“Herd immunity,” time consistency, and the epidemic yoyo

Saloni has long and detailed arguments against herd immunity.  Here is Caplan on Hanson.  Here is Kling on Hanson.  Here is the Taleb critique.  Here is the underlying Imperial College paper everyone is talking about.  The bottom line is that “locking everyone up to bend the hospital admissions curve” might have to last for at least a year to really choke off the coronavirus.

I’m not going to recap this complex debate, which most of you already have some inkling of.  Instead, I’d like to stress the issue of time consistency, noting that I’ll consider some extreme versions of policies to make exposition easier, even though no one advocates exactly those extreme versions.

Let’s say we expose lots of people to the virus rather quickly, to build up herd immunity.  Furthermore, we would let commerce and gdp continue to thrive.

Even if that were the very best policy on utilitarian grounds, it might not be time consistent.  Once the hospitals start looking like Lombardy, we don’t say “tough tiddlywinks, hail Jeremy Bentham!”  Instead we crumble like the complacent softies you always knew we were.  We institute quarantines and social distancing and shutdowns and end up with the worst of both worlds.

Alternatively, let’s say we start off being really strict with shutdowns, quarantines, and social distancing.  Super-strict, everything closed.  For how long can we tolerate the bankruptcies, the unemployment, and the cabin fever?  At what point do the small businesspeople, one way or another, violate the orders and resume some form of commercial activity?  What about “mitigation fatigue“?

Again, I fear we might switch course and, again, end up with the worst of both worlds.  We would take a big hit to gdp but not really stop the spread of the virus.

I also can imagine that we keep switching back and forth.  The epidemic yoyo.  Because in fact we find none of the scenarios tolerable.  Because they are not.

David Brooks postulates another possible form of time inconsistency:

What happens when there are a lot of people who’ve had the disease and become temporarily immune. They start socializing. Social distancing for the rest become harder if not impossible.


I greatly fear the epidemic yoyo.  And figuring out how to deal with it may be at least as important as calculating the numerical returns from various consistent policies.

I thank an anonymized correspondent for the term “epidemic yoyo.”


My upcoming column on Takimag.com considers whether, if we suppress the pandemic now, are we just doomed to be struck by a massive second wave late in the year? Are we just kicking the can down the road for 7 months?

I'm certainly not an expert, but I think there are a lot of reasons to be more optimistic than the Imperial College report suggests. E.g., by autumn we should finally have plenty of tests and facemasks like South Korea has. Our treatment regimens should be much improved. And so forth and so on.

There are a lot of reasons for hope.

Not just kicking the can, but kicking it none the less.


Hubei deletes a report claming COVID-19 damages male fertility.
In other words, herd immunity = herd infertility.

"The infection could result in “impairment of immune homeostasis in the testes”, which could cause orchitis – an inflammation of the testicles – which in turn could reduce a man’s sperm count and possibly lead to infertility, it said."


I would view this as speculative at best.

However, we're talking about a virus that has already mutated in a large way effecting its aggressiveness once already. Are we so certain that the next mutation won't make it kill 20% of children? Or increase r0? Or that it already has long-term health effects in mild cases that we haven't had time to see, like shingles for chicken pox?

In my view, we don't have enough information to bet everybody on the long-term effects of billions of cases being benign (even if they're 80% non-hospitalizing in the short-term). Especially when the downside is trading a 1% fatality rate for a crowded hospital 5% fatality rate.

When Kling talks about separating the young from the old, I'm not sure if he meant just two groups. But maybe we should be separating them both from the general population.

If we did do Hanson's plan of deliberate infection, we should put the young adults (volunteers, of course) into their own isolation to purposefully get the disease. We can slow mutation by purposefully infecting them from an original source. The young adults seem determined to catch it anyway by going to beaches and bars.

I want to emphasize: We should keep on with our current plans of closing schools and keeping people at home and doing monitoring at the border, because those are the best plans we have at the moment. But we need to be thinking of other options, and we can't discard other options merely because they are risky -- all roads are risky. My parents are extremely vulnerable to this, and I'm not sure "just isolate" will work for a year.

I've wondered if we should do things like deliberately infect military units or some subset of medical personnel.

But my understanding right now is that it's seen as too risky to infect even healthy young people, in part because they see long-term lung damage in cases that fall short of death and also because rehab is difficult. But that's based on medical accounts that are best described as anecdotal.

Here's my new column "Reasons for Hope:"


At least takimag will have no problem with Kung Flu jokes, unlike other more PC places.

+1, or the Wuhan Flu

Completely unrelated to PC but it's not a flu. Coronaviruses cause SARS and influenza viruses (Orthomyxoviridae) cause influenza. Now your non-PC jokes don't make any sense. Might as well call it the Walla Walla Washington Whooping Cough.

Yeah, I know that. And the mouse I'm using by my computer isn't actually a rodent.....

Good column. But you write:

In contrast, Japan, like the U.S., lags behind South Korea in testing capacity. So, while its case numbers are probably badly understated, there haven’t been many deaths in Japan: only 28 at latest count. And Japan has one of the oldest populations in the world, with a median of 48 years of age, a couple of years older even than hard-hit Italy.

Question: What is the difference between Japan and Italy?

Answer: Italians kiss, Japanese don't.


P.S. I haven't been to Japan, so I'm just going by Shogun and "word on the street." ;-)

The number one reason is that the Japanese have lots of experience with pandemics since every winter practically breeds a new strain in that part of the world. Social distancing and wearing masks are part of the routine over there. It's not just pandemics but earthquake, tsunami, typhoon, and heat waves which baked disaster preparedness into the cultural DNA of the Japanese. In America, we have prepper culture which has the right idea but isn't as battle tested and a little kooky. The Italians had they the same experience as the Japanese would cut back on contact but this was a new thing for them. The next generation will learn.

Nice article.
I remember reading a few weeks ago (academic journal, alas I looked but could not find it) that in Japan they instructed their entire society to avoid contact, to wear masks, and the paper traced the reduced incidence of influenza after these instructions were given.
I see that Feb 27 they closed schools for a month, with only 210 cases in a country of 100+M.
Source: American-born Japanese and have visited many times.

I remember reading a few weeks ago (academic journal, alas I looked but could not find it) that in Japan they instructed their entire society to avoid contact, to wear masks, and the paper traced the reduced incidence of influenza after these instructions were given.

Bowing can be done at six feet. It's tough to shake hands (wearing gloves of course) at a distance of six feet.

Cheek kisses are even more difficult to do at 6 feet distance.

Maybe we'll see people "air-shaking" hands and "air kissing" at six feet in the future. Though with a mask on and at six feet of distance, the "air kissing" would be particular difficult to figure out what was being conveyed.

P.S. And of course, this would need to be eliminated completely:


We'll have a lot more test results within a week.

After those results come out, it will be hailed as a major victory under the wise leadership of the president. who unlike other people, had the foresight to prepare the American nation for its upcoming challenges.

Gallows humor is always appropriate..

Of course, President Trump quickly figured out the CDC was an impediment and found a better way.

This is from March 15, detailing the experiences of someone returning from Europe after having flown earlier to another country that had yet to close its borders. "I recently flew to Qatar for a meeting. Immediately upon disembarking, passengers walked past a temperature measuring device to identify those with a fever, so they could be segregated out before entering public areas. Dulles had no such plan."

No impediments made by Trump in creating this wonderfully wise situation. "Instead, after the agent examined your passport, he pointed a thermometer device at your forehead. By that time, you would have spent three hours in close contact with hundreds of other people. Even the way the lines were organized, snaking around, might have been designed to ensure that one sick person would expose the maximum number of others."

The chocolate ration will be increased to 15 grammes soon too..

Prior is both against state capacity because someone from GMU said it’s important, and heavily criticizes the US response for not having sufficient state capacity.

Prior’s Schrödingeresque reasoning.

On March 1 we had 98 cases in the US. Now it's 35,000.

The average daily growth has been 32%. If that holds, by next Monday there will be 185,000 cases.

Philosophical trigger warning ;)

As mortal beings, kicking the can down the road is all we do since birth.

Getting vaccinated, beating a flu or surviving cancer is just "kicking the can down the road". Is it worth or not?

Well, sometimes we stop kicking the can, and we get in that car, we board that plan, we go to that concert, or we kiss that friend on the cheek. Life's full of tough choices.

The Four C’s of COVID-19: Coughs, Contacts, Communication, and Calm:

1. Cough Isolation: watch each new cough for 3 hours, isolate immediately for 14 days if persistent

2. Contact Tracing: Carefully trace your contacts from the moment of your first 3-hour persistent cough for the critically contagious first 5 days, and for a cautious full 14 days to make sure.

3. Communicate Transparently: phone your public health officials, tell your close contacts of your Self-Isolation, and tell anyone in your Contact-Tracing list to Self-Isolate as a precaution.

4. Calm: very few people get very sick but the true danger is people ignoring the first three C’s and causing undetected Community Transmission. Even with widespread Community Transmission, following the first three C’s will first slow then stop the spread.

#4 is the key to the whole thing. We have to be calm and realize this just is not a rough enough virus to shut down the world over.

The point of flattening the curve is not to squeeze under the horizontal lines of health care system capacity. The point is to give us a few weeks time to bend that capacity upward. This point is rarely made.

We ought to be able to add a lot to treatment capacity in six months.

In 1942 the United States built a 14,000 ton Liberty cargo ship in 4 days, 15 hours, and 29 minutes.

But it took time to work up to that.

Is the United States even laying out new ventilator production capacity at the moment? (My country isn't, but that is a very poor excuse.)

The US launched an immense number of warships in the last 30 months or so of WWII, but it took until something like May 1943 to get a new aircraft carrier, the most important ship, into the fight.

The USS Ventilator.

(It's actually a rather threatening name.)

But the jokes tend toward the obscene, what with the blowing and holes, connections to ventilation and destruction. Although, according to the testes option, Lil Covid may connect sex and morbidity in ways never before expected.

She blows up her enemies.

The free market will provide everything America needs.

Eventually, and not for everyone, but the invisible hand works in mysterious ways.

On one hand - yes, western society is more lethargic and ossified than it was in 1900-1970 period. However, the war economy during WW2 was mostly about minor adjustments to existing manufacturing capacities - i.e. instead of normal cars, make Willys MB, instead of ferries, make military ships, etc.

In the current situation, there simply isn't the manufacturing capacity to repurpose. Apparel industry is almost non-existent, and whatever is left is geared towards luxury that doesn't want to dilute the brands on some banal flu masks. Medical equipment might be in a better spot, but even there, most of it is manufactured in Europe and East Asia; out of components made in China.

There is a trade-off between optimization and resiliency. Western world has been optimizing for several decades now, leaving it very vulnerable to any adverse events.

Is the United States even laying out new ventilator production capacity at the moment?

The solution is not to need new ventilator production capacity. With sufficient tests, contact tracing, filter masks, and washable cloth gloves, there should be no need for additional ventilator production capacity.

This will look quite foolish in a few weeks, at least in terms of the people dying.

However, you are not wrong at another level - an extra 50,000 ventilators would require trained staff 24 hours a day (requiring constant replenishment whenever someone is infected, even assuming adequate use and supplies of PPE), and that bottleneck is considerably more intractable then simple production of machinery.

This will look quite foolish in a few weeks, at least in terms of the people dying.

No, not Really.

You misunderstood what I wrote. We will NOT have "...sufficient tests, contact tracing, filter masks, and washable cloth gloves,..." in the next few weeks.

I was referring to the hypothetical Fall recurrence. Right now, we need to practice social isolation in order to avoid the need for additional respirator capacity.

Or more appropriately, far too late to avoid the need for additional respirator capacity. Sad that we didn't spend the last 6 weeks in preparation for what is likely to occur, based on the experience of several other countries.

Or more appropriately, far too late to avoid the need for additional respirator(sic) capacity.

I had a typographical error in my last sentence. I wrote: "Right now, we need to practice social isolation in order to avoid the need for additional respirator capacity."

I meant "Right now, we need to practice social isolation in order to avoid the need for additional ventilator capacity.

We've got to be careful of standard US medical bureaucracy think in the face of death. How long does it take learn to tend a ventilator? A lot, lot, lot less than it does to become an RN.

Yeah, at this point I think our best option to throw billions at some sort of patient treatment capacity surge. Some of the social distancing measures will likely be heeded for months, maybe over a year. But at some point businesses are going to start defying orders to close.

Many people might start believing that the number of livelihoods at stake are worth more than the 1-2 million people who directly die because of COVID-19.

Many people might start believing that the number of livelihoods at stake are worth more than the 1-2 million people who directly die because of COVID-19.

I don't think there's even the slightest chance of 1-2 million people dying in the U.S. from COVID-19.

Agree. We need to slow the spread and use the time building up: 1) testing capacity, 2) ventilators, 3) masks, 4) beds, 5) other. I am not sure how much progress we are making on these initiatives, but I hope we are moving fast.

We also need time to develop a treatment flow system (like the Chinese did) to remove suspected cases from the population into temporary quarantine locations, then to centralized quarantine locations (if confirmed), and finally to hospitals if patient conditions require.

We also need time for businesses to develop operating plans. Staggered hours (when possible), health checks every morning (with temperature checks), regular sanitation/cleaning of workspaces, zero tolerance of coming to work sick. This is already done, in many cases.

We cannot shut down the economy the way San Francisco has for any extended length of time. Growth is important! Also, we may need things like hospital equipment, food processing equipment, transportation equipment. These things are made from many distributed components. If more places follow the lead of San Fran, with shelter in place mandates, we're going to run into serious shortages of this important stuff, and it will be hard to predict what we're running out of. Not to mention that we won't be able to buy Teslas.

Basically, we need to practice fairly extreme social distancing for a relatively short period of time, and then begin to relax the conditions somewhat once society is ready for care, case tracking, and preventative maintenance.

I suspect the treatment flow system is the most critical.

And yes, total shutdown is ridiculous. It can't last for more than a week or so. Find a level of curtailment that shows up on the numbers and can be sustained for a period of 60 days.

"total shutdown is ridiculous. It can't last for more than a week or so"

I would point out that there have been many cases in the past century when people had to stay in for extended periods of time, e.g. during WWII, in occupied Europe, in places where war was actively being waged etc. Not to mention the 1918 Spanish Flu where many places enforced various types of "stay at home" ordinances.

The fact that it never happened in the US is not really a valid argument in favor of this point.

Steve, I wish I shared your optimism but our health care system is good only for the top 10% but is a sh*thole for everybody else. It's safer to hole up than to depend on future treatment capacity that might not arrive.

He always likes a chance to show what letting the good times roll means, as the top 10% are by definition upright citizens who deserve all the benefits the state can possibly provide.

Exactly right. You cannot stop the spread of the virus, you can only slow it down to mitigate health system overload while iteratively scaling up diagnosis and treatment capacity. The time delay and unclear transfer function between action and consequences practically guarantees both over- and under-reactions leading to an "epidemic yoyo." It is a real shock so there is no perfect solution, only unpleasant tradeoffs.

It's quite possible to stop the spread of the virus. Australia could have but didn't. Australia is still capable of doing so, but so far our leadership has decided not to. If we decided to contain it, all we have to do is follow the methods used in China and South Korea that have proved effective. As we are in a far better position than China to contain it, success would be likely if we tried it.

Australia isn't China or South Korea
Once the economic consequences of pursuing such a policy become evident herding immunity will make sense
If the choice is economic suicide that impoverishes millions or adapting to a virus that kills less than 2% of those who contract it overwhelming majority of them elderly I would pick the former

We're feeling relieved about China's economic situation. Coke, thermal coal, iron ore... all at around the same price as before they clamped down on the virus.

"If the choice is economic suicide that impoverishes millions or adapting to a virus that kills less than 2% of those who contract it overwhelming majority of them elderly I would pick the former"

Current death rate in Italy is 8%, and Iran is officially at nearly 7%.

Trust the CCP's numbers the same way you'd treat the estimated number of lovers of a girl you meet at a bar on a Tuesday night.

These death rates don't take the dunkelziffer (figure of untested cases) into account, which at current (impossible) estimates is between 5 -10 times larger...

That's roughly what would be expected based on figures from China. About 12% of China is over 65. About 22% of Italy is over 65.

At the moment I'm putting more trust in China than you.

Most countries have death rates around 1-2% including South Korea and most of Northern Europe. The death rate in Germany is under 0.5%. Italy and Iran are outliers (and Iran most likely has many unreported cases).

The death rate in Germany is statistically questionable given the surrounding countries. Switzerland, France, Netherlands, Belgium and Italy are all neighboring countries and have a far higher rate. It's likely that the German death numbers are significantly understated.

Austria and Czech are reporting lower rates, but they are clearly at the beginning of the infection cycle and that's probably just lag.

The death rate should be low initially until the health system is overrun, then it climbs as people shift from survivable with treatment to nothing we can do. And if health workers start to die how do you replace them?

The numbers for Germany are at variance with Switzerland, France, Netherlands, Belgium and Italy.

To be specific, the more typical numbers are a CFR in the 1-3% range, but potentially going substantially higher if the medical system gets stressed.

South Korea remains a misleading example, since the country had only one infection vector to track. As for emulating China, I'm trying to imagine Crikey's reaction if the government he opposes were to have forcibly sealed Australians into their homes at gunpoint.

Is there evidence that worked and wasn't counter productive? A large number of people fled Hubei because they didn't trust the Chinese government because they do things like that.

In terms of vaccination, the concept is useful. Apart from vaccination, herd immunity to a circulating contagious disease is not particularly useful. Resistance and rate of spread are relevant. Measles is not a disease we ever developed herd immunity too - and clearly do not have as soon as vaccination rates drop a few percent. Further, it is not established how easy re-infection is, though it has occurred in well documented cases.

Italy has yet to yoyo - having so many deaths, dayafter day, makes it easier to focus on what is actually happening.

This is all very obvious stuff, I am surprised and a little disturbed intelligent people feel compelled to call it out. A path through hell is different than actually walking it.

Can we please focus on RCTs to test hypotheses regarding immunity?? That's the biggest uncertainty at this point, and without that answered we cannot even begin to speculate what the optimal policy going forward might be.

The UK government crumbled like complacent softies after only a few days.

It seems that the best way to avoid a possible epidemic yoyo is developing a vaccine sooner rather than later to create actual herd immunity.

I'm not following why the epidemic yo-yo is necessarily bad. Doesn't it imply lower peaks than doing nothing and less hit to GDP than shutting everything down for a year?

At this point, I think we need a new variant of Boxes quote: "All policies are awful, but some are more horrible than others"

The Chinese government is a brutally pragmatic one. If that means shutting down a whole state, they will do it. They don't waste precious time on "herd immunity" and dawdle with this too-clever-by-half stuff. Pragmatics beats pipe dreams 100% of the time. I'm disappointed in the quality of thinking in the West.

However if recovered people are indeed immune, it's possible Chinese "pragmatism" was overreaction. Too few people have been infected, thus too few people have immunity. They shut down their entire economy with the result that now less than 0.1% of their population is immune.

The wisdom of any country's decision is determined by your belief of how immune recovered patients are.

"However if recovered people are indeed immune, it's possible Chinese "pragmatism" was overreaction."

Anything is possible but that's a big "if". You make decisions using the best information you have. You can play armchair official after the fact but that's not how decisions get made going forward.

You assume that immunity is the goal but that might not be possible. There's plenty of viruses we don't have immunity for like herpes or HIV. Measles and smallpox we do have immunity for but through mass vaccination not through direct exposure like the herd immunity crowd is calling for. SARS, which COVID-19 is related to, doesn't have a vaccine or treatment. SARS came out in 2003, it's now 2020 and still nothing.

A decision maker can't bet on a cure that doesn't exist yet so containment, which has a proven track record, is the key to crisis management. The SARS epidemic of 2003 is the closest thing we have to the today's pandemic and China never developed immunity to it. It was contained and died out on its own.


Those who peddle herd immunity need to be spayed and neutered for the good of the species.

Well, I'd hate to know what Harold the Harsh has in store for those peddlers.

He keeps using that word, i don't think it means what he thinks it means..

How about quarantine and social distance buy time for a vaccine?
But I guess Tyler is ok letting more deaths as to maximize GDP, which is apparently helps everybody, well, except those dead folks, as long as they are not family or friends.

SARS-CoV-2 (the Covid-19 virus) is a close relative of the SARS virus, where, after 18 months of testing, no vaccine was ever found. Also recall HIV/AIDS and vaccine after countless decades of trying.

It was not found not because it was too hard to find it but because nobody cared anymore about a vaccine. The virus was extinguished.

The virus sat dormant in some asymptomatic carriers (could be human, could also be bats or civets) and then mutated into the main villain of this season COVID-19. We should have worked harder to find a vaccine.

Good job taking a complex problem and distilling it down to a personal attack on Tyler. You do realize that if we shut down everything for a long period of time, we're going to run out of food, fuel, hospital equipment, ambulances (and other transportation), plus many other things that we NEED (not want).

As I said above, we need to practice strict social distancing for a period of time to give us all time to prepare. Right now, we should certainly err on the side of stay home. But we can't "shut down" for 9-12 months. That will cause many more deaths.

It's astonishing to me that literate people are still talking about infection-borne herd immunity (as opposed to vaccine induced herd immunity, which is where the concept comes from). Assume the immune response is robust and lasting. Assume you can largely protect most of the high-risk populations to some extent while among everyone else you get the 80% illness rate you'd need (many guess 60% immunity in the population, but since we're trying to exclude the high-risk populations, a higher percentage of low-risk need to get sick). What does that look like in the real world?

Italy was overwhelmed by the time they had about 20k confirmed cases. Assume they didn't know about 90% of the cases because symptoms were too mild. So 200k cases. Assume all those people were only in Lombardy, population 10 million. So at a population infection rate of 2% the medical system began to collapse.

Hubei has a population of close to 60 million. What did they have, 60k confirmed cases? Assume it was 600k cases. 1% of the population, and they had to build entire hospitals in a week.

The whole reason people bring up this ludicrous plan is to "get it over with so it doesn't linger." We already know that there's no health care system in the world that can readily handle having even 1% of their population infected at the same time. We could devote 50% of the world's resources to beefing up the world's medical systems, so that every country could handle 2% of their population being infected. And then we'd have to stage a slow roll out of the infection, giving people time to recovery from each wave (about a month, minimum). Imagine managing to get 1% of the population newly infected per month, month after month, for 60-80 months. Imagine replacing the health care workers who are dying and becoming disabled as they many the front lines.

Does that seem reasonable to anyone? Is this really the best all these smart, ignorant people can come up with?

The optimal response to the virus hinges on questions that are not yet fully answered:

* does the virus spread through aerosols, or by droplets?

* some research suggests a chunk of population is simply immune to it. Is it 20%? 50%? 80%?

* how well can the virus spread through a/pre-symptomatic carriers?

* how susceptible is it to UV rays and temperature?

* does the infection cause any long term damage? To the kids? To the asymptomatics? To the severe cases?

So far, it seems the optimal response would consist of (basically - being Taiwan):

* ban on large events, avoiding crowds

* ban on non-essential travel

* population-wide mandate on mask usage or head covers

* large-scale testing, including random population sampling

* keeping proper hygiene and personal distance

* otherwise go on about the life as usual

*Droplet. This doesn't spread like the measles.
*What research? An Israeli chemist who doesn't understand virology?
*It spreads just fine. Not as much viral shedding as symptomatic people, but it's clearly been happening.
*It appears to be about as susceptible as similar viruses. It's not a fungus.
*Yes, lung fibrosis and damage to liver and kidneys is being reported in more severe cases. Asymptomatics very unlikely to have long term damage.

Large scale testing and tracking will go a long way to our improving our understanding. But this is going to take a while. Agree with your other notes.

+1, we are working in the middle of a crisis and you rarely have all the information to make a perfect decision in a timely manner.

The herd immunity people are the anti-vaxxers of the moment. They tend to be well-off, college educated and unfortunately disconnected from reality in a really naive way. Herd immunity might work as stated if this was just a serious flu or cold season. But it doesn't work for just about anything else unless there is mass vaccination which for COVID-19 isn't there yet.

You mentioned that they just want to "get it over with" and that's true. They are fundamentally an impatient lot. There is so much that can be done with a few convenient swipes and clicks that the thought of having to wait weeks to months could easily overwhelm their ability to endure.

Kevin, Keep commenting. I really like your access to FACT and your presentation of EVIDENCE. I know it takes time to write this, but keep it up.

You are providing a public good.

@W, +1. @Bill, thank you!

These are not two extreme cases. A more extreme case not considered above is to assume that we just live that way from now on, with social distancing, online communication, etc. I see at least two good points in that scenario:
- investment into the epidemic-time infrastructure will not be wasted;
- we will be one step closer - in terms of social interaction and possibly social organization - to Martian dome-cities.

I am glad that the comments above take this very seriously, but for anyone who missed it:

"How many is 4 million people? It's more Americans than have died all at once from anything, ever. It's the population of Los Angeles. It's 4 times the number of Americans who died in the Civil War...on both sides combined. It's two-thirds as many people as died in the Holocaust."


Keep well people, and keep calm. Netflix, chill, and let's get those checks in the mail for the financially insecure.

Thoughts and prayers to those suffering.

"How many is 4 million people?"

I have a bet with David Henderson on Econlog that less than 40,000 will die in the U.S. from COVID-19 in 2020. Where does the 4 million come from?

Regarding the Civil War versus COVID-19: I'm too lazy to calculate the average value (and I'd need to find more data), but from what I've seen, the median age of those who have died of COVID-19 in the U.S. has been over 60 years old. So that's like all the people in the Civil War who *lived* through the war, and simply died of what was then "old age" around the time of WWI.

The guy shows his math in that thread.

Note that it was an uncontrolled growth scenario, and not a bet about what what we will get.

The guy shows his math in that thread.

I've got some math, too. The moon has one-sixth earth's gravity. So if I played basketball on the moon, I could dunk the ball. With two hands, even. I think his math has about as much relevance as that.

Here, I think, is some much more realistic math:


Even in perfect conditions for the virus on the Diamond Princess cruise ship, only 17% even contracted the disease. In Wuhan, China, where the disease originated (as far as we know), the infection rate was only 3%.

The ultimate rate in the U.S. must certainly be lower than both of these cases, but assume 1-3%.

Covid-19 has had a fatality rate of about 0.5% in China. Let’s say it turns out to be a little higher in China, perhaps 0.6% or 0.7%. However, it should be lower in this country due to our better medical system. Let’s assume a range of 0.2% to 0.5%.

With 330 million people, I expect the 2020 death rate from Covid-19 to be between 6,600 (lower infection rate and lower death rate) and 49,500 (higher infection and death rates) deaths, or roughly equivalent to a regular flu season, in which about 40,000 Americans die each year.

"Even in perfect conditions for the virus on the Diamond Princess cruise ship, only 17% even contracted the disease."

I don't want to dip into this preposterous reasoning, but I'll note that these people were under a supervised medical quarantine and locked in their rooms. So while there were errors, it was hardly perfect conditions for spread. This thing is spectacularly intolerant of small mistakes.

The WHO thinks the death rate is 3.4%. We've been dealing with this for five months now and if you divide deaths by cases, you get 3.4%, suggesting that if anything, 3.4% is an underestimate.

I don't want to dip into this preposterous reasoning, but I'll note that these people were under a supervised medical quarantine and locked in their rooms.

Oh, really? They were locked in their rooms during the entire cruise? Bummer!

And how did the crew serve the passengers, with the crew locked in *their* rooms? Or did the crew roam around the ship (even crew who were infected)?

And speaking of the crew, where did they eat their meals...in their rooms? Or in a common area?

This thing is spectacularly intolerant of small mistakes.

Care to take some action where your mouth is, Lord Action? The CDC estimates a typical flu season has resulted in 12,000-61,000 deaths since 2010. That would be an average of approximately 37,000 deaths per year.


If this "thing is spectacularly intolerant of small mistakes" you must expect that there will be many multiples---perhaps a factor of 10?--times as many deaths from COVID-19 this year. Or do you think the U.S. hasn't make any small mistakes (of which "this thing" is allegedly "spectacularly intolerant")?

How would you like to bet say $40 that more than 400,000 people in the U.S. will be killed by COVID-19 this year. (That's a factor of 10 less than 4 million...so you ought to be pretty confident the deaths will be at least 400,000 in the U.S. this year.) (Or we could the time all the way to March 2021 if you'd like.)

My proposed death counter would be the Johns Hopkins website:


How about it? You wanna bet? Are you truly a man--a Lord, even!--of Action? Or just words?

Friend, calm down. I'm not questioning your manhood.

The vast majority of the spread on Diamond Princess occurred after they were under lock and key. It wasn't perfect, but you'd expect vastly less spread to have occurred than in an equivalently populated town not taking those measures. So it's not evidence of immunity.

Friend, calm down.

My friends don't write things to me like, "I don't want to dip into this preposterous reasoning..."

I'm not questioning your manhood.

No, I'm questioning your willingness to take action, "Lord Action." You wrote that you didn't want to "...dip into this preposterous reasoning..." From that, I infer you think you're pretty hot stuff. I'm wondering if you'd care to back your high opinion of yourself with a little ($40) wager? How about it? You feelin' lucky?

"The vast majority of the spread on Diamond Princess occurred after they were under lock and key."

1) You you know this how?
2) And what do you think was the "lock and key" situation?
a) Do you think that couples sharing adjacent rooms with balconies never stood out on those balconies and talked to each other?
b) Do you think there was never a situation where both infected and uninfected passengers from different rooms were in places in close contact in the same room (such as for medical exams)?
c) Do you think there was never a situation when, during feeding rounds, adjacent and cross-hallway doors were open? How about a situation wherein the people in rooms with balconies opened their doors to let "fresh air" from their rooms blow into the rooms of the poor devils stuck with completely windowless rooms?
d) When the aforementioned poor devils in the windowless rooms were allowed to get out and wander on deck for short periods of time, did they all come out of their rooms at once, and walk down the hall together?

It wasn't perfect, but you'd expect vastly less spread to have occurred than in an equivalently populated town not taking those measures.

You write, "you'd expect" but you don't know what I would expect. In fact, I expect exactly the opposite. What you mean of course mean is that you would expect that. But what do you know about the two situations? Do you know anything about, for example, what sort of precautions were taken when the crew at together in a common room, or brought food to the passengers? For example, did the crew even wear gloves each time they put food on plates and took the plates to passengers? What is the HVAC system on the Diamond Princess like? Is it possible for the HVAC system to transmit disease from one room to another?

Once the medical quarantine went into effect, all the passengers were locked in their rooms.

The ship has about 3700 passengers and 1000 crew. Letting the passengers mingle freely with half the crew gives you 17 million vector (4200 squared).

When you lock the passengers in their rooms, and have half the crew mixing freely feeding the rest of the crew and those passengers, you have about 250 thousand vectors (500 squared + 3700 vectors from the food service + 3700 vectors in the guest-to-guest sharing a room).

The deaths to this depend on our response. If people are convinced it's deadly and respond appropriately, we will have fewer deaths. If people believe it's no big deal, we will have lots of deaths.

Once the medical quarantine went into effect, all the passengers were locked in their rooms.

No, that's simply not true...at least in the sense that, "They never came out of their rooms nor opened their doors."

1) Passengers in balcony rooms were routinely going onto their balconies and talking to other passengers in adjacent balcony rooms.
2) Passengers were taken out of the rooms, and even clearly sick passengers from some rooms stood with passengers from other rooms who were not sick.
3) When food time rolled around, doors from balcony rooms would be opened up at the same time doors from interior windowless rooms were open, in order to let in "fresh air" to the rooms across the hall:

The disastrous Diamond Princess "quarantine"

"They never came out of their rooms nor opened their doors."

If this is the standard, it's pretty silly. Obviously the passengers were being fed and not by sliding single slices of cheese under the door.,

"1) Passengers in balcony rooms were routinely going onto their balconies and talking to other passengers in adjacent balcony rooms."

This is shockingly little contact. There is a barrier between the two balconies, and they are further outdoors where the germs very quickly dissipate into the air. The vectors here are minimal.

"and even clearly sick passengers from some rooms stood"

That sucks.

"doors from balcony rooms would be opened up at the same time doors from interior windowless rooms were open, in order to let in "fresh air" to the rooms across the hall"

Good. The coronavirus isn't mustard gas.

Quarantining on a boat sucks -- much better to do it on land -- but it still slows the spread.

Quarantining on a boat sucks -- much better to do it on land -- but it still slows the spread.

Have you actually read anything about the "quarantine"? If so, what?

What is your background?

Dr. Amesh Adalja told "Live Science" (well before the "quarantine" low point):


The ship quarantine, which has been going on for over a week, has already led to a major increase in the spread of the virus, even extending to a quarantine officer, he told Live Science. "The virus is being served up more and more victims by the Japanese government so long as this quarantine lasts," Adalja said.

Do you think Dr. Adalja doesn't know what he's talking about?

"How about it? You wanna bet? Are you truly a man--a Lord, even!--of Action? Or just words?"

You make a reasonable case, but this bit just comes off as obnoxious.

You make a reasonable case, but this bit just comes off as obnoxious.

I don't take well to people writing that I have "preposterous reasoning" that they don't want to "dip into." Especially when they hide behind a false name, and provide no evidence they have any expertise in the matters under discussion.

They were actually not locked in their rooms.

They were held under conditions that medical personnel thought would contain the spread. Conditions that medical personnel called "quarantine". I get that it didn't work, but it was a condition of limited movement and contact. So we can't use it as evidence for an immunity because there's a perfectly normal reason to explain why "only" 17% of the people came down with the disease in the span of a couple of weeks.

I get that it didn't work, but it was a condition of limited movement and contact.

What is your expertise in this matter? What articles have you read about the "quarantine"? What is your background?

Define "all at once" and "from anything". The current mortality in the United States is something like three million people each year. Most of the leading causes of death are, or would be, co-morbid with this virus, not least of which is "natural causes", i.e., old age. I'm not trying to downplay these deaths but it would be more helpful to look at the effects on average life expectancy.

It seems a cliche at this point, but it bears repeating that the flu kills up to 50,000 people in the United States each year. (How many is that? It's more than the number killed by traffic accidents each year. It's over twelve times the number who died in the 9/11 attacks...) Yet we don't enact, say, 1/80th of the controls we have for this pandemic, that, even given the exponential growth of the number infected, still has less than 7,000 cases in the US as of yesterday.

1940's Shark probably railed against FDR fighting the Nazis because "More Americans died from falling off their horse carriage than that evil German ever did to us Americans!"

Did you ever console a friend by telling them: Look you lost your loved one but its just one death there are more people that die from lung cancer each year from smoking cigarettes!

And did you ever wonder if it even worked ....

I'd go on but its shooting fish in a barrel at this point. Sorry nuking sharks from low earth orbit and killing it with plutonium laced fire ....

His guy tells him "The onslaught of the Chinese Virus is not your fault! Will be stronger than ever!"

Maybe that will get him on board. Xenophobia for the win.

Oh no. Mustn't be xenophobic! A little smidge of xenophobia would have kept the Wuhan flu off American shores altogether, but That's Not Who We Are™. Four million people may die instead, but by God we'll let 40 million die before someone can accuse us of xenophobia. Plus, it's as close as we've ever come to taking his guy down!

I'm not sure I should even try, but you do realize that sufficient cases came from Italy to kick-off exponential growth here, right?

This pandemic started in Wuhan China, but there was so much air traffic in and out, that it was everywhere in a couple weeks.

We would have had to close all borders with all countries even earlier, and none of this accepting back American citizens stuff. Not without 100% quarantine of those thousands and thousands of people.

We've been way to slow closing borders. In particular we dragged our feet on Europe despite that move being obviously necessary.

Right now we should close the border to non-US citizens, and we should tell US citizens they have two weeks to come home and directly into quarantine, or they're stuck too.

We should also have the NYC area shelter in place. What is it with those guys? It's the biggest problem in the US and rushing headlong towards Italy.

A domestic air travel ban also makes sense, although there I could live if it rolled out gradually, starting with CA, NY, WA, and maybe MA.

We did a horrible job closing our borders, as seen by the Europe closure. Poor presidential communication and a lack of preparation at the designated airports led to a large number of people being confined in an enclosed space among a number of carriers of covid-19. And after leaving the terminal, there was not even an attempt to impose isolation. This is crazy.

- When I asked a security guard about the other lines, he told me they were for people with a confirmed corona diagnosis. There was no separation for this group — no plastic sheets, not even a bit of distance. When your line snaked to the left, you were inches away from the infected.

And this - Some of the agents were asking people to use the fingerprint screen — all fingers, then the thumbs. Mine didn’t, but I watched the adjoining one and was astounded to see that the screen was not wiped, sprayed or in any way sanitized between individuals, or indeed at all during the hour I had it in my line of sight. My agent asked me how I felt (the true answer would have been upset by your colossal ineptitude) and if I had been to China or Italy. (I had not.)

This is the true cherry on top of a miasma of covid-19 particles - That was it. No instruction to maintain a two-week self-quarantine. No phone number to call if I felt symptoms — standard in Europe for several weeks. After being immersed in our three-hour virus incubator, we were unleashed on the American public, free to mingle. This had to be going on all day and further into the night, as flights kept landing and the immigration hall kept filling with new passengers.

Funnier in the sort of way black humour way that leads to coughing fits, anyone arriving from the UK was originally allowed entry without restrictions. Even if a passenger had started in Milan, flew to London, then flew to the U.S.

You have a pandemic with the president that won the Electoral College, regardless of how much heat they can stand in the kitchen.

Oh boo hoo. Some people had to stand in a line. The infamy!

We should have done it with zero warning and then arranged a quarantined repatriation, sure. But the complaints about how the service was less than first class fall a little flat. The error was on the side of doing too little not doing too much.

We created the perfect environment to spread covid-19 at multiple airports, then simply let everybody walk out without even talking about self-isolation, much less actually isolating them.

Boo-hoo, who cares about a bunch of people spreading covid-19? Not the Trump Administration, clearly.

I'm the one saying the ban could have been more effective. You're the one saying we shouldn't have had a ban. Get your story straight.

One of the worst things for border control is a surge. The confusing announcement that scared people into thinking they would be stuck outside the country unless they panicked made them all show up at once.

The tests should have been put into place first. Then an encouragement of citizens to come home. Then saying "okay, get home soon or else."

I've seen so many levels of government put out releases and not worry about whether it's going to cause panic or even if it's correct (Orange County yesterday seemed to be telling people to shut everything down and stay home, and then said "no we didn't mean to shelter in place") .

They should be announcing travel bans "effective immediately" and then implementing a repatriation mechanism that involves quarantine. Advance notice gives the people who want to avoid quarantine a way around it.

It isn't a question of closing or not closing borders, it is a question of competence. Something that at least some people seem to understand.

The way the European travel stop was handled will result in an increase - not decrease - of American cases in the next couple of weeks. No quarantine of returning passengers, mingling of confirmed infected people and non-infected people, shared touch surfaces - the list is surprisingly long, and saying it was hasty is being far too generous.

Many of the American cases were caused by American citizens on American cruise line ships bringing the disease with them.

If only we had a few more sharks with lasers, they could have all been sunk before having the numbers double because of one ship that wasn’t our fault.

While there were fairly large numbers, especially from the early days before it became obvious how deadly this thing is, at least we knew where they all were and did a reasonably good job containing them.

It's the air travel that was the problem, with people coming in in ones and twos without any time to recognize symptoms. We should have shut down travel from Europe far sooner than we did.

From Martch 14 - At least six Americans infected aboard the Asara returned to Maryland, according to Gov. Larry Hogan (R), potentially seeding their communities with the virus. Twelve others have reportedly tested positive in the Houston area. The Centers for Disease Control and Prevention is contacting dozens of other Americans who were aboard the Asara in February to test them or ask them to self-quarantine.

The whole Post article is considerably more detailed, and plenty of blame to spread around, but the article as makes clear we did anything but a good job of knowing where they all were and did not do a reasonably good job containing them.

Sharks cannot live in fresh water, so this cruise ship would have ended up being a leaker regardless.

The Centers for Disease Control and Prevention is advising travelers who have recently disembarked from a cruise to self isolate.

“Stay home for 14 days from the time you disembark, practice social distancing and monitor your health,” the agency said Tuesday. “Social distancing means staying out of crowded places, avoiding group gatherings, and maintaining distance (approximately six feet or two meters) from others when possible.”

Shame that the U.S. still seems to lack anything resembling adequate testing capacity involving the ten of thousands of potential cruise ship carriers to at least confirm who is infected.

Meanwhile, China shipped 2,300 cases of masks to Italy, inscribed, “We are waves from the same sea, leaves from the same tree, flowers from the same garden.”

Probably infected...

"We would take a big hit to gdp but not really stop the spread of the virus."

The "economy" is people selling and buying goods and services. Consumers stuck for several years with health services debt and dead people don't buy anything =)

Do you want food? Then I guess we better keep running food processing plants. Do you want those food processing plants to run, then we need spare parts. Do you want someone to make those spare parts? Then we need other factories to operate. Do you want to move food and critical goods from one place to another? Then we need transportation.

If you follow these chains, you will see pretty quickly that we can't just shut everything down. What we can do is pause (shut most things down for a period of time). We can use this pause to slow the spread of the virus and prepare to manage it. I don't know exactly how long to pause and how to do it, but right now I would lean towards having less economic activity rather than more.

We can't just focus on the long term. There is a cost in lives to stopping everything.


"it is very clear that capacity expansion has to be the priority now" !!!

Reading Tyler's post and this article from Robby Soave (https://reason.com/2020/03/18/coronavirus-quarantine-imperial-college-london-covid-19/) then it became pretty clear to me that dramatic expansion is the only feasible option for both near and long term targeting of this.

I agree with the time inconsistency assertion: we have a dynamic system of innovation, and buying time saves both present lives and future lives.

If you want to test your hypotheses, you might want to first look at Our World In Data website which collects data on this and other public policy matters: https://ourworldindata.org/coronavirus

What is interesting is the doubling growth rates of various countries which you can match against their policies or lack thereof. My criticism of Hanson (and I haven't read it but only the description) would be that he probably does not match data of variants to the real world, I too can create models, but why would I if I could look at real data and play with that.

This is a really good source of data. Go to the table Total tests performed by million people.

United States is below Vietnam, Armenia, and above the Philippines.

I think your point is the most important. Buy time. There is some known, but very little about this thing. As Dr. Birx said two days ago about her people, they all have their favorite models.

The 15 day edict on limiting contact was described by Birx and Fauci as a trial to see whether the curves could be changed. Time gives a chance for a trial for the tuberculosis treatment to be tested. A couple other things, one from Japan. A very interesting tidbit was the weekly updates of data from Europe simply due to the people who know the most are overwhelmed at the moment.

I don't think it is unreasonable to suggest that a Shelter in Place order in New York would stand in practice for not a long time. Extreme measures may be counter productive. But this again can be tried; what happens in San Francisco is a natural experiment.

Time also reduces the initial size and it reduces the replication rate.

And, don't there is heterogeneity of infection starting points across geography, so adopting a policy one place may have an even bigger effect another place, so that the virus doesn't get too far along.

What I think will be interesting is how firms will in the future disburse their intellectual property assets (people) across geography rather than consolidating them in one area.

"What I think will be interesting is how firms will in the future disburse their intellectual property assets (people) across geography rather than consolidating them in one area."

+1, good point.

I'm also curious to what happens to the Movie theater chains. Will they ever recover? They were already struggling to compete against, the Studios charging very high rates (90%) of ticket prices for the first two+ weeks, big screen 4K TVs, streaming services, Red Box, etc

It only has to work until there's a cure or vaccine or widespread accurate test.

@Christine, right you are but a cure or vaccine may be far off, recall coronavirus is like the common cold virus, and "there's still no cure for the common cold" as a pre-internet meme us Boomers can recall used to say. As for testing, I guess in theory we can have segregated societies between those that have Covid-19 and those that do not, kind of like leper colonies.

Oh look, Taleb wading into a discussion in a subject area about which he knows nothing and, surprise surprise, the experts in the area have made fundamental mistakes because they didn't use fat-tailed distributions!  Woah!

And Robin Hanson has a spreadsheet with some counter-intuitive results!  Nice job, chrome-dome.

In other shocking news: the sun rose this morning, the Pope woke catholic and the MR comment section is a veritable shitshow

Taleb can't even do basic reading. He states "The model they use appears to be in the general class of SIR differential equations used in epidemiology"

Page 4 of the paper states "We modified an individual-based simulation model developed to support pandemic influenza planning"

I mean, for fucks sake.

Tyler maybe just stick to pissing Peter Thiel's money away on pointless rubbish?

Well, we can all take comfort that there are probably as many models out there as cases of the influenza. It keeps people off the street.

We lost 3% of the population to Swine flu in 2019 yet mankind survived and flourished and life expectancy increased by 20 years since that epidemic . It is the wealth created that allows for the improvement in healthcare. If we were really concerned about lives saved and didn't resort to the snarky top 1 or 10% comments we would continue to create wealth instead of destroying it as we are currently doing and allocate more of it to medical care in impoverished nations. It would save many millions of more lives than we are going to do with current policies. Very few people on this board and have exhibited any inclination to do that in a meaningful way.

"Let's kill our mums" is not a great sell for the general population, who generally have pretty good relationships with said mums.

On the impoverished billions, they are seeing a lot of improvement. The tension there is less with our mums than our tariffs.

Swine flu (H1N1) was in 2009, not 2019.

H1N1 had a mortality rate of 0.01–0.03% similar to influenza while COVID-19 was much more lethal with a range from 2.3–3.4%. That's a couple orders of magnitude in death percentage-wise. Swine flu is not as a infectious with an R0 of 1.2–1.6 compared to COVID-19 at a much larger R0 of 3.

In other words, the 2009 swine flu was just a more serious flu while COVID-19 is a more infectious, more deadly, and altogether more serious disease. It is fair to say that the total possible number of deaths would be much, much higher than 2009 had the governments not done anything and the numbers bear this out.

I'm not sure the comment about destroying wealth. How can more wealth be generated when everybody gets sick? If you are an American, the fastest way to destroy your wealth is to get sick as you've probably heard. When your downside is unlimited and your upside is limited, you don't have an economic incentive. So the first order of the day is to contain the thing to build confidence so people can feel safe to work.

Thanks for calling out this guy’s bad stats. I’m also surprised he would think that killing 3% of the population (on top of all other deaths) would be a small deal. Holy cow. That would be devastating. Your numbers on the actual death rate from the 2009 swine flu are accurate from what I’ve seen but deaths from the seasonal flu are actually much higher than that at 0.1%. So he’s right that the swine flu didn’t affect the economy much because it was a weak flu.

Good catch. You are correct that influenza has a much higher death rate than the swine flu at 0.1%. So that line should be corrected to say that swine flu is a "less serious flu".

Another note to add is that H1N1 has a vaccine while COVID-19 and its related virus SARS have no vaccines.


In all seriousness, it is plausible we see some sort of profound societal permanent change occur.

- When in prior histories did we see something as galvanizing and global? What proxies can you identify?
- How long did such events occur until large shifts in attitude and policy were embraced?

- For example - after 9/11 we saw widespread adjustment in law (patriot act), and a war that lasted for over a decade. That was a relatively 'rapid' event, although terrible. What might be expect this time?


Has anyone read Taleb’s critique of the Ferguson paper? I don’t understand the following sentence: “ The actual minimal number for resurgence is larger than 1 because (1) a significant percentage of those in close contact with confirmed cases are not infected, indeed only 5% of close contacts of infected individuals traced in China subsequently tested positive”

The number he’s talking about is R0 and it only counts people who are effectively infected, so it already incorporates the 5% ratio. If my interpretation is correct, this sentence makes no sense. Did I miss something?

Taleb didn't even get the type of model that the imperial researchers used correct.

Nothing he has to say on the issue is of value.

Very good point.

Maybe he is arguing that Rnought declines as people die. You do not interact with the dead. Or, at least, I don't, but if I die, and can, I may interact with Taleb.

He is not talking about Ro in that sentence. He is saying you don’t have to get the actual number of remaining infected cases to 1 or less. Even more than one case left ( say 3) may still result in the virus being extinguished. I have seen a plausible model where even 4 independent cases lead to only a 50 % probability of the virus restarting.

Got it, thanks!

" indeed only 5% of close contacts of infected individuals traced in China subsequently tested positive”"

That's probably true, but Taleb should clarify the 5% close contact number. It doesn't mean close contact as in live in the same house, it means close contact as in worked in the same office or traveled the same bus. IE if one of the contacts rode a bus with 60 people and only 3 other people got infected, then it's 5%.

I think Taleb is providing evidence why peer review prior to publication and public dissemination is valuable thing. If indeed he is wrong, he is doing damage by pissing in the swimming pool.

Now, if we could just have peer review of comments and posts, but, I guess the comment section provides some of that, although, I can humbly state, very few can match me as a peer.

One approach to this is as a resource allocation problem: what is the cheapest way to get R0 below 1? Working from home clearly helps achieve this, and you can easily see if people are going to offices or not, but it is not necessarily cheap, because a reduction in communication will reduce productivity. Both now and for the indefinite future, we need to know the effectiveness of measures which allow people to work together, such as virus-inhospitable coatings on doorknobs ?copper?, screening methods such as thermometers for fever - other biometrics? cold plasma or bright UV inside air conditioners?

Maybe we should look at this from two points of view:

On the supply side how can suppliers alter their product offering or distribution model, or what new products that could be created to fulfill a demand

On the demand side: I think it is perfectly rational for persons to factor in the costs of getting sick and dying from consumption of a good or service.
The problem is when it is an irrational decision, which comes from either a lack of information or misinformation, aka, market for lemons. So, the question is: how do we provide information on safety. Maybe, if we have more testing, you would interact with someone who has been or is constantly tested. And, they would interact with you. I have an app for that: I will be wearing a forward facing Phone that says: I was tested negative on (fill in the date) and if you have a similar forward facing phone with the same data, I will give you a hug.

Maybe we can also identify this as a time where we can improve the efficiency of the economy: a way to push stores to have an online presence, something they may have been delaying for awhile, or developing new products or services that do not involve in person presence, which, maybe as society ages, might be a new market.

Also, if you become more resilient to these events--and they will happen again--maybe you have improved long term economic efficiency.

Sometimes when an organism is attacked by a virus it becomes stronger and more resilient to the next attack.

Is it just me or is Brooks bizarrely missing the point. "When there are a lot of people who’ve had the disease and become temporarily immune," then social distancing is already less important for "the rest," right?

I generally think that David Brooks misses the point. I don't know how he remains employed other than pure Tokenism.

I think the Taleb critique makes sense. Ro is probably 2.2 at the beginning when there's no awareness of the virus, but is very sensitive to social distancing. The serial interval is not that short ( ~ 7 days) ( time between onset of illness in 2 subsequent generations). Strong measures of tracking and distancing for up to a month is all that's needed to extinguish it as we saw in China. You don't have to reduce it to one case, just isolate in pockets with small numbers ( < 3) and the battle is won. Good tracing is important. So a month of pain then ongoing tracing looking for resurgence.

I'll aTEST to that. And, not only can we test more, but we may be able to track better. Location aware apps provide location data that can identify where you are or have been. If you test positive, you can not only tell who you met with, but where you have been. If another person also tested positive, and their data revealed they were at the same place during their period of transmission, and neither of you identified the same contact, you might have identified the hot spot...maybe a local bar or restaurant.

If you are interested in the use of location data in marketing and marketing based on location, you might want to read the book: Tap: Unlocking the Mobile Economy by Prof. Anindya Ghose at NYU.

I feel like you are conflating lock downs with social distancing. Lock downs at the China extreme were effective,. Quickly dealing with limited vectors and prompt testing was how SK tamped down on it.

But social distancing isn't quite the same. It might work, but we won't be certain of its efficacy for at least a few more weeks.

How much of this is binary thinking? Do we need to fight "the war" on all fronts at the same time?

The game seems to be about limiting the transmission rates. Isolating/quarantine helps that. Controlled burns to remove potential infections will as well.

There are probably several good approaches to blending the control efforts here that will tend to push R0 towards or under 1.

I’m not going to recap this complex debate, which most of you already have some inkling of.

Why in God's name would you consult a pair of economics professors (one of whom publishes no research in economics) and a publicity-grifter finance professor on a question of epidemiology?

I will reiterate my suggestion the other day that business NOT be shutdown, but instead move to temperature checking at entrances. Thermometers are cheap, read more or less instantaneously, and are easy to come by. It won't catch every "asymtomatic" carrier, but it will catch the vast majority of infectious patients. And people can keep going to restaurants and working and earning wages and paying rent. The R0 will decline, maybe not to <1, but significantly.
Also, it's something the Chinese have continued doing even after restarting their economy and they have not see a resurgence yet.

The businesses I work with are already doing this and more. Still, I think we need some kind of pause. People who don't need to travel shouldn't. People who can work from home should. We may even need temporary lockdowns, like in SF. But these measure can only be temporary in nature.

Work from home and curtailed air and rail passenger transport are low-cost no-brainers. Cargo should cross borders, but people generally shouldn't.

I'd expect schools and factories to reopen long before this is over. I'd expect air travel limits to last until most everyone is vaccinated. I'd expect working from home will be the norm for almost all knowledge workers forever. I don't think we're ever going back on that one. It will be disruptive for the commercial real estate market.

Most of my colleagues work from home. But they are experienced. When I think of the start of my career, I would have progressed quite slowly had I not been able to physically interact with peers and mentors.

It's been a learning experience for my crew. We have an older set of senior managers who look on working-from-home with disfavor. 24 months from now, they'll know for sure if it's viable. We rely on hiring a lot of people straight out of undergraduate and graduate programs. Onboarding and training those people is my biggest long-term concern.

But I am beginning to think it's likely I will never go back to the office full time.

I can't see that ever really working . Many "knowledge workers" in tech work with development hardware platforms that they need a lab for. We don't all just write software - even if you write software, you are often writing it for an embedded platform that you can't take home.

"I can't see that ever really working ."

Well, in part because people fought quick travel bans on China, we no longer have a choice.

"Many "knowledge workers" in tech work with development hardware platforms that they need a lab for."

Yeah, well, that's going to be a challenge. Clearly if you work with a computer and a phone, it's easier.

But frankly, it's 2020. If you work with a computer and a phone and you still spend 40 minutes in a car getting to an expensive downtown office, something is wrong with your company. This is an opportunity to fix it.

"We don't all just write software - even if you write software, you are often writing it for an embedded platform that you can't take home."

True, I do. But on the other hand you don't have to be in the office every day, for the full day. Download and check out a panel/device, go home.

Inefficient if you aren't using it for the whole day. Also sometimes what you have to "check out" is not one device, but a whole bunch of them connected together, with a bunch of other equipment attached to them.

There are cheaper and less risky ways of helping small businesses.

Here are some that I’ve recommended to a local foundation to help restaurants and small stores in a neighborhood.

1. A grocery store currently delivers. They could deliver products of other merchants in the neighborhood to other merchants.

2. The city could limit parking in front of a store that has pick up options to 5 minutes. No parking in front of the store that has a pick up option. Get out of the car, pick up your stuff and go. Next car goes in, etc.

3. I’ve suggested that local merchants could also have a common pickup point...that all the merchants drop off orders at one place and the customers pick up all their orders at that one location.

4. The Foundation establish a prepaid coupon book for restaurants .

rather than "to other merchants" it should be" for other merchants to customers."

Throw in the suggestion for extended hours. Longer hours. I see places limiting hours which in practice means larger and denser crowds.

Change lunch and dinner hours. Spread it out, less crowded.

Our dentist was still doing check-ups for the kids last week (some have stopped). They had my wife wait in the car and texted her when it was their turn to come in.

Hazel, this has been tried in airports. It misses most patients who are infected but don't already know they are sick (it likely catches some who know they're sick but go out anyway). When someone is infected, it takes time to develop a temperature, and viral shedding is happening in this interval. And there are many cases where the elevated temperature was transient while the infection continued. You can keep reiterating it, but this scheme of taking temperatures instead of social isolation is known to be ineffective.

I could imagine a world where money pours into testing resources and we test substantially everyone, every week. With a real test, not a thermometer. You could restore a lot of functionality in that world while we wait for a vaccine. You'd be putting out fires a lot with local breakouts, and would need to be able to quickly shut down localities.

But you probably can't have air travel.

Temperature checks is the world China has moved into after getting the disease under control. We will need to do something like this at some point (or have enough tests for everyone like Lord Action says).

For now, we need to practice fairly extreme social distancing, but it can't last forever.

We just need to get a handle on who is infected and trace their contacts, put them into self-isolation. Once that's done we can move to temperature checks and restore normalcy.


"We estimated that 44% of transmission could occur before first symptoms of the index. Disease control measures should be adjusted to account for probable substantial pre-symptomatic transmission."

Temperature checks are clearly not a significant portion of the solution. As for "getting a handle on who is infected," we aren't close to a rapid/accurate/scalable test, and if we were, deploying that and getting the info back would be an overwhelming logistical challenge. And if you did that, too, what do you do with those freedom-loving folks who say "FU" to self-isolation?

"For now, we need to practice fairly extreme social distancing, but it can't last forever."

10-14 days before people get bored and/or broke

Yeah, if they try to keep this going longer than that there will be rioting. There are plenty of people out there who will think that this is all some smokescreen for Trump to become a dictator or something. (Would be the same on the right if it was a Democrat in power). You'll need cops on the street to STOP people from going to the beach, etc.

I'd like to see some links on that.

I'm not talking about testing people once in the airport and then letting them walk into a shopping mall 3 days later. I'm talking about testing people at the door of *every* mall, store, bar, or restaurant. The information I have heard is that asymtomatic/pre-symtomatic transmission is relatively rare. If people have had a fever they should already be self-quarantining for the most part. Yes, there will be exceptions, but it's not going to miss most infectious individuals. This will flatten the curve rather than create waves. You don't actually want the R0 to go to <1 until after most people have been infected. I.e. in China, the two week quarantine pushed R0 to 0.3 - they're destined to have a second wave sooner or later since there are plenty of susceptible individuals out there.

Asymptomatic and presymptomatic are not synonyms, I think.

Asymptomatic means "people who get the virus but never appear to be ill." The WHO has said they don't think those people exist, or are rare, so their death rates are likely accurate.

Presymptomatic means "has COVID-19 and would test positive, but doesn't have symptoms yet". Those people apparently go on to develop symptoms a large majority of the time. There is plenty of scientific debate as to whether they are contagious in the few days before they develop a fever and become obviously ill.

Yes, I know, in the interests of brevity I'd rather not have to detail that out every time.

It's an important difference, because sometimes people confusing the concepts use them to argue that the WHO's death rate studies are wrong. I may be mistaken, but I thought I'd seen you doing that. Forgive me if I'm incorrect.

I'm arguing people need to be consistent with respect to the death rate statistics. EITHER the death rate is closer to 3.4% and there are NOT a lot of asymtomatic carriers, OR there's a lot of asymtomatic carriers AND the death rate numbers are much lower. See JWatts below. Contradictory information out there on whether there are asymtomatic people who are infectious - and people keep making up inflated predictions of future deaths based on contradictory assumptions.

Posted a link above that close to half (44%) the spread is coming from people who have not shown symptoms yet. The WHO is incorrect about there not being asymptomatic cases. Multiple reports from the field have shown this. This has not been a question for a few weeks, though it is disappointing that many health organizations have been reluctant to seem alarmist about advertising this reality. In particular, children and young people tend to show negligible or no symptoms, and yet have been demonstrated to be active in spreading the virus. This is based on series of patients reported from China and South Korea. Sadly, instead of journalists and smart economists looking at the preprints that are headed for publication, they want to talk about price gouging on face masks and toilet paper runs.

The death rate will not be accurate until we have a clear idea of how many people have been actually infected. It will surely be much less than 3.4%, and much higher than the death rate of the typical influenza. It will be months before we have an accurate assessment of death rate for the whole population. That said, we have a very good idea of the death rate of different age groups and risk groups among those showing symptoms, who are the only people being tested in most countries. It's high enough be a serious cause for concern. 20% of family practice doctors in northern Italy have the virus, and several of them have already died. A much larger number of them have been hospitalized with serious symptoms.


Your link is discussing presymptomatic people, not people who never show symptoms. Presymptomatic spread is a big concern. But presymptomatic spread does nothing to mess up the death rate estimates.

As far as I can tell, the idea of truly asymptomatic spreaders is not well supported.

Also, it seems like outside of China the idea that young people show negligible to no symptoms is turning out not to be true. It may be that it wasn't true in China either, but it will take time to sort that out. The US is seeing 20% hospitalization rates among 20-44 year old and even 6% rates among children. I gather Europe is seeing something similar.

"I'd like to see some links on that."

"Asymptomatic and mildly symptomatic transmission are a major factor in transmission for Covid-19," said Dr. William Schaffner, a professor at Vanderbilt University School of Medicine and longtime adviser to the CDC. "They're going to be the drivers of spread in the community."


I dunno I keep seeing different statements from different experts.
One day you hear one guy theorizing about all the asymptomatic carriers, and the next day someones saying 98% of transmission is from people who are symptomatic.

Mildly symptomatic would probably include a low-grade fever. We just need to be meticulous about keeping anyone with ANY fever out of public establishments. Like even 99.3 degrees. Anything above 99 degrees - no entry.

Incidentally, when I had my weird viral bronchitis last june (which may or may not have been an ancestral strain of COVID-19 *ahem*) I had a low-grade fever of 99-100.5 or so that lasted a week. I did take aceteminophen to keep it down, and that might be an issue if someone knows they've got a fever and they take a fever reducer. But still, I suspect a lot of people will cooperate with "don't go anywhere if you have any kind of fever", and not take aceteminophen and go out anyway.

And even if a few people slip through the cracks, that's not the end of the world. Id rather have a low rate of on-going transmission than a indefinite shutdown of the economy.

Checks at airports are different than checks every time somebody enters a building. Get an airport check wrong and you let in somebody who may not be infectious now but may spread the virus days later. Get a building check wrong and you may catch the mistake the next day, and there might not have been much spread in the day you missed. Shutting down all of the offices reduces spread more, but if we can drive R0 below 1 without this we can stop the health crisis without causing an economic crisis.

Do any of these models take into account the cure we will discover in 4-6 weeks? That is the only resolution that does not end in total disaster. It must be our focus. Social distancing buys us time, but does not solve the problem.

The flu pandemic of 1968 killed over 100,000 Americans. Why didn't we have martial law and shut down the economy for it? Something is odd about the reaction now.


The difference is that the 1968 pandemic was bog standard influenza (H3N2 swine flu). A low fatality rate, relatively low rate of infection, standard flu symptoms. SARS-CoV-2 aka COVID-19 is not a flu but SARS (severe acute respiratory syndrome) with higher rate of infection, higher rate of fatality, and severe reduced lung capacity. So the sickness is much more severe, the chance of death is greater, and it is easier to get.

Somebody ran the numbers from early statistical sampling and the numbers were scary big and the facts on the ground in Hubei showed this wasn't going to be just a regular flu season. The world demographically has gotten older so this means a higher possible death count than any previous pandemic and increased globalization compared to the 1960s is another risk factor for infection.

The reaction is different now because the characteristics of the disease requires a different response. We don't have martial law, though travel is restricted, and this isn't strictly a government led shutdown. The NBA and Disneyland for example shut down before governments in the US made their moves. Corporations have an interest in protecting their workers and their reputation while governments have an interest in protecting their citizens so I believe the reaction was proportional though somewhat delayed.

It is more infectious than SARS but LESS deadly, not more. SARS killed like 15% of everyone it infects. Also 80% of the time COVID-19 causes mild symptoms that do not lead to pneumonia.

Perhaps we are more afraid of death now.

@Tim, no "data" or available computation and much lower volume and frequency media.

And not in a good way. Computation drives models, models drive predictions, predictions drive "We must do something", "We must do something" drives implementations of the only thing you can do in a short time, which is extreme social chilling.

I would guess that 'mitigation fatigue' is more relevant to older cohorts, whereas millennials will experience little change since one of their defining group characteristics seems to be the willingness to opt for digital, rather than physical, interaction with peers.
An example:
Fortnite in most cases requires social distance between participants, shuffleboard on the other hand...

What’s funny is that these sophisticated models don’t much value. A very basic approach calibrated to 2009 h1n1 revealed that 2.5 million deaths was plausible/likely under a do nothing senario.

Tiny modifications to the assumptions baked into the simulations cause large changes. The variances are huge. I don’t really give them much credence.

IMO, using a simple logistic model with a small window of step ahead forecasts (like 2-3 out) is more useful.

If the forecast are accurate, we are still on the same exponential trend. By constantly monitoring the estimates of the growth factors we can simply say, x, y, z polices seemed to have worked or not.

Do this for another month or so while we consider the options.

MR coronavirus coverage is been top notch.

There is no risk of yo-yo: in the western countries since the virus has spread over too large a group of people to be contained, specially given the lack of state capacity and social distancing culture as it exists in East Asia. In Japan for instance I think they will maintain present levels of social distancing until the vaccine in about 15 months or so. In Western countries the lack of effective enforcement of quarantine means that herd immunity will arrive in a few months.

From my daughter, the pathologist, to her dad:

. "Unfortunately our curve looks like Italy's. Also, since we have had so many testing difficulties, including limited numbers of tests, we really don't know how many people are affected. We will know a lot more in the next few days.

For now, take full precautions"

Too many people seem to think their opinion on this is somehow valuable and needs to be aired simply by virtue of their above average IQs. Economists, computer scientists, Nassim Taleb, etc. Years of training and experience in public health, epidemiology, virology, etc can't be made up for simply by being bright and intellectually self assured.

There are at least two costs to using a public platform to wade into topics without qualification in a time of crisis. The first is simply the cost of promulgating wrong information in a time of crisis. The second and may most impactful is decreasing the signal to noise ratio. We're inundated with information and each bit of wrong or not useful information competes with important and essential information for our attention and ultimately, our actions.

My suggestion for public intellectuals is to focus on 1) sharing the information of experts and 2) sharing their views limited narrowly to their areas of expertise.

+1 And have persons with knowledge like KevinK comment in response to misinformation.

Thank you Kevin.

It's hard work to clean the stables.

Maybe MR should have an epidemiologist or physician watch the comment section.

The bottom line is that “locking everyone up to bend the hospital admissions curve” might have to last for at least a year to really choke off the coronavirus.

The other important question involves drugs: you're no doubt already aware of remdesivir and hydroxychloroquine (and variants), as well as favipiravir (results just announced in China).

If none of these work then yes, the yoyo is possible.

For those who thing China did a great job handling this, I'd like to point out that if they hadn't kept it a secret for weeks, there would have been a lot fewer travelers returning from China to other parts of the world with the virus, before we all knew about it. The reason we have "community spread" not limited to people who recently traveled to China is probably because of that.

Those travelers would have returned anyway. Right now, knowing what we know, infectious cases are still traveling. Currently, the vast majority of new cases in China are people returning from other places. By your logic, that should not happen.

Also, early on there was no understanding of how infectious a person was without symptoms (erroneously believed to be nil, but now understood to be substantial). Regarding the time to self quarantine, we have good evidence that a 14 day quarantine only gets about 88% of cases before the quarantine is over. You need a 21 day quarantine to get up to 98%.

Expecting China to have locked the country down and not allowed air travel for either their own citizens or visitors from abroad seeking to return home, starting in January, is not a rational expectation. Even after seeing the events in China, most nations chose to do next to nothing even as the first cases started showing up in their own countries. Most of the spread where I live is from cases in northern Italy. Should I blame Italy for not preventing travel to and from their country starting 6-8 weeks ago?

If we had known about it earlier we could have tested or quarantined travelers coming from China. I'm not talking about shutting down air travel. We also could have tracked people coming from China more easily. I'm just talking about keeping records and knowing who might be infectious.

A third way between the quarantine cludge and letting the disease pass through is what South Korea has done. A big part of that is due to quick, competent and widespread testing coupled with virus detectives that backtrace who infected people may have come in contact with.

Note the government has the authority there to collect and use mobile phone, credit card and other data from those who recently tested positive. That's shared anonymously on social media so others can be made aware and know if they might need to be tested.


A non-insignificant number of my friends and colleagues here in central London suspect they had mild coronavirus over the last two weeks (low fever for 1 day or severe fever for 2-3 days; tightness in the upper chest for a week; occasional dry cough for a bit longer).

We weren't exactly risk-free to start with... in late Feb, I passed through Madrid airport. My colleagues had holidays in France, northern Italy, and in Tenerife (staying next to the first hotel to be quarantined).

I feel a lot better today than yesterday.

So how do I know if I am now immune and can come out of self-isolation?

Or do I stay locked up for a month or two, just in case?

My wife was the same. A mild case of Flu B that was going around. How do you know this time of year without testing?

Everyone keeps talking about "flattening the curve" and "slowing the spread" but nobody seems to be carrying this out to it's logical conclusion. Where does any of this lead? The planning horizon seems to be weeks rather than months-years.

We are pulling the ripcord right out of the gate, hamstringing the economy, and I fear we aren't even going to flatten to curve, so much as shift it to the right. Plus all the mental health damage this is doing, and all those who will receive substandard care from the medical institution. Immune systems overreact all the time, is that happening on a global level?

Until a vaccination is developed, this will be like the smallpox of the past, an endemic disease with occasional "massacres"; the better is social isolation and wait for a vaccine.

Oh, and to classify the workers of IT firms, broadband supliers, etc, as "essential workers", like the health workers, soldiers, etc.


Comments for this post are closed