Tuesday, May 28, 2019

Cost divergence

Source: Marginal Revolution
This lovely picture is from Why are the prices so D*mn High? by Eric Helland and Alex Tabarrok. (It's covered in Marginal Revolution: The Initial post,  Bloat does not explain the rising cost of education, and an upcoming summary on health care.)

Bottom line: objects got cheap, people got expensive. Technology, automation, globalization (thank you China), and quality improvement made goods cheaper. People, especially skilled people, got more expensive. All of which should make you feel good if you're a person and especially a skilled person.

The source of the relative rise in the cost of education and health care is less clear. Looking around at  a typical university,  school system, or hospital suggests massive bloat and inefficiency. Alex suggests  not:
I assumed that regulation, bloat and bureaucracy, monopoly power and the Baumol effect would each explain some of what is going on. After looking at this in depth, however, my conclusion is that it’s almost all Baumol effect. 

In case you need to look it up (Wikipedia),
Baumol's cost disease (or the Baumol effect) is the rise of salaries in jobs that have experienced no or low increase of labor productivity, in response to rising salaries in other jobs that have experienced higher labor productivity growth.... 
The rise of wages in jobs without productivity gains is from the requirement to compete for employees with jobs that have experienced gains and so can naturally pay higher salaries, just as classical economics predicts.... 
The classic example:
the same number of musicians is needed to play a Beethoven string quartet today as was needed in the 19th century; the productivity of classical music performance has not increased. On the other hand, the real wages of musicians (as in all other professions) have increased greatly since the 19th century.
The basic idea, if productivity of skilled people rises in one place, then you have to pay more to the skilled people who stay behind even in an industry like education that has had no change in technology. In education, where it's pretty clear that technology has not improved, they find that we use fewer teachers per student, and teachers are paid more. Most of the rise in the cost of health care, they claim, is greater doctor salaries. My personal observation of the massive growth of school bureaucracies, and astonishing and growing waste in our health system is denied by their data.

I think the question is still open, but theirs is a challenging view which is why I point to it in the blog. They close with a suggestion that education and medicine are ripe for a big technological disruption.

As they point out,  "The cost disease is not a disease but a blessing." It is the result of the tremendous increase in productivity of skilled people over the last few decades. Compared to stagnant periods of human history, a chance to fight about divvying up a bountiful pie is not the worst thing that could happen to us.

Two minor quibbles: We should not state the "cost disease" proposition wages have increased while productivity has not increased. In the Baumol proposition, and to a good approximation in the huge change in musician wages from 1826 to now, musicians are paid their marginal products.  The physical marginal product of string-quartet playing is the same, as (apparently) is the physical marginal product of teaching, but the value marginal product has risen, which is why wages rise.

When skilled people leave music playing for other fields, music playing contracts, and people are willing to pay more to hear music played. (All of this may or may not be true about music. I'm just using it to illustrate the theorem.) You can charge more for a string quartet performance, so the amount of (real) dollars produced by playing a string quartet rises in parallel with the productivity of potential musicians in other fields.

The pure "cost disease" phenomenon, then, should accompany a dramatic reduction in the physically unchanged technology. And indeed, a few horse shoers and craft blacksmiths still exist, and charge modern prices to wealthy horse owners.

But health care and education have rapidly expanded, not contracted. We need a different source for the fact that the value marginal product has risen while the physical marginal product has not done so.   Evidently, demand for health care and education has risen as well. We need both observations to account for the rising relative cost.


  1. What if the providers of higher and lower education are simply better rent-seekers than others?

    Not to suggest that rent-seeking has played a role in the increase of medical costs....

  2. What is the way in which we are measuring everyone’s wages going up? If labor and capital share of gdp are declining, how are wages going up everywhere in step with marginal productivity? Im clearly missing something obvious...

  3. In thinking about what part of the problem is Baumol and what part is bloat, we should consider that the cost disease may induce bloat.

    There could happen for two reasons. First, as wages are pushed up in industries with stagnant productivity and skilled workers, we should expect non-pecuniary and tax free compensation to go up too. This might manifest itself in nicer offices, more administrative assistants and reduced workloads. Second, as the skilled workers receive higher wages, there is more value to increasing their productivity in any way possible—professors, for example, could be given fancier computers and musicians provided more skilled conductors.

    Again, I’m suggesting that the cost disease may play in important role but that it might lead to more of what looks like bloat.

  4. In the K-12 sector, the US average salary (in constant dollars) went from $55,411 in 1969 to $58,950 in 2016-17. That's an increase of 6.4%. (During that time period (judging from the graph above), the cost of "lower education" roughly tripled.)

    I find it puzzling that Helland and Tabarrok don't directly grapple with these sorts of statistics on salaries over time. (The education graphs are only cited to "NCES," which is like citing "Census" as a source . . . about as untraceable as possible.)

    1. I don't have time to read the paper right now but two things occur to me.

      First, is your 6.4% increase top line salary or total with benefits? Benefits have gone up as a proportion of pay over the last 70 years. But this probably means K-12 compensation increased 20%-40%, still far less than the tripling in the graph.

      The other question is how they determined the Baumol increase. If you look at total salaries in K-12 (or higher ed or medicine) then I'm sure most of the increase is due to salaries. But the total number of people collecting salaries seems to have ballooned. Not in the classroom but administration -- far more central office staff. When my kids started school the local elementary school had a principal and at least one assistant principal. When I was in elementary school (and afterwards worked briefly at the school) the "office" consisted of one principal and one or two secretaries.

      I'd also be interested in seeing comparable numbers for K-12 principals and other administrators. At my local school district the lowest paid administrator makes more than the highest paid teacher. Seems quite a difference which I'm not sure is justified by "productivity" (also, I'd guess, pulls many good teachers into administration for the money).

  5. In terms of health care costs, since doctors and hospitals bill separately it should be possible to look at the change in doctor's rates vs hospital charges to determine the source of the increase.

    Of course, the majority of health care is outpatient and I'd guess the majority of the expense is in small charges, huge hospital bills just get the headlines. A very large proportion of medical costs are also for chronic conditions. One book (from around 2008) by a former head of Kaiser said that half of medical costs are for five diseases (depression, heart disease, congestive heart failure, asthma, and diabetes if I remember correctly).

  6. The quibbles are well taken. It pays to keep in mind that US health care spending is exceptional by international standards, and spending on education is probably so. They are outliers.

    In health care demand is heightened by the deductablilty of health benefits from taxable income, and supply is limited by the Liaison Committee on Medical Education. Such institutions exist nowhere else in the other rich countries.

    Lower education is unionized, and higher education is supplying the weapons for an arms race because alternative quality measurement methods have been outlawed.

    In short, the first commentor, Erik, is spot on for both industries: These guys are Class A rent-seekers!

  7. People I pay for directly haven't increased in cost as much as people I pay for indirectly through taxes or health insurance. Compare the rise in veterinarian costs to doctor costs.

  8. I have always hated the metaphor of the string quartet that the advocates of the Baumol explanation use.

    "the same number of musicians is needed to play a Beethoven string quartet today as was needed in the 19th century; the productivity of classical music performance has not increased."

    When I want to listen to music, as I am at the moment am typing this comment, I do not hire a group of musicians to come to my house and play. I use a program on my desktop to load some MP3 files and play them. Playing music costs me much depreciated pennies.

    In the 1950s, my grandfather loved to listen to recorded music. But his LPs cost $5 a piece ($45 in 2019 money). He played them on 'stereo" equipment that cost him hundreds of 1958 dollars. (His McIntosh tube amps would be worth thousands now).

    The fact is that technology has made music that was the exclusive privilege of the truly wealthy and powerful --Counts and Dukes, not dentists-- into a democratic possession and has driven its price close to zero.

    Now lets talk about education. Most of the content provided by colleges to their inmates is delivered in the form of lectures. Lectures? Why lectures? haven't they heard of video? Virtual Reality?

    And yes, I acknowledge that the proliferation of administrators is a real problem. But the colleges have not even begun to restructure their teaching methods to use 21st century technology. Lectures? Are they kidding?

    No, the pathology is institutional sclerosis and way too much easy Federal money. $1.5 trillion of students loans that are largely worthless and will never be paid at par. If the colleges had been forced to make loans from their own funds and then raised cash by selling the paper at market without Federal guarantees, they would be bankrupt.

    1. Well said. yes, it takes the same 2.66 total hours of labor today to play Bach's whatever as it did in 1801. But today it takes only 1 minute more to reproduce a billion copies, and it need never be actually played again.

  9. I don't understand why the Baumol effect is cited.

    The link between productivity and wages was severed many decades ago. Even a cursory examination of manufacturing wages and output will verify this.

    There is no shortage of wannabe college professors willing to work for beans until they can get that position. Lots of studies show a lot of bloat of administrative staff at universities. I understand a few people charged Harvard a ton of money to mismanage their financial portfolio.

  10. There is also the relative comparison. If X, who goes to the University of Chicago, gets a high salaried job, while Y, who goes to Podunk University, does not, the demand for spaces in the University of Chicago will rise. That will lead directly to the increase in tuition costs. If real wages overall start to rise, the rate of increase may drop a bit.

  11. "My personal observation of the massive growth of school bureaucracies, and astonishing and growing waste in our health system is denied by their data."

    No it's not... Helland & Tabarrok's data shows a MASSIVE increase in "bloat." What they are proving is that this increase in bloat is (almost) perfectly correlated with the an equally massive increase in direct costs.

    There are a few ways to think about this. One is to dismiss bloat as a concern and think about education/healthcare spending at a more macro level (i.e., what Helland & Tabarrok seem to advocate). Another would be to explore whether bloat levels relative to direct costs were already too high back in the 1950s, when the data series begins (if bloat was already too large at the beginning of the data series, facts about its relative growth rate are sort of irrelevant). Or we could ask whether bloat should decrease on a relative basis as total spending increases, similar to what has happened with "plant" costs in higher education.

  12. Helland and Tabarrok observe that the cost of educational services has continued to rise even as the outcome (mathematical aptitude) has remained essentially constant. In this case, the marginal product (mathematical aptitude) shows no sign of improvement even though the marginal revenue product of educators has increased at a linear rate with time.

    This would imply that the Baumol effect is likely increasing the rate of compensation of mathematically-trained graduates as they enter the workforce in such a way that the students are willing to pay educators higher salaries in return for training in the skills that mathematicians require to succeed following graduation, hence the increasing marginal revenue product of educators.

    Automation in industrial occupations as well as in financial markets and in insurance is probably contributing to the demand for increased mathematical skills and hence the value of educators. But has there been a noticeable decline in the quantity of instructors of higher mathematics at universities and colleges? Probably not, but the competition for mathematics specialists by industry is likely contributing to the higher educator salaries, along the lines that Dr. Cochrane describes above.

    Explaining continuously increasing salary/wage compensation rates for police officers and firemen is not so easily achieved. Unionization likely has had a major influence. One argument advanced by union negotiators, in the course of contract negotiations with municipal officials, is that if the city fathers in jurisdiction B don't pay the same rate as the city fathers in jurisdiction A where contract negotiations were recently settled at an annual increase of xx%, then police (firemen) of jurisdiction B will depart for jobs as police (firemen) in jurisdiction A. The union's argument is specious, but generally effective except during economic recessions when municipal revenues plunge and cut-backs in cost expenditures must be rigorously effected to avoid municipal insolvency or bankruptcy. Is Baumol's effect in play, or is some other economic agency involved?

  13. I am a doctor. My revenue, for the same work, has stayed the same for 15+ years. Friends of mine say the same. Surgeons say their revenue has decreased.
    If Baumol cost disease is correct, then the percent of healthcare costs for doctors should rise, compared to hospitals and pharmaceutical companies. Shouldn't it? Is that what is happening?

  14. Something doesn't compute: Really? Teacher salaries are skyrocketing? Adjunct professors are living high on the hog? Class sizes are down to 10? Maybe it's the administrations? Or maybe it's prestigious schools competing for students by offering outlandish amenities.

    The doctor costs can be easily addressed by training more doctors. Pull the sharpy kids away from the bloated finance sector.

  15. From what I have seen, in the US, health care is less efficient than in the past due to the loss of the independent doctor and the resulting loss of any competition in the field.
    An example; in the early 1970's, I had a small growth on my arm. It was removed by my family doctor, in his office, for the cost of a standard office visit.
    Recently, I developed a lump on my finger. The Primary physician did nothing except refer me to a specialist. The specialist looked at it, and proclaimed that it was a harmless cyst. Why couldn't the primary physician tell me that?
    The specialist also told me that it is extremely simple to remove if it is bothering me, but I would have to go to a surgery center in the hospital to get this done! All at insanely high prices.
    How did it get this way?
    From my limited view, It seems that the independent doctors do not exist anymore; they have been driven out of business by malpractice costs, government regulations, and insane insurance complexities. All health care is now through major hospital chains, which is some cities are a monopoly, like UPMC in Western PA, or Christiana Care in Northern DE. Even where there is a choice, you are deciding between different similar big hospital chains all with similar policies and costs.

    Let me give you a more general observation. I call it the Architecturally Significant Building Index.
    I spent several decades working in the printing industry, which is a very competitive business. I observed that whenever a successful printer built a beautiful new building, they would be out of business in about 2 years. They could not compete with the businesses that were in a modest, cheaper, but functional building.
    Look now at the businesses that are currently in expensive Architecturally Significant buildings. Hospitals and College Campuses. These are businesses that currently either have a monopoly, or high enough demand (with significant business entry barriers) that they can raise prices far above what they would be if that industry was competitive, allowing them extreme profits, which of course they deny.
    But the buildings don't lie. The businesses which have competition show up on the bottom of your cost chart, and are normally in cheap concrete block buildings. The industries on the top of your cost chart are not competitive, and that can be seen in their overpriced, fancy buildings.

    Ed Kennedy

  16. Mr. Cochrane: I have submitted a comment on this item twice, did you receive it?

  17. The graph, I assume, shows the cost to the consumer. Instructor/professor wages are internal costs that the consumer does not directly see. The growth rate in quality educational institutes cannot keep up with the natural population growth in demand for quality education, the main reason for the high rate of increase in education cost.

    1. Nonsense. This is not a "there are only so many rocks on the river to fish from situation".

      With time, it is relatively easy to fabricate more teachers/professors.

    2. https://nces.ed.gov/fastfacts/display.asp?id=84

  18. I read somewhere, I don't remember where, that the Baumol's Cost Disease should really be called Baumol's Wage Bonus, since it means productivity growth in one sector will benefit people in all sectors. It also makes clear that it's not a disease, since by definition, it can't cause the cost of anything to rise faster than peoples' ability to pay for it.

  19. Also, the original Baumol's Cost Disease case study had already been solved. By filming productions once, then showing them over and over again on theater screens or on television, the productivity of actors has been increased several orders of magnitude.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.