|Home | About | Journals | Submit | Contact Us | Français|
Several myths about health insurance interfere with the diagnosis of problems in the current system and impede the development of productive reforms. Although many are built on a kernel of truth, complicated issues are often simplified to the point of being false or misleading. Several stem from the conflation of health, health care, and health insurance, while others attempt to use economic arguments to justify normative preferences. We apply a combination of economic principles and lessons from empirical research to examine the policy problems that underlie the myths and focus attention on addressing these fundamental challenges.
Several common myths about the benefits and design of health insurance undermine the development of a productive conversation on reform efforts. These misunderstandings both interfere with the diagnosis of problems in the current system and impede the development of a much-needed bipartisan consensus on how to engineer reform. Although many of the myths are built on a kernel of truth, advocacy for addressing real problems often simplifies complicated issues to the point of being false or misleading. In this paper we evaluate these myths using a combination of economic principles and lessons from careful empirical research. The misconceptions often arise from genuine policy concerns, and we hope that stripping them away will promote reforms that focus on the underlying challenges facing the U.S. health system.
Our choice of which misconceptions about health insurance to address is idiosyncratic to our experience. These misconceptions are pervasive enough that pointing to specific instances may be counter productive. Rather, we prefer to draw attention to the genuine underlying policy challenges. A common feature of several myths is the conflation of health, health care, and health insurance. The three are surely connected, but they are not the same. Others stem from attempts to use economic arguments to justify normative preferences. Our discussion is meant to give an economist’s point of view, rather than to introduce new analysis or to provide a comprehensive treatment of any of these important topics. We begin by discussing what health insurance is and is not, and then discuss five myths about health insurance in the United States.
Insurance, in its simplest form, works by pooling risks: many pay a premium up front, and then those who face a bad outcome (getting sick, being in a car accident, having their home burn down) get paid out of those collected premiums. The premium for health insurance is the expected cost of treatment for everyone in the pool. The key insight is that not everyone will fall sick at the same time, so it is possible to pay for the care of the sick even though it costs more than their premiums. This is also why it is particularly important for people to get insured when they are healthy—to protect against the risk of needing extra resources to devote to health care if they fall ill.
Uncertainty about when we may fall sick and need more health care is the reason that we purchase insurance—not just because health care is expensive (which it is). Lots of other things are expensive, too, including housing and college tuition, but we don’t have insurance to help us purchase them because they are not uncertain in the way that potentially needing very expensive medical care is. The more uncertainty there is, the more we value insurance.
Insured sick people and uninsured sick people present very different public policy challenges. People who have already purchased insurance and then fall sick pose a particular policy problem: insurance is not just about protecting against unexpected high expenses this year, but is also about protecting against the risk of persistently higher future expenses in the case of chronic illness. With this kind of protection, enrollees’ premiums would not rise just because they got sick, but this is not always the case today. In fact, insurers have an incentive to shed their sickest enrollees, which suggests a strong role for regulation in protecting such enrollees. Nor are insurers held responsible when inadequate coverage raises the costs for a future insurer, such as Medicare for those over age sixty-five. These problems highlight the limited availability of true long-run insurance offerings, a reform issue that is often glossed over in the confusion between health care and health insurance.
Uninsured Americans who are sick pose a very different set of problems. They need health care, not health insurance. Insurance is about reducing uncertainty in spending. It is impossible to “insure” against an adverse event that has already happened, for there is no longer any uncertainty about this event. (Insurance could still cover the uncertainty of other changes to health, but not this pre-existing condition.) Try purchasing insurance to cover your recent destruction of your neighbor’s Porsche: the premium would be the cost of a new Porsche. You wouldn’t need car insurance—you’d need a car. Similarly, uninsured people with known high health costs do not need health insurance—they need health care. Private health insurers will not charge uninsured sick people a premium lower than their expected costs. The policy problem posed by this group is how to ensure that low-income uninsured sick people have the resources they need to obtain what society deems an acceptable level of care—and ideally, as discussed below, how to minimize the number of people in this situation.
This highlights one of the many reasons that health insurance is different from car insurance: the underlying good, health care, is viewed by many as a right. Furthermore, we may want to redistribute money from the healthy to the (low-income) sick, in the same way that we redistribute money from the rich to the poor. This kind of redistribution is fundamentally different from private insurance—it is social insurance, and it is hard to achieve through private markets alone.1 Private markets can pool risk among people starting out with similar health risks, and regulations can ensure that when some members of those risk pools fall ill, insurers cannot deny them care or raise their premiums. Transferring resources from lower-health-risk groups to higher-health-risk groups, however, requires social insurance. There is a distinction between the public provision of a good and the public production of a good: social insurance may or may not be “socialized.” For example, providing subsidies for individuals to purchase private insurance or providing the insurance directly (as through Medicare) are both forms of social insurance.
How then do we provide the sick and uninsured with socially acceptable care? For starters, it would help to understand that unregulated private health insurance markets are unlikely to deliver this goal: no insurer will be willing to charge a premium less than enrollees’ likely health costs. Instead, they could be given health care directly or a premium subsidy equal to their expected health care costs. Alternatively, we could force sick people and healthy people to pool their risks, such as through community rating coupled with insurance mandates (to preclude healthy people from opting out of subsidizing sick ones). But such pooling implies a transfer from healthy people to sick people, and consequently is based on normative preferences about redistribution.
The advantage of social insurance programs, including a nationalized health care system, is that they can achieve redistribution that private markets alone cannot. They may also provide benefits with lower administrative costs (although, in the case of moving to a single-payer system, the size of administrative savings relative to overall health care cost growth is likely to be small).2
There are, of course, also costs associated with social insurance programs. First, there is the drag on the economy imposed by raising revenues to finance them. Second, there is the loss of competition and the inability to offer diverse insurance options for patients with diverse preferences. Third, social insurance programs may reduce innovation (because of administered prices or constraints on covered services). We acknowledge that it is difficult to quantify the size of these costs, but they make social insurance programs more expensive and less efficient—and thus impose an even larger burden on already strained public budgets. These fiscal pressures have, perhaps unsurprisingly, spawned additional myths that imply that social insurance programs have lower costs and higher benefits than a more detached analysis would suggest. We turn to some of these myths next.
This is a common and deceptively appealing argument for expanding insurance coverage: we could spend less and get more, and who could be against that? But as with most prescriptions that promise something for nothing, this misconception finds little empirical support.
Yes, emergency room (ER) care for the uninsured is inefficient and might have been avoided through more diligent preventive care and disease management. Diabetes treatment is a good example; it is much cheaper to manage diabetes well than to wait for a hospitalization that requires a leg amputation. Having health insurance may lower spending on ER visits and other publicly provided care used by the uninsured through better prevention and medical management. But empirical research also demonstrates that insured people use more care (and have better health outcomes) than uninsured people do—so universal insurance is likely to increase, not reduce, overall health spending.3
Why does insurance cause greater consumption of health care? Insurance, particularly insurance with low cost sharing, means that patients do not bear the full cost of the health resources they use. This is a good thing—having just made the case for the importance of the financial protections that insurance provides—but it comes with the side effect of promoting greater consumption of health resources, even when their health benefit is low.
This well-documented phenomenon is known as “moral hazard,” even though there is nothing moral or immoral about it. The RAND Health Insurance Experiment (HIE), the largest experiment in social science, measured people’s responsiveness to the price of health care. Contrary to the view of many noneconomists that using health care is unpleasant and thus not likely to be responsive to prices, the HIE found that it was responsive: people who paid nothing for health care used 30 percent more care than did those with high deductibles.4 This is not done in bad faith: patients and their physicians evaluate whether the value of the care exceeds the out-of-pocket costs, rather than the higher total costs.
The increase in care that individual patients use because of insurance has even greater systemwide ramifications. Research and development (R&D) in new medical technologies responds to the changes in aggregate incentives driven by health insurance. There is evidence of these systemwide effects from the introduction of Medicare in 1965: providers made spectacular investments in high-tech care, and hospital spending surged more than 25 percent in five years.5 These technologies may improve welfare, but they also raise premiums because of the larger armamentarium of treatments available to the sick.
Contrary to another common assertion, even increases in preventive care do not usually pay for themselves: in general, prevention is good for your health, not your wallet. Some preventive care has been shown to be cost-saving—such as flu vaccines for toddlers or targeted investments such as initial colonoscopy screening for men ages 60–64—but most preventive care results in greater spending along with better health outcomes. Indeed, some money spent on preventive care may not only cost money but may be no more cost-effective than some “high-tech” medical care. For example, screening all sixty-five-year-olds for diabetes, as opposed to screening only those with hypertension, might improve health but would cost much more than the standard estimates of our willingness to pay for improved health (at about $600,000 per quality-adjusted life-year, or QALY, compared to a standard value of $100,000), which suggests that that money might be better spent elsewhere.6
All of this suggests that insuring the uninsured would raise total spending. This does not mean that it would not be money well spent (we personally believe that it would be, but this reflects a normative preference). Spending more to extend insurance coverage is not a problem if it generates more value than it costs, and the view that health care is a right is not inconsistent with this framework. First, and sometimes overlooked, is the security that insurance provides against the uncertainty of unknown health care expenses. The value of this financial smoothing alone is estimated to be almost as much as the cost of providing people with insurance.7 Second, much of the additional health care that the newly insured would receive is clearly likely to improve their health. (But this is by no means automatic; as we discuss below, being insured is not enough to guarantee good health care.) Extending health insurance coverage may be well worth the cost for these reasons, but it would not save money.
Having insurance may increase the quantity of care you get, but it is no guarantee of getting high-quality care. A recent study found that Americans received less than 60 percent of recommended care, including preventive, acute, and chronic care—and including such low-cost interventions as flu vaccines and antibiotics for surgical patients.8 Thus, although gaining insurance would likely improve outcomes for the uninsured, it is by no means the only reform the health system needs.
“The likelihood of getting high-quality care has more to do with geography than with insurance status or spending.”
An exclusive focus on the uninsured often presupposes that the insured are receiving high-quality care, equating higher spending and higher quality. Yet the likelihood of getting high-quality care has more to do with geography than with insurance status or spending. Beginning with the work of John Wennberg at Dartmouth, an immense literature in medicine and economics has found that even among Medicare beneficiaries (all of whom are insured), there are enormous differences in the quality of care received: in fact, in areas where the most is spent on Medicare beneficiaries, they are the least likely to get high-quality care.9 Mammograms, flu shots, the use of beta-blockers and aspirin for heart attack patients, rapid antibiotics for pneumonia patients, and the use of simple lab tests to evaluate the management of diabetes are all lower in higher-spending areas.10 Higher spending is not even associated with lower mortality, which suggests that more-generous insurance provision does not necessarily translate to better care or outcomes.
When these results showing the lack of relationship between spending and quality were first reported, there were two predictable responses by skeptics: that high-spending areas had sicker patients who were (appropriately) less likely to receive these therapies, and that patients in high-spending areas had higher satisfaction even if their measurable health outcomes were the same. Neither claim is supported by the evidence. We should thus discard the simplified notion that more spending guarantees better care or even basic preventive care.
What, then, do patients in high-spending areas get? Evidence suggests that this higher intensity is driven by greater use of procedures of questionable clinical value. Patients in high-spending areas are no more likely than those in low-spending areas to receive surgery, but they see more specialists more frequently, have more diagnostic and imaging services, and get more intensive care at the end of life—none of which has been shown through clinical trials to improve health.11 “Coordination failures” in delivery may both raise costs and lower quality, even among the insured.
Insuring the uninsured will give them access to the sort of health care that everyone else receives: a combination of valuable care, overuse of some costly interventions with little proven benefit, and under use of some vitally important therapies—care that is sometimes coordinated but often fragmented. This is better than no care, but it highlights the problem of collapsing the entire debate about U.S. health care reform down to the issue of uninsurance: health insurance does not guarantee good health care.
Employees ultimately pay for the health insurance they get through their employer, no matter who writes the check to the insurance company. The view that we can get employers to shoulder the cost of providing health insurance stems from the misconception that employers pay for benefits out of a reservoir of profits. Regardless of a firm’s profits, valued benefits are paid for primarily out of workers’ wages.12 Workers may not even be aware of how much their total health premium is; however, employers make hiring and salary decisions based on the total cost of employment, including both wages and benefits such as health insurance, maternity leave, disability insurance, and retirement benefits.13 They provide health insurance not out of generosity of spirit, but as a way to attract workers—just like wages. When the cost of benefits rises, wages fall (or rise more slowly than they would have otherwise), leaving workers to bear the cost of their benefits in the form of lower wages.14
The uncomfortable arithmetic of this wage-fringe offset is seen in other contexts—for example, workers bear the costs of workers’ compensation, and mandated maternity benefits primarily reduce the wages of women of child-bearing age.15 When it is not possible to reduce wages, employers may respond in other ways: employment can be reduced for workers whose wages cannot be lowered, outsourcing and reliance on temp agencies may increase, and workers can be moved into part-time jobs where mandates do not apply. These adjustments are neither instantaneous nor dollar-for-dollar for every person (depending, for example, on wage rigidities, how much individuals value the insurance benefit, and how heterogeneous the employees’ incomes and health are)—a fact that obscures the underlying connection. This also means that the claimed connection between health care costs and the “international competitiveness” of U.S. industry is murky at best: higher health costs primarily lower current workers’ nonhealth compensation, rather than firms’ profitability (although the same trade-off cannot operate in retiree health benefits, making their effects more complicated).16
Why, then, do we have a private health insurance system based primarily on policies offered through employers? There is a preference in the tax code for premiums paid by employers relative to premiums paid by individuals or direct payments for health care. This tax preference drives both the predominance of employment-based policies and the prevalence of policies with low cost sharing, because care paid for in the form of higher employer premiums comes at a lower after-tax price than care paid for out of pocket. Of course, this tie between employment and insurance comes at a cost: workers who leave or lose a job risk losing their insurance or facing much higher premiums, sometimes forcing them to stay in a job to retain health insurance.17
This is not to say that there are not important advantages to getting insurance through an employer instead of through the individual insurance market (especially given the current state of the individual market), including better pricing and risk pooling. The employer market is the primary mechanism for maintaining cross-subsidization from low-risk populations to high-risk ones, with tax subsidies adding an element of social insurance (albeit one that is not particularly progressive).18 It is these benefits that are the main advantages of access to employer policies, not the fact that employers nominally pay part of the premium.
Greater patient cost sharing would help, but it is not the magic bullet that some make it out to be. It is certainly true that first-dollar insurance coverage (that is, insurance coverage for the first dollar of health care expenditures or insurance with very low cost sharing more broadly) encourages the use of care with very low marginal benefit and that greater cost sharing would help reduce the use of discretionary care of questionable value. But there is also evidence that patients under use drugs with very high value when confronted with greater cost sharing (whether because they lack resources or lack information). Worse, there is evidence that even $5–$10 increases in copayments for outpatient care can result in some patients’ being hospitalized as a result of cutting back too much on valuable care, partially offsetting the reduced spending.19 Capping total insurance benefits is also short-sighted and imprudent: not only does evidence suggest that such caps result in adverse clinical outcomes, worse adherence, and increased hospital and ER costs, but the presence of caps means that patients are not insured against catastrophic costs—exactly what insurance is supposed to protect against the most.
There is no reason to think that the optimal insurance structure would look like the typical high-deductible plan. Rather, it might subsidize high-value care such as treatments to manage diabetes or asthma, while imposing greater cost sharing on care of lower value, such as elective surgeries with limited health benefits. People would choose the insurance plans that offered them the best benefit mix—trading off higher premiums for plans that covered care of diminishing marginal value. Of course, what may be valuable to one patient could be wasteful for another, and the key challenge for “value-based insurance design” policies is to differentiate between these cases. Firms experimenting with these plans include the University of Michigan, Pitney Bowes, Aetna, and Active Health Management.20 We believe that focusing exclusively on high-deductible plans that rely on a blunt structure of patient cost sharing and perfectly forward-looking patients may forestall the development of even more innovative plans.
This does not mean that competition and cost sharing have no role in promoting higher-value spending, however. Competition between insurers to offer plans that have the mix of benefits that enrollees find most valuable could drive the kind of innovative plans described above. More sophisticated cost sharing could greatly improve the value of care. As the evidence from the RAND HIE discussed above shows, the low-cost-sharing plans fostered by the current preferential tax treatment of health insurance (which look more like prepaid health care than true insurance) promote the use of care that is of limited health benefit. Although most spending is indeed done by people with very high total costs, well-designed cost-sharing programs could still have substantial effects on spending decisions. Most spending is not done in emergency settings, and even limited cost sharing can affect a sizable share of total spending.21 This suggests that carefully designed incentives could have a big effect on improving the value of care delivered.
We know that our health care system is not delivering the consistently high-quality, high-value care that we should expect. There are many open questions in the design of the ideal system; however, with millions of uninsured people and rising costs threatening to swamp public and private budgets alike, we cannot afford to wait to act. This does not justify the adoption of simple remedies that promise to cure the system at no cost. Such promises may appeal to nonexperts, but support is likely to evaporate as soon as more realistic cost estimates are made.
Focusing on the kernels of truth that underlie the myths that we have discussed suggests that the fundamental problems facing our health insurance system are unlikely to be cured by the extremes of either a single-payer system or an unfettered marketplace. On the one hand, the unregulated marketplace is unlikely to provide long-run stable insurance. Private insurers will always have an incentive to try to shed their highest-cost enrollees, so without regulatory safeguards, even the insured sick will be at risk of losing the insurance protections to which they are entitled. Private insurance fundamentally cannot provide the kind of redistribution based on underlying health risk or income that social insurance can.
On the other hand, a single-payer system does not automatically provide high-quality care: the provision of low-value care is as pervasive in the single-payer Medicare system as it is elsewhere. Single-payer systems are also slow to innovate—as suggested by the fact that it took Medicare forty years to add a prescription drug benefit, long after most private insurers had done so. Nor do calculations of the costs of a single-payer system measure the utility loss from forcing people with different preferences into a monolithic health insurance plan, or the cost to the economy of raising taxes. The private facilities that have sprung up in Canada to meet the demands of those who want more health care than the public system provides fundamentally undermine the “single-payer” nature of the system. If the goal of policy is to direct subsidies to people with high health costs, a single-payer system is just one of many alternatives—all of which represent normative preferences that are hard to justify through improved efficiency or reduced costs.
How voters balance such trade- offs is likely driven as much by philosophy and politics as economics, and any reform will involve tough choices between competing values. Well-intentioned prescriptions with a “no pain, all gain” outlook are easy to sell, particularly in a political climate shy of nuanced discussion, but they are unlikely to produce miraculous cures.
Serious health reforms would focus on increasing the value that we get from health insurance and health care, instead of counting on self-financing expansions of coverage. Reforms that promoted higher-value insurance could both extend coverage so that more people benefit from the protections that insurance affords and ensure that those protections are secure for those who fall ill. These reforms would not be enough to achieve uniformly high-quality care, however. The frequent failure of the use of best practices and the tremendous geographic variation in the use of costly care of uncertain medical benefit are often obscured in the focus on the uninsured.
That many nations, including both the United States and Canada, struggle with these challenges suggests that reforms of the payment system alone are unlikely to solve all of these problems. We believe that a comprehensive reform proposal that aimed both to extend insurance protections to those who lack them and to improve the value of care received by those who are insured would be more likely to succeed at each goal than proposals that focused on just one.
The authors thank six anonymous reviewers, Don Metz, Ashish Jha, Alan Durell, Michelle Mello, Chris Jencks, Seth Seabury, Michael Chernew, Elliott Fisher, Jonathan Skinner, and Douglas Staiger for deeply insightful comments. The opinions in this paper are those of the authors and not those of any institution with which they are affiliated. Each blames the other for any remaining errors.