We propose a spatial epidemic spread model to study the COVID-19 epidemic. In our model, a city consists of multiple neighborhoods, each of which has five disease compartments (susceptible/exposed/infected clinical/infected subclinical/recovered). Due to the movement of individuals across neighborhoods (e.g., commuting to work), the infections in one neighborhood can trigger infections in others. We consider the problem of a planner who reduces the economic activity in a targeted way to curb the spread of the epidemic. We focus both on the regime with a small number of infections and the regime with a large number of infections, and provide a framework for obtaining the policies that induce the lowest economic costs.

We use the available data on individuals’ movements, level of economic activity in different neighborhoods, and the state of the epidemic to apply our framework to the control of the epidemic in NYC. Our results indicate that targeted closures can achieve the same policy goals at substantially lower economic losses than city-wide closure policies. In addition, to curb the spread of the epidemic in NYC, coordination with other counties is paramount. Finally, the optimal policy (under different scenarios) promotes some level of economic activity in Midtown Manhattan locations (due to their economic importance) while imposing closures in many other neighborhoods in the city (to curb the spread of the disease). Contrary to what might be intuitively expected, and due to the spatial aspect of the epidemic spread, neighborhoods with higher level of infections should not necessarily be the ones exposed to the most stringent economic closure measures.