openpolitics.com

The Contrarian Coronavirus Theory That Informed the Trump Administration

President Trump, who at one point called the coronavirus pandemic an “invisible enemy” and said it made him a “wartime President,” has in recent days questioned its seriousness, tweeting, “WE CANNOT LET THE CURE BE WORSE THAN THE PROBLEM ITSELF.” Trump has said repeatedly that he wants the country to reopen by Easter, April 12th, contradicting the advice of most health officials. According to the Washington Post, “Conservatives close to Trump and numerous administration officials have been circulating an article by Richard A. Epstein of the Hoover Institution, titled ‘Coronavirus Perspective,’ that plays down the extent of the spread and the threat.”

Epstein, a professor at New York University School of Law, published the article on the Web site of the Hoover Institution, on March 16th. In it, he questioned the World Health Organization’s decision to declare the coronavirus outbreak a pandemic, said that “public officials have gone overboard,” and suggested that about five hundred people would die from COVID-19 in the U.S. Epstein later updated his estimate to five thousand, saying that the previous number had been an error. So far, there have been more than two thousand coronavirus-related fatalities in America; epidemiologists’ projections of the total deaths range widely, depending on the success of social distancing and the availability of medical resources, but they tend to be much higher than Epstein’s. (On Sunday, Anthony Fauci, the director of the National Institute of Allergy and Infectious Diseases, estimated that there could be between a hundred thousand and two hundred thousand deaths in the U.S.) In a follow-up article, published on March 23rd and titled “Coronavirus Overreaction,” Epstein wrote, “Progressives think they can run everyone’s lives through central planning, but the state of the economy suggests otherwise. Looking at the costs, the public commands have led to a crash in the stock market, and may only save a small fraction of the lives that are at risk.”

Epstein has long been one of the most cited legal scholars in the country, and is known for his libertarian-minded reading of the Constitution, which envisions a restrained federal government that respects private property. He has also been known to engage with controversial subjects; last fall, he published an article on the Hoover Institution Web site that argued, “The professional skeptics are right: there is today no compelling evidence of an impending climate emergency.” Last Wednesday, I spoke by phone with Epstein about his views of the coronavirus pandemic. He was initially wary of talking, and asked to record his own version of the call, which I agreed to. During our conversation, which has been edited for length and clarity, Epstein made a number of comments about viruses that have been strongly disputed by medical professionals. We have included factual corrections alongside those statements.

What did you want to achieve with your pieces?

What did I want to achieve with my pieces? First of all, I am not a politician. What I did is that I looked at the standard model that was put out in the New York Times [in an Op-Ed by Nicholas Kristof and Stuart A. Thompson, published on March 13th], which was backed up by other models in other places, and it occurred to me that I just did not think that the underlying assumptions there were sound. The single most important thing to me was not to get my own estimate of what the number is. The most important thing was to look at that curve, which seemed to suggest that there would be ten million cases a day during a ten-day or so band in the middle of July, and to explain why, in relationship to all other things I know about evolutionary theory, that this just has to be wrong. The better way to have phrased the paper would have been to say that the traditional models, which were used for the last flu season, for the 2009 H1N1 situation, are much better approximations of what is likely to happen than these rather scary kinds of projections.

You wrote last week, “In the United States, if the total death toll increases at about the same rate, the current 67 deaths should translate into about 500 deaths at the end.” We are currently at eight hundred deaths—over eight hundred deaths. [This was true when we spoke; the number is now over two thousand.]

First of all, let me just say I wrote an amendment to that, the thing I regret most in that whole paper. But I was not so much interested in explaining why my number was right. I was interested in explaining why the other projections were wrong.

O.K., but your number was surpassed in about a week, and now we’re already—

The New Yorker’s coronavirus news coverage and analysis are free for all readers.

I understand that, but the point about that is that, first of all, there was a simple stupid error, which is you would never want to put it in a model that total deaths in the United States relative to the world would be one per cent. So if you just inflated it to five per cent or ten per cent, then all of a sudden you’ve got a number which is either five or ten times as high.

Secondly, suppose I should have been wiser in this and said, as I referred to the flu vaccine and later on to the H1N1 situation, if those are your benchmarks, then the number goes up to say between fifteen thousand and forty thousand deaths, as opposed to the one million-plus that are projected. [The Times model projected, without interventions by governments or citizens, a million deaths in the U.S.; with such interventions, the model showed that number dramatically decreasing.] And, remember, the one million-plus is on a model which is universal and worldwide, and you should expect to see something like that somewhere else. And there’s no evidence whatsoever that any of the situations, even in Italy, is going to approach the kinds of numbers that you had there. And so I am truly sorry about that [five hundred] number. I regard it as the single worst public-relations gaffe I’ve made in my entire life. But the question to ask, Isaac, is not whether I chose the right number but whether I had the right model.

Something else you wrote, in an earlier piece, was, “Why has there been such a dramatic mismatch in the responses to ordinary flu and the coronavirus?” Is that a question you’re still unsure about?

Look, the basic problem is, I think, in effect, that the tendency on the part of many people to treat this particular thing as unique is a mistake. There’s an underlying, standard model that you want to use, and the question is how you stuff it full of parameters. That is, numbers you add into it to make what’s going on. And, so, the situation that you get is you cannot use any exponential system because essentially then everybody is going to be dead, because things just keep doubling, doubling, and doubling.

So you have to develop a model which is going to explain why there’s a fairly rapid increase at the outset, and then why the thing starts to turn flat, ultimately down, and then disappears. That’s the strategy that you have to do. And so the theory here is one that I actually worked with and I thought worked pretty well in the AIDS stuff, which I worked on back in the early nineteen-eighties. And the model goes something like this. You start off with this virus, and there’s a range, some of which are very serious and some of which are less, so it’s a theory of natural selection with a normal distribution set. And, if you’re moving into a new environment like the [Life Care Center] of Kirkland or like the cruise ship, what happens is you have people who are completely unaware of what is going on. And so you take a population like in Kirkland, which is fragile and old, you get somebody in who has a strong version of the virus, and the thing will just absolutely rip through the population and kill everybody in it within a very short period of time. [Daniel Kuritzkes, the chief of the infectious-diseases division at Brigham and Women’s Hospital, in Boston, said, after being read this passage, “There’s no evidence that there are strong and weak variations of the coronavirus circulating. There may be minor variations person to person or location to location in the actual genetic sequence, but there is no evidence that they have different virulence or that a less virulent version is overtaking a more virulent version.”]

And Washington State had a very high peak early on. And, probably, most of the deaths that we have today, if you went through and chased them down, were either people who are in the facility or people who have family members who hugged and kissed people in the facility. So they got large doses of intense viruses. [Albert Ko, a professor of epidemiology and medicine and the department chair at the Yale School of Public Health, responded, “There is no evidence that you have a higher risk of dying from contact from someone in the facility than if you had contact from some other source.”] But then adaptation starts to set in. And, in my view, adaptation is a co-evolutionary process in which things change, not only in human behavior but also change in genetic viral behavior. So, on the human side, once you see that these things are really going to happen in this particular form, people take steps to avoid contact.

You write, “I fully understand the need for immediate responses to immediate threats, like fires, but not for crises that may last for three months or more.”

Yes. Well, I’m saying in effect, by this particular point—this is not the medical side—is after you start declaring emergencies you have time for reflection and adaptation and modification, which you don’t have in a fire case. So the political point is one which essentially says when you see governors of three major states putting out statements that their experts have said this, that, and the other thing is a result, and you don’t see the studies and you can’t question the assumption, I regard that as a serious breakdown in the political process. So my view on that particular point is I’d like to know which of these studies they’re relying on. If it’s the New York Times studies, then I thought that that study was mistaken for the reasons that I was trying to give you a moment ago, which is that as the virus becomes more apparent, adaptive responses long before government gets involved become clear.

You wrote, “The adaptive responses should reduce the exposures in the high-risk groups, given the tendency for the coronavirus to weaken over time.” What tendency are you talking about, and how do we know it will weaken over time?

Well, what happens is it’s an evolutionary tendency. [“There is absolutely no evidence for that,” Ko told me. According to Kuritzkes, “There is no proof that this is the case. To the extent we see that evolution taking place it is usually over a much vaster timescale.”] So the mechanism is you start with people, some of whom have a very strong version of the virus, and some of whom have a very weak version of the virus. If the strong-version-of-the-virus people are in contact with other people before they die, it will pass on. But, if it turns out that you slow the time of interaction down, either in an individual case or in the aggregate, these people are more likely to die before they could transfer the virus off to everybody else.

On the other hand, those people have the more benign version of the virus that will allow them to live longer, which means that they have the chance to make a connection with somebody else. And so what happens is, if it turns out you think something like the coronavirus is ten times as strong as another virus, what that means is that the distancing is going to be more violent, which means that the evolutionary process should be more rapid than that for the ordinary flu.

But you stated as fact that the coronavirus has a tendency to weaken over time.

Isaac, let me just explain it. This tendency takes time. It could be a week. It could be a month. It could be longer. But, in the end, you should expect something of this particular sort to take place. More importantly, on the other side, where there’s no dispute, it’s clear that people will start to evolve away from these things so that the rate of transmission will start to go down. And, as that starts to happen, whether you have just private or social responses, you’d expect the rate of transmission to go lower. So the question then is, How quickly will this thing peak? And, if you looked at the standard model, it basically postulated a very long buildup and a peak which is about three months away from now. And that just strikes me as being too high and too far relative to everything else that we have.

So, in the United States, if you start looking at yesterday’s figures, it was about two per cent higher than the day before, which is already indicating that the speed-up is slowing down. We’re going to have to see what the next days do. But we’re talking ten thousand cases a day at the current maximum, and the flu was vastly larger in terms of its number and its extent. And my sense is, given the reactions that you’re going to have, this thing will peak earlier and start to decline earlier than the common models start to say, because they don’t seem to build in anything by way of adaptative responses.

You keep talking about your “sense.” I think that’s the word you’re using. But you’re stating as a fact that the virus is going to weaken over time. It seems like we do not know that. We can turn to other viruses and how they’ve—

No, that is not what I said. I said there’s a long-term tendency in these ways. Over time, yes. And is this a hundred-per-cent tendency? No. Is there any known exception to it? No. [“We did not see SARS or Ebola weaken over time,” Kuritzkes said. “It is only appropriate public-health measures or vaccines that have helped to control those epidemics.”] It seems to me that if you do this, what you’re trying to do is to figure out what the probabilities are, and I think the answer is, if you look at all of the cases that we’ve seen, no matter what’s going on, even if you subtract out the coercion that was used in China and in Korea and so forth, you cannot come up with a credible story that those places in Korea would have had, say, a half a million cases a day. Or in China you would have had, say, thirty million cases a day. And so I do think that the tendency to weaken is there, and I’m willing to bet a great deal of money on it, in the sense that I think that this is right. And I think that the standard models that are put forward by the epidemiologists that have no built-in behavioral response to it—

And you’re not an epidemiologist, correct?

No, I’m trained in all of these things. I’ve done a lot of work in these particular areas. And one of the things that is most annoying about this debate is you see all sorts of people putting up expertise on these subjects, but they won’t let anybody question their particular judgment. One of the things you get as a lawyer is a skill of cross-examination. I spent an enormous amount of time over my career teaching medical people about some of this stuff, and their great strengths are procedures and diagnoses in the cases. Their great weakness is understanding general-equilibrium theory.

O.K., so your expertise in the subject I guess comes in part from your work with AIDS, which you just referenced, is that right?

AIDS, and I’ve worked on evolutionary theory for forty years in its relationship to law.

You write, “There are two factors to consider.

1. One is the age of the exposed population, and the other is the
2. rate of change in the virulence of the virus as the rate of transmission slows, which should continue apace. By way of comparison, the virulent AIDS virus that killed wantonly in the 1980s crested and declined when it gave way to a milder form of virus years later once the condition was recognized and the bath houses were closed down.” [I read this passage to Kuritzkes, who responded, “It’s completely inaccurate. It had nothing to do with the change in the virus. We were able to do it by safe-sex practices and the like, and we saw the explosive growth of H.I.V. during the nineteen-nineties in sub-Saharan Africa and more recently in Eastern Europe. There is nothing about the virus that has become less virulent.”] What milder form of the virus are you talking about?

Look, all it is is it’s a distribution. What you do is you figure out what this toxicity strength is and if it’s X at one point, then it’s going to be some fraction of X down the road. And it’s quite clear that that is what happened with AIDS. And then, when it comes along and you start getting [the antiretroviral drug] AZT and other conditions, it’s easier to treat them because all of a sudden AIDS is evolved in much the same path as syphilis. If you go through the history of syphilis, it starts off, it’s essentially a deadly disease and kills most people. And then those who survive have the milder version of it. And so after a while what happens is it becomes a tamer disease. [Syphilis is a bacterial infection, not a viral infection. “One doesn’t have anything to do with the other,” Kuritzkes said. Ko told me, “That’s not something that is based in empirical evidence, so the fallacy in his argument is the over-all lack of scientific rigor in his analysis.”]

You’re saying that there’s a milder form of the H.I.V. virus than what was around in the eighties and nineties. That’s what you’re saying?

Well, I mean, there’s always been a continuous distribution from severe to less severe. What I’m saying is the probability distribution switches so that the medium becomes a little more mild. And, if it becomes more mild, it kills fewer people. And, after a while, it becomes something that becomes a chronic disease of some sort or another instead of something that’s virulent.

O.K., but you used the phrase a “milder form” of the virus, which I could find no scientific backup for, so I wanted to clarify that point.

Well, I’m just telling you, I’m giving you this as a theory.

Oh, it’s a theory.

No, look, I’m not an empiricist, but, again, let me just be clear to you, because you’re much too skeptical. The evolutionary component has not been taken into account in these models, and so before one is so dismissive, what you really need to do is to get somebody who’s an expert on this stuff to look at the evolutionary theory and explain why a principle of natural selection doesn’t apply here.

What I’m doing here is nothing exotic. I’m taking standard Darwinian economics—standard economic-evolutionary theory out of Darwin—and applying it to this particular case. And, if that’s wrong, somebody should tell me. But what happens is I just get these letters from people saying, “You’re not an expert. The H1 virus differs from this one in the following way.” What I don’t get from anybody is a systematic refutation which looks at the points parameter by parameter.

I guess my point is that shouldn’t you be careful about offering up these theories before they’re printed?

No. It turns out in this particular world if you become quiet about this stuff it never gets heard. And what we’ve had now is very loud talk on one side. I think most of it is incorrect. I’m always willing to debate somebody on the other side who wants to say this is the way the model works. In fact, I have several of my Hoover colleagues who have done exactly that.

Richard, with all due respect, your article is apparently circulating in conservative circles in Washington and the White House.

I didn’t write it as a conservative article. Donald Trump’s name is not mentioned in it.

I grasped that.

I’m not interested in politics.

I know, but we have a responsibility when we put our name on something, no?

I absolutely do. And I told you I’m willing to debate anybody and anywhere at any time on this particular topic, and we’ll see how it comes out. What was the last sentence in that article? Would you care to read it again?

I’ve got it right here, I believe. “Perhaps my analysis is all wrong, even deeply flawed. But the stakes are too high to continue on the current course without reexamining the data and the erroneous models that are predicting doom.” Are the stakes too high to publish articles with basic errors?

This is not a mistake. It’s an open challenge. I’ve spent my entire life as a lawyer taking on established wisdom. My view about it is what you’re asking me to do is, when I think everybody is wrong, to remain silent, and the stakes are too high. So my view is there’s all these experts on the other side. Somebody come up and explain why it is that they think the results are going to be different. Looking at the data thus far, both theories tend to predict a sharp rise at the beginning, mine less sharp than the one that’s coming out.

In the next week or so, we’ll see. I will be, shall we say, much more compromised if we start to see a continuing explosion of deaths going on for two or three weeks. But, if the numbers start to level off, the curves will start to go downward.

I was just asking about—

I’m saying what I think to be the truth. I mean, I just find it incredible—

I know, but these are scientific issues here.

You know nothing about the subject but are so confident that you’re going to say that I’m a crackpot.

No. Richard—

That’s what you’re saying, isn’t it? That’s what you’re saying?

I’m not saying anything of the sort.

Admit to it. You’re saying I’m a crackpot.

I’m not saying anything of the—

Well, what am I then? I’m an amateur? You’re the great scholar on this?

No, no. I’m not a great scholar on this.

Tell me what you think about the quality of the work!

O.K. I’m going to tell you. I think the fact that I am not a great scholar on this and I’m able to find these flaws or these holes in what you wrote is a sign that maybe you should’ve thought harder before writing it.

What it shows is that you are a complete intellectual amateur. Period.

O.K. Can I ask you one more question?

You just don’t know anything about anything. You’re a journalist. Would you like to compare your résumé to mine?

No, actually, I would not.

Then good. Then maybe what you want to do is to say, “Gee, I’m not quite sure that this is right. I’m going to check with somebody else.” But, you want to come at me hard, I am going to come back harder at you. And then if I can’t jam my fingers down your throat, then I am not worth it. But you have basically gone over the line. If you want to ask questions, ask questions. I put forward a model. But a little bit of respect.

O.K. Let me ask you this question. All my questions are asked with respect.

That’s not the way I hear it.

O.K. Let me ask you this question. This is an important point, so I want us to come to some sort of resolution on this, because I think it gets to the crux of what you’re saying. You write, “Moreover, it is unlikely that the healthcare system in the United States will be compromised in the same fashion as the Italian healthcare system in the wake of its quick viral spread. The amount of voluntary and forced separation in the United States has gotten very extensive very quickly, which should influence rates of infection sooner rather than later.” Here is what confused me. Are you saying we are overreacting, which seems to be the tone of the rest of the article, or are you saying that the measures we are taking are going to prevent an Italy-like disaster, in which case, why the complaints about the overwhelming response?

Well, I think the answer is all of the above. If what I’m saying is correct, we should not expect to see many replications of what has happened in New York, where in fact the caseload here is actually lower than it was during the height of the flu season last year. [At the time of this interview, Epstein was correct. Now, however, the number of COVID-19 cases surpasses it.] What’s happening is the amount of quarantine that we’re putting in has created strain on the system, because what we do now is we spend much more money per case than we did in the summer, because we have more respirators, more isolation, more special uniforms, more special ambulances, and so forth.

But also on the other side we’re going to see a very strong reaction. The self-quarantine that you’re seeing taking place in New York, where I live, this is hugely voluntary. I have so many friends, not only are they going into quarantine, anybody who has a house in Connecticut or New Jersey, in the suburbs, they go out and they’re spreading this thing out there. It’s like the days in England when people would leave London during the plague and head off to Oxford.

So I think all of these forces are going to take hold fairly quickly. And the Italian situation, well, what’s the problem there? This is a national health-care system, and the way in which national health-care systems ration, since they typically do not devote enough resources and they charge no prices, is they ration by queuing, which is they have people wait. And so the stories that come out of Italy, there was one in the Jerusalem Post, said that they had a consistent policy of not giving a respirator treatment to anybody who got the virus who was over sixty. Well, that’s going to clearly change the mortality rates. [Italy does not have a consistent policy on ventilators but has had to ration their use. New York City officials have warned that they, too, may run out of equipment. On Sunday, Mayor Bill de Blasio said, “We’re going to need at least several hundred more ventilators very quickly.”]

And, if you start looking at the mortality rates in Italy, ninety per cent or eighty-five per cent of the deaths are from people over seventy, which suggested something like this is going on. [The number appears closer to seventy-five per cent.] In the United States, we don’t have quite the skew, and we have a little bit better supply on things. I think we have a greater ability to ramp up in terms of things. So my guess is you will see real stresses in the short run in the New York area, but that given the time that’s available, I first of all did not think that the runoffs will be as great anywhere else throughout the system. All right? But, on the other hand, I think there’ll be more preparedness. My fear is that there will be overpreparedness if it turns out that the numbers that they’re preparing for are too high. Illinois declared that the state was on lockdown when they had twelve cases. By the way, Bill Gates agrees with me, I’m happy to say.

What does he agree with you on?

He thinks you have to relax the economic, lockdown restrictions. They are too severe.

Just to clarify, this is from a Vox Recode article, “Bill Gates rebuked proposals, floated over the last two days by leaders like Donald Trump, to reopen the global economy despite the Covid-19 coronavirus outbreak, saying that this approach would be ‘very irresponsible.’ ” Gates said, “There really is no middle ground, and it’s very tough to say to people, ‘Hey, keep going to restaurants, go buy new houses, ignore the pile of bodies over in the corner. We want you to keep spending because maybe there’s a politician who thinks GDP growth is all that counts.’ It’s very irresponsible for someone to suggest we can have the best of both worlds.”

I misread him then. Whatever Mr. Gates said, that’s fine. That strikes me as more populist than I am. The question is we managed to survive all of these things during the periods of flu vaccine. Nobody wants to trivialize death. You’re trying to minimize them. The point that I’m making is, you shut down the economy and put it on lockdown, you’re going to get medical deaths, when people don’t get adequate supervision and care for other kinds of conditions that also kill.

Right. I think the point that Bill Gates and others are making is that if we don’t shut down the economy or large parts of it, we’re going to get the economic devastation anyway down the road, because we’re not going to stop the virus.

That presupposes that the basic model is correct, that they’re relying on it. I think it is wrong. And the whole point of writing that article was to attack that model. Let’s put it this way: if in fact that model is right, it’s not at all clear that shutting down the economy will do any good, because it may well just be so devastating. And, if it’s wrong, it’s absolutely clear that the model is devastating.