State Capacity and Covid-19 Responses: Comparing the U.S. States

• In Event: Plural Approaches to Understanding COVID-19 Response

Abstract

How do variations in state capacity among U.S. states account for the speed and success of Covid-19 responses? We aim to create a measure of subnational state capacity that goes beyond traditional measures of wealth and resources. In line with Berwick and Christia (2018), we approach state capacity in terms of three underlying domains: extraction, coordination, and compliance. Extractive capacity is the ability of states to mobilize resources and raise revenue. Coordination relates to the ability to organize for collective action or the relationship between state agents and civil society. Compliance refers to the ability to overcome principal agent dilemmas and involves interactions between state agents, bureaucrats, and non-state actors. We view state capacity as the latent ability to accomplish these three tasks, namely, the state’s ability to enact its will (which is separate from the political will to accomplish a particular action).

We use a latent trait modeling approach to measure state capacity. To do this, we leverage different indicators of outcomes caused by a singular dimension of state capacity that manifests in a variety of domains. This measurement approach allows us to focus on the downstream consequences of differential capacity, as opposed to theorizing about the necessary conditions. We will therefore create a singular framework to test if Covid-19 responses, such as vaccine distribution, fit this measure or not. Using a variety of indicators to identify state capacity will ensure that we do not predict health policy success from previous health policy successes. We will combine health and non-health indicators, such as maternal mortality, violent crime rate, and auditing capacity. These items reflect goals that are largely shared across states, allowing us to tap into capacity rather than particular instances of will.

We predict that state capacity will lead to more effective vaccine distribution. This is an excellent case to assess the predictive capacity of our state capacity measure because we can largely remove political will from the equation. We make the mild assumption that all states want to vaccinate the largest possible proportion of their citizens and that they want to use as many of their allotted shot units as possible. We will also control for institutional factors, such as the delegative structure of the health system and political parties in state government, that could influence the efficacy apart from general capacity. Our work will contribute to the literature on U.S. state politics and the literature on state capacity by creating a valid and comprehensive measure of state capacity at the subnational level. In addition, our research will help to better understand which aspects of state health infrastructure matter most for health outcomes.

Authors

• Kiran Rose Auerbach, University of Bergen

• Joshua Yoshio Lerner, NORC at the University of Chicago

• Hannah Ridge, University of Chicago