Imagine a system in which health plans offer networks of doctors and hospitals in return for fixed premiums. People who are seriously ill and need specific, expensive medical treatment will select in a very different way from other people. As I discuss in my book Priceless: Curing the Healthcare Crisis, take a heart patient in need of cardiovascular surgery. The individual has a self-interest in finding the best cardiologist and the best heart clinic. Armed with this knowledge, the patient will try to learn which health plan employs that cardiologist or has a contract with that clinic. The premium is of secondary importance, since the value to the patient of receiving the best cardiovascular care will far exceed any premium payment.
The incentives facing healthy people are different. Since their probability of needing any particular service in the near future is small, they are unlikely to spend much time investigating particular doctors and clinics. To the degree that they do investigate, they are likely to inquire only about the primary care services they are likely to receive.
Thus, the people who carefully compare the acute care services offered by competing health plans are likely to be the people who intend to use them. These are the very people health plans want to avoid. By contrast, those who choose a plan based on the quality and accessibility of nonacute services are more likely to be healthy.
As Alain Enthoven has noted (disapprovingly), “A good way to avoid enrolling diabetics is to have no endocrinologists on staff. . . . A good way to avoid cancer patients is to have a poor oncology department.”
To see how managed competition affects the incentives of insurers, imagine two competing HMOs. In the first, enrollees can see a primary care physician at any time, but there are cumbersome screening mechanisms and waiting periods for knee and hip replacements, heart surgery, and other expensive procedures. In the second, joint replacements and heart surgery are available when needed, but primary care facilities are limited. Given a choice, most of us would enroll in the first HMO if we were healthy and switch to the second if we had a serious health problem. But if everyone acted in this way, the second HMO would attract only expensive-to-treat patients. It might seem that the second HMO could compete successfully by offering more primary care services. But to be truly competitive, it would have to change its strategy completely. To cover its costs, it would have to charge a premium many times higher than the first HMO. The premium would have to equal the cost of the expensive procedures, but few people could afford the premium, and they might be better off to simply buy their medical care directly. In any event, the HMO would face financial ruin.
A survey by the Kaiser Family Foundation discovered how HMOs were competing for seniors on Medicare. The HMO ads in print and on television showed seniors snorkeling, biking, and swimming but did not feature the sick or disabled. In addition, nearly one-third of HMO marketing seminars were held at sites that were not wheelchair accessible. The following are just a few other examples of how managed competition works, uncovered by The Washington Post:
The term medlining is sometimes used to describe the practice of avoiding the sick. It is healthcare’s version of redlining, the banking and insurance practice of avoiding deteriorating neighborhoods. The other side of the coin, of course, is attracting the healthy. In addition to health club memberships, health plans also have offered dental benefits and vision care. The theory is that anyone who will switch health plans to get a free pair of eyeglasses cannot be very sick.
 Alain Enthoven, “The History and Principles of Managed Competition,” Health Affairs (1993 Supplement): 35, doi: 10.1377. On the practice of encouraging high-cost patients to “disenroll,” see Jonathan E. Fielding and Thomas Rice, “Can Managed Competition Solve the Problems of Market Failure?” Health Affairs (1993 Supplement) 222; Joseph Newhouse, “Is Competition the Answer?” Journal of Health Economics 1 (1982): 109–116.
 The HMO would receive premiums only from people who were about to undergo expensive medical procedures. Thus, the average premium would have to equal the average cost of the procedures. It is precisely because most people cannot easily bear such a financial burden that health insurance is desirable in the first place.
 Reported in Natalie Hopkinson, “Study Finds Medicare HMOs Target Active Seniors but Not Disabled in Ads,” Wall Street Journal, July 14, 1998.
 David Hilzenrath, “Showing the Sickest Patients the Door,” Washington Post, National Weekly Edition, February 2, 1998.