1994-96 Advisory Council

APPENDIX II:

COMPARISON OF FINANCIAL EFFECTS OF ADVISORY COUNCIL PLANS TO MODIFY THE OASDI PROGRAM By Stephen C. Goss, Deputy Chief Actuary
Social Security Administration The estimates and analysis presented in this and the following appendix represent the collective efforts of all of the members of the long-range division of the Office of the Chief Actuary since the council began in 1994. Many new methodologies have been developed for this analysis. It is our hope that this analysis, and the use of the new methodologies in the future, will assist in the process of developing and assessing alternative approaches to restoring long-range financial soundness for the Old-Age, Survivors, and Disability Insurance (OASDI) program.

The Advisory Council has developed three plans that would eliminate the long-range (75-year) OASDI actuarial deficit under the intermediate assumptions of the 1995 Trustees Report. Each of the plans would also result in a stable trust fund ratio (i.e., a constant ratio of trust fund assets to annual program cost) at the end of the long-range period, meaning that the long-range actuarial balance of the program would no longer tend to deteriorate with the passage of time. Estimates for the Council proposals are based on modified intermediate assumptions of the 1995 Trustees Report; they have not been updated to reflect the 1996 Trustees Report assumptions because the long-range financial status of the OASDI program changed very little.

One of the three plans is designed to increase tax revenue and to increase the yield on trust-fund assets through partial investment in stocks so that benefits would be reduced very little from the level specified in present law. This plan is thus named the Maintain Benefits (MB) plan. The other two Council plans include provisions for individual savings accounts funded through mandatory payroll tax contributions. The Personal Security Account (PSA) plan would redirect 5 percentage points of the 12.4 percent payroll tax to these PSAs, financing a reduced OASDI system with the remaining payroll tax plus transition taxes and borrowing from the United States Treasury. The Individual Account (IA) plan would reduce benefits and increase tax revenues for the OASDI program so that the 12.4 percent payroll tax would not need to be increased, and would add an additional 1.6 percent payroll contribution to fund the individual accounts.

This memorandum provides a brief description of the components of these plans, a brief analysis of some of the key assumptions made in developing the plans, and estimates of the long-range financial implications of their enactment. Analysis of the effects on money's worth ratios, internal rates of return, and benefit replacement rates for current and future beneficiaries, as well as the estimated effects on the unified budget of the United States are also included. Assumptions made for these proposals are of particular importance because all three packages depend upon changes in the Consumer Price Index (CPI) that have been implemented by the Bureau of Labor Statistics (BLS). In addition, all three plans depend upon certain assumptions about the performance of stock investments in the future.

Because the Council has requested that proposals be developed and estimates be made reflecting modifications to the CPI announced by the BLS earlier this year, the baseline assumptions differ from those specified as the intermediate alternative in the 1995 Trustees Report. These changes in assumptions result in a reduction in the estimated long-range actuarial deficit of 0.31 percent of taxable payroll, from a deficit of 2.17 percent of taxable payroll to 1.86 percent of taxable payroll. Assumptions other than the modified growth rate for CPI are the same as used in the intermediate alternative for the 1995 Trustees Report. In addition, assumed yield rates for stocks were specified by the council for use in these calculations. No assumption for stock yields was included in the Trustees Report. See the section on critical assumptions (below) for more detail.

Plan Specifications

Tables 1.MB, 1.IA, and 1.PSA provide a listing of the component provisions of the three Advisory Council options. These tables also provide the estimated change in the OASDI actuarial balance for each provision (generally separately, as if applied alone to present law) and for the plan as a whole.

Maintain Benefits Plan (MB)

This plan increases revenue to the OASDI program by covering all newly hired State and local government employees after 1997, changing the provisions for the taxation of benefits, investing 40 percent of the OASI and DI Trust Funds in equities by 2015, and increasing the payroll tax rate by 0.8 percent each for employees and employers beginning in the year 2045. The provision to cover State and local government employees would result in a partially offsetting increase in benefits. Benefit levels would be reduced about 3 percent on average by increasing the number of benefit computation years from 35 to 38 by 1999. The changes in benefit taxation would (a) tax benefits in a manner consistent with that currently prescribed by the Internal Revenue Service for private and government-employee contributory defined-benefit pensions, (b) phase out the current thresholds ($25,000 and $32,000) between 1998 and 2007, and (c) gradually redirect revenue from benefit taxation that is transferred to the HI Trust Fund to the OASI and DI Trust Funds, with 10 percent redirected in 2010, 20 percent in 2011, ..., and 100 percent redirected in 2019 and later.

A variation of the MB plan, called MB', would replace the provision to increase the benefit computation period with a provision for an additional increase in the payroll tax rate of 0.15 percent for employees and employers, each in 1998. This variation would result in no change in OASDI benefit levels from those specified in present law. Estimated financial effects of the MB' plan would be very similar to the estimates presented in the balance of this appendix for the MB plan.

Individual Account Plan (IA)

This plan increases revenue by covering newly hired State and local government employees after 1997 and changing the taxation of benefits with (a) and (b) described above. Reductions in the cost of the OASDI program are achieved by increasing the Normal Retirement Age (NRA) by 2 months per year beginning in 2000, reaching 67 in 2011 and indexing thereafter (see section on indexing retirement age below), modifying the PIA formula by reducing the latter two factors in the formula gradually from 32 and 15 percent to about 22.4 and 10.5 percent, respectively, between 1998 and 2030, increasing the number of benefit computation years for the benefit formula from 35 to 38 by 1999, and reducing spouse benefits from 50 to 33 percent of PIA between 2000 and 2016. Paying surviving spouse benefits equal to 75 percent of the amount payable if both spouses were still alive would increase benefits for many two-earner couples (this provision would be phased in gradually between 1998 and 2037). The individual account would be based on the accumulation of an mandatory additional employee contribution of 1.6 percent of taxable payroll for all workers in 1998 and later. The individual account accumulation would be required to be annuitized upon retirement, with annuity payments being tax exempt.

The provision to modify the 32-percent and 15-percent PIA factors would multiply these factors successively by 0.995 for years 1998 through 2011 and by 0.985 for years 2012 through 2030. The slower rate of reduction through 2011 was specified to reduce the possibility of creating "notches" (the condition where benefits for newly eligible beneficiaries are lower than benefits for similar workers who were eligible in one year earlier).

Personal Security Account Plan (PSA)

This plan reduces the cost of the OASDI program by increasing the NRA to 67 by 2011, indexing the NRA thereafter (see below), increasing the Earliest Eligibility Age (EEA) for retirement benefits to 64 by 2011 and indexing up to a level of 65 thereafter, reducing from 100 percent to 70 percent the percentage of PIA payable to disabled worker beneficiaries as the NRA rises, gradually replacing the current PIA formula with a flat benefit for retirees, and finally by setting spouse benefits equal to one half of the flat benefit level. Surviving spouse flat benefits would be modified as in the IA plan (75 percent of the sum of the flat benefits payable to the couple if they were both alive). Phasing out the earnings test at NRA would increase benefits to some workers between NRA and 70, but lower their benefits thereafter due to the loss of the delayed retirement credit.

The flat benefit ($410 in 1996 for 35-year career workers) would be phased in; the flat benefit plus a past service credit based on earnings before 1998 would be provided for retired workers who were between 25 and 54 in 1998, and a flat benefit only for retired workers who were under 25 in 1998. Retired workers aged 55 and over in 1998, disabled workers prior to reaching the age of conversion to retirement status (65), and survivors other than aged surviving spouses would receive benefits based on the present law benefit formula. Children and aged spouses of retired workers who are eligible for a flat benefit would be eligible to receive one-half of the full flat benefit, subject to family maximum benefit limits.

The level of the full flat benefit ($410 in 1996) would be indexed after 1996 and up to the year of initial benefit eligibility based on increases in the average wage level. After initial benefit eligibility, the flat benefit would increase at the rate of increase in the CPI-W through the cost-of-living adjustment (COLA). Ultimately, 50 percent of the full flat benefit would be payable to workers who worked for only 10 years, with the percentage increasing by 2 for each additional work year up to 35. A work year for this purpose is attained by earning 4 quarters of coverage.

As mentioned above, retirement benefits for workers aged 25 through 54 in 1998 would be a combination of a past service credit for years of work prior to 1998 and the flat benefit/individual account accumulation for work after 1997. The past service credit is intended to reflect the proportion of the present-law retirement benefit that was earned during potential work years prior to 1998. The flat benefit, along with the accumulated personal security account, is intended to provide a retirement benefit reflecting the proportion of potential work years after 1997. Potential work years are specified for the sake of simplicity to be the years between attaining age 22 and attaining age 62. Thus, the flat benefit described above would be multiplied by a "proportion factor" for those aged 25-54 in 1998. The proportion factor would be the number of years starting with 1998 up to (but not including) the year in which age 62 is attained. For workers aged 25-54 in 1998, the ultimate 10-work-year and 35-work-year requirements for the flat benefit would be reduced to reflect the proportion of potential work years that fall after 1997.

Past service credits (PSCs) for workers aged 25 through 54 at the beginning of 1998 would be computed in three steps. First, a present law PIA would be computed as if the worker had become entitled as a disabled worker at the beginning of 1998. (The special recent-work requirement for disability insured status would be waived for this consideration.) Second, this PIA would be indexed with the average wage up to the year of retired worker benefit eligibility. Third, the indexed PIA would be multiplied by a "proportion factor" reflecting the proportion of "potential work years" that passed prior to 1998. This proportion would be computed as the number of years after reaching age 22 up to 1997, divided by 40. The PSC would be indexed with the COLA after initial retirement eligibility, actuarially reduced for early retirement, increased by the delayed retirement credit, and recomputed based on additional work years after initial eligibility.

Disabled worker beneficiaries would receive a present law PIA up to age 65 (the age for conversion to retired worker status--this would not change under the PSA plan). However rather than paying the full PIA, as under current law, the PIA would be reduced by the same percentage as for a retired worker beneficiary becoming entitled at age 65 in the same year the disabled worker becomes entitled to disability benefits. (As the NRA increases, the percentage of full benefit payable to retired workers entitled at age 65 declines.) The reduction for disabled workers would not be allowed to exceed 30 percent (which would be attained the NRA reaches 70.) Accumulated assets in the individual account would not be available to disabled workers until they convert to retired worker status at age 65.

Disabled workers, upon attainment of age 65, would convert to retirement status and who were under age 55 in 1998 would receive a combination of a partial flat benefit, individual account accumulations, a partial continuation of their disability PIA benefit, and for those who were between ages 25 and 54 in 1998, a past service credit. The past service credit would be computed as described above, except that for those disabled before 1998, (1) the proportion factor would reflect potential work years between attainment of age 22 and entitlement to disability benefits and (2) the disability PIA would be the actual PIA payable at the beginning of 1998, indexed by the average wage thereafter to retirement eligibility. The disability PIA benefit would be continued at age-65 conversion but multiplied by a proportion factor of years of entitlement to disability benefits divided by 40. (If recovery from disability occurs before retirement eligibility, then the PIA at the time of recovery is indexed by the average wage level up to retirement eligibility.) The flat benefit would be computed reflecting the number of years between 1998 (age 22 if later) and age 62, excluding any years for which disability benefits were received.

Additional revenue for the OASDI program under the PSA plan would come from State and local coverage (as in the IA and MB plans).
Further modifications to revenue for the OASDI program are described in the remainder of this section.

The current 6.2 percent employee OASDI payroll tax (and the current 12.4 percent tax for net self-employment income) would be reduced by 5 percentage points for workers under 55 in 1998 with this contribution being redirected to fund the individual PSAs (those 55 and over would not participate). Distributions from the PSAs would be tax exempt.

The revenue loss to OASDI from redirecting 5 percentage points of the payroll tax to PSAs is much greater than the reduction in OASDI cost until about 2070. This additional cost through 2069 is referred to as the transition cost of the plan. This transition cost is financed on an average basis by a transition tax of 1.52 percent of payroll, which would be required for all covered earnings during the period 1998 through 2069. However, this transition tax would be less than needed through 2034 and more than needed between 2035 and 2069. As a result, additional revenue would be borrowed from the general fund of the U.S. Treasury in years 2002 through 2034, with repayment of all borrowed amounts with interest from the Trust Funds to the Treasury between 2035 and 2069. Table 2 provides an analysis of the borrowing and repayment.

Taxation of benefits under the PSA plan would be modified considerably. No more than 50 percent of present-law benefits payable to those 55 and older in 1998 (or to disabled workers and survivors in 1998 and later) would be subject to Federal income taxation. This would eliminate the transfer of revenue to the Medicare Hospital Insurance Trust Fund based on benefit taxation as of 1998. Up to 50 percent of past service credits would be taxable, flat benefits would be fully taxable, and PSA distributions would be tax exempt. As with the other plans, the $25,000 and $32,000 thresholds would be phased out between 1998 and 2007.

Additional Plans

Table 1A provides a brief listing of the provisions of the MB, IA, and PSA plans, along with a comparison of estimates of the cost and income under the plans. Table 1B provides the same information for four additional plans, PL PAYGO, MB PAYGO, MTR, and MB+, which are useful in the analysis of the financial implications of the Council plans.

PL PAYGO (present law benefits on a pay-as-you-go financing basis) retains the present law OASDI benefit structure, and maintains financial solvency by raising payroll tax rates as needed (beginning 2025) to keep the OASDI combined trust fund ratio from dropping below 100 percent after that point. The PL PAYGO, therefore, represents one possible policy option for restoring long-range solvency for the OASDI program through tax rate adjustments alone. Other possible theoretical approaches to this end would include, for example, raising the combined payroll tax rate immediately by 2.2 percent.

MB PAYGO (MB provisions without investment in equities, with PAYGO tax rates) includes the taxation of benefits, 38-year computation period, and State and local coverage provisions of the MB plan, along with pay-as-you-go tax rate increases as needed (in 2040, 2060, and 2070) to move to a steady 100 percent contingency reserve trust fund by 2070. This plan represents a variation of the PL PAYGO plan which allows more time before tax rate increases would be needed, and reduces the size of the increases.

The MTR (maintain tax rates) plan is the IA plan without the individual account and its supporting employee contribution of 1.6 percent of taxable payroll. This plan would primarily reduce benefit levels in order to reduce the cost of the OASDI program sufficiently so that sustained financial solvency would be achieved without any need to raise payroll tax rates. This plan may be viewed as the residual OASDI program that would exist, separate from the individual accounts, under the IA plan.

The LTR (lower tax rates) plan, which is not shown in table 1B, is the PSA plan without the personal security accounts or their supporting 5 percent payroll tax contribution. This plan represents the residual OASDI system, separate from the individual accounts, under the PSA plan.

The MB+ plan is similar to the MB plan but with three significant changes. First, the payroll tax increase in 2045 is replaced with an increase of 1.52 percent for the period 1998 through 2069 and borrowing from the General Fund of the U.S. Treasury with repayment by 2070, in the same amounts as under the PSA plan. Second, OASDI Trust Fund assets would be 50 percent invested in private equities by 2014 (approximately the proportion of PSA balances assumed to be invested in equities). Third, benefit levels would be increased by 28 percent between 1998 and 2037, using the additional revenue from the first two changes while providing for a stable trust fund ratio for the OASDI program at the end of the long-range period. This plan was developed to explore the implications of providing the same increases in revenues specified under the PSA plan for a government-operated defined benefit plan.

Total Contribution Rates Under the Plans

Table 1.C provides a comparison of contribution rates under the plans described in tables 1A and 1B. The MB+ plan, not shown, would have the same contribution rates as for the 5% PSA plan. Both payroll tax contributions and mandatory contributions to IAs and PSAs are included (except, of course for MTR and LTR, which omit the individual accounts). Borrowing under the PSA (and MB+) plan from the General Fund of the Treasury through 2034, and repayment through 2069 is not reflected in these rates.

Indexing Retirement Age

Both the IA and PSA plans call for indexing the normal retirement age (NRA) after an NRA of 67 is reached by schedule for those age 62 in 2011. Thereafter, the NRA would be increased as necessary in order to maintain a constant ratio of retirement years (i.e., life expectancy at NRA) to potential years of work (i.e., the number of years from age 20 to NRA).

Table 1.D illustrates how the principle of this constant ratio might be applied. The top section provides life expectancy at a variety of retirement ages for years 1940 through 2070, based on the intermediate assumptions of the 1995 Trustees Report. The middle section provides the ratio of retirement years to potential work years for each age in each year 1940 through 2070. The bottom section indicates the NRA series required to maintain a constant ratio (middle section) as compared with the ratio at age 65 in a series of base years (1940 ... 2030). That is, it shows the extent to which the NRA would need to be increased to keep up with life expectancy changes if initially set at age 65 in 1940, 1960, etc. This section indicates that the NRA must rise by about 5 months per decade, or about 1 month every 2 years, in order to maintain a constant ratio of retirement years to potential work years.

Table 1.E provides a comparison of the expected NRA and earliest eligibility age (EEA) patterns for the three council plans. The PSA plan would increase the EEA in tandem with NRA increases, until an EEA of 65 is attained.

Critical Assumptions in the Design of the Plans

Each of the plans depends upon reductions in the future growth rate in the CPI to result in benefits that grow more slowly after initial eligibility. The plans also depend upon a continued higher yield for equities than that for long-term U.S. securities. Finally, for the IA and PSA plans, the Council specified assumptions for the distribution of assets in individual accounts.

Consumer Price Index

The Council members have agreed to specify that the "baseline" for estimates for the proposals should include future reductions in the rate of growth in the consumer price index (CPI). The Bureau of Labor Statistics (BLS) has stated its intention to correct the "formula bias" in the CPI with changes made in January 1995 and July 1996. The change made in 1995 is expected to reduce measured growth in the CPI by about 0.11 percent per year on average. Changes for 1996 are estimated to further reduce growth by about 0.10 percent on average. On this basis the Council assumes a reduction in the December 1996 OASDI cost-of-living adjustment (COLA) of 0.10 percentage point and reductions on an average of 0.21 percentage point for December 1997 and later from the levels assumed in the 1995 Trustees Report.

There has been some debate about the extent to which changes in the CPI might influence future nominal interest rates and future growth in nominal wage levels. For the purpose of these estimates, it is assumed that nominal interest and wage levels would not be affected by the assumed changes in the CPI. This assumption appears to be consistent with the thinking of most economists, although the view is not universally held. Changes in the computation of the Gross Domestic Product price index by the Bureau of Economic Analysis (which were made late in the deliberations of the Council) and the relationship between this index and the CPI were not addressed by the Council.

Reducing the growth in the CPI without affecting the nominal levels of interest and wages has a significant effect on the financing of the OASDI program. The assumptions specified above reduce the projected long-range actuarial deficit by 0.31 percent of taxable payroll, i.e., by one seventh. This improvement in the deficit would be less if nominal interest and wage levels were assumed to be affected by the change in the CPI.

Yields on Stocks

Analysis for the Advisory Council by Joel M. Dickson of the Vanguard Group, Inc., indicates that real yields on stock investments have averaged about 7.0 percent for the period 1900 to 1995. The Council members agreed to specify this rate as the assumption to be used in estimating average future yields on stocks. This real yield for stocks exceeds the 2.3 percent ultimate real yield on long-term U.S. Government securities used in the intermediate assumptions of the 1995 Trustees Report by about 4.7 percentage points. However, the greater volatility of returns on stocks has been the subject of considerable discussion. Also, questions have been raised as to whether it is consistent to assume the same yield on stocks in the future as in the long-term past, when the Trustees intermediate assumptions include slower growth in both real GDP and labor productivity (real output per hour worked) than has occurred so far this century. On the other hand, increased national saving that might result from these plans could increase productivity and real GDP growth. Estimates provided in this appendix do not reflect the possible changes in net national saving and the possible effect on interest rates and economic growth of the plans.

Yields on the Trust Fund Assets

The MB plan specifies that over a 15-year period, beginning 2000, the composition of the OASDI Trust Funds would be gradually altered to ultimately retain about 40 percent of assets in stocks by the end of 2014. Assuming that stocks have yields about 4.7 percentage points higher than for U.S. securities and that transaction costs for trust fund investment in equities consume about 0.005 percent of assets annually, the ultimate trust fund real interest rate would be increased from 2.3 percent to 4.175 percent (0.4 x 7.0 + 0.6 x 2.3 - 0.005 = 4.175). Expected future variation in the yield on stocks increases the uncertainty of trust fund projections. (Note that in order to allow for comparability with Trustees Report values and with historical data, real yields discussed above do not reflect the Council's assumption for lower future growth in the CPI. Ultimate real yields used for analysis are 0.21 percentage point higher than shown, reflecting the lower assumed CPI growth).

The IA and PSA plans specify that trust fund assets would continue to be held solely in special issue United States securities.

Yields on Individual Accounts

Both the IA and PSA plans anticipate investment in stocks for the individual accounts, but not for the OASDI Trust Funds. For the purpose of money's worth analysis (both money's worth ratios and rates of return) and replacement rates, the average real yield on individual accounts has been estimated using the specified 4.7 percentage point difference between stock and U.S. security yields described above.

For the intermediate return, assets in the IA and PSA individual accounts are distributed consistent with 401(k) experience. The portion in equities is assumed to have a real return that is 4.7 percentage points higher than the return for the Trust Funds under present law. Annual administrative expenses of 0.105 and 1.00 percentage points are assumed for the IA and PSA plans, respectively. The administrative expense for the IA accounts is assumed to be largely the additional cost of maintaining individual records of the desired transactions and balances of each participant. IA funds are assumed to be collected by the U.S. government and invested collectively, thus incurring minimal transaction costs. PSA funds are assumed to be invested individually, thus incurring higher annual transaction costs.

The distribution of assets between stocks and U.S. securities, by age group, was based on analysis by Joel Dickson of the Vanguard Group, Inc. Distributions for the PSA plan are based on 401(k) plan experience. Distributions for the IA accounts reflect the 401(k) experience with a modification toward less stock investment as retirement age is approached, because of the requirement to annuitize the account balance upon retirement.

Ultimate Net Real Yield on Individual Accounts-- Intermediate Return
PSA--401(k)			IA--401(k) annuitized
Age Group	Percent in Equities	Admin Expense Factor	Net Real Yield	Percent in Equities	Admin Expense Factor	Net Real Yield
<40	55	1.00	3.885	55	0.105	4.780
40-49	52	1.00	3.744	50	0.105	4.545
50-59	48	1.00	3.556	40	0.105	4.075
60-69	43	1.00	3.321	20	0.105	3.135

Note that in order to allow for comparability with Trustees Report values and with historical data, real yields above do not reflect the Council's assumption for lower future growth in the CPI. Ultimate real yields used for analysis are 0.21 percentage points higher than shown, reflecting the lower assumed CPI growth.

Actual yields for individual accounts depend not only on the average, or even the fluctuating annual yields in the bond and equities markets, but also on individual choices of the distribution of assets and the timing of changes in this distribution. Thus, even in periods of good market performance, many individuals may "mistime" market fluctuations, or simply invest solely in U.S. securities or money market instruments. Such unlucky, or conservative, choices, along with administrative loads, may result in considerable variation in the actual yield experienced by different individuals.

In order to better understand the implications of these variations, and the sensitivity to interest rates, money's worth analysis for the IA and PSA plans has been done using three different assumed yields on the individual accounts: (1) the intermediate, or 401(k) yields shown in the table above, (2) low yields associated with conservative or poorly timed investment, or with poorer than assumed average yields on stocks (net real yields of 2.3 percent for IA accounts and 2.0 percent for PSAs), and (3) high yield associated with better performance in equities (net real yields as in the intermediate assumption but with an assumed 7.0 percentage point difference between U.S. securities--at 2.3 percent--and stocks--at 9.3 percent; plus lower administrative expenses--at 0.50 percentage points--for PSAs).

Note that in order to allow for comparability with Trustees Report values and with historical data, real yields above do not reflect the council's assumption for lower future growth in the CPI. Ultimate real yields used for analysis are 0.21 percent higher than shown, reflecting the lower assumed CPI growth).

OASDI Financial Operations under the Options

Table 1A provides a brief summary of the financial operations of the OASDI program (excluding income and outgo for individual accounts in the IA and PSA plans) under present law and each of the three Council plans. It should be noted that estimates for present law baseline differ from those for intermediate assumptions in the 1995 Trustees Report because of the modified CPI assumption specified by the Council.

Each of Advisory Council plans being considered would eliminate the long-range (75-year) actuarial deficit for the OASDI program. Each of the plans has also been designed with the intention to avoid deterioration in the actuarial balance as successive valuations are made for future Trustees Reports. Currently, the actuarial deficit rises by about 0.08 percent of payroll each year for new valuations due to the inclusion of an additional year with a large deficit in the long-range valuation. By designing plans that have roughly stable trust fund ratios at the end of the long-range period, each of these plans would eliminate this problem well into the future.

Maintain Benefits Plan (MB)

Under the MB plan, the OASDI combined Trust Fund would stabilize at around 450 percent of annual outgo by about 2015. Maintaining a relatively large fund balance permits reduced tax income when the rate of interest on the fund is greater than the growth in the cost of the program. (If the interest rate is equal to the rate of growth in the program costs, then retaining all interest in the fund will be just enough to maintain a stable trust fund ratio. But if the interest rate is higher, then the excess interest is available to help cover current costs, permitting lower taxes.) With a gradual (15-y ear) transition to investment of 40 percent of Trust Fund assets in equities, which are assumed to earn about 4.7 percent higher returns on average than do U.S. Government long-term securities, the ultimate real trust fund interest rate would be about 4.175 percent, or nearly 1.9 percent higher than the 2.3 percent real interest rate assumed in the intermediate assumptions of the 1995 Trustees Report.

As a result of the relatively high yield expected for stock investment, annual deficits (the extent to which program costs exceed tax income) of over 2 percent of payroll are sustainable after 2060 under the MB plan. This means that beginning in 2015, the trust funds would generally be expected not to be purchasing additional stock, but would be able to be sell shares, or at least cash dividends to help cover annual program costs.

Market fluctuations would, however, result in periods in which the yield on stocks would fall short of the yield on U.S. securities. In these periods, additional stock would need to be purchased in order to a maintain 40-percent share in stocks. Similarly, when stock yields are unusually high, more stock would be sold. This practice might provide a counter cyclical influence on the market.

Individual Accounts Plan (IA)

Under the 1.6% IA plan, trust fund assets are invested in U.S. Government securities, as under present law, at an assumed ultimate real interest rate of 2.3 percent. Relatively large trust fund balances around 4 times annual outgo by 2015 provide excess interest that permits annual deficits slightly over 0.5 percent of payroll due to the higher expected yield on these securities than would be needed to maintain a constant trust fund ratio.

Personal Security Accounts Plan (PSA)

The 5% PSA plan is designed to accumulate only a contingency reserve fund for OASDI of between 100 and 150 percent of annual outgo, with investment in U.S. Government long-term securities. Sustainable annual deficits for this plan are small because there is relatively little excess interest from the funds to apply to the current cost of OASDI benefits.

OASDI Program "Size"

The "size" of the OASDI program is indicated roughly by the cost rates (OASDI cost as a percentage of taxable payroll) in table 1A. By 2070, the cost rate under the MB plan is nearly 18 percent of payroll, only slightly lower than present law. The 1.6% IA plan results in an OASDI cost rate of about 14 percent, for a reduction of about 25 percent from present law. The 5% PSA plan reduces the OASDI cost rate for 2070 by about 55 percent, to 8.2 percent of taxable payroll. It must be noted, however, that for the IA and PSA plans, retirement benefits are also provided from distributions or annuities based on the accumulated individual account balances. Money's worth analysis, described below, provides a more complete measure of the intended protection under the plans.

Personal Security Account (PSA) Financing

Table 2.PSA provides estimates of the annual operations of both the OASDI trust funds and the aggregate PSA accounts under the 5% PSA plan. Annual operations for both OASDI and the PSAs reflect income entering and expenditures leaving the funds. Thus, internally compounding interest is not reflected in the "cash flow" net annual operations. Interest is, of course, reflected in accumulated end-of-year balances in the funds. Table 2.PSA presumes that PSA account holders will withdraw funds gradually, after retirement, with an average rate of disbursement similar to that provided under annuitization. Of course, the PSA plan does not require annuitization, so individuals might withdraw balances either faster or slower. Individuals with high marginal tax brackets might leave PSA balances intact as long as possible due to the favorable (tax exempt) tax policy specified for accumulations at disbursement.

Table 2.PSA specifically illustrates that in addition to the normal OASDI annual balance, which includes payroll taxes and benefit payments, (1) an additional payroll tax of 1.52 percent would be required over the period 1998-2069 to finance the transition costs associated with the 5% PSA plan, and (2) annual borrowing from the general fund of the Treasury, initially at a level equivalent to nearly 2 percent of payroll, would be required through about 2034. Between about 2035 and 2069, the additional 1.52 percent payroll tax would provide for repayment of the earlier borrowing from the general fund, with interest.

Table 2.PSA also indicates that aggregate PSA balances would amount to 172 percent of GDP by 2070. However, Table 2.PSA' indicates that aggregate PSA balances would amount to at most 114 percent of GDP by 2070 if all account holders decided to disburse their accounts at retirement in order to purchase life annuities. Actually, balances would be lower in this case because account holders would be assumed to hold less stocks as they approached retirement age (as assumed for the IA plan). Reality would likely fall between these two outcomes.

Individual Account (IA) Financing

Table 2.IA provides a similar analysis of accumulated aggregate balances in individual accounts under the 1.6% IA plan. As specified in the plan, all account balances are assumed to be disbursed at retirement to purchase life annuities.

Unified Budget Effects

Each of the Council plans would have significant effects on the United States Government unified budget deficits. First order estimates of these effects are presented in table UB (in current, or nominal, dollars) and in table UB' (as a percent of GDP). This analysis benefits from the advice of David Koitz of the Congressional Research service. The estimates are referred to as first order for several reasons: (1) no assumption is made regarding changes in Federal government spending or taxes other than the direct effects on the OASDI program and reductions in revenue to the Hospital Insurance (HI) program (from taxation of benefits in the MB and PSA plans), (2) no assumption is made regarding possible changes in aggregate national saving and investment, and thus possible effects on economic growth and taxable income. Because distributions from PSA (and possibly IA) accounts would be tax exempt, Federal tax losses would occur to the extent that contributions to the accounts were accompanied by reductions in other savings. If, in the case of the IA plan, employee contributions to the individual accounts were made tax exempt, with later distributions taxable, then additional reductions in Federal income tax revenue beyond those reflected in these tables would occur starting in 1998.

Columns 1 through 4 provide estimates of annual noninterest income less outgo (plus revenue to HI from taxation of OASDI benefits) for present law and each of the three Council plans. Columns 5 through 7 indicate the change from present law estimates for each plan. Columns 10 through 12 present estimates of additional U.S. Government borrowing from the public under each plan assuming that taxes and spending will be unchanged other than plan specifications. Adjustments needed to produce estimates of additional U.S. Government borrowing from the public include PSA borrowing from the Federal Government (column 8) and MB net equity purchase (column 9), which would be scored as an expenditure for budget purposes. Estimates of accumulated change in U.S. Government borrowing from the public including interest (columns 13 through 15) were developed in order to estimate interest on the accumulation (columns 16 through 18), the last component needed to estimate the effect on the annual unified budget balance (columns 19 through 21).

It should be noted that unified budget effects for each plan are compared to the estimated status under the present law OASDI program, assuming that shortfalls in program income relative to outgo (column 1) would be accommodated by borrowing from the general fund of the Treasury, and thus, indirectly from the public. Present law does not permit such borrowing. If present law were interpreted to mean that OASDI benefits would not be payable beyond funds available, then the magnitude of negative values in column 1 would be greatly reduced, with commensurate negative effects on columns 19 through 21.

The MB plan would have a modest negative effect on unified budget balances for years 2000 through 2014, when substantial purchases of stock are needed to gradually transfer 40 percent of the OASDI program assets into stocks. The magnitude of these negatives is small because the basic provisions of the MB plan (increasing revenue and reducing benefits) provide most of the funds needed for the purchase of stock. Thereafter, positive effects are substantial as revenue from the yield on stock holdings increases.

The IA plan has initially modest, but increasing positive effects on the unified budget balance because of the combination of benefit (expenditure) reductions and income (revenue) increases. The magnitude of the positive effect is somewhat larger than that for the MB plan throughout the next 35 years.

The PSA plan would create substantial unified budget deficits until 2030. Annual borrowing from the general fund of the Treasury through 2034 would more than offset improvements in OASDI (and HI) annual balances through 2018 (column 12), and additional interest on borrowing from the public (column 18) would result in worse unified budget balances through 2029 (column 21). Thereafter, the effect on unified budget balances would improve, as OASDI repays its debt to the general fund, thus indirectly reducing borrowing from the public.

Stock Holdings

Each of the Council plans would result in substantial new gross investment in the stock market. The effects of these purchases on private demand for and supply of stocks is uncertain. Estimates of the extent to which gross investment by any of these plans might be offset by reductions in investment of other forms is beyond the scope of the work included in this appendix. Therefore, it is not clear what affect these investments would have on prices and yields of stocks (and government and corporate bonds). For the purpose of this exercise, stock prices and yields are assumed to be unaffected.

Table Stock. includes estimated total balances in the OASDI Trust Funds under the MB plan, and in the individual accounts under the IA and PSA plans (in both current and constant 1996 dollars). Stock holdings in the MB trust funds gradually grows to 40 percent of total assets by the end of 2014, and remains at that level, as specified in the plan.

Stock holdings in the individual accounts were calculated based on the Council specified assumptions for investment choice by age (see earlier critical assumptions). Under the IA plan, about 48 percent of IA balances would be held in stocks, declining to about 43 percent around 2040, as the holdings of relatively older account holders become larger (older account holders are specified to hold less stocks).

Stock holdings in PSAs would start at about 53 percent of balances and decline to about 45 percent, assuming that accounts are maintained past retirement and disbursed gradually.

Through about the year 2020, stock holdings in PSA accounts are about double the amount held in the MB Trust Funds, which, in turn, are about double the amount held in IAs. Thereafter, gross stock holdings rise more rapidly for the IA and PSA plans than for the MB plan.

OASDI Unfunded Obligations

Under present law, the OASDI program has obligation over the long-range period (1995 through 2069) of $19.9 trillion expressed in present value as of the beginning of 1995. Combined current assets and income over the same period have a present value of $17.6 trillion. The OASDI program thus has an (open group) unfunded obligation of about $2.3 trillion for the period 1995-2069. This unfunded obligation is essentially the dollar-amount analog of the long-range OASDI actuarial deficit of 2.19 percent of taxable payroll for the same period.

Table.UO presents equivalent analysis of the OASDI unfunded obligation under each of the Council plans. The unfunded obligations are negative in each case, consistent with the fact that income to the OASDI program is estimated to be more than sufficient to provide for obligations through out the 75-year period in each case. Values for the MB plan are shown both with present value computed using the actual yield assumed for fund investments (the customary approach) and, under MB", using the yield for U.S. bonds, consistent with the investment practice under present law and each of the other plans.

Values are also estimated for the succeeding 75-year valuation period (1996-2070) in order to indicate the rate of change in the unfunded obligation in each case from one valuation period to the next. All plans fare well, that is, unfunded obligations become increasingly negative. This is consistent with the requirement that the plans maintain constant trust fund ratios at the end of the 75-year period, thus avoiding deterioration of financial status due to the passage of time well into the future.

Money's Worth Comparisons: Money's Worth Ratios

Methodology

Tables 3A, 3As, and 3Am provide comparisons of "money's worth" ratios derived from contributions and benefits from both the OASDI program and the individual accounts under the various plans. The "moneys worth" ratio is defined as the ratio of expected lifetime benefits (in present value) to expected lifetime contributions (in present value). The values in these tables were computed by Orlo Nichols using a program that projects the full range of possible contingencies that may affect a worker after entering the workforce at age 22.

Table 3A includes "composite workers" which represent a weighted average of the benefits and contributions of a variety of marital-status and/or earnings-level categories. The composite worker may be viewed as the expected outcome for a worker entering the workforce at age 22 with uncertain future marital status and, in the case of the "total" composite worker, uncertain future earnings level. The weights used for developing composite workers were based on beneficiary distributions estimated for the 1995 Trustees Report, and are presented in table W, following tables 3A.

Table 3A also represents married couples with both spouses working, at a variety of earnings levels. Table 3As represents single workers who never marry. Table 3Am represents married couples with one lifetime earner and one nonearner.

Married couples are assumed to be the same age with 2 children born in the couples mid 20s. Workers are assumed to enter the workforce at age 22 and to be steady workers until retiring at age 65, or until death or disability, if earlier. Steady earnings levels illustrated include Low (45 percent of the average), Average (equal to the SSA average wage indexing

series), High (equal to 160 percent of the average), and Max (greater than or equal to the OASDI taxable maximum).

All dollar values are expressed in present value (discounted to age 22, and then wage indexed to 1995, for comparability across cohorts). All present values are based on discounting by the actual past and expected future real yields on the Special Public Debt Obligations issued by the United States Treasury to the Trust Funds. This rate, equal to the market yield on long-term U.S. Government securities (ultimately assumed to be 2.3 percent real), is often referred to as the "riskless rate of return".

The present value of expected benefits includes disabled
worker and retired worker benefits. For married couples, spouse, child and survivor benefits are also included. The probability of death and disability at each age after 22 is reflected in expected benefits. Nondisabled persons who survive to age 65 are assumed to retire at that age, with retired worker benefits beginning at age 65.

The present value of expected contributions includes both employee and employer payroll taxes and mandatory contributions to IAs and PSAs, and reflect the probability of dying and becoming disabled between ages 22 and 65.

Present law values are based on tax rates and benefit levels scheduled in present law (modified to reflect the Council assumption of lower future CPI change), even though income is not projected to be sufficient to allow for timely payment of full benefits after 2035. PL-PAYGO illustrates the effect of increasing tax rates periodically, beginning when the combined OASI and DI Trust Funds decline to 100 percent of annual outgo (around 2030), by enough to maintain the combined fund at 100 percent of outgo (see tables 1A and 1C for tax rates).

Values for the IA and PSA plans with the designation "401K" are presented assuming contributions to PSAs and IAs compound at the intermediate net yields described earlier in the critical assumptions section. Values for the IA and PSA plans are also presented assuming ultimate real net yields of 2.3 and 2 percent, respectively. These are designated as "USBond" and are described earlier as the low yield alternative. Finally, IA and PSA ratios are presented using the "Hi Yld" described earlier. The range of possible yields on individual accounts is useful because many individuals may invest conservatively (as in money-market funds) or may time their movement in and out of equities poorly, thus resulting in disappointing yields, even if the overall equities market performs well throughout their working lifetime.

At retirement (age 65), IAs and PSAs are assumed to be converted to life annuities (joint and survivor annuities in married couple cases) based on general population mortality and an ultimate real net interest rate of 2.3 percent. The assumed valuation interest rate may be conservative, possibly offsetting the lack of an explicit assumption for dministrative costs.

It should be noted that comparisons among some proposals are affected by the use of hypothetical workers who, in the absence of death and disability, are assumed to work steadily from age 22 to age 64 (43 years) at a consistent wage-indexed level of earnings. For example, for actual workers with between 35 and 38 years of earnings, PL PAYGO benefits would be essentially the same, but OASDI benefits under the IA and MB plans would be reduced (by up to about 3 percent) due to the provision to increase the benefit computation period from 35 to 38 years. IA and PSA individual accounts would also accumulate to less, providing smaller retirement benefits.

If, for another example, earnings were scaled to follow the pattern of average earnings by age through the working career, benefits based on accumulations in individual accounts would be reduced (by up to about 5 percent) because account contributions lwould be shifted to later in life and would thus have less time to grow. However, benefits based on the PIA formula or the flat benefit would be little affected. These considerations should be kept in mind when analyzing money's worth and replacement rate values.

Money's Worth Ratios

Money's worth ratios for the three Council plans are depicted graphically in charts 1 through 5 and charts 1a through 4a. Ratios for cohorts born as early as 1930 are included to illustrate the historical trend. All charts include the PL PAYGO plan to provide a comparison of results with the approach where present law benefits are continued by simply increasing tax rates as needed (beginning in 2025).

PL PAYGO

Three distinct periods of change in the money's worth ratio
are evident for the PL PAYGO ratios. First, money's worth ratios drop steadily through cohorts born around 1950. This results from two conditions: (1) through this period the OASDI program was maturing and increasing payroll tax rates resulted in reductions from initially extraordinarily high ratios of benefits to contributions, and (2) interest rates from the late 1970s through the late 1990s exceeded the growth rate in wages to an historically unusual extent--as a result, cohorts with earnings during this period and benefits after this period have extraordinarily high present value of contributions as compared with the present value of benefits. The latter condition largely explains the temporary low level of money's worth ratios clustered around cohorts born in 1950.

The second period shows an increase in ratios between cohorts born around 1950 and those born around 1970. This increase occurs because payroll tax rates are essentially stable among these cohorts, life expectancies are increasing--yielding increasing lifetime benefits, and the depressing effect of high interest rates relative to annual wage growth during the last two decades (described above) disappears as ultimate assumptions return to a more historically usual relationship between interest rates and wage-growth rates. The third period, for cohorts born after 1970, finds declining money's worth ratios due to the

increasing payroll taxes starting 2030 that are specified for the PL PAYGO option.

Council Options

Charts l through 5 and 1a through 4a indicate that the periods of change described above for the PL PAYGO plan are mirrored to some extent in each of the Council options. However, ratios under the Council plans for cohorts born after 1970 do not diminish as they do for PL PAYGO because these plans do not include ever increasing tax rates.

Ultimately, the PSA plan provides the highest composite money's worth ratios because it produces a program that is increasing in size and it provides for the most advance funding and investment in relatively high-yielding stocks. The MB plan produces ratios somewhat higher than the IA plan for all but the low earner composite worker. The MB plan provides ratios that are closer to the PSA for steady single workers, and produces the highest ratios for married one-earner couples, due to the more generous spouse and aged survivor benefits.

The IA and PSA plans produce somewhat lower ratios than PL PAYGO for cohorts born around 1950 in some cases due to the reductions in benefits, changes in normal retirement age, or increases in taxes, before individual accounts are in effect. However, ultimate money's worth ratios for these plans rise back to higher levels as the value of the individual accounts begin to appear. In the case of the PSA plan, an additional shift in the level of money's worth ratios would begin for cohorts born after 2010. This shift results from the elimination of the 1.52 percent transition payroll tax starting 2070. For those unaffected by the transition tax, i.e., those born after 2050, money's worth ratios would be about 12 percent higher (i.e., multiply by 1.12) than shown for cohorts born around 2005.

The MB plan has ratios that follow those for PL PAYGO for cohorts born through 1965, with higher ratios thereafter because higher trust fund yields and additional income from State and local coverage and from taxation of benefits do not affect these ratios the way that increased tax rates affect the ratio for PL PAYGO. The tax rate increase in 2045 under the MB plan does begin to affect ratios for persons born after about 1980, resulting in leveling of money's worth ratios similar to the trends for the IA and PSA plans.

Sensitivity to Individual Account Yield

Charts 6a and 6b provide money's worth ratios additionally for the IA and PSA plans assuming low (US BOND) and high real yields on individual accounts. Although historical analysis indicates that average yields on equities over long periods of time may be expected to exceed the yield for long-term U.S. Government securities, the yields that individuals may experience might vary widely. First, individuals cannot necessarily depend upon the long-term average yield to result for them because of fluctuations and cycles in the equities market as a whole. Some cohorts may experience significantly higher average yields while other cohorts may experience significantly lower yields. Second, many individuals may, even with education, choose to invest conservatively, and not take advantage of the higher potential yield from equities. Third, some individuals may actively invest in equities but make poor (unlucky) choices in their selection of particular stocks or in the timing of their purchase and sale of equities, while others will choose well. Thus, while the intermediate real yield for individual accounts may be the average expected experience for those who invest about half in equities and half in U.S. securities, many individuals will experience substantially higher or lower yields.

The values in Charts 6a and 6b (tables 3A) may also be used to approximate results for other assumed yields on stocks. For example, the intermediate (401K) yield incorporates the assumption of an ultimate real yield on stocks of 7 percent, while the low yield (US BOND) may be interpreted as achieving only a 2.3 percent real yield on stocks. If, for example, results for a 6 percent real yield on stocks were desired, ratios about one-fifth of the way down from the intermediate yield toward the low yield would be appropriate.

Benefits Net of Taxes

The Council plans tax benefits in varying ways. The PSA plan was designed to follow the "consumption tax" principle, or "only tax once". In essence this approach requires that investments made with after-tax money never incur further tax, not even on investment yield accumulations. This exclusion of investment yields (unearned income) is without precedent in United States Tax Code. However, tax treatment that is roughly equivalent to this approach is currently available in 401(k) and IRA plans where before-tax contributions are fully taxable upon distribution (including investment yields). As with the approach specified for IA and PSA accumulations, the IRA 401(k) procedure taxes the plan contribution only once, at the time of distribution rather than at the time of contribution. The financial effect of these approaches is only roughly equivalent because an individual's marginal tax rate may be quite different at the time contributions are made from the marginal rate at the time distributions are taken.

The IA and MB plans provide for taxation of OASDI benefits in the manner currently provided for private and government-employee contributory defined benefit plans. In these plans, all distributions are taxed except for the original employee after-tax contributions. This is equivalent to the exclusion of the original basis for general savings and investment. However, this approach does result in higher taxation than the 401(k) approach, or that applied in the PSA plan. (Note that the same tax principle as used in the PSA accounts is also used in the IA accounts.)

Chart 7a provides a comparison of the Council plans with the
PL PAYGO where all but the PSA have been adjusted to lower benefits to the extent that taxation of benefits exceeds the "tax only once" principle. The reduction factors developed for this purpose are shown in table TA.

Constant Valuation Rate

The money's worth ratios have been developed, as earlier described, by expressing all values in present value, using the actual past and expected future yields on the OASDI Trust Funds (U.S. long-term bond yields). These yields include historical variation, including low real yields before 1980, much higher real yields between about 1980 and 2000, and an assumed movement to a real yield of 2.3 percent by 2010. Use of these actual and expected yields allow the money's worth ratios to answer the question "How do benefits under the plan compare to what I could have gotten if I had invested all my and my employer's contributions in U.S. bonds?" The question answered by using a constant 2.3 percent real discount for present value is, "How do benefits compare to a secure investment at a constant 2.3-percent real yield, when such yields may not have been available across many past periods.

Chart 7b shows the ratios for the Council plans with the constant 2.3-percent real discount rate. The primary effect of this illustration is to eliminate the "dip" in ratios for the cohorts born around 1950, that was caused by the actual variation in U.S. bond yields.

PSA Plan with All Tax (No Borrowing)

Chart 8a compares the money's worth ratios for the Council plans with a variation of the PSA plan where the borrowing needed between 2002 and 2069 to cover transition costs is replaced with additional payroll tax. The additional tax, including the 1.52- percent transition tax, would be about 3.5 percent initially, declining to zero by 2070. This approach results in lower ratios for the PSA plan for cohorts born through about 1990, and higher ratios thereafter.

MB+ Plan

As described earlier, the MB+ plan is essentially the MB plan except that the proportion of trust fund assets invested in stocks is increased to 50 percent, additional taxes and borrowing are added equivalent to those specified under the PSA plan, and benefits are increased for future cohorts as permitted by the additional revenue.

Chart 8a illustrates that the effect of the greater advance funding and higher investment in stocks under the MB+ plan results in composite money's worth ratios that are more comparable with those for the PSA plan. In fact, the ratios are generally higher than those for the PSA because the MB+ plan incurs less administrative expense on its investments.

IA and PSA Residual OASDI Programs

Money's worth ratios have been presented for the IA and PSA plans in their totality. However, the individual accounts and the residual defined benefit OASDI program would be essentially separate programs. Therefore separate analysis of money's worth is appropriate. The money's worth ratios for the individual accounts would of course be high for the intermediate and high yields because the accounts are fully advance funded at higher than U.S. bond rates.

Chart 8b illustrates the relatively low ratios for the residual OASDI program under the IA and PSA plans, designated MTR and LTR, respectively.

Money's Worth Comparisons: Internal Rates of Return

Tables IRR1 through IRR4 provide estimated internal real rates of return for the Council plans. The internal rate of return is computed as the constant real rate of return that is needed to equate the present value of contributions under the plan with the present value of benefits received under the plan.

The ordering of plans by internal rate of return is essentially the same as for the money's worth ratios. Doing better under the one implies doing better under the other. However, where the money's worth ratios provide an automatic comparison with the actual historical and assumed future yields available in the market place for bonds, the internal rates of return allow the reader to make his/her own comparison with the yields available on other investments, but require that the reader has a knowledge of what consistent yields would be over varying time periods.

Benefit Replacement Rates

Benefit levels and replacement rates (ratio of first-year benefit to prior-year earnings) are useful indicators of the value and adequacy of benefits across generations. For example, the present OASDI program has been designed to produce constant

replacement rates across generations for persons with equivalent relative (on a wage-indexed basis) past earnings levels.

However, for the purpose of comparing different plans, benefit levels and replacement rates are only useful if the level of contributions is kept in mind. A plan with higher contributions should be expected to produce higher benefit levels and replacement rates.

Tables RR.1 through RR.4 and Charts R1 through R4 provide estimated benefit replacement rates under present law, the MB, IA, and PSA plans, and the MTR and LTR plans (the OASDI component of the IA and PSA plans, i.e., without the individual account benefits). Replacement rates are presented for hypothetical steady single workers (worker benefit only) with steady low, average, high, and maximum taxable earnings. In 1995, these levels are equivalent to annual earnings of about $15,000, $25,000, $40,000, and $61,000.

Replacement rates are presented for three cases:
(1) workers retiring at age 65.
(2) workers disabled at age 50, and
(3) workers at age 65 who were disabled since age 50.

Under the IA and PSA plans, account balances are assumed to be used to purchase life annuities with CPI indexed payments beginning ate age 65. It should be noted that the IA plan specification suggests that life annuities with a "certain payment" period might be either available or required. Replacement rates presented here assume no certain period, i.e., they assume that annuity payments would terminate at death. Inclusion of a certain period would lower the benefit replacement rates, but would have no affect on money's worth ratios.

Individual account distributions are not available for the disabled worker at age 50 in case 2. The third case is presented to indicate the effect of the availability of IAs and PSAs for the disabled worker by age 65. The account balances are assumed to be used to purchase a life annuity at age 65 in this case. ("Replacement rates" for the third case are computed as the ratio of benefits for the year after reaching age 65 to the earnings level in the year before disability, CPI indexed up to age 64.) All life annuities are assumed to be based on general population mortality expectations with a net interest rate equivalent to the assumed yield on long-term U.S. bonds.

For the retired worker, the PSA plan has the highest replacement rates, consistent with the fact that it develops the most advance
funding with the greatest investment in stocks (Table Stock). Replacement rates for the IA and MB plans are relatively close together.

For the disabled worker at age 50, the IA and PSA plans show lower replacement rates than the MB plan, due to significant benefit reductions. However, upon attainment of age 65, the "replacement rate" for the worker disabled since age 50 improves markedly for the IA and PSA plans, as the balances in individual accounts becomes available. The availability of the wage-indexed past service credit at age 65 is also beneficial for disabled workers who were between 25 and 54 in 1998, under the PSA plan.

Table 1.MB. Estimated Long-Range OASDI Financial Effect
of Advisory Council Maintain Benefits Plan

Ultimate Net Real Yield on Individual Accounts-- **High Return**
PSA--401(k)			IA--401(k) annuitized
Age Group	Percent in Equities	Admin Expense Factor	Net Real Yield	Percent in Equities	Admin Expense Factor	Net Real Yield
<40	55	0.500	5.650	55	0.105	6.045
40-49	52	0.500	5.440	50	0.105	5.695
50-59	48	0.500	5.160	40	0.105	4.995
60-69	43	0.500	4.810	20	0.105	3.595

Proposal	Estimated Change in Long-Range OASDI Actuarial Balance 1/ Proposal (percent of payroll)
0. Assume that changes by the BLS will result in 0.21 percent lower COLAs by Dec. 1997.	0.31
1. Cover under OASDI all State and local government employees hired after 1997.	0.22
2. Increase the number of benefit computation years to 36 in 1997, 37 in 1998, and 38 in 1999 and later, effective by eligibility.	0.28
3a. Beginning 1998, subject OASDI benefits to Federal Income taxation in the same manner prescribed for private and government employee defined benefit pension plans, but with $25,000/32,000 thresholds.	` 0.15
3b. Phase out the $25,000/32,000 thresholds for taxation of benefits between 1998 and 2007.	0.16
3c. Redirect revenue for taxation of benefits from
Subtotal for Proposals 0 through 3	1.37
4. Invest a portion of the OASDI Trust Funds in stocks beginning in 2000, reaching 40 percent of assets in stocks for 2015 and thereafter.	0.82
Subtotal for Proposals 0 through 4	2.19
5. Increase the OASDI payroll tax rate in 2045 by 0.8 percentage points for employees and employers, each.	0.22
Total for Proposals 0 through 5	2.42
Estimated Long-Range OASDI Actuarial Balance	0.24

1/ Estimates for individual proposals do not reflect interaction, except proposal 4 (0.92 without interaction) and proposal 5 (0.39 without interaction).

Based on the intermediate alternative II assumptions of the 1995 Annual Trustees Report.

Office of the Actuary/Social Security Administration October, 1996

Table 1.IA. Estimated Long-Range OASDI Financial Effect

of Advisory Council 1.6 Percent Individual Account Plan

Proposal	Estimated Change in Long-Range OASDI Actuarial Balance 1/ (percent of payroll)
0. Assume that changes in the CPI by BLS will result in 0.21 percent lower COLAs by Dec.1997.	0.31
1. Cover under OASDI all State and local government employees hired after 1997.	0.22
2. Increase the NRA by 2 months per year for 2000 to 2011, reaching NRA of 67. Index NRA thereafter to increases in life expectancy (by about 1 month every 2 years).	0.50
3. Reduce the 0.32 and 0.15 PIA formula factors by 0.5 percent (multiply by 0.995) for 1998-2011, and by 0.15 percent (mult by 0.985) for 2012-30. Factors for 2030 and later, 0.224 and 0.105.	1.32
4. Increase the number of benefit computation years to 36 in 1997, 37 in 1998, and 38 in 1999 and later, effective by year of eligibility.	0.28
5a. Beginning 1998, tax OASDI benefits in the same manner prescribed for private and government employee defined benefit pension plans (retain $25,000/32,000 thresholds).	0.15
5b. Phase out the $25,000/32,000 thresholds for the taxation of benefits between 1998 and 2007.	0.16
6. Reduce the 50 percent aged spouse benefit by 1 percentage point each year 2000 to 2016, reaching an ultimate level of 33 percent.	0.17
7. Gradually replace the current surviving spouse benefit with the highest of own PIA, spouse's PIA, and 75 percent of combined benefits if both were still alive, phased in over 1998 to 2037.	-0.32
8. Starting 1998, require all workers to contribute 1.6 percent of their OASDI taxable earnings to an individual (retirement savings) account. This contribution is in addition to the OASDI payroll tax (6.2 percent for employees, and 12.4 percent for self-employed workers). Proceeds from the accounts would not be taxed. --
Total for Proposals 0 through 8	2.57
Estimated Long Range Actuarial Balance	0.40

1/ Estimates for individual proposals do not reflect interaction, except that proposal 7 reflects interaction with proposal 6.

Based on the intermediate alternative II assumptions of the 1995 Annual Trustees Report.

Office of the Actuary/Social Security Administration October, 1996

Table 1.PSA. Estimated Long-Range OASDI Financial Effect of Advisory Council 5-Percent Personal Security Account Plan

	Estimated Change in Long-Range OASDI Actuarial Balance 1/ Proposal (percent of payroll)
0. Assume that changes in the CPI by BLS will result in 0.21 percent lower COLAs by Dec. 1997.	0.31
1 Cover under OASDI all State and local government employees hired after 1997.	0.22
2. Increase the NRA and EEA by 2 months per year from 2000 to 2011. Index thereafter (limit EEA to 65). Reduce disabled worker benefits, eventually to 70 percent of PIA.	1.25
3. Gradually replace the current PIA formula with a basic flat benefit ($410 in 1996, wage- indexed thereafter), for workers under 55 in 1998. Provide past service credits for workers 25 to 54 in 1998. Disability and young survivor beneficiaries retain current PIA formula. Aged spouses get 50% of the full flat benefit.	3.82
4. Beginning 1998 50% of, OASDI benefits subject to income tax (100% of flat benefit), no revenue transfer for the HI program. Phase out $25,000/ 32,000 thresholds between 1998 and 2007.	0.16
5. Eliminate the earnings test at the NRA.	2/
6. Gradually replace the current surviving spouse benefit with the highest of own PIA, spouse's PIA, and 75 percent of combined benefits if both were still alive, phased in over 1998 to 2037.	-0.39
7. Increase the payroll tax by 0.76 percent for employees and employers, each for 1998 to 2069 (or equivalent tax of another form).	1.42
8. Borrow additional funds from the General Fund of the U.S. Treasury in years 2002 to about 2034, pay back with interest by 2070.	2/
9. Redirect 5 percentage points of the OASDI employee payroll tax to Personal Security accounts for workers under 55 in 1998. Proceeds from the accounts are not taxed.	-4.60
Total for Proposals 0 through 9	2.18
Estimated Long-Range OASDI Actuarial Balance	0.01