessed with a denser rail network than the United States.

To illustrate that the requisite travel would have rendered supplemental in-season jobs less viable for American than English players, we contrasted the geography of the National League with that of the Football League upon their respective inaugurations. The average separation between the eight original National League teams was ten times that between the 12 original Football League teams—595 miles (958 km) versus 59 miles (95 km).⁴²

Over the history of the Football League, the furthest two teams have ever traveled for a league game is 410 miles (660 km), the distance between Plymouth and Newcastle. That trip is about the same as that from Boston to Baltimore, or Los Angeles to San Francisco. It is less than the average separation between the original National League teams, and less than a third of the average separation between today's major league teams. Manchester and Liverpool are 35 miles (55 km) apart. Either one is less than 95 miles (155 km) from Birmingham, roughly the distance between Chicago and Milwaukee. These three English cities are closer than the minimum separation between National League teams in 1876, the 101 miles (163 km) between Cincinnati and Louisville. English compactness led to lower travel times and thus lower total (as distinct from marginal) travel cost.

Travel time is a reasonable proxy for the average player's opportunity cost, and the time that would have been required during the late nineteenth century to travel round-trip to another city and play an average game can be estimated. Assuming that compensation foregone per hour was equivalent in baseball and football, we took the intercity mileage, assumed average train speeds of 35 mph (55 km/h), then added six hours (three hours for getting to and from the stations at each end, dressing, warm-ups, undressing, and treating injuries, plus a generous three hours for playing the game).⁴³

None of the 66 city pairs in the original Football League would have required as much as 14 hours for that round trip, and 54 (82 percent) would have required less than 12 hours. In contrast, among the 28 original National League city pairs, 24 (86 percent) would have required more than 14 hours, and none could have been completed in less than 12 hours. Indeed, nearly two-thirds of the National League round trips would have required more than 24 hours, nearly half would have required more than 48 hours, and a round trip between St. Louis and either Boston or Hartford would have taken almost three full days.

The average round trip in the Football League (including game time) was less than half a day and well under the National League's minimum. Plymouth didn't play a league match against Newcastle until 1934, but even in 1888 that round trip would have required only 30 hours, well under the 40-hour average in the original National League.

Because strong competitors were nearby, English football teams could leave for most away games Saturday at noon (the end of the British workweek until the 1950s), play a game against a closely matched opponent during the afternoon, and return home Saturday evening. Separating the teams by enough distance to form monopolies would have meant paying the players more, because the semiprofessionals among them would have experienced more broken time. Separating the teams into territorial monopolies would have prevented many fans from traveling with a team, reducing revenues. European fans have always attended away games in large numbers, and from the beginning most Football League games were scheduled to accommodate them. In contrast to all North American sports, night games remain rare in English football because so many of them would require visiting fans to spend a night away from home. Thus, when the Football League formed, there was good reason to eschew territorial monopoly.

By contrast, a quality baseball team had to travel so far that its entire roster became fully professional of necessity; no player could maintain outside employment during the season. Few supporters could travel to away games, so attendance was almost exclusively local fans. When the National League formed, there were few potential cost savings to be netted against territorial monopoly advantages. With fully professional players on both continents and the availability of air travel, the travel cost difference has fallen substantially, but even today in the United States, visiting teams and fans generally must incur at least an overnight stay.

Excitement Level of Games

The excitement level of games is a function of parity and variety/rivalry. The literature has discussed the benefits of parity extensively—any league's success depends on staging games with uncertain outcomes.⁴⁴

If outcomes become too predictable, games become boring, and fewer fans attend. We characterize parity as having positive, but decreasing, marginal value. Under normal circumstances, sports fans prefer a greater than 50 percent probability that their favored team wins, but attendance wanes when the likelihood of a win becomes too great.⁴⁵ Consider one team in a two-team market. The other local team offers low travel costs, but two teams playing repeatedly expose parity differences. So, variety is also a virtue, again with positive, but decreasing, marginal value. Variety and rivalry are negatively correlated. If rivalry also has positive, but decreasing, marginal value, an optimal league size exists where adding a team leads to a loss of passion that overwhelms the weak benefit of added variety, but the increased rivalry that results from subtracting a team fails to compensate for the ennui from repetition.

To produce well-played games with uncertain outcomes in the sparsely populated, relatively rail-poor United States, the National League restricted membership to one outstanding team in each of eight cities that spanned 1,207 miles (1,942 km). For the sake of parity, the league might have begun with even fewer teams, but that would have provided too little variety.⁴⁶

Although the Football League began in England's midsection with 12 teams spanning 126 miles (203 km), parity tended to be higher than the National League's had been at a similar age (see Figure 2).⁴⁷

Because comparable, high-quality opponents were much denser, variety was less costly to English teams, so, until reaching long-run equilibrium, any Football League division would be predicted to have more teams than would baseball. That requires a relatively strong assumption for cross-section analysis—omitted variables have similar impacts on baseball and football—though that assumption is unnecessary for time-series analysis. Testing that hypothesis is trivial. The Football League began with a larger membership and grew to 28 teams within four years. Despite the much larger U.S. population, well over a century passed before 28 teams were eligible to compete for the World Series championship. Even today there are only 30 such clubs, and only 22 have won a title. One hundred thirty-one teams have played in the Football League, 20 are in the Premier League and another 72 in the three Football League divisions today. Sixty-three have reached the Premier League at some point—an average of more than 29 separate teams during any five-year span since the Football League added a fourth division in 1921—and 23 have won its championship. From the birth of the Football League in 1888 through the leagues' seasons that began in 2005, the inverse Herfindahl yields an equivalent of 35.9 Premier League teams, but only 23.5 Major League Baseball teams.⁴⁸

If one grants our strong assumption that omitted variables have similar impacts, parity was less costly in England at the outset, so the Football League would have begun as the more balanced league. Even without such an assumption, falling travel costs would raise parity on both continents, though more strongly for baseball where the initial costs were greater. Testing the parity prediction is more challenging than testing the variety prediction. The parity measure we employ is the proportion of clubs that win between 40 percent and 60 percent of the maximum points (football) or games (baseball) available to a team in a season.⁴⁹

Figure 2 contrasts parity in the Premier League and Major League Baseball at similar league ages (the effect remains when we use chronological dates that begin each series in 1888). We end the analysis with the 1955/56 Football League season due to several exogenous events that seem to have affected football parity adversely—court decisions that eliminated maximum salaries, broadcasting's advent as an important revenue source, and the inauguration of continent-wide competitions. It is unclear a priori how salary cap elimination or broadcasting should be expected to alter football parity vis-à-vis baseball, as comparable phenomena occurred in North America. The continental competitions, however, would predictably reduce it.

In the continental competitions—principally the Champions Cup (subsequently the Champions League) and the UEFA Cup—Europe's best few teams according to the prior season's results play against each other. That provides a bonus for excelling—a lucky few teams play additional games against similarly skillful foreign clubs. Of necessity being interwoven between weekend league games, the continental contests are played on mid-week evenings against unusually distant clubs. Because fewer visiting fans will attend, those extra games are more valuable to teams with strong local followings, and, as of the 1956/57 season, they had an added incentive to invest in team quality. Perceiving little chance to “play in Europe,” the many teams that rarely play in football's upper divisions, and then for only a few seasons at a time, were given little added incentive to improve.⁵⁰

The number of teams admitted and the number of games played by the median team—and thus the influence of the continental competitions—have increased over the years.⁵¹ A regression of the ratio of Major League Baseball parity to Premier League parity against league age suggests convergence, the trend line's ratio increasing from 0.673 at the start to 0.910 at the end. The statistical confidence that the increase of 0.004 per year differs from zero approaches 99 percent (p = 0.015).⁵²

The National League was pinched on the variety margin despite including a few fringe members, as it was pinched on the parity margin because it included them. To produce a more attractive product that balanced variety with parity, the National League rejected open entry. Assuming that the league started with the appropriate teams, potential supplicants either were too far away (say, in San Francisco) implying travel that would be too time-consuming and costly, or they were in a market that was thought to be unable to support a suitable team.

The National League was prepared to expel teams that did not measure up, but tried to discourage autonomous exit by those that did. During the early league history, a league club might play as many “exhibitions” against nonleague foes as it played league games, but playing such games in league cities was banned from the outset. As the number of league games that comprised a season grew, exhibition games were virtually eradicated. Consequently, any high-quality team that left the league by choice or expulsion experienced difficulty scheduling worthy opponents, and often failed.⁵³

In contrast, frequent nonleague games remain staples of English football, institutionalized in the FA Cup, continental competitions, and various others.⁵⁴ The Football League could be more blasé about turnover because, with greater team density than the National League, a defecting team was relatively easy to replace.

The Football League could also be more relaxed about the danger that an owner would weaken his team by capitalizing the contracts of star players. The problem would be dealt with promptly as the weakened team would be relegated to a division more in keeping with its reduced ability. If a North American team attempts to capitalize its roster, however, the impact on parity, and thus the external cost on other teams, could prove long-lived.

Territorial Monopolies

In sharp contrast to the nascent Football League, National League founders seized the opportunity to create territorial monopolies from the outset. Initially the league replaced clubs between seasons when that suited the interests of those that remained, but no team has ever been permitted to locate in a city that is already the home of another National League team.⁵⁵

So from its birth the National League was closed to a degree that the Football League never has been. Following the American Association's demise, entry into the National League—even into a nonleague city—became extraordinarily difficult.

Monopoly successfully achieved, even if for meritorious purposes, may provide market power. When the first additional market grows to a size adequate for a major league caliber team, entrepreneurs face an expensive choice—organize an entire league, as the American League did, with members who either must face entrenched franchises or be sited in initially substandard markets (risking piecemeal invasion as a result of major league expansion or franchise movement as those markets grow); or offer to purchase a franchise from the league.

A European team of any level, coordinating with no other club, need only keep winning to reach the top. The Football League's mergers with the Football Alliance and Southern League as well as the more recent emergence of the Conference illustrate that credible alternative leagues are easy to form in England's compact, densely populated environment. When the better teams became fully professional, the multidivisional promotion-and-relegation system was firmly entrenched. By 1921 the Football League already had 86 teams, and 33 had played in the Premier League. It seems that the Football League could not have maintained an effective closed cartel, and it never tried.

Assume that the Premier League had decided to adopt North American-style territorial restrictions for the 2005/06 season. Further, assume each metropolitan area was limited to a maximum of two teams within a 30-mile (48-km) radius.⁵⁶

This means that only two of the six London teams in the league could stay there; four would have been forced to move to much smaller cities if they wanted to remain in the Premier League. Indeed, considering the number of teams separated by less than 30 miles in other parts of the country, nine of the 20 Premier League teams would have had to move to meet the territorial restrictions. Rather than move, there would be an incentive for the potentially displaced teams to remain in place and play in the Football League. Thus, the likely result would be two leagues of similar ability and credibility with no alteration in any team's capacity to control a territory. As the National League discovered when the American League moved into its territory, when faced with credible insurgents, cooperation makes more sense than exclusion.

But why did the initial teams institute the promotion-and-relegation system that permitted so many competitors to gain a foothold in the first place? To understand that, it is necessary to recognize that the interests of league teams are heterogeneous. Expansion of the number of teams in a single league became unattractive once the marginal values of variety and rivalry were offsetting, so a divisional structure was adopted. Random year-to-year fluctuation in team quality meant that promotion and relegation within a divisional structure would maintain a higher level of parity. Teams with a strong expectation of playing in the Premier League would benefit, though weaker league members might not. English football is definitely fluid, but a core has enjoyed frequent Premier League membership, and, while doing so, promotion and relegation has provided them with greater parity.⁵⁷

As long as travel costs are significant, our analysis predicts football team survival in smaller markets than are necessary to support baseball teams and, collaterally, English metropolitan areas that possess more teams for a given population. Given England's size, virtually every game is against a neighbor (at least from a North American perspective). The lack of territorial monopoly increases price elasticities in European football, but games draw fans from two markets so the more elastic demand curve is shifted to the right.⁵⁸

In contrast, most baseball attendance is drawn from the territory over which a team has a monopoly.

Intercountry differences in market size are difficult to quantify because the United Kingdom and the United States do not define metropolitan areas in analogous ways; the “conurbations” as defined in the 2001 U.K. Census cover substantially smaller areas than the “combined statistical areas” (CSA) defined in the 2000 U.S. Census. London's conurbation appears to be about as populous as the San Jose-San Francisco-Oakland CSA, the fifth largest in the United States. Whereas that CSA has two major league teams, 13 London teams have played in the Premier League, six during the 2005/06 season.⁵⁹

Using a road map for guidance, we calculated populations for ersatz English CSAs of roughly the same area as Chicago's. London's “CSA” has a substantially higher population than the London conurbation, but still only half that of New York City. Again, only two New York City teams now play in the major leagues. Chicago's CSA has a larger population than our imaginary Manchester-Liverpool CSA, but, as contrasted with Chicago's two major league clubs, 13 football teams in Lancashire and Cheshire have reached the Premier League, seven during the 2005/06 season.⁶⁰

Table 1 lists, by size over time, British conurbations and cities with teams in the Premier League. The number of teams in the original Football League and the number of teams today give an indication of league growth, with the last four columns indicating when it occurred. All conurbations plus the city of Nottingham have multiple teams with Premier League experience, and other Football League teams in and around nearly all of those places are yet to be promoted that far. The pattern is extremely varied in comparison with North America. A conurbation with over 7 million people hosts 13 past and present Premier League teams, but a town with a population of less than 90,000 also supports one.⁶¹

Beginning with the National League's founding and ending with the 1903 Major League Agreement, Table 2 shows, on a year-by-year basis, all teams playing in the National League (denoted by N), the American Association (from 1882 through 1891, denoted by A), the American League (as of 1901, also denoted by A), or that a city had a team in both extant leagues (denoted by 2). Urban areas have been ranked by the 1890 Census; Brooklyn is ranked with New York City throughout these years. Three noteworthy features are no city with more than two teams until 1903 when the American League placed a team in New York City (see note 24); headlong abandonment of small cities, both features in sharp contrast with English football; and league saturation of the largest markets, comparable to the Football League's expansion north to Tyne and Wear, east to Yorkshire, and south to London.

In Table 3, today's major league cities are ranked at the four census years that correspond as nearly as possible with the Major League Agreement (1900 Census data paired with teams in 1903), the beginning of relocation (1950 paired with 1953), the beginning of expansion (1960 paired with 1961), and today (2000 paired with the present).⁶²

At each date, markets without teams at that time are in italics. The process of transfer and expansion that led to teams in an increasing number of major markets is evident. Milwaukee, the smallest market with a team now in Major League Baseball, is nearly 19 times the size of Burnley, the smallest city that has had a team in the Premier League.⁶³ By 1903 teams were in nearly every major market of the northeast quadrant of the United States, but nowhere else. Only three unoccupied markets with populations that exceeded the smallest major league city's would have been available to an entrant in the same region—Baltimore, Milwaukee, and Buffalo (the eighth most populous in 1900 and a perennial bridesmaid in baseball). Before air travel became possible, San Francisco and New Orleans (the twelfth most populous) were too distant to be included. In sum, at the time of the Major League Agreement 16 teams played in ten cities, with the smallest ranked fifteenth (see Figure 1). Those teams remained in precisely the same cities for half a century, but by then the smallest market—Cincinnati—ranked seventeenth, and seven larger ones around the nation had no major league team. Though still tedious, air travel was rapidly improving. Soon the entire nation would comprise a single market for professional baseball. Major League Baseball formulated a plan to meet the coming threat—transfer “surplus” teams from multiteam markets to large cities that had none. In successive years beginning in 1953, one of the two teams in Boston, St. Louis, and Philadelphia moved to Milwaukee, Baltimore, and Kansas City, three sizable cities without major league teams.

For a few years the West Coast remained unworkably distant, but that need not have stymied a major league caliber entrant that confined itself to the region, as the existing leagues continued to confine themselves to the continent's opposite side. The minor Pacific Coast League (PCL) appeared to be on the verge of becoming a third major league, relying solely on the region's population and wealth. The PCL had four teams playing in urban areas that ranked among the nation's seven largest by 1950 and one more ranked nineteenth.⁶⁴

Moreover, all of those markets were growing rapidly, as Table 3 illustrates.

The PCL moved too slowly, being preempted by the 1958 transfer of two of the three New York City franchises to the two largest West Coast markets once long-distance air travel emerged. During the 1950s the smallest major league market had fallen to twenty-second, and air travel between New York and Los Angeles became more convenient than travel by rail between New York City and Chicago had been in 1900.

In 1960, after Major League Baseball had reduced its supply of multi-team cities to Chicago, the insurgent Continental League announced it would place new teams into many of the unoccupied large markets as well as New York City. As a consequence, in 1961 the American and National Leagues initiated a process of expansion that created the first new teams to be admitted to the major leagues since 1901. Once again baseball preempted an upstart, this time by quickly placing expansion teams in three of the Continental League's most attractive intended markets.

By 2000 major league teams were in all of the nation's top 22 markets (and two smaller ones). Except where going head-to-head with the existing leagues, no entrant could have even one team in any market ranking higher than twenty-third. By contrast, at the founding of the American League, its three unopposed franchises were in the seventh, thirteenth, and fifteenth largest markets. During any season, only a few large U.S. markets remain on the outside, hat in hand, hoping wanly for team movement or expansion, almost begging to finance fancy new facilities.⁶⁵