5

Competitive Balance
You’re doing all right … We’re ahead but you just can’t tell what those Cumberland players have up
their sleeves. They may spring a surprise. Be alert, men!
—GEORGIA TECH COACH JOHN HEISMAN, speaking to his team at halftime with his team up by 18
touchdowns in the worst blowout in college football history (222–0).1

Introduction
One of the oldest adages in professional football is that on any given Sunday,
each team has a chance to beat the other. But what if, year after year, a few
teams regularly win, while the rest almost always lose? No doubt, the games
would be less interesting. As early as 1956, economists noted that successful
leagues must be based on relatively even competition,2 but the degree of
parity within a sports league can mean different things to different people. To
some, it means close competition every year, with the difference between the
best and worst teams being relatively small. To others, it means regular
turnover in the winner of the league’s championship. Whatever the measure,
we call the degree of parity within a league competitive balance. This
chapter discusses competitive balance from the perspectives of the fan and the
owner. In addition, it explores how economists measure competitive balance,
how leagues try to alter competitive balance in a league, and why such efforts
might not be successful.
Learning Objectives
Understand why owners and fans may care about competitive balance.
Master the use and interpretation of the different measures of
competitive balance.
Be able to compare the tools that leagues use to promote competitive
balance and the limitations of those tools.
5.1 Why Study Competitive Balance?
Fans and owners alike have a conflicted relationship with competitive
balance. No Bears fan enjoys losing to the Packers, but each recognizes that an
occasional loss makes the games more interesting. Similarly, the McCaskey
family, which owns the Bears, gets a greater financial return—and greater
satisfaction—if the team is successful but does not always win. Thus, whether
competitive balance is a good thing depends on one’s perspective. On any
given day, hometown fans want to see wins, so owners want to provide them.
Over longer stretches of time, the viability of the league may depend on some
uncertainty of outcome. In this section, we look more deeply at the value of
competitive balance to fans and owners.
The Fans’ Perspective
Suppose you grew up in Wyoming without much exposure to professional
sports. On a visit to your uncle in Texas, you go to your first MLB game, one
between the Texas Rangers and the Minnesota Twins. The Rangers quickly
build a commanding lead, scoring on the overmatched Twins in almost every
inning. While the Rangers’ fans seem to enjoy the rout, as a casual fan, you
quickly become bored because the talent on the two teams is very uneven. If
that were your only exposure to baseball, you would probably leave the game
thinking it is all a waste of time. Your uncle, however, feels great about the
big win for the home team. On the way home, he says, “I’m glad the Rangers
won. A close game might have been more fun, but a Rangers loss would have
been awful.” Even this short example shows that fans view competitive
balance differently from casual observers.
From the fan’s perspective, an uncertain outcome in a game may be more
interesting than a foregone conclusion, but, as we discuss below, fans may
prefer an easy win to seeing their team lose in an upset. Historically, fans
have shown their displeasure with unbalanced competition, even when their
own team did most of the winning. The Cleveland Browns of the late 1940s
were so dominant in the All-American Football Conference that they became
less popular with their home fans.3 In baseball, the New York Yankees may
have dampened attendance at their own games and across the American
League when they won eight league pennants and six World Series between
1950 and 1958. Table 5.1 shows that between 1950 and 1958, a period generally
marked by prosperity and economic growth, attendance for both Yankee
games and the American League as a whole either stagnated or fell as the
Yankees dominated the league. In contrast, in the National League, which had
four different champions (Phillies, Giants, Dodgers, and Braves), attendance
grew substantially. This relationship between competitive balance and
attendance requires that we look more deeply into the question of the
desirability of competitive balance.
Table 5.1 New York Yankees’ Success and American League and National League Attendance,
1950-1958
Studies of fan behavior show interesting asymmetries in the benefits (utility
gained) of winning versus the costs (utility lost) of losing. Several empirical
studies show that fans enjoy an uncertain outcome and are most likely to
attend games in which the home team has a 60 to 70 percent chance of
winning.4 Fans do not want their teams to lose 30 to 40 percent of the time,
but the chance of losing makes the game more interesting and stimulates
demand. This is known as the uncertainty of outcome hypothesis (UOH).
Recent research by Dennis Coates, Brad Humphreys, and Li Zhou casts doubt
on the universal applicability of the UOH. They use concepts from behavioral
economics to investigate whether the UOH applies in some circumstances but
not others. Behavioral economics incorporates elements of psychology in
analyzing decision-making. Though elements of behavioral economics were
developed over 40 years ago, its power as an economic tool was popularized
by the work of Daniel Kahneman and Amos Tversky in the 1980s. More
recently, economists have found widespread applicability to their theories
regarding human behavior and economic action.
One of the most important concepts of behavioral economics is prospect
theory. Prospect theory was developed by Kahneman and Tversky to analyze
settings in which individuals treat gains and losses differently. Rather than
acting in accordance with the strict predictions of economic rationality,
individuals tend to show greater aversion to losses than desire for gains, a
concept known as loss aversion.5
Coates et al. construct a model that incorporates these elements of behavioral
economics. Unlike the strict UOH model, which posits an optimal probability
of winning, with increasing and decreasing probabilities having a symmetric
effect on a fan’s expected pleasure from attending a game, Coates et al.
present a more complex set of motives. They begin by claiming that,
consistent with reference-dependent preferences (i.e., preferences that may
differ depending on how the spectators feel about their team’s chances going
into a match), fans prefer a win of any kind to a loss. They then posit a
variation of the UOH by saying that unexpected wins are sweeter than
expected wins and that unexpected losses are worse than expected losses.
Finally, they incorporate loss aversion by hypothesizing that being upset
reduces utility more than staging an equally improbable upset increases
utility. Their estimates using game-level data for the 2005–2010 baseball
seasons support their hypotheses.6
The Owners’ Perspective
Perhaps the briefest explanation of why competitive balance matters to
owners and the leagues in which they compete is that, as seen in Table 5.1, it
matters to fans. This is particularly true for teams that rarely win. Overall
American League attendance fell by about 2 million during the era of Yankee
dominance in the 1950s, while National League attendance rose by 2 million.
While all owners would rather win than lose any given game (and so may
also be loss-averse), they are also concerned with the health of the league. If
the same team wins every game or every championship, fans may lose
interest. In Chapter 3, we saw that owners may have a variety of objectives,
such as maximizing profits or maximizing wins. No matter what objective
individual team owners have, consistently unbalanced competition alienates
fans and is not in the best interest of the league.
Recall that leagues perform functions that individual teams either cannot do
or have no incentive to do. Hence, leagues often adopt policies to promote
competitive balance because individual teams lack the means or the
motivation to do so. As we will see later in the chapter, the concern with
unbalanced competition and the will to adopt rules designed to increase
competitive balance vary greatly across leagues and national borders. In
particular, we shall see that European soccer leagues, such as La Liga, have
long been characterized by highly unbalanced competition.
One challenge of maintaining competitive balance is that it may not be the
natural state of the league. Leagues and professional associations do not need
to take specific action if they tend naturally toward equal strength. If, in the
absence of intervention, a few teams flourish while most teams languish, the
league must create a more competitive environment. For example, in auto
racing, NASCAR goes to great lengths to promote equal competition among
cars. It measures each car in a Sprint Cup race with a device known as “the
claw,” to ensure the dimensions of the car bodies comply with regulations that
prevent cars from having an aerodynamic advantage. Engines must meet an
exacting set of criteria and are even restricted to lower horsepower on larger
tracks, where speeds are greatest. To further ensure even competition, some
cars are retested at the conclusion of the race. NASCAR places so many
restrictions on cars to promote close competitions decided by the skills of the
drivers and their teams.
Intervention might be necessary in team sports for two reasons: the
motivation of team owners and market size.7 If some owners maximize wins
while others maximize profits, competitive balance can suffer, harming all
teams financially. Even if all owners have the same motivation, in the absence
of revenue sharing, the financial return to investing in talented players is
likely greater for teams from large cities. As the attendance data in Table 5.1
show, actions that maximize individual team profits or wins may be
detrimental to competitive balance.
The Effect of Market Size
There is considerable disagreement among those who have studied the impact
of market size on competitive balance. There appear to be three primary
sources of contention: what to use as a measure of success, how to
characterize market size, and how to measure the impact of policies designed
to alter competitiveness, such as revenue sharing. Studies that use different
measures of success or market size come to very different conclusions. For
example, MLB’s Blue Ribbon Panel reported in 2000 that low-payroll teams
(their measure of market size) rarely succeeded in the playoffs (their measure
of success) from 1995 to 1999. However, Berri, Schmidt, and Brook (Wages of
Wins, 2006) note that success in the playoffs is different from success in the
regular season, and, using a longer data-set, find only a weak correlation
between payroll and wins.8 Other work by Schmidt and Berri tests several
measures of market size and concludes that team quality is not strongly
related to population or per capita income.9
Determining market size is particularly complex, in part because the way in
which markets are defined continues to evolve. Before the television era,
teams relied almost solely on gate revenue, so attendance was most relevant
and the financial advantage of large cities was limited. The growth of regional
sports networks (RSNs) (and the revenue they provide) has created an income
differential that is not limited by the size of a stadium. As a result, teams in
large markets can now generate much more revenue for each additional win
than teams in small ones can.10 The impact of their greater revenue on
competitive balance, however, remains a point of debate.
To further complicate matters, even if fans desire some level of uncertainty,
our conclusions regarding how to best distribute success change when we
change our unit of analysis from the team to the fan. Because teams in large
markets (such as the New York Yankees) have many more fans than teams in
small markets (such as the Milwaukee Brewers), a profit-maximizing league
prefers that the teams in the largest markets win more often than teams
elsewhere.11 In a 30-team league, perfect parity means that the Yankees and
Dodgers—teams in the two largest markets—win the World Series only once
every 30 years on average. If championships were allocated so that they were
distributed equally on a per capita (per fan) basis, rather than a per-team
basis, the Yankees would win about once every 9 years, the Dodgers once
every 14 years, and the Milwaukee Brewers almost once a century. Over the
100 World Series from 1917 to 2016, the Yankees have won more than twice as
often as they would if wins were allocated evenly by population (26 v. 11), the
Dodgers come up a little short (6 v. 7), and the Brewers are still waiting, with
n0 World Series wins in their 48 years of existence.
To see why big-market teams gain more from winning than small-market
teams do, assume that each team gets its revenue only from tickets and local
television revenue, and that teams benefit from having a higher winning
percentage, but the additional benefits of increasing the winning percentage
become smaller as it approaches 1.000. Thus, the marginal revenue curve from
additional wins is positive but downward sloping. Because teams in large
cities enjoy greater increases in fan support from an additional win than
teams in small cities, an additional win generates more gate revenue, more
media revenue, and more venue revenue for a team in Los Angeles than it
does for a team in Indianapolis. Figure 5.1 illustrates the greater value of wins
for a team in a large market. It shows the additional (marginal) revenue from
one more win for a team in a small market (MRs) and a team in a large market
(MRL)12
Figure 5.1 The Benefit of Improving the Team Depends on Market Size
Because winning is more valuable to a big-market team, it pursues winning more avidly than a smallmarket
team does.
If all firms maximize profits then they operate where the marginal revenue of
wins equals the marginal cost of wins. To keep the focus on revenue, assume
that the marginal cost of a win is constant and equal for all teams. For a
small-market team, MR = MC occurs at WS wins. A large-market team has an
incentive to acquire more talent and wins WL > WS games. Thus, even in a
world without Mark Cubans or Jerry Joneses, where all teams maximize
profit, the model predicts that, all else equal, teams from big cities win more
frequently than teams from small cities.
The Influence of Diminishing Returns
While teams from large cities may have a greater incentive to acquire talent
and win games than teams from small cities, there is still a limit to how far
they go in pursuit of wins because of the law of diminishing marginal returns.
Diminishing returns to labor are found in every industry. In the short run, as a
firm adds units of labor, the marginal product (the additional output) of the
last unit of labor must eventually fall, even if labor is homogeneous. The
reason is straightforward: In the short run, capital is fixed, so the additional
workers eventually have less capital to work with and are less productive.
In the context of sports, diminishing returns may set in very quickly,
especially in basketball, where only five teammates play at a time, and, as the
saying goes, “there is only one ball.” The same logic applies across all team
sports. With rare exceptions, professional football teams have only one
quarterback on the field at a time. The New England Patriots would not
dispute that Cam Newton is a great player, but his value to the Patriots is
surely less than the salary that a team without a top quarterback would offer,
given that the Patriots already have Tom Brady.
Diminishing returns act as a brake on team behavior because they reduce the
incentive any one team has to stockpile talent. It does not make economic
sense for a team to spend large sums of money acquiring all the best players at
each position when some, perhaps many, of them will contribute very little.
As we will learn later in the chapter, even with the influence of diminishing
returns, variations in market size and owner motivations have led leagues to
seek additional ways to equalize team quality.
A Brief History of Competitive Balance
In MLB, five different teams won the five World Series played between 2012
and 2016. But the league has also seen long periods of dominance by the New
York Yankees. The Yankees have been a dominant franchise since the 1920s,
when they won six American League championships between 1921 (the year
after they acquired Babe Ruth) and 1928. It was even more pronounced when
they won five straight World Series between 1949 and 1953. Two of the three
other major sports have similar instances of a single team dominating the
league over many seasons. The Boston Celtics won every NBA championship
but one between 1959 and 1969. Between 1965 and 1979, the Montreal
Canadiens won the NHL’s Stanley Cup ten times. The Canadiens’ dynasty
was followed by that of the New York Islanders, who won the Cup the next
four years in a row. Only in the NFL has no team ever won the league
championship (the Super Bowl) more than twice in a row, but even there, the
Steelers, Cowboys, Patriots, and 49ers have each won at least five times while
two teams—the Browns and the Lions—have never appeared in a Super Bowl,
despite being in the NFL for all 51 of them.13
Internationally, unbalanced competition is the norm in the elite European
soccer leagues. A few dominant teams have long runs of championships. For
example, from the 2000–2001 season through the 2015–2016 season, FC
Barcelona and Real Madrid combined to win 13 of 16 La Liga championships
in Spain, and Bayern Munich and Borussia Dortmund won 13 of 16
Bundesliga championships. The EPL in England and Serie A in Italy show
similar patterns of dominance by one or two teams over many seasons.
Competitive imbalance also appears at the amateur level. For example, since
the formation of college football’s Big 10 Conference in 1896, the University of
Michigan has won or shared 42 titles, more than one-third of the total. In
contrast, Indiana University has won only two titles, five fewer than the
University of Chicago, which gave up football in 1939.
Changes in the relative importance of the various revenue sources and the
growth of the sports industry in general have increased the concerns that the
financial consequences of unbalanced competition are becoming more severe.
Before we analyze the potential implications of unbalanced competition, we
first need to consider the various ways in which it is measured.
5.2 Measuring Competitive Balance
There are two approaches to measuring competitive balance.14 The first
focuses on team performance over the course of a given season. A wide
dispersion in winning percentages means that some teams are much better
than others in that season. The second approach looks across several seasons.
It measures the concentration of championships or turnover in the league’s
standings over a given period. Leagues with a high concentration of
championships have a small set of teams winning year after year. No single
approach is necessarily better than the other. To fully evaluate competitive
balance in a league, one should use both measures. While fans may prefer
lopsided wins to unexpected losses in individual games, fans and owners are
likely to desire both tightly contested seasonal competition and regular
turnover in champions.
Within-Season Variation
In a way, sports fans resemble opera or classic rock buffs. Both are attracted
by the absolute quality of the performers. A serious music fan would much
rather see Eric Clapton than even the best local band. Similarly, the demand
to see a big-time college football game between Texas and Oklahoma far
outstrips the demand to see a game between Ivy League powers Harvard and
Penn. However, absolute quality is not the only factor; relative quality matters
too. As seen in Table 5.1, a tight race for the conference championship will
generate more demand for games than a season in which the winner is a
foregone conclusion. Within-season variation in winning percentage focuses
on the evenness of competition over the course of a season.
Measures of within-season variation start with the standard deviation of
winning percentage. The standard deviation is the distance of the typical
observation from the sample mean. In the case of a professional sports league,
ignoring ties, there is a loser for every winner, so the mean winning
percentage for a league must be 0.5.15
The formula for the standard deviation of winning percentages within a single
season is:
where WPCTi, t is the winning percentage of the ith team in the league in year
t, .500 is the average winning percentage of all teams for the year, and N is the
number of teams in the league. The larger the standard deviation, the greater
the dispersion of the winning percentages. For example, consider the final
standings for MLB’s 1996 season. We use this season because it was before
inter-league play began in 1997, ensuring that the mean winning percentage
in each league is exactly 0.500. Table 5.2 shows the final standings for both the
American and National Leagues. One can get an impression of how balanced
the two leagues were by looking at the highest and lowest winning
percentages. In the American League, the Cleveland Indians had the highest
winning percentage (.615), while the .327 winning percentage of the Detroit
Tigers was the lowest. The Atlanta Braves led the National League with a .593
winning percentage—with three fewer wins than the Indians. At the bottom of
the standings, the last-place Phillies won 14 more games than the Tigers.
Computing the standard deviation of winning percentage confirms this
impression. The standard deviation in the American League was 0.067,
meaning that the typical team’s winning percentage varied by 0.067 from the
mean. In contrast, the standard deviation in the National League was 0.054,
about three-fourths that of the American League.
Table 5.2 Winning Percentages for the National and American Leagues, 1996
The standard deviation of winning percentages is a useful summary of
competitive balance, but it has significant limitations. In particular, the
standard deviation of winning percentages varies with the number of games
in a season. To see why, try flipping a fair coin—one with an equal chance of
coming up heads or tails—four times. This is the equivalent to two equally
matched teams playing one another four times, with only random factors
determining the outcome. More than 12 percent of the time (about once in 8
tries), you will get an extreme outcome of all heads or all tails.16 As you
increase the number of flips to 40, 400, or 4,000, the chance of an extreme
outcome becomes more and more remote. This experiment tells us that, even
if a league were perfectly balanced, we might see some teams with many
more wins than others in a short season. As the season gets longer, winning
and losing streaks begin to offset one another, just like runs of heads and tails
with coin flips.
Table 5.3 Dispersion of Winning Percentages, 2015–2016
League Actual Ideal Ratio
MLBa .065 .039 1.67
NFLa .197 .125 1.58
NBA .166 .056 2.96
NHL .077 .056 1.38
English Premier League .132 .081 1.63
Bundesliga .149 .086 1.73
Serie A .147 .081 1.81
La Liga .155 .081 1.91
a MLB and NFL results are for the 2016 season.
b Because teams receive one point for overtime losses and EPL teams receive one point for draws,
winning percentages in the NHL and EPL are computed as the percentage of possible points and means
are not equal to 0.5.
Sources: All data are generated from the standings on the official league websites and skysports.com.
Table 5.3 shows the standard deviation of winning percentages in the four
largest North American leagues and four major European soccer leagues. If
we consider only the standard deviations, it appears that the NFL has the least
balanced competition. Alternatively, it might simply be that the NFL has only
a 16-game season, by far the shortest of any league. Because the standard
deviation is larger for sports with shorter seasons, we do not directly compare
standard deviations. Instead, we first compute what the standard deviation of
winning percentage would be for a league that had completely equal teams.
The ratio of the actual standard deviation to this “ideal” standard deviation
tells us how far out of balance the league was. Because the ratios for MLB and
the NFL account for the length of the season, we use them to compare
competitive balance in the two leagues.
The standard deviation that corresponds to a world in which each team has a
0.5 chance of winning each game is:
where 0.5 indicates that each team has a 0.5 probability of winning, and G is the number of games
each team plays.17 Because each MLB team plays 162 games per season, the ideal is 0.039. Because
NFL teams play only 16 games, a randomly occurring string of wins or losses has a greater impact on
a team’s final winning percentage, so the ideal standard deviation is much larger, 0.125. In the NHL
and NBA, whose teams play 82-game schedules, the standard deviations are 0.056. In all but the
Bundesliga, which has a slightly shorter season, the ideal standard deviation in the European leagues
is 0.081.
To measure competitive balance within a single season, we use the ratio (R) of
the actual standard deviation of winning percentages (σw) to the ideal
standard deviation (σl18
Thus, for the NBA in 2015–2016:
R = 0.166/0.056 = 2.96
Based on this result, the standard deviation of winning percentages in the
NBA is close to three times what it would be in a world with absolutely
balanced teams. Again, this result is consistent with our casual observation
that competition appears unbalanced in the NBA, as two teams had winning
percentages of over .800, five teams had winning percentages over .670, and
four teams had winning percentages less than .300.
In addition to the actual standard deviations, Table 5.3 presents the ideal
standard deviations for each of the eight major sports leagues and the ratio, R.
In 2015–2016, the NHL was the most equally balanced North American
league, with an R-value of 1.38. The NFL and MLB followed with R-values of
1.58 and 1.67. Competition in the top divisions of European soccer was less
balanced than most North American sports. It is notable that the EPL was the
most balanced European league as it has the most extensive revenue sharing
of the European leagues, a subject we return to later in the chapter. As noted
previously, the NBA, with R = 2.96, is by far the least balanced of the eight
leagues.
In The Wages of Wins, Dave Berri, Martin Schmidt, and Stacey Brook discuss
why competition in the NBA is so unbalanced.19 Their theory is based on the
old adage “you can’t teach height.” In basketball, taller players have a distinct
advantage over shorter ones. There are good players who are not tall, but if
we compare two players of equal skill but substantially different heights, the
taller player will be more effective. The number of very tall people who are
also very gifted athletes—and whose athletic skills are well suited for
basketball—is extremely small. Thus, a team lucky enough to get an unusually
gifted player, such as LeBron James, has a competitive advantage. The small
number of players who are on the court at any one time adds a premium to
rare talent in basketball. In soccer, even an extraordinarily gifted player such
as Lionel Messi represents only one-eleventh of his team’s resources at any
one time.
Between-Season Variation
For baseball fans everywhere, spring is a special time of year that brings the
promise of a new baseball season and the chance that “this could be the year”
that their team wins it all. Across seasons, competitive balance implies that
each team can realistically compete for the championship. This type of
competitive balance is called turnover, or team-specific variation. It differs
from within-season variation in that it considers the change in the relative
positions of the teams in the standings each year rather than the distance
between teams in each season. Brad Humphreys (2002) defines team-specific
variation for a team as:
where T is the number of seasons, and WPCT is the team’s average winning
percentage over the T seasons.20 The larger σT, i becomes, the more a team’s
fortunes change from year to year. If team i always finished with the same
record, sT, i would be zero. If fans support only teams that have a reasonable
chance of winning their division or conference, variation across seasons is
vital to maintaining fan interest over long stretches of time. If σT, i were zero
for all teams, we would know how all teams would perform before the season
even started. Such a situation would surely reduce demand for all teams.
One frustrating aspect of using the variation between seasons is that, unlike
the within-season standard deviation, there is no obvious standard of
comparison. We cannot say whether fans or owners care more about how
much their team’s winning percentage varies across the years or how their
team’s position changes relative to other teams. For example, would
Philadelphia hockey fans be happier if the Flyers had a very good record
instead of a mediocre record but finished second to the New Jersey Devils
every year or if they won the Stanley Cup once every few years but finished
in last place every other year? Though turnover is important, the absence of
an absolute standard means that team-specific variation is useful only as a
relative measure of dispersion (when comparing one time period with another
or one sport with another).
Frequency of Championships
We can also evaluate competitive balance by looking at the frequency with
which teams win successive championships. At one extreme, if the San Jose
Sharks win the Stanley Cup almost every year, then the winning percentages
of the teams in the league may not matter as much, since the ultimate
outcome is seldom in doubt. At the other extreme, if a different team wins the
Stanley Cup every year, then one can argue that competition is balanced
regardless of how unequal the teams’ records are. This criterion is similar to
the turnover criterion discussed above, but it relates to championships rather
than regular season standings.
The data in Table 5.4 provide an interesting contrast to the within-season data
in Table 5.3. While the NHL appears to be the most competitive league within
a single season, it is the least balanced league in North America when it
comes to championships. Just 3 of the 30 teams won 70 percent of the
championships in the 10-year period ending in 2016. In the NBA, the withinseason
competitive balance was by far the worst of any league in Table 5.3,
but Table 5.4 shows no team winning more than two championships. As noted
in the previous section, the distribution of championships in the European
soccer leagues is highly skewed toward a few dominant teams. In all four
leagues, just one team won at least half the championships over the ten-year
span. In many of those championships, the runner-up is also one of the teams
featured in the table. For example, in La Liga, nine of the ten championships
listed in the table featured either Barcelona or Real Madrid as the runner-up.
No matter how balanced the competition was over the course of the season,
either Barcelona or Real Madrid probably won the league championship. To
the extent that teams can “buy championships” because they have more
revenue than their competitors, differences in market size and team popularity
may be to blame. Recall from Chapter 3 that most European leagues share a
much lower percentage of revenue than North American leagues. We return
to the influence of revenue sharing on competitive balance later in the
chapter.
Table 5.4 Distribution of Championships, 2006–2007 to 2015–2016
NBA NFL MLB NHL
Miami 2 NY Giants 2 San Francisco 3 Chicago 3
LA Lakers 2 New England 1 St. Louis 2 Pittsburgh 2
San Antonio 2 Denver 1 Chicago Cubs 1 LA Kings 2
Cleveland 1 Seattle 1 Kansas City 1 Boston 1
Golden State 1 Baltimore 1 Boston 1 Detroit 1
Dallas 1 Green Bay 1 NYYankeesi Anaheim 1
Boston 1 New Orleans 1 Phillies 1
Pittsburgh 1
Indianapolis 1
HHI = 0.16 HHI = 0.13 HHI = 0.18 HHI = 0.20
Bundesliga La Liga EPL Serie A
Bayern Munich 6 Barcelona 6 Man. United 5 Juventus FL 5
Borussia Dortmund 2 Real Madrid 3 Man. City 2 Internazionale 4
VflWolfsburgi Atletico Madrid 1 Chelsea 2 AC Milan 1
Vfb Stuttgart 1 Leicester City 1
HHI = 0.42 HHI = 0.46 HHI = 0.34 HHI = 0.42
The Herfindahl-Hirschman Index
Counting the number of teams that win a championship in a given period
suggests that the NHL is more balanced than the NFL and that the EPL is
more balanced than Serie A. Unfortunately, simply counting the number of
teams that have won in a given time period does not provide a single value
that measures this variation. Fortunately, there is a statistic that does. The
Herfindahl-Hirschman Index (HHI) was originally developed to measure the
concentration of firms in an industry, but sports economists use it to measure
the concentration of league championships.
We calculate the HHI by counting the number of championships (ci) team i
won within a given period, dividing by the number of years in the period (T),
squaring this fraction, and adding the fractions for all teams:
The maximum, 1.0, indicates perfect imbalance. If the number of years
exceeds the number of teams in the league or conference, then the minimum
value of the HHI is 1/N, where N is the number of teams in the league. For
shorter periods of time, the minimum value is 1/T. Thus, the value of the HHI
for the NHL (0.13) is very close to the minimum possible level (0.10) for a tenyear
period. The NHL could not have been much more competitive by this
standard. To see this, consider two leagues, each with five teams. In one
league, each team has won two championships over the last ten years. In the
other league, one team has won all the championships. The HHI for each
league is:
The HHI for the eight leagues in Table 5.4 appears beneath each list of
champions. As expected, the HHI values for the European leagues are far
greater than for any North American league. This lack of competitive balance
combined with the extraordinary popularity of European soccer provides
additional evidence that fans may be less concerned with competitive balance
than one might think.
Illustrating Competitive Imbalance
Thus far, we have measured competitive balance with purely statistical
measures. In this section, we show how to express competitive balance
graphically. To do so, we use an economic tool known as the Lorenz curve.
Initially created to show income inequality, the Lorenz curve illustrates how
evenly distributed any resource or characteristic is in a population.
The Lorenz curve is often used to show income inequality by illustrating the
percentage of income earned by a given percentage of the population. For
example, if the poorest 10 percent of the population earns just 3 percent of
total income and the richest 10 percent of the population earns 25 percent of
total income, we can see that income is highly concentrated among the
wealthy. In this application, we show the cumulative winning percentage of
the teams in a league, divided into deciles. In a league with perfectly equal
competition, 10 percent of the teams would earn 10 percent of the wins (or
points), 20 percent of the teams would earn 20 percent of the wins, and so on.
The more unbalanced the competition, the smaller the percentage of wins or
points earned by the weakest teams.
Consider the 2015–2016 regular season in the NHL and La Liga. Because
draws are possible and NHL teams earn points for overtime losses, we analyze
the number of points earned in each decile as a fraction of the total points
accumulated in each league in that year rather than counting wins. We know
from Table 5.3 that in 2015–2016, the NHL was the most balanced and La Liga
was one of the least balanced of the eight leagues we considered. The three
weakest teams (i.e. the lowest 10 percent) in the NHL combined to earn about
7.8 percent of the total points earned in the NHL that season, and the weakest
20 percent earned 16.3 percent. In La Liga, the weakest 10 percent of teams
earned only about 6.5 percent of the points, and the weakest 20 percent earned
just 13.8 percent, far less than the 20 percent we would observe with perfect
equality. Plotting these points for each decile up to the 10th (100 percent of the
wins) creates the Lorenz curve. In Figure 5.2, we see the Lorenz curves for
both the NHL and La Liga. Notice that the curve for La Liga lies below that of
the NHL. The less balanced the competition, the more the curve bows
downwards. For comparison, Lorenz curve graphs include the line of perfect
equality, a straight 45-degree line from the lower left to the upper right. Along
this line, each decile of teams earns 10 percent of the wins. However, while
the Lorenz curve provides an easy-to-read graphic of competitive balance, it
makes no claim regarding the desirability of this outcome. As discussed
earlier, the optimal distribution of wins may well favor larger-market teams.
Figure 5.2 The Lorenz Curves for La Liga and the NHL, 2015–2016
The lower Lorenz curve for La Liga shows that competitive balance was worse than in the NHL in 2015–
2016.
In sum, there are many ways to measure competitive balance, and no single
method should be regarded as most appropriate. To fully grasp the state of
competitive balance in a league, one must consider intra-season balance—the
spread of winning percentages across teams—and inter-season balance, which
includes both the turnover of teams in the standings and the frequency of
championships. In the next section, we discuss ways in which leagues attempt
to alter competitive balance.
5.3 Attempts to Alter Competitive Balance
All major North American sports leagues have developed policies designed to
promote competitive balance. In this section, we discuss three such policies:
revenue sharing, salary caps and luxury taxes, and the reverse-order draft.21
The draft is meant to equalize competition over time by allocating new talent
in a systematic way toward weaker teams. Revenue sharing and salary
limitations are designed to limit the advantages of big-market teams by
reducing the benefits or increasing the costs that big-market teams face when
pursuing talent. Whatever impact these policies have on competitive balance,
many players believe that the true goal is to depress salaries. We analyze the
impact of these policies on salaries in Chapter 9. In this section, we focus on
their impact on competitive balance. Before we begin, we discuss an
important economic principle that speaks directly to the effectiveness of such
efforts.
The Invariance Principle
Many of the tools we consider in this section are designed to limit the impact
of free agency (the right of a player to sell his services to the highest bidder)
on the distribution of talent. Team owners have consistently asserted that free
agency and competitive balance cannot coexist. Economic theory, however,
says that free agency should have no impact on competitive balance. A basic
principle of economics is that freely functioning markets distribute resources
to where they are most highly valued. Changing property rights—the
ownership or control of resources—affects who gets paid but not where the
resources are employed. The fact that the allocation of resources does not vary
when property rights change is known as the invariance principle, which
was first applied to sport over 50 years ago by Simon Rottenberg.22
To see how the invariance principle works, consider how the International
Olympic Committee (IOC) might allocate tickets for the 2020 Summer
Olympics in Tokyo. Japanese officials are confident that the opening
ceremonies can attract far more than the 50,000 tickets available at the new
National Stadium in Tokyo. As seen in Figure 5.3, the opening ceremonies
would sell out if tickets cost $150. (For simplicity, we assume that all seats are
equally valuable.) Suppose the IOC decides that $150 is too much to charge
and declares that tickets will cost only $10. At a price of $10, however, 500,000
people are willing and able to buy tickets. At that price, the IOC must ration
the tickets. It could, for example, institute a lottery in which 50,000 people are
chosen at random. Those lucky winners can buy tickets for $10.
Figure 5.3 The Market for Tickets to the Opening Ceremonies at the 2018 Olympics
At the equilibrium price of $150, the IOC can sell exactly 50,000 seats. At a price of $10, there is an
excess demand of 450,000.
The invariance principle says that, as long as people are free to transact with
one another, the final allocation of tickets will be the same under both the free
market and the lottery. Suppose, for example, that Barry is willing to pay $300
for a ticket, while Ann won a ticket, which she values at $20. As we saw in
Chapter 2, Barry and Ann can become better off if Barry buys the ticket at a
price between $20 and $300. For example, if Barry pays Ann $150, he enjoys a
consumer surplus of $150 and Ann enjoys a producer surplus of $130.
More generally, if people are free to transact with one another, there are two
results. First, the market price of a ticket rises from the official price of $10 to
the free market price $150. Second, because the market price is $150, only
consumers who place a value of at least $150 on a ticket attend the game, just
as with the free market solution. By transferring the property rights to the
ticket holders, the IOC has not altered the ultimate distribution of tickets. It
has simply changed who gets paid for them.
In the context of professional sports, consider the case of Bryce Harper, the
star outfielder for the Washington Nationals, who will be a free agent in 2018.
One of the top young stars in baseball, Harper is a popular draw and would
add greatly to the gate and media revenue of any team for which he plays. In
a midsize city like Washington, DC, he might generate $15 million per year in
additional revenue. However, in a much larger, wealthier metropolitan area,
such as greater Los Angeles (which has twice the population of the
Washington, DC, consolidated metropolitan area), Harper would add
considerably more. As a free agent, Harper will be able to sell his services to
the highest bidder, and the Nationals may well lose out to the Dodgers
because they will be willing to pay Harper as much as $30 million per year.
Consider now what would happen in a world in which players are not free to
move but profit-maximizing owners were free to sell the rights to players’
services for cash or payment in kind. The Dodgers would still get Harper
because they were willing to pay the Nationals more than he was worth to
DC but less than he was worth to Los Angeles. As a result, both teams profit.
This example shows that property rights do not affect where top players like
Bryce Harper play. They do, however, determine whom the Dodgers pay to
obtain his services. Under free agency, the player owns the rights to his
services. In such a world, the Dodgers lure Harper from DC to Los Angeles by
offering him a higher salary. If Harper were unable to move freely, the
Dodgers would not have to worry about enticing Harper to Los Angeles.
Instead, they would have to pay the Nationals.
In some situations, the invariance principle fails to hold. For our purposes, the
most important situation occurs when there are substantial transaction costs.
As the term suggests, a transaction cost is the expense of dealing with a
buyer or seller in a market in addition to the price one pays for the good or
service. Returning to our Olympics ticket example, suppose Barry had to take
out a $200 advertisement in the newspaper to find Ann. He would face a
market price of $150 plus a transaction cost of $200. Because the ticket is
worth only $150 to Barry, the transaction cost would discourage Barry from
buying the ticket. More generally, when there are substantial transaction
costs, resources might not flow to their most valued use. In this section, we
analyze several ways in which leagues have imposed transaction costs that
prevent the free flow of players to their most valued positions.
Revenue Sharing
As we showed in Chapter 3, revenue sharing equalizes teams’ profits. Some
team owners also claim that they need revenue sharing to equalize the
demand for talent. As the magnitude of revenue and the variety of revenue
sources continue to increase, leagues have initiated an increasingly
sophisticated array of revenue-sharing mechanisms that typically treat leaguewide
revenue differently from local revenue. Because revenue sharing is
explained in detail in Chapter 3, we review it only briefly here.
The NFL has extensive revenue sharing. Teams share equally in the large
national television revenue. With very little local media revenue, this has been
a strong equalizer. While less important than it used to be, the 60–40 split of
gate revenue described in Chapter 3 still has a big impact.
The most recent NBA contract substantially increased the amount of revenue
shared by teams.23 In addition to sharing national television and sponsorship
revenue, all teams submit up to 50 percent of their revenue, less certain
expenses, to a common pool, from which they receive an amount equal to the
average payroll. Teams that contribute less than the average payroll are net
receivers of income, while high-revenue teams, which contribute more, are
“net payers.”24
Since the early 2000s, MLB has significantly increased the degree of revenue
sharing. Beginning with the 2002 labor agreement, teams now pay 31 percent
of their net local revenue (which includes gate and local broadcasting
revenue) into a pool that is divided equally among the 30 teams. They also
redistribute a portion of the Central Fund, which includes all non-local
revenue (i.e., national television revenue), which helps to even out revenue
disparities. Teams with larger local revenues receive smaller shares, while the
teams with smaller local revenues receive larger shares.25 The 2012 agreement
phased in additional provisions. Teams in the 15 (later reduced to 13) largest
markets are ineligible for shares even if they have low local revenues.26
Further, teams are not allowed to use these revenues to pay down team debt
(and so increase profits), and have to report how the funds increase team
quality.27 To the extent that these new restrictions induce net recipients to
increase player quality, they increase competitive balance.
Following the 2004–2005 lockout, the NHL adopted a complex formula that
makes transfers from high-revenue teams to low-revenue teams to equalize
payrolls. In order to qualify for these revenues, a team must be among the
bottom 15 in revenue and play in a city with a media base of less than
approximately 2 million people.28
Interestingly, the model shown in Figure 5.1 predicts that if teams are profit
maximizers, revenue sharing will not increase competitive balance, as shared
revenue represents a tax on wins for all teams, shifting MRL and MRS
downwards equally and leaving the distribution of wins unchanged. Research
by John Solow and Anthony Krautmann supports this prediction.29 Revenue
sharing increases competitive balance if teams that receive revenue use it to
improve the quality of the team (i.e., maximize wins).30
From a league-wide profitability standpoint, revenue sharing is problematic
because the entire league benefits when large-market teams win more often.
To see why, think of revenue sharing as a tax on big-market teams and a
subsidy to small-market teams. At its most extreme, the tax/subsidy system
equates all incomes, effectively directing all revenue to a central pool from
which each team takes an equal share. Even if post-tax revenues are all equal,
the central pool is larger if big-market teams win more often. If the central
pool is larger, then the post-tax payment to all teams—small-market as well as
big-market—is larger. Thus, all teams have a financial interest in the success
of big-market teams.
Salary Caps and Luxury Taxes
Of the major North American sports leagues, only MLB does not have a salary
cap. A salary cap restricts the amount a team is allowed to pay its players. In
Chapter 9, we will see how caps affect payrolls. Here, we focus on their
impact on competitive balance. The term “salary cap” is a misnomer for two
reasons. First, its principal target is payrolls, not individual salaries. Second,
salary caps are accompanied by minimum salary levels. Together, the two
create a band, setting both a ceiling and a floor on what teams can spend on
players. Most of the attention has focused on the ceilings, which were $155.27
million in the NFL (exclusive of carry-overs from the previous season), $71.4
million in the NHL, and $70 million in the NBA for the 2016 (2015–2016)
season. The minimum payrolls are approximately $138 million in the NFL,
$52.8 million in the NHL, and $63 million for the NBA.31 While the caps
restrict spending, the minimum payroll values prevent teams from
substantially under-investing in talent. In contrast to most North American
leagues, the major European soccer leagues do not control team salaries. In
the EPL, for example, 2016–2017 payrolls varied from a high of $335 million
(Manchester City) to a low of just $25 million (Hull City).32 The disparities in
other soccer leagues are similar.
The big difference between the salary restrictions in the NFL and the NBA is
that NFL payrolls have a hard cap, while NBA payrolls have a soft cap. A
hard cap is an absolute limit. NFL teams must stay within the limits of the
cap. A soft cap has exceptions to the limits that the cap imposes. The three
most important exceptions in the NBA’s soft cap are the mid-level exception,
the rookie exception, and the Larry Bird exception.
The mid-level exception allows each team to sign one player to the average
NBA salary even if the team is already over the salary limit or if signing a
player to such a contract puts the team over the limit. Under the rookie
exception, a team can sign a rookie to his first contract even if doing so puts
the team over its cap limit. The Larry Bird exception permits teams to re-sign
players who are already on their roster even if doing so exceeds the cap limit.
This rule got its name because its first use permitted the Boston Celtics to resign
their star player in 1983. That year, the Celtics were loaded with stars,
and a hard cap would have forced the Celtics to break up this very popular
team to re-sign Larry Bird. To avoid this, the NBA softened the cap,
permitting the Celtics to keep their team intact. The soft cap may explain why
competitive balance in the NBA is so low. A soft cap makes it easier for teams
to stock up on and retain talented players.
Because the exceptions have undermined the soft cap, the NBA has developed
a number of supplemental measures to reinforce it. These include caps on
salaries that teams can pay individual players and a luxury tax. The cap on
individual salaries classifies players according to their years of experience and
specifies maximum and minimum salaries that a team can pay players with a
given amount of experience.
A luxury tax has nothing to do with luxury boxes; it is a surcharge the league
imposes on teams whose payroll exceeds a specified level. The 2011 collective
bargaining agreement (CBA) significantly tightened the NBA’s luxury tax.
Prior to 2012–2013, teams that exceeded the cap had to pay a $1 tax for every
$1 by which they exceeded the cap. Now, teams must pay a tax rate that
increases for every $5 million by which they exceed the cap. For example, a
first-time offending team that is $15 million over the cap used to pay a $15
million tax but now must pay a $28.75 million (1.5 × $5million + 1.75 ×
5million + 2.5 × $5million) tax. Penalties are greater still for teams that
exceeded the cap in three of the four previous seasons.33
Baseball’s luxury tax (now known as a “competitive balance tax”) also charges
teams for having a payroll above a pre-specified threshold (starting at $178
million in 2011–2013, and rising over time to $197 million in 2018, then up to
$210 million by 2021). Like the NBA, MLB charges a higher tax rate to repeat
offenders. Teams that exceed the threshold for the first time pay a 20 percent
tax rate. Subsequent violations increase the tax rate to 30 percent for a twotime
violator and then to 50 percent for teams exceeding the threshold three
or more times.34 In addition, beginning with the 2017 CBA, teams that exceed
the threshold by more than $20 million pay another 12 percent in tax, while
teams with payrolls more than $40 million over the maximum pay an
additional 45 percent competitive balance tax. In total, a team with a large
payroll and multiple violations could pay as much as a 95% tax on any payroll
above the threshold. In 2016 a record-high six teams were subject to the tax:
the Yankees, Tigers, Red Sox, Giants, Cubs, and Dodgers. Between 2003 and
2016, the Yankees have paid the luxury tax every year, with total payments
exceeding $325 million.35 The Yankees’ continued willingness to exceed the
threshold indicates that the luxury tax has had a limited impact on their
behavior. Another important difference between the luxury taxes in the NBA
and MLB is in the distribution of the collected tax revenues. In the NBA,
luxury tax revenues are distributed to non-violators (though there is room in
the contract language for the league to keep 50 percent or more for “league
purposes”). These distributions increase the revenue of teams who follow the
salary cap. In MLB, tax revenues are not redistributed to teams but instead go
to player benefits and the Industry Growth Fund.
The Reverse-Order Entry Draft
The reverse-order entry draft allows teams to select amateur players in
reverse order from their finish of the previous season. The team with the
worst record chooses first, the second-worst team chooses second, and so on,
until the team that won the previous season’s championship chooses last. The
same procedure continues through subsequent rounds. All players not chosen
in the draft are free to sign contracts with any team.
The origin of the draft can be traced to 1934, when two NFL teams—the old
Brooklyn Dodgers and the Philadelphia Eagles—competed for the services of
Stan Kostka, an All-American player at the University of Minnesota. The
resulting bidding war drove salary offers to the then-unbelievable level of
$5,000 (what Bronko Nagurski—the greatest player of the era—made). At the
next annual league meeting, Bert Bell, the Philadelphia Eagles owner,
proposed a unique way to avoid future bidding wars. Teams would select the
rights to unsigned players, with the order of selection determined by each
team’s performance in the previous season.
The dubious impact of the reverse draft on competitive balance and the clear
limitations drafts place on the market power of drafted players have led critics
to claim that the reverse-order draft is nothing more than a tool to keep
players’ salaries low. It might be no coincidence that Bert Bell’s Eagles were a
last-place team when he proposed the draft. Still, some teams gave up a
considerable advantage by agreeing to the draft. The New York Giants and the
Chicago Bears, two teams from the biggest markets at the time, dominated the
NFL’s early years. Sacrificing the right to bid against other teams was not in
their immediate self-interest, as the Giants and Bears could have outbid all
other teams for any player they desired. However, the owners of the Giants
and Bears recognized that a larger issue was at stake. In the words of Tim
Mara, the owner of the Giants at the time, “People come to see a competition.
We could give them a competition only if the teams had some sort of
equality.”36
All the major North American leagues now have a reverse-order draft. Partly
because NBA rosters are so small—each team has only 12 players on its active
roster—the NBA draft lasts only two rounds. Because other leagues have
larger squads, and because success at the professional level has traditionally
been harder to predict, they have longer drafts: the NHL draft lasts 5 rounds,
the NFL draft lasts 7 rounds, and the MLB draft lasts 40 rounds or until all
teams decide to stop drafting players.
Evaluating the Reverse-Order Draft
In theory, reverse-order drafts promote competitive balance by allocating the
best new players to the weakest teams. However, its success in equalizing
talent depends on the teams’ motivations—a draft is most effective if all teams
are win-maximizers—and on the ability of teams to identify and develop
talented players. Some teams (such as the NFL’s New England Patriots)
consistently find talented players despite having poor draft picks, while other
teams regularly fail to find good players despite having excellent draft
positions.37
Because the reverse-order draft rewards failure with high draft picks, it can
worsen competitive balance. As teams fall out of competition, they could
begin to lose deliberately so as to improve their position in a future draft.38 In
recent years, this strategy has been made somewhat famous (or infamous) by
the Philadelphia 76ers, who developed a long-term strategy that became
known as “the process.” Unfortunately for Philadelphia fans, the process
involved winning just 19 games in 2013–2014 and only 10 games in 2014–2015.
During this time period, the team was widely regarded as making little effort
to win in an effort to secure the top draft picks. In response to concerns that
teams may be losing intentionally, the NBA and NHL have instituted lottery
systems in which teams with the worst records have the best chance of
securing the top draft pick but are not sure to get it. A lottery reduces the
incentive to lose games intentionally, but it also reduces the possibility that
the weakest teams benefit most from the draft. For example, the NBA’s
Orlando Magic won the 1993 draft lottery despite having the best record of
any team eligible for the lottery.
Evaluating Talent: The Oakland Athletics and Moneyball
In the early 2000s, the management of the Oakland Athletics baseball team
found a new way to evaluate talent. Despite being a quintessential smallmarket
team that regularly had among the lowest revenues and payroll in
MLB, the Athletics became a consistent winner. They finished first in their
division four times between 2000 and 2006 and finished second the other three
years.39
Michael Lewis analyzed the Athletics’ astonishing run in the best-selling book
and hit movie Moneyball.40 Lewis attributes much of the Athletics’ success to
their GM, Billy Beane. Lewis claims that Beane stood much of the
conventional wisdom regarding player evaluation on its head.41 Teams
traditionally looked for “five-tool” players who could run fast, throw far, field
well, have a high batting average, and hit the ball hard. The standard
yardstick for the two latter tools was slugging percentage, which measures a
player’s total bases per at bat. A player who hits a lot of home runs will, all
else equal, have a much higher slugging percentage than a player who hits a
lot of singles. Beane, a follower of baseball statistician Bill James, claimed that
the key to success lay in a team’s getting on base as often as possible, which is
best measured by on-base percentage.42 Lewis asserts that, while traditional
scouts look for physically gifted athletes, Beane looked for players who were
disciplined and received a lot of bases on balls (walks), even if they were not
much to look at in a uniform.
In an economic analysis of Moneyball, Hakes and Sauer find evidence to
support Lewis’s claim that Beane found an unexploited market imperfection.43
Applying multiple regression analysis, they show that, prior to the success of
the A’s, increasing a team’s on-base percentage contributed more to wins than
increasing its slugging percentage. Using salaries to measure the value that
teams place on a player’s characteristics, they also show that teams generally
valued slugging percentage more highly than on-base percentage.
Emphasizing on-base percentage allowed Beane to acquire players whose
contribution to success had been undervalued by other teams. This gave the
Athletics a competitive advantage and allowed them to win consistently with
far lower payrolls than other teams.
Unfortunately, the Athletics could not translate their success in the regular
season to a World Series appearance. As Beane himself admits, his strategy
was better suited to a 162-game season than to a 7-game series.44 More
importantly, as in any competitive market, the first mover enjoys a relatively
brief advantage. Hakes and Sauer demonstrate that the relative valuation of
slugging percentage and on-base percentage had begun to shift by 2004. With
wealthier teams now applying Bill James’s principles, the Athletics were
relegated to also-ran status by 2007.
The ability of teams in other sports to adopt and profit from the advanced
analytics—Moneyball—approach has been widely debated. The advantage of
such analysis in baseball is that performance of the players is often separable.
For example, no other players directly assist a batter in hitting a home run.
Bill Gerard argued that to transfer the analytical approach to what he calls
“complex invasion sports,” such as ice hockey, basketball, and soccer, analysts
must resolve three measurement problems: tracking, attribution, and
weighting. These amount to tracking the actions of various players both in
and out of possession of the ball or puck, attributing productivity to those
actions, and then weighting that contribution to success.45 David Berri, Martin
Schmidt, and Stacey Brook published the first significant contribution to the
body of work on this subject in their well-known book The Wages of Wins.
Similar to Moneyball’s view of baseball, they conclude that NBA team owners
make systematic errors in signing and paying players because they are too
focused on scoring totals instead of efficient use of possessions.
Schedule Adjustments in the NFL
The NFL has a unique approach to increasing parity across seasons, which is
unrelated to player movement or salaries. Each team’s schedule is determined
in part by its performance in the previous season. Each team plays 16 games,
14 of which are independent of its prior success. It plays the other three teams
in its own division twice, all four teams in another division of its conference,
plus all four teams from one division in the other conference. The remaining
two games are intended to increase balance and are based on the team’s
performance in the previous season. The first-place team in each division
plays the first-place teams in the two divisions (within conference) that the
team is not scheduled to play, the second-place team plays the two other
second-place teams, and so on. As a result, stronger teams play tougher
schedules and weaker teams play easier schedules the following year, creating
a natural tendency toward parity.
Promotion and Relegation
The promotion and relegation system provides an additional incentive
mechanism that may increase competitive balance. We saw that top European
leagues, such as the EPL and the Bundesliga, have competitive balance ratios
that are second only to the NHL, North America’s most balanced league. Yet,
if we use frequency of championships as the measure of balance, Table 5.3
provides strong evidence that these same leagues are among the least
competitive. The promotion and relegation system helps explain this puzzle.
When a team in a closed league is having a bad season and stands to finish
near the bottom of the standings, it may not have much incentive to win.
Once eliminated from the playoffs, the team may use the remaining games to
invest in future seasons, trying out new players from its minor league system
and trading or selling off top players to playoff-bound teams in search of an
edge. Thus, the winning percentage of the poor team may erode further,
increasing the standard deviation of winning percentage.
In an open league, teams near the bottom of the standings have no such
luxury. If they let their performance slide further, they may be relegated to a
lower league. Teams that are near the bottom of the standings have an
incentive to play to win right to the end. Thus, promotion and relegation may
not increase turnover of the league champion, but it may decrease the
standard deviation of wins within the league.
Biographical Sketch
Bud Selig (1934–)
Selig listened and questioned and murmured empathetically, all of the things he did best.
—John Helyar 46
Perhaps no person symbolizes the struggle over competitive balance
more than baseball Commissioner Alan H. (“Bud”) Selig. Selig served as
the commissioner of Major League Baseball from 1998 to 2015. After
stepping down, he was given the title of Commissioner Emeritus.
Selig’s Wisconsin roots run deep. Born in Milwaukee, he graduated from
the University of Wisconsin, Madison in 1956 and, after serving in the
military for two years, joined his father’s automobile business. Business
proved so good that, when major league baseball came to Milwaukee,
Selig was able to act on his love of baseball by becoming a stockholder
in the Milwaukee Braves. Selig’s ties, however, were to the Milwaukee
Braves. When the team moved to Atlanta in 1965, Selig promptly sold
his stock and formed a group dedicated to bringing a new team to
Milwaukee. His efforts bore fruit when the Seattle Pilots, a badly
financed expansion team, went bankrupt after the 1970 season. Selig
immediately bought the team for $10.8 million and moved it to
Milwaukee, renaming it the Brewers.
With Selig as their president, the Brewers gained a reputation as an
exemplary organization, and the team came within a game of winning
the 1982 World Series. The Brewers’ performance on and off the field led
Selig to play a growing role in the governance of MLB’s affairs. When
the owners forced Fay Vincent to resign as commissioner in 1992, Selig,
as chairman of the owners’ executive council, effectively took over the
duties of commissioner. For the next six years, Selig walked a tightrope,
serving the interests of all of baseball while working to advance the
interests of his own Milwaukee Brewers. Finally, in July 1998, Selig’s
fellow owners elected him as commissioner. Selig then put his holdings
in the Brewers into a blind trust and turned over operations of the
Brewers to his daughter Wendy Selig-Preib.
Selig’s popularity with his fellow owners and his insistence on
consensus brought the owners unprecedented cohesion. That enabled
him to introduce a variety of innovations designed to bring greater
excitement to the game. Under his tenure, baseball raised the number of
divisions per league from two to three, increasing the number of teams
in the post-season. The number was further increased by the
introduction of a “wild-card” playoff team (which has won the World
Series six times since it was instituted). He also oversaw greater
consolidation of the American and National Leagues, whose war of the
early 1900s did not fully end until Selig brought both leagues under the
authority of the commissioner’s office in 2000.
Most importantly, by bringing the often-fractious owners together, Selig
reversed a trend of over 20 years. All labor stoppages prior to Selig’s
becoming commissioner had ended with the owners capitulating. The
1994–1995 strike effectively ended in a draw, with neither side achieving
its aims. In the near-strike of 2002, the owners forced the players’
association to blink and to approve a revenue-sharing plan and luxury
tax that it had bitterly opposed. This marked ownership’s first outright
victory in negotiations since labor negotiations began in 1972. Sadly, the
1994–1995 strike caused the cancellation of the 1994 World Series,
something two world wars had failed to do. The willingness of owners
to forgo the rest of the season severely tarnished the game’s reputation.
The resurgence of baseball’s popularity in the late 1990s has since been
marred by allegations that many of the period’s greatest stars used
performance-enhancing drugs. The allegations of drug use and the weak
antidrug stance by MLB led to Congressional hearings at which Selig,
union representatives, and star players were subjected to embarrassing
questions on national television. In response, Selig played an
instrumental role in the development and enforcement of a strong antidoping
policy, including a first-ever in US professional sports policy on
human growth hormones in 2010.
Selig stepped down as commissioner in 2015, turning the reins over to
Robert Manfred. Because of his controversial record, some see Selig as a
man who saved the game, while others feel that he was ill-suited to be
anything other than the owner of a small-market team. Whatever one’s
opinion of him, few can deny that he has had a major impact on Major
League Baseball.
Sources: Andrew Zimbalist. 2013. In the Best Interests of Baseball?
Governing the National Pastime (Lincoln: University of Nebraska Press);
MLB.com. “Alan H. ‘Bud’ Selig,” Commissioners, at
http://mlb.mlb.com/mlb/official_info/about_mlb/executives.jsp?
bio=selig_bud. Accessed March 7, 2017.
Summary
For a league to succeed financially in the long run, it must have a semblance
of even competition among teams. At the same time, recent research shows
that teams’ aversion to losing causes them to prefer lopsided wins to
unexpected losses. Given that the value of a win is much greater in large
cities, it is unlikely that leagues would maximize revenue from perfect parity
across teams, and they would likely do better to have better teams in cities
where demand for the sport is greatest.
Fans and owners both have an interest in competitive balance. The perception
that fans are less likely to follow a league in which a few teams win most of
the games or championships gives owners a financial stake in maintaining
competitive balance. Policies aimed at increasing competitive balance are
generated by the central league office as individual teams always prefer
winning to losing.
There are several ways to measure competitive balance, no one of which is
necessarily better than the others. A popular measure of within-season
balance is the ratio of the standard deviation of winning percentages to the
“ideal” standard deviation that would prevail if all teams were equally
talented. A popular measure of cross-season balance is the Herfindahl-
Hirschman index, which shows how narrowly concentrated championships
have been over a given time period. By both measures, MLB is relatively
competitive compared to other sports. This contradicts the claims of MLB
owners that baseball has had a competitive balance crisis since the advent of
free agency.
Economic theory predicts that free agency will not affect the distribution of
talent in a sport as long as the team owners maximize profit, players
maximize income, and transaction costs are low. Sports leagues have
implemented several policies—such as revenue sharing, salary caps, luxury
taxes, and the reverse-order draft—to increase transaction costs and limit the
movement of players from small-market to big-market teams. These measures
have met with uneven success.
Discussion Questions