[An updated version of this article can be found at Health Care in the 2nd edition.]
Health care is one of the most rapidly growing sectors of the economy, with expenditures in 1988 totaling $539.9 billion, or 11.1 percent of GNP, up from 5.3 percent of GNP in 1960. For 1992, health care spending was projected to be over 13 percent of GNP. [Editor's note: this article was written in 1992.] This amounts to $2,124 per capita, of which $1,882 was for personal health care—medical services and supplies for individuals. The rest was for research, construction, administration, and public health activities. Hospital expenditures accounted for 39 percent of the total spent for personal health care, physician services for 19 percent, nursing home care for 8 percent, and other personal health care for 22 percent.
Is Health Care Different?
Health care differs from other goods and services in important ways. The output of a shoe factory is shoes. But the output of the health care industry is less well defined. It is unpredictable and imperfectly understood by producers, and still less by consumers. Also, third-party payment and government intervention are pervasive. None of these characteristics is unique to health care, but their extent and their interaction are. Nevertheless, health care markets obey the fundamental rules of economics, and economic analysis is essential in appraising public policy.
The ultimate output of medical care is its effect on health. This effect can only be assigned probabilities before the care is provided and is difficult to measure even after the fact. Medical care is not the only determinant of health; others include nutrition, exercise, and other life-style factors. Efficient allocation of private and public budgets to health requires equating marginal benefit and marginal cost for each of these inputs (see Marginalism).
Risk and insurance. The risk of illness naturally leads people to demand health insurance. But in the United States the demand for health insurance is distorted by the fact that employer contributions are tax-exempt compensation to employees (see Health Insurance). This implies an open-ended subsidy at the employee's marginal tax rate, including income and payroll taxes at the federal and state levels. This "tax expenditure," which does not appear in any government budget, was estimated at over $50 billion in 1990. Assuming an average marginal tax rate of 33 percent, this subsidy more than offsets the administrative expense built into insurance premiums. Consequently, the average employee is better off insuring even routine medical services.
Since 1960, third-party payment for health care has increased dramatically. The share paid directly out of pocket by consumers fell from 49 percent in 1960 to 21 percent in 1988. At the same time, public financing increased from 24.5 percent to 42.1 percent, and private health insurance increased from 22 percent to 32 percent.
Pervasive third-party payment profoundly affects the structure of the medical care industry and the quantity, cost, and quality of services offered. Because insurance companies pay a large percent of the cost of medical care, the insured consumer's point-of-purchase price is necessarily lower. If the doctor charges forty dollars and the insurance company pays 80 percent, for example, the consumer's price is only eight dollars. As in any market the quantity demanded increases when price falls.
A five-year randomized trial of alternative insurance plans for the nonelderly population conducted by the Rand Corporation found strong evidence of the responsiveness of demand to insurance coverage. Some patients in the experiment were given totally free care. Others were required to pay 95 percent of the cost of medical services, up to a stop-loss. (A stop-loss is a limit on out-of-pocket expenses. In the experiment, it was set at 5 to 15 percent of income, up to a maximum of a thousand dollars in 1976 dollars.) Beyond the stop-loss they too received free care. Total expenditures for the group given free care were 45 percent higher than for the group that paid 95 percent up to the stop-loss. Free care increased total expenditures by 23 percent relative to a plan in which patients made a 25 percent copayment up to a stop-loss. For the great majority of participants, the difference in expenditures had no measurable effect on health, whether judged by objective measures or by the patients themselves.
Far harder to measure is the effect of insurance on technological change and on the "quality" of services available. Insured consumers (or physicians on their behalf) have incentives to use any new medical service if the expected benefit exceeds their private out-of-pocket cost, which is less than the full social cost. Thus, medical technologies can be profitable even if their expected benefits are below their cost. Overinvestment in technology is reinforced by provider incentives to compete on nonprice dimensions of service in markets where consumers are insulated from prices. In recent years third-party payers have become more aggressive as cost-conscious purchasers on behalf of insured consumers (see below), but the tensions remain.
Asymmetric information. Consumers typically have less information than providers do about the risks and benefits of alternative treatments, and therefore rely on physicians to advise as well as treat them. Such mixed roles are common in many professional and other contexts. They are, however, more complex in medical care because the provider is an agent not only for the individual patient but also for the third-party payer, who in turn is ultimately the agent for policyholders/patients as a group. Each individual patient would like to consume any service that has any expected benefit at all if the out-of-pocket cost is zero. But in the long run patients as policyholders are better off if insurers control moral hazard (the increase in quantity and "quality" of services in response to insurance) because insurance premiums must ultimately rise to cover the costs.
Insurers compete by devising better ways of controlling moral hazard. Thus, devising contractual incentives for providers to make the right trade-offs between the short-run desires of individual patients and long-run insurer/policyholder interests is at the heart of the ongoing revolution in health care markets, both in forms of reimbursement and organizational structure. Preferred provider organizations (PPOs), health maintenance organizations (HMOs), and various forms of managed care give doctors incentives to control insurance-induced overutilization.
Government. Government is more pervasive in health care than in almost any other industry, though less so in the United States than in most other developed countries. Such interventions are rationalized on grounds of assuring either access or quality. Government is the largest insurer, through Medicare and Medicaid, and public hospitals act as provider of last resort for those who cannot pay for care. Licensure, accreditation, and other regulations either directly or indirectly affect entry of physicians, dentists, and other medical professionals, as well as hospitals, nursing homes, and other institutional providers. New pharmaceuticals and medical devices must first be approved by the Food and Drug Administration.
The Growth in Costs—Why, and Is It Worth It?
Health care expenditures as a percent of GNP have grown more rapidly in the United States than in other countries. How much value we get for these expenditures and whether governments should further intervene to control costs have become major issues in public debate.
The growth in real health care costs per capita, net of economy-wide inflation, can be split into three components: medical price increases (in excess of other prices); increases in volume of services per capita; and increases in intensity of resource use per unit of service. Intensity reflects changing technology, "quality," and other factors that make any given service, such as a diagnostic test, more resource-intensive than it was in the past. In practice it is virtually impossible to construct quality-adjusted and technology-adjusted price indexes that meaningfully separate pure medical price increases from increases in intensity. Moreover, even an accurate accounting for cost growth does not illuminate the underlying causes.
Nevertheless, technology appears to be the single most important factor driving health care costs currently. A standard economist's presumption, based on theory and evidence, is that technology is not introduced unless it produces benefits at least as great as the costs. This presumption does not necessarily apply to new medical technologies, however. The reason is that massive government subsidies, directly through tax-funded government insurance programs and indirectly through the tax subsidy to private health insurance, cause medical providers to use technology that consumers may value less than the cost.
Although other countries with more centralized government control over health budgets appear to have controlled costs more successfully, that does not mean that they have produced a more efficient result. In any case, reported statistics may be misleading. Efficient resource allocation requires that resources be spent on medical care as long as the marginal benefit exceeds the marginal cost. Marginal benefits are very hard to measure, but certainly include more subjective values than the crude measures of morbidity and mortality that are widely used in international comparisons.
In addition to forgone benefits, government health care systems have hidden costs. Any insurance system, public or private, must raise revenues, pay providers, control moral hazard, and bear some nondiversifiable risk. In a private insurance market such as in the United States, the costs of performing these functions can be measured by insurance overhead costs of premium collection, claims administration, and return on capital. Public monopoly insurers must also perform these functions, but their costs tend to be hidden and do not appear in health expenditure accounts. Tax financing entails deadweight costs that have been estimated at over seventeen cents per dollar raised—far higher than the 1 percent of premiums required by private insurers to collect premiums.
The use of tight physician fee schedules gives doctors incentives to reduce their own time and other resources per patient visit; patients must therefore make multiple visits to receive the same total care. But these hidden patient time costs do not appear in standard measures of health care spending.
Both economic theory and a careful review of the evidence that goes beyond simple accounting measures suggest that a government monopoly of financing and provision achieves a less efficient allocation of resources to medical care than would a well-designed private market system. The performance of the current U.S. health care system does not provide a guide to the potential functioning of a well-designed private market system. Cost and waste in the current U.S. system are unnecessarily high, because of tax and regulatory policies that impede efficient cost control by private insurers, while at the same time the system fails to provide for universal coverage (see below).
Industry Structure and Competition
Despite barriers to entry, the health care industry has become extremely competitive in recent years. This is because of the large number of firms in most market segments, a more aggressive role of public and private payers in attempting to control costs, and antitrust enforcement.
Hospitals. Prior to the eighties hospitals were paid largely on the basis of costs incurred. In 1983 Medicare introduced a system of "prospective" payment according to diagnosis-related groups (DRGs), whereby hospitals are paid a fixed fee per admission, based on the patient's diagnosis. In contrast to retrospective cost-based reimbursement, the hospital bears the marginal cost of all expenses incurred. In addition, employers and private insurers also have ceased to be passive payers; now, they actively attempt to control price and utilization through such strategies such as HMOs, selective contracting with PPOs for fixed, discounted fees, utilization review, and required second opinions. These attempts to reduce costs have been effective. Since 1981 the number of hospital admissions and the average length of stay have declined for both the over-sixty-five and under-sixty-five population, and average hospital occupancy fell from 75.9 percent in 1980 to 64.5 percent in 1988, despite a reduction in the number of beds. Changing technology has also contributed to the decline in length of stay, but aggressive buyers have certainly played a role.
The categories of customer for whom hospitals must compete have also increased. Traditionally, hospitals competed primarily for physicians who, as independent contractors with admitting privileges at multiple hospitals, have critical influence over the volume and cost of hospital admissions. Now, hospitals must also compete for contracts with third-party payers who restrict their policyholders' choice of facilities, and must market directly to patients, particularly for elective services, where patients choose the hospital. Moreover, technological advance has increased the number of surgical and major diagnostic procedures that can be performed on either an inpatient or outpatient basis. Hospitals, therefore, also compete with ambulatory surgery and diagnostic centers.
Physicians. The number of physicians active in patient care almost doubled from 237,500 in 1965 to 455,700 in 1987, or from 124 to 189 physicians per 100,000 population. This increase reflects the response of medical schools to federal subsidies introduced to increase the supply of physicians after the introduction of Medicare and Medicaid in 1965.
In competitive markets an increase in supply is expected to lead to lower prices and, hence, increased quantity. Total expenditures may increase or decrease, depending on whether demand is elastic or inelastic. Many commentators express concern that, in the medical context, more physicians means increased volume of "supplier-induced" services, rather than price reductions.
The evidence on this issue is mixed and likely to remain inconclusive. Many physicians have moved to rural areas that were previously unserved. Presumably, they would not have done so if they had unlimited ability to induce demand in cities. The increased willingness of physicians to accept alternatives to unconstrained fee-for-service payment is also consistent with increased competitive pressures. And many physicians have agreed to capitation (a fixed payment per patient per month, that puts the physician at risk for volume of services) and fixed-fee arrangements with utilization review. There is a persistent positive correlation between number of physicians per capita and frequency of physician visits or surgical procedures. While this is consistent with supplier-induced demand, it is however, also consistent with the commonsense idea that physicians tend to locate in areas where demand for their services is high.
Since the abolition of the antitrust immunity of physicians and other professions, antitrust has been applied to challenge such activities as maximum price schedules, preferred provider organizations, peer review, and denial of staff privileges. Similarly, antitrust has been applied to hospital mergers and contractual arrangements with physicians, medical supply companies, and insurers. Such cases require a delicate balancing of the need to protect against anticompetitive practice while at the same time permitting the contractual freedom needed to effectively control costs and quality in a market with pervasive insurance and asymmetric information.
Government intervention in the health care sector typically addresses either quality or access. Regulations to assure minimum quality can potentially enhance efficiency in markets with asymmetric information, infrequent purchase, and potential for catastrophic mistakes. But often the regulations take the form of licensing, which limits entry and therefore limits competition. For some professionals, replacing licensing with certification, so that consumers who want a minimum quality can be assured of it, might achieve quality control while interfering less with competition. Moreover, reputation and other market forces are increasingly powerful stimuli to quality (see Brand Names). As the market evolves in the direction of competition among alternative medical plans that compete on all dimensions of quality (including technology, amenities, and choice of providers) as well as price, the appropriate role of government in setting minimum quality standards should be reassessed.
Government intervention to assure access includes public insurance and government subsidies to hospitals and clinics. Economic theory generally concludes that government intervention to stimulate the consumption of particular commodities is undesirable. People will consume, without subsidy, what they regard as the optimal amount of various items. A subsidy for consuming hamburgers, for example, causes people to consume too many hamburgers relative to other goods. One exception to this rule is for goods whose consumption by some consumers confers benefits on others. If, for example, people would not get vaccinated against polio unless subsidized or required to do so, a strong case could be made for subsidizing or requiring polio vaccinations. But the great majority of health expenditures are now devoted to purely private services that benefit only the recipients of the services.
Another type of external effect, however, is often used to justify a public subsidy to health insurance for low-income individuals, rather than simply a cash transfer. Many view health care as a "merit" good. That is, people derive satisfaction from knowing that everyone has access to a minimum level of health care, and they therefore are unwilling to deny care to anyone in extreme need. Addressing this concern through private charity creates a free-rider dilemma: I have little incentive to give to the poor if I think you will take care of them. Moreover, the not-so-poor have little incentive to buy insurance if they expect to receive charity care should the need arise. In such cases government intervention can be efficient.
A strong public demand for some minimum medical safety net argues for assuring universal access to insurance, but this does not require public provision of either insurance itself or medical services. Universal access to insurance can be achieved through a system of vouchers with income-related subsidies. But subsidies may not be sufficient to assure that everyone buys coverage, unless the subsidies are set at a very high level. Subsidies high enough to induce everyone to buy insurance voluntarily would provide large benefits to those who would have bought coverage with a lower subsidy. And such subsidies would entail large dead-weight costs of raising tax revenues. Subsidies alone are therefore an inefficient means of assuring universal coverage.
If the policy objective is universal coverage, then the simplest and most efficient approach is to make coverage compulsory, with income-related tax credits if necessary to assure affordability. Placing the requirement to obtain coverage on the individual entails less distortion of labor markets than the more widely discussed alternative, of mandating that employers provide coverage for employees. Mandating that employers provide coverage is equivalent to imposing a fixed tax per worker. Because insurance is a fixed cost per worker, the implicit tax rate is higher on low-wage and part-time workers. Unless wage rates of such workers fall to offset the cost of insurance to the employer, employment opportunities must fall. The cost of employer mandates will therefore be borne largely by currently uninsured workers, many of whom are in low-wage jobs. Some costs may also fall on small employers. In either case, this approach to covering the uninsured entails unnecessarily high total costs.
By contrast, placing the requirement to obtain coverage on the individual does not preclude that employers provide insurance. Most insurance would probably continue to be obtained through employment because of the savings in administrative costs. But other sources of group insurance are more likely to develop than under the status quo, which distorts relative prices heavily in favor of providing insurance through employment. Despite the efficiency and equity arguments in favor of requiring individuals to obtain coverage, however, politicians favor employer mandates because the costs of such an approach, although higher, are largely hidden.
Patricia M. Danzon is Celia Z. Moh Professor of Health Care Systems and Insurance and Risk Management at the Wharton School in Philadelphia.
Danzon, Patricia M. "Hidden Overhead Costs: Is Canada's System Really Less Expensive?" Health Affairs (Spring 1992): 21-43.
Frech, H. Edward, III, ed. Health Care in America: The Political Economy of Hospitals and Health Insurance. 1988.
Manning, Willard, et al. "Health Insurance and the Demand for Medical Care: Evidence from a Randomized Experiment." American Economic Review 77, no. 3 (June 1987): 251-77.
Pauly, Mark, Patricia Danzon, Paul Feldstein, and John Hoff. "A Plan for 'Responsible National Health Insurance.'" Health Affairs (Spring 1991): 5-21.