UPDATE: A few people have reasonably asked about my methodology and data sources. All of that’s included at the blog post links, but here it is again.
I should also note that my own methodology, while not as sophisticated as I’d like, concluded as early as February that around 71% more GOP voters likely died of COVID between the 2020 & 2022 elections than Dem voters.
Last month, a more sophisticated analysis by professional analysts at the National Bureau of Economic Research concluded that around 76% more GOP voters likely died of COVID than Dem voters.
In other words, if anything, it looks like I (slightly) underestimated the impact, although the timeframe for my analysis is also longer than theirs.
While the NBER study only applies to the national estimates, not the state-level ones, it makes me pretty confident that I’m well within the ballpark there as well.
A lot of people have been speculating about this question over the past two weeks:
Several folks keep citing Jonathan Last’s quote from the Bulwark article cited in the last tweet, in which he says:
So again, 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. Part of this is because of vaccine hesitancy; part of it is because of the age profile of voters.
I’m not going to burden you with the math here, but if you want to read up on it, the data is quite striking, all the way to the county level.
To take just one example: between January 2021 and this month, 9,400 people in Nevada died of COVID. The data suggests that the majority of these people would have been Republican voters. Keep that number in mind.
It’s been incredibly frustrating to read all of these hot takes and speculation seeing how yes, I’ve been tracking and projecting the likely real-world impact of the GOP/Dem COVID Death Rate Divide on the midterms for a solid year now.
I first attempted to tackle the issue back in September 2021, where I concluded:
Let's assume it ends up being something like 150,000 more Americans dying of COVID between now and [November 2022], at roughly the same ratios as in the June - September data above (that is, roughly 80% GOP & 15% Dem). That would boil down to something like 74,000 additional Republican voters and 14,000 additional Democratic voters, for a net loss of 60,000 more GOP than Dem voters nationally.
The grand total would then stand at somewhere around ~105,000 Democratic voters, ~220,000 Republican voters and ~26,000 swing voters having died nationally between 11/04/20 - 11/08/22...or a net loss of perhaps 115,000 more Republicans than Democrats overall.
I also warned at the time, however, that:
...while that may sound like a high number, keep in mind that most of those either live in counties (and therefore, indirectly, districts) which are already pretty Blue or in counties/districts which are solidly Red to begin with. A GOP Congressperson used to winning with 75% of the vote isn't going to lose much sleep over that number being knocked down to 70% or even 65%.
There may be a handful of swing districts which will be decided by just a few hundred votes where it ends up making a difference...but if losing 115,000 more GOP voters results in, say, 500,000 independent/swing voters flipping back to the GOP (or at least not bothering to vote at all) because "Biden said he'd stop the pandemic and he hasn't" etc etc, then that would still result in a net gain for Team Red.
In October 2021 I ran an updated version of my national projection.
In December 2021 I ran another update, this time attempting to break it out by state...and again, I reiterated that it was unlikely to play a major role. I even made that the headline:
Elephant In The Room: Yes, More Trump Voters Have Likely Died Of COVID Than Biden Voters. No, It Likely Won't Make Much Difference.
In February 2021, I re-ran my projection, but this time I also attempted to adjust the raw state-level death toll by age using data from the CDC, U.S. Census Bureau and NY Times 2020 Exit Polling.
I ran updated versions of this in March, April, and one more time in September 2022.
At this point my estimate of the 2022 Midterm Red/Blue Voter COVID Death Gap was up to roughly 150,000 nationally, ranging from ~4,700 more likely Democratic voters having died of COVID in the prior 2 years than Republican voters in California to a whopping 19,200 more Republican voters having died of COVID than Democratic voters in Texas.
And yet, I still urged caution:
Again: The net gap between Trump & Biden COVID deaths will likely continue to increase slightly more between now and November 8th, but aside from a handful of extremely close races, I doubt that will be enough to make the difference in who ends up winning.
Well, as it happens, my projections have been proven exactly right.
At the statewide level, nationally, there’s only 15 statewide races where a Democrat won (or is likely to win) by a smaller margin than my estimates of the GOP COVID Death Cult Gap:
Of those, as you can see, only one of them is actually within the high-end estimate: Arizona Attorney General, where Democrat Kris Mayes is up by just 850 votes over Republican Abraham Hamadeh. There’s a few thousand ballots left to report later today, but they’re unlikely to change the outcome unless Hamadeh wins them by something like a 4:1 margin.
UPDATE: Welp. There it is:
Mayes won by 510 votes prior to the automatic recount (which is unlikely to move more than a few dozen votes at most):
OK, but what about HOUSE races?
Well, those get a bit trickier due to the way in which legislative districts cut through multiple county borders, although there’s only around 400 counties (out of over 3,100 nationally) where that happens.
However, I’ve been able to roughly approximate the percent of each of these counties which are parts of various Congressional Districts. It’s hardly ideal but it should be reasonably close.
Based on that data, as of this writing, there’s only four House races which haven’t been called for either the Democratic or Republican candidate; as far as I know, the other 431 races were either won by the GOP or, if the Democrat won, they did so by over 1,000 votes...well over the several hundred voter COVID death margin I’ve estimated for each:
“But Brainwrap,” you may be asking...”aren’t there districts where thousands more Republicans died of COVID than Democrats?”
Yes, that’s true...but again, as I said over a year ago:
...most of those either live in counties (and therefore, indirectly, districts) which are already pretty Blue or in counties/districts which are solidly Red to begin with. A GOP Congressperson used to winning with 75% of the vote isn't going to lose much sleep over that number being knocked down to 70% or even 65%.
Sure enough, here's two perfect examples of this. AL-04 and KY-05 are two of the districts with the highest GOP/Dem COVID death gap...and also two of the reddest House districts in the country. As you an see, while I estimate that around ~1,900 more Trump voters died of COVID since the 2020 election than Biden voters in each district, it's barely a rounding error compared to the ~137,000 vote difference between the Republican and Democratic candidates in each race:
To be clear, there probably are a few dozen local races around the country where the GOP COVID Death Cult factor determined the winner (county executives, city council, mayor, local sheriff, school board, etc), but those are so numerous and the numbers so granular I’d have no way of proving or even speculating about most of them.
UPDATE: It’s important to clarify that I’m not saying it didn’t have any impact; clearly it did. It just wasn’t the determining factor in any races other than AZ Attorney General & a handful of other local races.
Going forward, while the trend will likely continue to some degree, as long as the raw COVID death numbers remain at the ~300/day range that they’re currently at or lower (and/or there’s some newer, even uglier variant or different disease which follows a similar pattern, which is certainly possible), I wouldn’t expect the impact to be significantly more in 2024 or beyond...at least not in a clear, measurable way.