Social Security: Different Approaches for Addressing Program Solvency
(Chapter Report, 07/22/98, GAO/HEHS-98-33).

Pursuant to a congressional request, GAO provided information on issues
related to social security financing, focusing on: (1) the various
perspectives that underlie the current solvency debate; (2) the reform
options within the current program structure; (3) the issues that might
arise if social security were restructured to include individual
retirement accounts; and (4) the likely impacts on national saving of
reform proposals that call for changes in how Social Security benefits
are funded.

GAO noted that: (1) many options exist for restoring long-term solvency
within the current program structure; (2) these possibilities include
raising the retirement age, altering the benefit formula, reducing the
cost-of-living adjustment, investing Social Security Trust Fund
surpluses in the stock market, and mandating participation of workers
who are currently excluded; (3) some combinations of these changes could
restore program solvency while retaining the program's social insurance
features; (4) while these options generally require reducing benefits or
raising revenues, their effects on workers and retirees might be
mitigated if adjustments were made sooner, not later; (5) proposals for
more fundamental program changes have the potential to increase returns
overall but would entail increased risk; (6) moving even part of social
security to individual accounts would require careful consideration of
the issues raised by such a fundamental change; (7) the consequences for
the insurance aspects of the current social security system would
require close scrutiny if social security were wholly or partly
privatized; (8) most of the reform proposals envision substituting
advance funding for the largely pay-as-you-go system that exists today;
(9) in principle, advance funding of social security benefits could lead
to an increase in national saving; (10) increased saving could lead to
higher rates of economic growth and better enable future generations to
support themselves and future retirees; (11) moving to an advance funded
system would entail substantial transition costs that could offset any
potential savings for a number of years; (12) over the years, social
security has evolved to be more than a retirement program; (13) social
security not only provides the floor for an adequate retirement income,
it also insures families in the event of the death or disability of the
earner and helps provide retirement income security for low-income
workers; (14) restoring the system to financial solvency will require
fundamental choices about such issues as the strength of guarantees of
retirement income to the nation's elderly, levels of insurance for
working families, and the role of government in providing retirement
income; and (15) because such decisions will affect the nation and its
economy for years to come, they should be made with full knowledge and
debate of the trade-offs inherent in each proposed change.

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

 REPORTNUM:  HEHS-98-33
     TITLE:  Social Security: Different Approaches for Addressing 
             Program Solvency
      DATE:  07/22/98
   SUBJECT:  Future budget projections
             Budget deficit
             Social security benefits
             Economic analysis
             Trust funds
             Deficit reduction
             Federal social security programs
             Retirement benefits
             Investment planning
             Employee retirement plans
IDENTIFIER:  Social Security Program
             Social Security Trust Fund
             Supplemental Security Income Program
             Old Age Survivors and Disability Insurance Program
             Social Security Disability Insurance Program
             Federal Thrift Savings Plan
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** 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.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Committee on Finance, U.S.  Senate

July 1998

SOCIAL SECURITY - DIFFERENT
APPROACHES FOR ADDRESSING PROGRAM
SOLVENCY

GAO/HEHS-98-33

Social Security Solvency

(207447)


Abbreviations
=============================================================== ABBREV

  AIME - average indexed monthly earnings
  COLA - cost-of-living adjustment
  DI - Disability Insurance
  EBRI - Employee Benefit Research Institute
  ERA - early retirement age
  ERISA - Employee Retirement Income Security Act
  GPO - government pension offset
  HI - Hospital Insurance
  IA - individual account
  IRA - individual retirement account
  MB - maintain benefits
  NRA - normal retirement age
  OASDI - Old-Age, Survivors, and Disability Insurance
  OASI - Old-Age and Survivors Insurance
  PIA - primary insurance amount
  PSA - personal security account
  SSA - Social Security Administration
  SSI - Supplemental Security Income
  TSP - Thrift Savings Plan
  WEP - windfall elimination provision

Letter
=============================================================== LETTER


B-277092

July 22, 1998

The Honorable William V.  Roth, Jr.
Chairman
The Honorable Daniel Patrick Moynihan
Ranking Minority Member
Committee on Finance
United States Senate

The Social Security program has successfully provided support for our
nation's elderly and disabled populations for nearly 60 years. 
However, it is expected that the program will face a financing
shortfall in the future as the baby boom generation retires. 
Numerous groups have studied the problem and are recommending
solutions.  Some of the proposals involve adjustments to the
provisions of the existing program, while others discuss more
fundamental structural reform of the program.  This report provides
an overview of the financing problem and examines the different
approaches that are being advanced to address the program's future
solvency. 

As agreed with your offices, unless you release its contents earlier,
we plan no further distribution of this report for 30 days.  At that
time, we will make copies available to the Commissioner of Social
Security; relevant congressional committees and subcommittees; and to
others, on request. 

This report was prepared under my direction.  Please contact Barbara
D.  Bovbjerg, Associate Director, at (202) 512-7215 if you have any
questions.  Other major contributors to this report are listed in
appendix IV. 

Cynthia M.  Fagnoni
Director, Income Security Issues


EXECUTIVE SUMMARY
============================================================ Chapter 0


   PURPOSE
---------------------------------------------------------- Chapter 0:1

The aging of the "baby boom" generation, lower fertility rates, and
increasing longevity have eroded the long-term solvency of the Social
Security program.\1

The system's annual cash surpluses are currently projected to decline
substantially beginning around 2008, and by 2013 benefit payments are
expected to exceed cash revenues.  The Social Security Trust Funds
are forecast to be depleted by 2032, and from that point on revenues
are expected to be sufficient to pay no more than 75 percent of
promised benefits.  While the financing problem is not immediate,
analysts agree that it should be addressed soon to mitigate the
impacts of whatever means are chosen to correct it by spreading them
out over a longer period and allowing participants more time to
adjust. 

A national debate is now under way to identify how best to solve
Social Security's long-term financing problem.  To increase
congressional and public understanding of the issues related to
Social Security financing, the Senate Finance Committee asked GAO to
discuss (1) the various perspectives that underlie the current
solvency debate, (2) the reform options within the current program
structure, and (3) the issues that might arise if Social Security
were restructured to include individual retirement accounts.  This
report also discusses the likely impacts on national saving of reform
proposals that call for changes in how Social Security benefits are
funded. 


--------------------
\1 In this report, "Social Security" refers to the Old-Age,
Survivors, and Disability Insurance (OASDI) program. 


   BACKGROUND
---------------------------------------------------------- Chapter 0:2

The Social Security program was enacted in 1935 in response to the
economic deprivations of the Depression.  Originally created as a
benefit system for retired workers, over time Social Security has
been expanded to insure disabled workers and the families of retired,
disabled, and deceased workers.  Today, Social Security provides
income support to 44 million retired and disabled workers and to the
dependents and survivors of covered beneficiaries.  Since Social
Security's creation, poverty rates for the elderly have fallen from
an estimated 50 percent in 1935 to 11 percent today. 

Social Security's pay-as-you-go financing structure was adopted to
keep the federal government from building up large cash reserves.\2
There were concerns that doing so could prolong the Depression, and
there was also a desire to alleviate the financial plight of the
elderly as soon as possible.  Given the high ratio of workers to
eligible recipients, it was possible to give the earliest retirees
relatively large benefits, as compared with their contributions, and
the benefits were able to have a measurable impact on the recipients'
financial well-being.  The program's creators foresaw the need for
increased revenues as the system matured, and legislative actions
over the years have maintained its solvency.  In the future,
financing will be more difficult because more people will be living
longer and relatively fewer workers will be supporting them.  Over
the next 75 years, Social Security's total shortfall is projected to
be about $3 trillion in 1997 dollars. 

In early 1997, the Advisory Council on Social Security reported on
the program's long-term financing problem.  While the Council members
generally agreed on the size of the problem, they could not reach a
consensus on how to resolve it.  Instead, three packages of proposals
were advanced by different groups of Council members.  One package
consisted largely of revenue enhancements and benefit changes but did
not recommend fundamental changes to the structure of the existing
program.  In contrast, the other two packages, while also
incorporating adjustments to the current program, called for a degree
of "privatization" through the creation of mandatory individual
accounts.  A number of other proposals to address Social Security's
long-term financing problem have been advanced by various research
organizations, academics, and members of the Congress.  For the most
part, these other proposals contain provisions similar to those found
in the Advisory Council's report. 


--------------------
\2 Under a pay-as-you-go financing structure, payments to current
beneficiaries come from payroll taxes paid by current employees. 


   RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3

The need to ensure long-term solvency drives the current Social
Security debate.  Many options exist for restoring long-term solvency
within the current program structure.  These possibilities include
raising the retirement age, altering the benefit formula, reducing
the cost-of-living adjustment (COLA), investing Trust Fund surpluses
in the stock market, and mandating participation of workers who are
currently excluded.  Some combination of these changes could restore
program solvency while retaining the program's social insurance
features.  While these options generally require reducing benefits or
raising revenues, their effects on workers and retirees might be
mitigated if the adjustments were made sooner rather than later. 

Proposals for more fundamental program changes, featuring the
creation of individually owned retirement accounts, have the
potential to increase returns overall but would entail increased risk
for individuals--risk now borne by the government.  Indeed, moving
even part of Social Security to individual accounts would require
careful consideration of the issues raised by such a fundamental
change.  Further, the consequences for the insurance aspects of the
current Social Security system--disability insurance and
survivor/dependent benefits--would require close scrutiny if Social
Security were wholly or partly privatized.  Individual accounts also
raise issues such as how benefits would vary among individuals and
groups at greater risk of falling into poverty, how to finance any
transition costs, and how best to administer such plans. 

Most of the reform proposals envision substituting, to some extent,
advance funding for the largely pay-as-you-go system that exists
today.  In principle, advance funding of Social Security benefits
could lead to an increase in national saving.  Increased saving, in
turn, could lead to higher rates of economic growth and better enable
future generations to support themselves and future retirees. 
However, moving to an advance funded system would entail substantial
transition costs that could offset any potential savings, at least
for a number of years. 

Over the years, Social Security has evolved to be more than a
retirement program.  Social Security today not only provides the
floor for an adequate retirement income, it also insures families in
the event of the death or disability of the earner and helps provide
retirement income security for low-income workers.  Restoring the
system to financial solvency will require fundamental choices about
such issues as the strength of guarantees of retirement income to the
nation's elderly, levels of insurance for working families, and the
role of government in providing retirement income.  Because such
decisions will affect the nation and its economy for years to come,
they should be made with full knowledge and debate of the trade-offs
inherent in each proposed change. 


   GAO'S ANALYSIS
---------------------------------------------------------- Chapter 0:4


      FUNDAMENTAL DIFFERENCES
      UNDERLIE REFORM PROPOSALS
-------------------------------------------------------- Chapter 0:4.1

Supporters of changes within the existing system cite the need for a
"social insurance" approach, whereby workers participate
collectively, through the federal government, in a program that pools
risks arising from the loss of earnings caused by retirement,
disability, or death.  Such reforms would thus preserve the
long-standing structure of Social Security. 

Alternative approaches to reform largely focus on strengthening
individual choice and responsibility.  Supporters of these approaches
argue for a greater emphasis on linking returns to contributions and
for separating the retirement income purpose from the social
insurance goal.  Advocates of this "annuity-welfare" approach
emphasize providing retirement income through private institutions
(that is, privatization) and support a more limited government role
in providing a basic level of support.  Reforms they support
incorporate individual accounts, which, they argue, offer the
possibility of higher returns on contributions and improvements in
national saving. 


      OPTIONS EXIST FOR RESTORING
      SOLVENCY WITHIN THE EXISTING
      PROGRAM STRUCTURE
-------------------------------------------------------- Chapter 0:4.2

Social Security's future financing problem could be resolved within
its existing structure by increasing revenues, reducing benefits, or
some combination of the two.  Options that would increase revenues
include raising the payroll tax, extending coverage to currently
excluded workers, increasing the maximum taxable earnings level,
increasing the income tax on benefits, and investing the Trust Funds
in higher-earning assets.  Options for reducing expenditures include
reducing or eliminating benefits for certain groups of retired
workers, survivors, dependents, or disabled workers; increasing the
normal or early retirement age; and limiting COLA increases.  The
impact from any one of these options on the long-term actuarial
deficit would depend on the extent to which the option was
implemented.  Although changes such as these would retain both the
social insurance and redistributive aspects of the current system,
all except changing the Trust Funds' investment policy could directly
and negatively affect the economic well-being of groups of workers or
beneficiaries.  The results from these changes would be less severe
if they were spread out over a number of years.  However, some
analysts express concern that ongoing demographic trends will
continue to affect the program, requiring additional adjustments in
the future. 


      INDIVIDUAL ACCOUNTS ARE AN
      OPTION TO RESTRUCTURE THE
      PROGRAM
-------------------------------------------------------- Chapter 0:4.3

Other proposals would alter the system more fundamentally to restore
solvency.  These proposals have in common the advance funding of
future benefits (as opposed to the pay-as-you-go approach) through
the creation of individual accounts.  System participants would own
and, to varying extents, manage their own individual accounts, whose
returns would provide some or much of these contributors' future
retirement income.  Some proposals would finance these accounts with
new or increased taxes, while others would shift some portion of
current Social Security taxes to this purpose.  Most such proposals
retain some features of the current system.  Advocates of individual
accounts point to the potential for increased returns for
participants, although others have raised concerns that the risks of
investing would be borne by each individual rather than collectively
by the government. 

While individual accounts offer the potential of higher retirement
incomes through investing in stocks and bonds, important concerns
surround such proposals.  The primary concern is risk.  There is a
much greater potential for significant deterioration of an
individual's retirement "nest egg" under a system of individual
accounts.  Not only would individuals bear the risk that market
returns would fall overall but also that their own investments would
perform poorly even if the market, as a whole, did well.  Further,
shifting to individual accounts could disadvantage certain groups. 
Women, for example, have historically earned less and tended to
invest more conservatively than men and so would likely receive less,
on average, than men from their individual accounts. 

Other issues also must be addressed, including what level of basic
benefits would continue to protect those without adequate resources,
how the existing disability and ancillary benefits provided through
Social Security would be treated, how other sources of retirement
income might be affected, how much the accounts would cost, whether
and how the accounts would be annuitized, and whether the accounts
would be administered publicly or privately.  Such issues would need
to be addressed before implementing a system incorporating individual
accounts. 


      RAISING NATIONAL SAVING
      COULD HELP ALLEVIATE THE
      BURDEN OF RETIREMENT COSTS
-------------------------------------------------------- Chapter 0:4.4

A larger economy in the future would help ease the burden of meeting
retirement costs while sustaining a rising standard of living.  Thus,
Social Security reforms that promote increased national saving while
restoring solvency would advance complementary national goals.  For
example, reform proposals that involve advance funding of future
benefits could raise saving if they were not offset by additional
public borrowing or additional spending.  However, changing the
manner in which current Social Security funds are invested, whether
publicly or privately, might not, by itself, increase saving.  For
example, while shifting current Trust Fund surpluses to the stock
market or to individual accounts could potentially raise overall
returns, it would represent an asset shuffle rather than new saving. 
Currently, the Treasury uses the Social Security surpluses to finance
other government activities.  If the government had to borrow to
replace the lost surplus funds, no net increase in national saving
would occur. 

Efforts to increase advance funding are complicated, because
explicitly recognizing and funding some or all of the $9 trillion in
the program's unfunded promised benefits could require some
generations of workers to "pay twice"--once to cover the
already-accrued benefits of current retirees and again to advance
fund their own retirement benefits.  To the extent that advance
funding requires such transition costs, the positive effects on
saving are postponed well into the future.  Moreover, raising saving,
while an important national goal, should not, by itself, determine
the approach to restoring the solvency of the Social Security
program. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:5

GAO obtained comments on a draft of this report from the Social
Security Administration (SSA) and from several subject matter
experts.  SSA generally agreed with GAO's treatment of the issues and
offered a number of technical comments.  The subject matter experts
also offered technical comments.  GAO made changes throughout the
report to respond to these comments, as appropriate. 


INTRODUCTION
============================================================ Chapter 1

The Social Security program is the foundation of the nation's
retirement income system.  Since 1940, Social Security has been
providing benefits to the nation's eligible retired workers and their
dependents.  In addition to retired worker benefits, Social Security
also provides protection for covered workers with severe disabilities
and their dependents.  Also, spouses and children of deceased workers
may receive Social Security survivor benefits.  The program is
financed largely on a pay-as-you-go basis, with payroll taxes from
today's workers paying the benefits of today's beneficiaries. 

Demographic trends indicate that the Social Security program will
begin to experience a long-term financing problem after about 2013,
when benefit payments will start exceeding cash revenues.  The aging
baby boom generation will be followed by a relatively smaller work
force that will have to support a relatively larger group of
retirees.  This trend, combined with the increasing longevity of the
elderly, will significantly drive up the costs of maintaining the
program.  Without action to raise program revenues or cut program
spending, the Social Security Trust Funds will be exhausted by 2032. 

Proposals being considered for resolving the future solvency problem
range from making adjustments to the tax and benefit structure of the
current program to introducing features such as individual accounts
that could substantially alter the existing program structure. 
Despite these differences, policymakers and Social Security experts
agree that taking action soon is desirable to alleviate impacts on
workers and beneficiaries. 


   THE SOCIAL SECURITY PROGRAM
---------------------------------------------------------- Chapter 1:1

About 44 million people receive Social Security benefits today, and
about 147 million covered workers pay Social Security payroll taxes. 
More than 40 percent of the cash income of those aged 65 and older
comes from Social Security benefits, and over 60 percent of this
population receives at least half their income from Social Security
benefits.  For 15 percent of this population, Social Security
benefits are the only source of cash income.  The Social Security
program is one reason that poverty rates among the nation's elderly
have fallen dramatically--an estimated 39 percentage points since
1935.\3


--------------------
\3 In 1996, the poverty thresholds were $7,525 for an aged individual
and $9,491 for an aged couple. 


      REVENUE STRUCTURE
-------------------------------------------------------- Chapter 1:1.1

Social Security revenues come from three main sources:  (1) payroll
taxes of 12.4 percent on covered earnings (up to $68,400 in 1998)
split equally between employees and their employers and paid in full
by the self-employed,\4 (2) income taxes on up to one-half an
individual's or couple's Social Security benefits when total income
exceeds certain thresholds,\5 and (3) interest earnings on U.S. 
Treasury securities held by the Trust Funds. 

Program revenues in 1997 totaled $457.7 billion, of which almost 90
percent came from payroll taxes, about 1.7 percent from the income
taxation of Social Security benefits, and 10 percent from interest on
the Trust Funds' assets.  The share coming from the income taxation
of benefits is expected to grow because the income thresholds at
which benefits become taxable are not indexed.  The portion coming
from interest on the Trust Funds will increase until about 2020 and
then fall dramatically as the Trust Funds redeem securities to help
pay benefits. 


--------------------
\4 Payroll and income taxes paid by the self-employed are adjusted so
they can receive the favorable income tax treatment given employers
for their payroll tax contributions. 

\5 Up to 85 percent of Social Security benefits can be subject to the
income tax.  Some of this income tax revenue is dedicated to
Medicare's Hospital Insurance (HI) Trust Fund. 


      BENEFIT STRUCTURE
-------------------------------------------------------- Chapter 1:1.2

Social Security's benefit structure has evolved and expanded
considerably over time.  Under the original 1935 Social Security Act,
only retired workers meeting specified conditions were eligible for
monthly benefits.  Benefits under the original act had a strong
"individual equity" component--that is, individual benefits were
positively related to lifetime earnings.  Benefits also contained a
"social adequacy" component--that is, they were proportionately
larger, but absolutely smaller, for those with relatively low
lifetime earnings.  Currently, benefits are calculated using the 35
years of highest earnings, not total lifetime earnings, and benefits
are provided to workers' spouses, children, and survivors, who may
not have worked for pay.  These changes improved the social adequacy
component of the benefit structure.  The appropriate balance between
individual equity and social adequacy is a fundamental issue
surrounding Social Security's benefit structure and reflects the
extent to which the program redistributes income among workers and
beneficiaries. 


         AUXILIARY BENEFITS
------------------------------------------------------ Chapter 1:1.2.1

Social Security was originally designed to provide benefits only to
retired workers.  Major expansions were made to the program in 1939,
when the Congress provided "auxiliary" benefits for workers' eligible
wives, children, and survivors.  In 1956, it provided benefits for
disabled workers and their eligible dependents.  Other amendments to
the act have extended benefits to husbands, widowers, divorced
spouses, and mothers and fathers (spouses under age 65 with
benefit-eligible children in their care).  Some beneficiaries are
eligible to receive retired worker benefits on the basis of their own
work record and are also eligible to receive a higher benefit on the
basis of their current or former spouse's work record.  Essentially,
these beneficiaries, who are called "dually entitled," receive their
own retired worker benefit and the difference between that and the
higher auxiliary benefit.  Table 1.1 shows the current benefit
categories and the number of beneficiaries in each category. 



                         Table 1.1
          
          Social Security Benefit Types and Number
             of Beneficiaries Receiving Them in
                       December 1996

                                                 Number of
Benefit type                                 beneficiaries
----------------------------------------  ----------------
Retired workers                                 26,899,170
Dually entitled                                  5,629,780
Not dually entitled                             21,269,390
Disabled workers                                 4,386,040
Spouses                                          3,194,950
Of retired worker                                2,971,650
Of disabled worker                                 223,300
Children                                         3,811,600
Of retired worker                                  442,010
Of disabled worker                               1,467,490
Of deceased worker                               1,902,100
Survivors                                        5,445,710
Widows and widowers                              5,204,220
Mothers and fathers (with eligible                 241,490
 children in their care)
----------------------------------------------------------
Source:  SSA. 


         BENEFIT CALCULATION
------------------------------------------------------ Chapter 1:1.2.2

Calculating Social Security benefits is a three-step process.  First,
a worker's covered earnings over his or her 35 years of highest
earnings are identified.  Social Security uses average indexed
monthly earnings (AIME) as its measure of these "lifetime" covered
earnings.\6 Second, a progressive benefit formula is applied to these
lifetime covered earnings to determine the benefit that will be
payable to the worker at the normal retirement age (NRA), currently
age 65.\7 This NRA benefit, or primary insurance amount (PIA), is the
basic amount used to determine the actual benefit for those receiving
benefits on the basis of a worker's earnings record.  Finally, the
benefit is adjusted for the age at which the beneficiary first
receives the benefit.\8

Auxiliary benefits are based on the worker's PIA.  The benefits for
dually entitled people are based on their own PIAs.  If the spouse or
widow(er)'s benefit is higher, the dually entitled person's benefit
is supplemented to raise it to the amount of the spouse or
widow(er)'s benefit. 


--------------------
\6 The AIME is calculated by multiplying a worker's actual earnings
in a given year before he or she attains age 60 by the ratio of the
average national earnings level for the year he or she attains age 60
to the average national earnings level for the year being indexed. 
Earnings received after age 60 are not indexed but can be used in the
benefit calculation formula.  The total amount earned during the 35
years of highest indexed earnings is divided by 420 (35 years x 12
months/year) to arrive at the AIME. 

\7 The benefit formula is progressive in that it replaces a
relatively larger portion of lifetime earnings for people with low
earnings than for people with high earnings.  According to SSA,
workers with low lifetime covered earnings will have benefits that
replace approximately 56 percent of their AIME, workers with average
lifetime covered earnings will have about 42 percent replaced, and
workers with lifetime covered earnings at or above the maximum
taxable level will have only about 28 percent replaced.  The NRA
benefit, or "primary insurance amount," equals 90 percent of AIME up
to the first threshold + 32 percent of AIME between the first and
second thresholds + 15 percent of AIME above the second threshold. 
The thresholds are wage-adjusted and set yearly.  For 1998, the first
threshold is $477 and the second is $2,875.  The thresholds apply to
those attaining age 62 (or becoming disabled or dying) in a given
calendar year, regardless of their age when benefits are first
received. 

\8 To ensure roughly comparable expected lifetime benefits regardless
of the age when benefits are first taken, benefits are actuarially
reduced if first taken before the NRA and increased if first taken
between the month one attains the NRA and the month one attains age
70.  Benefits are also reduced if postretirement earnings exceed
certain thresholds. 


         BENEFIT INDEXING
------------------------------------------------------ Chapter 1:1.2.3

Currently, automatic benefit indexing provisions generally increase
the worker's PIA by an annual COLA.\9 The COLA is equal to the rise
in the consumer price index over a congressionally established period
of a year.  Indexing allows Social Security benefits to maintain the
same purchasing power over the beneficiary's retirement.  Retirement
income from most other sources is not fully indexed and thus tends to
decline in real terms over time. 


--------------------
\9 The COLA is payable for each year that has passed since the worker
first became eligible for benefits (age 62 for retired worker
beneficiaries).  Thus, someone who waits until age 65 to claim
retired worker benefits will use the benefit formula in place when he
or she was age 62, and his or her PIA will be increased by the total
of all COLAs provided beginning with the year he or she became 62. 


      FINANCING STRUCTURE
-------------------------------------------------------- Chapter 1:1.3

Social Security is financed largely on a pay-as-you-go basis.  Under
this type of financing structure, the payroll tax revenues collected
from today's workers are used to pay the benefits of today's
beneficiaries.  Under a strict pay-as-you-go financing system, any
excess of revenues over expenditures is credited to the program's
trust funds, which function as a contingency reserve.\10 Social
Security's Trust Funds reserve allows the government to manage the
inevitable differences over time between revenues and expenditures. 

One reason the pay-as-you-go approach was initially used is that it
required relatively small contributions at a time when the program
was young and benefit payments were small.  However, this structure
required increasing contribution levels as the program matured and
more beneficiaries with higher average benefits were added to the
beneficiary rolls.  In addition, the pay-as-you-go structure leaves
the program and the federal government susceptible to financing
problems when costs increase more than expected or revenues fail to
meet expected levels, such as might occur with changing short-term
economic conditions.  Every year, Social Security's Board of Trustees
estimates the financial status of the program for the next 75 years
using three sets of economic and demographic assumptions about the
future.\11 According to the intermediate set of these assumptions,
the nation's Social Security program will face solvency problems in
the years ahead unless corrective actions are taken. 


--------------------
\10 Annual program revenues currently exceed annual program
expenditures, resulting in partial advance funding of the program. 
Thus, Social Security is not strictly a pay-as-you-go program, and
the Trust Funds are not simply a contingency reserve. 

\11 The 75-year projection period was established as an indicator of
whether the system would have sufficient resources to provide
benefits over the lifetime of a worker.  The Trustees make three sets
of projections:  one using a high-cost (pessimistic) set of
assumptions, one an intermediate-cost set, and one a low-cost
(optimistic) set.  We have used the intermediate cost projections as
the basis of our analysis.  Unless otherwise indicated, we refer to
the estimates for the 75-year period from 1998 to 2072 (see The 1998
Annual Report of the Board of Trustees of the Federal Old-Age and
Survivors and Disability Insurance Trust Funds [Washington, D.C.: 
U.S.  Government Printing Office, 1998]). 


   SOCIAL SECURITY FACES A
   LONG-TERM FINANCING PROBLEM
---------------------------------------------------------- Chapter 1:2

The Social Security program is not in long-term actuarial balance. 
That is, Social Security revenues are not expected to be sufficient
to pay all benefit obligations from 1998 to 2072.  Without changing
the current program, excess cash revenues from payroll and income
taxes are expected to begin to decline substantially around 2008.  By
2013--15 years from now--these cash revenues will be insufficient to
pay all program costs.\12 After 2013, Social Security will have to
start redeeming some of its assets to obtain the cash needed to pay
benefits.  The Trust Funds are expected to be exhausted in 2032. 

The anticipated revenue shortfall over the next 75 years is estimated
at $3 trillion, or an average annual shortfall of $40 billion (in
1997 dollars).  This $3 trillion shortfall is based on the assumption
that the Social Security program will continue under its current
structure.  That is, new workers will enter the system, pay payroll
taxes (which will be matched by their employers), accrue benefit
credits while working, and receive benefits when they retire. 

Even if revenue or expenditure adjustments necessary to reach 75-year
balance were achieved, the financing problem still might not be
permanently resolved.  For the foreseeable future, each new 75-year
projection period will have a higher long-term financing shortfall
than the last.  For example, suppose the payroll tax was raised
sufficiently to reach balance, and the current actuarial assumptions
were realized for the period 1998 through 2072.  Under this scenario,
the Trust Funds would have only about 1 year's worth of benefits
remaining in 2072.  If the same actuarial assumptions continued to be
used in each of the years between 1998 and 2072, the Trust Funds
would continue to be expected to be exhausted shortly after 2072,
but, beginning in 1999, the 75-year projections would show a
long-term revenue shortfall for the program that would grow over
time. 

The program has another, higher revenue shortfall estimate--about $9
trillion, as of October 1, 1997.  This is the amount of the program's
unfunded benefit obligations--the accrued future benefit obligations
that will not be able to be paid with assets currently in hand.  A
large unfunded liability in a government program financed primarily
on a pay-as-you-go basis is generally not considered a problem
because of the government's authority to tax current workers to pay
current benefits.  Thus, current unfunded liabilities are passed onto
future generations.  However, if the current Social Security program
were ended or changed to an advance funded system, all $9 trillion of
accrued benefit obligations would have to be paid if the government
honored these obligations in full. 

Social Security does not face an immediate financing crisis because
its cash revenues are expected to exceed its expenditures until 2013. 
However, the substantial size of the anticipated 75-year shortfall
($3 trillion if the program remains a pay-as-you-go system and $9
trillion if it is terminated or becomes a system that is funded in
advance) suggests the need for reform action in the near future. 
Social Security is currently building up some Trust Funds reserves,
which can help offset some of the revenue shortfall after 2013. 
Interest earnings on and redemption of these reserves, along with
payroll and income tax revenues, are expected to provide sufficient
resources, under the Trustees' 1998 intermediate assumptions, to pay
program obligations until about 2032.  Without action to improve the
system's financial outlook, the program is expected to have revenues
sufficient to cover only about 75 percent of anticipated benefit
obligations in 2032, and this will decline to about 68 percent by
2072. 


--------------------
\12 About 98 percent of program costs are for benefit expenditures. 


      DEMOGRAPHIC CHANGES WILL
      STRAIN SOCIAL SECURITY'S
      RESOURCES
-------------------------------------------------------- Chapter 1:2.1

An important factor affecting Social Security's pending financing
problem is the rapidly approaching retirement of the baby boom
generation.  The oldest of this generation will reach early
retirement age (62) in 2008, and the youngest will reach it in 2026. 
This large number of retirees would substantially increase program
costs and strain the ability of the program to pay benefits even if
it were the only factor affecting future costs.  (See fig.  1.1.)

   Figure 1.1:  Historical and
   Projected Growth in the Number
   of Social Security
   Beneficiaries, 1945-2070

   (See figure in printed
   edition.)

Source:  1998 Annual Report, Social Security Board of Trustees, 1998,
Tables II.H12 and II.H43, pp.  156-57 and 162-63. 

Exacerbating the problem of the retirement of the baby boom
generation is the relatively smaller generation that follows it.  The
post-baby-boom generation, which resulted from the rapid decline in
fertility rates from the mid-1960s to the mid-1980s (see fig.  1.2),
will result in relatively fewer workers to support a larger number of
retirees.  The number of workers whose payroll taxes will support
those on Social Security will fall from today's about 3.4 per
beneficiary to an anticipated 2.0 per beneficiary in 2030. 

   Figure 1.2:  Historical and
   Projected Fertility Rates,
   1940-2070

   (See figure in printed
   edition.)

Source:  1998 Annual Report, Social Security Board of Trustees, 1998,
Table II.D2, pp.  60-61. 

Another factor that will raise program costs is the increase in life
expectancies.  Life expectancy for 65-year-old men increased from
11.9 years at the program's inception to 15.3 years in 1995 and for
65-year-old women, from 13.4 years to 19.0 years.  Life expectancies
are expected to continue to increase to 18.7 years for men and 22.0
years for women in 2070.  This increase will further strain the
program's financing, requiring revenue increases or benefit cuts to
keep the program solvent.\13


--------------------
\13 1998 Annual Report, Social Security Board of Trustees, 1998,
Table II.D2, p.  60. 


      OTHER FACTORS CAN AFFECT
      PROGRAM SOLVENCY
-------------------------------------------------------- Chapter 1:2.2

Other factors, primarily economic and behavioral aspects of Social
Security's actuarial assumptions, can also affect its costs and
revenues.\14

Factors that increase program costs include the following: 

  -- automatic COLAs, which maintain the real purchasing power of
     benefits but increase both nominal benefit levels and program
     costs geometrically over time;

  -- the relaxed earnings test, which allows benefit-eligible workers
     to receive Social Security benefits even though they have
     considerable earnings;\15 and

  -- rising real wages, which increase real benefits over time. 

Factors that constrain program revenues include the following: 

  -- an earlier average retirement age, which reduces the period
     during which workers pay payroll taxes;\16

  -- lower than expected rates of real economic growth, such as occur
     with recessions, which constrain the growth of covered wages and
     make paying taxes to support Social Security beneficiaries more
     onerous than if the economy had grown at a faster rate; and

  -- the growing share of total employee compensation that is not
     subject to payroll taxes. 


--------------------
\14 The impacts of these factors would usually occur regardless of
the program's financing structure. 

\15 Because Social Security originally insured against a loss of all
covered earnings by those 65 and older, an earnings test was
instituted to ensure that such a loss had occurred before benefits
were granted.  The earnings test is discussed more fully in ch.  3. 

\16 An earlier average retirement age also lengthens the time
benefits are received.  However, the program actuarially adjusts
benefits received before the NRA, so this does not, on average,
increase program costs. 


   ANALYSTS DIFFER ON APPROACHES
   TO THE PROBLEM BUT AGREE ON THE
   NEED TO ACT SOON
---------------------------------------------------------- Chapter 1:3

The crucial role Social Security plays in providing income support to
the nation's elderly and disabled populations makes the program an
ongoing policy focus of the Congress and numerous nonfederal groups
and organizations.  In the past when financing problems have been
encountered, the Congress has acted to alter the revenue and benefit
provisions of the program to maintain its solvency.  While the
program has been modified on an ongoing basis, major legislative
reforms, such as those enacted in 1977 and 1983, have been made less
frequently. 

Because the Advisory Council could not reach consensus on how to
fully restore solvency over the 75-year period, it brought forward
three packages of proposals, including two that combine elements of
individual accounts with other program changes, such as adjustments
to the benefit formula.\17

One of the three Advisory Council proposals--the "maintain benefits"
(MB) proposal--involves mainly traditional reforms\18 that would
operate within the existing structure of the program.  A second
proposal would significantly change the system by creating individual
retirement accounts--"personal security accounts" (PSA)--that would
be privately managed and invested as directed by the individual
worker.  The third proposal--"individual accounts" (IA)--is
essentially a hybrid of the other two proposals and includes both the
creation of private individual accounts administered by the federal
government and traditional-style reforms to the current program. 

The MB proposal would maintain most of the existing benefit structure
of the program.  However, since it used only traditional reforms,
this proposal did not fully close the financing gap and restore
actuarial balance.  To achieve long-term actuarial balance, the MB
group considered an option that involved investing about 40 percent
of the Trust Funds' assets in private securities, such as through
stock and bond mutual funds.  This approach, in essence, would have
expanded the extent to which the program was advance funded.  In the
end, the MB group simply recommended that this Trust Funds investment
option be studied further. 

The other two Advisory Council proposals include systems of
individual accounts.  The PSA proposal would divert a portion of the
existing payroll tax into accounts that would be managed privately,
while the remainder of the payroll tax would go to finance a public
benefit that would be smaller than current benefits are for most
beneficiaries.  Under this plan, 5 percentage points of the
employee's share of the current OASDI tax rate would be diverted to
an individual account.\19 The accounts would be individually owned
and privately managed, and individuals would choose from a variety of
investments in private financial instruments.  The accounts would be
tax-deferred, and individuals could begin drawing from them at age
62.  Any funds remaining in an account upon the death of the owner
would become a part of the estate.  The individual accounts would
represent a second tier of benefits, with a modified version of the
existing Social Security program benefits maintained as the smaller
first tier.\20

The IA option would essentially maintain the structure of the
existing system, with adjustment, as a large first tier and add an
individual account component as a supplemental second tier.  Under
this proposal, workers would be required to contribute an additional
1.6 percentage points of taxable payroll to fund the individual
accounts.  The accounts would be invested in private securities, and
workers could choose among such investments as stock and bond mutual
funds and government securities.  However, these accounts would be
administered largely through the existing Social Security program. 
The account accumulation would also be required to be annuitized
through Social Security, a feature not included under the PSA plan. 

While these three Advisory Council options tend to dominate the
current debate, numerous other proposals and options have also been
advanced by various organizations, academics, and members of the
Congress.\21 For example, in the 104th Congress, proposals were
advanced by Senators Kerrey and Simpson (S.  824, S.  825, and S. 
2176) and Representative Nick Smith (H.R.  3758) and, in the 105th
Congress, by Senator Judd Gregg (S.  321), Senator Daniel Patrick
Moynihan (S.  1792), Representative Mark Sanford (H.R.  2768 and H.R. 
2782), Representative John Porter (H.R.  2929), and Representative
Nick Smith (H.R.  3082).  Numerous other proposals have been offered
recently by organizations such as the National Taxpayers' Union
Foundation and the Committee on Economic Development, as well as by
various economists and analysts (see bibliography). 

Although the Board of Trustees has indicated the program is expected
to have sufficient assets and revenues (including interest on the
Trust Funds) to pay all benefit obligations for the next 3 decades
with no changes to the program, most analysts believe early action to
reduce the actuarial imbalance is important for a number of reasons. 
First, the longer action to address the program's financing problem
is delayed, the larger the per-year cost of the solution because the
shortfall in revenues will still have to be addressed, but over a
shorter period of time.  Second, some of the possible solutions to
the solvency problem--such as raising the program's NRA, reducing
benefits for future beneficiaries, or increasing the program's
advance funding--will take time to implement or phase in, once
enacted.  Third, if certain changes, especially those that reduce
benefits, are made, workers will need time to adjust their saving and
retirement goals to help mitigate the personal impacts of these
changes.  Thus, the sooner the changes are made, the less disruptive
they are likely to be. 


--------------------
\17 Vol.  1, app.  II, of the Advisory Council's report contains an
evaluation of the three proposals.  Report of the 1994-1996 Advisory
Council on Social Security, Volumes I and II (Washington, D.C.: 
Advisory Council on Social Security, 1997) can be accessed at the
following Internet site: 
http://www.ssa.gov/policy/adcouncil/toc.htm. 

\18 We define "traditional" reform proposals as those aimed at
improving Social Security's solvency while maintaining the basics of
the program's revenue and benefit structures and preserving the
federal government's role in the program's administration.  A
proposal to adjust the revenue provisions of the program is an
example of a traditional proposal. 

\19 The current OASDI tax rate is 12.4 percent of taxable payroll, or
6.2 percent for employer and employee each.  Thus, 5 percentage
points represents about 80 percent of the employee portion and about
40 percent of the combined payroll tax. 

\20 This first tier would also encompass modified ancillary benefits. 
A number of adjustments would be made to both benefits and the
retirement age so that this first tier would continue to be
self-financing.  Full eligibility would result in an indexed basic
benefit equivalent to $410 per month in 1996, and the benefits
arising from individual account accumulations would supplement this
level.  This new first tier of Social Security benefits would be
equivalent to about 65 percent of the current poverty level. 

\21 For a detailed summary of several major proposals, including the
Advisory Council proposals, see Employee Benefit Research Institute
(EBRI), Assessing Social Security Reform Alternatives, App.  A,
Dallas L.  Salisbury, ed.  (Washington, D.C.:  EBRI, 1997); EBRI
Notes, Vol.  19, No.  4, (Apr.  1998), pp.  6-10; and the National
Academy of Social Insurance, Social Insurance Update, Vol.  1, No.  3
(Dec.  1996). 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
---------------------------------------------------------- Chapter 1:4

The Chairman and Ranking Minority Member of the Senate Finance
Committee asked us to discuss (1) the various perspectives that
underlie the current solvency debate, (2) the reform options within
the current structure, and (3) the issues that might arise if Social
Security were restructured to include individual accounts.  We also
discuss the likely effects on national saving of reform proposals
that call for more advance funding of Social Security benefits. 

Because of the wide-ranging nature of the numerous proposals being
advanced, our report focuses on the common, or generic, elements that
underlie various proposals to reform Social Security financing rather
than a complete evaluation of specific proposals.  In conducting this
study, we reviewed literature on Social Security's long-term
financing problem and related issues as well as a number of proposals
that would address this problem.  We held discussions with SSA
officials and with other subject matter experts from government, the
policy community, and academia about these issues.  We also drew on
our own previous work.  We obtained comments on a draft of this
report from SSA and subject matter experts and made revisions as
appropriate.  We conducted our work between October 1996 and February
1998, in accordance with generally accepted government auditing
standards. 


DIFFERENT CONCEPTUAL RATIONALES
UNDERLIE SOCIAL SECURITY REFORM
PROPOSALS
============================================================ Chapter 2

The need to ensure Social Security's long-term solvency has sparked a
debate that has roots in the program's creation.  Both at the
program's inception and today, the discussion has centered around
different frameworks for providing social insurance.  The many and
varied proposals for addressing Social Security's future solvency
problem--including those put forth by the 1994-96 Advisory
Council--reflect these fundamentally different perspectives on the
appropriate structure for Social Security.  As a result, these
proposals range from traditional reforms of the current program to
significant restructuring.  Increased advance funding forms a core
element of many solvency proposals. 


   THE HISTORY OF THE PROGRAM
---------------------------------------------------------- Chapter 2:1

The Social Security program emerged in the 1930s as the nation sought
to address hardships created by difficult economic conditions.  Some
historians of Social Security point out that prior to the Great
Depression there was considerable resistance to involving the federal
government in providing economic security and creating a federal
social insurance program.  Despite this view, there was also a
developing realization that individual and voluntary actions were not
adequate to address poverty among the elderly, and a number of state
programs to assist the elderly were instituted.  With the coming of
the Great Depression and as various social movements gained
attention,\22 President Franklin D.  Roosevelt appointed the
Committee on Economic Security to devise what came to be the Social
Security program.  Throughout the legislative deliberations leading
to passage of the Social Security Act in 1935, the theme of attaining
a consensus on the balance between government and individual
responsibility was prevalent. 

Over the years, the debate about the role of government has largely
centered around three models:  the social insurance model, the
tax-transfer model, and the annuity-welfare model.\23 Given the
structure of the program as it emerged in the 1930s, the social
insurance model (and, to a lesser degree, the tax-transfer model) has
provided the most frequently used framework for analyzing the
program.\24 Some analysts, however, view the annuity-welfare model as
a more appropriate approach for reform. 


--------------------
\22 One of these was the Townsend Movement, which focused on
funneling income to the elderly and encouraging consumption to spur
economic expansion.  See David V.  Bryce and Robert B.  Friedland,
"Economic Security:  An Overview of Social Security," in Assessing
Social Security Reform Alternatives, Dallas L.  Salisbury, ed. 
(Washington, D.C.:  EBRI, 1997), p.  32. 

\23 Lawrence H.  Thompson, "The Social Security Reform Debate,"
Journal of Economic Literature, Vol.  21 (Dec.  1983), pp.  1425-67. 

\24 The main difference between the social insurance and tax-transfer
models is that the social insurance model focuses on life cycle or
intergenerational transfers, while the tax-transfer model focuses on
Social Security simply as a current period tax-transfer program
(Thompson, "The Social Security Reform Debate," 1983, pp.  1436-38). 
The tax-transfer model is probably most useful for conducting certain
types of economic analyses and for analyzing Social Security in a
federal budget context, but it does not provide as much insight into
the underlying philosophy for social insurance as the social
insurance framework does. 


      THE SOCIAL INSURANCE MODEL
-------------------------------------------------------- Chapter 2:1.1

Workers face a variety of risks arising from the loss of earnings
that can result from retirement, disability, or death.  Consistent
with the social insurance model, Social Security represents a way for
workers to pool these risks; it offers a package of benefits that can
be obtained for a given price in the form of taxes.  Since the risk
primarily involves the loss of earnings the taxes to finance such a
program are earnings-related, as are the benefits received.  In
general, because such a package of benefits may not easily be
obtained in private markets, the government is involved in providing
the benefits.  In addition to this market failure rationale for
involving the government in administering such a pooling of risks,
related rationales include reducing uncertainty about individuals'
future retirement income; alleviating insurance market failures, such
as adverse selection; addressing social concerns about income
redistribution; reducing the social burden imposed by nonsavers and
the short-sighted; and institutionalizing the compact between
generations (filialism).\25

In constructing the program along the lines of the social insurance
model, two important--and apparently conflicting--objectives were
addressed:  individual equity and social adequacy.  Linking benefits
directly to the tax price paid, or to contributions, invokes the
standard of a market return, or an "actuarially fair return," and
demonstrates the individual equity principle.  But pooling risks
against earnings loss also involves the concept of need or a desired
minimum level of benefits.  Thus, the program is designed to also
embody the principle of social adequacy, which involves
redistribution among participants within the program.  Balancing
these seemingly conflicting objectives through the political process
has resulted in the design of the current Social Security program. 


--------------------
\25 Lawrence H.  Thompson and Melinda M.  Upp, "The Social Insurance
Approach and Social Security," in Social Security in the 21st
Century, Eric Kingson and James H.  Schulz, eds.  (New York:  Oxford
University Press, 1997), pp.  3-21. 


      THE ANNUITY-WELFARE MODEL
-------------------------------------------------------- Chapter 2:1.2

Some analysts advocate an alternative approach for restructuring
Social Security:  the annuity-welfare framework.  The emergence of
this model is linked with the debate that took place in the 1930s and
with various economic critiques that have emerged since the 1960s.\26
The fundamental basis of this model is the view that the different
components of Social Security--individual equity and social
adequacy--should be addressed separately; that is, the part of Social
Security that pays benefits related to contributions by workers
should be separated from the part of Social Security that relates to
adequacy, or maintaining a minimum level of income to alleviate
poverty.  This view generally leads to a rather different approach to
providing retirement income. 

Several key points about the annuity-welfare model and its relation
to Social Security are worthy of note.  First, while the individual
is required to participate in Social Security, the annuity-welfare
model emphasizes maximizing voluntary arrangements whenever possible. 
Nevertheless, the annuity-welfare model generally recognizes that
because some individuals may choose to "free ride" on society by not
saving adequately, and others may experience conditions during their
lifetime that leave them without adequate resources, a role for
government involvement may be justified.  Second, Social Security is
not advance funded in the manner of private pensions and does not
grant contractual rights to individuals as does, for example, a
pension trust arrangement.  Rather, the pay-as-you-go financing
structure means that current workers pay for the benefits of current
retirees and that benefits are promised largely on the basis of the
ability of the government to pay them in the future.  Third, the
connection under Social Security between benefits and contributions
is loose, mainly because of the redistributive nature of the system. 
As a result, some individuals will receive less than a market return
for their contributions, which has raised concerns among proponents
of the annuity-welfare model about the value provided by the program. 

There is an important fourth issue.  Some see the existing Social
Security structure as leading to further difficulties because
decisions about the program and its impact on individuals are made
through the political process.  This is known as "political risk."
According to this view, the design of Social Security creates the
potential for program expansion because there will always be
political incentives to promise higher benefits, which will be paid
for disproportionately by certain groups, such as high earners or
future generations.\27 In addition, higher benefits that may need to
be paid for by future workers can be promised in the near term, even
though the ability of the government to raise funds in the future to
make good on these promises may be dependent on the political
situation at the time. 

Proponents of the annuity-welfare model view obtaining adequate
retirement income as a matter of individual responsibility and
believe that this private decision should be separate from the social
decision about providing an adequate or minimum level of retirement
income for those who otherwise would fall into poverty in old age. 
Thus, under this model, the individual may have greater control,
through the political process, of the level of minimum or basic
income to be provided by society because he or she is not required to
participate in a larger program of social insurance that is subject
to legislative and political actions. 

The emergence of privatization and individual account plans as an
element of the current Social Security financing debate can in large
part be tied to the annuity-welfare model.  Two key features of this
framework are its emphasis on advance funding and on a more direct
linkage between the contributions made to the system and the benefits
received from it.  While proponents of both the social insurance and
annuity-welfare approaches agree that those who contribute more to
the system should receive more from it, the existence of income
redistribution in the current Social Security program weakens this
linkage.  Individual account proposals could strengthen the program's
equity goal by establishing a system in which the returns on
investments would accrue to individuals themselves.\28 \29


--------------------
\26 In the 1960s, a critique of Social Security emerged that is
closely associated with the ideas and writings of Milton Friedman and
James Buchanan.  Each criticized the system and advocated changing to
a system that was voluntary, fully funded, and market-based. 
Subsequent work by others, such as Edgar Browning, refined the
critique and emphasized the tendency of the political process to
expand the program.  The critique was extended further by Peter
Ferrara, who emphasized the need for a voluntary and private system,
and proposed that the individual retirement account (IRA) concept be
applied to the government-managed provision of retirement income
(Social Security). 

\27 While these groups will usually be higher-income groups, this is
not exclusively the case given the structure of auxiliary benefits. 

\28 However, depending upon the design of the individual's account,
its interaction with any remaining government-sponsored benefit,
family composition, and market returns, the equity goal could be
enhanced or diminished.  See app.  II. 

\29 Individual accounts generally require a defined contribution-type
benefit structure wherein contributions and earnings on these
contributions determine the amount of money available to fund
benefits. 


      THE ROLE FOR GOVERNMENT IN
      THE TWO MODELS DIFFERS BY
      DEGREE
-------------------------------------------------------- Chapter 2:1.3

In general, the frameworks discussed here reflect differences in
philosophies about the appropriate balance between individual and
government responsibility.  While both frameworks include a role for
government in providing retirement income and some mandatory
contribution toward it, the degree of support to be provided through
government is a major source of contention. 

Concerning the issue of linking benefits to contributions, supporters
of the current Social Security program structure argue that
redistribution is a desirable goal and a major reason for a social
insurance program.  They object to the separation of the individual
equity and social adequacy elements, as this holds the potential, in
their view, for undermining the consensus for redistribution and
support of the less fortunate elderly.  Further, they assert, the
commitment of government under a social insurance system precludes
the need for contractual arrangements and, because risks are borne
collectively, reduces many of the risks that would otherwise be faced
individually.  Supporters of the current system also argue that a
primarily pay-as-you-go system is an appropriate way to finance
transfers intergenerationally.  Thus, these advocates propose
solutions to the financing problem that essentially maintain this
structure and preserve government's primary role. 

Others offer proposals that would fundamentally restructure the
Social Security program to reduce the role of government and increase
individuals' returns.  They particularly focus on increasing
individual choice and responsibility and emphasize private market
returns on contributions, such as could occur with individual account
proposals.  Consistent with this focus, they emphasize that it is
important that government address the unfunded liabilities of Social
Security, and they recommend moving toward a greater reliance on
advance funding and away from the primarily pay-as-you-go approach
now in use. 


   ADVANCE FUNDING IS A CENTRAL
   ELEMENT OF THE CURRENT DEBATE
---------------------------------------------------------- Chapter 2:2

Advance funding involves saving real assets to finance benefits
promised today but paid in the future.  Applying such a financing
approach to the Social Security system, which is currently financed
primarily on a pay-as-you-go basis, would require a period during
which contributors paid twice--once for current beneficiaries and
again to "advance fund" some part of their own retirement benefits. 
Despite this potential drawback, most proposals to reform Social
Security's financing build in some degree of advance funding, arguing
that the long-term economic benefits could offset short-term costs. 


      ADVANCE FUNDING OFFERS AN
      ALTERNATE MEANS TO FINANCE
      FUTURE PROMISES
-------------------------------------------------------- Chapter 2:2.1

The ability to finance future benefit promises, regardless of the
financing method chosen, depends fundamentally on the capacity to
generate a given amount of resources that will be sufficient to meet
future obligations.  This can be done through a social insurance
program wherein the government makes a political commitment--which
may or may not include issuing debt--or, alternatively, through
advance funding. 

As an element of most private pension plans, advance funding involves
a contractual obligation under which real assets sufficient to meet
the future payments are placed in a legal trust arrangement.  In
contrast, pay-as-you-go requires a political commitment to levy taxes
in the future.\30 Proposals for advance funding Social Security
usually involve investing some portion of current Social Security
contributions in private sector securities (stocks and corporate
bonds) owned by the individual contributors.  It would also be
possible for the government to hold government securities or private
securities, and this approach has been proposed as well.  In both
approaches, increasing Social Security's advance funding has the
potential to capture returns from investment of assets; these returns
could help mitigate the benefit reductions or tax increases that
would otherwise be necessary to restore solvency to the system. 

Supporters of advance funding point out that it offers a way to
increase national saving, investment, and economic growth.  They also
assert that increased economic growth could raise both wages and the
national standard of living, which would reduce the burden of setting
aside a given level of income for retirement.  Thus, they advocate
reducing current consumption in order to increase future consumption. 

Others suggest that the claims of those favoring advance funding may
not be realized.  The linkage between national saving and economic
growth is not certain.  Because future market returns, inflation, and
life expectancies are uncertain, there is no guarantee that a given
level of contributions paid into an advance funded plan would
necessarily be sufficient to provide an expected, or even an
adequate, benefit that would last throughout an individual's
retirement.  Also, an increase in personal or government saving from
advance funding Social Security would not necessarily translate into
an increase in national saving--for example, if the government used
some current Social Security revenue to fund additional personal
saving and then borrowed to continue paying current benefits. 


--------------------
\30 The noncontractual basis of Social Security benefits was
established by the Supreme Court decision in the Nestor case (Fleming
v.  Nestor, 363 U.S.  603, 1960). 


      THE NEED FOR A TRANSITION
      PERIOD COMPLICATES ADVANCE
      FUNDING
-------------------------------------------------------- Chapter 2:2.2

To achieve full advance funding, a transition period might have to
occur during which workers would have to fund both their own future
Social Security benefits and the benefits for those who had already
earned unfunded credits under the current program.  Funding the
program's currently unfunded promises through taxation could place a
large burden on the first group of workers who financed their own
benefits.  Debt financing could reduce the burden on this group and
place some of the burden on later generations that paid off the debt. 
The transition costs could be substantially reduced if some of the
unfunded future benefit obligations were eliminated by reducing the
benefits of current and future beneficiaries. 

Once the transition period had passed and advance funding was fully
implemented, future workers would no longer need to finance the
Social Security benefits of those who were currently working. 
Theoretically, enough money would have been set aside by workers and
employers (in either individual accounts or a collective account) to
secure the benefits of each worker throughout retirement--and,
depending on the proposal's design, perhaps those of his or her
dependents and survivors as well.  In addition, as the burden of
supporting older generations decreased and investment returns funded
an increasing portion of the growth in individual accounts, reducing
individual account contribution rates to a level below today's OASDI
payroll tax rate would be possible. 


      SOME ADVANCE FUNDING IS
      PRESENT IN MOST REFORM
      PROPOSALS
-------------------------------------------------------- Chapter 2:2.3

The Advisory Council has proposed three packages of options.  These
packages capture most of the essential features that are found in
other reform proposals.  While the packages include adjustments of
the current structure (traditional reforms), such as increasing the
retirement age, changing the benefit formula, and lowering the
postretirement COLA, each also contains nontraditional reforms
involving increased advance funding. 

Although all three of the Advisory Council proposals would increase
the system's advance funding, only the PSA and the IA options call
for individual account plans.  The MB proposal instead would increase
the system's advance funding within the current structure, and the
government would invest at least some portion of the additional
assets in the stock market.  Thus, the Advisory Council has indicated
that, to restore solvency, the element of advance funding in private
investment markets should be increased, whether the Social Security
program is strengthened within its current structure or fundamentally
altered. 

The Advisory Council's two individual account proposals represent
what is generally referred to as the "privatization element" in the
current debate.  Precisely defining privatization in relation to the
Social Security debate is difficult, but privatization is usually
associated with two key elements:  advance funding of retirement
income through investment in private financial assets and greater
individual control of decisions about investing those assets.  The
PSA and IA proposals would change the current benefit structure of
Social Security.  Individuals would receive part of their future
benefit from a modified Social Security program and part from the
accumulations from the individual account.  These individual accounts
would be, essentially, advance funded retirement income arrangements,
as are private pensions, and would be similar to defined contribution
pension plans, or 401(k) plans.  These accounts would earn a return
that depended solely on the investment performance of the assets
held, and historical data suggest that the gross returns to these
funded arrangements could be higher than the amounts beneficiaries
could expect to receive under the current system.  The opportunity
for higher returns, however, would come with increased investment
risk that would be borne by the individual owning the account. 


A VARIETY OF OPTIONS COULD RESTORE
PROGRAM SOLVENCY WITHIN THE
EXISTING PROGRAM STRUCTURE
============================================================ Chapter 3

Resolving Social Security's long-term financing problem within the
program's current structure would require increasing the program's
revenues, decreasing its expenditures, or both.  By combining various
options, it would be possible to restore Social Security's actuarial
balance for the next 75 years without changing the program's benefit
or financing structure.  A summary table on the estimated effects of
various options appears as appendix I. 

The options for increasing revenues include expanding coverage to
additional workers, raising the payroll tax rate, expanding taxable
payroll through increasing the maximum taxable earnings level or
including nonwage compensation as covered earnings, increasing the
income taxation of Social Security benefits, using general revenues,
and changing investment policy to earn a higher rate of return on the
Trust Funds' assets. 

The options for reducing expenditures include eliminating or reducing
some existing benefits; reducing initial benefits through changing
the current benefit formula or increasing the NRA, the early
retirement age (ERA), or both; and controlling the growth of benefits
after entitlement through improving COLA calculations, limiting COLA
increases, limiting the recomputation of benefits, restrengthening
the earnings test, disallowing most "new dependent" benefits,\31 or
reducing benefits because of other income.  A number of these options
have been used in the past to ensure the solvency of Social Security. 

Increasing the element of advance funding within the current program
structure is also a means of addressing the solvency problem. 
Increasing the Social Security Trust Funds' assets would require
determining how the government might best reserve those funds for
future benefits. 


--------------------
\31 "New dependents" are those who become entitled to dependent
benefits because of either (1) a marriage that occurs after the
primary beneficiary becomes entitled to benefits or (2) the birth of
a dependent beneficiary that occurs more than a specified period of
time (for example, 9 months) after the primary beneficiary becomes
entitled. 


   OPTIONS FOR INCREASING PROGRAM
   REVENUES
---------------------------------------------------------- Chapter 3:1

Revenues can be increased by expanding coverage, raising additional
revenues through the existing payroll tax structure, and raising
revenue from other sources. 


      EXPANDING COVERAGE
-------------------------------------------------------- Chapter 3:1.1

One way to increase revenues is to expand the number of jobs covered
by Social Security.  This option was first used in 1950.  The
original Social Security Act covered about 60 percent of the U.S. 
workforce.  Today, about 96 percent of the workforce is covered. 
This option increases revenues relatively quickly and improves
solvency for some time, since most of the benefits for the newly
covered workers are future obligations. 

Most beneficiaries have received more in lifetime benefits than they
have paid in payroll taxes.  This would suggest that increasing
coverage would have a long-term negative impact on the program's
solvency.  However, the Advisory Council estimated that covering most
of the remaining noncovered jobs would actually have a positive
effect on program solvency because many of the newly covered workers
would already be eligible for Social Security benefits because of
earnings in other covered employment. 

A majority of the members of the 1994-96 Advisory Council recommended
that all newly hired state and local government workers, who would
not otherwise be covered by Social Security, be covered.\32 They
estimated that this change would represent a net improvement in
actuarial balance equivalent to 0.22 percent of taxable payroll over
the next 75 years, or about 10 percent of the currently estimated
long-term revenue shortfall. 


--------------------
\32 Implementing this recommendation would effectively eliminate the
expansion of coverage as a source of additional program revenues
because almost all jobs would be, or soon would be, covered by Social
Security.  It would also increase employment costs for some states
and localities because they would have to pay the employer's share of
the payroll tax.  This increase in costs would require states and
localities to reduce their expenditures or increase their revenues. 
To offset some of these cost increases, these government entities
might consider modifying the pension systems for newly hired workers. 
The disposable incomes of many newly hired workers would likely be
reduced relative to those of current workers because these new hires
would have to pay both the employee's share of the payroll tax and
any required or voluntary pension contributions to their modified
pension plans. 


      RAISING THE PAYROLL TAX RATE
-------------------------------------------------------- Chapter 3:1.2

Revenues could also be raised by increasing the OASDI payroll tax
rate paid by workers and their employers (currently 6.2 percent of
covered earnings for each) and by the self-employed (currently 12.4
percent).\33 Until 1978, this action was taken quite regularly,
usually by announcing scheduled increases some years in advance to
give workers and employers time to adjust.  The 1977 amendments to
the Social Security Act were the last to raise the OASDI rate for
workers and employers (to 6.2 percent, effective in 1990).  The 1983
amendments raised the payroll tax rate for the self-employed to 12.4
percent, effective in 1990.  No future increases are scheduled even
though the retirement of the baby boom generation is imminent. 
Raising the payroll tax rate by about 1.1 percentage points for both
employees and employers could eliminate the program's currently
projected long-term revenue shortfall. 

One advantage raising the payroll tax has over several other
revenue-enhancing options, from both programmatic and federal budget
perspectives, is that it would not result in higher future benefits
because benefits are based on covered earnings, not total
contributions.  Raising revenues by expanding coverage or expanding
the definition of taxable earnings, on the other hand, would result
in future benefit increases for the affected workers, thereby
reducing the net long-term gains to the program and to the federal
budget. 

Disadvantages of raising the payroll tax include lower disposable
income for workers and higher labor costs for employers.  Moreover, a
higher payroll tax would also lower the value of the program to
workers because future benefits for them and their dependents and
survivors would not increase.  Because employers' additional costs
would be tax-deductible, their business income taxes would fall, but
by less than the payroll tax increase.  The end result would be that
employers' net incomes would fall somewhat, and federal income tax
revenues would decline. 

In addition, the Congress might be reluctant to further increase the
payroll tax rate because (1) it and other tax rates are already
considered too high by many, (2) many workers already face higher
payroll taxes than income taxes, and (3) the payroll tax is
regressive.  The Advisory Council concluded that there is little
political support for bringing the program back into financial
balance through payroll tax rate increases alone.  However, all three
Advisory Council proposals contained payroll tax increases as a part
of their recommended solution to the program's solvency problem.  One
recommended an immediate and permanent payroll tax increase, one a
permanent increase beginning in about 50 years, and one a temporary
(70-year) increase.  Moreover, the Medicare program faces a more
immediate solvency problem than does Social Security, and increasing
the payroll tax rate to improve the long-term financial solvency of
one program limits the extent to which this option can be used to
improve the long-term financial solvency of another. 


--------------------
\33 The OASDI tax rate was initially set at 1 percent of the first
$3,000 of earnings for both the employee and the employer.  The rate
increased 20 times between 1937, when the tax was first collected,
and 1990, when the rate reached its current level. 


      EXPANDING TAXABLE PAYROLL
-------------------------------------------------------- Chapter 3:1.3

There are two ways to expand the taxable payroll base:  raising the
maximum level of earnings subject to the payroll tax and including
some nonwage compensation in the definition of taxable payroll. 


         INCREASING THE MAXIMUM
         TAXABLE EARNINGS LEVEL
------------------------------------------------------ Chapter 3:1.3.1

Over the years, the maximum taxable earnings level has risen from
$3,000, initially, to $68,400 in 1998.  In 1995, covered earnings
accounted for about 88 percent of all earnings for employees and
about 72 percent of reported self-employment net earnings.  Overall,
about 87 percent of all earnings were covered by Social Security. 
The maximum taxable level is automatically adjusted to the growth in
national wages, and this generally increases program revenues over
time. 

While increasing the taxable earnings level would generate additional
program revenues immediately, it would also increase future costs by
raising benefits for those high earners who would pay the additional
payroll taxes.  However, because the additional covered earnings
generally would increase the benefits of high earners only modestly
(recall that the rate of earnings replacement for the highest
increments of the AIME is only 15 percent), raising the maximum
taxable earnings level could increase revenues in both the short and
long run.  Social Security actuaries estimated that raising the
maximum taxable earnings level in 1997 and later so that 90 percent
of all earnings were taxable (a 3-percentage point increase over
current levels) would improve the program's long-range actuarial
balance by 0.48 percent of taxable payroll, or the equivalent of
about 22 percent of the program's estimated 75-year financing
shortfall. 


         INCLUDING SOME NONWAGE
         COMPENSATION AS COVERED
         EARNINGS
------------------------------------------------------ Chapter 3:1.3.2

Over the past few decades, the proportion of total compensation paid
in the form of wages and salaries has declined, and nonwage
compensation (payments for pension contributions and health
insurance, for example), which is not subject to the payroll tax, has
risen to about one-third of payroll.  This increase in the benefits
portion of total compensation has reduced the relative amount of
total compensation subject to the payroll tax.  Social Security
revenues could be increased if some or all of these nonwage
compensation costs were included in the definition of taxable
payroll.\34 Estimates made for the Advisory Council suggest that
including employer-provided group health and life insurance or
pension and profit-sharing contributions in OASDI taxable earnings
would improve the program's long-term actuarial balance by 0.80 and
0.37 percent of taxable payroll, respectively.  Combined, these two
options represent about one-half of the anticipated financing
shortfall. 

This option could present some difficulties in implementation,
however.  Employee benefits generally are greater for highly paid
workers whose wage compensation may already exceed the maximum
taxable earnings limit.  Thus, subjecting their nonwage compensation
to the payroll tax would not raise additional revenues.  Also, it
could be difficult to separate nonwage benefit costs on an individual
basis.\35 If such an individual allocation could be made, the
increase in taxable payroll would increase the future benefits of
many workers.  An alternative would be for only employers to pay the
additional tax on nonwage compensation.  Subjecting all
employer-sponsored private pension and profit-sharing contributions
to a 3-percent payroll tax and crediting these contributions as
earnings to individual workers would improve OASDI's long-term
actuarial balance by an estimated 0.15 percent of taxable payroll. 


--------------------
\34 However, subjecting these nonwage forms of compensation to
payroll taxation could reduce their attractiveness to both employers
and employees.  If this occurred, there could be a reduction in the
provision of these forms of compensation, leading to a decrease in
total retirement income for some workers in the future. 

\35 Employers often make their payments for these benefits in lump
sums (for example, they might pay $1 million as a premium to provide
their workers with health care coverage) and do not generally
determine how much of this payment should be allocated to each
employee.  Indeed, determining an appropriate allocation for each
employee would prove to be a daunting task in many instances. 


      INCREASING THE INCOME TAX ON
      SOCIAL SECURITY BENEFITS
-------------------------------------------------------- Chapter 3:1.4

Up to one-half of Social Security benefits have been subject to
individual income taxes since 1984.\36 \37 These revenues are
returned to the Social Security Trust Funds.  Taxing Social Security
benefits can be considered either a form of means testing
benefits--because one's total Social Security benefit is effectively
reduced as income rises--or a way to partially fund the program out
of general revenues. 

Increasing revenues by taxing Social Security benefits could be
accomplished by several means, including lowering or eliminating the
income thresholds at which benefits become taxable, taxing all
benefits above the amount of the employee's contributions,
redistributing to Social Security the portion of benefit taxation
currently going to Medicare, and treating all Social Security
benefits as normal taxable income subject to the current income tax
rules.  Eliminating the thresholds but otherwise keeping the benefit
taxation provisions as they are is estimated to improve the program's
long-term actuarial balance by 0.21 percent of taxable payroll. 
Lowering or eliminating the thresholds would require increased income
tax payments from some lower-income beneficiaries; higher-income
beneficiaries would not contribute more unless the proportion of
benefits subject to this tax was also increased.  Taxing all Social
Security benefits that exceeded the worker's own contributions would
save another 0.15 percent of taxable payroll.  Shifting the HI
portion of benefit taxation to OASDI would save 0.36 percent, but at
the expense of worsening Medicare's solvency problem.  Finally,
making all Social Security benefits subject to the income tax while
keeping the current thresholds in place would increase income taxes
for both those higher-income beneficiaries currently paying the tax
on Social Security benefits and those whose total incomes are close
to, but below, the current thresholds. 


--------------------
\36 The 1993 amendments to the Social Security Act made up to 85
percent of Social Security benefits subject to income taxation. 
However, the additional revenues collected from this source are
dedicated to the HI Trust Fund and do not increase OASDI revenues. 

\37 Individual income tax filers pay the tax if their adjusted gross
income plus tax-exempt interest income plus one-half their Social
Security benefits exceed $25,000.  A married couple filing jointly
will pay the tax if this income exceeds $32,000.  The threshold for
married couples filing separately is $0 (half of all Social Security
benefits are automatically subject to taxation) if the couple lived
together at any time during the tax year.  The thresholds are not
indexed, so the percentage of beneficiaries subject to this tax will
rise as the nominal amount of their total income, as taxable for this
purpose, increases. 


      USING GENERAL REVENUES
-------------------------------------------------------- Chapter 3:1.5

The program's revenues could also be increased by partially funding
the system with money from other government revenue sources.\38
General revenue funding of the program has been used in the past,
most notably during the program's 1982-83 financing crisis.\39
General revenue financing of a portion of Social Security expenses
could be accomplished by dedicating a portion of existing general
revenues to the Social Security program; creating a new tax, such as
a national consumption tax, with proceeds dedicated to Social
Security; and reducing expenditures on other federal programs and
using the cost savings to help fund the program. 


--------------------
\38 Some of the original designers of the program assumed that the
government would eventually share in the costs of the program. 

\39 The 1983 amendments directed the Treasury to make payments to the
OASDI Trust Funds from general revenues for unfunded gratuitous
military service credits for military service after 1939, the value
of uncashed benefit checks issued in the past (including interest),
revenues from the income taxation of up to 50 percent of Social
Security benefits paid, and tax credits given for Federal Insurance
Contributors Act and Self-Employment Contributions Act taxes paid by
workers from 1984 through 1989. 


      EARNING A HIGHER RATE OF
      RETURN ON THE TRUST FUNDS'
      ASSETS
-------------------------------------------------------- Chapter 3:1.6

Currently, the Trust Funds are invested in Treasury securities that
earn a relatively low rate of return.  Investing a portion of Social
Security Trust Funds in the stock market could increase the return to
the fund, albeit with a risk of capital loss.  While stocks and other
investments do not outperform Treasury securities every year, they
have, over the long term, performed much better. 

Higher investment earnings could extend the life of the Trust Funds
without other program changes.  As we reported previously, investing
the projected Trust Funds' surpluses, absent other changes to the
Social Security program, could extend the life of the Trust Funds by
almost 11 years, assuming stock returns remained at the historical
average.\40 If this were implemented in isolation, the Trust Funds
would inevitably have to liquidate the stock portfolio to pay
promised benefits and would be vulnerable to losses in the event of a
general stock market downturn.  While stock investments alone would
not completely address the program's long-term solvency, they could
lessen the size of other program changes needed to bring the program
to solvency.  This option is addressed in greater detail in the
advance funding discussion later in this chapter. 


--------------------
\40 See Social Security Financing:  Implications of Government Stock
Investing for the Trust Fund, the Federal Budget, and the Economy
(GAO/AIMD/HEHS-98-74, Apr.  22, 1998). 


   OPTIONS FOR REDUCING PROGRAM
   EXPENDITURES
---------------------------------------------------------- Chapter 3:2

Until the 1970s, most attempts to address financing problems focused
on increasing program revenues.  But expenditures can be controlled,
or reduced, in numerous ways, including eliminating or reducing some
existing benefits, reducing initial benefit levels, and slowing the
increase in benefits once they have been initiated. 


      ELIMINATING OR REDUCING SOME
      EXISTING BENEFITS
-------------------------------------------------------- Chapter 3:2.1

Eliminating benefits has been used only sparingly in the past, most
notably in the early 1980s when the following benefits were
abolished:  the minimum Social Security benefit for those attaining
age 62 after 1982, child benefits for students aged 18 to 22, and
benefits for (widowed) mothers and fathers whose youngest nondisabled
child has attained age 16. 

Reducing benefits for selected beneficiaries has been used a little
more often.  In 1967, a limitation of $105 per month was placed on
spousal benefits, but this limit was quickly removed in 1969.  The
process for determining Social Security benefits was modified in 1977
to offset unintended increases in initial benefit levels that
resulted from a benefit calculation process first used in 1975.  In
1980, the method of computing the applicable family maximum benefits
on the basis of the earnings records of those who became disabled
after June 1980 was changed in a way that effectively limited the
total benefits the spouses and children of disabled workers could
receive.  Social Security benefits were also reduced in 1977 and 1983
for those who had pensions from noncovered government employment at
the federal, state, or local level.\41 In addition, the 1983 program
amendments reduced benefits by delaying the COLA for 6 months and by
raising the NRA for those born in 1938 or later. 


--------------------
\41 These reductions are the result of the windfall elimination
provision (WEP), which is intended to reduce the retired worker
benefits of the affected workers, and the government pension offset
(GPO), which is intended to reduce or eliminate the Social Security
spouse or survivor benefits the worker might have been entitled to on
the basis of his or her spouse's earnings record.  The WEP and GPO
were enacted to keep workers with substantial work in noncovered
employment from taking advantage of the progressiveness of the Social
Security benefit formula, which is intended to boost the Social
Security benefits of long-term, low-wage workers and workers who have
only marginal attachment to the labor force. 


         SPOUSAL, SURVIVOR, AND
         DEPENDENT CHILD BENEFITS
------------------------------------------------------ Chapter 3:2.1.1

The spouses, children, and parents of retired and disabled workers,
as well as survivor beneficiaries, receive Social Security benefits
that are based at least in part on the covered earnings record of
retired, disabled, or deceased workers.  These benefits were added in
1939 to ensure that a worker's family had adequate benefits once the
worker retired; died; or, after 1956, became disabled.  These
benefits currently account for more than 25 percent of all program
expenditures.  No absolute measure of need or adequacy has ever been
applied to these benefits.  For example, eligible spouses receive a
benefit based on one-half the worker's PIA regardless of the amount
of the worker's benefit.  At the end of 1996, 73 percent of the
spouses of retired workers had their benefits based on PIAs of $800
or more, fewer than half of all retired workers had their benefits
based on PIAs this high, and less than 40 percent of disabled
beneficiaries but more than 50 percent of their spouses had benefits
based on PIAs of $800 or more.\42 The average PIAs on which
children's benefits were based also exceeded those of retired or
disabled workers. 

Limiting spousal benefits could be accomplished by, for example,
capping them at one-half the average retired worker's PIA, or by
phasing them out if the combined benefits of the worker and spouse
exceeded a given threshold.  The benefits of workers with low
lifetime earnings and those of their spouses would continue to be
paid as under current law, but the benefits for spouses of workers
with higher than average PIAs would be reduced.  Limiting spousal
benefits to one-half the average PIA of retired workers as of
December of the prior year is estimated to improve the program's
long-term actuarial balance by 0.21 percent of covered payroll. 

At the end of 1996, benefits for most types of survivors were also
based on average PIAs that were higher than the average PIAs of
retired workers, although not as high as PIAs for spouses.  The
maximum monthly benefit for a worker retiring at age 65 in 1996 was
$1,284 in December of that year.  More than 1 million beneficiaries
receiving only survivor benefits at that time had their benefits
based on PIAs of $1,100 or more, and 38 percent of these had benefits
in excess of $1,250 that month.\43 At the same time, about 200,000
beneficiaries were entitled to combined retired worker and survivor
benefits in excess of $1,200 (averaging about $1,400).  Thus,
hundreds of thousands of survivor beneficiaries received benefits in
excess of what a 65-year-old worker retiring in that year could have
received.  If it were desirable to do so, this situation might be
addressed by, for example, capping survivor benefits at some
percentage above the poverty threshold, at the average retired worker
benefit level, or at the maximum benefit available to a worker
attaining age 65 in the year the survivor became widowed.\44

Costs could also be reduced by modifying children's benefits.  For
example, eliminating benefits for nondisabled children of retired
workers is estimated to save 0.05 percent of taxable payroll.  Also,
the level of benefits for children of disabled and deceased workers
could be made dependent on the earnings that continue to come into
the household from the nondisabled or nondeceased parent and not just
on the child's own earnings.  There is already a precedent for this
type of reduction, in that the benefits of auxiliary beneficiaries
can be reduced not only by their own earnings but also by those of
the retired worker.  This action would save about 0.04 percent of
taxable payroll over the 75-year period. 

Capping or eliminating certain spousal, survivor, and dependent child
benefits, or tying them to the amount of household income, could
ensure that lower-earning families continue to receive adequate
auxiliary benefits while higher-earning families do not receive
benefits that are difficult to justify on adequacy grounds. 


--------------------
\42 At this time, the average PIA for a retired worker was $753 and
for his or her spouse, $939.  The average PIAs for disabled workers
and their spouses were $711 and $834, respectively. 

\43 This occurred primarily because the recomputation of benefits for
those with covered earnings after age 65 increased their PIAs and
their survivors' benefits above the maximum $1,284 a 65-year-old
worker could receive in December 1996. 

\44 This last action would save only about 0.01 percent of covered
payroll over the 75-year evaluation period. 


         DISABLED WORKER BENEFITS
------------------------------------------------------ Chapter 3:2.1.2

The disability insurance (DI) program has been one of the fastest
growing Social Security-administered programs over the past 10 years. 
Controlling the growth in the DI program would be an important way to
control overall program expenditure growth.  This could be done by
tightening program eligibility requirements; making determinations of
eligibility at various review levels more consistent; taking action
to encourage DI beneficiaries to return to work; limiting how long DI
beneficiaries can be on the rolls;\45

reducing DI benefits by lowering initial levels of all benefits; and
limiting the initial disabled worker benefit to the retired-worker
benefit available at age 65, using the current law's increasing
retirement ages and adjustment factors.  This last means of reducing
disabled-worker benefits is estimated to improve the program's
long-term solvency by 0.40 percent of taxable payroll. 


--------------------
\45 See Social Security Disability:  SSA Must Hold Itself Accountable
for Continued Improvements in Decision-making (GAO/HEHS-97-102, Aug. 
12, 1997); Social Security:  Disability Programs Lag in Promoting
Return to Work (GAO/HEHS-97-46, Mar.  17, 1997); and Social Security
Disability:  Improvements Needed to Continuing Disability Review
Process (GAO/HEHS-97-1, Oct.  16, 1996). 


      REDUCING INITIAL BENEFITS
-------------------------------------------------------- Chapter 3:2.2

Expenditures for retired-worker benefits will increase rapidly once
the baby boom generation begins to retire.  To help control these
anticipated expenditure increases, initial benefits for all
beneficiaries could be reduced through (1) changing the current
benefit formula and (2) increasing the NRA or the ERA--or both. 
Reducing the growth in benefits once they are received is also an
option. 


         CHANGING THE CURRENT
         BENEFIT FORMULA
------------------------------------------------------ Chapter 3:2.2.1

Benefits for those born in 1929 or later are based on the average of
a worker's 35 years of highest indexed covered earnings.  Earnings
received before age 60 are wage-indexed to the year the worker turned
age 60.\46

Once the average indexed monthly earnings are determined, a formula
converts them to the PIA.  Benefits equal 90 percent of average
earnings up to a threshold ($477 for 1998), plus 32 percent of
average earnings above this first threshold until a second ($2,875)
is reached, plus 15 percent of average earnings the worker might have
above this second threshold.  The PIA is then adjusted for the age
the worker first receives benefits.  The benefit is lowered if
benefits are first taken before the NRA (currently age 65) and
increased if benefits are first received after the month the worker
attains the NRA but before age 70. 

Initial benefits could be reduced by changing the values of
components of the benefit formula--for example, increasing the number
of years of earnings included in the computation period from 35 to
38, as a majority on the Advisory Council advocated.  The indexed
earnings of the additional 3 years would, by definition, be no larger
than the indexed earnings of the year of lowest earnings included
under current rules.  This change would result in a decrease in both
average indexed earnings and benefit amounts for all new
beneficiaries. 

The reductions from extending the computation period would be larger
for those with limited or intermittent attachment to the labor force
than for those with continuous attachment, because more years of $0
earnings would be included in the computation formula--for example,
women would be more affected than men.  According to the Advisory
Council's report, increasing the computation period would reduce
benefits by 3 percent, on average, and improve the program's
long-term actuarial balance by 0.28 percent of taxable earnings. 
Those with 35 or fewer years of earnings, however, would experience
about an 8-percent decrease in AIME, and many beneficiaries with
fewer than 36 years of earnings already have relatively low AIMEs. 
This change would reduce the benefits for those with low lifetime
covered earnings more than for those with high lifetime covered
earnings.  A $1 decrease in AIME could reduce the PIA of a low earner
by 90 cents, while the PIA of the highest earners would be reduced by
only 15 cents. 

Another way to reduce initial benefits would be to lower either the
rates of earnings replacement or the bend points that convert average
earnings to benefits.  Reducing all replacement rates would reduce
benefits for everyone, including those with the lowest AIMEs and
benefits.\47 Gradually reducing each of the three replacement rates
by 0.5 percent between 2020 and 2029 and maintaining them at the new,
lower levels thereafter is estimated to improve the program's
long-term actuarial balance by 0.29 percent of taxable payroll.\48
Reducing the bend points would protect the benefits of those with the
lowest benefits but reduce benefits for everyone with average
earnings above the new (lower) first bend point.  Indexing the bend
points in the benefit formula by either the current consumer price
index or the annual wage index minus 1 percentage point rather than
by the average wage index would be expected to reduce the new benefit
rate of growth.  Either index adjustment would improve the program's
long-term actuarial balance by 1.54 percent of taxable payroll, about
70 percent of the long-term financial imbalance. 

Initial benefits could also be reduced by increasing the reduction
factor for early retirement and reducing the incremental increase for
first receiving benefits after the NRA.  In addition, the benefit
formula could be reduced by indexing benefits to a younger age than
age 60 or by using an index that grows more slowly than national
wages.  These last changes would reduce Social Security's measure of
lifetime covered earnings which, in turn, would reduce calculated
benefits. 


--------------------
\46 Earnings at age 60 and older are not indexed. 

\47 Reducing only the middle and lowest replacement rates would
preserve the PIAs for all those whose AIMEs are at or below the first
bend point.  Reducing only the lowest replacement rate would reduce
PIAs for only those with the highest AIMEs. 

\48 Such a reduction would lower the current replacement rates of 90
percent, 32 percent, and 15 percent to 85.5 percent, 30.4 percent,
and 14.25 percent, respectively. 


         INCREASING THE NRA, THE
         ERA, OR BOTH
------------------------------------------------------ Chapter 3:2.2.2

An increase in the NRA would be tantamount to a graduated benefit
reduction for all affected beneficiaries.  Some policymakers are
concerned that this additional reduction in benefits for those who
retire early--especially for those who have health problems and for
those who are widows--would reduce the adequacy of their benefits and
result in an impoverished retirement. 

The NRA has already been increased once.  The package of program
changes used to resolve the program's 1982-83 financing crisis
included a provision to gradually increase the NRA from age 65 to age
67 beginning with those born in 1938 (and attaining age 62 in the
year 2000).  The NRA increase will be fully phased in for those born
in 1960 or later.\49 However, the ERA of 62 was not changed. 

Increasing the NRA further can be justified because life expectancies
at age 65 are longer now than they were in 1940, the year benefits
were first paid.\50 The longevity trend is an important reason for
the growth in Social Security costs.  Increasing the NRA would be one
way to control program costs because benefits available at all ages
would be lowered, and this could provide an incentive for some
workers to delay their initial receipt of retired worker benefits. 

How much to increase the NRA would depend on the goal of the
increase.  If the goal was to keep the program solvent, the increase
in the NRA could be calculated once the other actions to maintain
solvency had been decided on.  However, the goal of increasing the
NRA could also be either to keep life expectancy at the NRA constant
(using life expectancy at age 65 in 1940 or some other year as a
base) or to maintain a life expectancy at the NRA that is a constant
proportion of one's life expectancy as an adult (life span after age
20).\51 For example, in 1940 at age 65 the average life expectancy
was just under 13 years.  To keep the same 13-year life expectancy at
the NRA in 1995, the NRA would have had to be age 72.  Alternatively,
in 1940 the average person aged 65 would have expected to spend about
22 percent of his or her adult life older than the NRA.  In 1995,
spending 22 percent of one's adult life above the NRA would require
an NRA of age 70, using the Social Security Actuary's projections of
life expectancies.  Given either of these two goals, the NRA would
need to be increased as life expectancies continue to improve. 

More than 50 percent of newly retired workers elect to receive
benefits at age 62.  Increasing the ERA would preclude workers from
claiming benefits between age 62 and the new ERA and could,
therefore, increase the incentive to apply for DI benefits at those
ages.  Social Security would receive some short-term financial
savings because these potential beneficiaries would have to delay the
receipt of benefits.  However, because benefits are adjusted on an
actuarial basis, the initial benefits of affected workers would be
larger than if the ERA had remained at age 62, and long-term program
savings would be low. 

Raising the NRA, the ERA, or both could place a large burden on the
DI program and result in lower net savings than might be expected. 
Raising the NRA would increase the reduction factor applicable to
those retiring at the ERA, giving them lower benefits than they
currently receive.  Raising the NRA would not reduce the amount of
the DI benefit, however, unless DI benefits were reduced
independently.  The benefit gap between DI benefits and the new,
lower retirement benefits for everyone below the new NRA would rise,
providing an incentive for some, who would not otherwise do so, to
apply for DI benefits.\52 DI caseloads and costs would grow if the
number of applicants increased and, if some of these additional
applicants were allowed on the DI rolls, DI benefit costs (and total
OASDI costs) also would increase. 


--------------------
\49 The legislated change in the NRA of 65 increases it by 2 months
each year for those born from 1938 to 1943.  That is, the NRA of a
person born in 1943 is now 66.  The NRA will not be increased for
those born from 1944 through 1954 (the hiatus)--it will remain at age
66.  Those born from 1955 to 1960 will again see their NRA increase
by 2 months each year.  The NRA for those born in 1960 and later will
be age 67.  For example, those born in 1938 will have to wait until
they are 65 years and 2 months old to receive their PIA or "full
benefit." If they retire during the month they attain age 62 or 65,
for example, they will receive only 79 or 99 percent of their PIA,
whereas they would have received 80 or 100 percent if they had they
been born 1 year earlier.  These reductions will continue to grow for
those born from 1938 to 1943 and again for those born from 1955 to
1960.  Those born after 1959 will receive only 70 percent of their
PIA at age 62 and about 87 percent at age 65, compared with today's
80 percent and 100 percent, respectively.

Eliminating the currently scheduled hiatus so that an NRA of age 67
would be reached for those born in 1949 and indexing the NRA
thereafter to keep the proportion of the average adult lifetime that
is above the NRA constant are estimated to improve OASDI's long-term
actuarial balance by 0.50 percent of taxable payroll. 

\50 Life expectancies at age 65 have increased by about 30 percent
for men and about 40 percent for women since 1940. 

\51 We calculated this proportion by dividing the life expectancy at
the NRA by the NRA plus the life expectancy at the NRA, minus 20. 

\52 For example, under current law, retired worker benefits taken at
age 62 after the NRA has increased to age 67 will be 70 percent of
the worker's PIA.  Disability benefits will remain at 100 percent of
PIA.  If a 62-year-old worker in marginal health decided to apply for
disability benefits rather than reduced retired worker benefits and
was approved for DI benefits, his or her monthly benefits would be
about 43 percent higher, for life.  This level of potential increase
in monthly benefits could provide a strong incentive for many
retirement-eligible people with health problems to apply for
disability benefits. 


      CONTROLLING THE GROWTH IN
      BENEFITS AFTER ENTITLEMENT
-------------------------------------------------------- Chapter 3:2.3

In addition to reducing the level of initial Social Security
benefits, controlling the growth of benefits after initial receipt is
another way to reduce program expenditures.  Various possible actions
are discussed below. 


         IMPROVING COLA
         CALCULATIONS
------------------------------------------------------ Chapter 3:2.3.1

Since 1975, Social Security benefits have been automatically
increased to keep pace with inflation using the consumer price index
as the inflation index.  This automatic increase allows benefits to
maintain their purchasing power over time.\53 However, COLAs are
costly.  Social Security currently pays about $370 billion a year in
benefits.  Each 1-percent increase in the COLA costs the program an
additional $3.7 billion.  Because COLA increases are cumulative,
their impact on program expenditures grows rapidly.  For example,
those who first received benefits in the first half of 1975 currently
receive monthly benefits that are 187 percent higher (in nominal
terms) than their original monthly benefit; that is, for each $100
received in early 1975, $287 is received in 1998. 

Recently, a congressional commission reported that the consumer price
index overstates the true rate of inflation on average by about 1.1
percentage points yearly, and that this may result in
overcompensation of beneficiaries.\54 Many economists agree that the
consumer price index probably overstates the rate of inflation but
differ on the degree.  Even the Bureau of Labor Statistics, which
calculates the increase in the index, consistently states that it is
not a measure of inflation. 

Improving the calculation of the COLA, either by making the consumer
price index a more accurate measure of inflation (which is
technically difficult to do) or by adjusting it after the fact to
better measure true changes in inflation, is a desirable option. 
Given the direction of the current bias in the index, such an
adjustment would lower yearly COLAs and result in long-term
improvements in the program's solvency. 


--------------------
\53 For most retirees, Social Security benefits are the only source
of retirement income that maintains its purchasing power through
time.  Other sources generally fail to keep pace with inflation or
disappear at some point after retirement.  Thus, most retirees become
more and more dependent on Social Security benefits as they age. 

\54 Michael J.  Boskin and others, Toward a More Accurate Measure of
the Cost of Living, final report to the Senate Finance Committee from
the Advisory Commission to Study the Consumer Price Index
(Washington, D.C.:  Advisory Commission to Study the Consumer Price
Index, Dec.  4, 1996). 


         LIMITING COLA INCREASES
------------------------------------------------------ Chapter 3:2.3.2

Reducing COLAs could control the growth in Social Security benefit
expenditures.  Expenditure savings would be apparent immediately, and
savings in 1 year would carry forward in later years in a cumulative
manner.  In addition, COLA reductions would affect current as well as
future beneficiaries, spreading the burden of the program's financial
reform over a broader population.  Not all other actions to resolve
the program's long-term solvency problem would affect current
beneficiaries. 

COLA reductions could be achieved by several means, including

  -- lowering the COLA to less than the measured rate of inflation
     (for example, consumer price index minus 1 percentage point);\55

  -- capping the COLA (increasing benefits by the consumer price
     index increase or, for example, 2.5 percent, whichever is less);

  -- delaying the COLA;

  -- eliminating the COLA;

  -- changing the index used to measure the COLA;

  -- not providing a COLA until cumulative inflation since the
     previous COLA increase exceeds a specified threshold, such as 5
     percent; and

  -- allowing a full COLA up to some specified threshold (for
     example, the average PIA amount) and then reducing or
     eliminating COLAs for benefits above that threshold. 

These alternative ways of reducing COLAs would have differing impacts
on certain individuals and households.  For example, changing the
COLA by reducing the consumer price index by 1 percentage point
forever would gradually reduce the purchasing power of benefits as
beneficiaries age.  A reduction in the COLA from, for example, 3.5
percent to 2.5 percent annually would reduce the purchasing power of
benefits by about 9 percent after 10 years, 22 percent after 25
years, and 32 percent after 40 years.  Alternatively, giving full
COLAs for benefits below some threshold (the average PIA amount, for
example) and giving reduced or no COLAs for benefits above that
threshold would fully protect the purchasing power of benefits for
those with low benefit levels while gradually reducing it for those
with higher benefit levels. 

Reducing COLAs would have an important drawback, however.  The
purchasing power of Social Security benefits would gradually shrink
over time.  As they age, some beneficiaries with little or no
additional retirement income could be pushed into poverty as a result
of COLA cuts.  This could be a particular problem for single
(widowed, divorced, or never married) elderly women who already have
one of the highest poverty rates of any population subgroup in the
nation.  In 1994, 22 percent of single women aged 65 or older lived
in poverty, and another 12 percent had incomes between 100 percent
and 125 percent of the poverty line.  As more beneficiaries fell into
poverty, more would become eligible for government-provided safety
net programs, such as Supplemental Security Income (SSI).  Increases
in the costs for these safety net programs would partially offset the
savings to Social Security from the COLA reductions. 


--------------------
\55 Reducing the COLA to equal the consumer price index minus 1.0 or
0.5 percentage points beginning in 1998 is estimated to improve
OASDI's long-term actuarial balance by 1.39 or 0.72 percent of
taxable payroll, respectively. 


         LIMITING THE
         RECOMPUTATION OF BENEFITS
------------------------------------------------------ Chapter 3:2.3.3

The benefits of those who continue to work after age 62 are
recomputed to account for their new earnings, even if they receive
benefits while working.  If their current earnings are larger than
the smallest earnings currently used in calculating their current
benefit level, the new earnings will replace those smallest earnings,
and their benefits and those of their dependents will increase for
all future years. 

Another way to reduce future program costs would be to limit the
recomputation of benefits, which could be done by allowing
recomputation of the benefits of only those who did not receive any
benefits during the year they worked; capping benefits at the maximum
benefit payable to someone in that worker's birth cohort who first
drew benefits at age 65, adjusted for subsequent COLAs; or applying
any benefit recalculation only to the worker's own benefit and not to
any dependent benefits based on his earnings record.  However, those
who currently work and receive Social Security benefits could argue
that they are paying payroll taxes on their current earnings and that
these earnings should be included in the benefit recalculation if it
is to their advantage. 


         RESTRENGTHENING THE
         EARNINGS TEST
------------------------------------------------------ Chapter 3:2.3.4

The earnings test was originally designed to control program costs by
ensuring that only those who lost their earnings because of
retirement would receive benefits.  However, the earnings test has
been relaxed many times over the past 60 years.\56 This relaxation of
the earnings test has been very costly to the program.  SSA estimates
that, in 2000, it will pay about $80 billion to working beneficiaries
and their dependents, about 20 percent of the program's estimated
total benefit expenditure.  This does not mean Social Security
benefits would be reduced by $80 billion yearly if a draconian
earnings test were reintroduced, however, because many of those who
currently work and receive benefits would choose to forgo their
earnings rather than their benefits. 

The earnings test could be strengthened by reducing the threshold at
which the test first applies by (1) either increasing the amount
Social Security benefits are reduced for each dollar of earnings
above the threshold or reducing benefits by a given percentage for
each dollar of earnings above the threshold\57 or by (2) increasing
the age at which the test no longer applies, perhaps in line with any
increase in the NRA. 


--------------------
\56 Today, there are two annual earnings tests.  Those under age 65
can earn $9,120 annually without penalty, after which point benefits
are reduced $1 for each $2 of additional earnings.  For those aged 65
through 69, the threshold level is $14,500 (rising to $30,000 in
2002), and benefits are reduced $1 for each $3 of additional
earnings.  Once a person attains age 70, the earnings test no longer
applies. 

\57 This would result in all benefits being withheld from all
affected beneficiaries once their earnings exceeded the threshold by
a given dollar amount.  Currently, those with higher benefit levels
can earn more income while receiving some benefits than can those
with lower benefit entitlements. 


         DISALLOWING MOST "NEW
         DEPENDENT" BENEFITS
------------------------------------------------------ Chapter 3:2.3.5

Disallowing dependent benefits for those who were not dependents when
the beneficiary became entitled to his or her current benefits is
another means of controlling the growth in benefits after
entitlement.  Exceptions might be made for newly born children who
were being carried by a pregnant beneficiary or spouse when the
beneficiary became entitled to benefits and for dependents who are
not yet eligible for auxiliary benefits because they do not yet meet
all eligibility requirements, such as age requirements. 


         REDUCING BENEFITS BY
         MEANS TESTING
------------------------------------------------------ Chapter 3:2.3.6

Some have suggested reducing program costs by means testing Social
Security benefits.  To an extent, means testing is already being done
via the income tax on benefits and the earnings test.  Means testing
via these options could be enhanced as discussed earlier in this
chapter. 

Benefits for some could also be eliminated or reduced further by more
traditional means testing, which would act essentially as a tax. 
Means testing works by determining whether a beneficiary has other
income above a specified threshold and then either eliminating the
benefit if the "income from other sources" threshold is exceeded
(implying an infinite tax rate) or reducing the benefit according to
some formula related to how much the other income exceeds the
threshold (the formula determines the tax rate, which could be 100
percent or even higher).  A means test need not be based on all the
non-Social Security income of a beneficiary.  Social Security
benefits could also be reduced, regardless of the beneficiary's gross
income level, if the beneficiary had income from a specified source,
such as savings income or a pension--an alternative already being
used to reduce the Social Security benefits of many federal, state,
and local government workers who receive pension benefits from
employment not covered by Social Security. 

But a means-test tax could lead to economic inefficiencies by
changing individuals' behavior.  For example, if having any other
retirement income could cause a reduction in Social Security
benefits, some workers might be reluctant to save for retirement,
whether through employer pensions, individual savings, or any other
means-tested vehicle.  Such workers might prefer to spend their
earnings before they retired rather than have their saved earnings
reduce retirement income they otherwise would have received.  Such a
reallocation of consumption from the future to the present could
reduce our already near-historically-low national saving rate.  This
type of behavior can be seen when people shift their income, assets,
or both to family or other entities so they can qualify for
government-provided Medicaid, SSI, or long-term care. 

Means testing benefits would eliminate or further reduce Social
Security benefits for many higher-earning beneficiaries.  But these
individuals tend to pay the largest amount of payroll taxes and
receive the smallest percentage return on those contributions.\58
Moreover, means testing the benefits of these individuals could
undermine their political support for the Social Security program,
and their support is essential if Social Security is to maintain its
financing and benefit structures. 


--------------------
\58 High earners receive the largest monthly benefits, in dollar
terms, and tend to live longer than those with lower incomes.  Thus,
the lifetime returns on their contributions tend to be closer to the
returns for workers with lower incomes than Social Security's
replacement rate estimates would indicate. 


   ADVANCE FUNDING WITHIN THE
   CURRENT PROGRAM STRUCTURE
---------------------------------------------------------- Chapter 3:3

Although Social Security's long-term financing problem could be
addressed without significant change to the primarily pay-as-you-go
approach currently in use, some have proposed that the solvency
problem could be better addressed with greater reliance on advance
funding.  Two main mechanisms for advance funding exist within Social
Security's government-managed structure:  advance funding through a
buildup of Treasury securities and advance funding through government
investments. 


      ADVANCE FUNDING THROUGH A
      BUILDUP OF TREASURY
      SECURITIES
-------------------------------------------------------- Chapter 3:3.1

Currently, the Treasury issues its securities to the Trust Funds in
exchange for the program's excess revenues.  These securities are
backed by the U.S.  government and have virtually no risk of default. 
However, they also represent obligations the government issues to
itself.  From the Social Security Trust Funds' perspective, these
securities represent program assets--they signify a reserve budget
authority that can be used to meet future benefit obligations. 
However, from the perspective of the rest of the government, these
securities are not assets but claims against the Treasury. 

One method of advance funding Social Security would essentially
retain the program's current financing, Trust Funds, and benefit
structures.  Indeed, the current program is already building up a
sizeable, but temporary, level of assets that could be used to pay
some of the benefits the baby boom generation will need once it
retires.  The degree of buildup could be enhanced by increasing the
program's excess cash revenues through increasing revenues or
decreasing expenditures.  For example, program revenues could be
raised by increasing the total payroll tax by 2.19 percentage points
and "investing" all excess revenues in Treasury securities.  This
change would increase the Trust Funds' buildup and extend the
program's solvency by more than 40 years.  However, at the end of the
75-year period, the Trust Funds would be expected to contain only
about 1 year's worth of benefits.\59 The estimated impact of a
2.19-percent increase in the payroll tax rate on the Trust Funds is
compared with the expected impact on the Trust Funds under the
current payroll tax rate in figure 3.1. 

   Figure 3.1:  OASDI Trust Funds
   Assets-to-Expected-Expenditure
   Ratios Under Current Law and
   With a 2.19-Percentage-Point
   Increase in the Payroll Tax
   Rate Beginning in 1998,
   2000-2075

   (See figure in printed
   edition.)

Source:  SSA data. 

This change would result in higher excess program revenues in the
near term and a maximum Trust Funds balance-to-expected-expenditure
ratio that would almost double from about 3.2 under the current law
to 6.35.  But this higher Trust Funds-to-expenditure ratio would
present a formidable challenge to future Congresses when they needed
to redeem these assets. 

Increasing the program's excess revenues and, thus, the amount of
Treasury securities held by the Trust Funds could exacerbate the
concerns that are voiced today about whether the monies in the Trust
Funds are really saved.  The Treasury uses the cash received from
issuing securities to the Trust Funds to finance other government
activities, thereby reducing the Treasury's need to borrow from the
public.  Some are concerned that this action both masks the size of
the deficit in the non-Social Security component of the federal
budget and allows the Congress to spend these Social Security
revenues on other programs in the short term without addressing the
long-term consequences of this action.  Under these conditions, the
improvement in Social Security financing would not contribute to
increased national saving.  It would only allow the Trust Funds to
build up more claims against the Treasury without enhancing the
nation's future ability to meet these increased claims. 

One way that a buildup of Social Security's excess revenues could
contribute to national saving would be to use these revenues to buy
down the nonfederally held portion of the gross debt (the debt held
by the public).  This action would not only free up resources and
allow them to be used more productively in the private sector of the
economy but also reduce the size of future cash interest payments the
government would otherwise have to pay.\60

The resulting enhanced economic growth could increase the size of the
future economy and make the government's efforts to collect taxes or
borrow to fund future Social Security benefits easier than if the
economy had not grown.  However, if, after a number of years, Social
Security's excess revenues were more than sufficient to pay off the
nonfederally held portion of the national debt, then additional
productive means of investing these excess revenues would have to be
identified. 

Advance funding Social Security through increasing purchases of
Treasury securities would allow the current, familiar benefit
structure to be maintained.  Benefits could still be determined using
the progressive benefit formula, which provides relatively higher
benefits to those with low average lifetime earnings than to those
with high average lifetime earnings.  The protections beneficiary
families now experience through disability, dependent, and survivor
benefits could also be retained.  Thus, the adequacy focus of the
current program could be maintained.  However, if benefit cuts were a
part of the reform package, the adequacy goal of the program could be
weakened. 


--------------------
\59 The 2.19-percentage-point payroll tax increase is projected to
keep the program solvent for the next 75 years under the Trustees'
intermediate assumptions and, thus, is sufficient to only partially
advance fund the program.  Fully funding the program would require
additional revenues or expenditure cuts. 

\60 Currently, $1 of every $7 of government expenditure is used to
pay the interest due on nonfederally held Treasury securities. 


      ADVANCE FUNDING THROUGH
      GOVERNMENT INVESTMENTS
-------------------------------------------------------- Chapter 3:3.2

Additional excess revenues created by financing reform could also be
invested by the federal government in the private equities market.\61
Such a move would have two distinct advantages over using these
excess revenues to purchase Treasury securities.  First, insofar as
Social Security's excess cash revenues were invested in the private
equities market, they would not be available to the federal
government for other expenditures.  Second, these investments could
improve the rate of return the Trust Funds earn because, over the
long term, investments in equities have historically outperformed
investments in Treasury securities. 

However, such investments, while offering the opportunity for greater
returns, also carry higher risks.  For example, equity investments
could expose the federal government to the risk associated with asset
loss should there be a general market downturn.  Should the Trust
Funds' equities need to be quickly liquidated to pay benefits, there
is no guarantee of the prices they would bring.  In contrast, Trust
Funds Treasury securities can be readily liquidated, should the need
arise, with no uncertainty about their value. 

From a federal budget standpoint, investing Trust Funds in the
private sector would increase the federal deficit (or reduce the
surplus), because the purchase of equities would be counted as an
outlay under current budget rules; therefore, the funds used to
purchase these equities would no longer be available to the rest of
the federal government.  If the deficit in the non-Social Security
portion of the federal budget was not otherwise eliminated, the
government would need to borrow an additional sum, up to the amount
of the program's excess revenues, from the public to pay for all its
then-current expenditures.  However, the increase in the federal
deficit that would result from borrowing additional monies from the
public would not increase the federal government's debt.  The
Treasury securities would be held by the public rather than the Trust
Funds. 

Equity investing by itself would not change the impact of federal
finances on national saving if the equity purchases were offset by an
equivalent issue of Treasury securities to the public.  In the short
term, such an asset shuffle could result in higher equity prices and
higher interest rates.  Even with higher equity prices, however, the
returns to equities would generally be expected to remain above the
rates of return from investing in Treasury securities.  The increase
in interest rates would raise interest income from new Treasury
securities held in the Trust Funds, but it would also raise future
interest expenditures for the non-Social Security component of the
federal government.\62

Equity investing would necessarily result in additional
administrative costs for handling the investments:  costs for hiring
and training a staff to carry out the daily operations of the
organization that oversees these investments, hiring a board and
financial advisers to determine how to invest the Trust Funds, hiring
fund managers to be responsible for actually investing the funds, and
hiring and training staff to carry out certain oversight
responsibilities.  However, the increase in the government's costs
could be manageable because the majority of the operating and
administrative needs of such a modified Social Security program are
already in place. 

Other concerns about government investing in the equities market are
that (1) the funds might not be invested with the goal of minimizing
risks and maximizing returns; (2) the government might be tempted to
steer these investments for politically motivated purposes, such as
aiding financially troubled companies or industries or achieving
socially desirable purposes; and (3) even if the government did not
select an equity portfolio on the basis of political or nonfinancial
objectives, the government might be able to affect corporate
management decisions by exercising its stock voting rights.  To
minimize the first and second concerns and to control transactions
costs, the government could direct its fund managers to select
equities using a broad-based market equity index.  However, the third
concern would remain unless the government either assigned its stock
voting rights to its fund managers or forbade itself from exercising
these rights.  In this latter case, the power of the voting rights
held by the remaining large stockholder groups would be enhanced. 


--------------------
\61 See GAO/AIMD/HEHS-98-74, Apr.  22, 1998. 

\62 However, if policymakers reacted to an apparently higher unified
deficit by cutting spending or raising taxes, the resulting fiscal
improvement could contribute to higher national saving. 


FEATURES AND IMPLICATIONS OF
INDIVIDUAL ACCOUNT PLANS
============================================================ Chapter 4

Most proposals to restore long-term solvency to Social Security
include the creation of a system of individual accounts.  Some
proposals have the government managing the accounts, but others leave
it largely to the individual to make the investment decisions.  The
key question raised by these proposals is how well individuals and
households might do if part of their retirement income that now comes
from Social Security depended on the performance of their individual
accounts.  Such a movement to individual accounts involves a
trade-off between higher returns and higher risks.  Historically,
stocks and bonds have yielded higher returns than the implicit return
that current workers can expect from Social Security.  Nevertheless,
consideration should be given to the added risks associated with
individual accounts.  The Congress would need to decide how the
social adequacy goal would continue to be met under such a system and
determine how the social insurance elements of the current program,
such as disability and survivor benefits, would be provided. 

Implementing individual accounts raises other issues as well.  Making
the transition to advance funded, individual accounts would require
some to "pay twice"--once for current beneficiaries' retirement
benefits, and once for their own.  In addition, major issues, such as
whether beneficiaries would be required to annuitize their accounts
and what changes would be necessary for administering the program,
would need to be addressed.  These issues would need attention
regardless of whether the accounts were managed by individuals or by
the government. 


   INDIVIDUAL ACCOUNTS INVESTED IN
   STOCKS AND BONDS WOULD LIKELY
   GENERATE A BETTER RATE OF
   RETURN, ALBEIT AT SOME
   INCREASED RISK
---------------------------------------------------------- Chapter 4:1

Individual account systems generally aim to add to the retirement
income provided by Social Security.  Proponents of individual
accounts argue that the returns to payroll taxes have fallen and will
continue to do so.  Returns in the early years of the program were
high because, for adequacy reasons, the benefits received far
exceeded what could be justified given the contributions the earliest
retirees made to Social Security while they worked.  As the program
matured and workers spent increasing time in the covered workforce,
the high initial benefit subsidies declined as did the implicit rate
of return on contributions.  At the same time, because average real
returns to stocks and bonds are higher than the return from Social
Security, individuals have the potential to be better off if their
contributions to Social Security are invested in individual accounts. 
There have been a number of studies aimed at demonstrating the
advantages of individual account proposals.  The Advisory Council
presented, for various individual and household configurations,
estimates of the returns on contributions for its three proposals. 
In general, the estimates suggest that the PSA plan, which most
closely represents the annuity-welfare concept, might provide
superior retired worker benefits for many individuals.  (See app. 
II.)

A primary concern in moving to individual account plans is the
increased risk to the security of retirement income.\63 Historically,
Social Security has offered near certainty regarding benefit receipt. 
The uncertainty that can surround the amount and lifelong receipt of
nonannuity, privately provided retirement income is, in fact, one of
the major rationales for public provision of retirement income. 
Individual accounts introduce elements of market risk and other risks
currently borne by the federal government. 

Markets are volatile, and, while they can generally be expected to
provide better returns than bonds over the long term, they have had
periods of substantial downturn that lasted for some years.  Pensions
hold a majority of their assets in stocks, and even individuals hold
substantial amounts of their savings in accounts that are invested in
stock market equities--such as IRAs, mutual funds, voluntary 401(k)
plans, and so on.  Thus, if a significant portion of Social Security
income also depended on the market's performance, a broad and
long-lasting market downturn could have a negative impact on a large
portion of retirement income. 

Even if the market experienced no dramatic or long-lasting downturns,
the normal market cycles will create "winners" and "losers,"
depending on when and how workers invest their "Social Security"
assets in the market and when they liquidate their holdings. 
Individuals with similar work histories could receive substantially
different benefits.  As long as workers are aware of and accept this
risk, there will probably not be calls to fix the "unfair benefit
outcomes." However, if such large differences in outcomes become
commonplace, many participants could become dissatisfied with the
program. 


--------------------
\63 This concern raises the questions of whether government should
guarantee a minimum benefit from individual account accumulations and
whether such a guarantee would encourage greater risk-taking. 


   INDIVIDUAL ACCOUNT PROPOSALS
   RAISE QUESTIONS ABOUT HOW TO
   MAINTAIN THE GOAL OF INCOME
   ADEQUACY AND HOW TO PROVIDE
   ANCILLARY BENEFITS
---------------------------------------------------------- Chapter 4:2

If individual account proposals were implemented, the question of how
to preserve the goal of income adequacy would need to be answered. 
Many proposals based on the annuity-welfare model seek to minimize
the redistributive aspect of Social Security and focus on providing a
basic income floor or minimum benefit.  Thus, one issue involves
determining the appropriate level of "social adequacy" for the social
insurance system.  Proposals for individual accounts focus primarily
on the retirement benefits portion of the program, but the current
Social Security system also includes ancillary benefits that may not
be easily obtained or duplicated in the private market.  It is
important, then, to consider how creating individual accounts would
affect these other elements of the benefit package--in particular,
disability benefits and benefits for dependents (spouses, children,
and survivors).  Social Security also has important interactions with
other retirement income sources:  pensions, personal savings, and
earnings play substantial roles in determining the level of income
that individuals and households will have in retirement.\64


--------------------
\64 Retirement Income:  Implications of Demographic Trends for Social
Security and Pension Reform (GAO/HEHS-97-81, July 11, 1997). 


      LEVEL OF SOCIAL ADEQUACY
      WOULD NEED TO BE DETERMINED
-------------------------------------------------------- Chapter 4:2.1

The annuity-welfare concept of social insurance leads to questioning
the appropriate role for the government in providing retirement
income.  The emphasis under this approach is on separating the
annuity part of the program, in which benefits are directly linked to
contributions, from the redistributive or welfare part of the
program, in which the benefits of the less fortunate are raised to a
"more adequate" level.  The existing Social Security program embodies
the idea that these decisions should be made jointly in the context
of a universal program of retirement income (social) insurance. 
Ascertaining the real difference between these opposing conceptions
of social insurance may be difficult, but a key part of the
difference relates to the "process" for deciding the relative
importance given to the components of redistribution and contributory
insurance.  While under each of these concepts the political process
would sort out the relative importance of the components, the main
thrust of the annuity-welfare view is to make the redistributions
more explicit--that is, more visible to program participants, voters,
and political decisionmakers--than is the case under the existing
structure. 

While discussions of social adequacy often address the poverty issue,
it does not necessarily follow that these discussions determine the
level of support that should be provided.  The provision for
retirement income spans an individual's entire lifetime, and it is
particularly important to consider various incentive and efficiency
effects of any social adequacy level that is provided.  The obvious
consideration is that if the safety net benefit level is set too
high, then work and savings disincentives could arise, and some
workers could be encouraged to "free ride." But if the level is set
too low, then some individuals could live out their retirement years
in extreme poverty. 

Incentive effects are a major rationale for the contributory aspect
of Social Security.  An individual's benefit must be "earned" by
making contributions.  In considering the social adequacy level in
the context of the program structure, several ideas have been
advanced.  Some favor a better targeting of the redistributive
component through means testing.  One idea behind proposals for means
testing is that the existing design of Social Security provides
benefits to all income groups, and often the redistributive aspect is
not well focused on the needy.  Advocates for the existing structure
of Social Security point out that the program is, in fact, designed
to avoid the pitfalls of means testing, which create both stigma and
work and savings disincentives for low earners. 

Some proponents of the annuity-welfare concept have raised the idea
of a flat benefit, or "demogrant." Since everyone would receive the
demogrant, many of the work disincentive effects would be minimized,
particularly if the demogrant was not set at too high a level.  With
the demogrant, the redistribution would be addressed in a way that
was visible politically.  It could even be financed with general
revenues.  This type of financing would be consistent with
strengthening the linkage between contributions and benefits in the
annuity part of the program.  Also, the demogrant could avoid the
stigma that means testing would introduce, since it would go to
everyone.  The current program and the demogrant approach are similar
in their effects, with the major difference being how the decision
about the social adequacy level is arrived at in the political
process. 

The fundamental issue for social insurance, then, is what level of
social support society wants to provide to its elderly.  Even
providing a level of support far below the poverty level is likely to
carry substantial cost.\65 Another important aspect is the notion of
minimizing the stigma that is usually associated with the receipt of
transfers (that is, "welfare").\66 Also, an important consideration
that is often overlooked is the role of the SSI program.  Depending
on the design of reforms, the existing SSI program might be expanded
to serve more people.  Proposals could be devised to include a
demogrant, which might absorb the role played by SSI. 


--------------------
\65 Thompson, "The Social Security Reform Debate," 1983. 

\66 Thompson, "The Social Security Reform Debate," 1983. 


      TREATMENT OF ANCILLARY
      BENEFITS WOULD NEED TO BE
      CONSIDERED
-------------------------------------------------------- Chapter 4:2.2

Disability and dependents' benefits are often not included in the
discussion of individual accounts because it is, in principle,
possible to separate them from retirement benefits.  Retirement,
disability, and auxiliary benefits, respectively, account for
approximately 68.1 percent, 10.5 percent, and 21.4 percent of all
benefits paid.  Separating the "price components" of the various
parts of Social Security would mean that disability and auxiliary
benefits could be maintained in the presence of individual accounts
for a part of the retirement benefits portion of Social Security. 
However, it would also imply that the administrative apparatus of
Social Security, including the reporting of earnings by employers,
would have to be retained.\67

There is also the question of whether the disability and dependent
portions of OASDI could be better provided through private markets. 
Disability insurance is provided by private insurers and through
group insurance arrangements financed by employers.  However, a key
feature of the benefits provided by Social Security is that they are
universal--that is, they are available to everyone regardless of age
or occupation.  This would generally not be the case with individual
disability insurance policies, and even the current employer-provided
group arrangements might be subject to certain restrictions. 

A voluntary private disability insurance program, combined with
insurers who might want to avoid the problem of adverse selection,
suggests that comprehensive disability protection would be available
to some only at a high price.  At the same time, it is difficult to
assess how private markets might perform in providing various
insurance substitutes given that Social Security today plays such a
major role in providing such benefits.  If the private sector were to
play a larger role in providing disability benefits, it might be
necessary to enact laws that require private providers to offer
certain benefits or features.  An example of this is the recent
preexisting condition legislation in the health care area. 

While disability benefits would largely be unaffected under the
Advisory Council's MB proposal, the IA and PSA proposals reduce these
benefits.  Under the IA proposal, the essential structure of DI would
remain intact, but the benefits for DI beneficiaries would be reduced
because individual investment benefits needed to offset the reduction
in program benefits would not be available until age 62.  DI benefits
would also be heavily affected under the PSA proposal and could be
reduced by as much as 30 percent from today's DI benefit levels. 
These DI beneficiaries would not have access to their individual
accounts until age 65, the proposed early retirement age under the
PSA proposal. 

With respect to dependents' benefits, individual account proposals
imply reduced spousal benefits.  With individual accounts, much of a
person's retirement benefit would depend on how well his or her own
investments performed.  Thus, unless spouses had their own individual
accounts, they could be worse off than under current law.  Those
spouses who would not accumulate substantial assets in individual
accounts might be eligible for a reduced spousal benefit or a
demogrant.  But it is also important to realize that the role of
spousal benefits within the existing program structure may be
declining in importance because of changes in women's labor force
participation.\68

Survivor benefits would also be affected under proposals to create
individual accounts.  Currently, when a retired worker dies, the
dependent spouse is eligible for a survivor benefit if it is higher
than his or her own retired worker benefit.  With individual
accounts, the survivor could inherit the asset accumulation in the
retired worker's individual investment account.  These assets could
supplement any other Social Security benefits the survivor might
receive.  However, the deceased worker could also bequeath these
assets to others.  Even if these assets were left to the surviving
spouse, the survivor could have a lower or higher benefit amount than
under current law, depending on the survivor's individual
circumstances.  The IA proposal would lower spousal benefits in order
to increase survivor protection for two-earner couples. 


--------------------
\67 Administrative expenses for DI are higher than those for OASI. 
In 1996, OASI expenses were 0.6 percent of benefit payments, and DI
expenses were 2.6 percent. 

\68 For further discussion, see Social Security:  Issues Involving
Benefit Equity for Working Women (GAO/HEHS-96-55, Apr.  10, 1996);
Social Security Reform:  Implications for the Financial Well-Being of
Women (GAO/T-HEHS-97-112, Apr.  10, 1997); and Social Security
Reform:  Implications for Women's Retirement Income (GAO/HEHS-98-42,
Dec.  31, 1997). 


      OTHER RETIREMENT INCOME
      SOURCES COULD BE AFFECTED
-------------------------------------------------------- Chapter 4:2.3

The post-WWII era has seen a general rise in living standards and a
substantial evolution in the retirement income system.  Social
Security has provided the foundation for the retirement living
standard of the population and has largely fulfilled its original
intent in alleviating elderly poverty.  But private pension coverage
has also increased and now provides a substantial portion of
retirement income for many of today's elderly.  Increases in home
ownership and personal savings have meant greater wealth in
retirement for many households.  Incorporating individual account
features in Social Security would have important implications for the
entire framework that provides retirement income to the elderly.  The
debate over how to resolve Social Security's financing requires
recognition of the broader changes that may take place in response to
any actions taken.  Here we suggest but a few of the issues that
might arise. 

The existing private pension system has traditionally provided a
voluntary, private source of retirement income.  Creating individual
accounts is essentially aimed at further expanding the role of
private institutions in providing retirement income.  If this role
were expanded, it is hard to imagine that the existing private
pension system would not be affected.  One obvious change would be in
private pension plans' "integration" with Social Security. 
Currently, some employers agree to provide a benefit that is adjusted
by any amount received through Social Security.  If Social Security
benefits were reduced, then private employers with integrated plans
might experience an increase in their pension costs that could prompt
them to redesign their plans.\69

It is also unclear how workers' personal savings behavior might be
influenced by a new system involving individual accounts.  Economists
have long debated various theories of savings behavior in the context
of the effects of Social Security.  This debate has largely focused
on the theory that the promise of Social Security benefits would be
viewed by individuals as a form of "social security wealth" that
could result in lower saving.\70 This fundamental debate about the
behavioral effects of social insurance on personal saving is ongoing
although, on balance, the prevalent view is that funded social
insurance is more likely to be consistent with higher saving.  Much
recent research focuses on the effects of individual savings plans,
such as IRAs and, particularly, 401(k) plans, and this research may
yield useful insight into the possible effects of introducing
individual accounts.\71

Finally, individual accounts could affect workers' decisions on when
to retire.  A number of factors affect an individual's decision to
retire.  If the age at which a worker becomes eligible for full
benefits is further increased, individuals might stay in the
workforce longer.  If individual accounts fulfill their promise of
higher levels of retirement income, then workers may retire early
despite the increase in the retirement age. 


--------------------
\69 The Tax Reform Act of 1986 changed pension plan integration
rules.  See Geoffrey Kollmann, Ray Schmitt, and Michelle Harman,
Effect of Pension Integration Rules on Retirement Benefits, report to
the Congress (94-974 EPW) (Washington, D.C.:  Congressional Research
Service, Dec.  6, 1994). 

\70 Martin Feldstein, "Social Security, Induced Retirement and
Aggregate Capital Accumulation," Journal of Political Economy, Vol. 
82, No.  5 (Sept./Oct.  1974), pp.  905-26. 

\71 For example, see 1994-1996 Advisory Council Report, 1997, Vol. 
II, pp.  41-48. 


   MAKING THE TRANSITION TO
   GREATER ADVANCE FUNDING WOULD
   POSE FUNDING CHALLENGES
---------------------------------------------------------- Chapter 4:3

Because by definition individual accounts are advance funded, a
significant shift toward such a system would raise transition
questions.  The practical problem that would occur is that, because
most of the benefit obligations of current retirees and workers are
unfunded under pay-as-you-go, any diversion of current workers' taxes
to fund their own benefits would leave less with which to pay current
and accrued retirement benefits.  As a result, current workers would
need to be asked to "pay twice"--once for the accrued benefits of
current and future retirees and again for their own retirement
benefits.\72 In subsequent generations, workers would have to pay to
fund only their own benefits.  Although advance funding is generally
associated with individual accounts, advance funding could be
introduced without them.  The system is already partially advance
funded, and government's investing a part of the Trust Funds in the
stock market would represent an increase in advance funding. 

If the current program was terminated and a new fully advance funded
one that included all current and future workers was started, the
amount necessary to pay the accrued benefit obligations under the
current Social Security system would be about $9 trillion.  In
principle, transition costs do not pose any greater cost than already
exists under Social Security.  Concerns about transition costs arise
primarily because of the timing of paying for benefit entitlements. 
Under pay-as-you-go, the costs of paying accrued benefits occur in
the future and represent an unfunded promise--a type of implicit
debt.  In moving to an advance funded system, the future benefits,
which would need to be paid for eventually, would be recognized
today.  Paying these benefits could involve significant payroll tax
increases or a sizable increase in government debt. 

Making a transition to an advance funded system would also present
political difficulties.  Reneging on benefit obligations or requiring
current workers to "pay twice" could significantly disadvantage many
individuals, and showing that a funded system was superior to
pay-as-you-go would do little to ease the pain.  It has been
suggested that the current pay-as-you-go program has created a
"lock-in" effect that was largely intended when the program was
designed.\73 That is, transition costs could prevent individuals from
supporting a potentially superior alternative that offers higher
benefits or returns until the benefits under the existing system
become considerably worse than they would be under the
alternative.\74 Voters could continue to support the current program
structure, which could increase the political costs of making a
transition, and the existence of strong interest groups could add
further to the political costs and difficulty of making such a
transition. 

Nonetheless, the problem of "paying twice" could be mitigated in
several ways.  One way would be to reduce accrued benefits for the
current retired generation.  The IA plan attempts to address the
long-term financing problem, for the most part, by reducing future
benefits, including those already accrued, for current workers. 
Another way to mitigate the economic and political costs of
transition for a particular generation of workers would be to push
the costs of accrued benefits into the future.  This would be similar
to what a pay-as-you-go system does.  Two ways to avoid putting
responsibility for all of the transition costs on one generation of
workers would be to levy special taxes on the entire population or to
finance the transition through borrowing.\75

Under the first approach, the accrued obligations of the Social
Security system that came due, and that were in excess of the
financing available, could be financed by levying an array of taxes. 
Each type of tax that could be used would have different impacts on
individuals and the economy.  A payroll tax would be consistent with
the current financing of the program, but because of its regressive
impact on lower earners, it might not be seen as desirable.  Income
taxes could reduce the impact on lower-earning workers and families
but could have undesirable effects, such as increasing taxes on
savings.  Levying taxes to pay the accrued benefits, moreover, could
still leave a substantial burden on the current generation of
workers. 

Another way of financing the accrued obligations in transition to an
advance funded system would be to use government borrowing.  The
government would issue bonds to finance the payment of benefits, and
the bonds would be paid off in the future, which would spread the
cost to future generations.  Because the interest and principle on
the bonds would be paid with future taxes, the use of bond financing
would be a more effective way of spreading the cost of the transition
than taxation.  There are a number of ways to implement bond
financing.  One of the earliest ideas was to issue "recognition
bonds," which could be issued to individuals in recognition of the
government's intention to honor its benefit obligations.\76
Individuals could redeem the bonds at retirement to provide a
retirement benefit.\77


--------------------
\72 The original decision to finance the Social Security system on a
pay-as-you-go basis was the result of the need to get funds to the
elderly and needy quickly, the need to spur the economy, and the fear
of a large reserve in government hands.  This decision led to the
first cohorts of retirees receiving transfers considerably in excess
of their contributions.  Subsequent cohorts, however, have
increasingly "paid" for their benefits but must still depend on
future generations of taxpayers to fund them. 

\73 David V.  Bryce and Robert B.  Friedland, "Economic Security:  An
Overview of Social Security," in EBRI, Assessing Social Security
Reform Alternatives, 1997, p.  83, note President Franklin D. 
Roosevelt's famous explanation for his insistence on contributory
taxes:  "We put those payroll contributions there so as to give
contributors a legal, moral, and political right to collect their
pension and unemployment benefits.  With those taxes in there, no
damn politician can ever scrap my social security program."

\74 James M.  Buchanan, "Social Security Survival:  A Public Choice
Perspective," Cato Journal, Vol.  3, No.  2 (fall 1983), pp.  339-53. 

\75 Note that the three Advisory Council proposals all include an
element of partial advance funding and an additional tax to meet
these costs. 

\76 This option was used in privatizing the Chilean system. 

\77 One important technical issue could arise in using recognition
bonds to finance the transition to a new system.  While the concept
of an accrued benefit is common to private pension plans, the Social
Security system is not designed in a way that workers can easily be
"cashed out" when they leave the system.  Thus, imputing an accrued
Social Security benefit to individuals could be a subject for
technical debate. 


   INDIVIDUAL ACCOUNT PROPOSALS
   WOULD NEED TO ADDRESS
   ANNUITIZATION ISSUES
---------------------------------------------------------- Chapter 4:4

Social Security is structured in a way that, upon reaching
eligibility, workers receive a monthly benefit--that is, an
annuity--for the remainder of their lives.  With individual account
plans, the worker might be able to choose one of several options for
receiving benefits.  Depending on the plan design, an individual's
account accumulation could be converted to an annuity, taken as a
lump sum, left in place, or used for any purpose desired.  While this
approach offers greater freedom of choice, it also raises several
concerns. 

One concern is whether the account accumulation would be intended to
constitute a source of retirement income as opposed to simply a
savings accumulation device that might not be fully used for
retirement income.  Another concern is the process of annuitization
itself.  Obtaining annuities individually or on a group basis in the
private market could be more costly than having the government
provide them.  Related to this issue is the question of whether an
individual could obtain an annuity that provided features similar to
those currently provided by Social Security. 

One of the major goals of proponents of individual account plans is
to ensure that individuals have as much freedom as possible in
choosing how to allocate their own resources.  Individual accounts
can offer a large amount of freedom and choice and, in principle,
there is no inherent reason why individuals should be required to
receive their retirement income through an annuity.  Thus, a
fundamental issue of retirement income policy is how much the
individual's choice should be restricted in order to ensure that he
or she does not become a burden on society in old age. 

The social insurance approach seeks to provide a "socially adequate"
benefit, not a minimal benefit, that protects a substantial portion
of the preretirement living standard.  This perspective suggests that
restricting an individual's choice is justified to achieve a more
socially desirable outcome.  This perspective is embodied in
proposals that restrict the individual to investing through the
government and require annuitization of the account accumulations so
that there is more certainty that an adequate retirement living
standard will be achieved. 

An additional complication arises when individual borrowing
provisions are considered as a feature of the individual account
plans.  This is an important issue with private pensions,
particularly with 401(k) plans.  Restricting borrowing from the
accounts helps ensure that the funds constitute retirement income. 
Allowing borrowing provides more freedom of choice but does not
ensure that the accounts will be used for retirement.  In this case,
the accounts of many might represent more of a tax-deferred saving
vehicle than a retirement saving vehicle.\78

A second major concern is whether those individuals who chose to
convert their account accumulations to retirement income by
purchasing an annuity would actually be able to do so.  One of the
major advantages of Social Security is the provision of a lifetime
annuity.  Private pension plans also provide an advantage because
they are able to offer annuities through group arrangements.  But
those who have an individual account plan might have to obtain
annuities in the individual annuity market.  While individual
annuities are available, they can be costly, especially relative to
annuities provided through Social Security.  This issue is
compounded, since private annuities might not generally contain the
same features as a Social Security annuity--features such as
dependents' benefits, inflation protection, and the use of unisex
life tables (that is, the same assumed mortality rates for both men
and women). 

It is difficult to predict, however, what would occur if an
individual account system were put in place.  Some retirees would
prefer individual annuities over other payment options, and
competition among financial institutions to provide such annuities
would ensue.  This could be a positive force in driving down the cost
of annuities, but the possibility remains that firms would compete
for what they perceived to be the best risks, which in this case
could be those who are likely to have shorter lifetimes.  Competition
directed at avoiding adverse selection problems could result in
market imperfections wherein certain individuals might not be able to
obtain annuities at reasonable cost, and this might lead to calls for
legislation or regulation restricting the ability of financial
institutions to deny individuals an annuity contract.  The government
would potentially play some role in either (1) ensuring that
insurance markets worked efficiently or (2) continuing to provide
annuities when private markets failed to do so. 


--------------------
\78 401(k) Pension Plans:  Loan Provisions Enhance Participation but
May Affect Income Security for Some (GAO/HEHS-98-5, Oct.  1, 1997). 


   INDIVIDUAL ACCOUNT PLANS WOULD
   RAISE OTHER IMPLEMENTATION AND
   ADMINISTRATION ISSUES
---------------------------------------------------------- Chapter 4:5

Implementing individual account plans would raise a number of
implementation issues regarding the cost of managing accounts and
investments and how to manage financial flows and protect investors. 


      COSTS OF MANAGING ACCOUNTS
      AND INVESTMENTS
-------------------------------------------------------- Chapter 4:5.1

Individual account plans would require creating financial accounts
for each worker.  This would be a huge undertaking, although arguably
it should be feasible since existing financial markets and SSA are
already able to handle large numbers of individuals and transactions. 
Nevertheless, depending on the design of the program--whether the
accounts were managed by individuals or were managed for them by the
government--the scale of new resources required could be large and
could imply a significant expansion of the administrative structure
of either the current program or investment firms (and employers) in
the private sector. 

There are significant differences in the relative costs of publicly
managed social insurance systems and privately managed individual
account arrangements.  Social Security is a large, centrally managed
public system.  The costs of Social Security were high relative to
benefits paid during the early years of the program.\79 As benefit
payments have grown, the administrative costs of OASI as a percentage
of expenditures has fallen to the rather low 0.6 percent of benefit
payments experienced today. 

Administrative costs for individual account plans would depend
greatly on the specific design.  There would be initial costs in
setting up the necessary systems.  But the experience with Social
Security suggests that while moving to an individual account system
might involve large start-up costs, the ongoing costs of the system
might fall as a percentage of assets as the accounts grew over time. 
The ongoing costs of an individual account system would involve two
major elements:  the cost of managing and maintaining the accounts
(that is, record keeping costs) and the costs associated with
investing funds. 

Concerns have been raised about the amount of funds that would be
held in the individual accounts and how transaction and
administrative costs would affect them.\80 It has been noted that the
account balances for many individuals could be quite small and that
there could be a large number of rather small transactions.  These
factors could make it costly for private institutions to maintain the
accounts.  If the administrative and transaction costs were charged
to individual account holders, they could greatly reduce, or even
eliminate, any gains small accounts might otherwise receive.  This
could be one area in which a government-managed individual account
plan might have an advantage.  The government would be able to
collect deposits through the existing payroll tax collection system
and perhaps reduce transaction costs.\81 However, it is not certain
that the size and number of individual accounts would be a
significant problem for the private sector, which already manages
401(k) plans that are similar to the individual account plans
proposed. 

Another issue concerns the specification of investment alternatives
for the accounts.  Under a privately managed system of individual
accounts, individuals or employers might contract directly with
financial institutions.  This could mean a wide array of investment
choices for individuals and, at the same time, a wide variation in
potential financial outcomes.  Some individuals might not be familiar
with basic investing strategies, and they would have to sift through
a potentially large amount of information that financial institutions
sent them as they competed for clients.  Expanding investment
education for workers has been suggested as one way to address this
concern.  However, who would provide this education is an open
question. 

As already noted, financial institutions would incur costs in
managing the accounts and would charge fees as part of making
transactions.  Data suggest that administrative expenses are higher
in mutual funds that are more actively managed, whereas funds that
are more passively managed--such as index funds, which tend to make
fewer transactions--have substantially lower costs.\82 The extent to
which an individual account system would result in large transaction
fees, as has been the experience in the early phases of the Chilean
privatized system, is unclear.\83

In estimating outcomes for the Advisory Council proposals,
assumptions were made about relative administrative costs.  Both the
IA and PSA proposals assume that at least a part of the activities of
the current Social Security program would continue after the new
advanced funding mechanisms were in place.  Thus, much of the costs
of the current program would be retained.  The proposal that
recommended the larger individual accounts was estimated to have new
administrative costs that would be considerably higher than more
limited individual accounts.\84


--------------------
\79 Olivia S.  Mitchell, "Administrative Costs in Public and Private
Retirement Systems," working paper no.  5734 (Cambridge, Mass.: 
National Bureau of Economic Research, Aug.  1996). 

\80 Robert J.  Myers, "Social Security:  Myths and Realities," in
EBRI, Assessing Social Security Reform Alternatives, 1997. 

\81 Mitchell, in "Administrative Costs in Public and Private
Retirement Systems," 1996, notes that some of the costs of collecting
payroll taxes and subsequent monitoring and enforcement are borne by
the IRS and that these costs are not included in the operations of
SSA. 

\82 Mitchell, "Administrative Costs in Public and Private Retirement
Systems," 1996, p.  20. 

\83 Robert J.  Myers, "Chile's Social Security Reform After Ten
Years," Benefits Quarterly, Vol.  8, No.  3 (third quarter 1992), p. 
56. 

\84 The proposal that would result in larger individual accounts is
the PSA option, which was assumed to have an administrative expense
factor of 1.0 percent for the privatized portion of the proposal
alone.  Expenses that would continue to be borne by the remaining
components of SSA are not included in this figure.  Thus, if the
expected gross yield on an individual account was 7.0 percent, the
expected real yield would be 6.0 percent.  The IA option would be
managed by the government through SSA.  The administrative expense
factor assumed for this option was 0.105 percent.  Thus, an expected
gross yield of 7.0 percent on individual account balances would
result in an expected 6.895 percent real yield. 


      MANAGING FINANCIAL FLOWS AND
      PROTECTING INVESTORS
-------------------------------------------------------- Chapter 4:5.2

Implementing private account systems would also raise questions about
the management of financial flows and how the individual investor
might be protected, both of which relate to the role of monitoring
and regulation of private account systems. 

Individual accounts could be maintained under the auspices of the
government, and the financial flows would not need to change
significantly.  Employers could deposit the required contributions
directly with the Treasury, and SSA could make the appropriate
distributions to the individual accounts.  SSA and the Treasury would
have to arrange procedures for allocating the funds to accounts. 

If the system was run much like the government's Thrift Savings Plan
(TSP) for federal workers, the government would contract with a
financial institution to manage several funds.  However, given the
size of the contributions involved, it would probably not be wise or
feasible to have only a few institutions manage these assets.  Thus,
it would probably be the case that the range of institutions and
investments would have to be expanded.  As the number of financial
institutions participating expanded, so would the administrative
complexity.  Arguably, a system of direct deposits from employers to
financial institutions might be feasible and efficient.  However, it
is unclear whether a private system might still be more costly than
funneling the funds through SSA, in part because a centralized
operation might more efficiently handle such functions. 

Individual account systems could require substantial monitoring, as
would any system of financial transactions.  For example,
transferring funds between employers and financial firms creates
opportunities for fraud.  Private pension plans are covered under the
Employee Retirement Income Security Act of 1974 (ERISA), which
provides a broad framework of pension law that includes codification
of fiduciary responsibilities for handling pension assets, disclosure
to plan participants, and other provisions aimed at protecting
workers' benefit rights.  It is not currently clear whether
individual account plans would need ERISA-like provisions, although
many of ERISA's provisions might prove useful in protecting
individual account assets. 

Certain monitoring and regulatory concerns--such as those regarding
the provision of investment advice and financial education for
investors--would need to be addressed.  Under a government-managed
individual account plan, handling the accounts through the Treasury
and SSA might require few additions to the current regulatory
apparatus.  However, under individually managed individual account
plans, new or expanded monitoring and regulatory functions might be
necessary.  These functions would affect the cost of implementing the
new individual account systems.  Evidence suggests that regulatory
requirements have added significantly to the cost of private
pensions.\85 \86

While administrative issues are not necessarily decisive criteria for
determining whether a new Social Security system should be
implemented, they do represent an important consideration as reforms
are debated.  Some evidence suggests advantages from a centralized
approach based on the Social Security model, but other evidence
suggests that relying on individuals and their brokers has advantages
in terms of efficiency and service to the participant.\87 This aspect
of the reform debate requires careful scrutiny and additional
attention. 


--------------------
\85 See Mitchell, "Administrative Costs in Public and Private
Retirement Systems," 1996, and Committee for Economic Development,
Who Will Pay for Your Retirement?  The Looming Crisis (New York: 
Committee for Economic Development, 1995). 

\86 Since individual account systems involve stock ownership, moving
to such a system would raise the issue of corporate governance and
proxy voting.  This could be addressed, however, by allowing
investment managers to vote shares subject to certain guidelines and
criteria, as is done in the TSP. 

\87 Mitchell, "Administrative Costs in Public and Private Retirement
Systems," 1996, p.  40. 


CHANGES TO THE SYSTEM'S FINANCING
WOULD HAVE IMPORTANT IMPLICATIONS
FOR THE FEDERAL BUDGET AND THE
NATIONAL ECONOMY
============================================================ Chapter 5

Decisions to increase the advance funding of the Social Security
system, whether or not accompanied by individual accounts, could have
significant consequences for the federal budget.  Changes to the
status of the federal budget, in turn, could have implications for
the level of national saving and future economic growth.  Advance
funding could be done either through the public or private sector,
although advocates of privately held individual accounts believe that
funding through private institutions is more likely to lead to
capital formation and enhanced economic growth.  Regardless of
whether the advance funding was done through the public or private
sector, the cost of the transition to the new system would need to be
addressed, and the way the transition was accomplished could
determine the impact of the shift to advance funding on national
saving. 


   FEDERAL BUDGET POLICIES WILL
   INFLUENCE THE ECONOMIC EFFECTS
   OF REFORM
---------------------------------------------------------- Chapter 5:1

Social Security's current financing structure and the Trust Funds
have important interactions with the federal budget and government
finance.  The status of the federal budget, in turn, can affect
national saving and future economic growth, which could determine the
ability of future workers to provide for their own retirements and
for beneficiaries. 


      SOCIAL SECURITY TRUST FUNDS
      AFFECT THE FEDERAL BUDGET
-------------------------------------------------------- Chapter 5:1.1

The Social Security Trust Funds were designed to maintain a
short-term contingency reserve, not to provide advance funding for
future obligations.  Amendments to the Social Security law in 1977
and 1983 have allowed the Trust Funds to accumulate a reserve beyond
what is considered necessary to meet contingencies.  This reserve,
however, is still well below what would be needed for full advance
funding.  The Trust Funds' excess cash revenues are, by law, invested
in U.S.  Treasury securities.  In effect, these revenues are loaned
to the Treasury, reducing the Treasury's need to borrow from other
sources to finance non-Social Security federal spending. 

The Social Security cash surplus is expected to remain at about $50
billion annually for another decade, after which the surpluses will
get smaller.  Without changes to current policy, the program's cash
surpluses are expected to disappear in 2013.  To cover the subsequent
annual cash shortfall, the Trust Funds will begin drawing on the
Treasury, first relying on its interest income and, eventually, on
its assets.  This will have a direct and increasingly negative impact
on the federal budget.  By around 2032, the Trust Funds will be
effectively exhausted--at that time, without government action,
program revenues will pay only about 75 percent of total benefits. 

While the Trust Funds' Treasury securities are assets of the Social
Security program, they are also liabilities for the rest of the
federal government that, when redeemed, will have to be financed by
raising taxes, borrowing from the public, or reducing other federal
expenditures.  Thus, not only will the government no longer have
access to Social Security's surplus, but the need to cover the
system's cash shortfall could force difficult budget and tax
decisions in the non-Social Security portion of the budget. 


      THE FEDERAL BUDGET
      INFLUENCES NATIONAL SAVING
-------------------------------------------------------- Chapter 5:1.2

The realization that there will be relatively fewer workers in the
future to produce the goods and services to support not only
themselves, but also a larger number of retirees, has led many to
focus on the potential contribution of Social Security financing
reform to long-term economic growth.  Future national income and
output depend on, among other things, the level of capital stock
available.  Capital accumulation, in turn, depends on national saving
that can be used for investment.  National saving is composed of
personal saving by individuals, business saving (undistributed
profits), and government saving.  When the government runs deficits,
it subtracts from national saving.  National saving rates in recent
years have been at historically low levels. 

A purely pay-as-you-go system has little, if any, direct effect on
saving.  The current Social Security system is running cash surpluses
that reduce the size of the unified budget deficit and, all else
being equal, should increase national saving.  However, to the degree
that the existence of the Social Security surpluses undermines fiscal
discipline elsewhere in the budget, the potential positive effect on
national saving is mitigated.  If the non-Social Security part of the
budget were balanced, the buildup in the Trust Funds would mean
positive government saving and could result in larger national
saving.  These resources would be available for investment and could,
presumably, enhance economic growth.  Moreover, a larger economy
could lighten the future burden of maintaining Social Security.\88
Higher rates of economic growth would mean higher real wages and
living standards, and future workers, even if they had to pay higher
payroll tax rates to maintain benefit levels, would be in a better
position to do so. 


--------------------
\88 See H.  Aaron, B.  Bosworth, and G.  Burtless, Can America Afford
to Grow Old?  Paying for Social Security (Washington, D.C.:  The
Brookings Institution, 1989) and Budget Issues:  Analysis of
Long-Term Fiscal Outlook (GAO/AIMD/OCE-98-19, Oct.  22, 1997). 


   ADVANCE FUNDING COULD FOSTER
   NATIONAL SAVING
---------------------------------------------------------- Chapter 5:2

The economic importance of advance funding is that it could foster
saving.  These savings would then be available for capital formation. 
Retirement programs, such as pensions, are essentially savings for
long-term capital formation.  Pensions transfer the portion of
current income that is not consumed today into income that will be
consumed in retirement.  In an advance funded pension arrangement,
the savings put into pension funds provide capital for business
investment, and the returns generated accrue both to the businesses
that invested the funds and to the individuals saving for their
retirement.  Thus, the return available to pension savers is related
to the real growth of the economy, and pension saving provides an
important basis for capital formation and economic growth.\89

One of the objections to pay-as-you-go financing is that it is mainly
a tax-transfer mechanism that extracts resources from current workers
and redistributes them to current retirees and has no direct impact
on saving.\90 In order for the government to save and contribute to
capital formation, it must extract resources from the economy and
either invest them productively on its own or use them in a way that
frees up other (private) resources for investment. 


--------------------
\89 The implicit return available to participants under a
pay-as-you-go financed plan is also related to the growth of the
economy (that is, the growth in the productivity of the workforce
(wages) and the growth in population), but this return may differ
from that earned by private capital. 

\90 Pay-as-you-go financing may have secondary effects on saving. 
For example, one could compare the overall savings/consumption
balance of the income that was taxed to the savings/consumption
balance of retirees' use of the income to attain a sense of net
impact on saving and the economy.  There may be a host of behavioral
reactions by individuals and households to any actions associated
with financing social insurance programs, whether they are
pay-as-you-go or advance funded.  For example, coverage under an
advance funded pension plan could cause some individuals to save less
through other saving vehicles.  And under pay-as-you-go Social
Security, it has long been theorized that individuals believe they
hold "Social Security wealth," which may cause them to save less
outside the Social Security program. 


      ADVANCE FUNDING THROUGH THE
      PUBLIC SECTOR
-------------------------------------------------------- Chapter 5:2.1

Should policymakers choose to increase advance funding through the
current Social Security program structure, the Trust Funds could
continue to invest rising surpluses in Treasury securities.  Under
such policies, the federal government could use this capital to
retire outstanding debt held by the public, thus freeing up resources
to be invested in private sector capital, or it could undertake
"public investment," such as building or maintaining infrastructures,
which could provide economic benefits that improve efficiency in
other areas of the economy.  Alternatively, Social Security Trust
Fund investment policies could be altered to permit investing surplus
funds outside the federal government, such as in the stock market. 
The various uses of surplus Social Security funds could have
different impacts on national saving. 


         RETIRING DEBT
------------------------------------------------------ Chapter 5:2.1.1

When the government runs a budget surplus, resources have been taken
out of the economy.  If these resources were used to retire
outstanding public debt, instead of to fund other government
programs, some of the resources of investors who had purchased the
debt would be freed up.  To the extent that these funds were
reinvested, the government would have increased private investment,
which, in turn, creates the potential for higher economic growth.\91
The ability of the government to retire debt would depend on
congressional spending decisions.  It is important to note that the
annual Social Security surpluses themselves represent only a small
fraction of the future unfunded promises of Social Security.  To
advance fund all these promises would require running much larger
annual Social Security surpluses. 


--------------------
\91 In testimony before the Senate Budget Committee, we noted that
extended periods of fiscal surplus could increase per-capita gross
domestic product significantly in the long term (Budget Issues: 
Long-Term Fiscal Outlook [GAO/T-AIMD/OCE-98-83, Feb.  25, 1998]). 


         INVESTING IN CAPITAL
         PROJECTS
------------------------------------------------------ Chapter 5:2.1.2

Increasing public capital investment would require budgetary actions. 
Actual investments that could contribute to economic growth would
have to be identified, and funds for them would have to be allocated
in the budget.  The traditional concern with public capital
investments is that political processes introduce considerations
other than purely economic returns into the decision-making process. 
Some believe such considerations can be used to impart a "social
return" to a particular allocation of resources, which many view as
highly desirable.  However, there is disagreement about this, and
others hold that such decisions may be less subject to the discipline
of market forces and, hence, undermine rather than enhance economic
efficiency and capital formation. 


         INVESTING IN PRIVATE
         FINANCIAL ASSETS
------------------------------------------------------ Chapter 5:2.1.3

A third way for the government to engage in capital formation would
be to invest the Trust Funds' assets in private securities.\92 This
would create the potential for larger Trust Funds, which could then
earn higher returns, further improving the program's solvency.  The
contribution of such a proposal to capital formation would be
contingent on a number of factors.  If the non-Social Security
portion of the unified budget was in deficit, such an action would be
unlikely to change national saving.  The purchase of stocks could
result in an equivalent issue of government bonds to provide
substitute financing for the payroll tax revenues that would have
been used concurrently to finance government expenditures.  The most
likely way for this proposal to represent funding that would lead to
higher saving and capital formation would be in the context of a
budget surplus, particularly one that had arisen from a balanced
non-Social Security budget.  But even if there were a budget surplus,
the proposal would generate only a small portion of the amount
necessary to fully advance fund future Social Security benefits. 
Such a Trust Funds investment proposal would also raise questions
about how a large, government-controlled fund would be managed and
whether political considerations would be introduced into the
management of the funds or the entities in which the funds were
invested.  Mechanisms could be designed to limit involvement of the
political process in allocating investments; however, it would be
likely that such involvement could not be completely precluded. 


--------------------
\92 This approach is treated in greater detail in
GAO/AIMD/HEHS-98-74, Apr.  22, 1998. 


      SOME SUGGEST PRIVATE ADVANCE
      FUNDING HAS ADVANTAGES
-------------------------------------------------------- Chapter 5:2.2

Many analysts believe that advance funding of Social Security would
increase the likelihood that the resources contributed to social
insurance programs would result in increased capital formation. 
While conceptually this could occur regardless of whether the funding
was done publicly or privately, in practice, there may be important
differences between public and private saving and investment
decision-making.  There is a substantial body of thought that
questions the ability of political institutions to ensure that the
resources raised through Social Security, in fact, contribute to
national saving and capital formation, and some suggest that funding
through private institutions or individuals is more likely to lead to
increased saving and capital formation than funding through public
institutions. 

The argument for private funding is essentially based on the notion
that private markets allocate capital efficiently.  The main
motivation underlying private market decisions regarding investment
is the generation of profit, or return.  The discipline that the
profit motive places on markets is key to the efficient allocation of
capital.  It is generally agreed that well-functioning, efficient
markets are fundamental for healthy economic growth. 

Proponents of the annuity-welfare model believe that funding through
private institutions would enhance the likelihood that resources
devoted to providing for retirement income would lead to increased
saving.  Private saving, especially for retirement, involves legal
arrangements that explicitly recognize ownership of, or benefit
rights to, contributed resources.  In the private pension field, such
arrangements are backed by legal fiduciary restrictions and
guidelines such as those that attempt to preclude noninvestment uses
of saved resources.  Advance funding Social Security would require
that sufficient resources be allocated to generate a future expected
benefit.  Thus, when sufficient resources were allocated in advance
and invested efficiently and productively in real assets, the
likelihood that these resources would represent saving, contribute to
capital formation, and generate economic returns would be maximized. 
While advance funding of retirement benefits is viewed as having
economic advantages over pay-as-you-go financing,\93 the choice
between public and private institutions hinges on judgments about
both the role of government and the relative weight given to the
adequacy and equity goals. 


--------------------
\93 Thompson and Upp, in "The Social Insurance Approach and Social
Security," 1997, sum up the evidence on saving in the following way: 
"a number of economists have examined the effect of pay-as-you-go
Social Security in the United States on individual saving.  On
balance, their results do not support the fear that such a system
will seriously erode savings and capital formation.  On the other
hand, studies in this country also suggest that funded pension plans
do have a positive effect on savings.  National savings may increase
by 30 to 40 percent of any increase in the amount of assets being
held in pension and other retirement accounts.  Taken together, these
two results suggest that, in the absence of offsetting changes in
government fiscal operations, shifting from a pay-as-you-go Social
Security system to an advance funded system would have a positive
effect on national savings."


      FINANCING TRANSITION COSTS
      COULD REDUCE SAVING
      RESULTING FROM ADVANCE
      FUNDING INDIVIDUAL ACCOUNTS
-------------------------------------------------------- Chapter 5:2.3

While increased advance funding of Social Security by the government
could potentially have a positive impact on national saving, the
impact would depend in part on what happened in the non-Social
Security part of the budget.  Advance funding through individual
accounts could also have a limited initial impact on national saving,
depending on how the transition was financed.  If the transition was
financed through more borrowing from the public, then the impact on
national saving would be reduced. 

The transition to full advance funding could mean that one generation
of workers would face a potentially staggering payroll tax rate.\94
While this might result in an eventual rise in saving, workers'
consumption would significantly drop in the near term, which could
negatively affect the performance of the economy for a considerable
period of time.  In addition, accumulating a significantly larger
stock of capital could have implications for financial markets, and
the prices of securities, the return to capital, or both could be
affected.  Thus, the advantages of advance funding hinge on the
likelihood that the higher saving would result in increased
productive investment and future economic growth.  If so, long-term
increases in the standard of living might be deemed to be worth the
disadvantage of reduced consumption in the near term. 

Incorporating a lesser degree of advance funding suggests a lower
transition cost.  As noted in chapter 4, if the current program were
terminated and a new fully funded one were started, the amount
necessary to pay the accrued benefits of current workers and current
beneficiaries would be $9 trillion.  These unfinanced costs of the
current system are the transition costs of moving to a fully funded
system.  The question is when, not whether, such costs will be
addressed.  Alternatives to paying for the full transition now
include benefit reductions, tax increases, or borrowing.  Extending
the period of time over which such costs must be met would spread the
burden over several generations. 


--------------------
\94 Feldstein and Samwick's recent work shows that the cost of paying
for the transition could be considerably smaller than previously
thought.  See Martin Feldstein and Andrew Samwick, "The Transition
Path in Privatizing Social Security," working paper no.  5761
(Cambridge, Mass.:  National Bureau of Economic Research, Sept. 
1996), and Martin Feldstein, "The Case for Privatization," Foreign
Affairs (July/Aug.  1997), pp.  24-38. 


OBSERVATIONS
============================================================ Chapter 6

Social Security has provided the basis on which most Americans have
built their retirement incomes for nearly 60 years.  The program has
been highly effective at reducing the incidence of poverty among the
elderly, and the disability and survivor benefits have been critical
to the financial well-being of millions of others.  While the
economy's recent performance has extended the projected life of the
Social Security Trust Funds, there is general agreement that Social
Security's revenues eventually will be inadequate to pay all promised
benefits.  The nation is now engaged in a debate about how best to
ensure the long-term solvency of the program.  A number of proposals
have been put forward and, while they share the goal of restoring
solvency, they contain significant differences reflecting alternative
perspectives as to the appropriate structure of Social Security in
the 21st century.  The approach chosen by decisionmakers will affect
nearly every American's retirement income and could be critically
important to the economic welfare of many, especially those relying
on survivor and disability benefits.  Moreover, the way we choose to
address the financing issue also could have important implications
for the long-term performance of the national economy. 

Many elements of the debate that surrounded the creation of the
program in the 1930s are resurfacing today.  The proposals that are
being advanced not only address the relatively narrow question of how
to restore solvency but also go to the larger question of what role
Social Security and the federal government should play in providing
retirement income.  The proposed reforms all include both individual
equity and income adequacy goals, but the balance struck between them
differs widely.  Today's social and economic environment is very
different from what prevailed in the 1930s when Social Security was
enacted.  Social Security originally was designed to replace a
portion of earnings lost because of retirement or unforeseen
circumstances.  However, Social Security is now viewed by many as the
most significant source of retirement income, and for many it is
their only source.  Because Social Security provides a lifetime
annuity that is indexed for inflation, it becomes an increasingly
important source as retirees grow older and exhaust other income
sources. 

Supporters of the existing program argue that Social Security's
financing problems could be addressed without changing the current
structure of the program.  A combination of revenue increases and
benefit reductions, similar to those that have been used in the past
to preserve solvency, equal to about 2.19 percent of taxable payroll
would be sufficient to restore long-term actuarial balance over the
next 75 years.  In addition, some supporters of maintaining the
existing structure propose to invest a portion of the Social Security
Trust Funds in the stock market to improve the flow of revenues.  Our
analysis shows that there are a number of adjustments that, in
combination, could restore long-term balance while leaving the
structure basically intact. 

Those who seek fundamental changes to the system do not believe that
a sustainable solution to the financing problems can be found within
the current structure of the program.  They argue that any
restoration of actuarial balance within the current pay-as-you-go
structure will be short-lived, as demographic trends continue to
cause future revenues to fall short of future expenditures. 
Maintaining the current system, they assert, would thus require
periodic increases in revenues, reductions in benefits, or both. 
Those supporting fundamental change generally call for replacing the
primarily pay-as-you-go system with one that relies more heavily on
advance funding and replacing, at least in part, the centralized
Trust Funds with individual accounts that are owned and managed by
the program participants.  These accounts could be invested in
securities that offered the potential for higher rates of return than
the implicit rate of return earned on Social Security contributions. 
Those advocating fundamental changes rely on historical stock market
performance to support their view that the increased risks associated
with individual accounts are unlikely to outweigh the benefits. 

Moving even part of Social Security to individual accounts would
raise many questions and challenges.  While individual accounts offer
the potential benefits of higher returns, they also expose
individuals to risks now borne collectively through the government. 
The nature of these risks and their potential impacts on different
groups of individuals, such as low earners, would need to be
carefully considered.  It would also be important to consider how
important ancillary benefits, such as disability and dependents'
benefits, would be treated and how other sources of retirement income
might be affected under a restructured Social Security program. 
Moreover, moving to an alternative program structure that included
advance funded individual accounts would require a decision regarding
how best to finance the transition costs.  Funding this transition
would require either supplementary taxes on current
generations--asking them in effect to "pay twice"--or a substantial
increase in government debt.  Further, a host of program design,
administrative, and oversight issues would need to be addressed.  The
costs of implementing the new program design and its administrative
requirements could offset some of the advantages of higher investment
returns associated with individual accounts. 

Another key element is the relative impact of different program
financing structures on aggregate saving and the national economy. 
Saving is critical to the economy's long-term growth, and a larger
economy in the future would help ease the burden of meeting
retirement costs while sustaining rising standards of living. 
Advocates of moving toward a system of individual accounts argue that
such a system would increase the nation's saving rate, although the
substantial transition costs associated with these proposals offset
the positive effects on saving in the short and medium term, pushing
positive economic effects even further into the future.  Raising
saving is only one of several important goals addressed in Social
Security financing reform proposals.  But because saving is so
important to societal goals, proposals that have the potential to
encourage saving should be carefully considered. 

While the debate continues over whether the existing system should be
maintained or whether fundamental restructuring is desirable, there
is broad consensus that action is needed soon to dilute the impact of
the changes and to give workers and their families time to adapt to
them.  Nonetheless, because such action will affect the nation and
its economy for years to come, decisions should be made with full
knowledge and debate of the trade-offs inherent in each proposed
change. 


ESTIMATED EFFECTS OF SELECTED
OPTIONS FOR REDUCING SOCIAL
SECURITY'S 75-YEAR ACTUARIAL
DEFICIT
=========================================================== Appendix I

                                                                                Estimated
                                                                           improvement in
                                                                                actuarial
                                                                             balance as a
                                                                            percentage of
                                                                                  taxable
                                                                             payroll over
Options for maintaining solvency                                                 75 years
-------------------------------------------------------------------------  --------------
Increasing revenue
-----------------------------------------------------------------------------------------
Expanding coverage
Requiring coverage for state and local workers hired after Dec. 1977                 0.22
Raising the payroll tax rate                                                       Varies
Expanding the taxable payroll
Increasing the maximum taxable earnings level to cover 90% of total                  0.48
 earnings
Including employer-provided group health and life insurance as covered               0.80
 earnings
Including pension and profit-sharing plans as covered earnings                       0.37
Subjecting employer-provided pension and profit-sharing contributions to             0.15
 a 3% payroll tax
Increasing the income tax on Social Security benefits
Eliminating current thresholds                                                       0.21
Taxing all benefits that exceed an employer's own contributions                      0.15
Crediting tax income currently allocated to the Hospital Insurance (HI)              0.36
 Trust Fund to the Old-Age, Survivors, and Disability Insurance (OASDI)
 Trust Fund
Using general revenues                                                             Varies

Earning a higher rate of return on Trust Funds' assets
-----------------------------------------------------------------------------------------
Investing 40% of OASDI Trust Funds in equities from 2000 to 2015 at a 7%             0.92
 real return

Reducing expenditures
-----------------------------------------------------------------------------------------

Eliminating or reducing certain spouse, survivor, and child benefits
-----------------------------------------------------------------------------------------
Limiting the spouse benefit to one-half the average primary insurance                0.21
 amount (PIA) of retired workers
Capping survivor benefits at the worker's maximum benefit in the year the            0.01
 individual was widowed
Eliminating benefits for nondisabled children of retired workers                     0.05
Relating benefits of disabled and deceased workers' children to household            0.04
 earnings

Reducing disabled worker benefits
-----------------------------------------------------------------------------------------
Limiting the initial disabled worker benefit to the retired worker                   0.40
 benefit available at age 65

Reducing retired worker benefits
-----------------------------------------------------------------------------------------
Increasing the number of years of earnings included in the benefit                   0.28
 computation period from 35 to 38
Reducing each of the three replacement rates by 0.5% between 2020 and                0.29
 2029
Indexing benefit formula bend points with either the consumer price index            1.54
 or the annual wage index minus 1 percentage point
Raising the normal retirement age (NRA) for those born between 1944 and              0.50
 1954 and indexing the NRA to maintain a constant proportion between the
 average adult lifetime and the NRA
Controlling the growth in benefits after entitlement
Reducing the cost-of-living adjustment (COLA) to equal the annual                    1.39
 increase in the consumer price index minus 1 percentage point
Reducing the COLA to equal the annual increase in the consumer price                 0.72
 index minus 0.5 percentage point
Limiting benefit increases that are based on recomputation of benefits                 \a
Restrengthening the earnings test                                                      \a
Disallowing most "new dependent" benefits                                              \a
Reducing benefits because of other income                                              \a
-----------------------------------------------------------------------------------------
\a Estimate not available. 

Source:  SSA. 


ESTIMATED INDIVIDUAL OUTCOMES FOR
THE THREE ADVISORY COUNCIL
PROPOSALS
========================================================== Appendix II

The following tables summarize selected estimates of outcomes under
the three Advisory Council proposals:  the maintain benefits (MB),
individual accounts (IA), and personal security accounts (PSA) plans. 
The tables show estimates of three measures:  the ratio of benefits
to taxes, the internal rates of return, and the benefit replacement
rates for single males and couples.  The tables also include
estimates of outcomes under present law. 



                         Table II.1
          
           Ratio of Present Value of Benefits to
            Taxes for the Three Advisory Council
                         Proposals

                     Current
                         law        MB        IA       PSA
------------------  --------  --------  --------  --------
Single male
----------------------------------------------------------

Low earner
----------------------------------------------------------
Born 1949, aged 22        79        79        73        79
 in 1971
Born 1973, aged 22       103       103        95       116
 in 1995
Born 1977, aged 22       112       110        98       121
 in 2019

Average earner
----------------------------------------------------------
Born 1949, aged 22        59        59        54        57
 in 1971
Born 1973, aged 22        77        77        70        84
 in 1995
Born 1977, aged 22        83        81        73        88
 in 2019

Maximum earner
----------------------------------------------------------
Born 1949, aged 22        43        43        40        43
 in 1971
Born 1973, aged 22        51        51        50        66
 in 1995
Born 1977, aged 22        55        54        53        71
 in 2019

Married couple
----------------------------------------------------------

One average earner
----------------------------------------------------------
Born 1949, aged 22       123       123       107       109
 in 1971
Born 1973, aged 22       151       151       121       127
 in 1995
Born 1977, aged 22       159       156       123       128
 in 2019

Two average earners
----------------------------------------------------------
Born 1949, aged 22        70        70        68        70
 in 1971
Born 1973, aged 22        86        86        90       101
 in 1995
Born 1977, aged 22        92        90        95       106
 in 2019
----------------------------------------------------------
Notes:  MB = maintain benefits plan, IA = individual accounts plan,
and PSA = personal security accounts plan.

Estimates assume a 0.21-percent lower COLA starting in Dec.  1997
with no change in nominal wage or interest. 

Source:  1994-1996 Advisory Council Report, 1997, Vol.  I, App.  II,
Tables 3A, 3As, and 3Am, pp.  200, 202, and 203.  Estimates prepared
by the Office of the Actuary, SSA, on the basis of the intermediate
assumptions of the 1995 Trustees Report. 



                         Table II.2
          
           Internal Rates of Return for the Three
                 Advisory Council Proposals

                     Current
                         law        MB        IA       PSA
------------------  --------  --------  --------  --------
Single male
----------------------------------------------------------

Low earner
----------------------------------------------------------
Born 1949, aged 22     2.43%     2.43%     2.18%     2.40%
 in 1971
Born 1973, aged 22      2.51      2.50      2.26      2.95
 in 1995
Born 1977, aged 22      2.68      2.61      2.30      3.00
 in 2019

Average earner
----------------------------------------------------------
Born 1949, aged 22      1.40      1.40      1.14      1.22
 in 1971
Born 1973, aged 22      1.48      1.48      1.20      1.77
 in 1995
Born 1977, aged 22      1.66      1.59      1.28      1.86
 in 2019

Maximum earner
----------------------------------------------------------
Born 1949, aged 22      0.19      0.15     -0.07      0.05
 in 1971
Born 1973, aged 22      0.07      0.06      0.04      0.85
 in 1995
Born 1977, aged 22      0.28      0.19      0.20      1.03
 in 2019
Married couple

One average earner
----------------------------------------------------------
Born 1949, aged 22      3.90      3.90      3.47      3.53
 in 1971
Born 1973, aged 22      3.78      3.77      3.11      3.32
 in 1995
Born 1977, aged 22      3.83      3.78      3.04      3.26
 in 2019

Two average earners
----------------------------------------------------------
Born 1949, aged 22      2.05      2.05      1.83      1.85
 in 1971
Born 1973, aged 22      2.09      2.09      2.09      2.47
 in 1995
Born 1977, aged 22      2.23      2.16      2.18      2.55
 in 2019
----------------------------------------------------------
Notes:  MB = maintain benefits plan, IA = individual accounts plan,
and PSA = personal security accounts plan. 

The internal rate of return is computed as the constant real rate of
return that is needed to equate the present value of contributions
under a given plan with the present value of benefits received under
the plan. 

Source:  1994-1996 Advisory Council Report, 1997, Vol.  I, App.  II,
Tables IRR1 and IRR3, pp.  219-21.  Estimates prepared by the Office
of the Actuary, SSA, on the basis of the intermediate assumptions of
the 1995 Trustees Report. 



                         Table II.3
          
          Benefit Replacement Rates for the Three
                 Advisory Council Proposals

                     Current
                         law        MB        IA       PSA
------------------  --------  --------  --------  --------
Low earners
----------------------------------------------------------
Born 1930, aged 65     58.2%     58.2%     58.2%     58.2%
 in 1995
Born 1960, aged 65      48.7      48.7      49.1      54.9
 in 2025
Born 1990, aged 65      48.7      48.7      49.4      60.4
 in 2055

Average earners
----------------------------------------------------------
Born 1930, aged 65      43.2      43.2      43.2      43.2
 in 1995
Born 1960, aged 65      36.2      36.2      35.6      38.0
 in 2025
Born 1990, aged 65      36.2      36.2      36.5      41.9
 in 2055

High earners
----------------------------------------------------------
Born 1930, aged 65      34.2      34.5      34.5      34.5
 in 1995
Born 1960, aged 65      29.9      29.9      29.6      31.9
 in 2025
Born 1990, aged 65      29.9      29.9      31.0      36.3
 in 2055

Maximum earners
----------------------------------------------------------
Born 1930, aged 65      23.8      23.8      23.8      23.8
 in 1995
Born 1960, aged 65      24.1      24.1      24.5      27.3
 in 2025
Born 1990, aged 65      24.0      24.0      26.5      33.2
 in 2055
----------------------------------------------------------
Notes:  MB = maintain benefits plan, IA = individual accounts plan,
and PSA = personal security accounts plan.

The replacement rate is the percentage of earnings in the last year
of work that is replaced by benefits in the first year.  Estimates
assume a 0.21-percent lower COLA starting in Dec.  1997 with no
change in nominal wage or interest. 

Source:  1994-96 Advisory Council Report, 1997, Vol.  I, App.  II,
Tables RR.1 through RR.4, pp.  223-26.  Estimates prepared by the
Office of the Actuary, SSA, on the basis of the intermediate
assumptions of the 1995 Trustees Report. 




(See figure in printed edition.)Appendix III
COMMENTS FROM THE SOCIAL SECURITY
ADMINISTRATION
========================================================== Appendix II



(See figure in printed edition.)



(See figure in printed edition.)


MAJOR CONTRIBUTORS TO THIS REPORT
========================================================== Appendix IV

Francis P.  Mulvey, Assistant Director, (202) 512-3592
Kenneth J.  Bombara, Senior Economist
Michael D.  Packard, Senior Economist


BIBLIOGRAPHY
============================================================ Chapter 1

Aaron, H., B.  Bosworth, and G.  Burtless.  Can American Afford to
Grow Old?  Paying for Social Security.  Washington, D.C.:  The
Brookings Institution, 1989. 

Advisory Council on Social Security.  Report of the 1994-1996
Advisory Council on Social Security, Volumes I and II.  Washington,
D.C.:  Advisory Council on Social Security, 1997. 

Baker, Dean.  Saving Social Security With Stocks:  The Promises Don't
Add Up.  New York:  The Twentieth Century Fund, 1997. 

Bipartisan Commission on Entitlement and Tax Reform.  "Staff Report
on Entitlement Reform Options:  Reference Materials." Washington,
D.C.:  Bipartisan Commission on Entitlement and Tax Reform, Dec. 
1994. 

Board of Trustees, Federal Old-Age and Survivors Insurance and
Disability Insurance Trust Funds.  The 1998 Annual Report of the
Board of Trustees of the Federal Old-Age and Survivors Insurance and
Disability Insurance Trust Funds.  Washington, D.C.:  U.S. 
Government Printing Office, 1998. 

Boskin, Michael J.  Too Many Promises:  The Uncertain Future of
Social Security.  Homewood, Ill.:  Dow Jones-Irwin, 1986. 

_____, and others.  Toward a More Accurate Measure of the Cost of
Living, final report to the Senate Finance Committee from the
Advisory Commission to Study the Consumer Price Index.  Washington,
D.C.:  Advisory Commission to Study the Consumer Price Index, Dec. 
4, 1996. 

Brown, J.  Douglas.  An American Philosophy of Social Insurance. 
Princeton, N.J.:  Princeton University Press, 1972. 

Browning, Edgar.  "The Politics of Social Security Reform," in
Financing Social Security, Colin D.  Campbell, ed.  Washington, D.C.: 
American Enterprise Institute, 1977. 

Bryce, David V., and Robert B.  Friedland.  "Economic Security:  An
Overview of Social Security," in Assessing Social Security Reform
Alternatives, Dallas L.  Salisbury, ed.  Washington, D.C.:  Employee
Benefit Research Institute (EBRI), 1997. 

Buchanan, James M.  "Social Insurance in a Growing Economy." National
Tax Journal, Vol.  21, No.  4 (Dec.  1965). 

_____.  "Social Security Survival:  A Public Choice Perspective."
Cato Journal, Vol.  3, No.  2 (fall 1983), pp.  339-53. 

_____, and Colin D.  Campbell.  "Voluntary Social Security." Wall
Street Journal (Dec.  20, 1966). 

Carter, Marshall N., and William G.  Shipman.  Promises to Keep: 
Saving Social Security's Dream.  Washington, D.C.:  Regnery
Publishing, Inc., 1996. 

Committee for Economic Development.  Who Will Pay for Your
Retirement?  The Looming Crisis.  New York:  Committee for Economic
Development, 1995. 

Congressional Budget Office.  Implications of Revising Social
Security's Investment Policies.  Washington, D.C.:  Congressional
Budget Office, Sept.  1994. 

_____.  Reducing the Deficit:  Spending and Revenue Options. 
Washington, D.C.:  Congressional Budget Office, Aug.  1996. 

Council of Economic Advisors.  Economic Report of the President,
1997.  Washington, D.C.:  U.S.  Government Printing Office, Feb. 
1997. 

Diamond, Peter A.  "Proposals to Restructure Social Security."
Journal of Economic Perspectives, Vol.  10, No.  2 (summer 1996), pp. 
67-88. 

_____, David C.  Lindeman, and Howard Young, eds.  Social Security: 
What Role for the Future?  Washington, D.C.:  National Academy of
Social Insurance, 1996. 

Employee Benefit Research Institute.  Assessing Social Security
Reform Alternatives, Dallas L.  Salisbury, ed.  Washington, D.C.: 
EBRI, 1997. 

_____.  EBRI Notes, Vol.  19, No.  4 (Apr.  1998), pp.  6-10. 

Federal Reserve Bank of Boston.  Social Security Reform Conference
Proceedings:  Links to Savings, Investment and Economic Growth,
conference series no.  41.  Boston:  Federal Reserve Bank of Boston,
June 1997. 

Feldstein, Martin.  "Social Security and Savings:  New Time Series
Evidence." National Tax Journal, Vol.  49, No.  2 (June 1996), pp. 
151-64. 

_____.  "Social Security, Induced Retirement, and Aggregate Capital
Accumulation." Journal of Political Economy, Vol.  82, No.  5
(Sept./Oct.  1974), pp.  905-26. 

_____.  "The Case for Privatization." Foreign Affairs (July/Aug. 
1997), pp.  24-38. 

_____.  "The Missing Piece in Policy Analysis:  Social Security
Reform." American Economic Review, Vol.  86, No.  2 (May 1996), pp. 
1-14. 

_____, and Andrew Samwick.  "The Transition Path in Privatizing
Social Security," working paper no.  5761.  Cambridge, Mass.: 
National Bureau of Economic Research, Sept.  1996. 

Ferrara, Peter.  A Plan for Privatizing Social Security, SSP No.  8. 
Washington, D.C.:  The Cato Project on Social Security Privatization,
Apr.  30, 1997. 

_____.  Social Security:  The Inherent Contradiction.  Washington,
D.C.:  Cato Institute, 1980. 

_____, ed.  Social Security:  Prospects for Real Reform.  Washington,
D.C.:  Cato Institute, 1985. 

Friedman, Milton.  Capitalism and Freedom.  Chicago:  University of
Chicago Press, 1962. 

Goss, Stephen C.  "Measuring Solvency in the Social Security System,"
presented at the 1997 Pension Research Council Symposium. 
Philadelphia:  The Wharton School of the University of Pennsylvania,
May 12, 1997. 

Grad, Susan.  Income of the Population 55 or Older, 1994. 
Washington, D.C.:  U.S.  Government Printing Office, Jan.  1996. 

Gramlich, Edward M.  "Different Approaches for Dealing With Social
Security." American Economic Review, Vol.  86, No.  2 (May 1996), pp. 
358-62. 

Harris, Seymour.  The Economics of Social Security.  Westport, Conn.: 
Greenwood Press, 1941. 

Hieger, Melissa, and William Shipman.  Common Objections to a
Market-Based Social Security System:  A Response, SSP No.  10. 
Washington, D.C.:  The Cato Project on Social Security Privatization,
July 22, 1997. 

Howe, Neil, and Richard Jackson.  National Thrift Plan Project: 
Draft Final Report.  Washington, D.C.:  National Taxpayers Union
Foundation, June 21, 1996. 

Kingson, Eric R., and James H.  Schulz, eds.  Social Security in the
21st Century.  New York:  Oxford University Press, 1997. 

Kollmann, Geoffrey, Ray Schmitt, and Michelle Harman.  Effect of
Pension Integration Rules on Retirement Benefits, report to the
Congress (94-974 EPW).  Washington, D.C.:  Congressional Research
Service, Dec.  6, 1994. 

Kotlikoff, Laurence J.  "Privatizing Social Security at Home and
Abroad." American Economic Review, Vol.  86, No.  2 (May 1996), pp. 
368-72. 

Lopez, Eduard A.  "Constitutional Background to the Social Security
Act of 1935." Social Security Bulletin, Vol.  50, No.  1 (Jan. 
1987), pp.  5-11. 

Meriam, Lewis.  Relief and Social Security.  Washington, D.C.:  The
Brookings Institution, 1946. 

Miron, Jeffrey A., and David N.  Weil.  "The Genesis and Evolution of
Social Security," working paper no.  5949.  Cambridge Mass.: 
National Bureau of Economic Research, Mar.  1997. 

Mitchell, Olivia S.  "Administrative Costs in Public and Private
Retirement Systems," working paper no.  5734.  Cambridge, Mass.: 
National Bureau of Economic Research, Aug.  1996. 

_____, James M.  Poterba, and Mark J.  Warshawsky.  "New Evidence on
the Money's Worth of Individual Annuities," working paper no.  6002. 
Cambridge, Mass.:  National Bureau of Economic Research, Apr.  1997. 

_____, and Stephen P.  Zeldes.  "Social Security Privatization:  A
Structure for Analysis." American Economic Review, Vol.  86, No.  2
(May 1996), pp.  363-67. 

Myers, Robert J.  "Chile's Social Security Reform After Ten Years."
Benefits Quarterly, Vol.  8, No.  3 (third quarter 1992), p.  56. 

_____.  Social Security, Fourth Edition.  Philadelphia:  Pension
Research Council and the University of Pennsylvania Press, 1993. 

_____.  "Social Security:  Myths and Realities," in Assessing Social
Security Reform Alternatives, Dallas L.  Salisbury, ed.  Washington,
D.C.:  EBRI, 1997. 

National Academy of Social Insurance.  Social Insurance Update, Vol. 
1, No.  3 (Dec.  1996). 

National Conference on Social Welfare.  50th Anniversary Edition: 
The Report of the Committee on Economic Security of 1935 and Other
Basic Documents Relating to the Development of the Social Security
Act.  Washington, D.C.:  National Conference on Social Welfare, 1985. 

Packard, Michael D.  "The Earnings Test and the Short-Run Work
Response to Its Elimination." Social Security Bulletin, Vol.  53, No. 
9 (Sept.  1990), pp.  2-16. 

Poterba, James M.  "The History of Annuities in the United States,"
working paper no.  6001.  Cambridge, Mass.:  National Bureau of
Economic Research, Apr.  1997. 

Schieber, Sylvester J., and John B.  Shoven.  "Social Security
Reform:  Around the World in Eighty Ways." American Economic Review,
Vol.  86, No.  2 (May 1996), pp.  373-77. 

Social Security Administration (SSA).  "A Brief History of the Social
Security Administration." Baltimore:  SSA, Mar.  1995. 

_____.  Annual Statistical Supplement to the Social Security
Bulletin.  Washington, D.C.:  U.S.  Government Printing Office, Dec. 
1997. 

_____.  Social Security Handbook, 1993, 11th Edition.  Washington,
D.C.:  U.S.  Government Printing Office, Aug.  1993. 

Steurle, C.  Eugene, and Jon M.  Bakija.  Retooling Social Security
for the 21st Century.  Washington, D.C.:  The Urban Institute Press,
1994. 

Thompson, Lawrence H.  "The Social Security Reform Debate." Journal
of Economic Literature, Vol.  21 (Dec.  1983), pp.  1425-67. 

_____, and Melinda M.  Upp.  "The Social Insurance Approach and
Social Security," in Social Security in the 21st Century, Eric
Kingson and James H.  Schulz, eds.  New York:  Oxford University
Press, 1997, pp.  3-21. 

U.S.  General Accounting Office.  Budget Issues:  Analysis of
Long-Term Fiscal Outlook.  GAO/AIMD/OCE-98-19, Oct.  22, 1997. 

_____.  Budget Issues:  Long-Term Fiscal Outlook. 
GAO/T-AIMD/OCE-98-83, Feb.  25, 1998. 

_____.  Federal Pensions:  Thrift Savings Plan Has Key Role in
Retirement Benefits.  GAO/HEHS-96-1, Oct.  19, 1995. 

_____.  401(k) Pension Plans:  Extent of Plans' Investments in
Employer Securities and Real Property.  GAO/HEHS-98-28, Nov.  28,
1997. 

_____.  401(k) Pension Plans:  Loan Provisions Enhance Participation
but May Affect Income Security for Some.  GAO/HEHS-98-5, Oct.  1,
1997. 

_____.  401(k) Pension Plans:  Many Take Advantage of Opportunity to
Ensure Adequate Retirement Income.  GAO/HEHS-96-176, Aug.  2, 1996. 

_____.  Retirement Income:  Implications of Demographic Trends for
Social Security and Pension Reform.  GAO/HEHS-97-81, July 11, 1997. 

_____.  Social Security:  Analysis of a Proposal to Privatize Trust
Fund Reserves.  GAO/HRD-91-22, Dec.  12, 1990. 

_____.  Social Security Disability:  Improvements Needed to
Continuing Disability Review Process.  GAO/HEHS-97-1, Oct.  16, 1996. 

_____.  Social Security Disability:  Management Action and Program
Redesign Needed to Address Long-Standing Problems. 
GAO/T-HEHS-95-233, Aug.  3, 1995. 

_____.  Social Security:  Disability Programs Lag in Promoting Return
to Work.  GAO/HEHS-97-46, Mar.  17, 1997. 

_____.  Social Security Disability:  SSA Must Hold Itself Accountable
for Continued Improvements in Decision-making.  GAO/HEHS-97-102, Aug. 
12, 1997. 

_____.  Social Security Financing:  Implications of Government Stock
Investing for the Trust Fund, the Federal Budget, and the Economy. 
GAO/AIMD/HEHS-98-74, Apr.  22, 1998. 

_____.  Social Security Funds:  Additional Measures Could More Fully
Indicate the System's Financial Condition.  GAO/PEMD-88-11, Feb.  5,
1988. 

_____.  Social Security:  Issues Involving Benefit Equity for Working
Women.  GAO/HEHS-96-55, Apr.  10, 1996. 

_____.  Social Security:  Past Projections and Future Financing
Concerns.  GAO/HRD-86-22, Mar.  11, 1986. 

_____.  Social Security Reform:  Implications for the Financial
Well-Being of Women.  GAO/T-HEHS-97-112, Apr.  10, 1997. 

_____.  Social Security Reform:  Implications for Women's Retirement
Income.  GAO/HEHS-98-42, Dec.  31, 1997. 

_____.  Social Security:  Restoring Long-Term Solvency Will Require
Difficult Choices.  GAO/T-HEHS-98-95, Feb.  10, 1998. 

_____.  Social Security Retirement Accounts.  GAO/HEHS-94-226R, Aug. 
12, 1994. 

_____.  Social Security:  The Trust Fund Reserve Accumulation, the
Economy, and the Federal Budget.  GAO/HRD-89-44, Jan.  19, 1989. 

Warshawsky, Mark J.  "The Market for Individual Annuities and Reform
of Social Security." Benefits Quarterly (third quarter 1997). 

Weaver, Carolyn L.  "The Economics and Politics of the Emergence of
Social Security:  Some Implications for Reform." Cato Journal, Vol. 
3, No.  2 (fall 1983), pp.  361-79. 

Wheeler, Peter M., and John R.  Kearney.  "Income Protection for the
Aged in the 21st Century:  A Framework to Help Inform the Debate."
Social Security Bulletin, Vol.  59, No.  2 (summer 1996), pp.  3-19. 

World Bank, Averting the Old Age Crisis, policy research report.  New
York:  Oxford University Press, 1994. 


*** End of document. ***