[I am happy to turn this space over to my former colleague and (I trust) lifelong friend Romans Pancs, who offers what he describes as
a polemical essay. It has no references and no confidence intervals. It has question marks. It makes a narrow point and does not weigh pros and cons. It is an input to a debate, not a divine revelation or a scientific truth.
I might quibble a bit — I’m not sure there’s such a thing as a contribution to a debate that nobody seems to be having. I’d prefer to see this as an invitation to start a thoughtful and reasoned debate that rises above the level of “this policy confers big benefits; therefore there’s no need to reckon with the costs before adopting it”. That invitation is unequivocally welcome. ]
| —SL |
The Main Argument
It is a crime against humanity for governments to stop a capitalist economy. It is a crime against those whom the economic recession will hit the hardest: those employed in the informal sector, those working hourly customer service jobs (e.g., cleaners, hairdressers, masseurs, music teachers, and waiters), the young, the old who may not have the luxury of another year on the planet to sit out this year (and then the subsequent recession) instead of living. It is a crime against those (e.g., teachers and cinema ushers) whose jobs will be replaced by technology a little faster than they had been preparing for. It is a crime against the old in whose name the society that they spent decades building is being dismantled, and in whose name the children and the grandchildren they spent lifetimes nourishing are subjected to discretionary deprivation. Most importantly, it is a crime against the values of Western democracies: commitment to freedoms, which transcend national borders, and commitment to economic prosperity as a solution to the many ills that had been plaguing civilisations for millennia.
Capitalism and democracy are impersonal mechanisms for resolving interpersonal (aka ethical) trade-offs. How these trade-offs are resolved responds to individual tastes, with no single individual acting as a dictator. Governments have neither sufficient information, nor goodwill, nor the requisite commitment power, nor the moral mandate to resolve these tradeoffs unilaterally. Before converting an economy into a planned economy and trying their hand at the game that Soviets had decades to master (and eventually lost) but Western governments have been justly constrained to avoid, Western governments ought to listen to what past market and democratic preferences reveal about what people actually want.
People want quality adjusted life years (QALY). People pay for QALY by purchasing gym subscriptions while smoking and for safety features in their cars while driving recklessly. Governments want sexy headlines and money to buy sexy headlines. Experts want to show off their craft. But people still want QALY, which means kids do not want to spend a year hungry and confined in a stuffy apartment with depressed and underemployed parents; which means the old want to continue socialising with their friends and, through the windows of their living rooms, watch the life continue instead of reliving the WWII; which means the middle-aged are willing to bet on retaining the dignity of keeping their jobs and taking care of their families against the 2% chance of dying from the virus.
Suppose 1% of the US population die from the virus. Suppose the value of life is 10 million USD, which is the number used by the US Department of Transportation. The US population is 330 million. The value of the induced 3.3 million deaths then is 33 trillion USD. With the US yearly GDP at 22 trillion, the value of these deaths is about a year and a half of lost income. Seemingly, the country should be willing to accept a 1.5 year-long shutdown in return for saving 1% of its citizens.
The above argument has three problems that overstate the attraction of the shutdown:
- The argument is based on the implicit and the unrealistic assumption that the economy will reinvent itself in the image of the productive capitalist economy that it was before the complete shutdown, and will do so as soon as the shutdown has been lifted.
- The argument neglects the fact that the virus disproportionately hits the old, who have fewer and less healthy years left to live.
- The argument neglects the fact that shutting down an economy costs lives. The months of the shutdown are lost months of life. Spending a year in a shutdown robs an American of a year out of the 80 years that he can be expected to live. This is a 1/80=%1.25 mortality rate, which the society pays in exchange for averting the 1% mortality rate from coronavirus.
It is hard to believe that individuals would be willing to stop the world and get off in order to avert a 1% death rate. Individuals naturally engage in risky activities such as driving, working (and suffering on-the-job accidents), and, more importantly, breathing. Allegedly, 200,000 Americans die from pollution every year. Halting an economy for a year would save all those people. Stopping the economy for 15 years would be even better, and save all the lives that coronavirus would take. Indeed, stopping the economy is a gift that keeps giving, every year, while coronavirus deaths can be averted only once. Yet, with the exception of some climate change fundamentalists, there were no calls for stopping the economy before the pandemic.
The economy shutdown due to coronavirus seems to be motivated by the same lack of faith in progress and society’s ability to mobilise to find technological solutions (if not for this strain of the virus then for the future ones), and by the Catholic belief in the virtue of self-flagellation of the kind sported by climate-change fundamentalists of Greta’s persuasion. This lack of faith is not wholly the responsibility of governments and is shared by the citizens.
How We Got Here
Every schoolboy can compute and plot the flatten-the-curve meme. Few—especially few economists in a position of power—seem to understand or wish to discuss how economics works. Is the shut down economy supposed to restart itself on a whim of world governments? Do we understand how bankrupt businesses that were built for years will be reincarnated? Do we understand what the effect on competition will be? The government will bail out (large) incumbents, not small businesses (for both economic and political reasons), and large incumbents are in a better position to weather the storm to begin with. While economic consequences are dire, they are harder to quantify and illustrate in a catchy manner than the flatten-the-curve meme. And in the 21st century, quantities are the terms in which debates are cast, and rightly so. As a result, those not bold enough to throw around numbers are not heard. Economists should be bolder in sharing their back-of-the-envelope calculations, while making clear the provisional nature of these calculations and the assumptions on which they rely.
No sane economist would defend the view that each of 10,378 major economic malaises can be treated by fine-tuning one instrument: the interest rate. Yet somehow lax monetary policy and checks mailed to all citizens emerge as the cure du jour while most production is shut at the same time. Where are all the economists?
The individuals who will be hit hardest by the economic collapse are, if not invisible, then neglected and at the mercy of those who are wealthier, older, and more educated (and, so, in particular, able to work from home). If you are rich, you can sacrifice half of your wealth in order to reduce the probability of death by 1% or, for those who are older, by 10%. If you are poor, you will not accept halving your income; if you do, you die. While mortality overall appears to be cyclical, suicides are countercyclical.
Most do not understand how the economy works. (They do not have to; this is the luxury of the division of labour.) The generations now alive (apart from immigrants from alternative regimes) do not remember what a truly dysfunctional economy looks like. They think the economy can take a gap year at the expense of the government. They do not realise that they are the government, and if they do not work (and their neighbours do not work), then there is no one at whose expense to vacation.
Individuals are scared of correlated risks. The media makes big news of correlated risks (a plane that crashes or a terrorist attack). Coronavirus deaths are not as numerous as some others (e.g., due to pollution) but they are highly correlated in time and space, and, therefore, are salient in everyone’s mind.
Of course, one should not expect the government policy to be fully optimal. But, perhaps, it is not too much to expect the policy to be informed and consistent with the values that the society has revealed to hold sacred. One should not expect government policy to be optimal because, first, the political pressure is always not to act too soon, because the layman does not understand the exponential, which governs the rise in the incidence of infections. Then the pressure on the government is to act in a particular way (e.g., lockdowns), to show that it cares, even if alternatives could have been better (e.g., prioritise mass testing, convert hotels into hospitals, invite foreign experts in containment). Finally, the pressure will be to lift unpopular policies too soon, thereby undoing the benefit they could have delivered. This is democracy. It has some advantages.
Conclusion
The current government interference with freedoms of movement and contract is—for lack of a better term—criminal. If you want to interfere, interfere with zoning rules and build makeshift hospitals. Interfere with FDA approval guidelines, with patents on critical components for life support machines and face masks, and with employment restrictions imposed by professional associations. Motivate the industry to produce life support machines. Announce prizes for medical advances. Pay the nurses double. But do not gratuitously interfere with the freedoms that are the core values of the civilisation. Western democracies have already been exhibiting isolationist tendencies, and the recent travel bans run the risk of normalising nationalism.
But what scares me most is the layman who neither questions nor protests. And the economist who refuses to calculate.
This is just a variant of Schumpeter’s gale. Only instead of capitalism being the creatively destructive force, it is nature (and that includes human nature).
If you were old enough to have observed and thought about it, you would know from the 9/11 attacks that Americans over react, and over react wildly. America is a fearful society, having never really had to fight an enemy on its shores…having exempted most of its people from military service. You should not expect anything else from Americans during something like this.
But is “this” really that bad?
I imagine many other economies in history were hit just as hard if not harder. They collapsed, and in due time arose out of the ashes.
We might have a generation or so of residual fallout, but as the poet said…this, too, shall pass.
This is a excellent essay that I certainly hope starts a good debate!
My random thoughts: We are not really stopping the economy just slowing it down a lot. Most people would willingly give up some income for reduced risk and logically there must be an equilibrium point for every new risk where the marginal cost of minimizing it equals the benefits. I know this is stating the obvious but sometimes the obvious needs to be stated.
The problem is that we have no ready way to ascertain in the current situation where that equilibrium is. Our central planners (who appear to have had no plans at all for the current situation) produce ad-hoc plans that take into account only benefits and not costs. They then don’t even bother enforcing these plans efficiently so wiping out most of the benefits while extending the costs into the distant future.
Also I’m hoping that the writer is being overly pessimistic about the resilience of the economic system. Both during the crisis and in the recovery phase human creativity plus a free market may reduce utility loss more than he fears especially given my hunch that the economy may have the ability to work quite efficiently ‘online’.
Assuming a complete shutdown rather than a slowdown does seriously overestimate the costs. If we only lost 25% of output for a year and that really did prevent 3.2 million deaths then that trade off does not sound on-its-face unreasonable to me. But that is exactly the discussion we need to have in order to design the right policy.
In other words, we should be asking, What is optimal level of health precautions in a pandemic?
Much of the argument resonates with me, particularly the call for a complete cost-benefit analysis. As we know from debates on security (e.g., TSA) and climate change, it is easy to marshal impressive facts for one side but there is never just one side.
The assessment, though, strikes me as overly apocalyptic. Surely “shelter in place” is less bad than prison which is less bad than death or non-existence. Also, why such pessimism about the future? Given the numerous times societies have recovered from wars, tsunamis, earthquakes, etc., why expect less this time around? I have more faith in individual initiative to resurrect the economy from its moribund state once free to do so.
>Given the numerous times societies have recovered from wars, tsunamis, earthquakes, etc., why expect less this time around?
Exactly! Well said.
I think this is hyperbolic to the point of invalidating this part of the argument: “The months of the shutdown are lost months of life. Spending a year in a shutdown robs an American of a year out of the 80 years that he can be expected to live.”
First, there are people like me who are actually getting more done. I work in computer security, and normally my job involves finding security holes, and then getting buried in meetings and paperwork trying to identify the right people to fix the holes. Since working from home, I can spend more time finding the security holes, and since the company realizes that the holes have to be fixed anyway, it turns out, lo and behold, they can be fixed without the paperwork. While mine is an unusual case, I’m hearing from a lot of people saying unironically that they’re getting more done working from home.
And even for people who are getting less done, surely their productivity has not dropped by 100%.
As for non-work activities, even with large gatherings prohibited, people have reported enjoying the time with their families and 1-on-1 get-togethers with friends.
Like blink said, being dead or in prison really is being robbed of a year of your life. If this is better, then that part of the argument fails.
Good points, but what we’re seeing here in practice is a set of competing trade offs.
What is the duration and intensity of shutdown needed to minimize damage to the economy and to minimize loss of life? We’d need to use QALY to arrive at this answer.
However, once we arrive at that answer, there’s the tough part of politically selling it. UK tried to do a herd immunity strategy but backed out due to public outcry. We can’t restart the economy even with strict isolation of at risk groups without a sizeable # of people getting sick enough to need hospitalization or ICU level care. Once you overwhelm the system, you have to implement some kind of wartime-esque triage as is going on in Lombardy by denying treatment to someone who could otherwise be treated (and potentially saved).
There will be massive public outcry against that level of triage. I can already see people calling it “death panels” and “people dying in the streets”. The limits here are mainly political.
There is a point in the near future when the pressures will shift. Enough people will have gotten the first wave of illness and recovered that we can slowly lift restrictions and let low-risk/recovered people go back to work while continuing to quarantine at risk groups before slowly releasing them back into the community in waves so as to not overwhelm the health care system.
Ultimately, the disease will probably be endemic in the human population.
Furthermore, if we assume the goal of policy should be to generate herd immunity, one could argue keeping the economy running -and indeed even encouraging more freedom, not less – could get us there faster. For instance, consider the issue of whether we should encourage or discourage college students to celebrate Spring Break as usual.
I could make a very good argument encouraging college students to go out and party more, not less, could derive the greatest social benefits, since all of them will become immune at a low cost and risk. Obviously this assumes they don’t infect older people when they come back. But, again, making the assumption they isolate sounds more plausible when they think they are infected (go partying), than when they don’t.
Of course, this is the same argument Dr. Landsburgh made in his book “more sex is safer sex”. I have yet to hear anyone making this argument, and I wonder whether the reason is not that no one has thought about it but rather because we are scared of making it. That, to me, is the more scary thought.
#1 ” America is a fearful society, having never really had to fight an enemy on its shores” What about the war of 1812?
“8 “UK tried to do a herd immunity strategy but backed out due to public outcry.” I think it was more an epidemiologists outcry. You would need something like 70%-80% infection to get herd immunity (depending on exactly how infectious it is – measles needs over 90%). That is not something that can really happen in a selective way. Herd immunity without a vaccine is not a sensible policy.
#9 “Obviously this assumes they don’t infect older people when they come back.” This is the point. You cannot infect over 70% of the population without spreading it indiscriminately.
“Every schoolboy can compute and plot the flatten-the-curve meme.” The meme is illustrative rather than modeled. The origin seems to be here
https://stacks.cdc.gov/view/cdc/11425
This is from the CDC in 2007. This pandemic is a category 4 or 5. Almost certainly 5 as the paper is describing influenza for which there is some immunity already and assumes a 30% illness rate. The chart is on page 18 and is illustrative rather than calculated based on hard data.
We knew this was coming in some form. The CDC described the measures that may be necessary and emphasized the importance of the secondary effects and how we must be ready to cope with the secondary impacts. “Communities must be prepared for the cascading second-and third-order consequences of the interventions,such as increased workplace absenteeism related to child-minding responsibilities if schools dismiss students and childcare programs close.”
They say:
“It is recommended that the measures be initiated early before explosive growth of the epidemic and,in the case of severe pandemics,that they be maintained consistently during an epidemic wave in a community.”
What happened? We were not prepared, no plans were in place and measures were not taken soon enough.
The fact is that if everyone gets it at once, the death rate will be higher than if it is spread out. There will be many more deaths from other causes as the healthcare system is overwhelmed. The 1% death rate we are probably stuck with unless treatments become available or a vaccine is produced. What the Govt is trying to do is keep it down to 1%.
I don’t know if the costs and benefits are correctly balanced by the measures taken. Plans were not in place and the over-reaction is going to happen. But it is not the correct argument to make that the measures are trying to stop that 1% of fatalities.
“Motivate the industry to produce life support machines…” In the UK there was a call last week for manufacturers of components of ventilators, such as compressors, valves etc., to get grants to scale up manufacture and coordinate production. This is the sort of measure that is taking place but is not reported widely. Even so, it will take some time to get the finished products out there and a delay in the peak will help keep people alive, so when they do get to hospital there will be a ventilator for them and staff to administer the help. I don’t know if similar measures are being taken in the USA. There was talk of war-time powers to compel manufacturers, but why not just pay them to do it? These measures are being taken, but they also need the social isolation to allow them time to work.
In the UK the Govt. has said it will pay 80% of the wages of workers who cannot get to work. They are trying to prevent firms closing down and laying off workers so they can start up again. There are problems for self-employed and “gig economy” workers without contractual wages, but it is a start.
“Governments have neither sufficient information, nor goodwill, nor the requisite commitment power, nor the moral mandate to resolve these tradeoffs unilaterally.” Neither do individuals. Individuals are in a far worse position since they do not have access to experts. Experts do not know everything about individual preferences, but they do have a much better idea of the consequences of specific actions. Individuals cannot know that if they carry on as normal the healthcare system will collapse, whereas experts are in a much better position to predict the consequences. What the experts do not know is whether the individuals would prefer the healthcare system to collapse so they can carry on as normal for a while. Experts tend to over-react to address the problem they see in front of them, as they do not know the consequences in areas other than their expertise. This is why we need coordinators, who can take the view of many experts. This particular division of labor I don’t think can happen via the market.
We trust our politicians to do the right thing, which is a misplaced trust in every case but particularly so in the USA at the moment. I am no fan of Boris Johnson, but he does seem to be reacting to this situation in a good way. I may be wrong. I don’t know how many extra people would die if we carried on as normal. I don’t know how many people will be saved by the extra ventilators that will come on stream in a few weeks. I don’t know how many people would die of other conditions if the health service collapses due to over-demand and ill staff. Maybe we would be better letting things play out, but I think on balance not.
There is another factor-mutation of the virus. There seems to be little at present but the more virus is out there the more chance of mutation to a more virulent form. That is a real risk, but unquantifiable at present.
#10 “The [flatten-the-curve] meme is illustrative rather than modeled.” The meme can be computed rigorously by calibrating the SIR (susceptible-infected-recovered) model: https://en.wikipedia.org/wiki/Compartmental_models_in_epidemiology#The_SIR_model
The discrete approximation of the governing differential equations can be plotted numerically in one’s favourite programming environment in five minutes. So, a school boy can indeed do that, and rigorously. (The model also admits an explicit solution: https://arxiv.org/pdf/1403.2160.pdf).
#11. Thank you for those links. I agree the curve can be calculated, and thanks to you I now see how. Still, I don’t think the initial meme curve was calculated. Not that important, but perhaps it is useful to know that the meme curve is very generalised one and not one calculated for Covid19 specifically. On an almost comedic point, I think your estimate that *every* schoolboy could do it suggests an overconfidence in the education system.
It does not help that both recovered population and basic reproduction number are represented by R.
I somewhat miss-spoke above when I said the point was not to reduce the 1% fatality. That is very much part of the objective, but keeping it to 1% (or whatever the real fatality rate is) is also part of it by delaying the peak. The fatality rate will be higher if we have a sharper peak.
The Wikipedia article says that at the end of an epidemic is caused by the decline in the number of infected individuals rather than an absolute lack of susceptible subjects. Fewer people will become infected with a flattened curve and more of the infected will survive. This does not necessarily mean that the measures taken to try to ensure this are worth it. I notice Cuomo saying today that you cannot put a value on a life, but we do have to do this.
FYI: I wrote up a “natural rights” critique of the current economic shutdown here: https://priorprobability.com/2020/03/24/my-natural-rights-critique-of-the-economic-shutdown/
Bennett Haselton: “there are people like me who are actually getting more done… I’m hearing from a lot of people saying unironically that they’re getting more done working from home.
As for non-work activities, even with large gatherings prohibited, people have reported enjoying the time with their families and 1-on-1 get-togethers with friends.”
The Japanese automakers were able to boost their productivity, in
the decades following the war, because they enjoyed the advantage
of bombing back to the Paleolithic, hence forced to rebuild, with
new equipment.
Whereas, the Detroit clique had no such luxury.
#14 I, too, am an optimist. Of course, one should have a certain kind of a home and a certain kind of a job to work from home comfortably and productively. Many of those with other kinds of jobs and homes may still be worthy of preservation. Time with the family can cut both ways (e.g., https://www.globaltimes.cn/content/1181829.shtml). But change is good. So-called bullshit jobs (https://en.wikipedia.org/wiki/Bullshit_Jobs), if indeed problematic, may wither. And a nudge towards change (if it does not kill one) may generally be a good thing (https://www.theatlantic.com/business/archive/2016/08/quitting-your-job-and-other-life-choices/495122/).
#13. from your post “In brief, I will sum up my natural rights approach to the pandemic as follows: at some point you have to pay me to stay at home; that is, if you (the government) are going to order me to stay home for the greater good, then you are also morally required to pay me “just compensation” (including lost wages) in exchange for my cooperation.”
The UK Govt is paying 80% of wages for employees that their employer registers as furloughed to avoid laying them off.
I think the author’s problem #3 would be more impactful if he focused on literal lives – based on estimates I’ve seen a severe recession can easily cost tens or even hundreds of thousands of lives (depression, drugs etc.). So we’re really trying to flatten the sum of two curves that can move in opposite directions, quite a challenge. But naturally medical experts will tend to focus on one of them. So absolutely — where are the economists? We need their expertise in order to save lives too. In fact I’d argue problems like this are really problems for economics, as informed by the medical science – how do we best organize our resources to respond to this threat?
P.S. Harold – why would we need to worry about herd immunity if we had a vaccine? I thought of the former as a second-best strategy in the absence of the latter.
btw the mortality estimate was for the US alone
#17. Herd immunity protects the otherwise susceptible by preventing the spread of the disease. Vaccines protect the un-vaccinated due to the herd immunity principle. For measles, with 95% of the population vaccinated then those that cannot (e.g. immune suppressed) or will not vaccinate are still protected by herd immunity as the disease cannot spread.
I have had another look regarding Covid19. Rates are based on the basic reproductive ratio, R0, which is how many people one infected person would infect in a susceptible population. For measles this is very high about 12-18. This is why we need such a high vaccination rate. For seasonal flu it is 1-2 and for Covid19 it is estimated at between 1.4 and 3.9. The fraction of the population needing to be immune is 1-1/R0. From the estimates this would be from 30 to 75% of the populations who would need to have caught the disease to provide herd immunity. If the reproductive ratio is at the bottom if this range we would need to have 30% of the population recovered from the disease to provide herd immunity. I do not think even at this most optimistic rate we could somehow hold the number to 30%, and if we had 30% infected at the same time we would overwhelm the healthcare system.
Let’s take this argument further. Under utilitarian philosophies, we should seek the greatest benefit for the most people. This disease is primarily killing elderly folks with multiple other serious ailments, like obesity and diabetes. These people are a huge drain on our medical system, Social Security, and other govt services. Our society could benefit by killing them off.
The Wuhan virus could have been genetically engineered to infect the world, and to target the biggest public liabilities. The virus is fairly harmless to productive members of our economy. It efficiently kills those who need to be euthanized anyway. Think of the global economy as one big organism, and it just took a medicine to kill harmful parasites.
The officials who ordered the lockdowns did not get the memo. NY Gov. Cuomo is adamant that no one would ever put a price on human life, and that we should continue to shutdown the economy as long as it slows the spread of the virus. This is exactly backwards. We need to put a price on everyone’s head, and eliminate the negatives.
The Wuhan virus has gotten to the point where it cannot be stopped. We should accept the inevitable, and celebrate the benefits of infecting everyone.
I offer this as another input to the debate, and not as my personal opinion. I do believe that we need more reasoned analysis of ALL the pros and cons of the drastic actions that have been taken.
“The Wuhan virus has gotten to the point where it cannot be stopped.” I think you mean Covid19.*
That is not quite right. We need to get the R0 below 1 and the spread will end before everyone is infected. That is not to say we necessarily should do that, but we certainly can do that.
“for the first time since the beginning of the outbreak, there have been 0 new cases in Wuhan and in the Hubei province, and no new and no existing suspected cases in Wuhan and in Hubei.”
https://www.worldometers.info/coronavirus/country/china/
Whilst data from China may not be totally trusted, it is clear that they have for now massively slowed down the spread with a total number of cases so far of 81,000. This will increase, but it is far short of everyone.
If we were to consider the approach of letting it rip, we would have to decide who to save as there would not be anywhere near enough hospital beds available, even if we simply denied medical attention to the over 50’s.
“In France, 50 percent of the country’s COVID-19 patients in intensive care are under the age of 60.
“In the United States, among almost 2,500 patients with COVID-19 with a known age, 29 percent were 20 to 44 years old, according to the Centers for Disease Control and Prevention (CDC)Trusted Source…Of patients in the ICU, about 12 percent were below the age of 45. Few died.”
Without social isolation measures, most people will catch the disease (as there is no or little immunity) and many of them will have it at the same time. If we let it rip, tens of millions of people could be infected at the same time and many more than 1% will die, of all ages.
Restricting hospital treatment to the under 50’s would not be anywhere near enough. The fatality rate for the under 50’s would rise to well over the current rates.
The officials who ordered the lockdowns did get the memo.
In the article, Romans Pancs said what if 1% of Americans die of the virus. We know more now than when the article was written, even in these few days. We know that without social isolation the fatality rate would be much higher than 1%.
*With the most cases in the world, we could call it the American virus now.
Harold — this piece suggests herd immunity can develop in the absence of a vaccine, and that this virus may be particularly amenable to such a process:
https://www.nytimes.com/2020/03/20/opinion/coronavirus-pandemic-social-distancing.html
P.S. while I have no interest in getting into “what’s in a name” and the Chinese-preferred “covid-19” works just fine for me, as I’m sure you’re aware there is a long tradition of naming pandemics per their point of origin, however I’m not aware of any examples of retroactively labelling them by where they came to be more widespread.
Roger — I know you’re not endorsing that position, which is why it serves as another good example of the cold calculus of utilitarianism (denial of individual rights) the ultimate consequences of which I find hard to believe anyone truly believes as a standard of justice (although there are those trolley-pushers…)
However your point may be in part that in the debate over how much economic damage to sustain in our response to this crisis, at some point we are forced to think about horribly difficult issues like how to weigh “duration of loss” e.g. young people losing jobs, or businesses they’ve invested their lives in, and all the morbid consequences that can follow from that.
Note this blog is one of the few places I feel comfortable even talking about something like this. Which may be the answer to “where are the economists?”
@21 (Harold)
‘In the article, Romans Pancs said what if 1% of Americans die of the virus. We know more now than when the article was written, even in these few days. We know that without social isolation the fatality rate would be much higher than 1%.’
Can you validate this statement? I see no data that backs this up. Even the worse case scenarios I have seen suggest only a 1% total death rate even with no (enforced) social isolation.
#22 It seems that an awful lot of folks must be infected for herd immunity to kick in, and a lot of luck is required for the infected folks to be of the right kind, under 65 years old. Let me do a back-of-the-envelope calculation. Take the workhorse Kermack-McKendrick SIR model, described here: https://mathworld.wolfram.com/Kermack-McKendrickModel.html (The model generates the flatten-the-curve meme, among other insights.)
Combining equations (1) and (3) in that model to eliminate the variable “I” and solving the implied differential equation gives the long-run relationship between R, the number of recovered individuals by the time the virus has died out, and R0, the ubiquitous “basic reproduction number”:
1 – R = exp (-R0 * R).
For coronavirus, R0 is estimated to be between 1.4 and 3.9 by Wikipedia. Assume that R0=2.22. Then, the long-run equation above implies R=0.85; that is, 85% of the population must have been infected and now recovered (the model assumes no deaths) for the remaining 15% to be immune thanks to the so-called herd immunity conferred upon them by the 85% who have experienced the disease and can no longer transmit it. These 15% is exactly the fraction of the US population above 65 years old. (When R0=3, R=0.94; and when R0=4, R=0.98.) What is the chance that only the young will be infected and everyone above 65 will be spared? Zero.
For the herd immunity to work, first, we should have dispatched all senior Americans to Idaho for isolation and should have deliberately infected all the remaining Americans with coronavirus. Then we should reintroduce all senior Americans to the society by mixing them evenly with the young population, so that the young population serves as a barrier between the virus and the seniors. Sending the seniors back to their retirement communities and bridge clubs, where seniors disproportionately interact with other seniors, will not work. And, god forbid, if coronavirus were to make it to the senior colony in Idaho, then 85% of the seniors would have to get the virus for the remaining 15% to enjoy the herd immunity, and the entire scheme would be ruined.
#3 “The argument neglects the fact that shutting down an economy costs lives. The months of the shutdown are lost months of life. Spending a year in a shutdown robs an American of a year out of the 80 years that he can be expected to live. This is a 1/80=%1.25 mortality rate, which the society pays in exchange for averting the 1% mortality rate from coronavirus.”
This may be true in theory but the fact remains we actually haven’t shut down many parts of our economy. There are “essential” industries/jobs where the folks are still working. And just because we’re told to stay inside and/or away from crowds, that doesn’t mean life is lost for that year, at least not for everyone.
I concede it is perhaps mostly true for some. But for me for example, I’ve been able to do my job from home. I haven’t lost pay yet. I’ve actually benefited in some sense because I don’t have to drive to work anymore. I have more flexibility in my day in some sense as well. It’s not been so bad. Others, I concede, it’s not that way, especially for non-essential people who can’t work remotely. But I guess my point is thanks to modern technology a large portion of society can.
Continuing on my #25 above, I also concede it’s easier for me because I live in a walkable suburban community where one can walk outside without coming close to anyone else. I’m also at a point in my life where, being in my late 30s and married with a young baby, going out on the town and socializing every weekend with folks isn’t as important as it used to be, though admittedly a year of not socializing would be tough. We still have friends!
so all this to say I think the costs of this are highly dependent on a lot of personal factors.
#23. The increase in fatalities is a simple extrapolation of jospitalisation and icu rates together with the fact that such places are limited. The model described by Romans above shows that many people would be infected at the same time with an R0 of 2.2
Many of these would not get ventilators they need to stay alive whilst they recover.
#27. In the SIR model I described at #24, the maximal number of infected individuals at any given point of time will be about 19% of the total population for R0=2.22. (Of course, only a fraction of these will require hospitalisation.) When R0=4, the maximal fraction of infected rises to 40%.
This report suggests about 3 for the Italy and China outbreak before measures were taken,
https://www.ecdc.europa.eu/sites/default/files/documents/RRA-seventh-update-Outbreak-of-coronavirus-disease-COVID-19.pdf
“Basic reproduction number (R0): Recent modelling of the basic reproductive number (R0) from Italy estimate R0 between 2.76 and 3.25. Researchers from Lombardy who analysed the early phase of the outbreak in their region reported a reduction in R0 shortly after the introduction of mitigation measures [23]. This is consistent with findings from China. A recent review of 12 modelling studies reports the mean R0 at 3.28, with a median of 2.79.
R0 is proportional to the contact rate and will vary according to the local situation. Further research is needed to
get a more accurate estimate of R0 in the various outbreak settings.”
Assumption: Exposure + Recovery = Immunity
Assumption: A vaccine is not imminent.
We are all different. Some of us will exhibit few symptoms and will recover quickly. Some will have more severe symptoms that, if treated, can recover. Others are simply doomed.
Our choices are to do nothing (Roger #20), do “everything” or do something in between the extremes.
Do Nothing: Herd immunity builds rapidly and the return to “normal living” is quicker. Death count is likely higher because health care facilities are overwhelmed. Not that we have a “cure” but we can treat symptoms which would increase the recovery rate.
Impose Draconian Measures: Simply delays the inevitable; ultimately we will all get infected and recover, or die, until we attain herd immunity. Delay until a vaccine is available is unlikely. But delaying can alleviate overwhelming healthcare facilities such that we can treat symptoms and increase recovery rates.
But at what cost? I’ll leave that one to the economists.
Do something in between: Sweden appears to be having good luck. They are educating and advising but they are not imposing or demanding. I’m an old guy. I can read. I have no reason to go out so I simply volunteer to stay in.
My son is younger. He still needs to work. Distances, sneeze shields, soap and water. He can simply volunteer to be cautious.
Here’s to Sweden and their “R rated” movies.
Going forward; will anti-vaxers ever have a leg to stand on again?
The article GREATLY underestimates the damage the epidemic would do.
In the absence of a shutdown, the death rate would not be 1%.
All patients that require ICU care and some of the non-critical, but severe cases would die. That would be somewhere around 3-5% of all infected.
More damage would occur to doctors and other medical workers who have to deal with the massive avalance of death – a total of maybe 3 million medical workers would be involved dealing with this, and most of them would likely be severely traumatized. Add another 1 million USD * 2 million people to the total cost.
Of course a lockdown of 1.5 years is not feasible; but that is not what any country is (or should be) aiming for.
The right approach is “the hammer and the dance” approach, where you first hammer down the virus with a lockdown (maybe 6 weeks) and then keep it under control with strict social distancing, mask wearing, and mass-testing and contact tracing, similar to what Taiwan and South Korea have been doing successfully.
It is detailed here:
https://medium.com/@tomaspueyo/coronavirus-the-hammer-and-the-dance-be9337092b56
#24 — FWIW an estimate I saw based on similar range for Ro was 50-65%; I’ll try to remember where. But clearly the piece I referenced was describing a more “surgical” approach that focuses medical resources and isolation strategies on thoee most at risk. The purpose of which was to balance the two sources of *death* (and pain and suffering), which seems in keeping with Dr. Pancs’ concerns. So I’m not sure where we have the good for the perfect here. Clearly any strategy is only as good as human behavior, or authoritarian governments, will allow. Seems safe to assume Dr. Pancs would prefer we maintain our reliance on the former even where it is not without cost.
Also FWIW today CNN reported that Lancet Infectious Diseases estimates the mortality rate from covid-19 adjusted for unreported incidence at 0.66%, heavily skewed to the elderly (virtually non-existent in children)
#33 The Lancet paper illustrates jargon problems. The epidmiologists often talk of case fatality ratio whilst the public assume they are talking about infection fatality ratio. The public are much more concerned about infection fatality ratio because that actually makes sense.
The case fatality ratio is the proportion of people dying that have been diagnosed with the disease rather than who have had the disease. The numbers for flu are usually given as the case fatality ratio rather than the infection fatality ratio, so we should be careful of comparing apples to oranges. This is an important point and one I think the experts have failed utterly to explain. Trump famously said he felt the fatality ratio was lower than that cited by the WHO. In this instance I can understand him being confused, as the infection fatality ratio, which he was talking about, will be lower than the case fatality ratio, which the WHO was talking about. It is harder to forgive him for the way he communicated this confusion as he does have the experts on hand.
Paper here
https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(20)30243-7/fulltext
I think it is fantastic that we can access quality papers so quickly and easily. Just a word of caution – I presume this has been rushed through much faster than the normal peer review process would allow.
The estimated *infection* fatality ratio for 30-50 year olds is about the same as for *case* fatality ratio seasonal flu overall (about 0.1%). What we don’t know is the difference between case fatality ratio and infection fatality ratio for flu. I cannot find this information with a quick search. fatality ratio for flu is estimated anyway, and may be based on estimated number of infections. Anybody have any further information here?
This is comparing two different values – the case fatality for flu and the estimated infection fatality for Covid 19. In a best case, the fatality ratio for flu is actually the infection rather than the case ratio, or there is no difference between the two.
Given that we have a large number of flu deaths each year, but these are concentrated among the elderly, this may be a way to get a feel for the consequences of the disease. Best case, we could expect a similar proportion of deaths in the 30-50 group as we usually get for seasonal flu overall. Given that we have no immunity or vaccine we would expect the number of infections would be higher, so the absolute number in this age group to be higher than the usual number of flu deaths overall.
That again assumes the healthcare system is able to provide the same quality of care as normal.
I have posted my “property rights approach to the coronavirus pandemic” here: https://priorprobability.com/2020/04/02/a-property-rights-approach-to-the-coronavirus-pandemic/
#35. Do you think the approach of Denmark and the UK for example, of paying 80% of wages for employees is compliant with your proposal?
In the UK, the employer must register the employee as “furloughed” in order to access the funding.
This avoids the employer laying off the worker, so things can get back to normal faster, and compensates the worker for being forced to stay at home.
A similar scheme for the self employed is under construction for those that filed a tax return in 2019, but the mechanisms to pay them do not currently exist. As such, the payments will likely not happen until later – maybe June or July.
This seems a more sensible approach than sending everyone a one-off payment.
In the USA context, would this need to be administered by the States, but could be underwritten by the Federal Govt.? It may present problems in allocating funds which currently are unknown.
#35. If the Govt pays me, it gets the money from tax payers. Are you saying that other tax payers should pay me to stay at home?
What if they don’t want to?
I think I agree with your conclusion, but we can arrive at the same place from a socialist starting point. The Govt. decides that some action is for the collective good, so it raises taxes from those that can afford it to pay the people to do that action.
Is this a reasonable analysis, and if so, is it odd that socialist and property rights approaches come to the same conclusion? It seems odd to me.
Harold 34 thanks and I think I understand and agree with the points, certainly an important distinction for purposes of properly comparing medical risks, I would just say for the purposes of comparing (say) deaths for virus versus recession and more generally the types of concerns Dr. Pancs was expressing, we would want to focus on the absolute risk of the virus i.e. infection fatality rate, yes?
“we would want to focus on the absolute risk of the virus i.e. infection fatality rate, yes?”
Absolutely. This is what matters to us.
There’s a lot of assumption in the business-as-normal case as well. As we’ve seen the panic hoarding of toilet paper, when the hospitals are overcrowded and people are fearful, we may end up with a form of breakdown that disrupt lives in ways that are suboptimal than a more orderly scaling down of activities, as what we’ve seen here.
#40 Indeed.
Here is a more carefully thought through vision, by Martin Eichenbaum, Sergio Rebelo, and Mathias Trabandt:
https://www.kellogg.northwestern.edu/faculty/rebelo/htm/epidemics.pdf
Consider their Figure 3 in Appendix A. Without forced social distancing, consumption falls by up to 10%. With forced social distancing, consumption falls by up to 28%. In their model, in order to avoid dying from the virus, individuals voluntarily socially distance by working and shopping less. Because individuals do not internalise the beneficial effect that their social distancing has on others, optimal social distancing is more drastic than equilibrium social distancing, and leads to a larger recession. Optimal social distancing is defined to maximise the sum of (non-dead) agents’ utilities.
The model does not distinguish the orderly scaling down from the disorderly breakdown.
#41. Thanks once again for informative the link.
The take-away is that it is worth enduring a significantly larger recession to maximise utility.
You say “Optimal social distancing is defined to maximise the sum of (non-dead) agents’ utilities.”
Does not the model also include the loss of utility of the dead? They say on p10 “The expression for Uit embodies a common assumption in macro and health economics that the cost of death is the foregone utility of life.” I confess I did not go through the math, but I took it that this means the lost utility is taken into account. They say that this implies the value of a life is about $10 million.
In the medical preparedness section, where fatality rates rise with number of infected due to healthcare constraints: “Since the
costs of becoming infected are much higher, people cut back on consumption and work to reduce the probability of becoming infected. The net result is that fewer people are infected but more people die.” That is, I think, more people die than if death rates were static with numbers of infected, rather than more people die than if they had not cut back on consumption. Is that right?
“Recall that in the benchmark model we choose our baseline parameters so that, in the beginning of the infection episode, economic decisions account for 1/3 of the infection rate.”
The containment policy section models the containment as a tax on consumption. However, if 2/3 of infections arise from non-economic interactions, how does this realistically model the effect of social isolation measures? If we re-ran the model allowing normal economic activity but curtailed non-economic interactions we would arrive at a much higher utility. Basically, go to work and the shops but otherwise stay at home. This would reduce infection by 2/3, which seems unrealistically high. Many social interactions require someone to be working, eg restaurants, but we could reduce interactions with lower costs.
“Why not choose initial containment rates that are sufficiently high to induce an immediate, persistent decline in the number of infected? Absent vaccines, the only way to prevent a recurrence of the epidemic is for enough of the population to acquire immunity by becoming infected and recovering. The optimal way to reach this critical level of immunity is to gradually increase containment measures as infections rise and slowly relax them as new infections wane.”
Ultimately, everyone will catch it until herd immunity kicks in. However, if the infection rate is reduced sufficiently and there is sufficient testing the containment could be more local. This would be more akin to the “smart containment” section, but with regional rather than demographic targeting.
This would require extensive testing.
In order to have effective containment with minimal damage we must know who is infected and who has had it. We also need to know the most vulnerable, but I think we have a reasonable handle on this already, although some vulnerabilities will be at present unknown.
Testing is where our attention needs to be. If current containment successfully reduces the infection rate, but before most people have become infected, we can relax the containment whilst monitoring massively. This will pick up new outbreaks early and spread can be contained with regional measures and contact follow-up, as initially happened in S. Korea and Taiwan.
Taiwan has done a very effective job, considering the large movement of people to and from China.
The virus mutates relatively slowly, so a vaccine is a very realistic possibility. The model suggests that with a possibility of a vaccine, the initial constraints should be higher. If we can really do a smart, local containment and a vaccine actually does appear, this would seem to be even more effective.