Do Legal Restrictions Account for the Downward Trend in New COVID-19 Cases?

The New York Times thinks so, but the data do not fit that hypothesis very well.


Three weeks after newly identified COVID-19 cases began falling in the United States, The New York Times is acknowledging the downward trend, which it attributes to "effective restrictions." That explanation fits neatly with the paper's reflexive enthusiasm for lockdowns, but it does not fit the data very well.

Consider Arizona, where the seven-day average of daily new cases, according to Worldometer's tallies, rose more than tenfold between Memorial Day and July 7. Alarmed by that increase, Gov. Doug Ducey ordered gyms, bars, movie theaters, and water parks to close on July 23, while indoor dining in restaurants continued at 50 percent of capacity, a cap Ducey imposed on July 11. But the downward trend in new cases, which had fallen by 82 percent from the July 7 peak as of yesterday, began well before the new restrictions could have had a measurable impact (taking into account the typical five-day lag between infection and symptoms that might cause people to seek testing). That suggests other factors are at least partly responsible for the decline.

Newly confirmed cases are also falling in Georgia, which did not respond to this summer's spike with new business restrictions. The seven-day average, which rose fivefold between May 25 and July 29, has fallen by 35 percent since then.

In Texas, the seven-day average of newly identified cases rose tenfold between May 25 and July 15. It has since fallen by nearly half. Gov. Greg Abbott closed bars and reduced the cap on indoor dining from 75 percent to 50 percent of capacity on June 26. Yet cases continued rising for nearly three weeks, longer than the maximum incubation period. The story is similar in Florida, where Gov. Ron DeSantis closed bars the same day as Abbott. The seven-day average of daily new cases peaked three weeks later, when it was nearly 16 times higher than it was on May 25, then fell by two-thirds as of yesterday.

California Gov. Gavin Newsom closed bars, zoos, and museums on July 13, when he also banned indoor dining in restaurants. The seven-day average of daily new cases peaked 12 days later, when it was nearly five times the number on May 25, then fell for two weeks before rising again. The decline resumed in mid-August, and as of yesterday the average was down 42 percent from the July 25 peak. That pattern likewise does not easily fit the hypothesis that new legal restrictions account for most or all of the decline.

California's restrictions, including a total ban on indoor dining, are more severe than the statewide limits in Arizona, Florida, or Texas, which enforce a 50 percent cap. Yet the latter three states have seen bigger declines in confirmed cases, although they also saw bigger increases in June and July.

It is plausible that limiting the options for people to get together, especially indoors in close quarters, would reduce virus transmission. But it is difficult to disentangle the impact of government edicts from the impact of increased voluntary precautions, both of which can be expected in response to spikes in cases. Just as it is hard to assess the additional contribution of general lockdowns at a time when Americans were already responding to the COVID-19 epidemic by dramatically changing their behavior, it is hard to say how much credit relatively modest legal restrictions should get for reversing the recent surge in cases.

Given the timing of the downward trends in these Sunbelt states, attributing them mostly or entirely to new legal restrictions seems more than a little hasty. The Times alludes to voluntary changes in behavior just once in its story, citing a University of Arizona professor of public health policy who mentioned "news media coverage of the heightened risk" as one factor that helps explain the decline in cases. The implication, which the Times predictably fails to unpack, is that people responded to that "heightened risk" by being more careful—e.g., by limiting social interactions, avoiding crowds, following physical distancing guidelines, and wearing face masks in public places.

What about face mask mandates? Even if businesses were already requiring customers to wear masks, a legal requirement could help reduce disputes about those policies and increase compliance. Yet Newsom mandated masks in California on June 18, more than a month before new cases peaked in that state, while Abbott followed suit in Texas on July 2, nearly two weeks before that state's peak. Neither Arizona, Florida, nor Georgia has a statewide mask requirement, although some local governments in those states have imposed their own mandates.

"Of the states that are driving the decrease" in new cases, the Times says, "all have at least some local mask mandates, and most have paused or reversed statewide reopening policies." That gloss, which dismisses the absence of statewide mask mandates, elides the difference between imposing restrictions and delaying their removal, and ignores the states without new restrictions, makes the Times seem desperate to credit government policies for positive trends that ultimately depend on how people decide to behave.

Granting that legal restrictions have some impact on virus transmission, the extent of those restrictions is still a matter of controversy, given their substantial costs and uncertain benefits. The Times describes a "flattening" of the COVID-19 curve in states where restaurants are still serving customers indoors, for example. Does that mean Newsom went too far by banning indoor dining? Is there an important distinction between bars and restaurants, as DeSantis, Abbott, and Ducey seem to think, or is that premise dangerously mistaken? One thing seems clear from recent COVID-19 trends: The curve can be flattened without general lockdowns, without statewide mask mandates, and even (as in Georgia) without new, post-lockdown restrictions on businesses.

In all of these states, the decline in new cases has been followed by a decline in daily COVID-19 fatalities since earlier this month. Nationwide, according to Worldometer's numbers, the seven-day average of daily new cases, which exceeded 69,000 on July 25, had fallen to about 43,000 as of yesterday. The seven-day average of daily deaths fell from nearly 1,200 on August 4 to fewer than 1,000 yesterday. Independent data scientist Youyang Gu, who has a good track record of predicting COVID-19 fatalities, is currently projecting a nationwide death toll of about 227,000 by November 1, compared to about 176,000 now.

Is that number lower than it would have been without last spring's lockdowns, and will the ultimate death toll likewise be lower than it otherwise would have been? On that question, the Times seems to be hedging.

In late June, the paper noted that "the shared sacrifice of millions of Americans suspending their lives—with jobs lost, businesses shuttered, daily routines upended—has not been enough to beat back a virus whose staying power around the world is only still being grasped." The implication of that story was that lockdowns were lifted too soon and too quickly. But since the sweeping social and economic restrictions that most states imposed could not feasibly have been sustained until effective vaccines were deployed, it was inevitable that we would eventually face the situation we now confront.

The virus is still here, as everyone knew it would be. And to the extent that lockdowns were effective in reducing transmission, they left a population more vulnerable to infection than it otherwise would have been. In this context, it makes sense to concentrate on protecting Americans in high-risk groups rather than younger, healthier people who face a negligible risk of dying from COVID-19.

Recent data suggest we have had some success on that score. Even as cases spiked this summer, the nationwide crude case fatality rate—deaths as a share of confirmed cases—continued to decline, dropping by nearly 50 percent since mid-May. That trend largely reflects a younger, healthier mix of patients. The key to minimizing the death toll while we wait for a vaccine is preventing new cases among older, frailer people. The debate should focus on the best ways to do that rather than recriminations about what went wrong with lockdowns. And in weighing our options, it would be foolhardy to assume that government policy is all that matters while overlooking the crucial role of the decisions that each of us makes every day.