undefined | Better HN

0 pointscurun1r8y ago0 comments

Let's say your odds of getting that cancer are 5% and you know that ahead of time. So, knowing that it would bankrupt you, you gather 20 other people who are also at a 5% risk of getting cancer to form a group where you all chip in no matter who gets sick. You've now spread the risk out across a pool of people and, yet, everyone is still paying their fair share because no one knows who will be the unfortunate cancer patient at the time the agreement is made.

But in your little group, there's still a 36% that no one will get cancer and that's offset by a small chance that multiple members of you group will get the disease. If even 4/20 members get cancer, that would now come pretty close to bankrupting everyone in the group. So what do you do? You increase the size of the group and lower the chances that no one will get cancer while also increasing the chances that only a proportional number of people (5%) in the group get the disease.

But then along comes Danny. We also know that Danny has a 10% chance of getting cancer. He also wants to join our group. If we just let him join, it won't be fair...after all he's twice as likely to need us all to pay for him. So we can either tell him to get lost, or we can just make sure that when the bills for someone's cancer show up, he pays twice as much as any other member. He's bringing twice the risk with him so he pays twice as much.

In poker, players learn to avoid being outcome-oriented when evaluating their decision making. Rather than looking at a decision they made and seeing whether they won or lost, it's more important to look at the decision and ask themselves if they'd made the same decision a million times, how many times would it lose and how many times would it win. Because any individual hand where a correct decision is made can still lose due to variance.

Insurance is the same concept. Actuaries use statistical analysis to determine an individual's risk so insurance companies can price that risk accordingly. Over a million lifetimes, your insurance premiums (less the insurance company's profits) would come very close to the overall insurance payouts. But since you've only got the one life, you'll be under- or over-paying. But you're still only paying for your own risk, you're not paying for anyone else's risk.

Where the ACA comes in is when Danny can't really afford the double stake. It forces us to let him join the group and pay less than double whenever someone gets cancer. That's a bad deal for all of us 5%ers, despite the fact that there's still a 90% chance that Danny doesn't get sick.

Now try to answer your question...do you see the difference between paying for your own portion of the risk pool vs lower risk individuals subsidizing higher risk individuals? If you were in one of those groups of 5%ers, would you just let Danny join without adjusting for his added risk or would you be looking for more 5%ers like yourself?

0 comments

b1daly8y ago

I’m repeating myself, but you are wrong in at least two ways.

Health insurance in the US is not conducted in any way like the over simplified, text book example you provide. There exists no real world example of health insurance that functions as you describe.

There is the obvious problem that there is nowhere near that level of precision in anticipating the given level of risk for any individual getting a certain disease, especially as most populations are fairly heterogeneous when it comes to health risk profile.

The insurance companies use statistical tools to estimate the average risk of insuring a pool of people with some element of similarity, like a pre existing conditions. This is a far cry from being able to evaluate the risk profile of a given individual.

The complaints people have about being rejected for coverage based on a pre-exsisting condition, when they think that condition doesn’t really make them that risky reflect the imprecision of the actual tools.

This is also illustrated by these blanket denials: insurance companies are rejecting populations that the feel cannot be adequately evaluated for risk.

There is another aspect of such complaints (about blanket denial of coverage) which come from the issue that some people simply do have a health condition that they cannot cover the expected cost of.

This is a societal issue, and different socities deal with it differently. It inevitably involves some sharp elbows between the various interests involved.

In the US, for historical reasons, a big method of covering the costs of those who can’t afford the full premium was through the patchwork system of employee health plans, complementing government run plans like social security, Medicare, Medicade, local government programs, and charitable programs.

As imperfect as the system is, it remains the case that the health care coverage of high risk individuals, meaning those who can’t cover the full cost of a risk based premium, relies on what is known as “health insurance.”

wyager8y ago

> There is the obvious problem that there is nowhere near that level of precision in anticipating the given level of risk for any individual getting a certain disease

This doesn’t matter at all, on account of the way variances from independent RVs add up. Look up the Bates distribution as an example. Variance over mean goes down as the square root of the number of independent RVs (insurees). That’s like half the point of insurance.

> was through the patchwork system of employee health plans,

These were the result of highly nonlinear tax policy during WWII, not the reasons you’re claiming.

j / k navigate · click thread line to collapse