COVID-19 has killed around 800,000 Americans. Joe Biden won the 2020 election by a margin of less than 45,000 votes in three states. There’s a morbid, tragic, and uncomfortable question that may linger in the back of your mind when you read those two sentences back to back. Did the number of voters killed by the coronavirus tip the result of the election?

COVID-19 does not discriminate by party, and yet Trump’s vote share in states — and especially in counties — correlates with a higher rate of deaths from COVID-19. But of course, it was not always this way. The pandemic initially hit Democratic-leaning parts of the country the hardest and for nearly a year, despite more mask wearing, more social distancing, and less travel, this trend persisted. There are some obvious reasons for this contradiction — Democrats are more clustered in high-density cities, rural residents (which lean Republican) do travel more than city-dwellers out of necessity, and COVID-19 has disproportionately affected African-Americans and Latinos, who lean Democratic.

But as the COVID-19 vaccine was rolled out and the Delta variant began washing over the country, the death rate flipped. Counties with large Trump vote shares now have much higher death rates than those with large Biden shares. It doesn’t take a genius to put two and two together — Republicans are the demographic group most likely to respond they will “definitely not” get the vaccine, while Democrats are the group most likely to say they’ve already been vaccinated. As the summer dawned, Trump-voting states lagged Biden-voting states dramatically in vaccination rates. “Polarization is killing people… it didn’t have to be this way,” noted one astute analysis by German Lopez of Vox on this disparity, and it’s true. Expertise, fair elections, science, journalism — essentially, the trust in institutions that guides a normal society — have been rejected by Republicans. As a result, they’re dying and they’re killing others. 

There’s no excuse for this. Any Republican trying to “own the libs” need look no further than this piece for a reason to get vaccinated. You can find a location for the free COVID-19 vaccine near you right here.

Did COVID-19 Deaths Swing the 2020 Election?

Let’s start with the macabre question first posed at this article’s onset: did enough voters die of COVID-19 to seriously impact the margin in the closest states? We’ll turn to the tipping point state in 2020, Wisconsin, to cut to the chase.1This is the state that provided the decisive 270th electoral vote to the winning candidate. As in, Biden won states totaling less than 270 electoral votes by a larger margin than Wisconsin, but his smaller margin victory in Wisconsin set him over the edge, and then any states he won beyond that were electorally superfluous. Biden won Wisconsin by 0.63% of the vote — only 20,682 votes — something that should make Democrats incredibly nervous. 

By November 5, 2020 — Election Day — the total number of reported deaths in Wisconsin was around 2,200. So, even if we extrapolated the actual death rate from estimated excess deaths (which is around 20% above the official death toll in the United States), that would still be fewer than 3,000 deaths in the state overall by Election Day.2There is yet another awkward issue to confront here — Wisconsin allowed early voting in the 2020 election. What happens if a voter cast a vote in mid-October and then died of COVID-19 a few days later, before the election? There’s actually an answer for this, and in Wisconsin, that vote is to be discounted — though that’s not the case in every state. Even if every person who died in Wisconsin of COVID-19 were a voter, which was certainly not the case,3 Only about 67% of the voting-eligible population voted nationally in 2020, which means only around 72% of the US is included in the voting-eligible population — which excludes children, some immigrants, and felons and even if each of them were die-hard Republicans, had they voted it would have only reduced Biden’s margin to around 17,700 votes — still a close one, but well beyond a comfortable margin in this hypothetical.

So while the pandemic certainly affected the 2020 election in the overarching sense (as in, do 20,000 voters in Wisconsin still go for Biden if there is no pandemic?), we cannot say that it swung it by a fatal purge of the voter rolls in the state that mattered.

But What About Next Time?

But what about the next presidential election? Now that the propensity of COVID-19 death rates have moved over to Republicans as Democrats have proactively vaccinated themselves to prevent, you know, death, will this asymmetry — given enough time — meaningfully dent the results in 2024?

This will take some grim modeling. Every death from coronavirus is tragic, and so many were completely avoidable, but that death toll will continue to increase. Many will be deaths among the unvaccinated, many too will be deaths among the vaccinated. There’s nothing amusing about this — especially when one party is looking the other way as its own supporters die and when so many who are able to get the vaccine are refusing to do so — and this projection is being constructed expressly to demonstrate the potential, raw number, impact of that.

First we need to estimate how many people will have died of the coronavirus between Election Day 2020 and Election Day 2024. Roughly 500,000 Americans have died of COVID-19 in the last year, so we will add that to our projection over the next three years. We’ll base this estimate on average deaths since the widespread availability of the vaccine. Though vaccines were available to all adults starting in April, we will use July 1 to give plenty of time for as many people as possible to get vaccinated, and because this is also a good midpoint for when vaccination rates slowed down. From July 1 to November 1 of 2021, an average of 1,108 people have died from COVID-19 per day. We can run this daily average up to November 5 of 2024 (the date of the next presidential election) and estimate that, from November 1, 2021 to November 5, 2024 (1,100 days), there will be a total of around 1,218,800 more Americans who die of COVID-19 by that point. That’s a death rate of an additional 370 out of 100,000 Americans in the next 3 years. Then we add the 500,000 over the last year and our estimate is that from Election Day 2020 to Election Day 2024 there will have been 1.7 million deaths from COVID-19 in the United States, a death rate of approximately 516 out of 100,000.

The real trick is to estimate proportionally how many of these deaths are Republican voters. To start, we’ll use the death rate per 100,000 per state as of October 29, 2021. It may be intuitive to use data only since the vaccine emerged (which skews strongly in favor of our hypothesis, as more deaths have been in red areas) but we’re actually going to go against that instinct because the underlying factors in the blue states have not changed. Democrats are still more densely clustered in states like New Jersey and New York and some of the most rural states (like Alaska and Wyoming) do not have above-average death rates despite being below-average in vaccination levels. This is why we are explicitly not controlling for urbanization — it can make a difference in overall death rates and there is no way to control for that without meticulously controlling other variables that, at the end of the day, we want to see from a topline overview anyway. Just remember that, if anything, this is a more conservative estimate of the impact of COVID-19 on Republican states than we might expect going forward.

Regardless, running a regression for COVID-19 deaths per 100,000 people and the percent of the vote Trump got in the 2020 election validates that the two variables are correlated:

The percent of the vote Trump received in a given state is a significant variable in the COVID-19 death rate. For each increase by 1% in the Trump vote, there are an additional 1.87 COVID-19 deaths per 100,000 people. Our linear equation is: 

COVID-19 death rate per 100,000 = 119.4 + 1.87 * (Trump vote percentage)

In other words, if a state cast 100% of its votes for Trump, we would predict 189 out of 100,000 more individuals would have died than in a state that cast 0% of its votes for Trump. It’s not all-encompassing in terms of predictive power, but it’s conservative enough to not get carried away. To estimate how many deaths will be Republicans compared to Democrats, we will compare the number of deaths from our formula if we assume a state cast 0% of its vote for Trump with one that cast 100% of its vote for Trump. A state with no Trump votes has a death rate of 119.4, and a state with only Trump voters has a death rate of 306.2. The ratio is about 2.56 to 1, which implies roughly 72% of deaths would be Republican compared with 28% for non-Republicans. This is the most liberal assumption we’re making in this projection, but remember that this number is based off of data that primarily comes before widespread availability of the vaccine. And, only about 10% of Democrats have not been vaccinated, compared with about 40% of Republicans — an even larger ratio of 4 to 1 — but considering how miniscule the risk of COVID-19 becomes once you’ve been vaccinated, our slightly more conservative ratio seems like a fair balance.

Let’s pull out the 12 states4For this analysis, we are removing from consideration the five congressional districts that cast electoral votes separately, those being the three in Nebraska and the two in Maine. that had a margin below 100,000 votes in the 2020 election and run the numbers. We know that about 49% of the American population voted in 2020, and we’ll assume the same for 2024, so, to be conservative, we will assume 51% of COVID-19 deaths are non-voters, and approximately 2% of voters did not vote for either major party. 

In the three closest states in 2020, the disproportionate number of COVID-19 deaths by Republicans is more than 50% of the total margin between Trump and Biden. In fact, total GOP voter deaths in Arizona and Georgia exceed the margin by which Biden won the state. Were the election to be rerun without these voters, Biden’s margin would be roughly twice as large in Georgia, 75% larger in Arizona, and about 30% larger in Wisconsin.

Biden already won all three of these states but, all else being equal, Republican voters are dying at a rate high enough to pad Biden’s narrow margins. Considering Trump would have won the election had he carried these three states where his party is now losing voters to COVID-19, Republicans would be apt to get vaccinated.