Virus Transmission Estimates Provide More Evidence That COVID-19 Lockdowns Are Overrated

Two models suggest that broad restrictions had less impact on the epidemic than commonly thought.

States began lifting their COVID-19 lockdowns more than a month ago, and so far the results have not been nearly as disastrous as many people predicted. By and large, according to estimates by two teams of researchers, virus transmission in Florida, Georgia, and Texas—three states with large populations that loosened their restrictions at the end of April—has either declined, stayed about the same, or risen slightly since then. Those numbers reinforce other evidence that broad business closure and stay-at-home orders deserve less credit for curtailing the epidemic than they commonly receive.

The two models estimate the reproductive number for the COVID-19 virus: the average number of people a carrier infects. A reproductive number higher than one indicates an ongoing epidemic. When the number falls below one, the epidemic is waning. The daily number of new cases can be expected to decline, and eventually so will the total number of active cases as previously infected people recover.

After Florida's lockdown was lifted, according to a model by researchers at the University of Utah, the reproductive number went up and down before rising to 1.29 as of May 29, compared to 0.98 on April 30. But in Texas, according to this model, the reproductive number fell from 1.13 on April 30 to 0.45 on May 28. In Georgia, the number barely changed during this period, remaining slightly below one.

A model by independent data scientist Youyang Gu is generally more pessimistic but probably more accurate, judging from its projections of COVID-19 deaths. The Gu model shows Florida's reproductive number rising from 0.97 on April 30 to 1.07 today. The picture is similar in Texas and Georgia, where the reproductive number rises from slightly below to slightly above one during the same period.

Concern is justified whenever the reproductive number rises above one. But with the exception of the University of Utah estimate for Florida, these numbers do not indicate that lifting the lockdowns had a big impact on virus transmission.

While the two models diverge in their estimates since April 30, they tell basically the same story about trends before these states closed "nonessential" businesses and told people to stay home except for officially approved purposes. In all three states, the models indicate, the reproductive number fell precipitously in March. Those downward trends began before the lockdown orders were issued.

Judging from the estimates in both models, statewide lockdowns had little or no perceptible impact on virus transmission in these states. And Oklahoma, which never imposed a lockdown, saw essentially the same drop in transmission around the same time. These estimates are consistent with cellphone and foot traffic data, which show that Americans were moving around less in response to the COVID-19 epidemic well before politicians made it mandatory.

None of this means that lockdowns were completely ineffectual. It is plausible that they had an impact on people who were not already taking precautions, which is consistent with the Gu model's post-lockdown estimates and the University of Utah's recent estimates for Florida. But voluntary changes in behavior are clearly more important than lockdown supporters typically imagine.

University of Hong Kong epidemiologists Dillon Adam and Benjamin Cowling, writing in The New York Times, note that a virus' reproductive number, which represents an average across all carriers, can be misleading, because "it doesn't convey the vast range between how much some infected people transmit the virus and how little others do." Their research in Hong Kong found that "just 20 percent of cases, all of them involving social gatherings, accounted for an astonishing 80 percent of transmissions." Another 10 percent of carriers "accounted for the remaining 20 percent of transmissions," meaning that 70 percent of people infected by the virus did not pass it on to anyone.

Other studies cited by Adam and Cowling confirm the outsized role played by "superspreaders," which is relevant in evaluating the cost-effectiveness of broad control measures. They recommend that policy makers focus on "stopping the superspreading" through social distancing guidelines and restrictions on large, crowded gatherings, especially indoors, rather than trying to regulate everyone's movements. They argue that the experience with COVID-19 in Hong Kong and Japan, neither of which imposed general lockdowns, suggests "the epidemic's growth can be controlled with tactics far less disruptive, socially and economically, than the extended lockdowns or other extreme forms of social distancing that much of the world has experienced over the past few months."