Highlights From The Comment Thread On School Choice

Several people including Yehoshua K and Freddie deBoer point out that “nonprofit” and “for-profit” are potentially meaningless terms in situations like these.

IrishDude adds some context to the for-profit hospital scene by noting that companies are not allowed to open new hospitals until they apply for and are granted a Certificate Of Need, apparently on the basis of a theory that an oversupply of hospitals would increase (?!) costs.

Algorizmi describes his work with a private school that costs less than half as much as most public schools, including how it saves money

Half a dozen people yell at me for saying that the grocery industry worked well for the poor, objecting that I had forgotten about Food Deserts. Various other people save me some time by pointing out that most of the claims about Food Deserts are kind of fake (1, 2, 3). There’s a defensible version of the term, which is that in very poorly planned car-centered zoning-regulated cities without good public transportation it’s not always possible for someone without a car to easily get to the stores they want, but just posing the problem that way makes the solution pretty obvious.

Spotted Toad is great as usual, pointing out among other things that Obama has already led the most pro-charter administration in history.

Douglas Knight proposes that college demand curves are upward-sloping, which is kind of terrifying but seems to have at least anecdotal support.

Robinson: “The things needed by poor people, if done well, will never be money-makers.” Matt M: “A bold claim considering that the biggest company on Earth, more than double #2, serves the needs of poor people almost exclusively.”

MDP has worked in the payday loan industry and explains how their interest rates can be so high and their profits so low.

Many people (1, 2, 3) point out the high cost to schools of misbehaving kids, and try to explain the rise in education costs by saying that kids today are raised wrong (or not at all) which makes them harder to control. But everybody who looks for this kind of thing finds the opposite – kids today have less teenage pregnancy, crime, dropouts, et cetera. I don’t know if anyone has specifically looked at classroom misbehaving, but it would be weird for that to be getting worse in such an isolated way.

Lots of people are very angry at me for posting the graph from the Cato Institute for various reasons. A few people object that it is dishonest because it didn’t adjust for inflation, even though it did adjust for inflation and is very clear about that. A few people object that it is dishonest because it puts cost increases and score increases on the same scale as each other, instead of skewing them to make the effect look bigger than it really is – this is a definition of “dishonest” I haven’t heard before (maybe I am being uncharitable here?). Several others say that test scores have increased more than they give credit for. You can look at a very good summary of test score changes here. You’ll see there are some gains among younger students, but much fewer among older kids – for example, 17 year olds’ math scores went from 304 in 1971 all the way to 306 in 2012. This is not just a race-related Simpson’s Paradox – among white students alone, for example, the gain was 4 points (though blacks did gain 18, as I mentioned before). I am not sure what good it is to have high gains in early years if those gains are all lost by the time kids leave school. Overall I think the Cato graph comes out looking pretty good. But even if you disagree, I would ask you to pick whatever metrics you want – your favorite test, racial group, axis scheme, whatever – and tell me whether it really looks like the doubling-to-tripling of education costs during the relevant time period has been money well-spent. If not, then as happy as I am to debate details in the comments, it all seems like basically nitpicking.

Jonah Katz brings up a really complete analysis of increasing costs in higher education:

The rising cost of higher education isn’t quite so mysterious, at least for the last 10-15 years. The Delta Cost Project has put together some fairly comprehensive data about this. What you see across most categories of post-secondary institutions is that basically *everything* is becoming more expensive, but ‘student ‘life’ and ‘academic support’ are rising fastest, followed by ‘institutional support’. Student life is all of the bells and whistles (athletic centers, movie theaters, etc.) that colleges use to try to entice prospective students into paying huge amounts of money to enroll in their institutions, and I believe it also includes health and mental health services, which I would imagine have become exponentially more expensive over the past couple decades (this is probably unavoidable, because health costs are going up in general and universities are enrolling a far wider range of students with more mental and physical health issues who wouldn’t have gone to college in the past). Academic support includes a mix of stuff that is crucial to the academic mission of a university (libraries, IT systems), stuff that is arguably not part of the core academic mission at all (Dean’s Office personnel, museums), and stuff that is well intentioned but tends to be useless in practice (central offices for teaching and curriculum development). Institutional support is administration proper. Note that these data come from 2003-2013, so they don’t capture the explosion in university administration that is generally agreed to have occurred from roughly the 1970s to 1990s. I’ve never been able to find categorized data that goes back that far, but I imagine the change in spending on administration during that period must have been astronomical. The cost of instruction is still the largest single category of expenditure, and accounts for the majority of absolute price increases, but proportionally it is not rising as fast as these other categories. Also, the NY times Op-ed piece you link to is either selectively pulling misleading data or is just plain ignorant about the state of public financing for higher ed. There has not been a ‘modest’ reduction in per-student funding: it has dropped around 30% in inflation-adjusted dollars since 2000.

Static brings up the role of pensions as a driver of schooling cost increases.

Swing finds there have not been similar cost increases in the Netherlands. And Politifact rates as true a similar claim – that we greatly outspend other First World countries in per pupil spending. This article notes that we spend about 25% more than Britain and almost 50% more than Germany. On the other hand, the Netherlands is only a little better than we are, so this doesn’t match a scenario where the Netherlands’ spending goes up by only a little but America’s goes up by 200%. I don’t know where the discrepancy comes from. {EDIT: Douglas Knight points out that most countries have per pupil spending as a similar percent GDP)

Various people chime in with their favorite anecdotes about school vouchers working very well (DC) or working very poorly (Sweden). Murphy describes a personal bad experience with school privatization. At some point I do want to go through and sum up all the empirical literature on this, but not now.

Many people mention the possibility for bad incentives or market failures in schooling. 1soru1 thinks that, absent better signals of quality, schools will compete on shininess and raise prices to have the biggest and most breathtaking stadium (which I think is what the post above was saying happened to colleges, so certainly plausible). Tanagrabeast describes finding the private schools in Arizona heavily politicized: “What Scott worries about is already happening. I was skeptical until I took my son to an open-house at a fast-growing chain of charters where they tried very hard to play conservative buzzword Bingo and did all but lead us in a prayer to the Founding Fathers.” In contrast, Doctor Mist says that as a rightist, he feels like going to a private school lets him escape what he sees as public schools’ existing liberal politicization.

EarthSeaSky is a purist and reminds us that the free market which can be named is not the eternal free market.

Steve Sailer argues that for-profit colleges are a natural comparison group for for-profit primary schools, and they are very bad.

I talked to Education Realist on Twitter. Their position is complicated but they recommend their posts The Fallacy At The Heart Of All Reform and Charters: The Center Won’t Hold as introductions/summaries. I am still not entirely clear on their position – the objection seems to be that successful charters succeed only by taking the best students who would get good test scores anywhere, then claiming charters raise test scores. Obviously charters are trying this, but every halfway-decent study on charter schools has tried to control for this possibility. Also, none of my points involved empirical claims that charter schools raise test scores, so I don’t see why this discredits me in particular. They also note that US education is already pretty good both compared to other countries and compared to its own past, something else I agree is true and have never denied.

Levarkin brings up James Tooley’s fascinating work on private schools for the poor in Third World countries.

And Justreggedthis on the subreddit makes what I find the most convincing argument in this whole discussion:

Sweden’s experiment with school vouchers showed a different problem: the market delivers what you want, not what you need. What (stupid) parents want is good grades. What kids need is good education. So precdictably, voucher schools ended up diluting grades. You can probably imagine how it works. We live in an age of narcissism. Many parents want to hear their kids is super, special, and a genius, and get straight-A grades for a performance that is at best average. Few parents have the character left to stand up to it, and want challenging education and honest grading.

Of course this is a problem with people, not vouchers. I am sure the very same narcissism in modern culture also rears its head in public schools as well.

The classic solution was school principals having low time preference and interested in preserving the long-term good name of their school. So they would not agree to grade dilution, they would not encheapen the brand of their school.

Seems like today time preferences are high.

Grocery stores are a good parallel. You need healthy food. You want (a stupider version of you wants) gallon buckets of ice cream. Hence, you get all kinds of special offers and discounts on tasty and cheap gallon buckets of ice cream. The market delivers what you want. Hence, obesity epidemic.

I don’t really know any solution that is acceptable within a democratic framework. Obvious someone somewhere should override personal preferences, but that someone should have a very good set of incentives and that is what we don’t get in this framework.

Overall reading this has made me somewhat more pessimistic about charter schools. But I’m still uncertain enough that I want to look into the empirical literature more, and I still think careful experimentation is the way to go.

…so maybe I should end with shadypirelli’s comment from the subreddit pointing out that Betsy DeVos’ policies cannot be described as “careful experimentation”.