So much for that: Looks like the Red/Blue COVID Death Rate Divide didn't disappear after all


Last month I noted that the partisan COVID death rate gap, which had been significantly higher in the reddest U.S. counties than the bluest counties every month since July 2020, appeared to be on the verge of finally disappearing entirely:

The initial COVID wave in March - May 2020, of course, devastated Blue America, primarily densely populated (and heavily Democratic) New York and New Jersey, while leaving Red America (mostly sparsely populated, rural Republican counties) relatively unscathed.

This situation quickly began to reverse itself only a few months later. Starting in July 2020--the same month the Vanity Fair expose was published, as it happens--the situation had already reversed itself: The reddest fifth of the U.S. was already experiencing a higher COVID death rate than the bluest fifth...and it has stayed that way ever since.

That's right: While the COVID death rate ratio between the reddest and bluest parts of the country has grown larger and smaller at various points (reaching an all-time high of nearly 4.5:1 during the height of the Delta wave in September 2021), there has never been a single point in time that the bluest quintile has seen a higher death rate than the reddest quintile since July 2020. And this graph only includes official COVID deaths, not the "likely COVID excess deaths" which were never officially recognized as such...and which ran over 11 times higher in the reddest fifth of the U.S. than the bluest fifth in 2021.

However, while the lines haven't reversed themselves again yet, they've been getting closer for awhile now--the Red/Blue divide dropped below 2:1 in June 2022 and hasn't risen above that since. Most notably, in December, it dropped to virtually even (3.87/100K in the reddest quintile vs. 3.79/100K in the bluest quintile...the closest it's been since May 2021...which, as it happens, is when the original COVID-19 vaccine first became widely available to all U.S. adults nationally.

I went on to note that the reason for the convergence of COVID death rates in both red & blue counties alike seemed to mostly be due to Democratic-leaning residents joining their GOP neighbors in abandoning pretty much all mitigation measures across most of the country: Not only is there hardly a mask to be seen in most public places these days, but only 15.5% of the U.S. population has gotten an updated bivalent booster vaccine shot to date. Combine this with the spread of the new XBB.1.5 COVID-19 variant and it's no wonder that COVID deaths are on the rise again across the political spectrum.

I wrapped up my December update by assuming that the blue and red lines would officially reverse themselves in January, but that:

Assuming the patterns of the past continue with the spread of XBB.1.5, I'd expect to see a spike in rural/red America a couple of months after it peaks in urban/blue America...probably sometime in March or so. The big unknown is how much of a spike either one will be this time around.

Well, the lines didn't flip after all in January--the reddest quintile jumped up faster than the bluest quintile after all--two months earlier than I expected:

  • Bluest Quintile: 4.70 per 100K residents
  • Reddest Quintile: 5.33 per 100K residents (13% higher)

The January gap wasn't that significant by itself...except that it had looked like the rate in the reddest quintile might be lower last month.

Again, the graph below looks at the reddest 20% and bluest 20% of the total U.S. population based on the 2020 Presidential election results at the county level. The reddest quintile includes roughly 65.1 million people living in counties where Donald Trump received 62.4% or more of the vote in 2020; the bluest quintile includes around 66.8 million living in counties where Trump received less than 31.6%. While the raw numbers aren't quite equal, both have been adjusted for actual population.

In short: Get vaccinated, get your bivalent booster, and still wear a mask when attending crowded indoors public events if possible.