One big reason for this result is that doctors are taught that people should not be able to do such cost-benefit calculations, that anything with possible benefits should be done, that it is in fact unethical to take costs into account.

In our system, insurance pretty much pays whatever the doctor orders. It would be shocking if costs did not constantly increase.

Expand full comment
Aug 1, 2022·edited Aug 1, 2022

I understand your interest in the moral dyad but there's also some practical and political drivers that are probably even more important. The preferential tax treatment of health benefits held over from WWII (another bad economic decision by FDR's brain trust) has long created an incentive for US companies to offer expanded health benefits in place of real wage increases. I like to say what companies offer is better understood as 'prepaid health care' than true insurance. In that light, the only way to unlock the value of the benefit is to utilize health services even when they are of low to no benefit. The advent of HSAs has pushed us away from that model a little bit but they really should be made universal, and adapted to Medicare/Medicaid. What we actually got wasn't a transition to a more stable and viable system but a reform designed to make providing employer-paid health insurance as painful as possible, the better to Cloward-Piven the system so companies would dump the entire mess back on the government.

Expand full comment
Aug 1, 2022·edited Aug 1, 2022

Not sure what to make of it all, but there do seem to be alternate sources that offer a different perspective than that of many of your assertions. I don’t know what your copays and annual deductibles are, but I am pretty sure most Americans would kill for whatever plan you happen to have.

Kling: “In the U.S., we pay a smaller share out of pocket than just about any other industrialized nation, including Canada.”

World Health Organization: USA per capita out-of-pocket spending 2019 in US dollars - $1,235; Canada $753, France $416, Germany $697. (see: https://www.who.int/data/gho/data/indicators/indicator-details/GHO/out-of-pocket-expenditure-(oop)-per-capita-in-us )

Interestingly democratic Switzerland had a figure nearly double the USA, $2,445.


Kling: “The reason that we spend so much on health care is that we obtain a lot of procedures that have high costs and low benefits.”

JAMA: “Contrary to some explanations for high spending, social spending and health care utilization in the United States did not differ substantially from other high-income nations. Prices of labor and goods, including pharmaceuticals and devices, and administrative costs appeared to be the main drivers of the differences in spending.” (See: https://jamanetwork.com/journals/jama/article-abstract/2674671 )


Kling: “Next time you go for an MRI, try to calculate the probability that the result will affect the treatment plan and multiply by the likely benefit of the treatment plan. My guess is that the answer will be much less than the cost of the MRI.

Because we pay for health care collectively, as individuals we far over-consume medical services.”

Commonwealth Fund: USA has 110.8 MRI’s per 1,000 population, France 114.1, Japan 112.3 (see: https://www.commonwealthfund.org/international-health-policy-center/system-stats/mri-exams-per-1000 )

Interestingly, democratic Switzerland only has 74.1 per 1,000.

The Commonwealth Fund (see: https://www.commonwealthfund.org/sites/default/files/2021-08/Schneider_Mirror_Mirror_2021.pdf ) offers what appears to my plebeian eyes to be a more persuasive summary of the situation:

“The U.S. continues to outspend other nations on healthcare, devoting nearly twice as much of its GDP as the average OECD country. U.S. health spending reached nearly 17 percent of GDP in 2019, far above the 10 other countries compared in this report. Moreover, high U.S.

out-of-pocket health spending per person, the second highest in the OECD, makes it difficult for many Americansto access needed care.

The U.S. has managed to keep pace with or exceed other

countries on several measures of care process included

in the report, such as influenza vaccination rates for

older adults, lower rates of postoperative sepsis after

abdominal surgery, and more use of patient-facing health

information technology for provider communications and

prescription filling. But the U.S. still lags other nations on

measures of health care outcomes, access to care, equity,

and administrative efficiency. “

In terms of moral dyads, this notion of the little people use too much health care seems to be more of an extension of World Economic Forum moralizing about how the little people don’t need cars, heating, air conditioning, or to eat meat or drink alcohol, versus dynamist proponents of human flourishing.

Expand full comment

The MRI that has more cost than benefits is thinking in dollar terms, not utility terms. If an MRI means that a doctor can be confident in confirming a diagnosis, or excluding a potential disease, then that reduces uncertainty, which can be a positive in utility terms, even if it is more costly.

Expand full comment

One indication that health insurance arrangements, copays, moral hazard and other similar factors are not big cost drivers is that in affluent countries expenditures on pet medicine grow at the same or higher rate as expenditures on human medicine. This is true both in US and in Japan. Health insurance for pets is growing fast but the proportion of insured pets is still small, on the order of 5% in US.

Expand full comment

It is hard to do a rational analysis on medical cost when the "costs" are unknown. For example, I had a 4X bypass re-plumbing of my heart with the cost of almost $250K+ if I didn't have insurance. The actual amount paid was about $50K.

If I could have obtained the same price as the insurance company did, it would have been optimal to save the almost 10K per year I was paying for my insurance from age 40 to 64 (24 yrs). I had made the mistake of being diagnosed with a blocked LAD when I was 40 and my insurance went through the roof as Prudential cut my existing contract and put me in a high risk pool.

I was forced to by insurance (very high deductible so most years it covered nothing) to get the "discount" the insurance companies get from the medical providers. Saving $240K to cover my risk would make sense.

We need open fixed pricing if you ever want to do something about medical cost.

Expand full comment

I think doctors are in a better position to do cost benefit analysis than patients.

Expand full comment

Well, there ARE a lot of supply side constraints. We could be drawing in many more foreign doctors and residencing more medical students, allowing more use of medical practitioners to do things that do not require 7 years of training. But more fundamentally, DFA, Medicare, and individual doctors need to apply benefit cost analysis to their decisions. I don't think higher co-pays would help very much. If we shifted more people away from health insurance "provided" by their employers and toward individual plans financed with partial tax credits, THAT might help in the long run by living insurance companies to put pressure on health care suppliers to come up with more cost effective innovations.

Expand full comment

It seems two changes would have a big effect on health care efficiency and costs: (1) greatly diminish the regulation of health care insurance, and (2) require transparency on health care provider pricing.

On the first point, many people, if they had the option, would opt for real health insurance, i.e., insurance to protect against true financial catastrophe, similar to the way life and property insurance markets operate. Overall, it seems such policies would be much more affordable.

On the second point, people with real health insurance (or no health insurance) would be far more price-sensitive. Health care provider pricing transparency would thus permit something closer to a free market in medicine, with its consequent competition among providers, to develop.

Expand full comment
Aug 1, 2022·edited Aug 1, 2022

>>Next time you go for an MRI, try to calculate the probability that the result will affect the treatment plan and multiply by the likely benefit of the treatment plan. My guess is that the answer will be much less than the cost of the the MRI.

As John Hall has already pointed out, this is the wrong way to calculate the benefit of the test, because it excludes the value of the information itself. How long do you expect to live? Are you likely to become disabled soon? How much should you be saving for future health and long-term care costs? For most people, this sort of information is extremely valuable.

This is just a basic error in these Robin Hanson-esque arguments about the costs and benefits of health care, and I'm disappointed to see you making it. I don't believe there is anywhere near as much non-beneficial treatment as you suggest, when factors like the long-term planning and peace of mind of the patients are included in the calculus.

Expand full comment

"My guess is that the answer will be [the benefits will be] much less than the cost of the the MRI."

That's assuming the hospital actually knows what the costs of the MRI are. I recently had a procedure (not an MRI) that was billed at $4,600. The "negotiated price" that the hospital had with my insurance company was $491. When the actual price paid is 10% of the sticker price, something is way out of line. It seems to me that many hospital prices are not determined in any way similar to that how market prices are determined.

Expand full comment

It might be interesting to compare bundling of (a) catastrophic insurance and (b) insulation from costs of non-catastrophic events in two large sectors: (1) health care and (2) autos.

My intuition is that most people choose "insulation" when they buy auto insurance. They choose policies that have low deductibles and broad coverage of non-catastrophic events (glass, minor collisions, etc).

Does auto insurance exhibit anything like coverage of "high cost, low benefit" service, which greatly increases costs in health care?

Here is a website with various aggregate data about auto insurance premiums:


Arnold shows that *insulation from costs* and *high prevalence of high cost, low benefit services* go hand-in-hand in health insurance.

By contrast, auto insurance seems to exhibit substantial insulation, but relatively low prevalence of high cost, low benefit services. This is an empirical question, I'm reporting my impression.

Perhaps "moral dyad" psychology plays little role in the auto-insurance industry. Maybe most people believe that medical patients lack agency, but that car owners/drivers have agency because car owners/drivers usually aren't ill when they have to reckon with property damage and auto-insurance adjusters.

Expand full comment

There will be no healthcare reform. It will continue to eat larger and larger portions of the economy until a fiscal crisis causes a collapse.

Can you think of any sustainable healthcare system today? Is it in a large democratic western country?

Expand full comment

True for too many “services” influenced or provided by government—medical services but also military services, regulation, and debt?

Expand full comment
Aug 1, 2022·edited Aug 1, 2022

‘ As individuals, we would like unlimited access to medical services without having to pay for them.’

That’s because since WWII we have been told it is a ‘Right’, provision of healthcare for all the mark of a civilised society, we fear death and are encouraged to believe medical care can eliminate death, that better medical care means a healthier nation, and if only you vote for me I shall ensure you get all the medical care you want, if need be, paid for by somebody else.

By the end of six weeks, assuming adequate supply of fresh water, without any food 100% of the population would be dead. By the end of six weeks without any medical care nearly 100% of the population would be alive.

If only half of what we in developed Countries spend on medical care, were spent on resources to provide cheap food for people in poorer Countries, our health would be measurably better and so would theirs.

Expand full comment