A bunch of folks got together at Cato for lunch to gang up on Alex Tabarrok. Recall that he and Eric Helland want to claim that the high prices of health care and education are almost entirely due to the Baumol Effect.
I offered as an alternative hypothesis that much of the higher prices can be accounted for by the government subsidizing demand and restricting supply. Here are some notes, based on the discussion.
1. Health care spending has been rising at the same rate in most developed countries. Can government policy be the cause everywhere? On the other hand, in most developed countries the proportion of health care spending paid for out of pocket is low, to perhaps there really are not such significant differences in policy across countries.
2. Veterinary care prices have been rising even faster than human health care prices. Yet there are no government subsidies for pet care.
3. Incomes for other high-skill professions, such as accountants and lawyers, also have risen sharply. Point (3) sounds like it could support the Baumol-effect story, but on closer examination it is more problematic.
4. A pure Baumol Effect would raise wages in every occupation where productivity growth is slow, including for barbers and waiters. That has not taken place.
5. It is difficult to account for the vast difference in pay between adjunct professors and tenured professors. At one point, I asked “Are adjuncts idiots?” That is, are their skill levels so dramatically lower than those of tenured professors?
6. Another question to ask is, “Are college administrators idiots?” In theory, it would seem that you could create a university with all courses only taught by adjuncts and offer a low-cost degree. That this does not happen shows how difficult it is to compete in higher education.
Overall, I still suspect that the story of “It’s all a Baumol effect” is an intellectual swindle. It is tautologically true that in a two-good world, if the relative price of good X falls, then the relative price of good Y goes up. But it is not necessarily the case that the price of good Y has to rise relative to *the* wage rate. In fact, the opposite seems more likely. But the actual ata seem to show that prices in health care and education have gone up faster than wages. I have a hard time coming up with a two-good, homogeneous-worker general equilibrium model that can exhibit that behavior.
In fact, when it comes to talking about wages, Tabarrok pulls a switch and starts talking about the wages of high-skilled workers, so we are no longer talking about “the” wage rate. Instead, we are talking about a skill premium. So Tabarrok has already added an epicycle, as it were, to the Baumol Effect story. That is, he has grafted on a skill premium.
I can more easily fit the data to a story that includes a skill premium as well as a Baumol effect. But then one can argue that this skill premium depends in part on regulations that protect credentialed workers. It is amplified by demand subsidies for education and health care, which put government in the role of enhancing the skill premium.
Arnold,
Thanks for sticking with this topic, which might be a watershed in policy analysis.
Thanks also for your helpful clarification of the concatenation of mechanisms: subsidize demand, restrict supply.
As you and other scholars have noted, there are several additional mechanisms that dilute the explanatory power of the Baumol effect, re: high prices in education, health care, and some other sectors:
• Demand is delinked from outcomes. (Arnold Kling)
• Category creep; e.g., amenities in education. (Arnold Kling)
• Government production of the output, especially in the education sector, induces supply of gold-plated private alternatives. (Bryan Caplan.)
• Status affiliation: Prosperity increases demand for services of prestigious professionals. (Robin Hanson)
The description seemed ill defined, it did not hint at a mechanism, thus no solution.
My approach was to narrow down the definition to a single mechanism, the rescaling that happens in the economy.
These are helpful points to consider. The Baumol effect likely depends as well on whether service providers hold direct responsibility for their real estate costs, which only become more substantial in regions where such costs go well beyond actual construction value. By way of example, barbers would need to make relatively more than waiters due to their responsibility for building rent, but not near as much as physicians who are particularly responsible for real estate costs and valuations, even if not directly. As to the latter, while such costs weren’t always an important factor, they now are entrenched to the point it would be difficult for governments to backtrack on subsidized demand and restricted supply, since this revenue not only reimburses human capital investment but also maintains real estate value. Put another way…the cost of the knowledge conduit is too high!
Your approach was to assume that a complex system had a single and easy to identify cause?
In some labor markets there is a large “incumbency bias.”
One example offhand: the public school teacher with 15 years on the job has the position and apparently can keep it. No one can say “hire me for his/her position and I’ll do the job next year for $10k less than your current teacher is paid.” This seems to result from a variety of things simultaneously–it’s not just unions, but in addition staffing arrangements, licensing, and employment contracts in which people “work cheap” in their early years and tunnel toward a defined benefit retirement pension at the end.
Presumably some teachers are worth every penny, some are definitely not, and it’s hard to tell. The best students, especially smart ones from high-functioning families, often are good at educating themselves–so it’s hard to tease out the impact of teacher quality.
Plus “stickiness” in the assignment of students to schools based on residence in a district, and the lack of price competition in terms of a price paid by the consumer for public school education–stickiness plus opacity plus other factors I can’t unpack and articulate on the fly.
The incentives to lower the cost of higher education aren’t really there in the best districts (Brighton and Pittsford NY). And apparently, the incentives to improve the results aren’t there in the worst districts (Rochester, NY). Meanwhile the public just hears “We need more money”–and putting in more money is like pushing on a string, it seems to get sucked up by the vast assortment of “producer interests,” and results don’t seem to materialize in any predictable manner.
= – = – = – =
The adjunct situation is a bit different–I think we would have to analyze things on a case-by-case basis, going one department at a time. The English literature adjunct could replace the tenured professor in English more easily than the chemistry adjunct could replace the tenured professor in Chemistry and continue to run an expensive lab and keep it going with a steady stream of research grants. Do you want to have a pipeline of grants on into the future?
Tyler Cowen said something about chess instruction, perhaps in _Average is over_. Computer aided instruction spread very rapidly in chess teaching in part because there were no Chess Departments at universities. It didn’t happen at the same rate in university Foreign Language Departments. I would add that we are gradually seeing this happen now, but with long lags and with the abrupt collapse of some foreign language departments.
In the USA I think you can/will see this both in German (because it’s a bit hard and low-demand) and (Italian because it’s low demand and the easiest Romance language for a highly motivated native English speaker to learn on their own). The place to look for evidence about best practices relative to the cost of instruction is going to in places like Foreign Language Institute and Defense Language Institute–what are they doing there in terms of automation vs. retaining highly paid warm bodies.
And look abroad: does Israel Defense teach Persian to soldiers using (1) poorly paid adjunct staff, with (2) highly paid professors, or (3) with computers. Or do they not have the problem yet because of a large pool of native speaking Persian Jewish immigrants.
It may be in part that chess geeks are highly competitive and the tournament rankings help in chess. Meanwhile, perhaps much language instruction in primary / secondary / college simply aims to get people over a low bar in order to check a box for distribution requirements. Maybe.
= – = – = – =
A peculiar example of non-incumbency is musicians in an orchestra, where the 2d violinist can challenge the 1st violinist for the top position and call a matchoff. This may be a isolated case.
= – = – = – =
Health care economics is beyond me. Offhand I like the argument of the “Grumpy Economist” (John Cochrane) that there is so much n-tier pricing and cross-subsidizing in the USA that we see large organizations charging every penny they can get–from anyone who can/will pay it– in order to cover state mandated losses elsewhere in the system.
Perhaps this is apropos
https://johnhcochrane.blogspot.com/2019/01/the-death-of-healthcare-market.html
There is a gigantic difference between people who want to learn to be good chess players or Israelis in the IDF who want to learn Persian and the ordinary high school student.
The ordinary high school student does not have a burning desire to learn what the school (and the state Ed Department) say he is supposed to. His desire is to get a good enough grade to pass and/or get into a good enough college. There is no reason to assume that what will work for the first group will work for the second.
And, of course, the chess players and the IDF members will actually use their knowledge. The high schoolers pretty much won’t, and will forget most of it fairly quickly.
Why is it that economists love to discuss explanations of overall sector behaviors like this when so many of them also believe some version of Hayek’s ideas, which would indicate that such explanations are impossible?
It seems so much more useful to focus on individual policy decisions and vectors, which are usually much easier to analyze and explain, rather than to come up with oversimplified explanations of complex systems like this.
When dealing public policy, changes in entire sectors actually can be affected by relatively simple causes, precisely because public policy usually isn’t Hayekian and often applies rules or subsidies broadly across a somewhat arbitrarily defined sector, creating sector-wide patterns. There’s no reason to limit oneself to highly particular policies. If a million departments and agencies regulate a million different kinds of goods or services in a given sector, perhaps in a million different ways, but usually subtly different ways, I see no reason to avoid examining general sector-wide patterns and treat each particular as though there were no similarities between them.
*This was supposed to be in reply to Tom DeMeo
The discussion is on the claim that the high prices of health care and education are almost entirely due to the Baumol Effect.
It is not whether there a Baumol effect is present in these sectors. It isn’t whether there are sector-wide patterns.
Notice the words “almost entirely”. There is no way you can buy Hayek and proceed to argue that it is possible to unpack the underlying causes of pricing levels across a major sector.
Swindle is a rather uncharitable word, isn’t it?
My feeling is that the Baumol effect is a necessary pre-condition for rising real prices, but it isn’t sufficient by itself.
It’s hard to find an example of a sector with rising prices which *isn’t* influenced by the Baumol effect to a large degree, so it seems obvious that it’s a driving force.
You could probably prove its necessity as a logical construct. If an industry has some sort of pricing bottleneck with regards to its labor inputs and isn’t captured by the Baumol effect, then said bottleneck will simply be traded or automated away.
So it’s hard to have rising real prices with labor inputs that require a skill premium in a way that can’t be arbitraged.
But there are enough edge cases that it seems a little bit of a leap to attribute the *entire* cause of rising real prices to Baumol.
Considering this is a long term trends stating it is only X (higher demand/lower supply) versus say Baurol Effect misses trends come multiple micro sources.
1) Higher healthcare:
a) Yes Higher Demand/Lower Supply with various government and private programs. Haven’t most health insurance moved to high deductible program?
b) Baurol Effect of higher end employment.
c) The reality of healthcare increased productivity is more lives saved which does increase healthcare cost. There is cost savings productivity but it is very limited. Cancer survival rates have increased the last several decades but cancer is expensive. Throw in other diseases as well.
“…much of the higher prices can be accounted for by the government subsidizing demand and restricting supply.”
I’ve heard exactly this argument explaining why house prices and rents are so high in places, i.e., the government makes programs to make buying a house easier thus increasing demand, than puts zoning restrictions that won’t allow an increase in the supply of houses.
Robust supply response seems pretty predictive of less severe price increases. Perhaps you think something other than housing policy is holding down supply in some places where government promotes and subsidizes housing? But it seems like there’s pretty strong concordance between cities that have weak supply response to increasing demand and cities with relatively stringent housing restrictions.
Regarding point 5 and 6, isn’t tenure (and the associated pay/status) the prize that keeps adjunct pay low? If you replaced all the tenured professors with adjunct professors, many of the adjunct professors will quit and go do something else. There’s a similar dynamic with law firm associates and partners (though there it’s more hours worked than pay).
I think there are a number of fields which are “tournament” fields, where all the rewards accrue to a few individuals at the summit, but where there are many more people competing to get to the top. Entertainment, athletics, higher education. In most of these fields, life below the top tier is unpleasant. It’s great to be a major league athlete. It’s terrible to be a minor league athlete.
University professorships are a tournament field, and behave exactly like you would expect a tournament profession to behave.
How’s this for a possibility:
* The US, in both public and private sectors, is expanding debt at an extreme rate and keeping interest rates very low. This has predictably inflationary effects in non-tradeable sectors of the economy.
* China is maintaining higher interest rates for its own reasons, resulting in lower inflation.
* China is keeping its currency pegged at about 7 renmimbi per $.
* This arrangement would quickly implode if China weren’t willing to accumulate enormous dollar reserves (currently over $3.1 trillion (yes with a “t”)).
I wrote two pieces that are somewhat relevant, although I’m not an economist.
First, Charles says:
the public school teacher with 15 years on the job has the position and appaIrently can keep it. No one can say “hire me for his/her position and I’ll do the job next year for $10k less than your current teacher is paid.” This seems to result from a variety of things simultaneously–it’s not just unions, but in addition staffing arrangements, licensing, and employment contracts in which people “work cheap” in their early years and tunnel toward a defined benefit retirement pension at the end.
It doesn’t result from any of those things. the minute Job security is a major attraction of teachers. Without it, don’t kid yourself: all teachers, great or otherwise, would be fired FIFO. And then no one would be teachers anymore, cheap or not.
And by the way, why is it we have a teacher shortage?
I wrote about this here:
Second, Arnold wonders why a college administrator wouldn’t hire a bunch of low-status adjuncts and run a low cost college. Why not wonder instead why low status adjuncts stay teaching college when tenure is no longer a possibility?
I wrote this in response to Alex’s article (not really about Baumol):
“Why don’t these adjunct professors rebel against the crappy pay and insecurity and move down to high school level teaching?”
Partly because they can’t. They don’t have K12 teaching degrees / teaching certificates (they’d need to return to school for those), and their advanced degrees make them unattractive candidates for many public school districts with unionized teaching corps. Districts often don’t want to hire inexperienced teachers with advanced degrees because union contracts require teachers with advanced degrees to be paid significantly more. Put yourself in the shoes of a school district administrator. Would you rather hire an (expensive) 35-year-old burned-out academic who’s applying for a high-school teaching job not because that’s what she really wants to do but because she has run out of options? Or a (cheaper) new 22-year-old graduate who genuinely wants to be a high-school teacher?
Note that this administrator is selling education as the key to financial success to politicians who hate taxes. The guy with 9 years of college who applied for a lower-middle class job because he’s out of other options can’t help but clash with the sales pitch.
As a PhD who has gone through periods of unemployment, I can give some responses:
* People usually don’t get PhDs because they want to teach. We usually want to do research. We generally have had very little training in how to teach (although we’re usually excellent learners) – my own “teacher training” was about an afternoon and basically boiled down to “don’t get sued”.
** We usually aren’t extroverts. We can usually handle teaching eight hours a week, but maybe not eight hours a day.
** In a college environment, we usually get by. We have deep knowledge of the subject area and students are both experienced and above average (but still they sometimes horrify us). Teaching kids to read or dealing with a classroom of unruly teenagers require skills we’ve generally never cultivated. Presumably.
* STEM PhDs who leave academia usually have more attractive options elsewhere. A lot of my friends went into banking or technical sales, or started companies.
* PhD programs are extremely unstructured. You can do whatever you want, as long as nobody’s ever done it before. You can work any hours you like, as long as you’re clearly working far too much. It’s a bad culture fit to public education.
* We often have little experience of, interest in, or (frankly) sympathy for the academically disadvantaged populations that the education bureaucracy seems obsessed with. Our political views are often idiosyncratically heterodox.
Slocum and Jay above are missing the point. I’m well aware of the reasons why failed tenure PhDs don’t become teachers, although the idea that a k-12 administrator would turn down a PhD because it interferes with the message is kind of hilarious. As for the fat that they don’t have a credential, the teaching shortage is so dire that they could get a full time job while getting it.
What I am saying is that k-12 teachers salaries can’t be considered in terms of pay based on education or selectivity, because any pay comparison of college adjunct teaching and k-12 teaching shows that the latter is far superior in every method of considering benefit, and the former is far more selective. And that people who think they understand pay issues in k-12 generally don’t know what the hell they’re talking about.
Fair enough. I don’t pretend to understand pay issues in K-12 education, other than to say you couldn’t pay me enough.
Let’s step back. The question posed is, “Why are the prices so damn high?”.
Which prices? Consider, the Palladium price has spiked in the last few years, and Bitcoin is back up. But these aren’t the prices people are complaining about: the ones that are matters of widespread and major social concern.
Ok, those are pretty volatile, but another example close to my heart is beef, which on average has been gradually increasing for about 20 years now, at about double the inflation rate. Also, cable TV prices have beaten inflation for about as long, despite the fact that there has been a steady and significant drop in bandwidth prices at the wholesale / backbone level. Plenty of complaints about both of those prices, sure, but not the kind of things worth this much fuss.
So, what’s worth the fuss? The stuff that we spend big chunks of our total incomes on, the prices of which have been growing faster than typical wages for a long time.
1. Health Care
2. Education
3. Government (both “spending per household” and “prices of goods and services mostly bought with taxes”)
4. The Rent
Obviously #4 is not generated by any Baumol-type explanation. Tabarrok would have a distinct explanation for that, and would agree with “subsidize demand, restrict supply”, especially the “restrict supply” part.
You could go to some heroic lengths and try to squeeze some of #3 into a purely Baumol story, but really it doesn’t work. “Restrict supply” doesn’t make a lot of sense here either. You could come up with yet more distinct explanations, like “political distortion”, or “regulatory burdens”, or “monopsony problems”, etc., and even “all of the above”.
And of course Kling has provided those possible alternatives for even #1 and #2, the least implausible-on-their-faces Baumol-story sectors, to include the claim that these are simply “normal goods” the spending on which should be expected to increase as people get richer. Also there is the general “Third Party Payer” problem.
But if we had a single story with the explanatory power to account for all those sectors, then that should be preferable to one which only accounts for a few, and which, as one can tell from the response and all the commentary, is full of holes and simply wasn’t ready for prime time.
An a single story which does this is “Ricardian Rents”. In a Ricardian Rent situation, some important product is both scarce enough and indispensible enough such that the price will rise to whatever level buyers can afford. So the consumers can become much more productive in real terms, and the real prices of other goods can fall substantially, but the prices in the Ricardian sectors will simply rise to neutralize any new surplus, because people will have to use their higher real incomes to bid up the prices in those sectors, or simply go without.
The classic example is of course land rents when calories are scarce in Malthusian circumstances. They’re not making any more land, increasing agricultural productivity was tough, fertility rates were high, and people gotta eat. Hence, to a first approximation, arable land was wealth.
Ok, so what’s indispensable today? Well, certain kinds of “Health Care” when needed, especially if one has a life-threatening but treatable ailment. Why are those prescription drugs under patent (literally a government-granted license to monopoly rents) so expensive? Well, how much would you pay for that insulin or epinephrine shot when you need it?
Also, how much would you pay to stay out of prison for tax evasion? Well, for most people, “If I knew I’d get caught evading, I’d pay my taxes, whatever they are” is the answer. It’s kind of like being mugged at gunpoint while unarmed: what else are you really going to do besides hand over the money?
What about education? Actually, the price of acquiring knowledge and skills has dropped substantially. The internet: check it out. But no employer cares about how many Wikipedia articles you’ve read, and none of your social acquaintances are going to be impressed with the number of YouTube DIY videos you’ve watched, even if you really can remember everything about Roman History, and even if you now have the skills to repair everything in your house.
The alternative to educational credentials, is, for most people, to be effectively socially disqualified from the vast majority of lucrative professional opportunities, and the status of those credentials correlates with the number and value of those opportunities. Hence, high status educational credentials are like a license to have a good shot at earning a lot of money, and how much would you pay for that license?
And now we come to The Rent. Just like in education, the rent actually isn’t high at all, and in many areas it’s dirt cheap and has even fallen in the past few decades. Ah, but people aren’t really complaining about all the rents, they are complaining about what is actually a miniscule portion of livable geography, which is quality neighborhoods relatively close to the center of the Big Winner Cities.
Again, one can talk about zoning and building permits all the day, but the overall pattern is almost the same everywhere in the world. What’s more likely, that it’s a regulatory issue, or that some kind of global economic phenomena is at work, which is best explained by changes in technology, which tend to quickly produce effects everywhere.
And everybody knows that an increasing number of opportunities for higher paying jobs are located in the Big Winner Cities across the world. Why? Because technological change and globalization mean that the allocation of the labor force is shifting towards things that can only be done by local humans, which means tasks which require those humans to stay in close geographic proximity to each other, giving rise to economies of agglomeration, and a tendency for industries to become centralized in one place, as a kind of national or regional “capital” for that sector.
As an aside, the “Handle-Baumol Theory” (thanks to Kling for naming it) says that because of this technological development, real labor productivity growth rates should noticeably slow down, because activities which can’t automated can only be done at human speed, which doesn’t increase over time. The barber does the same number of haircuts per hours, and so forth. If more and more people are doing that kind of work, then most people aren’t getting any more productive than their parents or grandparents doing similar work. Also, if there’s less opportunity to increase output by augmenting labor with lots of capital (what are you going to give the barber, super-fast snipping scissors?), then overall demand for investments should decrease, which impact on real interest rates and savings and lending patterns.
You are in a Big Winner City if one of the sectors with a large portion of regional or national revenue has its “capital” in your town, as the anchor industry and “primary producers” of the wealth that supports the rest of the local economic ecosystem. E.g., NYC: Finance, LA: Entertainment, DC: Government, Bay Area: “Tech”, Houston: Petroleum. In many countries, most everything ends up even more centralized in one place.
So, in order to have access to most of those opportunities, most people will have to live relatively close to the center of gravity, because economic connections between humans who have to work with each other in close proximity – for which closer is often better since it saves scarce time – are like a spiderweb with tension in the strands tending to pull everything in.
And living in a place means paying “The Rent” there, as a license to have access to those opportunities? What would you pay? “How much is that license?” – “How much you got?” “I’ve got the gap between the opportunities here, and the next best opportunity not in a Big Winner City”. That gap is large and growing, so The Rent is high and rising.
The overall picture is that the reason there is so much spending in these sectors is that there is no good alternative to paying the pipers, and so the pipers are able to suck up any potential surplus, whatever it happens to be, or however it changes over time.
This is the Neo-Ricardian Theory (thanks to Kling again) explanation for “Why the Prices Are So Damn High” in the sectors we really care about, and it works for Tabarrok’s industries, and the others too.