Reducing Government Subsidies for Wealthier Medicare Enrollees

What This Paper Covers
This paper examines current means-testing policies in Medicare and potential approaches for expanding it. Means testing in Medicare is not new, but its current scope is limited.

What We Found
Although Medicare is often described as a “universal” entitlement program (in contrast to welfare programs such as Medicaid), higher income beneficiaries are currently subject to higher premium payments. However, these “means-testing” policies apply to only about 8 percent of Part B beneficiaries, who still receive significant government subsidization of their health care costs. Roughly one in five Medicare beneficiaries have lower costs due to various assistance programs.

Why It Matters
Given that Medicare represents the nation’s primary fiscal challenge, and given rising federal debt and interest payments, building on existing but limited means-testing policies for wealthier beneficiaries could help shore up the program’s finances and improve the country’s budgetary outlook. Despite current means-testing policies, wealthier Medicare enrollees currently have 15 to 65 percent of their estimated Part B costs subsidized by taxpayers, and enrollees across income levels receive far more in benefits than they pay in payroll taxes during their working years and in premiums during retirement.

Policy Suggestions
Policymakers should reduce government subsidies for wealthier beneficiaries. This can be accomplished a few different ways, including by (1) adjusting the income thresholds at which means-testing applies or the additional premium amounts paid under current law (as proposed by numerous organizations in the past), (2) expanding the measures of wealth, including lifetime income, that are used to target means-testing policies, and (3) considering alternative mechanisms for means-testing besides beneficiary premiums.

Federal health programs are the biggest driver of government spending besides interest payments on existing debt. They are expected to increase as a share of the economy from 6 percent in 2023 to 9 percent in 2052.¹ Annual spending on Medicare alone will increase by about $930 billion over the next decade, from 14 percent to 17 percent of federal outlays.² This current trajectory increases not only costs for Medicare enrollees but also the potential for fiscal crisis. Medicare Part A is projected to become insolvent in 2031, which would necessitate an 11 percent across-the-board payment cut.³ Rising interest rates could also prevent the government from covering its debt obligations, leading to broader economic consequences.⁴ To avoid a debt crisis, policymakers should reduce federal health spending. One potential policy approach that would not cut benefits and has been endorsed by prominent members of both political parties is to reduce the public resources going to wealthier Americans by means-testing federal programs.

Means-tested programs usually target benefits to lower income and less wealthy participants. In 2019, about a quarter of all mandatory spending was means-tested.⁵ Medicaid and other welfare programs are entirely means-tested in that typically only low-income individuals are eligible, unlike social insurance programs such as Social Security and Medicare, which are available to anyone meeting age or disability criteria. However, some aspects of these “universal” programs are more generous for needier recipients than others. For example, Social Security Old-Age and Survivors Insurance benefit levels are based in part on a beneficiary’s income history and replace a higher percentage of pre-retirement income for those with lower lifetime earnings.⁶ By contrast, Medicare benefits do not change based on income.

Traditional Medicare has two major components: a hospital insurance program (Part A) and a supplemental medical insurance program (Part B), which mainly covers outpatient and clinician services. Enrollees can also enroll in private insurance plans for Part A and B coverage through Medicare Advantage (Part C) and for prescription drug coverage (Part D). Americans become eligible for Medicare at age 65 or earlier if they meet certain disability or health criteria. Those who have paid payroll taxes for a requisite period receive Part A coverage automatically without paying monthly premiums, while Part B coverage requires a monthly premium.⁷ Beneficiaries can choose to not enroll in Part B but may face penalties if they enroll later. Part C and D coverage are also voluntary, and premium costs depend on the plan selected; some Part C plans charge more than the Part B premium, while others are able to reduce it.

Premiums

The main form of means-testing in Medicare is the variance in premium amounts for Parts B and D. Standard beneficiary premiums cover only about 25 percent of the cost of these benefits overall, with the rest subsidized by taxpayers. Starting in 2007 for Part B and 2011 for Part D, Congress reduced these subsidies for higher income beneficiaries, requiring a modest increase in premiums based on the income-related monthly adjustment amount (IRMAA).⁸ Table 1 below shows the additional Part B and D premium amounts for beneficiaries in each income tier in 2024. The 2024 IRMAA starts for individuals with incomes greater than $103,000 per year ($206,000 for married couples with joint filing status). Beneficiaries with income below these levels pay the “standard premium” for each program—in 2024, $174.70 per month for Part B coverage and the premium for the Part D plans they select. Since 2020, the income thresholds have been updated annually for inflation.

As Table 1 shows, despite current means-testing, Part B and D premiums as a percentage of income decline as household income increases. Furthermore, the share of Medicare beneficiaries subject to these higher premium tiers is quite small. In 2023, the Centers for Medicare and Medicaid Services projected that the higher 2024 premiums would apply to about 8 percent of enrollees with Part B coverage and 8 percent of enrollees with Part D coverage.⁹ Figure 1 shows this breakdown for each income tier using income data from 2021, on which the 2023 premiums were based. For comparison, median annual household income among Part B beneficiaries that year was $35,000, and only 25 percent of them had income higher than $74,900. Overall, about 7 percent of Medicare beneficiaries would have faced higher premiums in 2023 based on their 2021 income, and only 0.2 percent of beneficiaries were in the highest premium bracket. Overall, this translates to roughly $13.4 billion in additional 2023 premium revenue from means-testing.¹⁰

Premium levels have important implications for Medicare’s overall financing. Similar to how Medicare benefits are separated into different parts, so are its funding sources. Part A benefits are financed by the Hospital Insurance (HI) trust fund, which is primarily funded by payroll taxes from current workers. (Contrary to a widely held belief, workers’ taxes do not “pre-fund” their future benefits.) Part B and D benefits are financed by the Supplemental Medical Insurance (SMI) trust fund, which is funded primarily by beneficiary premiums and transfers of general tax revenues. Since the SMI trust fund does not maintain a balance and the HI trust fund lends its excess balances to the Treasury, neither functions as an actual trust fund.¹¹

Medicare’s financing structure has changed over time. Originally, premiums from enrollees financed about 50 percent of Part B costs. Congress reduced this “premium percentage” for standard premiums to about 33.3 percent in 1994 and about 25 percent in 2005.¹² For higher income beneficiaries, premiums cover between 35 and 85 percent of average program costs.¹³ Overall, premiums have gone from funding 50 percent of the SMI trust fund in 1970 to 26 percent in 2022, and general tax revenue has gone from 49 percent to 75 percent during that time.¹⁴ Figure 2 shows that Medicare’s reliance on general tax revenue has significantly increased while payroll tax revenue funds a declining fraction of the program.

Taxes

Federal taxes fund most of Medicare, which also impacts the degree to which households at different levels of income finance the program. Despite proposals to increase Medicare funding through tax increases on the wealthy, such taxpayers already fund much of the program under current policy. The Medicare payroll tax is the primary source of HI trust fund revenue. This is a flat 2.9 percent tax split equally between employers and employees (although economic incidence shows that employees bear most of that amount in the form of lower compensation).¹⁵. Unlike the Social Security payroll tax, the Medicare tax applies to income at all levels with no cap. The Affordable Care Act also enacted an extra 0.9 percent Medicare surtax for individuals with annual compensation exceeding $200,000 ($250,000 for married couples) and a net investment income tax for those same income thresholds that goes to general revenues. These thresholds do not update for inflation and therefore apply to more households over time.¹⁶

As mentioned above, general revenues have become a more important source of Medicare funding. Excluding receipts earmarked to specific social insurance programs (i.e., payroll taxes), individual income taxes accounted for about 77 percent of federal receipts in 2023.¹⁷ Because individual income taxes are explicitly designed to be progressive, those with higher incomes tend to pay a higher share of such revenues, meaning they indirectly finance more of the SMI trust fund, in addition to paying higher premiums and Medicare-specific surtaxes. The top 50 percent of taxpayers by income paid 98 percent of all federal income taxes in 2020, and the top 10 percent of taxpayers paid 74 percent of all federal income taxes. Therefore, these groups financed about 54 and 40 percent, respectively, of all 2020 expenses for Part B and D benefits.¹⁸

Low-income Assistance Programs

Within Medicare, there are also several programs that provide additional financial aid or coverage to lower income beneficiaries. For example, Medicare beneficiaries who are eligible for Medicaid coverage in their states can receive assistance for their Medicare expenses and eduction in cost-sharing.¹⁹ Such “dual” Medicare-Medicaid enrollees also have access to Medicare-Medicaid plans or dual eligible special needs plans in Medicare Advantage, which can offer integration of benefits to varying degrees.²⁰ There were 11.8 million dual enrollees in 2021 (about 18 percent of all Medicare beneficiaries).²¹

Medicare beneficiaries below certain income and resource limits may also qualify for several Medicare Savings Programs (MSPs) in traditional Medicare. Dual enrollees are usually eligible for MSPs, and states typically administer them. The Qualified Medicare Beneficiary program helps pay for Part A and B premiums and Part B cost-sharing for the lowest income beneficiaries. The Specified Low-Income Medicare Beneficiary program helps pay for Part B premiums. The Qualifying Individual Program also does this, but participants must apply every year, cannot participate if they qualify for Medicaid, and receive state approval on a first-come, first-served basis. Finally, the Qualified Disabled and Working Individual program helps with Part A premiums for disabled beneficiaries who lose Social Security disability benefits and premium-free Part A because they are working. Table 2 below summarizes these MSPs.²² Overall, about 9.9 million Medicare beneficiaries (15 percent) were in MSPs in 2021.²³

Part D beneficiaries can also receive assistance with prescription drug costs through the Low-Income Subsidy (LIS) program, also called Extra Help. There were about 13.1 million Medicare beneficiaries enrolled in or deemed eligible for LIS (21 percent) in 2021.²⁴ LIS participants get federal support for Part D premiums, deductibles, and other cost-sharing, plus lower copayments for on-formulary prescription drugs (but must pay full cost for off-formulary drugs). Dual enrollees, those enrolled in MSPs (except the Qualified Disabled and Working Individual programs), and those receiving Supplemental Security Income are automatically eligible for LIS without applying. Those with incomes below 135 percent of the federal poverty level and resources below $17,220 ($34,360 for married couples) with burial expenses can apply in 2024.²⁵ LIS spending was $42 billion in 2022.²⁶

The Current Limitations of Means-Testing in Medicare

The rising portion of Americans over 65—a function of the increasing number of elderly Americans, longer life expectancy, and declining birth rates—will further strain the program’s finances. The mismatch between retired and working-age Americans is demonstrated by the decline in the number of workers per Medicare beneficiary—from about four from 1980 through 2008 to 2.9 in 2021 to an expected 2.5 in 2030.²⁷

The net surplus in lifetime benefits that Medicare enrollees receive (i.e., the value of benefits minus premium payments and payroll tax payments over the course of a lifetime) is large and expected to increase. A single worker with average earnings who turned 65 in 2020 is expected to receive $176,500 more in lifetime Medicare benefits than his total tax and premium payments. This is expected to rise to $248,500 for those who turn 65 in 2030. This net surplus persists for both “low earners” and “high earners” who do not pay higher premiums.²⁸ Although higher income individuals pay more in payroll taxes throughout their lifetimes and typically have lower health care expenses, their longer life expectancies allow them to receive benefits for a longer period.²⁹ Figure 3 shows that lifetime Medicare benefits (net of premiums) have significantly exceeded lifetime Medicare taxes for average two-earner households at every point in the program’s history and will continue to do so for the foreseeable future. This trend holds for other household types (single and married one-earner) across income levels.³⁰

At the same time, the share of Medicare beneficiaries subject to higher premiums due to their incomes—about 8 percent of those with Part B in 2024—is relatively small, and higher income beneficiaries still receive federal subsidies of 15 percent to 65 percent of expected Part B costs. By contrast, about 15 percent of Medicare beneficiaries were in MSPs in 2021, and 21 percent were in LIS in 2021.

Opportunities to Expand Means-Testing in Medicare

Increasing means-testing in Medicare can help to target resources toward those with the most need and put the program on a more sustainable trajectory by alleviating Medicare’s pressure on the federal budget.

The general approaches to increasing means-testing in Medicare may include adjusting (1) the income thresholds at which means-testing applies or the additional premium amount paid under current law, (2) the measures of wealth, including lifetime income, used to target means-testing policies, and (3) the mechanisms for means-testing.

The first approach would mean incrementally modifying the current means-testing structure for Part B and D premiums. When higher premiums were implemented in 2007, they applied to about 4 percent of Part B enrollees.³⁴ This rose to about 6 percent in 2015 and 8 percent in 2024.³² On several occasions, Congress has made changes to income thresholds and the percentage of Medicare costs they must cover. For example, in 2015 Congress reduced the premium subsidy for certain income brackets for 2018, and in 2018 it increased the number of brackets from five to six starting in 2019. Figure 4 shows these changes, with 2024 inflation-adjusted thresholds for comparison.³³

Expanding the share of beneficiaries owing higher premiums would reduce subsidies for a broader number of enrollees. For example, Congress could lower the income threshold above which reduced government subsidies would apply, which is currently $103,000 for individuals ($206,000 for married couples) under current law. Alternatively, increasing the premium percentage for enrollees in new or existing income tiers would increase premiums at higher tiers. For example, the top premium tier could have no subsidies (i.e., a premium percentage of 100 percent).

The second approach would introduce means-testing based on alternative measures of wealth besides annual income.³⁴ Means-testing based on annual income is likely not the best way to determine a household’s ability to pay additional premiums. Relying on income may also discourage responsible retirement savings practices, as seniors face an implicit marginal tax on these savings if more retirement income leads to higher premiums. Although some means-testing programs (such as MSPs) incorporate levels of resources, such wealth or asset tests are administratively more complex, may discourage retirement saving, and can be more easily gamed. For example, many individuals find ways to “hide” their wealth by purchasing exempt assets to become eligible for Medicaid long-term care coverage.³⁵ Another option would be to means-test based on lifetime earnings, which is how the federal government currently calculates Social Security retirement benefits. This would reduce the current disincentive on retirement savings and would be easier to administer since Social Security data could be used rather than more malleable measures of wealth.³⁶

The third approach would means-test outside of Part B or D premiums. For example, policymakers could apply Part A premiums on the basis of income or another measure. Means-testing of cost-sharing requirements such as deductibles and coinsurance could also be an option, although the impact of this policy on beneficiaries would vary based on health care utilization and may be more difficult to administer at the point of service.

Recent Proposals to Expand Medicare Means-Testing

There have been several proposals over the past decade to increase means-testing in Medicare. Although policy specifications and savings estimates vary, they all would have impacted income thresholds or premium percentages rather than using alternative measures of wealth or mechanisms besides premium amounts. Table 3 summarizes these proposals and their estimated budgetary impacts.

In its fiscal year (FY) 2013 and 2014 budgets, the Obama administration proposed increasing the share of expected Part B costs that higher income beneficiaries pay through their premiums—so that they would range from about 40 percent to 90 percent rather than 35 percent to 85 percent—and freezing income thresholds until those higher premium rates applied to 25 percent of beneficiaries.³⁷ The Congressional Budget Office (CBO) estimated that the FY2014 budget proposal would save $56 billion over 10 years.³⁸

Former Senator Alan Simpson and former White House official Erskine Bowles (the Simpson-Bowles Fiscal Commission) proposed lowering income thresholds so that approximately 15 percent of beneficiaries would pay higher premiums, freezing those thresholds through 2030, and increasing the percentage of Part B costs covered by premiums by 15 percent (similar to the 2013 budget). This was estimated to save about $66 billion.³⁹

In 2022, CBO estimated that freezing the income thresholds from 2024 to 2032 would increase the share of enrollees paying higher premiums to 10 percent in 2024 and 17 percent in 2032, saving $57 billion.⁴⁰

Think tanks of various ideological outlooks have also proposed increased means-testing. The Center for American Progress in 2012 proposed that the share of beneficiaries paying higher premiums remain at around 10 percent but that their premium payments increase by 15 percent, saving $25 billion.⁴¹ The Bipartisan Policy Center in 2013 recommended lowering income thresholds for three years so that 17 percent of beneficiaries would receive smaller premium subsidies, which was estimated to save $66 billion.⁴² More recently, the Heritage Foundation’s FY2023 budget blueprint recommended lowering the threshold for premium subsidy reductions to $55,000 for individuals ($110,000 for married couples) while eliminating the subsidy entirely for the wealthiest enrollees. Given that this change would roughly halve the income thresholds for paying higher premiums, it estimated significant savings of about $1.2 trillion.⁴³

Another approach to indirectly adjust income brackets is to incorporate a different measure of inflation. Specifically, using the chained consumer price index (CPI) would account for changes in spending patterns and other statistical biases that exist in traditional CPI and would thereby slow the increase in income thresholds.⁴⁴ If incomes rise faster than inflation, then more households would fall into higher income thresholds over time. The Obama administration proposed this approach in its FY2014 budget, when President Biden’s current chief of staff was the White House budget director.⁴⁵ CBO estimated at the time that this proposal would save $7.8 billion in Medicare. In 2022, it projected that making this change across programs would save $41.3 billion in federal health spending.⁴⁶

Unlike welfare programs, Medicare eligibility is not tied to wealth or income, but there are numerous policies that adjust enrollee contributions based on these factors. However, means-testing policies in Medicare are currently limited as only about 8 percent of Part B and Part D enrollees face higher premiums for their coverage. Improving Medicare’s finances should be a priority for policymakers. Medicare beneficiaries tend to consume much more in benefits than they pay into the program, and the ratio of current enrollees to taxpayers is increasing. As such, Congress should reduce government subsidies for wealthier Medicare beneficiaries. There are numerous potential bipartisan approaches to updating the current means-testing structure. Policymakers could also account for alternative measures of income, or apply different mechanisms for means-testing. Such policies will be a critical tool for stabilizing America’s largest health care program while ensuring that benefits are available and affordable for seniors.