Today I’m going to talk about how insurance works. I used to work in the industry (on the life side, but in the American industry) and I studied it in detail at the time.
Insurance is making book
One of my friends was an underwriter, and she used to say “I feel like a bookie. I figure out the odds on people’s lives.” I used to joke “when a life insurer writes you a letter saying they’re sorry about your relative dying, you can be sure they mean it from the bottom of their cold shriveled hearts”.
Which is to say, what insurance companies do is look at how risky a person or a group of people are, assign odds to them taking a loss and figure out how much the average loss will be. That’s the basic cost of the insurance. To vastly oversimplify and pull numbers out of the air, say you get insurance for getting your hand cut off, then replaced. The odds per year are 1 in 1,000 (you work in a machine shop) and the cost of reattaching your hand is 100,000 dollars. To break even they will have to charge each person $100.
Variability: God plays dice
Well, they would if there were no variability. Those may be the odds, but what if one year there number of machinists getting their hands chopped up is higher than normal? The fewer machinists the company is insuring the more variable this will be. Insure one machinist and you either pay out nothing or $100,000. Sure the odds of paying out are low, but if you do, your loss is 99,900. Ouch!
The more machinists the company insures, the less the variability. So they will want to insure as many machinists as possible. The larger the population they insure, the closer actual claim experience will match with theoretical claim experience. And if they insure every machinist in the country, well the experience should be very close to exactly what they originally calculated.
So the larger the population an insurer serves, the closer its experience matches what would be expected. The smaller, the less it does: the more the variability increases. This has a cost, the smaller the population you insure, the more money you have to put aside to cover variability. (1)
When you divide up an insurance market you increase variability, and therefore you increase the cost of insurance. It is impossible to avoid.
So at this point, the price of insurance is the cost of covering average losses plus the cost of covering variations from normal claims.
Underwriting. Or some people don’t have average experience.
So, since you don’t have a monopoly, and you aren’t the government, you don’t have the entire market. That mean your experience could vary for reasons other than just luck. Say you happen to insure machinists in a shop which has bad safety procedures and where machinists gets their hands chopped off at twice the normal rate? If you were to give them standard rates, and assume standard deviations from the norm, you’d take a huge loss.
That’s not a risk you can take. So you hire people to get all the records on the shop, someone to inspect it and someone to make book on the shop. How likely is it that this shop will have more (or less) than the normal experience? How much more, or less, should you charge them based on their safety precautions or lack thereof?
And if individual machinists are applying it gets even worse. Not only do you have to look at where they’re working, you have to make sure the applicant isn’t accident prone. How’s his eyesight? What’s his personal safety record? Does he have drug or alcohol problem? Any medical problems which might make him more likely to slip up? How old is he and what is the relationship between age and accidents? Etc…
All of this underwriting costs money. You’ve got the people getting the records, doing the medical tests, visiting the sites and so on.
So the price of insurance is now the cost of average losses modified by your specific population’s characteristics + variability + underwriting costs.
Anti-Selection: or, people who need insurance are bad bets (2)
One of the biggest problems in insurance is that an applicant knows more about his situation than you do. When everyone isn’t automatically enrolled, studies find that those people who do insure themselves are more likely to have adverse events. Think about it. If you have reason to believe you won’t get sick, you’re more likely not to get health insurance. If you live in a geologically stable area with no woods nearby in a stout well built house you’re far less likely to apply for catastrophe insurance, and so on.
People who want insurance are bad bets. So you have to underwrite them and investigate them. What it is that they know that you don’t know, that you need to know? (This is why insurance companies see recissions as moral: they figure that people held out on them, and holding out on them is effectively cheating them.)
But even after you underwrite, the people you underwrite will still have more incidents than they should have. So you have to put in some extra money to pay for that. If you insured the entire population you wouldn’t have to do that, but you don’t.
This, by the way, is one reason why insurers prefer to insure groups, like companies. A company hires employees for a purpose other than buying insurance. There shouldn’t be any anti-selection. If anything, a corporation’s employee’s experience should be slightly better than the general population, because people who are really ill (or likely to die) are less likely to be employed. Sick people don’t make hiring managers day.
So, the price of insurance is now up the cost of average losses modified by your specific population’s characteristics + variability + underwriting costs + anti-selection costs.
Insurance Companies Aren’t Non-Profits
Well, most of them aren’t. Companies that are mutual companies (fewer and fewer are) are sort of non-profits. But the rest of them need to make a profit. Traditionally 5% was considered a reasonable profit in the insurance industry—it was a widows and orphans sort of industry which invested in the dullest of dull, safest of safe securities. But, alas, when you’re a stock company, and Wall Street is demanding 15%, well, you’ve got to make that amount of profit, because if you don’t the stock price won’t go up and the executives options and bonuses won’t be as high as they could be. And we all understand that would be an unacceptable disaster.
So, the price of insurance is now up the cost of average losses modified by your specific population’s characteristics + variability + underwriting costs + anti-selection costs + profits.
Investing for the future
Health insurance costs can be roughly divided into two parts. There are ongoing health care costs which are about even over the life of your insured, and there are catastrophic end of life expenditures. About half of the cost of health care for Americans comes in the last year of their life (the problem being guessing which year is the last year. My father has been through his last year twice.) Insurance companies take the premiums they gain and the invest them in various securities. Roughly half of them go into short and medium term duration investments, the other half go into long term bonds.
This may not seem like a cost, but the point is that they have to take in more money than they are spending any given year, so they can invest that money for the future. That is an expense. It is an expense which a national insurance system which simply pays out of pocket each year does not have.
So, price of insurance is now the cost of average losses modified by your specific population’s characteristics + variability + underwriting costs + profits + investment costs.
Private Insurance Costs More
Why? Because it is the cost of average losses modified by your specific population’s characteristics + variability + underwriting costs + profits + investment costs.
Public insurance is either the cost of average losses + (very small) variability
or, if you do invest rather than paying year to year
Public insurance is the cost of average losses + (very small) variability + investment costs
There’s more to it than this, of course. But those are the basics embedded in the way that insurance itself works.
Let’s talk about some other issues.
A Sure Thing Isn’t Insurance
If you aren’t insuring the entire population another issue crops up. What about individuals who you know are going to take a loss? In medical insurance, what about someone who’s an asthmatic, or a diabetic, or has Parkinson’s, or whatever? The odds of that loss aren’t (population X incidence rate), they are already known. Insurance is about odds, about making book. When someone is already a claimant before they even sign up that’s not insurance. It’s just a known cost and you can’t provide it for less than the cost of providing it. You can’t make a profit on someone who’s already sick, unless you charge them more than the cost of their healthcare. At which point, why bother? If instead, you were to average their costs over the population of people who haven’t suffered a loss, well, that increases the cost to everyone. And it isn’t insurance. Insurance is about losses which haven’t yet occurred
This is why insurance companies won’t take people with pre-existing conditions. Either such people are a straight loss, or they increase cost and price for everyone else, and that damages the insurance companies competitive position. Either way it damages profits.
The Gatekeeper problem, aka: doctors and hospitals and drug companies want more money
Or to put it another way, between the insurer and the insured are gatekeepers who decide (or try to) how much care those people get. The more care the person receives, whether they need it or not, the more the gatekeeper gets. The insured isn’t in a position to deny care in most cases. If your doctor tells you you need a test, how often do you argue. So as an insurance company you have two choices. You can just accept the bills, and pay them or you can try and manage the care the insureds receive. Even if this is totally good faith (you actually do want people to get care they need, just not care they don’t need), well, denying care costs money in terms of staff and so on. And once an insurer realizes that staff saves you money, well perhaps you should deny even more care and turn into a profit center.
Of course, somehow costs do keep rising fast, and part of it is that some actors don’t get negotiated with very hard, for example drug and medical appliance companies. The question of bargaining comes down to difficulty: it’s easy to screw over a sick individual. He or she doesn’t have the power, money, time or health to fight back effectively. A big pharma company does. So the small actors get squeezed, and the larger actors make outsize profits (for most of the 00’s the only sector making more profit than the pharmaceutical companies were the financial companies.)
Lessons for health reform and the public option
Look again at the formula:
cost of average losses modified by your specific population’s characteristics + variability + underwriting costs + profits + investment costs
The public option, should eliminate profit costs, at least for it.
Removing the ability to deny applicants + not allowing different premiums due to history eliminates underwriting costs.
Anti-selection costs are reduced but not eliminated if one of the plans is seen as better than the others when it comes to actually delivering care (i.e. if insurance companies continue to make it hard to make claims and the public option doesn’t, the public option will suffer from anti-selection
Variability costs due to population size are not eliminated
Investment costs are not eliminated
The gatekeeper issue does not go away. Especially if the public option is specifically forbidden to negotiate with pharmaceutical companies.
The sure thing problem goes away, but could lead to an increase in prices depending on whether gains from reducing uncovered catastrophic care are enough to offset sure costs for actually treating people
Concluding Remarks
This is a fairly simplified look at how insurance works. But it does suggest a number of things.
- Insure the entire population’s so you eliminate variability, anti-selection and the need for underwriting
- Remove the need for profits
- move to pay-go rather than pay-in and invest
- Allow negotiation with large actors
- Move away from fee-for-service(4)
Next week I’ll talk more specifically about what is required for a public option to work to reduce the number of uninsured people, under-insured people and to reduce health care costs at the same time. While single payor automatically eliminates all of these problems except for fee-for service, public option needs to be done very carefully to actually work.
Notes
1) Note that in health insurance terms a population could be “all 66 year olds”. Medicare has an entire population, everyone above a certain age. There are still disadvantages to this because if you don’t have the population through its lifespan you can’t institute preventative care and other insurers can try and dump costs on you, but you do get the variability taken out to as large a degree as possible.
2) Also known as adverse selection. Same thing.
3) Mutual companis had consistently lower costs for insurance than stock companies. They are better for people who are insured, the data is dead clear on this. One model of reform is forcing all stock companies in the insurance business to go back to the mutual model.
4) There are some arguments against moving away from fee for service. It seems to work well enough in various other countries and inside some parts of the US health care system. When economics fails it tends to fall back on bad sociology, but the fact is that a health culture amongst health care providers doesn’t take advantage of fee for service to do necessary procedures and tests. Unfortunately, so much of US medical culture is sick that messing with incentives (the last refuge of bad management) may be necessary.
