Last weekend, the Cleveland Plain Dealer released a 50-state survey of state responses to the COVID-19 pandemic. State policy is rapidly changing and by the time this blog post is published many of these state policies will have changed, but in the meantime this is an interesting snapshot of the US’s patchwork quilt of social distancing measures enforced at the state level.
From a qualitative standpoint, it’s easy to make conclusions from this map. Liberal California and New York anchor the hotspots of state response to the pandemic. It is telling that New York, California, and Illinois, home to the country’s three largest metropolitan areas (New York, Los Angeles, and Chicago) and controlled by Democrats, also have adopted the highest level of restriction at this point. Also notable is the fact that Texas, home to the next two largest metropolitan areas (Dallas and Houston) and controlled by Republicans, has less restrictions.
Something that interested me, though, was whether these restrictions were related to the incidence of disease in a given state. Theoretically, a function of federalism is to allow different jurisdictions to react to local conditions. Therefore, we should expect to see states that have larger outbreaks adopt stricter measures and states with smaller outbreaks to adopt laxer measures.
Using data from the Johns Hopkins Center for Systems Science and Engineering, we can compare the cases identified in a state to the level of restrictions imposed by the state. In the chart below, higher restriction levels correspond to more restrictions as reported by the Cleveland Plain Dealer.
Data from Johns Hopkins Center for Systems Science and Engineering and Cleveland Plain Dealer. New York excluded as a high-case outlier.
Overall, I find that there is a very weak relationship between the number of cases in the state on the day the Cleveland Plain Dealer article was published and the restrictions they have in place. The number of cases in a given state only explain 7.5% of the variation between states. Adjusting for per-capita case rates only improve this number slightly and using older data to account for a lag only weakens the relationship. Note that Washington state, the original epicenter of the coronavirus outbreak in the U.S., has relatively weak restrictions in place despite a large number of cases.
Another approach I was interested in taking was to compare death rates to restrictions. If policymakers weren’t responding to cases of coronavirus, maybe they were responding to loss of life associated with the disease.
Data from Johns Hopkins Center for Systems Science and Engineering and Cleveland Plain Dealer. New York and Washington excluded as high-case outliers.
The relationship here is nonexistent, with only 0.3% of the variation in restrictions explained by death cases in the states. Like the relationship between restrictions and cases, the relationship gets slightly stronger if deaths are measured on a per capita basis and slightly weaker if lagged, but neither of these changes are enough to significantly strengthen the relationship. With this chart we can even see the variation among higher-death states, with California adopting strong restrictions, Hawaii and New Mexico adopting laxer restrictions, and Maine and Georgia acting even more lax.
Overall, what the relationship between outbreak data and state restriction data says is that state decisions are not being driven by conditions on the ground, but by other factors. While often federalism allows for state-by-state specialization and adaptation to local conditions, a quickly-moving global threat like COVID-19 can make patchwork policy less effective and even creates the potential to undermine efforts from state to state.