Testing and reopening in an SEIR model

We quantify how testing and targeted quarantine make it possible to reopen an economy in such a way that output increases while deaths are reduced. We augment a standard Susceptible-Exposed-Infectious-Recovered (SEIR) model with (i) virological testing, (ii) serological testing, (iii) permanently asymptomatic individuals, (iv) incomplete information, and (v) a reduced form behavioral response of reopening to changes in health risks. Virological testing allows for targeted quarantine of asymptotic spreaders. Serological testing allows for targeted release of recovered individuals. We fit our model to U.S. data. Virological tests every two weeks accommodate more aggressive reopening that more than halves output losses while keeping deaths below forecasts under the status quo. Serological tests are much less effective. Implementing testing against a fixed budget, low sensitivity tests that are cheap and used frequently, dominate perfect tests that are expensive and used less frequently.

Duke Scholars

Author David Berger Economics