CBO Analyzes 'LMS' Social Security Reform Plan

As requested in your letter of July 12, 2005, the Congressional Budget Office (CBO) has analyzed the Social Security proposal developed by Jeffrey Liebman of the Kennedy School of Government, Maya MacGuineas of the New America Foundation, and Andrew Samwick of Dartmouth College. Their LMS proposal would reduce traditional Social Security retirement benefits, increase revenues, and create a new system of individual accounts.

The proposal would reduce traditional Social Security retirement benefits through adjustments to the benefit formula and an increase to 68 in the normal retirement age. It would also increase the earliest eligibility age from 62 to 65 and create a low-earner supplement.

The proposal would boost revenues by raising the taxable maximum (the maximum amount of earnings that is subject to the payroll tax) and by requiring workers to deposit an additional 1.5 percent of their taxable earnings into individual accounts. Also, 1.5 percentage points of the payroll tax would be diverted to the accounts from existing revenues, so a total of 3 percent of taxable earnings would be deposited in the accounts.

The proposed benefit reductions and revenue increases would generally result in improved federal budget balances. However, the proposal would initially result in increased federal budget deficits because the redirection of payroll taxes to individual accounts would appear in the budget as outlays.

Total benefits received by participants, including payouts from individual accounts, would be slightly lower than under current law until about 2035. They would then be approximately equal to current- law benefits until 2052.

Under current law, the trust funds will become exhausted in 2052, CBO projects, at which point annual benefits would be limited to annual revenues. After that year, total benefits would be substantially higher under the proposal. The trust funds would become exhausted in 2089 under the proposal, but total benefits would remain substantially higher than under current law throughout the 100-year projection period (through 2105).

This analysis examines revenues and outlays on the basis of current law, meaning that outlays are limited to current revenues in years after trust fund exhaustion. Appendix A contains a similar analysis based on scheduled benefits, which assumes that those benefits are paid in full even after trust fund exhaustion.

Analysis of the LMS proposal is based on a description of it provided by the authors; no bill has been introduced. CBO made its projections using the 2004 Social Security trustees' demographic assumptions and CBO's January 2005 economic assumptions. The analysis does not reflect any considerations of the potential effects on the macroeconomy that may occur under the proposal. CBO has not produced a 10-year cost estimate.

The underlying data for the figures in this analysis are available on CBO's Web site (www.cbo.gov). If you would like any additional information on these analyses, we will be pleased to provide it. The CBO staff contact for the analysis is Noah Meyerson.

The proposal would reduce traditional Social Security retirement benefits through adjustments to the benefit formula and an increase to 68 in the normal retirement age (NRA). It would also increase the earliest eligibility age from 62 to 65 and create a low-earner supplement.

The proposal would boost revenues by raising the taxable maximum (the maximum amount of earnings that is subject to the payroll tax) and by requiring workers to deposit an additional 1.5 percent of their taxable earnings into individual accounts. Also, 1.5 percentage points of the payroll tax would be diverted to the accounts from existing revenues, so a total of 3 percent of taxable earnings would be deposited in the accounts.

The proposed benefit reductions and revenue increases would generally result in improved federal budget balances. However, the proposal would initially result in increased federal budget deficits because the redirection of payroll taxes to individual accounts would appear in the budget as outlays.

Total benefits received by participants, including payouts from individual accounts, would be slightly lower than under current law until about 2035. They would then be approximately equal to current-law benefits until 2052.

Under current law, the trust funds will become exhausted in 2052, CBO projects, at which point annual benefits would be limited to annual revenues. After that year, total benefits would be substantially higher under the proposal. The trust funds would become exhausted in 2089 under the proposal, but total benefits would remain substantially higher than under current law throughout the 100-year projection period (through 2105).

This analysis examines revenues and outlays on the basis of current law, meaning that outlays are limited to current revenues in years after trust fund exhaustion. Appendix A contains a similar analysis based on scheduled benefits, which assumes that those benefits are paid in full even after trust fund exhaustion.

Analysis of the LMS proposal is based on a description of it provided by the authors; no bill has been introduced. CBO made its projections using the 2004 Social Security trustees' demographic assumptions and CBO's January 2005 economic assumptions. The analysis does not reflect any considerations of the potential effects on the macroeconomy that may occur under the proposal. CBO has not produced a 10-year cost estimate.

• Changes in Benefit Formula: The replacement factors would be reduced under the proposal. The 90 percent replacement factor would be reduced by 0.75 percent per year from 2013 through 2050. (Specifically, the rate would be equal to 0.9925 times the rate in the previous year.) The 32 percent and 15 percent replacement factors would be reduced by 1.8 percent annually. (Each year, the rate would be equal to 0.982 times the rate in the previous year.) The reduction in the 32 percent replacement factor would begin in 2013 and end in 2050. The reduction in the 15 percent replacement factor would begin in 2008 and end in 2045. By 2050, the replacement factors would fall from 90, 32, and 15 percent to 67.6, 16.0, and 7.5 percent.

• Retirement Age: Under current law, the normal retirement age will first rise to 66 for people who turn 62 in 2005. After 12 years, it will again climb in two-month steps for six years, reaching 67 for people who turn 62 in 2022 or later. The proposal would eliminate the 12-year hiatus, and the NRA would ultimately increase to 68. Specifically, the NRA would be 66 and 2 months for people born in 1944 and then increase by two months per birth year, so people born in 1955 and later would have an NRA of 68. (The analysis did not consider the administrative feasibility of changing the NRA for people turning 62 in 2006.)

• Earliest Eligibility Age: The earliest eligibility age would increase from 62 to 65. It would increase to 62 and 2 months for people born in 1955 and then increase by two months per birth year, reaching 65 for those born in 1972 and later.

• Conversion to Old-Age Benefits: Disability benefits would not be affected by the proposal. However, as under current law, once a disabled beneficiary reached normal retirement age, he or she would receive retirement benefits. Under current law, benefits do not change when a disabled worker beneficiary reaches the NRA, but they would fall under the proposal. The reductions to retirement benefits for workers who were entitled to disability benefits at some point in their lives would increase proportionately to the time, between ages 22 and the earliest eligibility age, that they were disabled. For example, a worker who was entitled to disability benefits for most of his life would receive retirement benefits that were approximately equal to current-law benefits. In contrast, someone who became disabled at age 60 would be subject to nearly all of the retiree benefit reductions specified under the proposal.

• Low-Earner Benefit: A "low-earner enhanced benefit" would be established for workers with low earnings over many years of work. This provision was originally proposed as an element of Plan 2 of the President's Commission to Strengthen Social Security. The enhanced benefit would be phased in over 10 years, from 2011 to 2020. The provision would increase traditional benefits by up to 40.4 percent (relative to those that would be paid under the proposal without this provision). That 40.4 percent increase would apply to a worker who earned the minimum wage for at least 30 years; the increase would be smaller for eligible workers with higher earnings or fewer years of work. The provision would also apply to disabled worker beneficiaries, with parameters adjusted to account for their fewer potential years of work.

• High-Earning Spouse Benefits: Under current law, a dependent spouse's benefits are based on an amount equal to 50 percent of the worker's primary insurance amount (PIA). Under the proposal, spousal benefits (before any actuarial reduction) would be capped at an amount equal to 50 percent of the PIA of the average worker in the cohort. This provision would be phased in over 15 years. Benefits for spouses born in 1950 and earlier would not be affected, and the provision would apply fully to spouses born in 1965 and later. For spouses born between 1951 and 1964, the spousal benefit would be based on a weighted average of the two.

• Widow(er) Benefits: Under current law, a surviving spouse is eligible for between one-half and two-thirds of the total Social Security benefit that would have been paid to the couple if the deceased spouse were still alive. Under the proposal, benefits for some survivors would be increased to 75 percent of the couple benefit. The benefit for the surviving spouse (before any actuarial reduction) could not exceed the average PIA for retired-worker benefits in the December before the month of entitlement to the widow(er)s benefit (or, if the month of entitlement is December, then that same month). The proposed change would be implemented for those who apply for a surviving spouse's benefit beginning in 2010.

Figure 1 displays Social Security outlays and revenues as a share of gross domestic product (GDP). Revenues include payroll taxes (including the additional mandatory contribution to individual accounts) and income taxes on benefits, but they exclude interest credited to the Social Security trust funds. Outlays include Social Security benefits, administrative costs charged to the trust funds, and outlays to individual accounts.

• Under the proposal, transfers to individual accounts would begin in 2008. Revenues would increase sharply in that year as participants began to make the additional mandatory contributions to the accounts. In 2008, those mandatory contributions would equal 0.3 percent of GDP. Outlays would increase by more than twice as much, however, since an amount equal to the primary surplus plus the mandatory contributions would be transferred from the trust funds to the individual accounts.

• Under current law, outlays are projected to exceed revenues beginning in 2020 and to exceed the sum of revenues and interest credited to the Social Security trust funds beginning in 2033. The trust funds would be exhausted in 2052; thereafter, outlays in each year would be limited to current revenue.

• Under the proposal, outlays would equal revenues from 2008 through 2020. Thereafter, outlays would be larger than revenues, but the annual deficits would be substantially lower than under current law. Outlays would first exceed revenues and interest in 2035, and trust fund exhaustion would occur in 2089.

• Trust fund exhaustion would occur later under the proposal than under current law, so outlays would be higher under the proposal from 2053 through 2089, and annual balances would be lower.

Figure 1.

 

 

Social Security Revenues and Outlays as a Share of GDP Under

 

Current Law and the LMS Proposal, 1985 to 2105

 

 

Figure 2: Net Effect on the Federal Budget

The effects on the total federal budget balances as a share of GDP are illustrated in Figure 2. Negative numbers mean the proposal would increase the deficit (or reduce the surplus). Positive numbers indicate that the proposal would reduce the deficit (or increase the surplus).

The dotted line shows the effect of the proposal on the primary budget balance -- the balance excluding interest effects, which are the budgetary cost of additional debt held by the public. The solid line includes interest effects.

• Initially -- from 2008 to 2020 -- the additional outlays to individual accounts under the proposal would result in larger primary deficits. (Including increased interest costs, deficits would be larger until 2029.) After 2021, the savings from the reductions to traditional Social Security benefits and the additional revenue from the higher taxable maximum would be larger than the costs of the additional outlays.

• Benefits are assumed to be automatically reduced upon trust fund exhaustion. Under current law, exhaustion will occur in 2052, CBO projects. Under the proposal, scheduled benefits could be paid through 2089, when the trust funds would become exhausted. As a result, from 2053 through 2089, outlays and the deficit would be higher under the proposal than under current law.

• Following trust fund exhaustion, outlays would be automatically reduced so that they would equal revenue, as is the case under current law. Once the trust funds were exhausted, government debt would be the same as under current law, so interest costs would be the same as well.

Figure 2.

 

 

Effects of the LMS Proposal on Total Annual Budget Balances as a

 

Share of GDP Relative to Current Law, 2005 to 2105

 

 

Table 1: Effect on Social Security Finances for Specific Years and Provisions

The top panel of Table 1 shows snapshot measures of Social Security finances under current law at 20-year intervals. Following trust fund exhaustion, benefits are automatically reduced so that annual outlays equal annual revenues. The fourth line shows the size of the automatic benefit reduction.

• Under current law, automatic benefit reductions begin in 2053 (as shown in Figure 1). By 2065, automatic benefit reductions total 1.68 percent of GDP and by 2105 amount to 2.15 percent of GDP.

• Because outlays to the individual accounts would be greater than the additional mandatory contributions, the introduction of those accounts would worsen the annual balance. As more participants reach retirement age and begin drawing from the accounts, the amount of additional revenue from the taxation of the benefits paid from the accounts would increase. As a result, the negative effect of that provision on budget balances would diminish over time.

• Increasing the normal retirement age is effectively equivalent to a reduction in benefits, so that provision would improve the annual balance.

• Increasing the earliest eligibility age would have little effect on the present value of lifetime outlays. Retirees would receive higher monthly payments, but they would receive them for fewer years. As a result, this provision would shift the timing of outlays: it would reduce outlays in earlier years and increase them in later years. Over the long term, it would result in a small increase in the present value of outlays and in federal debt. The increase would occur for two reasons: First, it would induce more workers to apply for and receive disability benefits. Second, for the average beneficiary, lifetime benefits currently are approximately the same regardless of the age at which they are claimed. But even as life expectancy increases, the adjustments made to annual benefits for claiming before or after the normal retirement age will remain constant. As a result, claiming at a later age will result in higher lifetime benefits. In the future, therefore, forcing beneficiaries to claim at a later age would result in higher lifetime benefits.

• Reducing retirement benefits and increasing the taxable maximum would improve annual balances; the low-earner benefit enhancement would worsen balances. The widow and spouse adjustments would have little effect on overall finances.

• Under the proposal, automatic benefit reductions would start following trust fund exhaustion in 2089. In 2105, for example, scheduled outlays would be 0.43 percent of GDP higher than revenues, so there would be an automatic benefit reduction of 0.43 percent of GDP.

The trust fund ratio -- a measure of the adequacy of the trust funds -- is the ratio of the total trust fund balance at the beginning of the calendar year to total Social Security outlays during that year. After the trust funds are exhausted, outlays are limited to Social Security revenues, holding the ratio at zero.

All trust fund assets are included, regardless of their source. For example, an intragovernmental transfer to the trust funds would increase the trust fund ratio but would have no direct effect on the total federal budget. However, this proposal does not include any intragovernmental transfers.

• The elimination of the surplus in earlier years, followed by lower deficits in later years -- as shown in Figure 1 -- would result in lower trust fund balances in earlier years, followed by higher balances in later years. Under current law, the trust funds become exhausted in 2052, while that would not occur until 2089 under the proposal.

Figure 3.

 

 

Social Security Trust Fund Ratios Under Current Law and the LMS

 

Proposal, 1985 to 2105

 

 

 

Figure 4: Total Social Security Benefits

The top panel of Figure 4 shows total Social Security benefits. For the proposal, this includes the annuitized payouts from individual accounts received by beneficiaries.

The bottom panel shows only the traditional Social Security benefits received by beneficiaries.

• From 2006 through about 2035, total Social Security benefits would be slightly lower under the proposal than under current law because of the proposed benefit reductions. They would then be approximately equal to current-law levels through 2052.

• Under current law, benefits would be automatically reduced upon trust fund exhaustion; so from 2053 through 2089, total benefits would be significantly higher under the proposal.

• Trust fund exhaustion is projected to occur under the proposal in 2089, after which traditional benefits would be approximately equal to current-law levels. (Under the proposal, the additional revenues generated by the increase in the taxable maximum plus the additional revenues from taxation of payouts from the individual accounts would approximately equal the cost of the "carve-out" portion of the individual accounts. As a result, the amount remaining to pay traditional benefits would be effectively the same as under current law.)

• The payouts from the individual accounts would be about 7 percent of total benefits in 2030, 20 percent in 2050, and 30 percent in 2080.

Figure 4.

 

 

Social Security Benefits as a Share of GDP Under Current Law and

 

the LMS Proposal, 2005 to 2105

 

 

 

Table 2: First-Year Retirement Benefits

Table 2 shows first-year benefits, including payouts from individual accounts (net of income taxes paid on benefits and credited to the Social Security trust funds) for the median retired worker in three lifetime earnings quintiles. This table shows results only for retired workers. For example, the effects of changes to widow(er) or disabled-worker benefits are not shown.

For ease of comparison, benefits are computed assuming all workers claim retirement benefits at age 65, even though most workers claim at earlier ages. First-year annual benefits are computed for all workers who under current law are eligible to claim Old-Age Insurance benefits at age 62 and who have not yet claimed any other benefit, on the basis only of earnings through age 61. Benefits are adjusted to put them into 2004 dollars.

Payouts from individual accounts are computed by setting the rate of return on account balances equal to the rate of return on Treasury bonds. (For a discussion of the relationship between risk and return of alternative investments, see Congressional Budget Office, Analysis of H.R. 3304, Growing Real Ownership for Workers Act of 2005 (September 13, 2005), Attachment 4: Alternative Investments, Risk, and Return.)

• Under the proposal, the normal retirement age would be higher than under current law for workers born in 1944 and later. As a result, average benefits for workers born in the 1940s would be lower than under current law.

• For many lower-earning participants, the additional benefits from the low-earner enhancement plus the payouts from individual accounts would be greater than the reductions in benefits from other provisions. As a result, total benefits for the median retired worker in the lowest household lifetime earnings quintile would generally be greater under the proposal than under current law. The increase would be even greater for later cohorts, who would turn 65 after 2052, and who would therefore, under current law, have first-year benefits reduced because of trust fund exhaustion.

• Total first-year retirement benefits for middle and higher earners would generally be lower under the proposal for cohorts from the 1980s and earlier. Because trust fund exhaustion would occur much later under the proposal than under current law, benefits would be higher than under current law for those claiming them after 2052, most of whom were born in the 1990s and later.

• As shown in the final column, there would be no intragovernmental transfers under the proposal.

Figure 5 compares the present value of total Social Security benefits received (from both Old-Age and Survivors Insurance and Disability Insurance, including payouts from individual accounts, net of income taxes paid on benefits) to the present value of total payroll taxes paid (by both employers and employees, including mandatory contributions to individual accounts) over a lifetime for all individuals who live to at least age 45.

• The Social Security system has already paid benefits to previous generations that exceed the taxes they paid into the system. Thus, the benefit-to-tax ratio for current and future generations will average less than one, as shown in Figure 5. The ratio exceeds one for the lowest earnings quintile mainly because of that group's high disability receipt rates.

• The proposal raises revenues and adds add-on individual accounts. Thus, as shown in Figure 5, it increases the benefit-to-tax ratio for almost all cohorts and income groups.

• The proposal lowers the benefit-to-tax ratio slightly for earlier cohorts of the middle and higher earnings groups because their taxes increase more than their benefits. The proposal raises the ratio slightly for later cohorts in those earnings groups because the arithmetic boost from the additional revenue more than offsets the fact that the taxes they pay have increased more than their benefits.

• The proposal raises the ratio for all cohorts in the lowest earnings group because it increases their benefits more than their taxes. (This offsets a slight reduction in the ratio that follows arithmetically from the addition of the accounts.)

• The ratio under the proposal would be lower than shown for very high earners -- those in the highest 5 percent or 10 percent of the earnings distribution -- who would be affected the most by the increase in the taxable maximum.

Figure 5.

 

 

Ratio of Lifetime Dedicated-Tax-Financed Benefits to Lifetime

 

Taxes Under Current Law and the LMS Proposal, by Birth Cohort and

 

Lifetime Earnings Level

 

 

 

10-Year Birth Cohort Beginning in Specified Year

 

The following section contains range estimates that are based on 500 stochastic simulations. Those simulations are based on a probability distribution of possible future outcomes for the various demographic and economic inputs used in the projections. The distribution of each assumption is centered at its most likely value, but the variation around those values is based on historical experience.

For the uncertainty analysis, CBO assumes that participants would invest their individual accounts in the following portfolio:

 Investment        Share of Portfolio      Annual Real Expected Return

In its results, CBO gives its estimate of the 80 percent range of uncertainty in figures and the 10^th and 90^th percentiles in tables. There is an 80 percent chance that the actual outcome will fall in the displayed range, a 10 percent chance it will be higher, and a 10 percent chance it will be lower. Likewise, there is a 10 percent chance that the outcome will be below the 10^th percentile and a 10 percent chance that the outcome will be above the 90^th percentile.

In some cases, CBO presents the median -- or middle -- of the range of outcomes in the uncertainty analysis. Those median values and the results in the first section both indicate "typical" results; however, the median results may differ somewhat from the single-simulation results presented earlier.

Figure U1: Effect on Social Security Finances

• The uncertainty about Social Security revenues as a share of GDP results primarily from uncertainty about the level of taxable earnings as a share of GDP. For example, if a larger-than-expected share of compensation is paid in the form of health benefits, then taxable earnings -- and thus Social Security revenues -- will be a lower-than-expected share of GDP.

• Under current law, it is very likely that the trust funds will become exhausted within the projection period. After exhaustion, annual outlays will be limited to revenue in that year, and the uncertainty about outlays will be approximately equal to the uncertainty about revenue.

• Uncertainty about outlays in the years before trust fund exhaustion is driven by uncertainty about future economic and demographic conditions, such as productivity growth and fertility. (See Congressional Budget Office Background Paper, Quantifying Uncertainty in the Analysis of Long-Term Social Security Projections, November 2005.) As shown in the figure, outlays could be lower than revenues under the proposal.

• Under the proposal, trust funds could remain solvent throughout the projection period. If that occurred, scheduled outlays could be paid even if they were higher than revenues.

Figure U1.

 

 

Potential Range of Social Security Revenues and Outlays as

 

a Share of GDP Under Current Law and the LMS Proposal,

 

2005 to 2105

 

 

 

Figure U2: Net Effect on the Federal Budget

Recall that negative numbers mean that the proposal would increase the deficit (or reduce the surplus). Positive numbers indicate that it would reduce the deficit (or increase the surplus).

• Initially -- from 2007 to around 2035 -- the additional outlays to individual accounts would be clearly specified and the changes to traditional benefit outlays would be relatively small, so there is little uncertainty about the proposal's effects on future budget balances.

• In later years, the differences relative to current law would be larger, so there would be greater uncertainty about the proposal's effects on budget balances.

Figure U2.

 

 

Potential Range of Effects of the LMS Proposal on Total Annual

 

Budget Balances as a Share of GDP Relative to Current Law, 2005 to

 

2105

 

 

 

Table U1: Effect on Social Security Finances for Specific Years

The top three lines show the 10^th, 50^th, and 90^th percentiles of Social Security annual balances under current law at 20-year intervals. Following trust fund exhaustion, benefits will automatically be reduced to reach a balance of zero. The next three lines show the size of possible automatic benefit reductions.

• Under current law, the range for the balance includes nonzero numbers in every year because there is some probability that the trust funds will not be exhausted, allowing the system to run a deficit. There is also some probability that the system will run a surplus in any given year.

• Although in the single-simulation analysis presented above the trust funds are projected to become exhausted in 2089, in the uncertainty analysis, they remain solvent throughout the projection period at the 50^th percentile of the probability distribution. Therefore, no automatic benefit reductions are required at the 50^th percentile, though they would be required for most of the period at the 10^th percentile of the distribution.

• Because contributions to the individual accounts are drawn from the trust funds beginning in 2008, the probability of early trust fund exhaustion is greater under the proposal than under current law.

• Because benefit reductions grow over time, there is a significant chance that the trust fund ratio would be large and growing at the end of the projection period.

• Under current law, there is more than a 90 percent chance that the trust funds would be exhausted by 2100. Under the proposal, there is about a 55 percent chance that they would have been exhausted by then. (In some cases, however, the trust funds would fall to zero and then become solvent again as the system ran primary surpluses.)

Figure U3.

 

 

Potential Range of Social Security Trust Fund Ratios Under Current

 

Law and the LMS Proposal, 2005 to 2105

 

 

 

Figure U4: Total Social Security Benefits

• There is more uncertainty about the payouts from the individual accounts than about traditional benefit payments because the accounts are assumed to be invested in a mix of stocks and bonds. Stocks have higher expected returns than government bonds, but they carry greater risk.

• Payouts from the individual accounts are likely to result in total benefits that are at least the current-law level. In later years, total benefits are potentially much larger than under current law. This occurs both as a result of the additional mandatory contributions to the individual accounts and because returns on investments made in the accounts may be higher than returns on government bonds.

Figure U4.

 

 

Potential Range of Social Security Benefits as a Share of GDP

 

Under Current Law and the LMS Proposal, 2005 to 2105

 

 

 

Table U2: First-Year Retirement Benefits

• Within the projection period, most workers would generally receive higher benefits under the proposal than under current law. As explained above in the context of Figure U4, this occurs both as a result of the additional mandatory contributions to the individual accounts and because returns on investments made in the accounts may be higher than returns on government bonds.

• For many lower-earning participants, the additional benefits from the low-earner enhancement plus the payouts from individual accounts would be greater than the reductions in benefits from other provisions. As a result, total benefits for the median retired worker in the lowest household lifetime earnings quintile would generally be greater under the proposal than under current law.

• Total first-year retirement benefits for middle and higher earners would generally be lower under the proposal for cohorts from the 1960s and earlier.

• This table generally shows the proposal to be more advantageous than does Table 2. For example, Table 2 shows that for a middle earner in the 1970s cohort, total benefits would be 10 percent lower under the proposal ($16,900 rather than $18,600). But at the 50^th percentile, Table U2 shows that worker to be slightly better off under the proposal, with benefits of $17,800 rather than $17,600. This occurs mainly because the expected return on private investments used to calculate the figures in Table U2 is greater than the market return used to calculate the figures in Table 2.

• As explained above (see Figure 5), the generally higher benefit-to-tax ratio under the proposal is driven mainly by the introduction of additional revenues.

• The uncertainty is greater under the proposal because of the additional investment risk that would occur in the individual accounts. Those investments also have higher expected returns, which explains in part the higher benefit-to-tax ratios under the proposal.

Figure U5.

 

 

Potential Range of the Ratio of Lifetime Dedicated-Tax-Financed

 

Benefits to Lifetime Taxes under Current Law and the LMS Proposal, by

 

Birth Cohort and Lifetime Earnings Level

 

 

 

 

Appendix A: Scheduled-Benefits Scenario

 

However, the exhaustion of the trust funds would not affect a beneficiary's legal right to full benefits. The analysis in this appendix presents future spending for benefits under an alternative "scheduled benefits" scenario, in which outlays after trust fund exhaustion are assumed to include full benefits, despite any shortfall in the system's annual revenues. This would result in a negative trust fund balance, as if the system borrowed money.

When comparing currently scheduled benefits to those that would be scheduled under the LMS proposal, readers should remember that scheduled benefits are not fully financed under either currently scheduled benefits or under the proposal as scheduled. However, the funding gap would be much lower under the proposal.

Figures 5 and U5 compare taxes paid to benefits received, but this analysis cannot be done without fully specifying the source of the taxes. Because the financing source of some benefits is unspecified under the scheduled-benefits scenario, Figures 5 and U5 are not included in this appendix.

Figure 1 (Scheduled): Effect on Social Security Finances

• Under the proposal, transfers to individual accounts would begin in 2008. Revenues would increase sharply in that year as participants began to make the additional mandatory contributions to the accounts. In 2008, those mandatory contributions would equal 0.3 percent of GDP. Outlays would increase by more than twice as much, since an amount equal to the primary surplus plus the mandatory contributions would be transferred from the trust funds to the individual accounts.

• Under currently scheduled benefits, outlays are projected to exceed revenues beginning in 2020 and to exceed the sum of revenues and interest credited to the Social Security trust funds beginning in 2033. The trust funds would be exhausted in 2052.

• Under the proposal as scheduled, outlays would equal revenues from 2008 through 2020. Thereafter, outlays would remain above revenues, but revenues would be higher and outlays lower than under the scheduled-benefits scenario. Outlays would first exceed revenues and interest in 2035, and trust fund exhaustion would occur in 2089.

Figure 1 (Scheduled).

 

 

Social Security Revenues and Outlays as a Share of GDP Under

 

Currently Scheduled Benefits and the Scheduled LMS Proposal,

 

1985 to 2105

 

 

 

Source: Congressional Budget Office.

 

Recall that negative numbers mean that the proposal would increase the deficit (or reduce the surplus). Positive numbers indicate that it would reduce the deficit (or increase the surplus).

• Initially -- from 2008 to 2020 -- the additional outlays to individual accounts under the proposal would result in larger primary deficits. (Including increased interest costs, deficits would be larger until 2029.) After 2021, the savings from the reductions to traditional Social Security benefits and the additional revenue from the higher taxable maximum would be larger than the costs of the additional outlays.

• The higher revenues and lower outlays under the proposal would improve the annual budget balance (excluding interest). By the end of the projection period, the improvement would approach 2 percent of GDP. This would result in a large reduction in federal interest payments; by the end of the projection period, the improvement including interest would be nearly 9 percent of GDP.

Figure 2 (Scheduled).

 

 

Effects of the Scheduled LMS Proposal on Total Annual Budget

 

Balances as a Share of GDP Relative to Currently Scheduled

 

Benefits, 2005 to 2105

 

 

 

 

Source: Congressional Budget Office.

 

• This table is nearly identical to Table 1. The main difference is that under the scheduled-benefits scenario, there are no automatic benefit reductions.

• Under currently scheduled benefits, after trust fund exhaustion, revenues are about 0.1 percent of GDP higher than under current law (shown in Table 1). This is due to higher revenues from the taxation of benefits, which result from higher outlays.

Under the scheduled-benefits scenario, the trust fund balance could become negative, representing net borrowing by the Social Security program.

• Under currently scheduled benefits, the trust funds would first become negative in 2052, whereas that would not occur until 2089 under the proposal.

• Because annual deficits would be much smaller under the proposal as scheduled, the trust fund ratio would decline more slowly than under currently scheduled benefits.

• At the end of the projection period, the trust fund ratio would be - 22 with currently scheduled benefits, but -1 under the proposal as scheduled.

Figure 3 (Scheduled).

 

 

Social Security Trust Fund Ratios Under Currently Scheduled

 

Benefits and the Scheduled LMS Proposal, 1985 to 2105

 

 

 

 

Source: Congressional Budget Office.

 

• Total benefits paid under the proposal as scheduled, including payouts from the individual accounts, would be approximately equal to the currently scheduled level. They would be slightly lower than currently scheduled benefits before 2040 and slightly higher in later years. (However, the gap between currently scheduled taxes and benefits is much larger than under the proposal.)

• Traditional scheduled benefits would be much lower under the proposal. (Scheduled disability benefit outlays, which are not shown separately in the figure, would be about 20 percent higher under the proposal. The proposal makes no direct changes to scheduled disability benefits, but the low-earner supplement would result in higher benefits for many disabled beneficiaries.)

• The payouts from the individual accounts would be a substantial portion of total benefits: for example, they would be about 7 percent of the total in 2030, 20 percent in 2050, and 30 percent in 2080.

Figure 4 (Scheduled).

 

 

Social Security Benefits as a Share of GDP Under Currently

 

Scheduled Benefits and the Scheduled LMS Proposal, 2005 to 2105

 

 

 

 

Source: Congressional Budget Office.

 

• Under the proposal, the normal retirement age would be higher than under current law for workers born in 1944 and later. As a result, average benefits for workers born in the 1940s would be lower than currently scheduled.

• For many lower-earning participants, the additional benefits from the low-earner enhancement plus the payouts from individual accounts would be greater than the reductions in benefits from other provisions. As a result, total benefits for the median retired worker in the lowest household lifetime earnings quintile would generally be greater under the proposal as scheduled than under currently scheduled benefits.

• Total scheduled first-year retirement benefits for middle and higher earners would generally be lower under the proposal. (As noted above, aggregate scheduled benefits under the proposal would be approximately equal to currently scheduled benefits. Lower outlays for retirement benefits would be offset by an increase in scheduled disability benefit outlays.)

• In the current-law analysis, annual outlays equal annual revenues following trust fund exhaustion, so uncertainty about outlays is limited after exhaustion (see Figure U1). There is no such restriction in the scheduled-benefits scenario, so there is much greater uncertainty about outlays in later years than in the current-law analysis.

• The partial shift to individual accounts leads to less uncertainty about outlays from the trust funds under the proposal. Outlays to individual accounts are a fixed portion of revenues, so there is little uncertainty about that share of outlays.

Figure U1 (Scheduled).

 

 

Potential Range of Social Security Revenues and Outlays as a Share

 

of GDP Under Currently Scheduled Benefits and the Scheduled LMS

 

Proposal, 2005 to 2105

 

 

 

 

Source: Congressional Budget Office.

 

Recall that negative numbers mean that the proposal would increase the deficit (or reduce the surplus). Positive numbers indicate that it would reduce the deficit (or increase the surplus).

• Initially, the possible range of effects of the proposal on the annual budget balances would be the same as under the current-law analysis shown in Figure U2.

• In later years, there would be much greater uncertainty about the effects than is shown in the current-law analysis, because there is much greater uncertainty about outlays, as shown in Figure U1 (Scheduled).

Figure U2 (Scheduled).

 

 

Potential Range of Effects of the Scheduled LMS Proposal on Total

 

Annual Budget Balances as a Share of GDP Relative to Currently

 

Scheduled Benefits, 2005 to 2105

 

 

 

Table U1 (Scheduled): Effect on Social Security Finances for Specific Years

• The data shown here are similar to those shown in Table U1, but there are no automatic benefit reductions in this scenario.

• In general, the balance under the scheduled-benefits scenario is equal to the sum of the balance and the automatic benefit reductions in the current-law scenario.

• Note: This is not necessarily true of the data shown in the tables because data are sorted strictly numerically. In some simulations, the trust funds become exhausted in the 2030s. Under current law, annual balances automatically equal zero after trust fund exhaustion. Since the annual balance in most simulations is negative, a scenario with a balance that falls to zero will move very high in the distribution. Most other values in the distribution will then shift downward. As a result, for example, in the currently scheduled benefits scenario the 50th percentile of the balance in 2045 is -1.60. The corresponding value under current law is -0.65.

• While the trust fund ratio would generally be higher under the proposal than under currently scheduled benefits, the range of uncertainty about the ratio would be almost identical.

• The uncertainty ranges are wider under the scheduled-benefits scenario than in the analysis shown in Figure U3 only because the trust funds are allowed to fall below zero in the scheduled-benefits scenario.

Figure U3 (Scheduled).

 

 

Potential Range of Social Security Trust Fund Ratios Under

 

Currently Scheduled Benefits and the Scheduled LMS Proposal,

 

2005 to 2105

 

 

 

Figure U4 (Scheduled): Total Social Security Benefits

• The amount of uncertainty about traditional benefits as a share of GDP under the proposal is similar to the amount of uncertainty under currently scheduled benefits (see bottom panel).

• There is greater uncertainty about total benefits as a share of GDP under the proposal than under currently scheduled benefits, primarily because of the additional uncertainty about investment returns on account holdings (see top panel).

Figure U4 (Scheduled).

 

 

Potential Range of Social Security Benefits as a Share of GDP

 

Under Currently Scheduled Benefits and the Scheduled LMS Proposal,

 

2005 to 2105

 

 

 

Table U2 (Scheduled): First-Year Retirement Benefits

• There is greater uncertainty about scheduled benefits under the proposal than under currently scheduled benefits, primarily because of the additional uncertainty about investment returns on account holdings.

• The range between the 10th and 90th percentiles is greater under the proposal than under currently scheduled benefits in dollar terms. However, this is partly because the overall range is shifted higher. For example, for workers in the middle earnings quintile born in the 1980s cohort, the range between the 10th and 90th percentiles is 25 percent higher under the proposal than under currently scheduled benefits (a range of $13,700 rather than $11,000). But median benefits (the 50th percentile) for that cohort are higher under the proposal, so relative to the median level, the range is only 14 percent larger.