Bruce Pyenson: Adverse selection describes a situation where an insurer enrolls an unexpectedly higher concentration of less healthy individuals. Of course, the idea of insurance means the insurer will lose money on some people and make money on others. But people, like insurers, naturally make choices that maximize their own financial interests, and certain insurance product characteristics and market conditions may create situations that result in a greater concentration of less healthy (or more expensive) individuals covered under an insurance policy. In its most severe form, adverse selection is the insurance equivalent of a run on a bank.

If this seems unclear, consider the 80/20 rule. Something close to the 80/20 rule applies to healthcare--about 20% of the people generate about 80% of the cost. Adverse selection can happen because not all of us are average from a health cost perspective; as a matter of fact, almost none of us are average. Most healthcare costs come from relatively few people. In any year, the people who are low cost may not feel they need good coverage, if any at all. The people who are high cost really need the financial protection of coverage.

When insurance coverage is optional, which is generally the case in the United States, the high-cost people naturally pay for insurance but will use more than their share. In order for this insurance mechanism to work, the insurer must essentially receive an offsetting subsidy from healthier, lower-cost individuals. If too few of those healthier lives buy insurance, the subsidy is not sufficient. In that situation, the 80/20 rule becomes a "70/20" rule, and each of the healthier people then needs to pay a greater subsidy or the insurer will lose money. That extra subsidy paid by the healthy people then becomes an incentive for them to go elsewhere to seek lower cost insurance. In some cases the healthier lives may forgo insurance altogether.

None of us are very good at predicting the future, especially when it comes to our health. Doctors are not very good at it and the various insurance predictive models that try to do it are not very accurate on an individual basis. People may believe they are healthy this year, but they really don't know if they are going to be healthy next year or the year after that. This uncertainty drives the need for an insurance system. Most of us feel we need good coverage and will pay for it, even when chances are that we don't need it--"just in case." The fact that most of us get health insurance with employer subsidies through a group policy also helps. The 80/20 subsidy works pretty well most of the time.

Q: Can you provide an example of adverse selection?

Bruce Pyenson: Market conditions can lead to adverse selection. For example, in the 1980s, many Blue Cross Blue Shield plans sold insurance to small groups with community rating--all groups were charged the same per-capita premium regardless of the age, sex, or health status of the members. "Shuffleboard Seniors"--a group of older sicker members--would be charged the same as "Beachball Bums"--a group of young healthier people--so the latter would subsidize the former.

Unfortunately for the Blues, other insurers set rates based on the age and sex of members and thereby lured away Beachball Bums. The subsidy that community rating required from healthier lives started disappearing and the Blues lost money. When they raised rates, even more of the healthier groups shifted to the competition, which led to greater financial losses for the Blues. Eventually, most Blues plans threw in the towel and moved away from community rating.

In a broader example, people at one time could purchase noncancelable policies. Over time, the healthier lives left, leaving the very expensive lives behind. Insurers were left with very small blocks of business that they couldn't cancel. Attempts to raise rates (subject to regulator approval) could not account for the full cost of these lives.

This dynamic is known as the insurance death spiral: The more you raise rates the more you drive out the healthier, lower-cost individuals. The ones that stay behind are really bad, and you have to raise the rates again, and charge more from the healthy individuals, which encourages them to seek coverage elsewhere.

In employee benefits, there is a rule (COBRA) that says that when someone leaves a group insurance program they have the right to buy into that insurance program for a couple years using a rate differential that is set by the government. They have at least 60 days to sign up. The long sign-up period opens the door to adverse selection. If someone has a condition that emerges in those next couple of months, they are much more likely to sign up for insurance, which may explain why the experience of the people who sign up for COBRA benefits is surprisingly bad.

Q: How does adverse selection relate to healthcare reform?

Bruce Pyenson: Critics of any particular kind of healthcare reform are often quick to point to the potential for adverse selection. Insurers understand that you don't sell fire insurance to a homeowner whose house is on fire, and you don't sell life insurance to someone with a terminal illness. Rules like these are essential to make insurance programs viable. That said, adverse selection doesn't need to be a death knell for reform. As I started out saying, insurance is about low-cost people subsidizing high-cost people. Payers can manage adverse selection if they have the right structures and the right tools.

Q: How do you manage adverse selection?

Bruce Pyenson: Over the years, commercial and government insurers have developed a handful of techniques to manage adverse selection.

First, make sure the insurance is reasonably attractive for the buyer. You want to have a broad base of insureds who buy the insurance in order to appropriately spread the risk among the high-risk people and the low-risk people. For example, participation is very high in group insurance programs, where the employer subsidizes a lot of the cost. That's also the case with Medicare Part B, where the premium is set at about 25% of the actual cost.

Another technique involves penalizing people who can sign up for insurance but don't do so when they are first eligible. These kinds of rules, including preexisting condition exclusions, prevent situations where people wait to sign up for insurance until they need it. Medicare uses a rule with Part D where, if you don't sign up when you are first eligible (generally age 65), the person has to pay 10% more for each year they delay signing up, and they pay the additional amount forever. Rules like this encourage early participation and give the insurer flexibility in how rates are set so that the rates charged can bear some reasonable association with the risk being written.

Q: Are there basic lessons about adverse selection that we can apply to the larger healthcare reform conversation?

Bruce Pyenson: Yes, we've identified a couple of basic lessons. First, do not expect to set up a system that is going to work for years to come without flexibility. We are in a dynamic healthcare system and change continues to accelerate. Any healthcare reform proposal should assume that rules made today may not work the same five years from now (let alone 20 years from now). An independent board with authority to make and change rules is one approach. The payers--and this is true for both a government payer and for multiple private payers--need flexibility to deal with situations as they emerge.

We can also manage adverse selection by getting as close to universal coverage as possible. I think the smart money says that you are not going to have 100% coverage. There will always be issues where people, for one reason or another, do not get coverage. But making coverage as broad as possible will help prevent adverse selection.

The dead bear in the room: the cost problem. Part of the reason adverse selection is such an issue is that the cost of health benefits is too high and is getting worse. As costs grow, some families may succumb to the temptation to use money spent on insurance premiums for other purposes and go uninsured. A lot of healthcare reform proposals are just window-dressing unless we face up to that reality.

Bruce Pyenson, FSA, MAAA, is a principal and consulting actuary in Milliman's New York office.