The Economist covers Why Are the Prices So D*mn High?

The Economist does a very nice job covering Why Are the Prices So D*mn High.

Baumol’s earliest work on the subject, written with William Bowen, was published in 1965. Analyses like that of Messrs Helland and Tabarrok nonetheless feel novel, because the implications of cost disease remain so underappreciated in policy circles. For instance, the steadily rising expense of education and health care is almost universally deplored as an economic scourge, despite being caused by something indubitably good: rapid, if unevenly spread, productivity growth. Higher prices, if driven by cost disease, need not mean reduced affordability, since they reflect greater productive capacity elsewhere in the economy. The authors use an analogy: as a person’s salary increases, the cost of doing things other than work—like gardening, for example—rises, since each hour off the job means more forgone income. But that does not mean that time spent gardening has become less affordable.

It’s an implication of the Baumol effect that everyone ends up working in a low productivity industry!

The only true solution to cost disease is an economy-wide productivity slowdown—and one may be in the offing. Technological progress pushes employment into the sectors most resistant to productivity growth. Eventually, nearly everyone may have jobs that are valued for their inefficiency: as concert musicians, or artisanal cheesemakers, or members of the household staff of the very rich. If there is no high-productivity sector to lure such workers away, then the problem does not arise.

Misunderstanding the Baumol effect can lead to a cure worse than the “disease”:

These possibilities reveal the real threat from Baumol’s disease: not that work will flow toward less-productive industries, which is inevitable, but that gains from rising productivity are unevenly shared. When firms in highly productive industries crave highly credentialed workers, it is the pay of similar workers elsewhere in the economy—of doctors, say—that rises in response. That worsens inequality, as low-income workers must still pay higher prices for essential services like health care. Even so, the productivity growth that drives cost disease could make everyone better off. But governments often do too little to tax the winners and compensate the losers. And politicians who do not understand the Baumol effect sometimes cap spending on education and health. Unsurprisingly, since they misunderstand the diagnosis, the treatment they prescribe makes the ailment worse.

My only complaint is that the excellent reviewer has not followed our lead and called it the Baumol effect–cost disease is a misleading name!

Addendum: Other posts in this series.


"And politicians who do not understand the Baumol effect sometimes cap spending on education and health."

So the conclusion of your Baumol cost disease analysis is that politicians should subsidize demand in healthcare and education?

Arnold Kling won't be happy.

The notes above state that politicians should redistribute the gains, not that those specific industries are subsidized.

Note that redistribution also decreases the wealth and economic impact of the winners, reducing the cost increases they create.

Donald Trump is the greatest President in the history of the USA. If you love America, vote for your God Emperor Trump in 2020 and make the salty cucks cry again. #maga2020

Yang as the only politician that is even remotely addressing this issue.

Yang is arguing gdp must go down, just like Alex, except Alex reffuses to state GDP must go down?

After all, if all prices go down, then incomes must go down, thus consumption must go down, thus production must go down, ie, GDP goes down.

Food costs have gone down even with industry finding way to hike the price of cheap commodities, and thus food is a much smaller share of GDP, probably the lowest in the world in the US.Yet increased consumptions of calories per person has not offset reducced costs of food and thus preventing total food consumption from falling as share of the economy.

Even massive diversion of food into a higher cost mandated replacement of fossil fuels has failed to prevent food GDP from falling.

If the cost cutting in food over the past century were applied to housing, transportation, energy, health care, GDP would be half of today's level., even as quantities consumed of everything are cut.

Typo in the first line.

As I'm sure someone has pointed out, the cure for unproductive segments is substitution. Sure, the productivity of concert musicians has not increased, but I haven't been to a concert in years. My favorite version of Beethoven's 9th was recorded in 1981 and I listen to it on my phone.

Likewise, MRUniversity is a substitute for unproductive higher education.

The problem with Federal money flowing to these sectors is that they subsidize inefficiency. That and that alone supports the increase in employment. So grants to higher education are jobs programs and nothing more. They don't improve education, the discourage the rise in more effective education.

You favor lower GDP and your lower income that goes with it?

Ie, why should you be paid any more than the musiciaans, the ticket sellers, and takers, and ushers, and janitors, and construction workers building theaters, you see virtue in not paying?

Surely you don't want them producing something that costs you more than what you would have paid them a century ago when there was no cheaper substitute, other than singing, dancing, etc yourself.

In his column today, economics journalist Robert Samuelson (not related to Paul) says we need to tear up all economics textbooks and start over, arguing that the textbooks have been overtaken by events. The Baumol effect is but one gap; indeed, I should mention that the Baumol effect serves as a rationalization for the high level of inequality.

There's no doubt that the economy has experienced an enormous shift to so-called tech, but has the shift actually accomplished much? Sure, there's digital advertising and the promise of artificial general intelligence, but has the shift made America great again (to borrow a phrase)? Does Samuelson's criticism of economics textbooks extend to the fundamental principles that underlie the textbooks, such as that markets produce the best results and the appropriate "winners", "best" in the sense of the most efficient allocation of a scarce resource (i.e., capital)? Or to rephrase Samuelson, have the theories that underlie economics textbooks been overtaken by events?

I'm quite certain that the priniciples of the economy circa the Americaan Revolutionary War, eg, Adam Smith, Locke, et al, and circa WWII, eg, J. S. Mill, Marshall, Pigou, Keynes, et al, apply today.

I doubt that a any of several first year economics textbooks from circa 1960 would provide a great basis for teaching economics for business people, politicians, bankers, the educated citizen, and with reductions in costs in printing and electronic media, would cost the same or lower dollar price of 1960. About 5%-10% of the Tyler-Alex textbook, which is the best example of "cost disease" around.

Where is the analysis of the massive cost bloat in economics textbooks from the authors of a massively bloated cost Econommics textbook?

Will Mankiw's textbook sell for printing cost in paper for, or be the cost of download in electronic forms now that he's retired and no longer producing anything? Or he merely drive the cost-disease by demanding payment for doing zero work, as if farmers deserve payment for producing the food we ate 20 years ago today based on we being alive today only because farmers producted the food preventing our starving to death 19 years ago.

Prices are so darned high because people pay them.

Generally, I don't pay above what I want - except that new gun . . . The wife handles that BE for both of us. It's OK! She's spending my money. Same same for higher prices in education and health - it's perceived as other people's money.

Governments face little or no budget constraints. Real time: each year, our local school budget expenditures rise above the so-called inflation rate and is voted down. Damn the empty nesters! Then, the PTA et al gear up and it passes on the re-vote.

Governments and school districts have limits on what can be 'produced' and consumed.

Health care, education, etc. present common features: lack of competition, opaque pricing, heavy subsidies, regulatory constraints on entry/supply. Solutions - (Never happen!) government controls of spending for education and health; and wage controls.

In addition to education and health, child car seats, cribs, etc. serve as examples of "price disease" - government regulation and "What About The Children!!!" Sorry. I was in accounting and finance where cost is different than price.

Off-topic: How many guns you own, DtheB? You one of these gun freaks that has 1000 guns in the basement? And do you carry one in the chamber or empty? The ASP channel had a good video the other day about a crazed robber who charged the store owner despite being fired at in broad daylight (rare) and the defender almost lost his life (the robber was armed with a gun) since he had an empty chamber, something about 'tap and rack' (defender didn't have to tap but out of habit he did, costing him a precious half second). Good outcome for the defender however (he lived, the robber took the "room temperature challenge" as the ASP host John calls it). All caught on video, good red-neck stuff, somewhat educational.

"The Baumol effect" is just supply and demand for skilled labor. It's not new or insightful. This "spin" of making it sound like having massive and wildly unproductive sectors of the economy like education and health care is a good thing is nuts. There is much more going on in those sectors than competition for skilled labor.

"The Baumol effect" is just supply and demand for skilled labor: I think you might be right, see if you wish to answer my question below.

Mystery to me how someone as prolific, innovative, solid and long-lived as Baumol got passed over for a Nobel.


I've been struggling with the issue of why barbers and bus drivers have seen their wages stagnate even though there has been no productivity gains in their sector. Their pay gap relative to doctors and nurses has been multiplied by 3 (fig. 19).

The Economist's article offers a different explanation than Baumol, it seems. Instead, the author claims that "teachers and engineers compete in the same labour market." The idea is that, since productivity has increased in engineering, so have the wages of engineers and, by extension, of teachers. This is a story of increased demand for skilled labor vs unskilled labor, not a story of higher productivity gains (engineers) vs lower productivity gains (teachers).

Another way to see it is that fig. 22 shows a simplified economy producing only cars and education. Bus drivers and barbers are like teachers: they have had no productivity gains so one would expect their wages to increase relative to cars. But this is not the case, contrary to what a Baumol effect would lead to expect.

Am I missing something?


@ Stephane with an accent mark - it's hard to tell what you're missing, since I don't have the article in front of me, but it seems you're making various assumptions. Perhaps your difficulty is semantics? Consider your sentence: "The idea is that, since productivity has increased in engineering, so have the wages of engineers and, by extension, of teachers. This is a story of increased demand for skilled labor vs unskilled labor, not a story of higher productivity gains (engineers) vs lower productivity gains (teachers)." - now substitute "higher productivity" for "skilled" and "lower productivity" for "unskilled", and you get Baumol's effect. Simply put, Baumol's effect is why farming in the USA is something like 4% of GDP despite the enormous importance of being fed: productivity in the agricultural sector has made farming irrelevant, hence no "national income" in it (nevertheless I bought some ADM stock the other day, and so should you, reader).

Bonus trivia: the beef I have with AlexT's reading of cost disease is this: usually, like Stephane implies above, Baumol's effect is deployed to show why low-productivity people--call them 'teachers' or 'musicians' end up with a high salary compared to high-productivity people, call them 'engineers', but I'm not sure AlexT brought out this distinction (without carefully reading him, that's my 15 second take).

Dunno, but in one of the several other posts describing this supposedly simple one-size-fits-all concept, he said "In just the same way, to the extent that greater spending on education, health care, or car repair is due to the rising opportunity costs of inputs we should not expect any increase in quality." Teachers could be engineers but have to be dissuaded somehow. And: "Similarly, higher productivity in say goods production increases income at any given production level and people choose to take some of this higher income in services."

But our resident physician Sure replied with a pretty convincing post that medical care even in the 1960s compared to now is hardly the same service, nor, in a sense, serving the same population.

"Barbers" include licensed ladies' hairdressers, yes? It's true that many women pay large sums for stylish haircuts and treatments like having their hair smoothed and flattened, or things like weaves (and now, with particularly unfortunate results among the hoi polloi, the application of a great red spot, or a purplish wave). But on the other hand, overall there has been a trend toward more natural-looking styles ... my grandmothers spent an hour each week at the beauty shop having their hair coiffed and frosted, even little boys used to have their hair trimmed or buzz cut like clockwork, but no longer ... coupled with an expansion of other ways to treat-yo'self. The styling of hair was all my grandmothers paid for (albeit weekly). Now there's massages and chiropractic and acupuncture all available within a mile of me. There's a place around the corner where you pay to do stretches like ballerinas (next door to the place where you can get your infant's brain balanced - I haven't had the heart to learn more about that - I have to live among these people). Nearby I could get facials, waxing, or lasering off of hair, eyebrows shaped with needle and thread (! and ?), lips plumped, fake eyelashes interspersed with existing ones. There's a salon where the hairdresser doesn't cut the hair of those who don't cut it themselves - she only washes and blows it out! Nails, fingers and even toes, polished by immigrant ladies; teeth whitened. Skin firmed with botox.

Not to mention tattoos. I would be surprised if there's another tattoo-less woman under fifty in my zip code.

When I think I've grasped Baumol - musicians! that one's easy! - I haven't, and now I can't remember the point of this post: I guess I'm wondering if this expansion of ways to pamper and prettify yourself has any bearing on the case.

@ Ray
Indeed, there is probably an issue with semantics: does “increasing productivity” mean increasing quantity of stuff produced or increasing value of stuff produced? It seems to me that Baumol focuses unduly on quantities: we can now produce more cars with fewer hours of labor. But if this caused the Baumol effect, we would expect the wages of teachers and barbers alike to go up since it still takes the same time to produce a one hour lecture or a haircut as it did 40 years ago. It’s not what we see, that’s why I am puzzled. If instead we focus on the value of the stuff produced, then productivity has increased in teaching because the value of almost all degrees has gone up. On the other hand, the value of a haircut has not changed.

A good part of the supposed inflation in healthcare is that we now do much more of it. If you drill down and examine the cost of many procedures you will find they have decreased in both nominal and real terms over the last forty years. The example I am most familiar with is cataract surgery. It is now done at a fraction of its 1970s cost with vastly superior results. And eye surgeon pay has remained steady because the typical caraactbsurgeon is five times more productive than his 1970s predecessors.

Total spending on eye surgery has gone up significantly because we do a lot more of it. This is due to demographics, earlier interventions, and the treatment of disease that forty years ago had no treatment (e.g. macular degeneration, macular holes, etc).

The whole topic of healthcare “inflation” really needs to be rethought. The areas of real inflation and inefficiencies (hospital care) need to be distinguished from the many areas where total spending has increased not due to per unit inflation, but due to volume increases and innovation (i.e. new procedures and therapies).

Yeah, I always find this one odd. I mean the original iPhone retailed for $500 bucks but a new iPhone X costs $900. CPI inflation suggests that the original iPhone was $614.57 in today's dollars. 46% real cost increase in 12 years. Medical inflation from 2007 to 2019 is just barely south of 40%. Clearly there has been even less productivity gained by iPhone manufacturers.

Yet everyone understands that the iPhone X does more than than the original. Quality gains represent productivity gains and we forget about it.

In health care, nobody cares. Kidney transplants, for instance, cost (everything except maintenance; includes 30 days of care pre-op), around $414,000. There were 17,000 kidney transplants last year in the US. We spent $7 billion last year on mostly successful kidney transplant (the first of which occurred in 1960 between identical twins). If you want to have those then every man, woman, and child needs to fork over around $20 each year. And kidneys are cheap. Hearts run around $1.3 million per (or about $5 billion per year). And on and on it goes.

So why don't we all just agree that medicine has improved our lives and all its productivity is showing up in quality? Because our life expectancy has stopped going up in last decade or so.

And yet, for some cohorts - those who basically live a middle American life from 1960 (is not obese, married with married parents, attends church/synagogue at least monthly, does not drink heavily, has children, is involved in one or more social clubs/activities, and does not use recreational drugs) still see lifespan and healthspan growth.

Nowadays healthcare innovation and resources are increasingly used to compensate for social changes. No spouse to change bandages and drains at home? Fine, we will send you off to a skilled nursing facility that will only cost a few hundred bucks per day. No kids to act as guardians if you present through my doors with a GCS of 6? Fine, I will intubate unless you have a DNI on file, then start the very expensive paperwork to find your next-of-kin or an appointed guardian to eventually ascertain that you actually do not want to live on the vent and withdraw care. It is only tens of thousands of dollars. You want to drink at 1970s levels? Great we can keep you alive. A liver transplant only runs $812,000 (2017 average). You want to sleep around using condoms or heck even prep at typical use rates and efficacies? Great we can expect to spend tens of thousands of dollars on your lifetime care. Have no friends, church members, or anyone else who will check in on you when you are aged? Fine, nursing care is only tens of thousands per year ... or if you fall and break a leg we can just move the decimal point over one space for simple break vs a protracted time of immobility and dehydration.

In no sane world is healthcare less productive. It has dumped all of its productivity into increased quality. It has dumped most of its increased quality into making up for falling social determinants of health.

People are smoking, drinking, and sleeping around less but costs continue to climb. Why? I'm perfectly willing to settle for 1990s era medicine or anything off patent to be more cost effective but I'm forced through insurance to subsidize for everyone's latest and most expensive treatment. At this point, I'd rather have universal healthcare because the government can cap prices and I can move around jobs easier.

People are smoking less. Drinking depends entirely upon your window of choice and problem drinking has been going up steadily for a decade (e.g. we have seen a 7 fold increase in acute ethanol poisoning). For the 1960 - 2018 window, drinking is up substantially. As far as sleeping around; somehow syphilis rates have basically doubled in the last decade and gone up three fold in the last two decades. The British NatSal survey suggests that number of sex partners is increasing even as sexual frequency is declining (e.g. number of female's sex partners basically doubled from 1990 to the last NatSal in 2013). More importantly we seeing a rise in both extremes - more people with zero partners and more partners for the top 5%.

A lot of modern society has seen clustering in the extremes. Which is why policies focusing on average patients are terrible. If you want to predict the impact of alcohol consumption on health you can take the bottom 70% of the population for alcohol consumption and completely ignore them. You can look at only the top 5% and account for over half of all alcohol related deaths.

For sex, the importance of the top decile is even more important. Highly promiscuous individuals not only catch more diseases, but as their partner count climbs the density of the transmission graph increases. Even if 90% of the population had exactly 3 partners, but the top 10% doubled their partner count ... we would see a much higher rate of transmission and disease.

Capping prices is a joke. The AHP has basically written every bit of reform legislation and somehow hospital charges continue to be the fastest growing medical expense even though they are among the least effective treatments. The odds that whatever regulatory structure you put in place will not get captured by the AHP or its ideological descendants are nil. Removing the US from the data shows zero correlation between healthcare delivery system (single payer, Beaveridge, etc.) and cost control.

So you show that it is Baumol’s disease since 1980 but wasn't the big growth in administration between 1960 and 1980?

Cutting administration in an industry being effected by Baumol's disease can still reduce cost. It seem like nothing works in schooling and so should slash overhead and send the money to principals and let them pretty much run their schools with minimal (not zero) oversight. Also pushing fewer people to go to college might be a godo idea.

Wait, what? "But governments often do too little to tax the winners and compensate the losers"

Is Tarabak actually for more taxing rich to give to the poor? Or taxing brumal beneficiaries to compensate brumal losers?

The feeling of declining livings standards that people feel, and is visible comparing millennials lives to those of their parents comes in large part from:

1. The falling labor share; you really could raise a family working at a supermarket or driving a bus in the 1970s. Food is cheaper and better now sure, as are gadgets, housing, education, healthcare, not so much.

2. The eating up of Malthusian goods, caused by incredible population growth. The most desirable places to live are heavily populated, safe areas to live and good schools are only had by mortgaging oneself near other people able to afford expensive houses. This is madness, there as to be a way to keep underclass predation at bay other than filtering based on income/wealth.

A practical approach to amelioration is to do to universities, what's been done to journalism. Erase it as a significant expenditure category, abolish non-elite university employees as a class. There's no reason workers need to spend four years buying the very expensive labor of college administrators, it's simply become too expensive to pretend middle-class kids are rich kids.

Outstanding, academic track students can go to top unis as normal, the rest of us can sign apprenticeship contracts with major firms, and do a few math and writing courses at the company, with the professors who got laid off from the shuttered second-tier schools. 64 credit hours should be plenty for white collar work, augmented with business-specific course work the apprenticeship companies can give. If you gave corporate America smart 18 year-olds, they could be trained up to do office work, STEM work, PR, HR, IT whatever, in a few years. This already happens as college is such a poor approximation of corporate life. This is why wage growth is so strong from age 22 to 30, you're getting the skills needed to work, skills you didn't get in college.

"1. The falling labor share; you really could raise a family working at a supermarket or driving a bus in the 1970s."

You still can, at least to the same degree you could in 1970.

"Food is cheaper and better now sure, as are gadgets, housing, education, healthcare, not so much."

"New US homes today are 1,000 square feet larger than in 1973 and living space per person has nearly doubled"

Also the price per square foot for a house in 1973 is the same as the price per square foot for a house in 2014 when adjusted for inflation.

The average person in 1973 didn't go to college. Furthermore, a typical employee working in a supermarket in 1973 was not in the middle class. A bus driver working then had very nearly the same amount of income as one working today.

Now granted, a working white male doesn't make as much money relative to what he would have made in 1973, but females and minorities make a lot more.

"New US homes today are 1,000 square feet larger than in 1973 and living space per person has nearly doubled"

This is such a weird, dumb talking point. It's literally illegal to build houses under a certain size in much of the country. Of course new houses are bigger—that's the only kind of single-family home allowed. Moreover, "living space per person" has gone up primarily because family formation has cratered.

It's such a facile, tired objection at this point it's hardly worth a response. Some boomers built "dream homes" out of particle board in the late 90s-2006, pushing house sizes up. I bet they had marble counter tops too. Meanwhile they're struggling to sell these to yonger people who recognize the increasingly Malthusian nature of our economy.

Load the CoreLogic house price indexes and plot the ratio of house prices to personal income. Housing is substantially less affordable in major metro areas. It doesn't do any good if housing is affordable in rural Missouri where there's little work.

The characteristic weakness in Rat's reply, and you see this with all these libertardian "everything is better" types, is a failure to address my core point: The Labor Share! What about the bloody Labor Share! Labor is relatively cheaper since the 70s, and most citizens are in the business of selling labor. It's an oligarch's world.

Wages = Marginal Productivity of Labor + Cost of Capital

If capital is easily substituted for labor that keeps wages down.

Skills in the workforce are not evenly distributed. The ability to reallocate workers from a depressed sector to another sector meets barriers. Capital investment adjusts rather easily.

Baumol effects seem to assume greater labor mobility across sectors and that highly specialized workers are easily produced. Inside of some sectors of the economy that appears true, cutting across sectors seems more difficult.

Health care, the sectors of health care seeing the most cost increases, attract a great deal of capital investment and human investment. But the supply of skilled health care workers seems to have a supply constraint.

Plus as health care solves one problem it creates other problems. You increase the life span first through expensive care but over time those costs fall. But then you have an elderly population that gets new diseases in greater numbers. You prevent cardiac death in the 60's but they then develop Alzheimer's in greater numbers in their 80's.

Education has so many cross-subsidies, so much government intervention, etc that analysis of the "education market" is difficult. Certainly, you need to compensate highly skilled teachers at a time when those skills are in high demand. That explains about 1/34 teachers.

"you really could raise a family working at a supermarket or driving a bus in the 1970s"

But in the U.S. poor people have more children than the middle class, don't they?

That is true, fertility is U shaped wrt income. My point is to note the drop in perceived living standards and comment on the drop in the labor share, especially labor share accruing to the low end, not necessarily fertility. If what was once the working class, has become the lumpenproletariat, having 2.1 children out of wedlock, living on transfers that doesn't mean living standards haven't fallen.

Congratulations on good study that has become a big story.

Maybe you should have called it the Helland-Tabarrok Curve!

Take that Art Laffer

The Economist states "In America the share of all education spending that goes on administration has been roughly steady for decades." repeating an assertion from the Helland and Tabarrok book.

Helland and Tabarrok provide a figure 8 depicting shares of spending in three categories - "administrative," "research," and "plant" and attribute the data to NCES. They do not however state or show whether or not "instruction" as a share has increased or decreased. Nor do they explain what they deem as "admininstrative." The Institute of Education Sciences (IES) data at the NCES breaks down ed spending into categories inlcuidng "public service," "academic support," "student services" and "institutional support" and "auxilliary enterprises" and others as well as "instruction." IES does not have an "administrative" type of expense. Adding the spending types listed above other than instruction and dividing by total expenditures using the 4 yr figures at IES DataLab table 24 ( ) we get 37% of spending on these overhead categories rather than the less than 20% Helland and Tabarrok ascribe to "administrative." Fiddling with the numbers a little, it looks like they are defining "administrative share" as only spending on academic support and institutional support, ignoring trends in other spending type categories.

Looking at the most recent Delta Cost Project (has an NCES web page) reported in Table 5 Spending per FTE Student by Standard Expense Categories, FY 2003–2013 (in 2013 Dollars) reported for public research institution increases of 9.5% for instruction, 22.3% for student services, 25.5% for academic support, and 15.4% percent for institutional services. See p. 11 at:

Thus the share of spending on instruction, that is " Activities directly related to instruction, including faculty salaries and benefits,
office supplies, the administration of academic departments, and the proportion of faculty salaries going to departmental research and public service" is actually decreasing relative to other spending types typically considered "bloat."

The bloat hypothesis is thus very far from being ruled out.

And Helland and Tabarrok do their reputations no favors with what appears to be a straight up case of hiding data unfavorable to their desired outcome.

And contradictory longer term data:

"A comprehensive look at all campus employment also shows the familiar shift toward administrative positions (see Figure 5b). There were at least three times as many FTE faculty and staff for every administrative position in 1990. By 2012,
this figure had declined by roughly 40 percent, to an average of 2.2 to 2.5 faculty and staff per administrator at public institutions, and two or fewer faculty and staff positions per administrator at private institutions. "

I like the book overall, but I really, really think it needs a rebuttal to the excellent points raised by Scott Alexander on Slate Star Codex. His criticisms are not trivial.

Baumol effect:


SMBC on Baumol:

Comments for this post are closed