Cost/Benefit analysis of School Closures in the US

post by · 2020-07-10T23:35:47.865Z · LW · GW · 8 comments

TLDR: Covid 19 school closures may do more harm than good, because years of schooling affects overall life expectancy and quality of life for millions of children.

This is my very first post here, so if I apologize if I say something that has been said many times before or that does not seem to belong on this forum. That said, here goes...

Coronavirus school closures and the move to online schooling has resulted in a very significant decrease in the quality of education in the US. In person schooling provides a lot of important resources, especially for low income students, that cannot be easily accessed in a poor home environment. Online education also makes it much easier to cheat, and this provides students with less incentive to learn. Additionally, since many parents work full-time, this also means that students have limited adult supervision throughout the day. Thus, I'm going to assume for the purposes of this argument, that online schooling is roughly equivalent to no schooling at all.

There have been many studies showing that additional years of schooling correlate with longer life expectancy, higher income, and lower incarceration rates. Now obviously, you might worry about a correlation/causation problem with that kind of statement. However, there have also been several twin studies that help eliminate this bias. These studies follow identical twins who grow up in the same home but have different levels of educational attainment. The results of theses studies still show a statistically significant difference between the incomes of twins who have different educational status. Now, these studies are not perfect, but there have been several of them and they tend to get similar results. Thus, I think it's fair to conclude that years of schooling, particularly at the high school level, have a positive impact on income. There are no twin studies that I know of regarding education and life expectancy specifically in the US, but I think that if education impacts income, it likely impacts life expectancy in a statistically significant way.

Another thing to note is that the impact of education, and higher quality education is often most significant for low income children.,of%20adult%20poverty%3B%20effects%20are

Now, it stands to reason that if years/quality of schooling has a significant impact on both the quantity and quality of life of students, then we must accurately take that into account when deciding the costs and benefits of school closures. It is difficult given the currently available data to accurately estimate exactly how much school closures will affect children.

Let's say hypothetically, one year of school closures takes 1 year off of every child's life on average. There are 56 million school children in the United States right now. That means that 56 million years of life are lost with such a policy.

On the other hand, scientists estimate Covid 19 kills approximately 1% of all people who get the virus. (In the US the number is higher but that is likely due to lack of testing and asymptomatic cases) Herd immunity is also estimated to be approximately 66% of the population. 66% of 350 million is 231 million. 1% of 231 million is 2.31. This is a total of 2.31 million deaths. However, 8 out of 10 covid deaths are in people 65 years or older. Many of the deaths occur in people with very serious pre-existing health conditions. 45% of deaths occur in people over the age of 80. Given that life expectancy in the US is approximately 79 years, let's assume that on average people who die of covid would have lived 15 years longer had they never contracted the disease.

Now let's assume school closures reduce corona cases overall by 85% until the development of a vaccine. That means the total number of lives saved due to school closures is 0.85*2.31million=1.96 million lives. Now, to calculate total years of life saved we multiply 1.96*15=29.45 million.

29.45 million years of life saved. At the expense of 56 million years of life lost for today's school children. Thus, it would seem that the better policy would be to keep schools open.

Now I'm aware that I am making a huge amount of assumptions in this analysis. But the overall point of this post is to show that it's possible that school closures will do more harm than good. Nobody can perfectly predict the future, and there is a lot of data that is missing that I would need in order to make a more accurate analysis. If you think my analysis is missing some important details please let me know. I would love to hear some other opinions.


Comments sorted by top scores.

comment by wolajacy · 2020-07-11T20:37:42.035Z · LW(p) · GW(p)
Now, these studies are not perfect, but there have been several of them and they tend to get similar results. Thus, I think it's fair to conclude that years of schooling, particularly at the high school level, have a positive impact on income.

This study takes into account only university attendance, on a yes/no scale, measuring impact relatively to non-attending twins.

(side note: the second link is broken, but judging from the address, it points to the first study, probably a typo).

From the second study (about K-12 funding)

Event-study and instrumental variable models reveal that a 10 percent increase in per-pupil spending each year for all twelve years of public school leads to 0.27 more completed years of education, 7.25 percent higher wages, and a 3.67 percentage-point reduction in the annual incidence of adult poverty.

This paper again talks about the boost relative to underfunded students.

I would also say those two say very little about high-school level.

As far as I understand, the whole argument against schooling (which for example Caplan makes) is that it serves mostly as signalling, i.e. benefits from schooling are relative to the position of other members of the society. By limiting it across the society, you are thus not losing much - even if you find the correlation between school-ness and success in life later in the normal conditions, you should not expect it to be present if everyone is handicapped in the same way.

Also, no school is probably qualitatively different than heavily-underfunded school, as more funding can just remove the horrible-ness of the environment. The paper talks about the relative advantage of students left to do their own thing ("unschooled"), compared to normal ones - and finds that going to school actually gives you just 1 year boost in education, which is, to say, not much.

Replies from:
comment by · 2020-07-13T11:02:35.404Z · LW(p) · GW(p)

Interesting points. However, I think the study you referenced is a bit small to draw any significant conclusions from, given that there were only 12 unstructured homeschoolers. I also think that those homeschoolers are probably not representative of your average American middle class/low income student, who might have their parents (or single parent) working full-time and thus unable to homeschool them much at all. As far as I could tell, even the the "unschooled" kids still got a lot of guidance from their parents.

I think the signaling theory has some truth to it, especially at the university level, but I'm personally very skeptical of it at the the k-12 level. Schools allows societies to more efficiently distribute childcare, which is economically advantageous. It allows parents (especially mothers) to return to the workforce. It effectively allows adults to pool their resources together to make educating and looking after children a lot less time consuming and expensive.

Nevertheless, I think you bring up an important point. No school might not be that damaging for some kids. If schools continue to stay closed, I would guess we would see the following two things happen. 1) kids with rich/upper middle class parents might actually do better because their parents can afford to custom tailor their education. 2) Kids with poor/working parents will do worse because there will be basically no incentives for them to learn on their own and their parents will be unable to teach them. Thus, inequality between the two groups will increase.

Replies from: wolajacy
comment by wolajacy · 2020-07-18T18:27:41.448Z · LW(p) · GW(p)

You're right about the small sample, I read the paper some time ago and forgot/didn't notice that.

I think you're also right about the K-12 school purpose as being more a daycare facility than anything else. I wouldn't necessary expect custom-tailoring the education on a large scale - parents still have to work (from home or not) or need to take care of many other issues araising from the current situation.

My prediction would be that the inequality araising from that factor would stay ~the same, because first, I feel that education on that level doesn't matter that much in the long run, second, because many extracurricular activities for middle/upper class are not available, and third, when it does matter, there is a greater focus on producing quality learning materials online, which further bridges the gap.

comment by korin43 · 2020-07-11T22:02:45.702Z · LW(p) · GW(p)

The thing I wonder about these studies is the correlation vs causation question. Is it the case that people who do more schooling make more money and live longer, or do people who are always going to make more money and live longer more likely to stay in school?

Replies from: Pablo_Stafforini
comment by Pablo (Pablo_Stafforini) · 2020-07-12T14:16:24.790Z · LW(p) · GW(p)

The post addresses this worry:

you might worry about a correlation/causation problem with that kind of statement. However, there have also been several twin studies that help eliminate this bias.

There is, however, another worry unaddressed by those studies, which wolajacy raises in their comment [LW(p) · GW(p)]. This is the debate between the 'human capital' and 'signaling' theories of education, covered extensively in Bryan Caplan's book, The case against education. Even if years of education cause—rather than correlate with—increased quality of life and length of life for individual people, reducing years of education for the population as a whole may not reduce those measures much if signaling is the main causal mechanism.

Replies from: korin43
comment by korin43 · 2020-07-12T18:55:03.488Z · LW(p) · GW(p)

I don't think twin studies entirely answer this since genetics isn't everything. I've known several genetically-identical twins and none of them had identical personalities.

Replies from: Pablo_Stafforini
comment by Pablo (Pablo_Stafforini) · 2020-07-12T21:27:22.068Z · LW(p) · GW(p)

Of course genetics isn't everything. This is recognized in the third law of behavioral genetics. Researchers who rely on twin studies do not assume otherwise.

comment by Pattern · 2020-07-11T17:17:15.764Z · LW(p) · GW(p)

Interesting username. (The email.)

Let's say hypothetically, one year of school closures takes 1 year off of every child's life on average.

I wonder what the actual number would be, absent corona.