It is understandable that many Americans are coming to the conclusion that the problems of the federal government’s deficit have at last been laid to rest. After all, the President’s budget package enacted last fall made a much publicized “down payment” on eliminating the deficit; and Clinton’s insistence in his fiscal year 1995 budget proposal on maintaining the cuts he earlier promised suggests a clear departure from the complacency of the last twelve years. The near-term projections are encouraging: it is possible that for the first time in half a century the annual federal deficit may shrink (in dollars) for four consecutive years. With the economy improving, and with health-care reform now attracting far more attention than fiscal policy, we are losing our sense of urgency about the need to stop the ongoing growth in our national debt.1

That is a serious mistake. In fact, the debt is a far more intractable problem than most politicians and policy professionals are willing to admit—at least in public. True, last year’s budget deal was designed to save $500 billion over the next five years (with the “savings” calculated as reductions from a sharply rising projection of deficits), and it is projected to reduce the annual deficit from 4.0 percent of the Gross Domestic Product (GDP) in 1993 to 2.2 percent in 1998. Seen from a different angle, however, the prospect for the deficit during the 1990s doesn’t look so bright. Even with the recent improvement, the money borrowed by the US Treasury to finance our deficits will still siphon off nearly half of all net savings made by US businesses and households—a far greater share of savings than deficits typically consume in any other industrial nation. Thanks to Clinton, America has been spared an explosion in the publicly held national debt in the near term—but the debt is still due to rise in dollars more over the next ten years than over the previous ten.

What is more remarkable is how little Clinton’s first budget has altered the long-term outlook. After 1998, the year the deficit is expected to sink to its low point, its share of GDP is expected to grow again each year. Deficits will rise to 3.3 percent of GDP by 2004, to 5.0 percent by 2010, and to 10.0 percent by 2020.2 As the economist Benjamin Friedman of Harvard recently wrote in these pages, no nation can expect to prosper if its public debts continuously increase faster than its income. Yet on our current budget trajectory, the national debt will keep climbing as a share of the economy into the next century: from a postwar low of 25 percent of GDP in 1974 to 55 percent in 2004, to 67 percent in 2010, and to 112 percent in 2020. The last figure approximately equals the previous record reached in 1946, after a ten-year depression and a five-year world war.3

Some may ask how anyone can make such long-term projections for the federal deficit. The answer is simple: the part of the budget that is called “discretionary” (32 percent of the total in 1993), but that pays for such services as national defense and public health, is already scheduled for deep cuts, and it is extremely doubtful that it can be reduced much further. Interest on the national debt (14 percent of the budget in 1993) must be paid to avoid a devastating financial panic. The rest of the budget—over half of all federal spending—consists of “entitlements,” the word for all federal benefits that are not controlled by Congress’s yearly appropriations. These include Social Security, Medicare, Medicaid, Federal pensions, food stamps, and AFDC. Since entitlement payments are made automatically according to a preset schedule, their costs can be projected by taking account of the economic, demographic, and social trends that seem most likely during the next several decades.4

As it turns out, nearly all of those trends are projected to raise the cost of our current entitlements system—from 10.3 percent of GDP as recently as 1990, to 17.5 percent in 2020, and to 21.3 percent in 2040. The cost of Social Security and Medicare (both hospital and non-hospital care) alone is projected to reach an unthinkable 38 to 53 percent of the average worker’s taxable wages by the time children born this year reach their mid-forties. Assuming that spending commitments and tax policy remain unchanged, a future of widening federal deficits is virtually assured.

All of these projections are based on calculations by the Congressional Budget Office and the Social Security and Health Care Financing Administrations. One might hope that the projections are an overstatement of the likely consequences of current policies. But, if anything, they are an understatement.

To begin with, the projections assume that all the savings promised in Clinton’s first budget will actually be made, even though (a) the administration’s estimates of future revenues from taxes have been seriously questioned, (b) its spending cuts include $116 billion in “unspecified reductions,” and (c) three quarters of its five-year spending cuts are conveniently scheduled to be carried out after the 1996 election. Second, the projections assume that no major new spending commitments will be made during the next fifty years, a questionable assumption (even within the next few years) now that Clinton has gone on record in favor of an “economic security system” for every citizen, which includes broad new entitlements in health, employment, retirement, and education. Clinton’s health plan alone promises that many new benefits—from prescription drugs to long-term home care to large subsidies for early retirees—will be made universally available with little or no regard to financial need.


Finally, the projections assume that the economy will grow steadily, with no new downturn in the business cycle and no meaningful rise in long-term interest rates. They also assume long-term rates of productivity growth and demographic change that are significantly more favorable (from a fiscal point of view) than the historical record over the past twenty years.5 Even in the near term, a sudden dip in the economy—together with faulty “savings” forecasts—can change a favorable economic outlook into a nightmare. This happened repeatedly during the 1970s and 1980s. Most recently, after the Budget Enforcement Act of 1990 was passed, the Congressional Budget Office projected in January of 1991 that the deficit would fall to $57 billion by 1995. Two years later, when Clinton took office, the CBO projection of the deficit for 1995 had climbed all the way to $284 billion. Over the long term, it seems questionable, to say the least, to assume that the country’s economic performance would not be negatively affected by increased spending on benefits, the interest costs that will result from large federal borrowing, and widening deficits. The coming explosion in entitlement spending will not only siphon a rising share of the economy’s resources away from workers’ paychecks (or company profits) and toward federal beneficiaries, it will also siphon those resources away from investment.

In my view, the only way to change the long-term trend of the deficit is by reforming the entire system of entitlements—the fastest-growing part of the federal budget. Reform will mean rethinking how public policies toward everything from health care and housing to retirement and farm aid help or hurt the public interest.6 Most of all, it will mean putting an end to the vast and largely unearned windfall we now give to the more affluent half of all American households.


The hopes of most Americans for an improving standard of living are fading. For objective evidence, we can point to the fact that average real wages, household wealth, and rates of home ownership have all remained virtually stagnant over the last twenty years, while during the same period the overall poverty rate has risen. What is even more discouraging, the wellbeing of younger Americans, as measured by each of these indicators, has become considerably worse. It is not surprising that according to opinion surveys a rising share of adults doubt that the next generation will be “better off” than they are. Once again, this attitude shift is especially marked among the young. A new Roper poll shows that, among eighteen-to-twenty-nine-year-olds of both sexes, the number of those who think they have a “very good” chance of achieving “the good life” has declined from 41 to 21 percent over the last fifteen years.

An extremely revealing appendix to the Fiscal Year 1995 Budget of the US Government entitled “Generational Accounting” supports these forebodings. In it, projections made by the economist Laurence Kotlikoff of Boston University show how fiscal policy will affect future generations if tax and benefit schedules affecting all of today’s living Americans remain unchanged. The projections reveal that Americans now in their mid-thirties to mid-sixties would end up paying between 31 and 36 percent of their employment income in net taxes (“net” means all taxes paid in excess of all benefits received) throughout their lifetimes. For all Americans born after 1992, however, their lifetime net taxes would consume 82 percent of their employment income. Accounting showing the impact of current policy on future generations has been included in the last three US budgets. Yet hardly any political leader has called attention to the startling implications of this analysis.

What accounts for such blindness? The answer seems to lie in a recent and unprecedented erosion in the sense of responsibility toward future generations that used to characterize US policy-making. Throughout our history Americans have engaged in fierce social conflict over the two central questions of political life: Who benefits? Who pays? Farmers against bankers, consumers against the trusts, unions against corporations, New Dealers against Liberty Leaguers—in all these struggles, no one questioned that the issue of costs would eventually have to be settled, one way or the other, among living generations. It was understood that, as a matter of equity, huge public debts would not be passed on to our children and grandchildren. Today that understanding no longer exists. One can, for example, observe in Washington an unspoken truce between “supply-siders,” determined to oppose any tax increase, and “liberals,” determined to oppose any cut in domestic spending. More often than not, both get what they want, at the price of a larger national debt.


To be sure, politicians often express their concern for future generations: conservatives talk of “family values,” liberals of “environmental stewardship.” But when the moment comes for them to specify, in dollars and cents, what they are willing to sacrifice so as to avoid imposing heavy debts on future generations, politicians on all sides have little to say. (Conservatives especially are guilty of unseemly hypocrisy when they shun all tax hikes in favor of vast spending cuts—but then offer nothing more specific than unenforceable “caps.” Liberals do something similar when they make exaggerated claims about the increases in revenue that will result from “taxing the rich.”) Huge fiscal burdens are now being shifted to the citizens who will grow up in the next century. Nowhere is this unacknowledged transformation—this replacement of intragenerational burden-sharing by intergenerational burden-shifting—more marked than in our acquiescence in the huge growth in entitlement expenditures.

What should be clear is that no one can morally justify encumbering our children with public obligations that erode the savings we should be putting aside for their future and that threaten to tax away an ever-rising share of their income. No one can know just what American society would be like if it had to accept static or declining living standards as a permanent way of life. But it is easy to imagine that it would be a far less generous and humane society than the one in which we now live.7

As most economists understand it, the connection between deficits and living standards is based on three propositions. First, higher living standards require that labor be more productive, i.e., that the output per worker increase over time. Second, such productivity depends critically on society’s stock of capital investment, consisting not only of plant and equipment but also of “human capital” in the form of education, skills, and technological knowledge. Finally, such investments are directly reduced by the extent to which borrowing to finance public deficits absorbs society’s supply of savings.

There is nothing controversial about this basic logic. No doubt it is technically possible for living standards to improve at a faster pace than labor productivity. A society that lends more money abroad than it borrows can do somewhat better than its productivity allows, just as a society that is a net global debtor can do somewhat worse.8 But for an economy as large as ours, this exception can rarely change the outcome by much in either direction. And it reinforces the overall argument by emphasizing the critical importance of national savings. Even if we assume that the level of domestic investment somehow remains constant, a larger deficit will increase the net savings that must be imported from foreign savers; a smaller deficit will increase the net supply of savings that will be available to export abroad.

That a nation’s productivity and capital stock depend on each other is, in fact, one of the very few propositions on which generations of economists—from Adam Smith to Karl Marx to Alfred Marshall to John Maynard Keynes to Paul Samuelson—have all emphatically agreed. This proposition is also supported by many empirical studies of economic performance. Since the early 1950s, for instance, the rate of productivity growth in each of the major industrial market economies has been very closely correlated with its rate of investment. (See chart 1.)

Chart 1

Foreign credit aside, the total amount of new productive investment in any society can never exceed the total amount of new domestic savings. It can be less, however, to the extent that savings are borrowed for unproductive purposes. It may not matter to private savers whether their dollars end up building a machine, training a worker, or purchasing a Treasury bond. But it matters a great deal to the economy. New machines or trained workers are likely to raise the economy’s productivity and income. A Treasury bond will not. (See Chart 2.)

Chart 2

The connection between lower deficits and rising living standards is sometimes misunderstood. First, it is necessarily a long-term relationship: its decisive impact shows up not over two or three years, but over two or three decades. Second, it is not the only thing that matters. Other global events or economic forces—technological discoveries, oil embargoes, or changes in private savings rates or the global cost of capital—can all have an independent impact on living standards. What makes deficit spending unique is that it’s the only force that representative government can fully control. Congress cannot legislate the price of energy, the discovery of room-temperature fusion, or even a rise in household savings. But Congress can legislate a balanced budget.

Since the early 1970s, the growth of national output for each fully employed American has averaged roughly 0.5 percent per year—a rate so low that it will require nearly 150 years for output per worker, the most fundamental measure of living standards, to double. If we were to aim at a sizable yet prudent increase in our long-term rate of productivity growth, say to 1.5 percent per year,9 this would reduce to forty-seven years the time it would take for living standards to become twice as high as they are now, clearly a large improvement. It wouldn’t return the US to its growth rates of the 1950s and 1960s or match Japan’s record during the 1970s and 1980s. But it would return US productivity growth close to its average rate over the entire last century—and it would bring the US to a rate close to most of its European competitors.

How much extra capital would we need to meet this goal? Most economists believe that it would require shifting 6 to 8 percent of GDP (between about $375 and $500 billion of current GDP) from consumption to savings available for investment. Using the late 1980s as a baseline, a shift of this magnitude would raise our net national rates of savings and investment to roughly 10 percent of GDP.10 Again, this is both very close to the average US net investment rate of the entire last century and close to the average rate in Western Europe.

But where is this amount of savings to come from? We could get most of it by finishing the job, which President Clinton has tentatively begun, of balancing the budget—and then keeping it balanced past the year 2000. Again using the late 1980s as a baseline, bringing the consolidated budget into cash balance would boost national savings by roughly 4 percent of GDP. Moreover, if we were to balance the budget not on a cash basis but on an accrual basis—which, as any good accountant knows, means looking at the entire federal balance sheet of assets and liabilities—we would boost national savings by at least 6 percent of GDP.11 Balancing the budget on this basis makes available a substantial cash surplus that could be used to pay back some of the Treasury debt owed to the public or could be spent on such public investments as preschool education, job training, infrastructure, or basic research. In my recent book, Facing Up, 12 I argue that 1 percent of GDP could be effectively spent on these kinds of investments by the year 2000. But however we use the cash surplus, the essential point is the same. The single most important step we can take today to improve the outlook for our own and our children’s future living standards is to plan for long-term budget balance according to honest accounting rules.


Why isn’t Clinton’s approach to deficit reduction adequate? He has, after all, raised taxes, ensuring that, as a share of GDP, average yearly federal revenues during the 1990s will be higher than during any prior decade in US history. He is also aggressively cutting back on the Pentagon’s budget, so that by 1998 we will be spending less on defense, as a share of GDP, than in any year since Pearl Harbor. As for future domestic discretionary spending—for R&D and space exploration, highways and bridges, parks and the environment, immigration control and law enforcement—this too is expected to decline as a share of GDP, and by 1997 will sink beneath the postwar record low it reached late in the Reagan administration.

The problem of the deficit lies elsewhere: in Washington’s failure to find large and credible savings in entitlements. (See Chart 3.)

Chart 3

In 1993, entitlements amounted to $761 billion, or 54 percent of all federal spending, with 40 percent of the total going to Social Security, 8 percent to federal pensions, and 29 percent to Medicare and Medicaid. And this doesn’t include $150 billion more in consumption benefits dispensed through the tax code—so-called tax expenditures—such as the deduction for home mortgage interest and the exclusion from employees’ taxable income of health care paid for by employers.

Before Clinton’s first budget was enacted, direct entitlement spending was projected to grow by $418 billion in real 1992 dollars over the next ten years. Now, with the President’s deficit reduction plan passed into law, it is still projected to grow by $373 billion between 1994 and 2004—which, along with interest costs, would account for more than the total of all growth in federal spending during those years. This result is possible because the combined outlays for all other functions of government are, under current law, scheduled to decline by $29 billion in real dollars during the next decade. Such an outcome should be deeply troubling to liberals who hope for an expansion in social programs that would address the needs of poor and disadvantaged Americans. In addition to crowding out investments from private capital markets, federal deficits bloated by entitlements increasingly crowd out public-interest spending from the federal budget. In doing so, they put a moratorium on any vision of progressive government.

President Clinton did not promise to leave all entitlements untouched, but he made a sweeping social promise that was just as unfortunate in its effects. He assured the public that nothing in his budget plan would take much away from any household belonging to the “middle class.” He even implied, in his loose campaign rhetoric, that middle-class Americans were victims deserving reparations. According to Clinton’s definition, moreover, only 1 percent of all US households—those with gross incomes over $200,000—are too rich to be included in the middle class.

As a guide to tax reform, Clinton’s promise was unfortunate, since it ruled out most measures that might have raised more revenue from the 87 percent of all pretax income earned by households beneath this high threshold. The small increase in the gasoline tax asks the middle class to “sacrifice” by paying 4.3 cents per gallon (or an average of $27 per driver per year), but this sum is a mere one fiftieth of the total gasoline tax typically levied in other industrial countries.

Clinton’s promise to protect the middle class, however, was most damaging in its impact on entitlement reform, for this promise precluded major savings from the largest and fastest-growing portion of the budget. Americans with incomes over $200,000 received just 1 percent ($5 billion) of all federal entitlement outlays in 1991. No one is or should be talking about significant entitlement cuts for lower-income Americans. This leaves all households with incomes above the US household median—roughly $30,000 in 1991—yet below $200,000, 13 i.e., that part of Clinton’s “middle class” which, however hard-pressed, cannot claim to be destitute. In 1991, such households received no less than 40 percent ($265 billion) of all federal benefit outlays, including 38 percent of Medicare and 42 percent of Social Security—an institution originally founded to provide “a minimum” and to “prevent destitution and dependency.”14 They received an even greater share—67 percent ($93 billion)—of all the consumption benefits that come from tax subsidies. It is often observed that even cutting the Pentagon’s funding by half would still not balance the budget. It is rarely but just as truly observed that cutting off entitlement outlays to just the more well-to-do half of American households would balance the budget, and would do so with a comfortable surplus to spare.

On entitlements no less than on taxes, Clinton fudged a bit on his promise to the middle class. Some middle-class entitlements were trimmed—notably Social Security (with a new tax on benefits) and Medicare (with provision for smaller increases in reimbursements to doctors and hospitals). But like the four-cent gas tax, these measures are unlikely to produce serious savings. The new tax on Social Security benefits only affects members of the upper-middle class. The health-care savings claimed in the President’s 1994 budget are guesswork at best (and don’t, of course, take into account the new spending on middle-class health entitlements proposed in Clinton’s Health Security Act). In fact, federal health-care spending projections typically overestimate the savings that will result from reforms. Most recently, an effort to save $43 billion by limiting Medicare charges as part of the 1990 budget deal ended up saving nothing at all.

President Clinton is to be commended for concentrating attention on the need for far-reaching healthcare reform, but he has so far failed to encourage a genuine public debate on the kinds of long-term savings measures that are necessary. Most experts believe that achieving lasting savings in health care will require patients and doctors to accept a system with new incentives that cause them to weigh the costs and benefits of medical care. Many further believe that cost containment will ultimately mean accepting some form of “rationing” of care that has high costs but low benefits.15 Eliminating “waste” in health care will alone not be sufficient, since real medical costs are rising primarily because of the development of expensive new technologies and procedures that have at least some benefit. Yet, thus far, the administration has publicly maintained that its health plan can achieve large-cost reductions without sacrifices, even as it expands benefits.

Some of Clinton’s supporters nonetheless argue that the administration’s health plan will be able to force real cuts in medical services, thereby controlling costs. The plan, after all, is to limit the total of health-care dollars spent by restricting increases in insurance premiums; it is also supposed to limit costs by prohibiting people from purchasing care outside of its new health alliances. Serious practical questions have been raised about whether the plan’s immensely complex regulatory apparatus will work as intended. More fundamentally, it is a political fantasy to believe we can use regulation to clamp down on costs without a public consensus about the need to be cost conscious and make sacrifices.

The current health-care debate illustrates how the politics of entitlements in America has become a politics of denial heavily influenced by special interest groups representing “entitled” claimants—such as the powerful lobby of senior citizens—and only weakly influenced by citizens representing the general interest. That is why the typical household with an income over $100,000 actually receives more in total federal benefits than a household with an income under $10,000. That is also why the claims of current generations are strongly favored over the interests of future generations—as is shown by an unfunded future federal benefit liability totaling some $14 trillion,16 four times larger than the official national debt. Meanwhile, an increasingly affluent generation of people sixty-five years old and older continues to receive eleven times more in per-capita federal benefits than an increasingly impoverished generation of children below the age of eighteen17—a far greater disparity than in other industrial countries.

No one should blame Clinton’s party, still less President Clinton himself, for the social and political forces that have caused the cost of entitlements to expand from one quarter to more than one half of the federal budget during the last thirty years. Most of the mistakes that created this growth—including overindexing of benefits, naive cost projections, and absurd overconfidence about our future prosperity—occurred in the late 1960s and early 1970s during the Nixon as well as during the Johnson administration. Ever since, the central problem has been the inability of both parties to admit that we cannot afford the growing costs of what we pretend we cannot control. (See Chart 4.)

Chart 4

At the same time, we are losing a political (and economic) opportunity that won’t soon repeat itself. Already, the US economy is severely burdened by the lowest savings rate and highest consumption rate in the industrial world. During the coming decades, if there are no radical reforms, technological, social, and demographic changes are bound to put public budgets under severe strains. More and more expensive medical treatments will become available as the result of improvements in technology. Disability and workers’ compensation will broaden their coverage as claims for treatment of job stress and mental illness, for example, are recognized and accepted. If the recent sharp increases in the AFDC and food stamps rolls are any indication, families will continue to break up into smaller units of needy dependents. Seriously compounding all these trends will be the aging of the US population, as the proportion of people sixty-five and over grows from 13 percent today to between 21 and 25 percent in the year 2040.

The long-term fiscal projections allow only one conclusion: sooner or later we will have to change course. If we act soon to control the costs of entitlements, the changes can be voluntary, gradual, and humane, with enough time remaining to adjust our personal habits and expectations. If we act later, the change is more likely to be involuntary, sudden, and draconian—and made in the midst of economic crisis and social upheaval.

The usual solutions no longer apply. In his 1995 budget, President Clinton proposes eliminating 115 separate discretionary federal programs—for a total savings of $3 billion. Meanwhile, automatic cost-of-living adjustments to federal cash entitlements will alone add roughly $12 billion to the deficit in calendar year 1994. Even if we immediately and permanently cut defense spending by half, twenty years from now we would again face the same deficit burden as we do today. We can keep raising tax rates, but to keep up with the growth of entitlements alone we would have to enact a new tax hike, equal to the one sponsored by Clinton, every four years for the next half century. What must now be faced is the need to cut entitlements; and this cannot be done without sacrifices by the middle class.


Many politicians agree that controlling the growth in entitlements is essential, but then they protest that achieving this goal is impossible without ravaging the poor, inflicting unbearable hardships on the elderly, or confiscating benefit claims that are an “earned right” for all classes. Such a view, however, rests on three powerful myths: (1) that entitlements are primarily a “floor of protection” for the needy, even though the 21 million US households with incomes above $50,000 received about $115 billion in benefit outlays under major entitlement programs in 1991; (2) that the elderly are disproportionately poor, even though older Americans now have the lowest total-income poverty rate of any age bracket18—and the highest levels of per-capita household income and per-capita household wealth; and (3) that beneficiaries are just “getting back what they paid in,” even though most retired Americans today receive two to five times more in Social Security benefits—and at least ten times more in tax-free Medicare benefits—than what they and their employers contributed, including the interest on their contributions.

Politicians hesitate to challenge these myths. They understandably fear the political consequences if they do anything that might antagonize the members of America’s vast middle class or that might threaten any benefit that well-organized interest groups, particularly the “senior citizens” lobbies, are committed to defend.

If the myths I have listed can be exposed as false, the prospects for eliminating the deficit distinctly improve. We could control the growth in entitlements without hurting the poor or imposing undue hardship on any American. We could create an equitable and sustainable social welfare system. But to do this, we would need a program of reform with four basic components: tax-code changes, so that federal benefits are treated in much the same way as other kinds of income; a strategy for containing health-care costs; a revision of retirement programs to take into account longer life expectancies of Americans; and, most important, an “affluence test,” which would scale back the benefits to higher-income households.19

Tax-code reform should begin by following the example of other industrial nations and make public benefits just as taxable as any other income. Until 1994, only 50 percent of Social Security cash benefits were taxable, and even then only for families with incomes above certain thresholds ($25,000 for a single beneficiary and $32,000 for a couple). As President Clinton proposed in his first budget, we should tax 85 percent of all Social Security benefits as if they were regular income. (The 15 percent exemption can be justified as a generous estimate of the dollar value of previously taxed personal FICA contributions.) But we also need to go a step further and eliminate the thresholds.20 As for Medicare, which is now entirely tax-free, I propose making 25 percent of its insurance value taxable to beneficiaries (about the share of private employer-paid health insurance that would also be taxable in my plan).

The changes in the taxability of benefits I propose would reduce the deficit by about $19 billion in the year 2000 and by $25 billion in 2004. But since these changes would not affect beneficiaries with an income beneath the taxable minimum ($12,300 in 1993 for an elderly couple), they would affect no one in the lowest income bracket.21 At higher incomes, beneficiaries of Social Security and Medicare would be required to pay some tax. Small or large, the tax on their benefits would be at the same rate as all Americans in the same tax bracket pay on their incomes. An equitable and progressive tax code is a “means test” from which no citizen should be exempted.

We should also cut back other costly and regressive entitlements that are now dispensed through the tax code, for example the home mortgage interest deduction—a $46 billion tax subsidy that effectively steers scarce US capital away from improving the equipment in our factories and toward home improvements in the suburbs. Currently, a taxpayer can deduct interest on a home mortgage up to $1,000,000. My plan would further limit the size of a home mortgage on which interest is tax deductible, with the limit set at $250,000 for a couple. This reform would cut the deficit by about $6 billion in 2000 and by $10 billion in 2004—a savings that will steadily grow since the limits would not be indexed for inflation.

Another candidate for reform is the tax exclusion for employer-paid health care. This perverse subsidy, which in 1994 will cost the Treasury $76 billion in lost tax revenues, benefits relatively well-paid Americans the most while offering nothing to those who are poor, unemployed, or uninsured. And it is most generous to employees who sign up with the most expensive and permissive insurance plans. My plan would continue to allow this exclusion but would limit it to $410 per month for family coverage and $170 per month for individual coverage—about the current average cost of insurance. Above these levels, anything employers spend on health care will be taxable income to employees. This would cut the deficit by about $40 billion in the year 2000 and $56 billion in 2004.

So we arrive at the necessity of controlling health-care costs—a major cause of the growth in entitlements. A workable approach would be very different from the President’s. Rather than promises of cost-free savings, it would emphasize the specific policies needed to make our system more cost-conscious. In addition to deep and immediate cuts in tax exclusions for employer-paid health care, it would raise the deductibles and copayments (the share of medical bills the patient has to pay out of his own pocket) required for publicly funded programs. Also needed are (1) a complete overhaul of the malpractice system, which now imposes unreasonably high insurance costs on doctors, encourages “defensive” medicine, and inflates the cost of customary practice; (2) standardized forms (and possibly financial incentives) to popularize the use of living wills, which, as the Clintons have rightly stressed, would help cut down on unwanted and costly medical interventions during the last days of life; and (3) improved research into the effectiveness of medical treatment so that doctors and patients will have a clearer idea what procedures work, and at what cost, and in order to develop uniform and more cost-effective guidelines for doctors. An effective health plan would also be as economical as possible in its new spending commitments, making affordable insurance coverage available to all those who now lack it (as my plan does),22 while avoiding the creation of new open-ended, middle-class entitlements.

Setting limits on the current openended tax exclusion for employer-paid health care is critical. Such limits will not only raise new revenues, but, by creating incentives to enroll in cost effective plans, will also cut the excess demand for health care now induced by overinsurance. At the same time, cost consciousness among Medicare beneficiaries should be encouraged through new incentives in Supplementary Medical Insurance (or SMI), the fast-growing part of Medicare which covers expenses outside hospitals. I propose raising the SMI annual deductible from $100 to $150 and raising current SMI copayments from 20 percent to 25 percent (while requiring 20 percent SMI copayments for the use of clinical laboratory services, home-health services, and skilled-nursing facilities, none of which now requires copayments). The current SMI premium that Medicare beneficiaries pay would also be increased so that it covers 30 percent of the program’s costs, up from 25 percent under current law. Together, these changes in Medicare would directly save the budget about $28 billion in the year 2000 and $42 billion in 2004.

Although the health-care reform strategy I have outlined would increase cost consciousness throughout the health system, no one, frankly, can claim to know how much it will slow the growth in health-care costs or how quickly. That is why I advocate an incremental and experimental strategy that explicitly awaits the actual effect of one set of policy changes (not just an optimistic guess) before planning the next. After putting into effect the changes in incentives I’ve described, in other words, we would wait for the cost results, then decide if stricter measures are necessary or if further expansions of benefits are affordable.

In addition, to ensure that we meet a savings target within a reasonable time frame, my plan includes a further reform: as in most other industrial countries, Congress would establish an overall budget for federal health-care spending with annual limits on each federal health-benefit program. For Medicaid, the limits could be enforced by lump-sum grants to states. For Medicare, they could be administered by giving a fixed per-capita voucher to each beneficiary with which to purchase a certified insurance plan.

Since 1980, federal health-care spending has grown at the staggering average annual rate of 11 percent. The goal of the federal health-benefit budget I propose would be to bring the rate down below current projections by a modest one quarter of a percentage point a year. Still, the budget savings from meeting even this goal would be substantial: about $20 billion in the year 2000 and $77 billion in 2004.

Along with setting limits on federal health-care benefits, we should be preparing our entitlement system for the inevitable aging of our population. Instead of encouraging people to retire earlier, often with inadequate financial resources, we should follow a more responsible, and possibly more rewarding model: later retirement, larger family savings, and the better use of the experience of older employees in the workplace.

In 1983, Congress raised the Social Security full-benefit retirement age from sixty-five to sixty-seven. The increase is to go into effect, in stages, between 2000 and 2022. This step is useful but too small and too slow. (In fact, Congressman Jake Pickle observed at the time that the distant date was chosen precisely because it would affect so few Americans who were yet old enough to vote.) By 2022, life expectancy at age sixty-five will be not two years longer than it was when Social Security was founded, but at least six years longer. And no one needs, as the 1983 act allows, half a lifetime to plan for such a change. Accordingly, my plan stipulates that the rise in Social Security’s retirement age should begin in 1995 and proceed by three months per year until a new retirement age of sixty-eight is reached by 2006. Eligibility for “early retirement” would be simultaneously adjusted so that today’s “reduced” benefits at sixty-two will become available at sixty-five. This would reduce the federal deficit by about $16 billion in 2000 and $34 billion in 2004.

At the same time, we should trim overly generous federal employee pensions to bring them more in line with the (already generous) plans of Fortune 500 corporations.23 We should scrap policies, such as the Social Security earnings test, that deter older Americans from working. Within an overall strategy of entitlement reform, we should also use some of the budget savings and expand the safety net—Supplemental Security Income or SSI—for those older people who are in genuine need. Along with higher SSI benefit levels, I advocate lowering the SSI eligibility age to sixty-two, which should provide better income protection to those elderly people who are unable to work as the Social Security retirement age rises.

Taken together, the reforms I have thus far outlined will ensure substantial and permanent budget savings; the savings from health-care and retirement-age provisions in particular will not only be permanent but will compound as the US population ages and the number of federal beneficiaries grows.

On both fiscal and moral grounds, however, we need another reform that is specifically designed to scale back entitlement payments to higher-income households. This can be done—without hurting the poor or setting up a huge new federal bureaucracy—by applying an “affluence test.”

How would it work? Although such a test amounts to a cut in benefits, it would be administered through the tax code. Each year all taxpayers with incomes above the US household median—which my proposal assumes will be $35,000 in 1995, when this reform would begin to take effect—would be required to report all their estimated combined federal benefits for the coming year. The relevant federal agencies will then withhold some fraction of those benefits, based on each beneficiary’s total household income. Reductions in benefits would of course be adjusted in the event people faced changed economic circumstances during the year or in subsequent years.

Households with incomes under $35,000 would not lose a penny in benefits. Beginning in 1995, higher-income households would lose 7.5 percent of all benefits that cause their incomes to exceed $35,000, plus 5 percent for each additional $10,000 in income. Thus, a household with $40,000 in total income and $10,000 in federal benefits would lose 3.75 percent of its benefits—only $375.24 A household with $55,000 in income would lose 12.5 percent of its benefits: $1,250. And a household with $95,000 in income would lose 32.5 percent of its benefits: $3,250. For most types of benefits, the maximum benefit reduction rate would be set at 85 percent, applicable to incomes over $185,000.25 All income brackets would be adjusted upward each year to take inflation into account.

The affluence test has clear advantages over other methods that have been proposed to reduce entitlements. We might, for example, scale back subsidies for the relatively well-to-do by rewriting the benefit formulas and eligibility criteria of Social Security and every other entitlement program. But that would require amending every federal benefit statute—a process that would have to make its way through dozens of congressional committees and would be endangered at every turn by lobbies bent on ensuring that their constituency is not the only one singled out for sacrifice. Or we could, as some experts have proposed, use benefit-related cuts in cost-of-living allowances in order to reduce welfare for the well-to-do. But such a plan could not distinguish between the retired civil servant with two pensions and a minimum Social Security benefit and a retired widow living alone who subsists entirely on a bigger than average Social Security check.

The affluence test overcomes these obstacles to reform. It would include all federal benefits—not just Social Security and Medicare, but everything from farm aid to veterans benefits to federal pensions. It would thus ensure that everyone who is able to make a fair contribution does so. No single group, most obviously the elderly, need fear bearing an unfair burden. The affluence test, moreover, always takes into account a household’s total current income, not just the size of its federal benefit checks. It therefore cannot inadvertently hurt households below the national median, and among higher-earning households, the progressive sacrifices it calls for are always based on beneficiaries’ true financial circumstances in any given year. Finally, it could be enacted in a single piece of enabling legislation.

During the last year a number of criticisms of affluence testing have been made, mostly by writers who have long opposed any significant changes in established entitlement programs.26 It has been said that, since the withholding rate rises with total household income, an affluence test amounts to a “tax” on savings and would therefore discourage thrift. But the net impact (plus or minus) of such a “tax” is theoretically ambiguous and, in any case, its magnitude is small.27 (By the critics’ perverse reasoning, any attempt to favor the needy is necessarily a “tax” on the non-needy.) It has also been said that reducing benefits in a social insurance program cannot possibly shrink the deficit because the selfish public will demand an instant and equal reduction in earmarked payroll taxes. This claim has no historic merit.28 But it perhaps does reflect a broader fear that an affluence test will ultimately lead to a disintegration of the entire social safety net. After all, the reasoning goes, if we don’t somehow bribe the affluent with munificent “universal benefits,” won’t we undermine political support for any assistance to anyone? I believe that Americans can rise to a higher sense of public purpose, especially if the disastrous consequences of longterm fiscal trends are made clear.

The affluence test, in my view, is a policy reform of primary importance, one that would result in enormous budget savings: $90 billion in 2004, a sum that will then climb to about $207 billion a year in 2020 and $524 billion a year in 2040.

These entitlement reforms, of course, assume we are already taking action on the rest of the budget. Specifically, they assume that the deep cuts already scheduled by last year’s budget deal for spending on national defense and domestic discretionary programs can and will be achieved.29

Because of the huge projected growth in entitlements, savings in benefit and interest outlays will have to exceed new taxes in any workable budget plan. In mine, these savings greatly exceed new taxes—by 1.7-to-1 in 2004, a ratio that will grow to 4.9-to-1 by 2020 and 11.4-to-1 by 2040. Already by 2004, my plan will be saving 1.7 percent of GDP in entitlement spending; by 2040 it would be saving 3.0 percent of GDP, or $380 billion in today’s dollars. Still, balancing the budget does require higher tax rates or entirely new taxes beyond those enacted by Congress last year. My own plan includes higher taxes on tobacco and alcohol, a 5 percent across-the-board consumption tax (ideally a progressive “consumed income” tax),30 and a phased-in forty-cent gasoline tax, not only to raise revenues, but to limit an environmentally harmful type of consumption.

The best way to bring the budget into balance is the one that will best meet the economic and social needs that have caused us to be concerned about the deficit in the first place. While it is essential that the pool of savings available for private investment be expanded, it is also essential that we devote more public resources to productive public investments in parts of the economy where reliable private capital markets do not exist. I suggest that the federal government invest an additional 1 percent of GDP in America’s physical and human infrastructure—not just high-speed trains or deep-water ports but full funding for Head Start—and that we make sufficient budget cuts to pay for it. In general we should do whatever we can to encourage more private savings (thus my emphasis on consumption taxes) and eliminate subsidies that reward unproductive uses of private capital (thus my proposals to reduce tax breaks for housing and much health care).31 By curbing public largesse for the relatively well-off, we can also make our entitlement system itself more equitable. This is why I propose expanding benefits for those who are in the most serious need—especially the elderly poor, the disabled poor, and the children of low-income families.


No doubt vigorous action on the deficit could present risks to the economy if all we do is suddenly raise taxes and cut government spending. No one is proposing anything of the kind. A workable plan must be gradual and should not try to raise national savings faster than investment—or exports—can expand. It must provide for budget outlays and tax incentives that will promote investment—not just rely on lower interest rates—in order to ensure that investment purchases accelerate at least as fast as consumer purchases slow. If we want households to become bigger savers32 and businesses smarter investors, we must also commit ourselves to changing tax and benefit policies that push them in the wrong direction.

Still, for many politicians intent on postponing today’s problems to tomorrow, deficits will always be the least threatening of evils. Time and again, the politicians and others who oppose a long-term program to reduce the deficit defend their inaction by invoking the fear of short-term economic collapse and by stressing the need to bolster confidence among consumers and investors. But the reality is quite the opposite: we desperately need a national commitment to a long-term savings goal not only to achieve long-term prosperity but to create the very consumer and investor confidence needed to shore up the current recovery and, importantly, to bring down still historically high real long-term interest rates. (A more responsible fiscal policy will also mean that the Federal Reserve can safely pursue a more expansionary monetary policy, further helping to bring down interest rates, and thus providing more “push” behind the investment spending needed to encourage growth and keep the economy on the path toward full employment.) Indeed, continued failure to address our long-term problem could set off far more acute short-term problems—a financial crisis or a new surge in interest rates—that themselves may pose the largest risk of recession.

The plan I have offered makes the needed commitment. It would go into effect in 1995, balance the budget by the year 2000, and generate a budget surplus of 1 percent of GDP by 2005. (See Chart 5.)

Chart 5

By that year, this surplus, together with the 1 percent of GDP in new federal investment spending, would put the budget in balance on an accrual basis. For decades thereafter, the plan would keep outlays roughly in line with revenues. Under current law, the national debt is projected to rise as a share of GDP every year after 1998, and even pass 200 percent of GDP by 2030. Under my plan, it is projected to be just 31 percent of GDP in 2030.

Why choose the year 2000 to achieve a balanced budget? I believe the goal must be near enough both to encourage political leaders to act and to enable the public to anticipate the economic payoff. It must also be far enough ahead (as the year 2000 is) to allow us to raise our savings rate at a year-by-year pace that we have matched earlier in American history. Moreover, the goal must be to put the budget completely in balance—not just make another down payment.

Each year adds another deficit to the national debt, saddling us with another $10 to $20 billion in annual interest charges; each year, meanwhile, US living standards continue to stagnate, a result traceable to a failure of investment. Each year, finally, brings us closer to the arrival of a huge increase in the numbers of retired people—with all of its explosive implications for entitlement payments. For the next fifteen years, the unusually large Baby Boom generation will be entering its peak income years and the unusually small generation born during the depression will be retiring. Beginning around 2010, however, this quantitative relationship between workers and retirees will begin to swing the other way.33 By 2040, we will see at least a doubling in the number of retired Social Security and Medicare beneficiaries age sixty-five and over—and the tripling to quadrupling of the number of very old, age eighty-five and over, who consume twice as much medical care per capita as the younger elderly, and twenty times as much nursing home care. Meanwhile, the number of tax-paying workers may have grown by no more than 15 to 30 percent. It seems obvious that an aging society must invest heavily in the education and productivity of the young. What is less obvious—when the impending claims of the retired are added to a budget already deeply in the red—is how we will finance this investment.

A clear choice must be made between a society fated to a bitter decline in living standards and one that is willing to confront reform of entitlements. At the recent Bryn Mawr conference on entitlements, President Clinton reaffirmed his early commitment to doing something about the budget deficit. He also emphasized the importance of the newly appointed Kerrey-Danforth Commission on entitlements reform, which is scheduled to issue a report by December 1, 1994, and offers hope for a bipartisan solution to fiscal problems that have too long been deferred.

There really is no decent alternative to reducing entitlements, but if it is to be done, it must be based on a consensus on the need for broadly shared sacrifice. No special interests can be singled out for privileged treatment nor can any be spared. If older workers must postpone their retirement so too must young professionals lose some of their mortgage and health-insurance tax privileges. If well-off seniors must forgo some fraction of their Social Security benefits, so too must well-off farmers, veterans, and federal pensioners lose a fraction of their entitlements. And if every group—rich and poor, old and young—must bear some part of the burden, the contribution of the middle class, because of its size, must be the largest of all.

George Bernard Shaw said: “I have to live for others and not for myself; that’s middle-class morality.” Today, America’s middle class has understandably adopted a kind of siege mentality as its expectations of higher living standards have faded. Eliminating deficits will hardly require the degree of self-abnegation Shaw called for; but it will mean a willingness on the part of the middle class to rely somewhat less on privileges financed by public borrowing. Which politicians will have the courage to make this clear?

This Issue

April 7, 1994