The fundamental issue with health insurance is that it isn't insurance; if it was, then getting liver cancer on insurance would mean that the insurance company would "own" that liver cancer, regardless of whether I continued being on their plan or not; they'd pay out for the treatment as long as it was necessary / I continued to survive.

It's a mechanism for spreading cost amongst its pool, rather than a mechanism for spreading risk amongst its pool.

And the better an insurance company is able to distinguish between high-cost and low-cost individuals, the less well it will perform this function; if all the high-cost individuals are in a pool, then you haven't actually spread the cost around at all. But since all the low-cost individuals would prefer not to pay for high-cost individuals, even if an insurance company -wanted- to spread costs around, it could not actually do so; their customers would vote with their wallets to go with the plan that costs them, personally, less money.

Imagine, for a moment, that some ingenious individual at a property insurance company creates an oracular AI that can, within a 5% margin of error, calculate exactly how much money it would cost the insurance company to insure a specific house; if it will get hit with a hurricane next year, the AI knows this, and calculates the damage. This insurance company then tailors every policy to guarantee a margin of 5%. The AI spreads, and property insurance companies all begin using it.

Property insurance becomes impossible - nobody who would be taken as a customer should agree to be taken as a customer, because whatever their premium is, they should just put that in a savings account and cover the costs themselves, and save themselves the profit margin of the insurance company. You'd need regulations prohibiting property insurers from using this information, in order for property insurers to have a job to do - which would become spreading cost, instead of risk.

If our regulation was too heavy-handed, and forbade insurers from using -any- information - costs would skyrocket, because properties now regarded as uninsurable, because the risks are too high, become mandatorily insurable. People would buy properties they wouldn't otherwise buy; the insurance company, after all, has to pick up the tab.

So now you're in an uncomfortable position. Imagine yourself the regulator of the industry, seeing these problems. You decide you have to do something; so you decide that "Imminent collapse" can now be used as a reason not to insure a building. Costs go down.

Now you have regulators doing the job that the market is supposed to be doing - and doing exactly as bad a job as you'd expect, because, in truth - there is little to be seen of a free market here. The companies role-playing as participants in a free market will, in the end, have to compete mostly on the costs (of many kinds, including how difficult it is to actually get a claim filed) of their relative bureaucracies, which, granted, isn't nothing, but it's quite far removed from having much at all to do with property itself.

Meanwhile, the most important things, in terms of cost, all happen at the regulatory level. Whether or not you can move people who smoke in bed into their own risk pool. Whether or not insurance companies are allowed to force homeowners to pay for upgrades to their home to prevent electrical fires. Whether or not insurance companies are allowed to force homeowners to pay to replace their roof, or separate roofing out into risk pools. Odds are, the actual way it would work out is that insurance companies have to pay for all the incidental costs - they pay for your roof to be repaired, they pay to have your foundation shored up, they pay to repair your electrical system. And probably they have to pay for whatever the homeowner chooses to do on their own cognizance, as opposed to the insurer acting by fiat - and if it is the homeowner making the decision, costs go up, because, beyond whatever deductible we employ, the insurance company is footing the bill anyways, so might as well choose the roof that lasts twice as long and costs ten times as much, since they're already maxing out their deductible anyways, and that's the economically rational thing for the consumer to do.

And if you're not behaving that way, well, better not get property insurance, because you're paying for everybody else who does, and it'd be cheaper to just go without. Which drives costs higher still, so now the government, considering the catastrophe, mandates that everybody buy property insurance, to spread the costs around.

That's more or less where health insurance is.

Expand full comment

It seems like it would be pretty easy to mandate a pool for each combination of sex and birth year, and then to vary the UBI by combination of sex and birth year. Would such a change overcome your objection?

Expand full comment

Wouldn't that require the UBI for the elderly to be vastly higher than for the young to cover their insurance costs? If so, that doesn't really solve the problem; if not, it makes UBI a giant subsidy for insurers of elderly healthcare costs, which is pretty much the problem we have now with Medicare anyway.

Expand full comment

Possibly. It depends on how much of the premium the UBI attempted to cover. I have to admit that I don’t have a great sense of literally how much the average 80-year old costs vs the average 25-year old.

Expand full comment

I knew that answer once, probably nigh on 10 years ago... it was close to 1.5 orders of magnitude, although it didn't include things like birth control or some other stuff that wasn't emergency care, so the ratio might be on the lower end now as things have become a bit more like "my car insurance pays for oil changes" over time. Still, the difference was pretty staggering, although in retrospect it isn't too surprising. When I was in my 20's and early 30's I broke a bone or two and needed stitches, and past that it was yearly checkups and and antibiotics for ear infections. Mid 30's to 40's outside kids and getting my sinuses drilled out (all of which were uncovered largely), a few more stitches, and a skin cancer getting cut off. Then I think of my dad, who is in really good shape for a guy in his 70's, who has had a knee replaced, hip replaced, some cancer business, eye surgery, all within a few years. Throw in a few chronic problems, a handful of different prescriptions, or two handfuls, and yea, it gets super expensive super fast. The last 5 years of life especially wind up representing something like 40% of the total life time for most people (take that with a grain of salt, I am remembering that from a LOOONG way back.)

Anyway, yea, typically in group insurance plans the young heavily subsidize the old. Even having kids doesn't change it much, as kids don't have many chronic issues as well. My three kids have, somehow, managed to go to the hospital twice in total (not counting their very first.)

I think Robin Hanson wrote on this pretty extensively a while back, but for some reason I am not thinking he is the one who put together the stats on the difference in cost between age groups. Sorry, it is late and I am really tired :D

Expand full comment

The differences are huge. The health care for children and young adults is very inexpensive. After age 60 it mounts quickly.

I’d you tried to make up the difference with UBI, it would not be at all Universal.

Expand full comment

Generally it is cost prohibitive for private insurers to cover 91-year olds - it basically collapses to prepaid medical care. That's why we have Medicare. It is also cost prohibitive to cover many chronic and terminal conditions, for the same reasons. The purpose of insurance (private or otherwise) is to (1) subsidize the sick and subsidize on-demand emergency services, (2) insure against getting sick, (3) create a vehicle for pretax payment of routine/predictable medical care, (4) provide accountability on private billing, and (5) provide incentives for preventative care. #s2-5 are held out as benefits to soften the blow for the cost to healthy people. However, #2 is a bit of a mirage, because of annual underwriting and pre-existing condition mandates/non-mandates. #4 is really an embedded financial penalty for being uninsured, as providers are often required to upcharge those not on plans (though it can be difficult to collect). For health insurance to work (private or public), you need large pools and/or everyone to participate. That means your premiums are really a tax. As long as its optional, health insurance implodes, much like voluntary taxes.

Expand full comment


"Insurance" and "cross subsidies" are heavily mixed.

In part because "insurance" is hard to provide because of underwriting cost, and when the government offers to solve the underwriting problem its very tempting for it to introduce cross subsidies for political purposes.

Expand full comment

Better to underwrite every individual and let the premiums reflect that assessment. Then subsidize accordingly.

Expand full comment

Things like AGE are easy to adjust for via things like Risk Scores.

The difficult thing in underwriting is pre-existing health conditions. It's easy to verify someones age and model a likely cost that is accurate for that bucket. It's harder to model the cost for all possible health conditions.

How much does a "diabetic" cost? That's a huge range, and at a certain point trying to define that down to ever smaller categories just becomes cost prohibitive from an underwriting perspective. And that doesn't even get into the issue of individuals knowing more about their health status than insurers.

If you have a competitive private health insurance system they will indeed pay smart people to try and min/max whatever the current payout system is for every condition code. That's more of an argument for universal healthcare, you simply take "selection" off the table as a competitive advantage.

Expand full comment

1) Universal Healthcare along Murray-ish lines is already practiced somewhere. It's called Singapore. This Substack comment isn't the place to flush out the entire Singapore healthcare system, but it's an Actuaries Dream and it works very well.

I could offer other health systems that work along some similar lines to Singapore, its just the best of them.

2) What you're getting at is that *Underwriting* is a critical problem in the health insurance sphere.

Age isn't actually that hard because you don't have any selection/asymmetric knowledge issues.

Private insurance can't manage Underwriting because it can't handle pre-existing conditions (no, pre-existing condition insurance doesn't work and has failed many times).

You basically have two choices.

A) Accept that you can't offer pre-existing condition insurance (very unpopular and usually fails the democratic test).

B) Come up with a Singapore-esque technocratic solution involving universal government Underwriting.

The problem is that even though its POSSIBLE for a government to provide genuine value as universal underwriter, in practice a lot of large democracies (not all) use Underwriting as a backdoor way to introduce subsidy for favored groups since its so opaque.

There is also the issue of "judgementalism". There is a difference between "I got cancer and now I have a pre-existing condition" and "I got AIDS from fucking around in gay bathhouses and now I have a pre-existing condition." Places like Singapore are more BASED about calling it like it is on that stuff.

3) I agree that health insurance and old age retirements is where Murray struggles. UBI would require OLDS to take a cut over the status quo, which is why its such a hard sell.

Expand full comment

The Australian private health system broadly operates like this. Insurers must offer a set of standard policies, must accept all customers and must charge the same price to all customers.

They don't seem to try and push out the elderly.

There are some tax rules which encourage young people to take out coverage.

The model is starting to break down though with fewer young people taking coverage (and overpaying) shifting the pool and increasing costs which in turn deters more young people from taking coverage.

Expand full comment

I am not sure that I see the point of trying to force all ages to pay the same premium.

At least until age 60, income and health care costs covary. Young people make less money, and they pay lower premiums. As young people age their income increases as well as their health care costs.

Why not just say that an insurance company must charge everyone the same rate to persons of the same age?

It is simple and transparent, and it seems pretty fair. Plus it overcomes the problem that Caplan mentions.

Expand full comment

I generally agree with your criticism, but I disagree with your claim of "In Murray’s world, no insurer wants to be known as a geriatric specialist."

Imagine a world where every insurer is offering the same product, competing for the young customers. That opens market space for the "Geriatric Specialist Insurance". It just costs a lot more than the companies that compete for young customers. But it offers better quality coverage for heart disease, cancer, Alzheimer's, all that stuff. Sure, a young person *can* buy the geriatric specialist insurance. But why would they? It's not a good deal for them. If nobody else was doing this, it would be a profitable product.

This doesn't make Murray's proposal work, though. It's just that it isn't going to have the intended effect to say "medical insurers must treat the population as a single pool". Insurers will still be able to charge different prices to the different segments of the population by offering different products at different price points. And the problem with the Murray proposal is that this geriatric specialist insurance, to be as good as average US health care is today, would cost more than $10k a year, so you wouldn't be able to buy it on the UBI.

Just like selling Pokemon cards and golf clubs, technically they are not restricted to 10-year-olds and 40-year-olds, but in practice....

Expand full comment

That would be a tricky, but possibly functional solution, if you were to insure procedures instead of costs, and then offer plans that include or exclude certain procedures and vary price that way. Young people would get the "Car accident, broken bones and reproductive health" package, and then old people would skip the reproductive health and add on the "old as hell" package.

The trouble I see is that some (most) old age procedures become almost essentially certain, so you run into the same issues as we do now with the adverse selection issues. Not to mention the problem of doctors inflating prices because they know patients are not paying, and patients getting all sorts of recommended tests and procedures that do very little (and perhaps are net negatives) because they are not paying for them.

Expand full comment