Abstract:
This article explores if ticket revenue increases due to a particular team’s performance, or if it increases on how much Stars team might have. This article takes a more comprehensive look then the study done by Hausman and Leonard (1997) that looked at attendance changes when a team added or played a team with stars. The structure of the study first shows the empirical model they use to examine the demand for professional basketball. Then the authors discuss a review of the econometric of the empirical model.
Dependent variable
GATE=Gate revenue (in $millions)
Independent variables
Team performance characteristics
WINS= Regular season wins
WPLAY= Playoff wins
WPLAY(-1)=lagged playoff wins
WCH20= Championships won, weighted
(Prior championship gets 20, year before 19…)
STARVOT=All-Star votes received
Franchise characteristics
SCAP= Stadium capacity
DCAP=Teams at capacity
OLD=Age of stadium
DEXP5=Expansion team, dummy variable equal to one if the team is
less than five years old.
RSTAB=Roster stability
WHITERRATIO= White ratio (WHITEMIN / WHITEPOP)
ATTEND = Team attendance
Market characteristics
CB=Competitive balance in conference
COMPTM=Competing teams
RYCAP= Real per-capita income
POP=Population
Racial variables
WHITEMIN=Percentage of white minutes
WHITEPOP=Percentage of whites in population
1.) Competitive Balance formula
CB= Competitive balance
σ(wp)it=Standard deviation of winning percentages within league (I) in period (t)
μ(wp)it= League (I)’s mean
N= the total number of games.
CBit= σ(wp)itactual/σ(wp)itideal ; where σit ideal=μ(wp)it/√N
For this study the competitive balance equation was calculated for each conference, Eastern and Western, and each year, for this study.
2.) Econometric issues and estimation-The list of the dependent and independent variables was utilized to construct the following model:
Yn= 4Σ i=1αi+19Σ k=1α kXkn+εn n = 1,2...,108.
3.) Marginal value of team wins
Y= The value of gate revenue
Xi= The value of team wins
αi=The estimated coefficient from the double-logged model.
X1=[Y/Xi]αi
2a.) Relationship between revenues and the values of wins multiplicative model:
Yin=[α I19Σ k=1Xα k]
Results:
Table 2: Estimated coefficients for equation 1, dependent variable is GATE, multiplicative model.
T-Statistics, significant at the 1% level:
WPLAY=2.718
WPLAY(-1)=3.896
WCHM20=3.547
SCAP=3.761
OLD=-2.940
T-Statistics, significant at the 5% level:
WNS=2.334
STARVOT=2.474
DEXP5=2.319
POP=2.197
The R2 value is .815 meaning that 81.5% of the data is explained.
R2 = .815
Adjusted R2=.767
F statistic= 17.038
P value: F statistic = .000
Table 3: Estimated coefficients for equation 3, dependent variable GATE, Linear Specification.
T-Statistic, significant at the 1% level:
WINS=3.544
WPLAY=3.578
WPLAY(-1)=3.466
WCH20=3.089
SCAP=3.519
T-Statistic, significant at the 5% level:
DHILL=-2.217
DCAP=2.048
The R2 value is .760 which means 76% of the data is explained.
R2= .760
Adjusted R2=.698
F statistic=12.231
P value: F statistic = .000
Conclusion:
Given the results, the study concludes that gate revenue is most responsive to changes in stadium capacity and wins. A single win generates $83,037 in revenue while one All-star vote produces $.22, which the team would need 370,000 votes per team to have the equivalent of one win. Which proves that performance on the basketball quart is what attracts the fans not star power. One thing that was not taken into account for revenue was that star power might generate, is the revenue that they generate for their opponents.
Source:
Berri, David J. Martin B. Schmidt, Stacey L. Brook. "Stars at the Gate". Journal of Sports Economics, 2004, Vol. 5, no. 1, Pages 33-50.