The Washington PostDemocracy Dies in Darkness

No, Republican underperformance in 2022 wasn’t due to covid deaths

The stage remained empty most of the evening at a Republican watch party on Election Day in Topeka, Kan., as the candidates for governor and attorney general waited until nearly midnight to speak. (Reed Hoffmann/AP)

It was probably inevitable, given the extent to which the response to the coronavirus pandemic overlapped with political partisanship, that the effects of the virus would also be viewed in reverse. If Republicans were less likely than Democrats to be vaccinated (which they were) or to do things like wear masks (same), wouldn’t that mean that more Republicans would die of covid-19? And given that more than 1 million Americans have died because of the virus, wouldn’t that necessarily have some effect on politics?

Those rumblings, particularly common in the post-vaccine era when the gap between red and blue widened, have grown louder ever since Republicans fared unexpectedly poorly in the midterm elections. Was this a manifestation of the disproportionate toll on Republican voters?

The short response is no. The longer one is that this line of argument is often nothing more than an effort to whitewash a macabre I-told-you-so sentiment among Democrats with a veneer of quantitative objectivity.

Sign up for How To Read This Chart, a weekly data newsletter from Philip Bump

One paper published earlier this year showed the gulf in covid-19 deaths by party. In Florida and Ohio, a comparison of party registration with confirmed pandemic deaths revealed a big increase in Republican deaths in the period after vaccines became widely available. This comports with our understanding: The vaccine became a culture-war marker, leaving Republicans more vulnerable from infections. More than a quarter-million deaths probably could have been prevented had the infected been vaccinated — and many of those who weren’t vaccinated were Republicans.

That paper, though, is useful not only because it shows the divide but also because it gives us a rough rubric for testing the question at hand: Were there races that might have been affected by the partisan gap in covid deaths?

What the researchers found was that there was a wider partisan gap in deaths in less-vaccinated counties, using vaccination rates at the start of June 2021 as a metric. Here’s the paper’s chart comparing the period before vaccines were widely available with the period once they were made available to all adults.

The gap is clear.

We can do a rough reverse-engineering here. The research looked not at covid-19 deaths but “excess deaths” — that is, the number of deaths of members of each party above a pre-pandemic baseline in a county. So, if we make five assumptions about the figures, we can come up with estimates of the pandemic’s effect on voters. Those are:

  1. That the results from Ohio and Florida can be extrapolated nationally.
  2. That all of the excess deaths captured in the curves above are due to covid-19.
  3. That the curves accurately capture the relationship.
  4. That everyone who died would have otherwise voted.
  5. That everyone who died would have voted for one of the two parties.

You can immediately see some problems here. Setting aside the first three assumptions, none of which is particularly defensible, we can say with certainty that the last two aren’t. Maybe about half of eligible voters cast ballots this year. Even recognizing that older Americans vote more heavily and have been more likely to die of covid-19, we should still assume that no more than two-thirds of those who died would otherwise have voted.

But we can set that aside because it doesn’t really matter. The fundamental issue (as I wrote back in February when I first addressed this question) is that 1 million deaths nationally translates into far fewer deaths at a state level. And when you then apply the partisan ratios sussed out from the above assumptions, the net difference in votes is often tiny.

Imagine a state in which 50,000 people died of the virus, for example. Imagine that those deaths were demonstrably 2 to 1 Republicans. That means that the net loss for Republicans was about 16,500 votes — a small margin in a statewide race.

I took county-level data (compiled by The Washington Post), determined the vaccination rates on June 1, 2021 (using federal data and in keeping with the paper’s assumptions) and applied a distribution of those deaths by party in keeping with the curves above. Then I tallied the number of deaths in each state between May 1, 2021 (after vaccines were broadly available) and Election Day and figured out the partisan gap.

Unsurprisingly it was in Texas and Florida — red states with large populations — where the net losses for Republicans were largest. But in both of those states, the margins in Senate and gubernatorial races (that is, state-level contests) were far wider than the 14,000 or 10,000 net differences, respectively.

In no state, in fact, was there any Senate or gubernatorial race that was likely to have been affected by the partisan divide in coronavirus deaths. The closest was in Nevada’s Senate race, where Sen. Catherine Cortez Masto (D) was projected the winner with a current lead of about 6,600 votes. The partisan difference in covid-19 deaths would have narrowed that margin by about 2,000 votes, less than a third. And that’s with assumptions 4 and 5 above, about perfect turnout and no third-party voters.

The next closest Senate race was the one in Wisconsin, where Sen. Ron Johnson (R) leads by about 12 times the net partisan difference in covid deaths. The closest gubernatorial race was in Nevada, where the Republican is winning by 16,000 votes. The next closest Senate and gubernatorial races where Democrats lead — and therefore might have benefited from deaths — were in Georgia (35,000-vote lead, 5.5 times the estimated death gap) and Kansas (20,000-vote lead, 16 times the estimated gap). These are not races that were going to be affected by the death toll.

In fact, there is no state in which a Democrat leads in the Senate or gubernatorial race and where, if we assume that all of the covid-19 deaths since May 2021 were of certain Republican voters, the results would have been different.

There will be a tendency to drill down a level deeper: Well, what about House races? To which the response is straightforward: The effect of covid-19 deaths will necessarily be smaller in those cases, so only the most narrow of results might have been affected by those deaths. That’s setting aside the fact that close races are determined by hundreds of factors.

Again, it’s worth examining why this argument keeps cropping up. What’s the point? The point, clearly, is to suggest that Republican politics came back to bite the GOP. But the numbers here are so small, there’s no reason to assume that’s what happened — except that it allows one to believe that fate had, at last, weighed in on their partisan side.

If one’s response to this article is disappointment, that, too, is worth examining.