"Economic Perspective" column reprinted with permission.
Copyright 1997 TAX ANALYSTS
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
Part One: From General Theory to Practical Design
As part of its budget package, the Senate attempted to bring Medicare more into balance for the long term. One of the specific methods chosen was to try to charge high-income taxpayers for some of their Medicare benefits. While a strong economic and political case can be made for steps in this direction, little attention was paid in advance to exactly how such a proposal could be implemented. Such issues of practical design, in the end, can make or break any proposal, and this one is no exception. Assessing a fee on the basis of annual income carries along the complications that would be associated with adding a new income tax onto the one that already exists.
Of all the long-run budgetary problems confronting the nation, none is larger in magnitude than Medicare. Medicare faces the same demographic pressures affecting social security. Rapid increases in the number of beneficiaries begin around 2010, when the first of the baby boomers become eligible for Medicare. Even without this aging of the population, recent projections show costs per household rising at an alarming rate. A couple retiring in 1980 was scheduled to receive about $87,000 in lifetime Medicare benefits, one retiring in 1995 about $186,000, and one retiring in 2030 about $423,000 (in 1993 dollars, after adjusting for inflation). Even these projections assume arbitrarily that the annual cost increases within the system somehow stabilize relative to the size of the economy within another 15 years or so.
Much of the cost increase is due to a health insurance design that knows no realistic bounds on what it will cover. Within certain limits on fees for particular services, doctors and patients still essentially bargain between themselves over how much care will be provided and, therefore, how much taxpayers will need to pay for those services. This type of design -- not that much different from the design of most private health insurance until recently -- encourages high-priced professional help and cost-increasing technological development. For the most part, the budget bill does not address those features of Medicare that drive up costs.
Partly because these rapid cost increases keep shifting burdens forward to tomorrow's taxpayer, no cohort of retirees to date has come anywhere close to paying for his or her Medicare benefits. Under current law, even the $423,000 in lifetime benefits received by a couple retiring in 2030 is matched by lifetime contributions of only between $131,000 and $276,000 for low- and middle-income taxpayers depending on income level and whether or not there are two workers in the family. Even many high-income, one-earner couples do quite well -- receiving more benefits than they paid in taxes -- for decades to come. For high-income taxpayers who are single or two-earner couples, however, the situation turns around sooner and Medicare taxes eventually begin to exceed the benefits that will be received. Compared to previous generations, what makes their taxes exceed benefits are (1) the higher hospital insurance tax rate that these individuals pay over most of their lives and, (2) in some cases, the higher tax base they face since there is no longer a cap on the amount of earnings subject to this tax.
We shouldn't be surprised that high-income taxpayers would eventually have to pay for more than they receive in a system that by its very nature is meant to be redistributive. The questions we should be asking are exactly how this transition should take place, and how should beneficiaries be made more aware of the cost of Medicare relative to the fees they are paying.
These issues are confused by the extent to which programs for the elderly rely on general revenues -- that is, income and other taxes -- rather than on social security taxes alone. Even some social security cash benefits are paid for out of general revenues -- for instance, income taxes paid on social security benefits are transferred back to social security. Assistance to the low-income elderly flows not only through social security but also through Supplemental Security Income, which is not supported by social security taxes. Medicaid payments for long-term nursing home care are paid for out of general revenues. Finally, Medicare, Part B or Supplementary Medical Insurance (SMI) is paid for mainly out of general revenues. Only about one-quarter of SMI is paid for out of a premium collected from Medicare beneficiaries.
Reformers wanted the elderly to recognize how much is being transferred to them in the way of health benefits. With Medicare far out of balance, any incremental reform moves the nation one step closer to attaining a viable system. Delay only keeps the budgetary straitjacket tied tighter than necessary and potentially lowers saving for the long run. Why not, therefore, at least start with SMI? Unlike hospital insurance, SMI is supported out of general revenues rather than the social security tax, so there is less pretense that previous social security taxes paid for current benefits. At one time, moreover, beneficiaries were expected to pay for one-half of the benefits from this selected part of the multi-tiered set of programs for old age. As the system became more generous (and more expensive), other taxpayers were required to pay a larger share of the total bill.
Therefore, the argument proceeded, if somebody has to give up something to get this system more into balance, why not start with the high-income elderly? After all, this means that few elderly would be affected at first, which makes the system progressive and the political cost small. In addition, the current high-income elderly are individuals who are receiving substantial net Medicare subsidies even after taking into account any social security tax they might have paid for Medicare, Part A, or hospital insurance. When looking only at winners and losers or who pays and who gains, this looks likes as viable a political compromise as any, doesn't it?
Not really. The logic is all right up to a point. What it ignores, however, is that the phase-in of a charge on the basis of income is nothing more or less than a new income tax system that must be administered. Designing tax systems is not easy. They require some thought. There are principles involved, such as the equal taxation of equals, and, as the calculations above demonstrate, what may be equitable for one generation may be inequitable for the next.
Most important, the whole thing has to be easily administered. That is, taxpayers must know how to pay the tax, there must be withholding systems in place to collect the money over time, and the tax authorities must be able to enforce the law. Information systems must be developed. Reports and other administrative data must pass back and forth between the collecting agency (IRS or Social Security Administration), the taxpayer, and Medicare.
These last issues typically are not discussed until too late. If Congress wants to charge more for Medicare, then it has to devote some more attention to the pluses and minuses of various alternatives, especially from an administrative standpoint.
Part Two: Some Alternatives
Medicare is becoming an extraordinarily expensive part of the nation's social welfare system. Its growth rate is so high that one can argue that it is pushing aside most other activities of the federal government. As one method of trying to control costs, the Senate recently tried to assess a higher fee on the high-income elderly. A modest additional fee for these individuals would be phased in as income increased above certain levels. Such a fee not only would reduce the net impact of the system on the deficit, but would reduce the net transfers being made by workers to support Medicare, and make more apparent to some of the elderly the societal cost of the benefits they are receiving.
Like so many proposals, however, this one got tied up in technical problems of implementation, which led to its rejection by the conference committee as much as did the rising political opposition. The method contained in the original Senate bill threatened to add significant additional bureaucracy and taxpayer complexity for only a tiny reduction in the deficit -- and could actually add to the gross cost of Medicare over the long run. There are different and better ways, however, to achieve roughly the same end without all the additional difficulties.
Here we will compare three different ways to reduce the net cost to taxpayers of providing Medicare to elderly individuals with above-average means. Our focus is on that part of Medicare (Part B) primarily paid for out of general revenues, and for which there is already an administrative apparatus that collects a modest fee that typically covers about 25 percent of costs. First, one can simply try, like the original Senate bill, to assess an additional fee on the basis of annual income. As one simplification along these lines, an additional income-based fee might be coordinated with the income tax on social security benefits already in current law. Second, one can attempt to assess a fee on the basis of lifetime earnings, as measured by social security, rather than annual income, as reported to the IRS. In this case the additional fee would simply be collected along with the fee already taken out of social security checks. Finally, one can reduce the cost of Medicare, while providing additional cash benefits to lower-income individuals so that on net they suffer no loss. Only the last two methods can be performed without significant administrative cost, and only the last really gets at most of the economic impact of Medicare growth on the economy.
Method 1: A Fee Based on Annual Income
Advocates of means-testing and income-based fees focus their attention on getting fees more adequately to reflect costs. Seldom, however, do they give full attention to the implications of phasing in charges (or phasing out benefits) on the basis of income. From an administrative standpoint, the proposed phase-in of a higher supplementary fee on Medicare, Part B operates like an additional income tax on benefits. The income base has to be defined. Withholding has to take place. Enforcement efforts are required to collect fees that haven't been paid. The relationship among individuals in a household must be taken into account, as marriages, divorces, and deaths affect the composition of the household, the average fee to be charged, and the months of eligibility for Medicare. For technical reasons I will not describe here, almost any income-conditioned fee will also create marriage penalties, and two individuals will sometimes pay a lower fee if they stay unmarried or divorce. For households who are no longer working, it is quite easy to make large adjustments in the measure of the tax base. For example, assets can be transferred to children; investment can be shifted over to stocks with low yields but high amounts of accrued capital gains that do not fall into the tax base; withdrawals from pension accounts can be slowed down or speeded up. And so on. Does this sound familiar? It should. It reflects all of the problems of an income tax.
To try to simplify the matter somewhat, policymakers typically try to make the tax base for this new system match closely the base for the income tax, but this seldom leaves them completely satisfied. For instance, a millionaire collecting tax-exempt bond interest would not pay the fee charged other higher-income individuals if the normal income tax base is adopted. Adjustments are then sometimes made to create yet another tax base.
An income-based Medicare fee would be only one of many phase-in or phase-out procedures that operate like a separate little tax system. Taxpayers already may face such additional calculations for phase out of itemized deductions, or phase out of personal exemptions or phase in of income taxation of up to 50 percent of social security benefits, or the additional phase in of income taxation of up to 85 percent of social security benefits (at another income level), or a whole slew of other income-based patches, sewn on one at a time.
In the case of Medicare fees, some members of Congress have indicated that they do not want the fee to show up as just another schedule to be filed with income tax returns. Then it will look like just another tax, no matter what it is called. If they try to have social security or the Health Care Financing Administration handle the new fee, however, they then require all sorts of reports on income going back and forth between the IRS and the administering agency. Only by April 15, 1999, however, does even the IRS know the income of most, but not all, individuals in 1998, but it has little idea of their income or family composition for 1999 during 1999. Fees assessed incorrectly for Medicare in 1998 would have to be adjusted for rebates or further assessments against taxpayers, who would have to file papers with whatever office was responsible.
To minimize the number of taxpayers affected, those with income less than some amount might be totally exempted. A related, partial simplification from the taxpayer's standpoint would be to have the phase in somehow fit closely with the income taxation of social security benefits, so that all these calculations could be fitted onto one form. The form is already very confusing and unpopular with those affected, but at least some coordination could reduce administrative hassle of yet another form based on yet another income level and income tax base. However, taxpayers are unlikely to perceive any additional patch as simply a fee for their Medicare benefits.
Method 2: A Fee Based on Lifetime Earnings
An alternative way to apply the new fee in a progressive manner would be to base it not on annual income, but on the lifetime earnings of individuals, as reported to the Social Security Administration. Most of these lifetime earnings are currently applied to the calculation of something called average indexed monthly earnings, from which social security cash benefits are determined. The social security formula itself is progressive, and grants a higher rate of return to those with lower lifetime earnings.
There are three advantages to this approach. First, the Medicare beneficiary would easily perceive the charge as an additional fee for Medicare, not as something else. Second, the additional charge would be easily administered. Those with high lifetime earnings and higher social security checks simply would find that a higher fee was assessed against them for Medicare. They already pay the rest of their Medicare fee out of their monthly social security check, and the amount withheld is exact. No additional forms need to be filed, and no reconciliation is required at the end of a year. Next year, instead of all beneficiaries bearing an identical additional fee of say, $3, only those with lower lifetime earnings would see a $3 increase, but those with higher incomes would see a larger increase, say, up to $20.
A third advantage is that past lifetime earnings are not as easily manipulated as is annual income for those who are already retired. Like annual income, however, no one measure is perfect, and some adjustments would need to be made -- as they are in the case of social security -- for those who are measured as having low lifetime earnings because they worked in federal or other employment that was exempted from social security tax. While these adjustments would not be perfect, on average I would guess that they still would be more equitable than annual income as a base for assessing the fee.
Method 3: Cut Medicare Costs and Raise Cash Benefits for Those With Modest Means
While method 2 eliminates the extraordinary complexity inherent in method 1, it still does little to solve the problem of higher and higher costs associated with Medicare. Assessing a small additional fee on the higher-income elderly really does nothing to lower the growth rate. It simply is an additional method of financing that growth. Call it a fee or a tax or what you will, it does not lower Medicare costs. It only provides an offsetting revenue source. Now in budget accounting backdoor fees and taxes are treated as negative expenditures -- which makes them more popular with lawmakers. In truth, however, they are the economic equivalent of fees that got added to revenues rather than subtracted from expenditures.
Put another way, Medicare still displaces private and other public activity, and its growth would not be inhibited by a higher fee. Indeed, the additional revenues might make it even easier to maintain a higher-cost system, since now an additional portion of those costs would be covered.
Lawmakers need to realize that cutting back on the value of Medicare by $100 on average is equivalent to a fee increase of $100, at least as far as the deficit is concerned. Only the former approach leaves a smaller Medicare system in place. A slightly larger cutback in Medicare payments for health services could also be accompanied by an increase in the cash benefits paid to individuals of less modest means. One could base this additional cash benefit on lifetime earnings, as discussed above, or one could simply adopt a much higher minimum benefit in social security or Supplemental Security Income for the poor. Many low- and moderate-income individuals would probably be happy with this trade-off as well, especially if they value additional cash more than a higher-cost Medicare system.
Methods 2 and 3 accomplish most of what lawmakers seek in reducing Medicare subsidies to individuals with above-average means. However, these alternative methods do so in a way that recognizes up front that a new system needs to be easily administered and understood.