Abstract
High and rapidly rising health care costs in the United States and growing ranks of uninsured persons have brought health care reform to the top of the U.S. Administration’s policy agenda. This paper describes the health care financing system in the United States, highlights what are viewed as its most serious shortcomings, and explores possible reasons for high and rising medical care costs. After brief descriptions of alternative reform proposals, the paper discusses universal coverage under managed competition and its ability to deal with the equity and efficiency problems in the U.S. health care system.
I. Introduction
High and rapidly rising health care costs in the United States and growing ranks of uninsured persons have brought health care reform to the top of the Administration’s policy agenda. The United States spent 13 percent of its gross domestic product on medical care in 1991, a larger percentage than any other industrial country. That share has been rising steadily over the last several decades and is projected to reach 18 percent by the year 2000 and 32 percent by the year 2030 unless government policies or private behavior are changed. 2/ At the same time, a sizable share of the U.S. population lacks any form of health insurance coverage.
Growth in U.S. health care expenditures is viewed with concern for two principal reasons. First, it is believed that the resource costs of providing health care are greater than the benefits generated in terms of improved health or patient satisfaction. 3/ In other words, the production of health care is believed to be inefficient. The performance of unnecessary procedures, excess capital investment, and the use of high-cost specialists instead of primary care providers are cited as examples of these inefficiencies. Second, it is believed that the distribution of health services is inequitable. Fourteen percent of the population lacks any form of health insurance and as much as another 20 percent is considered to have inadequate coverage. These inequities and inefficiencies are not new developments, and reform of the U.S. health care system has been under debate for some time. However, the pressure that escalating costs have and increasingly are expected to put on government and private budgets has heightened the urgency of health care reform.
Recent proposals for health care reform attempt to address these efficiency and equity issues. With regard to equity, a number of proposals include provision for universal health insurance coverage, although the extent of that coverage is subject to some debate. There is even less consensus regarding the best ways to eliminate inefficiencies in the health care market without impeding access, since health insurance that ensures access also creates incentives for the (inefficient) overconsumption of health care services. The major controversy in health care policy is whether adequate control over resource allocation is best achieved through market mechanisms (which tend to focus on the demand side of the market by increasing the cost-consciousness of consumers) or through regulatory controls on the supply side. 4/
This paper describes the financing of health care in the United States, highlights what are viewed as its most serious shortcomings, and draws some comparisons with other OECD countries (Section II). It then examines the special characteristics of the health care market (in particular the existence of health insurance) that may contribute to an excessive level of spending (Section III). Evidence regarding the various possible reasons for high and rising health expenditures in the United States is then presented (Section IV). Finally, after brief descriptions of alternative reform proposals, the paper discusses universal coverage under managed competition and its ability to deal with the equity and efficiency problems in the U.S. health care system (Section V).
II. Overview of the U.S. Health Care System
Although the health care system in the United States generally is viewed as providing high quality care, it is considered to suffer from a number of shortcomings: health care spending relative to GDP is high and rising rapidly and a significant percentage of the population lacks any form of health insurance, while many others are considered to have inadequate insurance. 5/
1. Health insurance in the United States
In contrast to most other industrial countries, the United States does not have a single nationwide system of health insurance, but relies instead on a complex mixture of public and private insurance. Most insurance is provided by employers to their employees and dependents. Government programs provide insurance to the elderly and disabled and some of the poor. 6/
Most of the population (74 percent) is covered by private health insurance, of which over 80 percent receives insurance through an employer. Although employer-provided health insurance is voluntary, it is encouraged by tax policy since it is tax deductible to the employer and tax free to the employee. There are over 1,000 private health insurance companies, which traditionally have provided coverage based on the indemnity insurance model, in which inpatient care is given generous coverage but preventive services and the management of chronic conditions that could reduce the need for inpatient care generally are not covered. 7/ Providers of medical care are paid on a fee-for-service basis, in which they are reimbursed for the services they provide (“retrospective” payment). Beginning in the 1970s, health maintenance organizations (HMOs), which combine the financing and delivery of medical care and give greater emphasis to preventive care, began to proliferate. During the 1980s, membership in HMOs rose by 300 percent, reaching 15 percent of the population in 1989. 8/ HMOs accept financial responsibility for a defined set of health care benefits in exchange for a fixed monthly premium (“prospective” payment). 9/
Medicare is a national health insurance program for the aged and disabled administered by the federal government. It is the single largest health insurer, covering about 13 percent of the population, including virtually all of those aged 65 and over and certain persons with disabilities or kidney failure. Medicare outlays were an estimated $138 billion (2 ¼ percent of GDP) in 1992. Payroll taxes represent the primary source of funding for the Medicare program and are supplemented by premiums and general government funds. Medicare is oriented toward acute care and does not provide coverage for long-term nursing home care and outpatient prescription drugs. These exclusions, combined with deductibles and copayments, have meant that Medicare covers less than one half of the total medical care expenses of the elderly. Many of the elderly have supplemental private insurance plans to cover these expenses, but out-of-pocket payments (including insurance premiums) still represented about 11 percent of their income in 1991, compared to 4 ½ percent for all age groups. 10/
Medicaid is a joint federal-state health insurance program that covers preventive, acute, and long-term care services for certain groups of the poor—about 10 percent of the population who are aged, blind, disabled, or members of families with dependent children. 11/ Each state designs and administers its own Medicaid program within broad federal guidelines. The Federal Government pays an average of 57 percent of the cost of the program (50-80 percent depending on state income). Combined federal and state expenditures were an estimated $129 billion (2 ¼ percent of GDP) in 1992.
2. Gaps in health insurance coverage
It is estimated that roughly 35 million persons (14 percent of the population) lack any type of health insurance coverage. 12/ Most of the uninsured are poor or near poor. In 1990, one third of the uninsured were in families with incomes below the federal poverty line ($13,254 for a family of four), while two thirds were in families with incomes below twice the poverty line. These gaps reflect the narrow designation of beneficiaries under the Medicaid program. Childless, nondisabled adults under 65 years of age, no matter how poor, are not eligible for government programs. 13/ As a result, Medicaid assisted only 45 percent of those with incomes below the poverty level in 1990. Even among the extremely poor (family incomes below 25 percent of the poverty line), nearly one quarter are not covered by Medicaid or any other program.
Most of the uninsured are directly or indirectly attached to the labor force but do not receive health insurance from their employers. In 1990, nearly 90 percent of the uninsured were workers or their dependents. About 50 percent of the uninsured were in families headed by full-year, full-time workers, while about 20 percent were in families headed by part-year, full-time workers. The remaining one quarter of the uninsured had only limited or no work force attachment. Most uninsured workers were in small firms—about half were in firms with 25 or fewer employees. This largely reflects the barriers that small firms face in obtaining insurance because of much higher administrative costs per worker 14/ and the underwriting and rating practices of the health insurance industry. 15/
Most of the uninsured are also relatively young. In 1990, 43 percent of the uninsured were under the age of 25 and 26 percent were under 18. Since many of these people are also relatively healthy, these statistics seem to suggest that the gaps in insurance coverage are not as serious as they would first seem. Moreover, most of those without health insurance who need medical care still receive some care. They may pay for the care directly or receive it free from providers, primarily hospitals. However, many of the young uninsured are very young—12 percent of children 5 years of age and under do not have insurance coverage. And while the uninsured do receive some medical care, they are less likely to receive preventive care and more likely to receive expensive emergency room care.
In addition to the gaps in coverage described above, few are insured against the high costs of long-term care necessitated by disability, including disabilities associated with old age. In particular, long-term care at home tends to be provided by family members without compensation because of this lack of coverage. Of the $53 billion spent on long-term care in 1988, only 18 percent was spent on home care even though 80 percent of the disabled and 60 percent of the severely disabled live at home. 16/ Spending on nursing-home care thus represents the bulk of long-term care expenditure and is financed about equally by the Government and from personal resources. 17/ Medicaid provides more than 40 percent of all nursing-home revenues, but finances nursing-home care only after personal savings have been exhausted. Some argue that the key to solving the long-term care problem is to encourage more self-insurance. But given the high cost of long-term care and the uncertainty surrounding the likelihood and length of a nursing-home stay, many argue for an increased role for public insurance. 18/
3. Comparisons with other OECD countries
The United States spends a larger share of its GDP on health care than any other OECD country. Compared with an OECD average of 7 ½ percent in 1990, the United States spent 12 percent of GDP on health care; that share rose to 13 percent in 1991 and an estimated 14 percent in 1992. Chart 1 compares the ratio of health care spending to GDP in the United States and the six largest of the other OECD countries. Health care spending relative to GDP has stabilized or declined in Germany, Japan, and the United Kingdom. 19/ Although the share of GDP devoted to health care has continued to rise in Canada, France, and Italy, those shares have risen more slowly and remain well below U.S. levels.
Health Care Expenditures of the Seven Largest OECD Countries
Citation: IMF Working Papers 1993, 093; 10.5089/9781451951189.001.A001
Source: OECD.The higher income share devoted to health care in the United States translates into a much higher level of per capita spending. In 1990, U.S. per capita health care spending was $2,566 compared with $1,770 in Canada (with the second highest level) and an OECD average of $1,204, converted to U.S. dollars using PPP exchange rates for GDP. Differences in income can explain some of the variation across countries. Chart 2 shows a simple linear relationship between per capita health spending and per capita GDP. The United States is a clear outlier, with spending exceeding by 40 percent the amount implied by this simple relationship. Had this relationship held for the United States, per capita health spending would have been $1,792 in 1990, about the actual level of spending in Canada, and total health care spending relative to GDP would have been 3 ½ percentage points lower. 20/
Health Care Expenditures and GDP: 1990
(Thousands of U.S. dollars per capita)
Citation: IMF Working Papers 1993, 093; 10.5089/9781451951189.001.A001
Health spending in the United States has been rising rapidly, at an annual compound rate of 10 ¼ percent during the period 1980-90. In real per capita terms (using the GDP deflator) U.S. health care spending rose at an annual compound rate of 4 ½ percent over the period, compared to an OECD average of 3 ¼ percent. 21/ However, while other countries spend a smaller percentage of their GDP on health care, many are facing increasing cost pressures. For example, Canada, whose national insurance system is often cited as an alternative to the U.S. private insurance system and whose health spending stood at 9 ¼ percent of GDP in 1990, experienced annual compound growth rates of 10 ½ percent in nominal terms and 4 ¼ percent in real per capita terms during the period 1980-90.
4. Shares of health care financing and expenditure
Chart 3 shows the share of U.S. national health expenditure financed by public and private sources, as well as the shares of various types of expenditure. Compared with an OECD average of 75 percent, the public sector in the United States financed about 42 percent of total health care expenditure in 1990, of which federal financing represented two thirds. 22/ Following a sharp increase beginning in 1965 with the introduction of the Medicare and Medicaid programs, the public sector share has remained relatively stable since the mid-1970s, but is projected to rise to just over 50 percent by the year 2000 if policies remain unchanged. 23/ Rising health outlays are expected to put substantial pressure on federal and state budgets in future years. The U.S. Congressional Budget Office (CBO) projects that by the year 2000, federal health care outlays will absorb over 28 percent of federal revenues, compared with about 18 percent in 1992. 24/ In fact, escalating health related outlays are the principal source of the deteriorating medium-term fiscal outlook foreseen by the CBO and the Administration.
U.S. National Health Expenditure: 1990
Citation: IMF Working Papers 1993, 093; 10.5089/9781451951189.001.A001
Notes: Other personal care includes home health and medical durables. Other spending includes includes public health, research, and construction.Source: HCFA.Private financing of health care represented about 58 percent of national health expenditure in 1990, with private insurance accounting for 33 percent and out-of-pocket expenses accounting for 20 percent of total expenditure. 25/ Private health insurance has grown as a source of financing since 1970, but is projected by the HCFA to decline to 28 percent of expenditure by 2000. Out-of-pocket expenditures have declined steadily since 1960 when they financed about 50 percent of health expenditure, and are projected to decline further to 17 percent of health expenditure by 2000. This decline in the share of costs directly financed by consumers has been put forth as one possible reason for the high rate of growth in health care spending over the period (see below).
Chart 3 shows the shares of total expenditure received by different sectors of the health care market. Hospitals (38 percent) and physicians (19 percent) together received over one half of total health expenditure in 1990, suggesting that an examination of the rise in total spending should look closely at these, sectors. These shares have remained remarkably stable over the past two decades (see Table 1) and do not differ markedly from those in other OECD countries. 26/ Drugs have exhibited a declining share of total expenditure (from almost 12 percent in 1970 to a little over 8 percent in 1990), while nursing-homes have shown a rising share (from 6 ½ percent in 1970 to 8 percent in 1990). Administrative costs accounted for 6 percent of national health expenditure in 1990, a level considerably higher than in other OECD countries. 27/
United States: Health Care Spending by Category
(Percentage of total health care expenditures)
Includes medical nondurables.
Includes “vision products”.
Includes construction, noncommercial research, government public health activities, program administration, and the net cost of private health insurance.
United States: Health Care Spending by Category
(Percentage of total health care expenditures)
| HCFA Projections | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1960 | 1965 | 1970 | 1975 | 1980 | 1985 | 1990 | 1991 | 1992 | 2000 | 2030 | ||
| Personal | 88.1 | 85.6 | 87.3 | 87.7 | 87.7 | 87.5 | 87.9 | 88.2 | 88.7 | 90.4 | 92.4 | |
| Hospital care | 34.2 | 33.7 | 37.6 | 39.4 | 40.9 | 39.8 | 38.4 | 38.7 | 39.5 | 40.3 | 41.8 | |
| Nursing home care | 3.6 | 4.1 | 6.5 | 7.5 | 8.0 | 8.1 | 8.0 | 8.1 | 8.1 | 8.4 | 9.3 | |
| Physicians | 19.5 | 19.7 | 18.3 | 17.5 | 16.7 | 17.5 | 18.9 | 18.7 | 18.5 | 19.8 | 24.1 | |
| Dentists | 7.2 | 6.7 | 6.3 | 6.2 | 5.7 | 5.5 | 5.1 | 5.0 | 4.7 | 3.6 | 1.8 | |
| Drugs 1/ | 15.7 | 14.2 | 11.8 | 9.8 | 8.6 | 8.6 | 8.2 | 8.1 | 7.9 | 7.2 | 6.5 | |
| Medical durables 2/ | 3.0 | 3.0 | 2.7 | 2.3 | 1.8 | 1.7 | 1.8 | 1.7 | 1.6 | 1.4 | 0.8 | |
| Other | 4.9 | 4.2 | 4.1 | 5.0 | 5.8 | 6.4 | 7.5 | 7.9 | 8.3 | 9.6 | 8.1 | |
| Nonpersonal 3/ | 11.9 | 14.4 | 12.7 | 12.3 | 12.3 | 12.5 | 12.1 | 11.8 | 11.3 | 9.6 | 7.6 | |
| Total (in billions of dollars) | 27.1 | 41.6 | 74.4 | 132.9 | 250.1 | 422.6 | 666.2 | 736.5 | 819.9 | 1,739.8 | 15,969.6 | |
| Percent of GDP | 5.3 | 5.9 | 7.4 | 8.4 | 9.2 | 10.5 | 12.1 | 13.0 | 13.9 | 18.1 | 32.0 | |
Includes medical nondurables.
Includes “vision products”.
Includes construction, noncommercial research, government public health activities, program administration, and the net cost of private health insurance.
United States: Health Care Spending by Category
(Percentage of total health care expenditures)
| HCFA Projections | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1960 | 1965 | 1970 | 1975 | 1980 | 1985 | 1990 | 1991 | 1992 | 2000 | 2030 | ||
| Personal | 88.1 | 85.6 | 87.3 | 87.7 | 87.7 | 87.5 | 87.9 | 88.2 | 88.7 | 90.4 | 92.4 | |
| Hospital care | 34.2 | 33.7 | 37.6 | 39.4 | 40.9 | 39.8 | 38.4 | 38.7 | 39.5 | 40.3 | 41.8 | |
| Nursing home care | 3.6 | 4.1 | 6.5 | 7.5 | 8.0 | 8.1 | 8.0 | 8.1 | 8.1 | 8.4 | 9.3 | |
| Physicians | 19.5 | 19.7 | 18.3 | 17.5 | 16.7 | 17.5 | 18.9 | 18.7 | 18.5 | 19.8 | 24.1 | |
| Dentists | 7.2 | 6.7 | 6.3 | 6.2 | 5.7 | 5.5 | 5.1 | 5.0 | 4.7 | 3.6 | 1.8 | |
| Drugs 1/ | 15.7 | 14.2 | 11.8 | 9.8 | 8.6 | 8.6 | 8.2 | 8.1 | 7.9 | 7.2 | 6.5 | |
| Medical durables 2/ | 3.0 | 3.0 | 2.7 | 2.3 | 1.8 | 1.7 | 1.8 | 1.7 | 1.6 | 1.4 | 0.8 | |
| Other | 4.9 | 4.2 | 4.1 | 5.0 | 5.8 | 6.4 | 7.5 | 7.9 | 8.3 | 9.6 | 8.1 | |
| Nonpersonal 3/ | 11.9 | 14.4 | 12.7 | 12.3 | 12.3 | 12.5 | 12.1 | 11.8 | 11.3 | 9.6 | 7.6 | |
| Total (in billions of dollars) | 27.1 | 41.6 | 74.4 | 132.9 | 250.1 | 422.6 | 666.2 | 736.5 | 819.9 | 1,739.8 | 15,969.6 | |
| Percent of GDP | 5.3 | 5.9 | 7.4 | 8.4 | 9.2 | 10.5 | 12.1 | 13.0 | 13.9 | 18.1 | 32.0 | |
Includes medical nondurables.
Includes “vision products”.
Includes construction, noncommercial research, government public health activities, program administration, and the net cost of private health insurance.
III. The Economics of Health Care
The health care market is characterized by a number of inefficiencies that imply that the socially optimal amount of health services may not be provided. 28/ First, a high level of spending is encouraged by the fact that consumers of health care bear little of the direct cost of medical treatment. Rather, the cost is covered by a distant third-party payer that typically does not participate directly in the treatment decision and is likely to reimburse for the cost of treatment as long as it is recommended by the provider. 29/ The welfare loss associated with the overconsumption of health services stemming from these inappropriate incentives is illustrated in Figure 1 below.
The Benefits Curve for Health Care
Citation: IMF Working Papers 1993, 093; 10.5089/9781451951189.001.A001
The demand curve or marginal benefits curve for medical care is represented by DD1, with the added benefit of health care declining with each successive dollar spent. Suppose the supply of medical care were perfectly elastic at price P0, the cost of providing a unit of care. If consumers were to face the full cost per unit of care they would demand Q0, the quantity of health care at which marginal cost is equal to marginal benefit. However, health insurance and government subsidies lower the marginal cost faced by consumers, driving a wedge between social and private cost and increasing the demand for health services to an excessive level. As discussed above, the average out-of-pocket cost to consumers in 1990 was 20 percent of national health expenditures. This figure probably overstates the marginal cost faced by many consumers, however, since the incremental cost of health care falls to zero under most policies once a certain level of expenditure has been reached. In Figure 1, a marginal cost of P1 would cause consumers to demand Q1. At point C, marginal social cost exceeds marginal social benefit, producing a welfare loss of ABC.
The overconsumption of health services may be encouraged further by a second problem in the market for health care, namely that consumers lack information about the price and quality of services they receive. The inability of patients to assess the value of the treatment they receive causes them to delegate responsibility to health care providers, putting these providers in the position of acting both as agents for consumers and suppliers of medical care. This position has raised concerns about “supplier-induced” demand: providers are both ethically bound and financially motivated to encourage their patients to receive treatments no matter how small the benefit and regardless of cost. 30/ 31/ The overconsumption of health services stemming from these information problems and from supplier-induced demand is illustrated in Figure 2. Suppose a consumer believes (possibly through encouragement by a health care provider) that the marginal benefit of a treatment is greater than its true social marginal benefit, causing the demand for health care to shift outward to DD2. Facing price P1, the quantity Q2 will be demanded, resulting in an even greater welfare loss of abc.
The Benefits Curve for Health Care
Citation: IMF Working Papers 1993, 093; 10.5089/9781451951189.001.A001
These problems in the health care market, which prevent the market from providing a socially optimal amount of health services, point to a need for reforms that focus on ensuring greater information to consumers and providers about appropriate care and increasing the cost awareness of consumers and providers. However, the possibility of physician-induced demand implies that demand side strategies such as raising deductibles and copayments 32/ may not be very effective in controlling spending since consumers base their decisions on the advice received from their physicians as well as the costs they face. 33/
Moreover, a small percentage of the population accounts for a large percentage of total health expenditures. This high concentration of spending suggests that there may be limits to raising the cost awareness of consumers while providing universal access to medical care. In any given year, 5 percent of the population accounts for more than one half of all health care outlays, while 1 percent of the population accounts for over one fourth of outlays. Almost all of national health care outlays (90 percent) are accounted for by 30 percent of the population. Extrapolating from past trends, the per capita spending of $2,566 in 1990 would have implied average per capita outlays of $72,500 and $27,500 for the most costly 1 percent and 5 percent of the population, respectively. 34/ This concentration of expenditure implies that to have an appreciable effect on total health care spending, the marginal cost of health care faced by consumers would have to be raised to a point that would impede access.
IV. Possible Reasons for High and Rising Health Care Costs
The inefficiencies in the health care market discussed above are thought to result in a level of spending for health services that is higher and rising faster than is socially optimal. In evaluating the sources of high and rising expenditures on health in the United States, it is useful to distinguish between factors that may affect the level of spending and those that may affect the rate of growth in spending.
1. The high level of U.S. health expenditure
Excess administrative costs (of providers as well as insurance companies) are estimated to represent 5 percent to 13 percent of total health care spending (as much as 1 ¾ percent of GDP in 1991). 35/ Administrative expenses in the United States consist of insurance overhead, hospital and nursing administration, and physicians’ overhead and billing. Estimates of excess administrative costs associated with the private insurance system in the United States have been made by comparing the share of national health expenditures that are devoted to administrative expenses in the United States to the share in countries such as Canada with national health insurance systems. It is estimated that in 1987, the United States spent 19-24 percent of health care spending on administrative expenses compared with 8-11 percent in Canada. 36/ Further evidence of the high administrative expenses associated with private insurance may be found within the United States itself: administrative expenses of the federally run Medicare program represented only 2 percent of program expenditures in 1990, compared with administrative expenses of approximately 14 percent of premiums for private insurance companies. 37/ The estimates of excess administrative costs presented above include the costs of excessive paperwork faced by health care providers. Such costs are difficult to measure, but anecdotal evidence suggests that they represent a significant burden to providers in the United States stemming from the hundreds of insurance companies and claim forms that are used.
Unnecessary spending on facilities and equipment and spending on health care that is medically inappropriate are estimated to represent 5 percent to 20 percent of total health care outlays (as much as 2 ½ percent of GDP in 1991). 38/ The possible existence of excess capacity is suggested by marked differences in the level of capital investment in the United States compared to that in other countries. Compared with other countries, the United States has invested heavily in high-technology equipment and facilities: CBO (1991a) found 0.7 cardiology units per million people in Germany, 1.2 in Canada, and 3.3 in the United States. 39/ Similarly, Germany had 0.9 magnetic resonance imaging machines per million people compared with 0.5 in Canada and 3.7 in the United States. Further evidence of excess capacity is provided by the high proportion of U.S. physicians who are specialists compared to that in other industrial countries and the percentage (about 30 percent in 1990) of vacant beds in U.S. hospitals. 40/
The possible presence of inappropriate medical care in the United States is suggested by significant geographic variations in practice patterns such as average length of hospital stays, hospital discharge rates, and surgical procedures. These variations cannot be fully explained by differences in the health status of the populations. Rather, the most important factor seems to be differences in the practice styles of physicians. Further evidence on the appropriateness of care has been found in audits of medical records. A review of the literature by the GAO 41/ concluded that 14 percent to 32 percent of surgical procedures were unnecessary, while 7 percent to 19 percent of hospital admissions were unnecessary. Inappropriate medical care poses risks to patients, as well as wasting resources. However, some inappropriate care involves the misuse or underutilization of procedures and therefore might not suggest higher-than-appropriate spending. 42/
2. Rapid growth of U.S. health expenditure
From 1960 to 1990, real per capita health care spending (using the GDP deflator) in the United States rose by a factor of four. 43/ Reasons put forth to explain this rapid growth include: rising incomes; a decline in the average coinsurance rate paid by consumers; rising administrative expenses; demographic factors; the rising costs of medical malpractice; medical price inflation in excess of general inflation; and the advent of new medical technologies and procedures.
One would expect spending on medical care to rise with income. Estimates of the income elasticity of health spending vary widely, however. Cross-sectional estimates for household demand within the United States yield estimated income elasticities of 0.2 to 0.4. However, these estimates may be biased downward by the endogeneity of income at the household level—an illness may simultaneously lower income and raise medical expenditure. 44/ International cross-sectional studies of OECD countries yield estimated elasticities of 0.8 to 1.5. 45/ Assuming a unit elasticity of demand, the 80 percent increase in real per capita GDP between 1960 and 1990 would explain a little over one fourth of the rise in real per capita spending on health care.
As described in Section III above, the mere existence of insurance for health care expenditure can contribute to excessive spending on health care. Similarly, the growing importance of insurance may explain some of the increase in real health care spending over the past several decades. Private out-of-pocket payments have declined significantly over the period 1960-90, from 49 percent to 20 percent of national health expenditures, reflecting the growing importance of both private insurance and government health programs. 46/ However, using widely cited estimates of price elasticities from the RAND Health Insurance experiment, this decline in the average coinsurance rate can explain only one twentieth of the increase in real per capita spending over the period. 47/
In addition to their effects on the level of expenditure, discussed above, administrative costs have been cited as an important reason for increases in real health care outlays. Between 1980 and 1990, administrative expenses of private insurers rose by 277 percent compared with an increase of 185 percent in benefits. 48/ Some argue that some of these higher administrative costs may have actually helped to control spending, however, since they may be associated with “utilization controls,” such as pre-approval and second opinions, put in place by insurance companies in order to control the overconsumption of medical services and eliminate unnecessary treatment.
Other commonly cited factors such as malpractice insurance and demographics are unable to explain much of rapid growth of expenditure. Malpractice insurance premiums represent only one percent of total health care spending, so although they have grown rapidly over the period, they can explain little of the growth in total health care outlays. Estimates of the costs of inappropriate medical care to avoid successful malpractice claims (“defensive medicine”) are only slightly higher, at about 3 percent of total health care spending. 49/ Demographic shifts can explain only one twentieth of the increase in real per capita spending between 1965 and 1990. Although spending on health care for those aged 65 and over is roughly four times the average spent for the rest of the population, the share of the population that is in this age group is relatively small and rose by only 3 percentage points over the period. 50/
One often cited source of rising health care spending is escalating prices of medical services. The discussion so far has focused on per capita health expenditure deflated by the general price index. However, between 1960 and 1990, medical price inflation averaged 7 ¼ percent a year versus a 5 percent annual average general rate of inflation as measured by the GDP deflator. 51/ As a consequence, the relative price of medical services rose by 92 percent over the period and thus could potentially account for about one fourth of the increase in real (adjusted for changes in the general price level) per capita spending. However, there are serious problems with the price index for medical services since it does not allow for quality changes. A price index should measure the price of the same product over time and should adjust for any changes in quality in that product. While other price indices suffer from similar problems, the pace of technological change in the health care industry (see below) renders the medical price index particularly flawed. For example, the use of computerized tomography (CT) scanners rather than X-ray equipment represents a significant quality improvement, but the current index treats the additional expenditure as pure medical price inflation. 52/
whether the rise in the relative price of medical care stems from price or quality changes has important implications for reform of the health care system. 53/ To the extent that movements in the medical price deflator actually reflect price rather than quality changes, reform might focus on eliminating rents received by providers of medical services. For example, some have pointed to what they view as excessive incomes received by physicians and profits in the pharmaceutical industry as sources of cost pressures. Conversely, if quality changes are being mislabeled as medical price inflation, attempts to control health care spending will need to focus on limiting the quality or quantity of care that consumers receive. As discussed in Section III above, several factors, including the existence of health insurance, imply that the quantity of medical care consumed is greater than would be implied by the equality between marginal social cost and marginal social benefit.
A steady rise in the relative wages of health care employees lends some support to the view that the relative price of medical care has increased. Between 1963 and 1989, compensation of health care workers relative to average compensation rose by 47 percent. During the 1960s and 1970s, the ratio of physician incomes to average compensation fluctuated around 4 ½. Beginning in the 1980s, physicians’ relative incomes began to rise, reaching over 5 ½ times average compensation by 1989. 54/ This figure compares with 4 ¼ in Germany, 3 ¾ in Canada, and 2 ½ in France, Japan, and the United Kingdom. 55/ Further evidence of rents in the provision of health care has been found by comparing pharmaceutical prices across countries. The U.S. General Accounting Office (GAO, 1992) found that manufacturers’ prescription drug prices tended to be significantly higher in the United States than in Canada, with a median differential of 43 percent and a maximum differential exceeding 500 percent. The report concluded that the differences stemmed primarily from the regulation of drug prices in Canada rather than from differences in costs of production.
Even after taking into account the factors cited above, there is a significant unexplained residual in the increase in health care spending. Most analysts agree that this residual should be attributed to the advent of new and expensive medical technology. 56/ The practice of medicine today consists largely of diagnosis and treatment devices and techniques that were unknown 40 years ago. 57/ Moreover, the rise in hospital expenditure, the largest component of national health expenditure, seems to reflect to a significant degree an increase in the “intensity” of services, as opposed to an increase in the volume or price of the same services over time. Between 1971 and 1986, the ratio of hospital employees to patients rose from 2 ¾ to almost 4, while the average length of a hospital stay declined. 58/
Weisbrod (1991) argues that the pace of technological change in U.S. medical care is closely linked to the health care financing system. In particular, the third-party payer system provides incentives for the adoption of new technologies irrespective of cost and no matter how small the benefit. Consequently, technological change in the health care industry has tended to be cost increasing, as compared to technological innovations in other industries that have largely resulted in declining costs. Weisbrod concludes that if the current retrospective cost-based payment system encourages cost-increasing technological change, then a move to a prospective fixed-fee system should encourage more cost-saving innovations. 59/
V. Managed Competition and Other Proposals for Reform
The most commonly cited proposals for reform of the U.S. health care system (to provide universal coverage and control costs) include tax credits, “play or pay” mandates, managed competition, and national health insurance. The proposals vary widely in the degree to which they would change the current U.S. system. Tax credits merely make the current system more affordable to lower-income groups without changing the voluntary nature of employer-provided health insurance. “Play or pay” mandates require that employers provide health insurance to their employees or pay a payroll tax to cover their employees’ enrollment in a public health care plan. Critics of this proposal point to the incentive for employers with high-cost plans (stemming from the experience-rating practices of insurance companies) to choose the “pay” option and shift their employees to the public plan, resulting in rising public costs. Neither of these proposals includes incentives for cost control, although, for example, a “play or pay” mandate could incorporate changes in the public plan that encourage cost savings.
Managed competition is the model of reform that has been the focus of the recent health care reform debate and would represent a more fundamental change to the current system. It is presented as a combination of regulation and competition in the market for health care, which would provide for universal coverage and would focus more closely than some other reform proposals on cost control (see below). National health insurance would represent the largest departure from the current system and would involve the replacement of more than 1,000 private insurance companies with the government as a single insurer. 60/ Health services would continue to be provided privately but would be publicly financed, eliminating the excess administrative costs associated with the current private payer system and lending the Government considerable scope, as the sole source of financing of health services, to control costs. Such a reform would involve a significant redistribution of financial flows as private insurance premiums would be substituted by taxes to finance health care coverage.
Under managed competition, government regulations are designed to ensure an equitable distribution of health care resources and to deal with the special characteristics of the health care market, such as information asymmetries and institutional limits on competition, while elements of competition are introduced to increase cost-consciousness among consumers, providers, and insurers and thus enhance efficiency. Health cooperatives (or alliances) are the managers of managed competition 61/—they are regional agencies that would contract with health plans (fee-for-service as well as HMOs and other managed care plans) and bargain over the price of a standard set of benefits while overseeing quality. Groups of providers or insurers would offer plans that compete with each other based on price (the premium of the standard benefit package) and quality. The role of the Government would be to oversee the system, ensuring quality care and universal access.
Currently, consumers have few incentives to choose low-cost health plans because they usually bear little of the incremental cost of more expensive plans. 62/ Furthermore, complex differences in the coverage and quality (if providers are restricted) of health plans give consumers little basis on which to discriminate among plans based on price. Managed competition would encourage consumers to choose low-cost health plans by creating a standardized health benefit package and making available indicators of quality of care under each plan, and by requiring consumers to buy more expensive plans with their after-tax incomes.
Managed competition would prompt insurers to compete by offering low-cost plans, rather than by trying to attract only relatively healthy enrollees (“cream-skimming”). Insurers would be required to announce premiums based on community ratings 63/ and agree to offer the standard benefit package to anyone who wished to purchase it without exclusions for pre-existing conditions. These community-based premiums would be collected by the regional health cooperative, which would redistribute them to the plans after adjusting them for the average health status of the plans’ enrollees. 64/ Insurers would compete for enrollees by arranging with networks of providers for cost-effective high-quality care.
Providers, in order to be affiliated with a health plan, would in turn face stronger cost-saving incentives under managed competition. They would be encouraged to compete based on the cost-effectiveness of their care while maintaining quality, rather than by offering the latest technologies (regardless of benefit) and amenities (such as spacious private hospital rooms). This increased cost-consciousness of providers would cause them to be less likely to perform procedures that offered little benefit relative to their cost. In particular, the use of new high-cost technologies would face a stronger market test under managed competition.
Cost control under managed competition would occur through changes in the financing of health care, but is expected to change the way health care is delivered. The incentives for reduced consumption of medical care are thus indirect. Consumers economize on their purchase of health insurance, causing insurers to compete for enrollees and to put pressure on the providers with which they are affiliated to practice medicine conservatively. However, incentives to control costs under managed competition could be weakened if, for example, large firms were allowed to self insure and opt out of the system or supplemental insurance beyond the standard benefit package were permitted.
In order to complement the incentives for cost control under managed competition, some would advocate limits on national health expenditures or global budgets. A market-determined global budget could be defined as the average premium in each region, multiplied by the number of people residing in the region, summed up over all regions. 65/ If national health expenditures grew faster than desired from a public policy standpoint, the Government could reduce benefits under the standard package or raise copayments and deductibles. 66/ An alternative approach is for the Government to set an overall budget to control spending and then allow managed competition to allocate resources within that budget. 67/ Whether health care reform based on managed competition will need to rely on global budgets to control spending over the longer term depends in large part on the extent to which it can encourage cost-saving rather than cost-increasing technologies.
References
Aaron, Henry J. (1991), Serious and Unstable Condition: Financing America’s Health Care, Washington: Brookings Institution.
Burner, Sally, Daniel Waldo, and David McKusick (1992), “National Health Expenditures Projections through 2030,” Health Care Financing Review, Vol. 14, Number 1, pp. 1–29.
Congressional Budget Office (1991a), Rising Health Care Costs: Causes. Implications, and Strategies, Washington, April 1991.
Congressional Budget Office (1991b), Universal Health Insurance Coverage Using Medicare’s Payment Rates, Washington, December 1991.
Congressional Budget Office (1992a), Economic Implications of Rising Health Care Costs, Washington, October 1992.
Congressional Budget Office (1992b), Projections of National Health Expenditures, Washington, October 1992.
Congressional Budget Office (1993a), Managed Competition and Its Potential to Reduce Health Spending, Washington, May 1993.
Congressional Budget Office (1993b), Trends in Health Spending: An Update, Washington, June 1993.
Consumers Union (1992a), “Health Care Dollars,” Consumer Reports, July, pp. 435–48.
Consumers Union (1992b), “Are HMOs the Answer?,” Consumer Reports, August, pp. 519–31.
Consumers Union (1992c), “The Search for Solutions,” Consumer Reports, September, pp. 579–92.
Council of Economic Advisors (1993), “The Economics of Health Care,” Economic Report of the President, pp. 119–68.
Davis, Karen (1989), “National Health Insurance: A Proposal,” American Economic Review, Papers and Proceedings, Vol. 79, May, pp. 349–52.
Davis, Karen, et al. (1990), Health Care Cost Containment, Washington: Johns Hopkins University Press.
De Lew, Nancy, George Greenberg, and Kraig Kinchen (1992), “A Layman’s Guide to the U.S. Health Care System,” Health Care Financing Review, Vol. 14, No. 1, pp. 151–69.
Ebrill, Liam (1992), “The U.S. Health Care Industry: Performance and Issues,” in The United States Economy: Performance and Issues, edited by Yusuke Horiguchi, Washington, D.C.: International Monetary Fund, pp. 499–530.
Enthoven, Alain C. (1993), “The History and Principles of Managed Competition,” Health Affairs, Supplement, pp. 24–48.
Levit, Katharine, Gary Olin, and Suzanne Letsch (1992), “Americans’ Health Insurance Coverage, 1980-91,” Health Care Financing Review, Vol. 14, Number 1, pp. 31–57.
Newhouse, Joseph P. (1992), “Medical Care Costs: How Much Welfare Loss?,” Journal of Economic Perspectives, Vol. 6, Number 3, pp. 3–21.
Newhouse, Joseph P. (1993), “An Iconoclastic View of Health Cost Containment,” Health Affairs, Supplement, pp. 152–71.
Organisation for Economic Cooperation and Development (1992a), The Reform of Health Care: A Comparative Analysis of Seven OECD Countries, Paris: OECD.
Organisation for Economic Cooperation and Development (1992b), U.S. Health Care at the Cross-roads, Paris: OECD.
Pepper Commission (1990), A Call for Action: Final Report of the U.S. Bipartisan Commission on Comprehensive Health Care, Washington D.C.: U.S. Government Printing Office.
Phelps, Charles E. (1992a), “Diffusion of Information in Medical Care,” Journal of Economic Perspectives, Vol. 6, Number 3, pp. 23–42.
Phelps, Charles E. (1992b), Health Economics, New York: Haper Collins.
Reinhardt, Uwe (1989), “Economists in Health Care: Saviors, or Elephants in a Porcelain Shop?” American Economic Review, Papers and Proceedings, Vol. 79, May, pp. 337–42.
Rich, Spencer (1993), “A ‘Living Example’ for Health Care,” Washington Post, April 5, p. A19.
Schieber, George, Jean-Pierre Poullier, and Leslie Greenwald (1992), “U.S. Health Expenditure Performance: An International Comparison and Data Update,” Health Care Financing Review, Vol. 13, Number 4, pp. 1–87.
Starr, Paul (1993), “Design of Health Insurance Purchasing Cooperatives,” Health Affairs, Supplement, pp. 58–64.
Starr, Paul and Walter A. Zelman (1993), “A Bridge to Compromise: Competition Under a Budget,” Health Affairs, Supplement, pp. 7–23.
U.S. General Accounting Office (1991), Improvements Needed in the Identification of Inappropriate Hospital Care, Washington: GAO.
U.S. General Accounting Office (1992), Prescription Drugs: Companies Typically Charge More in the United States than in Canada, Washington: GAO.
Zelman, Walter A. (1993), “Who Should Govern the Purchasing Cooperatives?,” Health Affairs, Supplement, pp. 49–57.
The author would like to thank Robert Anderson, Jessica Banthin, Robert Ford, and colleagues in the Western Hemisphere Department for useful comments and discussion.
Estimates by the U.S. Health Care Financing Administration (HCFA), which administers the Government’s Medicare and Medicaid programs (Burner et al. (1992)).
Davis et al. (1990), p. 4.
Reinhart (1989), p. 339.
Estimates of the percentage of the U.S. population who are “underinsured” vary widely, from 8 percent (Pepper Commission (1990)) to 20 percent (De Lew, et al. (1992)). The underinsured are defined as those who do not have a limit on their out-of-pocket health expenses and are at risk of being impoverished should they experience a costly, major illness. In addition, few Americans are insured against the high costs of long-term care and the strong link between health insurance and employment implies that many are at risk of losing their insurance coverage.
For further details of the U.S. health care system see De Lew, et al. (1992), pp. 151-60.
Enthoven (1993), pp. 25-6.
CBO (1991a), pp. 14-5.
Prospective payment is thought to help to control health care spending by removing some of the incentives created by insurance for the overconsumption and overprovision of medical care. (These incentives under insurance are discussed further in Section III. HMOs and other types of so-called managed care arrangements and their ability to control spending are discussed further in Sections IV and V.)
CBO (1993b), p. 3.
The sum of the percentages of the population covered by private insurance, Medicaid, and Medicare and the percentage that is uninsured exceeds 100 because some persons have more than one type of coverage.
Moreover, states can set their own eligibility requirements for Medicaid. For example, in Alabama, a family of three qualifies for Medicaid only if its income is less than 13 percent of the federal poverty guidelines (Pepper Commission (1990), pp. 30-31).
Administrative expenses represent 40 percent of claims for the smallest plans (1-4 workers) and 25 percent for groups of 20-49, compared to 5 percent of claims for very large groups (10,000 or more).
Private insurers use experience rating in an attempt to deal with adverse selection. Adverse selection arises in insurance markets when individual policy holders differ in their riskiness and the insurer is unable to fully distinguish those differences, thus attracting a disproportionate number of bad risks. Insurers attempt to avoid adverse selection by attracting only low-risk customers (“cream-skimming”) and charging premiums based on their assessment of individual risk (experience rating) (Pepper Commission (1990), p. 27).
Public programs provide little assistance with long-term care expenses in the home or community. Medicare only provides in-home care to those who are unable to leave their homes and require skilled or professional services (i.e. custodial care is not covered). Medicaid spending on home care represents only about 10 percent of its spending on nursing-home care.
Private insurance finances only 1 percent of nursing-home care.
Pepper Commission (1990), pp. 92-3.
Figures for Germany are for the former West Germany.
The OECD (1992b, p. 17), using an estimated equation that excludes the United States, concludes that predicted health care spending would have been 4 ½ percent of GDP lower. Conversely, a slightly smaller difference between predicted and actual health care expenditure is found if the equation (including the United States) is estimated in log form, but the difference is insignificant when presented as a percentage of GDP.
The GDP deflator is used to calculate real health care spending because of serious problems with the price index for medical care, which are discussed below.
The calculated share of health care expenditures financed by the public sector does not take account of the deductibility of health insurance premiums for employers. For the Federal Government, that deduction was equivalent to an outlay of over $45 billion in FY 1991 (approximately 6 percent of FY 1991 total health care outlays).
HCFA projections presented in Burner, et al. (1992), p. 22.
CBO (1992a), p. 50.
The remaining 5 percent includes philanthropy.
See OECD (1992b) for data on other OECD countries.
This figure does not include the administrative costs faced by providers and masks the significantly higher administrative costs of private insurance relative to public insurance. See Section IV below for further discussion.
See CBO (1992a), pp. 11-20, for a more detailed discussion. It is important to note that the following discussion focuses on economic efficiency. CBO (1993a), p. 34, makes the important point that other criteria could be used—that care be provided in the most cost effective way possible (cost minimization) or that only benefical care be provided (positive marginal benefit). Following these criteria would generate some cost savings, but they would be smaller than those generated by requiring that marginal benefit equal marginal cost.
HMOs, by integrating the provision and financing of health care, have attempted to overcome this problem.
The supply side of the health care market is also characterized by informational problems since providers may lack accurate information about the benefits of certain procedures. See discussion in Section IV on inappropriate care.
Licensing requirements for health care providers limit the ability of sellers to enter the market and therefore further reduce the incentive for providers to practice medicine cost effectively. To the extent that providers do compete, it tends to be based on the availability of the latest technologies and amenities (such as large private rooms in hospitals).
A deductible is a fixed amount of health care spending (say, $200) that an enrollee must pay before insurance will reimburse for expenses. A co-payment (say, 20 percent) is the share that an enrollee must pay of expenses covered by insurance. A stop-loss provision limits the amount of out-of-pocket expenses, often by reducing the co-payment requirement to zero once a certain level of expenditure has been reached. The marginal cost faced by the consumer is the full cost of treatment until the deductible is reached, falls to 20 percent of the cost of treatment once the deductible is reached, and falls further to zero once the stop-loss provision is reached.
CBO (1991a), p. 21.
See Aaron (1991), pp. 51-2.
See OECD (1992b), pp. 33-4 and Rich (1993), p. A10.
Consumers Union (1992c), p. 586.
Administrative expenses of the Medicaid program, which is jointly run by the federal and state governments, were 4 ¼ percent of program expenditures in 1990.
See Consumers Union (1992a), pp. 436-37 and Rich (1993), p. A10 for a summary of estimates.
Figures for Germany refer to the former West Germany.
CBO (1993a), p. 36.
GAO (1989).
Pepper Commission (1990), pp. 40-1.
The GDP deflator is used to calculate real health care spending because of serious problems with the price index for medical care, which are discussed below.
Newhouse (1992), pp. 7-8, makes this point.
See Scheiber et al. (1992), p. 7 and CBO (1992a), pp. 26-7. Newhouse (1992), p. 8, highlights a potential source of upward bias in aggregate time series estimates of income elasticities, namely, that technology is not held constant over time and would itself have a positive income elasticity.
The share of national health expenditure financed by private health insurance increased by 11 percentage points, while that financed by all levels of government rose by 18 percentage points.
The RAND Health Insurance experiment was a large, federally funded empirical study of variations in health care spending under alternative insurance plans. The authors found health care spending to be highly inelastic, with estimated price elasticities of -0.1 to -0.2 (Phelps (1992b)).
CBO (1992a), p. 25.
Estimates are presented in Newhouse (1992), p.9; CEA (1993), p. 150; and Consumers Union (1992a), p. 443. Estimates of the costs of defensive medicine would be included in broader estimates of the costs of inappropriate care. Note, however, that high premiums and the risk of malpractice may also result in underutilization of medical services, as evidenced by the decline of obstetrical services in some areas (premiums are the highest in this specialty, averaging $35,300 in 1987 compared with a national average of $15,900).
Changes in the age distribution of the population are expected to become more important in the coming years, however, as the baby-boom generation reaches retirement age. In 1990, those 65 and older represented about 12 percent of the population; that share is projected by the CBO to reach 20 percent by the year 2030 (CBO (1992a), pp. 25-6).
This differential widened during the 1980s, with medical price inflation averaging 8 ½ percent a year and general price inflation averaging 5 percent a year.
The consumer price index for medical services suffers from additional problems: it uses list prices that do not incorporate discounts and reflects the shares of out-of-pocket expenses faced by consumers (giving, for example, a disproportionately small weight to hospital costs since most are covered by insurance).
CBO (1992a), p. 22.
However, the increase in physician incomes in the 1980s was not too different from that of other post-college-educated workers (OECD (1992b), p. 56).
OECD figures for 1987 cited in Consumers Union (1992a), p. 444.
See for example Aaron (1991), Newhouse (1992), and Weisbrod (1991). Newhouse attributes 50-75 percent of the increase in real spending over the past 50 years to technological change.
Weisbrod (1991), p. 523.
Aaron (1991), p. 42.
The fact that the rate of increase in spending by HMOs, which use a prospective payment system, is not too different from fee-for-service plans would suggest that potential savings in this regard are not great. However, HMOs have little incentive under the current system to control premium increases by eliminating low-benefit care because HMOs must compete with the dominant fee-for-service sector and consumers usually bear little of the incremental cost of more expensive insurance plans. See further discussion in Section V.
Private companies could continue to offer supplementary policies and could be subcontractors to manage the government program.
Managed competition does not refer to managed care, although HMOs and preferred provider organizations (PPOs) are expected to increase in importance under this type of system (Starr (1993)).
The following discussion draws on CBO (1993a).
Premiums set by community ratings would reflect the average health status of the population in a region, rather than the health status of enrollees in a particular plan (experience rating).
The use of such “risk-adjusters” prevents plans from being penalized if they randomly receive enrollees who have poorer-than-average health.
Enthoven (1993) advocates using the lowest premium in each region, thus putting downward pressure on spending.
This approach assumes that copayments and deductibles can be used to limit the demand for health services without impeding access.
See for example Starr and Zelman (1993).
