Philippe Lemoine is a PhD candidate in philosophy at Cornell University, with a background in computer science. He’s also a blogger, a research fellow at the Centre for the Study of Partisanship and Ideology, and a lockdown sceptic. During the pandemic, he’s written several detailed articles about the efficacy of lockdowns. I interviewed him via email.
On December 4th, you published an article on your blog titled ‘Lockdowns, science and voodoo magic’, which criticised the well-known paper by Flaxman et al. That paper (which has been cited more than 1,300 times) concluded, “major non-pharmaceutical interventions – and lockdowns in particular – have had a large effect on reducing transmission”. Could you briefly summarise your criticisms?
I made two main points against that paper. First, the model assumed that only non-pharmaceutical interventions affected transmission, so any observed reduction in transmission could only be ascribed by the model to non-pharmaceutical interventions. Since in fact transmission went down quickly everywhere during the first wave, the only question was how much of that reduction would the model attribute to each intervention. But the fact that non-pharmaceutical interventions were jointly responsible for the entire reduction in transmission was not something the model inferred from the data, it was assumed at the outset by the authors when they defined the model. A consequence of this fact is that, when they compute a counterfactual scenario in which there weren’t any non-pharmaceutical interventions to estimate how many lives were saved by lockdowns and other restrictions, the authors just assume that cases would have continued to rise until the herd immunity threshold was reached and would only start to go down then. Although the authors did not deem it necessary to reveal this small detail, this meant that, in their counterfactual, more than 95% of the population was already infected by May 3, which is preposterous. Even one year and a half after the beginning of the pandemic, there isn’t a single country where the proportion of the population that has been infected even comes close to such a figure, not even in countries where restrictions were extremely limited. So when the paper finds that non-pharmaceutical interventions in general and lockdowns in particular saved three million lives in Europe alone during the first wave, they only reach that conclusion by comparing the actual number of COVID-19 deaths to the number of deaths in a ridiculous scenario where essentially everyone had been infected. Yet this preposterous estimate was taken seriously by the entire scientific establishment and, as you noted, the paper became one of the most cited studies on the COVID-19 pandemic.
The second point I made is that, not only was this result based on totally unrealistic assumptions, but the authors failed to disclose a key result that completely undermined their conclusion. As I explained above, the model was bound to attribute the entire reduction in transmission that was observed in Europe during the first wave to non-pharmaceutical interventions, the only question was how much of it would be attributed to each intervention. Their headline result was that, apart from lockdowns, nothing else had any clear effect, which meant that lockdowns were responsible for the overwhelming majority of the 3 million lives that, according to this study, non-pharmaceutical interventions had collectively saved. However, Sweden was included in the study and never locked down, yet only a tiny fraction of its population was infected during the first wave. How is that possible if only lockdowns have a substantial effect on transmission? I knew this made no sense, so I downloaded the code of the paper to reproduce their analysis on my computer and take a closer look at the results. Their model allowed the effect of the last intervention, which happened to be a lockdown everywhere except in Sweden, where it was a ban on public events, in each country to vary. What my analysis of their results showed is that, in order to fit the data, the model had to find that banning public events reduced transmission by ~72.2% in Sweden but only by ~1.6% elsewhere. In other words, according to the model, banning public events had somehow been 45 times more effective in Sweden than anywhere else. Now, unless you believe there are magical anti-pandemic fairies in Sweden that somehow made banning public events 45 times more effective than elsewhere, this obviously never happened. Rather, what this means is that the model was garbage, which in turn means that we have no reason to believe the paper’s headline result that lockdown had a huge effect on transmission. There is a lot more in my piece about that paper, which I methodically demolish, but those are the main points.
Then on March 4th, you published a report for the Centre for the Study of Partisanship and Ideology titled “The Case against Lockdowns”. This was followed by an op-ed in the Wall Street Journal titled “The Lockdowns Weren’t Worth It”. Could you briefly summarise the case against lockdowns, as you see it?
First, I think it’s impossible to estimate precisely the effects of non-pharmaceutical interventions because too many factors contribute to transmission, and we lack the kind of background knowledge we’d need to be confident that the statistical techniques people use to estimate those effects are reliable, so people who claim to be able to do that are full of it. I just published another piece in which I take a very close look at a study which found that non-pharmaceutical interventions had a substantial effect on the number of cases and deaths in the US during the first wave. This study is far more sophisticated than Flaxman et al.’s paper and, in particular, the authors did not assume that only non-pharmaceutical interventions affect transmission, and tried to model the effect of voluntary behavioural changes. Nevertheless, as I show in my article, when you look at it closely and perform various sensitivity analyses, the conclusions no longer hold. So we have no way to estimate precisely the effects of non-pharmaceutical interventions and we should be honest about this. However, whatever their precise effects, they can’t be huge because otherwise they would be much easier to detect. The contrast with the effect of vaccination is particularly striking in that respect. In the case of vaccination, the effect is so obvious that you can see it on a simple chart, whereas in the case of non-pharmaceutical interventions you have to squint and use very complicated statistical techniques that, although they impress people because they look scientific, we have no reason to think are reliable in this context. Now, if you do a cost-benefit analysis, even if the only costs of lockdowns you take into account is the immediate effect they have on people’s well-being and you make ridiculously optimistic assumptions about how much stringent restrictions reduce transmission, they don’t pass a cost-benefit test. In fact, not only do they fail to pass a cost-benefit test, but it’s not even close. The costs of lockdowns, by which I mean just their immediate effect on well-being, so far outweigh their benefits that one cannot reasonably doubt a more rigorous cost-benefit analysis would reach a different conclusion.
According to some people, claiming that lockdowns don’t have a large effect on the spread of COVID-19 is tantamount to “denying germ theory”. What do you say to those people?
Nobody is denying that transmission occurs during physical interactions, but it doesn’t follow that lockdowns have a large effect on transmission, so people who make this argument simply haven’t thought things through. In theory, lockdowns could even increase transmission, so this argument is very confused. For instance, it could be that, although lockdowns decrease between-household contacts, the effect on transmission at the aggregate level is more than compensated by the increase in within-household contacts they produce. To be clear, I don’t believe this is the case, I’m just saying it’s a theoretical possibility that obtains in some models, even though nobody denies the germ theory of diseases. There are many possible explanations for why lockdowns don’t result in the very large reduction in transmission that one might have expected. For instance, we don’t expect lockdowns to be equally effective at reducing all types of contacts and, as I just noted, they even increase the frequency of some types of contacts, such as within-household contacts. So it could be that the types of contacts that lockdowns manage to reduce a lot don’t contribute a lot to transmission, while the types of contacts they aren’t very useful for preventing contribute a lot to it. Another important point is that, even in the absence of a lockdown, people change their behaviour in response to the pandemic. So it could be that the types of contacts that contribute the most to transmission are the same types of contacts that people tend to reduce voluntarily even in the absence of a lockdown. Anyway, whatever the explanation, it’s pretty clear that lockdowns don’t have a very large effect. It would be very surprising if such a fact were inconsistent with the germ theory of diseases, but fortunately it isn’t. It’s just that people who make this argument are confused. The effectiveness of lockdowns and restrictions in general is an empirical question that cannot be solved by theorizing from the armchair.
Much of your writing about lockdowns has dealt with the deficiencies of epidemiological models. Why have most models done so poorly at predicting the epidemic’s trajectory?
This is a difficult question and I’m not sure what the answer is. I’m very confident that part of the story is that most of those models don’t take into account the kind of voluntary changes of behaviour I was just talking about. If your model is based on the assumption that people’s behaviour only changes in response to government interventions, it should be no surprise that it performs terribly. But I don’t think it’s the whole story and I increasingly suspect that the fact the models don’t adequately model population structure is another factor. Most epidemiological models that have been used to make projections assume that, within large age groups, people mix homogeneously. But this is totally unrealistic since, for instance, a 55-year-old is not equally likely to run into any other person in the 50 to 59 age group. Rather, a particular 55-year-old is very likely to have contacts with some people in that age group (such as friends and family), but very unlikely to meet many other people in that age group and has essentially no chance of running into the vast majority of people in that age group. Anyway, nobody really knows why those models perform so terribly at larger scales, but in order to investigate the problem epidemiologists would first have to acknowledge it. Unfortunately, they mostly ignore it and act as if their models had not proven incapable of explaining the data, except in the sense that you can always “explain” any data if you are willing to make enough purely ad hoc hypotheses, so they don’t even get started.
As far as I’m aware, no Western government has published a cost-benefit analysis of lockdowns. Why were these far-reaching policies implemented with so little regard for costs?
As I noted above, any serious cost-benefit analysis would immediately show that lockdowns are not worth it. Yet as you say no Western government has published any to justify their policy. This is particularly surprising when you know that, in most Western governments, the use of cost-benefit analyses is largely institutionalized and the authorities are often required to make one before they can embark on projects as banal as building a bridge. Yet they apparently didn’t feel the need to publish a cost-benefit analysis to justify what are effectively the largest attacks on individual freedoms in the West since the end of World War 2. One interpretation is they realise that, as I noted above, no cost-benefit analysis would ever vindicate lockdowns. But this wouldn’t explain why they are pursuing lockdowns and I don’t believe in that explanation for a second anyway. In a way, if that were really the explanation, I would almost find that reassuring because it would at least imply a level of competence and understanding which I think is entirely lacking from our political leaders. Rather, I think their decisions are the result of a combination of cluelessness not just on their part but also on the part of their advisors and a variety of bad incentives that conspire to create absurd policies, such as the desire not to leave themselves open to the accusation of not having done anything to curb the epidemic. This desire must be strong as they are constantly under pressure from the largely pro-lockdown media to enact more restrictions. In order to answer this unremitting call to “do something”, they do something, even if that’s completely absurd, as long as they have something to show to the people who constantly ask them to “do something”. The idea of measuring the costs of their decisions against their supposed benefits often doesn’t even enter their heads because their decision-process is not governed by rational considerations, but rather by this ungodly combination of emotion, illusion of control, bad incentives and even worse advice.
You’re a Frenchman. Given what we know now, what should Emmanuel Macron have done in March of 2020?
With the benefit of hindsight, I think he should have just told people to try to limit their contacts to reduce the amount of stress on hospitals, but leave them free to make their own choices and focus his efforts on preparing government services to respond as best as possible. I think there are lots of reasons to blame Macron and French officials for their conduct at the beginning of the pandemic, especially for their lack of preparation and their carelessness in the weeks leading up to the explosion of cases in the country, but if we put aside the lies they told repeatedly during that period and since then, they at least had the excuse that we didn’t know much about the virus and how different policies would affect spread. I was in favour of the first lockdown and, while I now think that I was wrong and that I should have predicted lockdowns would become entrenched after we had used them once, it was a genuinely difficult decision because we didn’t know much. But after the first wave there was no longer any excuse and Macron should be judged harshly for keeping us more or less locked down for months after the first wave, even though it was already very clear by that point that restrictions did not make a very large difference to transmission, yet had a very negative impact on the population’s well-being.