Skip to content

Did vaccine denial play a role in Republican midterm losses?

A reader asks what I think of the theory that Republican losses in the midterms were partially due to Republicans killing off their own voters by railing against COVID-19 vaccines. He points to this piece by Jonathan Last as an example:

Lots of factors were at play. Including one that doesn’t get talked about much: excess COVID deaths. There’s been an ongoing study of the Republican resistance to the COVID vaccines and the preliminary findings suggest that post-vaccine, Republicans accounted for about 80 percent more of the excess deaths than Democrats.

....Many of the Republican losses were by extremely small margins. For example: Adam Laxalt lost the Nevada Senate race by 6,000 votes. (Go look at that COVID number again.)

The best estimate of partisan differences in COVID deaths that I know of is from an NBER paper that Last links to here. Three Yale researchers concluded that in the post-vaccine era (starting around May 2021) the excess death rate for Republicans was 5.4 percentage points higher than it was for Democrats. That's everything we need to know. All that's left is a bit of arithmetic:

  • The overall excess death rate in the US during this period was about 600,000, or 12%.
  • Therefore, the D excess rate is 9.3% and the R excess rate is 14.7%.
  • Assuming a roughly even split of D and R voters, that comes to 232,000 excess D deaths and 367,000 excess R deaths.
  • That's a difference of 135,000 extra Republican deaths.

That's not enough, no matter how many what ifs you toss out. Nevada, for example, accounts for about 1% of the total US population, so it probably accounts for about 1,300 extra Republican deaths. Even in the ultra-tight Senate race, it means only that Adam Laxalt lost by 6,000 votes instead of 5,000.

The COVID theory is an appealing one, but the numbers just aren't there. In a race decided by a tenth of a percent or so, you might argue that vaccine denial played a role. But that's how close it would have to be.¹

¹The closest Senate race was Nevada, with a winning D margin of 0.82%. The closest governor's race was Arizona at 0.78%. The closest House race at the moment is NM 2 at 0.66%, though that may change. None of these are close enough to have been plausibly decided by COVID deaths.

31 thoughts on “Did vaccine denial play a role in Republican midterm losses?

  1. NeilWilson

    I calculated it to be between 250 and 1,000 people per congressional district.
    So it probably will change 1 or 2 house races.

  2. Joseph Harbin

    This syncs with Charles Gaba, who did an in-depth analysis. The only race where Covid deaths may have swung an election is Attorney General of Arizona, where Democrat Kris Mayes leads GOP-er Abraham Hamadeh.

    Covid deaths deficit for GOP: 4,100 (estimated)
    D margin in AG race: 2,269 (as of 1pm ET, 11/15/22)

    Gaba: https://acasignups.net/22/11/15/update-elephant-room-redux-gops-covid-death-cult-made-difference-exactly-one-statewide-race
    AZ tally: https://www.nytimes.com/interactive/2022/11/08/us/elections/results-arizona-attorney-general.html

  3. greyhair

    Kevin, you make a fundamental assumption. You've divided the "death votes" equally between the parties. The article suggests, (not unreasonably) that more of the "death votes" would be Republican than Democratic. Given the deaths are predominately among seniors, this is quite possible.

  4. climatemusings

    One quick check: didn't the NBER paper say that "Post- vaccines, the excess death rate gap between Republicans and Democrats widened from 1.6 pp (22% of the Democrat excess death rate) to 10.4 pp (153% of the Democrat excess death rate)." Which makes the D excess rate 6.8% and the R excess rate 17.2%. So 200,000 more excess R deaths. (I think you are also underestimating total excess deaths, but I also think that not every death is a voter, so I'll call those even)

    1. climatemusings

      (and I recognize that this doesn't change the bottom-line that the chances that COVID deaths would be enough to alter an election are very very small, but maybe a bit bigger than your top post would suggest)

  5. jvoe

    I think we are going to see that Republican morbidity has increased overall. Rural hospitals got hammered and many doctors left. There is long covid, which is walking death for many people and probably makes voting seem unimportant. And as pointed out in an earlier Drum post, there is widespread vaccine hesitancy among Republicans. Even the freaking flu shot is being questioned now. It's full on anger/skepticism toward the health care apparatus and its gonna have widespread consequences for years.

  6. Citizen99

    It's not a binary choice (either Thing A was the entire cause, or it was nothing). It's more complicated than that. Maybe some traditional GOP voters in some districts who lost loved ones to COVID changed their votes. What about age? In some districts, virtually all of the GOP voters are old and all of the Dem voters are young, so the impact could much larger than a proportional calculation would indicate. Regardless, it surely was a factor.

    1. kaleberg

      There's also the matter of whether someone might be too sick to get out and vote. COVID doesn't kill everyone, but it can make one sick enough to not want to go out and do things like go to work, run errands or vote. If you are vaccinated, you are less likely to get that sick, though I have known a few vaccinated folks who were pretty miserable.

  7. jobywalker

    You assumed all those excess deaths were voters -- even among the elderly turnout in a non-presidential year is much lower than 100%.

    1. Jerry O'Brien

      Of course, it's hard not to feel some satisfaction when those advancing unwisdom are rewarded with unsuccess.

      There can still be a humane impulse, though, toward looking at the human costs of following bad guidance, and emphasizing those costs by noting how such misguided leadership even hurts the demagogues.

  8. sonofthereturnofaptidude

    Watching all the GOP resistance to vaccines and the consequences of them may have had effects on voting, too.

  9. D_Ohrk_E1

    Skip the rigamarole that gives you a broadly generalized conclusion.

    Just go to Nevada's (or any given state's) data on COVID-19 deaths.

  10. kenalovell

    It will be a different story next time. After the death jabs have wiped out Fauci liberals, Republicans will win in a landslide.

  11. Austin

    Ugh. Thanks for another half assed analysis Kevin.

    You assume Republican (and Democratic) deaths are equally and proportionally distributed across all House districts and Senate districts (“states”), when you know death rates were higher in some regions vs others. It’s entirely possible redder states managed to kill off more than their “fair share” of either or both parties’ voters, while bluer states managed to collectively act and save more than their “fair share” of either or both parties’ voters despite their efforts to destroy themselves through poor choices. You should’ve looked up what excess deaths were for each state - since some states punched above their weight in killing their own residents - and then applied the party split to those figures to find out if, say, NV really had enough excess deaths to make a difference.

    1. jimminy

      One of the problems, including Gaba's analyses, is that the deaths in a "reddish" district are distributed uniformly among the population in that district. This ignores individual behavior. If the Ds in the district masked, distanced, and vaccinated, then as individuals they'd have better survival rates than their R neighbors who did none of these things.

      Unfortunately, it's hard to do analyses at the right granularity to make sense.

      Note that according to the CDC, some states (FL, MI, MS) have been reporting
      a "pneumonia not related to COVID deaths" epidemic that exceeds the number of COVID deaths they report. Suspicious, no? I tried to get Charlie Crist's campaign to
      mention this. Still, FL went Red due to election manipulations.

      See https://www.cdc.gov/nchs/nvss/vsrr/covid19/index.htm
      to see the numbers. Scroll back a few months and look at various states.

  12. memyselfandi

    "The overall excess death rate in the US during this period was about 600,000, or 12%." Where did Kevin get this number? Worldometer lists 500k covid deaths in this period so excess deaths were 1 million. (Excess deaths are double covid deaths since many deaths caused by covid such as hen the virus attacks the heart or other organs are not called covid.

    1. illilillili

      "Excess deaths" is an estimate of the number of deaths in a year over the expected number of deaths if there were no Covid. It's another way of measuring Covid deaths. Many Covid deaths are unreported or mis-reported, and :"excess deaths" is a way of estimating the actual number of deaths attributable to Covid.

Comments are closed.