Sign in / Join

Estimates of MLB's "Stadium Effect"

A quick follow-up to Take Me Out to the (New) Ballpark.  I appreciate the comments and I find the shared experiences and emotions from teams/stadiums a fascinating topic on both sociological and personal levels.

In my follow-up here, I briefly address the questions arising about the boost in attendance to newer parks.  I referred to a consistent and measurable "bump" to newer stadiums.   Rather than a "bump," the effect that I have in mind is more accurately described as a drag to older stadiums.  When looking at the last 20 years of MLB data by season and team, a strong and consistent effect of stadium age can be found in statistical models of attendance that take into account other influences such as winning percentage, last year's winning percentage, population, distance to competitors, domes, precipitation, and temperature.  I'm allowing for a continuous, negative effect of stadium age that diminishes as a stadium gets older.  The difference between a brand new and 30 year old stadium is about 5000 fans per year game if team-specific and year-specific effects are not included and about 8000 fans per year if team and year effects are included.

In contrast, an "off-on" treatment of an effect for new stadiums -- for example, on for the first two years and then off -- generates elusive and inconsistent results.  In my data, if the "on" period includes 2 years or 4 years, no "bump" appears; however, if the "on" period is stretched to 7 years, then a 5000 fan impact per game appears.   The measure associated with a stadium effect depends largely on what whether a researcher is focusing just on new stadium "bumps" or a broader question of stadium effects on attendance.  A continuous measure of stadium age distinguishes new parks from 10 year old parks and 10-year old parks from 40-year old parks in a way that an off/on metric cannot.

Interestingly, the stadium effect shows up even if using game-to-game data within a season over the 20 years.  The interpretation differs, however.  Now it suggests that most of the fan loss from an older stadium occurs during the first half of the season.