New NBER working paper with @jwswallace and @jasonlschwartz on Covid mortality: “Excess death rates for Republicans and Democrats during the COVID-19 pandemic” nber.org/papers/w30512
Ungated on arxiv here: arxiv.org/abs/2209.10751
Political affiliation has emerged as a potential risk factor for COVID-19, amid evidence that Republican-leaning counties have had higher COVID-19 death rates than Democrat-leaning counties and evidence of a link between political party affiliation and vaccination views.
2/ A popular commentary on the Covid crisis has been how much higher the Covid death toll has been in Republican vs. Democratic counties in the U.S.
U.S. Covid Deaths Get Even Redder
The partisan gap in Covid’s death toll has grown faster over the past month than at any previous point.
3/ There are natural reasons to believe that there are strong differences by Republicans vs. Democrats: survey data suggests that there are big differences by party ID on Covid-19 vaccination:
KFF COVID-19 Vaccine Monitor Dashboard
Using a combination of surveys and focus groups, the KFF COVID-19 Vaccine Monitor will track the dynamic nature of public reactions as vaccine development unfolds, including vaccine confidence and ÷
4/ The challenge, of course, is whether it’s really about Republicans vs. Democrats living in these areas, or just the areas where individuals sort into are different.
5/ This statistical analysis runs into a serious challenge, driven by the fact that publicly available data on Covid deaths, and measures of political party, are typically only available at the county level.
6/ The focus on Covid deaths and counties has lead researchers to try to account for these locational differences (by controlling for features at the county level), but are still limited by the aggregated nature of the data:
The Association Between COVID-19 Mortality And The County-Level Partisan Divide In The United States | Health Affairs Journal
Partisan differences in attitudes toward the COVID-19 pandemic and toward the appropriateness of local policies requiring masks, social distancing, and vaccines are apparent in the United States.
7/ The other issue with this approach is that it focuses on reported Covid deaths as an aggregate measure. This measure may not fully capture the “counterfactual” deaths in the absence of the pandemic. Our world in data does an excellent discussion:
Excess mortality during the Coronavirus pandemic (COVID-19)
Excess mortality is a term used in epidemiology and public health that refers to the number of deaths from all causes during a crisis above and beyond what we would have expected to see under ‘normal’…
8/ Intuitively, calculating excess death rates requires a prediction of death rates in 2020 and 2021 based on previous years for the group of interest: namely Democrats and Republicans. Fortunately, we have mortality data with party affiliation, age, and location in this paper!
9/ We construct data using individual-level voter registration in 2017, linked to death records from 2018 to 2021, for Ohio and Florida. We then construct excess death rates that control for differences in mortality rates (pre-Covid) at the age-by-party-by-county-by-month level
This lets us ask and answer three questions:
11/ Q1: Does excess death in 2020 and 2021 differ by political party, how much and when does this occur?
A1: Yes, the excess death rate for Republicans was 5.4 p.p., or 76%, higher than for Democrats. The gap was exclusively in the post-vaccine period (10.4 pp or 153%).
12/ Q2: Is this difference explained by geographic or age differences in political party affiliation?
A tiny share of the difference is explained by differential impacts of age-by-county during Covid (recall that excess deaths already controls for pre-Covid differences):
13/ Q3: How much can we point to vaccines?
A3: This is harder, since we don’t have individual-level data on vaccines. However, two facts emerge:
A. The association between the Rep.-Dem. gap and county-level vaccination rates grows significantly after they become available:
B. Moreover, pre-vaccine, the relationship across counties between realized vax rates and excess deaths was identical for both groups.
Post-vaccine, the Democrat rate fell and Republican rate climbed; and the gap between the two was near zero in high-vax counties.
16/ If this is really a story about vaccines, the continued story of low take-up of vaccines + boosters among Republicans may perpetuate some of these differences:
KFF COVID-19 Vaccine Monitor: September 2022
Our latest Vaccine Monitor survey finds that half of the public has heard either “a lot” or “some” about the newly-available bivalent COVID-19 boosters, and a third (32%) say they’ve already gotten…
17/ We’re working on expanding this out now to contrast our results with the existing literature a bit and highlight some more points, but would welcome any comments or suggestions.
fin/ It is important to reiterate that our results hold fixed differences in mortality by age, location, and party pre-Covid, and can account for location-by-age differences post-Covid. Hence these are within-age-and-location differences in mortality outcomes by political party.