Implications of Social Security's Long-Range Financial Projections, The

Number 6 in Series, "The Retirement Project"

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

Note: This report is also available in the PDF format, which many users find more convenient when printing.

---

Table of Contents

Introduction
The Baseline Pension Cost Projections
Key Demographic and Economic Assumptions
Life Expectancy and Pension Costs
The Impact of Longevity Increases Under Different Reform Proposals
The Impact of Increased Disability Incidence
Summary
Endnotes
References
About the Author
About the Series

List of Tables

Table 1.	Projected Social Security Cost Rates and Key Components
Table 2.	Impact of Alternative Demographic Assumptions on Cost Projections: Difference between the Present Value of Income and the Present Value of Costs by 25-Year Subperiod under Different Demographic Assumptions (percentage of taxable payroll)
Table 3.	Impact of Increased Life Expectancy on Pension Financing

---

Introduction

A major shift is projected in the demographic structure of the population, leading, among other things, to serious financing problems for the nation's public pension and health financing systems. Whereas today each Social Security beneficiary is supported by some 3.4 workers, by 2030 each beneficiary will be supported by just 2.1 workers. In the official 1999 projections, scheduled benefits exceed available resources by the equivalent of 2.07 percent of taxable payroll over the next 75 years.

These demographic changes are generally understood to be linked both to the retirement of the baby boom generation and to other longer-term demographic developments. Less well understood, however, is the relative importance of the various projected demographic and economic developments and their implications for adjusting the current program. This brief explores the assumptions about future demographic and economic trends that underlie the current Social Security financing projections and notes some of their implications for Social Security reform.

The Baseline Pension Cost Projections

Social Security's costs are driven by demographic and economic trends that influence the relative size of the worker and beneficiary populations, the average benefit payment, and the average earnings against which payroll contributions are assessed. An easy way to summarize these impacts is to focus on the ratio of beneficiaries to workers and the ratio of the average benefit to the average wage. Taken together, these two ratios determine the fraction of the covered payroll that must be used to cover future Social Security spending.

Table 1 shows how these key relationships evolve over the next 75 years in the projections prepared for the 1999 Trustees' Report (Board of Trustees 1999). The 44 million beneficiaries and 148 million workers in 1998 resulted in a ratio of 0.297. Also that year, the average expenditure per beneficiary was $8,507 while the average earnings per worker was $23,582, for a ratio of 0.368. The product of the two ratios was 0.109, indicating that 1998 total Social Security spending amounted to some 10.9 percent of taxable payroll.

In the current projections, virtually all of the change in the program cost occurs as a result of changes in the second column, the ratio of beneficiaries to workers. The impact of the retirement of the baby boom generation appears in the changes occurring between now and 2030, the year the last boomers turn 65. During that time, the ratio of beneficiaries to workers rises by 60 percent, from 0.297 to 0.477, producing essentially the same percentage increase in costs as a percentage of payroll.

The retirement of the baby boomers is not a permanent phenomenon, however. As the boomers die off after 2030, the smaller successor cohorts will replace them on the benefit rolls. Yet, the projections assume that the ratio of beneficiaries to workers does not decrease with the passing of the baby boom generation. The ratio continues to drift upward, to 49.9 beneficiaries per 100 workers in 2050 and to 54.4 beneficiaries per 100 workers in 2075, and the ratio of program costs to taxable payroll continues to drift upward also. This projected long-term increase in the beneficiary-to-worker ratio results from assumptions about future trends in birth rates, death rates, and disability incidence that transcend the retirement of the baby boom generation.

Key Demographic and Economic Assumptions

Each year's Trustees' Report also includes an analysis of the independent impact of a number of key demographic and economic assumptions. The data reported in table 2 are based on this sensitivity analysis.

The first row in table 2 shows the difference between the present value of projected receipts and the present value of projected outlays, expressed as a percentage of projected taxable payroll, under the baseline intermediate projections. Separate calculations are shown for each of the 25-year periods and for the 75-year period as a whole. Revenues are projected to exceed outlays over the first 25 years, but fall increasingly short of outlays over the subsequent two 25-year periods. Thus, the baseline projection shows a surplus of 0.28 percent of payroll in the first subperiod, followed by deficits of 4.67 percent and 5.71 percent in the two subsequent subperiods. Over the entire 75-year period, the present value of projected outlays exceeds the present value of projected revenues by 2.45 percent of payroll.¹ Note, however, that the revenue-expenditure gap continues to grow throughout the projection period, so that the gap in the final years is more than twice as large as the 75-year average indicates.

The second and third rows show the separate impact of two demographic assumptions that play a major role in the current long-range estimates and that, if true, have major implications for the current pension debate. The second row shows how these projections would look if the life expectancy of the aged is assumed to remain constant at today's level; the third row shows the additional impact of assuming that the disability incidence rates also remain essentially constant at today's rates. A comparison of the first and second rows shows that almost half of the projected long-range deficit can be traced to the assumption that the aged will live longer in the future than they do today. The baseline assumes that mortality rates among the aged will drop by some 34 percent over the next 75 years. If life expectancy did not change in the future, the 75-year deficit would fall from 2.45 percent of payroll to 1.28 percent of payroll.²

The third row shows that another 10 percent of the projected long-range deficit results from the assumption that disability incidence rates will rise. The baseline assumes that disability incidence rates will rise by about 25 percent for men and by about 47 percent for women over the next 75 years. If disability rates among men do not increase and those for women rise only modestly, long-range costs will be correspondingly lower. The combination of a constant life expectancy for the aged and nearly constant disability incidence rates shrinks the 75-year average deficit by almost 60 percent.

Once the impact of increased disability and longevity has been accounted for, most of the rest of the projected deficit can be attributed to one of two additional assumptions. The first is that future fertility rates will fall to an average of 1.9 children per woman in the future. The second is that future wage increases will average just 0.9 percent above future price increases. Both involve assuming that the next 75 years will be less favorable to Social Security financing than the previous 75 years, though not necessarily less favorable than the most recent 25 years.

In combination with the two changes noted previously, assuming that fertility rises to 2.2 percent reduces the remaining projected long-range deficit by one-third, to 0.65 percent of payroll. In combination with the other two changes, assuming that wage increases average 1.4 percentage points above the inflation rate removes half of the remaining gap in the final subperiod and reduces the 75-year average deficit to 0.39 percent of payroll.

Life Expectancy and Pension Costs

Since the projected increase in life expectancy among the aged is such an important contributor to the projected long-range deficit, it is appropriate to focus more closely on how changes in life expectancy affect pension programs. On average, beneficiaries will live to collect more monthly pension checks, and if no other changes are made, aggregate pension expenditures will increase. The increase in life expectancy can be accommodated by (1) increasing the revenues into the system, (2) reducing the monthly benefit paid to each beneficiary, or (3) forcing beneficiaries to delay taking benefits long enough to offset the increase in life expectancy.

Two of the basic assumptions underlying the 1999 financing projections and their implications for pension benefits are shown in table 3: that male life expectancy at age 65 will rise from 16.0 years for those turning 65 in 1995 to 18.4 years for those turning 65 in 2050, and that female life expectancy will rise from 19.5 years to 21.9 years in the same period.³

The second row of table 3 shows the annual annuity that could be purchased for $100,000 by each of the age and gender cohorts if all transactions costs are ignored, and the third and fourth rows show the percentage reduction in annuities directly attributable to longer life spans. Basically, by 2050 increased life spans for men will have caused a 23 percent decline in annuity benefits relative to the 1965 cohort and an 11 percent decline relative to the 1995 cohort. Because of their longer life expectancies, the annuities that women can buy are significantly smaller than those that men can buy, but the decline is projected to be somewhat less owing to slower growth in female life expectancy.

These calculations illustrate the approximate magnitude of two of the three possible adjustments to longer life expectancies. Mortality reduction of the size projected to occur between the 1900 and 1985 birth cohorts can be offset by reductions in the range of 17 to 20 percent in monthly benefits, the amount by which the annuity that can be purchased with a given lump sum has declined. Alternatively, monthly benefits can be maintained if the revenues are increased by enough to allow a 20 to 25 percent increase in the amount accumulated at retirement.

The third possible adjustment is to increase the age at which benefits become available. As a first approximation, both the monthly benefit and the contribution rate can be held constant in the face of rising life expectancy if the retirement age is adjusted to maintain a constant ratio of years spent in retirement to years spent working. Under the current mortality projections, the retirement age must be raised to about 69.5 if men reaching 65 in 2050 are to have the same ratio of retirement years to working years as those reaching 65 in 1965; the equivalent retirement age for females turning 65 in 2050 would be roughly age 69.

The Impact of Longevity Increases Under Different Reform Proposals

Any refinancing plan that balances projected revenues and expenditures must deal either explicitly or implicitly with the projected longevity increase. Reform proposals differ as to whether the essential adjustment is through forcing a delay in retirement dates, reducing monthly benefits, or increasing revenues.

Reducing Monthly Benefits

Although several financing proposals have included explicit reductions in monthly benefits, such provisions are not very common. Where they are found, moreover, they tend to be coupled with proposals to institute individual accounts that the proposals' authors believe will help to offset the impact of the reduction.

Raising the Normal Retirement Age

At present, age 65 is defined as the normal retirement age, but actuarially reduced benefits are available as early as age 62. The normal retirement age is scheduled to rise gradually to age 67 in the first quarter of the 21st century. Benefits will still be available at age 62, but the reduction from full-rate benefits for those retiring at 62 will be greater.

Several reform proposals include further increases in the normal retirement age as a way of dealing with the aging of the population. Whether such proposals actually cause people to delay retirement is likely to depend, however, on whether parallel changes are made in the age of first eligibility. Proposals to raise the age of first eligibility are somewhat less common than proposals to raise the age for full benefits.

At present, some 57 percent of newly retiring men and 63 percent of newly retiring women file for their benefits at age 62, and roughly 80 percent of each gender files before reaching age 65 (Social Security Administration 1998).⁴ Research on retirement behavior suggests that a modest reduction in the benefits available to those first filing at any given age is likely to have only a minor impact on retiree behavior. Results suggest that a 10 percent reduction in benefits would lead to a one-to-five month delay, on average, in the date at which benefits are first taken (Magnussen 1996). If the normal retirement age is raised without an adjustment to the age at which benefits are first drawn, the majority of the adjustment will most likely occur through benefit reductions rather than postponed retirement.

On the other hand, if an increase in the retirement age includes increases in both the normal retirement age and the age at which benefits are first drawn, a substantial impact can be expected on the average age at which people file for benefits. Given the large fraction of the population taking benefits before age 65, an increase in the age of initial eligibility will necessarily lead to a delay in benefit receipt for the majority of the beneficiary population. Moreover, some fraction of the people forced to delay filing for their retirement benefits will continue to work, increasing payroll tax revenues for the program.

Increasing Revenues

Most reform proposals include at least some increase in revenues available to the pension system, implicitly using the higher revenues as part of the strategy for adjusting to longer life spans. Some of these proposals involve generating additional payroll tax revenue through such changes as extending coverage to the state and local government employees who are not now covered, increasing the fraction of earnings that are taxable under Social Security, or increasing the payroll tax rate. Others involve increasing the use of revenue sources other than the payroll tax to finance future retirement benefits. The most common proposals involve changing the investment strategy to increase returns or using a portion of the currently projected general fund surplus to increase the portfolio of financial assets held for retirement purposes.

Those favoring the use of increased investment returns to help finance future Social Security benefits differ as to whether this result is best achieved by creating a new set of individual accounts or by changing the investment policy of the current central fund. Regardless of the approach chosen, however, the implicit assumption underlying the proposals appears to be that the public would rather see the fraction of the gross domestic product (GDP) devoted to the Social Security program rise than see benefits cut or retirement ages increased, as long as payroll taxes are not raised.

Accelerating Economic Growth

Some advocates of increasing the pool of investment funds available for financing future benefits hope the approach can be implemented in a way that increases the national savings rate, leading to faster economic growth and lessening the burden of financing future retirement benefits. Estimates of the impact of higher savings on GDP growth and of faster growth on Social Security costs suggest, however, that faster growth is unlikely to provide more than a partial offset.

Table 2 shows the impact if the growth of real wages rises by 0.5 percent per year, an increase that after 75 years would cause real wages to be some 45 percent higher than under the base case. Growth of this magnitude is estimated to reduce costs by about 1 percent of taxable payroll in the third 25-year period and by 0.6 percent of payroll over the entire 75-year period. By comparison, several recent analyses conclude that replacing the current Social Security program with a fully funded alternative might lead eventually to an increase in GDP of between 4 and 10 percent, a far more modest impact than that assumed in table 2 (see, for example, Feldstein 1996; Hviding and Mérette 1998).

Faster economic growth would have a more dramatic impact on future costs if the retirement system were organized so that rising real wages did not also produce rising real retirement benefit levels. In that case, an acceleration in growth that produced a 10 percent increase in real wage levels would cause a 9 percent reduction in the ratio of benefits to wages. The information in table 1 suggests that this would be enough to reduce the 2075 cost rate of 19.9 percent of payroll to 18.1 percent.

Switching to a Defined Contribution Approach

One of the most contentious aspects of the current Social Security reform debate is whether to shift all or part of the responsibility for financing future retirement benefits from the current defined benefit approach to a defined contribution approach. This debate is usually phrased in terms of social philosophy, distributional impact, and predictions about the behavior of future politicians. The issue also has important implications for how retirement benefits will be adjusted as life spans continue to increase, however.

If life expectancies continue to improve (or the improvements occur more rapidly than is assumed in the current projections), the kind of revenue and benefit changes now being discussed will eventually prove insufficient, triggering the need for further program adjustments. The nature of these further adjustments will depend critically on whether by that time the program is operating on a defined contribution basis or a defined benefit basis.

If the program is shifted to a defined contribution basis, the volume of assets accumulated as of the retirement age will remain constant as the life expectancy at that age increases. The most probable adjustment to increased longevity, therefore, will be made through a reduction in monthly retirement income.

If the program continues under a defined benefit basis, further increases in longevity will produce another funding imbalance, leading to another round of debate about how to deal with rising costs. The most likely outcome of such a debate is similar to the likely outcome of the current debate: some combination of increased revenues and lowered monthly benefits.

The Impact of Increased Disability Incidence

The current financing debate has all but ignored the impact of projected increases in disability incidence. Several proposals advanced in the 1995-96 time frame would have reduced disability benefits, but largely as a by-product of reducing retirement benefits. Other proposals implicitly seek to cover the increased cost through revenue increases.

Summary

The deficit currently projected for Social Security is primarily the result of assumptions about future demographic and economic developments. By far the most important of these assumptions is that life expectancy will continue to increase.

Pension programs can adjust to increased life spans in one of three ways: through lowering retirement benefits, forcing delays in the age at which benefits begin, or increasing the revenues used to support the system. In the case of our national pension program, revenue increases mean increasing the share of GDP that is accounted for by retirement payments to the aged.

For a time, many Social Security reform proposals contemplated some delay in benefit entitlement or reduction in monthly benefits or both. More recently, however, the emphasis has been on methods to generate additional revenues for the program. At the same time, a separate debate involves whether to shift all or part of the program to a defined contribution basis.

If the program is shifted partially or entirely to a defined contribution basis, a future debate about how to adjust to further longevity increases may well have a different outcome. Under the defined contribution approach, benefit reductions are the more likely response to longevity increases.

---

Endnotes

References

About the Author

Lawrence H. Thompson is a senior fellow at the Urban Institute. He has served as principal deputy commissioner and chief operating officer of the Social Security Administration, assistant comptroller general of the United States, and chief economist of the U.S. General Accounting Office. He is president of the National Academy of Social Insurance.

About the Series

The Retirement Project is a multiyear research effort that will address how current and proposed retirement policies, demographic trends, and private-sector practices affect the well-being of older individuals, the economy, and government budgets. The project is made possible by a generous grant from the Andrew A. Mellon Foundation.

---

The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.

Usage, posting and reprint of materials on the UI web site:

Most publications may be downloaded free of charge from the web site in PDF format. This information may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required.

Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact paffairs@urban.org.

If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.