In my previous post on Tampa Bay’s attendance woes, I established that the Rays attendance has not responded to the team’s on field success as well as we would expect. Most teams see large attendance bumps when they win a lot of games and reach the postseason, but this has not been true for the Rays. Potential explanations for the attendance discrepancy vary widely. Many point to the stadium’s poor location and the newness of the franchise, others blame Florida’s snowbird population for keeping their [old] hometown allegiances, some cite the economy and low median income in the area, while others claim Tampa Bay is just not a baseball town. All of these and more could be factors, but we can simplify the attendance problem by lumping all potential explanations into one of two categories: (1) the location/ambiance of Tropicana Field and (2) the size of the fan base.
If the biggest problem is the stadium, it is fixable. Getting a stadium built is certainly not an easy or cheap process, but most teams — even the Marlins — manage to get it done through some combination of public and private financing. If the problem is the lack of fans though, the team could be facing years of low attendance, low television ratings, and as a result, low payrolls even if they get a new stadium.
How can we know if a new stadium would solve the Rays’ attendance problem? Ideally (warning: entering social scientist mode), we’d randomly assign half the Rays’ home games to a new, centrally located stadium, while playing the other half at Tropicana Field. We could compare attendance across the two venues and be able to make accurate causal inference. A controlled experiment such as this would allow us to parse out the “true” effect of Tropicana Field on the Rays’ attendance. Unfortunately, my proposed experiment would likely cost in excess of $500 million dollars and is entirely unfeasible.
Given that we cannot get experimental data, we are left to do the best we can with observational data. Luckily, we have some data sources that can help us assess the size of the Rays’ fan base independent of Tropicana Field. Though the games are not randomly assigned, we do have two sets of games each year that both involve the Rays and are not played at Tropicana Field: spring training games and regular season road games. Unfortunately, the relationship between on-field performance and road attendance is weak and inconsistent from season to season, so I’ll focus here on spring training attendance.
Using Grapefruit League attendance data from 2000-2011, I find a relatively strong relationship between wins the previous season and spring training attendance. The correlation between lagged wins and spring training attendance is 0.56, whereas it is 0.55 between regular season attendance and lagged wins. For Tampa Bay, the data are mixed. As the table below reveals, spring training attendance has increased since the Rays began their winning ways, but this increased attendance corresponds with a move of the spring training site from Al Lang Field in St. Petersburg, FL to Port Charlotte, FL in 2009. Ironically, moving the spring training site out of the Tampa Bay/St. Petersburg metropolitan area is associated with increased spring training attendance.
|Season||Avg. Spring Attendance||Wins in Previous Season|
To more fully explore the relationship between on-field success and spring training attendance, I regressed averaged spring training attendance on the previous year’s spring attendance, wins the previous regular season, an indicator variable for whether a team made the previous postseason, and market size (average of 2000 and 2010 populations). All of these factors are positively associated with spring training attendance. As for the Rays, they fit the model of spring training attendance much better than the regular season attendance model. In the regular season model the Rays always underperform expectations, sometimes drastically (i.e. more than 6800 fans a game in 2011). In the spring training model the Rays data is more mixed. Spring attendance was slightly higher than expected in 2010 (147 more per game) and slightly lower than expected in 2011 (523 fewer per game).
Would a new stadium solve the Rays attendance problem? Unfortunately for the Rays, the data do not provide clear answers. The combination of winning on the field and a new spring training home are associated with an increase in spring training attendance, but, to realize this gain, the site was moved from St. Petersburg to Port Charlotte. In fact, there is a distinct possibility that spring training is part of the problem. Why pay full price for a regular season game in April when you can see the team up close for less money in March? It is possible that a new stadium would allow the Rays to have a higher payroll and less margin for error in player acquisitions, but it is far from certain that there are enough fans in the area to sustain an MLB franchise.