For most of my teaching career, I was an economics professor at the Naval Postgraduate School. My students were generally motivated, reasonably hard-working military officers from the Navy, other US services, and other countries. I loved teaching them. But one day in the mid-1990s I told them a true anecdote whose shock value required them to know, for comparison, the population of the United States. I usually read students’ faces well and I saw little shock. I immediately wondered if they knew the US population. So I told them to write their estimate of the US population on a piece of paper and hand it in. At the time the US population was about 270 million. Finally there was a shocked face: mine. Although the median answer was within 10 percent of the right answer, only about 40 percent of the students gave a number close to the median. The estimates ranged from—are you ready?—1.5 million on the low end to 2 billion at the high end. I pointed out that 1.5 million was close to the number of Americans in the armed services. So that would mean that almost everyone in the United States, including little babies and eighty-year-olds, was in the military. At the other end, I noted, 2 billion was about one-third of the world’s population.
This is the opening paragraph of my latest article for Hoover’s Defining Ideas, “What Are You Afraid Of?” Defining Ideas, December 2, 2021.
And on to Covid:
And now we come to the issue that has dominated American society since March 2020: COVID-19. From almost the start, the United Nations, echoed by various media outlets, told us that “we are all in this together.” That’s one of those rare statements that are literally true but figuratively false. Yes, we’re all in this together, but the probability of death from COVID-19 varies dramatically by age. The younger you are, the less at risk you are, and if you’re really old, your risk is three orders of magnitude greater than the risk for the very young. This matters because the very young are bearing substantial costs of government measures on COVID-19 in return for tiny benefits to themselves. COVID-19 lockdowns, shutdowns of schools, and masking requirements not only indoors but also, sometimes, outdoors, are taking a huge toll on the very young that could well scar them for the rest of their lives.
Consider the latest risk data reported by the Centers for Disease Control and Prevention (CDC). As of November 24, 621 people in America age seventeen or younger had been killed by COVID-19. That’s only about 1 percent of all deaths (62,999) of people in that age bracket. People sixty-five years old and older, by contrast, suffered 580,605 deaths. That was a substantial 12.6 percent of all deaths (4,609,094) of the elderly. In 2019 there were approximately 54.1 million American residents age sixty-five and up. For people sixty-five or older, therefore, the risk of dying from COVID, over the almost two years we have dealt with it, was 1.1 percent. In 2019, there were 73 million American residents age seventeen or younger. So their risk of dying from COVID over the same almost two years was only 0.00085 percent. The elderly people’s risk of dying from COVID, therefore, is 1,294 times the risk of that of children.
Read the whole thing.
READER COMMENTS
Andrea Mays
Dec 2 2021 at 6:22pm
“He noted that some sharks had recently been seen near where the training occurred, and he wondered if maybe the trainers should pick another spot.” Only if the spot has fresh water!
Paraphrasing a CSU marine biology professor (who runs the Shark Tank on my campus) “If you can taste salt in the water, there are sharks in it.”
Phil H
Dec 2 2021 at 8:36pm
I’ll make the classic paternalist argument:
Premise: We know that a large fraction of the population has poor numeracy.
Premise: Making a good decision on what kind/level of Covid protections to take requires good numeracy.
Premise: Making the wrong choice on Covid protections can cause significant harm to yourself and others.
Therefore, if significant harm is to be avoided, some form of intervention is necessary.
Obviously what kind of intervention that should be is the bone of contention; I think some form of government intervention is valuable, and many here disagree.
But the underlying insight is right, I think: modern life is just too hard for individuals to handle alone. We need extensive cooperative mechanisms in order to do what we want to achieve. And the rationale for government is that sometimes those mechanisms need to be enforced.
KevinDC
Dec 2 2021 at 10:10pm
This argument, as you have presented it, is invalid, in the technical sense that even granting all the premises, the conclusion does not follow. Your argument is structured as a modus pones deduction, but it falls short of supporting the conclusion according to standard rules of logical inference. In order to entail that conclusion, you need to add in additional premises and assumptions, such as “intervention from politicians reliably causes more good than harm” and “the current politicians and bureaucrats operating within existing political structures are reliability capable of designing and implementing interventions that meet this criteria”, along with a few others, in order to reach your conclusion. Of course, what’s usually under dispute is whether politicians are in fact capable of reliably designing interventions that do more harm than good, both in general and with regards to Covid in particular. If your argument requires using that point as a premise, then it’s little more than a textbook case of question-begging.
KevinDC
Dec 2 2021 at 10:12pm
That last sentence should say “do more good than harm.” Ugh, seriously Kevin, learn to proofread.
Phil H
Dec 3 2021 at 2:09am
What’s the name of that law that says when you criticise someone’s grammar on the internet, you’ll immediately make the same mistake?
“you need to add in additional premises and assumptions, such as “intervention from politicians reliably causes more good than harm””
The part you quote doesn’t mention any politicians, so no amount of rigorous logic will require that they be inserted.
And the point of my reciting this somewhat tired and very well-known argument was precisely that. We have to go through it quite carefully, and see at what point it breaks down, and requires additional premises.
KevinDC
Dec 3 2021 at 9:54am
I’ve never heard of this law before but now I’m very much curious, it sounds like one of the internet’s more fun adages.
I’m aware that politicians didn’t appear anywhere in the quote I provided – in fact, that was literally my point. It seems I was unclear in making it, however, so I’ll take another stab at it.
First, the premises you listed out don’t entail the conclusion you stated. That is, even granting the premises you put forward at the start, the conclusion “Therefore, if significant harm is to be avoided, some form of intervention is necessary”, which you reached in the first half of your comment, is not logically entailed from those premises. There’s nothing in your premises about the effectiveness of intervention in avoiding harm, either in general or in this specific case, so your conclusion “if significant harm is to be avoided, some form of intervention is necessary” cannot be inferred from the premises you gave. This isn’t particularly “rigorous logic”, as you say, it’s just elementary rules of inference – your conclusion can’t contain anything not also contained in the premises. (This, incidentally, is why you can’t derive “ought” from “is” – in order for an “ought” to appear in a conclusion, it must be somewhere in the premises as well. An argument with only “is” statements in the premises, and an “ought” statement in the conclusion, is automatically invalid, because conclusions can’t contain anything not also found in the premises.) Even setting aside the issue of whether the intervention comes from government, the “classic paternalist argument” which you laid out is logically invalid at the outset.
But you also made it clear in the second half of your comment the specific type of intervention you wanted was government intervention specifically. In order to get to that further conclusion you endorsed, you would need to add in more assumptions and premises – that intervention designed and implemented specifically from state actors reliably provides more good than harm in situations like this, and that the specific state actors who will be implementing the actual policies are capable of designing these sorts of interventions, and so on. And policies do not originate from and are not implemented by abstract “governments” but are rather made and implemented by specific politicians and bureaucrats – hence why I phrased it that way when spelling out the additional premises needed to conclude that government intervention, which you explicitly said you would call for, was necessary or justified. It’s not enough to establish whether a specific, Platonic Ideal form of government could design effective interventions. In order to reach the conclusion that actual, existing government actors should take those actions, you need separate arguments to establish that the specific people empowered to design and carry out those interventions would do so effectively in the real world.
Also, I would note how you’ve said this before:
The prospect of how to design effective interventions in situations like the one we face now, while being able to be confident that the short and long term benefits will outweigh the overall costs, seems orders of magnitude more difficult to design and implement than the comparatively trivial task of “identifying skilled baseball players.” And yet, you also believe tasks like the second are so overwhelmingly complicated that the very idea of differences in aptitude is epistemologically irrelevant. If you don’t think institutions can be trusted to do something as relatively simple as identifying aptitude, I’m not sure how you can be so confident that institutions can reliably design interventions which will do more good than harm in cases involving far greater scope and complexity than that.
robc
Dec 3 2021 at 3:09pm
On the Reason H&R blog (RIP) it was known as joez law, after a leftist poster named joe. He was a lot of fun during the Bush years, but left in a huff in January of 2009 when overnight everyone turned from being anti-Bush to being anti-Obama.
MarkW
Dec 3 2021 at 8:31am
But the politicians (who also usually have poor numeracy skills) need to attract votes from the large fraction of population with poor numeracy. So they favor policies that please those voters, regardless of whether those policies are what their experts recommend. AND their experts become political operators who recommend and defend policies that will please their politician bosses (else they will find themselves on their way out). So….government is simply no solution to the problem of an innumerate populace.
zeke5123
Dec 3 2021 at 10:29am
I would say it is even worse than what you wrote.
There is, I think, a notion in intervention that all other things are equal. But, like everything in a complex system if you change one variable there are numerous variables that respond against that variable in often unpredictable ways.
We often make arguments, and they even may be logical and sound. It doesn’t follow they are correct. To quote a wise man from a bygone year — it is hard to make a prediction, especially about the future.
The key insight of de-centralized system is that given that unpredictable future result having numerous bets means more organisms within the structure die but the system is better off. Centralized systems in this regard work great when the bet hits, but terrible when it misses.
KevinDC
Dec 3 2021 at 11:47am
Recently, I read the book The Revolt of the Public and the Crisis of Authority in the New Millennium by Martin Gurri and he makes a similar point:
It is, of course, logically possible that the overall unintended consequences will be beneficial rather than harmful. But as a matter of sheer probability, this is very unlikely, as it’s far easier to make things worse than to make things better.
Oliver Wendell Holmes made a similar observation when arguing against the idea that judges should rule cases in a way they personally think would promote the “good of society” as opposed to simply applying a straightforward interpretation of the law, when he said:
While I have many gripes I could throw at Holmes, it would be nice if modern interventionists showed a semblance of this humility.
BC
Dec 2 2021 at 10:11pm
“Making a good decision on what kind/level of Covid protections to take requires good numeracy.”
That’s not necessarily true. For example, one does not need to know anything about valuing stocks to invest wisely. One needs only to understand that informed investors’ actions tend to cause stocks to become valued fairly relative to their risk and, therefore, that the index portfolio is a reasonable investment. Understanding the incentives of informed persons and the implications for interpreting their actions is often enough to determine one’s own actions. In fact, it’s often better than trying to determine one’s actions from first principles because one can learn and take advantage of what others know.
In the case of Covid, for example, we know that many public officials have consulted extensively with experts about the risks and those same public officials have also been seen many times not wearing masks in public, especially when they don’t know that they are being observed. From that, one can conclude that the risks of going maskless are probably not that great, at least if one is of similar health risk as those officials. One does not need to pour through the data oneself.
One can learn a lot just by observing other people.
Phil H
Dec 3 2021 at 2:12am
Two points to make:
(1) I see what you mean, that you don’t have to do the calculations for yourself, but then you do need some way of telling who is smart enough to do the calculations so that you can copy them. Either way, some heuristic mechanism is needed, and they seem to have been lacking.
(2) Masks don’t do that much to protect you. They do significantly reduce the risk of transmission to people around you.
Gene
Dec 2 2021 at 10:38pm
How about this govt intervention: announce, loudly & relentlessly, the stats that Henderson just presented? Low cost, harmless to freedom & our rights, and able to eliminate innumeracy enough to lead to better individual decisions. Even a modest 20% reduction in media & public hysteria since March 2020 would have had tremendous benefits.
Mark Z
Dec 2 2021 at 11:20pm
What if people are specifically innumerate in that they tend to overestimate very small numbers (such as the risk of shark attacks or plane crashes)? In that case, intervention may be akin to putting shark nets on beaches: more an expression of voter innumeracy than a solution to it.
Pajser
Dec 3 2021 at 2:54am
“Yes, we’re all in this together, but the probability of death from COVID-19 varies dramatically by age … the very young are bearing substantial costs of government measures on COVID-19 in return for tiny benefits to themselves. COVID-19 lockdowns, shutdowns of schools, and masking requirements not only indoors but also, sometimes, outdoors, are taking a huge toll on the very young that could well scar them for the rest of their lives. ”
I do not believe that online school or the obligatory wearing of masks causes noteworthy bad consequences for normal children in normal families. It is possible that there are children who for some other reason are on the verge of dysfunction and a small disorder can push them to the unwanted side of the edge. I don’t know how to quantify that problem.
But all normal children will suffer damage if someone in their family gets covid19 and dies or has very severe consequences. If children, by any chance, have infected that person, it will cause them occasional remorse and suffering for the rest of their lives.
Patrick Peterson
Dec 3 2021 at 9:44am
David,
You use the term and official stats “dying from COVID.”
But these statistics are HIGHLY suspect.
The confusion with the totally different thing “dying with COVID” can change the statistics (and real causal connection) by at least an order of magnitude, and the fear of Covid by an even higher number.
No?
David Henderson
Dec 3 2021 at 10:04am
You write:
Yes, that can change the statistics but I don’t believe it’s by close to an order of magnitude.
Todd Kreider
Dec 3 2021 at 10:14am
The odds that a healthy American child, 0 to 17 years old, dies from Covid is less than 1 in a million.
German doctors recently reported that there were no Covid deaths in healthy 5 to 17 year olds during the first 15 months. The odds that a healthy 5 to 11 year old needed hospitalization was 1 in 50,000 and for 12 to 17 year olds 1 in 8,000. Eight infants and toddlers died from Covid, five had preconditions. (The flu is also more deadly for 0 to 4 year olds than it is for 5 to 17 year olds.)
I estimated that an unvaccinated, healthy American has about a 1 in 20,000 chance of dying from Covid based on CDC statistics although maybe better to use a range of 1 in 10,000 to 1 in 30,000 based on the assumption of “healthy.”
————–
With respect to nuclear war: “Although the probability of such a war breaking out in any given year is almost certainly below 1 in a thousand, the hundreds of millions of deaths it would cause make nuclear war a suitable candidate for worry.”
Under 1 in 1,000 per year seems overly optimistic, although I hope it is sill under 1 in 100. While hundreds of millions might die in an all out nuclear war, there are also many scenarios with 1 million to 50 million dead – still a lot of people.
Todd Kreider
Dec 3 2021 at 10:21am
Oops, I meant that for a healthy 50 year old so the sentence should be: “I estimated that an unvaccinated, healthy 50 year old American has about a 1 in 20,000 chance of dying from Covid based on CDC statistics although maybe better to use a range of 1 in 10,000 to 1 in 30,000 based on the assumption of “healthy.””
Thomas Lee Hutcheson
Dec 3 2021 at 9:47pm
Policies and messaging about COVID have been marred by inadequate attention to population heterogeneity with respect to the costs and benefits of spread prevention measures.
Grand Rapids Mike
Dec 6 2021 at 9:07pm
The press is constantly hyping Covid deaths without citing relevant statistics. This is aided and abetted by politicians who can’t wait to install means to control everyone. It would be nice if the the press would provide coherent set of data and informing the public on who is at risk etc., and providing simple but informative info on the probability of dying by age and other sickness issues, due to Covid. However, that would be too rational. Right now the press is a tool of politicians.
Remember how rediculous the tulip mania was. This is in the same category.
Comments are closed.