To be fair, we should also add the time spent under lockdown. For many, existing in such conditions is 'not living', it's the time crossed out from their lives. Not to the same extent as dying, of course, but given that many millions are under such conditions, this will add up to a very significant number.

This is all very well laid out but what it says to me is not "the quarantine was a bad idea" as the article's author seems to have intended to imply, but "The US healthcare and social support systems are an absolute horror show"

What about people who gained exesive weight during the lockdown, and those who (re-)started being addict to alcohol and other drugs ?

Brilliant answer, but that last paragraph about what this means strikes me as something of a "buried lede": as with many questions on this site, the answer is "the claim is true, but doesn't imply what you might think as a casual reader", and that should probably be right at the top of the answer. One way to put it is that the claim is true of the number of years lost, but doesn't calculate the number of years saved, so isn't actually useful for measuring the impact of the lockdown measures.

@Shadur that is off-topic, but IMO it would be hard to draw that conclusion considering how the US has fewer COVID-19 deaths per capita than UK, Spain, France, Italy and lower unemployment than those countries. And not significantly more negative mortality effect of unemployment than other countries studied (Scotland, Sweden).

@LаngLаngС, it's conceivable that without quarantine, coronavirus deaths would (say) quadruple and outweigh anything discussed here. That is, the net effect of the quarantine is saving lives/years (versus not having it), though of course the situation overall is still quite poor.

What about the drop in flue cases, the drop in work accident, the drop in sports accident, the increase in work at home (reducing CO2, travel accidents, etc..), the general future drop in sicknesses due to higher hygiena habits, and the lot of people who felt a reduction in stress during quarantine, as well as started exercising more frequently ? This paper contains only argument in one direction, and do not try to disprove his point. Also, WTF in comparing the number of COVID real years lost instead of comparing the number of years saved because of quarantine.

@LаngLаngС The key point is that this claim looks on the face of it like it is a comparison of two things, but in reality it's a breakdown of one figure - the number of years lost in the current scenario. To be a meaningful comparison, there would need to be a baseline or alternative projection; without such a comparison, it is incorrect to conclude from this evidence that the result is "sub-optimal". To conclude that because a number is large it must be larger than a different number you haven't attempted to calculate is a fallacy.

@PaulDraper of course, the question does not ask that.

I am of the general opinion that the best answers on this site are the ones that challenge implicit claims, rather than allowing them to stand. It is similar to cases that confuse correlation and causation: a claim might be worded such that it doesn't explicitly mention causation, but an answer analysing the claim should not leave the reader with the impression that causation might be proven if it is not.

@Shadur there's nothing here that couldn't be said about other health care systems. People are avoiding hospitals and not seeking preventative treatment at all. That's what is contributing to the mortality.

@Shadur That may be true, but we have to deal with the healthcare system we have, we can't do anything now about mistakes in the past. So the most appropriate question to ask is what we should be doing now in response to the pandemic, not what we could have done years ago to have prevented it from becoming a disaster in the first place (some politicians like to play blame games, but it's not productive).

The "years remaining" numbers under "Death share by age" are way off. Just because a 79-year-old is past the overall life expectancy doesn't mean they are expected to live 0 more years. The age-specific life expectancy for 79-year-olds is nearly 10 years. For 89 year olds, it's almost 5 years. No matter how old you are, you are expected to continue living. Those numbers undercount the expected number of years remaining for the elderly population by a wide margin.

@NuclearWang you are correct which is why I called it a "naive average." Unfortunately, a great answer isn't possible since there is a big (proportionally speaking) life expectancy difference between 79 and 89 but COVID-19 statistics never separate those. I may go back and made some assumptions on the contrary side and see how much it can change.

@Spoutnik16 you’re arguing against the premise of the paper, which is entirely irrelevant to the discussion, which is simply “are the numbers of this paper correct?”, not “are the numbers of this paper useful?”.

@PaulDraper The stats at ssa.gov/oact/STATS/table4c6.html can be useful for this. If we assume a uniform distribution of cases by age in each age bracket, the 70-79 group has a life expectancy of ~11 years, while 80-100 is ~5 years. Of course, the uniform assumption isn't great since there's clearly a higher proportion of cases the older you get, so those number would have to be revised down somewhat to account for the fact that there's likely more cases that are 75-79 than 70-74. But even taking the extreme case of all 79-year-olds in the 70-79 bracket, it comes out to ~9 years.

A key flaw in the paper's analysis is imputing all the costs to the lockdown policies, when it's clear some sizable portion are due to aggregate individual behavior and self-quarantining. OpenTable data was showing a plunge in restaurant dining before widespread lockdowns. Demand for air travel collapsed. Demand for entertainment venues -- sports, movies, amusement parks -- were all impacted to some degree by people not going to crowded places. Blockbuster movies aren't coming out because they can't reach economies of audience scale.

Extending @IMSoP's fallacy argument, the numbers aren't taking into account what the loss of life would have been without lockdown. I.e. would it have doubled the years lost due to COVID19? Tripled? Hundredfold? It is the difference between with-lockdown-COVID19-years-lost and without-lockdown-COVID19-years-lost that should be compared to the lockdown-years-lost. In other words, so long as the lockdownprevents > 1.5 million additional COVID19-years-lost (i.e. the difference made by having the lockdown, both up until now and in the future), it is a net positive.

@Flater: Indeed, by analogy, in cases where a sprinkler system causes more property damage than a fire, many people are prone to view the system as having done more harm than good. A person who sees a burned wastebasket and partially-burned chair, along with a room full of water-soaked and ruined furniture, may easily miss the fact that the only reason the fire appears to have been small and harmless is that the sprinklers suppressed it.

@Zeus, ...I know some disabled people who get extremely offended by the claim that staying inside, socially distancing, and working from home where feasible is "not living" -- that's the lives they've always had, and they'd very much rather keep that than having literally no life at all, thank-you-very-much.

@jeffronicus: Indeed, one could argue that the too-brief lockdowns actually reduced some of those costs, by tricking people into trusting the reopenings. ("The governor says foo is safe now!")

Hmm I did only a quick read, but it seems you might add that years lost due to COVID are not only the years people who died lost. There is some evidence that it leaves damage even if you survive ecowatch.com/…

@Zeus: On the other hand, for some people the "lockdown" was actually a pretty nice vacation. While it didn't affect me (I've been 100% telecommuting for over a decade), people I know used it as time to catch up on chores, relax, work on personal projects, &c. For those of us who don't live in cities, it really didn't affect our ability to get out and hike, bike, cross-country ski (at least while we had snow), ride horses, or whatever we normally do. (Note that my normal life does not include any significant amount of time hanging out in bars or other crowded places.)

I am deeply skeptical that modelled estimates of marginal* increases in deaths due to marginal increases in unemployment or decreases in GDP are reliable in major downturns outside of the normal. (* Marginal, as in at the end, rather than important.)

There is a lot to read, and I won't pretend to have read everything perfectly. But from reading this answer and skimming the research materials, it looks like the conclusion is largely based on the assumption that one can simply multiply "unemployment mortality" figures by the number of months unemployed. I haven't found support of such an assumption at this point in my reading. Are there peer-reviewed studies that demonstrate that you can just multiple the numbers together to get a meaningful result?

There is so much wrong with saying (as the article seems to) "the unemployed have a 60% higher mortality rate than the employed" (with seems uncited) implies "becoming unemployed increases your risk of mortality by 60% (forever?)". Statistics just doesn't work that way. People more likely to become unemployed are probably already at higher risk (meaning the increase is far from 60%) and people less likely to become unemployed might be at lower risk, but are probably also at lower risk while unemployed. Never mind the atrocious potential assumption that people will remain unemployed forever

I would suggest making much more prominent the minor footnote saying that the article doesn't prove (nor in any reasonable way attempt to prove) we would've been better off with no lockdown. The authors of the article appear to be very much want to imply that, and so does answering "yes" to this question.

@Charles, it doesn't matter what they think. It matters that the affected people feel that way. There are, after all, people who'd rather die than be disabled (in some ways), and their opinion deserves no less respect than the opposite one you quote. It's always a personal evaluation.

@jamesqf, true, me too wasn't affected that much. But millions were; and mere inconvenience or millions does amount to lives lost.

The use of "EDIT" monikers is not necessary. Every Stack Exchange post already contains a detailed edit history; the edit history for this posted is located here.

@Zeus, the point isn't that they're offended; the point is that they're right. Once someone is accepts that that's what their life is, it's a lot better than nothing. Humans, in general, have a baseline level of happiness that they generally end up at regardless of their current situation; deviations happen around that baseline based on variances in current events, but in the long term it always trends back. Give a person who tends towards depression a sudden windfall and they'll still trend back that way; give a generally happy person a turn of bad luck and they'll eventually recover.

@Zeus: Ny point is that assuming that everyone affected was affected negatively in the same ways and to the same extent, as the authors seem to be doing, leads to a considerable over-estimation of the years of life lost.

@jamesqf, my understanding was that the authors don't include this metrics at all. They are talking only about (years lost due to) death. Given that we have difficulties even calculating mortality, I agree that it would be next to impossible to calculate 'moral' losses. Yet while there are some who were not much affected or even benefited, it would be hard to deny that the overwhelming majority were affected negatively (to different extent); thus the estimation can only grow from the authors' figure.