> What I meant, though, is that across a big population’s entire lifetime, there will be a ton of high healthcare cost events. And with technological progress, new treatments will always be coming out. Which is a great thing, just not what is typically thought of as an “insurable risk”.

Sorry, how does that follow? Insurance works any time you have a function with predictable average but high variance. Is the total health care expenditure across a relevant subscriber base in 2023 very close to 2022? Then you can make insurance work. It's just math.

Insurable risk as in charging someone an appropriate premium that is based on their specific expected loss. Property and casualty, term life, etc.

Non insurable risk as in charging someone a premium unrelated to their specific expected loss (which is what health “insurance” is).

How exactly are health insurance premiums "unrelated" to care outflows (what you're calling "expected loss")? Are you saying that health insurers books don't balance and that they're losing money (they aren't) or making too much profit (not unless they're criminally hiding it)?

What you're saying doesn't make sense. There's no difference between health insurance and any other insurance in the way it works. You collect reliable and regular premiums from everyone, pay out unreliable/bursty (but statistically very predictable in aggregate!) losses as contracted, and pocket the remainder as profit. And it works.

Really, I don't know what you're talking about here. Health insurance is "expensive" in the US, sure. But it's not failing.

I meant the premium for a specific person is not related to their expected loss.

For example, if you carelessly drive and get into car collisions where you are at fault, your premiums go up, because your expected loss goes up.

In health “insurance”, it does not matter what you do, because your premiums are not based on your expected health costs. Hence it is more akin to a tax (or subsidy).

You're really not understanding this. That's got nothing to do with the insurance model. That's just a regulatory thing. All those choices do is change the specific population that gets insured in a single pool, such that their specific computed premiums are different. But for ANY such population, the total premiums paid == the total loss outflow + a reasonable profit. And you can tell that's true because the accounting for those companies appears in their SEC filings.

In practice, car insurers are allowed to partition their customers this way because it's felt to be "fair" and because it encourages safe driving. Trying to partition health insurance customers like that feels "unfair", and has minimal net benefit as health expenses aren't as controllable-by-the-subscriber as car accidents are. So we pass laws about how the partitioning gets done.

But again, "insurance" as a business model (and mathematical model) works EXACTLY THE SAME WAY. The only difference is how you draw the lines around who gets insured at what rate.

> All those choices do is change the specific population that gets insured in a single pool, such that their specific computed premiums are different. But for ANY such population, the total premiums paid == the total loss outflow + a reasonable profit. And you can tell that's true because the accounting for those companies appears in their SEC filings.

I do not dispute this. As I wrote in a sibling comment:

>Insurance and tax/wealth redistribution is a spectrum.