for the notion of a positive production network externality in professional sports.
An original data set is collected that contains many potential factors of demand for
baseball games. Questions of customer discrimination, marketing ability, marginal
productivity, and public financing of stadiums can be addressed using this data set.
The assumptions of the demand specification are empirically verified in both a
fundamental model and a team fixed effects model. The ex ante optimal probability
that the home team wins is about 66%. Thus it is between 0.5 and 1.0 as assumed by
the model. The answer to the quote "Are the Bulls so good they're bad for the NBA?"
is a cautious yes. If the optimal probability that the home team wins is similar in
basketball to baseball, and if the Chicago Bulls were to continue dominating, then fans
across the NBA would likely stop coming to games. As Danny Ainge, coach of the NBA's Phoenix Suns, puts it "it'll be better when the Bulls break up. More teams will
feel they have a chance to win it all." Further, the fans will realize this as well and
begin showing up at the gate again.
This analysis also reconfirmed Scully's findings of customer discrimination against
black pitchers, and found owners are correct in their use of weekend games and new
stadiums to attract additional fans. The probability of the home team winning using
odds data outperforms a similar variable using winning percentage as both predictors
of attendance and of winning.
The author would like to thank Severin Borenstein, Clair Brown, Ken Chay, Elizabeth
Gustafson, Larry Hadley, Terry Kennedy, and Jimmy Torrez for helpful comments. However,
the usual caveat applies.
The previous literature failed in this respect (see the next section for details). Additionally,
imagine two sets of teams. One set contains a high-quality and a low-quality team. The other
set consists of two teams each having equal amounts of player quality, but whose player quality
sums to the total player quality of the teams in the first set. Previous models of demand for
games within the two sets had the same quantity demanded. The current model allows for the
closeness of the game between the teams in the second set to have a positive effect on demand.
It is possible for each team to have a greater than 50% chance of winning its home games
because of home field advantage. In fact, it is in the interest of sports leagues (given the findings
here) to create a home field advantage for each team in their league.
Cairns et al., 1986, for a brief summary of long run demand studies of professional
Interestingly, these studies have had problems getting a negative sign on the ticket price
variable. Some have concluded that owners are not profit maximizers when setting ticket prices.
Likely, it is because of a lack of proper control variables or a simultaneity problem.
Kerr ( 1990) find similar results using a limited baseball data set.
Of the 2,267 games, 51 were sellouts.
The power of the internet was evident in the creation of this data set. Every variable came
directly off of an internet web page for free except the pitcher race variable.
The source of the team wins variables and the score variables is www.usatoday.com. They
had a listing by team of each game played and the score.
These data were found at www.sportsfaxnews.com.
This information was found at www.totalbaseball.com.