Come November, COVID-19 will have killed enough Americans that it could have an effect on the outcome of some elections.

We don’t mean that people who vote for presumptive Democratic presidential nominee Joe Biden because they think President Trump screwed up the response to the pandemic, or people who vote Republican because they are mad about Democrats advocating the wearing of masks. Although those effects may be massive, we’re talking about the voters who won’t be able to vote by Election Day. Will the 250,000 to 300,000 voters that, according to at least one projection, will have joined the invisible electorate in the sky have an impact on close races?

We mean no disrespect to the dead or to their grieving families. COVID-19 is a national tragedy and has resulted in tremendous suffering. 

But COVID-19, now the third biggest cause of death in this country, is so serious that it merits analysis. And with the way we are polarized today, COVID-19 will have a death effect precisely because it hits Americans unequally. In America in 2020, even death is polarized.

We’ve run some of the numbers and have identified several races where the grim reaper of COVID-19 could help determine the outcome.

Consider two harsh demographic realities of COVID-19: It is much more fatal to the elderly, and to Hispanics, African Americans and Native Americans. When it comes to voting, the white elderly are a vital part of the Republican base, particularly in the age of Trump. Conversely, non-white voters strongly favor Democrats. 

We analyzed states in which polls show Biden and Trump are close, and also toss-up Senate races. Then we adjusted the COVID-19 numbers for likely voter turnout and projected the number of dead forward to November. These are rough estimates based on publicly available data; our aim is to take the best data from epidemiology, public health and political science and see if COVID-19 will influence the elections.

Overall, the death effect seems to give a modest boost to Republican chances. This is because Black and Hispanic voters favor Democrats at much higher rates than the white elderly favor Republicans. So, for every 10 deaths among African American voters, the Democrats lose eight or nine votes, while the Republicans lose only one or two. Among the white elderly, when 10 die, the Republicans lose six votes, and the Democrats lose four. Even though the white elderly are much more reliable voters, the turnout differential doesn’t come close to the gap in partisanship. Consequently, nationwide, we predict a roughly 22,000-vote advantage for Trump in the popular vote as a result of COVID-19 deaths. 

If it seems cruel to even suggest that mishandling a pandemic so badly could help Trump win, Democrats can take comfort in a few things. In an election that could see more than 150 million votes, 22,000 is miniscule. And Democrats know better than anyone that the national popular vote doesn’t matter. 

So, let’s look at battleground states. The state where the death effect is most likely to matter is Georgia. Georgia has been badly hit by the pandemic, particularly among its sizable Black population. We begin with the current CDC death tolls for Georgia Blacks, Hispanics and whites: 2,906; 522; and 2,175. It should be noted that CDC numbers for deaths have about a one- to four-week lag, so these are really numbers from the start of July. We adjust for the differential in deaths by men and women, because men and women have a gender gap in voting. Then we adjust for voter turnout rates, and assume that the July numbers will have doubled by November 3, based on Institute for Health Metrics and Evaluation projections. When we subtract expected deceased Democratic votes from expected deceased Republican votes, we find that Republicans end up with a 2,258-vote advantage.

North Carolina, another Southern battleground state, could also come down to COVID effects, with the margin there raising Republican hopes by 1,368 votes. Texas, which some Democrats think is in play, will be made 3,060 votes harder for the Democrats to take. 

But again, there is good news for Democrats. If Georgia and Texas are close enough for these small numbers to matter, then Trump will almost certainly lose Florida, Pennsylvania, Michigan, Wisconsin and other key states.

Are there any states where the death effect could help Democrats? Yes, in states with lots of old white people but few minorities. But the effect is so tiny – a few hundred votes – even in a small state like Iowa.

In Senate races, Georgia is again the big story. With two competitive Senate races, the death effect could be the difference in either or both of them. Similarly, the reelection hopes of Republican Sen. Thom Tillis in North Carolina are slightly boosted by the death effect of COVID-19. 

Our results are only as predictable as the virus. If we suddenly get better at handling it, and fewer Americans die between now and November, it’s unlikely to have as large an effect. Conversely, if it gets significantly worse, it will affect more races. 

But right now, COVID-19 could help Trump in a few states, and will help Senate Republicans by giving a small boost to incumbents in states such as Georgia and North Carolina.

Of course, there are many other variables that we do not take into account. For one, our numbers assume that living in a community where many people are dying from COVID-19 doesn’t affect how the survivors vote. It seems likely that as deaths from the pandemic climb in a community, Republicans will pay a higher price for Trump’s ineptitude in the face of this crisis. That “accountability effect” could outweigh the death effect by many thousands of votes.

Maybe one positive thing to take away from this horrible national tragedy would be if every relative of a COVID-19 victim pledged to do what their loved one can no longer do: vote — as if their life depended on it.

Jeremy D. Mayer is an associate professor at the Schar School of Policy and Government, George Mason University, and the coauthor most recently of “The South and the Transformation of U.S. Politics” (Oxford 2019). Laurie A. Schintler is also an associate professor at Schar, where she directs the Masters Program in Transportation Policy and Logistics, and researches and publishes about Big Data, emerging technologies and other topics.