How Will Attendance Change When Baseball Returns?

Baseball is going to return at some point, even if it’s 2021. If it returns in 2020, it will look wildly different. Perhaps it’ll be played in just two or three states. Maybe it’ll have virtual fans and robot umps. We really don’t know because the situation changes every day. Whatever it is will drastically affect attendance. I’m not necessarily interested in that type of return today. I’m more interested in what baseball will look like when it returns closer to normal, when teams like the Cardinals are playing in their own ballparks. I’ve recently gone down the rabbit hole of looking at the 1918 influenza pandemic- how it spread, how long it took to contain, and the way it changed society. From there, I found other dips in MLB attendance in the past. That sounds like as good a starting place as any.

Sean Lahman’s brilliant database has attendance data going all the way back to 1892. The further back we go, the fuzzier the data gets, so we’ll instead start at a nice round number- 1900. We’ll look at league-wide attendance trends. First, I’ve calculated the three-year rolling average, then the percent increase or decrease in the following year. Here’s how it graphs out, with peaks and valleys labeled:

Most of these valleys make perfect sense. The worst unemployment rates of the Great Depression were in 1933. 1943 was the middle of World War II while 1946 was the first season after it was over, explaining the biggest peak on the graph. 1981 had a work stoppage and 1995 followed one. There was a recession in 2009.

The valley in 1952 is a little confusing, though the Korean War (June 1950-July 1953) likely explains it, particularly so soon after the existential crisis of World War II. That attendance had also collapsed in 1950 and 1951 lends further evidence. 1961 is even murkier. The best answer I can offer is that per-game attendance was possibly diluted with three new teams- the Los Angeles Angels, Minnesota Twins, and Washington Senators Part Deux. The attendance in Minnesota was very good (sixth best per-game average out of 18 teams), but the Angels and Senators were the two worst in the league. Compounding matters, the presence of the new Angels seemingly leeched fans away from the Dodgers, who lost the third most fans from the previous season. In other words, 1961 was probably a perfect storm of negative expansion ramifications.

1914 and 1918

Only two seasons saw attendance plummet more than 40% from the previous three seasons- 1914 and 1918. Technically, 1933 was close- 39%- but just missed the cut. The 1914 dip can be explained by two factors. The Federal League began play in 1913 and increasingly went hard after MLB players and fans. If we look specifically at cities that had both MLB and Federal League teams in 1914 (Brooklyn, Chicago, St. Louis, Pittsburgh, Washington which competed with a team in Baltimore), attendance dropped 1,467 per game. If we exclude those seven teams in five cities, the league-wide change in attendance still would’ve been a massive 30% dip, but it wouldn’t approach the 1918 collapse. A larger factor was likely a spike in unemployment in 1914, which doubled according to the Social Security Administration.

That brings me to the reason I’m even writing this article- 1918. There are a lot of reasons attendance cratered that year. First and foremost, it was right in the middle of World War I. As we saw from the World War II and Korean War eras in the graph above, a war of that magnitude was surely one reason attendance dipped in 1918. As for the influenza pandemic, it wasn’t at its most deadly until the fall and winter but it was still present in the preceding summer months. The following graph is from a blog last maintained in 2009 by Rob Dorit, a professor of biological sciences at Smith College. It’s the best graph I can find to demonstrate how influenza spread in 1918.

Rob Dorit’s Flu Blog, https://sophia.smith.edu/blog/rdorit/

Influenza was initially discovered in Kansas early in the year. Most social distancing measures weren’t put in place until early September. We don’t know for certain if the presence of influenza cases in June and July prevented people from going to the ballpark. We do know that baseball cut its regular season short by nearly a month, ending on September 2nd, though the truncated season was a result of the war rather than the pandemic. Untangling the effects of war and influenza with regard to the 1918 attendance drop isn’t possible, at least not for someone with my limited math skills.

How this will affect attendance when baseball resumes is hard to know. What we do know is that baseball has seen valleys caused by:

• Depression/Recession and increasing unemployment rates (1914: -44.1%, 1932: -34.7%, 1933: -39.1%, 2009: -6.3%)
• War (1943: -25.65%, 1918: -46.3%, 1952: -22.6%, 1951: -22.2%)
• Strike (1981: -5.9%, 1995: -17.8%)

There aren’t a lot of examples of pandemics, particularly ones that require social distancing as a safety measure, occurring during a baseball season. There was another influenza strain during the 1957 and 1958 seasons, but it was fairly well contained in the United States (116,000 deaths total) and a vaccine was ready just two months after it arrived in the US. The same is true for H1N1 in 2009.

Rarely has baseball dealt with two of the four items we’ve mentioned- massive unemployment, war, a strike, and a pandemic- at the same time. The only year that qualifies is 1918. With unemployment soaring since we entered the quarantine, you don’t have to squint to see a scenario where baseball is trying to overcome both the pandemic and a recession/depression at the same time. If that happens, don’t be surprised to see the third 40% attendance decrease in baseball history. To put that in real world terms, baseball would lose an average of 7,600 fans each time they open the gates, down to 20,500 per game. Even if the pandemic effect is overstated, the resulting economic conundrum has historically led to decreases of 30% or more on several occasions. Unsurprisingly, ballparks are about to get much emptier.