General revenue financing of Medicare: who will bear the burden?

Two recent national advisory committees on Social Security recommended major shifts in Medicare financing to preserve the financial viability of the Social Security trust funds. This paper estimates the income redistribution consequences of the two proposals, in contrast to current law, using a micro-simulation model of taxes and premiums. These estimates show that while the current Medicare financing package is mildly progressive, the new proposals would substantially increase income redistribution under the program. Two insights provided by separate estimates, for families headed by the elderly (persons age 65 or over) versus those headed by the non-elderly, are: 1) the surprisingly large Medicare tax burdens on families headed by the elderly under the current financing package of payroll taxes, general revenues, and enrollee premiums; and 2) the substantial increases in these burdens under proposed shifts toward increased general revenue financing.

Introduction

Two recent national advisory committees on Social Security recommend major shifts in Medicare financing to preserve the financial viability of the Social Security trust funds. This paper estimates the income redistribution consequences of the two proposals, concentrating especially on the often-ignored financing burdens imposed upon the elderly.

Background

Medicare benefits and administrative costs are paid from two trust funds created under Title XVIII of the Social Security Act, passed in 1965. The Hospital Insurance (HI) Trust Fund, covering hospital and other institutional benefits, is now financed mainly through payroll taxes on employers and employees. The Supplementary Medical Insurance (SMI) Trust Fund, which pays for physicians' and other outpatient services, is financed through Federal general revenues and through monthly premiums paid by SMI participants. In fiscal year 1982, these trust funds will finance expenditures of approximately $33 billion and $15 billion, respectively, providing basic insurance coverage to over 28 million elderly and disabled beneficiaries.

The Old-Age and Survivors Insurance (OASI) program and the Disability Insurance (DI) program, together with the Medicare program, form the Social Security system that provides financial security for workers and their families. Cash benefits under OASI and DI are paid from separate trust funds financed by employer-employee payroll taxes. This mode of financing creates a highly visible link between benefits received and “contributions” paid, giving the system the politically appealing aura of private insurance. The initial decision in 1965 to use earmarked payroll taxes to finance the HI trust fund can be viewed as a logical expansion of the Social Security system that was already in place. With few exceptions, HI eligibility is dependent upon eligibility for Old Age or Disability cash benefits; that is, past payroll tax contributions are linked to the promise of future health insurance benefits upon disability or retirement.

Adverse economic conditions since the 1977 Social Security amendments are leading toward a short-run financing crisis for the OASI trust fund. Moreover, long-run demographic and medical cost trends lead to projections of financial crisis for the HI trust fund within a decade (Bartlett, 1980). Two recent national advisory committees have studied the overall viability of the Social Security system and made a number of recommendations for reform. These recommendations include the following proposals for major changes in Medicare financing:

The National Commission on Social Security—a bipartisan commission of private citizens which was established under the 1977 amendments to report to the President and the Congress—recommends in its 1981 report that, beginning in 1983, half of the HI trust fund be financed from general revenues, and that the freed HI payroll tax revenues be applied to the OASDI programs (“National Commission on Social Security: Recommendations,” 1981).

The 1979 Advisory Council on Social Security—appointed in 1978 as the fifth quadrennial advisory council to the Congress and the Trustees of the Social Security trust funds—recommends in its 1980 report that the HI trust fund be financed entirely from Federal general revenues, and that a portion of the freed payroll taxes be directed to OASDI, with the balance repealed (“Reports of the 1979 Advisory Council…,” 1980; and “Social Security Financing,” 1980).

While these proposals are not unprecedented (the 1971 and 1975 Advisory Councils also proposed substitution of general revenues for HI payroll taxes), they are only now being taken seriously by the Administration and the Congress.

One important dimension for analysis of these major financing shifts is their effect upon the income distribution. The 1979 Advisory Council made its values quite clear in saying that a “major shortcoming” of the payroll tax as the sole source of funds for OASDHI benefits is that it “falls more heavily on persons with relatively low incomes than does the personal income tax” (“Social Security Financing,” 1980, p. 22). The primary purpose of this paper is to estimate the magnitudes of Medicare financing burdens for families of various income levels, under current law and under each of the two recent proposals, demonstrating precisely how each proposal redistributes these burdens.

Recently, other observers of Medicare have come to question whether the program has fulfilled its initial promise to the elderly (U.S. House of Representatives, 1980; and Health Research Group, 1980). In particular, substantial concern exists over the total amount the elderly and disabled pay through cost-sharing, unassigned physician bills, and SMI premiums. The analysis in this paper widens the scope of that concern by estimating the Medicare tax burdens on families headed by the elderly. These burdens, which are strikingly large—especially under the proposed shifts to general revenue financing—have not previously been estimated.

Methodology

The methodological approach of this paper involves a micro-simulation model. We processed survey data on individuals and families, weighted to represent the U.S. population, through a set of micro-simulation routines to estimate the distribution of family burdens from each of the relevant taxes and premiums. We then apportioned the fiscal year 1982 projected Medicare revenue requirement from each financing source among families in proportion to their estimated tax burdens. We then aggregated these total family contributions to Medicare by quintile of the family income distribution. This same approach is used for each of three financing packages: that of current law, that recommended by the National Commission on Social Security, and that recommended by the 1979 Advisory Council on Social Security. Table 1summarizes the respective options.

Table 1

Medicare Financing Packages

	Current Law	National Commission on Social Security Recommendations1	1979 Advisory Council on Social Security Recommendations2
Hospital Insurance Trust Fund	Payroll Taxes3	½ Payroll Taxes	General Revenues4
		½ General Revenues
Supplementary Medical Insurance Trust Fund	General Revenues	Unchanged	Unchanged
	Enrollee Premiums

“National Commission on Social Security: Recommendations,” (1981).

“Reports of the 1979 Advisory Council on Social Security,” (1980) and “Social Security Financing,” (1980).

The current financing includes about 9 percent general revenues.

The recommendations involve an earmarked surcharge to the Federal personal income tax to preserve worker impressions of a contributory fund, with the balance coming from Federal general revenues. The council gave ambiguous percentage contributions from each tax.

Microdata

The 1976 Survey of Income and Education (SIE) is the primary data source on which the tax and premium simulations are performed. The SIE is a file of 151,170 household interview units, containing the records of 440,815 individuals residing in the 50 States and the District of Columbia (U.S. Department of Commerce, 1978). The March questionnaire for the Current Population Survey forms the core of the SIE questionnaire, providing information on labor force status, work experience, household structure, and demographic characteristics. The SIE also covers more detailed breakdowns of calendar year 1975 income by type, transfer program participation (for example, Aid to Families with Dependent Children), health insurance coverage, disability conditions, languages spoken, and educational characteristics. When survey case weights are used, the sample represents the total civilian noninstitutionalized population of the United States. These features of the SIE allow the micro-simulation of tax burdens according to their appropriate income bases and they also allow construction of the family income distribution of the U.S. population.

Microsimulation of Tax Incidence

Federal tax burdens are simulated using a version of the FEDTAX (that is, Federal personal income taxes) module of the Transfer Income Model (TRIM), originally developed by the Urban Institute (Moeller, 1973; and Sulvetta, 1976) and modified by the authors to address special issues in health care financing. In apportioning taxes (which may statutorily fall on firms and other organizations) to specific individuals or families, assumptions regarding the shifting of tax burdens (incidence assumptions) are crucial. We follow a widely accepted approach which presumes perfect competition, price flexibility, and factor mobility (Pechman and Okner, 1974, pp. 25-43). Moreover, throughout the paper we assume that the only budgetary change taking place is in the package of Medicare financing sources; that is, we employ the comparative static methodology common to many micro-simulation exercises. This section highlights the modified FEDTAX procedure for each important Medicare revenue source.

We assume that payroll taxes are ultimately borne by workers—the employee tax borne directly and the employer tax borne in the form of lowered wages. Using appropriate payroll tax rates and tax base ceilings, we performed calculations for private wage and salary earners, the self-employed, and railroad workers. Although the Social Security system does not mandate coverage of government employees, about 70 percent of all nonfederal government workers participate in the system through voluntary State and local government buy-ins. A random 70 percent sample of the SIE individuals working for State and local governments, then, were also charged payroll taxes.

We also assumed that SMI premiums were borne by the individual premium payor, except in the case of Medicaid “buy-ins” for the low income elderly. Thus, we assigned Medicare enrollees not covered by Medicaid their per capita share of the projected FY 1982 premium income to the SMI trust fund. We assigned the premium sum for Medicare enrollees covered by Medicaid to Federal general revenues.

Federal general revenues derive from a number of sources, although the bulk (94 percent) can be attributed to just three—personal income taxes, corporate income taxes, and excise taxes. In FY 1982, the Federal personal income tax will account for about 70 percent of general revenues. Incidence of this tax is assumed to remain with the tax filing unit. A complex FEDTAX routine reorganizes SIE household interview units into realistic tax filing units and determines the type of return to be filed (single, joint, or unmarried head of household). The FEDTAX routine then computes adjusted gross income (AGI) by aggregating taxable income sources across members of the filing unit. The FEDTAX routine also assigns personal exemptions by examining the financial relationships in the family (such as considering money earned by dependents) and checking for additional old age and blindness exemptions. Based upon IRS statistics on actual returns (Internal Revenue Service, 1978), a proportion of filing units in each income class is randomly chosen to itemize and is assigned the average itemized deduction for an itemizer in that income class. The standard deduction formula is applied to all other filing units. Taxes due are determined by calculating AGI, less exemptions and deductions, and referring to the appropriate Tax Rate Schedule (X, Y, or Z). For certain filers, this tax liability was reduced by an earned income credit.

In FY 1982, the Federal corporate income tax will account for about 20 percent of general revenues. The incidence of this tax is assumed to ultimately rest with all owners of financial and real property. Property income is assigned to SIE individuals on the basis of the following items: interest, dividends, rent, self-employment income, and pension income, according to Department of the Treasury (1975) estimates regarding the proportion of each source that represents a return to capital. Total corporate taxes are then assigned to individuals proportionally to their property income.

The remaining 10 percent of general revenues is comprised chiefly of excise taxes, but is also comprised of estate and gift taxes, custom duties, and other miscellaneous receipts. We assume that these taxes fall on consumers in proportion to their disposable income. Disposable income for SIE families is calculated as net income after personal incomes taxes.

Estimates of Family Financing Burdens

The final steps of the methodology involve apportioning specific program revenue requirements across families by tax and premium incidence and aggregating the results by income, age, or other classification. The estimates in this paper are based upon actuarial projections of program benefits, administrative costs, and receipts for fiscal year 1982 (Board of Trustees of the Federal Hospital Insurance Trust Fund, 1981; and Board of Trustees of the Federal Supplementary Medical Insurance Trust Fund, 1981). Our HI and SMI estimates are based upon total program costs (benefits plus administration) of about $33 billion and $15 billion, respectively. Each program total is divided among the financing sources as indicated in Table 1. We based current law amounts upon the actuarial projections of the share contributed by each revenue source (for example, SMI is estimated to depend upon general revenues for 79 percent of its financing and upon premiums for 21 percent of its financing in FY 82). General revenue financing is further divided into its three major component taxes by their estimated shares in total Federal general revenues in FY 82.

Once we calculate the control total for each financing source, that total is distributed over all families in proportion to their simulated share of the total burden of the respective financing source. In effect, we use the micro-simulation results as weights for allocating aggregate program costs. To summarize the results, we aggregate the burdens over families classified by characteristics of interest, reporting the results as an average burden per family or the burden as a percent of family income. In this latter calculation we projected the SIE 1975 income levels to 1982 to correspond to the FY 82 control totals. In classifying families by income, we order families into quintiles by total family income (that is, in five groups from the lowest 20 percent to the highest 20 percent). This approach, which is common in the literature on income distribution, avoids putting the results in dollar denominations for a particular year, thereby stating results for relative income that remain approximately correct over several years, since the relative position of a family in the income distribution is slow to change, even though money incomes change rapidly in periods of inflation.

Findings

Tables 2 through 4 present a summary of the micro-simulation results. These tables compare the distribution of Medicare financing burdens under current law, under the National Commission proposal requiring 50 percent general revenue financing of the HI trust fund, and under the Advisory Council proposal requiring 100 percent general revenue financing of the HI trust fund. Tables 3 and 4 separate the results for families headed by elderly and non-elderly persons.

Table 2

Medicare Program: Distribution of Financing Burdens by Family Income, FY 82 Dollars Per Family and Percent of Family Income

Family Income Quintiles1	Current Law			National Commission Proposal		Advisory Council Proposal
	Total	HI	SMI2	Total	HI	Total	HI
All Families	$ 729 (2.6%)	$ 501 (1.8%)	$228 (0.8%)	$ 729 (2.6%)	$ 501 (1.8%)	$ 729 (2.6%)	$ 501 (1.8%)
1	116 (2.2)	41 (0.8)	75 (1.4)	111 (2.1)	36 (0.7)	105 (2.0)	30 (0.6)
2	312 (2.3)	193 (1.4)	119 (0.9)	289 (2.1)	170 (1.2)	260 (1.9)	141 (1.0)
3	593 (2.6)	439 (1.9)	154 (0.7)	540 (2.4)	386 (1.7)	476 (2.1)	322 (1.4)
4	945 (2.7)	713 (2.1)	232 (0.6)	875 (2.5)	643 (1.9)	790 (2.2)	558 (1.6)
5	1,679 (2.7)	1,118 (1.8)	561 (0.9)	1,829 (2.9)	1,268 (2.0)	2,013 (3.2)	1,452 (2.3)

Quintile 1 represents the lowest 20 percent of families ranked by total income. Successive quintiles represent the next lowest 20 percent of families, up to quintile 5, which includes the highest income families.

SMI financing remains unchanged under the National Commission and Advisory Council proposals.

Table 4

Medicare Program: Distribution of Financing Burdens by Family Income for Families Headed by Non-elderly Persons, FY 82 Dollars Per Family and Percent of Family Income

Family Income Quintiles2	Current Law			National Commission Proposal		Advisory Council Proposal
	Total	HI	SMI3	Total	HI	Total	HI
All Families	$ 796 (2.6%)	$ 588 (1.9%)	$208 (0.7%)	$ 773 (2.6%)	$ 565 (1.9%)	$ 744 (2.5%)	$ 536 (1.8%)
1	80 (1.5)	63 (1.2)	17 (0.3)	62 (1.2)	45 (0.9)	41 (0.8)	24 (0.5)
2	312 (2.2)	253 (1.8)	59 (0.4)	259 (1.8)	200 (1.4)	193 (1.4)	134 (1.0)
3	595 (2.5)	475 (2.0)	120 (0.5)	518 (2.2)	398 (1.7)	424 (1.8)	304 (1.3)
4	952 (2.7)	740 (2.1)	212 (0.6)	864 (2.5)	652 (1.9)	758 (2.2)	546 (1.6)
5	1,669 (2.7)	1,141 (1.8)	528 (0.9)	1,783 (2.9)	1,255 (2.0)	1,922 (3.2)	1,394 (2.3)

Non-elderly heads of families include all those under 65 years of age.

Income quintiles are based on the income distribution for all families. See footnote 1, Table 2.

SMI financing remains unchanged under the National Commission and Advisory Council proposals.

Table 3

Medicare Program: Distribution of Financing Burdens by Family Income for Families Headed by Elderly Persons, FY 82 Dollars Per Family and Percent of Family Income

Family Income Quintiles2	Current Law			National Commission Proposal		Advisory Council Proposal
	Total	HI	SMI3	Total	HI	Total	HI
All Families	$ 449 (2.6%)	$133 (0.8%)	$ 316 (1.8%)	$ 549 (3.1%)	$ 233 (1.3%)	$ 670 (3.8%)	$ 354 (2.0%)
1	174 (2.9)	10 (0.2)	164 (2.7)	188 (3.1)	24 (0.4)	205 (3.4)	41 (0.7)
2	336 (2.4)	60 (0.4)	276 (2.0)	383 (2.8)	107 (0.8)	441 (3.2)	165 (1.2)
3	545 (2.4)	178 (0.8)	367 (1.6)	651 (2.9)	284 (1.3)	778 (3.4)	411 (1.8)
4	863 (2.5)	371 (1.1)	492 (1.4)	1,050 (3.0)	558 (1.6)	1,277 (3.7)	785 (2.3)
5	1,822 (2.7)	801 (1.2)	1,021 (1.5)	2,450 (3.6)	1,429 (2.1)	3,217 (4.8)	2,196 (3.3)

Elderly heads of families include all those 65 years of age or older.

Income quintiles are based on the income distribution for all families. See footnote 1, Table 2.

SMI financing remains unchanged under the National Commission and Advisory Council proposals.

As Table 2 shows, the average family in fiscal year 1982 will contribute $729, or 2.6 percent of family income, to the Medicare program, should current financing provisions be continued. This rather substantial amount reflects both the continued aging of the population and health care costs which continue to rise faster than consumer income. The distribution of total program costs across income classes is progressive. Families in the lowest income quintile contribute an average of 2.2 percent of their income, while families in the highest quintile contribute 2.7 percent of their income.

Comparing the distribution of financing burdens for Medicare Part A and Part B highlights their differing tax bases. The HI trust fund is currently financed almost entirely by proportional payroll taxes. However, estimated HI family contributions as a percent of income will not be constant across quintiles because of variations in the proportion of wages in total family income and because at very high earnings levels payroll tax base ceilings are met. The SMI trust fund is currently financed by general revenues, which are progressive, and by per capita premiums, which are regressive. The distribution of SMI costs across quintiles in Table 2 combines these separate effects. However, for the premium-paying elderly (shown in Table 3) the SMI contribution is more clearly regressive, while for the non-elderly (shown in Table 4) the general revenue contribution is progressive.

Financing the identical Medicare benefit levels under the National Commission and Advisory Council proposals leaves the average dollar contribution of $729 per family unchanged, but the increased role of general revenues for HI substantially alters the distribution of costs across income classes (Table 2). Under the National Commission alternative, costs are shifted exclusively to the highest quintile (quintile five) and all other quintiles show reduced contributions. This effect is amplified under the Advisory Council proposal. While families in quintile five contribute 23 percent more per dollar of income than families in quintile one under current law in FY 82, families in quintile five would contribute 60 percent more per dollar of income than families in quintile one under total general revenue financing.

A surprising result emerges when we compare financing distributions for families headed by the elderly versus those headed by the non-elderly. Families headed by the elderly participate equally with those headed by the non-elderly in the on-budget financing of Medicare. This equality is demonstrated by equal 2.6 percent contributions of family income (Tables 3 and 4). Furthermore, as general revenue financing increases, the proportion of family income contributed by families headed by the elderly rises above that contributed by families headed by the non-elderly. Under the Advisory Council proposal, the average household headed by an elderly person pays 50 percent more on every dollar of family income than does the average household headed by a non-elderly person (3.8 percent versus 2.5 percent, respectively).

Viewing the financing burdens on the elderly in a slightly different way, families headed by persons 65 years of age or older comprise 20 percent of all families and receive 12 percent of all family income. Under current law applied to FY 82, these families will pay 12 percent of all Medicare program costs, largely as a result of their premium and general revenue contributions to SMI. This percentage is, coincidentally, their proportionate share of budgetary costs, but this does not include the contribution the elderly make to their own health care through cost-sharing provisions. Under the National Commission proposal, families headed by the elderly will pay over 14 percent of all Medicare program costs, and under the Advisory Council proposal they will pay almost 18 percent. The increased burden on the elderly under general revenue financing results from the differing components of family income for families headed by the elderly and non-elderly, which Table 5 illustrates. Families headed by non-elderly persons receive a disproportionate share (96 percent) of all wage income, while families headed by elderly persons receive a disproportionate share of all property income (34 percent under the special definition of property income cited previously). Approximately 68 percent of all pension income, and 38 percent of all asset income accrue to families headed by the elderly. Each of these components bears a burden of corporate and personal income taxes, but not payroll taxes.

Table 5

Sources of Income for Families Headed by Elderly and Non-elderly Persons

Income Source	Share of All Families' Receipts from Income Source			Income Source as a Share of Family Income
	Elderly	Non-elderly	All Families	Elderly	Non-elderly	All Families
Income From All Sources	11.8%	88.2%	100%	100%	100%	100%
Wages and Self-Employment Income	4.5	95.5	100	31.4	88.2	81.6
Asset Income2	37.8	62.2	100	18.7	4.1	5.8
Government and Private Pensions	68.1	31.9	100	45.7	2.9	7.9
Transfer Income3	10.4	89.6	100	4.2	4.8	4.7

Source: 1976 Survey of Income and Education

Non-elderly heads of families are under 65 years of age; elderly heads of families are 65 years and older.

Includes interest, dividends, and rents.

Includes AFDC, SSI, veterans' benefits, workers' compensation, unemployment compensation, and miscellaneous income.

What can be learned about the overall increase in financing progressivity, when the results are separated for the elderly and non-elderly? Table 3 shows that the distribution of SMI costs across income classes for the elderly is regressive (due to premiums). In contrast, the distribution of HI costs is progressive, even under current financing provisions. Rather than this progressivity increasing dramatically with the move to HI general revenue financing, all elderly income classes take on more of the financing burden (as a result of shifting the tax base from wages and salaries toward property income). However, the impact of the Advisory Council's proposal on progressivity for families headed by the non-elderly is dramatic. (See Table 4.) For this group, families in the highest quintile contribute an average of 400 percent more per dollar of income than families in the lowest quintile under full general revenue financing.

Conclusions

Two recent national advisory committees on Social Security recommend major changes in Medicare financing which would shift HI trust fund reliance from payroll taxes to general revenues. A move toward general revenues obviously increases the progressivity of Medicare financing; our analysis quantifies the redistributive impact of the two advisory committee proposals relative to current law for FY 82. We show that financing burdens under current law are moderately progressive, with families in the lowest quintile of the income distribution contributing 2.2 percent of income and families in the highest quintile contributing 2.7 percent. The National Commission proposal of 50 percent general revenue financing for HI shifts costs exclusively to the highest quintile, and lowers the burden on all other quintiles. This redistribution is furthered under the Advisory Council's proposal of full general revenue financing, in which contributions as a percent of family income range from 2.0 percent in the lowest quintile to 3.2 percent in the highest quintile. Under this proposal, the highest income families pay over 1.5 times the percent of income contributed by the lowest income families.

When separate income redistributive impacts are estimated for families headed by the elderly, our estimates reveal surprisingly large financing burdens. Under current law the average family headed by an elderly person is expected to contribute $449 toward Medicare financing in FY 82. Less than one-third of this amount is attributable to SMI premiums; the rest arises from payroll and general revenue tax contributions of the elderly that have not previously been estimated. The $449 contribution represents 2.6 percent of the average income of families headed by an elderly person, the same percent of income that is contributed by the average family headed by a non-elderly person. With shifts to general revenue financing of HI, the burden on the elderly rises dramatically. Under full general revenue financing, families headed by the elderly pay 18 percent of all Medicare costs. This averages to $670 per family—50 percent more per dollar of income than families headed by the non-elderly. This $221 increase will exceed the magnitude of the SMI premium under the current law. Thus, ironically, the study committees' proposals for increasingly progressive Medicare financing will dramatically increase the burden on elderly families, for whose benefit the program was enacted.