How Long Should Social Distancing Last? Predicting Time to Moderation, Control, and Containment of COVID-19
7 Pages Posted: 30 Mar 2020 Last revised: 22 Apr 2020
Date Written: March 28, 2020
Lockdowns and stay-at-home orders in response to the Covid-19 pandemic have raised an urgent question in peoples’ minds, “How long must these restrictions last?” We propose two metrics of the spread of disease to answer this question: daily growth rate and time to double cumulative cases. These metrics enable three simple, intuitive, and actionable benchmarks to target: Moderation, Control, and Containment (growth < 10%, 1%, and 0.1% respectively). In addition, we define action or intervention as massive testing and quarantine, stay-at-home orders, or lockdowns.
An analysis of top 36 countries and 50 states of the US affected by the epidemic as of end-March yield the following results. Any moderation or slowdown has so far been due only to aggressive intervention. Countries take an average of about three weeks to act. However, even aggressive intervention does not show immediate results. Countries take an average of about three weeks to moderate, four weeks to control, and over 6 weeks to contain the spread of the disease, after aggressive intervention. Substantial differences exist between large and small and Asian and European countries in time to act. Using these findings, we predict the likely dates of moderation and control for specific countries and States of the US.
In the absence of a vaccine, cure, or massive testing and quarantine, lockdowns and stay-at-home orders will need to last for months. However, the US faces a unique challenge because only half the states have adopted aggressive intervention, and done so at varying times. Even if these states achieve control or containment, they may be vulnerable to contagion from other states that were late to do so.
Keywords: Coronavirus, COVID-19, restrictions, lockdown, metrics, prediction, model, social distancing
JEL Classification: I1, I18
Suggested Citation: Suggested Citation