A Closer Look at Artificial Markets for Health Insurance

"Everyone would pay the same price, regardless of expected health costs."

How Managed Competition Differs from Real Competition.

• A person who has AIDS would be able to join a health plan for the same price as someone who does not.
• People in hospital cancer wards would be able to join a plan for the same price as people who do not have cancer.

In a competitive insurance market, two similarly situated individuals could expect to purchase health insurance for the same premium before either of them becomes sick. Under managed competition, they would be able to join a health plan for the same premium after one of them becomes sick. If the same procedure were applied to other forms of insurance, people would be able to purchase life insurance from any carrier at normal premiums after they became terminally ill. People would be able to purchase fire insurance on their home while it was burning or casualty insurance on an automobile after it had been destroyed in an accident.

Placing a Bet after the Dice Are Thrown.

One way to think about managed competition is to see it as a game in which people are allowed to place bets after the dice have been thrown. Premiums do not buy insurance against risky events; instead they constitute prepayment for the consumption of medical care. In such a system, there is no relation between the amount people pay and the value of services they receive. People can’t make rational choices between self-insurance and third-party insurance, since under managed competition there is no such thing as an actuarially fair premium. Every premium is artificially set either too high or too low relative to the risk and expected cost the individual brings to the insurance pool.

"Someone with AIDS would pay the same premium as someone who is healthy."

In a competitive insurance market, people are encouraged to make choices that are socially good, in the sense that they lead to an efficient allocation of risk. Under managed competition, people would be encouraged to engage in opportunistic behavior to the detriment of everyone else.

The Consequences of Choosing Insurers after People Get Sick.

• About 4 percent of the group (two people in this example) account for 50 percent of health care spending.
• About 16 percent of the people account for 80 percent of the spending.
• And about 30 percent of people account for 90 percent of all of the spending.

If the example were broadened to include a much larger group, the extremes of the distribution would become more evident. A few people would have medical expenses of several hundred thousand dollars, while many others would have no medical claims. The characteristics of the distribution, however, would be similar to those shown in Figure I.62 Indeed, some health economists have estimated that 1 percent of the population as a whole accounts for 30 percent of the spending and 10 percent accounts for 70 percent.63 For private insurance, 1 percent apparently can account for 50 percent64 and the concentration remains even if considered for as long as a decade.65

Now imagine that these 50 people are allowed to purchase health insurance in a competitive market. If they are all in similar health at the time they insure, then the premium will tend to be the same for each of them. If we ignore administrative costs, an actuarially fair premium would be $1,200 - since, on the average, they are expected to spend $1,200. In this example, we discovered after the fact that eight people were unprofitable because they incurred expenses in excess of $1,200. By contrast, 42 people turned out to be profitable because their expenses were less than $1,200.

"About 4 percent of the patients spend half the health care dollars."

Under managed competition, people would have an option they do not have in competitive markets. They would be able to reselect among health plans after they got sick and still pay the common premium of $1,200. Thus, under managed competition both the insured and the insurers would be allowed to recontract in full knowledge (for the most part) of who the eight unprofitable people were.

When Sellers Compete to Avoid Buyers.

"Instead of competing to attract people who are sick, health plans would compete to avoid them."

In the example illustrated in Figure I, eight people paid premiums equal to $9,600 and consumed medical services worth $54,000. The health plans that attracted them incurred a loss equal to $44,400 (again, ignoring administrative costs). By contrast, 42 profitable people incurred only $6,000 of medical expenses but paid $50,400 in premiums. The plans that attracted these people realized a profit of $44,400. Under managed competition, each health plan would have incentives to compete vigorously for the profitable 42 while avoiding the unprofitable eight. Competing would be easier if the plans knew who the eight unprofitable people were. But even if they did not, they could successfully avoid them so long as the eight knew. People who know before they select an insurer that they need expensive medical treatment will use this knowledge in selecting a health plan. And since the insurers understand this, they can structure their product so as to discourage the eight. Let’s look at some ways this might happen.

How Perverse Incentives Affect the Behavior of Buyers.

Take a heart patient in need of cardiovascular surgery. The individual has a self-interest in finding out who the best cardiologist is and what heart clinic has the best services. Armed with this knowledge, the patient will try to learn which health plan employs the cardiologist or has a contract with the clinic. The premium charged is unlikely to matter very much, since the patient knows that the price of his care will exceed his premium many times over.

Notice that this behavior is different from the likely behavior of a healthy person choosing a health plan. Healthy people must consider all the illnesses they might have and all the medical attention they might need. Since their probability of needing any particular service such as heart surgery in the near future is very small, they are unlikely to spend much time investigating particular doctors and clinics. To the degree that healthy people investigate the services offered by competing health plans, they are likely to inquire only about primary care services such as vaccinations, mammograms and general checkups - since these are the services they are likely to receive.

"Heart patients would search for plans with the best heart doctors; cancer patients would seek out the best oncologists."

Another consideration is likely to influence the choices of healthy people. Under managed competition, people can change plans if their health condition changes. Thus, if the need for heart surgery arises, odds are that the patients will have an opportunity to switch insurers before the surgery is performed. They can wait until the need arises before choosing an insurer based on its cardiology services.

The bottom line is: The people who carefully compare the heart services offered by competing managed care programs are likely to be the people who intend to use the services and that insurers will want to avoid. By contrast, people who choose a plan based on the quality and accessibility of nonacute services are more likely to be healthy.

How Perverse Incentives Affect the Behavior of Sellers.

Now it might seem that the second HMO could successfully compete by offering more primary care services. But in order to be truly competitive, it must change its strategy completely. The easiest way to keep costs down is to have only healthy enrollees. And the easiest way to accomplish that feat is to have none of the doctors and facilities that sick people want. As Alain Enthoven has noted (disapprovingly), "A good way to avoid enrolling diabetics is to have no endocrinologists on staff in the county. A good way to avoid cancer patients is to have a poor oncology department."67

To successfully compete for profitable enrollees, a different strategy is needed. It might work something like this: In order to attract healthy enrollees, a health plan might offer services that are not attached to a risky event but are simply consumption items. For example, the plan might offer preventive medical services, free or relatively inexpensive diagnostic tests and access to a health club. The plan also might offer services at more convenient times and locations, free parking and other amenities. Of course, these services might be attractive to all potential applicants, but they are more likely to be decisive for healthy people. People who are sick tend to put more value on other services.