Risky Selection: What does competition in the health sector really mean?
Much of the health care debate over the past couple of weeks has been fueled by the possibility that Obama would drop the so-called “public option,” and allow it to be replaced with a co-operative. Like many aspects of the debate over health reform, the “public option” has been fiercely defended and fiercely criticized, but little understood. This stems, I believe, from some fundamental misunderstandings about the nature of competition in the health sector. These misunderstandings are sustained by simplistic arguments about “markets” and “competition” that have been adopted from normal markets for consumer goods. But health care is not a consumer good, and it is characterized by “abnormal” markets. Both advocates and opponents of a government role in health care reform need to take the abnormalities of the health sector seriously if reform is going to work.
Let’s start with the common assumption that most people seem to have about competition among insurance plans. It is assumed, by advocates and critics alike, that a public option will increase cost pressures on private insurers. Advocates like this idea because they believe that private insurers make big profits and provide shoddy service. By putting cost pressure on them (in addition to certain regulations, like prohibiting discrimination against people with pre-existing conditions), these firms will be forced to reduce their overhead and improve their value for money to consumers. Opponents dislike this idea because they think the competition will be “unfair,” since government can subsidize an unprofitable “public option” to keep prices lower to consumers, thus driving private plans out of business. Everyone seems to agree, however, that the introduction of a public option and a health insurance exchange will increase competition and cost pressure on private plans.
This idea that competition will reduce costs seems so obvious, it is rare to hear anyone question it, and yet, there is no clear evidence to support it. In the first place, we should ask what we can say about competition in general in the health sector. For example, we have a system of private insurers competing right now (without a public plan), and they do not seem to be able to hold costs down. Why should that be? If competition lowers prices, why aren’t the existing health plans competing away profits and improving value for money for consumers?
There are several possible answers to this question. One is that there is really not that much competition in the insurance sector, because, in any one town, an individual does not have very many choices. In most states, one insurer dominates over 50 percent of the health care market. It follows that the introduction of a public option will not just tighten competition in some markets, it will create competition for the first time. To the extent this is true, we should recognize that the public option is actually an instrument of anti-trust within local markets.
But the structure of the American insurance market raises deeper issues. Why is it that local markets are unable to support more private competition in the first place? There are various theories about this problem. A popular view among conservatives is that regulations at the state level prevent insurers from competing across states, which undermines competition. There is probably some truth to the claim that such regulations dampen competition. However, it should be noted that even with such restrictions: states are pretty large markets in which monopoly/oligopoly can be observed even within sub-state regions, and big insurers like Wellpoint and United Health Group do sell policies in different states.
Another part of the answer, which in my view is more important, has to do with the way that insurance companies compete. Rather than compete on price, insurers have invested a great deal of money in “risk selection.” Risk selection simply means that insurers work hard to avoiding issuing a policy to people they expect to have high health care needs. If they fail to pre-screen people out, they can always deny coverage later for expensive treatments.
If one insurer successfully screens out people with pre-existing conditions, and those likely to incur high costs, this leaves an unattractive pool of residual clients for other insurers to serve. The highest cost patients are left behind, and a second insurer can only make money on this population by charging very high premiums that will cover the costs of patients likely to be very ill. But many people cannot afford such premiums, and, at any rate, high premium, high benefit plans tend to attract even sicker people, making the insurance unsustainable. (Remember that insurance has to pool risks to stay in business. Suppose it costs X/2 dollars on average to cover the costs of the population of uninsured. Once the first insurer has skimmed the low risks off the population, the average cost of covering the remaining population is X. Now suppose that a second insurer charges X. At this price, the insurer will tend to attract patients who are certain that their own expenses will be more than X. As a result, this insurer will end up covering a population whose average costs are actually higher than X, which is unsustainable.)
This is one reason the market for health care is not like the market for peanut butter. It is also the reason why the proposed “health insurance exchange” is supposed to change the nature of competition in the insurance market. To participate in the exchange, insurers will have to give up the practice of risk selection, and this should force them to take price and value for money more seriously.
Supposing that all insurance in America had to be sold through an exchange, and the rules forbid risk selection or differential premiums, then what would happen to the market? This is the key question. If insurance competition would “work” properly in this market, then there is no need for additional measures, such as a public plan. If the insurance market would not work, then additional measures are warranted.
The creation of a regulated exchange could lead more insurers to jump in to provide coverage, to compete away the profits of monopolistic or oligopolistic firms that control large market shares in local markets. In order for such firms to make money, they have to attract some of the healthier clients that have signed up with other firms, which they will do by lowering their premiums. This could lead to healthy competition.
However, we have been assuming something else about the health care market which is not necessarily true. When consumers choose a peanut butter, they can relatively easily try out a few different kinds of peanut butter and decide which one they like best, and how much they are willing to pay for that preference relative to other options. This is rarely true in health care. In the first place, the insurer is offering a bundle of services, not a single service. It is harder for consumers to distinguish between complicated bundles than between individual services. Second, since illness is, for most people, a rare event, it is unlikely that a patient has used many of the services offered by the insurance, or will know their likelihood of using them. Third, average people tend not to know much about the relative value of different treatments or even what they mean until they do get sick and have to make decisions. At that point, it is too late to be choosing an insurer.
The result of all of this is that patients tend to prefer insurance that covers desirable services, not all of which may contribute to their health. Desirable services may include things like private hospital rooms, new technology, and the reputation of institutions associated with the insurer. They also include reduced constraints on doctors to order tests or provide services, even if these tests and services are unnecessary. (This problem is known as “induced demand,” since patients demand services because the doctor indicates that they should, rather than because of their own intrinsic demand.)
It is not unreasonable for consumers to want these things, just as it is not unreasonable for me to want to live in a mansion instead of a cottage. But in an ideal market with consumer sovereignty and full information, these things would be balanced against actual information on consumer needs, health care quality, actual costs, and the true effectiveness of various covered interventions. Actual health care markets tend to lack this balance.
Thus the question arises whether competition among private health insurers will actually reduce costs. Many analysts would argue that such competition will not reduce costs, particularly if consumers are not forced to bear the bulk of these costs. Under our current system, insurance transfers the bulk of the expense to a third party (the insurer), and patients are protected by subsidies that transfer much of the remaining expense to their employers or the taxpayer. This is known as “moral hazard” by economists.
(Interestingly, neither the problem of moral hazard nor the problem of incomplete information has been resolved by the fact that much insurance is purchased by corporations for their employees. Corporations receive tax benefits that reduce the cost to themselves of providing care, which in turn induces some moral hazard on their part. And benefit managers at corporations also know little about what makes health care effective, which means they are not sovereign consumers either.)
Economists disagree about the relative importance of the lack of patient information/induced demand versus the moral hazard problem. (For differing points of view, see a debate about Singapore’s health system in this journal.) But they would tend to agree that unless at least one of these problems are solved, there is no reason to expect that competition among insurers will necessarily lead to reduced costs. Singapore’s own experience suggests that, even when moral hazard was reduced, there was a need for considerable additional government regulation to manage cost inflation. (Others have suggested that private competition cannot yield cost savings for another reason: providers are so powerful that they can dictate prices in local markets.)
Given the possibility that private competition will not yield reduced costs, it is perhaps not surprising that the Obama administration has introduced a complementary reform: the public option. But we must now consider whether or not the public option really could control health care costs. Should the introduction of a public competitor resolve the problems of information/induced demand or moral hazard?
There is no reason to expect the government to do anything about moral hazard. If the government plan followed the design that is currently being discussed, it would function just like private insurance (i.e., be based on premiums), in which case it would have no advantage with respect to moral hazard. If the plan functioned the way its critics intimate, meaning that it received subsidies from the government beyond those generated by premiums, it would introduce even more severe moral hazard. So under no scenario does the public option improve the moral hazard problem.
What a non-profit, public insurance plan could do is to create a cheaper and more effective insurance plan by overcoming individuals’ information problems and trying to provide coverage that was more cost effective, while reducing excessive and unnecessary care that largely lines provider pockets. Since the public option would not need to compete for profits like a for-profit insurer, it could take the high road and cover only really useful care, focusing on outcomes rather than flashy technology and private rooms that ultimately add a lot to costs, but little to health improvements.
But this immediately raises a question: if the market is skewed in such a way that patients prefer flashy technology and private rooms, won’t they be willing to pay more for them? And since the moral hazard problem has not been solved, they only bear a small fraction of the costs of these extra perks. And so, how exactly will the public option reduce costs?
The president, in defending the public option, has often used the example of the post office and FedEx to argue that the private sector will not have trouble competing with a public plan. His point is well-taken, but a better example of the kind of market likely to develop in health care would be that of higher education. In the United States, there are public universities offering in-state tuition at vastly reduced prices compared to private universities. No one would argue that these private universities are unable to compete. To the contrary, in spite of the big cost advantage of the public universities, they are the ones that struggle to compete for the best students. The market for higher education has become segmented, with the best students and those with money largely focused on private, elite institutions, while less successful students or those without money rely on public universities.
Why shouldn’t health insurance turn out the same way? If the public plan is perceived to be less desirable, because it provides less desirable non-medical services or less access to needless but attractive technology, people will tend to resort to it only when they cannot afford a “better” private plan. This could result in a variety of outcomes. One that already exists in the case of Medicare is that many people purchase complementary coverage, using Medicare for basic access, and a private Medigap plan for more extensive coverage. Alternatively, the market could be completely segmented, with some people relying entirely on public, and others on private, insurance. The result of this kind of segmentation would be that the public option’s primary purpose would be to provide insurance to everyone, but it would do little to control costs (except, perhaps, when compared to subsidizing everyone to buy costlier private insurance).
So if competition won’t reduce costs, what will? At various times, President Obama has alluded to a need for changes in the delivery system. This means changing how doctors are reimbursed, and how patients seek care. It seems likely that such changes are necessary to reduce costs, because private competition, with or without a public option, is insufficient. However, the issue of how to change the delivery system has not gotten much attention, in part because if we continue to rely on fragmented arrangements with providers, driven in part by our fragmented private insurance payment system, we cannot really do much to change the delivery system. Thus it is that the principal tool for “bending the cost curve” has gotten little attention in the current debate about reform, while a beguilingly simple debate on competition among insurers has largely missed the boat.
An alternative way to control costs is for the public option to take over such a large share of the market that it can negotiate lower fees for everything from provider care to technology. In other words, the principle force behind cost control would not be competition but monopsony power on the part of the government. It is fear (and hope) for this outcome that has led some to view the public option as a gateway to “single payer.” These analysts believe (rightly) that competition is not enough to bring down costs, and (more controversially) that the government will either desire or be forced to take over a larger share of the market to impose cost controls.
Is this what President Obama thinks? He has said that it is not. And proposed legislation would not lead to this outcome on its own. But the question of whether a public option is a gateway drug is ultimately one of trust. Either you believe that the government is intent on moving to single payer through subversive means, or you don’t. What can be said with some certainty is that there is no reason that a public option has to lead to single payer, because there are countries that have such an option and do not have single payer. The Netherlands, for example, provides public coverage to part of the population, while the wealthier third or so of the population pay for private coverage.
At the same time, the government could regulate the private insurance market in more effective ways, even without introducing a public option. This approach would draw on the experience of other nations as well. In addition to the policies under consideration already, such as forcing insurers to cover anyone and preventing them from charging a premium based on risk, it would be worth considering whether insurers should have to be non-profits. Non-profits, which have exclusive rights to the private insurance market in Switzerland, may have reduced incentives to provide costly but ineffective care, particularly if they do not face competition from for-profits. Converting our private market into a non-profit exchange would elicit a lot of opposition from for-profit insurers, but it would not be susceptible to claims about a government takeover and it could have an impact on the nature of health insurance competition.