The baseball season is the longest in terms of games played of any major sport in America. This means that the 81 home dates each team has per year comprises a greater percentage of their individual revenue source as opposed to football (eight home games) or basketball (41 home games). Revenues from ticket sales account for about 35 percent of total revenues in MLB, while NFL ticket sales comprise only around 18 percent of league revenue. For baseball clubs to increase revenue in MLB, they must put fans in the seats.
To quantify which factors are associated with increasing attendance at baseball games, I used a multiple linear regression model. After brainstorming independent variables and running them through many regressions, I arrived at the following model presented in standard form with the standard error of the coefficient directly beneath it (see box below).
Where attendance is average attendance per game for all 30 teams in 2005, payroll is team payroll in millions, metpop is the population of the metropolitan area in millions that the team plays in, prevwins is number of wins in the previous year (2004), and newstad is if the team is playing in a stadium which is at most three years old (newstad = 1 if less than three, = 0 if greater than three). The coefficients payroll, metpop and prevwins all had t-statistics greater than 2, showing their individual significance. Newstad had a t-statistic that was less than 1, and I left it in the model because its small sample size of six made it hard to prove significant and I feel it is valuable in this model.
The coefficients can be interpreted as a one-unit increase in an independent variable (payroll, metpop, prevwins or newstad) will increase the dependant variable (attendance) by its coefficient, on average - all else held equal. For example, signing a superstar for $15 million per year projects to increase attendance by 15*99.7=1,495.5 fans per game. That player can also help the team win more games and therefore further increase attendance.
By now you might be wondering why wins from the previous season would affect the current season's attendance, and why wins in the current season have no significant effect on attendance in that season. The reason for this is that all season ticket sales and most group ticket sales happen prior to the season and account for the bulk of a club's attendance. To show an extreme example, the Boston Red Sox sold a total of 2,813,354 tickets this season, with over two million of those tickets being sold before the first ball was thrown at Fenway.
Looking at the coefficient on the variable metpop in this model, I see why revenue sharing occurs in baseball. The population of the metropolitan area that each team plays in has great influence over how many people pass through the gates. While the Yankees can draw on the 21.2 million metropolitan New Yorkers, the Brewers out in Milwaukee only have 1.7 million inhabitants to draw from: a difference of 19.5 million. The impact? 19.5*452.1=8,816 fans per game. Multiply that out over 81 games, along with the increased marketing and advertising dollars being thrown around in the major cities, and you begin to see the divide between baseball's rich and poor.
The argument from the rich teams is that the poor teams must spend more money, and then win more games, to draw more fans to their stadiums. This is circular logic because to spend more money, you must increase attendance to increase revenue. Teams in the bigger cities have a structural advantage, because they have a larger fan base and therefore more revenue, simply due to their location.
