Did Covid-19 matter for the 2020 election?

Well, maybe not! A simple scatterplot by our friends at NPR shows no real relationship (although they want to make it look positive)

But that data is noisy! How can we know? Well, one thing we can do is take a look at the last pandemic election, back in 1918. So @LArroyoAbad and I did just that

And what did we find! Well, voters in 1918 swung away from the incumbent governor or the Democrats in congress the more the Spanish Flu hit their county.

That relationship is pretty tight! (There was a brief moment when "tight" replaced "fly" which replaced "fresh" which replaced "cool" so this works on multiple levels, at least in mid-1990s Brooklyn)

The result held up when we used the underlying disease environment or distance to military camps as instruments. And it didn't depend on our choice of controls, either

Pretty cool, hey! We just blew the entire conventional wisdom about the 1918 election of the water! It was the pandemic, stupid!

Er ... "out of the water"

But maybe I should not get so excited. Take a look at those coefficients. They are significant. They are robust. They were like cockroaches, they could not be killed. But like (most) cockroaches, they were also very small.

If you doubled excess mortality in a county, then the governor lost about 0.6 points in the next election, and congressional Democrats lost about a point

But wait! Maybe people don't vote based just on what happens in their county. Maybe ... hear me out ... they care about people elsewhere. Turns out there's an app for that!

Well, a statistical technique. You can capture spatial interactions to see if deaths elsewhere influenced vote shares. Consider these two maps of New York State

The first map show how much deaths in Manhattan influenced Democratic vote shares in other counties. As you can see, nobody upstate cared at all

Conversely, deaths in Syracuse impacted votes all over the state, but for people in NYC or Long Island it might as well have been in Nunavut

Especially apropro because Nunavut didn't exist yet

Anyway, if people in Erie County had cared about deaths in Flatbush as much as the cared about deaths in Lackawanna, then the whole map would have been dark blue. They weren't. In fact, our estimates barely changed when we took spillovers into account

Which means we have the best sort of null result --- a very precisely estimated teeny tiny itsy bitsy effect! Voters cared about the Spanish Flu and they held politicians accountable. They just cared about other stuff a lot more

This is surprising because the Spanish Flu was way way worse than Covid

And contrary to a strange false belief, the flu pushed World War I off the front pages

Plus the federal government printed up millions and millions of fliers

In fact, although the Wilson Admin ignored it, the flu was an issue. Some politicians were clever, like the governor of New Jersey, who said: “The only campaigns I am converned with right now are the Liberty Loan campaign and the campaign to suppress the influenza epidemic”

But at the end of day, voters cared about the Spanish flu and held politicians accountable. They just cared about other things (like agricultural price supports and the war) a whole bunch more. And we suspect that's going to be the story of 2020 as well

So go here and read an earlier version of the paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3680286

Enjoy!