September 2004
Federal Reserve Bank of Dallas

Social Security and Medicare: No Free Lunch

Public attention has recently focused on the federal budget outlook for the upcoming decade. But, as Alan Greenspan has often noted, the real budget challenge is the long-run growth of Social Security and Medicare.[1]

In this article, we first discuss why these programs are big and getting bigger, outpacing the growth of revenue. We show that large tax increases or benefit cuts will occur to address this shortfall, no matter how much we might wish they could be avoided. We explain that Social Security and Medicare involve transfer payments from the young to the elderly rather than actual saving. We then explain that scaling back these transfer payments would increase national saving and give future generations a better standard of living. Doing this would, however, impose a transition cost on current generations.

Many people hope for, and some people promise, a free lunch that will allow this transition cost to be avoided. Unfortunately, there is none. It is possible to shift the burden from one group of people to another, but no policy proposal—including privatization—offers an escape from that burden. If future generations are to be made better off, the transition cost must be paid.

Programs Are Big—And Getting Bigger

Figure 1 shows federal spending, other than interest on the debt, as a share of GDP. From 1960 through 2004, such spending has fluctuated around an average value of 17.3 percent of GDP. But the Congressional Budget Office’s December 2003 long-run budget projection paints a much different picture for the future. Non-interest spending is projected to rise relentlessly, to 23.4 percent of GDP by 2050, with no letup in sight.[2]

Figure 1

The federal budget includes thousands of spending programs. But Figure 2 shows that the spending surge is primarily driven by just two—Social Security and Medicare—with some contribution from a third—the federal portion of Medicaid. In fact, CBO projects that non-interest programs other than Social Security and Medicare will shrink from 11.1 percent of GDP in 2004 to 9.0 percent in 2050.[3] If these other programs don’t shrink, then the total spending growth will be even more dramatic.

Figure 2

How large will Social Security and Medicare become? From 2004 to 2050, Social Security spending rises from 4.2 percent of GDP to 6.2 percent. Over the same period, Medicare grows explosively, from 2.5 percent of GDP to 8.3 percent.

A variety of factors contribute to this growth. One such factor is the retirement of the baby boom generation, which will swell the ranks of retirees for the next few decades. That’s a temporary phenomenon, though. A law adopted last December adds a prescription drug benefit to Medicare, starting in 2006. That will raise costs as well, but it too is a secondary factor.

The two forces that account for most of the long-run spending surge are longer lifespans and rising medical costs.

The Social Security trustees project that life expectancy at age 65, which is now around 17 years, will steadily rise almost half-a-year per decade (Figure 3). CBO uses this same assumption in its long-run budget projections mentioned above. The Census Bureau, like many private demographers, projects increases about twice as rapid—nearly one year per decade. And the faster lifespans rise, the more Social Security and Medicare must pay.

Figure 3

The second force driving up program spending is the ongoing rise in medical costs. The Medicare trustees projects that spending per beneficiary in Medicare Part A (the hospital part of the program) will quintuple over the next 75 years, even after adjusting for overall inflation (Figure 4). Of course, medical costs are hard to predict, but some experts believe that costs will rise even more rapidly—placing an even greater strain on Medicare.[4]

Figure 4

Costs Will Outpace Revenue

Although spending is scheduled to grow sharply under current law, revenue is not scheduled to keep pace.

The nonhospital part of Medicare is financed from general revenue. Social Security and Medicare Part A are financed by earmarked taxes—primarily a payroll tax on employee compensation and an accompanying tax on self-employment income. The tax rate is 15.3 percent, up to a threshold ($87,900 in 2004) linked to national average wages, and is 2.9 percent thereafter.

This tax rate is not automatically adjusted for increases in lifespan or medical costs, even though these factors do automatically increase spending. As a result, future payroll tax revenue will not be sufficient to cover future benefit costs. The trustees estimate that Medicare Part A will be unable to pay full benefits after 2019 and that Social Security will be unable to do so after 2042. Of course, the exact years depend on various assumptions, but the day will come when revenue no longer covers costs.

How can this financial shortfall be addressed if we are to maintain promised benefits? We will have to come up with more money. How much more? Well, if we continue to rely on the payroll tax and if we keep revenue and spending in balance each year, the tax rate would need to rise ever higher to keep up with rising costs. By 2080, the tax rate would have to roughly double, to 31 percent, to cover that year’s Social Security and Medicare Part A benefits.[5]

Or, the shortfall could be addressed through income-tax hikes and discretionary spending cuts. For example, we could raise income tax revenue by about one-third, but such a large tax increase would likely reduce economic output and have other undesirable consequences. On the spending side, even the complete elimination of discretionary spending (excluding only Social Security, Medicare, Medicaid and interest) wouldn’t be enough to cover the shortfall. But substantial tax hikes could be combined with substantial spending cuts to raise the required amount of money.

The alternative is to reduce promised benefits, and there are many ways to do this. Eligibility ages for Social Security and Medicare could be raised by several years in line with longer lifespans. Means tests could be imposed on either or both of these programs, making them more like welfare. Social Security cost-of-living adjustments could be trimmed by using a more conservative measure of inflation, as Alan Greenspan and others have proposed.

Two other possibilities would change the rate at which future benefits rise. Social Security benefits for each cohort of retirees are currently tied to average wages in the economy at the time that the cohort attains age 60. Since prices generally rise more slowly than wages, we could reduce future spending by tying those benefit levels to prices rather than wages. This “price indexation” option has been suggested as a leading option (though not definitively endorsed) by the President’s Commission to Preserve and Strengthen Social Security and the Council of Economic Advisers.[6] Price indexation has also been mentioned favorably by others, including the Cato Institute, the Wall Street Journal and the Concord Coalition.[7]

A somewhat similar proposal can be applied to Medicare. Under current law, Medicare benefits are tied to rapidly rising medical costs. We could reduce future spending by linking those benefits to wages or even to prices.[8]

Reducing promised benefits doesn’t necessarily mean future retirees would receive smaller monthly benefit checks than current retirees do. But it does mean they’d receive less than current law now promises them—about fifty percent less in 2080, if the books are to balance in that year.

Reform plans can be simple or complicated, can raise taxes or cut promised benefits, can build up a trust fund or privatize the system—there are as many plans as there are economists (maybe even more). But the major economic effect of any reform plan depends on one simple feature: whether the plan imposes additional burdens on the young or reduces transfer payments to the old. Reducing these transfers helps future generations enjoy a better standard of living but requires current generations to bear a transition cost. Maintaining the transfers helps current generations avoid sacrifice but requires future generations to pay the tab in the form of a permanently lower standard of living. That feature, and that feature alone, determines the gains to future generations and the transition cost imposed on current generations.

To understand these conclusions, let’s look at how Social Security and Medicare operate.

Pay-As-You-Go Retirement Programs

Social Security and Medicare are pay-as-you-go retirement programs. In such programs, each generation provides benefits to its parents and each receives benefits from its children. Because workers’ contributions are transferred to the preceding generation, there is no actual saving. The programs accumulate no assets; they are merely a sequence of transfer payments from young to old.

Because no actual investment occurs, no generation earns an investment return. Nevertheless, a rate of return can be computed for each generation based on the size of the transfer payment that it receives from its children compared to the payment that it makes to its parents.

The long-run rate of return equals the growth rate of national labor income. If each generation pays a fixed percentage of its labor income to its parents and gets back the same percentage of its children’s labor income, the return depends on how fast labor income grows between generations.[9] Of course, a higher rate of return is possible as long as the tax rate continues to rise, but it can’t rise forever (certainly not above 100 percent).

This return is lower than what can be obtained through private saving. From 1929 to 2003, the growth rate of labor income averaged about 3.4 percent per year, after adjusting for inflation.[10] In contrast, economists have estimated an average pretax marginal product of about 6 percent on the capital—such as plants and equipment—that can be financed with private saving. [11]

(To be sure, both the 3.4 and 6 percent figures are historical averages, not mathematical certainties. These rates can be higher or lower in any given year, but since we are talking about long periods of time it is appropriate to use averages.)

The difference between 3.4 and 6 percent may not seem very large, but it is. Over a 28-year period, less than a working lifetime, it cuts the average person’s retirement payout in half, as can be seen in Figure 5. Going forward, the falling birthrate is likely to reduce both returns, but the gap between them will remain significant.

Figure 5

This below-market return explains why future generations would be better off if they could put less money into the pay-as-you-go system and invest more elsewhere. Although each generation would receive a smaller transfer from its children, it would come out ahead because it could earn market returns on the money it would otherwise transfer to its parents. The macroeconomic effects of this change would also be beneficial because it would permanently increase national saving and enlarge the nation’s capital stock.

Bigger reforms would provide higher returns and greater national saving. Complete replacement of pay-as-you-go—an extreme option—would allow each generation to avoid below-market returns entirely, earning market returns on its entire savings.

Alternative Retirement Systems

If workers transfer less money to their parents, what should be done with the money instead? One possibility would be for workers to just save the money on their own and receive market returns. But letting individuals fend for themselves would undermine the social protections Medicare and Social Security are intended to provide, such as a safety net for the elderly poor. Under a truly voluntary system, some workers might earn too little to save very much, some might choose not to save, and some might lose their savings through bad luck or bad investment decisions. This could potentially raise the poverty rate among the elderly.

The current system protects against these contingencies by providing benefits to all retirees, spreading benefits throughout their retirement lifetimes and providing more generous benefits relative to taxes paid (though not in absolute terms) for retirees with lower lifetime earnings. These social protections are important—but the current system is not the only way to provide them. At least two other approaches could preserve social protections while still providing better returns for future generations.

Government Saving
The first approach is for the government to save the money on behalf of each generation, collecting taxes from each generation while working and saving the money to pay benefits to that generation when retired. The government would save either by paying down debt or by creating a centralized stock-and-bond fund.

The government would choose diversified investments and distribute the proceeds to the elderly. Transfers would be made from high-wage workers to low-wage workers within each generation. This approach would have very low administrative costs, but it risks increased political interference in capital allocation—a prospect Alan Greenspan and others view with great trepidation.[12] There’s also the risk that government might spend the money instead of investing it.

Regulated Individual Saving (“Privatization”)
The other approach is a partial privatization in which individuals invest some of their payroll-tax dollars in mandatory IRA-like accounts. This approach would have higher administrative costs, but it would give individuals more control over their retirement funds. It would also deter government from simply spending the money. Of course, government would still maintain a safety net for the elderly poor. All workers would be required to save, would be limited to diversified investments and would be required to withdraw the money only gradually during retirement. Transfers would be made from high-wage workers to low-wage workers within each generation. Due to this governmental role, this option would not be a complete privatization, but would provide some of the benefits of unregulated private saving while avoiding some of its pitfalls.

Transition Cost

To summarize, then, a strong argument can be made for reducing transfer payments from the young to the elderly rather than compelling the young to pay ever-higher taxes in perpetuity. Doing so would allow future generations to earn a much higher rate of return than they can at present, without undermining social protections.[13]

But there is an elephant in the room: the benefits owed to current retirees.

Simply put, current retirees have been promised benefits for which they did not save. A severe reduction in benefits would inflict a catastrophic transition cost on those retirees, who are depending on others to fund their retirement. And indeed, even the most ardent advocates of reform would leave those in or near retirement largely untouched. For this reason, reforms would likely target current workers rather than current retirees. Those workers would then bear the transition cost, making full transfers to their parents while working but receiving reduced transfers from their children upon retirement.

Of course, the cuts could be delayed by another generation or even another. But eventually, some generation has to bear the transition cost if the system is to be reformed. That generation pays full benefits to its parents but does not receive full benefits from its children. In effect, that generation pays twice—funding its parents' retirement while saving for its own.

As things stand today, future generations are slated to bear a heavy burden indeed. Figure 6 shows the lifetime net tax rate faced by current and future generations. The lifetime net tax rate is the present value of federal, state and local taxes minus the present value of federal, state, and local transfer payments (including Social Security and Medicare), divided by the present value of labor income. While current generations face lifetime net tax rates between 25 and 32 percent, as can be seen from the bars on the left, future generations (those born after 1995) face a lifetime net tax rate of almost 50 percent.[14] That’s high by almost any standard. Unfortunately, we can reduce their burden only by shouldering some of the burden ourselves.

Figure 6

No Free Lunch

It’s important to understand that there is no free lunch. A formal mathematical analysis reveals that the transition cost imposed on current generations must equal in present discounted value (when discounted at the pretax marginal product of capital) the gains enjoyed by subsequent generations.[15] In layman’s terms, someone must pay, and the only question is who that "someone" will be. The following discussion explains why various proposals for avoiding this burden fail to do so.

No Free Lunch from General Government Revenue
Some reform plans call for the use of general government revenue during the transition. Under this approach, benefits are reduced one generation before a reduction in payroll taxes, with general revenues covering the financing shortfall. For example, today’s workers might receive a reduction in payroll taxes, while today’s retirees still receive full benefits (financed from general revenue rather than from payroll taxes). Benefit reductions would be deferred until today’s workers retire.

At first glance, this might seem to avoid saddling any generation with a transition cost. Today’s retirees would be protected. Although today’s workers would receive lower benefits when they retire, that burden would be offset by the lower payroll taxes they would pay while working.

But the transition cost is still present. The general government revenue used to pay benefits to today’s retirees would not appear from nowhere. Like all government revenue, it would come from the American people. One or more generations would have to bear tax increases or spending cuts to provide the general revenue and they would thereby pay the transition cost.[16] The size of the transfers between young and elderly is what matters, not whether they are financed with payroll taxes or general government revenue.

No Free Lunch from Debt Issuance
While the above discussion assumes that general revenue would be obtained from current taxes or spending cuts, some plans call for the general revenue to be obtained through borrowing rather than taxes. Debt issuance offers no free lunch, however, because the debt must either be serviced or retired. If the debt is retired, the generations that retire it bear the transition cost through higher taxes or lower spending. If the debt is serviced, future generations bear the burden of servicing it, which turns out to match the burden they would have borne from continuing the pay-as-you-go system. And there would be no increase in national saving because the extra debt would offset the saving increase that would otherwise occur.

No Free Lunch from Privatization
Some people think the transition cost can be avoided through mandatory individual accounts. It can’t be, though, because the money has to come from somewhere. If the money going into the accounts would otherwise have been transferred to the elderly, then national saving increases and future generations gain—but those generations that receive the smaller transfers after having paid the larger transfers bear the transition cost. If the money going into the accounts is obtained by issuing government debt, then (as explained above) servicing that debt causes the hoped-for gains from the accounts to evaporate. By themselves, private accounts do nothing to increase national saving or increase rates of return—those effects occur only if transfers from young to elderly are reduced.[17] The form of ‘privatization’ under discussion here offers a way to maintain social protections while minimizing government control of the economy—it does not offer a way to avoid the transition cost.

Inescapable Reality
The inescapable reality is that the pay-as-you-go system has promised benefits without accumulating assets to pay them. Someone must pay—the only question is who. If the system is maintained in its present form, every future generation must bear below-market returns to service this liability—just as you would have to do if you maxed out a credit card and decided to make minimum monthly payments from now to eternity. If the transfers from young to elderly are scaled back, on the other hand, current generations must bear a large transition cost as the burden is repaid—just as you would do if you paid off the balance on your maxed-out credit card.

While we might wish it were possible to pay current benefits in perpetuity without raising taxes, it is impossible to do so. This is the reality that must be faced.

Conclusion

“Why should I care about posterity?” asked famed comedian Groucho Marx—“What’s posterity ever done for me?” While obviously meant in jest, Groucho captured the essence of the tough choice facing policymakers today. Current generations can either sacrifice or not for the sake of future generations. Time will tell which choice we make.

The only certainty is that there is no free lunch.

Notes See Alan Greenspan’s Sept. 8, 2004 testimony to the Senate Budget Committee, at www.federalreserve.gov/boarddocs/testimony/2004/200409082. These numbers are taken from the study’s intermediate spending trajectory, which assumes that Medicare and Medicaid spending per beneficiary will, in the long run, grow 1 percentage point per year faster than per-capita GDP. The federal portion of Medicaid grows from 1.5 percent to 3.3 percent, while all other non-interest programs shrink from 9.6 percent to 5.7 percent. The trustees, like the Congressional Budget Office, assume that spending per beneficiary eventually grows 1 percentage point per year faster than per-capita GDP. From 1970 to 2003, the actual rate of “excess” spending growth for Medicare was 3.0 percentage points per year, Congressional Budeget Office, The Long-Term Budget Outlook, Dec. 2003, p. 5. Also see Henry Aaron, “The Truth About Social Security and Medicare”, Challenge, 47(3), May/June 2004, pp. 27–41, at p. 36. 2004 Social Security Trustees Report, p. 165. The President’s Commission to Strengthen Social Security, Strengthening Social Security and Creating Personal Wealth for all Americans, Dec. 2001, p. 15; 2004 Economic Report of the President, pp. 142–43; 2002 Economic Report of the President, p. 90. Michael Tanner, The 6.2 Percent Solution: A Plan for Reforming Social Security, Cato Project on Social Security Choice Paper 32, Feb. 17, 2004, p. 7; Wall Street Journal, “Social Security Showdown,” Oct. 26, 2000, p. A26; testimony of Robert L. Bixby, Concord Coalition executive director, to Senate Finance Committee, Oct. 3, 2002, www.concordcoalition.org/socialsecurity/021003senfintestimony.htm. Laurence J. Kotlikoff and Scott Burns, The Coming Generational Storm: What You Need to Know About America’s Economic Future, p. 169, propose linking Medicare benefits to wages. The fact that the long-run rate of return equals the long-run growth of labor income was originally noted by Paul A. Samuelson, “An Exact Consumption-Loan Model of Interest with or without the Social Contrivance of Money,” Journal of Political Economy, 66(6), December 1958 and Henry Aaron, “The Social Insurance Paradox,” Canadian Journal of Economics and Political Science, August 1966, pp 371–74. Employee compensation plus proprietors’ income grew from $65.3 billion in 1929 to $7,123.1 billion in 2003 in nominal terms, or from $563 billion to $6,751 billion in 2000 dollars (using the PCE deflator). The real growth rate is 3.41 percent per year, compounded annually. Estimates of the pretax marginal product of capital, which cluster around 6 percent, are collected by Alan D. Viard, "Pay-As-You-Go Social Security and the Aging of America: An Economic Analysis," Federal Reserve Bank of Dallas Economic and Financial Policy Review, Vol. 1, No. 4, 2002, www.dallasfed.org/research/efpr/pdfs/v01_n04_a01.pdf, p. 4. Alan Greenspan, “Statement Before the Committee on the Budget, U.S. Senate, January 28, 1999,” Federal Reserve Bulletin, 85(3), March 1999, pp. 190–92. The social protections impose some efficiency cost on the system. That cost would still be present with reform, assuming that social protections are maintained. The below-market rates of return are a separate phenomenon, arising from the fact that each generation pays for its parents’ retirement, and have nothing to do with any social protections provided by the system. The computations do not literally refer to current law, which is unsustainable, but rather to a particular method for restoring sustainability. Specifically, the computations assume that the government imposes a uniform lifetime net tax rate on all generations born after 1995 that is sufficient to close the fiscal gap. For a thorough discussion, see John Geanakoplos, Olivia S. Mitchell, and Stephen P. Zeldes, “Would a Privatized System Really Pay a Higher Rate of Return?” in Framing the Social Security Debate, ed. R. Douglas Arnold, Michael J. Graetz, and Alicia H. Munnell, pp. 137–57. Also see Congressional Budget Office, How Pension Financing Affects Returns to Different Generations, Long-Range Fiscal Policy Brief No. 12, Sept. 22, 2004, p. 4. Although general revenue does not erase the transition cost, it may be a desirable source of finance. For example, Larry Kotlikoff and Scott Burns, Coming Generational Storm, pp. 156–58, argue that a national sales tax would be a good way to finance the transition cost. The fact that privatization does not offer a free lunch has been noted by many observers, including Alan Greenspan, “Statement Before the Task Force on Social Security, Committee on the Budget, U.S. Senate, November 20, 1997,” Federal Reserve Bulletin, 84(1), January 1998, pp. 32–35. For a thorough analysis, see Geanakoplos, Mitchell, and Zeldes, supra note 17. These sources also explain that issuing debt to buy equity offers little or no real economic gains, even when equity has higher expected returns than debt. About In Depth This article is based on a presentation by Alan D. Viard, senior economist and research officer, and Jason Saving, senior economist, in the Research Department of the Federal Reserve Bank of Dallas. The views expressed are those of the authors and do not necessarily reflect the positions of the Federal Reserve Bank of Dallas or the Federal Reserve System.