Revised July 22, 2004


by Edwin Park and Robert Greenstein

PDF of analysis

View Related Reports

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.


In a major health policy address on July 12, Senate Majority Leader Bill Frist promoted a proposal to create new, tax-advantaged “Lifetime Health IRAs.”  Also known as Retirement Medical Benefit Accounts (RMBAs), these accounts are intended to encourage financial security for retirees by helping them pay for long-term care costs and other health care expenses.[1]  This proposal was originally designed by Fidelity Investments and first surfaced during negotiations in the fall of 2003 over the Medicare prescription drug legislation.

The RMBA proposal would alter existing tax-favored retirement savings accounts, such as traditional Individual Retirement Accounts (IRAs) and 401(k) plans, by allowing individuals to establish special RMBA “subaccounts” within IRAs and 401(k)s.  Individuals who contributed to IRAs or 401(k)s could place a designated percentage or dollar amount of their deductible contributions in these RMBA subaccounts and receive an extremely lucrative tax break for those funds.[2]

Under current law, contributions to traditional IRAs and 401(k)s are tax deductible and earnings on these accounts accrue tax-free.  When account-holders withdraw funds from IRAs and 401(k)s after they retire, the withdrawals are taxed as ordinary income.  By contrast, under the RMBA proposal, all funds withdrawn after retirement from the RMBA subaccounts would be tax free as long as they are used to pay for out-of-pocket health and long-term care costs.

The RMBA proposal would extend to retirement accounts the type of lucrative tax benefits provided by the new Health Savings Accounts (HSAs) included in the recently enacted Medicare drug legislation.  Under HSAs, individuals enrolled in high-deductible health insurance plans may make both tax-deductible contributions into and tax-free withdrawals from HSAs if the withdrawals are used to pay for out-of-pocket medical costs.  HSAs breached what had been a longstanding bright-yellow line in the U.S. tax code — that savings accounts may not feature both tax-deductible contributions and tax-free withdrawals. [3]  The RMBAs would extend this new principle, under which accounts may offer both tax-deductible contributions and tax-free withdrawals, to IRAs and 401(k)s as well, without any HSA-like requirement that participants be enrolled in a high-deductible health insurance policy.[4]  Any individual eligible to contribute to a traditional IRA or 401(k) plan would be able to establish a RMBA subaccount.

The RMBA subaccounts would provide their greatest tax benefits to high-income households.  Since high-income households are in the highest tax brackets, they would secure the largest tax deductions for contributions to these subaccounts and also garner the largest tax benefits from being able to withdraw funds from the accounts on a tax-free basis.  Moreover, they also are the households that can afford to make the largest deposits into retirement accounts.

Indeed, data on participation in, and contributions to, IRA and 401(k) accounts show that the majority of low- or moderate-income households have no such accounts.  By contrast, most high-income households do have such accounts, and those at high income levels make the largest average contributions to them.  As a result, the RMBA proposal would primarily function as an attractive and highly lucrative tax shelter for affluent retirees, even though these are the people who least need new government subsidies to help cover their costs in old age.

Proposal Would Have Large, Damaging Long-Term Fiscal Impacts

Allowing a portion of the funds in retirement accounts to be withdrawn tax free after reaching age 65 would make swelling federal budget deficits markedly worse.  There are very substantial amounts of assets in traditional IRA and 401(k) accounts.  In present value terms, $3.8 trillion in federal revenues are expected to be collected on withdrawals from these accounts between now and 2040.[5]

These anticipated revenues are reflected in the long-term budget baseline.  Even with these revenues, the long-term fiscal picture is bleak; budget deficits are expected to rise eventually to levels dangerous to the economy.  Stern warnings about the fiscal dangers that lie ahead have been voiced recently by the International Monetary Fund, the Comptroller General of the United States (the head of the General Accounting Office), the investment house Goldman Sachs, and such luminaries as former Treasury Secretary Robert Rubin, former Senator Warren Rudman, and former Congressional Budget Office director Robert Reischauer.

For example, former Treasury Secretary Robert Rubin, in collaboration with Brookings economist Peter Orszag and Wall Street economist Allan Sinai, wrote earlier this year that “the scale of the nation’s projected budgetary imbalances is now so large that the role of severe adverse consequences must be taken very seriously...”[6]  Similarly, the New York Times reported that an IMF report issued in early January “sounded a loud alarm about the shaky fiscal foundation of the United States, questioning the wisdom of the Bush Administration’s tax cuts and warning that large budget deficits pose ‘significant risks’ not just for the United States but for the rest of the world.”[7]  In strong language usually reserved for developing countries struggling with international debt obligations, the IMF report disapprovingly noted that the “United States is on course to increase its next external liabilities to around 40 percent of GDP within the next few years — an unprecedented level of external debt for a large industrial country.”[8] 

If the nation begins to allow some of the funds deposited in retirement accounts into which contributions have been made on a tax-deductible basis to be withdrawn tax free, an already grim long-term fiscal outlook will become considerably worseIn a recent scholarly assessment of the nation’s long-term fiscal problems, Alan Auerbach of the University of California at Berkeley, one of the nation’s leading public finance experts, and Brookings economists William Gale and Peter Orszag warn that “…proposals to reduce the taxation of withdrawals from retirement accounts could significantly and adversely affect an already bleak fiscal outlook.”[9]

In the fall of 2003, the Joint Committee on Taxation reportedly estimated that the RMBA proposal then being discussed in the closed-door deliberations on the Medicare drug bill would cost $150 billion over the first ten years.  Moreover, the costs in the first ten years constitute only a fraction of the ultimate costs of the proposal.  To a substantial extent, the upfront tax deductions that would be provided for contributions to RMBAs would represent deductions for contributions that would be made to IRAs and 401(k)s anyway.  As a result, revenue losses in the initial years would be limited.  But over the long term, as Americans retired and withdrew on a tax-free basis some hundreds of billions or even trillions of dollars of what otherwise would have been taxable income, the revenue losses would burgeon.  Policymakers who assume that the cost of RMBAs over the first five or ten years will accurately reflect the costs in later years will be sadly mistaken.  (See box on page 4.)

The RMBA proposal also would impose a substantial revenue drain on state budgets.  State income tax codes generally conform to the definition of taxable income in the federal income tax code.  Many states would consequently experience large revenue losses if the RMBA proposal is enacted.  State finances also could be adversely affected in a second way — through increased borrowing costs.  The new tax-free subaccounts would make traditional IRAs and

401(k) plans more lucrative, which could force state and local governments to pay higher interest rates to attract sufficient investment in the tax-exempt bonds they offer.

Why Five- or Ten-Year Cost Estimates Will Understate Long-Term Costs

The “official” cost of the RMBA proposal over the next five or ten years is likely to understate dramatically the proposal’s long-term cost. Since the proposal does not change existing income limits or annual contribution limits for IRAs and 401(k) plans, it would not be likely to result in large increases in the amounts currently being contributed to such plans, except possibly among those who already are 65 or over (see discussion below). Some increase in contributions might ultimately occur because of the added tax advantages, but it generally takes a number of years before understanding of such tax changes becomes widely known and affects taxpayer behavior. (Affluent taxpayers with financial advisers would tend to learn of the tax implications more quickly, but they generally already are at the maximum contribution limits for tax-favored retirement accounts.)

Because the amounts being contributed to retirement accounts on a tax-deductible basis would not be expected to rise substantially in the initial years except among the elderly (as described below), and because most funds deposited into RMBA subaccounts would not be withdrawn in the first five or ten years that the accounts were in effect (since most contributors would not yet be retired), the proposal necessarily carries a price tag in the 10-year budget window that is substantially lower than the projected cost in subsequent decades. In the future, when individuals began withdrawing substantial sums from these accounts after turning 65, hundreds of billions of dollars would be withdrawn on a tax-free basis. Revenue losses would swell sharply.

States would be absorbing these large revenue losses at the same time that the cost of maintaining state government would be rising substantially.  The retirement of the baby-boom will increase the fiscal burdens that states face in financing long-term care, as well as the share of prescription drug costs they continue to bear under the new Medicare law for low-income Medicare beneficiaries who also qualify for Medicaid.  States can ill afford the level of state revenue losses that enactment of RMBAs would engender.

Proposal May Pose Danger to Medicare

Over time, the RMBA proposal also could result in adverse effects on non-affluent Medicare beneficiaries.  The large long-term revenue losses that the proposal would engender would intensify budgetary pressures to cut Medicare and other programs down the road.  In addition, the existence of RMBAs could facilitate the emergence of proposals to increase Medicare premiums, deductibles, and other cost-sharing charges quite substantially over time. 

The RMBA proposal is being marketed partly as a way to help Medicare beneficiaries pay for Medicare premiums, deductibles, and cost-sharing (including cost-sharing associated with the new drug benefit), as well as other benefits such as long-term care that Medicare does not cover.  In the future, when the long-term effects of the Bush tax cuts in shrinking the nation’s revenue base collide with the mounting costs for retirement and health care programs for the elderly as the baby boom generation retires, the existence of RMBAs could be used to advance controversial proposals to reduce Medicare services markedly or increase substantially the charges that Medicare beneficiaries must pay.  The argument would be that Medicare beneficiaries could absorb increased charges and decreased coverages in Medicare because they could draw funds tax free from RMBAs to help defray such costs.

As explained below, however, RMBAs are likely to be of little help to most low- and moderate-income beneficiaries; they would primarily be a windfall for the more well-off.  A combination of RMBAs and increases in Medicare beneficiary charges would likely have a net positive effect on the pocketbooks of high-income Medicare beneficiaries, but a decidedly negative effect on senior citizens of modest means.


Proposed Accounts Would Expand Tax Shelter Benefits of Greatest Value to Higher-Income Households

RMBAs almost certainly would be used more as an expanded tax shelter for higher-income individuals than as a mechanism that enables households of more moderate means to accumulate greater savings to defray medical expenses when they retire.  There are several reasons why this is the case.

First, RMBAs would have limited effect in helping most low- and moderate-income individuals accumulate additional savings to pay for out-of-pocket medical and long-term care costs during retirement.  As with any tax deduction, the tax-cut value of a tax-deductible contribution to a RMBA would rise with an individual’s tax bracket.  Nearly 75 percent of households are either in the 10 percent or 15 percent tax brackets or do not earn enough to incur federal income tax liability.  When they file their tax returns, these households would receive at most a 10-cent or 15-cent tax subsidy for each dollar they contributed to a RMBA; low-income workers who have no income tax liability would derive no benefit at all.  By contrast, taxpayers in the top income tax bracket would receive a 35-cent subsidy for each dollar the taxpayer deposited into a RMBA.

In addition, low- and moderate-income households are unlikely to have significant income or resources available to make sizeable deposits into IRAs and 401(k) plans generally and thus are unlikely to build up significant savings in a RMBA.  IRS data show that in 1995 (the latest year for which these data are available), only seven percent of eligible taxpayers made any contribution to traditional IRAs.[10]  Similarly, a recent Congressional Budget Office analysis finds that in 1997, only six percent of workers with incomes under $20,000 made any contribution to a 401(k), and only 27 percent of those in the $20,000-$40,000 range did.  Even in the $40,000-$80,000 family income range, only a minority made 401(k) contributions.  Moreover, those lower- and moderate- income households who do have IRA or 401(k) assets tend to have quite small accounts.  By contrast, participation is high at high-income levels, and these individuals tend to have very large retirement accounts.[11]  (There are no income limits on who can make tax-deductible contributions to 401(k) plans.) 

In short, the people who are most likely both to make contributions to retirement accounts and to make the largest contributions — and hence to be the principal beneficiaries of RMBAs — are people at high income levels. 

Furthermore, RMBAs would represent a substantial new tax break for high-income individuals.  The tax breaks associated with RMBA subaccounts would be much greater than the tax breaks provided for IRAs and 401(k) plans generally, since amounts withdrawn after retirement from traditional IRAs and 401(k)s are counted as taxable income.[12] Creating a retirement-related feature of the tax code under which deposits made during working years are deductible, amounts earned on the accounts (which could be invested in the stock market) are tax free, and withdrawals upon retirement also are tax free is unprecedented.

RMBAs thus would be very attractive to higher-income individuals, and these substantial tax advantages would be expected to lead to heavy use of RMBAs by such individuals, increasing the cost of RMBAs to the Treasury.  To be sure, the requirement that funds in RMBA accounts be used solely to pay for out-of-pocket medical and long-term care expenses is a restriction that IRA and 401(k) accounts otherwise do not have.  But that should not pose much of an obstacle to widespread exploitation of RMBAs as a lucrative tax-shelter opportunity by higher-income households.  Nearly all individuals can expect to incur substantial out-of-pocket health care costs after they reach age 65; the incidence of illness and disability rises with age.  Moreover, Medicare requires beneficiaries to pay substantial cost-sharing and does not cover services such as long-term care (or under the new Medicare legislation, prescription drug costs inside the so-called “doughnut hole”).  A Commonwealth Fund study found the average Medicare beneficiary incurred about $1,500 in Medicare cost-sharing charges in 2002 and bore about $3,800 in overall out-of-pocket medical costs.[13] 

Retirees already can — and many do — pay for Medicare deductibles and coinsurance and other out-of-pocket health care costs such as long-term care expenses with funds withdrawn from an IRA or 401(k).  As explained earlier, funds withdrawn from these accounts are treated like other income for income-tax purposes.  With the availability of RMBAs, higher-income retirees with substantial RMBA balances would be able to pay for out-of-pocket health and long-term care costs with IRA and 401(k) funds that have been placed in RMBAs and never taxed.

The tax breaks that RMBAs would provide for both deposits and withdrawals, coupled with the expectation of inevitable out-of-pocket health care costs during old age, would likely induce heavy use of RMBAs by affluent individuals, even though such individuals are the most likely already able to afford such costs. 

RMBAs Not Related to Extending Coverage to the Uninsured

The RMBA proposal would not extend health insurance coverage to the uninsured; it is not intended to do so. The proposal is focused on retirement savings to pay out-of-pocket health and long-term care costs for individuals aged 65, nearly all of whom are insured through Medicare. The proposal should have no effect on the number of non-elderly Americans who are uninsured.


Effects on Current Medicare Beneficiaries

One argument that proponents of RMBAs may make is that RMBAs would help current Medicare beneficiaries meet out-of-pocket health and long-term care costs they face.  This argument is weak; RMBAs would be unlikely to provide much help to current low- and moderate-income Medicare beneficiaries in paying out-of-pocket costs.  If current beneficiaries are allowed to make both tax-deductible deposits into their subaccounts and tax-free withdrawals from them, the primary result would be a new tax-avoidance mechanism for affluent retirees.

But a Medicare beneficiary in the 35-percent tax bracket could use RMBAs to receive a 35-cent subsidy for each dollar that he or she deposited in a RMBA and then quickly withdrew.  Considering that higher-income beneficiaries are the people most capable of paying out-of-pocket health care costs out of existing income and resources, the use of RMBAs in this manner would offer an unwarranted tax windfall to those who least need it.

Some Medicare Beneficiaries Could Be Harmed

RMBAs also could make some low- and moderate-income retirees who receive health care coverage through their former employers worse off, because it would likely encourage some employers to scale back or drop retiree coverage.  Employer-provided coverage for retirees already is eroding; among younger retirees aged 65-69, the proportion of Medicare beneficiaries with retiree health coverage through a former employer declined from 46 percent in 1996 to 39 percent in 2000.[14]  (In addition, a new survey by the Kaiser Family Foundation and Hewitt Associates finds that more than 70 percent of large private firms that offer retiree coverage increased premiums for such coverage last year, and more than half of those firms increased beneficiary deductibles or co-payments.[15])

The Congressional Budget Office has estimated that the new Medicare prescription drug benefit will exacerbate this trend by inducing more employers to drop retiree drug coverage.  CBO estimates that 2.7 million retirees will lose employer coverage for drugs because of the new law.[16]  The availability of RMBAs could encourage additional employers to drop or further scale back retiree health coverage on the grounds that retirees could use funds in RMBAs to offset health care costs after retirement.

Finally, the availability of RMBAs may be used to justify opposition to future efforts to make the Medicare prescription drug benefit more adequate and affordable, as well as to promote proposals in future years to increase premiums, deductibles, and co-payment charges levied on Medicare beneficiaries or to scale back the health care services that Medicare covers.  Arguments may be made that there is little need to improve the drug benefit — and that it would not cause serious problems to raise various Medicare fees and charges significantly — because beneficiaries could turn to RMBAs to help offset Medicare cost-sharing charges and fill holes in Medicare drug coverage.  As explained earlier, however, it is unlikely that many current or future low- and moderate-income Medicare beneficiaries would get much help from RMBAs.  Those who would secure the greatest tax advantages and build the largest RMBA accounts would be higher-income individuals who already can afford to pay most of their out-of-pocket health care costs.



Due to the lucrative tax break that RMBAs would create, the cost of the RMBA proposal would escalate substantially over time, worsening the already-dire long-term fiscal outlook.  The proposal also would set a dangerous precedent by allowing substantial amounts deposited on a tax-deductible basis in retirement accounts such as IRAs and 401(k) plans to be withdrawn tax free in retirement for designated purposes.

Because of the adverse long-term fiscal effects of RMBAs, the RMBA proposal would further constrain the resources available to finance Medicare and other basic programs in the future.  In addition, RMBAs could be used, as budget problems mounted, as a justification for cutting Medicare fairly heavily through the imposition of substantially higher premiums, deductibles, and/or cost-sharing charges or the elimination of coverage for some significant health care services. 

Over the long term, the likely result of RMBAs would be further deterioration of an already grim budget picture, further enrichment of the most affluent members of society, and more intensive pressure to scale back the quality or affordability of the recently enacted Medicare drug benefit, Medicare coverage overall, or both.

The Center on Budget and Policy Priorities is grateful to the
Nathan Cummings Foundation for its support of this report.


End Notes:

[1] “Frist Addresses National Press Club,” July 12, 2004, available at:


[2] Certain details of the RMBA proposal have not been made available, including the percentage or dollar limit that would govern how much of annual IRA or 401(k) contributions could be placed in the RMBA subaccounts.

[3] For an analysis of Health Savings Accounts, see Robert Greenstein and Edwin Park, "Health Savings Accounts in Final Medicare Conference Agreement Pose Threats Both to Long-Term Fiscal Policy and to the Employer-Based Health Insurance System," Center on Budget and Policy Priorities, revised December 1, 2003.

[4] Withdrawals made from the subaccounts for out-of-pocket medical costs before age 65 and withdrawals for non-medical purposes would apparently be subject to income tax and a financial penalty.

[5] Alan J. Auerbach, William G. Gale, and Peter R. Orszag, “Reassessing the Fiscal Gap:  The Role of Tax-Deferred Savings,” Tax Notes, July 28, 2003.  This estimate is for the period between 2004 and 2040.

[6] Robert Rubin, Peter Orszag and Allen Sinai, "Sustained Budget Deficits: Longer-Run U.S. Economic Performance and the Risk of Financial and Fiscal Disarray," AEA-NAEFA Paper, January 4, 2004.

[7] Elizabeth Becker and Edmund Andrews, “I.M.F. Warns that U.S. Debt is Threatening Global Stability,” The New York Times, January 8, 2004.

[8] International Monetary Fund, "U.S. Fiscal Policies and Priorities for Long-Run Sustainability," IMF Occasional Paper 227, January 7, 2004.

[9] Alan J. Auerbach, William G. Gale, and Peter R. Orszag, “Reassessing the Fiscal Gap:  The Role of Tax-Deferred Savings,” Tax Notes, July 28, 2003.

[10] Robert Carroll, “IRAs and the Tax Reform Act of 1997,” Office of Tax Analysis, U.S. Department of Treasury, January 2000.

[11] Congressional Budget Office, "Utilization of Tax Incentives for Retirement Savings," August 2003.

[12] Withdrawals from Roth IRAs are not taxed after retirement, but individuals cannot claim an up-front deduction for contributions they make to Roth IRAs.

[13] Stephanie Maxwell, Matthew Storeygard and Marilyn Moon, “Modernizing Medicare Cost-Sharing: Policy Options and Impacts on Beneficiary and Program Expenditures,” The Commonwealth Fund, November 2002.

[14] Bruce Stuart et al., "Employer-Sponsored Health Insurance and Prescription Drug Coverage for New Retirees: Dramatic Declines in Five Years," Health Affairs (Web Exclusive), July 23, 2003.

[15] Kaiser Family Foundation and Hewitt Associates, "Retiree Health Benefits Now and in the Future: Findings from the Kaiser/Hewitt 2003 Survey on Retiree Health Benefits," January 2004.

[16] Congressional Budget Office, "Letter to the Honorable Don Nickles Providing Additional Information about CBO's Cost Estimate for the Conference Agreement on H.R. 1," November 20, 2003.