The most ridiculous metaphor of education courtesy of an economics professor

Send to Kindle

Acclaimed academics have policy agendas just like anybody else. And often they let them interfere with a straightforward critical analysis of their output. The monumental capacity for blindness of highly intelligent people  is sometimes staggering. Metaphors and analogies (same thing for metaphor hacking) make thinkers particularly prone to mis-projection blindness. Edward Glaeser, a Harvard economics prof, is just the latest in the long line of economists and blowhards, who think they have the education system licked by comparing it to some -free market gimmick. They generally reveal that they know or care precious little about the full extent of the processes involved in the market and Glaeser is a shining example of this. His analogy is so preposterous and only needs so little thought to break down, I can’t believe he didn’t take a few minutes to to do it himself. Well, actually, I can believe it. It’s pretty,  neat and seductive. So who cares that it is pure non-sense. Here’s what he said:

Why Cities Rock | Freakonomics Radio: I want you to just imagine if for example, instead of having a New York restaurant scene that was dominated by private entrepreneurs who competed wildly with each other trying to come up with new, new things. The bad restaurants collapse; the good restaurants go on to cooking show fame. You have these powerful forces of competition and innovation working. Imagine instead if there was a food superintendent who operated a system of canteens where the menus were decided at the local level, and every New Yorker had to eat in these canteens. Well, the food would be awful. And that’s kind of what we decided to do with schooling. Instead of harnessing the urban ability to provide innovation, competition, new entry, we’ve put together a system where we’ve turn all that stuff off, and we’ve allowed only a huge advantage for a local, public monopoly. It’s very, very difficult to fix this. I think the most hopeful signs, and there’s been as you know a steady stream of economics papers on this, the most hopeful signs I think are coming from charter schools, which are particularly effective in urban areas. And it’s not so much that the average charter school is so much better than the average public school, but rather that in charter schools, because they can go bankrupt, because the can fail, the good ones will succeed, and the bad ones will drop out of the market. And certainly we’ve seen lots of great randomized studies that have shown the ability of charters to deliver great test score results.

As we know, metaphors (and their ilk) rely on projections from one domain to another. Generative metaphor (of which this is one) then try to base a solution on a new domain which is the result of the blending of the source domain.

So this is how Glaser envisions the domain of New York restaurants: there is competition, which drives up the quality of the food (note that he didn’t mention driving down prices, lowering expense per unit, and other tricks used by ‘wildly competing entrepreneurs’). Restaurateurs and chefs must strive to provide better food than others because there is so much choice, people will flock to their competitors for the better food.

This is how he wants to project it into schooling: give people more choice (and means to exercise that choice by using the intra city’s short commutes) and this will result in competition, the competition will increase experimentation and as a result the quality of education goes up. He also mentions test scores at the end but these have little to do with education (but why should somebody at Harvard know that?).

Of course, he makes most of the argument through a reverse projection, where he asks us to imagine what the New York restaurants would look like if they were run like a centralized public school system. He envisions the end process as similar to Apple’s 1984 commercial: a sea of bland canteens with awful food. But this is just so much elitist blather. Glaeser should be ashamed of himself for not thinking this through.

First, what he describes is only true of the top tier of New York restaurants. The sort of places the upper-middle glass go to because of a review on Yelp. The majority of places where New Yorkers (and people everywhere) eat their lunches and the occasional dinner are either real canteens, some local greasy spoon, or a chain that makes its consistent prices and dining experiences possible through resolute mediocrity. The Zagat guide is for special occasions, not daily nutrition.

Second, Glaser never asks how this maps onto schooling or education, in general. Probably because the answer would be that it doesn’t. Glaeser certainly refused to say anything useful about his analogy. He went far enough to promote his shallow ideology and stopped while the stopping was good. Let’s look at a few possible mappings and see how we fare.

So first we have the quality of the food. This would seem to map quite nicely onto quality of education. But it doesn’t. Or at least not in the way Gleaeser and his like would like.  Quality of the food that can impact on competition is a surface property. We cannot also always trust people that they can judge the quality apart from the decor of the restaurant or its reputation – just like with wine, they are very likely to judge the quality based on a  review or the recommendation of a trusted acquaintance. In Glaeser’s analogy, we’re not really talking about the quality of food but the quality of the dining experience. And if we project this onto the quality of a school, we’re only increasing the scope of the problem. No matter how limited and unreliable, we can at least judge the quality of the overall dining experience by our own reaction to our experience. But with schools, the experience is mediated through the child and the most important criterion of quality – viz an educated human being at the end – is deferred until long after the decision on quality has been made. It’s like judging the quality of a restaurant we go to for an anniversary dinner by whether we will be healthy in 5 years. Of course, we can force such judgements but arbitrarily ranking schools based on a single number – like the disastrous UK league tables that haven’t improved the education of a single child but made a lot of people extremely anxious.

The top restaurants (where the competition makes a difference) don’t look at food from the perspective of what matters for life, namely nutrition. It’s quite likely the most popular restaurants don’t serve anything particularly healthy or prepared with regard to the environmental impact. Quality is only important to them as one of many competitive advantage. They also use a number of tricks to make the dining experience better – cheat on ingredients, serve small portions on large plates, etc. They rely on ‘secret recipes’ – the last thing we want to see in education. And this is exactly the experience of schools that compete in the market. They fudge, cheat and flat out lie to protect their competitive advantage.  They provide the minimum of education that they can get away with to look good. Glaeser also conveniently forgets that there is a huge amount of centralized oversight of New York restaurants – much more, in some ways, than on charter schools. Quality is only one of the results of rampant competition and oversight is necessary to protect consumers. This is much more important in schools than in restaurants (but it almost seems that restaurants have more of it, than schools – proportionally to their importance).

But that is only one part of this important mismapping, which is the process of competition. Many economists forget that the market forces don’t work on their own. They work off the backs of the cheated and underserved. Bad restaurants don’t go out of business by some market magic. They go out of business because enough people ate there and were cheated, served bad food or almost got poisoned. And this experience had to have been bad enough for them to talk about it and dissuade others from visiting. With restaurants the individual cost is relatively minor (at least for those comfortably off). You have to contribute one or two bad meals or ruined evenings a year to keep the invisible hand doing its business among the chefs of New York. (This could be significant to someone who only goes out once every few months but still something you can get over.) Also the cost of making a choice is virtually nill. It takes no effort to go to a different restaurant or to choose to eat at home. Except for the small risk of food poisoning, you’ve made no investment in your choice and the potential loss is minimal.

However, in the case of schooling, you’re making a long-term commitment (at least a semester or a year but most likely at least four years). You can shrug off a bad meal but what about a wasted half-a-decade of your child’s life? Or what if you enrolled your child in the educational equivalent of Burger King serving nothing but giant whoppers. Everything seems fine all along but the results are clogged intellectual arteries. Also the costs of a school going out of business (and here Glaeser is one of the honest few that admit to bankrupt schools as a desirable outcome of competition in education) are exceedingly high. Both financial and emotional. Let’s say a school goes out of business and a poor parent has invested in books, school uniform and transportation choice only to have to start this again in a new school. Or how about the toll that making new friends, getting used to new teachers, etc. takes on a child. How many ruined childhoods is Glaeser willing to accept for the benefits of his ideology? As far as I know, the churn among New York restaurants is quite significant – could the education system sustain 10% or even 1% of schools going out of business every year.

And more importantly what about great schools going out of business because of financial mismanagement of capitalist wannabes? Not all market failures (maybe even not most) are due to low quality. Bad timing, ruthless competition, impatient investors and insufficient growth have killed many a great product. How many great schools would succumb to one of these? And won’t we see the same race to mediocrity once the ‘safe middle ground’ of survival is discovered? How many schools will take on the risk of innovation in the face of relentless market pressures? For a Chef, one bad recipe is just a learning experience. For a school, one unsuccessful innovation can lead to absolute disaster.

But all that is assuming that we can even map the “quality of education” onto quality in any sphere of commercial activity whatsoever. What business do you get a product or service from for four or eight years that requires a daily performance of a complex and variable task such as caring for and educating a young person is? Not your electricity provider who provides a constant but a non-variable service, nor your medical care provider who offers a variable but only periodical service. Also, “the consumers of education’s” requirements keep changing over time. They may have wanted a rounded and fulfilling education for their child at the start but just want them to get to university at the end. You can measure quality by test scores or graduation rates but that still doesn’t guarantee success for roughly 10-20% of students even in the best of schools.

To conclude, fine food plays a role in the prosperity of restaurants but so does convenience and habit. The quality of education is too complex to map successfully on the quality of food (and possibly any single commercial product). And even if that was possible, the cost of making the market forces work is incomparably higher in education than in dining. Glaeser’s proposed model for reform is just as likely to produce pinnacles of excellence as ruthlessly competitive MacDonald’s-like chains of garbage.

There’s nothing wrong with using metaphors to try to look for ways to improve education. But generally, these should be very local rather than global and always have their limits carefully investigated. That means detailed understanding of both domains and meticulous mappings between them as well as the relationships between them. Not all mappings need to be perfect and some need not be there at all (for instance, computer virus is still useful metaphor even though it doesn’t spread through the air), but this should be done consciously and with care. Steve Jones once said of evolution that metaphor is to it like bird excrement is to statues. The same often goes for education, but it doesn’t have to.

Finally, this analysis didn’t necessarily imply that the current system is the best there can be or that it is even any good (although I think it’s pretty good). Just that reforming it based on this cock-a-maney metaphor could be extremely dangerous. New solutions must ultimately be judged on their own merit but with the many market metaphors, very many their merit is irretrievably tied to the power of the initial metaphor and not any intrinsic solution.

UPDATE: It seems I may have a been a bit too harsh on Glaeser. Obsevational Epidemiology posts this quote form his book (the one he was promoting on the Freakonomics podcast):

All of the world’s older cities have suffered the great scourges of urban life: disease, crime, congestion. And the fight against these ills has never been won by passively accepting things as they are or by mindlessly relying on the free market.

Ok, so he’s not just a mindless free-marketeer. So why on earth would he suggest the above as a suitable metaphor to base educational reform on?

Send to Kindle
  • Jospeh

    I made the same comment over on my website, but I figured it was appropriate here as well.

    I have a dark suspicion that Dr Glaeser will accurately diagnosis many of the problems (with things like educational policy) but will then go on to suggest overly simplified solutions. Chapter 1 seemed good but I am already worried at the direction on Chapter 2 (as the shift towards the metaphor you gave is already starting in the way he diagnoses failure).

    But the metaphor above is absolutely braindead . . .

    • Dominik Lukeš

      Yes, this is an incredibly common problem with metaphor-based policy. The beauty and revelatory power of the superficial fit is too strong for a mapping analysis. Market-based metaphors in particular seem to be impervious to all evidence because of the perceived success and moral superiority of the source metaphor.

  • Pingback: Hagwons as restaurants « Surprisesaplenty's Blog

  • Michael Bishop

    I’m sorry, but when you’re arguing against the government giving people freedom of choice (whether in schools, restaurants, sexual practices, cars, houses, etc.) the burden of proof is on you.

    • Dominik Lukeš

      Hm. I’m not sure that there is any inherent rule of argument that places the burden of proof on a specific ideology, as opposed to the other. Why not request the burden of proof from the position that choice is a good thing? (See here:

      But that’s actually not what I’m doing. I’m just showing 1. that it’s silly to expect the same kind of choice in school selection by parents as we see in dining choices, 2. that, even if we did, we need to properly understand how the choice works in the restaurant/public dining domain before we can use it as a source domain for reforming schools. Just saying “restaurants + choice = better food and therefore schools + choice = better education” is not very productive. We also need to understand schools enough to show that there actually isn’t choice already that is comparable to restaurants. The restaurant metaphor completely excludes people who don’t care about their quality of food because they can’t afford to eat in them. So what sort of choice do they have? And we know that people who can afford private tuition have all the choice they can buy. So how is it different at all? Why aren’t there calls for voucher or charter programs to help people eat in fancy restaurants? Well, there are food banks and soup kitchens but who’d want to eat in those?

  • Wonks Anonymous

    I agree that metaphors are a bad way to argue (I’d say Glaeser is implicitly relying on an economic model which abstracts from the particulars of restaurants), but this was a really poor argument. If you think his randomized studies don’t actually show education matters for test scores (an argument I could find plausible), actually argue for that proposition rather than asserting it when he claims to have those studies backing him up. And you brought up the “dining experience” without considering that the “educational experience” (not measured by test scores, but important to parents and children) might be a dimension of competition! Finally, we do have competitive institutions which perform the daily variable task of educating young people, an even closer analogy than daycares (which first popped in my head): higher education! They tend to be non-profits, but so are charter schools. The American higher education system has a much better reputation than K-12.

    On Michael’s burden-of-proof argument: there are many, many, many areas where we prefer allowing people to choose their provider of goods/services. In most areas where we don’t it’s simply impractical (standard example is national defense, which is non-rival and non-excludable). I have a hard time thinking of any areas where we regard choice as a bad thing and think it essential that people be assigned based on residence rather than getting to choose if that’s an option. So what’s the argument against choice? Does even Barry Schwartz think people should be assigned choices rather than making them themselves?

    • Dominik Lukeš

      First, I don’t think that metaphors are a bad way to argue. I think they are the only way to argue (see the About page) but just like any argument, they can be used badly or well (D. A. Schön is probably best on metaphors and policy).

      Second, I’m not necessarily saying that we could not use a market metaphor for education. Just that to use it like economists tend to, doesn’t make any sense. But there are many reasons why universities are not a very good example for this. 1. They are only marginally educational institutions – going to one is just as important as learning something in it (a place to grow 4 years older near a library, some research and someone telling you something they seem to feel is important). 2. They are essentially boarding schools – which makes the competition much more feasible. Prim/secondary schools are more like US cable companies – they have to have reasonably local monopolies. Also universities are not judged by standardized performance tests – for them, profit is actually a good measure of success (non-profit institutions my ass!). 3. Related to 2, students at universities are trying to achieve a speciality distinguishes them from others, whereas high school students are learning to be like everyone else. So they can only differentiate to an extent.

      Day care is also not a very good example because it doesn’t try to achieve economies of scale. If you had schools that would only take 50 students each year, you could have local competition – as long as they did not compete only on price.

      I’m absolutely certain that schooling matters a great deal for test scores. You go to a better prep for a test, you’re going to do better on the test. But that means pretty much nothing in terms of education as presented by proponents of test scores. But Glaeser’s randomisation is a joke – he can never control for all the factors that are involved and never include all the other nexuses of which the test scores are just one. We’re at the point of statistics being used to flat out lie, here.

      About choice: I grew up under a communist regime where we supposedly didn’t get any choice. But we did in lots of respects. There are always choices to be made albeit within a proscribed context. But this limitation was no more limiting than being poor in a capitalist society (choice of travel, education, etc.)

      In the “democratic” societies we limit choices in as many places as we give them. We just don’t pay any attention to the limits. No choice to pay taxes, no choice not to have a social security number, no choice to drive on a particular side of the road, no choice in the interpretation of the green traffic light. No choice to hear swearing on broadcast TV, no choice to elect a politician outside a given number of parties. No choice to urinate in public. Many of these may seem ridiculous but are they? The lack of choice is given by external forces or by the configuration of the situation. And to introduce choice in school selection without reconfiguring the situation is going to be a bad thing.

      But even where you think there is a choice, many people don’t. I spent many years preferring reading and lazy thinking to making money. So going to a supermarket full of rows of choices presented no options for me at all. I had to buy the cheapest stuff. I could not go to the friendly local shop, I had to cycle 3 miles to a supermarket with discounts. I never ever went to a restaurant! About a third of the people in any country live like that. You give them a choice of schools, you give them no more choice than they have now. Vouchers are a joke unless you spend at least twice as much per student on a student from a poor family – CHOICE COSTS MONEY!

      But most importantly, I still don’t believe that school choice would make any difference at all to the quality of education overall. You cannot take the success of charter schools (of which there is a lot less than the hype indicates) and say we would have better education if we only had charter schools. Education reform based on zeal does not scale up! If you want to improve education, balance out the choices. Take away some from the rich (taxes and enforcement of white collar crime) and give some more to the poor – free school lunches, free universities, free healthcare, less incarceration, more development of local communities, more jobs. You can then ignore schools altogether. Historically, education has followed economic progress not the other way around!

    • TGGP

      “Bad restaurants don’t go out of business by some market magic”
      A common story for how markets aren’t perfectly competitive is that established businesses already have a brand-name and it is difficult for new entrants to get customers in the first place, as people by default may just continue making the choice that has been satisfactory in the past rather than taking a risk. None of this establishes that choice is actually a problem, since less choice wouldn’t actually prevent bad experiences (it would merely prevent the feedback from bad experience leading to bankruptcy).

      “They are only marginally educational institutions – going to one is just as important as learning something in it”
      Makes sense if you accept the signalling model of education. That might then segue to edu-nihilism, which suggests that simply spending less on the wasteful practice makes sense (price competition uber alles!). From what I’ve heard, charters cost less, so score one for them. The signalling model also explains why universities advertise the ex ante standardized test scores of their students rather than any ex ante performance (which they could presumably affect).

      Why is competition more feasible for boarding schools?
      Why must prim/secondary schools “have reasonably local monopolies”? The fact that they are monopolies now is an artifact of public policy. The same is true of cable monopolies. Even roads, which you refer to later, are not “natural monopolies” but were provided competitively in the past before policy changed. See Thomas Dilorenzo’s “The Myth of Natural Monopoly”, which should be available online.

      “whereas high school students are learning to be like everyone else”
      Says who? Again this seems like an artifact of current policy whose desirability hasn’t even been attempted to establish.

      “If you had schools that would only take 50 students each year, you could have local competition – as long as they did not compete only on price”
      Why couldn’t you have local competition if they competed only on price? If edu-nihilism is correct, that may be the only reasonable criterion to judge schools.

      “But Glaeser’s randomisation is a joke – he can never control for all the factors that are involved and never include all the other nexuses of which the test scores are just one”
      The whole point of randomization is you don’t have to know what factors are involved and statistically control for them. You create a meaningless distinction between subjects, tie that distinction to the treatment, and then any differences between groups should be the result of the treatment plus random noise (canceling out with large sample sizes). The issue of how meaningful test scores are is orthogonal to that of randomization. Part of the idea behind the “choice” movement is that people themselves would choose which nexuses they think relevant.

      “profit is actually a good measure of success (non-profit institutions my ass!)”
      From what I’ve heard the really elite schools can entirely rely on their endowments and don’t actually have to seek out other means of profit. There is of course lots of differentiation in the field of higher ed, competition can tend to produce that.

      Since I’m fairly ignorant of the daily reality of communist societies, I would be interested to hear in what sort of ways you did and didn’t have choice.

      “No choice to pay taxes”
      That would be the case if we had a head-tax (which I could understand serious arguments for), but most of our taxation is highly dependent on decisions we make. So what we lack is the option of making certain decisions without paying taxes for them. Still, a meaningful restriction on choice.

      “No choice to hear swearing on broadcast TV”
      Carlin’s dirty words seem relatively tame now, and even broadcast tv has many of them. And of course we can choose other sources of tv than broadcast.

      “no choice to elect a politician outside a given number of parties”
      There are electoral rules that usually require some number of signatures to get on ballots, but generally in national elections there are a number of parties on the ballots which hardly anybody chooses to vote for (regarding it as a waste). This tends to happen in a first-past-the-post system, but even Duverger’s Law has been dismissed by political scientists recently, particularly after the previous general election in the U.K.

      The examples you gave were mostly about how you are treated by the law and your inability to avoid the law. If we are going to discuss the appropriateness of metaphors, is that like schooling? Or is schooling (which I would suggest thinking of in the abstract to better perceive other possibilities than the status quo) more like attending a church (which has tended to be linked to private k-12 education), or daycare, participating in little league, or the other metaphors discussed above?

      “And to introduce choice in school selection without reconfiguring the situation is going to be a bad thing.”
      It could be good, bad, or indifferent. I haven’t seen you give any argument for one over another.

      “I spent many years preferring reading and lazy thinking to making money”
      Sounds like a choice!

      “About a third of the people in any country live like that”
      That leaves two thirds, which is not only not nothing, it’s twice as much as one third! And furthermore, the bottom tertile does make choices as consumers, and their choices may even form a significant component of the identities they craft for themselves.

      “Education reform based on zeal does not scale up!”
      True enough. I suppose the hope is that somebody will figure out a method other than zeal, whether that hope is well-founded I don’t know. It may call for experimentation.

      “If you want to improve education, balance out the choices. Take away some from the rich”
      You seem to regard this as zero-sum. It is in a pure signalling model, but I don’t know if that’s your reason. For a purely positional good (an extreme variant of the signalling model) we would tax education rather than spending taxes on it, and even a complete ban might be an improvement.

      “Historically, education has followed economic progress not the other way around!”
      I believe that’s also correct, and much of the enthusiasm for education is cargo-cult. But another possibility other than edu-nihilism, is that greater wealth has made possible greater education, which doesn’t grow on trees.

    • Dominik Lukeš

      Hey, thanks for taking the time. Just a few points:

      1. “cargo cult” – see my somewhat stagnant project on Education as Voodoo (

      2. I am most emphatically NOT saying that choice is a problem (although, it is not an unadulterated good or a moral sine qua non). But I am saying that frequently, choice is an illusion. Charter schools are pretend choice.

      3. I don’t think primary/secondary schools have to be monopolistic intrinsically (also in any one areas they tend to be duopolies) but at current economies of scale, it is pretty much the only choice (thus the analogy to US cable providers in the 80s and cell companies at the moment). So while there may be some boutique alt edu providers in many areas they do not put any competitive pressure on the monopolistic ‘catchment’ schools. Many suggest that going back to the system of small schools would be a good idea – but I suspect it would also be a really expensive thing to do. And the problem with the market approaches to education is that they are also obsessed with saving money.

      Nothing is intrinsically monopolistic – not even the police, armies or tax collection (see Ancient Rome) – but there are areas where it is beneficial for the state to have at least a monopoly of regulation (electricity provision, spectrum interference, POTS, etc.) even if the provision of a service allows for competition on certain surface aspects of the service.

      4. You make valid points about the weakness of parts of my argument about lack of choice in some areas. I’ll have to think on this some more .

      5. I’ll leave the description of growing up in a “communist” country for another blog post.

    • TGGP

      I wrote a comment (currently in moderation) that grew rapidly. But I still forgot to include one bit: a reference to Albert O. Hirschmann’s “Exit, Voice and Loyalty”. I left the book as I came to it, an Exit chauvinist dismissive of Voice and Loyalty. I won’t link to my discussion of it at my blog, for fear of tripping the spam filter.