Price of Delay in COVID-19 Lockdowns: Delays Spike Total Cases, Natural Experiments Reveal

24 Pages Posted: 5 May 2020 Last revised: 16 Jun 2020

See all articles by Gerard J. Tellis

Gerard J. Tellis

University of Southern California - Marshall School of Business, Department of Marketing

Nitish Sood

Augusta University

Ashish Sood

University of California Riverside

Date Written: May 4, 2020


Massive protests against COVID-19 lockdowns across many states of the US raise the question: Were lockdowns effective? Were they even necessary? The authors aim to answer these questions. Since the US did not have a mandatory lockdown across the nation, governors of US States ordered lockdowns on different dates. The authors identify natural experiments consisting thirteen cohorts, each of two similar neighboring states with different lockdown dates. The authors estimate the difference in COVID-19 penetration within these cohorts, adjusting for the differences in penetration on the date of the first lockdown (Difference in Difference). Here are the major results:

• In eleven of the thirteen cohorts, the state that lockdown later or did not lockdown suffered a substantially higher penetration in Covid-19 as of April 30th 2020. In only two cohorts, the penetration of the disease was higher in the state that did lock down.

• For the states that lockdown later or never locked down, the penetration of the disease shows an average increase of 53% over the similar neighboring state that did lock down.

• Formal statistical analysis using difference in difference regression shows a significant effect of lockdown equal to about 232 less cases per million population of a state in the US, for the observation period March 10th to April 30th. This amounts to a decrease of 50%.

• The strength of these results resides not only in the care of identifying the natural experiments but also in the fact that results from eleven of the thirteen cohorts are consistently in the same direction and mean estimates of reduced cases from lockdown are similar from model free results (52%) and modeling results (50%).

In addition, the authors identify one natural experiment between one US Gulf state that had a huge social mixer (Mardi Gras) and five other Gulf States without such an event. The state with the social mixer shows a big increase of over 4,000 cases per million of population, amounting to a huge 296% (4 times) increase in the penetration of the disease, relative to the other five states.

Thus, population in US states paid a high price in terms of higher infections due to a failure to lock down a state or ban big social mixers. These findings suggest that lockdowns, in addition to better hygiene and voluntary social distancing, seem to be effective in controlling the spread of the disease. Thus, policy regarding response to pandemics should be informed by federal agencies and scientists with the resources to forecast pandemics and ascertain the need for defensive measures, rather than be left solely to state governors or local officials who lack such resources. Governors are now reopening their states. Learning from the price of delay in locking down the state would help them assess the true value of lockdowns and make an informed decision about unlocking the state.

Keywords: Natural Experiments, COVID-19, Disease Spread, Penetration, Lockdown, COVID-19, US Governors, Social Distancing, Healthcare Policy, Infectious Diseases

JEL Classification: I1, I18, M30, O33, D80, P16, Z1

Suggested Citation

Tellis, Gerard J. and Sood, Nitish and Sood, Ashish, Price of Delay in COVID-19 Lockdowns: Delays Spike Total Cases, Natural Experiments Reveal (May 4, 2020). USC Marshall School of Business Research Paper, Available at SSRN: or

Gerard J. Tellis

University of Southern California - Marshall School of Business, Department of Marketing ( email )

Hoffman Hall 701
Los Angeles, CA 90089-0443
United States
213-740-5031 (Phone)
213-740-7828 (Fax)


Nitish Sood

Augusta University ( email )

1120 15th Street
Augusta, GA 30912
United States


Ashish Sood (Contact Author)

University of California Riverside ( email )

United States
6782059931 (Phone)


Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics