Health Insurance: How Does It Work?

Four Questions to Ask About Health Care Reform

We rarely have to ask how car insurance or home insurance work. What is it that makes health insurance so much more complicated — with individual mandates, single or multiple payers, and so on? Since Congress is (once again) talking about changing how health care works in America, it may be a good time to explain what makes it so strange. This thing we sometimes call “health insurance” is actually a fusion of three different things, only one and a half of which are “insurance” at all. I’m going to walk through the basics you need to know, and end up with a set of four questions you can ask about any health care proposal that will let you understand what people are really talking about.

(I’m not going to tell you what I think the right answers are, though. This article is to help you ask the right questions — and I’m interested in your answers.)

Not as bad as it looks: this is a mass casualty drill at the Naval Medical Center, San Diego.

Let’s start by talking about what insurance actually is. The first modern version of insurance was invented in the Netherlands in the 1600’s, when people were finding that sending a ship on a trading voyage could be very profitable. The good news was that you could make back the whole cost of your investment (a ship, a crew, some things to start trading with) in just a few voyages; the bad news is that a lot of voyages also ended up with shipwrecks, which would take your entire investment to the bottom of the ocean.

Let’s say for simplicity that it would take you ten voyages to make back the cost of a ship, and one voyage out of twenty got shipwrecked. Suddenly, your life depends a lot on luck: if that shipwreck happens on your eleventh voyage, you’ve got enough saved up to buy a new ship and keep going. If that shipwreck happened on the tenth voyage, though, you were screwed.

Pretty soon, some fairly wealthy men figured out a way to fix this: they would offer an “insurance contract,” where before each voyage, a shipowner would pay them one-twentieth the cost of a ship plus a fee, and if they were shipwrecked, they would get paid the full cost of getting a new ship. Both sides came out happy: on the average, the insurer was earning the fee for each voyage, and the shipowners had turned an unpredictable, possibly catastrophic, expense into a regular, predictable one.

This is the simplest kind of insurance: it takes a large (but known) expense that might happen at an unpredictable time, and splits it up into a lot of small expenses that you can plan for. Your homeowner’s insurance today is basically the same kind of thing.

There have been lots of improvements to this idea over the years. One that’s particularly relevant to us is actuarial science: the development of professional experts in figuring out the odds. In the story above, the insurer was charging one-twentieth the cost of the ship plus a fee, because he knew the cost of a ship and could estimate that one voyage in twenty shipwrecked. But if he predicted those odds wrong, he could be in trouble: too high an estimate would mean he was charging people too much, and they would go to his competitors (often just a few tables down at the café). Too low an estimate, and he’d be paying out more than he collected in premiums, and could go bankrupt within the week. So hiring a lot of clever people to figure out just how likely these events were became very important.

One of the first things actuaries realized is that not all voyages are equal. A trading voyage from Eemshaven to London is a lot less risky than one to Shanghai. A well-crewed boat in good repair is less risky than one that looks like it may sink if someone sneezes too loudly. An experienced crew is safer than a green one. This means that insurers could — and should — charge different voyages different rates. The perfect insurance contract is one where everyone knows the exact odds of something going wrong.

Each of these cases has something in common: everyone knows and can agree upon the price of the thing being insured. But sometimes, you want to insure against something that you know will be very expensive, but whose actual cost you don’t know ahead of time — say, the liability insurance on your car, which is insuring against the medical bills of someone injured in an accident. How does an insurer know how much to charge? And what if it’s a “catastrophic” event, one that ends up being so expensive that the total premiums that the person buying the insurance would never pay that much over their entire lives?

It turns out that you can meaningfully insure these too, so long as the actuaries can figure out the average costs, and figure out the odds of different possible expenses; you just spread these costs over the whole pool of insured people. If you know that one person out of every thousand will have a million-dollar expense, and you’ve got 10,000 customers, then you can expect to make $10M in payouts over the years, and price your premiums at $1,000 apiece plus a fee. This works out even though none of those ten people will ever pay $1M in premiums; the lucky ones pay, the unlucky ones get paid back.

At first glance, this kind of “catastrophic event insurance” works a lot like ordinary insurance; by spreading risk over the entire pool of insured people, it lets people insure not just against events which they could pay for if the cost were spread out over their entire lifetime (like ordinary insurance), but even against events which you couldn’t pay for.

But catastrophic insurance is only sort-of insurance: it has a fundamental problem. Imagine that one day, someone figures out that the one-in-1,000 odds above aren’t as simple as we thought: there are actually two groups of people (maybe good and bad drivers, or maybe people with different risks for some disease) that have different odds. That one-in-1,000 actually breaks down to 90% of people having a one-in-10,000 chance, and 10% of people having a one-in-100 chance. If you were charging people based on their odds, like ordinary insurance does, then 90% of people would be paying $100 instead of $1,000, and the other 10% would be paying $10,000.

What this means is that, when everyone is paying the same rate, the “safe” people are really subsidizing the “unsafe” people. And so long as nobody can tell which pool anyone is in, that’s how all insurance works. But as soon as people get a better understanding of risk, things change. If catastrophic insurance is being sold on the free market, things change quickly: even if one insurer decided to ignore it and keep charging everyone $1,000, someone else could show up and insure 90% of their customers for only $100. Pretty soon they’d have all of those customers, and the original insurer would be left with just the 10% with higher risk; they’d have to raise their rate to $10,000 just to avoid going bankrupt!

That is, everyone would end up charging the same rates that they would if this were ordinary insurance; there’s no way (in the free market) for the insurers to spread the costs around. It becomes impossible to insure against events which cost more than any one person could afford.

The only way to prevent this is to move away from a perfectly free market. If insurers were forbidden to change their rates or refuse customers based on this criterion, then you could keep everyone paying the same $1,000, and everyone would still be insured against this risk.

Of course, this doesn’t come for free; what you’ve effectively done is say that we will cover this risk, and everyone has to pony up to cover it. That is, this is a kind of tax; it can happen the way described above, with insurers required to cover something and the cost being spread across everyone insured, or it could be done as a literal tax, with everyone paying in and some central service paying out and no private companies involved, or as anything in between. Whether this is a good or bad idea depends on the situation.

The upshot is this: catastrophic insurance (which spreads people’s costs across the whole population) is unstable as a free-market good, because as knowledge of risks improves, it turns into “ordinary” insurance (which spreads an individual’s costs across time). If its cost as ordinary insurance is more than people can afford — that is, if the thing being insured against is something which people simply couldn’t pay for if it happened, like cancer treatment or accident liability — then that risk becomes uninsurable. The alternative is to make a social decision that payment for this (we can’t really call it “insurance” anymore) is going to be made out of a fund that everyone chips in to; which is to say, a tax, which pays for the benefit that people are now protected from a risk which can’t be protected against via insurance. That decision amounts to saying “everyone is chipping in to pay for X when it happens to people.”

So whenever there’s a kind of catastrophic event that people might want to insure against, we should think about it the same way we do about a tax. Is it worth it? How will the costs be distributed among people? How will the benefits be distributed? Is it better to do this with insurance rates or a literal tax? This is the question we have to ask about any kind of catastrophic insurance, whenever we either have figured out a way to split up the risk pool, or when we think that’s likely to happen in the near future.

There’s one other thing that turns out to work almost the same as catastrophic insurance: pre-existing conditions. These are just like ordinary medical conditions (either of the sort that require ordinary or catastrophic insurance), but it’s not about people with a high risk of having something, but people with a certainty of already having it. That means that the only reasonable “premium” you could charge them is the full cost of treatment plus a fee. Since that’s more than the cost of treatment, you would have to be an idiot to pay that; pre-existing conditions aren’t insurable at all.

Like with catastrophic insurance, we have two basic options: leave it to the market, with pre-existing conditions uninsurable, or require insurers by law to accept people with pre-existing conditions. (In fact, we always did this in some ways, because otherwise an insurer could say you had a “pre-existing” condition the moment you got sick, and kick you off insurance before paying a dime. That way they would get to keep all your premiums and never have to pay out anything, which would be fantastic from their perspective. Insurance contracts don’t allow this — but insurers are quite known for finding ways around that, such as pressuring employers to fire people whose family members are seriously ill, or face having their own premiums spike.)

The problem with requiring that insurers accept people with pre-existing conditions is what happens when people move from insurer to insurer. If insurers were required to accept you, you would just not bother having insurance at all until you got sick, and then get insurance, pay a small premium, and have everything paid for; the insurer would go bankrupt overnight. This is just the same as the situation with catastrophic insurance, except now instead of two insurers competing and all the low-risk people going to the cheap one, each insurer would be competing with “no insurance at all.”

This is the origin of the “individual mandate” in the ACA: if you want to require insurers to cover pre-existing conditions, you also have to ban individuals from not having insurance. And since literally banning it (with a prison sentence?) seems like a terrible idea, the ACA compromised with a fine.

If we put this together, there are two options for how you handle pre-existing conditions:

• Option 1: Insurers aren’t required to cover pre-existing conditions; anyone who has to change insurers when they have an illness is uninsurable.
• Option 2: Insurers are required to cover pre-existing conditions; we spread the cost out over everyone by requiring everyone to buy insurance.

There are three similar options for catastrophic insurance:

• Option 1: Insurers can charge different risk groups different amounts for this kind of catastrophic event; catastrophic insurance turns into ordinary insurance as our understanding of risk gets better, and people at high risk for this event are uninsurable.
• Option 2: Insurers are required to charge different risk groups the same for this kind of event; the cost of handling it is spread out over everyone through the premiums.
• Option 3: Ordinary insurers don’t handle this kind of event at all; we set up a separate system that takes in a fee spread across everyone (i.e. a tax) and pays for treatment.

The first two options are just like the ones for pre-existing conditions. The analogue of the third one would be if we said that for any condition that can be pre-existing and that individuals couldn’t usefully pay for on the spot, the idea of “private insurance” doesn’t really work, and so we have a central system to handle it.

Summarized another way, “insurance” for catastrophic events or pre-existing conditions doesn’t really behave like insurance; if you try to sell that on the free market, it quickly ceases to exist. So if you want it to exist — that is, if you want there to be a payment mechanism other than cash on the barrelhead — for any such conditions, you have to create it outside the market, by law, and require the costs to be spread out. The difference between options 2 and 3 is just about whether private companies operate it day-to-day.

Any health care proposal that involves any spreading out of costs faces the same question as a tax proposal: how do we spread out that cost?

All three of these options have been used in real life. For pre-existing conditions, option 1 is what we did in the US before the ACA; option 2 is what we do today. For catastrophic conditions, most kinds of insurance start out as option 1 and gradually evolve to option 2 as our understanding of risk gets better; if something is completely uninsurable and no company wants to deal with it (like flood insurance for houses, uninsurable because everyone gets flooded at once) then we often move to option 3.

Option 3 is also popular in many countries for health insurance, where it’s called a “single-payer” system. This has the advantage of combining a solution to the catastrophic and pre-existing care system with the fact that health care access doesn’t work well on a market, either; it’s basically saying that health insurance, as a whole, doesn’t really work as a free-market good, and if we’re going to have it we have to build it as a big social project, much like we do the roads. It has the downside of pushing everyone into a single system, which is why many countries actually use a hybrid: there’s a central single-payer system that everyone uses, and you can also buy additional insurance (with maybe access to different kinds of hospital or doctor as well) on the open market.

We also don’t have to make the same choice for each kind of insurance. Your car’s liability insurance, for example, works by something between options 1 and 2, while if you have any flood insurance you bought it through a semi-governmental system using option 3.

One final thing to remember is that whenever we’re “spreading out the cost,” there’s no opt-out by definition. But some people can’t pay that cost; what happens to them? This is the same problem that we encounter with any tax — taxes are, after all, just us paying for something (be it roads or armies) that we’ve decided as a society we need to build, even though there’s no good way to do it individually. So any health care proposal that involves any spreading out of costs faces the same question as a tax proposal: how do we spread out that cost? (And if we’re using a fine, rather than the tax system, to charge it, what happens to people who can’t pay?)

So let’s look at health care. Health care includes the full range of event types: manageable and predictable events, like kids getting strep or breaking their arms, and giant and unpredictable ones, like cancer. It includes lots of “pre-existing conditions” if there’s any situation in which you might have to change insurers, especially if you can’t control that. (Such as due to a layoff)

From Wikimedia Commons

But the thing we call “health insurance” also contains a third item, which has nothing to do with insurance at all — instead, it has to do with the strange way medical payments work.

If you walk into a hospital and ask how much a procedure will cost, you probably won’t get a straight answer. If you do, it will be heart-stoppingly high: $250 for a small bottle of apple juice as a “rehydration supplement,” $1,000 for ten minutes of a doctor’s time, and so on. The reason is that health care is paid for through a complicated system which barely involves either the doctor or the patient; instead, the medical facility bills the insurance company. And a big part of the contract negotiation between the two, when that provider and that insurer agree to work together, is an agreement on how much each item will cost. But the resulting contract is secret; neither side wants its competitors to know what kind of deal it made. The “official” prices that a provider charges are nothing more than the starting point in a negotiation, where the insurer might want lower costs for blood tests and accept higher costs for hospital beds.

But if you walk into a hospital without insurance, you’re not covered by any of those deals — so the price you’ll get charged is the list price, a price never meant to be paid by anyone. And those are prices which would instantly drive someone into bankruptcy.

(Then this whole story is repeated when insurance companies negotiate with employers to sell them insurance for their workers; the cost you pay if you need to buy insurance individually is far higher than the cost a big company that’s buying policies for ten thousand people is paying.)

What ends up happening in practice is that if you don’t have insurance, you simply don’t go in for routine care; you can only go to the emergency room, where they are required to treat everyone who comes in the door. (Which is another example of a societal decision to subsidize something!) Of course, you then get billed a tremendous amount, and have to try to negotiate with the hospital to reduce this to something payable — but now they have all the leverage, since you owe a debt.

So this third aspect of health insurance really has nothing to do with insurance at all: it’s that being signed up with an insurer gives you access to that insurer’s negotiated prices, without which you can’t get access to medical care at all.

We’ve already seen that ordinary insurance works well on free markets, and catastrophic and pre-existing insurance doesn’t work there at all. Health care access sits somewhere in the middle: you could sell it on the free market, but that turns out to be bad for almost everyone. For example, if health care access were something you bought, but emergency rooms were still required to treat anyone who came in the door (changing which is likely to go down very badly with anyone who’s sworn the Hippocratic Oath), then this would be equivalent to saying “people who can’t afford health care access go to the ER.” But emergency treatment is a lot more expensive than preventive care, not least because by the time you’ve gone to the ER (and are facing the potentially ruinous bills that would follow), you’re a lot sicker than you would have been beforehand. That imposes costs on the hospitals (who have lots of indigent patients to treat), on the people who are getting much sicker, on their employers or customers who have people who are sicker and can’t work, on their families, on other patients at the ER who have to wait because there are a lot of seriously ill people showing up who wouldn’t have been as sick otherwise, and most dangerously of all, on society as a whole when contagious diseases start going around — pertussis doesn’t give a damn what kind of health care policy you’ve put in place or how much money you have.

That is, a sick population creates lots of deadweight costs for everybody around, and preventive care reduces those costs across the board, creating a net profit for society. This means there are some kinds of care which everyone has an incentive to make as widely available as possible. You could sell access to those on the free market, but it’s almost certainly a terrible idea.

To sum up, we have three different things, all bundled under the misleading name “health insurance:”

• Ordinary health insurance, which splits up the cost of your expected lifetime medical bills over time;
• Catastrophic health insurance, which splits up the cost of rare expenses so big that people couldn’t pay them across everyone; and
• Access to the health care system itself.

And these have different requirements. Most importantly, ordinary health insurance is something you can sell on the free market — but catastrophic health insurance, and insurance for pre-existing conditions, isn’t. It exists if and only if we decide to spread out its costs across the wider population, which means some kind of law (basically, a tax) to do it. Health care access can be sold as a free-market good, but there are some aspects of health care access which, if we make them available to the public as a whole, create huge across-the-board wins.

When you start seeing proposals for health care laws, here are some questions you should ask:

• How does the proposal handle people getting health care access? If it doesn’t simply give it to everyone by fiat, what happens for people without it? If it does, how is that managed and paid for? Which kinds of care, if any, do we actively try to make available for everybody?
• Does the proposal provide ordinary health insurance at market rates? If not, how does it change that?
• How does the proposal handle catastrophic and pre-existing coverage? Does it use option 1 (don’t insure it), option 2 (private insurers but we spread out the cost), or option 3 (one central insurer and we spread out the cost)?
• If we are spreading out the costs for any situation, what happens to people who can’t afford that cost?

These are the questions you have to ask no matter whose proposal you’re looking at. Do we insure people with a high risk of cancer? If so, how do we spread out the cost? What mechanism do we use to collect that — insurance premiums and fines, an overall tax payment, something else? If not, what do we do with people at risk for cancer?

I’m not telling you the answers to any of these questions — just showing you the questions you’ll have to ask. The answers are good things to argue over with your friends and Congressthings.

The law we have today in the US is a strange sort of mish-mash. Everyone is required to get insurance, with various exceptions, or pay a fine. That insurance provides a combination of health care access, ordinary health insurance, and catastrophic and pre-existing care insurance, all done using option 2. The requirement to get insurance is the way we spread out the cost, and there’s a system of refunds to reduce the cost for people who can’t afford it. (Unfortunately, because of Congressional roadblock, those refunds are set at the “first guess” values which were in the original bill, and nobody’s been able to adjust them to what people actually need — which is what had to be done for every other system like this in the world.) People in any of those exceptions who don’t get insurance get none of the above, including health care access.

There are various ideas about “health savings accounts” being floated, and if one gets suggested soon, you should ask these questions about it. HSA’s on their own provide ordinary insurance, but neither access nor catastrophic insurance; they literally mean that you save up your money against the chance of getting sick in the future. (In this way, they’re awfully similar to “no health insurance,” since you could always do that to begin with — unless you find yourself like that shipowner whose ship went down before he had ten successful voyages, in which case you’re SOL.) This means that HSA’s have to be coupled with something else… but what else? That’s the $100-billion question.

As for other ideas… who knows? Take a look at whatever comes up, and ask these questions about it. It’ll make it a lot easier for you to navigate between competing ideas.