Trends in Medicare Expenditures and Financial Status, 1966-2000.

In this article, the author reviews expenditure growth trends over Medicare's 35-year history and comments on how the program's long-range financial outlook has changed over time. The author focuses on the various legislative, economic, and demographic factors that have affected expenditure growth and financial status. In addition, Medicare's share of total U.S. health costs is briefly reviewed. In an appended comment, the author considers whether the impact of the Balanced Budget Act of 1997 (BBA) was greater than intended by Congress and the Administration. The author concludes with a plea for greater attention to correcting the projected long-range deficits for the Hospital Insurance (HI) Trust Fund.

Introduction

For 35 years, the Medicare program has helped cover the costs of medical care for most persons age 65 or over and (after 1972) for certain disabled persons. It is the Nation's second-largest social insurance program, with total expenditures in calendar year 2000 estimated to reach $228 billion. (For comparison, Social Security Old-Age, Survivors, and Disability Insurance [OASDI] expenditures are expected to total $410 billion in 2000.) The purpose of this article is to review the trends in Medicare expenditures since the program began operations in 1966 and to comment on how its long-range financial outlook has changed over time.

Medicare was enacted in 1965 as a sweeping compromise among competing proposals. Consequently, many of its coverage, eligibility, benefit, and financing provisions differ substantially between Medicare Part A (HI) and Part B (Supplementary Medical Insurance, or SMI). Because the circumstances leading to this outcome are not well known, Medicare's first chief actuary, Robert J. Myers, has graciously provided a very interesting account of these events, which appears in the ensuing article of this 35th Anniversary issue (Myers, 2000). A description of the HI and SMI provisions is also available in Hoffman, Klees, and Curtis (2000).

In 1967, the first full calendar year of operation for Medicare, program expenditures accounted for approximately 9.7 percent of all health expenditures in the United States (Health Care Financing Administration, 2000). This proportion increased steadily from 1972 to 1983 before stabilizing in the vicinity of 16 percent in 1984-1993. In recent years, the percentage increased further, reaching 19.4 percent in 1997, but it has declined somewhat following the BBA to an estimated 18.4 percent in 2000. From 1967 through (estimated) 2000, Medicare expenditures per beneficiary increased at an average annual rate of 10.0 percent, while the corresponding figure for per capita national health expenditures is an estimated 9.3 percent. Although these average growth rates are roughly similar, they have diverged significantly on occasion within this period (Levit et al., 2000).

For the population age 65 or over, Medicare paid for about 42 percent of total personal health care expenditures in fiscal year 1968, as shown in Figure 1 (Cooper, Worthington, and McGee, 1976). By calendar year 1997, this percentage had increased to 55 percent, with most of the balance covered by Medicaid, private health insurance, and the beneficiaries' own out-of-pocket payments. (These figures are not strictly comparable, because the percentages for 1968 are for everyone in the population at ages 65 or over, whereas those for 1997 are only for Medicare beneficiaries at such ages. Given that nearly everyone over age 65 was covered by Medicare in 1967, the difference has little impact on the comparison.) Medicare's increased share is in part attributable to the Part B deductible, which was $50 in 1968 and has been increased only three times since then, to $100 currently. Because covered costs increased far more rapidly, a greater proportion of covered costs is in excess of the deductible and is therefore reimbursable by Medicare. In 1968, only 38 percent of beneficiaries had Part B costs in excess of the deductible, but by 1997, this proportion had risen to 87 percent (Gornick, 1976; Board of Trustees of the Federal Supplementary Medical Insurance Trust Fund, 2000). Medicare's increasing share has also reflected rapid growth in the prices, utilization, and intensity of such covered services as physician, skilled nursing, and home health care. On the other hand, in some years, certain non-covered costs—such as for prescription drugs and long-term nursing home care—increased more rapidly than health costs generally, thereby adding to the portion funded by non-Medicare sources. Overall, the trend has been toward a greater Medicare share of the total personal health care costs of the aged.

Figure 1

Sources of Funding for Personal Health Care Expenditures for Persons 65 or Over, 1968 and 1997

Also noteworthy in Figure 1 is the relatively small decline in Medicaid outlays as a percentage of total personal health care expenditures for beneficiaries over age 65. The proportion of older persons with incomes below the poverty thresholds (who are the most likely to be eligible for Medicaid) fell from roughly 16 percent in 1966 to 11 percent in 1997 (Gornick et al., 1985; Social Security Administration, 1999). The impact of this trend on Medicaid expenditures was largely offset, however, by expansions in coverage, including the creation of Qualified Medicare Beneficiaries (QMBs) and Specified Low-Income Medicare Beneficiaries (SLMBs). (Medicaid pays the Medicare premium[s] on behalf of QMBs and SLMBs and also the beneficiary cost-sharing liabilities for QMBs.) In addition, during this period, Medicaid absorbed a substantial portion of the rapidly increasing expenditures for nursing home care.

The proportion of health care service costs paid directly by beneficiaries has declined significantly since the beginning of the program, from about 28 percent in 1968 to 20 percent currently. This change is attributable primarily to the increased shares covered by Medicare and private health insurance. It should be noted, however, that beneficiaries' premium payments to Medicare and private insurance plans are not included in the out-of-pocket costs shown here.

Review of Medicare Expenditure Trends

Figure 2 shows aggregate annual Medicare expenditures for calendar years 1966-2000. (Medicare began benefit payments in July 1966; therefore, the amount shown for the first year reflects only 6 months. The amount shown for 2000 is an estimate, based on data through July 2000.) The expenditures are shown using a semi-log scale to highlight periods of faster or slower growth, indicated by the slope of the expenditure curve. As shown in Figure 2, Medicare expenditure growth has been fairly volatile, ranging from very fast growth at program startup to slight decreases in expenditures in 1998-1999 and, until the last few years, tending to average at least 10 percent per year.

Figure 2

Medicare Expenditures and Average Annual Rate of Growth, by Era: 1966-2000

Five subperiods of relatively faster or slower Medicare growth trends are identified in Figure 2. Over the years, expenditure growth has been affected by many factors, including:

Increases in the number of beneficiaries.

Legislative and regulatory changes (including both program expansions and provisions designed to slow cost increases).

General economic factors, including the rates of inflation and wage growth.

Changes in the utilization and intensity of health care services covered by Medicare.

The primary factors affecting Medicare expenditure growth rates in each of the five subperiods are summarized in Table 1. More detail on these factors is available in Gornick (1976), Gornick et al. (1985 and 1996), Helbing (1993), Christensen (1991), and Davis and Burner (1995).

Table 1

Key Factors Underlying Periods of Faster or Slower Medicare Expenditure Growth: 1966-2000

Calendar Years	Average Annual Percent Increase	Type of Factor
		Cost-Accelerating	Cost-Decelerating
1966-1969	32.3	Program startup; pent-up demand; increased access for minorities; rapid increases in hospital costs and SNF utilization.	None of consequence.
1970-1973	10.0	Growth in outpatient hospital costs.	Imposition of wage and price controls; 1973 increase in Part B deductible; declining skilled nursing facility utilization.
1974-1982	20.0	Eligibility extended to certain disabled persons and individuals with end stage renal disease (effective July 1973); removal of wage and price controls, and rapid general and medical inflation; increased physician, outpatient hospital, and home health care utilization; inpatient hospital intensity growth.	Declining skilled nursing facility utilization.
1983-1997	9.8	Major increases in skilled nursing utilization and cost per day and in home health services (beginning in late 1980s); increases in inpatient case mix and outpatient utilization.	Substantial decline in general inflation; prospective payment system for inpatient hospital services (effective October 1983), and reductions in inpatient payment updates; 22- to 30-month freeze on physician payment levels (effective July 1984); physician fee schedule and volume performance standards (phased in 1992-1996); 1982 and 1990 increases in Part B deductible.
1998-2000	12.2	Balanced Budget Refinement Act of 1999.	Balanced Budget Act of 1997; intensified efforts to combat fraud and abuse; very low general and medical inflation.

Based on preliminary estimate of increase for 2000.

NOTE: SNF is skilled nursing facility.

SOURCES: (Gornick, 1976; Gornick et al., 1985 and 1996; Helbing, 1993; Christensen, 1991; Davis and Burner, 1995; Board of Trustees of the Federal Hospital Insurance Trust Fund, 1966-2000; Board of Trustees of the Federal Supplementary Medical Insurance Trust Fund, 1966-2000.)

The very low average growth rate during 1998-2000 stands in marked contrast to the significantly higher rates experienced throughout Medicare's prior history. This change reflects three very favorable developments affecting Medicare expenditure growth: the implementation of the BBA, the impact of intensified efforts to address fraud and abuse in the Medicare program, and very low rates of general and medical inflation. The overall average growth rate of only 2.2 percent during this period, however, reflects slight declines in expenditures in 1998 and 1999, together with a significant rebound of an estimated 6.9 percent in 2000. Although detailed data are not yet available for calendar year 2000, the expenditure growth in that year is associated with the Balanced Budget Refinement Act of 1999, which eased certain of the provisions enacted in 1997, and with increased utilization of services.

The factors underlying Medicare expenditure growth can be analyzed in greater detail by considering HI and SMI separately. Figure 3 shows annual increases in HI expenditures by (1) growth in the number of beneficiaries, (2) general inflation, (3) medical inflation in excess of general inflation, and (4) all other factors. This last category includes any increases in the utilization of covered services and in the intensity of services. Any errors in measurement of price change will also be reflected in the “all other factors” category, as discussed later.

Figure 3

Annual Increase in Hospital Insurance Expenditures, by Source of Growth: 1968-2000

The number of HI beneficiaries has generally increased by roughly 2 percent annually, with two notable exceptions. First, above-average growth occurred in 1973 and the following several years, when disabled individuals and persons with end stage renal disease became eligible. Also, during the most recent few years, annual enrollment growth dropped to about 1 percent as a result of the relatively low birth rates experienced 65 years earlier during the Great Depression. As shown in Table 2, until recently, enrollment growth for aged beneficiaries had been remarkably stable in the 2-percent range, while growth in the number of disabled beneficiaries had fluctuated dramatically and generally exceeded the aged growth rate by a significant margin. This latter variation is attributable in part to economic factors but is also associated with changes in law, regulations, and administrative policy for the Social Security Disability Insurance program (Zayatz, 1999). (Persons under age 65 qualify for HI benefits if they have received Social Security or Railroad Retirement disability benefits for at least 2 years.)

Table 2

Average Annual Growth Rates in the Number of Aged and Disabled Medicare Beneficiaries: 1967-2000

Calendar Years	Type of Beneficiary
	Total	Aged	Disabled2
	Percent
1967-1970	1.7	1.7	—
1971-1975	4.0	2.2	34.6
1976-1980	2.7	2.3	6.6
1981-1985	1.8	2.0	-0.2
1986-1990	2.0	1.9	2.2
1991-1995	1.9	1.4	6.2
1996-2000	1.2	0.7	4.4

Medicare beneficiaries are defined as the average number of persons with Hospital Insurance and/or Supplementary Medical Insurance eligibility during the calendar year. Growth rate for 2000 is preliminary.

Includes persons eligible because of end stage renal disease.

Represents average annual growth from 1973 (the first year of disabled eligibility) to 1975.

SOURCE: Health Care Financing Adminisitration, Office of the Actuary, 2000.

General inflation is a major contributing factor to growth in health care costs, as one would expect. The very rapid inflation associated with the two oil price shocks in 1973-1975 and 1979-1980 is a prime example. Inflation fell substantially following the 1981-1982 economic recession and rebounded temporarily during the Gulf War in 1989-1990. In 1992 and later, inflation was relatively low, although it rose somewhat in 2000, again as a result of energy costs.

The prices paid for medical services have frequently increased at a faster rate than general inflation. In Figure 3, this excess medical inflation is shown as the difference between a chain-weighted index of personal health care costs and the chain-weighted price index for the gross domestic product (Bureau of Economic Analysis, 2000). Such excess price growth has fluctuated considerably since 1966 but was generally above 2 percent per year during the period 1983-1993. This factor has diminished significantly in the last few years.

The price measures used in this analysis are intended to provide only a broad illustration of the impact of medical inflation on Medicare expenditure growth. Ideally, separate price indexes would be calculated for HI and SMI and would reflect the actual Medicare payment updates for each category of service (inpatient hospital, physician, and so forth). Unfortunately, creation of Medicare-specific price indexes poses daunting technical challenges, especially for SMI. In the absence of such indexes, the industrywide inflation factors shown in Figure 3 should be considered illustrative.

In practice, actual Medicare price increases have frequently been set below the prevailing level of medical inflation. For example, payment updates for inpatient hospital care are based on the inpatient hospital input price index or market basket, but Congress has reduced the update below the market basket increase in many of the years since the prospective payment system (PPS) was enacted. The intent is to provide a strong financial incentive for health care providers to maximize efficiency in the provision of care. As another example, increases in the Medicare fee schedule for physicians are currently based on changes in physicians' input costs (such as staff wages, practice expenses, and materials costs), but a penalty or bonus is added based on whether aggregate expenditures for physician services are above or below specified target levels. Thus, the prices paid by Medicare for medical services are not always closely associated with health providers' underlying input costs or the prevailing price changes in the health sector, as used in Figure 3.

The final growth category shown in Figure 3 for HI expenditures is labeled “all other factors” and represents the difference between total growth and the annual increases in the number of beneficiaries, general inflation, and excess medical inflation. This residual growth factor thus reflects changes in utilization and intensity of covered services as well as changes in Medicare reimbursement rates that differ from total medical inflation. As indicated, the “all other factors” growth category has been quite volatile. Its variation is related in large part to legislative changes, most notably the introduction of the PPS for inpatient services in October 1983 and the sweeping changes mandated by the BBA Judicially mandated regulatory changes implemented in the late 1980s for skilled nursing care and home health care also contributed substantially to residual cost growth through 1997. Recent efforts to curb fraud and abuse appear to have contributed significantly to lower growth, first in the home health category and more recently with inpatient hospital costs (Savord, 1998).

The growth factors for SMI expenditures have shown a generally analogous pattern, as indicated in Figure 4. The number of beneficiaries has increased very similarly for both HI and SMI, and the same general and excess medical inflation factors are used here for both. SMI residual growth has been significantly larger than for HI in most years (4.5 percent versus 2.7 percent on average from 1967 to 2000), although the pattern of variation through time is similar. Key exceptions occurred with the introduction of the inpatient hospital PPS in 1984 and the introduction of physician payment reform in 1992. During each period, the respective program residual growth factors declined substantially. Following the BBA, HI residual growth declined, and SMI residuals increased somewhat, as a result of the transfer of a majority of home health services to SMI (Foster, 1998).

Figure 4

Annual Increase in Supplementary Medical Insurance Expenditures, by Source of Growth: 1968-2000

The fact that SMI expenditures have grown at a significantly faster pace than HI expenditures, on average, is attributable to a number of factors. First, many procedures, such as cataract removal, can now be performed in an outpatient setting (covered by SMI) rather than inpatient (covered by HI). Second, the legislation governing Medicare payments has arguably been tougher on institutional providers (largely HI-covered) than on physicians (covered under SMI). During much of Medicare's history, for example, hospitals were reimbursed for their “reasonable costs,” while physicians received their “customary or prevailing charges.” Since 1984, the inpatient hospital PPS, as already noted, has frequently limited payment updates below medical inflation. Most recently, as part of the BBA, Congress shifted roughly two-thirds of home health care costs from HI to SMI, in an effort to help delay the depletion of the HI Trust Fund. In fact, one could reasonably suppose that, in general, lower HI cost increases have historically reflected congressional efforts to forestall HI asset exhaustion, whereas less pressure has been brought to bear on SMI costs because they are automatically financed (as described further in the following section).

Overall, this review of Medicare expenditure trends shows a pattern of relatively rapid growth in most years, with occasional periods of slower growth attributable to important legislative or administrative initiatives. The recent elimination of most remaining cost-based reimbursement provisions by the BBA should contribute to more restrained growth in the future. Growth in the number of beneficiaries will accelerate, of course, with the retirement of the post-World War II baby-boom generation starting in about 2010. Also, the public will continue to demand—and likely receive—the benefits of new medical technologies as they are developed. Under present rules, further increases in the number of beneficiaries in Medicare+Choice managed care plans will have little impact on cost increases, because payment updates for such plans are based largely on increases in national Medicare fee-for-service costs. Thus, future Medicare expenditure growth, while less rapid than in the past, is still likely to continue to outpace the growth in workers' wages, self-employment income, and Federal general revenues for many years.

Review of Medicare Financial Status

From the very beginning of the Medicare program, Congress has required an annual assessment of the actuarial status of the HI and SMI Trust Funds. This requirement is based on the recognition that Medicare makes important financial commitments to current and future beneficiaries and that the government has an obligation to ensure that these commitments can reasonably be met in practice. The annual reports of the Medicare Boards of Trustees to Congress include detailed short- and long-range projections of the trust funds' future financial operations, together with an assessment of each trust fund's financial status (Board of Trustees of the Federal Hospital Insurance Trust Fund, 2000; Board of Trustees of the Federal Supplementary Medical Insurance Trust Fund, 2000). (Technically, there are two separate boards—one for the HI Trust Fund and another for the SMI fund. In practice, the memberships of the boards have always been identical to each other and to the Board of Trustees for the OASDI trust funds.)

The trustees note that the purpose of these projections is not to predict the future with certainty, which is obviously impossible, but rather to illustrate how the Medicare program would operate under specified reasonable economic, demographic, and health cost trends. Projections are shown under three alternative sets of assumptions to illustrate the uncertainty inherent in the estimates and to provide a test of sensitivity to the various assumptions.

In recent years, a few individuals have criticized the long-range Medicare projections, stating that estimates beyond 10 or 25 years are too uncertain to be useful. I would argue, however, that this view undervalues the role of long-range projections and that it would be inappropriate not to make such projections. The HI program, for example, routinely makes financial commitments that can easily span the next 75 years. Young persons starting employment today at age 20 are assured that, in exchange for their HI payroll taxes now, they will qualify for HI benefits at age 65—some 45 years from now. Moreover, many of these individuals will live for as long as another 30 years or more thereafter. Accordingly, we should make every effort to ensure that the promises made today can actually be fulfilled in the future. To do otherwise could easily lead to overcommitment, future cutbacks in promised coverage, and (further) public distrust of government. I believe it is to Congress' credit, and that of every Administration since Medicare was enacted, that they have recognized and taken seriously the need to evaluate the long-range financial status of the program.

Without doubt, the long-range projections are sensitive to assumed future economic, demographic, and medical cost trends. However, we have an excellent idea of how many beneficiaries there will be for the next 65 years, because these individuals have already been born, and life expectancy tends not to change dramatically over time. We also have a fairly good idea about how the use of health care changes by age. Health expenditure growth can be volatile, as evidenced by the preceding discussion in this article. By relating costs to the underlying source of program income, such as taxable payroll or gross domestic product, we can obtain useful relative measures that minimize the volatile effects of future inflation. Finally, when the projections point to a serious long-range financial imbalance under a very wide range of reasonable assumptions, as is the case in the 2000 HI Trustees Report, then I believe it is sound public policy to address the imbalance and inappropriate to ignore it on the grounds that projections are inherently uncertain.

By way of comparison, long-range projections are routinely made for private pension plans. As actual wage increases, investment returns, mortality and disability rates, employee withdrawals, and other factors diverge over time from the actuarial assumptions, frequent small adjustments are made to the plan contributions to restore financial balance in the long range—and to prevent the need for potentially wrenching, last-minute changes. This model was also followed (albeit a little more loosely) for the Social Security program through the early 1980s.

The criticism of long-range projections may be based in part on the sensitivity of asset projections for the HI Trust Fund. The estimated year of depletion for HI has varied substantially throughout Medicare's history, most often attributable to new legislation affecting the program's financial status but sometimes following revisions in economic or other assumptions. In the short range, the estimated year of depletion is an important indicator of a trust fund's status. Over longer periods, however, it can change substantially as a result of even modest changes in assumptions. The change in assets during a year represents the difference between two very large amounts: total income and total expenditures for the year. In any such circumstance, as analysts have cautioned for centuries, the difference will be very sensitive to relatively small changes in either or both of the large numbers giving rise to the difference. Studies have indicated that the long-range “actuarial deficit” used by the trustees is a considerably more stable (and informative) measure of trust fund financial status than the year of depletion (Foster, 1989). Accordingly, the actuarial deficit is used in this article to summarize changes in the financial outlook for the HI Trust Fund over time.

Figure 5 summarizes the estimated long-range actuarial deficit for the HI program as it was presented in each of the Trustees Reports for 1966 through 2000. The deficit represents the amount by which projected HI tax income falls short of projected expenditures, on average, over the next 25 or 75 years. Prior to 1984, HI projections were shown only for the next 25 years. Beginning in 1984, the Board of Trustees adopted the practice of showing full 75-year projections to recognize the financial implications of the baby boom's retirement and to match the longstanding practice for the Social Security program.

Figure 5

Projected Long-Range Actuarial Deficits for Hospital Insurance Trust Fund: 1966-2000

As indicated in Figure 5, the HI Trust Fund was initially estimated to be in actuarial balance in the 1966 through 1968 Trustees Reports. As utilization and medical price growth rapidly exceeded assumed rates, however, the 25-year deficit rose to about 0.75 percent of taxable payroll, despite amendments to increase HI payroll taxes. The deficit was brought back to approximately zero by the Social Security Amendments of 1972, which raised the HI payroll tax rate significantly and indexed the amount of earnings subject to the tax for future years.

Soon after, the projected actuarial deficit rose again, as the rapid inflation and other factors described previously contributed to benefit growth that substantially exceeded the increases in workers' wages and salaries. This unfavorable situation was compounded by the economic recession of 1974-1975. The actuarial deficit stabilized temporarily during 1978-1980, primarily as a result of the significant increases in the maximum wage base for HI taxes enacted by the Social Security Amendments of 1977 and the recovery from the prior economic recession. Before long, however, growth in HI taxable payroll was adversely affected by the recessions in 1980 and 1981-1982, and expenditures were driven by the highest rates of inflation experienced since World War II.

The financial outlook improved considerably with the deceleration in inflation after 1982, together with the major legislative changes enacted as part of the Social Security Amendments of 1983. In addition to the inpatient hospital PPS, these changes included an increase in the HI payroll tax for self-employed workers, mandatory coverage of non-profit employees, and prevention of coverage termination by State and local government employees. (In addition, coverage of Federal Government employees was mandated by the Tax Equity and Fiscal Responsibility Act of 1982.) As the economy rebounded strongly and experience with the inpatient PPS developed, the projected 25-year HI deficit declined steadily until reaching about 0.50 percent of payroll in 1988. This improvement also reflected enactment of several provisions reducing PPS payment updates during this period and mandating coverage of newly hired State and local government employees.

As noted previously, the trustees introduced 75-year projections for HI in 1984. The projected deficits for the longer period were substantially greater than for the first 25 years, reflecting the financial impact of the baby boom's retirement. The trend in the estimated 75-year deficit, however, largely matched that for the 25-year estimates.

The Medicare Catastrophic Coverage Act of 1988 established extremely complex financing provisions, including a supplemental beneficiary premium in the form of an income tax surcharge. As a result of these complexities, it was not possible to prepare long-range financial estimates for HI for the 1989 Trustees Report. The Catastrophic Coverage Act was subsequently repealed, in large part because of beneficiary displeasure with these same financing provisions.

Projected deficits began rising once again in the early 1990s as expenditures for skilled nursing facilities, home health care, and hospice services increased by as much as 40 to 50 percent annually. Simultaneously, the average complexity of inpatient hospital admissions increased faster than anticipated, and the economic recession of 1990-1991 reduced growth in payroll tax income. Legislation in 1990 raised the HI maximum wage base to $125,000 (substantially above the level imposed for OASDI, for the first time), but this change was not sufficient to offset the other factors that were adversely affecting the projected actuarial deficit. The 25-year deficit during 1993-1997 climbed as high as 2.10 percent of payroll, approximately matching the highest level previously projected (in the 1982 Trustees Report). The corresponding 75-year deficits were more than 4 percent of payroll. If such levels had continued, HI payroll tax rates would have to have doubled or expenditures been cut in half (or some combination) to close the deficit.

This dire financial situation was addressed in a number of ways. First, the Omnibus Budget Reconciliation Act of 1993 significantly increased financing by eliminating the maximum wage base for HI, so that the HI payroll tax applied to all earnings without limit. The same act increased the amount of Social Security benefits subject to Federal income taxes and allocated the additional revenue to the HI Trust Fund. Payment updates to hospitals, skilled nursing facilities, and home health agencies were also trimmed. At about the same time, a combined initiative by the Health Care Financing Administration, the Office of Inspector General in the U.S. Department of Health and Human Services, and the Department of Justice sought to reduce fraud and abuse in the Medicare program—particularly in the provision of home health services. These efforts had a major impact. Growth in home health expenditures, for example, declined from an average annual rate of 37 percent in 1990-1995 to 8.5 percent in 1996 and to -2.3 percent in 1997. The financial outlook began to improve further as a result of the reduction in inflation from previous levels, together with strong growth in payroll tax revenues associated with the surging economy starting in about 1997.

Even with the beginning of these favorable developments, the financial outlook for HI was still poor. This situation, together with the desire to reduce overall Federal budget deficits, led to enactment of the BBA. This sweeping legislation reduced payment updates for virtually all health care providers in 1998-2002 and replaced the remaining cost-based Medicare reimbursement mechanisms with prospective systems. The net effect of the BBA, combined with the simultaneous low inflation, rapid economic growth, and gains in combating fraud and abuse, resulted in the lowest projected 25- and 75-year actuarial deficits in more than two decades. In the 2000 Trustees Report, the 25-year deficit is a virtually insignificant 0.12 percent of payroll. For the longer range, however, the 1.21-percent deficit remains well outside the trustees' allowable margin, despite the substantial improvement relative to the 1997 and earlier projections.

The financial status for the SMI Trust Fund is considerably easier to describe than that for HI. In short, beneficiary premiums and general revenue financing for SMI are revised annually to match the following year's estimated expenditures. A modest trust fund is maintained to cover differences between actual and estimated expenditures and to provide assets sufficient to meet any incurred-but-unpaid claims that would be left outstanding in the unlikely event that the program terminated. Thus, the SMI program is “automatically” in financial balance under present law over any future period. The official evaluation of actuarial status focuses on the relatively narrow issue of whether current assets are sufficient to meet outstanding claims and to provide a sufficient contingency reserve.

The Board of Trustees emphasizes that, despite the program's inherent financial balance, the rate of growth in SMI expenditures remains a serious concern. SMI costs have grown faster than those for HI in most past years, with the differential averaging 2.0 percent per year from 1967 through (estimated) 2000. The rapid increase in expenditures places a growing burden on beneficiaries, who finance approximately one-quarter of SMI through monthly premiums, and on Federal general revenues, where SMI has represented a steadily increasing share of the Federal budget.

Conclusion

Medicare has served the Nation well during its long history, despite the program's somewhat unusual coverage, eligibility, benefit, and financing provisions. The program's many benefits, however, have come at a cost that has grown quite rapidly more often than not. It seems unlikely that the Nation would be able to support such growth indefinitely, especially as the baby-boom generation reaches retirement age and becomes eligible for Medicare. The new payment systems established by the BBA will add to the pressures on the health care industry to provide care in a cost-efficient manner, especially compared with the prior cost-based reimbursement systems. Even so, technological advances and demographic changes will likely continue to drive costs at a faster rate than the taxable payroll or gross domestic product that underlie Medicare financing.

Therefore, our collective challenge will continue to be how best to balance the Nation's needs for high-quality and comprehensive health care with a cost that the Nation can afford. In the process of deliberating and deciding these issues, I would argue that greater attention should be placed on establishing long-range financial balance for Medicare. Although the HI program has been out of balance far more often than it has been in and has operated fairly successfully despite the prospective financial imbalance, there are significant consequences. First, delay in addressing deficits can lead to rushed action and inadequate time for consideration of how to address the problem most effectively. Working in crisis conditions does accomplish change but is not usually conducive to the most thoughtful or optimal solutions.

In addition, a change that is developed well in advance of a critical financial situation can be implemented more gradually, allowing us to avoid a sudden and drastic shift at the last minute with little warning. Even the range of possible solutions is greater when considered early. At the last minute, many potentially useful changes may not be feasible.

Finally, we risk doing a great disservice to program participants if we hold out the promise of Medicare benefits with specified eligibility, coinsurance, and tax requirements when, in fact, the promises cannot be realistically fulfilled without significant changes in one or more of these provisions. When the need for change is apparent, changes should be implemented with as much advance notice as possible, thereby allowing beneficiaries, workers, and health care providers an opportunity to adjust their expectations and plans. Public confidence in government and government programs is enhanced by their efficient operation and freedom from crises—especially those foreseeable many years in advance.

Did the Balanced Budget Act of 1997 Do Too Much?

It is often alleged that the BBA has had a far greater impact on Medicare payments to health care providers than Congress intended or anticipated. Is this accurate?

This question is not easy to answer. It is straightforward to know Medicare expenditures under the BBA, but no one can determine exactly what payments would have been without this legislation. For example, in the absence of the skilled nursing facility prospective payment system, effective July 1, 1998, what increases in the average cost per day would have occurred? And would the number of days of care have been the same? The best that can be done is to estimate these factors, but we would not expect an updated estimate of the BBAs financial impact—with one notable exception—to be substantially different from our original estimates in August 1997.

One can argue, fairly persuasively, that in many instances the actual impact of the BBA was exactly what Congress intended. For example, the legislation specified that Medicare payment rates for inpatient hospital admissions in 1998 would be frozen at their 1997 level—and this is precisely what occurred. Similarly, the payment update limitations mandated for other health providers were all implemented as specified. The actual savings from these provisions would be somewhat lower than originally estimated, because subsequent low inflation rates have reduced the “base level” against which these reductions apply. But again, we believe the overall impact would not be very far off from our original estimates.

It is certainly true that Medicare expenditure levels were significantly lower in 1998-2000 than we originally estimated at the time the BBA was enacted. To a great degree, however, this reduction is attributable to lower inflation rates and more success at addressing fraud and abuse than we had anticipated. Inflation, for example, averaged only about 2 percent in 1998-2000, rather than the 3.3 percent that was assumed at the time the BBA was enacted. The lower inflation carried through to many of the price indexes used to adjust Medicare reimbursement amounts, with the result that actual expenditures were significantly lower than previously estimated (roughly $6 to $7 billion in 2000), for reasons not associated with the BBA.

As another example, the inpatient hospital case-mix index decreased by 0.5 percent in 1998 and again in 1999, the first time this index had ever declined in the 16-year history of the inpatient prospective payment system. Analysis suggests that the decline is primarily attributable to changes in the coding of certain hospital admissions, particularly shifts in coding from “respiratory infection” to “simple pneumonia,” and from cases “with complications” to those “without complications” (Savord, 1998). Not coincidentally, these coding categories were the focus of a recent investigation by the U.S. Department of Justice. These behavioral changes had a very substantial impact on Medicare expenditures in 1998-1999 (roughly $3 billion in 1999) and had nothing to do with the BBA.

There is one clear area in which the impact of the BBA was, in fact, dramatically greater than anticipated. The number of home health care visits in 1999 was less than one-half of the level in 1997. Although a significant decline was expected—and appropriate in view of the excessive and often fraudulent billing for these services in recent years—this change is still dramatic. The reasons for the abrupt fall-off in services are not yet fully evident. The continuing program integrity efforts certainly have had an impact, and available evidence suggests that many home health agencies have misinterpreted the requirements of the BBA or are purposely erring on the conservative side to avoid the possibility of large retroactive settlements to Medicare. Even with these considerations, however, it is likely that the interim payment system has resulted in much of the abrupt decline in services and that this impact is significantly greater than Congress intended.

In view of the uncertainty associated with the factors underlying the slow growth in Medicare costs during 1998-2000, Congress and the Administration have focused on beneficiary access to care as an important indicator of whether reimbursement levels are too low following the BBA. This approach led to a number of targeted adjustments in the Balanced Budget Refinement Act of 1999, with a modest increase in Medicare expenditures in 2000 and later.