Social Security Reform: Potential Effects on SSA's Disability Programs
and Beneficiaries (Letter Report, 01/24/2001, GAO/GAO-01-35).

There has not been much analysis of how the various Social Security
reform proposals will affect the Social Security Disability Insurance
(DI) Program. GAO assessed to potential of these proposals on the
solvency of the DI trust fund. GAO's analyses indicate that most
disabled beneficiaries would receive higher benefits under the various
Social Security reform proposals it reviewed than under a solvency
scenario that maintained payroll tax rates while reducing benefits.
However, most disabled beneficiaries with the characteristics GAO
studied would receive lower benefits under three of the reform proposals
reviewed than under a solvency scenario that maintained current-law
benefits while raising payroll taxes. The proposals GAO studied treat DI
beneficiaries similar to Old-Age and Survivor Insurance beneficiaries.
However, the circumstances facing disabled workers differ from those
facing retired workers. The differences between disabled and retired
workers suggest that Social Security reform proposals should be viewed
not only in light of their effects on retired workers but also
explicitly for their effect on disabled beneficiaries and their
families.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  GAO-01-35
     TITLE:  Social Security Reform: Potential Effects on SSA's
	     Disability Programs and Beneficiaries
      DATE:  01/24/2001
   SUBJECT:  Social security benefits
	     Disability benefits
	     Retirement benefits
	     Disability insurance
	     Federal social security programs
	     Income maintenance programs
IDENTIFIER:  Social Security Program
	     Social Security Disability Insurance Program
	     Supplemental Security Income Program
	     Old-Age and Survivors Insurance Program
	     SSI

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-01-35

A

Report to the Ranking Member, Subcommittee on Labor, Health and Human
Services, Education, and Related

Agencies, Committee on Appropriations, U. S. Senate

January 2001 SOCIAL SECURITY REFORM

Potential Effects on SSA's Disability Programs and Beneficiaries

GAO- 01- 35

Letter 3 Appendixes Appendix I: The Social Security Reform Proposals and
Their

Provisions 40 Appendix II: Alternative Solvency Scenarios and the Social
Security

Simulation Model 42 Appendix III: Comments From the Social Security
Administration 50

Tables Table 1: Effect of Reform Proposals on Trust Fund Solvency and
Benefit Income 7

Table 2: Proposals That Maintain the Current Level and Structure of Benefits
12 Table 3: Proposals That Affect the Current Level and Structure of

Benefits 13 Table 4: Actuarial Balance of Social Security Trust Funds Under

Current Law and Selected Reform Proposals 16 Table 5: Estimated Change in
Current Law Actuarial Balance

for Selected Reform Provisions 20 Table 6: Social Security Reform Provisions
40 Table 7: Individual Account Provisions 41 Table 8: Reductions in Benefits
42 Tabl e 9: Payroll Tax Rates 44 Table 10: Economic and Demographic
Intermediate Assumptions

From the 1999 Trustees Report 45 Figures Figure 1: Lifetime Benefit Income
of Low- Earnings Worker and

Dependents Under Selected Proposals and Alternative Solvency Scenarios 22
Figure 2: Lifetime Benefit Income of Average- Earnings Worker and

Dependents Under Selected Proposals and Alternative Solvency Scenarios 23
Figure 3: Comparison of COLA Changes to Lifetime Insurance

Benefits of Disabled Workers With Alternative Solvency Scenarios 26 Figure
4: Comparison of PIA Changes to Lifetime Insurance

Benefits of Disabled Workers With Alternative Solvency Scenarios 27

Figure 5: Comparison of PIA Changes to Lifetime Insurance Benefits With
Alternative Solvency Scenarios for Adult Disabled Child Dependent of Retired
Worker 28 Figure 6: Lifetime Benefit Income of Disabled Workers Under IA
Provisions and Alternative Solvency Scenarios 30

Figure 7: Lifetime Benefit Income of Dependent of Disabled Worker Under IA
Provisions and Alternative Solvency Scenarios 31 Figure 8: Lifetime Benefit
Income of Adult Disabled Child

Dependent of Retired Worker Under IA Provisions and Alternative Solvency
Scenarios 32 Figure 9: Net Addition of IA to Current- Law Lifetime Benefit
Income Under Selected Reform Proposals: Low- Earnings

Worker Disabled at Various Ages 34 Figure 10: Net Addition of IA to Current-
Law Lifetime Benefit IncomeUnder Selected Reform Proposals: Average-
Earnings

Worker Disabled at Various Ages 35

Abbreviations

AIME average indexed monthly earnings COLA cost- of- living adjustment CPI
Consumer Price Index DI Disability Insurance IA individual account NRA
normal retirement age OASI Old- Age and Survivor Insurance PIA primary
insurance amount SSA Social Security Administration SSI Supplemental
Security Income

Lett er

January 24, 2001 The Honorable Tom Harkin Ranking Member, Subcommittee on
Labor,

Health and Human Services, Education, and Related Agencies Committee on
Appropriations United States Senate

Dear Senator Harkin: Although much has been written about the effects of
reform on the solvency of the Social Security program and on the benefits of
retired workers, little attention has been directed to the effects of reform
proposals on the Social Security Disability Insurance (DI) program or on

the benefits that disabled beneficiaries and their families receive. Yet, in
1999, disabled beneficiaries and their families accounted for about 17
percent of all Social Security beneficiaries. These 7.4 million
beneficiaries included disabled workers, families of disabled workers, and
adult disabled children who were dependents of disabled, deceased, or
retired workers. You asked us to assess the potential effects of Social
Security reform options on the solvency of the DI trust fund and the
benefits disabled beneficiaries receive. In this report, we analyze both the
potential effects of comprehensive Social Security reform proposals on the
solvency of the DI

trust fund and on the benefits disabled beneficiaries receive and the likely
contribution that individual proposal provisions would make to these
effects. You also asked us to examine the potential implications of Social
Security reform for the Supplemental Security Income (SSI) program, which
provides significant income support for low- income individuals with
disabilities.

In response to your request, we analyzed the five Social Security reform
proposals that we have previously reviewed. 1 These proposals include that
of President Clinton as well as four of the proposals discussed in the 106th
Congress: Archer- Shaw, Kolbe- Stenholm (H. R. 1793), Gregg- Kerrey-
BreauxGrassley (S. 1383), and Kasich. In these proposals, we identified 11
major

1 Social Security: Evaluating Reform Proposals (GAO/ AIMD/ HEHS- 00- 29,
Nov. 4, 1999). With respect to the President's proposal, we evaluated the
version that was presented in the President's fiscal year 2001 budget on
Feb. 7, 2000.

types of provisions and examined their effect on trust fund solvency and DI
benefits. 2 We compared the benefits under these reform proposals with
benefits under two alternative current- program scenarios that also achieve
trust fund solvency: one that would maintain current benefits while
increasing payroll taxes and another that would maintain current payroll tax
rates while reducing benefits.

To analyze the effects of Social Security reform on DI trust fund solvency,
we used estimates produced by the Social Security Administration's (SSA)
Office of the Chief Actuary as well as our estimates using the SSASIM policy
simulation model. 3 We also used the SSASIM model to examine the effects of
these reforms on benefits that disabled beneficiaries receive. Our estimates
using the SSASIM model are based on the intermediate assumptions reported in
the 1999 Social Security Trustees Report because the Office of the Chief
Actuary has scored only a few of the proposals we studied using the
assumptions in the 2000 Social Security Trustees Report.

We analyzed the effects of Social Security reform on the Old- Age and
Survivors Insurance (OASI) program as well as on the DI program because the
solvency estimates are generally available for the combined programs only
and because many disabled beneficiaries receive benefits from the

OASI program. For example, more than 90 percent of the adult disabled
children who receive Social Security benefits are dependents of deceased or
retired workers and therefore receive OASI benefits. Disabled workers who
have reached retirement age and their dependents receive benefits from the
OASI program because DI benefits are automatically converted to

retirement insurance benefits at the normal retirement age (NRA). Because
little information is available about the earnings levels and work history
of DI beneficiaries, we assumed the best possible case for disabled
beneficiaries. Such an assumption would generally provide an upper limit to
the benefits that most disabled beneficiaries could expect to receive under
the Social Security reform proposals. We assumed they work fulltime until
they receive disability benefits. We also assumed they received

2 See app. I for information on these provisions. 3 The SSASIM model,
developed by the Policy Simulation Group, Inc., can simulate a variety of
policy reforms to the Social Security program from incremental changes in
the Old- Age and Survivors Insurance (OASI) and DI programs to broader
structural reforms that would introduce an individual account (IA) component
to the Social Security system. Additional information on the model and the
assumptions we made in using it are in app. II.

earnings equal to either the average economywide earnings of men or 45
percent of the average earnings of men. Following the approach taken in the
Report of the 1994- 96 Advisory Council on Social Security, we chose a low
administrative cost for the individual account (IA), implicitly assuming

a centralized system of recordkeeping and a limited number of investment
options. We assumed that individuals self- annuitized by drawing down the
balance in the IAs through periodic withdrawals. As a result, the benefit
income for the disabled beneficiaries whom we studied in this report will
clearly be greater than that for disabled beneficiaries with intermittent or

less than full- time employment who pay annuitization costs and relatively
higher administrative costs on their IAs. We conducted our work between
September 1999 and November 2000 in accordance with generally accepted
government auditing standards. Results in Brief According to estimates of
SSA's Office of the Chief Actuary, the Social Security reform proposals that
we examined would improve the solvency

of the combined DI and OASI trust funds. Table 1 shows that the extent of
the improvement varies across proposals. Most of the individual proposal
provisions, such as those that call for general revenue transfers or benefit
reductions, would have a positive effect on the solvency of the DI trust
fund by increasing revenues and decreasing costs, respectively. Only a few
provisions, such as those redirecting payroll taxes to IAs, which would
reduce trust fund revenues, and those establishing a minimum benefit,

which would increase some benefits and therefore costs, would have a
negative effect on the DI trust fund.

With regard to benefits, our analysis shows that three proposals- Kasich,
Kolbe- Stenholm, and Gregg- Kerrey- Breaux- Grassley- would result in
reduced benefit income for most of the disabled beneficiaries with the
selected characteristics that we simulated when compared with a solvency
scenario that maintains current- law benefits by increasing payroll taxes or
other sources of revenue. 4 (See table 1.) However, most disabled

beneficiaries would receive greater benefit income under any of these three
proposals than under a scenario that would achieve solvency and maintain
current payroll tax rates by reducing benefits. 5 These proposals
distinguish

between the insurance benefits received from Social Security and the income
that would be received from the IAs created under these proposals. 6 The
reform proposals would reduce insurance benefits while creating IAs, with
the expectation that the income from an IA would largely

offset reductions in the insurance benefits. In our estimates, the income
from the IA was not sufficient to compensate for the decline in the
insurance benefits that disabled beneficiaries would receive. This is
especially true for the Kasich and the Gregg- Kerrey- Breaux- Grassley
proposals, which would further reduce the insurance benefit in response to
the potential income from the IA. Our estimates also indicate that the

longer a beneficiary works before becoming disabled, the greater the income
from the IA will be.

4 President Clinton's proposal maintains current- law benefits and does not
include an IA provision. The Archer- Shaw proposal differs from the other
proposals that contain an IA provision in that it guarantees that
individuals will receive at least the amount of their current- law benefits.

5 Following the approach used in the Report of the 1994- 96 Advisory Council
on Social Security, we adjusted the benefit formula so that only the
benefits at the higher levels are reduced while the benefits at the lowest
levels are unaffected. For additional information, see app. II. 6 In this
report, we look at two kinds of benefits: the traditional benefit provided
by Social Security, which we refer to as the insurance benefit, and the
income from the IA. We refer to

the sum of these two benefits as benefit income. The income from the IAs
depends on assumptions as to rate of return, administrative costs, and so
on, which we describe later in this letter and in app. II.

Table 1: Effect of Reform Proposals on Trust Fund Solvency and Benefit
Income Benefit income: average earner who

first receives DI at age 45 b OASDI trust fund

Maintain Proposal

solvency a benefits Maintain tax rates

Current law 0.07 c c President Clinton's 0.80 c c Kasich +0. 00 d 15. 6 +28.
1 Kolbe- Stenholm +0.07 17. 7 +25. 0 Archer- Shaw +0.09 c c Gregg- Kerrey-
BreauxGrassley +0. 28 4.2 +45.3 a This columns represents actuarial balance
as a percentage of taxable payroll. OASDI, Old- Age, Survivors, and
Disability Insurance, refers to the combined OASI and DI programs. b These
columns represent percentage change. c Proposal does not affect the level
and structure of benefits. d Assumes everyone opts for the IA, which is
voluntary in this proposal. If no one opts for the IA, the

actuarial balance is +0.24 Source: Estimates are based on 1999 Social
Security Trustees Report.

The Social Security reform proposals we reviewed could increase costs for
the SSI program. Individuals receiving DI or OASI benefits who have low
levels of income and assets could supplement their income with benefits from
the SSI program. The SSI program could be affected in two different ways.
First, for beneficiaries who already receive SSI benefits, a reduction

in Social Security benefits resulting from reforms could lead to an increase
in SSI benefits (up to the legislated maximum) and thus to an increase in
costs for the SSI program. 7 Second, a reduction in DI and OASI benefits as
a

result of reform would likely make some individuals not currently receiving
SSI newly eligible for its benefits.

In commenting on this report, SSA noted that we addressed an important topic
that has until now received little attention. The agency specifically
highlighted two points in our report as being important for policy makers

7 If the individual account accumulations are treated as assets with regard
to determining SSI eligibility, some low- income individuals may lose SSI
benefits. Whether this loss of benefits in the case of disabled
beneficiaries would occur when they first received benefits or when they
reached retirement age depends upon the proposal.

considering changes to Social Security: IAs might not fully offset Social
Security insurance benefit reductions for some beneficiaries, and SSI
benefits might increase as they compensate for the decline in DI benefits
resulting from Social Security reform. However, the agency had some concerns
about our use of a “best case” scenario to estimate the effects
of policy options and about the assumptions underlying this scenario. SSA
also expressed concern regarding our focus on lifetime benefits, a measure
that it believes does not adequately reflect differences in living standards
across policy options at specific points in time. Finally, it suggested that
we include a measure reporting on money's worth or internal rates of return
in

our comparison of costs and benefits of Social Security reform proposals.
Our use of a “best case” scenario demonstrated that even under
the best of circumstances, Social Security reform proposals could reduce
benefits to DI beneficiaries- people who would find it more difficult than
most nondisabled retirees to replace lost benefits with other sources of
income such as earnings. We did not examine “worse case”
scenarios because the

“best case” scenario demonstrates that most DI beneficiaries
would be adversely affected by the reform proposals we analyzed. With
respect to the agency's concern with our focus on lifetime benefits, we
acknowledge that we do not address the issue of variations across plans in
living standards before retirement age resulting from differences in the
accessibility of income from the IAs. As for the inclusion of money's worth
or internal rate of return measures, we agree that such analysis would be
useful, but these measures are beyond the scope of this report.

Background Working- age adults with disabilities may obtain cash benefits
from a number of private and public programs. 8 After the onset of a
disabling

condition, workers needing long- term cash benefits may receive assistance
from workers' compensation, private disability insurance, or DI. However, in
1996, only 26 percent of private sector employees had long- term disability
coverage under employer- sponsored private insurance plans. 9 Thus, the DI
program is an important provider of monthly benefits to workers who are no
longer able to work because of a severe long- term disability.

8 SSA Disability: Return- to- Work Strategies From Other Systems May Improve
Federal Programs (GAO/ HEHS- 96- 133, July 11, 1996). 9 Private Disability
Insurance (GAO/ HEHS- 00- 18R, Nov. 5, 1999).

Most Social Security disabled beneficiaries, including disabled workers and
their dependents, receive benefits from the DI program. However, adult
disabled children who are dependents of deceased or retired workers, and
disabled workers who have reached retirement age and their dependents,
receive monthly benefits from the OASI program. 10 In 1999, about 6.5
million beneficiaries received DI cash benefits totaling about $51.3
billion,

while about 38.0 million beneficiaries received OASI cash benefits totaling
about $334.4 billion. 11

Benefits Available Under Benefits for both OASI and DI beneficiaries are
based on the application of Current Law

the Social Security benefit formula to the worker's average monthly lifetime
earnings. The resulting monthly benefit is the amount payable to a worker
who becomes entitled to disability benefits or retires at the NRA. Because
monthly benefits for DI and OASI beneficiaries are based on the same benefit
formula, any change in this formula, as has been proposed in some Social
Security reform plans, could affect benefits disabled workers as well as
retired workers receive. Both DI and OASI monthly benefits will also be
affected by other proposed Social Security reform changes, such as

decreases in the cost- of- living adjustment (COLA). However, only OASI
monthly benefits are affected by proposed changes in the retirement age. 12

Under current law, the age at which an individual is first eligible to
receive full retirement benefits, or NRA, is gradually increasing from 65 to
66 for those who turn 62 in 2005 and to 67 for those who turn 62 in 2022.
Benefits retired workers take before NRA are subject to an actuarial
reduction.

Benefits taken by workers who postpone retirement and work between NRA and
age 70 are increased through a delayed retirement credit for each month
retirement is delayed. The benefit formula is weighted in favor of 10
Disability insurance benefits are automatically converted to retirement
insurance benefits at the NRA.

11 The 6.5 million beneficiaries receiving DI cash benefits do not include
adult disabled children who are dependents of deceased or retired workers
and, therefore, receive benefits from the OASI program. Consequently, this
number differs from the 7. 4 million disabled beneficiaries, reported
earlier, who do include adult disabled children who are dependents of
deceased or retired workers. The 38. 0 million OASI beneficiaries include
adult disabled children who are dependents of deceased or retired workers
and disabled workers whose benefits have been converted to retirement
insurance benefits.

12 The decrease in OASI benefits as a result of the increase in the
retirement age may prompt older workers to apply to the DI program in order
to receive what would be relatively greater disability benefits.

workers with lower earnings, so that benefits replace a larger proportion of
their earnings. Benefits are adjusted each year, based on increases in the
Consumer Price Index (CPI) in order to account for inflation. Auxiliary
benefits are paid to eligible dependents and are 50 percent of the Social
Security benefit that the disabled or retired worker receives, subject to a
maximum family limit on benefits. Upon the death of an insured worker, the
eligible spouse receives 100 percent of the worker's benefit (subject to
reduction for age) and the eligible surviving child receives 75 percent of
the benefit. 13 Individuals who receive low levels of DI or OASI benefits
can supplement them with benefits from SSI. 14 The SSI program, which was
authorized in

1972 under title XVI of the Social Security Act, is funded through general
revenues and provides monthly benefits to aged, blind, and disabled
individuals who have income and resources below specified thresholds.

The DI and SSI programs use the same criteria and procedures for determining
disability. However, unlike DI beneficiaries, SSI recipients do not need to
have a work history to qualify for benefits. The maximum federal SSI monthly
benefit in 1999 was $500 for an individual. This monthly benefit level is
reduced, depending on a recipient's income and

other sources of support, such as Social Security benefits. In 1999, 36
percent of SSI recipients also received Social Security benefits from either
OASI or DI. The average federal monthly benefit in 1999 was $249 for the
aged, $351 for the blind, and $364 for the disabled. In addition to the
federal SSI benefits, some states provide supplemental benefits that are
intended to reflect regional differences in living costs.

Social Security is financed primarily on a pay- as- you- go basis, which
means that the Social Security payroll taxes that current workers pay are
used to pay for current benefits. In 1999, there were approximately 3.4
workers for every beneficiary, but this number is projected to fall to 2.1
by 2030.

13 Spouses may also be eligible for benefits based on their own work
records. In these cases, spouses receive their own worker benefits plus the
difference between their spouse benefit and their own worker benefit.
Children younger than 18 are eligible for insurance benefits. In addition,
children who are full- time elementary or secondary students might receive
insurance benefits until age 19. Children older than 18 might receive
benefits if they are suffering a disability that began before age 22.

14 An individual might also supplement DI benefits with benefits from
private insurance. Benefits from private insurance might come from long-
term disability insurance or from pension plans with disability pension
features.

Because of this change in the ratio of workers to beneficiaries, and other
factors, the Social Security trust funds will have a projected financial
shortfall or funding gap of approximately $3 trillion over the next 75
years. According to estimates in the 2000 Trustees Report, the OASI trust
fund is projected to have sufficient funds to fully finance benefits until
2039, while

the DI trust fund is projected to have sufficient funds to fully finance
benefits until 2023. After the trust funds are exhausted- that is, after
2039 for the OASI trust fund and 2023 for the DI trust fund- the annual tax
revenues of the trust funds are expected to be sufficient to cover only
about 70 percent of annual expenditures. 15

Social Security Reform In order to address the solvency of the trust funds,
a number of Social Proposals Address Solvency Security reforms have been
proposed. We assessed five of these proposals,

by Affecting the Level and some of which maintain the level of current law
benefits and some of which

Structure of Benefits reduce and restructure these benefits. Table 6 in
appendix I lists the provisions in each proposal.

Two of the proposals we studied, President Clinton's proposal and the
Archer- Shaw proposal, maintain the current level and structure of benefits.
(See table 2.) Three of the reform proposals we studied- Kasich,
KolbeStenholm, and Gregg- Kerry- Breaux- Grassley- both reduce and
restructure current benefits. (See table 3.)

15 The OASI and DI trust funds, which together make up Social Security, are
set up as two separate accounts in the U. S. Treasury. However,
historically, there has been little real significance to this division with
respect to the financing of the program since, over the years, there have
been tax rate reallocations and loans between the two trust funds. The
assets in the combined OASDI trust funds will be exhausted in 2037,
according to projections by SSA's Office of the Chief Actuary as reported in
the 2000 Trustees Report.

Table 2: Proposals That Maintain the Current Level and Structure of Benefits
Proposal Description

President Clinton's a ? Maintains current- law benefits for current and
future retirees and disability insurance beneficiaries by requiring
additional general fund transfers to OASDI trust funds in each fiscal year
beginning

in 2011 through 2050. In addition, a portion of these transfers would be
invested in equities.

Archer- Shaw ? Maintains current law benefits for current and future
beneficiaries. Requires mandatory “add- on” individual accounts
financed through a refundable tax credit paid from the general fund. The
contribution to the IA equals 2 percent of taxable payroll. Beginning at
retirement or disability, the account balance is gradually returned to the
OASDI trust funds to finance benefits.

? Actual retirement or disability income could be higher, depending on the
account balance. The benefit amount paid each month is either the current
law amount or the payout based on the annuitized account balance, whichever
is higher. a The version of President Clinton's proposal that was presented
in his fiscal year 2001 budget on Feb. 7, 2000.

Table 3: Proposals That Affect the Current Level and Structure of Benefits
Proposal Description

Kasich a ? Reduces current law benefits by indexing initial benefits to
prices rather than average wages, as under current law. b

? Restructures benefits by offering voluntary individual “carve-
out” accounts with contributions financed by redirection of between 1
and 3.5 percent of individuals' taxable earnings, with a higher percentage
available to lower- income earners. For workers choosing the account option,
an additional benefit reduction is made at retirement to offset
contributions to their accounts. c Gregg- Kerrey- BreauxGrassley

? Reduces current law benefits by reducing the COLA, increasing the NRA to
67 sooner than under current law, and increasing the benefit computation
period. d ? Affects current- law benefits of both retired and disabled
workers by changing the primary insurance

amount (PIA) formula to increase the progressivity of this formula. e
Additional reductions in PIA formula factors apply only to retired worker
benefits.

? Restructures benefits by creating a mandatory individual “carve-
out” account financed by redirecting 2 percent of individuals' taxable
earnings. f ? At retirement age, insurance benefits for both retired and
disabled workers are reduced or offset by an

amount equal to the contributions plus interest that would have accrued had
these contributions been invested at the interest rate earned by the OASDI
trust funds. g

? Provides a minimum benefit for newly eligible aged surviving spouses of
retired workers. Kolbe- Stenholm ? Reduces current- law benefits by reducing
the COLA, increasing the NRA to 67 sooner than under

current law and indexing it to changes in average life expectancy, and
increasing the benefit computation period. h ? Reduces benefits for both
retired and disabled workers by modifying the PIA formula. Additional
reductions in benefits because of changes in longevity apply only to retired
workers. Requires a report to

the Congress in 2001 that may recommend similar reductions in DI benefits.
Formula changes may increase the progressivity of the benefit structure.

? Restructures benefits by creating a mandatory individual “carve-
out” account financed by redirection of 2 percent of individuals'
taxable earnings. i ? Provides a minimum benefit for newly eligible retired
and disabled beneficiaries.

? Insurance benefits are not reduced or offset by individual account
balances. j a The Kasich proposal does not explicitly refer to DI
beneficiaries.

b Because over time increases in wages have been, and are expected to
continue to be, greater than increases in prices, indexing the benefit
formula to prices rather than to wages would reduce initial benefits. Under
current law, once the beneficiary receives the benefit, further increases in
benefits are

based on changes in the CPI. These increases in benefits are affected by
changes in the COLA. c Disabled workers who choose the account option have
the additional benefit reduction when they receive DI benefits, according to
Congressman Kasich's staff.

d Current and near retirees are excluded from the reduction in the COLA. NRA
increases have no effect on benefit levels for those who receive disability
benefits up to the age they are eligible for retirement benefits. Disabled
beneficiaries are exempt from increases in the benefit computation period. e
The PIA is the monthly amount payable to a retired worker who begins to
receive benefits on reaching the NRA or to a disabled worker who has never
received a retirement benefit reduced for age. f The proposal also allows
additional voluntary contributions up to $2,000 annually. Lower- income and

middle- income earners are also eligible for a partial match. These features
were not included in our analysis. g KidSave accounts are established for
each child at birth with government contributions financed from the general
fund; these continue until the child is 5 years old. Half of the KidSave
contributions are included in calculating the offset when the individual
with the KidSave account retires. This feature is

not included in our analysis because the KidSave accounts begin with the
cohort born in 2000, whereas we analyze the cohorts born in 1986 and
earlier. h NRA increases have no effect on benefit levels for persons who
receive disability benefits up to the age when they are eligible for
retirement benefits. Disabled beneficiaries are exempt from increases in the
benefit computation period.

i The proposal also allows additional voluntary contributions up to $2,000
annually. Lower- income earners are also eligible for a partial government
match and may use the earned income tax credit to contribute. These features
were not included in our analysis.

j The worker may purchase an annuity or request a monthly pay- out. If the
monthly pay- out plus Social Security benefits guarantees a lifetime income
equal to the poverty level, then the balance in excess of this requirement
may be withdrawn.

The proposals we studied vary in the degree to which they explicitly refer
to disabled beneficiaries. President Clinton's proposal refers to
maintaining current- law benefits for both retired and disabled workers. The
ArcherShaw proposal implicitly refers to both disabled and retired workers
when it states that beneficiaries will be guaranteed at least current- law
benefits. However, it explicitly refers to disabled workers when it
discusses distributions from the IAs. Workers can receive distributions from
their IAs when they become entitled to either DI or OASI benefits.

The Kasich proposal does not explicitly refer to disabled beneficiaries when
discussing changes in benefits or the establishment of IAs, although
disability benefits are affected by the provisions in the Kasich proposal.
Rather, it emphasizes that the provisions described will not affect the
benefits of retired workers or those near retirement. The discussion of the
expected returns to the IAs clearly refers only to retired workers, with
their longer work history. Most of the provisions in the Gregg- Kerrey-
Breaux- Grassley and KolbeStenholm

proposals explicitly refer to disabled or retired workers. Under both
proposals, the benefits of disabled workers are affected by one reduction in
the PIA formula but are exempted from a second reduction. Benefits of both
disabled and retired workers are affected by reductions in the COLA.
However, the provision in both proposals that increases the

benefit computation period amends a clause in the Social Security Act that
refers only to retired workers. The provision increasing the retirement age
affects only the benefits of retired workers. Under both proposals, the
restrictions on IA distributions refer to receipt either at retirement age
or at the attainment of a particular level of funds in the IA. Under the
GreggKerrey-

Breaux- Grassley proposal, the insurance benefit is reduced by an offset
related to the amount of contributions to the IA. DI beneficiaries are
exempt from this adjustment to the insurance benefit when benefits are first
received. However, at retirement age, when they are able to gain

access to the income from their IAs, insurance benefits are reduced by the
appropriate offset.

Social Security Reform Estimates by SSA's Office of the Chief Actuary
indicate that all the Is Likely to Improve DI

proposals would improve the solvency of the combined DI and OASI trust
funds, with the extent of the improvement varying across proposals. In Trust
Fund Solvency addition, most of the specific provisions in the proposals,
such as transfers from general funds and reductions in benefit levels, would
have a positive effect on the solvency of the DI trust fund. However, a
provision such as the increase in the retirement age would have a negative
effect on the DI trust fund while at the same time improving the OASI trust
fund balance.

The Reform Proposals The reform proposals we studied had a range of effects
on the trust funds' Differ in Their Effect on solvency as measured by the
actuarial balance. The actuarial balance as Solvency

calculated by the Office of the Chief Actuary is the difference between the
present value of the Social Security program's revenues and costs over a
75year period and is expressed as a percentage of taxable payroll. 16 If
revenues exceed costs, the actuarial balance is positive; if costs exceed
revenues, the actuarial balance is negative, indicating a deficit. In 1999,
under current law, the Social Security program faced an actuarial deficit
equal to 2. 07 percent of taxable payroll. 17 This figure represents the
amount of the payroll tax rate increase in 1999 that would establish
actuarial balance in the Social Security trust funds over the subsequent 75
years. In

other words, increasing the payroll tax rate from the current 12. 4 percent
to 14. 47 percent of payroll would establish actuarial balance in the trust
funds.

The Office of the Chief Actuary provides annual estimates of the actuarial
balance for the combined OASI and DI trust funds under current law and, when
requested by the Congress or the executive branch, estimates of the 16 The
Social Security program's revenues include the sum of the trust fund balance
at the

beginning of the period plus the total income during the period. The costs
include the outgo during the period plus the targeted trust fund level at
the end of the period equal to the following year's outgo.

17 This is the actuarial deficit as presented in the 1999 Trustees Report,
based on the intermediate cost assumptions. We used it because the Office of
the Chief Actuary scored the proposals we analyzed using the assumptions of
the 1999 Trustees Report. According to the 2000 Trustees Report, the
actuarial deficit is 1.89.

actuarial balance under reform proposals. 18 The estimates of the actuarial
balance under current law and each of the reform proposals we studied are
presented in table 4. The actuaries estimated that the trust funds' deficit
of 2.07 percent of taxable payroll under current law would either be sharply
reduced or become a surplus for the combined trust funds under the reform
proposals we studied. A surplus in the combined trust funds could mean a
surplus in one trust fund and a deficit in the other. However, a
reallocation of payroll tax rates between the two funds would be expected in
this case.

Table 4: Actuarial Balance of Social Security Trust Funds Under Current Law
and Selected Reform Proposals

OASI and DI trust funds

Current law 2.07% President's proposal 0.80 Kasich +0.00/+ 0.24 a Kolbe-
Stenholm +0.07 Archer- Shaw +0.09 b Gregg- Kerrey- Breaux- Grassley +0.28 a
Under the Kasich proposal, the IA is voluntary. If everyone opts for the IA,
the actuarial balance is 0. If no one opts for the IA, the actuarial balance
is 0.24.

b Under the Archer- Shaw proposal, the payroll tax rate would be reduced
from 12.4 to 9.9 in 2050 and 8.9 in 2060. Source: Office of the Chief
Actuary, memorandums to the chief actuary on the effects of selected Social
Security reform proposals. Estimates are based on the 1999 Trustees Report.

The President's proposal would reduce the actuarial deficit but is not
expected to eliminate it. This would be achieved through general fund
transfers every year from 2011 to 2050 and by allowing some limited
investment in equities, which have a higher rate of return than do the
government bonds in which the trust funds have traditionally been

18 The Office of the Chief Actuary does not necessarily provide estimates of
the actuarial balance under the reform proposals separately for the DI and
the OASI trust funds. For major solvency proposals, the OASDI program is
generally considered on a combined basis, with the presumption of
reallocation of the tax rates as needed.

invested. 19 Estimates for two other proposals result in a small surplus for
the combined trust funds. The Kolbe- Stenholm proposal would generate its
surplus through benefit cuts and general fund transfers. 20 In the
ArcherShaw proposal, general fund transfers would finance the contributions
to the IAs that the proposal would establish. The proceeds from these

accounts would be transferred to the trust funds when benefits are received.
The proposal also calls for reducing payroll taxes in response to the
additional trust fund revenue expected to accrue from the proceeds of these
IAs. The other proposals we examined would result in larger estimated

actuarial surpluses for the combined trust funds. The Kasich proposal would
accomplish this by reducing the initial level of insurance benefits and then
further decreasing insurance benefits by a fixed percentage for each year of
contribution to an IA, as well as by borrowing from the general fund. The
Gregg- Kerrey- Breaux- Grassley proposal would achieve its surplus through a
mix of benefit cuts and revenue transfers that would offset the loss of
trust fund revenues resulting from the redirection of a portion of the
payroll taxes to the IAs.

The reform proposals we studied differ in the magnitude of the stipulated
transfers from general revenue. Transfers are smaller under the
KolbeStenholm and Gregg- Kerrey- Breaux- Grassley proposals, which contain a
number of provisions to achieve solvency by changing benefits or revenues.
Under the Kolbe- Stenholm proposal, general revenue transfers range from 0.
03 percent of taxable payroll in 2000 to 0.80 percent of taxable

payroll in 2060. General revenue transfers under the Gregg- Kerrey-
BreauxGrassley proposal range from 0.6 percent of taxable payroll in 2000 to
1.2 percent of taxable payroll in 2060. General revenue transfers are larger
under the proposals with fewer alternative provisions for attaining
solvency. Under the Kasich proposal, for example, the magnitude of the
transfers ranges from 1.17 percent of taxable payroll in 2000 to 1.57
percent of taxable payroll around 2030. 21 Under the President's proposal,
transfers

range from a high of 2.41 percent of taxable payroll to a low of 0.52
percent 19 The general fund means the accounts for receipts not earmarked by
law for a specific purpose, the proceeds of general borrowing, and the
expenditure of this money. Transfers between the general fund and the trust
funds do not affect the surplus or deficit in the government's unified
budget.

20 These benefit cuts and revenue transfers would offset the loss of revenue
to the trust funds that would result from the IAs.

of taxable payroll between 2011 and 2050. Finally, the Archer- Shaw proposal
calls for a general revenue transfer equal to 2 percent of taxable payroll
beginning in 2000.

Most, but Not All, Proposal Although most provisions in the proposals we
examined potentially have a

Provisions Would Affect the positive effect on the solvency of the DI trust
fund, some provisions would

DI Trust Fund Positively have a negative effect. The President's proposal
has two provisions- the transfer of funds from general revenue to the
combined OASI and DI trust funds and the investment of a portion of these
funds in equities. According

to the Office of the Chief Actuary, both provisions would be expected to
have a positive effect on the solvency of the DI trust fund. The Archer-
Shaw proposal calls for a gradual transfer of the income from the IA
balances, which are financed from general revenue, to the trust funds. In
the case of disabled workers, the income from the IA balances would be
transferred to the DI trust fund. This provision also would have a positive
effect on the DI trust fund. The Kasich proposal contains three provisions
that would have a positive effect on DI trust fund solvency: the indexing of
benefits to prices rather than to wages, which reduces benefits; the
reduction in benefits for individuals who opt to contribute a portion of
their payroll tax to an IA; and the borrowing of funds from general revenue.
22 The loss of payroll tax revenue associated with individuals opting for
IAs would increase the DI trust fund's deficit, and the general fund loans
are designed to compensate for this. 23

Both the Kolbe- Stenholm and the Gregg- Kerrey- Breaux- Grassley proposals
contain multiple provisions that would affect DI trust fund solvency.
Provisions that reduce the COLA and change the PIA formula so as to reduce
benefits for disabled workers lower program costs and, therefore, improve
the actuarial balance for the DI program. However, provisions such as the
redirection of payroll taxes to IAs and the establishment of a minimum
benefit have potentially a negative effect on the DI trust fund.

21 Under the Kasich proposal, this general revenue transfer is considered a
loan that will be paid off beginning in 2060. 22 Although the Kasich
proposal contains a provision describing a loan from general revenue, this
loan is, in effect, very similar to a general revenue transfer. There would
be no principal repayment or interest repayment on this loan until 60 years
into the future.

23 For the version of the Kasich proposal in which all individuals are
assumed to opt for the IA, the version we analyze in this report, the Office
of the Chief Actuary estimates only the effect of the entire proposal, not
the effects of the individual provisions.

Redirecting payroll taxes reduces revenues to the trust fund while
establishing a minimum benefit increases program costs for beneficiaries who
were receiving benefits below the minimum. Even provisions that appear to be
focused on retirement benefits can have an effect on the DI

trust fund. For example, increasing the retirement age also increases the
age at which disability insurance benefits are converted to retirement
insurance benefits. 24 As a result, disability beneficiaries remain on the
DI program longer, increasing costs to the DI program. 25 We were able to
use the SSASIM model to estimate the effects on solvency of certain of the
provisions in the reform proposals. Our estimates using the model are based
on the intermediate assumptions reported in the 1999 Social Security
Trustees Report because the SSA's Office of the Chief Actuary used these
assumptions to score the Social Security reform proposals we analyzed. Table
5 presents our results. Reductions in benefits have a positive effect on DI
trust fund solvency. The increase in the

retirement age results in the expected negative effect on solvency of the DI
trust fund.

24 Also, the resulting decrease in OASI benefit levels might cause some
individuals to apply for DI benefits rather than waiting to retire and apply
for OASI benefits, as they would otherwise do. If such an increase in
applications were to result in an increase in DI beneficiaries, there could
be an increase in costs to the DI program and a resulting increase in the
actuarial deficit. According to recent estimates that SSA's Office of the
Chief Actuary has made, this effect is relatively small.

25 Social Security Reform: Implications of Raising the Retirement Age (GAO/
HEHS- 99- 112, Aug. 27, 1999).

Table 5: Estimated Change in Current Law Actuarial Balance for Selected
Reform Provisions

OASI DI Actuarial balance

Current law 1.70% 0.38%

Change in actuarial balance

Reduction in COLA (Gregg- Kerrey- Breaux- Grassley) +0.58 +0.06 Indexing
benefits to prices (Kasich) +1.90 +0.40 Raising retirement age (Gregg-
Kerrey- Breaux- Grassley) +0.18 0.01 Source: GAO estimates using the SSASIM
model. These estimates are consistent with independent estimates produced by
the Office of the Chief Actuary. Certain Reform

Two reform proposals we studied either maintain current- law benefits-
Proposals Could the President's proposal- or guarantee that the beneficiary
would receive at least the amount of current- law benefits- the Archer- Shaw
proposal. Reduce Benefit Income

The remaining three reform proposals- Kasich, Kolbe- Stenholm, and for
Disabled Gregg- Kerrey- Breaux- Grassley- would affect the levels of
insurance Beneficiaries

benefits DI and OASI beneficiaries receive by changing the PIA formula for
calculating initial benefits, reducing the COLA, raising the retirement age,
or increasing the number of years of earnings used in computing benefits.
How a beneficiary's total benefit income (reduced insurance benefits plus IA
income) under these three proposals compares with the benefits received
under a maintain- benefits scenario or a maintain- tax- rates scenario
depends both on the extent of the decrease in the insurance

benefits and on the amount of income received from the IA. Our maintain-
benefits scenario achieves solvency through increased payroll taxes while
current- law benefits are maintained. 26 Our maintain- taxrates scenario
achieves solvency through benefit reductions while holding current payroll
tax rates at today's levels. These two scenarios represent a range of
benefit levels, with the maintenance of current- law benefits being

26 The maintain- benefits scenario could also be funded through transfers
from the general fund to the trust fund. For our analysis of the effect of
Social Security reform on benefits, the funding source is not material.
However, the source of funding- whether revenue transfers or payroll tax
increases- will have implications for individuals' net earnings and for the
economy as a whole.

at the upper end and the reduced benefits necessary for the maintenance of
current payroll taxes being at the lower end. We compared the benefit income
received under each of the three proposals with that received under the
maintain- benefits scenario and the maintain- tax- rates scenario for each
of three beneficiary groups with the selected characteristics that we
simulated: disabled workers, dependents of disabled workers (including
spouses, children younger than 18, and

adult disabled children), and adult disabled children who are dependents of
retired workers. 27 We made the comparisons under each of several different
assumptions about the year in which the worker was born, the worker's
earnings level, and the worker's age when the worker first received DI
benefits. 28 We chose the ages of initial benefit receipt to reflect SSA
data indicating that individuals are receiving DI benefits at younger ages.
29 For the IAs in our analysis, we assumed that individuals would have
portfolios with a smaller percentage invested in equities as they got older.
We assumed the return on equities would be a constant, inflation- adjusted 7

percent per year, which reflects the long- term historical average return on
equities.

Some Reform Proposals According to our estimates, the disabled beneficiaries
with the selected

Would Reduce Benefit characteristics we simulated would, in general, receive
higher benefits Income

under the maintain- benefits scenario than they would under the Kasich,
Kolbe- Stenholm, or Gregg- Kerrey- Breaux- Grassley proposals. Figures 1 and
2 present the results for workers as well as their dependents. The

workers were born in 1986 and have low or average earnings and, in the 27 In
the following analysis, we use the present value of all lifetime benefit
income received by the beneficiary to compare proposals and scenarios. The
present value of benefit income is the equivalent value, at a point in time,
of the entire stream of insurance benefits and IA income the individual
receives in his or her lifetime. In our analysis, lifetime benefit income
was discounted using the Treasury bond rate. 28 See app. II for a discussion
of these assumptions. 29 We focus on individuals who begin receiving DI
benefits at age 45 for the following reasons. The average age of new
beneficiaries in 1999 was slightly younger than 50, compared with 30 years
ago, when the average age was consistently above 50. Further, most
terminations of benefits result from death or retirement. Thus,
beneficiaries are staying in the program longer, and it is important to
understand how Social Security reform affects these younger disabled
workers. Additional information on the characteristics of DI beneficiaries
is in app. II.

case of disabled workers, first receive DI benefits at the age of 45 and
never work again.

Figure 1: Lifetime Benefit Income of Low- Earnings Worker and Dependents
Under Selected Proposals and Alternative Solvency Scenarios

Present Value of Benefit Income (Dollars in Thousands)

Disabled Worker 450

Disabled Worker Dependent 400

Adult Disabled Child 350 300 250 200 150 100

50 0

Maintain Benefits Gregg- Kerrey

Kasich Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: Assumes 1986 birth cohort. Disabled worker receives DI first at age 45
and never works again. Adult disabled child benefit is based only on retired
worker's benefit. The alternative solvency scenarios are the maintain-
benefits scenario, which achieves solvency through increased payroll taxes
while current- law benefits are maintained, and the maintain- tax- rates
scenario, which achieves solvency through benefit reductions while
maintaining current payroll tax rates.

Figure 2: Lifetime Benefit Income of Average- Earnings Worker and Dependents
Under Selected Proposals and Alternative Solvency Scenarios

Present Value of Benefit Income (Dollars in Thousands)

Disabled Worker 700

Disabled Worker Dependent Adult Disabled Child 600

500 400 300 200 100

0 Maintain Benefits

Gregg- Kerrey Kasich Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: Assumes 1986 birth cohort. Disabled worker receives DI first at age 45
and never works again. Adult disabled child benefit is based only on retired
worker's benefit. The alternative solvency scenarios are the maintain-
benefits scenario, which achieves solvency through increased payroll taxes
while current- law benefits are maintained, and the maintain- tax- rates
scenario, which achieves solvency through benefit reductions while
maintaining current payroll tax rates.

These reform proposals would reduce insurance benefits while providing
income from the IAs. Under these proposals, it is possible that the IA
income might compensate for the decline in insurance benefits resulting from
other provisions. However, this is less likely for disabled- worker
beneficiaries than for retired- worker beneficiaries because disabled
workers are likely to have shorter work histories and thus have smaller IA

balances. 30 The reductions in benefits resulting from the decline in the
COLA and the changes in the PIA formula are so great that the income from 30
In addition, under the Gregg- Kerrey- Breaux- Grassley and Kolbe- Stenholm
proposals, disabled workers would in most cases not be able to gain access
to the income from their IA accounts until they reached retirement age.
Therefore, depending on the age when they begin receiving DI benefits, they
could be receiving reduced insurance benefits for a long time before
receiving IA income. See table 7 in app. I and accompanying text for a
fuller discussion of this issue.

the IA would be insufficient to completely compensate for this loss for the
disabled- worker beneficiaries with the selected characteristics that we
examined. Disabled workers with low earnings and their dependents would
receive greater benefit income under the Gregg- Kerrey- Breaux- Grassley
proposal than under the maintain- benefits solvency scenario. However, this
higher benefit income is largely the result of changes in the PIA formula
that increase the progressivity of the benefit structure. 31 For the
proposals we examined, we included the income from the IA only in the
benefit income of the disabled or retired worker, not in that of the

worker's dependents, since apportioning the IA income among family members
is an individual matter and would vary by household. 32 Consequently,
benefit income for dependents of disabled or retired

workers would be reduced under the Gregg- Kerrey- Breaux- Grassley, Kasich,
and Kolbe- Stenholm proposals not only because of reductions in the
insurance benefit but also because it does not include income from
individual accounts. In addition, the insurance benefits of dependents

include only the amount received during the years in which the worker on
whose earnings record the benefits are payable is receiving insurance
benefits.

Contrary to the results of our comparisons with the maintain- benefits
scenario, in our comparison of each of the three proposals with the
maintain- tax- rates scenario, we found that in most cases the beneficiary
would receive higher benefit income under the proposals than under the
scenario. However, dependents of low- earner disabled workers under the
Kasich proposal would receive benefit income that is less than under the

maintain- tax- rates scenario. Also, adult disabled children of retired
workers would receive somewhat lower benefit income under all three

31 Adult disabled children who are dependents of retired workers also
benefit from these changes in the PIA formula that increase the
progressivity of the benefit structure. However, adult disabled children are
also, as OASI beneficiaries, subject to another PIA provision in the
proposal that decreases benefits.

32 In this report, we do not consider the effect on benefits resulting from
the application of the maximum family benefit. For further discussion of
this issue, see app. II.

proposals in almost all cases. 33 These results are presented in figures 1
and 2. The benefit income received under the three proposals would generally
be greater than the benefits received under the maintain- tax- rates
solvency scenario because the proposals have provisions for achieving
solvency, such as general revenue transfers, in addition to reducing
benefits. As a result, the insurance benefits would not have to decline as
much as in the maintain- tax- rates scenarios. Further, the benefit income
workers would receive under the proposals includes income from IAs.

Most Proposal Provisions We also examined individual provisions within the
three proposals to Would Reduce Insurance

assess their contribution to the change in the level of insurance benefits
Benefits

received. Reductions in the COLA instituted under the Gregg- KerreyBreaux-
Grassley and Kolbe- Stenholm proposals would decrease insurance benefits
relatively little compared with the maintain- benefits scenario for both
disabled workers and their beneficiaries and for adult disabled children of
retired workers. 34 Figure 3 presents the estimated effects of the decrease
in the COLA on workers born in 1986 who first receive disability benefits at
the age of 45 and never work again. The pattern of change in the present
value of benefit income for dependents of disabled workers and for adult
disabled children who are dependents of retired workers is similar to that
shown in figure 3 for disabled workers.

33 Under the Gregg- Kerrey- Breaux- Grassley and Kolbe- Stenholm proposals,
the insurance benefits of adult disabled children are subject to the same
reductions as those of disabled beneficiaries as well as to an additional
PIA reduction from which disabled beneficiaries are exempt. The Kasich
proposal adjusts the PIA formula so that benefits at all earnings levels are
reduced. Under the maintain- tax- rates scenario, we adjust the PIA formula
so that the benefits at the lowest levels are not reduced; only the benefits
at the higher levels are

reduced. Consequently, the benefits for lower earners would be smaller under
the Kasich proposal than under the maintain- tax- rates scenario. For a more
detailed description of how we adjusted the PIA formula, see app. II. 34 The
difference across proposals in the magnitude of the effect reflects the
extent of reduction specified- 0.33 percentage point in the Kolbe- Stenholm
proposal and 0. 5

percentage point in the Gregg- Kerrey- Breaux- Grassley proposal. The effect
of a COLA reduction is compounded over time. The cumulative effect of a 0.5
percentage point reduction over a period of 30 years is 13. 5 percent.

Figure 3: Comparison of COLA Changes to Lifetime Insurance Benefits of
Disabled Workers With Alternative Solvency Scenarios

Present Value of Insurance Benefits (Dollars in Thousands)

Average Earner 700

Low Earner 600 500 400 300 200 100

0 Maintain Benefits

Gregg- Kerry Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: Assumes disabled worker born in 1986, receives DI first at age 45, and
never works again. The Kasich proposal is not included because it does not
have a COLA provision. The alternative solvency scenarios are the maintain-
benefits scenario, which achieves solvency through increased payroll taxes
while current- law benefits are maintained, and the maintain tax- rates
scenario, which achieves

solvency through benefit reductions while maintaining current payroll tax
rates.

Changes in the PIA formula, however, generally result in large reductions in
insurance benefits relative to the maintain- benefits scenario. The one
exception is a provision of the Gregg- Kerrey- Breaux- Grassley proposal
that would increase benefits for workers with certain levels of earnings,
thereby

increasing benefits for low earners and decreasing benefits by a relatively
smaller amount for average earners. Figure 4 displays the effects on
disabled workers of changes in the benefit calculation formula. The pattern
in the present value of benefit income for the two other categories of
beneficiaries is similar to that shown in figure 4 for disabled workers.

Figure 4: Comparison of PIA Changes to Lifetime Insurance Benefits of
Disabled Workers With Alternative Solvency Scenarios

Present Value of Insurance Benefits (Dollars in Thousands)

Average Earner 700

Low Earner 600 500 400 300 200 100

0 Maintain Benefits

Gregg- Kerrey Kasich Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: For PIA changes that affect both disabled and retired workers. Assumes
disabled worker born in 1986, receives DI first at age 45, and never works
again. The alternative solvency scenarios are the maintain- benefits
scenario, which achieves solvency through increased payroll taxes while
current- law benefits are maintained, and the maintain tax- rates scenario,
which achieves solvency through benefit reductions while maintaining current
payroll tax rates.

The insurance benefits of adult disabled children who are dependents of
retired workers would also be significantly decreased by an additional
change in the PIA formula applicable only to OASI benefits under the Gregg-
Kerrey- Breaux- Grassley and Kolbe- Stenholm proposals. Figure 5 displays
the effects on the insurance benefits of adult disabled children resulting
from this PIA change. 35

35 We also analyzed the effect on benefits of increases in the NRA and in
the number of years of earnings used to compute benefits. These provisions
affect only the benefits of retired workers and their dependents. The
provision increasing the NRA had little effect on benefits because the NRA
(67) is the same for the 1986 birth cohort under the maintain- benefits
scenario and the Gregg- Kerrey- Breaux- Grassley proposal and only slightly
higher under the Kolbe- Stenholm proposal. The provision increasing the
number of years in the benefit calculation had little effect because we
assumed the retired worker was steadily employed with no years of low or no
earnings.

Figure 5: Comparison of PIA Changes to Lifetime Insurance Benefits With
Alternative Solvency Scenarios for Adult Disabled Child Dependent of Retired
Worker Present Value of Insurance Benefits (Dollars in Thousands)

Average Earner 120

Low Earner 100

80 60 40 20

0 Maintain Benefits

Gregg- Kerry Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: For PIA changes that affect only retired workers. Assumes retired
worker born in 1986. Kasich proposal does not have PIA changes that affect
only retired workers. The alternative solvency scenarios are the maintain-
benefits scenario, which achieves solvency through increased payroll taxes
while current- law benefits are maintained, and the maintain tax- rates
scenario, which achieves

solvency through benefit reductions while maintaining current payroll tax
rates.

As we stated earlier and as is shown in figure 3, reductions in the COLA
result in relatively small declines in the level of current- law benefits.
Consequently, the levels of insurance benefits that would be received under
this provision would be greater than the benefit income received under the
maintain- payroll- tax- rates scenario, in which benefits would be reduced
to levels supportable by current payroll tax rates. Despite the large
reductions in insurance benefits resulting from the changes in the PIA
formula, most disabled beneficiaries would be better off under this
provision in the

proposals than under the maintain- tax- rates scenario. The exception occurs
for all three types of low- earner beneficiaries under the Kasich proposal's
change in the PIA formula. This provision in the Kasich proposal indexes
initial benefits to prices rather than to wages, resulting in a sharp
decline in benefits. The effects of the PIA changes on the disabled worker
are shown in figure 4.

Provisions for Individual According to our estimates, the effect on the
disabled worker's benefit

Accounts Would Slightly income of the IA provision alone is positive under
the Gregg- KerreyBreaux-

Grassley, Kasich, and Kolbe- Stenholm proposals. 36 Benefit income Increase
Benefit Income for

would increase the most under the Kolbe- Stenholm proposal because the
Disabled Workers

IA income does not reduce insurance benefits. Benefit income under the
Gregg- Kerrey- Breaux- Grassley proposal would also increase but by less
because the proposal reduces insurance benefits by an amount that reflects
the present value of the government contributions to the IA plus the
interest that would have accrued had these contributions been invested at
the interest rate earned by the OASDI trust funds. 37 The benefit income
received under the Kasich proposal would also be less than that received

under the Kolbe- Stenholm proposal because under the Kasich proposal
insurance benefits would be reduced by a fixed percentage for each year of
contributions to the IA. 38 Figure 6 shows the effect of the IA provision
for

both the low- earning and the average- earning disabled worker. 36 Our
analysis assumes that the IA balance is used up by the account holder during
his or her lifetime so that the treatment of the balance at death is not an
issue. 37 DI beneficiaries are subject to this offset when they gain access
to their IA balances, which for most DI beneficiaries would occur at
retirement age. (See app. I.) Until the beneficiary reaches retirement age,
his or her benefits are unaffected by the balance in the IA.

38 IAs under the Kasich proposal would be voluntary. Our analysis looked at
beneficiaries who opt for the IA because we were asked to analyze the
effects of IA provisions in the proposals.

Figure 6: Lifetime Benefit Income of Disabled Workers Under IA Provisions
and Alternative Solvency Scenarios

Present Value of Benefit Income (Dollars in Thousands)

Average Earner 700

Low Earner 600 500 400 300 200 100

0 Maintain Benefits

Gregg- Kerrey Kasich Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: Assumes disabled worker born in 1986, receives DI first at age 45, and
never works again. Under the Gregg- Kerrey- Breaux- Grassley and Kasich
proposals, the insurance benefit is reduced by an amount reflecting the
contribution to the IA. We assume all IA income is received by the disabled
worker and not by dependents. The alternative solvency scenarios are the
maintain- benefits scenario, which achieves solvency through increased
payroll taxes while current- law benefits are maintained, and the maintain
tax- rates scenario, which achieves solvency through benefit reductions
while maintaining current payroll tax rates.

In our analysis, we assigned the income from the IA to the disabled or
retired worker, not to the worker's dependents, because the apportionment of
the IA income among family members is an individual matter and would vary by
household. Thus, our estimates reflect the most that the worker would
receive from the IAs, whereas our estimates for the dependents reflect the
most that their benefits would be reduced under these proposals.
Accordingly, for the dependent of the disabled worker and for the adult
disabled child, the IA will not increase benefit income because we assumed
that these beneficiaries, unlike the worker, receive no income from the IA.
Under the Kolbe- Stenholm proposal, there would be no

reduction in the benefit income of dependents because changes in IA income
do not affect the level of insurance benefits. However, the Kasich proposal
would decrease the insurance benefit of the worker by a set percentage for
each year of contributions to the IA. The Gregg- KerreyBreaux- Grassley
proposal would reduce the insurance benefit of the

worker by an amount that reflects the present value of the government
contribution to the IA plus the interest that would have accrued had these
contributions been invested at the interest rate the OASDI trust funds earn.
The insurance benefit that the dependent receives is a proportion of what
the worker receives. Consequently, the insurance benefit that the

dependent receives would be reduced under our assumption that dependents
receive no compensating income from the IA under the GreggKerrey- Breaux-
Grassley and Kasich proposals. (See figures 7 and 8.) 39

Figure 7: Lifetime Benefit Income of Dependent of Disabled Worker Under IA
Provisions and Alternative Solvency Scenarios

Present Value of Benefit Income (Dollars in Thousands)

Average Earner 350

Low Earner 300 250 200 150 100

50 0

Maintain Benefits Gregg- Kerrey

Kasich Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: Assumes disabled worker born in 1986, receives DI first at age 45, and
never works again. Under the Gregg- Kerrey- Breaux- Grassley and Kasich
proposals, the insurance benefit is reduced by an amount reflecting the
contribution to the IA. We assume all IA income is received by the disabled
worker and not by dependents. The alternative solvency scenarios are the
maintain- benefits scenario, which achieves solvency through increased
payroll taxes while current- law benefits are maintained, and the maintain-
tax- rates scenario, which achieves solvency through benefit reductions
while maintaining current payroll tax rates.

39 See app. I for how the IAs differ across proposals. The effect of IAs on
lifetime benefit income also depends on the legislated characteristics of
the IA (for example, how much can be invested and in what types of assets)
and the assumptions about the return on investments, administrative costs,
and so on.

Figure 8: Lifetime Benefit Income of Adult Disabled Child Dependent of
Retired Worker Under IA Provisions and Alternative Solvency Scenarios

Present Value of Benefit Income (Dollars in Thousands)

Average Earner 120

Low Earner 100

80 60 40 20

0 Maintain Benefits

Gregg- Kerrey Kasich Kolbe- Stenholm Maintain Tax Rates Breaux- Grassley

Proposals/ Scenarios

Note: Assumes retired worker born in 1986. Under the Gregg- Kerrey- Breaux-
Grassley and Kasich proposals, the insurance benefit is reduced by an amount
reflecting the contribution to the IA. We assume all IA income is received
by the retired worker and not by dependents. The alternative

solvency scenarios are the maintain- benefits scenario, which achieves
solvency through increased payroll taxes while current- law benefits are
maintained, and the maintain tax- rates scenario, which achieves solvency
through benefit reductions while maintaining current payroll tax rates.

The IA provision would increase benefit income for disabled workers compared
with the maintain- benefits scenario. Consequently, the benefit income of
the disabled workers we examined would also be greater than the benefits
available under the maintain- tax- rates scenario. In our

analysis, we assigned all the IA income to the disabled or retired worker
and none to the worker's dependents. As a result, under the Gregg-
KerreyBreaux- Grassley and Kasich proposals, dependents would experience the
reduction in insurance benefits related to the existence of an IA but would

not receive any compensating income from the IA, under our assumptions.
However, even the reduced insurance benefits that dependents would receive
would be greater than the benefits they would receive under the maintain-
tax- rates scenario.

In the analysis presented so far, we have provided graphs showing the effect
of the reform proposals on the worker who first receives DI benefits at the
age of 45 and never works again. 40 However, the income from the IA

is affected by the number of years for which contributions are made to the
IA and, therefore, by the age at which the worker leaves the labor force and
begins receiving DI benefits. To see how the income received from the IA
would vary by age of first receipt of DI benefits, we compared the income
received from the IA by workers who began receiving DI benefits at

different ages. Figures 9 and 10 provide the net addition of the IAs to
benefit income- that is, the addition to benefit income after reductions are
made in the insurance benefit in response to the income from the IA. Figures
9 and 10 indicate that the income received from the IA increases with the
age of first receipt of DI benefits. The later that DI benefits are

received, the greater the number of years in the labor force, the number of
years funds are deposited in the IA, and the number of years the IAs accrue
compound interest. The addition of IA income to benefit income across ages
would be greatest under the Kolbe- Stenholm proposal, which would not reduce
the insurance benefit in response to the income received from the IA. The
Gregg- Kerrey- Breaux- Grassley proposal would reduce the insurance benefit
by an amount that reflects the present value of the government contributions
to the IA plus the interest that would have accrued had these contributions
been invested at the interest rate earned by the OASDI trust funds. The
Kasich proposal would reduce the insurance benefit by one- third of a
percent for each year of participation in the IA.

40 We assumed in our analysis that the worker remains in the labor force
until he or she begins receiving DI benefits.

Figure 9: Net Addition of IA to Current- Law Lifetime Benefit Income Under
Selected Reform Proposals: Low- Earnings Worker Disabled at Various Ages

Present Value of Benefit Income (Dollars in Thousands)

40,000 Disabled at 35

35,000 Disabled at 45

Disabled at 55 30,000

25,000 20,000 15,000 10,000

5,000 0

Gregg- Kerrey- Breaux- Grassley Kasich Kolbe- Stenholm Reform Proposals

Note: Assumes 1986 birth cohort. Under the Gregg- Kerrey- Breaux- Grassley
and Kasich proposals, the insurance benefit is reduced by an amount
reflecting the contribution to the IA. We assume all IA income is received
by the disabled worker, not by dependents.

Figure 10: Net Addition of IA to Current- Law Lifetime Benefit Income Under
Selected Reform Proposals: Average- Earnings Worker Disabled at Various Ages

Present Value of Benefit Income (Dollars in Thousands)

90,000 Disabled at 35

80,000 Disabled at 45

Disabled at 55 70,000

60,000 50,000 40,000 30,000 20,000 10,000

0 Gregg- Kerrey- Breaux- Grassley

Kasich Kolbe- Stenholm Reform Proposals

Note: Assumes 1986 birth cohort. Under the Gregg- Kerrey- Breaux- Grassley
and Kasich proposals, the insurance benefit is reduced by an amount
reflecting the contribution to the IA. We assume all IA income is received
by the disabled worker, not by dependents.

Certain Reform Some Social Security reform proposals could increase costs
for the SSI Proposals Could program. Individuals receiving benefits from
both Social Security (DI or OASI) and SSI might become eligible for larger
SSI benefits if their Social Increase SSI Program

Security benefits decrease as a result of reform. In addition, some Social
Costs Security beneficiaries not currently eligible for SSI might become
eligible if their Social Security benefits declined as a result of reform.
As we stated earlier, we estimated that three Social Security reform
proposals- Gregg- Kerrey- Breaux- Grassley, Kasich, and Kolbe- Stenholm-

would lower Social Security benefit income, which includes income from IAs,
in most of the cases we studied. For DI and OASI beneficiaries who also
receive SSI, the decrease in Social Security benefit income would lower
their unearned income, which means that their SSI benefit would

increase. 41 This would have no effect on the number of recipients but would
increase the cost to the program. For the beneficiaries who receive only
Social Security and not SSI, the previously mentioned decrease in benefit
income would lower unearned income, which would make some eligible for SSI
benefits. This would increase both the number of beneficiaries and the cost
to the program. However, the full effect on SSI

would not be felt immediately because most of the individual provisions
within these proposals are to be phased in over time and in many cases are
not to be completely in effect until 2020. Given the complexity of the
interactions between Social Security and SSI and the difficulty of
projecting SSI caseloads so far into the future, it would be extremely

difficult to estimate precisely what the effects of reform proposals would
be on SSI program costs.

Concluding In the cases we studied, our analyses indicate that most disabled
Observations beneficiaries would receive higher benefits under Social
Security reform proposals than under a solvency scenario that maintained
payroll tax rates while reducing benefits. However, most disabled
beneficiaries with the characteristics we studied would receive lower
benefits under reform than

under a solvency scenario that maintained current- law benefits while
raising payroll taxes. This reduction in benefits under reform to levels
below that of current law would occur even though we assumed an optimal set
of conditions for disabled beneficiaries: full- time work until receipt of
DI benefits and low administrative costs and no annuitization costs for the
IAs. Consequently, the typical DI beneficiary could receive lower benefits
than the DI beneficiaries with the selected characteristics we studied.

The proposals we studied treat DI beneficiaries similarly to OASI
beneficiaries. However, the circumstances facing disabled workers differ
from those facing retired workers. For example, the disabled worker's
options for alternative sources of income, especially earnings- related
income, to augment the reduced benefits are likely to be more limited than
are those for the retired worker. Further, DI beneficiaries are entering the
program at younger ages and remaining in the program in most cases until
death or retirement. Thus, disabled beneficiaries could be subject to these

41 If the individual account accumulations are treated as assets with regard
to determining SSI eligibility, some low- income individuals may lose SSI
benefits. Whether this loss of benefits in the case of disabled
beneficiaries would occur when they first received benefits

or when they reached retirement age depends on the proposal.

reductions in benefits for many years. They will also have smaller balances
in their IAs because of fewer working years in which to make IA
contributions and accrue compounded interest. In addition, under several
proposals, disabled beneficiaries cannot gain access to income from
individual accounts until they reach retirement age. These differences
between disabled and retired workers suggest that Social Security reform
proposals should be viewed not only in light of their effects on retired

workers but also explicitly for their effect on disabled beneficiaries and
their families.

Agency Comments We provided a draft of this report to SSA. In commenting on
this report, the agency noted that we addressed an important topic that has
until now received little attention. Specifically, SSA highlighted two
points in our report as being important for policy makers considering
changes to Social Security: that individual accounts might not fully offset
Social Security insurance benefit reductions for some beneficiaries and that
SSI benefits might increase as they compensate for the decline in DI
benefits resulting from Social Security reform. However, the agency had some
concerns about our use of a “best case” scenario to estimate the
effects of policy options and about the assumptions underlying this
“best case” scenario, citing specifically earnings levels, life
expectancy, and investment return assumptions that SSA thought did not
reflect the actual situation of disabled beneficiaries. On the basis of
these concerns, the agency suggested that we give the report balance by
adding a “worst case” scenario. SSA also expressed concern
regarding our focus on lifetime benefits, a measure that it believes does
not adequately reflect living

standards at specific points in time. Finally, SSA suggested that we include
a measure reporting on money's worth or internal rates of return in our
table 1 that compares costs and benefits of Social Security reform
proposals. SSA also made a number of technical comments, which we

incorporated where appropriate. Our use of a “best case”
scenario demonstrated that, even under the best of circumstances, Social
Security reform proposals would reduce current- law benefits to DI
beneficiaries- people who would find it more difficult than most nondisabled
retired workers to replace lost benefits with other sources of income such
as earnings. We did not examine “worse case” scenarios because
the “best case” scenario demonstrates that most DI

beneficiaries would be adversely affected by the reform proposals we
analyzed. While including the “worst case” scenario SSA
suggested could provide a specific lower limit to a range of possible
benefit outcomes, that

lower limit would be useful only if accompanied by an evaluation of the
adequacy of that benefit level, which is beyond the scope of this report. In
building a “best case” scenario, we used the earnings of men
because they tend to have higher earnings than women do. To examine low-
wage

earners, we simulated workers who earn 45 percent of average earners, which
is the standard low level of earnings the Office of the Chief Actuary uses.
Benefits declined at this earnings level as they would for workers earning
even less. We assumed individuals lived until 79 because almost

one- third of individuals first receiving DI benefits at age 45 live that
long, and the number of these individuals is significant enough to warrant
study. With respect to SSA's concern about our use of an equity return of 7
percent, we note that this is a figure currently used in projections,
including

those of the Office of the Chief Actuary. We chose not to adjust for risk
because there is no one risk- adjusted measure that everyone agrees is the
best measure, and we believed that our analysis would be more clearly
understood with the simplifying “best case” assumptions.

With respect to SSA's concern with our focus on lifetime benefits, we
acknowledge that we do not address the issue of variations across plans in
living standards before retirement age resulting from differences in account
access rules. This is certainly an issue on which future reports could
usefully focus. As for the inclusion of money's worth or internal rate of
return measures, we agree that such analysis would be useful, but these

measures are beyond the scope of this report. SSA's written comments are
printed in appendix III. We are sending copies of this report to the
Commissioner of the Social Security Administration and others who are
interested. We will also make copies available to others on request. If you
or your staff have any questions concerning this report, please call me on
(202) 512- 7215. The

major contributors to this report are Carol Dawn Petersen, Assistant
Director, (202) 512- 7066; Barbara A. Smith, Senior Economist; Michael
Collins, Economist; and Kim Granger, Economist.

Sincerely, Barbara D. Bovbjerg, Director Education, Workforce, and Income
Security Team

Appendi xes The Social Security Reform Proposals and

Appendi x I

Their Provisions Table 6 lists the provisions in the five proposals we
studied.

Table 6: Social Security Reform Provisions Proposal Provision

P ASGKBG KS K

Reduce the cost- of- living adjustment (COLA) a x x Change the primary
insurance amount (PIA)

x x x formula a Establish a minimum level of benefits a x

Raise the normal retirement age (NRA) a, b x x Increase the benefit
computation period a x x Introduce individual accounts (IA) a x x x x
Transfer funds explicitly from the general fund x x x x x c Credit benefit
taxation revenue to the trust

x x funds Maintain coverage of taxable earnings x

Establish a minimum level of spousal benefits x Establish individual
accounts for newborns x Note: P = President's proposal, AS = Archer- Shaw,
GKBG = Gregg- Kerrey- Breaux- Grassley (S. 1383), KS = Kolbe- Stenholm (H.
R. 1793), K = Kasich. a Affects benefits received by disabled beneficiaries,
including disabled workers and their dependents

and disabled dependents of retired workers. b Raising the NRA refers to
either accelerating the already scheduled increases in the NRA or
immediately increasing the age of eligibility for retirement benefits. c
Although the Kasich proposal contains a provision describing a loan from
general revenue, this loan is in effect very similar to a general revenue
transfer. There is no principal repayment or interest repayment until 60
years into the future.

Source: Analysis performed and outlined in the SSA Office of the Chief
Actuary memorandums.

Table 7 shows that the access to the IA and the relationship between the IA
and the insurance benefit vary across the proposals we studied. Under the
Archer- Shaw and Kasich proposals, individuals can obtain funds from their
IAs at the age of retirement or when they become eligible for Disability
Insurance (DI) benefits. Under the Gregg- Kerrey- Breaux- Grassley and

Kolbe- Stenholm proposals, disabled individuals are able to obtain IA income
before retirement age only if the funds in the IA are sufficient to provide
a monthly income that, when added to the insurance benefit, is at least
equal to 1/ 12 of the current poverty line. According to the Social Security
Administration (SSA), this threshold for account access would be virtually
impossible for workers disabled at a relatively young age to meet

because they would not have the time to build up an IA. In addition,
insurance benefits are not affected by the presence of IA income under the
Archer- Shaw and Kolbe- Stenholm proposals. Under the Gregg- KerreyBreaux-
Grassley and Kasich proposals, there are reductions in the insurance benefit
because of the existence of an IA.

Table 7: Individual Account Provisions Adjustment to insurance Proposal
Beneficiary access to IA income benefit Effect on beneficiaries

Archer- Shaw When individual attains normal or No reduction in insurance

Beneficiary receives larger of early retirement age or becomes benefits.
Proceeds from IA are

current law benefit or monthly eligible for DI benefits. transferred to Old-
Age, income from IA. Survivors, and Disability Insurance (OASDI) trust
funds.

Gregg- Kerrey- BreauxGrassley When individual attains normal or

Insurance benefit reduced or When IA income is received, early retirement
age or when funds offset by amount of contribution beneficiary also receives
reduced

in IA are sufficient to provide plus interest accrued at rate insurance
benefit.

monthly benefit that, when added to earned by the OASDI trust

insurance benefit, equals or funds.

exceeds 1/ 12 of poverty line. Kasich When individual attains normal or

Insurance benefit reduced by At qualifying age, DI beneficiary early
retirement age or becomes 1/ 3 percent for each year receives both reduced
insurance eligible for DI benefits.

contributions are made to IA. benefit and income from IA. IA income could
make up for reduction in insurance benefit.

Kolbe- Stenholm Same as Gregg- Kerrey- Breaux None, no offset due to IA. But

Beneficiary receives income from Grassley. insurance benefit is less than IA
in addition to insurance benefit. current- law benefit because of IA income
could compensate for other provisions in bill.

decline in insurance benefit because of other provisions.

Alternative Solvency Scenarios and the Social

Appendi x II

Security Simulation Model The Calculation of Alternative Scenarios Maintain
Tax Rates and

This scenario maintains current payroll tax rates while reducing Social
Decrease Benefits Security benefits to levels supportable by these tax
rates. There are many ways to reduce benefits, including waiting until the
trust funds are exhausted and abruptly reducing benefits by the full amount
necessary to be supported by current payroll taxes. We decided to follow a
more gradual approach similar to that used in the “MTR (maintain tax
rates) Proposal” presented in the Report of the 1994- 96 Advisory
Council on Social Security.

The Council's proposal reduces the 0. 32 and 0.15 PIA formula factors by 0.5
percent for 1998- 2011 and 1.5 percent for 2012- 30. The PIA adjustments
used in this report also reduce the 0.32 and 0.15 formula factors but by 2.
0

percent for 2000- 13 and 3.0 percent for 2014- 32, which results in the
percentage reductions in benefits shown in table 8.

Table 8: Reductions in Benefits OASI and DI Year Women Men

2005 a 3.17% 5. 88% 2016 8. 68 16. 07 2022 11. 87 21.98 2034 16. 74 31.01 a
The year in which the benefit reduction amount is realized, based on the PIA
reductions. For example, a woman retiring in 2005 would receive a benefit
that is 3.17 percent lower than current law. Because

the SSASIM model has benefit relative to earnings by gender, reductions are
calculated separately for men and women. These percentage declines in
benefits result in trust fund solvency through 2074 under the 1999 Trustee's
Report intermediate assumptions. We assume no behavioral changes in response
to the decline in benefits because it is not clear how individuals will
respond to the decline in benefits- whether they will continue to retire at
younger ages or will postpone retirement to later ages in order to receive
larger benefits.

We instituted benefit reductions in the maintain- tax- rates scenario by
reducing only the 0.15 and the 0.32 brackets of the PIA formula, following
the approach used by the Advisory Council. (The PIA formula is described
below.) This is important to take into account when comparing benefits under
the maintain- tax- rates scenario with benefits for disabled

beneficiaries under the Social Security reform proposals. Kasich reduces all
three brackets, the 0. 90 bracket as well as the 0.15 and 0.32 brackets.
These reductions apply to both disabled and retired workers and their
dependents. This is why benefits for lower earners under Kasich's PIA
provision are below those calculated in the maintain- tax- rates scenario.
The Kolbe- Stenholm proposal, however, reduces only the upper two

brackets for disabled- worker beneficiaries and does not reduce these
brackets by as much as the maintain- tax- rates scenario does. 1 Therefore,
benefits for disabled low earners and their dependents under the
KolbeStenholm proposal are greater than benefits under the maintain- tax-
rates scenario. The Gregg- Kerrey- Breaux- Grassley proposal creates an
additional bracket and increases the 0. 32 bracket to 0.70. This explains
the

increase in benefits for low earners above the benefits received under the
maintain- benefits scenario. The full unreduced monthly benefit amount for
worker beneficiaries is

determined by using the PIA formula. This formula consists of three brackets
separated by two bend points. In 1999, these bend points were $505 and
$3,043 for newly eligible beneficiaries. A worker's PIA is calculated as
0.90 of the first $505 of career- average indexed monthly

earnings (AIME), plus 0.32 of any AIME amount between $505 and $3,043 and
0.15 of any AIME amount in excess of $3, 043.

Maintain Benefits and This scenario maintains current- law benefits while
increasing payroll tax

Increase Tax Rates rates to levels that support those benefits. There are
many ways to increase

payroll tax rates, including waiting until the trust fund is exhausted and
then abruptly increasing payroll tax rates to levels that would support
current- law benefits. We follow an approach similar to that used in the
“PL

PAYGO Proposal” presented in the Report of the 1994- 96 Advisory
Council on Social Security in which payroll tax rates are increased more
gradually. The PL PAYGO option modifies the present law payroll tax rate
schedule

1 Under Kolbe- Stenholm, retired workers face an additional benefit
reduction to all three PIA factors (0. 90, 0. 32, and 0. 15) of 0.5 percent.
These PIA reductions also affect benefits for adult disabled children of
retired workers.

from 12. 4 percent beginning in 1995 and reaching 17.1 percent in 2060. The
present law payroll tax rate adjustments used for this report are in table
9.

Table 9: Payroll Tax Rates Year a OASDI OASI DI

2000- 24 12. 4 10.6 1. 80 2025- 29 15. 4 13.12 2.28 2030- 49 16. 2 13.79
2.41 2050- 59 17. 1 14.55 2.55 2060- 73 18. 0 15.31 2.69 a The year the
adjustments are made to payroll taxes in the SSASIM model.

These payroll tax rates result in trust fund solvency through 2074 under the
1999 Trustees Report intermediate assumptions. Note that 85 percent of the
OASDI payroll tax rate is assigned to the OASI program, 15 percent to the DI
program.

The SSASIM Model To assess how the Social Security reform proposals affect
the solvency of the Social Security trust funds and the level of benefits
individuals receive, we conducted a variety of simulations using the SSASIM
model, developed

by the Policy Simulation Group. The initial version of the model was
developed under a series of contracts from SSA as part of the 1994- 96
Advisory Council on Social Security's activities. The model was

subsequently enhanced with major support from the American Association of
Retired Persons, the Employee Benefit Research Institute, and SSA as well as
other organizations. The model can simulate a variety of policy

reforms to the Social Security program, from incremental changes in the OASI
and DI programs to broader structural reforms that would introduce an IA
component to the Social Security system.

The SSASIM model simulates the dynamic interaction of the labor force, the
economy, and the Social Security programs and can be used to generate
aggregate program cost and income estimates as well as estimates for the
OASI and DI trust funds. Changes in program structure can be analyzed for
any specified future time periods. Consistent with SSA's annual projections,

we explored the effect of such changes on OASI and DI trust fund solvency
for the 75- year period 1999- 2074. The implications of a reform relative to

one of the alternative scenarios that achieve solvency are determined by
comparing the output results from a simulation that assumes the reform
policy with results from a simulation that assumes one of the two
alternative scenarios. Assumptions Used in the In our analysis, we made a
number of assumptions. With respect to Analysis of the Effects of

population and economic projections, we used the intermediate the Social
Security Reform assumptions in the 1999 Annual Report of the Board of
Trustees of the

Proposals on Solvency federal OASI and DI trust funds. We use the
assumptions in the 1999

Trustees Report because the Office of the Chief Actuary used these
assumptions to score the Social Security reform proposals we analyzed. (See
table 10.)

Table 10: Economic and Demographic Intermediate Assumptions From the 1999
Trustees Report Ultimate Year ultimate Assumption value value was attained a

Annual percentage

Labor force participation Women 60. 6 2075 Men 73. 8 2075 Unemployment rate
5.5 2009 Inflation rate 3. 3 2007 Labor productivity growth 1. 3 2008 Growth
rate of wages as share of compensation 0. 2 2008 Growth rate of hours worked
0. 1 2008 Nominal interest rate 6.3 2007 Mortality rate decline 0. 6 2023

Annual number

Total fertility rate 1. 9 b 2023 Net immigration 900,000 c 1999 Note: The
intermediate assumptions represent the trustees' “best
estimates” of likely future economic and demographic conditions. We
used these numbers throughout our analysis. a The ultimate value is
maintained for the remainder of the 75- year projection period. b Number of
children per woman. c Number of persons per year.

Assumptions in the Analysis We analyzed how the reforms affect individuals
born in 1946, 1966, and

of the Effects of the Reform 1986 in order to assess the effects of
provisions that are phased in over

Proposals on Individual time. We analyzed how the reforms affect individuals
with average earnings Benefits

and with 45 percent of average earnings to see how the reform provisions
affect workers at different earnings levels. The model contains information
on earnings separately for men and women. The user can specify a
genderrelated earnings pattern. Our analysis uses the earnings pattern for
men. These earnings are based on the national average annual earnings of
covered workers with earnings. Using 1998 data from SSA, we compared our
choice of earnings levels with the earnings levels of actual new
beneficiaries. We did so by calculating the DI benefit corresponding to our

selected earnings levels and comparing these benefit levels with the
distribution of benefits actual DI beneficiaries received in 1998. We found
that about 42 percent of all new beneficiaries in 1998 received benefits
that correspond to earnings that are less than 45 percent of average
earnings, about 38 percent of new beneficiaries received benefits
corresponding to earnings that are between 45 percent of average earnings
and average earnings, and about 20 percent of new beneficiaries received
benefits corresponding to earnings that are greater than the average level.
We analyzed how the reforms affected individuals with three different ages
of first receipt of DI benefits (35, 45, and 55) to compare the experiences
of people disabled at younger ages with those disabled at older ages. These

three ages reflect the experiences of individuals with different lengths of
time in the DI program and with different lengths of time in the labor
force. According to SSA, the average age of a new male DI beneficiary in
1999 was 49. 6 years, down from 51.2 years in 1980. In 1999, 19.3 percent of
men's new

benefits were awarded to individuals younger than 40, 24 percent to those in
their 40s, and 40 percent to those in their 50s. 2 DI benefits for disabled
workers are terminated mostly because of the death of the beneficiary or the
attainment of retirement age and conversion of benefits to the OASI program;
only half of 1 percent of DI beneficiaries leave the program each

year because of work. 3 2 2000 Annual Statistical Supplement to the Social
Security Bulletin. 3 2000 Annual Statistical Supplement to the Social
Security Bulletin and Work Incentives for Blind and Disabled Social Security
Beneficiaries (GAO/ HEHS unnumbered correspondence, Aug. 8, 2000).

According to SSA data on awards made to DI beneficiaries in 1998, the type
of disability that new DI beneficiaries claimed is somewhat associated with
age. In 1998, mental disorders were the most common diagnosis for new DI
awardees younger than 35, while diseases of the musculoskeletal system were
the most common diagnosis for those aged 50 and older. For new awardees
younger than 35, mental disorders accounted for 34 percent while diseases of
the musculoskeletal system accounted for 11 percent. For new DI awardees
aged 50 and older, diseases of the musculoskeletal system accounted for 27
percent while mental disorders accounted for

11 percent. 4 We assumed that individuals enter the workforce at age 22 and
work fulltime until disability or retirement with no years out of the labor
force. We chose these assumptions because they represent a “best
case” for the disabled individual. Many disabled individuals are
likely to work less than full- time and to have periods of time out of the
labor force. 5 However, little information is available on the wages,
earnings histories, and periods of nonwork of the disabled. This makes it
difficult to choose a “typical” earnings level and earnings
pattern for them. The benefit income for our “best case”
disabled individuals will clearly be greater than that for

disabled individuals receiving lower earnings from intermittent and less
than full- time employment. Our results, therefore, represent a maximum
level of benefit income that disabled beneficiaries could expect to receive
under the Social Security reform proposals that we modeled.

We also assumed that the nondisabled workers we simulated retire at age 67
and that all the individuals we simulated die at age 79. We made these
assumptions so that in our simulations the retired workers and all disabled
workers with a given age of first receipt of DI benefits would have the same
number of years of receiving benefits. Thus, differences in benefit income
across individuals would be the result of differences in reform proposals

and not the result of differences in individual characteristics. Because of
the possibility that actual disabled individuals might have a lower life 4
See SSA, 1999 Annual Statistical Supplement to the Social Security Bulletin
(Washington, D. C.: 2000). 5 In an earlier report, we found that an
increasing number of new DI applicants need supplementary SSI benefits,
which suggests that these applicants are less well off and might have less
extensive and less highly paid work histories than DI applicants had in the
past. See Social Security: Disability Rolls Keep Growing, While Explanations
Remain Elusive (GAO/ HEHS- 94- 34, Feb. 8, 1994).

expectancy than we assumed for our simulation, we asked SSA's Office of the
Chief Actuary to send us death rates for men who were born in 1986 and began
receiving DI benefits at age 45. We then calculated the proportion who would
still be alive at ages 46 to 79. According to our calculations, 49 percent
of these individuals would still be alive at 70, and 31 percent would still
be alive at 79.

We assumed that the benefits workers and their dependents received were not
affected by the application of the maximum family benefit. The maximum
family benefit refers to the maximum amount that can be paid on a worker's
earnings record. In the case of retired or deceased workers, the maximum
varies from 150 to 188 percent of the PIA. In the case of disabled workers,
the maximum family benefit is the smaller of 85 percent of the

worker's AIME or 150 percent of the worker's PIA. The family maximum cannot
be exceeded, regardless of the number of beneficiaries entitled on that
earnings record, although any benefit payable to a divorced spouse is not
included. Whenever the total of the individual monthly benefits payable to
all the beneficiaries entitled on one earnings record exceeds the

maximum, each dependent's or survivor's benefit is reduced in equal
proportion to bring the total within the maximum. For the analysis of IAs,
we assumed that administrative costs are 0.105 percent of assets. Our
estimate of administrative costs is that used in the Report of the 1994- 96
Advisory Council on Social Security. The Council considered an option to
create IAs alongside the Social Security system with a centralized system of
recordkeeping and limited investment choices. The estimate of 0.105 percent
of assets was a consensus of the Council members. We also assumed that
individuals do not annuitize but, rather,

draw down the balance in the IA through periodic withdrawals. 6
Consequently, the balance in the account is not reduced by the costs
associated with purchasing an annuity. We also assumed that individuals know
how long they are going to live and thus determine the schedule of periodic
withdrawals so as to use up the entire balance in the IA by the

time they die. These assumptions result in the largest balance possible in
the IAs.

6 Very few individuals in the United States purchase life annuities,
according to the studies we consulted.

For the Kolbe- Stenholm, Gregg- Kerrey- Breaux- Grassley, and Kasich
proposals, we used the same assumptions that SSA's Office of the Chief
Actuary used in scoring the Kolbe- Stenholm proposal. Following the approach
taken in the Report of the 1994- 96 Advisory Council on Social Security, we
varied the percentage invested in equities according to age. We assumed that
persons younger than 40 would invest 55 percent of their account in
equities, with an average real return of 4.8 percent for the portfolio. We
assumed that those 40 to 49 would invest 50 percent of their account in
equities, with an average real return of 4.5 percent. We assumed

that those 50 to 59 would invest 40 percent of their account in equities,
with an average real rate of return of 4. 1 percent. We assumed that those
60 to 69 would invest 20 percent of their accounts in equities, with an
average real return of 3.1 percent. We assumed the portion not invested in
equities would be invested in Treasury bonds and the return on equities
would be a constant, inflation- adjusted 7 percent per year, which reflects
the long- term historical average return on equities. We note that the
assumption of a 7 percent return on equities in the future has been
criticized by some as being optimistic. 7 We did not adjust the rates of
return on equities for risk. As we stated in a recent report, there are
numerous ways to adjust for risk but no clearly best way, and there is no
one risk- adjusted measure that everyone agrees is the correct measure. 8 As
a result, the returns on equity that we use are

likely to be higher than the risk- adjusted returns. 7 See the 1999
Technical Panel on Assumptions and Methods. 8 Social Security: Capital
Markets and Educational Issues Associated With Individual Accounts (GAO/
GGD- 99- 115, June 28, 1999).

Comments From the Social Security

Appendi x II I Administration

Lett er (207078) Lett er

Lett er

Ordering Information The first copy of each GAO report is free. Additional
copies of reports are $2 each. A check or money order should be made out to
the Superintendent of Documents. VISA and MasterCard credit cards are
accepted, also.

Orders for 100 or more copies to be mailed to a single address are
discounted 25 percent.

Orders by mail: U. S. General Accounting Office P. O. Box 37050 Washington,
DC 20013

Orders by visiting: Room 1100 700 4th St. NW (corner of 4th and G Sts. NW)
U. S. General Accounting Office Washington, DC

Orders by phone: (202) 512- 6000 fax: (202) 512- 6061 TDD (202) 512- 2537

Each day, GAO issues a list of newly available reports and testimony. To
receive facsimile copies of the daily list or any list from the past 30
days, please call (202) 512- 6000 using a touchtone phone. A recorded menu
will provide information on how to obtain these lists.

Orders by Internet: For information on how to access GAO reports on the
Internet, send an e- mail message with “info” in the body to:
info@ www. gao. gov or visit GAO's World Wide Web home page at: http:// www.
gao. gov

To Report Fraud,

Contact one:

Waste, or Abuse in

? Web site: http:// www. gao. gov/ fraudnet/ fraudnet. htm

Federal Programs

? e- mail: fraudnet@ gao. gov ? 1- 800- 424- 5454 (automated answering
system)

GAO United States General Accounting Office

Page 1 GAO- 01- 35 Social Security Reform

Contents

Contents Page 2 GAO- 01- 35 Social Security Reform

Page 3 GAO- 01- 35 Social Security Reform United States General Accounting
Office

Washington, D. C. 20548 Page 3 GAO- 01- 35 Social Security Reform

Page 4 GAO- 01- 35 Social Security Reform

Page 5 GAO- 01- 35 Social Security Reform

Page 6 GAO- 01- 35 Social Security Reform

Page 7 GAO- 01- 35 Social Security Reform

Page 8 GAO- 01- 35 Social Security Reform

Page 9 GAO- 01- 35 Social Security Reform

Page 10 GAO- 01- 35 Social Security Reform

Page 11 GAO- 01- 35 Social Security Reform

Page 12 GAO- 01- 35 Social Security Reform

Page 13 GAO- 01- 35 Social Security Reform

Page 14 GAO- 01- 35 Social Security Reform

Page 15 GAO- 01- 35 Social Security Reform

Page 16 GAO- 01- 35 Social Security Reform

Page 17 GAO- 01- 35 Social Security Reform

Page 18 GAO- 01- 35 Social Security Reform

Page 19 GAO- 01- 35 Social Security Reform

Page 20 GAO- 01- 35 Social Security Reform

Page 21 GAO- 01- 35 Social Security Reform

Page 22 GAO- 01- 35 Social Security Reform

Page 23 GAO- 01- 35 Social Security Reform

Page 24 GAO- 01- 35 Social Security Reform

Page 25 GAO- 01- 35 Social Security Reform

Page 26 GAO- 01- 35 Social Security Reform

Page 27 GAO- 01- 35 Social Security Reform

Page 28 GAO- 01- 35 Social Security Reform

Page 29 GAO- 01- 35 Social Security Reform

Page 30 GAO- 01- 35 Social Security Reform

Page 31 GAO- 01- 35 Social Security Reform

Page 32 GAO- 01- 35 Social Security Reform

Page 33 GAO- 01- 35 Social Security Reform

Page 34 GAO- 01- 35 Social Security Reform

Page 35 GAO- 01- 35 Social Security Reform

Page 36 GAO- 01- 35 Social Security Reform

Page 37 GAO- 01- 35 Social Security Reform

Page 38 GAO- 01- 35 Social Security Reform

Page 39 GAO- 01- 35 Social Security Reform

Page 40 GAO- 01- 35 Social Security Reform

Appendix I

Appendix I The Social Security Reform Proposals and Their Provisions

Page 41 GAO- 01- 35 Social Security Reform

Page 42 GAO- 01- 35 Social Security Reform

Appendix II

Appendix II Alternative Solvency Scenarios and the Social Security
Simulation Model

Page 43 GAO- 01- 35 Social Security Reform

Appendix II Alternative Solvency Scenarios and the Social Security
Simulation Model

Page 44 GAO- 01- 35 Social Security Reform

Appendix II Alternative Solvency Scenarios and the Social Security
Simulation Model

Page 45 GAO- 01- 35 Social Security Reform

Appendix II Alternative Solvency Scenarios and the Social Security
Simulation Model

Page 46 GAO- 01- 35 Social Security Reform

Appendix II Alternative Solvency Scenarios and the Social Security
Simulation Model

Page 47 GAO- 01- 35 Social Security Reform

Appendix II Alternative Solvency Scenarios and the Social Security
Simulation Model

Page 48 GAO- 01- 35 Social Security Reform

Appendix II Alternative Solvency Scenarios and the Social Security
Simulation Model

Page 49 GAO- 01- 35 Social Security Reform

Page 50 GAO- 01- 35 Social Security Reform

Appendix III

Appendix III Comments From the Social Security Administration

Page 51 GAO- 01- 35 Social Security Reform

Appendix III Comments From the Social Security Administration

Page 52 GAO- 01- 35 Social Security Reform

United States General Accounting Office Washington, D. C. 20548- 0001

Official Business Penalty for Private Use $300

Address Correction Requested Bulk Rate

Postage & Fees Paid GAO Permit No. GI00
*** End of document. ***