Fundamentals of Insurance (Part 3): The Bad (Adverse Selection)

The other complication for health insurance is that not everyone has the same likelihood of getting sick. The magnitude of this problem depends on how well the insurance company can identify these differences. Insurance works most effectively when the insurer is better informed (a fact that doesn't sit well with those behind the current reform).

Paying for what you get

Imagine an alternate universe in which insurance companies are allowed to review your health history and habits, and quote you a premium accordingly. (For health insurance, this is essentially outlawed in most states. For life insurance, however, it is completely allowed.) In such a world, your premium would reflect your average medical costs. If my family history reveals that I have a higher than average chance of prostate cancer, I will pay correspondingly higher premiums. If my non-smoker status indicates that I have a lower than average chance of lung cancer, I will pay correspondingly lower premiums.

In essence, I am simply paying for what I get. If the insurer is likely to pay out more on me, it's hardly unfair that the ask me to pay more for their service in my premiums.

I'm guessing that most people agree with that concept when it comes to health issues over which a person has significant control, such as smoking, obesity, or pregnancy. If you decide to smoke or have a child, you should be willing to bear your added health costs as part of that decision.

Where people get worked up is for things outside of personal control. Should I really be penalized with higher premiums because of the genes that I inherited? Admittedly, it seems unfair (I dealt with this issue in applying for life insurance, and it bothered me). But the real question is, what is the alternative?

Suppose we insist that insurers cannot discriminate by charging different premiums based on health history. In other words, we are saying that for some of their clients, the insurer must knowingly expect to spend more than they will earn. Would we force a restaurant sell dinner to some clients for less than it costs to produce? Doing so is a sure way to bankrupt and close down the restaurant.

Cross subsidizing

People justify this by saying we can cross-subsidize among clients: if everyone is charged the same premium, healthy people will pay more than their average cost, and that make up for unhealthy people who pay less than the average cost. The first issue with this argument is that we have now moved beyond insurance. It is now welfare.

Note that in using the term welfare, I intend it only in the technical sense: an involuntary transfer of wealth from one group to another. This is not meant to pass judgement (i.e. I'm not saying welfare = bad); the topic of redistribution will be addressed in a future post.

People's feelings on redistribution may differ. At least this seems to transfer the wealth in the right direction (to those who are worse off). However, mixing insurance with redistribution could make both less effective: it may not help the unhealthy as effectively as a direct subsidy, for instance. And it can significantly harm the way insurance operates.

Opting out

Typically, if the insurance company can judge your health risk, so can you. If I am a young person with a clean health history, I expect my health costs to be quite low. But then if I can only buy insurance under a cross-subsidization scheme, I will be charged far more than my expected health costs. Even if I can afford the premium, I could rationally choose to stay uninsured. Indeed, as many as half of the 47 million uninsured fall into this category, deciding that insurance isn't worth the premium.

Of course, this behavior will undermine the cross-subsidization scheme. The only people who actually sign up for the scheme are those with higher costs, which pushes the premiums higher still. Put another way, the insurer has to assume (and set premiums accordingly) that only the least healthy clients will ask for insurance. (Economists call this phenomenon adverse selection.) This leaves many people uninsured, and limits how much redistribution actually occurs too!

The left argues that adverse selection necessitates a single payer insurance program. The government has the advantage of guns and prisons — it can force healthy people to keep paying their premiums, even if they won't get their money's worth. But that is only true if one assumes (as the left does) that redistribution must be integrated with insurance. If insurers are able to customize their prices and their coverage, they can provide insurance (and only insurance) quite effectively.

In practice, since they are not allowed to charge different premiums, private insurance companies instead limit coverage on people with higher risk. In particular, pre-existing conditions are situations in which a disease is almost highly likely to be a continuing problem. Again, many people spit venom over these limitations. As for me, if I had a pre-existing conditions, I would prefer having coverage with a higher premium over having those conditions completely uninsured.

In it for the long haul

One more topic deserves to be addressed. What about the information we obtain about ourselves as life progresses? Should an insurance company be able to change rates or coverage on you once you have come down with a particular disease?

In some ways, this is merely a rephrasing of the original question about charging different premiums based on different health conditions. But here there is an initial uncertainty when you first buy the insurance. No one knows if you will have cancer in the future; and that sort of uncertainty is exactly what insurance was designed to handle.

At your initial enrollment, one can imagine two possible insurance contracts you could sign. In one, you are promised that your coverage and premiums will stay constant for the rest of your life. In the other, you are warned that premiums will rise when long term health issues arise.

The first contract would have to cost more than the second, because the insurer is taking on a bigger commitment. At the same time, clients would value the first contract more, since it smoothes out more of the risk. So it's a question of whether the added cost is worth the added benefit, and I'm sure some people would think so. Auto insurers are starting to offer long term contracts of the first style (AllState's accident forgiveness, for example).

Whatever the contract says, I firmly advocate enforcing it. If they promise lifetime coverage and then renege, they should be taken to court. But by the same token, if a client signed the second contract, he has no right to complain when 10 years later, something bad causes his premiums to rise. After all, for 10 years he benefited from a lower premium than in the first contract, precisely because the insurer didn't promise to hold rates constant. You can't have it both ways.

Long term insurance contracts are almost uniquely problematic to the health insurance industry — it is easy to find long term rate commitments in life insurance, disability insurance, and to a lesser extent, in auto and home. Much of this is due to the way health insurance is tied to employers (a strange accident caused by government intervention, to be discussed in future blogs). Most of us change jobs several times in our life, and thus won't be insured by the same firm long term. As we shall see, this single feature causes many of the problems at the heart of the debate.