In the long run we are all sick

US health costs are higher than others, but their growth rate is not.

This chart from the Kaiser Family Foundation, using OECD data, has been going the rounds of the blogs. You have probably seen it already:
Growth in Total Health Expenditure Per Capita
U.S. and Selected Countries, 1970-2008

Exhibit 3 Growth in Total Health Expenditure Per Capita, U.S. and Selected Countries, 1970-2008

A comment from Kevin Drum:

The United States, of course, has the highest spending, but it also has the highest growth rate.

The first part of that statement is true – it stands out a mile, or more precisely by $2,535 per person per year. This is the difference between the latest OECD numbers for annual real health spending per head in the USA and the next highest OECD country, Norway. A small, unrepresentative country with a population of 5 million scattered over 149,000 square miles of fjiords and mountains is unable to get economies of scale; as a super-rich petrostate it can afford a gold-plated universal system. And it still only spends 2/3 of the US level. This is outrageously high for the results it buys – 16% of GDP, a good 5% higher than the OECD peloton. The projections of future public health spending on Medicare and Medicaid are the only honest part of the deficit scare, but the overall cost picture is even worse as the public programmes are relatively more efficient.

So you’d expect the USA to be an outlier on the growth rate as well. Is it really? You can’t read that off from a non-logarithmic chart, and the overall inflation obscures what happened near the origin.

So I played around with the same OECD dataset to extract the growth rates by decade. Spreadsheet with source data and working here. [Update: corrected version, hopefully without bugs pointed out by commenters.] The results will surprise you.


[Update: corrected version thanks to commenters. Please do not reproduce the previous version]

Kevin is wrong on growth rates. In the 2000s, US costs have been rising slower than in Canada and Britain. US health spending has never grown as fast as it did in France, Japan and Switzerland in the 1960s. The big gap in absolute levels today results from a similar recent growth rate applied to an absolutely higher starting level. The real gap opened up in the 1980s.

Three things do stand out from the chart:
* long-term growth rates have converged (the range between highest and lowest has dropped in zigs and zags from 15% to 2.3%);
* growth rates have generally declined;
* growth rates are still much higher for health spending than overall GDP – about twice -, so the share of health in output continues to rise inexorably.

The obvious explanaton of these facts is that the growth rate is primarily determined by worldwide changes in medical practice and technology, not national differences in structure, which drive the wide differences in absolute levels.

* The convergence is economically explained by the globalisation of science through information technology and the Internet, the adoption of English as the scientific lingua franca, and freer movement of researchers and doctors. The hypothesis is testable: it predicts that the inter-country lags in the adoption of new medical techniques and equipment have gone down over time; an RBC-recommended research project for somebody.
* The decline is economically explained by the slowing-down in medical innovation, for instance in genuinely new prescription drugs coming to market.
* The stubbornly higher growth rate of medical costs than GDP is plausibly explained to a considerable extent by the cost-enhancing content of innovation. Some innovations, like the measles vaccine and keyhole surgery, have presumably saved money; others, like retrovirals for HIV, chemotherapy, and transplants have raised them. The latter have outweighed the former, so the net effect is up.

I’m much less sure about the force of the third hypothesis than the first two. Other forces are probably at work. Demographic ageing for one: but this is much higher in Japan then the USA, and the rate of growth in costs is lower. The use of data mining by insurance companies to discriminate against unprofitable high-risk clients is an innovation (a welfare-destroying, negative-sum one) limited to the USA, and cannot explain a wider trend. So pending More Research I can’t be fagged to do or seek out, I’ll assume the third factor is significant though not exclusive.

The fact that the US level is so much higher still has consequences. The boringly normal USA will still hit a health care spending crisis first. This just can’t grow faster than GDP indefinitely, or there will be nothing left for bread, circuses and the legions. Bending the cost curve sooner or later isn’t a policy option, it’s a logical necessity, and not only for the USA.

The USA does have one option that isn’t open to France, Britain, Japan or Canada: to copy one of their systems and bring absolute costs down to theirs. This would buy time, perhaps a decade. Clearly, the level of resistance to any such move would be enormous. It’s a safe bet that the existing patchwork, plus Obamacare, will provide US health care for a while. Obamacare, if its allowed to, includes lots of widgets to try to reduce costs, some of which will probably work. So it buys a little time, if less than single-payer would have done.

In the long run then all our economies are sick. If I’m right, the current course of medical innovation will drive us all into economic intensive care, so We Must Do Something.

The Something for the GOP, the Ryan Medicare plan, is to institute market rationing by income. The latest medicine will be limited to an ever-narrowing circle of rich families. The rest of us will buy the second-best we can afford, like the Nairobi poor today.

The socialist alternative – the GOP’s false vision of the British NHS today, and the reality of organ transplants – is administrative rationing of high-cost medical resources by need or desert.

Don’t like either option? Then we’ve got to bend the direction of medical innovation away from expensive treatments and towards inexpensive prevention and cure.

l think I’ve earned here the right to speculate, which I’ll abuse in a follow-up post.
Update: it’s here.

Author: James Wimberley

James Wimberley (b. 1946, an Englishman raised in the Channel Islands. three adult children) is a former career international bureaucrat with the Council of Europe in Strasbourg. His main achievements there were the Lisbon Convention on recognition of qualifications and the Kosovo law on school education. He retired in 2006 to a little white house in Andalucia, His first wife Patricia Morris died in 2009 after a long illness. He remarried in 2011. to the former Brazilian TV actress Lu Mendonça. The cat overlords are now three. I suppose I've been invited to join real scholars on the list because my skills, acquired in a decade of technical assistance work in eastern Europe, include being able to ask faux-naïf questions like the exotic Persians and Chinese of eighteenth-century philosophical fiction. So I'm quite comfortable in the role of country-cousin blogger with a European perspective. The other specialised skill I learnt was making toasts with a moral in the course of drunken Caucasian banquets. I'm open to expenses-paid offers to retell Noah the great Armenian and Columbus, the orange, and university reform in Georgia. James Wimberley's occasional publications on the web

17 thoughts on “In the long run we are all sick”

  1. I used the data from the Kaiser study that I linked to. It’s Exhibit 4a, and it extends from 1970. Discounting Spain, which was a very low income country in 1970, only Belgium and Norway have had higher heatlhcare growth rates than the United States. Thus my caveat: “There are individual countries that have had higher growth rates than us, but all of them started from a much lower base.”

  2. “direction of medical innovation towards inexpensive prevention and cure.” Unfortunately, past preventive intervention is generally not cost saving. That’s not to say that such actions are not cost-effective, just that they generally increase total costs.

    See, e.g., http://www.cbo.gov/ftpdocs/104xx/doc10492/08-07-Prevention.pdf and the citations within.

    For cost saving, my semi-informed instinct is to focus on cost-effectiveness using currently available interventions.

  3. Good work, James. Thanks for the effort that you put into this. I have just one (not so small) disagreement:…
    (James): “The latest medicine will be limited to an ever-narrowing circle of rich families. The rest of us will buy the second-best we can afford, like the Nairobi poor today.
    Not like the Nairobi poor. Substitute “automobiles” for “medicine” and “Ford, Kia, Nissan, and Volkswagen buyers” for “Nairobi poor”. We all benefit when “the rich” subsidize innovation. All of us except for those to whom relative inequality (i.e., envy) counts for more than absolute improvement.

  4. I did download the spreadsheet, but have trouble following the analysis (since the actual growth rate calculation isn’t active in the downloaded version.) Comparing a few countries: USA in 1960 is $148. Switzerland in 1960 $166, slightly higher. The chart above and the percentage growth rates in the spreadsheet show Switzerland having higher growth (averaged over all decades) than the US, but of course the final rates are $4627 for Switzerland and $7538 for US, so something is wrong. Similarly, Canada starts out at $123 in 1968 (slighty below US), chart and percentages show very similar growth rates averaged over all the decades, though Canada is a bit more front-loaded , but Canada ends up substantially cheaper ($4079) by 2008. One can also just look at the ratios of 1960 to 2008 spending as a measure of average growth – the US (x51) is worse than Canada, Switzerland, and the UK, tied with France. For growth from 1980 to 2008 the US is the worst though the UK is close.

    Some of the “check” values in the analysis tab for 1970 look off, so that might be a clue as to where the bug is.

  5. Yeah, Bruce is right. The Switzerland 1970 number should be 7.56%, not 19.83%. The 2008 number is ok though. I don’t have time to check them all but the his conclusion depends on the last decade, as the errors don’t propagate. So it still stands.

  6. I’m unclear as to why I should object to a larger and larger fraction of GDP going to health care, as long as we’re getting actual health for the money. What WERE you planning on spending the money on? Video games? SUV’s? Health care, it seems to me, represents a way GDP can grow without greater environmental impact. Isn’t that a good thing?

    A year and a half ago, I was diagnosed with Lymphoma. A few very nasty months of chemo, (Which cost as much as a modest house…) cataract surgery to deal with the side effects, and I’m reasonably healthy again, and probably a normal life expectancy. A couple of decades ago, my doctor would have told me to make out my will. Now, granted, that advice would have been cheaper than the chemo, but am I to regard this as a negative change?

    No, I think there’s nothing objectionable about spending a greater and greater fraction of GDP on health care, as long as the remaining fraction isn’t absolutely declining… And maybe even if it does, to some extent.

    Now, inflation in tuition, THERE’S something worthy to complain about, because the vast increase in expense hasn’t been accompanied by turning our population into encyclopedic geniuses. But health care? We’re getting something for the money.

    Further, I think there’s reason to believe the curve is going to bend itself eventually, as our medical knowledge grows to the point where we can prevent more illnesses, instead of treating them once they show up. But, even if this doesn’t happen, I don’t see the problem.

  7. Bruce, Russell: thanks a lot. Mea culpa for not checking the spreadsheet properly. I have done now and the checksums all look OK. There wa another error for the USA.

    Bruce: I solved for the growth rates using the magic black box of OpenOffice’s goal seek tool. That’s why I used the same formula in the check cells so the working is transparent and the answer verified.

    Kevin: I’ll try to find out the reason for the apparent disparity with the Kaiser analysis, but this will take some time and I wanted to fix the straightforward bugs first.

    Malcolm: I fear you are missing the argument. Medical innovation, whether its paid for by the rich (how?), a levy on everybody insured or sick through pharmaceutical profits, government agencies like the NIH, or charities like Wellcome can only help the poor if its’s affordable to them. The data indicate that this is becoming less and les the case, and on current trends wil eventually cease to be so entirely. Where do I go wrong?

  8. In the long run all of our economy should be health care…
    That’s not an economic illness, that’s the evolution of humanity.
    We actually want tomorrow’s economy to be focused on replacing knees and hearts and every other aspect of healing and life-span growing.
    The business of the future is healing ourselves…

    Bread, circuses and the legions will never employ large amounts of us again.
    Nor will manufacturing…
    Nor will education…
    Nor will politics…

    Humanity is being superannuated by machines. Bit by bit so to (algorithmically) speak.
    The next big boom a-coming is the personal robot era…
    And after that boom flattens its curve and robots are everywhere and unemployment reaches even higher…
    What then?

    If you are not going to build a future economy on doctors, medical scientists, nurses, etc.; what are you going to build it on?
    Surely not war…
    No…
    The future of mankind is all about healing mankind.
    There is no other forward economy.

  9. Kevin again:
    The only relevant Kaiser chart is Exhibit 5. If it’s correct, your proposition that the USA is at the top in growth rates only holds for the whole 1970-2008 period. What matters for policy and prediction are the current growth rates, where the USA is no longer an outlier.
    However, I don’t believe this particular Kaiser chart. The annual growth rate for Switzerland 2000-2008 is given as 1.8%, which would be an amazing achievement. The raw OECD data are $3,221 PPP per head exppenditure in Switzerland in 2000 and $4,627 in 2008, giving a still creditable annual rate of 4.63% – still faster than GDP per capita.
    I’ll ask the Kaiser people if they stand by Exhibit 5.

  10. Brett: I’ll concede your thought experiment, in which current non-medical consumption stays flat and health care absorbs all growth to infinity, converging asymptotically on 100% of GDP. But this is incredibly unlikely absent Stalinist planning to achieve it. Health care expenditure, if it continues to grow faster than GDP, will cross the line of absorbing all growth and not stop.

    BTW, sincere congratulations on the success of your chemotherapy. I know from my first wife’s experience that it doesn’t always have your happy result. Chemotherapy is a typical contemporary half-way medical house: an expensive, partly effective, long-drawn-out, unpleasant, treatment. What we all hope for is a real cure for major diseases, as in this remarkable case.

  11. “Health care expenditure, if it continues to grow faster than GDP, will cross the line of absorbing all growth and not stop. “

    On the contrary, absent your Stalinist planing, it will NOT do that. While health care is a very valuable good, and it is to be expected that it will absorb more of GDP as incomes rise, it is not an infinitely valuable good, and at some point people will simply refuse to purchase more of it. Indeed, the current level of health care spending is already a function of such decisions. It’s not like we couldn’t, even now, find some way to spend more of our incomes on health care.

    I was delighted that my insurance covered the chemo, but my insurance covered it, with a substantial deductible, because of market decisions that had already been made. There are other things it doesn’t cover, for the same reason. Health care inflation is not some independent variable subject to no constraint. It’s the result of such constraints.

  12. James): “Medical innovation, whether its paid for by the rich (how?), a levy on everybody insured or sick through pharmaceutical profits, government agencies like the NIH, or charities like Wellcome can only help the poor if its’s affordable to them. The data indicate that this is becoming less and les the case, and on current trends wil eventually cease to be so entirely. Where do I go wrong?
    Consider the analogy that I offered: automobiles. They began as a rich man’s toy, and improved in performance and fell in cost until most American families own at least one. Consider color TV, air travel, film and digital cameras, even books. Competitive markets generated cost reductions and improved performance until ordinary people today are wealthier than Andrew Carnegie. I have access to high quality chamber music beyond all but the nobles of Vivaldi’s time. I can put my finger anywhere on a spinning globe and say “I want to be there in one month”, and, with a few exceptions (K2 summit, the vault at Fort Knox, etc.) I can get there, for a price within my bank balance. And I’m poor by US standards.

  13. (Koreyel): “If you are not going to build a future economy on doctors, medical scientists, nurses, etc.; what are you going to build it on?
    As always, competition for mates. If we are lucky: art, music, poetry, athletic talent, wit, compassion. Someone said: I study war so my son can study politics so his son can study engineering, so his son can study history, so his son can study poetry.” Something like that, anyway.

  14. I’d also note that it’s questionable to even call it health care “inflation”, in as much as “inflation” means an increase in the (currency denominated) cost of the SAME article, and our more expensive health care today is better health care.

    Going again by my own example, last year about $50,000 worth of chemo got me a 95$ cure rate for my Lymphoma. According to my oncologist, ten years ago the cure rate for that same lymphoma might have been 25-30%. And most of that $50,000 went for two drugs that weren’t available 10 years ago, and hugely boosted the cure rate. So, how is that inflation?

    Again I say, if you want to go after a real case of market distorting, genuine inflation, why don’t you look into college tuition? It’s not like the educations have gotten ten times better since I went to college, to justify the cost going up so much.

    Let me suggest a reason academics aren’t interested in inflation in college tuition: THEY ARE THE INFLATION. But doctors are somebody else…

  15. As for progress “bending the curve”, I note that several approaches to preventing Alzheimer’s with a vaccination are approaching clinical trial, at least one already IN trial, with HUGE potential to save money avoiding institutionalization of Alzheimer’s patients. So the natural curve bending is already starting.

Comments are closed.