Peter G. Peterson’s two-part article on Social Security was published in the December 2 and December 16 issues of The New York Review. We here publish two critical comments on Mr. Peterson’s article along with his reply.—The Editors
A Calmer Look at Social Security
Peter Peterson’s obvious good intentions and apparent lack of a “vociferous constituency”1 have lent his two-part article on Social Security an aura of accuracy and intelligence that it does not deserve. Peterson has overstated the system’s financial problems, has found a link between the expansion of Social Security and the decline in the nation’s productivity that cannot be substantiated, and has used this purported link as a basis for advocating draconian benefit cuts. His analysis and recommendations must not go unchallenged.
1. The Magnitude of the Problem: No Crash for Social Security
The first step in understanding the Social Security financing situation is to separate the problems of the Old Age, Survivors, and Disability Insurance (OASDI) program from those of the Hospital Insurance (HI) system. The future financing requirements of HI are extremely uncertain, since rapidly escalating hospital costs have caused both public and private health insurance programs to become increasingly expensive. Restoring long-run balance to the HI system will undoubtedly require fundamental reform in the way we provide hospital care. For this reason, the National Commission on Social Security Reform and others have focused their attention on the OASDI portion of the program. The answer is not, as Peterson suggests, to accept past rates of increase for HI expenditures and then cut the OASDI program by an amount equal to four times its own deficit in the year 2000 in order to transfer those funds to the Hospital Insurance program. The next step, which is useful for sorting out the financing problems in the OASDI program, is to divide the future into three separate time periods—1983-1989, 1990-2014, 2015-2060.
1983-1989. Between now and 1989, the OASDI program is projected under the Trustees’ most pessimistic economic assumptions to run a deficit of about $200 billion (see Table 1 on page 42).
If the economy performs somewhat better, the shortfall may be close to $75 billion. In any event, when the interfund borrowing authority expires in July 1983, the OASI trust fund, the largest of the Social Security trust funds, will be unable to pay all benefits on time. Even if interfund borrowing were extended, all three funds, OASI, DI, and HI, together will be exhausted by mid-1984. The immediacy of the projected short-fall has caused many to characterize the Social Security program’s short-term problems as catastrophic and the press constantly refers to the impending “bankruptcy” of the system. In fact, the magnitude of the deficits forecast for the next seven years is relatively manageable, roughly 4 to 10 percent of annual outlays, and numerous options are available for restoring solvency. More importantly, the reasons for the current deficits are well understood and future problems of this type can be avoided by modifying the indexing procedure.
The most appropriate point from which to trace the origins of the current financing problems is 1977, since legislation passed in that year dramatically revised the Social Security financing and benefit provisions to restore fiscal balance to the program. However, payroll tax rates were established on the traditional assumption that the rate of growth of taxable wages would equal the rate of increase in prices plus an additional amount for productivity growth. This was a perfectly reasonable assumption, since it reflected the performance of the US economy over the entire postwar period. After 1977, however, the traditional relationship between prices and wages reversed and price increases exceeded wage growth. The projected balance in the trust funds is extremely sensitive to the relationship between these economic variables. The rate of wage growth determines the rate at which revenues grow, while the rate of increase in prices determines the rate at which benefit expenditures increase since benefits are indexed to the consumer price index. Moreover, the rapid inflation was accompanied by high unemployment which further worsened the financial outlook, since fewer people contribute revenue to the trust funds and more people, finding themselves unemployed, are likely to take early retirement.
Since the current financing problems can be traced to past experience with overly optimistic assumptions, the obvious question is whether the $75 to $200 billion deficit projected for the next seven years is realistic. Table 2 (on page 42) compares forecasts of productivity growth and the unemployment rate underlying the intermediate (II-B) and pessimistic assumptions (III) from the 1982 Trustees’ Report and two independent forecasters, Chase Econometrics and Data Resources, Inc.
Generally, the projections of the private forecasters fall somewhere between the intermediate and pessimistic assumptions, although considerably nearer the former. Hence, planning on a deficit somewhere between $75 billion and $200 billion for the next seven years seems like a very reasonable strategy.
1990-2014. In marked contrast to the next seven years, the outlook for OASDI financing is relatively favorable for the period 1990-2014. The primary reason is demographic. The low fertility rates during the late 1920s and the 1930s will be reflected in a considerable reduction in the rate of increase in the population over age sixty-five during the 1990s. As a result, the ratio of workers to beneficiaries, which has declined continually since 1940, is estimated to remain stable for the next twenty to thirty years at its current level of roughly three to one. With a stable ratio of workers to beneficiaries, even modest productivity gains will reduce the cost of Social Security as a percent of payroll.
If an upsurge in productivity occurs and wages rise by 1.5 percent more than prices, as assumed under the intermediate (II-B) economic assumptions, then revenues will exceed outlays over the entire period and the trust funds will accumulate surpluses rapidly, reaching 177 percent of the annual outgo by 2010. On the other hand, if the realwage differential is closer to 1 percent, as under the pessimistic assumptions, then outlays will slightly exceed revenues—an average of 12.9 percent versus 12.4 percent of taxable payrolls—and no balances will accumulate.
The future costs of the system and the total of trust-fund balances could be made considerably more predictable by revising the procedure for indexing benefits. As noted earlier, the OASDI program is thrown into deficit when unanticipated inflation and low wage growth cause outlays to increase more rapidly than revenues. This type of instability can be avoided only by linking post-retirement cost-of-living adjustments directly to the growth of wages. For example, indexing retirement and disability benefits by the rate of wage growth minus 1.5 percent would provide the same benefits as the current price indexing of benefits under the intermediate (II-B) assumptions. This reform would not only eliminate the short-run instability, but would also lock in the economic assumptions to insure the buildup of reserves during the 1990-2014 period.
2015-2060. The third period is characterized by rapidly rising costs as the baby-boom generation starts to retire. At the same time, the growth in the labor force slows markedly, reflecting the precipitous decline in the fertility rate which began in the mid-1960s. These two factors cause the ratio of beneficiaries to workers to increase dramatically. Assuming that the fertility rate will rise gradually from the current level of 1.8 to a long-run rate of 2.1 (the intermediate assumption), the Social Security Administration projects that the number of beneficiaries per 100 covered workers will rise from 31 in 1982 to 50 by 2035. If the fertility rate declines to 1.7 (the pessimistic assumption), then the number of beneficiaries per 100 workers will increase to 67 by 2035.
With a pay-as-you-go system the increase in this crucial ratio produces a substantial increase in costs as a percentage of payrolls. The question is, what is a reasonable estimate of the magnitude of the long-run problem? Is Peterson’s assertion that “to close the deficits in the Social Security system under the ‘pessimistic’ projection would take a payroll-tax rate of above 44 percent in 2035” even remotely realistic?2
Let us assume for the time being that the pessimistic demographic assumptions are borne out. Under these assumptions, the projected tax rate for the OASDI portion of the program for the year 2035 is 24 percent of payrolls. In order to bring the total to 44 percent, HI outlays must rise to 20 percent of taxable payrolls. Outlays for HI today account for only 18 percent of total expenditures under the Social Security program; it is difficult to believe that we will allow the HI program to grow to a point where the cost for hospital insurance (20 percent of taxable payrolls) roughly equals the total cost to support the aged, disabled, their dependents and survivors (24 percent of payrolls).
Moreover, the pessimistic demographic assumptions are not consistent with other forecasts. The pessimistic projections are based on the assumption that fertility rates will decline from the 1980 level of 1.83 to an ultimate long-run rate of 1.7 by 2005. The intermediate assumption is that fertility will increase gradually to a long-run rate of 2.1. Of the two assumptions, the evidence tends to support the higher. First, until October of this year, the Census Bureau’s “middle” series assumed a long-run fertility rate of 2.1. Although the Census now assumes that fertility rates will remain fairly steady, increasing slightly from 1.83 to 1.96 in 2000 and then decreasing to 1.90 births per woman in 2050, the fertility rate has increased over the last five years and the data on expected births indicate that young women continue to expect to have more than two children over their lifetimes. Thus, even with the downward revisions of the projections, current Census data appear more consistent with the intermediate than with the pessimistic assumptions. In addition, most observers, including Peterson, acknowledge that the conventional assumption for Social Security projections of net immigration of 400,000 persons a year may substantially understate the number of people entering the country each year.3 When illegal as well as legal immigration is considered, the working population is likely to be substantially larger than the fertility assumptions alone would indicate.
If the intermediate demographic assumptions are the more realistic over the long run and the system is stabilized by linking post-retirement indexing directly to wage growth, then the projected cost of the OASDI portion of the program is 15 percent for the year 2035, and current estimates of the costs for the HI program for the same year run around 11 percent. Thus, the cost for the entire Social Security program, assuming no major reform of Medicare, would be about 26 percent, that is, 13 percent for the employer and 13 percent for the employee. This compares to current rates of 6.75 each for the employer and employee.
It is important to note, however, that the 26 percent tax rate would be levied on a much smaller portion of the worker’s total compensation than is taxable today. According to the Trustees’ projections, the ratio of cash wages to total compensation is estimated to decline from 84.2 percent in 1980 to 67.4 percent by the year 2035.4 Since the payroll tax is levied only on cash wages, the expansion of fringe benefits reduces the tax base and boosts the percentage of taxable payroll required for paying benefits. If fringe benefits remain a constant percentage of total compensation, then the required tax rate for the OASDI portion of the program in the year 2035 would be 12 percent and for HI another 9 percent. In other words, in terms of the tax base we have today, the total required tax for OASDI and HI in the year 2035 would be 21 percent, or 10.5 percent each for the employee and the employer.
Moreover, the increased cost of Social Security after the turn of the century must be put into perspective. First, higher taxes do not mean that the Social Security program will be any more generous in the future than it is today, but rather reflect the inescapable burden of a very large dependent population in the twenty-first century. If the elderly are not supported through Social Security, the working population will probably end up providing equivalent support through some other program, since most people fail to save adequately on their own for retirement and many do not have private pension coverage.
Second, those concerned about higher Social Security costs often ignore the fact that long-term projections of lower fertility will result in fewer children per worker. If the economic burden on active workers is measured in terms of total dependents rather than just aged retirees, then the picture looks quite different. The total dependency ratio (the ratio of the number of people under age twenty and over age sixty-five per 100 people age twenty to sixty-four) will be lower in the year 2035 than it was in the 1960s.5 The rise in the aged will be more than offset by a decline in dependent children, thereby freeing resources which could be devoted to providing for the elderly.
Finally, while a projected tax rate of 15 percent for OASDI and 26 percent for the entire program for 2035 represents a substantial increase over the current levy, it is considerably below the present payroll tax rates in many European countries. Austria, West Germany, Italy, Sweden, and the Netherlands already have rates for programs comparable to OASDI in excess of 18 percent of payroll and total payroll taxes far in excess of any projected rate for this country.6
In summary, Social Security’s longterm deficits, like its short-term financing problems, are manageable. Costs are scheduled to increase, but Peterson’s assertion that payroll taxes will take 44 percent of the average worker’s wages is nonsense. A more realistic assessment reveals that the costs for OASDI will increase from the current level of 10.8 to 15.0 by 2035. Even assuming that HI taxes are allowed to increase from their current level of 2.6 to 11.0 percent, the overall required tax rate in the year 2035 would be 26 percent, 13 percent each for employees and employers. If fringe benefits did not continue to erode the tax base, the required rate would fall to 21 percent.
2. Eliminating the Deficits: Drastic Cuts Not “Salvation”
After his cataclysmic discussion of the magnitude of the problem, Peterson sets the stage for advocating extensive benefit reductions by 1) attributing our low productivity growth to Social Security’s financing problems, 2) dismissing out-of-hand the possibility of raising payroll taxes, and 3) dispelling the “myths” that have traditionally protected Social Security benefits from abrupt cuts. As we shall see, these premises provide a very weak foundation for Peterson’s ultimate recommendations.
Social Security and Productivity. Peterson argues that the rise in Social Security outlays has contributed to the large government deficits which, in turn, have absorbed savings that could have been used to finance plant and equipment. Greater investment would have increased the amount of capital per worker and a higher capital/labor ratio would have led to greater productivity.
While this argument sounds plausible, it is factually incorrect, since it presumes that 1) the Social Security program has been a substantial contributor to the federal deficit, 2) the slump in investment is the result of savings being diverted from the private sector to finance government deficits, and 3) the declining ratio of capital to labor is solely attributable to inadequate investment. None of these presumptions is correct.
Social Security revenues have exceeded outlays on balance between 1965 and 1982, the period of declining productivity in the US. Thus, over this period Social Security has not increased federal deficits but rather has been a net contributor to the government coffers. Even during the more recent period of 1970 to 1982, revenues have exceeded outlays by more than $2 billion. Between 1975 and 1982, the period of greatest fiscal pressure, Social Security outlays did exceed revenues, but Social Security deficits totaled only $20 billion, less than $3 billion per year, as compared to total federal deficits of $500 billion, or more than $60 billion per year, for the same period. Hence, to date, Social Security has not been an important factor in the deficits of the federal government.
More importantly, federal deficits have not been the cause of the decline in investment during the 1970s. Studies have shown that much of the slump in investment is directly attributable to the impact of repeated recessions on capacity utilization and corporate profitability. The persistence of excess capacity substantially undermines the incentive to build more plant and equipment, while the decline in corporate profitability reduces the expected return on new investment. The situation is worsened by the interaction of inflation and an unindexed corporate tax structure that leads to rising real tax burdens and even lower after-tax returns on investment.7
Finally, even if investment had proceeded at a steady pace, the capital/labor ratio would have declined during the 1970s because of the enormous growth in the labor force. As the baby boom generation matured, women and teenagers joined the work force in unprecedented numbers. Between 1970 and 1982 the civilian labor force increased at an annual rate of 2.5 percent, compared to 1.8 percent during the 1960s.
In short, Social Security has not caused federal deficits, deficits are not the reason for low investment, low investment is not the only reason for the decline in the capital/labor ratio and the resulting low productivity. No economic link exists between the substantial growth in the Social Security program over the past fifteen years and the decline in the rate of productivity growth.
Raising Payroll Taxes. While Social Security has not been a major source of federal deficits in the past, substantial shortfalls are projected in the next few years. As discussed earlier, the gap between revenues and outlays for OASDI between now and 1989 may be as much as $200 billion. To avoid deficits of this magnitude, either outlays must be reduced or revenues increased. Although many experts have suggested moving forward OASDI payroll-tax increases scheduled for 1985 and 1990 to January 1984, Peterson dismisses this option as totally unacceptable. He provides no justification for this conclusion, but the implication is that all the projected deficits must be eliminated by reducing benefits.
Exploding the Myths. In order to provide some justification for the disproportionate reliance on benefit reductions as opposed to tax increases, Peterson attempts to dispel several “myths” in order to establish that individuals should not feel entitled to their Social Security benefits and that the elderly do not need the current level of benefits provided by Social Security.
Myth I—It is “my money”: According to Peterson, the fact that people feel they have paid for their benefits is the “most damaging myth of all.”8 He documents the well-known fact that current retirees get back much more than they have ever contributed. He also reiterates the view that a system that allows retirees to receive benefits out of proportion to lifetime payroll-tax contributions plus interest “will remain fundamentally out of balance.”9 These two issues will be addressed in reverse order.
Estimates show that the average worker retiring today will receive benefits equal to more than twice the combined employer-employee payroll-tax contributions made on his behalf plus accumulated interest.10 These high ratios of benefits to contributions, however, do not imply that the system is fundamentally unsound. Rather, the current high ratio of benefits to contributions is the inevitable result of the start-up phase of a pay-as-you-go system. Although the original Social Security legislation in 1935 provided for the creation of a substantial trust fund, only a few years later, in the critical 1939 amendments to the original legislation, Congress decided to pay benefits in excess of contributions to the entire first generation of retirees. This was, in fact, a decision not to build up the trust fund that would have accumulated if current workers’ contributions had been stockpiled for their retirement.
However, it was felt at the time of the Great Depression that the needs of current retirees warranted the shift to a pay-as-you-go system. For the next few decades, benefits could be financed with modest payroll taxes because coverage continued to expand and the number of retirees were relatively few compared to the number of workers. Therefore, people retiring now and in the near future, who worked during the years when payroll taxes were low, will receive benefits considerably in excess of contributions. As the system matures, however, and the number of beneficiaries stabilizes relative to the number of workers, the average retiree will receive benefits which are roughly equal in present value terms to combined employee and employer contributions. This ratio will then tend to remain constant for all subsequent generations as long as population growth and real-wage increases remain relatively stable.11 In short, the current high ratio of benefits to contributions and the projected decline in this ratio over the next seventy-five years are to be expected and do not reflect any fundamental flaw in the system’s financial structure.
The other question, however, is whether the fact that current retirees receive benefits far in excess of contributions can be used to justify benefits cuts. It seems difficult to argue that individuals close to retirement who have planned on a particular level of income from Social Security should have their benefits reduced precipitously. These people have neither the time nor the resources to adjust their plans. Our essential collective commitment to the Social Security program will be undermined if benefit provisions can be changed abruptly.
The longer the lead time, the more reasonable it becomes to legislate benefit reductions. However, as we move into the next decade we find the ratio of benefits to contributions rapidly approaching unity. Thus, for the period for which benefit reductions are a reasonable option to consider, the “windfall” component disappears.
In short, contrary to Peterson’s arguments, the excess of benefits over contributions for current retirees neither signals a fundamental flaw with the financing nor provides an acceptable rationale for reducing benefits.
Myth 2—The elderly are by definition needy: Peterson quotes figures from a survey that showed that only 17 percent of the elderly regarded low income as a problem for them personally and 58 percent of the elderly thought that it was hardly a problem at all.12 Moreover, Peterson notes that the percentage of elderly below the poverty level declined from 35.2 to 15.7 in the last twenty years. He implies that Social Security benefits are superfluous not only because the elderly as a group are not poor but also because a high proportion of the benefits go to people above the poverty line.
It is true that the economic status of the aged has improved dramatically in the last twenty years. The major reason is that Social Security is our most successful antipoverty program. Approximately 90 percent of persons aged sixty-five and over are Social Security recipients, and for two-thirds of these recipients, Social Security accounts for more than half of total income.13 With such an enormous dependence on Social Security, any significant reduction in the program could adversely affect the economic well-being of a large portion of the elderly and reverse the gains that have been made over the last two decades.
If Social Security were simply a welfare program, then the percentage of funds channeled to the poor would be an appropriate criterion on which to evaluate the system. But this is not the case. Social Security is a national insurance program that protects all insured workers and their dependents and survivors against the loss of earnings that results from disability, retirement, or death. Since social insurance benefits do not require a proof of need, anyone who has contributed to the program for the required period of time is eligible for benefits. As a result, people feel no stigma attached to these payments, in marked contrast to their attitude toward means-tested welfare benefits. Survey after survey has shown that the American people approve of a government retirement program that provides benefits as a matter of earned right. Since Social Security is not designed as a welfare system, the fact that a significant proportion of benefits goes to individuals whose income is above the poverty line does not imply a weakness in the program.
Moreover, disproportionate cuts in the benefits for higher-paid workers would endanger public support for the program. Social Security’s progressive benefit formula already produces proportionally greater benefits for lower-paid workers than for higher-paid employees. Further reductions in the benefits for those with above-average earnings would mean that as the system matures, these people would receive benefits that are less than combined employee and employer contributions plus interest. If this were to happen, support for the Social Security program would decline.
Thus, Peterson fails to construct a convincing argument for either general benefit reduction or for benefit cuts for high-income workers.
Myth 3—The elderly are physically unable to work beyond age sixty-five: No one, to my knowledge, has taken this position. Certainly, it is difficult in the current environment with the great majority of people retiring well before age sixty-five and more than 10 percent of the labor force without jobs, to envision delaying retirement and keeping people in the work force until age sixty-eight. In the next decade, however, as life expectancy continues to increase, the rate of growth of the labor force slows, and the industrial structure shifts away from manufacturing toward the service industries, it may be possible and perhaps even desirable to encourage people to work longer.
In designing provisions to encourage later retirement, however, it is essential to remember that some older workers will not be able to engage in gainful employment past age sixty-two and must have access to some form of income support. If the decision is made to postpone the age at which retirement benefits are first available, then people who are prevented from working by physical disability will need access to an expanded disability insurance program. While current law makes some allowance for age in determining disability by applying a more liberal test to those aged fifty or older, more explicit recognition of the interaction of age and physical impairment may be required.
Summary. In considering alternative approaches to restoring fiscal solvency to the Social Security program, Peterson summarily rejects the option of raising taxes. He then attempts to establish the case for benefit reductions by arguing that individuals retiring now receive benefits far in excess of contributions, that the elderly as a group are not poor, and that substantial benefits go to upper-income people. It has been demonstrated that these arguments are unsound and do not provide support for the lopsided approach to deficit reduction that Peterson advocates.
3. A Balanced Solution to Social Security Financing
Three principles should guide reform of Social Security financing. First, the burden should be shared between workers and retirees by combining some tax increases with some benefit reductions. Second, tax increases and benefit cuts should be considered separately for the OASDI and HI portions of the program. Third, the instability in the system must be reduced in order to avoid repeated short-run financing crises. A reform package that incorporates these principles is presented in Table 3.
The most notable feature is that this package does not involve any drastic cuts or major restructuring of the program. The proposals are also very similar to those recommended by the majority of the bipartisan Commission on Social Security Reform.
To cover the projected shortfall of roughly $200 billion between now and 1989, the proposal increases revenues for the OASDI program by moving forward the scheduled 1985 and 1990 tax increases to January 1984. Additional revenues are also generated by extending coverage to employees of nonprofit institutions, new federal employees, and new employees of state and local governments not currently participating in Social Security. Such a reform is clearly desirable even if it did not have positive short-run revenue implications, since under the present system those workers who are entitled to civil service or state or local pensions can easily achieve insured status under Social Security and receive relatively large Social Security benefits in addition to their regular pension. These dual beneficiaries profit from the progressive benefit structure, which was designed to help low-wage workers rather than workers whose second career entitles them to benefits.
Finally, revenue could also be gained and the equity of the system improved by raising the tax on the self-employed to a level equal to the combined tax for employees and employers. Of course, to fully equalize the treatment of self-employed and salaried workers, the self-employed should be allowed a tax deduction under the personal income tax for the part of their payroll-tax contribution that corresponds to the employer’s share.
Abrupt and significant benefit cuts are neither desirable nor feasible in the short run, but the two selective changes included in the proposed package could substantially reduce outlays over the next seven years. The most straightforward proposal is to shift the date for the cost-of-living adjustment from July to October. This change would not only produce a reduction in outlays, but would also align the adjustments with the beginning of the new fiscal year.
Another form of benefit reduction, which would be desirable even if additional revenues were not needed, is the proposal to include 50 percent of Social Security benefits in taxable income for those individuals who have incomes in excess of $12,000 and couples with incomes in excess of $18,000. Such a reform would move toward equalizing the tax treatment of Social Security and private pension benefits, which under current law are taxed in full to the extent that they exceed the employee’s own contributions. To assure that no low-income individuals are adversely affected, the taxation of benefits can be phased in gradually by adopting the provisions currently applicable to unemployment insurance benefits, whereby only those single people and married couples with incomes over stated limits are required to include benefits in taxable income.
The combination of taxing benefits in this manner, shifting the indexing one quarter, extending coverage, raising the tax rate on self-employed, and accelerating the scheduled tax increases will generate more than enough revenues to cover deficits under even the Trustees’ pessimistic economic assumptions. These are not onerous changes, and most are desirable reforms in and of themselves.
As noted earlier, at least by 1990 a fundamental change must be made in the procedure for indexing benefits in order to eliminate the instability in the program. The repeated short-run crises can be avoided by linking the post-retirement cost-of-living adjustment directly to the growth in wages. Indexing retirement benefits by the rate of wage growth minus 1.5 percent would duplicate the benefits scheduled under the II-B assumptions. This reform would not only provide short-run stability, but would also ensure that long-run costs do not exceed those projected under intermediate (II-B) assumptions.
Even with the revised indexing procedure and the revenue increases and benefit reductions proposed for the 1983-1989 period, further changes will be required to completely eliminate the long-term deficit for OASDI. A logical and equitable approach would be to divide the burden equally by both lowering replacement rates and increasing tax rates. In all likelihood, replacement rates would probably decline by less than 5 percent and taxes would have to be raised by less than 0.5 percent each for employees and employers.
In short, Mr. Peterson has done a great disservice to the public by exaggerating the Social Security system’s financial problems, distorting the facts about the impact of Social Security on the economy, and misrepresenting the issues relating to the justification for the program. His analysis and his proposals to drastically cut Social Security retirement benefits by an amount equal to four times the OASDI deficit in the year 2000 should be rejected.
We Can Afford to Support the Elderly
The Social Security system, we are told, is a shambles. Peter Peterson’s two articles give us a detailed account of how he feels Social Security threatens our economy, and how he feels “reform” should be structured. The current wisdom about Social Security is riddled with fallacies—most of which appear in Peterson’s articles.
As background, the Social Security system is rather well designed to provide income with dignity to retired Americans. Its current problems—and there are some—stem from the shifting character of the US economy. The Ponzitype financing (in which growing numbers of new participants pay in, and old participants draw out more than they paid in during their participation) worked well when population was growing, the economy thrived, and real wages and productivity rose steadily.
Now, none of these conditions hold, and restructuring (not reform! which implies corruption) of the financing of benefits is clearly necessary. For optimists, the difficulty is temporary, the economy will eventually get back on the track, and financing must be rearranged accordingly. For pessimists, the difficulty is permanent—indeed, Peterson and others suggest that the growth of Social Security payments is a cause of the economic slowdown. Thus benefits to the elderly must be scaled back on a permanent basis, and people should be encouraged to save their own money for retirement.
The truth is more complicated, as usual. The slowdown in the US and the world economic growth will probably prove to be long lasting. But part of the cure involves maintaining or even increasing the income of the elderly and of others likely to spend it, for our problems are related to the maturing and stagnation of markets.14
Essentially, it boils down to a political and social decision which we Americans have to make. Can we afford, and do we choose, to keep our elderly in (some kind of) style? It is astonishing that Peterson solemnly warns us of one of our “highest moral obligations—to give our offspring a decent chance at prosperity.” This high moral obligation is to be met by cutting the real income of the elderly! What kind of moral conclusions will our children draw?
But Peterson thinks we have an either/or choice. The entire economy, he feels, will collapse under the strain of growing Social Security payments. This conclusion stems from a flawed tenet prevalent in much of today’s economic discussion: that consumption is bad for the economy. But consumption is good, not bad. Consumption does not take away from investment. It is the lack of promising markets that slows down investment, and so productivity and overall economic growth. With already-existing factories and offices, with current labor force, we can produce far more than we can sell. US Steel and the auto companies are understandably reluctant to build new and modernized plants when they can only sell half what they can currently produce. We do not have a problem of production but of distribution.
So how do we solve this problem? It is a thorny and worldwide problem, but there are a few things we should clearly not do. The worst mistake is to destroy existing markets—bad enough that their growth has slowed. At current income distribution, the demand for many products is already filled in the US and Europe. The apparent growth trend of the 1970s was demographic, not economic. On a per-household basis, there has been only replacement demand. If we attempt to cut consumption—as many economic “austerity” policies, among them proposals to curtail Social Security payments, do—things will only get worse.
Peterson argues that rather than transfer income from one group to another, as the Social Security system does, we should place more emphasis on pension funds. Each worker should accumulate a fund, individually or as part of a group, to finance retirement. The Social Security system could remain as a kind of welfare arrangement for those too poor to accumulate such funds, which would provide a pool of savings to finance increased investment, boosting economic growth. We could thus devote a greater proportion of GNP to investment—partly by cutting consumption by the elderly. This sounds inhumane, but more relevantly, it does not work. It at first appears surprising that in the US, investment measures a remarkably stable 10 percent of GNP, come slump or boom. But this is not surprising, for the simple and often-overlooked reason that the multiplier works! Attempts to raise investment as a percent of GNP will almost surely come to grief. If we start on the investment side, an increase in spending will raise employment, income, and consumption, keeping investment to its customary 10 percent of GNP. If we start on the savings side, using austerity, this will result in the first instance in a decline in sales of consumer goods. Business managers, with excess capacity and inventories piling up, will cut back on investment, rather than increase it. Investment will still claim 10 percent of output, but of a slump-level GNP.
And giving money to the elderly is a good way to stimulate business, for it fulfills a moral need as well. Peterson deplores the “myth” of the needy elderly. True, not so many of them live in poverty as used to be the case, and their lot has improved faster than that of the younger poor. In 1959, a disgraceful 35 percent of elderly Americans lived in poverty. By 1980, that figure had declined to 15 percent (compared with 13 percent for the overall population). But this is precisely because of Social Security—its broadened eligibility and the indexation of benefits to the CPI. If benefits are frozen, or indexation shifted to less than the full increase of the CPI, living standards of the elderly will erode.
In the first year or so, this erosion would not be severe—but Peterson has a seventy-year planning horizon! Real benefits could shrink to near zero if indexed at less than full CPI increases. Another popular suggestion that Peterson puts forth is that Social Security payments be indexed to wage gains, subtracting average productivity gains. He suggests that this will protect the system’s financing from cyclical shifts in revenues, which are tied to wages. But since wages move with the business cycle, this would add Social Security payments to the factors that aggravate the cycle. Currently, the system provides a small stabilizing force to the economy, which it can ill afford to lose.
It is often said that the elderly have lower expenses than other households. Of course! They have less money—40 percent less, on average. It is true that many of the elderly own their homes. But they pay, as others do, the rising costs of utilities, property taxes, groceries, transportation, clothing, and entertainment. Medicare has shielded many of the old from catastrophic medical expenses. But as others have pointed out, you can’t eat wheelchairs.
Peterson points out that Social Security is a transfer payment, but it is not paid only to the needy. Vast numbers of middle-class retirees—many with private pensions—and even the wealthy receive Social Security checks. Moreover, those with highest lifetime earnings and payroll contributions receive the most, while those with modest means get the smallest checks. The result is what looks like a massive inequity to Peterson: 30 percent of Social Security payments go to the most prosperous 20 percent of retiree households.
But US income distribution is terrifically skewed toward the top end of the scale, and what’s remarkable about Social Security is that only 30 percent of benefits go to the top 20 percent of households! Over 40 percent of total personal income goes to the top 20 percent of families, with 15 percent reserved for the wealthiest 5 percent. Of total Social Security payments, only 2.5 percent is paid out to the wealthiest 5 percent of families. Since a lot of money is involved—about 8 percent of personal income—the Social Security system makes our overall income distribution somewhat less unequal.
There is good reason to make Social Security payments more equally distributed—by bringing the minimum payment up to the average, for example. This would help to alleviate poverty among the elderly and would create demand for basic goods and services, giving some stimulus to the economy.
Benefits to well-off retirees could be reduced or eliminated, but with caution. The beauty of Social Security is its wide acceptance. No one is embarrassed to receive or to spend it. Little old ladies living alone in paid-off houses would (literally) die rather than apply for welfare or SSI. But Social Security payments, to which they feel they are entitled, are not demeaning. Moreover, it is probably necessary to continue to pay benefits to the well-off to retain their acceptance of paying into the system.
But it is quite possible to make the financing of Social Security progressive rather than regressive. In 1982, workers paid 6.7 percent of earnings up to $32,500. Thus workers earning $10,000 or $30,000 paid 6.7 percent of their pay-checks to Social Security, while those earning $70,000 paid 3.1 percent. It would be easy enough to eliminate the ceilings15 and to increase the tax rate at the upper end of the income scale. This would quickly solve the Social Security budget crisis.
But Peterson speaks of “the dead hand of taxation”—if we raise taxes, won’t the incentive to work decline, and the economy founder? Well, the hand of taxation is not very heavy. Overall, taxes are only slightly progressive in the US, and they are not high relative to other developed countries.16 And the hand of taxation is not dead—tax revenues are spent (often on socially desirable goals—in this case, consumption by the elderly) and so generate business, employment, sales, and profits. And making the tax system more progressive will generally increase the overall level of economic activity. For the poor are more likely to spend additional funds, and the well-off are more likely to save. Increased saving by the rich could stimulate investment, but not if businesses are saddled with excess inventories and excess capacity.
But can we afford to support the growing population of elderly? Peterson suggests that Social Security taxes would have to rise to a politically impossible 44 percent of taxable payrolls in the next century to continue current benefit levels, using conservative assumptions: small gains in life expectancy, low birth rates, slow real wage growth. One could tinker with Peterson’s arithmetic—he assumes no increases in the ceilings, for example—but it is more instructive to look at this in two other ways. First, we can produce the goods and services that the elderly might desire. GM, Eastern Airlines, the corner barbershop are only too happy to oblige potential customers. So by providing income to the aged, we raise overall GNP. Do we simultaneously reduce it by taxing younger households? Perhaps, though, if we make the tax progressive, the net effect is probably positive.
Second, look at the overall dependency ratios. Peterson makes much of his dire projection that by the year 2050 there may be eighty retirees for every 100 workers, up from thirty today. The elderly today make up 11 percent of the population, up from 9 percent in 1950. With low birth rates, that proportion will rise to 14 percent by the year 2010, a significant but not overly dire increase. Meanwhile, other dependent groups are shrinking. The number of children as a proportion of the population has plummeted, from over 30 percent during the 1960s to 24 percent today; it is projected to fall to 16 percent by the year 2010, if current low birth rates continue.
So can we afford to support today’s and tomorrow’s elderly? Of course we can! In the 1960s, when our economy was growing fastest, dependency was heaviest—over 40 percent of our population was of nonworking age. 17 Today that proportion is down to 36 percent, and in the decade of 2000-2010 it will be 30 percent.
It is often argued that people are retiring too young. Raising the retirement age gradually, reducing Social Security payments for early retirement, and correcting financial disincentives to working beyond age sixty-five are proposed. Peterson is convinced that these financial disincentives cause most people to retire by sixty-five. But a change in attitudes among employers and co-workers is more important than tinkering with benefit formulas. A humane and civilized society would surely encourage the elderly to work as long as they wished, and would respect and appreciate their contributions. But employers, if they don’t actually give older employees the shaft, are usually less than enthusiastic about keeping them on. Life at the factory or at the office often becomes demoralizing for older workers, and retirement—often early—becomes a sort of honorable discharge, before too many unexpected transfers, canceled projects, and reassigned responsibilities damage the ego irreparably.
Moreover, despite gains in life expectancy, illness and death loom large after sixty. While, on average, sixty-year-olds can expect to live another twenty years (men seventeen, women twenty-two), the proposition is risky. More than one out of every ten sixty-year-old men will die before they reach sixty-five (for women, one out of twenty). Among sixty-five-year-olds, one out of five will die before seventy. Needless to say, those who die, and many of those who don’t, will have to cope with severe and often prolonged illness—their own or their spouse’s—during those years. These are grueling prospects that all of the elderly have to consider. In order for a married couple to assure several years of healthy leisure and recreation together, early or on-time retirement is often advisable.
Two more points, parenthetically. Peterson says it would be intolerable for the federal share of GNP to rise from the current 22 percent to nearly 25 percent, and that a federal deficit of 5.5 percent of GNP would be simply “unworkable.” Government economic activity has been so maligned by economists and by the government itself that it is shocking to say that government activity is actually good for the economy. It helps to correct cyclical instability, shifts income distribution in favor of the needy, promoting market growth, and can use a long-term view of social and economic goals in planning public investment. Moreover, as a practical demonstration, in most advanced industrialized countries, including Japan and Germany, government plays a more active role, in both spending and in regulation, than in the US. Government spending and government deficits are higher as a proportion of GNP, public services are greater and more widely available, social welfare, job security and worker safety and rights are more widely protected. Despite these facts, economic growth is faster, productivity growth is higher, real wages are rising, and profits are as healthy as in the US. Moreover, the quality of life (as measured by such accepted statistical yardsticks as infant mortality and life expectancy, as well as the prevalence of tree-lined boulevards and dazzling urban parks) is higher in most of these countries than in the US. In sum, the size of our government sector is not the cause of our problems.
Finally, let us be a little reasonable in the length of our long-term outlook. While the view in some quarters that anything beyond the next three months is “the long run” is silly, planning for fifty or seventy years out seems a bit extreme. If we were to attempt to solve now all the dire problems that are likely to emerge over the next fifty years, we would be in trouble indeed! Let us concentrate on the next ten to fifteen years, and make the ethical, political, and economic decision to provide a decent standard of living for all the elderly, with full cost-of-living adjustments, and appropriate sufficient funds—by increasing taxes and by making them progressive—to pay for it.
Alicia H. Munnell
Vice-President and Economist, Federal Reserve Bank of Boston.
The views expressed are solely those of the author and do not necessarily reflect official positions of the Federal Bank of Boston or the Federal Reseve system.
Assistant Vice-President and Economist of a major New York bank.
A Reply to Critics
Peter G Peterson replies:
In a “Calmer Look at Social Security,” Alicia Munnell asserts that my two-part article on Social Security has an “aura of accuracy and intelligence that it does not deserve.” She accuses me of overstating the system’s financial problems, incorrectly linking the expansion of Social Security to the decline in America’s productivity growth, and advocating totally unjustified and “draconian” benefit cuts. These charges sound serious.18 In fact, Munnell is able to take what she calls a “calmer look” at Social Security only with the help of some powerful analytical anesthetics.
In the pages that follow, I shall explore those anesthetics and I shall respond as well to the letter from Rosemary Rinder, who takes a different critical position. She believes that the mushrooming Social Security program is desirable for precisely the reasons that I think it undesirable—that it will result in constantly increasing levels of deficit-financed consumption.
First I shall try to show that instead of analyzing the problem, Munnell tends to assume it away—that her apparent tranquillity is achieved at the expense of reality. I shall then try to address what seems to me the fundamental question in this controversy. What can we as a society afford? I should think that an economist of Munnell’s obvious abilities would accept as reality that our resources are limited and that our decisions about how to allocate them can affect the lives and hopes of future generations. Indeed, the textbook definition of economics is “the science of allocating scarce resources among competing needs.” Yet there is no hint in Munnell’s response that distributing scarce goods among competing interests is what economics and, increasingly, politics as well are about. In fact, our society is now facing some fundamental choices on how we wish to allocate our resources among seemingly implacable demands—this we see in a federal budget generally acknowledged to be out of control and to be generating unprecedented and unacceptable deficits.
Today Social Security already spends more each year than the combined net investment in plant, equipment, research, and development of all US companies. In the years to come, its expenditures are scheduled to increase dramatically. Is this a sound allocation of resources? Are public policies that will convert savings into immediate consumption appropriate to our current needs, or should we formulate policies that will encourage us to invest in future productivity? Any discussion of the enormous Social Security program that ignores these questions is bound to go around in circles.
I shall then discuss briefly the “solution” to the Social Security problem recently announced by the National Commission on Social Security Reform. The commission labored long and hard to reach agreement, and deserves considerable praise for dealing with an extremely difficult political situation. Its proposals are useful, if tentative, first steps. However, if they are not followed promptly by larger strides forward, these modest steps will merely postpone for a few years a crisis which they do little to avert. Like Munnell, the commission defined away huge chunks of the Social Security problem.
Finally, I shall offer some suggestions for the future—suggestions not specifically about Social Security, for I have set forth my proposals on that question in my articles—but for a society confronting scarcity, a democracy that can no longer take economic growth for granted.
Munnell begins by denying much of the problem with which she purports to deal. “The first step in understanding the Social Security financing situation,” she writes, “is to separate the problems of the Old Age, Survivors, and Disability Insurance (OASDI) Program from those of the Hospital Insurance (HI) system.”
The future financing requirements of HI are extremely uncertain…. Restoring long-run balance to the HI system will undoubtedly require fundamental reform in the way we provide hospital care.
Noting that the National Commission on Social Security Reform, among various others, has tended to focus on the OASDI portion of the program, Munnell follows suit. This is a crucial decision, since she thus ignores the component of Social Security that makes optimism in the near and medium term almost impossible; that component has been publicly termed three-fourths of the problem by A. Haeworth Robertson, chief actuary of the Social Security Administration from 1975 to 1978. 19
The case for including HI (or “Medicare Part A”) in any discussion of Social Security is a strong one. Like OASI and DI, it is supported, under the Federal Insurance Contributions Act, by a flat-rate tax applied to the same base of taxable wages of roughly the same population of covered workers. And like OASI and DI, HI represents an “intergenerational transfer” from current workers to current retired people.20 Indeed, some of the major factors driving the HI program toward a cost explosion are precisely the factors at the root of the OASDI crisis—in particular the growing ratio of beneficiaries to covered workers in the next century.
In addition to demographics, of course, the HI cost explosion is also being fueled by ever more sophisticated and expensive medical technology and perverse incentives that are driving medical costs to higher and higher levels. In 1982, for example, hospital costs rose three times faster than the overall rate of inflation. But these special effects—to which I shall return later—do not mean that HI is fundamentally different, rather that its problems are more serious. Perhaps the real reason why Munnell excludes HI from the present discussion is that it interferes with the “calmer look” that she promises.
Curiously, even after she defines HI out of her discussion, Munnell still finds complacency elusive. She spends much space trying to shore up the proposition that the OASI and DI components of Social Security are not in serious trouble. While accusing me of advocating “draconian” benefit cuts, she quietly proposes OASDI reforms similar to the ones I have advocated. She too favors a departure from perpetual indexation of benefits to the Consumer Price Index (CPI). She too thinks that a portion of Social Security benefits should be taxed and that some adjustment should be made in the initial benefit calculation. And, at least by implication, she too believes that there must be reductions in the cost of the HI program. Yet she also believes that there is no crash coming—that a few technical adjustments are enough to correct a few minor miscalculations.
Munnell has said all this before. In April 1978, Congress increased payroll taxes and eliminated a technical flaw in OASDI calculations by which benefits were overindexed to inflation. Munnell praised these actions as ones that would restore “financial soundness to the OASDI trust funds through the first decade of the next century.”21 Everyone now recognizes that her enthusiasm has not withstood the test of time.
In her current “calmer look,” Munnell remains undaunted. She starts by dividing the future into three periods. From 1983 through 1989, she recognizes a projected shortfall of $200 billion under the Trustees’ “pessimistic” assumptions and $75 billion under their “intermediate” assumptions. It is unclear which set of assumptions she finally accepts, though she seems tacitly to recognize that the evidence supporting the pessimistic assumptions is strong.
Mutely conceding this point, at least for the near term, Munnell seems amenable to something approaching the $200-billion figure. But she cloaks that figure in soothing words. The shortfall is “relatively manageable.” The reasons for it are “well understood” and future problems of the same sort can be avoided by “modifying the indexing procedure.” These reassuring words make the underlying message sound unexceptional; but readers should not be lulled, for she has recognized a serious problem. What she really means is (1) benefits have increased far faster than wages,22 (2) one key reason for this is automatic indexing to the CPI, and (3) it is therefore necessary to depart from CPI indexing. I agree. We need indexing changes, and more. But even in Munnell’s sanitized form, this message has not yet got through to most of this country.
Once certain minor modifications have been made to meet the shortfall between now and 1989, Munnell continues, the Social Security system will enter its second period: 1990-2014. This, she claims, will be the golden age of Social Security—a period of prosperity in which relatively fewer beneficiaries and a larger work force will generate booming surpluses. But these glowing predictions are made possible largely by Munnell’s decision to ignore the HI program—the component of Social Security that turns the period from 1990 to 2014 from a golden age to something like a disaster. The Social Security Trustees have published no set of assumptions under which the HI program will not be plunged into the red well before 1990 and remain there for the foreseeable future.
Munnell’s optimism about the period between 1990 and 2014 also derives from a subtle and unstated shift in her economic assumptions—a shift from something close to the Trustees’ pessimistic figures to their intermediate ones. For most of the period in question, the crucial assumptions are those concerning real-wage growth and unemployment. (Demographic assumptions do not achieve central importance until toward the end of Munnell’s middle period.) Yet for both real wages and unemployment, the pessimistic rather than the intermediate projections seem more plausible and certainly more appropriate for planning purposes.23
The pessimistic projection for average real-wage growth is for an average of plus 0.2 percent a year from now until 1995 and plus 1.0 percent thereafter. But even this pessimistic projection assumes a reversal of current trends: from 1973 to 1982, real-wage growth averaged minus 0.9 percent. From 1978 to 1982 it averaged minus 1.6 percent. And from 1960 to 1982, a period that includes the boom years of the 1960s, real-wage growth averaged only plus 0.6 percent. Moreover, there is little reason to expect a dramatic and sustained improvement in real-wage growth. On the contrary, we now face formidable obstacles that make it unlikely that the intermittent high growth rates of the past can be sustained over a long period; these obstacles include higher energy prices, the end of cost-free pollution, a marked slowdown in the growth rate of an increasingly competitive world economy, and the gradual shift in our economy away from manufacturing and toward services.
Munnell has no discouraging word to say about unemployment, but it too is a determinant of Social Security in her projected middle period. The Trustees’ pessimistic projection is for a declining unemployment rate, averaging 7.6 percent, from now until 1995 and a 6.0 percent rate thereafter. The intermediate projection is for an average of 6.6 percent unemployment between now and 1995 and 5.0 percent thereafter, a somewhat fanciful prospect. Unemployment during 1982 averaged 9.7 percent, and from 1978 to 1982 it averaged 7.2 percent. Currently it is running at between 10 and 11 percent. Moreover, many economists, including the current Council of Economic Advisers, now believe that structural changes in the American economy make a return to the low unemployment rates of the 1960s improbable, even during a period of sustained economic growth.
To put it bluntly, the economy has for some time been doing worse than the Trustees’ “worst case” mediumterm projection assumes. Yet Munnell is betting the future of Social Security on the Trustees’ considerably rosier intermediate assumptions. According to A. Haeworth Robertson, the former chief actuary of Social Security, this is not a good bet. In his letter to me, he writes:
I completely concur with your judgment that the “pessimistic” (alternative III) set of economic and demographic assumptions is far more plausible than the widely accepted “intermediate” (alternative II-B) set of assumptions. Certainly it makes sense to focus on the former as a prudent guide for policy making.
Had Munnell faced up to HI and to the real possibility that the pessimistic projections will be borne out, her golden age of Social Security would have looked something like this: by 2000 the combined deficit would be 6.05 percent of taxable payroll; five years later it would be 8.67 percent; and by the end of the period it would be 13.83 percent, or an incredible $2.2 trillion. Table 1 sets forth the intermediate OASDI projections on which Munnell relies—and shows what happens to those projections when HI is included and the pessimistic assumptions are used.
Munnell’s third period runs from 2015 to 2060. Here she is forced to admit that the program faces dramatic cost increases as the “baby-boom” generation retires and the number of beneficiaries per covered worker increases rapidly. She claims, however, that the problem is not so great as I suggest. She makes two points.
First, having decided not to discuss HI, she reintroduces it only to suggest that my projections are unreasonable. She accepts, for purposes of argument, the pessimistic projections, which I believe are the more realistic, and notes that under those projections the cost rate for the OASDI portion of the program for the year 2035 is 24 percent of taxable payroll (or a trifling $13.5 trillion). Noting that my estimate for the total cost of the program in that year is 44 percent of taxable payroll, she rightly concludes that this implies that HI will rise to 20 percent of taxable payroll.
This, she thinks, is not “remotely realistic.” Her argument seems to be that since HI today represents only 18 percent of total Social Security expenditures, “it is difficult to believe that we will allow the HI program to grow to a point where [its] cost…roughly equals the total cost” of the OASDI program. But here I have, if anything, been guilty of overoptimism myself. Since the official Trustees’ projections for HI do not extend past the year 2005, I simply used the official HI projection until that year and then assumed that HI costs will rise thereafter at the same rate as OASDI. Actual HI behavior since 1967, and every HI projection through the year 2005, show that HI costs have grown and will grow at a rate far faster than that of OASDI.
Indeed, the Social Security Trustees attribute a little over half of the annual growth in HI benefit payments to economic and demographic factors and most of the remainder to seemingly remorseless cost pressures peculiar to American medicine, which have been largely responsible for the growth in disbursements per beneficiary since Medicare began.24 Yet for purposes of exposition, I ignored the possibility that these would continue to accelerate after 2005, taking a conservative path from the point where the official projections leave off. To see how conservative this path may be—and to appreciate the magnitude of the HI cost explosion—we need only look at the history of HI costs per beneficiary shown in Table 2.
Although Munnell does not believe that HI costs could be nearly equal to those of OASDI by the year 2005, the figures I used simply reflect what the Trustees themselves estimate will occur by the year 2005 under the pessimistic assumptions. Indeed, even under the intermediate assumptions, HI costs will equal about two-thirds of the cost of OASDI in 2005.
Three years ago A. Haeworth Robertson made a projection based on pessimistic assumptions that were then more optimistic than current ones. It nevertheless indicated that the total cost of the OASDHI program in 2035 would be 38 percent of taxable payroll. Subsequent events—a worsening of pessimistic assumptions in general and rapid inflation in the medical sector in particular—make my 44 percent figure entirely reasonable.25
Munnell asks how much “we will allow” the HI program to grow. That, of course, is precisely the question. Surely she must recognize that new medical technologies are priming the HI program for a cost explosion without precedent, and that those technologies raise wrenching moral and economic questions, which need prompt and forthright attention.26 But she leaves the problem for others to resolve.
After dismissing my HI projections for the years between 2015 and 2060, Munnell raises her second point. She backtracks and once again criticizes the pessimistic assumptions themselves. Once again we differ in our judgments of the future.
Many readers will no doubt wonder at this point whether such highly speculative debate can possibly be of much use. If few of us trust an economist to make a three-year projection of trends in GNP, of what use is a seventy-five year projection of birth-rate trends? The answer is simple: the full consequences of the Social Security policies made today will take generations to be felt. Because the benefits for the elderly rest upon the confident expectations of the young, plans for Social Security must extend through the cycle of generations. To be sure, long-term projections are difficult and speculative—but they are nonetheless essential. Their purpose is not so much to predict what will happen (that is clearly impossible to do with precision) but rather to tell us what is likely to happen given a plausible set of assumptions. And preparing for likely possibilities is especially important if one believes, as I do, in “gradualism,” that it is far better to adjust benefits and taxes slowly and deliberately than quickly and without warning. If you wait for the future to happen, it will already be too late; you will have no choice but to make precipitous changes, including perhaps cuts so “draconian” as to make my own suggested reforms seem mild. For those who already face the vulnerability of old age and the unique uncertainties of retirement, this must never happen. One purpose of reasonable projections is to prevent this from happening.
One of the assumptions that continue to be important for the years between 2015 and 2060 is the rate of real-wage growth. In addition, the two demographic variables—mortality rates and fertility rates—become increasingly influential as the twenty-first century wears on. Table 3 shows that the effects of the demographic assumptions on the ratio of taxpayers to beneficiaries will be immense.
It also suggests why Munnell still foresees long-term financial problems even after jettisoning HI and choosing the intermediate projection. Even these aggregate statistics do not reveal the extent of the explosion of the segment of the population over age eighty-five, the age group whose care costs the most. This group is expected to grow fivefold between now and the year 2050—from about 1 percent of the population now to 5 percent or more by then.
Munnell ignores the real-wage growth assumption for the out years, perhaps because she recognizes that her intermediate assumption of plus 1.5 percent a year for fifty years is unrealistic. Similarly, she says nothing about mortality assumptions, nor is it easy to imagine what she could have said since even the pessimistic assumption seems unduly modest.27 She does, however, criticize the pessimistic long-term fertility assumption (1.7 lifetime births per woman) as being too low, favoring instead the intermediate assumption (2.1 lifetime births per woman). While Munnell claims that the evidence “tends to support” the intermediate fertility assumption, she fails to substantiate her claim.
She concedes that last October the Census Bureau revised its own intermediate long-term fertility estimate down-ward from 2.1 to 1.9 (its low, i.e., pessimistic, estimate is now 1.6); but she neglects to mention that the Census Bureau is using more recent forecasting data than the Social Security Administration’s actuaries. It is likely that the actuaries will soon follow suit and revise their own intermediate and pessimistic assumptions downward as well. Moreover, while Munnell cites Census surveys indicating that women of childbearing age now expect to have an average of more than two children, she fails to point out that women who respond to these surveys (especially single women) tend regularly to overestimate the number of children they later have.28
As I said in my first article, little can be said with certainty about future demographic trends. But it is instructive that fertility rates fell to 1.83 by 1980, and that recent trends—the education levels, incomes, and marital status of young women, and their participation in the labor force—could well depress that rate in the future. For these reasons, the pessimistic assumption of 1.7 lifetime births seems plausible. Why gamble the future of Social Security on Munnell’s vague urgings to the contrary? It would be far more sensible to take a prudent course for now and, if events turn out better than expected, to increase benefits or lower taxes at some later date.29
Looking back over past decades, we find that real OASDI benefits have been increased time and again, accompanied in each instance by solemn assurances that we could afford the increases under some seemingly reasonable set of assumptions. After each increase, however, the subsequent cost was higher than had been originally predicted. In 1977, for the first time, it was decided that the trend of future benefit growth had to be diverted from its steep trajectory, and the first Social Security “rescue package” was conceived. Yet only five years later this rescue has proved abortive, and everyone now recognizes that substantial new corrections are needed.
Oddly there seems to be widespread reluctance to ask why the assumptions and estimates have been so consistently wrong. I suspect the entire system of making projections may itself be flawed. There seem to be two primary sources for estimates of the Social Security program’s future, and you do not have to be an expert in econometrics to appreciate that each may have a built-in bias.
The first source is the political process by which each administration sets forth its projections for the nation’s expected economic performance and congratulates itself in advance for its success in managing the economy in the years to come. Surely it comes as no surprise that politicians err on the side of optimism, for politicians who are pessimists, particularly toward the outcome of their own economic programs, tend to have rather brief careers.
The second source of estimates comes from the Social Security Administration itself, where a different sort of innate optimism prevails, the optimism of the advocate. Each year, the Social Security Administration and the Health Care Financing Administration, working with staffs of the Board of Trustees of Social Security, turn out sets of optimistic, intermediate, and pessimistic assumptions. Historically, their “intermediate” assumptions have been as a rule over-optimistic. Yet the Social Security Trustees—hardly disinterested observers in view of the membership of the secretaries of the Treasury, Labor, and Health and Human Services departments—routinely choose the intermediate assumptions for planning purposes.30 To prefer the projections that make the program look less likely to unravel in the decades ahead is understandable, especially since it is always possible to go back to the public for “emergency” tax increases when the system is threatened by “unplanned” deficits.
Little in Munnell’s article therefore urges me to rethink my estimate that payroll taxes may have to be as high as 44 percent of taxable payroll in 2035 to finance the combined OASDI and HI programs. Moreover, even my so-called draconian scheme of benefit cuts would in fact eliminate little more than three-fourths of the long-term deficit. Unless additional steps are taken, including some of the HI modifications I have suggested but not quantified,31 the deficit for the year 2025 will still amount to roughly 8.2 percent of payroll or $2.4 trillion.
Munnell’s palliatives notwithstanding, my discussion of Social Security’s current problems is no more cataclysmic than the problems themselves.
Before addressing Munnell’s other points, I must first try to place the Social Security issue in a larger context, for Munnell’s case for a “calmer look” depends on the absence of such a context. Nowhere does she face the economy’s urgent need for savings and investment; nor does she ask how we can sustain growth and prosperity for any longer than a brief cyclical upturn without taking immediate steps to reduce the expanding federal deficits projected for 1985 and beyond. Nor does she recognize that we cannot afford everything we may want. Instead, she asks us to consider Social Security in an economic vacuum.32
If the last decade has taught us nothing else, it is that it is no longer possible to expect automatic prosperity. The growth rate of national productivity has dwindled over the last eight years to 0.1 percent per year. Unless something is done to reverse this trend, we will become a society virtually without hope, one which devotes ever more energy to bitter conflict over the distribution of increasingly scarce resources.
To escape this fate will be both painful and difficult, and “pain” and “difficulty” are not popular words these days. Yet it should be clear to an economist like Munnell that we have little prospect of escape if we don’t invest now in future productivity—in new tools and technologies, trained workers to use them, and an educational system whose graduates will develop the technologies of the future.
I have recently helped to put together a bipartisan coalition to address the nation’s fiscal crisis. Economists, university presidents, lawyers, accountants, bankers, businessmen, and former government officials of widely different political views all agree on the need to reduce the enormous “out-year” structural deficits we now face and to free resources for investment in future productivity. Munnell proposes simply to increase federally financed consumption.
Rosemary Rinder agrees with her and goes a large step further. While Munnell says little about the issue of consumption and savings, Rinder asserts that “consumption is good, not bad.” She feels that we need accelerated Social Security benefits and federal deficits. She stakes our future on the miracle of the Keynesian demand-accelerator relationship. Of course, economies need some present or anticipated level of consumer demand if they are to function at all. And, of course, the ultimate goal of investment is to make greater consumption possible. But as a proposal for future action, what she says is flatly wrong. Our economic performance in the decades to come will worsen, not improve, if we continue to encourage current consumption at the expense of savings and investment. In the long run, total consumption will be lower.
In the current substantial cyclical recession, some might reasonably prescribe a shot of demand stimulation, but the appropriate way to analyze Social Security is not in years but in decades. Even in the heyday of postwar neo-Keynesianism, demand stimulation was advocated only as a remedy for short-term cyclical downturns. Moreover, many economists today agree that we have used that remedy to excess over the past fifteen years,33 and no one I am aware of would seriously support building into the economy a half-century of accelerating deficit-financed consumption in the name of neo-Keynesianism.34 The biggest threat to our long-term future is insufficient investment, not insufficient demand.
Investment requires savings, and as a nation we are saving far too little. Social Security may already be contributing to the problem, and in the decades to come it is likely to contribute far more. This could happen for two reasons. First, expected Social Security benefits may actually reduce the national savings rate, since many adults save to provide for periods of low income—in particular, retirement or disability. People who are confident of high future Social Security benefits may tend to consume more and save less, to spend private assets which they would otherwise accumulate.
Economists are divided on this issue. Martin Feldstein believes Social Security displaces large amounts of private savings, while Henry Aaron disagrees, saying, “I conclude that the evidence does not support the position that reductions in social security benefits would be effective in increasing private saving.”35 Munnell herself has written extensively on the subject, and in a recent article is somewhat equivocal.36 While I have no illusions that I can settle this professional question, common sense does suggest that the growth of Social Security benefits can displace at least some private savings, given the magnitude of the benefits.
The statistics themselves are suggestive. As of 1981, the net OASDI cash wealth “possessed” by covered workers and retired people (that is, the present value of the amount by which their expected benefits exceed their expected future contributions) was officially calculated at nearly $6 trillion. That comes to $25,000 for everyone alive in the United States today. None of this wealth is “saved” in the economic sense of the word. Some of it presumably could be.37 Further, whether benefit levels that are ultimately consistent with Social Security taxes at a combined rate of 14 percent of payroll diminish savings substantially, slightly, or not at all is only one question. Another, more obvious question is: how much will wage earners be able to save if payroll tax rates are at two or three times that level?
Secondly, and far more immediately and demonstrably, Social Security is a key element in the federal government’s current fiscal crisis, which has profound implications for savings and investment. Whatever the success of long-run efforts to restructure incentives so that we save more and consume less, we must immediately ensure that a sufficient portion of our existing—and limited—pool of savings is available for investment. The most direct and effective way of doing this is to reduce permanently the portion of our precious savings that is, and will be, soaked up by grotesquely large government deficits.
While there is no magical percentage of GNP that should be allocated to government borrowing, current projections for the balance of this decade are totally unacceptable. When federal budget deficits run to the currently projected 6 and 7 percent of GNP, adequate investment will, in view of the low US savings rate, be arithmetically impossible. In 1983, the budget deficit will consume an unprecedented 100 percent of the net savings generated by individuals and families in the United States, and by the middle of the decade, the deficit will consume close to three-quarters of the net savings generated by the entire private sector.38
To reduce these deficits so that savings are freed for investment—and in the process to curb instability in financial markets, to bring long-term real interest rates to reasonable levels, and to combat the other debilitating consequences of fiscal extravagance—federal expenditures must be cut and federal revenues increased. Only a few components of the federal budget are large enough to yield the needed reductions. One is defense spending (28 percent of outlays). Another is non-need-related benefits (39 percent of outlays). Among the latter, Social Security alone accounts for 26 percent of federal outlays. Unless Social Security is modified, it will account for 30 percent by 1990 and by the beginning of the next century, 36 percent. 39 Thereafter, its share will explode.
Thus a meaningful discussion of Social Security is impossible if it does not take account of the entire economy. What we decide about Social Security will be vital and not only to the welfare of our aged but to our prosperity generally. With this in mind, I turn to the rest of Munnell’s criticisms. She accuses me of three additional sins: of attributing our low productivity growth rate in recent years to Social Security’s financing problems, dismissing the possibility of increasing payroll taxes, and falsely dispelling what I call the “myths” that surround the Social Security program.
First, I did not argue that current productivity problems are the fault of Social Security, but was addressing primarily the productivity problems of the future. Yet Social Security may have contributed indirectly to our current woes, for while the Social Security program has run no significant deficits in recent years, it has remained solvent only through large increases in payroll taxes.
We have recently seen a major tax-payer revolt across this country, which has led to sizable income-tax reductions. The leaders of this revolt were doubtless rebelling against the total of federal taxation, including mushrooming payroll taxes required to finance Social Security, which more than account for the increase in federal revenues as a percentage of GNP over the past twenty-five years.40 The result, of course, was to make income-tax increases politically impossible and an across-the-board income-tax cut virtually inevitable.
Thus I think it likely that current unified budget deficits can to some degree be attributed to Social Security even though the funds themselves have historically been in balance. Munnell’s tranquillity notwithstanding, Social Security has already begun to consume too much of the income that our citizens are willing to commit to government spending.
Munnell also attacks me for rejecting payroll-tax increases as a cure for Social Security’s ailments and asserts that I failed to justify this rejection. In fact, as I wrote, tax increases are not an adequate response to the long-term problem, for the simple reason that this society cannot afford, and its members likely will not tolerate, a program that allocates to support of the elderly anything close to 44 percent of taxable payroll by the year 2035. Munnell’s table of international comparisons indicates that other industrial countries rely more heavily than we do on earmarked payroll taxes to fund social programs in general. She does not point out that income and corporate profit taxes are generally lower or nonexistent in these countries. Also, a careful review of her tables reveals that few of these countries levy tax rates anywhere near those that would be needed in the US to pay for Social Security benefits under the pessimistic projection. Finally, it is worth noting that Japan, the industrial world’s foremost growth economy, levies a combined payroll tax of 10.6 percent to fund OASDI-equivalent benefits.41
Munnell’s defense of tax increases as a way to save Social Security can be explained only by her insistence on addressing the wrong problem. The question is not simply how to balance the OASDI books, or even the OASDI and HI books, but how to balance the country’s books. And that cannot be done in an economic vacuum.
Other factors militate against major reliance on tax increases as well. To begin with, payroll-tax increases can have a negative impact—at least transitionally—on employment, unit labor costs, international competitiveness, and profits. Moreover, payroll-tax increases fall disproportionately on the poor because of the regressive nature of the tax42 and tend systematically to benefit older workers at the expense of younger ones because they reduce the return the latter will receive on their investment. I return to this point below.
Finally, Munnell tries to salvage a number of myths about the Social Security system—myths that, as I said in my article, have tended to obscure the Social Security problem.
To take her points in turn, I first suggested that many Americans believe that Social Security benefits are a mere return of prior contributions made by retired people during their lifetime employment. I pointed out that a worker with a nonworking spouse and an average wage history who retires today will receive roughly seventy-two times his lifetime contributions, and twenty-six times his lifetime contributions plus interest.43 My point was twofold. First, I wanted to make it clear that no one has a permanent right to expect such a windfall, that no one even expects it, and therefore that benefit levels should be decided without rhetoric about rights and entitlements. Second, I questioned whether any system that on average pays each participant so much more than it receives from him or her can survive.
Munnell dismisses these problems with a tranquil hand. The current enormously high ratio of benefits to contributions is, she says, “the inevitable result of the start-up phase of a pay-as-you-go system.” Nothing in that ratio, she says, implies that the system is “fundamentally unsound” or “fundamentally flawed.” I am not a professional expert in these matters and Munnell’s credentials as a student of pension plans far exceed mine. But I note that other experts disagree deeply. For example, Messrs. James Capra, Peter Skaperdas, and Roger Kubarych from her sister bank, the Federal Reserve Bank of New York, conclude, as I do, that the Social Security system is “fundamentally flawed.” I also trust that I can bring to the discussion a certain amount of common sense; and common sense tells me that something is seriously wrong with any system that is scheduled to pay disproportionate returns to a rapidly growing number of beneficiaries out of revenues from a relatively smaller number of contributors.
Munnell asserts that “as the system matures…and the number of beneficiaries stabilizes relative to the number of workers, the average retiree will receive benefits which are roughly equal in present value terms to combined employee and employer contributions.” While this sounds soothing, the facts are unfortunately jarring. The Social Security program is no longer a fledgling; the last significant expansion of covered workers occurred more than thirty years ago. Yet for average wage earners the windfall aspect of the program will persist at least until the year 2040.
What is really happening, and what has been happening for years, is this: each time an increase in real benefits per beneficiary outstrips the growth in real wages per covered worker a new “start-up phase” begins. Since the 1960s we have experienced many such increases. Thus the chain-letter-style return ratios flow neither from the noble inception of a bold and worthy program nor from some deliberate and carefully calibrated expansion. Rather, they are the result of helter-skelter benefit increases, some of which were enacted in the euphoria of overly optimistic assumptions while others were the product of arithmetic quirks in the initial benefit and indexation formulas. To dismiss these current and future windfalls as the inevitable byproduct of Social Security’s “start-up phase” is simply misleading. In fact, the windfall is a vivid symptom of a serious ailment: benefit increases for which we ultimately cannot afford to pay.
Munnell’s response to the issues of intergenerational fairness raised by current, but declining, high rates of return is equally unsatisfactory. Basically she advocates doing nothing because “individuals close to retirement who have planned on a particular level of income from Social Security” should not find their benefit levels “reduced precipitously.”
There seems a curious political cynicism beneath Munnell’s approach. In her discussion of appropriate benefit levels and indexation, she clearly chooses to promise as much as possible and hope for the best. But when past instances of such optimism prove to have been unjustified, she insists that increased real-benefit levels resulting from these Panglossian excesses, economic accident, or just plain electioneering are suddenly inviolate because powerful groups have come to view such benefits as entitlements. This is not a satisfactory way of making policy. It means that no decision can be reexamined (even when sufficient information accumulates to make reexamination urgent), that major expenditures are systematically insulated from economic realities, that mistakes cannot be corrected.44
Though Munnell is surely right to oppose truly precipitous and immediate cuts in benefits, this is not what I proposed. On the contrary, I called for a series of gradual measures to reduce the growth of future benefits. There is nothing draconian about asking that such benefits increase no faster than the cost-of-living increases given the wage earners paying the taxes, and there is nothing precipitous about a series of gradual steps to increase the retirement age to sixty-eight (by three months a year) commencing in 1990. But it would be draconian to pretend that there is no problem and wait until the only choice left is a truly precipitous reduction in benefits.
Munnell also attacks my views of intragenerational fairness—the distribution of benefits between the elderly poor and the elderly well-off. I noted that a significant portion of Social Security benefits go to the elderly who are well-off and that the elderly as a group may not be as needy as we think. She distorts these comments and accuses me of implying that “Social Security benefits are superfluous.”
This does the reader a disservice. Our commitment to alleviating poverty among the elderly must be firm, and I explicitly supported it in my article. In particular, I made it clear that elderly beneficiaries near the poverty level should be exempt from benefit reductions, that the Supplemental Security Income program should, if anything, be made more generous, and that taxing benefits in excess of one’s contributions at graduated rates (a step which would not hurt the elderly with low incomes) was a desirable and significant reform.
Far from advocating that we abandon our commitment to the elderly poor, I propose to ensure that we can afford that commitment for decades to come. To do so—to maintain a sound and meaningful safety net for the poor—we must stop padding the hammocks of the comfortable. Someone, in other words, will have to give up something. My proposal is that the appropriate group should consist of those of us who are well above the poverty level—but certainly not the poor.
Munnell disagrees not because she has an alternative way of distributing the burdens, but rather because she wants to avoid tough decisions about who has to pay and to pretend instead that no one has to sacrifice anything. Her analysis is worth repeating in full:
Moreover, disproportionate cuts in the benefits for higher-paid workers would endanger public support for the program. Social Security’s progressive benefit formula already produces proportionately greater benefits for lower-paid workers than higher-paid employees. Further reductions in the benefits for those with above-average earnings would mean that as the system matures these people would receive benefits that are less than combined employee and employer contributions plus interest. If this were to happen, support for the Social Security program would decline.
This is stunning. Benefits to the elderly poor are, as they should be, sacrosanct. But Munnell wants to protect them by bribing the elderly well-off to get on the political bandwagon.45 Thus the rich must continue to receive at least the present value of their past contributions in order that the poor receive somewhat more. But if everyone gets on the bandwagon, who will be left to pull it? If everyone receives, who will be left to pay?46 Munnell is silent on this question.
After my articles were published, the National Commission on Social Security Reform released its proposals for dealing with the system’s current problems. Depending on how it is received and interpreted, the commission’s report can do a fair amount of good or a fair amount of harm.
First, the commission, in my view, deserves praise for dealing adroitly with a difficult political situation and coming up with a series of politically workable proposals. I have the highest regard for the chairman of the commission, Alan Greenspan, and many of its members, and I know they have struggled mightily to arrive at a consensus. That consensus ensures that the commission’s proposals have a good chance of being enacted swiftly.
What must be understood, however, is not only what the commission did, but also what it did not do. It made important proposals, including a delay in the next scheduled cost-of-living adjustment, a standby modification in the procedure for making future adjustments,47 a limited taxation of benefits, accelerated scheduled tax increases, and universal coverage of new federal workers and nonprofit employees by Social Security.48 Its most lasting contribution may well be that it put before the public, however tentatively, other possible reforms, including increases in the retirement age and more regular, annual restraints or limits on the indexing procedure. It did not recommend any of these steps, however. More important, it made a deliberate decision not to deal at all with the HI problem,49 and it did not identify, let alone attempt to solve, the vast problems that the combined OASDHI program will face in the next century.
These latter points have not received much publicity. Instead, the first reports in the press implied that the commission’s proposals will eliminate two-thirds of the Social Security deficit over the next seventy-five years. But if we include HI we find that they eliminate less than one-fifth of the projected deficit;50 and if we adjust our assumptions from the intermediate ones to the considerably more realistic pessimistic ones, the commission’s proposals would eliminate about one-fifteenth of the deficit.
At best, the commission’s proposals get the OASI and DI funds through the balance of this decade while leaving the HI fund deep in red ink. Even this achievement would be accomplished in part through transfers from other federal revenues, since about $50 billion to $60 billion of the $169 billion the proposal could inject into the OASI and DI fund will come out of another federal pocket—general revenues. This is so because portions of the proposed tax increases will be deductible from taxable income, or can be credited against taxes, and because another source of funds—the lump-sum reimbursement for military wage credits—will come directly from general revenues. Thus the net effect on unified budget deficits over the entire seven-year period 1983-1989 is not very large, only about $110 billion (averaging less than $16 billion a year), in the face of $250 billion annual deficits or more in each of these years.
At worst, the commission’s proposals will get the retirement and disability funds only through 1983 and 1984 and a new crisis could be upon us in 1985. The largest of the tax-rate changes designed to shore up the system would not take place until 1988 and 1989. In order for OASI and DI to make it through 1985-1987, the commission has recommended allowing the funds to “borrow” from the Hospital Insurance fund. However, Dwight Bartlett, former chief actuary of the Social Security Administration, has expressed doubts that even with this provision the system would remain solvent over those three years.51
The commission’s report can have the useful effect of relieving part of the short-term Social Security problem while helping us to focus on the balance of the short-term problem and the whole of the long-term one. But if, as with past reforms, we pretend that the report has restored long-term solvency to the system, we will find that it has become simply another example of unjustified optimism and an excuse for the inexcusable—a decision to do nothing more.
The greatest challenge for citizens of democracy is to make difficult decisions. Yet our legislative process often seems stalled, our leaders paralyzed before the formidable political force of special interest groups who view citizenship as lining up at a public vending machine to get their “entitlements.” We need as a society to stop consuming as if there were no limits and we need to start saving and investing more. As I showed in my first article, the US gross savings rate—the percentage of gross domestic product (GDP) allocated to investment rather than consumption—has lagged behind that of every other industrial country. During the 1970s our gross savings rate was 19 percent, as opposed to an average of 26 percent for all other industrial countries and a colossal 35 percent for Japan.
To increase our savings rate, we must lower our consumption rate. This is an inexorable law: There is no way to increase the one without decreasing the other. Perhaps it would help to make our international comparisons in reverse—to compare the 81 percent consumption rate in this country to the 74 percent rate in all other industrial countries and the 65 percent rate in Japan. These numbers may not be pleasing, but they are the economic reality.
Munnell seems to assume that our savings and investment problems are hopeless. If Social Security is cut back, she writes, “the working population will probably end up providing equivalent support through some other program, since most people fail to save adequately on their own for retirement and many do not have private pension coverage.”
There are two problems with this perspective. First, unless we invest now in the machines, technology, and training necessary to increase output per man-hour, the burden of consumption by the baby-boom generation when it retires will overwhelm the workers of the next century. We can only invest more now if we save more now. But—and this is the second problem—Munnell apparently believes that our society has become irretrievably addicted to current consumption and that the government’s key role is to ensure a steady supply of the drug we crave regardless of the ultimate consequences.
I could hardly disagree more. Governments can affect the habits and behavior of their citizens in countless ways. Our government has for years created an environment that is hostile to savings and investment. It has done so through economic policies that promote inflation, through benefit programs that reduce the need for private savings, and through taxes that reward consumption over investment. As to the latter, I have recently reviewed the tax systems of seven leading industrial countries and found that the United States collects a smaller proportion of its revenues from consumption-based taxes than any of the others. We have, I fear, the most pro-borrowing, pro-consumption, and anti-savings-and-investment tax system in the industrial world. If we change our tax system and otherwise modify the environment we have created, I am confident that we can change our spendthrift ways.
Investment in our future, of course, is not just a matter of building more plant and equipment in the private sector. It also means renovating our crumbling public infrastructure, pioneering new industrial technologies, and dramatically increasing the resources we devote to educating our young. Our efforts in those directions often compare lamentably with those of other nations. Japan, for instance, during the 1970s invested seven times as much of its gross domestic product in public infrastructure as we did. With one-tenth of the world’s GNP, Japan now possesses more than two-thirds of the world’s industrial robots, and it still orders more than one-half of the new ones produced.
Moreover, Japan pours far greater resources into scientific education for the young than we do—some estimate three times as much. Thirty percent of its youth study calculus and an equivalent number take computer-related instruction. Only 4 to 5 percent of our young people do so. Yet we continue to devote more than three times as much public spending to the elderly as to the young.52
Munnell argues that the burden of the elderly in the future will not be as great as it appears at first glance, because the working population will have fewer children to support than it does today. However, I believe our spending on education and training for our youth is now woefully inadequate. If there are to be fewer children in the future, I would hope that we would invest more per capita in them. We owe our young people the kind of education that will enable them to make the most of technological and scientific possibilities. As yet, we have not even conceived, much less implemented, a program of training and retraining our current work force for an increasingly technological, competitive, fluid global economy, particularly the young people who have been unable to find regular work.
In his book Governing America, Joseph Califano observed:
The proportion of the Federal budget spent on programs for the elderly has been rising steeply. In 1980, it was 25 percent; by the turn of the century it will approach 35 percent; and when the senior boom hits in 2025 it will hit about 65 percent.
Is this an appropriate allocation of our scarce resources? Was it arrived at by a deliberate decision made by an informed electorate that understood the choices and consequences? I think not.
But addressing our urgent need for investment need not cause us to weaken our commitment to alleviating poverty and hardship among the elderly;53 it will require some hard decisions about where to cut back. In my view, one significant place to cut back, in addition to the military, is in payments to the elderly with middle and upper incomes, which means cutting back the growth of Social Security, the bloated 100 percent indexed federal and military pensions (which are far more generous than those earned in the private sector) and other government entitlements and subsidies for which there is no means test.54
If I may quote Herbert Hoover here, he once said: “Blessed are the young for they shall inherit the debt.” At present, they are scheduled to inherit a great deal more than the widely proclaimed $1 trillion debt: They are to inherit nearly $7 trillion of unfunded liabilities from these entitlement programs for the elderly, nearly $6 trillion for Social Security, and $1 trillion for federal and military pensions. What this means is that we are mortgaging our future. And we cannot afford to do that any longer. The time has come to admit this and to chart a new course. The single-mindedness of single-issue politics will no longer do.
There is nothing liberal or humanitarian about pretending that the unaffordable can be afforded. One should hope that liberalism implies some boldness, some vision, some willingness to devise new solutions. This is the spirit in which the Social Security program was inaugurated. Yet today too many liberals seem to have dug themselves in for a last-ditch defense of the status quo. They have substituted hindsight for vision.
Munnell, Rinder, and others who have made similar arguments would do well to consider this. If their obvious knowledge and abilities were addressed to the deep and interdependent problems of a stagnant economy, these thoughtful people could make a valuable contribution to solving those problems. If they remain obsessed with single-issue politics, they will inadvertently prolong the stagnation that neither of them would want. For now, unfortunately, Rosemary Rinder promises a utopia of costless consumption and Alicia Munnell offers a “calmer look” at Social Security, which, I fear, is merely the most recent of several calms before the storm.
March 17, 1983
George F. Will, “Social Security: An End to Fantasy,” The Washington Post, December 19, 1982, p. C7; George V. Higgins, “A Call to Curb Social Security Raises,” The Boston Globe, December 11, 1982, p. 14; Anthony Lewis, “Social Security Alarm,” The New York Times, November 29, 1982, p. A19. ↩
Peterson, “The Salvation of Social Security” (NYR, December 16), p. 54. ↩
Ibid., p. 50, footnote 4. ↩
Yung-Ping Chen, “The Growth of Fringe Benefits: Implications for Social Security,” Bureau of Labor Statistics, Monthly Labor Review, vol. 104, no. 11 (US Government Printing Office, 1981), pp. 3-10. ↩
The total dependency ratio in 2035 will be 0.86 as compared to 0.91 in 1965. ↩
US Congress, Senate, Special Committee on Aging, Social Security in Europe: The Impact of an Aging Population, Committee Print, 97th Congress, 1st session (US Government Printing Office, 1982), Table 6, p. 13. ↩
For a survey of the issues and major articles in this area see Barry P. Bosworth, “Capital Formation and Economic Policy,” Brookings Papers on Economic Activity, 1982:2 (Brookings Institution, 1983). ↩
Peterson, “The Salvation of Social Security,” p. 52. ↩
Ibid., p. 50. ↩
National Commission on Social Security Reform, “Money’s Worth Comparison for Social Security Benefits,” Memorandum No. 45 (August 12, 1982), Table 1. ↩
In the long run, under a pay-as-you-go system people will receive benefits that amount to their combined employer and employee contributions plus a return that equals the rate of growth of population plus the rate of growth of real earnings per capita. See Henry J. Aaron, “The Social Insurance Paradox,” Canadian Journal of Economics and Political Science, vol. 32 (August 1966), pp. 371-374; Paul A. Samuelson, “An Exact Consumption-Loan Model of Interest with or without the Social Contrivance of Money,” and “Reply,” Journal of Political Economy, vol. 66 (December 1958), pp. 467-482, and vol. 67 (October 1959), pp. 518-522. ↩
Peterson, “The Salvation of Social Security,” p. 53. ↩
US Department of Health and Human Services, Social Security Administration, Office of Research and Statistics, Income of the Population 55 and Over, 1978, prepared by Susan Grad, Staff Paper no. 41 (US Government Printing Office, 1981), Table 32, p. 54. ↩
See Edward Nell and Rosemary Rinder, Beyond Austerity: Economic Policy and Economic Growth (forthcoming), for a more complete discussion of these issues. ↩
Indeed, the prosperous have been known to complain of the difficulty in budgeting resulting from lower take-home pay early in the year, when they are paying Social Security. Note that for someone earning $70,000, it only takes 5.5 months to gross $32,500. Most Americans would love to have this difficulty. ↩
See Okner, “Total US Taxes and Their Effect on the Distribution of Family Income” in The Economics of Taxation: Studies of Government Finance, edited by Henry J. Aaron and Michael J. Boskin (Brookings Institution, 1980). ↩
Since many working-age women were also “dependent” in the 1960s, at home caring for the numerous children, the proportion was actually higher. Then, only 37 percent of the total population held paying jobs. Today that figure is 44 percent. ↩
So much so that I sought the view of a detached and acknowledged expert, A. Haeworth Robertson, chief actuary of the Social Security Administration from 1975 to 1978 and now a managing director of the employee benefit and consulting firm of William M. Mercer, Inc. I showed Robertson my earlier article, Munnell’s response, and a draft of this reply to that response. I refer periodically in the pages that follow to Robertson’s response. Robertson is “in substantial agreement” with my “estimated, current-law projections of future OASDHI cost rates and deficits” and argues from this that my articles provide a “much needed and accurate perspective” on Social Security. Robertson has authorized me to quote from his letter. ↩
Referring to the National Commission’s decision not to consider the HI, or “Medicare, Part A” program, Robertson observed that “if the total Social Security problem is too large for the commission to tackle and it wants to handle only one-fourth of the deficit, that is understandable,” but then it should “tell the public that it is trying to resolve only one-fourth of the problem” (cited in Deborah Rankin’s “Personal Finance” column, New York Times, January 9, 1983). In his letter to me, Robertson suggested that any discussion of Social Security should also include the Supplementary Medical Insurance program (SMI or “Medicare, Part B”), which pays doctors’ bills for Social Security beneficiaries but is only 25 percent financed by premiums paid by beneficiaries. The share paid by federal “general revenues” is now 75 percent. I have not considered SMI costs because the program is not financed by payroll taxes. Had I done so, my estimate of future Social Security costs would have been even larger—roughly another 4 percent of taxable payroll, according to Robertson. ↩
It is true that HI benefit levels do not reflect previous earnings histories of recipients—but then again, OASI and DI benefits reflect such histories only imperfectly. It is also true that HI benefits are “in kind” as opposed to cash. But it is not clear why that justifies excluding them from a discussion of Social Security. ↩
“Why the Furor over Social Security?” Christian Science Monitor, April 12, 1978. As I indicated in Part I of my article, President Carter assured us in the same year that “from 1980 to 2030 the Social Security system will be sound.” ↩
Specifically, from 1970 to 1980, the average retired worker’s OASI benefit rose by 49.5 percent in constant dollars while the average wage of a covered worker increased by 6.4 percent in constant dollars. ↩
Even some of the government’s own actuaries have questioned the chances of a rapid reversal in real-wage trends like that assumed in the Trustees’ intermediate scenario. The actuaries of the Health Care Financing Administration, who prepare the HI projections, for example, suggest that a major reversal in current real-wage growth trends is unlikely. Their reasons are given in “A Critical Analysis of the Assumptions in the 1980 Social Security Trustees’ Reports,” by Roland E. King and Clifford K. Powell, Transactions of the Society of Actuaries, vol. 33 (1981). ↩
1982 Annual Report of the Board of Trustees of the Hospital Insurance Trust Fund. ↩
This long-range projection is reported in A. Haeworth Robertson, The Coming Revolution in Social Security (Reston Publishing Co., Reston, Virginia, 1982), p. 92. In Robertson’s letter to me, he wrote: ↩
The awesome dilemmas have been illustrated by Alan Greenspan, chairman of the National Commission on Social Security Reform, in the following clearcut language: “We cannot substantially constrain the cost of Medicare unless we slow the improvements in technology (a dubious goal) or choose not to employ the technology that is currently available. These decisions, of course, would raise very difficult ethical and moral questions.” (Greenspan’s testimony before the House Ways and Means Social Security Subcommittee, February 1, 1983.) ↩
In fact, a straight extrapolation of mortality rate trends, by cause of death, on the basis of the period from 1968 through 1977 suggests that the life expectancy of a sixty-five-year-old in the year 2000 will be one year longer than the Trustees’ most pessimistic assumption. See Eileen M. Crimmins’s article in Population and Development Review, July 1981. ↩
“Assessing Cohort Birth Expectations Data from the Current Population Survey, 1971-81,” Martin O’Connell and Carolyn C. Rogers (US Bureau of the Census, forthcoming). ↩
Because the decisions we make about the Social Security program in the next few years will have a significant impact on the performance of our economy, there is a certain circularity to the whole process of relying upon assumptions. For example, I think it quite likely that if we follow Munnell’s advice and act as though the intermediate economic assumptions will be borne out, we may, perversely, guarantee fulfillment of the pessimistic ones for the simple reason that lower investment will retard productivity growth and, with it, real-wage growth. Conversely, if we take action now on the basis of the pessimistic assumptions, we may be able to put the economy on a sounder course and make fulfillment of the intermediate assumptions somewhat more likely. Should that occur, of course, we could then make appropriate alterations in the levels of benefits and/or taxes. ↩
The article by King and Powell, cited in note 6, is most revealing in its discussion of the political bias toward optimism in Social Security assumptions and projections. ↩
See page 57 of my article “The Salvation of Social Security” (NYR December 16) for a list of some reforms that might reverse some of the perverse incentives that conspire to escalate costs. ↩
Munnell does suggest that our current productivity and investment problems can be blamed to some degree upon a decline in corporate profitability in recent years—a decline exacerbated by inflation and high corporate tax rates. While this may have been a factor in the past, it is not likely to be a driving force in the future. Recent tax changes have lowered the effective tax rate on corporate profits, and my estimates in my earlier articles of savings available for investment included an increase in corporate savings. That increase, those estimates show, will be swiftly consumed by voracious federal deficits unless the budget is brought under control. Munnell does not discuss these issues at all—her references to productivity and investment look only to the past, not the future. ↩
Many of our current problems can be laid quite directly at the door of excessive demand stimulation. In particular, such stimulation led to the upward ratcheting of “stagflation,” to a secular downturn in productivity growth, and to worsening investment performance. Rinder’s reliance upon a gross investment statistic (10 percent of GNP) is misleading for a variety of reasons. The story is far more starkly told by a look at the growth rate of our net stock of business plant and equipment, which has declined from 5.7 percent during the period from 1965 to 1969 to less than 3.3 percent during the period from 1975 to 1981. After subtracting capital expenditures to comply with government regulations, the growth rate for the most recent period is less than 3.0 percent. ↩
Only in the Great Depression would anyone have taken Rinder’s permanent demand stimulation thesis seriously. The economic circumstances we face today are, to say the least, very different from those of the 1930s. ↩
Henry J. Aaron, Economic Effects of Social Security (Brookings Institution, 1982), p. 52. ↩
In “Social Security, Private Pensions, and Saving,” New England Economic Review (May/June 1981), she concludes that “little evidence exists to support” the savings-displacement thesis in the past, but she also says that “at this stage an expansion of the funded private pension system would probably have a slightly more favorable impact on saving than increasing the pay-as-you-go social security program.” ↩
The idea that anticipated Social Security benefit levels are inversely related to private savings rates can be linked quite directly to virtually everyone’s insistence—Munnell’s and my own included—that benefit levels, if they must be reduced, be reduced gradually rather than precipitously. The usual explanation is that those who are affected ought to have time to “adjust” their financial plans. It is hard to imagine what this adjustment of plans would entail if not a decision to consume less and save more in some form—e.g., in a bank account, pension, IRA, life-insurance policy, etc. ↩
“Net savings” denotes the excess of total savings over depreciation of existing capital—plant, equipment, and residences. ↩
These figures assume that federal outlays remain at a constant percentage of GNP. ↩
See my article “The Coming Crash” (NYR, December 2) for the data that support this statement. I might also note that today over 25 percent of families paying any Social Security tax pay more in Social Security taxes than they do in income taxes. (Fifty-one percent do so if we count, as many economists argue we should, the additional taxes paid by the employer.) As the payroll tax increases, that percentage figure will as well, and the next taxpayer revolt will therefore flow even more directly from payroll taxes—particularly at levels not one or two percentage points higher than now, but two to three times what they are now. ↩
One of the problems with such international comparisons is that they apply only to one historical period. Munnell neglects to mention that none of these countries had a postwar baby boom comparable to ours and that while their ratios of retired people to workers are in consequence larger today, they are not destined to increase to the extent that ours will early in the next century. ↩
Munnell’s comments about projected increases in fringe benefits as a percentage of total compensation reflect her decision to ignore the issue of taxpayer equity. As fringe benefits increase as a percentage of average compensation, the effective tax rate on total compensation required to fund Social Security may, as Munnell suggests, diminish somewhat. However, since low-wage earners receive fewer significant fringe benefits, they will bear more and more of the burden of the tax. Thus the regressive nature of the payroll tax will be exacerbated. ↩
The figures are based upon the Trustees’ pessimistic assumptions, and count only the employee’s contributions; including the employer’s contributions (plus interest) would reduce these figures by half. ↩
The recent experience of the National Commission on Social Security Reform is a grim testimonial to this ratchetlike bias. The enormous real-benefit increases which lie at the root of our current problem passed through Congress in the late 1960s and early 1970s like a hot knife through butter. Yet in grappling with those problems during the past year, the commission as a whole found even modest benefit reductions politically unthinkable. ↩
The bribe is not an inexpensive one, for although Munnell correctly asserts that lower-paid workers receive, on a percentage basis, a higher return on contributions than higher-paid workers, a recent analysis by the Center for the Study of American Business has calculated that the absolute dollar value of the subsidy to the high-paid worker far exceeds the dollar value of the subsidy to his lower-paid counterpart. ↩
Munnell seems to imply that if benefit rules are not touched, every future worker can rest assured that he will receive benefits at least equal to his contributions. This is not true under any scenario if the system is to remain self-financed. The Samuelson model cited by Munnell does suggest that it is possible—using highly artificial assumptions—for each generation to have an average benefit-to-tax ratio greater than one. Declining birth and productivity rates, however, now render those theoretical assumptions utterly fanciful. ↩
Included in the commission’s long-term recommendations is a proposal to base automatic OASDI benefit increases after 1987 on CPI or wage increases, whichever is lower, if the ratio of OASDI balances to outlays for the year is less than 20 percent. If this OASDI fund ratio is greater than 32 percent at the beginning of the year, scheduled monthly OASDI benefits would be supplemented to make up for past wage indexation as long as the fund ratio is not less than 32 percent. ↩
Two things should be said about this coverage of federal workers. First, the 7 percent of salary contribution of these new federal workers and nonprofit employees to Social Security would be offset by a roughly equivalent amount that would no longer be contributed to civil service retirement funds. Thus there is essentially no net effect on the unified budget deficit, though of course this proposal would help the solvency of the Social Security system. ↩
Chairman Alan Greenspan has emphasized that this decision by no means implied that the commission was minimizing HI’s problems. “It is important to remember, as this committee grapples with the financing of the OASDI system, that there are very significant potential problems in Medicare. The financing problems, at least as now currently projected, are very large. But while OASDI and HI are two problems, they draw from one pot of revenues.” (Greenspan’s testimony before the House Ways and Means Social Security Subcommittee, February 1, 1983.) ↩
More recent press reports suggest that the Social Security actuaries will shortly raise their estimate of the long-term deficit, so that a smaller fraction of that narrowly defined deficit will be eliminated by the commission proposal than was initially thought. ↩
Testimony of Dwight Bartlett before the House Ways and Means Social Security Subcommittee, February 9, 1983. Robert Myers, another former chief Social Security Administration actuary and executive director of the commission, is reported to be in agreement on this point (New York Times, February 6, 1983, p. E2). ↩
Three times as much per capita, including outlays from federal, state, and local levels; see “The Influence of Low Fertility Rates and Retirement Policy on Dependency Costs” by Robert L. Clark (North Carolina State University, unpublished, 1976) and Stephen Crystal, America’s Old Age Crisis: Public Policy and the Two Worlds of Aging (Basic Books, 1982). ↩
Nor is it my point that we should as a society reduce our commitment to the general objective of providing for our elderly citizens a dignified and generous retirement. I would simply propose that we finance that retirement realistically, and not rely on deficit-financed public consumption. My alternative can provide a sure guarantee of a dignified retirement while increasing the nation’s pool of savings. Munnell’s does the opposite—it depletes our already shallow pool of savings and builds into the system levels of government spending and taxes that are simply not affordable. ↩
The Bi-Partisan Budget Appeal recommends a one-year freeze in the cost-of-living adjustments for these non-means-tested entitlement programs and a limit on future indexing as well as cuts in other subsidy programs not going to the poor. The Appeal also proposes a $25 billion cut in the 1985 defense budget, not including reductions in military pensions. ↩