A Dynamic Spatial Framework Linking Disease Transmission and Economic Activity

We
develop a disease-agnostic framework
for the co-evolution of an infectious disease, economic activity, and mortality, estimated via a multivariate
spatial dynamic panel with interactive fixed effects. The structure enables
decision-relevant counterfactual simulations before resistance is widespread. Applied
to COVID-19 in large U.S. counties in 2020, we find that cases and economic
activity are highly persistent; increases in activity raise infections, which in
turn depress subsequent activity. Simulations indicate that optimal long-run
policies include strict mask requirements and can include lockdowns; optimal
stringency increases with population and network centrality. In the COVID-19
case, suboptimal policies resulted in 10% less economic activity and 6,000 times more cases by December
2020. This framework is portable to other pathogens once re-estimated with
context-specific parameters.