SELF-RECOMMENDING!

# Book Review: Why Are The Prices So D*mn High?

Why have prices for services like health care and education risen so much over the past fifty years? When I looked into this in 2017, I couldn’t find a conclusive answer. Economists Alex Tabarrok and Eric Helland have written a new book on the topic, Why Are The Prices So D*mn High? (link goes to free pdf copy, or you can read Tabarrok’s summary on Marginal Revolution). They do find a conclusive answer: the Baumol effect.

T&H explain it like this:

In 1826, when Beethoven’s String Quartet No. 14 was first played, it took four people 40 minutes to produce a performance. In 2010, it still took four people 40 minutes to produce a performance. Stated differently, in the nearly 200 years between 1826 and 2010, there was no growth in string quartet labor productivity. In 1826 it took 2.66 labor hours to produce one unit of output, and it took 2.66 labor hours to produce one unit of output in 2010.

Fortunately, most other sectors of the economy have experienced substantial growth in labor productivity since 1826. We can measure growth in labor productivity in the economy as a whole by looking at the growth in real wages. In 1826 the average hourly wage for a production worker was \$1.14. In 2010 the average hourly wage for a production worker was \$26.44, approximately 23 times higher in real (inflation-adjusted) terms. Growth in average labor productivity has a surprising implication: it makes the output of slow productivity-growth sectors (relatively) more expensive. In 1826, the average wage of \$1.14 meant that the 2.66 hours needed to produce a performance of Beethoven’s String Quartet No. 14 had an opportunity cost of just \$3.02. At a wage of \$26.44, the 2.66 hours of labor in music production had an opportunity cost of \$70.33. Thus, in 2010 it was 23 times (70.33/3.02) more expensive to produce a performance of Beethoven’s String Quartet No. 14 than in 1826. In other words, one had to give up more other goods and services to produce a music performance in 2010 than one did in 1826. Why? Simply because in 2010, society was better at producing other goods and services than in 1826.

Put another way, a violinist can always choose to stop playing violin, retrain for a while, and work in a factory instead. Maybe in 1826, when factory owners were earning \$1.14/hour and violinists were earning \$5/hour, so no violinists would quit and retrain. But by 2010, factory workers were earning \$26.44/hour, so if violinists were still only earning \$5 they might all quit and retrain. So in 2010, there would be a strong pressure to increase violinists’ wage to at least \$26.44 (probably more, since few people have the skills to be violinists). So violinists must be paid 5x more for the same work, which will look like concerts becoming more expensive.

This should happen in every industry where increasing technology does not increase productivity. Education and health care both qualify. Although we can imagine innovative online education models, in practice one teacher teaches about twenty to thirty kids per year regardless of our technology level. And although we can imagine innovative AI health care, in practice one doctor can only treat ten or twenty patients per day. Tabarrok and Helland say this is exactly what is happening. They point to a few lines of evidence.

First, costs have been increasing very consistently over a wide range of service industries. If it was just one industry, we could blame industry-specific factors. If it was just during one time period, we could blame some new policy or market change that happened during that time period. Instead it’s basically omnipresent. So it’s probably some kind of very broad secular trend. The Baumol effect would fit the bill; not much else would.

Second, costs seemed to increase most quickly during the ’60s and ’70s, and are increasing more slowly today. This fits the growth of productivity, the main driver of the Baumol effect. Between 1950 and 2010, the relative productivity of manufacturing compared to services increased by a factor of six, which T&H describe as “of the same order as the growth in relative prices”. This is what the violinist-vs-factory-worker model of the Baumol effect would predict.

Third, competing explanations don’t seem to work. Some people blame rising costs on “administrative bloat”. But administrative costs as a share of total college costs have stayed fixed at 16% from 1980 to today (really?! this is fascinating and surprising). Others blame rising costs on overregulation. But T&H have a measure for which industries have been getting more regulated recently, and it doesn’t really correlate with which industries have been getting more expensive (wait, did they just disprove that regulation hurts the economy? I guess regulation isn’t a random shock, so this isn’t proof, but it still seems like a big deal). They’re also able to knock down industry-specific explanations like medical malpractice suits, teachers unions, etc.

Fourth, although service quality has improved a little bit over the past few decades, T&H provide some evidence that this explains only a small fraction of the increase in costs. Yet education and health care remain as popular (maybe more popular) than ever. They claim that very few things in economics can explain simultaneous increasing cost, increasing demand, and constant quality. One of those few things is the Baumol effect.

Fifth, they did a study, and the lower productivity growth in an industry, the higher the rise in costs, especially if they use college-educated workers who could otherwise get jobs in higher-productivity industries. This is what the Baumol effect would predict (though framed that way, it also sounds kind of obvious).

I find their case pretty convincing. And I want to believe. If this is true, it’s the best thing I’ve heard all year. It restores my faith in humanity. Rising costs in every sector don’t necessarily mean our society is getting less efficient, or more vulnerable to rent-seeking, or less-well-governed, or greedier, or anything like that. It’s just a natural consequence of high economic growth. We can stop worrying that our civilization is in terminal decline, and just work on the practical issue of how to get costs down.

But I do have some gripes. T&H frequently compare apples and oranges; for example, the administrator share in colleges vs. the faculty share in K-12; it feels like they’re clumsily trying to get one past you. They frequently describe how if you just use eg teacher salaries as a predictor, you can perfectly predict the extent of rising costs. But as far as I can tell, most things have risen the same amount, so if you used any subcomponent as a predictor, you could perfectly predict the extent of rising costs; again, it feels like they’re clumsily trying to get something past me. I think I can work out what they were trying to do (stitch together different datasets to get a better picture, assume salaries rise equally in every category) but I still wish they had discussed their reasoning and its limitations more openly.

The main thesis survives these objections, but there are still a few things that bother me, or don’t quite fit. I want to bring them up not as a gotcha or refutation, but in the hopes that people who know more about economics than I do can explain why I shouldn’t worry about them.

First, real wages have not in fact gone up during most of this period. Factory workers are not getting paid more. That makes it hard for me to understand how rising wages for factory workers are forcing up salaries for violinists, teachers, and doctors.

I discuss whether issues like benefits and inflation can explain this away here here, and conclude they can do so only partially; I’m not sure how this would interact with the Baumol effect.

Second, other data seem to dispute that salaries for the professionals in question have risen at all. T&H talk about rises in “instructional expenditures”, an education-statistics term that includes teacher salary and other costs; their source is NCES. But NCES also includes tables of actual teacher salaries. These show that teacher salaries today are only 6% higher than teacher salaries in 1970. Meanwhile, per-pupil costs are more than twice as high. How is an increase of 6% in teacher salaries driving an increase of 100%+ in costs? Likewise, although on page 33 T&H claim that doctors’ salaries have tripled since 1960, other sources report smaller increases of about 50% to almost nothing. Conventional wisdom among doctors is that the profession used to be more lucrative than it is today. This makes it hard to see how rising doctor salaries could explain a tripling in the cost of health care. And doctor salaries apparently make up only 20% of health spending, so it’s hard to see how they can matter that much.

(also, this SMBC)

Third, the Baumol effect can’t explain things getting less affordable. T&H write:

The cost disease is not a disease but a blessing. To be sure, it would be better if productivity increased in all industries, but that is just to say that more is better. There is nothing negative about productivity growth, even if it is unbalanced.In particular, it is important to see that the increase in the relative price of the string quartet makes string quartets costlier but not less affordable. Society can afford just as many string quartets as in the past. Indeed, it can afford more because the increase in productivity in other sectors has made society richer. Individuals might not choose to buy more, but that is a choice, not a constraint forced upon them by circumstance.

This matches my understanding of the Baumol effect. But it doesn’t match my perception of how things are going in the real world. College has actually become less affordable. Using these numbers: in 1971, the average man would have had to work five months to earn a year’s tuition at a private college. In 2016, he would have had to work fourteen months. To put this in perspective, my uncle worked a summer job to pay for his college tuition; one summer of working = one year tuition at an Ivy League school. Student debt has increased 700% since 1990. College really does seem to be getting less affordable. So do health care, primary education, and all the other areas affected by cost disease. Baumol effects shouldn’t be able to do this, unless I am really confused about them.

If someone can answer these questions and remove my lingering doubts about the Baumol effect as an explanation for cost disease, they can share credit with Tabarrok and Helland for restoring a big part of my faith in modern civilization.

This entry was posted in Uncategorized and tagged , . Bookmark the permalink.

### 384 Responses to Book Review: Why Are The Prices So D*mn High?

1. RomeoStevens says:

Robin Hanson posted today on the idea that as societies become wealthier they consistently shift excess funds into education and healthcare. If the next marginal dollar goes disproportionately towards these areas then they would consistently increase but at a faster overall rate than the other areas.

https://www.overcomingbias.com/2019/06/our-prestige-obsession.html

• janrandom says:

You beat me to it 😉 I think both posts complement each other nicely.

If I get it right Robin Hansons ‘affiliating with high-status people’ would explain

– why wages for teachers and doctors are not the primary driver: You can’t buy this prestige directly (hire the teacher) but need to go thru the system which allows other actors to participate. Can we say the market has commoditized prestige?
– why countries with less choice in education and health see less increase in cost – there is less opportunity to signal and thus no way by the market to exploit that.

• Doug says:

I like Robin’s take, but the next question would be why we don’t see a similar effect in fashion, gemstones, high-end cars, art, or other signaling sinkholes?

Apparel is really instructive here. If anything the total money being poured into brand signaling is diminishing. H&M and Zara have totally gutted the mid-to-high end brands, by duplicating their style at a fraction of the cost. There’s still some successful ultra-premium brands. But young people today spend a much smaller fraction of their income on fashion than prior generations.

From a cultural perspective, fashion brands as a status-marker is a dying trend. That doesn’t mean that clothes no longer signal social class. But that signaling is done through aesthetics and style choices moreso than brands and money spent. You can spend \$3000 on a “hood outfit” or \$50 and still look like a middle-class professional. Kind of like how we judge people based on their accent and grammar, even though speaking in a non-regional dialect doesn’t cost anything.

To digress, why is there no H&M in education or medicine? Why can Williams College still collect enormous rents as a class status-market gatekeeper, but Banana Republic no longer can? The only realistic answer I can think of is regulation and (primarily government imposed) barriers to entry.

• Maybe creating a close-but-much-cheaper facsimile of the Williams College signal is a lot harder in practice than doing so with the apparel signal?

• Doug says:

This is an interesting point of discussion. How hard would it be to clone Williams College? To a first approximation the prestige and value of an American college is very closely correlated with the US News and World report rankings.

The criteria for those rankings is pretty transparent. Most of those criteria are pretty easy to hack. And in fact there’s a long tradition of universities gaming the ratings, resulting in a dramatic rise in ranking and institutional prestige (as well as senior administrator salaries):

> Aoun lives in a 14-room townhouse overlooking Boston Common that the school purchased his first year, now worth \$8.9 million. A chauffeur drives him to campus… [I]n 2011, he was the second-highest-paid college president in the country, at \$3.1 million… In addition to Aoun, seven Northeastern employees made over half a million dollars that year.

The other interesting part of that article is that the USNWR rankings heavily bias towards higher spending. Much of the criteria is a direct measure of spending like faculty salaries of expenitures per student:

> [A] mid-30 school like the University of Rochester would have to spend an extra \$112 million on faculty salary and student resources alone to rise a total of two spots. To feed the beast, Aoun has tapped several new streams of revenue, including satellite graduate campuses in Charlotte, North Carolina, and Seattle, as well as online education. Nasella calls them “very important” and claims they are “very profitable” for the school. “There’s no rankings problem that money can’t solve,” says Michael Bastedo, director of the Center for the Study of Higher and Postsecondary Education.

• broblawsky says:

I think most people don’t actually care too much about what brand of clothes someone wears; they only care about whether they look good in them. That’s more about the person (and the amount of money invested in tailoring) than the clothes, frankly.

• nameless1 says:

Consider Scott’s zebra theory of status. If a middle class person can look decently elegant on the cheap (or even a working class person could) then the upper classes signal with looking shabby. It happened in the sixties, hippies, student revolutionaries, it happened in the 00’s, hipsters, yes there are still very expensive outfits at the Oscar gala but they look like a relic of olden times. While the shabby looking people may not be upper class economically, they are socially upper class. Elite universities etc.

• Freddie deBoer says:

Barriers to entry would seem to not fit given that there are now over 3000 two- and four-year colleges and that the number of colleges skyrocketed over the past 50 years.

• RalMirrorAd says:

I am inclined to agree. The natural ‘barrier to entry’ in education is the fact that to some extent, each combination of graduating class x university x degree is its own differentiated product. [By virtue of education being a positional good] meaning that if you have 100 or 1000 or 10,000 other universities degree mills that doesn’t change the number of universities that can realistically offer an identical product as far as the customers perceptions go.

The value of a degree comes to a large extent from the student body and the social networking that a university provides, which is a non-reproducible product. The only way to bring down the price is to suppress demand rather than increase supply.

• nkurz says:

> Why can Williams College still collect enormous rents as a class status-market gatekeeper, but Banana Republic no longer can?

I feel like I should have some insight into this, since I just attended my 25th reunion at Williams this last weekend, and thought a lot about similar questions. Still, I don’t think I have a good answer.

First, I don’t think “regulation” is a relevant factor. While it’s hard to start a new institution, many do get started. And once started, I can’t see how the regulations would prevent an institution from improving in status.

I do think money plays a large role. Williams has an endowment of about \$1 million per student (https://www.chronicle.com/article/If-Republicans-Get-Their-Way/241659, first chart), and operates their own very effective investment fund (https://en.wikipedia.org/wiki/Williams_College_Investment_Office). Having that amount of money in reserve lets you take a long term view that many new institutions cannot. But money itself is not enough.

I think the answer must be highly “path dependent”. Williams has an incredibly talented and dedicated faculty. They have an incredibly talented student body. From what I can tell, they have an equally impressive administrative team who really do put the best interests of the institution above personal gain. Once they have these things (and the money to back them) I think their ability to maintain their status become a lot easier. The real question is how they got to the point that they were able to attract such high quality and dedicated faculty, students, and administration.

An upstart hoping to overtake them would need to do all of these simultaneously. You aren’t going to attract the best faculty until you have great students; you aren’t going to attract great students until you have great faculty; and you probably aren’t going to get an ideologically driven administration until you have a stellar reputation.

I don’t think there is any strategy that can reliably produce this outcome. Trying to “game the metrics” might give small short-term boosts, but it’s never going to be a long-term winning strategy. A viable plan to create a comparable institution is probably going to require (literally) a century to enact, and tens of billions of dollars of funding. And even then, your chances of succeeding with a single attempt would probably be low. You might be better off putting your resources into trying to subvert a comparable existing institution, but that raises the question of what your end goal is in creating a competitor.

• emmajoey says:
• Eponymous says:

as societies become wealthier they consistently shift excess funds into education and healthcare

This would explain higher consumption of these goods, but not higher cost per unit of consumption.

• TDB says:

How do you figure? Supply curve stays constant, demand curve shifts out, price rises, no?

• Eponymous says:

Generally supply curves for a particular good are fairly flat in the long run.

Take a simple example: suppose all goods are produced by labor only, and labor is homogeneous (identical for all uses), but that moving people between different tasks can’t be done quickly. Then supply curves will slope up in the short run (if people decide to buy more education one year, prices have to go up because there are only so many teachers), but not in the long run (wages for teachers will go up, and people will leave other professions and go into teaching). Then in the long run the cost of all goods will just be inversely proportional to labor productivity in that sector, independent of demand.

(Of course, aggregate wages can go up or down depending on aggregate labor supply vs. demand, but that’s another matter).

It’s true that some sectors can have upward-sloping supply curves even in the longer run, for example if those sectors rely on very particular inputs that are in limited supply. But generally speaking supply curves are pretty flat in the long run.

By the way, this is one of the major conceptual errors I see people make discussing this topic, which is why I chose to reply to this comment in particular. The cause of the high prices has to be mostly on the supply side, either high production costs or market power.

• Shion Arita says:

My take is: why is education and healthcare so expensive? Why those two specifically? It seems like it has to be more than a coincidence that those are the things where we have enormous insurance and loan systems to make it so that people don’t have to pay up front and directly. One could argue that this was a result of high costs, rather than a cause of them, but I suspect it’s the other way around. If the patient had to pay out of pocket for the drug, they couldn’t make the price thousands of dollars. If people had to pay out of pocket tuition, the schools couldn’t maintain their enrollment count and charge 40000+ per year.

Now we can push it back and ask why did these things get into that weird insurance/loan space? and I don’t know the answer to that.

• David Condon says:

Housing doesn’t always require insurance, but is more expensive than either, and is growing in cost too (though not as quickly). Child care is growing in cost too despite not being subsidized. Cars have insurance, but car insurance is cheap relative to the running costs of cars, and cars keep getting cheaper.

• The Nybbler says:

There’s only a certain amount of desirable land, and the population is increasing. So that’s going to push housing up. I’d also argue that as a result of the return to the cities, the amount of desirable land has been dropping.

Child care is regulated out the wazoo and becoming more regulated; since 2017, Washington DC requires child care providers to have a college degree, for instance.

• Edward Scizorhands says:

It’s important to draw a distinction between the most urban areas, where housing is a positional good, from the rest of the country. There are lots of places in the country where housing is not insane but with perfectly fine economies, and jobs that are great for someone who doesn’t want to Be The Very Best.

• Plumber says:

@Edward Scizorhands,
Maybe.
But I’ve been down to San Jose, and out to “There be Dragons”, and the water tastes off, the air smells funny, my eyes start to itch, my sense of direction is askew, and/or the weather can go from below freezing to too damn hot in less than a day,

No thanks!

Experience has taught me that the further from my birthplace in Oakland, California a place is the more off it will be.

• sinvi says:

It’s important to remember that there is a pretty big distinction between car insurance and health insurance. Generally speaking, people use car insurance only for big ticket items – you’re not using it for an oil change or a tune up. Health insurance, on the other hand, is used for everything.

• Freddie deBoer says:

Don’t forget housing.

2. armorsmith42 says:

To add to your voice of “wait, this is confusing”: if non-education industries are getting more productive, then shouldn’t we see that either:

A) Spending the summer working in a non-education industry earns you *more* than a year’s worth of tuition, not less, than it used to be?

B) Paying student loans by working in a non-education industry after you graduate is much easier, not harder, than it used to be?

?

• roystgnr says:

Higher-skilled industries are mostly getting more productive, but it’s harder to get into one of those when you’re still a college freshman and you can only work for 3 months at a stretch. I think making the Baumol model just a little more fine-grained thus resolves (A): college professors are higher-skilled, so their wages rise in line with the jobs they could have chosen to do instead, but college students are not yet higher-skilled, so they don’t have those higher wage options for themselves.

(B) remains confusing.

• 10240 says:

Increase in productivity in some sectors manifests mostly as prices of products/services in those sectors getting lower (relative to salaries), rather than as salaries in those sectors getting higher relative to salaries in sectors with no productivity growth. We should expect that wages for work taking similar skill level are roughly the same in different sectors regardless of their productivity, as otherwise people would move to the higher-paying sector: this is the Baumol effect.

At the same time, wages for low-skill work (such as a summer job a student may do) have generally decreased somewhat compared to wages for high-skill work, which may make it harder to pay for tuition by a summer job. Note that the skill level required is more-or-less orthogonal to productivity growth: there are both high-skill and low-skill jobs in sectors with high growth (e.g. low-skill factory workers vs. engineers) as well as high-skill and low-skill jobs with little growth (e.g. food service vs. doctors).

• baconbits9 says:

Increase in productivity in some sectors manifests mostly as prices of products/services in those sectors getting lower (relative to salaries), rather than as salaries in those sectors getting higher relative to salaries in sectors with no productivity growth. We should expect that wages for work taking similar skill level are roughly the same in different sectors regardless of their productivity, as otherwise people would move to the higher-paying sector: this is the Baumol effect.

If this were the case then profits in the no productivity growth sectors would fall and investment would go intou the high productivity growth sectors. The question isn’t ‘why are some people payed X, its why do we keep paying more for the same level of productivity.

• 10240 says:

Why would profits necessarily fall in the no-growth sectors, and increase in the high-growth sectors? The amount of product/service a given amount of salary produces increases fast in the high-growth sectors and doesn’t increase in the no-growth sectors. But due to competition between the companies, consumer prices in the high-growth sectors also fall (so profits don’t increase much). Whereas if the price-elasticity of demand is low in the no-growth sectors, then consumers will pay the high salaries plus the profits that investors will demand to stay in the sector. If the price-elasticity of demand is high, then the no-growth sector will indeed downsize, as people put their money into sectors where they get more bang for their buck; profits will temporarily fall, then stabilize once the amount produced is decreased.

• baconbits9 says:

Its bee stipulated that productivity is increasing in one sector but not another. If wages in both sectors are increasing at the same rate then the increasing productivity sector must be increasing its profits relative to the flat productivity sector.

• 10240 says:

@baconbits9 Not if consumer prices drop in the increasing productivity sector or increase in the flat productivity sector enough to compensate for the changes in labor cost.

Again, at equilibrium, wages in comparable jobs must be similar. Likewise, rates of return on capital (adjusted for risk) must be similar. However, there is no such clear relationship between consumer prices in different sectors, as their products/services are not interchangeable from the consumer’s perspective. The new equilibrium (with changed prices) may arise after a change in the size of different sectors, depending on the price-elasticity of demand in each sector.

• eccdogg says:

Looks like the answer to B might be yes.

https://www.newyorkfed.org/medialibrary/media/research/current_issues/ci20-3.pdf

Looks like the return on investment in a college degree is higher now (2013 at least) than it was in the 70’s.

• baconbits9 says:

I think this is a bad use of averages.

• TDB says:

You seem to be assuming price of uni is constant, but it is the steep rise of this price that we’re trying to explain.

3. Radu Floricica says:

My understanding is that service industries as a whole suffered/benefited from Baumol, but health and education are special. After all, in the real world there doesn’t have to be a single neat explanation for everything. Health is expensive because it is mired in a web of regulations each designed to protect the consumer, but that ended up having too much inertia to allow for large efficiency increases. And education is a mix of always choosing the path of “buy more” instead of “expend less” because, well, it’s for the children; and state-guaranteed loans that are immune to bankruptcy coupled with a society that is telling everybody, teenagers and parents, that the best investment they can make is a college diploma, no matter which – but the more expensive the better.

4. The cost of a college education is free if you just use Kahn Academy. But the productivity of a Dartmouth education hasn’t changed at all in those years, so it is victim to the same thing as the string quartet.

• Joseftstadter says:

That is a good point. The cost of an actual college education has actually dropped precipitously since the 1970s, because you can find lectures on line, take advantage of resources like Kahn Academy, access most of the world’s scientific literature from your laptop, and even hire educated tutors to help you on topics you are having trouble mastering. What you are really paying for when you go to Haverford or U Penn is a certificate and a social network. If you accept that building a useful, reliable social network in real life probably requires a similar investment in time and engagement as it did in 1970, then it makes sense that the cost of that opportunity has risen dramatically relative to other resources.

• keaswaran says:

It’s certainly possible to get something a lot like the knowledge one gets from a college education from much cheaper means like Khan Academy and open access journal articles. But my understanding is that the number of people that actually go from the knowledge level of the average college matriculant to the knowledge level of the average college graduate by these means is basically a rounding error compared to the number of people that go to college.

While you’re right that a lot of what people are paying for in college is the social network, I think you’re missing out that the social network is also important for getting the actual learning done. (People don’t tend to stick with courses of study unless it seems like the default thing to do in their social circles, which it does if you’re a college student, but not if you’re just taking online classes without an educational support network.)

• James Banks says:

Maybe there could be “study clubs”, groups of self-educators who band together.

• Mark Atwood says:

groups of self-educators who band together.

Such things exist, I’ve been a member of a few of them, and have advised a few and taught at a few.

They can be extremely productive, in part because they self-select, and in part because they have have near-zero “the disruptive joker ruins the class” problem that plague public or socially-expected education.

I am given to understand that the European and the British old original universities flowered from these sort of study clubs.

They seem like the sort of thing that members of the “rationality community” would enjoy setting up and running.

• Freddie deBoer says:

Right. The problem with “the internet will teach you” is that it requires a level of self-starting and perseverance that most people simply don’t have. What you’re paying for in college is not the knowledge, which has been freely available in libraries for generations; you’re paying for the people who motivate you to work.

• RalMirrorAd says:

it’s worth pointing out that a significant portion of the high quality post-secondary education is largely self-directed.

*most people* probably cannot obtain [or benefit from] traditionally college level academic material but our efforts to push the degree’d population upwards have necessitated a dependence on direct instruction.

• Mark Atwood says:

significant portion of the high quality post-secondary education is largely self-directed.

My own personal experience getting my STEM degree was an example of both this, and the “study club” idea.

Students were expected to unofficially clump up into multi-year small (4 to 6 people) study groups. If you couldn’t find a study group that would have you, you were almost certainly going to fail out.

I spent more time with my study groups, seated around a table in one of the libraries, or working together on a slateboard in an unused lecture room, than I spent in formal lecture, or spent doing homework solo.

• Protagoras says:

Unfortunately, I can’t find a good way to apply this to learning entirely new things that I’d like to learn, but I do find that one of the best ways for me to learn a tremendous amount about a subject is to be assigned to teach it. My ego won’t let me look too foolish in front of an audience, so the preparation and research and thinking I do preparing to lecture is by far the most intense and productive learning I ever engage in, and I’m also alert and thinking about the subject while actually teaching (and on rare occasions the students will have insightful questions or even know about something I didn’t turn up).

• Matthew Green says:

Scott’s post gently raises a problem with this conclusion, but treats it as a “nit” rather than a fundamental and disqualifying problem with the Baumol hypothesis. Faculty salaries in the Ivy League have *not* increased at the same rate that tuition has. More importantly, 4-year private salaries have increased even more slowly than Dartmouth’s salaries (which indicates that some of Dartmouth’s salary rise is a symptom of that school’s rising tuition costs.) And certainly general wages have increased even more slowly, as Scott points out. (Here I won’t get into Universities’ increasing use of part-time adjuncts rather than tenure-track faculty to teach courses, which is certainly a very significant ‘productivity increase’ — albeit a crummy and unsustainable one.)

Scott presents these as details, but they’re really not. Until these numbers are grappled with properly, I find it hard to take the Baumol explanation seriously. The numbers just need to add up, and they don’t.

• AlexanderTheGrand says:

Can anyone link to a detailed breakdown of expenses for a single prestigious US institution? That seems like it would shed a lot of light on this issue.

• Freddie deBoer says:

Those institutions go to great lengths to hide these figures.

• Edward Scizorhands says:

All mature institutions hide their internal costs. Consumption get reclassified as “investment.” Feather-bedding gets filed under “infrastructure.” Nepotistic hires melt into “personnel.”

• CandidoRondon says:

I signed up to comment because I actually went to Dartmouth and fortunately for the college we had a blogger, Joe Asch, who would investigate these kinds of things until he passed away last year.

The rise in Dartmouth’s tuition can be almost wholly explained by the growth in non-essential administration. For example, Dartmouth, a college of 4000 undergraduates & approximately 2000 grad students had in 1999 about 2400 non teaching staff. By 2015 the college had 3500 non teaching staff. The number of non teaching staff increased ~500 people between 2010 – 2014 while the number of students only increased by 150 and number of professors only increased by 35. At the those rates of growth I imagine by next year there may be a non teaching staffer for each undergraduate. When you realize that each undergraduate is paying for the salary if a well compensated staffer with six weeks of pension and one of the fanciest health care plans in America it’s no wonder tuition has increased as much as it has.

http://www.dartblog.com/data/2016/01/012317.php

• Cliff says:

Yes… it seems obvious that an infinite number of students could watch the lectures on their computer, ask questions from their computer if they want, submit their homework electronically, come in person for the test if necessary (or there is plenty of software for online test-taking, some of which I used in college about 20 years ago).

So it should now be possible to receive THE EXACT SAME EDUCATION at a tiny, tiny fraction of the original cost. You could have 100,000 students in one class with a stellar teacher. For office hours you might have to see someone slightly less prestigious but for most people that’s no sacrifice at all since they otherwise wouldn’t have been able to be in the class with the stellar professor.

Why can’t that happen? Lazy rivers, administrators, U.S. News something something?

It seems to me education is a sector that should have the HIGHEST growth in productivity.

• keaswaran says:

I can tell you’ve never had to mark a homework assignment or moderate a classroom discussion in your life. Neither of those things scales well at all. Marking homework assignments can be scaled up to one grader per 100 students if the assignments are very simple, and can be scaled up to one grader per several hundred students if they’re only multiple choice. Moderating classroom discussions in which students actually develop any argumentation skills really doesn’t scale.

It’s certainly much cheaper and easier for anyone alive today to listen to Beethoven’s String Quartet No. 14 than it ever was before, if they’re willing to hear a disembodied version from computer speakers. But if you want to actually hear a live performance, it still takes just as many performer-hours per listener. Education is the same way.

• Cliff says:

Well, having majored in engineering, I rarely experienced a classroom discussion. So okay, English majors can have super-expensive college degrees, engineers can have super-cheap online courses. Sold.

Homework assignments- same. Marking engineering homework assignments is done by TAs, and with enough effort could be made automated- and for 1,000,000 students that effort would be well justified.

• Freddie deBoer says:

As I said above, people have been able to do this as long as there have been public libraries, and yet essentially no one does. Why? Because a big part of what you’re paying for when you pay for education is the accountability of someone who will impose consequences if you don’t do the work. Almost no one is an autodidact. It’s not how we’re built as a species.

• jermo sapiens says:

My guess would be that the main thing you pay for when you go to college is the certification from a prestigious institution, with the cost being greater as the prestige of the institution increases. I can watch a bunch of lectures on youtube which are equivalent to the classes one would take if they were to get a bachelors degree in economics, but no bank would hire me based on my browser history. The certification is pretty important.

Prestige is an even bigger deal. You can be an excellent law student from a state university and many big law firms will overlook you in favor of an average law student from Harvard. That’s not necessarily a bad move. The connections one makes at an Ivy League school are unbelievably valuable.

• tonyedgecombe says:

The obvious solution to that would be to remove responsibility for grading from the institutions. Force them to truly compete on the standard of education they deliver.

• jermo sapiens says:

Yes, I like that idea. Or, to truly change the model, some schools should only do grading (or provide a grading only option). So you watch all the youtube videos you want prior to taking the Harvard exam in psych 101. Harvard can provide the outline of the course so you know what to study in preparation. Then you go to Harvard for a weekend and write the exam and that’s it.

Maybe rich kids can still go to Harvard and attend classes, but the plebes watch youtube videos and only come on campus for the purpose of taking the exam.

• Lasagna says:

Prestige is an even bigger deal. You can be an excellent law student from a state university and many big law firms will overlook you in favor of an average law student from Harvard. That’s not necessarily a bad move. The connections one makes at an Ivy League school are unbelievably valuable.

This has come up here before. I want to push back on the “connections” and “networking” theory – I went to an Ivy League law school, and I’ve never used any connections I made there. The prestige of the degree was invaluable; the connections I made as a practicing attorney have also helped. But the fact that somebody went to the same school as me has always been unimportant.

My experience is pretty much the norm, I think. There isn’t any great secret society of ivy league lawyers bent on helping each other out. I’ve been hiring lawyers for almost 20 years (Jesus, 20 years), and at no time in any hiring committee meeting has anyone said “we should hire X – he went to Harvard, and those connections would be invaluable”.

• jermo sapiens says:

This has come up here before. I want to push back on the “connections” and “networking” theory – I went to an Ivy League law school, and I’ve never used any connections I made there. The prestige of the degree was invaluable; the connections I made as a practicing attorney have also helped. But the fact that somebody went to the same school as me has always been unimportant.

That’s surprising to me but I’ll take your word for it. I went to law school at a pretty unexciting Canadian university, and my classmates do provide me with a decent amount of referral work. It helps that my practice is in IP, while everybody else in my class mostly does corporate or family law.

• sourcreamus says:

I have been trying to do something similar and the answer is that there is a lack of feedback from online classes. If a student gets stuck on a point it is difficult to move past and easy to stop if they don’t have someone to help you get past Lab 1 step 5. In order to actually provide help the help desk would actually have to understand the material and if they do that they are likely too valuable to be answering stupid questions.

5. Auric Ulvin says:

Although we can imagine innovative online education models, in practice no matter how much technology we can, one teacher teaches about twenty to thirty kids per year.

While the meaning gets through, something’s sort of wrong.

6. InvalidUsernameAndPassword says:

Doesn’t the ability to record and subsequently replay the performance make the violinists massively more productive than in 1826?

• Simon_Jester says:

It sort of did, but in effect what happened is that “create and play recordings” became a whole separate industry, wherein everyone can have music whenever they want even if no one they know could carry a tune in a sedan chair.

But to get live performances, you still have to pay physical musicians, and then Baumol’s Cost Disease kicks in. That’s the direct comparison.

The difference is that the recording industry made live performances of music even more of a luxury good than they already were, which, yes, probably drove up salaries for violinists while also reducing demand.

• Randy M says:

The difference is that the recording industry made live performances of music even more of a luxury good than they already were, which, yes, probably drove up salaries for violinists while also reducing demand.

Wait, why would being a luxury drive up salaries? Live musicians being in less demand should imply that they are less expensive. Just because something is considered a luxury shouldn’t really impact how much can be charged for it; that caches out into an increased demand, which should increase the supply, since there is no relevant limit to the number of people that can learn to play instruments.

• Phigment says:

Lowered demand doesn’t drive down prices universally; it drives down prices, everything else being equal.

So, the hypothesis would be that previously you had lots of low-paid violinists, and probably a handful of highly-paid violinists, but now you only have a handful of highly paid violinists.

You still have the violinists who do huge concerts or work for major orchestras, but you no longer have very many violinists who are part of a mediocre string quartet that goes and plays dinner parties and weddings cheaply. Or those people have second jobs, because mediocre violin work won’t pay the bills any more.

Think of it like horses; pre-automobile, horses were in high demand, very common, and relatively affordable. Post-automobile, horses are in low demand, mostly as luxuries, and are actually pretty pricey. This doesn’t mean that the law of supply and demand has been disproven, it just means the new equilibrium price point is higher and the number of horses traded is lower.

• baconbits9 says:

This doesn’t address how prices are going up relative to people’s incomes. The average violinist’s income might rise but to total spending on violin music as a fraction of GDP would be expected to fall.

• Randy M says:

Ah, so in other words, the average salary has increased because it has basically dropped to zero for the lower end.

• 10240 says:

In other words, salaries or prices for a given quality of live violin performance haven’t increased compared to the average salary. Low quality performances have been largely driven out of business (though whatever remains pays less compared to the average salary), so the average violinist may make more.

• Phigment says:

This is speculation, of course; I haven’t done any research on violinists. I don’t even know if their average wages have changed, or how, in reality.

But you can easily imagine a situation in which the top 10% of professional violinists got raises because now they’re selling recordings and have larger audiences, and the bottom 50% of professional violinists got replaced with iPods playing recordings of the better violinists, and therefore the average income of a professional violinist increased, but the amount of money collectively spent on violin music went down.

• Mabuse7 says:

Of course, this all doesn’t take into account that the advent of recorded music coincided with (or possibly caused) the rise to mainstream prominence of very different styles of music from the traditional western musical canon, such that the sort of music industry that the average violinist works in is very different from the music industry of popular recording artists.

• J Mann says:

If you want to listen to or watch a recorded performance of a violinist, the price has dropped substantially. (You could argue it’s all but free, as you can in fact watch it for free on Youtube or can pay Spotify \$5 a month for all you can listen).

If you want to watch a violinist perform live, it’s gotten more expensive relative to buying a television.

7. eric23 says:

The problem is, prices for these services have gone up massively in the US, but not to the same extent in European countries which have had equally large increases in wealth.

• Simon_Jester says:

Much of the extra cost may be going into the creation of multiple large, mutually antagonistic bureaucratic systems. For instance, American health care is subdivided into health care providers, whose job is to make money by providing the service and charging you for any supplies you use, pharmacies, whose job is to sell you medicine, and insurance companies, whose job is to actually pay the bills for the above providers and pharmacists, while extracting as much money as possible from you the consumer to make up for it, and making this profitable by increasing your premiums and actually paying as little as possible.

The nature of the process creates incentives to engineer artificial scarcity or inefficient artificial gluts, incentives to expend time (and therefore wages) in squabbles that cancel each other out, and incentives to make the pricing system so complicated and involuted that nobody actually knows what anything really costs, only what purchasers can be arm-twisted into paying for it after the insurance companies get done playing middleman between the producers and the consumers.

Which is how finding a lump in your breast in Iceland can be “spend your lunch break going to a place that does mammograms, no you don’t need a referral from a general practitioner, what the hell’s that,” while finding a lump in your breast in America can be “spend three months running around trying to find a specialist who will take your insurance, an insurer who’ll let you see a specialist, or both, while the tumor grows to the size of a pack of cards and metastasizes.”

Obviously these represent the optimal and maximally suboptimal outcomes of the two systems, respectively, but the point remains.

• John Schilling says:

This has not been the case since the 1980s, at least in the realm of health care. The US was starting from a higher baseline for historic reasons that may be worth discussing, but for the last three decades the rate of price increase has been approximately the same in the United States and Western Europe, ~5%/year.

• Simon_Jester says:

Has either the US or Europe drastically restructured its health care system in the past forty years? Because the US still has its own system as I described in my last post, and my understanding is that most European systems instituted their various forms of health care system shortly after World War Two. Am I mistaken?

• eccdogg says:

I think it is wrong to say European countries have had equally large increases in wealth or at least that they are equally wealthy.

For instance in the US a top earner can make a whole lot more than you can in Europe. US doctors salaries are really high compared to other countries but if you compare doctor salaries to what other high earners make in a country US salaries are not particularly high. This is just what the Baumol theory would predict.

The issue of how physician incomes fit into the broader income distribution is discussed in a paper from earlier this year by David Cutler and Dan Ly that appeared in the Journal of Economic Perspectives. They compare incomes for general practitioners and specialists to the income of “high earners,” (people in the 95th to 99th percentile of the income distribution) in 13 other OECD countries. In the US, this ratio is about 1.37 for specialists. In the other OECD countries, the ratio ranges from 2.56 in the Netherlands to 0.8 in the UK, with the average being 1.45. For GPs, the ratio of income for US physicians to that of other US high earners is 0.92. In the other countries, this GP ratio ranges from 0.68 in Norway to 1.41 in Canada, with the average being 0.98. Thus, in both cases, US physician earnings seem to be in line with earnings of high earners, with the US being slightly below the OECD average for both GPs and specialists.

http://businesspublicpolicy.com/?p=1989

Now I am not sure this explains all of why US cost are high, but I do think the Baumol effects is likely one of the major drivers.

• baconbits9 says:

The Baumol cost explanation would be at a loss to give you any insight into why the range is so high. 0.8 to 2.45 is a huge range for OECD countries where as Baumol would give you a first expectation of a narrow range.

• eccdogg says:

Agreed, but the US is not at the top of the range or even above average.

• Alchemist says:

@eric23

The problem is, prices for these services have gone up massively in the US, but not to the same extent in European countries …..

Healthcare and university prices in Europe are constrained by government rationing. Naturally prices differ.

…which have had equally large increases in wealth.

No European countries have had the increase in wealth that the US has had over almost any time period you would like to compare post World War II. Even Germany has a far lower per capita GDP than the US has. Germany does have a higher GDP per capita than Alabama I think. (The Scandinavian countries did do very well before they tilted to welfare stateism)

• Clutzy says:

Also you have to adjust for immigration. Higher immigration from low GDP/Capita countries makes even having the same GDP/Capita gains even more impressive.

8. Frederic Mari says:

I’m glad you spotted the issue with their argument.

Using productivity gains is indeed misleading since productivity gains do NOT necessarily translate into higher wages. There was a tight relationship when unions were powerful but that relationship disappeared in the 70s and afterward.

Basically, they’re wrong. I don’t know Eric Helland but I follow Marginal Revolution so I’m a bit more familiar with Tabarrok and his more famous colleague, Tyler Cowen. They’re interesting and worth listening to but they’re a bit over eager to prove markets work, all is well and, really, our issues are mostly illusory/inevitable.

Can the Baumol disease explain some of the inflation in specific sectors? Sure. Assuming workers can move *somewhat* between certain industries then, yes, it’s not unreasonable to presume those industries will have to harmonize their compensation package [though, they can offer a different mix of things as compensation package so that too would confuse things].

If you want a general explanation for cost inflation in specific sectors, I think the right explanation is power.

Some sectors have inherent power. College used to be the guarantee of an upper middle class life. Now, it’s only the sine qua non for it, necessary but no longer sufficient. Still, if you aspire to live an upper middle class life, it’s your best bet. People will pay what is required for that lottery ticket. Indeed, one blogger (might have been Kevin Drum) calculated that, given the *expected* uplift in income you get from being college educated (i.e. even after accounting for the lower odds to get a high paying job after graduating), universities could still increase their prices without experiencing a drop in demand – there’s still surplus being created.

The US healthcare is just a weird sector for anyone not American. Doctors have power, that’s for sure and could frankly confiscate people’s wealth if they organised themselves into cartels and were allowed to charge variable price for a given medical treatment. Generally, that’s frown upon/regulated and the Hippocratic Oath/their entire social conditioning helps limit doctors’ ability to give in their natural greed and exploit their power over life and death.

So, yeah, power – and the ability to wield it without triggering violent pushback – seems to me a decent explanation for why some sectors are seeing cost inflation.

• Frederic Mari says:

re. power in economic affairs. Some economists are radically against introducing that into the equation b/c it automatically means politics get thrown in.

And keep in mind that the title “Marginal Revolution” Tabarrok chose with Cowen comes from the neoclassical construct of margin concepts (marginal cost, marginal utility, marginal revenue etc.). What is one of the key axiom for that neoclassical (from Alfred Marshall to Walras, Jevons etc.) thinking? The fact that all economic actors are price-takers, not price-makers i.e. they have to accept the market price, they cannot influence it.

Power assumes that some actors are very much price makers.

Imagine, if you will, that you’re living in a country where the garbage trucks/waste management is controlled by a single company or a cartel of company. In theory, that company could confiscate all the wealth create b/c, if you ever gone through a garbage collectors’ strike (I have, in Paris), you know a city can’t really function if its garbage is not being collected for around 2-3 weeks. It’s a vital service.

Obviously, electricity and water are even more direct necessities.

The way we make sure these guys can’t hold us to ransom is we regulate them and we make sure there’s enough competition to be able to go to another supplier.

For whatever reasons, there’s not enough competition/excess supply/space in Ivy Leagues colleges and in the medical sector in the USA.

• teageegeepea says:

Cowen & Tabarrok’s colleague Bryan Caplan is even more pro-market than them (he’s an anarcho-capitalist), and he disagrees with the book:
https://www.econlib.org/why-the-prices-are-so-damn-high-a-deeper-look/

• Frederic Mari says:

I really don’t like Caplan. He argues in bad faith.

But, here, at least, he’s true to himself and direct. He doesn’t like the Baumol explanation b/c it removes some responsibility from the government and blaming the government for all ills is Caplans’ bread and butter.

And, to be fair, while I don’t think “free market” would necessarily work well in education/health care (economies of scale and scope and coordination across providers offer too many practical benefits), I kinda agree with him on that topic.

i.e. past governmental interventions in those 2 sectors were badly thought out (possibly b/c of path dependencies/ having to satisfy too many stakeholders) and so I’m happy to concede government shares part of the blame for the issues.

I would argue for *better* intervention/regulation rather than removing all regulation but, of course, that’s not what Caplan is arguing for…

• Picador says:

Yeah, after reading Tabarrok’s summary on MR I rolled my eyes and filed it away in the libertarian crank file. “Prices increased by a factor of 20 but it’s okay because everyone is getting paid 20 times as much (on average*)” is libertarian economist code-speak for “prices increased by a factor of 20 and the five richest dudes in the country are getting paid 1000 times as much, and who gives a shit about the other 99.99 percent of the population whose wages remained stagnant during that dramatic increase in the cost of living”.

The conflation of “productivity” with “median wages” is a pretty typical bait and switch for these guys. It’s what lets the neoliberals tell everyone that “the economy” is booming when you and everyone you know is out of work and about to be homeless.

• Frederic Mari says:

Yep…

• Swami says:

I guess if you meant this in a theoretical way, it might be insightful. But let’s remember that today the unemployment rate is the lowest it has been in a half a century, consistently lower than most other other developed nations, and all tracking of actual people reveals that (contrary to the rhetoric to the opposite) the poor are the ones who have been gaining the most over time.

https://www.policyed.org/numbers-game/do-rich-get-all-gains/video

Just trying to keep things in perspective.

• Bamboozle says:

I call bull.

going to copy paste what i wrote about the unemployment figure. The economy is more segregated than ever, and headline figures are becoming more and more useless at describing what’s actually happening.

Here it is:

In western economies you need to ignore the official unemployment rate as (at risk of possibly sounding cw) it is a political football and unrepresentative of reality. Some examples of what i mean:

As measured by the BLS, the unemployment rate is defined as the percentage of unemployed people who are currently in the labor force. In order to be in the labor force, a person either must have a job or have looked for work in the last four weeks. A person only needed one hour in the prior week to be considered employed.

Doesn’t sound like the 9-5 40 hour work week most think of when they think of a job. It also doesn’t count people who have “given up” looking for full time work, work part time and would like to work full time. Sometimes people have turned to the gig economy often taking 2 or 3 jobs as a substitute. Here is an article looking in to this further, as gig economy jobs are very poorly measured and prone to being counted as 3 employed people instead of one person with 3 crap jobs.
This leaves out a ton of relevant people. According to the November 2016 data, over 5.5 million Americans said they want a job, but don’t have one, and are not considered a part of the labor force. If these people were included in the unemployment rate, it would jump to 8.2%. The BLS is not attempting to be deceptive. These folks are left out of the calculation because more than half of them have not done anything to find work in more than a year. Another 10% of this group say they are not available for work at the moment.

This is after a decade of QE where we have brought interest rates down to record lows for such an extended period that you’d imagine the unemployment rate should be negative somehow if extrapolated from historic trends.

In a clearer picture, here is the labour participation rate for the US here and you can see a falling share of total people in work. Yes people are retiring but not at this rate, indeed we often read about people working into old age as they are unable to retire. You can see we haven’t ever recovered from the great depression. The US even pads it’s unemployment figures by having the largest military in the world, which could be viewed as a giant government-sponsored work scheme.

You brought up France, and i’d say the only difference between France and the US is how they measure their employment rate. In the US they hide the true number of unemployed individuals from the baseline figure, where as in France and many European countries this doesn’t happen to nearly the same extend (excluding the UK which also likes to hide true unemployment whenever Hammond needs to pat himself on the back).

People in European countries are also more used to dealing with government and therefore more likely to respond to surveys. For example in the US the coverage rate (the percentage of targeted households who respond) for the Current Population Survey (CPS) is just 88 percent. (The CPS is the survey used in the United States for measuring unemployment.) It is considerably lower for people who are likely to be unemployed. For example it is less than 70 percent for young African American men. This means that the U.S. data almost certainly understate our true unemployment rate, if we assume that the people who don’t respond to the survey are more likely to be unemployed than the people who do. This fact will be more widely recognized as soon as an important economist decides to pay attention to it. Meanwhile in France it is close to complete.

TL:DR if you take the unemployed figures provided by Governments at face value you are being deceived. (I’d argue more generally if you take any of the metrics that are commonly used to evaluate a Government that are produced by the Government itself at face value you are being willfully deceived.)

A good blog post on the topic from someone smarter and with more time.

• The Nybbler says:

Cutting and pasting your previous claims, ignoring all the objections which were posted before, is not very convincing.

Not just the headline numbers but the alternate unemployment figures are quite low. The labor force participation rate stands at about 63%, a number first reached in the late 1970s. The employment/population ratio stands at about 60%, first reached in the early 1980s. The prime-age EPR is at about 79%, only 2 points off its all time high in 2000. The main difference is actually not retirees, it’s young people (16-25); more are in education.

There’s also this interesting series, “Not in labor force; want a job now”; it’s trending downwards to near its 2008 lows.

• Edward Scizorhands says:

Every generation has its brand new adherents to “the unemployment number is a lie!” and they always seem unaware, even after being told, that the US officially tracks at least 6 different measures, and has for a long time.

• Paul Zrimsek says:

In my experience, alternative unemployment stats are usually trumpeted by people who’ve bamboozled themselves into comparing the alternative stat now to the headline stat then, thereby convincing themselves that everything’s going to the dogs.

• Scott Alexander says:

“They’re interesting and worth listening to but they’re a bit over eager to prove markets work, all is well and, really, our issues are mostly illusory/inevitable.”

I’m against bias arguments in general (I might write about why later), but keep in mind that Tabarrok has done a lot of work explaining why government regulation is not causing as many problems as people think (including in this very book). I think it’s unfair to dismiss him as a libertarian ideologue.

“Indeed, one blogger (might have been Kevin Drum) calculated that, given the *expected* uplift in income you get from being college educated (i.e. even after accounting for the lower odds to get a high paying job after graduating), universities could still increase their prices without experiencing a drop in demand – there’s still surplus being created.”

I think this is a bad way to look at it.

Back before electronic medical records, I took notes when I was treating patients. This massively raised my productivity and let me see patients maybe twice as efficiently as if I had to do everything from memory. If a doctor produces \$200,000 in value, and half of that is because of having note-paper, paper manufacturers should be able to charge \$100,000 for their product. But paper is basically free, because competition brings prices down to costs.

In a well-functioning economy, nobody should have power. Everyone should have to bring prices down to costs like paper manufacturers do. If this isn’t working, and having power matters, 100% of the interesting explanation is “what part of the economy isn’t functioning like it should?”. I think medical cartels are a reasonable explanation with doctors (although if doctors’ salaries really have been rising so much more slowly than health costs in general, maybe not the true explanation here). Something’s probably going on with colleges too, but I think the locus of explanation has to be at why the usual process of competition doesn’t counterbalance the usual tendency to raise prices until you’ve eaten the entire value of your product.

• SCPantera says:

Could it maybe be the quantity of new grads depressing what otherwise might be higher salary growth?

I don’t know if this is an issue for doctors but it’s a hot issue among pharmacists where it wasn’t that long ago we were being told there’s an insane demand for pharmacists and within the span of a few years it flipped to being mind-bogglingly difficult to find a desirable job. Subjectively anyways, I guess I don’t have any numbers to cite but I see colleagues complain about it constantly, especially whenever a new school gets accredited.

Chains seem to be happy to pretty rapidly dump experienced pharmacists for willing-to-be-highly-self-motivated-for-less-pay new grads these days despite every aspect of the job benefitting explicitly from experience.

• Frederic Mari says:

I’m against bias arguments in general but keep in mind that Tabarrok has done a lot of work explaining why government regulation is not causing as many problems as people think. I think it’s unfair to dismiss him as a libertarian ideologue.

That’s fair enough inasmuch as an argument is valid/invalid regardless of the biases of the author. Still, I feel it helps inform the conversation when biases and tendencies are disclosed. I’m left leaning and (*somewhat/no PhD*) trained by economic professors aligned with the institutional school of thought/specializing in industrial economy (and labor economy). It’s fair to disclose those things to someone engaging me in an economic debate.

I think this is a bad way to look at it AND In a well-functioning economy, nobody should have power. Everyone should have to bring prices down. If this isn’t working, and having power matters, 100% of the interesting explanation is “what part of the economy isn’t functioning like it should?”

I’m not sure I follow the logic between the 2 statements. Because power matters, it is a good way to look at things even if the objective is/should indeed be to remove the constraints until everyone is a price taker. (It’s a bit more complicated than that. Sometimes, a balance of power between well matched entities leads to an “equilibrium of terror” where prices are very close to what would happen if the market was truly free and everyone was truly a price taker).

I mean, that’s my point. Power matters in education and health care b/c those sectors are fucked up. They’re fucked up for different reasons and I’m not entirely sure there’s much we can do about education (without the night impossible and liberticide demand that companies stop filtering entry jobs and initial assignments almost exclusively based on school prestige and alumni network). I’m less familiar with health care, esp. in the US, so I won’t comment too much except to note every other advanced country in the world seems to make it work better – i.e. costs are far more controlled – somehow.

• Bamboozle says:

@Scott

Are you comparing paper, which is a commodity, to a degree from an Ivy school?

No one cares if you use paper from one of the first paper producers ever, but people will forever use how old your university is as a heuristic for quality with degrees. If you doubled the population of the US overnight the quality of what you learned wouldn’t improve but the value of the degree would rise significantly.

i.e. It’s a positional good and paper is a commodity. Degrees will never be a commodity. Healthcare is somewhere in between, with trust being the influential factor between all 3.

9. Steve Sailer says:

It’s likely that string quartet quality has gone up considerably over the centuries. A richer society can afford more instruction and practice for violinists.

We have a big Baumol effect in sports, but quality of the athletes does go up: e.g., look at field goal kicking percentages in the NFL over the decades, which have been steadily improving at roughly the same annual rate for generations.

NFL placekicker salaries rank from \$500k to \$5 mil. So dads send their teen sons to Chris Sailer’s kicking academy and kickers no longer get offseason jobs down on the docks but instead just practice their kicking.

• Frederic Mari says:

The rise in star actors/star athletes compensation is mostly thanks to TV/mass media i.e. the ability to perform (and extract money) from tens of millions of people vs. just the people watching the game/the play being performed.

That said, you’re right – performance quality has improved.

• Steve Sailer says:

Right.

It would be interesting to compare economic niches that haven’t benefited from technology. For example, 40 years ago I used to go see bands like the Ramones and the Go-Gos play the Whisky a-Go-Go on the Sunset Strip. I presume the bands were paid a modest but not insubstantial fee. (Also their record labels subsidized their touring expenses in the hopes of them breaking through to a mass market.)

Today, the bands pay the Whisky to play there (or, I presume, the musicians’ parents pay).

That’s a fairly extreme example, but seems to be representative of a general trend toward parents subsidizing young people more. For example, my late father-in-law, a tuba player and musicians’ union leader, earned a nice salary in the summer of 1945 when he was 16 in a touring swing band. That was a halcyon year for unionized professional musicians, when every musician who wasn’t getting shot up on Okinawa could find good paying work in a big band.

He despised amplified rock and roll for only needing three instruments: “And their parents buy them \$20,000 worth of gear!” he lamented.

• vV_Vv says:

The rise in star actors/star athletes compensation is mostly thanks to TV/mass media i.e. the ability to perform (and extract money) from tens of millions of people vs. just the people watching the game/the play being performed.

Which means that the individual productivity of actors and athletes increased thousands times, thus the Baumol effect does not apply.

• Darwin says:

You’re talking about skill, but is there any reason to think these improvements in skill make the event itself a more valuable entertainment product to the consumer?

• Matt M says:

An interesting discussion indeed. It mostly takes place these days in regards to men’s tennis (the increasing amount of unreturnable serves does not necessarily make matches more entertaining to watch) and baseball (the increasing amount of at-bats that result in either strikeouts or home runs also generates less excitement).

I could see a similar argument for field goal kicking, in that the better kickers you have, the less likely teams are to “go for it” on 4th down, which is generally regarded as more exciting (and preferred by the fans) than kicking a field goal is.

• MartMart says:

The productivity of celebrities (I view football stars as sports celebrities) have gone up massively over the years. Their product is entertainment, measured in hours consumed. Modern media allows a the same performance/game to entertain vastly more people. It may be that since the 1800’s entertainers productivity have gone up more than any other profession.

• Bamboozle says:

At the expense of many of the local/smaller games that these people would have watched instead as that’s all they would have had access to.

Therefore it’s another example of the winner takes all effect. The top 0.1% of athletes / celebrities / violinists take all the winnings and everyone else is squeezed out of the market.

10. Steve Sailer says:

It’s an interesting question whether the standard of living goes up as, say, NFL placekickers’ pay goes up. In some ways, the NFL is now less entertaining because kickers have gotten so robotically good at their jobs. A field goal used to be a big adventure that was very exciting when the team managed to make one. That’s why placekickers won NFL Player of the Year awards in 1954 (Lou Groza), 1970 (George Blanda), and 1982 (Mark Moseley): it used to be a more heroic job when it was harder to convert a field goal. (The first two also played other positions, which is another big change.)

• Joseftstadter says:

Football is a weird sport because it has embraced hyperspecialization (I personally think football would be a more interesting sport if roster size was limited to 22). The Baumol effect is certainly visible in soccer and ice hockey, sports which still require top players to be good at a variety of key skills, including playing both offensively and defensively.

Otoh, American football is arguably far more representative of the economy, and the changes in the game seem to track the changes in society more than other sports do. For example, the way the head coach has become a Chairman of the Board, or the way that quarterbacks function like CEOs and are compensated far better than the “working class” offensive linemen. Football was always a funny game with embedded class differences (upper class QBs, working class white linemen, “dangerous” black men playing cornerback, and punters and kickers play the role in the football world that academics/artists/NGO leaders do in the real world), and those differences seem to be getting starker.

• Steve Sailer says:

Right. A lot can be learned about modern America from football.

• Conrad Honcho says:

I don’t know if what you said is true or not, but I really enjoyed the analogy.

• broblawsky says:

I think you’ve managed to sap some of my love for football with that analogy.

11. Darwin says:

Society can afford just as many string quartets as in the past. Indeed, it can afford more because the increase in productivity in other sectors has made society richer… Individuals might not choose to buy more, but that is a choice, not a constraint forced upon them by circumstance… College has actually become less affordable. Using these numbers: in 1971, the average man would have had to work five months to earn a year’s tuition at a private college. In 2016, he would have had to work fourteen months.

I think the claim here is that if all of society kept the same standard of living as it had in 19771 (ie, no nice new phones or anything), that standard of living would be so cheap now that we could afford it *and* a bunch of college and medical care.

Of course, in reality, everyone takes advantage of productivity gains to increase their standard of living. But the authors are saying no one is forcing them to do this. In theory, if an individual lived a 1971 lifestyle, over the course of their life they would save enough money from doing so to pay for college and healthcare and etc.

• Rana Dexsin says:

If you mean the “no one is forcing them to do this” straightforwardly, then it’s a pretty useless claim in terms of applicability to people’s choices in the world! Trying to live a 2000 lifestyle right now will get you effectively kicked out of many parts of society for not-directly-economic reasons, most obviously in the “dealing with you imposes weird things that I don’t have room to think about and I refuse to bother” category, for instance “I don’t care about trying to coordinate with you if you don’t have a cell phone with you at all times”. I can only assume trying to live a 1971 material lifestyle would make you even more of a pariah or even just be impossible. (This is from a relatively urban American perspective, as you might imagine.)

• 10240 says:

I’ve seen this objection a few times: that since computing devices and internet have become de facto necessary, it’s wrong to consider them purely as advances in living standards. However, a cheap-but-adequate laptop, cellphone and internet connection are so cheap that their cost has a negligible impact on what we can spend on other stuff: the remaining salary after these can be something like 97% of most people’s salary.

• baconbits9 says:

Additionally many of these things have options that are cheap that fit the minimum level. Cell phone plans with no data but unlimited text and talk are available for ~\$10 a month, which is less in nominal dollars than what landlines cost 20 years ago.

• nacht says:

I agree with this, being in an industry that has had a lot of productivity over the years. Take the \$399 price point, you can still buy a tool, sofa, appliance, musical instrument or bicycle that was \$399 in 1990 and is probably better or at least as good today. Granted, the people who make these products experienced “synergy” savings (no longer employed in the USA) – but that is economics. Those people had to re-train or drop out.

That \$399 will get you 1/4 college credit now, but would have bought a full credit in 1990.
Not sure how much \$399 gets you in medicine, but probably 4x times as many stitches or sessions.

Given “stagnant” wages, we can afford 25% as much education or therapy, but have no problem stocking our houses with like possessions.

• 10240 says:

Given “stagnant” wages, we can afford 25% as much education or therapy, but have no problem stocking our houses with like possessions.

Wages aren’t stagnating. They are stagnating (in a large range of income percentiles, according to some measures) in inflation-adjusted terms, which include the increasing costs of education and medicine; these are major drivers of the inflation. Since inflation in education and medicine are higher than average, if wages stagnate in real terms, that does mean that a fixed percentage of your wage pays for less education or medicine, but not as much if wages stagnated in terms of the prices of industrial products (where inflation is lower than average). Furthermore, if a fixed percentage of our wages pays for x% less education or medicine, that doesn’t exactly mean that we can afford x% less education or medicine: since the prices of some other things like industrial products are lower compared to wages, we can afford to spend a higher percentage of our wages on education and medicine, and that’s exactly what people do by-and-large.

• Alchemist says:

@10240

……the increasing costs of education and medicine; these are major drivers of the inflation.

Small distinctions matter. As do definitions of terms.

To an economist the term inflation is defined as a rate of increase in the supply of money beyond the rate of increase in economic output.

The result of inflation is a rise in prices. More money chasing about the same quantity of goods and assets.

Sometimes that increase in prices is found in consumer goods, sometimes in asset prices, sometimes all over, sometimes in some specific asset bubble (a fad. like tulips.)

Today we are seeing a tremendous rise in the price of non-elective medical procedures. At the same time the price of lasik surgery or cosmetic surgery is falling.

since the prices of some other things like industrial products are lower compared to wages, we can afford to spend a higher percentage of our wages on education and medicine, and that’s exactly what people do by-and-large.

Yes that is what people do. AND young people work longer and borrow more than their parents did to achieve the same standard of living. Because wages have dropped as a percentage of GDP.

• 10240 says:

I wasn’t aware of the distinction in definitions. It appears to me that inflation is frequently used to mean increasing prices; in more precise terms, that could be called price inflation, while your definition would be monetary inflation.

Regardless of definitions, the figures that show wage stagnation are adjusted for the consumer price index which, once again, includes healthcare and education. According to these figures, people at the median in general don’t have to work more for a given living standard, though people who buy a large amount of education or healthcare at a given time (such as students) are worse off.

Because wages have dropped as a percentage of GDP.

This doesn’t imply that wages have decreasing purchasing power, as the GDP itself has increased a lot in the meantime.

• Dedicating Ruckus says:

I think about this a lot in the context of medicine, particularly insulin and the other desiderata for Type I diabetes (as a family member has this).

The cost of this has, of course, increased extremely dramatically over the course of the past few decades. I’m pretty sure the increase is real as well as nominal, i.e. it’s greater than inflation (by a lot). Some would argue that the quality of care has also increased in a manner that might be commensurate. But, having been in a position to see the changes happen, I can’t really believe this is true. The modal change in costs for this stuff over that period is of the form: 1. drug companies put out a new product, notionally superior, that replaces their old product; 2. the new product is a few times more expensive; 3. soon afterward, the old product is discontinued and becomes unavailable anywhere; the new one never drops to near the old one’s price level. And the new products are basically never good enough to justify the delta; typically they duplicate the function of the old one, maybe with some new bells and whistles (e.g. electronic readout for blood sugar test rather than color-coded chemicals) that aren’t actually too useful.

The upshot is that yes, you are forced to “upgrade” your lifestyle to match changing technology; the cheaper products that worked perfectly well in 1971 (or 1991, for that matter) are simply unavailable anywhere, now. I have to consider this a central example of pure cost disease, i.e. the same benefits for steadily increasing cost, and I can’t see how Baumol effects can explain it. (Drug company scientists haven’t increased in productivity any? Surely they have. And anyway, nothing should prevent the old, cheap insulin from being manufactured now.)

• 10240 says:

I’m not sure how typical this example is of the cost increases in sectors like medicine or education in general.

In cases like insulin, we would have to look at what makes old insulin disappear. Regulation? Drug company cartels? Insurance pays for the expensive one anyway?

Or perhaps new insulin is actually significantly better? Let’s say (made up figures) that the simplest, cheapest insulin reduces one’s risk of early death from diabetes from 100% to 10%. Simple improvements reduced it to 2%. Then further improvements (that are expensive) reduce it to 1%. One could argue that the new, expensive insulin is only a tiny improvement: it increases effectiveness from 98% to 99%, by 1.01%, yet it’s many times as expensive. However, a wealthy society can afford to spend large amounts of money to eliminate a 1% risk of death. So it’s understandable that the market for older insulin evaporates, or (as the country isn’t libertarian) it may even get outlawed as substandard care.

• Dedicating Ruckus says:

As best I can tell, it’s a combination of 2 and 3. The drug companies switch about as often as is required to make sure the new offering is still under patent; and there’s no significant downward price pressure, because of insurance.

As far as I can tell there’s no specific regulatory barrier to selling the old insulin. However, regulations are still pivotal in preventing competition here: insulin is unusually difficult to produce for a drug, being a complex hormone that requires rDNA; coming out with a generic version would require a separate clinical trial to prove biosimilarity to the approved, patented version (AFAICT), which is far more expensive than producing a generic simple chemical drug (where you can just prove that you’re producing the same chemical).

I have no reason to believe that the new insulin is measurably better than the old, certainly by enough to explain the price delta. I am open to any information suggesting it actually is.

• Thomas Jorgensen says:

new insulin sells for vastly less in europe. Which.. is also where it is mostly manufactured and developed. Novo Nordic would keep on developing better insulin if the USA launched into space on spindizzies tomorrow. Their annual profits would just be a lot less healthy.
Near as I can tell, what is happening is that the US has just completely forgotten about the very concept of anti-trust, and is getting taken to the cleaners as a result.

This is not just medicine. Most of Vestagers billion euro fines are for behavior the relevant corporations are certainly also perpetrating on the US consumer, but US authorities never follow up on those giftwrapped cases.

• 10240 says:

Novo Nordic would keep on developing better insulin if the USA launched into space on spindizzies tomorrow.

Probably, but it would have much less incentive to develop further improved insulin or other medicines. As European countries push down the prices of patented medicines through price controls or monopsony buying, it’s my impression that Americans pay for the bulk of the fixed development costs.

Near as I can tell, what is happening is that the US has just completely forgotten about the very concept of anti-trust

In the case of patented products, the company explicitly has a monopoly, that’s the point of patents. I don’t see how antitrust would apply in that case.

• Scott Alexander says:

I’m not sure that’s right. In particular, the part where it’s 5 months of median salary to afford college in 1971, but 14 months today, seems independent of how nice a cell phone you’re buying.

• Frederic Mari says:

Yes!

That’s another thing some economists do – try and pretend that technological moves upward represents an improvement in economic terms vs. improvement in lifestyle/convenience.

The classic example is a car.

A 2019 car is better in every way (and thus relatively cheaper) than a 1970s car. True. Also irrelevant.

I want to go from A to B and I’ve got to use a car to do so. If I was able to do so for \$10,000 in the 1970s and I still have to spend \$10,000 now, I haven’t benefited financially from all those technological improvements. It just made my trips more entertaining/less polluting/safer etc. All important things but not things that fit on a family budget.

Basically – don’t look at “values” (it’s a rabbit hole!), look at “prices”. How many months of median income do I have to give up to satisfy a need? (car and its substitutes satisfy “transportational needs”)

• Swami says:

Perhaps in a world without used cars, Bird scooters and Uber. The cost of transportation has gone down relative to median earnings. In some cases by a lot.

• Frederic Mari says:

That’s fair enough – innovation can radically decrease prices and allow to satisfy existing or new needs at a better price, I’m not contesting that.

OTOH, if you were living in the center of a big city in the 70s, you may not have wanted a car either – even if Bird/Uber didn’t exist?

OTOOH, used cars is not a valid objection b/c used cars existed in the 70s too i.e. if you have a tight budget for transportation now and used cars are a good solution for you, the same is likely to be true for 1970s-you.

Ultimately, it’s an empiric question – has transportation gone down in terms of median income? Maybe, maybe not. I have no strong prior on the answer.

But therefore why I don’t think it’s okay to pretend health care and education are no more expensive than before b/c “Baumol cost disease” is clear?

• The Nybbler says:

OTOOH, used cars is not a valid objection b/c used cars existed in the 70s too i.e. if you have a tight budget for transportation now and used cars are a good solution for you, the same is likely to be true for 1970s-you.

This is not the case, precisely because the quality has improved. If the price of a used car now that is equivalent in quality and utility to a new car in the 1970s is lower than the price of that 1970s new car, then there has indeed been an improvement.

• acymetric says:

An improvement in quality, maybe, but not baseline affordability.

• The Nybbler says:

An improvement in quality, maybe, but not baseline affordability.

No, an improvement in baseline affordability. If in 1979 I can buy a new car for \$10K that will satisfy my transportation needs for 5 years, and in 2019 I can buy a used car for \$9K that will satisfy my transportation needs for 5 years, affordability has improved. Essentially the “manufacturing” process for a car good for 5 years is now “build a car good for 5+X years, then drive it around for X years”.

• sidereal says:

> that standard of living would be so cheap now that we could afford it *and* a bunch of college and medical care.

But really though? What specific standard of living expenses (in terms of of # of months worked) present in the 70’s have been obviated by modern industry? Our lives may be enriched by smartphones, but they do not represent a meaningful fraction of typical COL expense, that could even theoretically be forgone. (“Just buy fewer big screen TV’s LOL”)

The things that actually dominate expenses: housing, healthcare, childcare/education: which seem to already takes up a majority of lower/middle class income, are those to which some sort of cost disease applies.

12. Faza (TCM) says:

Re Baumol: it would probably be more intuitive if the effect was typically described in terms of “increasing-productivity industries’ output gets cheaper”, as opposed to “fixed-productivity industries’ output gets more expensive”.

I’m thinking that Baumol and Bowen may have chosen to focus on the wrong metric. Comparing hourly wages is a fine enough apples-to-apples comparison, but once you’ve brought productivity into the picture, you’re introducing quantities produced, which have the orange nature.

If we were to compare wage-per-output, the whole “industrial production got relatively cheaper” angle might be more apparent.

Re cost disease in general: I find that in economics most things get easier to understand when you go back to basics, i.e. supply and demand. If something’s getting more expensive, chances are demand is outpacing supply.

With healthcare, it should be obvious why demand is likely to be at least somewhat inelastic:
1. people really care about themselves or their loved ones getting better/not dying,
2. insurance provides a layer of insulation between what it costs and what people can afford at a particular moment in time.

For-profit healthcare providers are going to charge as much as the market will bear (which, due to the above factors, may be quite a lot) and if market entry is not easy (which may be a good thing, because most patients essentially need to accept on faith that their healthcare provider knows what they are doing), prices will continue to rise for as long as someone is willing to pay them. This need not be reflected in doctors’ salaries, of course; for that, you’d need a direct payment relationship between doctor and patient (insurer).

Education is in a similar boat, but for a different reason. The demand driver here is “diploma-as-CV-filter” and the “filter” bit is what holds back supply-side effect, too. Employers want to outsource candidate filtering to the educators (if a candidate hasn’t achieved a minimum level of education, they drop out at the first hurdle) and are savvy enough to know that if diploma=job, you’re gonna get diploma mills. As long as a college diploma is an indication of being part of a (somewhat) elite minority, you don’t need to go into that much detail of what kind of diploma it is. As soon as most people have one, you need to get a lot more picky if you just want to be dealing with the cream of the crop.

Education thus becomes something of a Veblen good, useful not for its intrinsic utility, but for signaling that you’re “the right sort of person for this job” (=better than everyone who is not the right sort). It needs to be unaffordable, ‘coz if everyone’s got it, it’s no longer a useful filtering signal.

• 10240 says:

It needs to be unaffordable, ‘coz if everyone’s got it, it’s no longer a useful filtering signal.

It can/could be hard to afford in terms of intelligence, hard work etc., not money.

• Faza (TCM) says:

Here’s the rub, though: it’s already supposed to be unaffordable in those terms. Having a degree from \$TOP_TIER_INSTITUTION is supposedly a signal for above-average intelligence and a good work ethic. To an extent, this is even true.

Problems arise if you try to take matters further. On the hard work end, there’s a point beyond which you start to see diminishing returns, so you can’t just pile on more work and expect only the best workers. You could try for a “let’s drive everyone past breaking point and see who’s the last to drop” approach, but it may well be that even those most resilient types will be outperformed in the long run by those who actually get enough rest.

On the intelligence end, you have a lot more leeway, but here you run into a different problem: you don’t need geniuses to do most jobs. In fact, you probably don’t want people who are too intelligent working for you. Intelligent people are known to get bored with repetitive tasks that do not challenge them. Best case: they bolt for greener pastures. Worst case: things go bad because the person who was supposed to be paying attention wasn’t. You can fire them, but the damage is done.

We could have higher education become a truly elite undertaking for only the best and brightest, but my guess is that a university education would quickly cease to be a valuable asset in getting a job, because most employers aren’t really interested in hiring a bunch of eggheads. The chief function of universities would henceforth likely be educating the future batch of university instructors, which I believe is currently the case with some of the more esoteric humanities departments.

If higher education is to remain a valuable assets for jobseekers, we can’t apply too strict a personal quality filter. If the intelligence requirement, say, is only that you are somewhat above average, a lot of people will qualify (same goes for hard work). If a lot of people qualify, qualifying no longer makes you stand out from the crowd.

It’s not like people consciously said “higher education should be an expensive status symbol”. The price is due to the fact that there’s demand for the status symbol and supply must be limited, ‘coz if everyone has it, it’s no longer a status symbol. There are other ways for prospective buyers of education to bid on who gets a share of the limited supply, but implementing some of them (a real high intelligence requirement) may well destroy its current use as a status symbol.

• 10240 says:

If a job doesn’t need the best and brightest, why would they prefer candidates from the most prestigious (and expensive) institutions, rather than have lower requirements, and hire someone who demands a lower salary? Companies have an incentive to hire people who have completed expensive universities if people from more affluent families perform better, even among those who have passed a reasonable threshold of intelligence, hard work, and have studied for a college degree. This may be true, but it’s not obvious.

• Faza (TCM) says:

If a job doesn’t need the best and brightest, why would they prefer candidates from the most prestigious (and expensive) institutions[?]

Two reasons off the top of my head:
1. You gotta make a short-list somehow and every filter you can apply easily is a good one – even if it isn’t necessary. No experience? In the bin. No degree? In the bin. Not prestigious enough degree? In the bin. Do that a couple of times and you’ve got a number you can manage. Mind you, none of those things hurt. It’s always better to have a somewhat smarter candidate.

2. Status symbols don’t stop once you sign on the dotted line. A company that can brag “we only employ Harvard graduates” could well be able to charge more for something that could just as well be done by the local State U graduate.

Notice that the second holds if, and only if, a Harvard degree is seen as an actual status symbol.

If it does, then – again! – people are going to be bidding for it and the price will rise, unless Harvard increases its enrolment, in which case the status symbol loses potency. It’s a vicious circle.

The fact that the current situation favours the rich isn’t necessarily seen as a bug by those in the position to decide. Again, the reasons are twofold:
1. It’s easy to convince yourself that the rich are smarter, on average. If everyone else was just as smart, they’d be just as rich, wouldn’t they? (Pre-empting Poe: no this is not a belief I actually hold.)

2. All things being equal, it’s better to have rich folks in your network than poor ones.

On a more optimistic note: round here (Poland) the state-funded institutions are considered more prestigious than private ones and they are (to an extent) free to attend. The competition for high-demand places is brutal, but at least it’s competition on merit, rather than size-of-pocketbook.

A different way is possible. Whether it is better is an open question.

• 10240 says:

Mind you, none of those things hurt.

They do: people with Harvard degrees will demand a higher salary than others, because they have more alternative opportunities.

• Bamboozle says:

Faza I couldn’t agree with you more, thank you for laying all of this out. Once you get into a sufficiently “elite” college degrees just become positional goods. I hadn’t heard of Veblen goods (only being familiar with Giffen goods) so thanks for that too.

13. Max J says:

“Third, the Baumol effect cannot explain things getting less affordable.”

I’m not sure I have the same conclusion. If the services in question are dependent on labor from workers with very high “human capital,” then the opportunity cost for that labor could be different from the opportunity cost for labor in general.

For example, take the case of university education and medicine, two areas that Scott singles out where cost disease seems to be making things less affordable. Both of these fields are very selective to enter and require a significant investment in education (which both costs money and has a significant opportunity cost of its own). Potential doctors or professors are generally highly capable people with high-prestige credentials that would make them competitive applicants for most top jobs. So I would think the relevant wage trends over time to look at, at least for these industries, would be wages earned by the top 1-5%, not the median wage. If elite wages and median wages decoupled, as I believe they have over the past several decades, then I guess we would expect for these services to become less affordable for most people if labor is a major component of their cost.

HOWEVER, if doctors’ and professors’ salaries are not significantly increasing over time in real terms, as Scott says, then that remains a problem for seeing the Baumol effect as the culprit here (or at least as all of the culprit).

• eccdogg says:

See my post up thread comparing international doctors salaries to the top 95% in their home countries.

• 10240 says:

The Baumol effect affects the high-skill, high-earning class, too. Some high-skill professionals such as many engineers and managers work in industries with high productivity growth; that in turn increases the salaries of other high-skill professionals such as doctors and professors as well.

14. zby says:

I think teaching is now harder than it used to be because of changed cultural norms.

• Simon_Jester says:

The converse of that is that if we still taught the way we used to, we’d be dismissing… I don’t know, let’s say 25-40% of the population as a worthless ineducable mass who can be taught to read, write, and tie their shoelaces and that’s it.

It’s relatively easy to teach the top 2/3 of your student body after kicking out the bottom 1/3 or pressuring them to drop out.

And back when there were boatloads of jobs working assembly lines, or failing that digging ditches, this was tenable.

Nowadays it’d result in an extra 10-20% of the population being effectively unable to contribute to the economy whatsoever.

• Cliff says:

I don’t think it’s that, I think it’s that people would rather do drugs and play video games than work a low wage job, since that is now allowed.

• Simon_Jester says:

Based on my experiences at various high schools, I think that there’s a significant fraction of the population that would have been dismissed as knuckleheads whose education doesn’t really matter as long as they can run a machine on an assembly line back in 1950, but who we are now expected to prepare for college and semi-professional career work today.

The kids haven’t gotten dumber, and behavioral standards haven’t actually gotten much worse, it’s just that the distribution of expectations has shifted and the kind of outcome we think of as “good enough” has risen so our definitions of “bad behavior” have correspondingly evolved.

There were a lot of baby boomers who sat in the back of the class in high school, copied notes from the board if that much, cut class when they felt like it, and didn’t learn a goddamn thing in high school. Some of them got a diploma, some of them didn’t, and for most of them it didn’t matter one way or the other because the US still had an economy capable of absorbing unskilled labor. Those of them that later regretted blowing off high school, had a fallback position in the economy from which they could re-educate themselves later in life when their brains finished growing in.

Similar behavior today yields different results.

People have been writing books decrying the parlous state of the disciplinary environment in American education for most of a century now; if it were really getting worse as fast as people say, we wouldn’t just be stagnating, we’d have backslid farther than test scores and other objective measures support.

• Dedicating Ruckus says:

People have been writing books decrying the parlous state of the disciplinary environment in American education for most of a century now; if it were really getting worse as fast as people say, we wouldn’t just be stagnating, we’d have backslid farther than test scores and other objective measures support.

Consider the possibility that continuing signal-spiral races around test scores at the high end may be masking major slides at the low. E.g. the rise of professional test prep should be expected to add a constant 100-200 points on an SAT (AFAICT), but it’s only the high end who are competing over college admissions that pursues this; if over the intervening period the low end has gotten much worse, the averages might remain the same. (I don’t have any detailed information here; this hypothesis should be checked with regard to actual data over time.)

In a related issue, with No Child Left Behind and similar, teachers and administrators now have strong incentives to push up test scores specifically, to the degree of large-scale cheating (i.e. teachers actually filling out test papers for their students) in some cases. Things like these, and “teach to the test” effects more broadly, might have the effect of masking significant degeneration in the actual quality of education if you’re only looking at test scores as a proxy.

• Cliff says:

the rise of professional test prep should be expected to add a constant 100-200 points on an SAT (AFAICT)

I’m really puzzled by how common this belief is, when there is a large literature showing effects to be much smaller then this (like 20-60 points at most if I recall?). The SAT is pretty close to an IQ test.

• Edward Scizorhands says:

You can add more points to the worse scorers, who’ve never been prepped on (or figured out on their own) the test-taking strategies that come second nature to the middle- and upper-class.

• bernie638 says:

Take heart, I don’t think it’s that bad. There is still a good living to be made for people who are willing to learn a skill that doesn’t involve the stuff taught in school. Dropouts are still accepted into the trade apprenticeships and learning a trade can still provide a good middle class life. Sometimes better!

Right now at Vogtle, electricians, pipefitters and welders are making piles of money. The boomtowns in the fracking fields are doing even better.

It’s not an easy life, but what is? The only real downside seems to be that it’s a feast or famine lifestyle where you have to save money during the good times and live off savings and unemployment benefits during the recessions.

All of the tradesmen (even the least skilled) can still stay above poverty level and they do contribute to the economy.

For the people that do graduate from high school, but aren’t going to succeed in college, the military is a good career option and they will train you.

The important thing is to stay drug free. Construction accidents are no joke and so a lot of people have drug testing programs.

• The Nybbler says:

The skilled trades require conscientiousness, which is another thing lacking in many of those who aren’t able to handle the college track. This is why employers of tradespeople (especially, but not exclusively, employers who try to cheap out) often complain about hiring people who just don’t show up, or show up drunk or high, etc.

15. Michael Watts says:

They claim that very few things in economics can explain simultaneous increasing cost, increasing demand, and constant quality. One of those few things is the Baumol effect.

How is the Baumol effect supposed to explain increasing demand? By my read, its effect on demand is negative.

The effect as characterized here works like so:

1. In 1800, one violinist plays a piece in the time it would take him to earn/produce two apples, and gets paid an amount approximating the value of two apples for doing so.

2. In 2000, one violinist plays a piece in the time it would take him to earn/produce 500 apples (because apples are much easier to produce).

3. The violinist of 1800 preferred playing music to growing or toiling for two apples, but the violinist of 2000 might prefer getting 250 times as many apples, even if the work is less pleasant than the music is.

4. So the violinist’s wage in 2000 is adjusted up to approximate the value of 500 apples.

This immediately implies that demand for violinists will decrease. If they can prefer 500 apples to playing the music, why wouldn’t I also prefer 500 apples to hearing the music?

I’ve used “demand” in the colloquial sense. In the technical supply-and-demand sense, I don’t see that the Baumol effect would affect demand one way or the other. But quantity traded will fall.

So where is the claim of “increasing demand” coming from?

How is an increase of 6% in teacher salaries driving an increase of 100%+ in costs?

Well, obviously it’s because before the increase teacher salaries accounted for 1670% of total costs. 😀

• nadbor says:

This immediately implies that demand for violinists will decrease.

– you’re confusing demand and quantity demanded. Demand is the whole relationship between price and how much gets bought at that price. Quantity demanded is one number – how much actually gets bought.

Demand may go up at the same time as quantity demanded goes down.

Baumol effect explains increasing demand – workers have more money overall so they are willing to spend more on violinists.

At the same time it explains increasing cost – violinists charge higher prices because they now have better options in other industries. In other words the supply of violin concerts decreases (fewer violinists are willing to play for any given wage).

With increasing demand and decreasing supply, the quantity demanded (which is equal to quantity supplied) may increase, decrease or stay constant depending on the specific numbers involved. (BTW. I learned about all this from David Friedman’s ‘Hidden Order’, which I recommend).

• Argos says:

I am not too sure about this:

As a consumer of violinist goods, has the opprtunity cost of my money not gone up? Sure I have more money now, but with a given amount of money I can buy more goods of sectors whose productivity *has* increased.

Basically, in 1960 instead of going to the opera every weak I could buy a shoddy tv; but now instead of going to the opera I can buy a flat screen tv, a high end gaming laptop AND a Switch. Or more realistically just a recording of Maria Callas for 9 dollars from itunes.

Whether this also applies to demand for education and healthcare is of course a different matter.

• Michael Watts says:

You know, I devoted a large part of my comment to discussing the exact distinction you want to blithely tell me I overlooked.

Would you mind justifying why you think the demand curve would shift up as the opportunity cost of buying violinist time rises?

• nadbor says:

You’re right. I read your comment too quickly and jumped to conclusions. Sorry.

I thought you were using apples as the currency – in which case people being willing to only pay two apples to listen to a violin in year 2000 constitutes constant, not decreasing demand.

I still think there is a good reason to think in practice (depending on availability of substitutes) the demand (relative to dollars, not apples) for violinists would rise in your example. I will try to reply later today.

16. sohois says:

Presumably objection 2 can be explained by an increase in absolute teacher numbers? In general it seems that demand for education and healthcare has risen consistently over the past few decades, so even though average compensation levels might not seem to have changed, because there are so many more doctors and teachers costs could still rise.

I’m guessing at this point someone is going to point out that actually teachers per pupil and doctors per patient has not changed at all and this is completely wrong

• Scott Alexander says:

I think that makes sense if you’re thinking of eg decreasing class size (and T&H do mention that).

But in terms of medicine, if there are more doctors, that should lower the price of medical care (increased supply), not raise it.

You could tell a story where now the average health problem requires more doctors than before (eg go to primary doctor, who refers you to a specialist for diagnosis, who refers you to a surgeon for treatment…) but (total anecdotal evidence here) I feel like even a single-doctor task like seeing your primary care doctor has gotten more costly.

• HeelBearCub says:

But in terms of medicine, if there are more doctors, that should lower the price of medical care (increased supply), not raise it.

Doctors need to to increase in supply directly in proportion to the population, because their productivity is not changing.

Other providers of goods only need to increase if population is increasing faster than productivity.

ETA: and of course the need for an increase in the supply of doctors has to to be met by …. education.

• Michael Watts says:

if there are more doctors, that should lower the price of medical care (increased supply), not raise it.

You have to be careful what people are referring to when they say “the price of medical care”.

Sometimes it’s the cost of purchasing some given bundle of medical care. This makes sense if you’re worried about affordability.

But sometimes, such as in cases where people are talking about government budgets, they’ll say “the price of medical care” when they mean “total government expenditures on medical care”. (Or, if they’re worried that people are wasting their money on the wrong things, “total expenditures on medical care”.) “The price of medical care” can still rise in this second sense even if there is no change to the price of any particular procedure.

• Edward Scizorhands says:

There are studies where increasing the supply of back surgeons in an area increases the money spent on back surgery but doesn’t improve outcomes. There is some name for the economic paradox that increasing doctor supply doesn’t act the way it should on prices.

https://qz.com/1010259/the-100-billion-per-year-back-pain-industry-is-mostly-a-hoax/ is probably overly critical but has a lot of links.

• Aapje says:

Probably because doctors can generate their own demand, to some extent.

• ADifferentAnonymous says:

Random Critical Analysis argues that price-per-same-procedure had not significantly increased.

17. Jack V says:

Assorted points:

* Oh gosh, that does make a lot lot of sense. It’s certainly like this for some things — the more “modern” white collar jobs there are, the harder it is to get jobs which require someone’s personal attention. “I never thought I’d be so rich I owned a car, or so poor I didn’t have any servants.”

* As someone else pointed out, how does this generalise internationally? Can you show a correlation with manufacturing productivity more than, say, regulation? That would tell a lot.

* I think higher education might be different, I think there HAS been a bubble there (in UK and US)

* It’s not just direct salaries, if the explanation is true, the budget for “stuff” from industries with increased productivity should rise and the budget for anything from any other industry should increase. Are there any breakdowns of budget of any institution over the long term that could be compared like this (maybe already in the book)?

* However… if this explanation is true. Violinists should have an overall better quality of life, but worse than people around them. Is that true? Since 1830, then yes, it probably is. But it’s really hard to compare — a lot of the benefit comes from “better medicine” and “more internet” and “better safety standards”, and it’s hard to say which of those can be totted up as worth a certain amount of money and which can’t. I’d like to see a closer analysis on this basis. Has life for teachers got worse (because the only people who do it are people who are sufficiently altruistic because it’s not getting paid commensurately)?

* How do you compare “are people paid more” on average, e.g. has the balance of GDP shifted between working age people to children and pensioners, or between men and women, separate to the change of the average “well off ness”?

• Aapje says:

Violinists should have an overall better quality of life, but worse than people around them.

Unless violinists have options and drop out of violining when quality of life becomes too low relative to other professions.

18. Joseftstadter says:

Conventional wisdom among doctors is that the profession used to be more lucrative than it is today.

My understanding is that that is a complaint about relative wealth. The earnings of lawyers, hedge fund managers and CEOs increased faster than what a doctor makes. A surgeon is relatively richer today compared to a high-school teacher, a police officer, or even a plumber, than he would have been in 1955. But he is no longer “elite” relative to a lot of the peers he might have gone to college with.

• Picador says:

Re: doctors and lawyers, average American income for doctors and lawyers are about \$294,000 and \$119,250 respectively. Doctors being envious of rising lawyer salaries seems a bit absurd.

• eric23 says:

Doctors probably pay more in office expenses and malpractice insurance.

• Picador says:

I’m skeptical that doctors pay more in office expenses than lawyers. Legal malpractice insurance is also not cheap. But in any case: no doctor is spending \$170,000 a year on these things. A doctors being envious of lawyers’ salaries rising is like a doctor being envious of plumbers’ salaries rising. (Actually, plumbers are pretty close behind lawyers — I see estimates of about \$90,000 a year for a licensed plumber.)

Of course, it may be that the subset of lawyers who count as being in the “elite” is the top of the distribution, where of course lawyers tend to make quite a bit more than the median income. So if we’re talking about e.g. graduates from Ivy League schools who end up becoming doctors and lawyers, it’s possible that the lawyers’ salaries are competitive with those of the doctors, assuming doctors’ salaries are fairly uniform (I have no idea) and assuming that lawyers’ salaries are highly variable and have a big spike at the high end (which I happen to know to be true).

• SamChevre says:

Lawyer salaries are really bar-belled, though. Lots of solo practitioners doing wills, bankruptcy filings, and family court making less than starting pay at a white-shoe firm, and a fair number of experienced lawyers at major firms making 7 figures.

19. fwiffo says:

The worrying thing for me here is that we are in a world where we can’t even agree on what has happened to wages in the last 50 years within an order of magnitude.

The zero wage increase of “production and non-supervisory employees” sounds like a cherry-picked category to me — further highlighted by the fact that the wages of people in this category were *exactly* at their 1973 peak when the chartist made the graph. We know manufacturing has been shrinking, so it’s not that surprising that these wages aren’t going up.

The sources you listed for low increases in physician pay seemed pretty random — not the articles that I would expect to be very careful with holding specialty constant and correctly adjusting for inflation. Tabarrok’s graph suggests physician and nurse pay have gone up 300% since 1960. I trust him a little more, but he may also be cherry-picking to make his point. Surely we can find an answer more precise than “somewhere between 0% and 300%” ?

• sohois says:

A poster on the subreddit pointed out that using wages only is misleading, and a graph of all compensation shows a continued upward trend, see:

https://fred.stlouisfed.org/series/COMPRNFB

• Wency says:

I’d think a big part of the difference is employer contributions to health insurance premiums, which is still going back to the same problem. Employers have to set aside a larger and larger share of salaries to subsidize healthcare, which makes it look like employees aren’t getting any more money.

• broblawsky says:

Although the ratio of compensation to output goes down more-or-less continuously from ~1980 to ~2010: see here.

• Cliff says:

Isn’t that what you would expect? Assuming all that increase in productivity is not a result of human capital, but also capital like machinery?

• Thomas Jorgensen says:

No. The historic trend was that output and wages tracked pretty closely, and increasing mechanization came out of the fairly constant percentage gap between those two. It is in fact a major problem this no longer holds – The massive rise in private and public debt can mostly be explained as “The output has to paid for somehow, and if wages does not cover it…”

• Cliff says:

This! I’m really surprised to see Scott using this terrible non-supervisory employee wage deflated by CPI stuff. Didn’t he have a post recently where he went over how deceptive this is? You have to use employee compensation, adjusted by the GDP deflator, and this completely solves the problem.

• The Nybbler says:

The zero wage increase of “production and non-supervisory employees” sounds like a cherry-picked category to me

I believe that is actually the broadest category for similar statistics, though it is for private sector employees only. It includes service sector as well as manufacturing sector employees. Figures for all private sector employees are only available since 2006.

It is wages only, however, not total compensation.

20. Steve Sailer says:

My guess is that real estate is a major driver of what is getting more expensive and what is not.

21. Argos says:

As usual, a big problem of this analysis is Americans being too focused on the US. At least for Germany, higher education costs have been falling since the 1980s
Translated with DeepL:

“In terms of a unit of achievement (education of a student), the financial situation of universities in Germany has deteriorated over the past two decades.
The Federal Statistical Office has calculated more precisely and compared the expenditure on teaching (2001: 11.7 billion euros) with the number of students.
The result is that since 1980 there has been a real decline in teaching expenditure per student of around 15 percent in real terms.”
from: http://www.bpb.de/apuz/28275/paradigmenwechsel-in-der-hochschulfinanzierung?p=all

• Thomas Jorgensen says:

Which is what you would expect during an information technology revolution. Except, possibly more so.

22. Hoopdawg says:

So, essentially, two libertarian economists just confirmed labor theory of value?

• JPNunez says:

Cultural marxism marches on, comrades!

• Cliff says:

You must not have understood it

• Hoopdawg says:

I originally considered replying with a one-liner, but these kinds of back-and-forths aren’t particularly productive. So, consider the following:

– Labor theory of value says that costs of goods and services depend solely on labor time necessary to provide them. (Y/N?)
– Salary is a cost of labor (per unit of time). (Y/N?)
– Baumol effect is an observation that salaries rise to match salaries in competing jobs even with no changes in productivity. (Y/N?)
– An economist looks into an increase in costs and concludes “that it’s almost all Baumol effect”. (Y)
– I.e. the rise of costs of goods and services can be explained (“almost”) solely by the rise of salaries. (Y/N?)
– I.e., the costs of goods and services can be explained solely by the costs of salaries, and thus the cost of labor. (Y/N?)

Which of the above have I misunderstood?

• JPNunez says:

It is mostly due to the move to the service economy, tho.

• The Nybbler says:

That the labor theory of value is supposed to extend to everything, whereas the Baumol effect extends only to certain industries. That certain industries are dominated by cost of labor does not mean that all value derives from labor.

• Hoopdawg says:

I accept the caveat that the work discussed herein only examined certain industries, but there’s absolutely no reason why the effect itself wouldn’t generalize.

23. bernie638 says:

I don’t think cost disease is accurate at all. The violinist may take the same time to learn and play, but from a listener perspective I can listen to the music for little to no money. If you compare furniture a handmade dresser using only traditional woodcraft without glue or nails and hand tools would be affordable by only the very rich, but I could afford a modern one made in a factory when I was a young and enlisted in the military. Why would you say that the productivity of dresser making has gone up, but music making hasn’t?

I also question the concept that the diseased industries can’t get more productive. One teacher per 20 students may be the same as it was in the 70’s, but welding a seam also takes the same amount of time. The difference is that factories have gotten better at having the welders spend more time welding and less time with the prep work and fit up. Teachers only spend about half their day teaching. There’s no reason why we can’t have three major curriculum vendors that supply textbooks along with worksheets and exams and electronic grading that maintains the grade book. The idea that a teacher is a craftsman that has to design the entire course from scratch each year is ridiculous for K thru 6th at least and likely further.

I don’t buy the argument for doctors either. In some areas they are getting more specialized and efficient, just slowly. I was happy to see that the person running the sonogram machine was a technician who only did that. Why can’t I just go to a bone setter trained specifically in setting broken bones without a 7 year medical degree and residency., after the x-ray technician verifies my bone is broken?

• JohnBuridan says:

You are the second person on this thread to claim that teachers can’t reuse material. I do not know where you get this idea. Could you show me some evidence?

Many curriculum vendors do supply worksheets and exams.
8:00-15:00 is more than half the day.
Also, electronic grading can’t well grade much of what k-6 grade teachers teach.

Edit: nevermind, bernie, that was just you claiming the same thing twice! 🙂

• bernie638 says:

Sure, I have two kids in public school right now and none of their teachers were teaching that grade and subject last year, and none of them are going to be teaching that grade and subject next year (their contracts for next year were signed last month). That’s a very small sample size, but when I talked to them they said that was normal everywhere (that’s not true, but it seems common).

Teachers don’t teach for 8 hours a day. There are four fourth grade classes, the fourth grade math teacher has four one hour classes each day. The rest of the time is homeroom, lunch monitor, lesson plan, grading, team meetings with the other fourth grade teachers and unscheduled other things.

Most vendors do Supply workbooks and exams, however the teachers don’t follow them completely. Some do the chapters in different orders , most skip worksheets to make room for projects or group activities (but not everyone skips the same ones or does the same group activities). Each child gets a weekly homework schedule, not an annual homework schedule.

I don’t have hard statistics, but somewhere between most and almost all of my children’s homework is short answer/fill in the blank/numerical which is easy to grade electronically. We are using websites to keep the kids learning for 1/2 hour per day over the summer in English and the grading seems fine. Example: identify the verb, pick the meaning of an unusual or made up word based on context clues in a paragraph.

Edit: sorry, was that a faux paste to post then reiterate in a reply to someone else? I rarely comment so I’m not familiar with the etiquette. I’m just a parent who is very frustrated with the school systems.

• JohnBuridan says:

No, no faux pas at all, I was just not paying attention to usernames.

Yeah, I understand your frustration, and if the worksheets are fill in the blank/numerical/multiple choice those are pretty easy to grade electronically. There are even phone apps which can do it if you program them right! Our Bio and Chem teacher uses them non-stop.

It’s definitely true that teachers sometimes move around, it is also true that a teacher should teach the same class or grade for many years and improve himself every year in that class. That’s what I try to do as a teacher anyway. I have taught all my subjects for at least three years, but will probably exchange one of my old hats to pick up one new.

Teachers make all sorts of judgment calls about what to use and not use from the publisher, what to supplement and what to recycle from previous years, and what type of homework to give. Good teachers of a subject develop good judgments and can give clear reasons for their decisions and the tradeoffs they entail. It takes about three years, IMO, to get good at teaching a particular class.

Teachers, as you point out, have more responsibilities than getting the subject across. They are also supposed to teach kids manners, morals, self-discipline; then they monitor, lesson plan (which for each class is about as much work as preparing a D&D session as a DM), grade, call and email parents of struggling students, take orders from admins, collaborate with other teachers, and, if they are good, trying to master their subject and adjust in order to challenge the capable and aid the challenged.

• Simon_Jester says:

There’s no reason why we can’t have three major curriculum vendors that supply textbooks along with worksheets and exams and electronic grading that maintains the grade book. The idea that a teacher is a craftsman that has to design the entire course from scratch each year is ridiculous for K thru 6th at least and likely further.

It’s not designing the course curriculum that’s a handicraft bespoke job. It’s maintaining the disciplinary environment, fine-tuning the lesson for kids who just need something the default curriculum doesn’t give them, actually explaining things, and providing grading for work that’s too complicated to grade by machine (so this kid wrote the first line correctly then half-assed the rest of the problem; do they get 0 credit, 1/4 credit, or 1/2 credit?).

I don’t buy the argument for doctors either. In some areas they are getting more specialized and efficient, just slowly. I was happy to see that the person running the sonogram machine was a technician who only did that. Why can’t I just go to a bone setter trained specifically in setting broken bones without a 7 year medical degree and residency., after the x-ray technician verifies my bone is broken?

The biggest reason is because there wouldn’t be enough bonesetting work for every little clinic to have its own bonesetters. You’d have to ship all your broken bone cases (without their bones set!) to a centralized hospital that gets enough broken bones to be able to afford to keep a bonesetter on staff.

This is basically how things already work for, say, brain surgery. The problem is that whereas people don’t need brain surgery very often, they do need bones set pretty often. Often enough that people aren’t comfortable waiting a few days to see a bonesetter specialist… but also not often enough to justify having a professional bonesetter in every doctor’s office.

• bernie638 says:

I’m not sure what you’re envisioning grading classwork or homework entails, but most of the ones my children do could easily be graded by machine. That makes sense also since there are four 20 student forth grade classes. If (on average) they have one 20 question worksheets each day, that’s 1600 lines for the teacher to grade each day. These worksheets are chosen to be graded quickly (numbers/ matching/single word/short answer). They don’t have graduate students to grade the papers and the teachers aren’t superheroes.

I don’t think we’re thinking of the same injuries for bonesetters either, if you’re imagining a compound fracture that arrives in an ambulance, then I agree with you.

• Simon_Jester says:

I’m operating at the high school level, where just giving the kids worksheets with extremely machine-scorable answers isn’t necessarily a great idea. It tends to strongly incentivize copying without comprehension, and it makes it harder to tell whether the child actually understands what’s happening on the worksheet. I’ve heard horror stories of kids practicing things like “add two digit numbers” dozens of times, only to encounter a word problem that involves adding two digit numbers and have to start using tally marks or something because they don’t KNOW that’s what they’ve been doing all this time.

On the other hand, you’re not fundamentally wrong- though it suggests another obvious way to improve educational outcomes: hire graders who work at significantly lower wages than the teachers.

• thisheavenlyconjugation says:

An even simpler solution: don’t set homework at all.

• Simon_Jester says:

The first problem with this ‘solution’ is that not all repetitive practice is homework. Students DO have to practice adding fractions a goodly number of times if they’re going to get good at it; teachers DO need to assign some sort of easily evaluated metric on whether a student has actually read the reading assignment and understood its core points. It’s not just busywork. Whether the work is assigned at home or in class, someone has to evaluate it and that does take time even if it’s not the biggest thing interfering with teacher performance.

The second problem is that when students aren’t doing homework, a lot more effort has to be put into finding methods to “do more with less” in terms of the amount of time they actually spend thinking about the content area. Adapting lesson plans that work as well without homework as traditional methods work with it is… nontrivial.

• Garrett says:

Part of the challenge for low-cost specialists is that there needs to be a sufficient demand available to support them.
You are always going to need a generalist physician who can do everything. The technician model works well if you have either a larger hospital with a lot of patients who might need that service internally, or you have a clinic where people go non-emergently to get the scan done.

It’s also going to be based on the consequences for getting something wrong. A technician who screws up an ultrasound will have the treating doctors say “that’s terrible – do it again”. In the case of your bone-setter, you have to worry not just about the act of setting the bone but also the complications which can arise from setting the bone, the management of those complications, and the liability if something goes wrong.

There are areas where lower-trained specialists for procedures do exist. Respiratory therapists and perfusionists come to mind. And then there are mid-level generalists like nurse practitioners and physician assistants who do a lot of the low-complication stuff.

At the end of the day, though, it’s being pushed down the pricing level as fast as it practically can be.

• keaswaran says:

With furniture and with music, there’s a question of substitution. If you’re willing to listen to a recording or a broadcast, then yes, musicians have gotten a *lot* more productive over the past century or so. Similarly with furniture, if you’re willing to use IKEA products, they’ve gotten a lot cheaper (much less manufacturer human labor per dresser). But the experience of actually being in the room with four chamber musicians working through Beethoven hasn’t yet been captured on tape (or digital format), and some people want that.

The set of wikipedia entries on advanced mathematics is really quite good, and completely free (once you have internet access). But it’s a really poor substitute for math grad school. IKEA products are a much more usable substitute for handmade furniture, and mp3’s are a pretty decent substitute for some live concert experiences (but not for others).

24. bitfu says:

Education: If you removed student loans from the equation, there’s no way prices would have gone up so far so fast. It would be interesting to look at the slope of monthly educational debt service as a percentage of overall income over the years. Maybe that hasn’t risen so much. Just keep ’em indebted for a long, long time.

Health Care: Is more complicated–as it’s difficult to lump together Gov’t plans, and Fortune 500 plans along with small business and the self-employed. Does an average worker with the Treasury Dept or VA experience these huge monthly payment increases for insurance? Same goes for your typical worker at Google or Microsoft. I know, however, that the Mom & Pops running small, local graphics businesses or doggie daycares are getting slaughtered. How much –as a percentage of overall revenue—has insurance risen for the larger companies with access to better payment plans?

• Mostly right, but health care and education are not that different. If there were no student loans, prices would be far lower. Similarly, if the government did not guarantee student loans, prices would be somewhat lower.

Likewise, if there were no insurance, health prices would be far lower. Similarly, if no one was required to provide insurance (e.g. employers do not have to pay for insurance, insurance agency can engage in as much price discrimination as it wants or refuse insurance for any reason), and no one was required to provide health services they did not expect to be paid for (i.e. hospital could refuse emergency patient if they did not expect to be paid), then health prices would be far lower, for similar reasons.

This is not necessarily to advocate those things. It is simply pointing out a perfectly obvious contribution to rising prices.

• Wency says:

“access to better payment plans”

I don’t think big companies have especially good sweetheart deals with insurers. They have some negotiating power, but they also have ability to pay. And insurance companies are also bound by regulations on “medical loss ratio” (MLR). In effect, insurers need to spend at least 80-85% of premiums on patient care.

The real difference you probably see is that a lot of big employers, particularly companies like GOOG and MSFT, have a lot of revenue per employee so this expense doesn’t hit them as hard. Also, an engineer making \$200k might cost about as much to insure as a dog washer making \$20k, so it’s a smaller proportional component of total compensation.

25. Chalid says:

But T&H have a measure for which industries have been getting more regulated recently, and it doesn’t really correlate with which industries have been getting more expensive (wait, did they just disprove that regulation hurts the economy? I guess regulation isn’t a random shock, so this isn’t proof, but it still seems like a big deal)

Tabarrok’s post on MR about this topic is titled Federal Regulation Is Not the Cause of Declining Dynamism, so their answer would be “yes.” (Note: this is just federal regulation, not state/local.)

A common comment-section objection to the basic thrust of their paper is that regulation has cross-industry effects – regulating industry A reduces dynamism in A and also in related industries B and C. But this is addressed in the paper.

Factory workers are not getting paid more. That makes it hard for me to understand how rising wages for factory workers are forcing up salaries for violinists, teachers, and doctors.

These workers’ relevant alternative isn’t factory work, it’s some kind of skilled white-collar professional work.

26. actinide meta says:

I can easily believe that the direct cause of high costs in these sectors is stagnant productivity. The important question is why these sectors’ productivity has not increased.

Imagine an alternate history where making shoes was still controlled by a shoemaker’s guild. Shoes must be made by journeyman shoemakers overseen by a master shoemaker. It takes eight years of apprenticeship to become a journeyman shoemaker and ten years as a journeyman to become a master shoemaker. Only the brightest can even get an apprenticeship. Tabarrok asks why the price of shoes has risen so much, studies the data carefully, and concludes that it’s entirely explained by the Baumol effect: the people who become shoemakers have many other choices, so the market has to pay them very high salaries ever though they can’t make any more shoes than in ancient times. But is there something missing from this explanation?

• thisheavenlyconjugation says:

How does this describe the situation in education?

• bernie638 says:

Easy, teachers are only hired to teaching positions if they have a minimum of a degree. That might make sense for the last two years of high school and above, but you don’t need to be a college graduate to teach 2nd grade math.

Additionally, teachers recreate the curriculum every year rather than using standardized materials.

Also, they don’t stay in the same position year over year. My middle child took 4th grade English from the teacher who is going to teach my youngest 3rd grade math. And that’s a smaller move than most of the other teachers at the school. They can’t even reuse what they did last year, much less improve on the parts that didn’t work as well.

Imagine how much better it would be to have a teacher using the same standardized materials that everyone else in the country has used and improved on to teach decimals and fractions for 10 years in a row. They would need very little preparation time, they would know what works and based on comparisons to the historical national exam database be able to recognize if any particular student was having exceptional problems.

• Simon_Jester says:

Easy, teachers are only hired to teaching positions if they have a minimum of a degree. That might make sense for the last two years of high school and above, but you don’t need to be a college graduate to teach 2nd grade math.

The thing is, 2nd grade math teachers aren’t required to have degrees in math in most school districts; they’re more likely to be required to have degrees in math education or early childhood education.

Because we don’t just expect our 2nd grade math teachers to stand there and explain how to add three digit numbers all day. We expect them to explain these concepts in ways that are developmentally appropriate for seven-year-olds. We expect them to be able to understand the body language and behavior of seven-year-olds very well. We expect them to be able to cope with all the little problems that arise when large groups of seven-year-olds are gathered together in the same place.

In short, there’s a lot of complexity here that isn’t about the math content knowledge the teacher is expected to have. If we hire teachers who aren’t qualified to handle the extra complexity, our education system tends to underperform. But it’s pretty hard to get teachers who can handle the complexity without some form of certification process, usually one that involves about this level of complexity.

Imagine how much better it would be to have a teacher using the same standardized materials that everyone else in the country has used and improved on to teach decimals and fractions for 10 years in a row. They would need very little preparation time, they would know what works and based on comparisons to the historical national exam database be able to recognize if any particular student was having exceptional problems.

Many teachers already do this. The problem is that this doesn’t actually export a lot of the labor of teaching.

It’s like, you can make doctors more efficient by streamlining the process by which a doctor fills out a prescription slip and explaining the side effects of a medication. But doctors don’t normally spend the bulk of their day filling out prescription slips and explain side effects. They spend that time figuring out what medication to subscribe in the first place.

So yes, increased curriculum standardization is efficient. On the other hand, it also leads to problems like regulatory capture- when there are only three textbook manufacturers in the country and they all make books towards a very complicated and highly specified inter-state common core curriculum standard, it’s easy to charge \$150 for an elementary school arithmetic book that would rightfully have cost 10% or 20% of that much.

• bernie638 says:

That’s a beautiful story and it sounds believable, but it’s just not what I’ve seen. The majority of teachers don’t have a degree in the subject or in a general childhood education. They just meet the requirement to have a degree. They also don’t stay in the same grade to have specific skills for 2nd graders since they are significantly different than 6th graders.

I just did a quick look on my phone and found a recent study that having an education degree didn’t get better results, but years of experience did. Could be wrong or a outlier but it fits my expectations.

• Cliff says:

We expect them to be able to understand the body language and behavior of seven-year-olds very well.

Hahaha! Do you have any experience with 2nd grade teachers? A certification program I can get on board with, a college degree seems grossly excessive.

• Simon_Jester says:

That’s a beautiful story and it sounds believable, but it’s just not what I’ve seen. The majority of teachers don’t have a degree in the subject or in a general childhood education.

This may be a product of different state requirements or different districts having different levels of success finding qualified teachers.

Now, that being said, I can certainly get behind the idea that “scored high on the SATs and completed a one or two-year program specifically training them in early childhood education” is a better credential for an elementary school teacher than “has completed a four-year college degree, ANY four-year degree.” I am 100% with you there.

At the same time, this is basically a subset of the general problem of the proliferation of college diplomas causing credential inflation. If you hire teachers, well… If someone is in your target hiring demographic for rookie teachers (people under 25, and usually closer to 20), and if they have high conscientiousness and an average or above-average IQ (both strongly desirable traits in teachers), they have PROBABLY graduated from college. Lack of a college diploma, in the current social context, strongly implies lack of intelligence, lack of conscientiousness, or both.

To fix this, we’d need to fix the general problem of credential inflation.

I just did a quick look on my phone and found a recent study that having an education degree didn’t get better results, but years of experience did. Could be wrong or a outlier but it fits my expectations.

I wouldn’t be surprised if the picture is cluttered by a lot of people who are temperamentally unsuited to the teaching profession who get education degrees, then wash out within a few years of actual contact with students. By contrast, people who are temperamentally unsuited to teach and don’t get an ill-advised education degree are unlikely to ever end up in front of a classroom.

• thisheavenlyconjugation says:

Most white collar jobs more or less require a degree, so you need something else to explain the high costs of teaching in particular.

• actinide meta says:

My hypothetical is focused on pointing out that low productivity growth in an industry requires explanation, rather than making a specific claim about the actual causes.

But if you want a concrete thesis, I think that both health care and education are low productivity because, for various reasons, entrepreneurs have not been allowed to destroy them. From the inside of a low productivity industry, big productivity increases look like the end of the world. Skilled shoemakers and everything about their craft are just gone. Farming went from something 95% of people did to something 2% of people do. I don’t know exactly what education would look like in a world where it had been exposed to creative destruction, but you might start by breaking it down into its true functions (babysitting, skill acquisition, knowledge acquisition, propaganda, credentialing, signaling, etc) and trying to figure out how to do each of these without O(N) skilled workers. But approximately everyone currently involved in every level of education would have to lose their profession.

• JPNunez says:

We have a fuckton of e-learning companies right now, but they cannot really make a dent on credentialism.

On the other hand, the basic problem of education is that it requires feedback, and thus runs the limit that teachers can only give feedback to so many students at once. That’s the main limit, not just allowing anyone to teach kids -although it would help to have a bunch of extra teachers to oversee little kids basic math class, I dunno if it’s really that good of an idea, for various reasons, among them that you may be underestimating how easy is to teach little kids-.

I think that e-learning can be largely automated; for example, the Deep Learning intro course in coursera, by Andrew Ng, is largely automated; even homework submissions have automated feedback, and they will point out your more probable mistake until you get a full score.

Full homework automation has some applications, mainly in math, programming, maybe some other hard sciences, but will be difficult to apply on other areas. But it may be a small productivity multiplier in the short term.

• keaswaran says:

There is a good question as to why no one has figured out a better way to improve student writing quality than by having students read a bunch and write essays, and then having trained professionals read those essays and give written feedback. But there’s no big mystery as to why the number of essays per hour a teacher can read and write feedback on hasn’t gone up.

27. Deiseach says:

In 1826, when Beethoven’s String Quartet No. 14 was first played, it took four people 40 minutes to produce a performance. In 2010, it still took four people 40 minutes to produce a performance. Stated differently, in the nearly 200 years between 1826 and 2010, there was no growth in string quartet labor productivity. In 1826 it took 2.66 labor hours to produce one unit of output, and it took 2.66 labor hours to produce one unit of output in 2010.

I’m probably not getting their point, but this example makes no sense to me. (Is this what happens when economists try to be witty?) By definition, a string quartet is played by four people. It was played by four people in 1826, it is played by four people in 2019, and if humanity still survives and plays string quartets in 2219 it will be played by four people.

Unless you’re thousand-armed Banasura playing the mridanga, or until someone decides to programme a computer to play all four parts at once, or orchestrions take over completely, this is how it’s going to be: you need four players and it will take them forty minutes. You can’t improve productivity in playing string quartets, though I suppose it’s technically feasible to have the aforesaid MIDI crank out the work in a compressed five minute burst, but that’s not more productive, that’s chaotic noise.

• Cliff says:

Yes that was the point

• 10240 says:

Everything you said is true, and that was the point that explains why the price of a concert has increased a lot.

You seem to have the attitude that the large price increase in some sectors (such as concerts) is a problem, and anyone who tries to explain why prices have increased is necessarily trying to find the cause of this problem, and then we must eliminate the cause. No, the Baumol effect is an explanation of the price increase, even if it’s impossible to improve productivity in a sector, in which case the conclusion is that the price increase is inevitable. Or course in some other affected sectors it may be possible to increase productivity.

• JPNunez says:

There may be other causes, tho.

For example, the price of the venue has also gone up, even if it is only in cost opportunity of keeping an old opera house in the center of the city where real state prices are higher.

28. JohnBuridan says:

Not here to restore faith exactly.
You say “teacher salaries today are only 6% higher than teacher salaries in 1970. Meanwhile, per-pupil costs are more than twice as high. How is an increase of 6% in teacher salaries driving an increase of 100%+ in costs?”
Well breakdown what we give students into categories: Teachers, Textbooks, Technology, Sports Programs.

[b]Teachers.[/b] The pay of teachers goes up 6% but the average middle – high school student has 6 teachers so the pay for the group of teachers who handle a student has risen by 36%, maybe more if the pay of auxiliary staff (counselors and nurses has risen as well).

Textbooks. Textbooks have risen in cost by ~800% since 1970. Most of this is some weird form of regulatory capture, as far as I can tell.
Although people focus on it at the level of college, I think that is only because the costs at public schools are more hidden from the eyes of taxpayers. In 1970 the cost of a textbook was 44 dollars, one can still buy textbooks at this rate, but your school would have to be free to buy textbooks that are not part of the Common Core Standards (or whatever variation your state has), also you would have to experiment with alternative systems and curricula than what The College Board and AP dictate High Schools should do.

Technology. Your school decides to become a 1:1 school, meaning one laptop per student because it’s modern, and in the modern world people use computers for everything regardless of negative learning outcomes. So you buy a laptop for every student. Your Pearson Textbook becomes an online system which is now by subscription. Since most administrators are totally tech illiterate, they pay whatever for whatever software is proposed to them. They pay per student plus a flat fee of 5k per month. I cry out that we would have saved money if we just stuck with the overpriced textbook. No one listens. A tech person has to be hired to repair and debug the laptops. He hates his life.

The costs of the digital systems are hidden, though. I am having trouble finding good data. I do notice though that my local government keeps buying computers in bulk at really horrible prices e.g. \$700 per desktop computer, no they aren’t going to be used to run anything more complicated than Excel.

Sports Programs. I don’t know enough about this to know how costs have changed overtime in sports and extracurriculars.

• JohnBuridan says:

I tried to edit with a further comment about technology costs based upon eyeballing this data. http://www.govwiki.info/pdfs/School%20District/ but it wouldn’t let me.

Based upon a cursorily look at data points from above, computers and new software seem to account for a 2% increase in the cost of instructional materials. (So not a lot. The bank is not broken by this one, but certainly enough to rake in many millions of dollars for Pearson et al. at tiny costs to them.) Technology has not lessened the cost of texts.

• bernie638 says:

I don’t think your math is correct, but I may have had poor teachers. If (using garbage numbers to make the math easier) each teacher used to make \$100 per year, and now they make \$106. A group of 10 teachers for a class used to cost \$1000 per year and now costs \$1060. That’s a 6% increase, not %60.

Also college textbooks may be outrageous, but my child’s fourth grade math textbook from last year is \$8.79 on Amazon.

• Randy M says:

[b]Teachers.[/b] The pay of teachers goes up 6% but the average middle – high school student has 6 teachers so the pay for the group of teachers who handle a student has risen by 36%, maybe more if the pay of auxiliary staff (counselors and nurses has risen as well).

No, the student only has 1.16 teacher at a time, only slightly more than elementary. Assuming each teacher as 1 paid conference period.

• Matthew Green says:

You’re saying that the teacher-student ratio in an average public school has gone up by a factor of 6. I don’t think this is true. Technology would certainly represent a one-time hit to the bottom line, but since IT/tech is a high-productivity area (meaning costs continue to decline rapidly), this shouldn’t account for continuous cost growth. The Internet and publishers like Amazon have made high quality course materials much less expensive — even free in many areas. Not sure what to tell you about sports.

• Simon_Jester says:

While schools do make use of valuable free resources (e.g. Khan Academy) they are ALSO making use of very expensive services. It’s a mix.

• JohnBuridan says:

Sorry, everyone, for my garbage math.

I somehow thought that this was a valid equation for the problem:
6(x * 1.06) = increase in cost of teachers

Wrong on the internet again… *sigh*

• 10240 says:

I don’t get the expensive American college textbook thing: how can they sell them? Being from Hungary, at \$180 a textbook would sell in approx. 1 copy, then it would be scanned, shared, and people would print it or read it on their computers. Even at much lower prices, there was hardly any textbook pirated copies of which didn’t circulate among the students.

• liate says:

In US public schools, textbooks are generally bought by the school system and loaned to the students for the school year, so the people who would actually view a pirated textbook as an option aren’t the ones (directly) paying for the textbooks

• Garrett says:

That would be a copyright violation and illegal to do. Not that it stopped anybody.
Also, the textbooks are updated frequently (sometimes merely to change the revision number) which encourages students to have the correct version. And that means that the required scanning would take lot and lots of time.

• Thomas Jorgensen says:

… Can anyone explain to me why the federal government does not just set up a department to make a set of standard text books for all subjects up to, oh, college 101 classes, keep them up to date and make them freely available as printable pdf / epub ? That should cost pennies, and be politically uncontroversial as long as you do not make them mandatory, but also bring down text book expediture very efficiently just because.. well, now school boards would have to justify not picking the free option.

• eric23 says:

That would be Big Government ™ and one of the two major parties would oppose it.

• 10240 says:

It may be that cheap or free textbooks are already available, though teachers don’t always choose those. bernie638 above wrote that a 4th grade math textbook was \$8.79.

• JohnBuridan says:

While the Houghton-Mifflin one is \$97 for just the student book. Then there’s the workbook, the test booklet, and the teacher’s guide…

I spend a lot of time studying different textbooks and curricula. A standardized government textbook would not even come close to solving our problems. Cf. https://xkcd.com/927/

It’s should be at the discretion of administrators and teachers to find the best and most reasonably priced for their students. There are plenty out there, and more to come… No government intervention needed.
Haven’t you heard the many millions cry out about their hatred of “Common Core Math”?

I think a school in a Minneapolis, Nashville, or Saint Louis can reasonably run a school on 10k per student. If you really hacked away at the budget, it can even be done at 7k, but that would be extremely bare bones (and would not include special education, sports programs, expensive fine arts (like orchestra), school sponsored transport, or lunches.).

Saint Louis has its special education system funded by the entire county instead of by school district, which makes our special ed system one of the best funded.

• Edward Scizorhands says:

Richard Feynman volunteered to review textbooks for his local schoolboard, in a story that everyone here will enjoy if they haven’t read it before.

https://www.apud.net/index.php/geknipt/273-richard-p-feynman-judging-books-by-their-covers

I have no reason to think the situation has improved. It’s principal-agent problem everywhere: bribe the decision-makers to go with you.

In all seriousness, this is a place that the government should be spending money on education. It’s a non-rivalrous good. The DoE has a budget of \$14 billion, with a B, and spending 1% of that on making public domain [1] textbooks would save local school boards craploads of money. (In theory, it should also lower the costs of even value-add [1] text books, but there are principal-agent problems here.)

Also: third-parties could do this, too. Maybe the next time Zuckerberg or Gates want to pour money into a school system, they could pour it into the Open Textbook Library and then spend a few million more reminding voters that they exist.

[1] There are debates about the best kind of licensing here, but until we start getting to actually going to this place, we don’t need to have them.

• Ghillie Dhu says:

But it’s still excludable (i.e., not a public good). Add in the fear of the delegated bureaucrats using the textbook channel to propagandize for their positions, and it’s a political non-starter.

The latter point applies to BillG & Zuck as well.

• Edward Scizorhands says:

What do you mean that BillG and Zuck can’t fund textbooks?

People refusing to use them will be an issue, but getting a few schoolboards to get, er, on board will help. Those schoolboards can see their costs drop significantly, and push other school boards to ask “why not you, too?”

Everyone wants their own version of history. If the books are public domain, individual groups can rewrite chapters as needed. One big help for history books in particular is getting all the IP agreements for photos and pictures bundled in one place. And there can be offers to help fund the public domain version that individual school boards in Kentucky want. They don’t have to wait for the textbook-opoly to write something for them.

https://open.umn.edu/opentextbooks/textbooks/trigonometry-2014

Here’s a Trig book, completely free. It could use some more production value.

• Ghillie Dhu says:

People refusing to use them will be an issue

That’s the crux of why I am skeptical it would be worthwhile for them to fund textbooks, since as left-leaning corporate tech billionaires they’re each widely distrusted from both directions (although BillG has significantly rehabilitated his reputation since the 90s).

• Edward Scizorhands says:

Then get some Texas philanthropist to take the existing CC-licensed history book and update it. Here is one to get started: https://openstax.org/details/books/us-history One of the 6 Senior Contributing Editors is from Oral Roberts University so it’s not like you would be starting from scratch.

The license allows you to edit and modify and sell a new version. The biggest headache is often licensing all the photos, and this version seems to have done that for you.

Oh, Bill and Melinda Gates are already funding it. So who is paying out the nose for textbooks these days? I’m sure someone is.

• Ghillie Dhu says:

That’s a pleasant surprise; thanks for sharing.

29. Icedcoffee says:

While I doubt this is an overarching explainer, don’t education and healthcare have fundamentally distorted economies? Both are areas where our fear of getting a poor product/service removes the normal downward pressure on prices, because we might not get another bite at the apple. (E.g. If I buy the cheapest TV, and get a crap TV: no big deal; if I buy the cheapest doctor, and get a crap doctor: potentially a very big deal.)

• keaswaran says:

Yes, they share those features. And they also share the feature that the customer has bad information about the product being provided (in the case of education this is almost by definition – someone who already knew the details of what they were going to get out of an education is someone who already has the education, not someone who is considering buying it). But those features of these industries haven’t changed over the past century, while the share of expenses paid to them has. Cost disease attempts to explain that.

• Icedcoffee says:

Its worth noting that customers arguably have bad information about lots of things. Guaranteed no one knows how their TV works, or why one is superior to another, but TV’s have still gotten much cheaper and much better.

30. baconbits9 says:

The critique part of this piece is one of your strongest economics posts to date, perhaps the strongest. Their entire position hinges on a single, completely untrue, assumption, that there is some level of production that can be measured across generations in the same sector. One example.

Doctors treat a gunshot wound to the leg today and 156 years ago. 156 years ago a doctor with several assistants ‘treats’ the wound in an hour, today a doctor with several assistants, treats the wound in an hour. That makes it sound as if the two are equivalent, but if you look at outcomes they are not at all equivalent. The outcome for the Civil War era doctor would frequently have been a lost leg to infection, and very often a lost life, while the outcome for the modern doctor is the opposite, rarely will a life be lost, or even a leg.

In terms of # of patients seen there is little change (in the example) in terms of outcomes the change is enormous, so even if we stipulate that a doctor can only see the same 10 patients a day (something that I wouldn’t do) there is still an enormous productivity gap between the doctor who can effectively treat 8-10 of those patients and the doctor who watches multiple patients die due to a lack of advancement in the medical field.

This is a variant of Arnold Kling’s complaint about the GDP factory, and he really needs to be taken more seriously within the economics profession.

• sohois says:

Cost disease is not being measured from the 19th century though. Everyone can agree that there were enormous strides in medical outcomes from the 19th through to the middle of the 20th centuries, but cost disease in medicine is often stared from quite a bit later than that – for example Scott’s original post looked at cost disease beginning from the 50s.

“Productivity” in medicine has not advanced nearly as quickly over the past half century, as the low hanging fruit has all been plucked and many of the advances are extremely niche.

• baconbits9 says:

Cost disease is not being measured from the 19th century though

The example still stands, if there is no ability to compare the productivity of doctors from 1860 to 2019 then why would we START FROM the assumption that we can compare doctors from 1950 to 2019 easily? You seem to be interpreting my post as saying ‘REAL productivity of doctors is being understated, if you used to proper metrics you would see higher productivity’, what is really happening is they are taking categories with the same names and assuming that they translate across generations when their position requires a rigorous analysis for every single subcategory before they even start to draw conclusions. Instead the authors use a rhetorical trick to draw equivalence between two things that have not been demonstrated to be equal.

To be crystal clear every single conclusion that they arrive at depends entirely on the fact that they are comparing like goods and they spend no time demonstrating that fact at all. Further they use completely inappropriate metrics, like standardized test scores.

We do nothave standardized test data from the 1950s and 1960s, but we
have standardized test data going back to 1978. Figure 3 shows math scores for 17-year-olds normalized to 100 in 1978 and real expenditures normalized to 100 in 1980.

Standardized tests are designed to demonstrate differences in ability within a generation, not across generations.

• JohnBuridan says:

Baconbits9, can I just say your comments have been really standout the last few weeks. +1

I feel like Tyler, Tabarrok, Caplan do compare broad categories a lot a lot. Is it just a common problem in macroeconomic models that can’t really be solved?

• baconbits9 says:

Well thanks!

I feel like Tyler, Tabarrok, Caplan do compare broad categories a lot a lot. Is it just a common problem in macroeconomic models that can’t really be solved?

I would say the fact that broad categories being imperfect comps is an issue that can’t be totally resolved in macro models, however I would also say that there are large differences in how much you can attempt to control for these issues and also large differences in how much you should attempt to control for them. Making broad pronouncements about the causes of pricing changes in major sections of the economy would fall under the category of needing a lot of work as there are many potential con-founders that have occurred over the past 70 years.

• sohois says:

But you can compare productivity from arbitrary periods. If we measure healthcare productivity with future life outcomes after interventions, then we can clearly see a huge increase from the 1850s to the 1950s. But using the same measure, productivity does not appear to have gone up much from the 1950s to the 2000s.

I’ve not read the book myself so I can’t speak for whether T&H make this explicit, but I don’t think that means we should assume it impossible to make comparisons across large time periods

31. JPNunez says:

As I understand, student loans for college will lead to higher prices in college tuitions overall, particularly once private colleges enter the equation. After all, colleges will compete for the same students, and will need to offer better teachers, and student loans will allow them to keep raising their tuition in search of these teachers. If the number of teachers is limited, their wages wil raise. At least this last part is true in Chile, dunno in USA. I suspect the demand for higher education in USA will let colleges to keep raising their prices as much as they want.

The solution may be either cut off the loans -probably unpopular- or seek some other means of financing, probably attached to cost controls. Nationalizing higher education may be needed too. Dunno if the Baumol effect has too much to do here; obviously to get good teachers, colleges have to steal them from industry, but this doesn’t explain countries with affordable colleges.

32. helloo says:

Ummm, might be a stupid question, but what exactly has changed between the discussion with this new book?
Note – I have not read the book, only the summary.

You specifically mentioned the Baumol effect in your piece back in 2017 and with basically the same criticisms as you did here.

You had different reasons for why it couldn’t be mostly regulation and some other possible reasons you mentioned before like increased fear of lawsuits and risk tolerances aren’t discussed or ruled out here.
Some other things that the commentators and others have mentioned (like luxury goods and the change in class structure) aren’t mentioned either.
So the things you aren’t convinced are the cause are still the things that you still aren’t convinced, and the things you aren’t sure about haven’t really been looked at.

The only new pieces that I’ve seen is evidence against administrative costs (which you complain about) and trying to link the rate of cost disease and the rate of growth in productivity. Which is something but I do not feel that’s really enough to persuade anyone.

• Matthew Green says:

Scott even points out some problems with the “evidence against administrative costs” component.

33. nobody.really says:

When comparing college costs/outcomes over time, we face at least two challenges.

1. We now educate a larger portion of the population. We might expect lower costs related to educating only that segment of the population that was most “suited to college,” using whatever criteria you like. We should expect some diminishing marginal returns to bringing an ever larger share of the population into college, just as we would expect from putting ever more acres under cultivation. This diminishing marginal return will sometimes manifest in higher costs (for aides/counselors, building greater accessibility ramps, etc.)

2. We need to look at actual college costs, not “list prices” for college tuition (and sometimes even for room/board/books). Colleges price-discriminate–charging more to affluent students and less to poorer ones. A minority of students pay list price. As colleges have grown more expensive, the disparities between list price and average cost have grown as well.

34. Alchemist says:

“If someone can answer these questions and remove my lingering doubts about the Baumol effect as an explanation for cost disease, they can share credit with Tabarrok and Helland for restoring a big part of my faith in modern civilization.”

Hi Scott, Good post. By ‘modern civilization’ do you mean the last several hundred years modern, or the last several decades modern?

The following oversimplifies of course: Prior to world war II (ish) and through 1973 to a lesser degree, the quantity of currency worldwide was limited by and linked to the quantity of gold/silver mined in the world. Gold mining rates rose and fell due to luck and cost variation which often produced short term or local rapid inflations in the quantity of currency, and corresponding lowerings of currency’s value. On the other hand, the productivity of the world – the growth in our ability to produce goods and services – grew slightly faster than the quantity of gold/silver kept as coinage or kept in vaults to secure notes. (not in service as decoration or technology.) As a result, the price of goods measured in workers wages fell gradually as living standards increased. By this I mean that even as a worker – bricklayer, attorney, cowherd, &c. earned the same wage in gold/silver year over year, the amount of material goods he could purchase with it increased slowly, prices fell slowly, as productivity increased and production grew. This trend ended precipitously in 1973.

Since 1973 inflation of the supply of currency has outstripped economic growth every year. And wages do rise following inflation, but FOLLOWING is the key here. Wages are set by a vast market of people who make the sort of mathematical calculations that make lottery tickets popular; ie. hopeful and imprecise. When the value of their wages is fixed to gold/silver and wages are historically somewhat constant or slightly growing, then they can easily realize when negotiations are going in their favor. When inflation eats away at their wages every year in a fashion that is difficult to calculate or see, they lose in their negotiations with capital owners who do better calculations. Since 1973 the bulk of economic growth rewards have gone to the owners of capital and the CEOs and executives who keep the game going, and a corresponding lower share has gone to anyone who works for a wage, whether hourly or yearly salary. The biggest transfer of wealth has been to government workers who are paid from government securities issued instead of taxes raised. The next biggest transfer has been to the growing army of high IQ people who work in the finance industry doing the increasingly complex calculations required to figure out what assets are worth when prices are measured in a funny sort of money that decreases in value steadily year over year.

You are quite correct; The Baumol effect does explain why the purchasing power of musicians’ wages rises even as their productivity remains constant. The Baumol effect has zero bearing on the very real rise in the cost of services in hourly or yearly wages.

• JPNunez says:

It’s possible the gold standard cannot survive a global economy.

Otherwise french generals start sending their navies filled with american dollars to cash in gold at Fort Knox, and USA ends up with a fiat currency anyway.

35. Alkatyn says:

Possibly stupid question, what makes the Baumol effect different from just the cost of labor rising?

• baconbits9 says:

Its specifically about an uneven increase in the cost of labor.

36. jbslattery says:

This may be an altogether too easy answer for the incredible cost increase in college but… simple supply and demand? There are only so many beds and chairs on college campuses, and more people than ever before have been applying to and attending (though not completing) college, thanks to the easy availability of credit (from the federal government and its partners). Is that just way too easy an answer?

37. Sebastian_H says:

I read one of the underlying discussions at marginal revolution, and my immediate objection was “but teacher salaries haven’t gone up”. I believe the Baumol effect can be a real thing, but it makes strong predictions about salaries in areas where the effect is a strong factor, and those predictions seem not to be borne out in education and health care.

• Alchemist says:

“but teacher salaries haven’t gone up”.

Not since 1973. But before that date they did rise.

What have teacher salaries to do with the Baumol effect?

Manufacturing and agriculture productivities have grown by astounding factors in recent history and continue to. A single farmer once couldn’t grow enough for a family without help from family members or other hands. Today a single farmer can feed a small town. So it is natural that a farmers wages have risen.

Musicians and teachers are slightly more efficient due to technology. Concerts and classroom lectures can be attended by more people because of electric amplifiers, the internet, &c. BUT the increase in productivity has been tiny compared to manufacturing or agriculture. Barbers and hairstylists probably take longer to tend to women’s hair then they did in the past. So why have hairstylist and teacher wages risen along with manufacturing wages? That is explained by Baumol et al.

Back to “but teacher salaries haven’t gone up”. In the past several decades they have probably gone down, at least in comparison to the wealth enjoyed by owners of capital, as capital has taken a larger share of the pie at the expense of wage earners.

• baconbits9 says:

What have teacher salaries to do with the Baumol effect?

It has everything to do with the Baumol effect, as the effect is aggregate not individual.

• JPNunez says:

If teacher salaries haven’t gone up, but you still can put teachers in every classroom, maybe the ideal candidates for teaching moved to other industries, and less qualified people took the positions, Baumoling some other jobs.

It may depend on other things, like the prestige of being a teacher too. Teachers used to be respected, but even in my youth I got told that teaching was a low pay, low prestige position.

I mean basic, mid education teacher here, not college, tho.

Baumol may apply in music because if you don’t pay enough to the flute player, JPNunez, hot out of high school and who can play an the sweet flute very poorly, will gladly take that job. And the orchestra will immediately go to hell.

But a high school may make do with cheaper and cheaper teachers and not notice a problem until way later. Baumol effect may apply there only as a second order effect, although proped up by the need to have a teaching diploma. Finishing credentialism among high school teachers may end up with more, cheaper teachers, but also way worse ones.

• Cliff says:

The reality is that teacher compensation has gone up. Getting things like this right is important and some minimal effort should be made to get there.

38. Alex Zavoluk says:

We discussed the question of “why aren’t teacher salaries actually going up” at a meetup recently. If my understanding is correct, the Baumol effect doesn’t predict a very big rise in real teacher/doctor salaries. It just says that the *relative* cost of those salaries goes up, when compared to *the service they provide*. I don’t think it says that those wages go up in comparison to other wages. Maybe this understanding is incorrect, but I don’t think that the Baumol effect predicts large increases in real teacher/doctor salaries. If anything, the whole point is that those salaries are dragged, kicking and screaming, up to the minimum possible level. Another factor we also discussed was support staff. A one-room, 20-child schoolhouse is run by one teacher, with basic support like repairs from the community. A 2000 student high school requires at least 100 teachers, but also involves medical staff, security, cafeteria staff, administrators, HR, guidance counselors, etc. Most universities are even bigger. It’s unclear to me which of these staff could be counted as “teachers” in the sense of being necessary for the production of education. If “most of them”, then productivity growth in education could easily be *negative.* If “none of them”, then I can only give Tabarok a confused look as he claims that administrative costs have stayed constant.

edit: I assumed class sizes were the same, but I suppose it’s also possible that they’ve gotten smaller. Again, that would make the productivity growth of teachers negative.

We didn’t discuss, and I can’t explain easily, why real salaries overall seem relatively flat. The obvious answer to this question is “benefits” but a large component of that is health care. You called this explanation circular, though a feedback loop is not impossible. Another possible answer is an incorrect choice of deflator, possibly one that doesn’t correctly take into account the massive drop in price and increase in quality of goods like computers.

• Cliff says:

why real salaries overall seem relatively flat. The obvious answer to this question is “benefits” but a large component of that is health care. You called this explanation circular, though a feedback loop is not impossible. Another possible answer is an incorrect choice of deflator, possibly one that doesn’t correctly take into account the massive drop in price and increase in quality of goods like computers.

There’s absolutely nothing circular about it. We all know health care prices have risen, the fact that teachers spend more money on it now should not be surprising or even notable in any way. The idea that we should measure some weird partial measure of compensation instead of total compensation is frankly absurd. The improper deflator is also a massive error.

39. notes says:

“These show that teacher salaries today are only 6% higher than teacher salaries in 1970. Meanwhile, per-pupil costs are more than twice as high. How is an increase of 6% in teacher salaries driving an increase of 100%+ in costs?”

Teacher-student ratios aren’t fixed. Take that into account, and that explains the majority of the cost increase, spent not on paying teachers more but paying more teachers. Imagine that due to changing tastes, we can only enjoy the orchestral version of a string quartet: this is Baumol + decreasing (per-pupil) productivity.

Compare the Lancastrian system to Horace Mann’s model, to today. The NCES thought that ratios across the US have gone from 26.9:1 in 1955 to 16:1 in 2015. That trend has been running for as long as we have statistics on American pupil teacher ratios (p. 29).

Taking 1970 on that chart and eyeballing it, it looks like it’s around ~24. The change from 24:1 to 16:1 + a 6% salary increase doesn’t explain an increase of 100%+ over the same time period… but it does explain an increase of 69%. If you suppose that the 16:1 is really 13:1 due to increased administration or similar (and UNESCO thinks the US has been sitting at 13-14 – but it’s from a different data set), that would explain all of it. Regardless, the remaining gap is small enough to be plausibly filled by other manifestations of productivity growth, as luxuries become necessities, from smartboards replacing chalkboards to better quality chairs, and so forth. For the same reason it would be difficult to lead a 1970 life if you wanted to, it’s also difficult to have a 1970 classroom.

• baconbits9 says:

This is basically tautological. We observe higher spending on education and then break down the categories until we find the ’cause’, the Baumol effect is supposed to explain why we are spending more on education.

40. Alex Zavoluk says:

(Second comment since it’s unrelated to the first)

Is music a good example of the Baumol effect? The marginal cost of listening to music is effectively 0, between spotify and youtube; recording devices mean that the productivity of musicians has increased tremendously due to technology. Recorded music is obviously not identical to live, but it is a good substitute. Moreover, even live shows aren’t necessarily that expensive (I don’t know how much they used to cost), and again technology like speakers means that you can play to many more people than you used to be able to.

• JPNunez says:

I think nearly all musicians in the modern day cite live performances as their main source of income. Spotify seems to pay the less among streaming providers of music, so the situation is not really getting better. Cost of entry is the lowest nowadays, so on top of that, there is a ton of competition.

On the other hand a classic musician will only have their orchestra as a source of fixed income. Think how common it has become to have violinists play in the subway stations, etc, so I dunno if they are really being Baumoled in other industries, as much as just accepting lower and lower pay. Dunno what the economy of the more popular orchestras looks like. Maybe it’s better in Europe.

41. Ttar says:

Re: regulations not having an effect — why do you think these highly regulated industries aren’t seeing the same productivity growth as other sectors of the economy?

42. m71 says:

Your third confusion is the easiest to dispell since it is one of terminology. The magic words “Society can afford” are doing the heavy lifting. Total societal GDP has definitely gone up, so “society” can afford more quartets. But you are thinking in terms of individual spending capacity, and considering a representative individual (median income from https://college-education.procon.org/view.resource.php?resourceID=005532). Obviously US income distribution has changed — maybe as much or more than average productivity. So “we” (people I imagine like myself) can individually afford less while “we” (Moloch) can afford more. Which is the better sense of “we” is a recurring issue in economics.

You other complaints/confusions are more substantial, with multiple and interacting causes and some do undermine the simplistic “Baumol is all” thesis.

But in support of Baumal causes:
– Factory work is not the best benchmark for alternatives to medical school or law. Try “Rising wages for computer programmers and MBAs are forcing up salaries for doctors, and teachers.” Doesn’t sound so strange.
– Teacher pay has increased in lots of countries. The US is a bit of an outlier in the OECD with your social hostility to public education.

Against:
– Nursing, doctoring, teaching, law, etc. are all hard to send overseas, so they have suffered less erosion of bargaining power than factory work.
– Barriers to entry (licensing, certification, etc.) have increased. Same.

• baconbits9 says:

– Factory work is not the best benchmark for alternatives to medical school or law. Try “Rising wages for computer programmers and MBAs are forcing up salaries for doctors, and teachers.” Doesn’t sound so strange.

This is ignoring half of the equation, if programmers are more productive then there will be competition across employers to hire more programmers and there will be increases in capital backing those hires. If other white collar professional wages are to rise from this competition it has to come with DECREASING investment in their professions. We would expect to see fewer doctors, but instead we see more doctors. We would expect fewer teachers not more teachers.

• m71 says:

I don’t understand your comment. Why should rising wages require decreasing investment? Or do you mean simply decreasing spending? For the latter, the reason is relative demand-elasticity between the alternatives. Baumol is about the supply side, not the demand side.

There are lots of services (e.g. personal servants) where there has been a large reduction as a consequence of Baumol.

• baconbits9 says:

Baulmol is about rising wages in areas without increased productivity. If wages are rising and productivity is staying the same then returns on investment must be declining.

• m71 says:

I still don’t get the connection. Yes — wages and investment returns are rival, but also rival with the consumer surplus. Who eats the cost increase is determined by relative bargaining power. In the case of higher-Ed, it’s been the consumer.

When oil hit \$130, consumer demand was generating good profits for extraction investment while short term worker productivity was flat or worse while wages rose. The consumer ate that increase too.

• baconbits9 says:

Dropping in consumer surplus doesn’t resolve the issue, if prices of these goods are going up relative to prices of other goods then you expect lower demand of the higher priced good. If you claim that demand is rising then the answer is no longer Baumol’s explanation, but an increase in demand.

• Alchemist says:

@m71
“The US is a bit of an outlier in the OECD with your social hostility to public education.”

Possible but unlikely IMHO.

More likely: US rewards public sector teachers with low salaries and very high health care and retirement packages for those who stay 20 years and vest all the benefits. Many comparisons carelessly compare ‘salaries’ without looking at overall compensation. On the other hand, many OECD nations pay students to attend all the university training required to get a teaching cert. instead of the US policy of loaning them huge \$\$ that they have to pay back. Also many OECD nations provide health care out of tax dollars, so it gets hard to compare.

43. Ninety-Three says:

My issue with using the Baumol effect to explain cost disease is that cost disease is mysterious. No one is totally sure why it’s happening, or where all the money is going, and that does not sound like the effect described. In the example of violinists, there would be no mysterious cost disease affecting the violinist industry. It would be plain to everyone that violin performances costing much more these days has something to do with violinist salaries costing much more these days. People might then go on to be confused about why that is the case but they’ve progressed one step past the “everything is more expensive and I don’t know why” of cost disease.

The Baumol effect acts on people’s salaries, and if cost disease were straightforwardly caused by rising salaries (nevermind why they’re rising), surely someone would have noticed that already and we’d be talking about the mysterious fact of doctor’s salaries increasing despite no productivity gains, rather than the mysterious fact that medicine costs more for no apparent reason.

44. baconbits9 says:

I’m sad to see another analysis on cost disease skate over the obvious discussion that has to take place at the opening off any analysis.

The US is set up to be an inflationary environment, the Federal reserve has been intentionally creating such an environment in the post WW2 world. As long as some categories of goods and services are decreasing in cost then other must be increasing in cost to compensate for that OR the Fed would have to fail to create inflation.

• Alchemist says:

@baconbits9

“As long as some categories of goods and services are decreasing in cost then other must be increasing in cost to compensate for that OR the Fed would have to fail to create inflation.”

This is a good point.

“Federal reserve has been intentionally creating such an environment in the post WW2 world.”

This is as good an explanation to the @Ninety-Three mystery of ‘why it is happening’ as any. You are on target.

The mechanism by which inflation results in falling real wages is not well publicized or discussed in the literature to the best of my knowledge. But the correlation is extraordinarily well documented, so the causation mechanism is really just icing on the cake as far as understanding the world is concerned. INFLATION of the amount of \$\$ >> much greater than >> ECONOMIC GROWTH => FALLING WAGES. Every. Single. Time.

• baconbits9 says:

The mechanism by which inflation results in falling real wages is not well publicized or discussed in the literature to the best of my knowledge. But the correlation is extraordinarily well documented

Of course there is a strong correlation as inflation is a component in real wages.

• Alchemist says:

Of course there is a strong correlation as inflation is a component in real wages.

Technicality, but perhaps important.

Economists define “real wages” as wages with the inflation component subtracted out. So technically, inflation is NOT a component of real wages. By definition.

The correlation I refer to is this: Wherever you look historically, periods of inflation correspond to periods of falling real wages. Because wage increases never keep pace with the falling value of currency.

Granted, correlation does not prove causation. But one needn’t understand how soot causes mutations in an oncogene, a tumor-suppressor gene and how that leads to cancers spreading to understand that smoking is bad for you. All you need is the epidemiological data that show the correlation between smoking and cancer rates.

• baconbits9 says:

The correlation I refer to is this: Wherever you look historically, periods of inflation correspond to periods of falling real wages. Because wage increases never keep pace with the falling value of currency.

Do you mean periods with any inflation, periods of high inflation, or periods of higher than normal inflation?

• baconbits9 says:

Economists define “real wages” as wages with the inflation component subtracted out. So technically, inflation is NOT a component of real wages. By definition.

No, its a component in calculating real wages, therefore you expect it to correlate with real wages. That it is subtracted out of nominal wages would lead you to guess that rising inflation would correlate with lower real wages.

• baconbits9 says:

To put it somewhat differently- Nominal wages and inflation are measures, real wages are our output when we combine the two. You only ‘subtract out’ inflation because it is reported as a positive number but you could report it as a negative number and ‘add it in’ with just as much validity.

• Cliff says:

periods of inflation correspond to periods of falling real wages

Periods of modest and expected inflation do not correspond to periods of falling real wages

• Alchemist says:

baconbits9,

Nominal wages and inflation are measures, real wages are our output when we combine the two.

You only ‘subtract out’ inflation because it is reported as a positive number but you could report it as a negative number and ‘add it in’ with just as much validity.

Adding a negative addend and subtracting a positive addend are the same operation.

I think your confusion comes because if you look at a series of wages over time, say from 2000 to 2010; Typically this will be a monotonic increasing series of numbers. Then you can choose a year as a reference year and convert the other years into ‘constant dollars.’ So if you choose year 2000 dollars as a reference, then you will decrease the 2010 figure, because year 2000 dollars were more valuable and one needs fewer of them to express the sum. Or you could choose 2010 as a reference, then you would increase the 2000 figure, because 2010 dollars are worth less than 2000 dollars, so you need more of them to express the quantity.

I used the term subtract because what you are attempting to do with this calculation is to remove the distortion caused by inflation. So my wording was less than precise.

45. jgr79 says:

Some things that might resolve your complaints:

– Looking at average hourly wages is misleading. People today are paid in a lot more non-cash stuff than 30 years ago (mostly in health insurance). This means that effectively real wages have risen more than you see if you only look at the cash portion of their wages. This could partially resolve a couple of your issues (how Baumol can take effect even when wages aren’t rising (they are rising); why some of these sectors don’t seem to have higher wages (the wages are higher); etc.).

– In most of these sectors, a really high percentage of the spending goes to pay someone’s salary (like 90% or something). So for prices to rise, the total salary paid per unit of service pretty much needs to go up. Which means either salaries have gone up or we have more people involved in providing each service. If it’s the latter, then productivity has actually dropped in these sectors. It’s like if today’s violinists can only play every other note so we need twice as many of them to play the same symphony. Prices are going to rise a lot more than the rise in violinist income. In which case Baumol is still the reason salaries don’t drop to match the lowering productivity, but now we also need to understand why we need more workers to provide the same number of services (which could be how regulations or other similar effects are manifesting themselves).

– Perhaps “affordability” for some products isn’t simply (income/cost of item). Maybe it’s more like (disposable income/cost of item). Disposable income can rise faster than income. E.g. some time in the past, an item cost \$10 and salary was \$100, so the naive affordability calculation gives affordability=10.0. Today, the item cost is \$40 and salary is \$200, so affordability is 5.0 (i.e it’s less affordable). But what if the cost of living was \$90 in the earlier period and \$150 today. Then disposable income increased from \$10 to \$50 and the affordability increased 25% from 1.0 (10/10) to 1.25 (50/40). If something like this is happening in health care and education, then these things are getting more affordable, not less, just as Baumol would predict.

• Alchemist says:

@Scott

“This matches my understanding of the Baumol effect. But it doesn’t match my perception of how things are working in society. College has actually become less affordable.”

@jgr79

“If something like this is happening in health care and education, then these things are getting more affordable, not less, just as Baumol would predict.”

Baumol and Bowen’s work was published sometime in the 1960s I think. Stuff was getting more affordable for wage earners then. Since 1973ish stuff has been getting less affordable. University education and health care are the most notable examples for a number of reasons.

Again, the Baumol effect describes a healthy wage market determining how much of society’s increasing wealth gets distributed to which individuals – Answer: a slight advantage goes to areas of expertise where the biggest productivity problems are being solved AND the bulk of the increase gets evenly distributed in increasing wages across all sectors of the economy, even the moribund parts.

So in the sixties Baumol et al used the example of auto workers as a place where productivity growth was hot and wages were rising. More recently it would be at Intel corp where Moore’s law was making electrical engineers rich. I’m not sure what to give as an example today, productivity seems to be rising in many sectors slowly, but not dramatically in any particular one.

The problem is that today we do not have a healthy labor market to set wages. Because wages are set not just across space but also across time.
> Wages need to vary across geographic space. Why? Because when a mine closes in one geographic area and oil wells are sunk in another, or a dam is being built somewhere else, or a hurricane demolishes an area…. There has to be a mechanism to entice workers to go through the discomfort of moving to another place. Hence higher wages for workers where they are too few and lower wages where they are too numerous. No one likes to move. Moving stinks. But moving is part of discovering and increasing the wealth in this world.

> Wages need to be set across time. Why? Because the cost of living varies for people across their lifetimes. It includes rare events like disease, flooded homes, house fires, and all manner of disaster. Only by comparing a bricklayers wages over a lifetime to his cost of living over a lifetime can people see what is a fair bargain for the bricklayer. When wages were paid in commodities this was possible. When wages are paid in fiat money that becomes worth less and less over time, then wage earners lose the ability to compare their wage with their parents wage or their mentor’s wages at a comparable point in their career (long ago.) . Wage earners can’t compare their wages to what they earned even a few years ago, and can’t anticipate how much their wages will buy them a year or more into the future. So the market can no longer establish a proper wage. It is just not possible to make the comparisons needed.

The purpose of markets is to establish proper prices via inputs from the largest possible number of participants. Today we have a wage market that is broken because the value of money drops over time, when it should remain constant or increase slightly. Asset prices also get screwed up, but that is a separate issue.

• Cliff says:

When wages are paid in fiat money that becomes worth less and less over time, then wage earners lose the ability to compare their wage with their parents wage or their mentor’s wages at a comparable point in their career (long ago.) . Wage earners can’t compare their wages to what they earned even a few years ago, and can’t anticipate how much their wages will buy them a year or more into the future. So the market can no longer establish a proper wage. It is just not possible to make the comparisons needed.

This seems plainly wrong to me. Laborers are mostly price takers, their negotiating ability and knowledge of past salaries is basically irrelevant. Is there any academic literature supporting your position? It seems very eccentric.

46. glorkvorn says:

Ah yes, the famous Baumol effect string quartet. This neatly explains why the salaries of classical string players have skyrocketed in recently, leading to all those stories you hear of millionaire violin-bros buying up all the real estate.

At one point, the London Symphony Orchestra tried to cut salaries for their violinists. They all left for higher-productivity growth industries, becoming CEOs, investment bankers, and so forth. It turns out that a lifetime spent practicing violin automatically makes you qualified for every other job, and that classical musicians are all only in it for the money.

(Seriously though, the “string quartet” seems like a terrible example, because it doesn’t match what happened in the real world at all. Isn’t there a better example they can find? It makes the whole explanation suspicious)

• Bugmaster says:

Agreed; Scott’s review of the book makes it sound like a giant “just-so” story.

• Steve? says:

The point isn’t that violinists who have already been trained can easily switch to higher growth industries. The point is that the opportunity cost to society to have a string quartet play a piece continues to increase. The violinist could have chosen to focus on some other pursuit earlier in life, but chose music because the expected salary was worth it. The only way for the expected salary to be worth it, assuming that productivity can’t be increased is to make sure that it is in line with other options when they choose their career path.

• 10240 says:

Ah yes, the famous Baumol effect string quartet. This neatly explains why the salaries of classical string players have skyrocketed in recently, leading to all those stories you hear of millionaire violin-bros buying up all the real estate.

The Baumol effect doesn’t predict that salaries in the affected sectors grow faster than salaries in other sectors. It predicts that prices (in this case concert tickets) grow faster than prices in sectors with high productivity growth, while salaries (of musicians in this case) grow at a similar rate to other sectors. Idk if concert ticket prices have grown faster than the general inflation rate in the past decades, but it’s plausible (just less talked about as they are less necessary than education or medicine, and they are a much smaller item in most people’s budgets).

• Nicholas Weininger says:

Concert ticket price comparisons across large time gaps are tricky for a similar reason to health care price comparisons. What was the quality of a “typical” string quartet’s live performance in 1826? I bet it was much worse than the world-class expensive-ticket superstar string quartet performance you get today, and not much better, if better at all, than the college (or even conservatory) student string quartet performance for which tickets typically remain quite cheap. Unlike in music, though, cheap lower-quality medicine practiced by unsupervised students or amateurs is not a legal option.

Another commonality among classical music, education, and health care is that in all three professions, intrinsic psychological satisfaction is a significant motivator to do the work– a factor not modeled by Baumol, AIUI, who seems to implicitly assume that all professions are compensated only in money. It is at least plausible that, as populations get richer, we would typically see a shift in labor supply toward professions with greater intrinsic satisfaction. This might be connected to lack of rising median wages for those professions, even as the cost of getting a superstar practitioner to attend to you live goes through the roof.

47. Epistemic_Ian says:

In addition to Education and Healthcare, the original cost disease post looked at Infrastructure and Housing. Can the Baomol effect also explain cost increases in these sectors? I imagine it takes as much time to lay concrete in 2019 as in 1950, but I’d think that in most areas, construction would have become much more productive over the past 50 years. The Baomol effect also can’t explain the massive price differences between Western Europe and the US for e.g. building a subway line.

Furthermore, if cost disease is largely caused by the Baomol effect, this begs the question—why aren’t these sectors more productive already, and how do we increase their productivity? If Caplan is right, than there is no way to significantly increase productivity in education, because people want to signal conformity, meaning that the traditional path is the only viable path. (Although decreasing education spending would reduce the overall amount of resources wasted on signaling.) However, I’m pretty sure that there are significant opportunities for productivity increases in healthcare. Consider this hospital, which has adopted assembly line-like techniques for surgery, and has costs dramatically lower than available in the US. Part of this is because labor is cheaper in India, but it is also dramatically cheaper than its domestic competitors. So, again, why don’t we have this kind of productivity across the whole industry? I’m inclined to think that healthcare, in the US and most other countries, just isn’t competitive enough.

48. meh says:

This is my request for a highlight of comments post on this post.

49. Guy in TN says:

There is nothing negative about productivity growth, even if it is unbalanced. In particular, it is important to see that the increase in the relative price of the string quartet makes string quartets costlier but not less affordable. Society can afford just as many string quartets as in the past. Indeed, it can afford more because the increase in productivity in other sectors has made society richer. Individuals might not choose to buy more, but that is a choice, not a constraint forced upon them by circumstance.

The problem here, is the author is jumping from “society” to “individuals”.

If the gains in wealth are concentrated at the top, then “society” gets richer while the median person does not. Hence while “society” (i.e., the wealthiest people) can afford more things, for most people they become less affordable.

Basically: More societal wealth -> more demand -> higher price. But if you aren’t one of the people who saw an increase in wealth (or if your increase was proportionally small), then you end up paying a higher price for no extra benefit.

If the gains in societal wealth were evenly distributed along the wealth spectrum, then perhaps the authors could be justified in breezily skipping past this little tidbit. But since it’s not, it is devastating to their whole “cost disease is a blessing” thesis. So its understandable why they would necessarily avoid addressing it.

Individuals might not choose to buy more, but that is a choice, not a constraint forced upon them by circumstance.

I just want to repeat this quote because of how insanely incorrect it is.

If someone else becomes wealthier, and this causes me to pay higher prices for things like healthcare and rent, then their increase in wealth has made me worse off. If this isn’t an involuntary, outside constraint that causes me to afford less, I don’t know what it.

• baconbits9 says:

You have made just as many errors in your assumptions as you claim the authors do.

Basically: More societal wealth -> more demand -> higher price. But if you aren’t one of the people who saw an increase in wealth, then you are paying a higher price for no extra benefit.

This only holds if the goods you would buy are the ones increasing in demand, so the lower class person who loves live classical music is SOL in this situation but not the ones who don’t have an interest in those performances. The concentration of wealth is unlikely to make rich people demand more bread, or copies of Transformers DVDs or other goods that those not gaining are spending on. Commodity prices are not, generally, rising with the concentration of wealth which is what your model would predict.

• Guy in TN says:

Commodity prices are not, generally, rising with the concentration of wealth which is what your model would predict.

No, the supply/demand model does not predict that for any given good prices will rise. It predicts that prices will rise on average across the entire market.

If you exclude housing, education, and healthcare, and focus only on commodities, then you are missing where rich people are primarily spending their money.

• baconbits9 says:

No, the supply/demand model does not predict that for any given good prices will rise. It predicts that prices will rise on average across the entire market.

As you described it that would be the prediction. More wealth = more demand for goods, therefore without productivity ALL of the basic components of goods should rise. You simply assumed the increase in productivity which is required to make that portion of your argument work, but you need to ignore it for the other half of it to work.

Your assessment only works if the increases in inequality are growing faster than productivity gains, not if there is simply an increase in inequality.

• Guy in TN says:

As you described it that would be the prediction. More wealth = more demand for goods, therefore without productivity ALL of the basic components of goods should rise.

It depends on where people want to spend this new wealth. If they don’t want to spend it as you suggest on Transformers DVDs, then that section of the economy won’t see an increase in prices. Even the “basic components” may remain unaffected: If more wealth leads to people switching from eating KFC to sushi, then the basic components of KFC (chicken, ect) won’t see an increase in price.

You simply assumed the increase in productivity which is required to make that portion of your argument work, but you need to ignore it for the other half of it to work.

I’m having trouble unpacking this. At what point am I ignoring an increase in productivity?

• Guy in TN says:

Your assessment only works if the increases in inequality are growing faster than productivity gains, not if there is simply an increase in inequality.

I think you are misunderstanding my assessment.

My assessment is NOT: increase in inequality-> increase in prices

Instead, its: increase in wealth/productivity -> increase in prices

• baconbits9 says:

@ GinT

I’m having trouble unpacking this. What do you mean by an “increase in productivity”? Are you talking about increase in supply? I don’t see where I am assuming an increase.

Your original statement again

More societal wealth -> more demand -> higher price

How are you getting an increase in societal wealth?

• Guy in TN says:

I edited for clarity, but not in time I see.

Increasing societal wealth obviously comes from an increase in total productivity.

So, at what point in my argument am I ignoring this increase in total productivity?

• Guy in TN says:

To condense my argument further, it is basically that these two things can simultaneously be true:

1. Prices are rising for x because of an increase in total societal wealth, resulting in an increase in demand.

2. The median person is getting a worse deal on x; paying more for less.

This is because the increase in wealth is not necessarily distributed equally throughout society, but the effect on prices is distributed to everyone.

This is why we cannot be so glib, saying “ah prices are rising yes, but that’s just demand being fulfilled, so nothing to worry about. If someone is buying less than they used to, it must be because they want to, not because they can’t afford it anymore”

• Cliff says:

increase in wealth/productivity -> increase in prices

It is quite well known that an increase in productivity results in a decrease in prices. If you don’t know that then it seems like your theory has a pretty fundamental problem. Indeed, productivity growth was driving down prices for hundreds of years prior to fiat money.

• Guy in TN says:

How does my neighbor, becoming more productive at his job and getting a pay raise, leading to more demand for rent, healthcare, ect., lower the prices of these things?

If it is “well known” that him becoming wealthier/more productive leads to lower prices, it’s still bullshit.

• Controls Freak says:

How does my neighbor, becoming more productive at his job and getting a pay raise, leading to more demand for rent, healthcare, ect., lower the prices of these things?

That depends. What is he being more productive at doing in his job?

• Guy in TN says:

It’s yall’s counter-argument, I’m not going to make it for you.

• Controls Freak says:

You selected the neighbor. I’m asking which one you selected.

• Guy in TN says:

Per Cliff, it shouldn’t matter because “an increase in productivity results in a decrease in prices.” All you know is that he’s more productive.

• Controls Freak says:

Ok. Then, I’ll pick. He’s more productive at apple farming. That means that for the same inputs, he’s able to bring more apples to the market… or that he’s able to bring the same number of apples to the market using less inputs. This reduces the marginal cost of producing apples. Assuming that he’s not able to hide his productivity secrets forever, market competition drives the price to the marginal cost of producing apples.

Q: Does the price of apples increase/decrease after your neighbor creates an increase in productivity in apple farming?

• Guy in TN says:

In the realm of pure hypotheticals, “more wealth” could obviously come in the form of more productive health and education sectors, but since I was talking about the real world, and the topic is the Baumol effect, it should be implicit that that’s not what I’m referring to.

Rent, for example. If I say that rent is going up because my neighbor is making more money, and you reply with “ah…but what if the reason he’s making more money is that he invented a land-creating machine, and is terraforming the oceans? Wouldn’t he cause the price of rent to go down?”

Oh damn, got me.

• Controls Freak says:

Forget terraforming. If your neighbor became more productive at constructing housing, or the materials used in housing, or in administration of rental properties… then, yes. The rent would go down.

You specifically talked about productivity, and you specifically proposed a causal mechanism for an increase in prices. You have not restricted this model to any particular subset of products or given any reason why we should categorize different products differently. The Baumol effect runs through a different line of causality, so it doesn’t seem to have much bearing on your claim here. You’re really going to have to be a bit more explicit, because it seems to be the case that you agree that I provided a response to what you actually said… it’s just that you’re possibly internally thinking something different than what you said.

• Guy in TN says:

Tell me where I’m going wrong. Let’s use healthcare as an example:

1. The price of healthcare is rising, due to increasing demand.

2. Demand is increasing, due to people having more money to spend on healthcare.

3. People have more money to spend on healthcare, because they have more wealth from being more productive. This increased productivity generally does not come in the form of increasing healthcare supply.

4. Therefore, increased productivity is causing the price of healthcare to rise.

• Controls Freak says:

The problem is subtle.

Therefore, increased productivity is causing the price of healthcare to rise.

The problem is that this appears to be a directly causal, and generally-applicable statement. You shortened it to something like “increase in wealth/productivity -> increase in prices” several times, which is a directly causal, generally-applicable statement. The directly causal, generally-applicable version doesn’t seem to be true, because otherwise, we could apply it to apples and other stuff.

Instead, it’s pretty easy to show a direct linkage, with productivity increasing on X -> decreases in prices for X. As soon as we change it to (X,Y), then it’s likely that other factors are playing important roles (and may be important enough to change the sign).

50. johnswentworth says:

I went through the NCES data on college costs myself here and here. Two of Scott’s questions have very clear answers from that data.

NCES also includes tables of actual teacher salaries. These show that teacher salaries today are only 6% higher than teacher salaries in 1970. Meanwhile, per-pupil costs are more than twice as high. How is an increase of 6% in teacher salaries driving an increase of 100%+ in costs?

I specifically looked at this, and it’s most likely decreased class sizes. The NCES doesn’t have direct data on class sizes, but the bar association did one study for law schools specifically, and we can also get a proxy measure by looking at number of courses listed in a course catalogue per student. Both of these point to class sizes shrinking by roughly 50% since the sixties, which would mean costs doubling.

Using these numbers: in 1971, the average man would have had to work five months to earn a year’s tuition at a private college. In 2016, he would have had to work fourteen months.

It looks like that source is using “sticker price” tuition numbers, rather than average amount actually paid. Based on the NCES data, actual tuition payments have grown at about half the annual rate of sticker price. This is just colleges playing games with their price – they get ranked higher if they have a high nominal price with very generous financial aid rather than a low nominal price. That said, of course college has still become less affordable even after accounting for this.

I also looked into medical expenditure data a few years ago, although I never wrote that up. The main thing I noticed was that demographic shifts explain most of the growth in US medical expenditures. After adjusting for inflation, old people don’t actually cost much more than they did 30 years ago, but there’s a lot more of them now (and practically all the expenditure is on old people). I’m extremely suspicious of Tabarrok’s discussion of medical expenditures, because I did not see any mention at all of adjusting for demographic shifts.

• eric23 says:

demographic shifts explain most of the growth in US medical expenditures. After adjusting for inflation, old people don’t actually cost much more than they did 30 years ago, but there’s a lot more of them now (and practically all the expenditure is on old people).

But there are just as many old people in rich European countries (probably more actually). Yet their medical expenditures are a fraction of ours, for similar results.

51. Bugmaster says:

Industries are not just getting more productive, in the sense of “number of widgets generated per work-hour”. People’s expectations of the widgets have also risen dramatically. In the good old days, a car had to have four wheels and an engine, and maybe some leather seats if you could afford them. Today, the car has to have anti-lock brakes, a music system, power windows, and good gas mileage. Similarly, in the old days you could go to college to learn classical Greek or Latin, plus perhaps some law or even natural philosophy. Today, you could learn how to wield fantastically complex machinery that uses raw fundamental particles in order to unravel the secrets of nature; and if you do end up studying law instead, you will do so with the aid of distributed knowledge engines, dealing with crimes that were literally unimaginable back in the late 1200s.

The surprising thing is that people can afford any of this stuff at all, not that it costs a little more (when adjusted for inflation).

• eric23 says:

Today, you could learn how to wield fantastically complex machinery that uses raw fundamental particles in order to unravel the secrets of nature

You mean like a telescope? Universities have been using those for centuries.

dealing with crimes that were literally unimaginable back in the late 1200s.

Why should a law student pay more just to learn about new crimes rather than old ones?

52. The violinist as Baumol example bothers me because it ignores a different but highly-relevant demand curve: the demand to be a professional musician.

It is a widespread lament in the contemporary American arts sector that “nobody but the stars can make a living wage anymore”. Everyone has stories of their uncle or whoever who, during the mid-20th century, made a solid middle-class living as a musician or stage actor or set designer or whatever. Whereas today in every sizeable performing-arts local economy (NYC, Chicago, Seattle, Bay Area, etc) the actual wages offered to the thousands of “working artists” who haven’t yet “hit it big” are lower (in real dollars) than they were in 1980 or 1950 or whatever. Mostly a _lot_ lower.

It’s not a matter of reduced audience demand: Broadway keeps setting new attendance records for itself, Chicago’s live-theater sector is vastly larger than it was in 1980, there are far more professional dance performances today than a few decades ago, there are today 200 salary-paying symphony orchestras in the U.S. compared to 10 in 1950, etc etc. Americans for the Arts’ annual data survey is full of statistics like that. And yet all those big cities are today full of hordes of highly-trained highly-skilled performing artists who are patching together a living with teaching or “day jobs” or temping or whatever in addition to their actual performing. If anything those hordes seem to be growing, and spreading to more and smaller cities. Huh?

The answer seems to be ongoing radical increase in the supply of highly-skilled and highly-trained performing artists. The literature of that sector is today rife with laments about the documented fact that more and more young people are graduating top-level conservatories of music, dance, and theater. And as with NFL placekickers the ongoing rise of the floor of professional performance is increasingly obvious…e.g. it was seriously difficult to win an audition to join a “top-6” symphony orchestra in 1980 or 1960 or 1940. Today it is a whole different insane order of magnitude harder. Seriously, _insane_.

So why would _that_ be? Why would, if it is harder and harder to actually make a living in the ever-more-competitive world of being a performing artist, more and more young people be trying to do so?

That answer, I hypothesize, is that (a) it’s always been a dream which certain fraction of young people would value most highly, but (b) we have in recent generations removed all the social stigmas or reluctances which previously artificially repressed a lot of that demand. Whereas 150 years ago proposing to be a professional actress was considered one small step above aspiring to success as a prostitute, and 75 years ago it was considered an unfortunate phase to be talked out of, and 25 years ago it was considered an impractical idea that the parents would hope you would grow out of, today it is if anything considered kind of cool. Even to most parental units, who are addicted to all those “reality contest” TV shows like “America’s Biggest Voice” or whatever it is. This has become cross-cultural too: there are literally Evangelical Christian musical-theater camps now.

Some fraction of the children born in any given year start life with the innate talent to become a singer, dancer, actor, whatever, if they end up gaining the requisite training to do so. We are today converting a much higher fraction of that potential starting pool into young people actually so trained. Because the countervailing social pressures are less and less salient.

So the latent or inherent level of demand for the life of a professional performing artist, which previously always had significant widespread social friction to slow it down in practice, has been unleashed. Which in turn is ultimately, though nobody involved wants to say so out loud, why the very best improv comedians in Chicago still get paid the same \$30 per show that they did in 1985.

Does that particular situation really lend itself to the Baumol cost disease analysis? Which would map onto subjects like education and health care? I dunno that it does very well.

• Ghillie Dhu says:

Those professions are typically characterized as tournaments; the vast majority of participants will experience disappointing outcomes, but the expected value a priori is still high enough to draw in new entrants.

• Indeed. And a lot of that expected value is nonmonetary in nature: lots of people would rather be a mildly- or intermittently-successful performing artist than a lawyer or consultant living in the comfortable upper middle class.

In other words there is a lively marketplace of desire for the fun and cool and entertaining life of the working artist. I have found however that my performing-artist friends (of whom I have many and who I deeply treasure and enjoy the company of) are not at all comfortable with that way of looking at it. Even the ones who I’ve come to view as stellar entrepreneurs, which honestly a _lot_ of working artists are, do not know how to respond to that compliment.

• Nicholas Weininger says:

It is also the case that the quality of amateurs who perform for free in their spare time has increased and this may also put downward pressure on artists’ monetary compensation. This seems especially true in vocal music, an uncomfortable admission for me as an amateur choral musician who aspires to do professional quality work and who has gotten to know many struggling professional singers through that aspiration.

• Yes very true as well. And I share your uncomfortable position, in fact I like your specific description of it and may plagiarize that. Except in my case it’s jazz piano…kids these days, damn.

53. Eponymous says:

I find their case pretty convincing. And I want to believe. If this is true, it’s the best thing I’ve heard all year.

Huh? Didn’t you mention Baumol in your initial post on this? Didn’t many people reply saying, “It’s Baumol”? Not to be too self-focused here, but this includes me.

First, real wages have not in fact gone up during most of this period.

Real wages have gone up. The widely quoted factoid to the contrary refers to median wages. Admittedly wages have increased less than GDP. Partly this is due to increases in other forms of labor compensation, but not all. Probably the best metric here is the labor share of income, which has declined a bit.

By the way, take a look at costs of veterinary care. Much less regulated, little insurance, and costs have mostly tracked human medicine.

Re: SMBC: most classes aren’t taught by adjuncts, and those that are (e.g. community colleges) are pretty cheap.

the Baumol effect cannot explain things getting less affordable

This is partly explained by rising inequality.

If someone can answer these questions and remove my lingering doubts about the Baumol effect as an explanation for cost disease, they can share credit with Tabarrok and Helland for restoring a big part of my faith in modern civilization.

Well, I certainly wouldn’t want to be responsible for doing that!

• baconbits9 says:

Re: SMBC: most classes aren’t taught by adjuncts, and those that are (e.g. community colleges) are pretty cheap.

The point of SMBC was to highlight that hiring 3 full time PhD students to tutor you was cheaper than college tuition.

• JPNunez says:

Also a cool cat on top of the tutoring.

• Alchemist says:

@Eponymous
You’re on the right track 😉

Real wages have gone up. The widely quoted factoid to the contrary refers to median wages. Admittedly wages have increased less than GDP. Partly this is due to increases in other forms of labor compensation, but not all.

The first chart you point to https://fred.stlouisfed.org/series/COMPRNFB
shows the very real inflection point in wage growth that occurs in 1973. The US dumps the gold standard completely.

I have some criticisms of the methods used to calculate ‘real wages’ by the bureau of labor statistics in the last few decades, but the data is still clear. Wage growth has stagnated.

Probably the best metric here is the labor share of income, which has declined a bit.

64% to 57% is not what I would describe as “declined a bit.” 7% of GDP is real money.

Also the effect compounds. Some portion of that 7% gets reinvested and leads to further year over year gains.

Owners of capital generally reinvest income and become even wealthier. Either because it is in their nature (See Warren Buffet) or because they simply can’t spend money on consumption fast enough to become poorer (See Bill Gates, Rockefeller &c.) Usually it takes a few generations to spend off the accumulated wealth. Picketty was correct in the sense that wealth can accumulate at the expense of the working class. He just has no understanding of the mechanisms that lead this. (Picketty enjoyed great success selling books and lecturing, so I won’t call him dumb. And he wrote for an audience that is not numerate. He himself may or may not be numerate. His theory is hogwash.)

the Baumol effect cannot explain things getting less affordable

This is partly explained by rising inequality.

Actually, this IS rising inequality. A few people have no problem affording university educations. They can even afford a few extra hundred k \$\$ to get their moron offspring through the admissions process.

• baconbits9 says:

The first chart you point to https://fred.stlouisfed.org/series/COMPRNFB
shows the very real inflection point in wage growth that occurs in 1973. The US dumps the gold standard completely.

The operative word here is completely, in 1973 the major change was to make previous changes permanent. Claims that only this last step should be viewed as causal here are very suspicious as the US took multiple major, and arguably larger, steps along the way. The US abandoning the standard in 1973 was also driven by economic conditions in the world, and viewing it as causal is suspect on its own.

64% to 57% is not what I would describe as “declined a bit.” 7% of GDP is real money.

GDP has risen by way more than 7% across that period though, nor is there anything special about 1973 as an inflection point for labor share of income.

54. Ghillie Dhu says:

There’s a lot of overcomplication in a lot of the discussion.

Starting from a normal market (downward-sloping demand, upward-sloping supply), Baumol is merely a description of a situation in which opportunity cost increases due to increased productivity elsewhere, thereby shifting the supply curve left. P goes up, Q goes down ceteris paribus.

That last is relevant for health care & education. Effective subsidization in these sectors simultaneously shifts the demand curve to the right. P goes up, Q goes up.

Throw in highly inelastic demand, and you see P going up a lot and Q moving indeterminately.

Finally, P is total cost (i.e., compensation); if the composition of individual compensation is shifting over time (which health insurance’s growing share of labor costs drives), isolating a single component (e.g., salary) will provide a biased estimate of effects.

• Alchemist says:

@Ghillie Dhu
Where did you get your masters degree?

• Alchemist says:

@Ghillie Dhu

From our last discussion on the topic:

And as the years and decades pass a generational knowledge of what proper wages should be for various tasks is completely lost. The proper value of assets gets screwed up by inflation as well, but that is a different essay.

“Proper”? No such thing.

Fair enough. Then I shall define Proper Wage. Then such thing will exist.

Proper Wage: A wage rate mutually agreed upon by buyer and seller of some service for some period of time whereupon each party is maximally wealthier at the end of said period of time. In mathematical terms, of the set of all possible wages select the wage that maximizes the wealth gain of all parties to the contract.

As an example of Not Proper Wages, see the Soviet Union where workers joked “We pretend to work and they pretend to pay us.” Also see most wages paid in Venezuela today.

• Ghillie Dhu says:

Previously you used “proper” normatively (i.e., coupling it with “should”); providing a positive definition for “proper wage” is an attempt to hide the ball.

Good day sir.

• Alchemist says:

Given the society that I was raised in, there did exist a pattern of wage rates relative to the size of the economy. One could certainly say that the continuing pattern was ‘normative.’ That pattern ended in 1972 [Bureau of Labor Standards] or 1973.

Of course it must be naive of me to think that discussing such is the game in play here. You have indeed bested me in comedy.

• Swami says:

Thanks for providing this, Alchemist.

A question though…. Why are you restricting a proper wage to “each party being maximally wealthier?”. Maximumly wealthier compared to what? Does this include (is it net) the costs of research to determine what is maximum?

Seems that a better and more useful definition would be a wage rate mutually agreed to by both parties subject to (relatively) free entry and exit into the market. This ensures win win interactions, and minimizes rent seeking and exploitation.

Correct me if you think I am missing or misunderstanding anything.

• Alchemist says:

Seems that a better and more useful definition would be a wage rate mutually agreed to by both parties subject to (relatively) free entry and exit into the market. This ensures win win interactions, and minimizes rent seeking and exploitation.

That is a very good start, and in line with classical liberal theory. But it turns out to be not enough.

tldr: Employees use historical data of the cost of living, subtract it from a wage offer to see what’s left over, and then decide what wage to accept for the future. If the government lowers the value of money in between the months in the past that tell the person what the cost of living is, and the months in the future when they get paid and have to make mortgage payments, then people get less than they expected to out of the bargain. If inflation is 2% annually that seems like no big deal compared to all of the other uncertainties in life. But it adds up and worse, it compounds over time.

Under a gold standard there were periods of some inflation, and mostly slight deflation as the economy grew faster than the money supply. Now we have always inflation. Economists are very happy to make the calculations that you see in their charts “in 2019 dollars” or “in constant dollars.” All of the ordinary people deciding whether or not to accept a job/salary offer do not include the 3.87% inflation compounded annually in their calculations. Because inflation always works against them, there is a constant drift toward lower wages. Year in and year out. A few percent here, a few percent there. This shows up very clearly in the data, as shown throughout this discussion. Gramm and Early have their heads in the sand. They don’t want to see it. The implication could be that they are doing a bad job, and that would feel bad to them. So their brains simply ignore the data.

• Swami says:

Alchemist,

We were discussing the definition for a “proper wage”, and for some reason you go off on a tangent on the effects of inflation on wage rates. Honestly, I have no idea why.

• Alchemist says:

Swami,

Questions are helpful. Thanks.

We were discussing the definition for a “proper wage”, and for some reason you go off on a tangent on the effects of inflation on wage rates. Honestly, I have no idea why.

So we have a definition, the interesting question is how does society go about discovering what wages should be?

The mechanism we have developed is the market. Markets exist for price discovery. All of the participants get a vote. And some consensus developes as to what the answer is to the question ‘what is the proper wage?’

The reason that I keep ranting on about inflation is this: Consistent inflation of the money supply over time breaks the wage market. The effect is very strong and very striking and it stands out clearly in the data. The details of how this happens are perhaps a distraction, but I offer them up in support of the notion that the well observed correlation between inflation of the money supply and stagnated wages is there for us to see because the former causes the latter.

55. danculley says:

A couple points on affordability:
(1) The Baumol effect would only have affordability remain constant if the quantity demanded was also constant. In education and health care, this is not the case. In education this is easier to see as we send an order of magnitude more people to college than we did in 1970. Notwithstanding that, I don’t think affordability has changed much for the marginal student. Sure, with ten times as many people competing and fixed places at Ivy League schools, they have gotten more expensive. I don’t see any evidence that community college — where the marginal college student is — is less affordable.
(2) Similarly, Baumol effect would only maintain relative affordability if labor productivity growth were evenly distributed with who is doing the consuming. That’s a reasonable assumption for say, haircuts, but not for education and health care. The substitute occupation for a college professor may be say, finance, where productivity is increasing strongly. But at the same time, the productivity in industries where the average person paying for college works may not be going up. There isn’t a necessary correlate between the two.

Also note that there is not really a conflict between other explanations and the Baumol effect. For example, administrative bloat may be the reason why education productivity is not increasing at the same rate as other industries. The Baumol effect is agnostic to what the cause of the differences in relative productivity are.

56. ConnGator says:

Wage stagnation is a myth. Three reasons explaining this:

1) Most charts only take into account money wages, not total compensation. Benefits have risen very substantially in the last 40 years.

2) Comparing the 70s to today does not take into account the much better quality of today’s products.

3) The Consumer Price Index (CPI-W) overstates inflation and understates real compensation.

Here is a good overview from last month: https://www.wsj.com/articles/the-myth-of-wage-stagnation-11558126174

• Alchemist says:

@ConnGator
I still get the Journal delivered. I read that article by Phil Gramm and John Early. I would like to believe in a world where it was true but as far as I can tell its bunk. Their article opens with:

Perhaps the most common indictment of America’s legendary prosperity is wage stagnation. Bureau of Labor Statistics data show that average hourly earnings of production and nonsupervisory employees peaked in October 1972 at \$23.26 (in 2019 dollars) and were at that same level in March 2019.But do these numbers remotely describe the life you have lived over the past 45 years, or square in any way with numerous official measures of changes in what Americans actually own and consume? In short, should you believe your eyes or government data?

I am reminded of the joke about the wife who comes home to find her husband in bed with another woman. The husband insists he’s innocent: “Who are you gonna believe, me or your lyin’ eyes?”

I started working in construction as a teenager between stints at college in the late 1980s. Real wages today in construction are far less. Obviously some other industries have fared somewhat better, others not. My mother is a PhD psychologist. Since I was a kid through today wages in her field have risen several fold in dollar terms 3 to 4 times. But price levels have gone up by a 7 to 8 times.

I am going to believe my lyin eyes over Gramm and Early.

• Cliff says:

I am going to believe my lyin eyes over Gramm and Early.

That’s a horrific mistake

• Scott Alexander says:

Benefits are mostly irrelevant. CPI considerations are relevant, but wouldn’t they be equally relevant in discussing eg the cost of health care?

• Scott, in that Feb 25 2019, post seems to have landed in the middle: that wage stagnation is not a myth, but is being exaggerated:

“Contrary to the usual story, wages have not stagnated since 1973. Measurement issues, including wages vs. benefits and different inflation measurements, have made things look worse than they are. Depending on how you prefer to think about inflation, median wages have probably risen about 40% – 50% since 1973, about half as much as productivity.”

Regarding the degree to which wages have lagged behind productivity, Scott concluded that about half of the effect is a matter of how increases are wages is being distributed:
“This leaves about a 50% real decoupling between median wages and productivity, which is still enough to be serious and scary. The most important factor here is probably increasing wage inequality.”

If I’m understanding his conclusions section correctly, Scott’s literature review concluded that only a quarter to a third of the wages lag is actual as opposed to being a matter of distribution. Right?

57. Loriot says:

This made me wonder if we simply ignoring things that aren’t subject to Baumol effects.

For example in education, Duolingo provides a basic introduction to language learning to people all over the world at essentially zero marginal cost. Wikipedia and Google and MOOCs make it easier to learn things then ever if you are sufficiently dedicated. Sure there are things that having an actual teacher do which are impossible to replace, but a surprising amount of instruction is available for free.

Likewise, live music is more expensive than ever, but recorded music is almost too cheap to meter (and the costs we do pay are only due to government intervention, rather than technical limitations)

58. Floccina says:

Using these numbers: in 1971, the average man would have had to work five months to earn a year’s tuition at a private college. In 2016, he would have had to work fourteen months.

Doesn’t Baumol’s disease say that it is just as affordable if the cost in time of everything else he used to buy has gone down in total cost by fourteen months?

59. Thomas says:

I use a \$10/mo phone. It allows text and has an internet connection. The phone itself cost about \$50.

60. Deej says:

It’s surely not just baumol. Also the more you have to dpend the more useless stuff you buy.

Civil service policy. Our policy making hasn’t gotten better but our publications are better produced and we spend more on consultants to provide “evidence” for whatever we were going to do anyway.

61. Michael Naunton says:

It’s hard to take theses like this very seriously.

String quartets are selling an inferior good versus a DJ with a pile of digital songs and a decent sound system. People want to dance to a variety of music that a string quartet can’t create.

Academia and health care are two of the most (self) regulated industries in America.

Music and professional cartels are hardly related in either barriers to entry or demand.

Certain economists will argue that this is all part of a grand objective reality that means teachers should be poor and bankers rich because of “productivity.” I was a banker, and wouldn’t even try to say that with a straight face.

• Freddie deBoer says:

String quartets are selling an inferior good versus a DJ with a pile of digital songs and a decent sound system.

says who

62. Carl Milsted says:

Imports are a huge factor. Thanks to cheap imports official inflation numbers are way below what they would be otherwise. If we paid American workers to make televisions, toasters, dresses, etc. the ratio of wages for medicine and education with respect to the rest of the economy would be much less dramatic.

Back in the 1970s high end department stores sold dress patterns along with fully assembled clothes. Upper middle class people (like lawyers) still felt the bite from the cost of buying clothes. This is also why clothing stores weren’t self service: they were more expensive relative to local wages.

Factory wage earners compete with cheap imports to sell their labor. Their wages have been driven down with respect to professionals which aren’t outsourceable.

This is why they are angry enough to vote for Trump.

(And many of them were angry enough to vote third party when H. Ross Perot ran.)

BTW, if we have “free trade” then imports don’t get taxed. We thus tax domestic labor to pay for the government services by the cheapened domestic laborers who must now depend on the government to pay for local professional services. The cheap merchandise at Wal Mart is subsidized by the tax system/welfare state.

63. Harmon Dow says:

The common denominator I see is government subsidies.

• Flavoryeagle says:

Figure 6 in T&H show that subsidies have substantially decreased for education from 1980 to today but more students are entering college.

Please explain how it is all subsidies when they decreased the subsidies, the cost went higher?

64. Stuart Buck says:

I had a discussion with Tabarrok on Twitter several days ago about one of these points, i.e., that salaries actually haven’t gone up very much in K-12. https://twitter.com/StuartBuck1/status/1133477305416847360 He pointed to an increase in benefits, but I said that health insurance and pensions couldn’t possibly explain the massive increase he had claimed. I also said that it was odd to pick the chart on instructional expenditures and divide it by the number of teachers, just to get a (very rough) proxy for salaries, when in fact the same government agency has been tracking actual salaries. If I were cynical, I might say that they found a way to manufacture a K-12 salary figure that showed the most dramatic and implausible increase, rather than using the actual numbers. But that might be unkind.

• Steve Sailer says:

My impression is that public K-12 spending in California is not that high on annual costs, but in recent decades has climbed considerably for capital costs: e.g., most Baby Boomer Californians were schooled in buildings that were basically big shacks. But lately public schools have built insanely expensive edifices, like the Giant Japanese Robot from Outer Space high school pointing its flamethrower at the L.A. Cathedral across the 101.

65. Stuart Buck says:

More broadly, I think that the Baumol effect is less plausible for health and education than the following:

1. In both sectors, consumers have few feedback loops or good metrics for quality. Huge asymmetries of information.

2. In both sectors, there are many third party payers — government, insurance, etc.

3. In both sectors, there are huge political disincentives for the actual payers (government, insurance) to ever crack down on expenses. (You can attribute this to public choice theory, or to Robin Hanson’s theory of how expenditures here are a signal of “caring,” or whatever.)

4. In both sectors, the people who make money find ways to make their services mandatory or close to it. School attendance is literally mandatory for many years. For college, many jobs (perhaps due to signaling) make college attendance mandatory as well. For health, there aren’t as many legal mandates (other than where children are involved), but doctors and hospitals do their best to persuade people that they absolutely must have such-and-such a test, service, etc. And given asymmetries of information, plus third-party payers, this works a lot of the time.

Put all those together, and it’s a perfect recipe for inefficiency.

• Nicholas Weininger says:

Then why, as another commentator points out, does veterinary care have the same cost disease as health care for humans? Or does it actually not?

• Clutzy says:

I’ve never seen that evidence, but there is good evidence that things like plastic surgery and lasik have not followed this pattern.

• gkai says:

I think it does not, especially for farm veterinary. Contrary to pet veterinary, for farm animals is no signaling and no attachment, it’s regulation and cost/benefit analysis. Regulation has obviously increased, but I doubt veterinary cost for farm animals has followed other medical expenses, it’s probably on a completely different slope.

Human and pet medicine have the “there is nothing too costly, you can not skimp on health” coupled with a largely opaque industry: As a patient, it’s difficult to have an idea of treatment efficiency, and impossible to know of practitioner efficiency. So a cost/benefit analysis is both difficult to do for psychological reasons (high personal stake) and for pratical reasons (lack of reliable information). And don’t forget that medical expenses get kind of unbounded for end of life treatments, basically it could absorbs all surplus/savings. I do not see how this could change, it’s inexorably linked to a wealthy society with low accidental death rate, and Pet-as-quasi-children status.

For education, there is the typical inefficiency of an old bureaucratic institution heavily sponsored by the state, and the fact that research is sponsored by education cost, in a way.
Absent from other western coutries, the exploding student debt in US is imho a ready to burst signaling bubble…

66. thevoiceofthevoid says:

Typo: “Maybe in 1826, when factory owners were earning \$1.14/hour and violinists were earning \$5/hour, so no violinists would quit and retrain.” seems to be missing a primary clause. Delete the “when” or the “so”.

• 10240 says:

Also, factory workers, not owners.

67. benwave says:

I don’t really know the Baumol effect, but if it indeed only acts on the price of labour, as your article suggests it does, then I think you’re right to conclude that it cannot explain cost disease.

On reading the article though, I did have a small niggling question about the CPI-indexing that is a common factor through these sorts of comparisons. If the CPI incorporates costs of healthcare, or other cost-disease goods such as housing, education, etc. in an inappropriate way couldn’t that throw off all of the calculations? And indeed, since it is a single number and different classes of people will definitely be buying different proportions of these types of goods, it seems like it must be a bad index for at least some significant group of people.

• Ghillie Dhu says:

Baumol can act on any input used in sectors where the rate of growth of output relative to that input diverges; labor is just the motivating example. It’s fundamentally a story about opportunity cost.

• benwave says:

Okay, I think I can imagine how it would work on an input of, let’s say, land. But how much of cost disease can be explained by Baumol operating on non-labour inputs? I’m still inclined to think that monopoly power plus financialisation contributes the greater part at this stage.

• Ghillie Dhu says:

But how much of cost disease can be explained by Baumol operating on non-labour inputs?

Probably very little which, I suspect, is why labor is the classic example of the effect. For the health care & education sectors, I agree that other factors (I’d genericize “monopoly power” & “financialization” as “inelastic supply” & “subsidized demand”, respectively) likely each contribute a greater share of the price increase than Baumol.

68. PeterDonis says:

So violinists must be paid 5x more for the same work, which will look like concerts becoming more expensive.

Yes, and that has happened. But something else has also happened with concerts: a much smaller number of people go to them. Which means there are a much smaller number of them, and hence a much smaller number of people who can actually make their living as professional musicians, even with the salaries of professional musicians forced to keep parity with other professions.

And the thing about education and health care is that they have not experienced this second effect. The number of people going to college has continued to rise as costs have risen. The amount of health care being provided has continued to rise as costs have risen. And, as you point out, much of the cost growth cannot be linked to growth in wages and salaries anyway.

It seems to me that this can only be because education and health care are subsidized–individual people are paying less of the cost (student loans and grants, employers paying a portion of health insurance premiums–and those premiums being pre-tax, etc.). This means people will want more of these things, and it also means they can’t judge whether the things are worth what they actually cost (because the people receiving them don’t see the actual cost). This seems like an obvious recipe for increasing costs uncoupled from productivity.

• Flavoryeagle says:

Except the percentage of costs borne by consumers has increased, not decreased. This is explicitly stated in T&H figure 6.

In 1980, 50% of costs were subsidized and today 25% are subsidized.

That did NOT decrease the enrollment…

69. wgkb says:

There are really two ways to look at the Baumol effect: (1) changes in relative productivity in one sector vs another cause the sector with increasing relative productivity to become relatively cheaper. And (2) inflation as measured with respect to a basket of goods which have mostly become much more efficient to produce is being massively underestimated by changes in price.

It’s this second interpretation of the effect that best positions itself to explain why college appears to be much more expensive today. Real wages are nominal wages deflated by a CPI deflator, but this deflator has been computed with respect to the prices of goods that have become much more efficient to produce. If industrial farming techniques make apples easier to produce by a large amount, but we see the prices of apples go up by a small amount, we shouldn’t think inflation is low; we should think inflation is so big that it swallows up the efficiency increase in the production of apples.

If this is the case (and it is if Baumol’s effect explains the significant increase in education costs as a stagnation in efficiency, rather than a marked decrease) then the real effect we have to explain is why real wages have gone down so much in the last half century, not why healthcare and education have gotten so much more expensive. This bears directly on Scott’s post on Technological Unemployment: unemployment may not be increasing as a result of increased automation but real wages certainly are.

70. Steve Sailer says:

Some things have gotten cheaper. For example, pre-Internet, for Scott to write essays citing as many sources as he does, he would have had to have had multiple research assistants with access to a university or major city flagship library.

On the other hand, the basics of a bourgeois life:

– Enough higher education for a respectable white collar job
– Unrationed health care for the family
– A house with a yard in a decent school district without an onerous commute

have gotten dramatically more expensive.

One obvious explanation is that the population has grown substantially, as has the cost of desirable real estate.

71. Swami says:

While I agree with the general discussion that costs of college have gone up, we need to remember that this has not necessarily led to decreases in affordability outside of elite universities.

The below study shows that two year colleges continue to be quite affordable (less than \$3k per year), especially considering the continuing increases in payoff to a good degree (outside of the humanities). In addition, many students are getting grants paid by others. And this doesn’t even count the benefits of switching to partially online degrees which MBAs seem to gravitate toward.

In most larger states, it is still possible to get a good college education for the cost of two years of community college and a transfer to two years at a state school while living with parents or friends.

https://trends.collegeboard.org/sites/default/files/college-pricing-2012-full-report-121203.pdf

Here is from the summary….

WHAT STUDENTS ACTUALLY PAY
In 2012‐13, full‐time undergraduates at public four‐year institutions receive an estimated average of \$5,750 in grant aid from all sources and federal tax benefits to help them pay the average \$8,665 published tuition and fees. The students pay an average net price of just over \$2,900.
• Full-time undergraduates at private nonprofit four-year institutions receive an estimated average of \$15,680 in grant aid from all sources and federal tax benefits to help them pay the average \$29,056 published tuition and fees. The average net price is about \$13,380.
• Full-time students at public two-year colleges receive an estimated average of \$4,350 in grant aid from all sources and federal tax benefits to help them pay the average \$3,131 published tuition and fees, as well as some of their other expenses.
• Average net tuition and fees paid by full-time students enrolled in public two-year and in private nonprofit four-year institutions are lower in inflation-adjusted dollars in 2012-13 than they were in 2007-08.
• About \$1,410 of the \$1,850 increase in average published tuition and fees at public four-year colleges between 2007-08 and 2012-13 was covered by increases in grant aid and tax benefits, and the student share increased by about \$440

72. NostalgiaForInfinity says:

How much of the increase in the cost of healthcare is people having more expensive health problems (or problems that were previously untreatable but now can be treated at great cost)? I don’t remember seeing this discussed in the previous cost disease post. Are there any articles someone can link to that do [attempt to] address this?
Example: a frequently quoted stat in the UK press is that ~1/10th the NHS budget goes on complications from type 2 diabetes. Given that the prevalence of type 2 diabetes has roughly doubled over the past 20 years, wouldn’t that disease alone explain a non-trivial increase in British health spending?

73. Spiritkas says:

I grasped an intuitive understanding/metaphor for the post which I didn’t see expressed elsewhere for this effect which doesn’t help take it away from a ‘just-so’ story, but helped me engage with the concept in a more relatable way than a stringed quartet vs a construction worker.

I took the Baumol effect to treat workers like a series of buckets. The size of each bucket might change, i.e. there are more doctors or teachers than in the past, but the amount that it can contain matches demand.

Instead of being filled with non-agentic water; each bucket is filled with frogs which have limited jumping power. If the bucket is overflowing then the frogs simply don’t fit and will spill out – matching when there is oversupply and a drop in wages. Crucially though, the frogs can choose to jump out of their bucket to another bucket if they feel they can earn more there / have more room to live-better conditions in that other bucket. Workers are not equivalent and there are barriers to access, but overall the Baumol effect is to keep the frogs happy. Tadpoles get to chose, to a limited extent, which bucket they want to join. If you are poorer you might pick locksmith over roofer, if you are richer you might pick accountant over secretary.

If you’re a software engineer and earn a high salary, you’re very unlikely to jump into a crowded bucket doing another job. The cumulative effect mixes markets and agency to seek a constant labour to wages productivity coupling scenario, but I’m not sure how this links to cost as labour costs are increasingly less important in determining cost, i.e. 20% of doctor wages are what patients pay.

If the productivity etc. works out that you can earn \$25/hr as a construction worker, but \$30/hr doing another job which is accessible to you, then at least some frogs will jump buckets/switch until a new equilibrium is reached through wage rises and declines, or continue for a while until new demand is met, etc. with waves of new educational choices evening this out over time for those jobs with higher barriers to access.

I’m taking the Baumol explanation to treat workers like a sort of chemical equilibrium equation mixed with worker agency and is fairly agnostic to industry with the only variations being caused by protectionism, regulation, changing market conditions on their way towards new equilibria, etc.

I think this means the ‘disease’ part of the cost disease somehow relates to scale where the importance of the individual diminishes as there are more people (Black death and the Renascence caused by increased value of individuals – killing peasant farmers was tougher because you needed them more?) . If there are only 5 doctors in a 10,000 person town vs if there are 50 doctors in a 100,000 person town, then each individual doctor is less valuable/more replaceable due to scale. I think the idea is that value isn’t entirely linked to a simplistic demand/supply equation but that there is some effect to the replacebility of any given person beyond the mere fact of their wages.

We often assume a limitless supply of workers or retrained workers which is only true in the modern context. Increased numbers of people lead towards more efficient ‘sorting’ of people into jobs, but this efficiency can depress wages precisely because it is so easy to find another worker. Your supply of frogs isn’t limited to those frogs currently in your bucket, but is expanded out in a wage-depressing way to all those other unhappy frogs in other buckets. Saying there are 3 million teachers in a country and counting that as your ‘supply’ is incorrect, there is a shadow factor caused by scale/replacability/bucket jumping frogs from other fields which impacts wages. There could be 5 million or 10 million other people who might also go into teaching if their jobs get worse or if teaching gets better. This highly efficient sorting and devaluation of people has increased worker supply and decreased worker value/wages to create the ‘cost/wage’ disease?

If you don’t want to work for a certain rate, then there are 49 other doctors to take your place? Or there are 3 million other teachers to take your place. There are wage limits to how much anyone can be paid, if you are the only teacher in a town you wont be able to demand super high wages, but somehow the dilution of your value in high numbers of people collectively dilutes wages, leading to a ‘cost’ disease which could be more of a ‘wage’ disease? Baumol talks a lot about productivity, wages, and labour and a lot less about costs, so it seems like the problem is a decoupling of wages and costs – so the answer could be due to disruption in coupling coming from costs, wages, or both.

I can’t say that’s super satisfying as an explanation for cost disease (but maybe for at least part of ‘wage disease’), but I think I’m understanding the Baumol explanation via the buckets of frogs metaphor? And the absolute scale of workers which are not being counted as a ‘shadow supply’ who can switch fields/jump buckets based on perceived wage advantages in other buckets?

74. Bram Cohen says:

Recently enrollment at liberal arts colleges has been plummeting, to the point of many of them going out of business, apparently because new college students are getting much savvier about what degrees are actually useful. This probably implies a meaningful increase in college productivity, because they’re delivering a more valuable product.

75. Jacobeus says:

Imagine there is a particular vitamin that all people need to include regularly in their diet in order to survive. The problem is, the vitamin is produced only from the milk of a type of goat that, while once extremely common, due to climate change is now an endangered species. Furthermore, the goat only produces the vitamin in its milk during three days a year, and only if the temperature is between 67 and 72 degrees F. This places a hard cap on the amount that can be produced (just barely enough to sustain the current population).

And while technological progress has continued over the centuries at a steady pace, unfortunately, scientists have not discovered a way to synthesize this particular enzyme. In fact, it is the only thing that, strictly speaking, is necessary for survival AND has not also been easily and cheaply mass produced. For that reason, the prices of nearly all other goods and services have fallen dramatically, while the price of procuring the goat nutrient has stayed the same – and dramatically increased in terms of the percentage of income people typically pay for it.

The result of this has been that, over the years, the elite of the small country in which the goats are farmed have been able to acquire a gargantuan share of wealth, almost mythical in splendor and scale – and in order to protect their claim have used this power almost entirely to ensure their own defense, via funding of various foreign conflicts and pulling the strings in regional politics. Their most dreaded nightmare is that someone, one day, will figure out a way to synthesize their vitamin, and suddenly the entire empire will collapse in a single day. As they’ve acquired so many enemies over the years, who curse the ground they walk on as they have been forced to pay them massive tribute or face death, they know that vengeance will be fierce and unstoppable if ever, one day, they lose that power.

Eventually, they decide to make a bold move: They will sacrifice their wealth in exchange for safety, by artifically lowering the price far below what they could potentially charge for it. They still make a good, solid, and secure income from it, while also seeming like saints for doing what seems to everyone like the morally correct thing to do. And in the process they also manage to use their political influence to engineer international treaties that govern and regulate the production and distribution of the vitamin, declaring it a human right for all. They end up trading wealth for prestige.

Eventually, the vitamin finally is synthesized, but at that point no one really notices – everyone had been getting it for free anyway. The small, rich country that used to be the sole owner of something the whole world needed is still small and rich, just not blindingly rich, and not an enormous target anymore. Even though it was their only industry at the time, they used their prestige to hire, I dunno, a lot of scientists and bankers or something.

The point is, the difference in consequences with Baumol effect stuff is that, for things you can live without, it may get really unaffordable, but just becomes a luxury item anyway. What we really should be asking is why education and healthcare are getting more necessary than they were in the past.

• The Nybbler says:

Imagine there is a particular vitamin that all people need to include regularly in their diet in order to survive. The problem is, the vitamin is produced only from the milk of a type of goat that, while once extremely common, due to climate change is now an endangered species.

This scenario is not at all credible. If such a goat existed, there’d be more of them than we have cows. If you come up with enough farfetched technical reasons this can’t be, the world looks so much different than it is now that there’s no analogy. Most likely you’d have a hydraulic empire.

76. Charles Engelke says:

Scott,
It’s very simple: when you look out the window of your parked car and see all the houses and people moving past you at a uniform speed, you aren’t parked. — You are the one who is moving!

The illusion that the costs of services rise is just due to viewing them relative to a similarly moving reference frame:
In terms of average compensation per hour (how long the average person works to pay for something) haircuts, rents, new homes, college tuition, all cost about the same in any era. It is the output of sectors with productivity growth that are continually getting cheaper.
But this is covered up by a ceaseless surge of inflation (which is actually nearly twice as big in my ‘real terms’ as economists typically assume, averaging near 5% a year since 1930).

Productivity increase means producing more output with less (or lower grade) work. Therefore the items produced become cheaper in terms of work.
In turn, prices paid for manufactured items, food, and so on, decline in terms of how long the consumer must work to buy them, because fewer man hours were used to make them. Money is basically a way that people can barter their work effort for the products of other people’s work effort. As US manufacturing requires only a quarter as many workers per ton of output in 2019 as it did in 1950 the real price of comparable items are about a quarter of what they were in 1950. Standards of living rise because we get more and more stuff for the same or less work.

As for the “paying for college” case you thought violated these rules: I think the problem must be in some inconsistent definition of salaries in the numbers you found for the two time periods:
I calculate \$19,300 as the true equivalent of the 1971 four year college nominal \$1410 from your link; the actual 2017 tuition was \$20,100, so obviously pretty constant in terms of compensation per hour

BTW, I’m a physicist, not an economist (thus the relativity lesson at top)

I could elaborate the economics that go with this at moderate length if you are interested
(and also provide plots of prices of various things from 1929 to present or 1910 to present, corrected for my “real” wage inflation )

• nkurz says:

> I could elaborate the economics that go with this at moderate length if you are interested
(and also provide plots of prices of various things from 1929 to present or 1910 to present, corrected for my “real” wage inflation )

I’m not Scott, but if you think you have what you think is a convincing argument with plots that mainstream economists are drastically underestimating the true rate of inflation, please post it! I think it would generate lots of discussion in an appropriate Off-Topic thread.

• Charles Engelke says:

Alright friend Nkurz,

I’ll try to offer a convincing argument that mainstream inflation adjustment must be drastically off

Two examples:
(1) How about the fact that CPI gives the equivalent modern food prices for 1930’s cities as equal to, or lower, than the price we would pay today.
And yet, 20% of the population was needed to produce that food, versus only 1% now (didn’t farmworkers then need to be paid?
he gas mileage of the tractors then was much worse than modern ones, so their fuel costs higher (and where horses were still used higher still).
Relatively profligate fuel consumption also disadvantages the 1930s food distribution system, which in every dimension was generally slower and less efficient than ours. Food refrigeration on the way to the city and in stores was, at best, primitive and spotty: much spoilage must have occurred. Yet the food is cheap according to CPI (though choices meager)
If the food was cheap, why didn’t people consume more calories and meat than has been recorded?
If there were shortages shouldn’t that have driven the prices up relatively high?
This all violates common sense!
But, the rejoinder must doubtless come:
The prices were low in CPI “2019 real dollars”, yes, but people were poorer as well by CPI so couldn’t buy it.
So what’s so “real” about this money? How is it supposed to relate to our present meaning of money?
Doesn’t this just sound like it should all be rescaled by some factor to give numbers that we would derive today in similar circumstances?
By what metaphysics do the economists know that this is the correct number to attach to money because its 1930. We couldn’t use all those old trucks and farm methods and had such prices in 2019!

(2) The average price of a new car in 1969 is found by CPI to be \$23,000 while the current average price is \$27,000.

However, a car in 1969 (full disclosure, my first car was a used 1969 Catalina convertible) was more massive: 4,200 lbs vs 2,800 lbs ( so more material cost) had more elaborate chrome fittings and decorations, more parts, more options, a larger more complicated engine, was built by unionized workers with higher pay and benefits with fewer robots and foreign completion to drive down prices.
So tell me again why the resulting vehicle is supposed to be even a little cheaper than today’s?
Or
If I could show my Excel plot using my patented “equal value for equal work” inflation adjustment on the data for average new car price from 1968 to present you’d see instead a gradual decrease in price (yeah! Progress does exist) from \$55,000 to \$27,000 now. That could explain why people own more cars per person now and use fewer car loans

So, why should I think that the conventional inflation adjustment is anywhere close to correct? By what magic can they prove that?

(BTW, Ford got his model-T down to \$30,000 2018 equivalence before the model-A boosted up to \$35,000 or something, so there really were cars cheaper than the 1968 ones further back in history)

PS
what’s an Off-Topic thread?

• nkurz says:

> PS what’s an Off-Topic thread?

Apparently it’s my personal and incorrect terminology. I meant that you should probably post in the most recent semi-hidden “Open Thread” as listed in the Recent Posts list in the top left of this page, currently 129.75. This post has been up for a while, and it’s possible that a post here will not be seen by too many people. A post in a more lively Open Thread will probably get more attention.

As for your actual argument, yeah, that makes sense to me! I’m not sure though how it meshes with your earlier comment that inflation is actually “nearly twice as big in my ‘real terms’ as economists typically assume”. If food and cars were the same relative price then and now, wouldn’t that make inflation between then and now just the difference in wages? Thus if they were relatively more expensive then (as your argument suggests) wouldn’t this make “real” inflation lower than economists claim rather than higher?

But I’m not really economically literate. Hopefully you’ll get a more learned critique from others in an Open Thread.

• The Nybbler says:

(2) The average price of a new car in 1969 is found by CPI to be \$23,000 while the current average price is \$27,000.

This does not appear to be the case.

https://fred.stlouisfed.org/graph/?g=ob89

Data for food prices in 1930 I do not have.

77. One point on the idea that some fields, such as education, can’t make much use of technology.

When I started my academic career almost fifty years ago universities had lots of secretaries, a large part of whose job was typing and retyping the professors’ papers, exams, etc. Currently, that job has almost entirely vanished, because professors can type and use word processors. But there still seem, by casual observation, to be about as many secretaries. That suggests that what may be happening is not that technology doesn’t increase productivity in such a field but that something about the incentives results in neutralizing the effect—what I don’t know.

• Matt M says:

I had a similar thought but with a different conclusion.

Scott rules out “administrative bloat” as a driver of cost by stating that total administrative expenses, as a share of all college expenses, are about constant. But there are two things we know:

1. Technology has produced productivity improvements that should result in dramatically decreasing costs for administrators specifically (typists being one category, accountants being another, etc.) But it hasn’t. Why not? Potential answer – for every “lost” administrator due to productivity improvement, the colleges used that money to “gain” an administrator in another area (like say, the diversity bureaucracy).

2. Give that costs are rising massively, the question of “where is the money actually going?” seems relevant. If administrative costs are basically constant and if professor salaries are basically constant… well, someone is taking this money. Is it just piling up in the endowment, or what?

• nkurz says:

> 2. Give that costs are rising massively, the question of “where is the money actually going?” seems relevant. If administrative costs are basically constant and if professor salaries are basically constant… well, someone is taking this money. Is it just piling up in the endowment, or what?

I think you are assuming that the number of professors is constant. I haven’t found figures I can point to, but I was told in a recent conversation that for Williams College, the cost of administration has remained almost constant as percentage over the last 30 years, but the student to faculty ratio almost dropped in half. In 1990, there was one faculty member for each 14 students. Today, there is one faculty member for every 7 students (and I’ve seen some sources claiming it’s as low as 6).

These exact numbers may be a little misleading, as the definition of “full time faculty” appears to have changed somewhat over this time period. Still, I think this is probably the right place to be be looking: while per-faculty costs have not risen, faculty cost per student is probably significantly higher. And by extension, if the faculty to administration ratio is constant, this means that administrative cost per student is probably a lot higher as well.

78. Flavoryeagle says:

As I understand it you made 3 disputes:

1. Real wages haven’t gone up so what hocus pocus is causing a price war for talented workers to drive up wages in all industries while leaving real wages unchanged.

2. Professor and physician wages don’t appear to be skyrocketing in any public data other than the specific source they chose

3. Back of envelope calculations demonstrate college is most certainly expensive nowadays relative to current wages.

1. As many have noted, the “correction” for inflation is where thinking gets very muddy when we are attacking this problem. Inflation necessarily includes Baumol effect for many industries, since healthcare and education will be included in the CPI. I attempted to find a place that simply broke out CPI by industries but the BLS breaks it into about 5000 groups which I did not have the time to decompose inflation by each factor. The best thing I could find was Figure 1 in WATPSDH (“Why are the Prices So D*mn High”). This most certainly doesn’t tell you all the contributions and education and healthcare might be massively under weighted in CPI calculations.

For the sake of argument, assume CPI is the best we can do and Figure 1 is a good representation of the decomposition. Wages staying in line with inflation implies that you are spending an ever greater portion of your lifetime income on education and healthcare and an ever decreasing portion on microwaves and dishwashers. Thus, you would expect a 1971 worker to work 24 months (arbitrary number) and 5 months goes to college tuition, 9 months goes to appliances and 10 months goes to car payments, the 2019 worker spends 14 months on tuition, 3 months on appliances and 7 months on car payments. They are both in the same spot.

The question of whether the world is in a better place due to this is hard to answer. Costs for education and healthcare certainly are higher. People are using more healthcare and education than ever before. Those are 2 clear cut facts. However, we can attempt to use consumer behavior to guess whether people are better off on average. If this increase in cost leads to more consumption (!) either 1971 people really hated education and loved goods and 2019 people love education and hate materialism (ie preferences drastically changes) or, as T&H argue, people are maximizing their utility and buying more healthcare and education with the plunge in cost of goods. That argument might be circular but I can’t find where I can disentangle it.

Another thing to think about is college graduation is not a median good. The current state of the world is probably an improvement for those that can easily afford college and it is hopefully an improvement for those that do not purchase college but, there will most certainly be a middle zone that it is unclear, hopefully government support and financial aid makes it positive across all populations.

2. The only real reason I posted this: I found a Dartmouth 1998 factbook with comparable college professor wages by rank. (https://www.dartmouth.edu/~oir/data-reporting/factbook/1998_factbook.pdf on pages 81-83) Using the factbook in addition to chronicle.com (https://data.chronicle.com/182670/Dartmouth-College/faculty-salaries/) I estimated that these colleges had about 3.75 – 4.25 CAGR on professorial wages from 1989 -2017. ie Professor’s nominal salaries tripled arithmetically in that time period. Full professors at Dartmouth were paid 61k in 1989 and 188k in 2017.

You can also look at tuition for 1989 and it was 13k and in 2017 it was 51k. That outpaces professor’s wages (which would imply 40k tuition) but not by a lot. I am guessing there are many reasons that is not a 1-1 correlation in the additional tuition but I would honestly believe a large portion is the fact that those are sticker prices and not the actual costs each student is paying. College’s have perfect price discrimination and “tuition” is just .. well you know all that.

Let’s assume professors’ wages are a good proxy for the population of similarly educated, high ability workers. They probably pay full tuition, but they also probably have good healthcare that is not included in the wage increase.

Who would you rather be professor in 1989 getting paid 61k + whatever benefits they had or professor in 2017 getting paid 188k + current benefits?

Personally, I am willing to pay the extra 10k in tuition for myself and maybe a higher deductible on healthcare for the remainder to be spent on current goods.

Even the assistant professor wages (36k to 96k) look like a very large improvement.

3. I don’t think this is at odds with the Baumol effect. Refer to above.

Since this is a random comment on a slightly old post and it will get very few reads, I am pretty hand wavy above. If anyone presses me for details I can try to improve this.

79. Charles Engelke says:

The overall “Productivity” increase that provides a rising standard of living is not about growing some NEW pool of money to distribute, it’s about getting more and more tangible stuff from less and less of our total potential work budget. This is clear when one looks through my filter: take mean compensation per hour as the standard by which to deflate nominal prices across time periods,

(This is not because work per se creates the value, instead, how long and hard one will work to get something indicates the current human valuation of that thing, whatever the period and context. At root, the’ cost’ of our self-generated effort is all we intrinsically possess with which to ‘pay’ for things, and naturally we don’t take that trouble unless the value received outweighs it.)

Look at agriculture: In 1930, 123 million people were fed by 30 million farmers (20% of US population) and agriculture composed ~20% of GDP. Today 330 million people are fed by less than 3 million primary agricultural workers (< 1% of population) with a much greater quantity & variety of food per person. However, these fewer farmers do not now get to split up 20% of GDP: agriculture also now represents less than 1% of GDP.

Why did the agricultural ‘wage’ decouple from productivity but stay coupled to the number of people involved? Because money is a marker for work value invested and can be traded for the tangible results of someone else’s comparable work effort invested. The way in which agricultural productivity lifts all ships is that food is now ubiquitous and super cheap in real terms (unless one insanely defines inflation to keep ‘real’ food prices constant across time, which is what the current deflators seem effectively to accomplish: i.e. defining a 1930 dime begged for a cup of coffee as equivalent to \$1.50 in real 2018 dollars, when in terms of time worked to earn the dime, it is \$10 such).

In parallel with the story of agricultural productivity above, US manufacturing has been generating around 3 metric tons/capita of physical output since 1947, yet the manufacturing work force has shrunk from near 30% then, to 8% now, meaning almost 4 times as much mass of physical stuff is now produced for the same number of work hours (presumably in the form of more numerous, lighter, more capable & sophisticated, devices). Comparable manufactured items are therefore now ~1/4 the 1947 price in terms of how long a consumer has to work to earn enough to buy them.

Supporting this expectation, the US chamber of commerce reports that from 1970 to 2010 manufacturing’s share of US GDP fell from 24% to 13%, (about halving) while the price of goods fell 52% relative to services (i.e., also halving, as I’d expect (USchamber.com/blog/2012/03/manufacturing).

That ‘relative to services’ used in the chamber of commerce quote is a proxy of my standard for judging inflation: if a ‘service’ always takes one person’s direct work to accomplish, while a manufactured object now takes a quarter as many people to generate, the one whose ‘real’ price has changed in the market place is the manufactured one.

The 'real' money that went to the more numerous workers in these sectors in the past now stays in the consumer’s hands, to be spent on new inducements, TVs, travel, books, lessons, music, the newest devices..

80. Charles Engelke says:

I’m new to this,
post it where?
excel plots don’t appear in these comment boxes

I don’t think there is any fundamental way to measure inflation except mine

”””””””””””””””””
The real data is the nominal data, the more economists put in different hidden ways of ‘adjusting’ or ‘correcting’ to ‘real’ current dollars before graphing trends, the more muddled things become, and the harder it is to reconstruct what objective conditions existed at a given time relative to another. It’s also hard to use earlier published results because every year of publication needs another scale; an old article might show history in ‘1978 dollars’ and have even used a different method to correct 1930 prices relative to 1978 than we use now (decades ago I remember the conversion being a factor of 20-25, while today’s CPI deflator it is 14, even though inflation should have made it even larger).

They are always changing the ‘basket of goods and services’ used, perhaps because everything about quality and quantity and relative need of the component goods and services change with time. But there is no way to quantify that (except mine, a measure of how dearly humans at the time value them), nevertheless they make up some official fudge or other.

Imagine if we used the price of a desktop computer to estimate inflation since 1973 (about zero change in nominal price range, while nominal pay per hour increased by factor of 11) you’d find mean wages today up 1000%. If you used the price of computing power as the standard, you’d find salaries up by a factor of 100 million. That would be a wage / productivity coupling to celebrate as it raises all incomes uniformly across quintiles!

The point? Economists are always comparing apples to oranges when they correct ‘value’ across time using items produced in an evolving economy. Food prices are a good comparison to make if you want to judge relatively how easy it is to feed a family at some point in time, but the real cost of food in a city a century ago had to be very different when half the country’s workforce was needed to produce it, transportation and preservation of goods were incredibly difficult prospects, and each item other than bread was in relatively lower supply.

They might think production of commodities like steel and useful industrial metals is a good indicator to use as a standard across time, or gold, but nothing about these substances in themselves carries a fixed value. NASA is sending a probe to the 200 mile wide metallic asteroid Psyche, a possible target of future mining. The principle investigator mentions that the iron content alone would be valued at ten thousand quadrillion dollars or over a billion dollars per human on earth, I estimate the gold would be 100,000 times as much as has ever been mined ‘worth’ 40,000 times US GDP.

So what happens to those valuations when the metals are brought to earth. How will you judge how big the change is (except by market price and that will be constrained only by how many people and how much energy are involved in mining and transporting them to earth.

Meanwhile the materials become more useful, not less, gold nanoparticles by the ton for new technologies, etc. How much does that raise prices? We’ll only know when we see the demand and how hard people will work to buy those products.

We have air and water almost free now, despite them being the substances most important to our existence
it could be a similar situation with metals in the future.
Moon colonies will have to mine and refine all their water and air.
How much will they value them?, Can you calculate the value from their intrinsic worth? You have to see what the market price is in my universal currency: work and time that people are willing to exchange for them

• Aapje says:

They are always changing the ‘basket of goods and services’ used,

Which IMO causes a structural undercounting of inflation, because it results in upward trends being removed from the data and downward trends to be included.

For example, imagine a situation where domestic help starts relatively cheap, but goes up in cost over time and washing machines are very expensive, but become cheaper over time. Then over time, more and more people will start buying washing machines instead of having domestic help & the basket of goods and services is going to be adjusted to weigh the price increases of domestic help less and those of washing machines more. So the price increases of domestic help is going to be counted less, as more people are being priced out, while the price decreases of washing machines is going to be counted more, as more people are priced in.

Yet this effect is equally going happen when people would prefer domestic help, but are priced out of having it vs them preferring washing machines. By changing the basket, the entire change to behavior is assumed to be motivated by desire, rather than people being priced out/in, so it will count changes in behavior that people are forced into due to financial reasons as changes in desire.

Of course, not changing the basket would have the opposite problem, where changes in behavior are assumed to be forced by financial reasons, rather than changed desires.

The most accurate might be to calculate both and then use these as the upper and lower bounds of actual inflation, but this would result in a huge range, probably.

• acymetric says:

They’re suggesting you post in in the “open thread” (a new one opens up roughly every few days). This is a general thread for discussing everything under the sun, more or less. You can see them in the left bar under “Recent Posts” titled (currently) “Open Thread 129.75”. If you go to Archives from the main menu bar you can also see them listed there.

Here is a link to the current one.

81. Charles Engelke says:

So, you’re saying that I could go there because this is one is no longer live,
but either way I have to put my plots online somewhere (make a google document?) and then provide a link for people to see them?

Presumably I could do that here, but no one is looking?

I really wanted to get the ideas to Scott, does he go back over these comments before closing them?
It is because his references to “cost disease” a few years ago stuck in my mind that I eventually came to this conviction that the cause was something deeper than job market incentives viewed in the conventional framework.
Instead it seemed to signal severe mistakes of practice at the foundation of economic measurement and the interpretation of money value and economic “growth.”

As a physicist who has worked

82. Charles Engelke says:

The document linked has some graphs showing results of my inflation adjustment of historical prices on the basis of “equal value for equal average work” across time, but not yet explanations.
https://docs.google.com/document/d/16rNp087cIPbiG6TlT5Lo4Xek-h8d-FV5ZtlVtGaahco/edit
I will not put it in the open thread until I have added much more to it, but I’ve already mentioned some of this data in earlier comments, so want it available.

HOWEVER, I realized that what I really have neglected to communicate clearly and directly is the error in the ‘Baumol Effect’ argument which Scott repeated at the beginning.

In the real world, a rise in productivity by a factor of 5 does NOT in fact lead to making 5 times as much money by selling the more numerous products at the same price and paying the worker 5 times more.

Instead, such an improvement in productivity, in the long run, always leads to selling items at close to 1/5 the original price while paying the worker about the same he has always received for his grade and difficulty of the job.

And if a service job involves a similar grade of work, it gets paid at that rate as well.
True, the market demand expands with the lower price, so more money is made that way by the company selling to an enlarged customer pool. The plant owner can pay the workers more from this profit for a while, but if it’s a salary much above the going price for the skill level involved, a competitor can hire similarly skilled workers at their usual rate and undercut the more generous owner/managers.
Basically a barber who switches to an assembly-line cutting wires instead of hair will earn at the same rate he always has. There was never any need for people to pay barbers more to keep them from running off to make their fortune. But it was a windfall for car factory owners when they could replace skilled master mechanics, rare and expensive and slow, with common diligent barbers and cab drivers who could do one small job repeatedly at near their usual rate of pay.

I’m amazed economists don’t seem to know that this is what happens. They inhabit a very different world from business people. When I first met my wife, three decades ago, she was working as an intern at a big management consulting firm, and every strategy revolved around “the experience curve.”
They had discovered that as predictably as a law of nature, every time a new product line or process was started up the cost per unit would decrease by some amount, maybe 20%, every time the number of units produced from the last price mark doubled, a self-similar power law.

So they would determine this coefficient in a particular industry’s case and could say, if we keep manufacturing at a high rate, in a month the price will be this much lower, so let’s set the price now at less than half what it is costing to make now, but the demand will get us through the doubling quickly and we’ll fall below that mark and be making a profit.
They’d then keep the price set there as long as the techniques developed during the continuous improvement phase are able to remain proprietary. But when competitors learn to do it and try to catch up, then the strategy is to keep lowering the price you sell at in proportion to the cost of production as it declines.
So continuous improvement = continuous price reduction.

That is the dynamic that positive feedback of the market place drives. The negative feedback classical economists concentrate on has nothing to do with productivity increase. If you don’t lower price, you don’t increase market share or make more profit, so you might as well not have bothered to manufacture more per person as , people rich enough to buy at that price are probably already doing so.

The workers might indeed be paid better as the profits come in from lower production cost, a lot of the improvements were skills and knowledge built-in to the worker expertise, that raises their work grade and necessitates protecting secrets. But it never rises in proportion to the ratio of product sent out per employee as Baumol assumes