Spatial econometric modelling of US COVID-19 policy stringency

In this paper, we study the influence exerted by a state's 'neighbours' in determining its COVID-19 policy stringency. In addition to traditional geographic neighbours, we also consider political neighbours (i.e. states related by the political affiliation of their governors). By employing a Spatial Durbin Model on a panel of 48 contiguous US states and Washington DC over 61 biweekly time periods, we find evidence for both endogenous and exogenous spillover effects. Our results suggest that both geography and politics represent significant endogenous transmission mechanisms for policy stringency. When considered individually, the political spillover effects are greater in magnitude than those driven by geography. However, our best specified model incorporates a convex combination of both geography and politics and indicates that, on average, about 30% of a state's policy stringency is attributable to its political and geographic neighbours.