Thursday, March 26, 2020

The Potentially Deadly Toll of a Law Professor's Libertarianism

by Michael C. Dorf

My Verdict column calling for a national lockdown and, if needed to ensure its enforcement, suspension of habeas corpus, continues to receive pushback. Some of that pushback has a through-the-looking-glass quality, like this piece in RT, which for those unfamiliar, is a Putin-backed propaganda outlet for pro-Trump and other trollish views. I shall wear being called an "authoritarian minded law professor" by an organ of a foreign authoritarian as a badge of honor.

That said, the RT article makes a valid point: people who raise doubts about lockdown policies could have a legitimate argument and thus should be given a hearing. Today I'll look at the most prominent law professor to make the case that we are overreacting: Richard Epstein. His views do not withstand even the most minimal scrutiny and may already be responsible for serious real-world effects.

In a column published on March 12, Professor Epstein predicted that US deaths from COVID-19 would  ultimately total about 500. The US exceeded that number early this week, leading Epstein to correct the number to 5,000 and acknowledge his error. That's fair. Anyone can make what Epstein acknowledged was "a stupid gaffe." Yet that revised estimate is still probably at least two orders of magnitude too low. Meanwhile, there are two further problems.

First, and most consequentially, Epstein's gaffe (and the rest of his analysis) apparently played a key role in persuading Donald Trump to make his deadly pivot back towards playing down the COVID-19 threat and set the absurd goal of restarting the economy and filling pews on Easter (seventeen days from today). Even if Trump walks that back, in the meantime, his airing of the view encourages his followers to go about their business heedless of the risk and encourages Republican governors to drag their feet. Epstein's "stupid gaffe" could end up being responsible for further overwhelming our health care system and thus causing numerous avoidable deaths.

Second, Epstein's follow-up column of March 23 makes additional, even stupider gaffes. In his opening paragraph, Epstein states:
In Spain, the death toll is 2,206.  Italy has taken the lead with 6,077 deaths, 85 percent of which are of people over 70, which stems, it appears, from a conscious decision not to supply ventilators to anyone over 60.  [China, Iran, Italy, and Spain] make up close to 13,000 deaths or about 82 percent of the total. Taken together, these . . . countries account for over 13,595 of the 16,097 deaths. The good news here is that the growth rates in both Italy and Spain have turned downward in the past 48 hours.
I'll focus on two absurdities. First, Epstein dismisses the high death rate in Italy because it appears to result "from a conscious decision not to supply ventilators to anyone over 60." Yet doctors in Italy are not wantonly cruel. That tragic decision is OBVIOUSLY a result of  a flood of cases and thus a shortage of ventilators. Can Epstein be unaware of that fact? It's hard to see how. To support his point, he links to a short Jerusalem Post article that expressly reports the ventilator shortage as the reason for the tragic triage decision. And yet Epstein somehow thinks that he has provided an argument against measures designed to flatten the curve.

Second, what are we to make of the "good news" that "that the growth rates in both Italy and Spain have turned downward in the past 48 hours" as of Monday? Here are the daily new cases in Italy through yesterday:

And here are daily new cases in Spain through yesterday:

Notice anything? Well, for one, Epstein's claim is not that the numbers of new cases are declining substantially, but that the exponential rate of growth is slowing. That does not appear to be true in Spain, although perhaps that's just bad timing on Epstein's part, because the spike of the last three days came after his column appeared. So let's focus on Italy, where the rate of growth does appear to have slowed, even though the situation remains catastrophic.

Slowing the rate of exponential growth of new cases is important but inadequate unless the exponent is less than 1. Otherwise, the health system continues to be overwhelmed and people who could receive life-saving treatment are left to die. It's still better to reduce the value of the exponent, even if it remains above 1, of course, because that spreads cases out over time and more people get the disease later in the pandemic, when better treatments may be available, but the goal should be to bring the growth rate down as close to zero as possible.

I have not yet named Epstein's biggest, stupidest gaffe of all: If Italy's new cases are in fact finally turning down, that is almost certainly because it has been a little over two weeks since the country went on lockdown. This is the pattern that appeared in Wuhan; given the incubation period, new cases continue to rise until about two weeks after lockdown. Then they fall. If Italy is a source of "good news," it is most likely due to the very policy that Epstein deems an overreaction. To borrow a phrase Justice Ginsburg used in another context, pointing to the slowing spread in Italy to justify a go-slow response to the coronavirus "is like throwing away your umbrella in a rainstorm because you are not getting wet."

What are we to make of Epstein's almost-criminal stupidity? Richard Epstein is hardly a stupid man. Nor do I have reason to think that he is a pathological liar like the President whose policies he has disastrously influenced. What then explains his making arguments that are either extremely stupid, extremely dishonest, or both? Ideological priors seem like the only possible explanation.

Epstein is a libertarian who distrusts government. Rather than thinking pragmatically, he falls back on his familiar truisms, saying in his March 23 column that "[t]he central Hayekian principle applies: All of these choices are done better at the level of plants, hotels, restaurants, and schools than remotely by political leaders." It is said there are no atheists in foxholes. Apparently libertarians are made of more stubborn stuff.

Now I want to acknowledge that there is a legitimate other side to this argument that need not rely on stupidity, dishonesty, or ideological rigidity. There is some evidence that the coronavirus is much less deadly but much more widespread than widely assumed. If so, that's good news over the medium to long run. It would mean that pretty soon we'll all have it and most of us will be fine. But this possibility does not undercut the case for lockdowns now, which aim to spread the still-large number of acute cases we will still have over a longer period in order to avoid overwhelming our health care system. I should really say in order to avoid continuing to overwhelm our health care system. With hospitals in New York City already experiencing an apocalyptic surge of cases and deaths, the need to slow the spread is beyond urgent.

Finally, I also want to acknowledge that the economic and social cost of shutdown is real and enormous. But let's be clear that the economic and social cost--as well as the cost in lives--will also be enormous if, as now seems inevitable, we end up doing too little. Moreover, effectively implemented, lockdown has long-run economic benefits. The point of lockdown is to reduce the transmission rate enough so that we can begin to shift to a South Korea-style approach that relies on widespread and rapid testing coupled with isolation and contact tracing. Increasingly, and with each passing day, however, the pandemic's expansion and the inadequate response of the Trump administration and those state officials and private actors following in its footsteps make that approach unavailable.


Joe said...

A takedown that is the sort of Takings that requires no compensation.

Greg said...

There is another, perhaps more fatalistic argument for allowing the healthcare system to be overwhelmed, and for not locking down too strongly:

What if it's truly too late to put the genie back in the bottle, and before this is all over, 3 million Americans will be dead, 60 million Americans will require hospitalization and 15 million Americans will need a ventilator.

The numbers on this are incredibly grim. If every hospital bed in the country were dedicated to only COVID-19 patients and nothing else, it would still take 3-6 years before we could come out of lockdown. I don't know what the stimulus package looks like that allows the economy to survive 6 years of being on lockdown.

Based on this fatalistic view, governments have to make a decision not about whether to stop the hospitals from being overwhelmed, but on how much we allow them to be overwhelmed, versus how long we allow the lockdown to happen. Maybe the right solution is going to triple hospital capacity and let the economy stagnate for a year. Maybe it's going to 6x capacity and cutting it to 6 months. (Ironically, policies that produce more transmission are also tolerable for longer periods.) I don't know nearly enough about epidemiology, crisis management, or economics to even begin to offer suggestions for how to handle that nightmare scenario, but I know enough to know that it could be where we're already living.

The few politicians who have been willing to face the numbers on this have been called callous or even murderers, but I don't think they're necessarily either.

Eric Charles said...

Suicides go up significantly during recessions.

The researchers found that there had been an increase in the relative risk of suicide associated with unemployment across all regions of 20% to 30%. There were an estimated 233,000 suicides a year between 2000-11, of which around 45,000 could be attributed to unemployment...

...The researchers found that there had been an increase in the relative risk of suicide associated with unemployment across all regions of 20% to 30%. There were an estimated 233,000 suicides a year between 2000-11, of which around 45,000 could be attributed to unemployment.

Eric Charles said...

...This study shows that the impact of unemployment was strongest in countries which enjoyed comparatively low unemployment rates before the crash. One possible explanation for this is that not only is the shock of losing a job much greater in these areas but the consequent loss of status also increased.

Ryan said...


The goal of a lockdown isn't to "put the genie back in the bottle," i.e., eradicate COVID in the U.S., it's to buy time to put in place a robust test-trace-and-quarantine system (possibly with other measures, such as bans on large public gatherings) sufficient to reduce the transmission of COVID to a manageable level while allowing a return to something resembling economic normalcy.

Oleg said...

Your noting of an error by Epstein regarding Italy's numbers is incorrect. In fact, Italy's rate of "exponential" growth has not only slowed, but daily cases are declining (on average). That is, the daily growth rate IS <1.0, if only slightly.

His point was, as he made in the vigorous podcast debate with Adam White on March 23, 2020, that virus outbreaks have an invariable pattern. They grow exponentially initially, inflect, and decline, usually at roughly the same rate as they went up.

To your credit, you partially acknowledge this. Infection rates in Italy are not just "probably", but almost certainly MUCH higher than recorded numbers (possibly a million or more). Given that fact, the effect of a lock down after the horse has bolted is less obvious than you suggest. It is widely, and likely correctly, presumed that individuals who have had even mild cases and recovered are no longer a vector for spread. Time will tell, but my guess is that the lifting of extreme restrictions in Italy will not lead to an immediate, or even certain, return to past exponential growth, due to the large number of convalescents. If that's true, then Epstein will have been more right than wrong.

Scott said...

If hospitals are going to be overwhelmed either way, doesn't "flattening the curve" just increase the number of people with unrelated issues who will be unable to access care?

CEP said...

A couple comments from a biochemist who has actually studied this sort of thing, admittedly a few decades ago:

(1) These graphs should look an awful lot like reaction profiles studied in chemistry-for-majors classes (virtually never in a watered-down non-majors "introduction to chemistry" class). The key point of a reaction profile is that it is not instantaneous — it takes place over time. (The contrast with the Efficient Market Hypothesis's ideological meme that all information is incorporated in a security's price instantaneously upon that information becoming available to the market bears careful consideration.)

And it's not just that the reactions themselves take nonzero time; that's the first-year chemistry approach. Reactions in systems also take time because not all reagents are exposed to the reaction conditions — including, in particular, to heat — at the same moment, with the same intensity, and/or in the same conformation. That's why advanced chemistry undergraduates look so much less certain about things than, say, the typical lawyer arguing an environmental-law case (for either side), even those few who made it all the way through second-year organic chemistry: The required coursework in physical chemistry and inorganic chemistry, which is restricted to chemistry majors at most schools, demonstrates otherwise in both the lecture hall and the laboratory.

And it's even more fun when dealing with the buffers of in vivo systems and how they inherently stretch out reactivity and results. Consider, as an obvious analogy (with all of the caveats that go along with reasoning by analogy), the difference between test-tube reactions to turn sugars into alcohols with yeast-based fermentation. Mmmmm, need some good ale right about now... which cannot be produced in a test tube without those yeasts. Which take time.

(2) More fundamentally, Epstein's analysis fails to consider reinitiation of infection by later exposure. That "fourteen day incubation period" is not a statement of absolute safety; it is an estimate, at some unstated statistical level based upon (highly educated guess-level) conjectures of actual population state and dynamics, that a substantial portion of the population will no longer be contagious fourteen days after symptom onset. But:
* Some of that population will self-incubate viral subpopulations to extend their virulence beyond the purported limit; and
* Some of that population will be misdiagnosed because testing isn't sufficient, and will only actually contract COVID-19 later in the respiratory-distress period (meaning that for that patient, the fourteen days began later); and
* A variety of other causation factors extend the contagion window beyond fourteen days.

What all of this means is that we can expect a spike of some kind beginning approximately a week to ten days after any lifting of our imperfect isolation measures. The closer we set that lift date to the bandied-about fourteen days, the larger that spike will be and more likely it will create a second wave. In short, we risk 1919 and the Spanish Influenza shape of things again...

The Epsteinian analysis is seeking to avoid a mineshaft gap, while implicitly advocating for the social conditions in the mineshafts posited by their proponent.

CEP said...

On another issue, I cannot resist jumping on Mr Charles's invocation of a horribly flawed article. If we're going to look at "death rates" that are purportedly caused by recessions, we have to look at the entire "death rate" profile and not reason that an increase in one part of that profile — suicides, and I am not denigrating them — necessarily means that the entire profile rises by a corollary amount. In fact, the rise in suicides cited in that article was more than offset by reductions in deaths caused by "economic activity" — traffic accidents, workplace injuries, etc. are just the most obvious ones. (Most citations I could provide for this are behind paywalls and some I've only seen on paper.) So any conclusion that the total death rate would be reduced by recession is just flat wrong... which would have been revealed by reading the whole article in The Lancet and not just the headline-grabbing "summary" at World Economic Forum.

Suicide rates deserve attention. But not in isolation; and the irony that "not in isolation" is the actual point of Dr Nordt's research seems to have escaped most people, especially the headline writer at World Economic Forum, who try to cite Dr Nordt's work in isolation.

Greg said...


Is it even possible to isolate COVID-19 to that extent while still allowing people to leave their homes, particularly in places that already have so many cases, like NYC?

Unfortunately, this is one of those areas where it really is tough to see through the media, and the people who know seem like they're being cagey because, well, they don't know either.

This is why some people (myself included) are starting to feel like if the only way to get through this is when everyone has had it, then we need to figure out how to let everyone have it and move on from there.

If there really is a way to get it tamped back down, then that's great. However the experts are consistently giving very open-ended numbers on how long these lockdown conditions are going to last, and if that means 6 or more years, then I currently have very little confidence that I'll survive to see the end of it.

To be clear: While I'm being alarmist, I'm also being completely honest. I'm genuinely concerned that I am going to starve in a few years when the money runs out and the economy continues to collapse while we remain in lockdown. That is even assuming that I'm not murdered in the interim over a bag of groceries that someone else who isn't willing to starve can't afford.

Eric Charles said...

CEP- I'm not sure anyone suggested the total death rate goes up during recessions including that article. The point is that not all deaths seem the same. Most would people are willing to accept tens of thousands of deaths per year on the highways than accept a 15mph speed limit law that would drastically reduce this at the cost of convenience. The same applies to sugar consumption that we accept a the expense of early death. Are skyrocketing suicides and possible starvation acceptable if it means a reduction in global warming that will arguably save millions? It seems qualitative differences in death rates should be considered.

Franklin Cox said...

Eric, you committing the fallacy known as "missing the point". Suicides are probably correlatable with high unemployment in some arcane way deep within the Koch "think tank" network. But please, are you claiming that unemployment *causes* suicides in the same way that coronavirus causes deaths?

I truly hope you aren't, because that would indicate that you are a loon.

If you should manage to establish that causal link, I think you'll agree that any company that downsizes and throws substantial numbers of workers into unemployment should be charged with negligent homicide, as you will have established a tight causal link between unemployment and suicide.

Franklin Cox said...

Greg, please read Tomas Pueyo's article in the Medium. One point of acting now is to prevent complete overloading of the medical system. The time to reduce infection rates is early on. The numbers we have of reported cases are going up 20% a day. If that continues until April 30, then about 50 million Americans will be sick, which is over 50 times as many people as hospital beds in the country. If, in mid-April, the rate of increase goes down the 15% a day, by the end of April there will be 30 million fewer people sick. If the rate is cut right now, the numbers by the end of April will be dramatically slower. IF you slow down the spread of the disease, you buy time to improve treatment, find effective remedies, and so forth.

Franklin Cox said...

"Blogger Scott said...
If hospitals are going to be overwhelmed either way, doesn't "flattening the curve" just increase the number of people with unrelated issues who will be unable to access care?"

This is a bizarre observation. If hospitals are less overwhelmed by the coronavirus, then they will almost certainly be able to handle more ordinary cases.