A few weeks ago, the NYT reported that “The Coronavirus Has Claimed 2.5 Million Years of Potential Life.” If you read the original study, you’ll discover one crucial caveat: The authors’s calculations assume that COVID victims would have had the standard life expectancy for Americans of their age. They freely admit that this is unrealistic and inflates their estimate:
The authors argue that fixing this problem would only modestly cut their estimates. I’m not convinced, but I’d rather focus on a much bigger issue: Taking quality of life into account, how many life-years has the reaction to COVID destroyed? To see what I’m getting at, ask yourself: “Suppose you could either live a year of life in the COVID era, or X months under normal conditions. What’s the value of X?” Given the enormous social disruption and dire social isolation that most people have endured, X=10 months seems like a conservative estimate. For what it’s worth, this Twitter poll agrees*:
Suppose you could either live a year of life in the COVID era, or X months under normal conditions.
What's the value of X that makes the AVERAGE AMERICAN indifferent?
— Bryan Caplan (@bryan_caplan) November 9, 2020
So what? Well, we’ve now endured 8 months of COVID life. If that’s worth only 5/6ths as much as normal time, the average American has now lost 4/3rds of a month. Multiplying that by the total American population of 330M, the total loss comes to about 37 million years of life. That’s about 15 times the reported estimate of the direct cost of COVID.
Casual readers will be tempted to declare that the cure has been much worse than the disease. The right cost-benefit comparison, however, is not to weigh the cost of prevention against the harm endured. The right cost-benefit comparison is to weigh the cost of prevention against the harm prevented. You have to ask yourself: If normal life had continued unabated since March, how many additional life-years would have been lost? I can believe that the number would have been double what we observed, even though no country on Earth has done so poorly. With effort, I can imagine that the number would have been triple what we observed. There’s a tiny chance it could have been five times worse. But fifteen times? No way.
Upshot: The total cost of all COVID prevention has very likely exceeded the total benefit of all COVID prevention.
Before you panic, note these key caveats:
1. This does not imply that zero COVID prevention was optimal. The lesson is merely that we went much too far.
2. Prevention includes both private and government efforts. The main lesson of the data is not merely that government overreacted, but that people overreacted.
3. As I’ve argued before, the initial costs of government action were moderate, because private individuals reacted strongly on their own. Over time, however, government’s share of the burden has increased because private individuals’ have a strong tendency to lose patience and return to normalcy.
4. If a vaccine suddenly became available today, my calculations for the story so far would still hold. Behavioral changes prevent deaths day-by-day. They also drain life of much of its meaning day-by-day.
At this point, you could protest, “Hey Bryan, I thought you weren’t a utilitarian.” So what if the cost of COVID prevention greatly exceeds the value of life saved? My answer, to repeat, is that I have a strong moral presumption in favor of human liberty. So while I respect individuals’ rights to overreact to moderate risks, I oppose any act of government that does not pass a cost-benefit test with flying colors.
And no, I don’t think that an asymptomatic person who walks down the street unmasked is “aggressing” against passersby in any meaningful way.
* You could object that my Twitter followers are self-selected to regard COVID prevention costs as high. In point of fact, they consider the personal costs markedly less serious than the average costs:
Suppose you could either live a year of life in the COVID era, or X months under normal conditions.
What's the value of X that makes YOU indifferent?
— Bryan Caplan (@bryan_caplan) November 9, 2020
READER COMMENTS
JFA
Nov 24 2020 at 11:43am
Given people’s tendency to revert back to a particular happiness level, I have become less enamored about the utility of these kinds of discrete choice surveys.
Michael
Nov 24 2020 at 12:05pm
“And no, I don’t think that an asymptomatic person who walks down the street unmasked is “aggressing” against passersby in any meaningful way.”
I am not sure how true that statement can be. If the person knows they are asymptomatic, or has might have been recently exposed I dont know how you can think its not aggressing when they would be knowingly putting someone in danger.
Even then, we know that masks aren’t a silver bullet but can drastically reduce viral load for the person with the virus and the person without making any infection less deadly so I am not sure this statement is valid. It seems it would take some willful ignorance to choose that action and place others at risk.
Masks Are a Tool, Not a Panacea – Reason
Student of Liberty
Dec 19 2020 at 4:49am
How many people know they are asymptomatic? Seriously, I do not even know how we ended up testing people that were asymptomatic.
Thanks to the scare on HIV to my cohort of similar aged morons, I got tested without symptoms for HIV and was happily negative but on COVID19? There is no way I am going to pay and get tested for it if I am not compelled to if I do not show symptoms!
suddyan
Dec 19 2020 at 6:49am
[I dont know how you can think its not aggressing when they would be knowingly putting someone in danger.]
When I get into my car and drive I am “knowingly putting someone in danger.”
As much as we like to think we are perfect, we are not. Each and everyone of us have concentration lapses, so behind the wheel of a car each and everyone us are “knowingly putting someone in danger.”
[It seems it would take some willful ignorance to choose that action and place others at risk.]
I manage my health through nutrition and a lifestyle that balances my health and my (rather sedentary) passion for pursuing interesting knowledge.
When I am outside, I wish to breathe in the air as I and my ancestors did.
It would seem like wilful despotism to force me to wear a very ineffectual piece of cloth for unproven, unambiguous, supposed “drastically reduc[tion of] viral load.”
If anyone is scared of me and my wilful, unhindered breathing of oxygen, they can indicate so and we pass each other at some distance.
This I am willing to do. Now kindly allow me to take this infernal mask of negligible, if any, statistical value off my face.
Thank you.
robc
Nov 24 2020 at 12:31pm
Will you accept 1.0 as the multiplier? Actually, it is a bit above 1 but it looks like policy did almost nothing:
robc
Nov 24 2020 at 12:33pm
Link got deleted for some reason:
https://www.frontiersin.org/files/Articles/604339/fpubh-08-604339-HTML/image_m/fpubh-08-604339-g006.jpg
And the full study:
https://www.frontiersin.org/articles/10.3389/fpubh.2020.604339/full
Scott Sumner
Nov 24 2020 at 12:44pm
This is a very plausible argument. However one should also consider the extra death toll with no social distancing, perhaps another million people. Then consider who much people dislike having a loved one die, compared to missing a few college parties.
I’m not saying you are wrong (I suspect you are correct), just that it’s a complicated question.
robc
Nov 24 2020 at 12:51pm
The kids going to the college parties have made clear their preference. Or they are really good at math. Probably the former.
Dave Baker
Nov 27 2020 at 10:07am
Along similar lines: you’d have to consider how many months the epidemic would’ve lasted without the reaction, and how a month of Italy-style case counts compares to a month of distancing in quality-adjusted terms. How much distress would it cause to wake up and read the news each morning under those conditions, or to wonder whether there would be a bed for you if you did end up catching a bad case? Etc etc.
Also note that death rates per infection have been higher in places with more infections, as even Ionnidis’s group eventually found. Could be an effect of higher viral dose, and/or it could be that more hospital crowding makes it easier to die if hospitalized at the margin, even in hospitals that aren’t full. I don’t think any realistic scenario would’ve been as bad as the Imperial College nightmare predictions, but it could’ve been a lot worse than one would imagine if you just extrapolate from the reality.
Vivian Darkbloom
Nov 24 2020 at 1:26pm
Tyler Cowen, on the other hand (mistakenly, I think) believes that the hardships we impose on ourselves or that are imposed on us today in order to avoid potential negative consequences from the virus represent an issue of “inter-temporal substitution”:
https://marginalrevolution.com/marginalrevolution/2020/11/intertemporal-substitution-remains-underrated-covid-in-scotland.html
I don’t really understand the argument, but he seems to suggest that those “lost quality years” are not really lost—they’re just put off until tomorrow.
Ghatanathoah
Nov 25 2020 at 2:59am
Vivian Darkbloom, I think he’s basically saying that people can just cram the stuff they couldn’t do because of COVID-19 this year into future years. For example, a family that goes to Disneyworld every other year could go two years in a row instead. Similarly, you could go to a few extra parties for the next few years to make up for this year.
Vivian Darkbloom
Nov 25 2020 at 7:21am
You may be correct; however, I don’t think it makes sense. That period of inconvenience can’t be reversed or replaced and this is not an example of “inter-temporal substitution”. If you buy that argument, it suggests that Caplan’s calculations of lost quality years are irrelevant.
Swimmy
Nov 24 2020 at 1:32pm
To me, the worst parts of COVID are the fear that I or someone I know will die of COVID, and especially the psychological costs of someone I know dying of COVID, which thankfully hasn’t happened yet.
Neither of those go away if you ease the various preventions, and they might get worse.
If I had answered your initial poll question, this is the only thing I would have been thinking about. I wouldn’t have thought about the costs of prevention at all! (They’re close to 0 for me.)
But doesn’t this analysis assume all the distress COVID causes is from the preventative measures? I think you need a differently worded poll question to make this case.
Jens
Nov 27 2020 at 6:20am
This is an important point. Does the article assume that the value of a year in months in a scenario where COVID exists and no preventive measures are taken is 12?
Aloha Jane
Nov 27 2020 at 1:37pm
Very hard to understand responses like this given what the numbers show. Do you also live in constant fear someone you love will die in a car crash, or from cancer? At various ages, either is likely more likely. Did you worry about pandemics before? How about solar storms? Meteors?
KevinDC
Nov 24 2020 at 2:06pm
This jumped out at me as illustrative of a common line of contention between people who advocate utilitarianism as opposed to Huemer/Caplan style common sense morality. If someone who advocates common sense morality starts talking in terms of costs and benefits, that is pounced upon as some sort of endorsement of utilitarianism. But I disagree entirely with this.
Common sense morality would agree that “consequences matter” is intuitively plausible. Consequentialism, by contrast, is the much less intuitively plausible idea that consequences are the only thing that matters. And utilitarianism is the even less plausible claim that just one particular kind of consequence is the only thing that matters. And even within utilitarianism there is disagreement if the one relevant consequence is average utility, or total utility, or whatever.
Now granted, those claims are coherent, in the sense of not internally self contradictory, but they at least seem wildly implausible on the face of them and seem to lead to patently absurd conclusions. One would think that utilitarians would have really powerful arguments to overcome this initial implausibility and to deal with the absurd seeming conclusions implied by utilitarianism, but all of the defenses of utilitarianism I’ve encountered fall wildly short of this. To any utilitarian who might be reading this comment section – what are the best arguments in favor of utilitarianism you can point me to?
Philo
Nov 24 2020 at 2:22pm
“I don’t think that an asymptomatic person who walks down the street unmasked is ‘aggressing’ against passersby in any meaningful way.” Well, how much of a risk would he have to pose to passersby in order to count as “aggressing”? What if, unmasked, he enters a store, a barber shop, a bar, a taxicab? What if the infection rate has doubled, tripled, etc., so that his probability of himself being infected (though asymptomatic) is greater? What if he has a slight fever?
It is hard to know where to draw the line on “aggression”; this region of law and commonsense morality is quite fuzzy and unsettled. You come across as overconfident.
Mark Bahner
Nov 24 2020 at 6:30pm
<blockquote>“I don’t think that an asymptomatic person who walks down the street unmasked is ‘aggressing’ against passersby in any meaningful way.” Well, how much of a risk would he have to pose to passersby in order to count as “aggressing”?</blockquote>
Walking up to someone and deliberately coughing, shouting, talking, or breathing in their face (from say, less than 2 feet away).
Transmission of COVID-19 in outdoor settings is incredibly rare:
Indoor vs outdoor transmission of COVID-19
robc
Nov 24 2020 at 7:34pm
New study in Nature out of Wuhan suggests that asymptomatic infection of covid doesn’t happen. Not rare, non-existent.
If so, there is no chance of an asymptatic person aggressing.
Scott Sumner
Nov 24 2020 at 10:48pm
Actually, asymptomatic infection is extremely common. The disease is at is most infectious in the two days before symptoms appear, and for three days after.
Mark Bahner
Nov 25 2020 at 12:42am
Actually, the people you are referring to (“in the two days before symptoms appear”) are “pre-symptomatic,” versus “asymptomatic” (never develop symptoms).
But per Harvard Health:
If both symptomatic and asymptomatic (never develop symptoms) people have the “similar amounts of virus in their bodies throughout the infection” it seems very unlikely to me (with no expertise in the matter ;-)) that asymptomatic people would never transmit infection.
BW
Nov 25 2020 at 2:56am
That’s an irrelevant semantic distinction. How is someone who hasn’t been tested supposed to know if they are asymptomatic or pre-symptomatic? If testing is the only way to know, then someone with no symptoms walking down the street unmasked could be contagious; in which case, we’re back to the point that Philo made.
Mark Bahner
Nov 25 2020 at 4:29pm
No, it’s not an “irrelevant semantic distinction.”
Commenter “robc” made the claim that asymptomatic (versus pre-symptomatic) people can’t transmit the virus, therefore asymptomatic people can’t commit a COVID-19 aggression.
Scott Sumner then stated that “asymptomatic” people can and do transmit the virus prior to coming down with symptoms. But I pointed out that the precise scientific term for what Scott was writing about was “pre-symptomatic,” versus “asymptomatic.”
If “robc” is correct that “asymptomatic” (versus “pre-symptomatic”) people can not possibly transmit the virus, then he is correct that they cannot commit a COVID-19 aggression.
IMPORTANT NOTE: I am highly skeptical of the claim that asymptomatic cannot possibly transmit the virus. Therefore, I am also highly skeptical that asymptomatic people cannot possibly commit a COVID-19 aggression.
BW
Nov 25 2020 at 9:19pm
It’s pretty clear that when Bryan Caplan, Philo and Scott Sumner used the term asymptomatic, they meant “not showing symptoms.” That’s a reasonable use of the word. It’s obviously semantics to point out that scientists use the word differently; or to point out that Philo and Scott Sumner’s arguments would be incorrect by the “scientific” definition. And because such a semantic argument doesn’t address the essence of their points (i.e. what they actually meant), it is irrelevant.
Mark Bahner
Nov 25 2020 at 9:26pm
I was very skeptical about asymptomatic (versus pre-symptomatic) people not transmitting any COVID-19, since a study showed asymptomatic (versus pre-symptomatic) people had similar virus loads in their bodies throughout infection.
But I see this from the WHO:
So even though I’m still very skeptical about absolutely zero transmission from asymptomatic (versus pre-symptomatic) people, it does appear that “asymptomatically-infected individuals are much less likely to transmit the virus than those who develop symptoms.”
Alan Goldhammer
Nov 24 2020 at 2:53pm
The focus on COVID-19 mortality statistics is but one part of the problem. Do you think hospital workers and staff enjoy seeing facilities packed with COVID-19 patients?
You should also factor in all those who have lingering side effects from the virus. We have good friends who both came down with COVID-19 two weeks ago despite all the precautions. They have been largely asymptomatic other than total loss of taste and smell. From all the reports I’ve read, it is not clear how long this anosmia will last. there are other neurological effects that can be lingering for some months. Not everyone is lucky and just has a mild case.
Jose Pablo
Nov 24 2020 at 7:44pm
Taking quality of life into account,
could have also another meaning:
“Suppose you could either live 10 years of life after your turn 70 or X years of life in your 20-30’s What’s the value of X?”
I guess X would be significantly less than 10.
Maybe we should use “in your 20-30’s equivalent years of potential life” instead of “years of potential life”. That would very likely reduce the 2.5 million years figure.
Thomas Hutcheson
Nov 25 2020 at 8:35am
So what is the bottom line? How does Caplan think we should be making decisions about what kind of restrictions we should impose on folks to prevent them from spreading infection to other people. [I’m assuming he would not interfere with people exposing themselves to risk.]
If, as is possible,many people are taking excessive precautions to avoid infection, does that not argue for greater expenditures on disease control (vaccine/treatments) than is justified just on externality grounds?
sk
Nov 25 2020 at 2:57pm
Seems to me you overlook risk of contracting Covid and end up suffering with some post Covid recovery symptoms: Compromised lung function, scarring of heart tissue ( and unclear implications and if heals over time).
This is a nasty disease that framing and performing some calculations of risk in death, no death scenario might be a bit simplistic
Philo
Nov 25 2020 at 5:37pm
“If normal life had continued unabated since March, how many additional life-years would have been lost?” Your answer to this question is the key to the whole post, but it is offered in a very offhand manner. Can you not provide some evidentiary support?
Richard
Nov 25 2020 at 7:20pm
Is the author comparing the years lost to Covid-19 with the current prevention measures in place to estimated cost of the measures? This would not be useful for any conclusion of this type.
Ethan
Nov 26 2020 at 8:14pm
No. He clearly says,
“The right cost-benefit comparison, however, is not to weigh the cost of prevention against the harm endured. The right cost-benefit comparison is to weigh the cost of prevention against the harm prevented.”
Student of Liberty
Nov 26 2020 at 4:34am
Some days, I really wonder if a month of normal life is not worth a year in 2020 (which is almost a complete write-off to me).
That said, I like the “years of life” cost measurement: some might have attempted at coming up with a dollar figure to the cost of restrictions but it does not compare well to a dollar figure of the cost/price of life, which for non-economists seems to be the worst horror they have ever seen (it needs to be an abstraction to make sense).
ConstantBees
Nov 26 2020 at 12:55pm
I live in the UK. Our government has effectively sunk the British economy, but insists that we have to continue with this living death through March. I’m in my 60s and one of the many who have contemplated suicide because of this wholesale destruction of our lives. Obviously, my life and my mental health is without value to our government – unless I happen to die of covid. Then my death is worth more than a thousand of the lives that have been destroyed by our hysterical government.
So I would happily give up the eight months that I have lived through this torment, especially the psychological damage done, for, say, a week of normal life. Frankly, at this point, I wish I had died at some point before the epidemic so I would not have had to live through this. Don’t fret, if you are someone who actually cares about people who die of something other than covid. I’m not actively suicidal right now, just doing a cost-benefit analysis, which seems entirely sensible in the current circumstances.
The Truth
Nov 27 2020 at 12:52pm
Peter Shaw did a similar, and I’d say better analysis more than 2 months ago.
https://shawfactor.com/2020/09/08/there-is-no-logic-in-lockdown/
AMT
Nov 27 2020 at 2:01pm
Would it surprise you to find that my answer is at least 13 months? I GET to work from home and save time (and money) commuting to and from work every day, which means I have more time to spend with my wife, and far more time spent with my dogs. (my commute wasn’t that long, but how many people are saving 2+ hours per day, five days per week now? 10 hours per week x 50 weeks=500 hours per year)
I also get to use an excellent excuse to avoid some social engagements that I would have felt pressure to attend, but didn’t feel were worthwhile, whereby declining in the past would have been seen as an insult/rejection and damaged the relationship.
Philo
Nov 27 2020 at 2:02pm
Tyler Cowen, Bryan Caplan on the cost of Covid – Marginal REVOLUTION, points out that government could have done some good things, such as subsidizing medical research, to enable a better response to the epidemic. He complains that Bryan has not advocated such governmental measures. My reaction: the government could do lots of good things, but if it is given the scope to do them it will do even more bad things, too. Better keep it, and its apparatus of coercion, small and rather weak, and rely on non-governmental (private) entities to do good things.
Josh Bremseth
Nov 28 2020 at 10:26am
Maybe I’m reading it wrong, but how did you go from rated value to years of life lost?
So what? Well, we’ve now endured 8 months of COVID life. If that’s worth only 5/6ths as much as normal time, the average American has now lost 4/3rds of a month. Multiplying that by the total American population of 330M, the total loss comes to about 37 million years of life. That’s about 15 times the reported estimate of the direct cost of COVID.
Wouldn’t it read something more like this:
So what? Well, we’ve now endured 8 months of COVID life. If that’s worth only 5/6ths as much as normal time, the average American has now lost (the potential worth of or value of) 4/3rds of a month. Multiplying that by the total American population of 330M, the total loss comes to about 37 million years of (potential value or estimated value or potential worth or rated value or something). That’s about 15 times the reported estimate of the direct cost of COVID.
In any case, thank you for the interesting essay,
Josh
Ari T
Nov 28 2020 at 11:22am
While I think all of these arguments have merit, I think it is too abstract. It’s like talking about war in Iraq or something. It’s like Bryan’s or anyone who is sitting 1st world comfy chairs views on war. People who have seen such hell how much more realistic perspective. Same thing I suppose would go for COVID or some other pandemic.
These things are way too important to be let be decided by ideological views. I do hope also liberty would be the “right” choice here, an I hope in big world there are many solutions and values.
Its too easy to favour either freedom or lockdowns based on pre-existing political beliefs. No skin in game. It is what Robin Hanson calls far-mode.
Also the talk about masks, aggressing and NAP are just plain silly. Lockdowns do have real and economic impact though.
Why no analysis on Sweden?
Sean
Nov 29 2020 at 8:37pm
QALY is a good way to frame these tradeoffs and it would be interesting to see some kind of break-even analysis on a strict, miserable lockdown that actually gets Covid under control vs. a longer but somewhat less miserable slog through Covid purgatory. My hunch is Melborne’s 111-day hard lockdown wasn’t “worth it” in QALY terms, but maybe it might have been if it were only 6 weeks?
In the US anyway, it’s unfortunate that the debate about such trade-offs devolved into “lives vs. the economy.” As soon as Trump said “we can’t let the cure be worse than the disease,” it became taboo to even raise the topic of trade-offs within a liberal social circle. This completely obscured the trade-off between lives and *life*.
Finally… if we’re talking QALY trade-offs… wear the damn mask.
Comments are closed.