Implementation and spillovers of local non‐pharmaceutical interventions

Abstract

Non-pharmaceutical interventions (NPIs) have been key to managing the Covid-19 pandemic. These policies may have important financial and non-monetary costs. Policymakers face a tradeoff between controlling the number of deaths and limiting the burden of containment policies (Alvarez et al., 2020). In this paper, we consider the economic costs and consequences of local NPIs. Specifically, we implement a difference-in-differences framework, estimating effects of municipal policies on mobility, consumer spending, and unemployment.Using difference-in-differences methods to evaluate effects of Covid-19 countermeasures is challenging for multiple reasons (Goodman-Bacon & Marcus, 2020). First, the timing of policies is likely a response to Covid-19 incidence. This policy endogeneity could impact analyses of economic outcomes if higher confirmed incidence has an independent effect on economic behaviors. Second, localities may experience multiple treatments: localities may implement a social distancing mandate when rates are rising, relax the rules when incidence is lower, then re-instate the mandate later if contagion levels go back up. Third, the localities that implement NPIs do so at different times; as a result, regression difference-in-differences may yield biased estimates when treatment effects vary over time (Baker et al.

Local restrictions, population movement, and spillovers during the pandemic: Evidence from Japan's restaurant restriction

“…In fact, few studies have estimated the spillover effects of local restrictions in the pandemic, with the exceptions being Godoy and Grotting (2023) and Holtz et al (2020). Although their focus was not restaurant restrictions, they found a spillover effect of local NPIs (Godoy & Grotting, 2023) and shelter-in-place orders (Holtz et al, 2020) by applying the DID approach to municipality-and county-level data, respectively.…”

“…In fact, few studies have estimated the spillover effects of local restrictions in the pandemic, with the exceptions being Godoy and Grotting (2023) and Holtz et al (2020). Although their focus was not restaurant restrictions, they found a spillover effect of local NPIs (Godoy & Grotting, 2023) and shelter-in-place orders (Holtz et al, 2020) by applying the DID approach to municipality-and county-level data, respectively. In contrast, our study employs a DDD approach by exploiting data with higher spatial resolution (i.e., the 500-m mash level) and a spatial difference in the policy impact (i.e., larger impact on restaurant areas), which allows us to clarify both the main and spillover effects of the RSBH.…”

“…In fact, few studies have estimated the spillover effects of local restrictions in the pandemic, with the exceptions being Godoy and Grotting (2023) and Holtz et al. (2020).…”