Competitive Balance

Competitive balance describes the degree of uncertainty about the outcome of sporting events. Economists posit that uncertainty about the outcome of sporting events plays an important role in determining fans’ interest in these events. In order to test this theory, some measure of the degree of uncertainty inherent in sporting events must be developed.

The interest of sports fans in sporting events depends on a number of factors. Fans enjoy watching athletes perform as well as the drama inherent in athletic competition at any level. Additional fan interest stems from watching gifted athletes perform at the highest level possible. But beyond these factors, a part of overall fan interest also depends on the perceived uncertainty of the outcome of the event. Sporting events with a predetermined outcome have no uncertainty and generate relatively little fan interest; closely contested sporting events have a high degree of uncertainty about the outcome and generate a relatively large amount of fan interest. Sporting events with a high degree of uncertainty of outcome are said to be “competitively balanced,” and sporting events with a low degree of uncertainty of outcome are said to be “competitively imbalanced.” Economists call the theory that fan interest varies with competitive balance the uncertainty of outcome hypothesis.

In general competitive-balance measures could be applied to any sporting event, from individual events like foot races, horse races, or figure skating to team competitions like the World Cup or America’s Cup yacht race to hybrid events that combine elements of both individual and team events like the Tour de France. In practice competitive-balance measures are most frequently applied to sports leagues like the National Basketball Association or the English Premiership League (football/soccer). In the context of a sports league, uncertainty exists about both the outcome of individual games or matches and the final standings in the league. Competitive-balance measures are commonly applied to end-of-season outcomes in sports leagues.

Competitive Balance in Sports Leagues

The owners of teams in professional sports leagues and the organizations that oversee amateur sports leagues have an interest in competitive balance, although the perspective of these individuals and organizations differs from that of sports fans. The long-term success of sports leagues depends on the interest of fans, and owners and regulators have a vested interest in organizing sports leagues in a way that actively promotes the staging of events with highly uncertain outcomes.

Unlike leagues, individual teams have no natural interest in promoting competitive balance.Winning is the primary objective of any sports team. Sports leagues throughout the world contain examples of “dynasties”— teams that enjoy prolonged periods of success— suggesting that success in sports leagues may have a self-perpetuating component. In professional sports leagues, successful teams generate greater revenues, allowing these teams to hire better players and coaches. In amateur settings successful teams attract more talented players than unsuccessful teams. Although diminishing returns to talent tend to mitigate these advantages to some extent—a team will not find it’s always advantageous to have the absolute best player at every position—some tension exists between the goals of individual teams and the goals of sports leagues. This tension often leads to a reduced competitive balance in sports leagues.

In addition professional leagues that grant individual teams exclusive geographic franchises—a common practice in North American sports leagues like the National Football League (NFL), the National Basketball Association (NBA), and Major League Baseball (MLB), as well as other professional sports leagues around the world—will have revenue disparities due to differences in market size that may also lead to competitive imbalances. These differences in market size may also be related to the existence of “dynasties” in professional sports leagues.

Research on Competitive Balance

Decision makers in professional and amateur sports leagues, sports management professionals, and economists have considerable interest in testing the validity of the uncertainty of outcome hypothesis and in understanding the effect of uncertainty of outcome on fan interest in sporting events. This interest spurs a large amount of research on competitive balance. The competitive-balance literature can be divided into two broad categories: measurement of competitive balance and tests of the uncertainty of outcome hypothesis.

Measuring Competitive Balance

Competitive balance depends on the perceived uncertainty about the outcome of sporting events. Uncertainty is a complex, multifaceted phenomenon that is difficult to reduce to a single quantifiable metric. Consequently, the literature contains no consensus regarding a single best measure of competitive balance. A large number of alternative measures of competitive balance exist, each with a different set of strengths and weaknesses.

The dispersion of winning percentages—the fraction of the total games played that are won by a given team within a sports league—is the most commonly used measure of competitive balance. At first glance average winning percentage may appear to be a good measure of competitive balance, but because each sporting event generates a win for one team and a loss for the other team (ignoring ties), the average winning percentage in a given league must always equal 0.500.The dispersion of winning percentages around this average reflects how much variation in winning percentage existed in the league, and it is also an indication of the extent of uncertainty of outcome in the contests played in the league.

Competitive Balance within a Single Season

From basic statistical analysis, variance is a common measure of the dispersion of a variable.The variance of winning percentages in a sports league is simply the average of the squared difference between each team’s winning percentage and 0.500. Leagues with a greater degree of competitive balance have a larger variance of winning percentages. In practice researchers studying competitive balance use the standard deviation of winning percentage—the square root of the variance— because the standard deviation is expressed in the same units: percent of games won as winning percentage.

A number of other single-season measures of competitive balance have been developed. These methods include Gini coefficients for wins, Lorenz curves, and Markov chain methods. All of these measures share a common feature: They describe the distribution of wins across teams in a sports league in a single season.

Between-Season Competitive Balance

Because sports leagues exist for many seasons, sports fans may also care about the degree of uncertainty of outcomes for periods longer than a single season. Between-season measures of competitive balance focus on the amount of turnover in the relative success of sports teams from season to season. Unlike singleseason measures of competitive balance, these measures of competitive balance focus on changes in the relative success of teams in a league, as measured by winning percentages or championships, over a number of seasons.

One approach to measuring between-season competitive balance examines the distribution of championships across teams in a sports league over some period of time. Sports leagues within which most championships are distributed among a small number of teams have a smaller degree of uncertainty of championship outcomes and have lower competitive balance; sports leagues within which championships are distributed relatively equally across teams have a larger degree of uncertainty of championship outcomes and have higher competitive balance.

A second approach to measuring between-season competitive balance calculates the dispersion of a single team’s relative level of success over a number of seasons and compares this dispersion with the average level of dispersion in success across all teams in the league.

Level of Competitive Balance in Sports Leagues

Professional sports leagues generally exhibit a wide range of variation in degree of competitive balance. Even within a particular sport, different leagues will have different levels of competitive balance.

Table 1 shows the standard deviation of winning percentages in a number of professional sports leagues for the 2002 or 2002–2003 seasons in cases in which the season spans two calendar years (with the ALF exception noted). Recall that the greater the dispersion of winning percentages, the less competitive balance in that league in the 2002 season.The dispersion of winning percentages varies from a high of 0.22 to a low of 0.08 in this sample. The general pattern across sports indicates that baseball leagues had the most competitive balance, American football and basketball leagues had the least competitive balance, while football and ice hockey fell in the middle of these two extremes.

Competitive Balance Indicators, 2002–2003 Seasons

Note: ALF standard deviation for all-time records, all others single season. Source: Author’s calculations.

A careful reader will note that the pattern of standard deviations across sports leagues in Table 1 varies with the number of games played in each season by the teams in each league. Baseball seasons consist of between 140 and just over 160 games; basketball seasons consist of about 80 games; ice hockey seasons consist of between 50 to 80 games; football (soccer) seasons consist of between 30 to 45 games; and American football seasons consist of between 12 to 16 games. Seasons consisting of more games will generally have a lower standard deviation of winning percentage, no matter what level of competitive balance occurs in the leagues. To account for this, the actual standard deviation of winning percentages can be compared with an ideal standard deviation that would result from an evenly matched league of games playing a season of a given length. This ideal standard deviation can be calculated based on the case in which each team in a league has a 0.500 winning percentage—the maximum possible uncertainty of outcome in a single season—for a given number of games played by the teams in the league by dividing the square root of the number of games into 0.5.

Comparing the actual standard deviation with the idealized standard deviation for each league leads to a somewhat different picture of the relative degree of competitive balance in the leagues shown on the table. On average, football (soccer) leagues have standard deviations about 40 percent larger than the ideal standard deviation, American football leagues are about 45 percent larger, ice hockey leagues about 85 percent higher, and baseball and basketball leagues have standard deviations twice the size of the idealized standard deviation based on length of season.

Remedies for Competitive Imbalance

All of the standard deviations shown on Table 1 are larger than the ideal standard deviation for the league. Because the actual standard deviations exceed this ideal value in all leagues, some degree of competitive imbalance exists in all these leagues. It appears that competitive imbalance is a constant feature of most professional sports leagues. Also, recall that the owners of teams in professional sports leagues have an economic incentive to maintain competitive balance. This tension leads sports leagues to impose rules in order to increase the level of competitive balance. However, most of these rules appear to have little effect.

One common rule aimed at enhancing competitive balance in sports leagues is the reverse-order entry draft. In these drafts the worst teams in terms of winning percentage in the previous season have the first choice of new players coming into the league. Presumably, bad teams will select the best new players and have the greatest chance to improve their performance.

Most evidence suggests that the institution of a reverse-order entry draft has no effect on the level of competitive balance in sports leagues.There are several reasons for this ineffectiveness. First, there can be a considerable amount of uncertainty about the quality of new players coming into a league. If coaches and managers have trouble determining the actual quality of incoming players, there is relatively little benefit to selecting earlier in the draft. Second, the success of the draft depends in part on the ability of the drafters. Bad teams are often run by bad managers and coaches, and less-able decision makers tend to make bad draft choices and bad trades involving draft choices.

Competitive imbalance often stems from imbalances in revenues. Revenue imbalances can come from differences in market size or differences in past success of teams. Some sports leagues attempt to reduce competitive imbalance by reducing the imbalance in revenues. Rules aimed at reducing revenue imbalance include sharing of revenues from ticket sales and revenues from television and radio broadcasts evenly among teams, rather than the team drawing the most fans or located in the largest market keeping these revenues. In some leagues, a “luxury tax” is imposed on teams with the largest payrolls. Finally, many sports leagues impose salary caps on teams in order to reduce competitive imbalance. A salary cap is a maximum payroll for players. Salary caps do not directly address revenue imbalances but instead equalize the amount of money that each team can spend hiring players. By limiting payroll, salary caps attempt to alter the distribution of talent across teams, thus changing competitive balance in the league.

Like the other rules aimed at reducing competitive imbalance, research suggests that salary caps do not reduce competitive imbalance, and in some cases they appear to increase it. One problem with salary caps is that it is very difficult to construct a salary cap system that cannot be manipulated by teams and players in some fashion. Players will often restructure the terms of contracts or agree to artificially low salaries in the early years of contracts and defer higher payments far into the future. In some instances teams have resorted to illegal side payments to players.

The Future

Competitive balance, and the related uncertainty of outcome hypothesis, have important implications for the behavior of sports fans, the owners of teams in professional sports leagues, and the organizers of amateur sports leagues. Fans, owners, and league organizers prefer a high degree of competitive balance in sports leagues. But competitive imbalance appears to be a feature of all sports leagues, and most rules aimed at reducing competitive imbalance appear to be ineffective. Future research on competitive balance will need to address this issue by (1) quantifying the relationship between the degree of competitive balance and fan interest in order to determine how much competitive imbalance should be tolerated, and (2) improving the understanding of why existing rules aimed at reducing competitive imbalance are ineffective and how to construct effective rules to improve competitive balance.

Brad R. Humphreys