Pro-Trump Partisanship and COVID-19 Mortality: A Model-Based Counterfactual Analysis

We show that a higher share of Trump voters (who are less likely to comply with COVID-19 public health guidelines than Democratic voters) in a U.S. county leads to significantly more COVID-19 deaths during times of high regional infection risk. Our model-based counterfactual analysis finds that about 15 percent of the cumulative death rates in pro-Trump counties after the first year of the pandemic can be explained by a pro-Trump partisanship effect. The analysis considers demographic and socioeconomic differences between counties, unobserved heterogeneity on county and interacted week x state level, and non-linear effects due to spatiotemporal differences in infection risks.