Measuring the Complexity of US State COVID-19 Policy Mixes

• In Event: Policy Design Complexity, Bundles, and Mixes

Abstract

COVID-19 highlighted two critical aspects of American political and policy systems: the limitations of the federal system in making effective response decisions under high uncertainty and the challenges of policy design on the complex and dynamic problem that requires the government’s multipronged strategies simultaneously. The need for multifaceted strategies that can address multiple aspects of disease response has been well recognized in public health (e.g., Legido-Quigley et al., 2020) best described with the term, policy mix, in the policy literature (Chapman, 2003; Howlett, 2019). Policy mixes, which utilize several policy instruments to address social problems, are increasingly scrutinized in areas such as innovation policy or environmental policy. These multipronged policy strategies have also raised policy design and evaluation challenges. In this paper, we structure, measure, and compare the complexity of US State COVID-19 policy mixes using two approaches: (1) The Average Instrument Diversity (AID) Index proposed by Fernández-I-Marín, Knill, & Steinebach (2021) for OECD countries’ environmental policy mixes and (2) A novel social network analysis inspired approach that treats instruments and the policy targets that they seek to address as two-mode networks or bi-partite graphs which will enable researchers to employ social network metrics, such as density, clustering coefficients, and cliques to identify commonly used implementation strategies or differentiate among implementation patterns. By reviewing the social distancing policy data from the University of Washington and associated original policy documents, we identified six policy instruments (emergency authority, restrictions, closure, public health safety requirement, case management, and quarantine), (b) 10 policy targets (state residents, state high-risk groups, out-of-state travelers, close contacts of suspected cases, COVID-19 positive cases, bar, restaurants, other businesses, schools, government agencies), and (c) instrument and target combinations defined in the policy documents. After constructing the policy mix dataset, we calculate and report the AID score as well as network-based measures from March to December in 2020 for all 50 states. Our preliminary results show that states’ AID scores were similar at the beginning of the pandemic, but the scores changed and diverged over time. Network measures of the patterns for utilization or employment of instruments for specific policy targets promise to provide a more nuanced model for describing and comparing policy mixes. The example of US State COVID-19 policy mixes provides a rich example of the role and challenges of states’ policy designs that rely on multiple policy instruments simultaneously for complex problems such as climate change in the post-pandemic era.

Authors

• Yushim Kim

• Elisa Jayne Bienenstock, Arizona State University

• Karen Mossberger, Arizona State University

• Chan Wang, Arizona State University