On the Herd Immunity Strategy
Before we have a vaccine for COVID-19, there are three alternatives to lockdowns:
Massive testing (where tracing can substitute to some extent for number of tests)—see for example “Seconding Paul Romer's Proposal of Universal, Frequent Testing as a Way Out”
Treatment improvements—for example, it is possible the monoclonal antibodies might work really well
Herd immunity of key subgroups of the population—see for example “How Does This Pandemic End?”
I have been frustrated by the relative dearth of forthright discussions of a strategy of going for herd immunity of key subgroups of the population. This relative dearth of forthright discussion is unfortunate, because I think that a herd immunity strategy is what most people who favor relative quick opening up of the economy have in mind, even if the focus of their rhetoric is simply on the high economic cost of lockdowns. One reason it is so important to think through various strategies is that, as I noted in “Two Dimensions of Pandemic-Control Externalities,” pandemic mitigation is not a concave problem: there is every reason to think that there are multiple local optima. If our metaphor is trying to keep damage low, we want a strategy that puts us in the lowest point; but even if we are at the bottom of one valley, over the neighboring mountains might be another valley with an even lower point.
If one is following a strategy that gets to herd immunity that allows the combination of declining COVID-19 prevalence and a return to some semblance of normality even before a vaccine, there is a very important strategic consideration: assuming one is keeping the timing and rate such that bad outcomes for each group given infection are about as low as they can reasonably be, however many infections there ultimately will be in the strategy, it is better to have those infections happen as early as possible, so that gradual improvements in herd immunity make it possible to begin opening up as soon as possible.
There are several things to unpack in what I just said. First, I think the lockdowns we have done so far can be justified by the need we had for greater scientific knowledge about the virus: both what affects its spread, who is at greatest danger, and how to treat it. I think we will have fewer deaths and other bad outcomes per infection as a result of that delay in some of the infections. Second, I am using “herd immunity” loosely to refer to what fraction of the population is immune. Using this loose definition, “herd immunity” is not an either/or thing. If more people in each subgroup are immune, COVID-19 can be kept in check with fewer restrictions on economic activity.
Along the lines of the scenario I discussed in “How Does This Pandemic End?” and in the strategy that Daron Acemoglu, Victor Chrnozhukov, Ivan Werning and Michael Winston study in their recent NBER Working Paper “A Multi-Risk SIR Model with Optimally Targeted Lockdown,” we might be muddling into a strategy of having those with high risk of a bad outcome given infection (especially the old) continuing to do strenuous social distancing, while those with a lower risk of a bad outcome given infection (especially the young) resuming social interactions that will let the disease spread fairly fast among them. If indeed, this is what we are going to do, it is better to do it now than later. It is only if we are going to try to avoid having the young avoid infection entirely and not get immune until the vaccine that we would want to force a strict lockdown on them. In other words, there is a discontinuity between the two strategies.
To get the benefit of strategy of trying to keep the number of infections low even for the young, one has to make it all the way to the finish line keeping those infections low. It is not enough even for that to be the right strategy. One would have to be quite confident one could politically pull it off. But trying to have few young people ever get infected might require such draconian and long-lasting lockdowns that they might not be politically feasible, even if they were, politically unconstrained, a good idea. Jesus made a good point about the low value of an unfinished tower (which can be a good metaphor for being stuck somewhere in the badlands between two distinct local optima):
Suppose one of you wants to build a tower. Won't you first sit down and estimate the cost to see if you have enough money to complete it? (Luke 14:28)
Here, it may not be literally “the cost” that needs to be estimated, but rather the political feasibility of a lockdown-heavy strategy. You might have to do a lockdown-heavy strategy all the way to have it be a good strategy. Are you really able to pull that off?
Some Wall Street Journal Perspectives on the Herd Immunity Strategy
I have seen some relatively forthright discussion of a herd immunity strategy in the news—and of course what is written about the pandemic is so vast that the absolute amount must be high, even if the percentage seems low.
In the May 14, 2020 op-ed “Scenes from the Class Struggle in Lockdown,” Peggy Noonan shows an awareness of the argument that, however many infections we are going to have in each subgroup, (once we get to more or less constant treatment effectiveness) better to have them come early and reopen the economy quickly than have them come late:
It’s not that those in red states don’t think there’s a pandemic. They’ve heard all about it! They realize it will continue, they know they may get sick themselves. But they also figure this way: Hundreds of thousands could die and the American economy taken down, which would mean millions of other casualties, economic ones. Or, hundreds of thousands could die and the American economy is damaged but still stands, in which case there will be fewer economic casualties—fewer bankruptcies and foreclosures, fewer unemployed and ruined.
They’ll take the latter. It’s a loss either way but one loss is worse than the other. They know the politicians and scientists can’t really weigh all this on a scale with any precision because life is a messy thing that doesn’t want to be quantified.
Aaron Ginn, who was interviewed by Allysia Finley for “The Lockdown Skeptic They Couldn’t Silence” (which appeared May 15, 2020), raises many issues relevant for a herd-immunity strategy. I’ll add bold italics to label different issues. Anything indented and set off from now on in this post is from this interview with Aaron Ginn.
One thing he says that I would counter is his suggestion that one meter (3.25 feet) might be enough social distancing. This makes it sounds as if distance is key. But duration is probably every bit as important as distance. Except in a retail context, indoor interactions tend to be of quite long duration and so are likely to pose quite a high risk of transmission. By contrast, most outdoor interactions by those who didn’t arrive together tend to be brief. (On indoor vs. outdoor, see this tweet.)
School Reopening:
One of his priorities is reopening schools. “When it comes to children, the data coming out of Europe is very, very strong,” he says. “You have, I would say, near-unanimous consensus among European scientists, public-health officials—including in Australia, South Korea and Japan—that children, for some reason, while they do get infected, they are not very infectious.”
A recent study from Australia identified only 18 cases (nine children and nine staff) across 15 schools, and only two of the infected children’s 863 close contacts at the schools became ill. Another review last month, published by the Royal College of Paediatricians and Child Health, couldn’t find an instance of a child passing on the virus to adults and noted that the evidence “consistently demonstrates reduced infection and infectivity of children in the transmission chain.”
Sweden Seems to Be Rapidly Moving Towards a Type of Herd Immunity that Allows “Segment and Shield” without a Disastrous Path There:
Mr. Ginn has been closely following Sweden, which has kept children under 16 in school and let most businesses stay open while restricting gatherings of more than 50 people. His daily briefings frequently cite Sweden’s state epidemiologist, Anders Tegnell, who has argued that government lockdowns lack a “scientific basis” and “people should be able to keep a reasonably normal life.” Dr. Tegnell recently estimated that 40% of Stockholm’s population would be immune to the virus by the end of May.
That could bring Sweden closer to “herd immunity” than countries that have sought to suppress spread altogether. “We need to ‘segment and shield,’ ” Mr. Ginn says, “and let the epidemic go through”: “The question is: How are you going to best protect those that are vulnerable in the larger population?”
…
A paper last week by Stockholm University mathematicians estimates herd immunity could be around 43% if young, socially active people mix more and gain immunity, protecting older, less socially active people. In other words, Stockholm may have already achieved herd immunity. Dr. Tegnell said this week that the declining number of cases in Stockholm supports this possibility.
Heterogeneity in Social Interactivity and Physical Reaction to the Virus Means We Might Only Need Herd Immunity for the Super-Spreader Subgroup to Be in a Much Better Position:
Some scientists say herd immunity would require 60% to 70% of the population to be infected, which would entail massive deaths. Mr. Ginn says those numbers are up for debate. A recent study from a large team of international researchers including some at Oxford and the National Institutes of Health (which hasn’t undergone peer review) estimates that “variation in susceptibility or exposure to infection can reduce these estimates” so that some populations may achieve herd immunity with an infection rate of only 10% to 20%.
All of these claims need to be questioned, but all claims anywhere in the ballpark of these claims need to be seriously considered. It could make a big difference to the optimal strategy.
Don’t miss these other posts on the coronavirus pandemic: