The Science of Immaculate Infection

Using Cherry-Picking and Data Manipulation to "Prove" That Lockdowns Don't Work

Yesterday I posted a piece looking in depth at the claim made by economist and Cato Institute fellow Steve Hanke and his colleagues, and boosted by populist YouTube commentator Jimmy Dore, that lockdowns didn’t work in reducing deaths from COVID. I got an error message indicating that my post was too long to be emailed to subscribers in full, so I wound up not sending out the email. So if you’re among my lengthy list of 20 subscribers, consider this your notification of the existence of that piece, titled Jimmy Dore Gaslights His Viewers about Lockdowns. Here I want to briefly summarize what I said there, and highlight my main point, which is: Are these guys nuts?

I’m no more an infectious disease specialist than Dore or Hanke are, but I’m pretty sure of one basic fact about communicable diseases: You can’t get one if you’re not near someone who has it. And the whole point of “lockdowns”—an umbrella term referring to measures that restrict individuals’ contact with each other such as closures of non-essential businesses and in-person schooling, limiting public indoor gatherings, limiting travel, etc.—is to reduce the amount of contact individuals have with each other (particularly indoors, where viruses spread more easily), and thereby reduce the possibility of their infecting each other. Perhaps, having grown up Catholic, Dore believes in immaculate infection?

Hanke, for his part, appears to have gotten cherry-picking data and massaging it until he gets the desired result down to a science (which is a far cry from being a scientist). He and his colleagues, fellow economists Jonas Herby and Lars Jonung, conducted a supposed “systematic review and meta-analysis” (meta-analysis is a statistical technique allowing the results of several studies to be mathematically combined, creating more power to find effects) in which, to make a long story short (see my previous post for the details), they cherry-picked some of the many studies of lockdowns that showed the smallest effects, then cherry-picked the data from one of them that showed the smallest effect, and then gave that one study a huge amount of weight in their “analysis,” thereby creating the false impression that they had “found” that lockdowns only reduced deaths by 0.2%. As one of the authors of that study, Carolyn Chisadza, put it:

They already had their hypothesis. They think that lockdown had no effect on mortality, and that’s what they set out to show in their paper.

In contrast, there are dozens of studies that show that—no shit, Sherlock—various government measures that serve to reduce physical proximity between strangers dramatically reduced incidence of, and mortality from, COVID-19, such as this one, and this one, and this one, and this one. A study of New Zealand’s lockdown during the first wave found that not only COVID deaths, but deaths from ALL causes, were reduced. But rather than dive into details of the studies, which I’ve linked for your convenience if you want to check them out, I’d like to look at the issue on more of a common-sense level, accompanied by some illustrative data.

According to Neil Ferguson and his colleagues, there are two general approaches countries could have taken to controlling COVID outbreaks: 1) mitigation, which is the sort of approach the UK and US took, and 2) suppression (or “zero COVID” as it’s come to be called, although it doesn’t literally mean a country can zero out COVID and keep it there when it’s still rampant elsewhere), which is the approach that China and until recently New Zealand took. Doing nothing to control COVID spread, as “immaculate infection” believers apparently would have preferred, isn’t an option that any countries pursued, which according to Ferguson et al.’s modeling, is quite fortunate: They predicted that the US would have had 2.2 million deaths and the UK 500,000 within a few months had they taken no action, whereas actually the US had about 190,000 deaths and the UK 40,000 between March and August of 2020.

While it seems intuitively obvious that doing something is much better than doing nothing, it is more compelling to compare countries’ outcomes based on what they actually did. Countries that did better implemented lockdowns sooner and more stringently, and brought cases down close to zero before lifting them (which in practice wasn’t much longer than other countries, due to their quicker action and greater stringency).

The UK (population 67 million) implemented a nationwide lockdown on March 23, at which time it had 12,707 cumulative confirmed COVID cases. The US (population 334 million), which never had a nationwide lockdown, on average implemented some form of lockdown on a state by state basis by March 28, at which point the US (which had a massive shortage of tests) had 127,417 cumulative confirmed cases. The earliest state to lock down was California on March 19, when there were 13,663 confirmed cases nationally.

In contrast, China (population 1.4 billion) locked down Hubei Province (where Wuhan is located), the epicenter of its outbreak, on January 23, at which point China had 639 confirmed COVID cases. New Zealand (population 5 million) implemented a nationwide lockdown on March 25, at which point it had 209 confirmed cases.

China’s and New Zealand’s lockdowns were also much stricter than the US’s or the UK’s. China sealed off all travel to and from Hubei, and international visitors were banned from entering the country entirely, with a 3-week quarantine period later adopted. New Zealand likewise banned all international travel into the country during its initial lockdown. The UK and US did not. The US did not even impose any restrictions on travel within the US. China and New Zealand both implemented extensive testing and contact tracing strategies and tracked every case, requiring anyone who tested positive to isolate immediately until they were no longer infectious; the US and UK were unable to implement such strategies because there were too many cases to test everyone by the time they implemented lockdowns. And whereas US states lifted their lockdowns while there were still tens of thousands of new cases per day, and the UK lifted its lockdown while there were still a couple of thousand new cases per day, China and New Zealand did not lift theirs until they had basically zeroed new cases out.

The proof that these measures worked is in the pudding. By the time the first wave of COVID deaths in the US and UK had ended (you might say China and New Zealand had “wavelets”), China had a total of 4600 deaths, a rate of only 3 deaths per million people, and New Zealand had 22 deaths, a rate of 4 deaths per million people. Compare these outcomes to those in the US and UK: 407 and 600 deaths per million, respectively—over 100 times higher.

One particularly annoying part of Dore’s June 23 interview with Hanke was when they ridiculed Ferguson et al.’s claim that lockdowns could reduce COVID deaths by 98% compared to what they would be if no attempts to mitigate the virus’ spread were implemented. Of course, compared to their predictions of 2.2 million deaths in the US and 500,000 in the UK in the absence of mitigation, 190,000 and 40,000 deaths were not lower by that amount. But China and New Zealand had death rates more than 99% lower than the US’s and UK’s, let alone in comparison to what would have happened with unchecked spread. (Additionally, the UK had a second lockdown from January to early May of 2021, reducing its daily death rate from 1200 to 6.)

Hanke and his colleagues excluded all data from outside the United States and Europe from their study. Dore, for his part, has never once mentioned anything about the outcomes of countries such as China or New Zealand that acted swiftly and decisively to contain their COVID outbreaks. In a nutshell, they are willfully blind.