Does Ideology Kill?
Interpreting the association between conservatism and COVID death rates.
There is mounting evidence of an association between conservative politics and COVID impacts. Indeed, the higher death rate among Republicans may even have swung some close elections. A recent study sheds light on how ideology and death rates interact. As the Washington Post reports, the results were striking: “Covid death rates were 11 percent higher in states with Republican-controlled governments and 26 percent higher in areas where voters lean conservative. Similar results emerged about hospital ICU capacity when the concentration of political power in a state was conservative.”
A deeper dive into the study is needed to fully understand these results. The researchers controlled for a host of variables, including demographic characteristics, and in some models for some health indicators (obesity and diabetes rates) and for COVID vaccination rate. The study is not easy to interpret because the paper is very terse and because the authors used multiple models testing a variety of interactions. Basically, the study found that three separate political variables seemed to significant effects: (1) the partisan lean of a district’s voters, (2) the stance of their congressional representative (gauged by votes on COVID measures and by a measure of ideology), and (3) the degree of GOP control of state government. The fact that the three variables seem to have somewhat independent connections with COVID outcomes is intriguing, because it suggests that there may be different causal mechanisms at work.
To start with partisan lean, one possibility is that ideology matters because it drives local public health decisions. Being in a GOP-leaning congressional district makes it more likely that you in live in a conservative county or town. Local-level health measures in turn could affect COVID outcomes. Alternatively, partisan lean might matter because it is tied with health behaviors. For instance, if you live in a district with a lot of Republican voters, you’re more likely to be Republican yourself. You may therefore be less likely to mask as a matter of tribal loyalty. You may also be surrounded by people who are less likely to mask.
If the politics of congressional representatives has a connection with health outcomes, it seems fairly unlikely that the link is direct. Individual members of Congress don’t make policy for their districts. (They might conceivably have some influence on the distribution of federal COVID funding, so the possibility of a direct link isn’t completely out of the questions, but that seems speculative.) Perhaps congressional representatives are sufficiently prominent that their views influence those of residents about issues like masking. The politics of representatives could instead be a signal of something else, like the nature of local Republican politics. Even in districts that have equal partisan make-ups, some could have differ in the degree to which voters or representatives are conservative, leading to different individual behaviors and local policies.
Assuming the results are valid, the connection between Republican control of a state and COVID outcomes seems to be independent of the district-level findings. This seems to indicate that state level policies had a substantial impact beyond their correlation with differences in individual behavior or local policies.
As the authors of the study say themselves, it is perilous to draw conclusions about causation from evidence that various factors are associated with worse outcomes. One reason for taking the findings seriously that we know conservatives have opposed health measures like masking and are more resistant to vaccination despite clear evidence of safety and effectiveness.
However, isolating the causal connections is likely to be very difficult because there are so many possible links between variables. Proving causation between individual ideology and health outcomes is tough, because ideology is related to many other factors and tends to be persistent. It may be easier to tease out the independent effect of state level and district level political variables on health outcomes.
If causation is truly at work, its implications may themselves be subject to ideological polarizations. Liberals will interpret the results as evidence that conservatism harms society. Conservatives may be more prone to think that a higher death rate is an acceptable price for greater personal liberty and smaller government. Researchers cannot change people’s minds about these deep normative issues, but at least they may be able to make the tradeoffs clearer.
It would also be worthwhile to look at the link between ideology and public health in other contexts. For instance, given similar city characteristics, it would be useful to see the extent to which differences in ideology at the state level impact pollution levels and therefore illness and mortality rates.