1994-96 Advisory Council Report
1994-1996 ADVISORY COUNCIL ON SOCIAL SECURITYEXECUTIVE SUMMARY OF REPORT BY TECHNICAL PANEL ON ASSUMPTIONS AND METHODS
1994-1996 Advisory Council on Social Security Technical Panel on Assumptions and Methods
MEMBERS OF PANEL
Howard Young, University of Michigan (Chair )
Barry Allen, Combined Insurance Company of America
Eileen Crimmins, Andrus Gerontology Center, University of Southern California
David Cutler, Department of Economics, Harvard University
Martin Holmer, HR&A
Diane Macunovich, Department of Economics, Williams College
Robert Myers, Actuarial Consultant
Samuel Preston, Population Studies Center, University of Pennsylvania
Eugene Steuerle, Senior Fellow, The Urban Institute
Michael Sze, Partner, Hewitt Associates
Kathleen Utgoff, Groom & Nordberg
Larry Wiltse, Consultant Actuary, Buck Consultants, Inc.
Barbara Wolfe, Professor of Economics, Preventive Medicine and
Public Affairs, University of Wisconsin - Madison
1994-1996 Advisory Council on Social Security
Assumptions and Methods Panel
The Panel's major conclusions are:
The "intermediate" projection of the Trustees Report for the Old-Age, Survivors, and Disability Insurance (OASDI) program provide a reasonable evaluation of the financial status. Although the Panel suggests that modifications be considered in various specific assumptions, the overall effect of those suggestions would not significantly change the financial status evaluation.
There should be evolutionary implementation of procedures to indicate more adequately the uncertainties involved in the projections. Even though such uncertainties are unavoidable, stochastic analysis should be used to examine more explicitly the probabilities of alternative projections. It is emphasized that there should be an extended period during which the new procedures would supplement, rather than replace, the current methods of considering high-cost and low-cost projections and individual assumption sensitivity analysis.
Evaluation of the long-range financial status should put less emphasis on the "75-year actuarial balance" and the "test of long-range close actuarial balance." Prior to enactment of legislation reforming the program, primary emphasis should be on the projected date the Trust Fund Ratio would fall below 100 percent; when definitive legislative revisions are adopted, subsequent long-range evaluation should compare up-dated projections with the intended results of the legislation.
There should be a substantial expansion of SSA's resources and its interaction with experts in related areas; increased recognition should be given to the inter-relationships between OASDI and many public and private programs as well as other aspects of the economy. Social Security Administration (SSA) staff does high-quality work, but is relatively small and works with inadequate resources. In addition to internal expansion, there should be greater use of outside consultants and contractual research; periodic comprehensive review by technical panels should be supplemented by ongoing arrangements for advice on specific matters.
Summaries of Individual Sections of the Report
Uncertainty of Projections
The Panel recommends evolutionary implementation of stochastic analysis procedures for presenting and evaluating the uncertainty in the OASDI projections. The current methods of considering high-cost and low-cost projections, and sensitivity analysis of individual variables, do not provide satisfactory indicators of the range of results and related probabilities that should be given consideration in evaluating program status.
It is emphasized that uncertainty about projections cannot be avoided, and that they must be based on assumptions (about the mean and variance of specific variables, and the correlations between them); in particular, there still will be the need to decide which past experience is most relevant. Nevertheless, such assumptions and the probability of resulting projections can be more explicitly examined than under current procedures.
Some interim arrangements are suggested for new procedures to supplement, rather than replace, the current methods of considering high-cost and low-cost projections and individual assumption sensitivity analysis. Illustrations of the stochastic analysis technique are provided in the appendices and are summarized later in this report.
This section discusses Mortality, Fertility, Marriage and Divorce, and Immigration. Emphasis is on the assumptions used for the "intermediate" projection. Although the Panel suggests that modifications be considered in various specific assumptions, the overall effect of those suggestions (including the economic assumptions discussed in the next section) would not significantly change the financial evaluation provided by the intermediate projection in the Trustees Report.
While the Panel has offered its best guesses as to the path of each demographic assumption for the 75-year period required for SSA projections, it is important to note the higher level of uncertainty in the latter part of the period.
Alternative II (intermediate cost) projections should more closely reflect long-run past experience. The current Alternative II assumption is for a lower rate of mortality improvement than has been experienced in the near-term (20-year) or long-term (90-year) past: such a decrease in the rate of mortality declines appears unwarranted. A mid-range projection that reflected continued mortality declines at the level experienced over the past century would be more appropriate. The Panel
recommends that the average rate of decline in age-specific death rates observed over the period 1900-1989 be reflected in Alternative II year-to-year projected mortality changes beginning in about 20 years. The Panel is not recommending a change in the procedures used for the earlier years, but the above-stated change would imply faster declines during this period, as well, because they grade into faster "ultimate" rates of decline.
Alternative methods of projection should be investigated. Cause-specific projections tend to produce conservative projections (that is, projections with slow mortality declines) because slowly declining causes become more prominent. Cause-specific projections also ignore the tendency for medical research and health intervention efforts to be targeted at diseases that are relatively more prominent. The use of relational models that impose some plausible age-pattern of mortality change should also be investigated. Also there should be further investigation of the financial impact of alternative patterns of age-sex improvement factors.
The Panel believes that fertility rates in the near future could be relatively volatile, and that the SSA should continue to monitor trends -- especially those among the younger age groups, to determine possible effects of birth-cohort size on fertility timing, and among baby-boom cohorts to identify trends in completed family size. In the meantime, the Panel recommends that the intermediate estimate of the long-term Total Fertility Rate be raised from its current level of 1.9 to 1.95. The high-cost and low-cost estimates of 1.6 and 2.2 are considered to provide an adequate range, in light of the stochastic effects of combining numerous demographic and economic assumptions. The Panel further recommends that an increase in the fertility rate should be assumed in the short-term, in the intermediate- and low-cost assumptions, before the long-term levels are reached.
Marriage and Divorce
The Panel recommends that the intermediate estimate of marriage rates should be raised from the current age-adjusted central rate of 5,730 to 6,000 per 100,000 unmarried of each sex, and that the estimate of divorce rates should be lowered from the current age-adjusted central rate of 2,140 to 2,000 per 100,000 married couples.
With regard to the high-cost and low-cost estimates, the Panel believes that the range provided by the assumptions used in the current Trustees Report is adequate. Consideration should be given to the anomaly created by combining low marriage rates and high divorce rates, however, with high fertility rates in the low-cost estimate (Alternate I), and vice versa in the high-cost estimate (Alternative III). In addition, current high levels of labor force participation, even among married women, suggest that the assumed link between high marriage rates and high OASDI auxiliary benefits may be outdated.
The Panel recommends no change in the current procedures used to make immigration assumptions. The procedures used by the SSA actuaries to update the assumptions seem appropriate. The level of the Alternative II projection seems appropriate.
This section discusses Real Wage Growth, Real Interest Rates, Inflation, and Unemployment assumptions, with emphasis on the intermediate projection. Although the Panel suggests that modifications be considered in various specific assumptions, the overall effect of those suggestions (including the demographic assumptions discussed in the preceding section) would not significantly change the financial evaluation provided by the intermediate projection in the Trustees Report.
The Panel split in its recommendations about assumptions for ultimate annual Real Wage Growth (RWG) and Real Interest (RI) rates. Half the Panel recommends RWG of 0.8 percent and RI of
2.8 percent; the other half recommends continued use of the assumptions used in the current Trustees Report: RWG of 1.0 percent and RI of 2.3 percent.
The Panel does not recommend any change in the ultimate Inflation (4 percent) or Unemployment (6 percent) assumptions.
Regarding short-range assumptions, the Panel recommends that considerable weight be given to the forecasts in the budget submissions of the Administration. However, a procedure for use when the long-range assumption for a variable is significantly different from actual recent experience is suggested.
The Panel does not recommend any explicit adjustment in assumptions attributable to possible changes in measurement of the Consumer Price Indices; implicit allowance for improved measurement procedures is reflected in the conclusions about each variable. It is emphasized that modifications to benefit formulas (for example, CPI minus 1 percent) have a result different from that of measurement changes; arbitrary adjustments would produce changes in the real benefits of the program.
Suggestions for future research are indicated; these should be considered in conjunction with the subsequent section on Research and Other Matters.
The Panel recommends periodic updating of the age-sex matrices used to project disability incidence and termination by recovery rates. In addition, use of different matrices for major
categories of disability should be considered. Analysis and projection of the factors, which reflect the overall levels of these rates, should give explicit recognition to the effect of unemployment and of claims administration practices.
Assumptions Regarding Retirement Age Under Current Law
The Panel recommends that the assumed pattern of retirement ages and the related benefits, and the sensitivity of the cost rates to such assumptions, be studied further and that detailed results be made available for review.
Presentation of Long-Term Status of the Trust Funds
Evaluation of the long-range financial status should put less emphasis on the "75-year actuarial balance" and the "test of long-range close actuarial balance."
The Panel does not recommend any change in the 75-year projection period, or in the concept of 100 percent Trust Fund Ratio (TFR) as an adequate contingency reserve.
The 75-year actuarial balance is an overall measure of changes in financial status, but should be less emphasized as a basis for evaluating the status of program or for designing reform proposals (especially if a substantial trust fund is to be accumulated).
The Panel suggests some revision in the presentation of annual balance projections (a different concept from the actuarial balance) and the treatment of trust fund interest.
Prior to enactment of legislation reforming the program, primary emphasis should be on the projected date the Trust Fund Ratio would fall below 100 percent.
When definitive legislative revisions are adopted, subsequent long-range evaluation should compare up-dated projections with the intended results of the legislation; based on the most frequently discussed proposals, such evaluation should consider whether the 75-year actuarial balance tends to deteriorate as the projection period moves forward; also whether the pattern of annual balances or TFR trend line departs significantly from legislative intent, or if the latter shows an apparent lack of stability beyond the 75-year valuation period. The Panel notes there is an important distinction between the financial adequacy tests appropriate for the Trustees Reports versus considerations for satisfactory legislative action.
Suggestions are also provided regarding indicators of future affordability of the program.
Research and Other Matters
The Panel recommends a substantial expansion of SSA's research capabilities, using additional in-house resources as well as outside consultants and contractual research. The prior (1990-91) Panel suggested an extensive list of research topics, many of which are still relevant; therefore, just a few
topics are highlighted in this report, but the important question of methodology is also discussed. The gradual erosion of support for the Office of the Actuary and the Office of Research and Statistics, in particular, pose fundamental problems to the system as a whole. These Offices operate with only a very small fraction of the resources that would be made available in private insurance companies and actuarial consulting firms to study matters of importance to clientele. Adequate funding and organizational support for these Offices is vital to the long-range status and effectiveness of the Social Security Administration.
To ensure periodic review and that the most appropriate assumptions and techniques are used for projections of the operations of the trust funds and other policy purposes, the Panel recommends that:
- Technical panels be appointed periodically (at least once every 5 years) to conduct comprehensive reviews of the assumptions and methods;
- An ongoing advisory committee of experts be established to provide, on an as-needed basis, advice on specific matters; and
- The SSA develop procedures to enable the staff easily to contract for extramural research and expert analysis to supplement ongoing staff activities.