Pigskin, tailgating and pollution: Estimating the environmental impacts of sporting events

This paper investigates the environmental impact of sporting events by analyzing a collection of cities---33 in total---which host at least one NCAA football team that competes in the Division I Football Bowl Subdivision (FBS). Using day-level data from the Air Quality System (AQS) database, PM2.5 is used as the measure of pollution. Supplementing the AQS data with college football data for the 2010 regular season from NCAA.org from September to November and daily weather measures from the Global Historical Climatology Network (GHCN), four questions are examined. First, do cities experience an increase in PM2.5 on game days? Second, do cities also experience an increase in PM2.5 the day before or after game days? Third, do these effects change over the course of the regular season? Fourth, do these effects pose serious health-risks?

Using city and month fixed-effects, baseline results for the simple day-of effect suggest that PM2.5 levels increase by .72 micrograms per cubic meter of air on game days relative to all non-game days. Adding greater flexibility by allowing for both commuting and day-of effects, fixed-effects results suggest an increase in PM2.5 by .85, .82 and .26 micrograms per cubic meter of air for days just before, on and just after game days relative to all non-game days, respectively. Lastly allowing for month-specific commuting and day-of effects, fixed-effects results were generally largest for the month of November with an estimated increase in PM2.5 by 1.40, 2.16 and .89 micrograms per cubic meter of air for days just before, on and just after November game days relative to all November non-game days, respectively. These latter estimates were robust to numerous specifications.

The marginal increase in exposure to PM2.5 from a game day is re-scaled using existing estimates in order to examine the percent change in mortality across various causes. For games hosted in November, the marginal exposure to PM2.5 translates to a percent increase in mortality that ranges from .06% (cardiovascular) to .47% (respiratory). Relative to baseline crude death rates and the value of a statistical life, the increase in mortality from respiratory disease of .47% translates to a monetary cost of $6,200 per November game.