Retirement Income: Implications of Demographic Trends for Social Security
and Pension Reform (Chapter Report, 07/11/97, GAO/HEHS-97-81).

Pursuant to a congressional request, GAO provided an overview of the
trends and key issues relating to retirement income, focusing on: (1)
demographic and economic trends affecting retirement income; (2) the
status of Social Security's long-term financing problems and proposals
to address them; and (3) the extent of pension coverage and retirement
saving and how to ensure that Americans can count on them throughout
their retirement years.

GAO noted that: (1) as the elderly live longer, they will need
retirement income over longer periods; (2) since fertility rates have
been declining, the number of the elderly will grow as a share of the
population; (3) while the income of the elderly has improved
considerably over the past 50 years, demographic trends may slow or
reverse such improvements; (4) the nation must confront how the trends
will affect the distribution of income between workers and retirees,
between the population's working and retirement years, and between high
and low earners; (5) these trends pose long-term financing challenges
for both Social Security and the federal budget; (6) currently, Social
Security revenues exceed expenditures; (7) the excess revenues are
invested by law in federal government securities and make the total
federal deficit lower than it would be otherwise; however, in just 15
years, expenditures are expected to exceed cash revenues, according to
Social Security Administration (SSA) projections; (8) at that point, the
government's general fund will have to make up the difference--in
effect, repaying funds owed to Social Security; (9) such repayment will
present a significant and growing challenge for the overall federal
budget; (10) by 2029, without corrective action, the trust funds will be
depleted, and Social Security's revenues will fund only 70 to 77 percent
of benefits; (11) to restore Social Security's long-term financial
balance, a number of reform options are available within the current
structure of the program; (12) some proposals go beyond restoring
financial balance and call for fundamentally restructuring the Social
Security system; (13) these proposals attempt to shift more of the
responsibility for retirement income from the federal government to
individuals; (14) solutions to Social Security's problems will
inevitably affect other sources of income and could give them an even
more significant role; (15) the proportion of workers covered by
pensions has not increased substantially since 1970; (16) ensuring that
Americans have enough retirement income to meet their needs will require
that the nation and the Congress make some difficult choices; (17)
Social Security has been an effective agent for ensuring a reliable
source of retirement income and greatly reducing poverty among the
elderly; (18) the effect of changes to the system on other retirement i*

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

 REPORTNUM:  HEHS-97-81
     TITLE:  Retirement Income: Implications of Demographic Trends for 
             Social Security and Pension Reform
      DATE:  07/11/97
   SUBJECT:  Population statistics
             Social security benefits
             Retirement benefits
             Elderly persons
             Future budget projections
             Aid for the elderly
             Budget deficit
IDENTIFIER:  OASDI
             Old Age Survivors and Disability Insurance Program
             Social Security Trust Fund
             
******************************************************************
** 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 Chairman and Ranking Minority Member, Special Committee
on Aging, U.S.  Senate

July 1997

RETIREMENT INCOME - IMPLICATIONS
OF DEMOGRAPHIC TRENDS FOR SOCIAL
SECURITY AND PENSION REFORM

GAO/HEHS-97-81

Retirement Income Issues

(207444)


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

  AIME - average indexed monthly earnings
  COLA - cost-of-living adjustment
  CPI - Consumer Price Index
  DB - defined benefit
  DC - defined contribution
  DOL - Department of Labor
  ERA - early retirement age
  ERISA - Employee Retirement Income Security Act
  GDP - gross domestic product
  IRA - individual retirement account
  IRS - Internal Revenue Service
  NRA - normal retirement age
  PBGC - Pension Benefit Guaranty Corporation
  PIA - primary insurance amount
  SEP - simplified employee pension
  SIMPLE - savings incentive match plan for employees
  SSA - Social Security Administration
  TSP - Thrift Savings Plan

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


B-276326

July 11, 1997

The Honorable Charles E.  Grassley
Chairman
The Honorable John B.  Breaux
Ranking Minority Member
Special Committee on Aging
United States Senate

Our nation's elderly population is growing dramatically.  In 1940,
when Social Security first paid benefits, it was 7 percent of the
U.S.  population; today it is 13 percent.  By 2030, the elderly
population will be 20 percent of the population.  This demographic
trend has profound implications not just for Social Security but for
all sources of retirement income, which also include pensions,
savings, and earnings.  While currently most of our nation's senior
citizens enjoy a reasonable standard of living, this trend raises
concerns that our society will continue to be able to maintain it. 
This report provides an overview of the trends and key issues
relating to retirement income. 

We are sending this report to the Commissioner of Social Security and
relevant committees and subcommittees.  The report will be available
to others on request. 

This report was prepared under my direction.  Please contact Francis
P.  Mulvey, Assistant Director, at (202) 512-3592 if you have
questions. 

Jane L.  Ross
Director, Income Security Issues


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


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

The U.S.  elderly population has tripled since 1940 and will more
than double by 2050, according to Bureau of the Census projections. 
The very old population (aged 85 and over) will increase fivefold. 
The elderly are expected to make up 20 percent of the U.S. 
population as early as 2030 compared with 13 percent today and just 7
percent in 1940.  These dramatic demographic trends pose serious
concerns about the future financing, availability, and protection of
retirement income for the nation's elderly.  Numerous reform
proposals under discussion could fundamentally change the nation's
retirement system. 

The Chairman and Ranking Minority Member of the Senate Special
Committee on Aging, concerned about these issues, asked GAO to
provide information on (1) demographic and economic trends affecting
retirement income, (2) the status of Social Security's long-term
financing problems and proposals to address them, and (3) the extent
of pension coverage and retirement saving and how to ensure that
Americans can count on them throughout their retirement years. 


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

The four major sources of retirement income are Social Security,
employer pensions, income from saved assets, and employment earnings. 
While Social Security provides income to 90 percent of elderly
households, it provides just 42 percent of their aggregate cash
income.  Pensions, savings, and earnings provide income to
considerably fewer households but together provide more than 50
percent of the elderly's aggregate income.  They largely determine
which households have the highest retirement incomes. 

As people live longer and if retirement patterns do not change,
Social Security and employer pensions will cost more or will have to
provide reduced benefits or both.  Personal retirement savings will
have to last longer, and workers may find themselves needing to work
longer.  Federal policies play a major role in all these sources of
retirement income and will play a significant role in helping adjust
to demographic trends. 

Title II of the Social Security Act established the Old-Age,
Survivors, and Disability Insurance program, which is generally known
as Social Security.  Initially, it provided old-age benefits to
nearly all workers in commerce and industry, or about 60 percent of
the workforce.  Over the years, coverage has been expanded to
dependents and survivors of insured workers, disabled workers, and
workers in other types of employment. 

With regard to pensions, the federal government has an interest in
promoting expanded coverage of workers so they can better meet their
retirement income needs.  The Employee Retirement Income Security Act
(ERISA) of 1974 focused new attention on making pension promises more
explicit and benefits more secure.  In addition to its minimum
funding requirements and other provisions, ERISA established a system
of pension insurance, administered by the Pension Benefit Guaranty
Corporation (PBGC).  PBGC assumes liability for certain types of
pensions when they are terminated and pays the pension benefits,
subject to certain limits. 

Federal tax policy also affects how people save for retirement. 
Since the 1970s, federal laws have granted favorable tax treatment to
individual retirement accounts (IRA) and the closely related 401(k)
pension plans.  Nevertheless, personal saving rates are at a historic
low.  Finally, federal policies also affect postretirement earnings. 
Most notably, provisions in the Social Security program affect
incentives regarding when workers take benefits and how much
beneficiaries choose to earn after retirement. 


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

As the elderly live longer, they will need retirement income over
longer periods.  Since fertility rates have been declining, the
number of the elderly will grow as a share of the population.  While
the income of the elderly has improved considerably over the past 50
years, demographic trends may potentially slow or reverse such
improvements.  Economic growth, if strong, could ease some of the
pressures these trends create.  Still, the nation must confront how
the trends will affect the distribution of income between workers and
retirees, between the population's working and retirement years, and
between high and low earners. 

These trends pose long-term financing challenges for both Social
Security and the federal budget.  Currently, Social Security revenues
exceed expenditures.  The excess revenues are invested by law in
federal government securities and make the total federal deficit
lower than it would be otherwise.  However, in just 15 years,
expenditures are expected to exceed cash revenues, according to
Social Security Administration (SSA) projections.  At that point, the
government's general fund will have to make up the difference--in
effect, repaying funds owed to Social Security.  Such repayment will
present a significant and growing challenge for the overall federal
budget.  By 2029, without corrective action, the trust funds will be
depleted, and Social Security's revenues will fund only 70 to 77
percent of benefits. 

To restore Social Security's long-term financial balance, a number of
reform options are available within the current structure of the
program, such as expanding the number of covered workers or further
increasing the retirement age.  However, some proposals go beyond
restoring financial balance and call for fundamentally restructuring
the Social Security system.  These proposals attempt to shift more of
the responsibility for retirement income from the federal government
to individuals. 

Social Security is meant to provide only a foundation for retirement
income.  Pensions, savings, and earnings also represent significant
sources of retirement income and largely determine which households
have the highest retirement incomes.  Solutions to Social Security's
problems will inevitably affect these other sources and could give
them an even more significant role.  More workers will eventually
receive income from pensions because of legislative changes regarding
vesting and spouse benefits and also because of women's increasing
labor force participation.  Nevertheless, the proportion of workers
covered by pensions at a given moment has not increased substantially
since 1970.  In addition, the national saving rate is at a historic
low.  Moreover, some trends threaten to diminish retirement income
from pensions and savings.  For example, workers increasingly borrow
from their pension savings, cash out their pensions when they change
jobs, or, at retirement, take their pensions as lump-sum payments
rather than a guaranteed life annuity. 

Ensuring that Americans have enough retirement income in the
twenty-first century to meet their needs will require that the nation
and the Congress make some difficult choices.  Social Security has
been an effective agent for ensuring a reliable source of income in
retirement and greatly reducing poverty among the elderly.  The
effect of changes to the system on other retirement income sources
and their effects on various groups within the elderly population
should be well understood before decisions are made.  Further, the
interplay of budget and savings effects will have to be carefully
considered before any reform proposal is adopted. 


   PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4


      DEMOGRAPHIC AND ECONOMIC
      TRENDS PRESENT SERIOUS
      CHALLENGES
-------------------------------------------------------- Chapter 0:4.1

Increasing life expectancy and declining birth rates are responsible
for the substantial growth in the number of the elderly as a share of
the total population.  The baby-boom generation will simply
accelerate this trend.  The elderly will need retirement income over
longer periods and will consume an increasing share of the national
output.  By 2030, only two workers are projected to be paying into
Social Security for each person receiving benefits, compared with
more than three workers today, according to SSA actuaries. 

Healthy economic growth could ease the pressures of supporting more
retirees with fewer workers.  Unfortunately, the rate of national
saving and the growth in real wages and productivity, factors that
relate to economic growth, have slowed notably in the past two
decades.  Even with strong economic performance, decisions affecting
retirement income policies still have to be made and will
fundamentally influence how the national output will be divided. 
They influence how much of total compensation workers consume now and
how much they will consume during retirement.  They also influence
how income is redistributed between current workers and current
retirees and between high earners and low earners. 

Social Security and pension benefits have helped increase the income
of the elderly over the past 50 years.  For example, the share of the
elderly living in poverty has fallen from 35 percent in 1959 to less
than 11 percent today.  Still, pockets of poverty remain, especially
among those aged 75 and over and among unmarried women. 


      ADDRESSING SOCIAL SECURITY'S
      LONG-TERM FINANCING PROBLEM
-------------------------------------------------------- Chapter 0:4.2

The growth of the elderly population as a share of the total
population poses serious long-term financing challenges for Social
Security.  In January 1997, the Social Security Advisory Council
issued its report on the system's long-range financial status, but
the council members could not reach agreement on a comprehensive
solution.  Still, five individual reform options enjoyed considerable
support and, if enacted, would eliminate about two thirds of the next
75 years' financing shortfall.  These reforms would modify (1) the
retirement age, (2) cost-of-living adjustments (COLA), (3) the
benefit formula, (4) income taxation of benefits, and (5) which
workers are covered under Social Security. 

While these reform options would work within the current structure of
the program, "privatization" proposals go further and would
significantly restructure Social Security.  Such proposals would
achieve financial balance by raising revenues or reducing costs while
pursuing other goals, such as increasing national saving or giving
individuals more responsibility for their retirement income.  For
example, such proposals would typically increase the role of
individual retirement saving while reducing government-provided
benefits.  For many of these privatization proposals, the transition
from the current program to the new system would be quite expensive. 
The contributions needed to fund both current and future liabilities
would clearly be higher than those currently collected. 


      THE SOCIAL SECURITY TRUST
      FUND BUILD-UP AND THE
      FEDERAL DEFICIT
-------------------------------------------------------- Chapter 0:4.3

Solving Social Security's long-term financing problem would not
necessarily address another significant challenge that the program
presents for the overall federal budget.  Currently, Social Security
cash revenues exceed expenditures by roughly $30 billion each year. 
Under current law, the Department of the Treasury issues
interest-bearing government securities to the trust funds for these
excess revenues.  In effect, Treasury borrows the excess revenues and
uses them to help reduce the amount it must borrow from the public. 
Moreover, the trust funds earned $38 billion in interest last year,
which Treasury pays by issuing more securities.  If Treasury could
not borrow from the trust funds, it would have to borrow more in the
private capital market and pay such interest in cash. 

However, 10 years from now, these excess cash revenues are expected
to start falling, and in just 15 years, Social Security's
expenditures are expected to exceed its cash revenues.  The
government's general fund will have to make up the difference, in
effect repaying Social Security.  As a result, the federal deficit
will increase, unless offset by spending reductions or revenue
increases.  In 2028, repayments from the general fund to Social
Security are expected to reach about $183 billion in 1997 dollars. 
In that year, this amount would equal 1.4 percent of gross domestic
product (GDP), which is the same share of GDP as last year's deficit
for the entire federal government. 


      STRENGTHENING EMPLOYER
      PENSIONS AND PERSONAL
      RETIREMENT SAVINGS
-------------------------------------------------------- Chapter 0:4.4

Employer pensions and personal retirement savings are also
significant sources of retirement income.  About 47 percent of
workers had private pension coverage in 1993.  Although this coverage
rate has changed little since the 1970s, more workers will eventually
receive income from pensions because of legislative changes in the
1980s regarding vesting and spouse benefits and also because of
women's increasing labor force participation.  Still, these changes
also imply smaller average pension benefits because they will reflect
the shorter job tenures of these workers who would not have received
pensions previously.  Also, joint-and-survivor annuities yield
smaller monthly benefits than worker-only annuities.  In addition,
complex pension regulations may deter employers from sponsoring
pensions.  Current proposals to expand pension coverage focus on
simplifying some of the regulations. 

Efforts to increase personal savings outside pension arrangements
seem to have had only marginal success.  Many households do not save
in any systematic way, and personal saving rates have declined by
nearly half since 1970.  Preferential tax treatment for 401(k) and
individual retirement accounts seems to encourage saving, but such
saving may substitute, at least partly, for other types of saving. 

To help meet the nation's retirement income challenges, pensions and
savings must be secure and wisely managed.  While tax laws discourage
using pension assets for nonretirement purposes, the laws do not
strictly prohibit such use.  Some workers cash out their pension
accounts when they change jobs, despite penalties and tax
liabilities.  At retirement, most take their pensions as a lump sum,
which they must carefully manage to provide income for the rest of
their lives.  In addition, most plans allow workers to borrow from
their accounts for specified purposes.  The loan option may encourage
workers to contribute more to these accounts and, in some cases, such
as borrowing for educational expenses, could result in greater
retirement income security.  However, if they do borrow, they may
reduce the savings available for retirement. 

In the case of retirement savings and some pension plans, workers
control how their savings are invested.  However, many workers may
lack the information necessary to get higher investment earnings. 
Research shows that educating employees in how to manage their funds
can increase both their retirement contributions and investment
returns. 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5

GAO is not making recommendations in this report. 


   AGENCY COMMENTS
---------------------------------------------------------- Chapter 0:6

GAO obtained comments on this report from SSA and from subject matter
experts.  They did not express any disagreement with the overall
message of the report.  They did suggest some technical corrections,
which have been incorporated as appropriate. 


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

As the U.S.  population ages, there will be growing pressures on
Social Security and pension programs to provide retirees with the
retirement income they need.  For more than 60 years, the Social
Security program has helped protect America's elderly from living in
poverty, but it now faces a long-term financing problem.  Proposals
to address the problem are wide-ranging, and many, if adopted, could
fundamentally change the nation's retirement system. 


   THE SOURCES OF RETIREMENT
   INCOME
---------------------------------------------------------- Chapter 1:1

When individuals retire from their principal employment, they may
receive income from one or more of several sources, primarily Social
Security, pensions, personal savings, and earnings from subsequent
work.  Federal policies play a major role in all these sources of
retirement income.  Moreover, they affect how income is distributed
between workers and retirees, between high and low earners, and
between individuals' working and retirement years. 

The Social Security Act was signed into law by President Franklin D. 
Roosevelt in August 1935 in response to the crisis Americans faced
during the Great Depression.  The retirement component of the act
initially provided benefits upon retirement to workers in commerce
and industry who had contributed to the program over a portion of
their working lives. 

Social Security was meant to provide a retirement income foundation
upon which individuals could build for their retirement years.  It
was a social insurance program.  Over the years, the size and scope
of the program has changed.  In 1939, coverage was extended to
provide benefits for some family members and survivors.  In the
1950s, state and local governments were given the option of covering
their employees.  The Disability Insurance program was added in 1956,
providing income for disabled workers.  The Medicare program was
added in 1965.  Beginning in 1975, benefits were automatically tied
to the Consumer Price Index (CPI) to ensure that the purchasing power
of recipients' income was not eroded by inflation. 

Because the Social Security program was designed as a pay-as-you-go
system with current benefits paid out of current income (primarily
from payroll taxes), it experienced periodic financing difficulties
as coverage and benefits increased and as the number of retirees
increased relative to the number of contributors.  Payroll taxes
increased from an initial 2 percent on the first $3,000 of earnings
(1 percent each from employers and employees) to 12.4 percent on the
first $65,400 of earnings today.  For many Americans, the payroll tax
is now the largest tax they pay. 

Some of the increases in payroll taxes were always anticipated as a
natural result of the maturing of the Social Security system and its
pay-as-you-go financing.  The first benefits were paid only to
individuals who had already made some contributions to the system. 
In the early years, a very small share of the elderly qualified for
benefits, and larger payroll taxes were not required.  As more and
more of the elderly received benefits, higher payroll taxes were
required.  While such increases had always been anticipated, some of
the payroll tax increases resulted from benefit increases, especially
cost-of-living adjustments, and eligibility expansion such as
survivor and disability benefits. 

Pensions are the second major source of retirement income.  Private
pensions grew rapidly in popularity between 1900 and 1920 until about
one worker in six was covered--typically those working in the
transportation, banking, mining, utility, and manufacturing
industries.  Pension coverage basically held its own during the Great
Depression and again grew rapidly between 1940 and 1960 and continued
to increase into the 1970s when coverage leveled off at just under
half the workforce.  Workers nearing retirement are the most likely
to have pension coverage. 

The federal government has an interest in promoting expanded pension
coverage of workers so they can better meet their retirement income
needs and because expanded coverage increases national saving.  In
addition, the Employee Retirement Income Security Act of 1974 (ERISA)
focused new attention on making pension promises more explicit and
benefits more secure.  This act established guidelines for operating
pension funds and set specific requirements for various pension plan
provisions.  These new rules and subsequent regulations have raised
the costs of employer-provided defined benefit plans and have
affected the growth of defined contribution plans, including 401(k)
plans.  Under defined contribution plans, the benefits are not
certain, as they are under defined benefit plans; they vary depending
on the level of contributions made to the pension and the performance
of the investment portfolio. 

The rate of personal saving, the third source of retirement income,
is at a historic low.  Moreover, the very low rate of national saving
has raised concern among many economists about the long-term effects
on national economic growth because saving helps spur investment,
which in turn contributes to economic growth. 

Finally, the last source of retirement income, employment earnings,
may become increasingly important if the other sources fail to
provide sufficient income to meet the needs of future retirees. 
Federal policies affect such earnings in that Social Security
provisions affect when workers retire and how much they choose to
earn after retirement.  Also, federal policies regarding health
insurance, whether concerning Medicare, Medicaid, or
employer-provided retiree health insurance, influence when workers
retire. 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
---------------------------------------------------------- Chapter 1:2

The Chairman and Ranking Minority Member of the Senate Special
Committee on Aging asked us to review current and emerging issues
relating to retirement income.  First, we identified the economic and
demographic trends affecting retirement income.  Second, we examined
the status of Social Security's long-term financing problems and
proposals to address it.  Third, we examined the extent of pension
coverage and retirement saving and how to ensure that retirees can
count on them throughout retirement. 

In conducting this study, we reviewed the literature and public
record relating to retirement income policy.  We also examined data
on retirement income sources and the demographics of the elderly
population.  Our data sources included official Social Security data,
especially from the 1996 Trustees' Report, Census data, and data from
the Health and Retirement Survey, prepared by the University of
Michigan Survey Research Center.  We looked at relevant legislation
and proposals affecting retirement income.  We discussed retirement
income issues with several experts from government, research
institutes, and benefit consulting firms.  We conducted our work
between March 1996 and May 1997, using generally accepted government
auditing standards. 


DEMOGRAPHIC AND ECONOMIC FACTORS
AFFECT RETIREMENT INCOME
============================================================ Chapter 2

The U.S.  elderly population has tripled since 1940 and will more
than double again by 2050, according to Census projections, and is
accounting for a growing share of the total population.\1 As people
live longer and have fewer children, the number of workers per
retiree is declining.  Moreover, the elderly will consume an
increasing share of the national output.  A healthy rate of real
economic growth could help ease the pressures created by these
challenges, but the prospects for such growth are unclear.  Whatever
the outcome, federal retirement policies will help determine how
national income will be distributed.  In fact, the growth in Social
Security and private employer pensions since 1940 has contributed
substantially to increasing income and reducing poverty among the
elderly. 


--------------------
\1 In this report, we use "elderly" to refer to people aged 65 and
over.  However, the "retired" population is not so clearly defined;
many retire before reaching age 65 and some continue working
afterward.  We use age 65 primarily because age group data often use
age 65 as a break point. 


   THE ELDERLY POPULATION IS
   GROWING IN NUMBER AND
   PROPORTION
---------------------------------------------------------- Chapter 2:1

In the United States, the elderly population grew from about 9
million in 1940 to about 34 million in 1995, and it is expected to
reach 80 million by 2050, according to Census projections.\2
Moreover, the very old population (those aged 85 and over) is
expected to increase fivefold, from about 4 million in 1995 to nearly
19 million in 2050.  (See fig.  2.1.) As a share of the total U.S. 
population, the elderly population grew from 7 percent in 1940 to 12
percent in 1990; this share is expected to increase to 20 percent by
2050.  Other nations, both developed and developing, are experiencing
similar and often more pronounced trends. 

   Figure 2.1:  Population Aged 65
   and Over, by Age Group,
   1940-2050

   (See figure in printed
   edition.)

Note:  Data for 2000-2050 are midrange Census projections. 

Source:  U.S.  Bureau of the Census, 65+ in the United States
(Washington, D.C.:  1996). 

Although the baby-boom generation will contribute heavily to the
growth of the elderly population, other demographic trends are also
important.  Increasing life expectancy, for example, is another major
factor.  Figure 2.2 shows that life expectancy has increased
continually since the 1930s and that further improvements are
expected.  In 1940, the life expectancy of persons at age 65 was 12
years for men and 13 years for women.  By 1995, life expectancy at
aged 65 improved to 15 years for men and 19 for women and, by 2040,
it is projected to be 17 years and 21 years, respectively, according
to SSA's intermediate actuarial assumptions.\3

   Figure 2.2:  Trends and
   Projections of Life Expectancy
   at Age 65 , 1940-2060

   (See figure in printed
   edition.)

Note:  Projections for 2000-2060 are based on the trustees'
intermediate actuarial assumptions. 

Source:  1996 Annual Report of the Board of Trustees of the Federal
Old Age and Survivors Insurance and Disability Insurance Trust Funds. 

A falling fertility rate is the other principal factor underlying the
growth in the elderly's share of the population.  Fertility rates
rose from about 2.2 children per woman in 1940 to a peak of about 3.6
children per woman around 1960.  Since then, the rate has declined to
around 2.0 children per woman today and is expected to level off at
about 1.9 by 2020, according to SSA's intermediate assumptions. 

The dependency ratio, the number of working-age adults (aged 20 to
64) divided by the number of elderly adults, illustrates the
society's increasing burden of supporting the elderly.  The
dependency ratio declined from 5.8 in 1960 to 4.7 in 1995 and is
expected to decline further to 2.7 in 2040.  Looking just at Social
Security, there were 3.3 workers for each aged or disabled
beneficiary in 1995, but by 2030, there are expected to be only 2.0
workers for each beneficiary.  (See fig.  2.3.)

   Figure 2.3:  Historical and
   Projected Dependency Ratios,
   1960-2040

   (See figure in printed
   edition.)

Note:  Here, the "aged" dependency ratio is the number of people aged
20-64 per aged person (aged 65 and over).  The Social Security
dependency ratio is the number of workers paying Social Security
taxes per aged or disabled beneficiary.  Projections use SSA's
intermediate actuarial assumptions. 

Source:  1996 Annual Report of the Board of Trustees of the Federal
Old Age and Survivors Insurance and Disability Insurance Trust Funds. 

In the future, there will be relatively fewer younger persons to work
and support a growing number of aged retirees unless retirement
patterns change.  In particular, there will be relatively fewer
workers to pay the Social Security taxes needed to fund benefits.\4

At the same time, Social Security, pensions, and other retirement
income will have to provide income over longer periods as life
expectancies rise.  As a result, contributions to Social Security and
to pension plans must increase or benefits must be reduced or both. 
Yet increasing payroll taxes or constraining wage growth could make
it harder for workers to save for their own future, while cutting
benefits will increase their need to save. 


--------------------
\2 In its population projections, Census uses alternative assumptions
of high, middle, and low rates of fertility, mortality, and
immigration.  These projections reflect the middle assumptions set. 
Some demographers project even more dramatic growth in the elderly
population than Census or Social Security actuaries do.  In
particular, the 1994-96 Social Security Advisory Council's Technical
Panel on Assumptions and Methods noted that Social Security's
mortality assumptions reflect a lower rate of mortality improvements
than may be warranted. 

\3 For the annual report of the Board of Trustees for the Social
Security Trust Funds, SSA actuaries project future revenues and
benefits.  For these projections, they use alternative assumptions
regarding economic and demographic trends, including average
earnings, mortality, fertility, and immigration.  The intermediate
assumptions represent the board's best estimate of future trends. 

\4 Social Security benefits are expected to increase from 4.7 percent
of gross domestic product today to 6.4 percent in 2030, according to
SSA's intermediate projections. 


   ECONOMIC TRENDS AFFECT EARNINGS
   AND FUTURE RETIREMENT INCOME
---------------------------------------------------------- Chapter 2:2

Our nation's ability to meet its retirement income challenges depends
substantially on how well the economy performs.  Retirement
contribution rates and benefit levels mostly affect how the nation's
output--its economic pie--will be divided.  High rates of economic
growth would increase the size of the pie.  In particular, growth in
inflation-adjusted wages (real wages) affects all types of retirement
income.  Earnings largely determine Social Security and pension
benefit amounts and affect how much workers can or want to save.  In
addition, real wage growth increases the ability and possibly the
willingness of workers to absorb any tax increases or benefit cuts
that might be necessary for Social Security's long-term financial
balance. 

Unfortunately, real wage growth has slowed over the last two decades,
largely reflecting slowing productivity growth.  This contrasts with
the 25 years after World War II when the standard of living doubled,
helping improve the economic status of today's elderly relative to
that of previous generations.  However, various measures of real wage
growth exist.  Measures that incorporate fringe benefits, such as
pensions, and that use different inflation estimates suggest that
real compensation has slowed but not necessarily stagnated.\5

As long as real wage growth is positive, living standards do improve. 
In fact, a Congressional Budget Office study reports that both the
real incomes and wealth of the baby-boom generation are notably
better than those of their parents at a similar age.\6

Positive real wage growth, increasing labor force participation by
women, and changes in ERISA rules in 1986 that increased the number
of workers who will ultimately receive a pension all contribute to
the prospect of higher inflation-adjusted retirement incomes for the
baby-boom generation. 

Income inequality, which has implications for the distribution of
retirement income, has widened in recent years.  For example, in
1979, of male workers 25 years old and older, the top tenth earned
3.2 times more than the bottom tenth, but by 1993, the top tenth was
earning 4.1 times as much as the bottom tenth.  Explanations for this
include, among others, an increased demand for a relatively small
number of highly skilled workers as a result of technological change. 
At the same time, the part of the service sector that employs
relatively low-skilled workers has also grown.  For persons at the
low end of the income distribution, their retirement incomes will be
lower, and more of them might qualify for means-tested income support
and other programs available to the elderly poor. 


--------------------
\5 Daniel Sullivan, "Trends in Real Wage Growth," Chicago Fed Letter,
No.  115 (Mar.  1997). 

\6 Congressional Budget Office, Baby Boomers in Retirement:  An Early
Perspective, (Washington, D.C.:  Sept.  1993). 


   THE INCOME OF THE ELDERLY HAS
   IMPROVED
---------------------------------------------------------- Chapter 2:3

Since 1940, Social Security, pensions, and savings have dramatically
improved the income of the elderly.  Accordingly, poverty rates for
the elderly have declined substantially, but pockets of poverty
remain.  Social Security provides a strong foundation of retirement
income, but pensions, savings, and earnings determine which
households will be relatively better off in their retirement years. 
Figure 2.4 shows the increase in total income and Social Security
benefits for the elderly relative to the poverty threshold.  Because
the threshold varies by household composition, this figure uses
unmarried persons aged 65 and over only as an example.\7

   Figure 2.4:  Median Income
   Levels and the Poverty
   Threshold, Unmarried Persons
   Aged 65 and Over, 1978-94

   (See figure in printed
   edition.)

Source:  Susan Grad, Income of the Population 55 and Over
(Washington, D.C.:  SSA, Office of Research and Statistics, 1981-96). 


--------------------
\7 About 60 percent of elderly households were nonmarried persons. 
Of these, 77 percent were nonmarried women.  In contrast to
nonmarried persons, married couples received median Social Security
benefits that exceeded the poverty line in each year and by an
increasing amount.  This reflects the lower poverty rates of married
couples. 


      SOURCES OF RETIREMENT INCOME
-------------------------------------------------------- Chapter 2:3.1

In 1994, about 91 percent of all elderly households received Social
Security benefits, 67 percent received some income from saved assets,
just over 40 percent received pensions, and 21 percent received
earned income.  (See fig.  2.5.) The percentage of elderly persons
receiving Social Security benefits has remained stable over the past
20 years while the share receiving income from assets and pensions
has increased by about 10 percentage points each.  The percentage
receiving income from earnings has fallen slightly, reflecting the
reduced labor force participation of elderly men. 

   Figure 2.5:  Share of Elderly
   Households Receiving Various
   Types of Income, 1994

   (See figure in printed
   edition.)

Source:  Susan Grad, Income of the Population 55 or Older, 1994
(Washington, D.C.:  SSA, Office of Research and Statistics, 1996). 

Social Security benefits contribute the most to the elderly's
aggregate cash income, accounting for 42 percent.  (See fig.  2.6.)
The three other sources contribute about 18 percent each to aggregate
elderly income, even though the share of elderly households receiving
each type of income varies considerably.  Their contributions to
aggregate income have fluctuated.  For example, savings have ranged
between 18 and 28 percent of total elderly income since 1978. 

   Figure 2.6:  Share of Elderly
   Households' Income by Types of
   Income, 1994

   (See figure in printed
   edition.)

Source:  Susan Grad, Income of the Population 55 or Older, 1994
(Washington, D.C.:  SSA, Office of Research and Statistics, 1996). 

While Social Security provides a strong foundation for retirement
income, it is only a foundation.  In 1994, Social Security provided
an average of roughly $9,200 to all elderly households.  Figure 2.7
and table 2.1 show the dollar contributions from each source of
elderly income by income level for 1994.  Social Security's
contribution plateaued at roughly $10,000 in the highest three
quintiles.  Social Security provided 81 percent of the aggregate
income for each of the bottom two fifths of the income distribution
but just 23 percent for the top fifth.  The second largest source of
income for those in the bottom fifth was public assistance, which
provided about 11 percent of their total income.  Pensions, savings,
and earnings determine which households have the highest retirement
incomes.  They each contributed an average of more than 20 percent of
income to the top fifth of the income distribution but less than 4
percent to the lowest fifth.  Saved assets, for example, provided
income to more than 90 percent of elderly households in the top fifth
but to only 30 percent in the lowest fifth. 

   Figure 2.7:  Pensions, Savings,
   and Earnings Determine Who Will
   Have Highest Retirement
   Incomes, 1994

   (See figure in printed
   edition.)

Note:  Median incomes for each quintile are GAO estimates.  Social
Security income for the highest fifth may be lower than for the
previous fifth because, among other possible reasons, some elderly
workers or their spouses may not yet be collecting benefits. 

Source:  GAO analysis of data from Susan Grad, Income of the
Population 55 and Older, 1994 (Washington, D.C.:  SSA, Office of
Research and Statistics, 1996). 



                         Table 2.1
          
            Elderly Households' Median Income by
               Types of Income and by Income
                     Distribution, 1994

                      Income level (quintile)
          ------------------------------------------------
             First    Second     Third    Fourth     Fifth
--------  --------  --------  --------  --------  --------
Social      $4,600    $7,936    $9,934   $11,401   $10,215
 Security
Savings        153       528     1,553     3,399    10,980
Pensions       181       724     2,126     5,642     9,495
Earnings        11       215       935     2,573    12,825
Other          720       382       528       590     1,485
==========================================================
Total       $5,665    $9,785   $15,075   $23,605   $45,000
----------------------------------------------------------
Note:  Median incomes for each quintile are estimates.  Social
Security income for the highest fifth may be lower than for the
previous fifth because, among other possible reasons, some elderly
workers or their spouses may not yet be collecting benefits.  Totals
may not add because of rounding. 

Source:  GAO analysis of data from Susan Grad, Income of the
Population 55 and Older, 1994 (Washington, D.C.:  SSA, Office of
Research and Statistics, 1996). 


      THE POVERTY OF THE ELDERLY
      HAS DECLINED, BUT POCKETS
      REMAIN
-------------------------------------------------------- Chapter 2:3.2

Poverty rates for the elderly declined dramatically from 35 percent
in 1959 to under 11 percent in 1995.  (See fig.  2.8) In 1994, 11.7
percent of persons aged 65 and over were poor compared with 11.9
percent of adults aged 18-64 and 21.2 percent of children under 18.\8

   Figure 2.8:  Poverty Rates by
   Age Group, 1959-94

   (See figure in printed
   edition.)

Note:  Estimates for adults 65+ and adults 18-64 are not available
for 1960-65. 

Source:  U.S.  Bureau of the Census. 

Social Security has contributed substantially to reducing poverty
among the elderly.\9

Excluding Social Security benefits, the incomes of about 54 percent
of persons aged 65 or older would have been below the poverty
threshold in 1994.\10 Still, even with various government benefits,
almost 30 percent of elderly households are poor or near-poor; here,
the "near-poor" are persons with incomes between 100 and 150 percent
of the poverty threshold. 

Poverty tends to affect particular subgroups of the elderly
population disproportionately.  For persons 75 and older in 1992, 35
percent were poor or near-poor, roughly the same percentage as for
children.  Poverty status also varies significantly by gender and
marital status.  Single, widowed, or divorced women are much more
likely to live in poverty than couples or unmarried men.  About 22
percent of unmarried women aged 65 or older are poor, compared with
15 percent of unmarried men and 5 percent of married couples. 
Unmarried women make up over 70 percent of poor elderly households. 


--------------------
\8 The poverty threshold in 1994 was $8,967 for a two-person elderly
household with no related children.  The poverty index is based
solely on money income and does not reflect noncash benefits such as
food stamps, Medicaid, and public housing. 

\9 Means-tested public assistance programs such as Supplemental
Security Income, Food Stamps, Medicaid, and housing programs also
attempt to alleviate poverty for the elderly and others.  While these
programs are beyond the scope of this report, they are an integral
part of income support policies for the poor, including the elderly. 
Changes in retirement income policy implicitly affect these programs
as well as the poverty of the elderly. 

\10 However, without Social Security, people might save more or
continue working. 


      HEALTH COSTS DEMAND A
      GROWING SHARE OF RETIREMENT
      INCOME
-------------------------------------------------------- Chapter 2:3.3

Health care costs are substantially greater for the elderly than for
others and place growing demands on their incomes, even with public
and private health care insurance.  Out-of-pocket health costs
consumed 21 percent of elderly household income in 1994, according to
an analysis of the 1987 National Medical Expenditure Survey and data
from other government and insurance industry sources.\11 The 1994
values were projected from the 1987 survey data, using trend data
from other sources.  The elderly pay 42 percent of their
noninstitutional health costs, not including premiums.\12 Including
premiums, the elderly's out-of-pocket health costs in 1994 were
projected to average $2,519 per person, which is nearly four times
greater than for the nonelderly.  After adjusting for inflation, the
elderly spent more than twice as much in 1991 on out-of-pocket costs
as they did before Medicare was enacted in 1965. 


--------------------
\11 Out-of-pocket expenses include premiums, deductibles, and
copayments as well as costs not covered by insurance at all. 
However, the data presented here do not include Medicare Part A
insurance premium contributions.  AARP Public Policy Institute and
the Urban Institute, Coming Up Short:  Increasing Out-of-Pocket
Health Spending by Older Americans, No.  9507 (Washington, D.C.: 
Apr.  1995). 

\12 In the case of long-term care, consumers paid 33 percent of the
costs out-of-pocket in 1993. 


SOCIAL SECURITY'S FINANCING
CHALLENGES
============================================================ Chapter 3

Social Security plays a major role in providing retirement income. 
It pays benefits to more than 90 percent of elderly households and
provides more than 40 percent of total elderly income.  However, the
elderly population is growing in both number and proportion, and as a
result, the program faces serious long-term financing problems. 
Because of previous efforts to address long-term financing issues,
the Social Security trust funds are building up substantial reserves
to help pay future benefits.  These reserves are invested in
government securities.  However, drawing down the trust funds will
have significant implications for the federal budget when the time
comes to do so. 


   ADDRESSING SOCIAL SECURITY'S
   LONG-TERM FINANCING PROBLEMS
---------------------------------------------------------- Chapter 3:1

Although Social Security's revenues currently exceed its
expenditures, revenues are expected to be about 14 percent less than
total projected expenditures over the next 75 years.  In 2031, the
last members of the baby-boom generation will reach age 67, when they
can receive full retirement benefits under current law.  In 2029, the
Social Security trust funds are projected to be depleted.  In 2030
and each year thereafter, Social Security revenues are expected to be
sufficient to pay only 70 to 77 percent of benefits, given current
law and SSA's intermediate assumptions about demographic and economic
trends. 

Restoring Social Security's long-term financial balance will require
some combination of increased revenues and reduced expenditures.\13 A
variety of options is available within the current structure of the
program.  However, some proposals would fundamentally alter the
structure of the Social Security program.  These more dramatic
changes go beyond restoring financial balance and attempt to achieve
other policy objectives as well. 

In addition to ensuring program solvency, a variety of policy
objectives underlie Social Security's program structure or proposals
to reform it, such as

  -- helping ensure that the elderly do not live in poverty;

  -- promoting benefit "equity"--that is, reducing the variation in
     the implicit rates of return that participants receive on their
     contributions;

  -- giving individuals greater control and responsibility for their
     retirement income;

  -- increasing personal and national saving; and

  -- reducing future federal budget deficits. 

In addition, financing reforms could affect the nation's economy in
various ways.  For example, increasing national saving or reducing
tax burdens could promote economic growth.  Reforms that would invest
Social Security revenues in the stock market or require increased
federal borrowing could affect stock prices and bond interest rates. 
Raising the retirement age could affect the labor market for elderly
workers. 

Reforms would have effects on other sources of retirement income and
related public policies as well.  For example, increasing payroll
taxes could affect the ability of workers to save for retirement. 
Raising Social Security's retirement age or cutting its benefit
amounts could increase costs for private pensions that adjust
benefits in relation to Social Security benefits.  They would also
interact with other income support programs such as Social Security's
disability insurance or the Supplemental Security Income public
assistance program. 

Reforms could have both immediate effects and effects far into the
future.  For example, bringing newly hired state and local government
workers into the Social Security system would immediately increase
revenues but would increase benefit payments only when the newly
covered workers retire.  However, even changes that take effect years
from now can affect how workers plan for their retirement now,
especially how much they choose to save.  Therefore, the sooner
solutions are enacted, the more time workers will have to adjust
their retirement planning.  Acting sooner rather than later also
means that the funding shortfall can be addressed over a longer
period at a lower annual cost. 

Finally, any financing reforms would implicitly have distributional
effects.  For example, increasing Social Security taxes would reduce
the disposable income of current workers but would help sustain
retirement benefits for current retirees and possibly for themselves
when they retire.  Cutting benefits instead of increasing payroll
taxes would have the opposite distributional effect.  Also, Social
Security redistributes income from high to low earners to some
degree; some reforms would change this redistribution. 


--------------------
\13 Decreasing the administrative costs of the program will not save
much money.  Current administrative expenditures are less than 1
percent of program expenditures for Old Age and Survivors Insurance. 


      OPTIONS WITHIN SOCIAL
      SECURITY'S CURRENT STRUCTURE
-------------------------------------------------------- Chapter 3:1.1

A wide range of options is available for reducing costs or increasing
revenues within the current structure of the Social Security program. 
Previously enacted reforms have used many of these in some form. 
Current reform proposals also rely, at least in part, on many of
these more traditional measures, regardless of whether the proposals
largely preserve the current program structure or alter it
significantly. 

Examples of reducing program expenditures include

  -- reducing initial benefits by changing the benefit formula for
     all or some beneficiaries;

  -- raising the retirement age, which implicitly reduces initial
     benefits;

  -- lowering or eliminating the annual COLAs; and

  -- means-testing benefits. 

Examples of increasing revenues include

  -- increasing income taxes on Social Security benefits,

  -- increasing Social Security payroll taxes, and

  -- investing trust funds in the stock market. 


         REDUCING INITIAL BENEFITS
         THROUGH THE BENEFIT
         FORMULA
------------------------------------------------------ Chapter 3:1.1.1

One way to reduce benefits would be to modify the formula used to
determine benefit amounts for each retired or disabled recipient. 
Determining benefit amounts starts by computing a measure of average
lifetime earnings.  For each year, the worker's covered earnings are
updated, or indexed, to reflect the change in average earnings for
the national economy.\14 For retired worker benefits, the average is
computed using the highest 35 years of indexed earnings.  Some
propose increasing the computation period to 38 years.\15 Because the
current formula uses the years with the highest earnings, additional
years would be those with lower earnings and would thereby lower the
average.  This change would likely decrease benefits more for women
than for men, because women usually have more years with zero
earnings. 

In addition, the benefit formula provides that benefits are higher
for workers with higher lifetime earnings but less than
proportionately so.\16 This "progressive" formula redistributes
income from high earners to low earners to help keep low earners out
of poverty.  At the same time, the formula attempts to maintain some
degree of equity for higher earners by providing that benefits
increase somewhat with earnings.  The specific parameters of the
formula implicitly reflect a particular balance between these
adequacy and equity concerns. 

Modifying the formula could include changing the parameters that
determine the progressivity of the formula.  Such changes could
reduce benefits for high earners but leave low earners' benefits
unchanged.  However, this could raise equity concerns because high
earners would get lower benefits relative to their earnings than they
do now.  Alternatively, parameter changes could reduce benefits
across the board, which could raise concerns that persons with the
lowest benefits might not get the retirement income they need. 

In addition, initial benefits could be reduced by changing the
adjustments for early or delayed retirement.  Such changes should
increase the incentive to postpone retirement and continue working. 


--------------------
\14 Noncovered earnings include any over the maximum taxable earnings
($65,400 per year for 1997) and any earnings in noncovered
employment.  Changes in average earnings reflect changes in real
wages as well as wage inflation.  Therefore, this indexing captures
some of the improvement in the standard of living. 

\15 For example, the Ball proposal included in the recent Social
Security Advisory Council Report.  (See "Current Proposals Mix
Several Elements.)

\16 Specifically, the "primary insurance amount" (PIA) is the full
monthly benefit payable to retired workers at age 65 or to disabled
workers when first entitled.  For those entitled to benefits in 1997,
the PIA equals (1) 90 percent of the first $455 of average indexed
monthly earnings (AIME), plus (2) 32 percent of the next $2,286 of
AIME, plus (3) 15 percent of AIME over $2,741.  The bend points in
this formula (dollar amounts of AIME defining each bracket) are
indexed to increases in average national earnings. 


         RAISING THE RETIREMENT
         AGE
------------------------------------------------------ Chapter 3:1.1.2

Raising the retirement age would reduce benefit costs and could
involve increasing the normal retirement age (NRA), the early
retirement age (ERA), or both.  The appropriate age for retirement
has arguably changed because life expectancy has increased and the
health of the elderly has improved.  However, these improvements have
been enjoyed primarily by those with higher education and
socioeconomic status. 

Under current law, persons who retire before the NRA receive reduced
benefits; those who retire after the NRA receive increased benefits. 
Therefore, raising the retirement age implicitly has the effect of
reducing initial benefits for all retirees.  For example, under
current law, persons who retire at today's NRA of 65 would get the
basic benefit amount, without adjustments for either early or delayed
retirement.  As the NRA increases to age 67, persons who still retire
at age 65 will have their basic benefits reduced for early
retirement.  Those in the same age group who retire at other ages
would see larger reductions for early retirement or smaller increases
for delayed retirement than they would under current law. 

Raising the ERA as well as the NRA would eliminate retired worker
benefits entirely until workers reach the new ERA.  More than 50
percent of newly retired workers currently receive benefits at age
62, and more than two thirds retire before age 65.  Therefore, the
cost savings from increasing the ERA would be substantial.  However,
some workers who would have retired before age 65 would still qualify
for Social Security under the Disability Insurance program. 

Reforms relating to the retirement age could provide incentives to
work longer, but their effects also depend to some degree on the
labor market's response.  Having people work longer would help with
the demographic problem of the declining ratio of workers to
retirees.  Working longer could increase workers' Social Security and
pension benefits and allow them to increase their savings rather than
spend them.  Still, it remains unclear whether workers will want to
work longer, whether they will need to because of benefit reductions,
and whether employers will provide attractive opportunities to
continue working rather than incentives to retire early.  In recent
years, workers have been retiring earlier, not later, and not always
by choice.  Less than one sixth of men aged 65 and over are in the
labor force today, compared with nearly half in 1950. 


         REDUCING COLAS
------------------------------------------------------ Chapter 3:1.1.3

Since 1975, Social Security benefits have been automatically
increased to keep pace with inflation and maintain their purchasing
power.  However, some believe that the CPI, which is used to
determine COLAs, overstates the true inflation rate.  A recent report
estimated that the CPI exceeds the true inflation rate by about 1
percentage point.\17 If the CPI overstates inflation, then Social
Security COLAs increase benefits too much.  Moreover, COLAs are
costly.  Social Security currently pays about $350 billion in
benefits; therefore, each 1-percent COLA raises outlays $3.5 billion. 
COLA increases and any errors in them are cumulative, with each
year's adjustment compounding the previous years' adjustments.  For
the same reason, however, COLA reductions would have a compound
effect on benefits.  Many retirees rely on Social Security COLAs to
help keep up with inflation because most pensions have no inflation
adjustments or only partial ones. 

Proposals to modify the COLA include the following:  making technical
adjustments to the CPI itself; limiting COLAs, for example, by using
the CPI minus 1 percentage point or capping the COLA at, say, 2.5
percent; delaying the COLA for a specified number of years or until
cumulative inflation reaches, say, 5 percent since the last
adjustment; eliminating the COLA; or limiting a full COLA to persons
with relatively low benefits. 


--------------------
\17 Advisory Commission to Study the Consumer Price Index, "Toward a
More Accurate Measure of the Cost of Living," Final Report to the
Senate Finance Committee, Dec.  1996. 


         MEANS-TESTING BENEFITS
------------------------------------------------------ Chapter 3:1.1.4

Means-testing Social Security benefits would also reduce program
costs and direct benefits to persons who need them most.  It would
reduce benefits for those with incomes above a certain threshold or
from specified sources, such as other pensions.  However, persons
losing benefits would tend to be those who pay the highest Social
Security taxes and already implicitly receive the lowest rates of
return on their contributions.  Means-testing would further reduce
benefit equity for them and could diminish whatever political support
they give to the system.  Means-testing could raise perceptions of
Social Security as a welfare program rather than a program that
ensures a basic retirement income to persons who work and contribute
to the system their entire working lives. 

Means-testing could also be applied specifically to dependents'
benefits.\18 Social Security provides benefits for spouses and
children of retired, disabled, and deceased workers.  For example,
eligible spouses currently receive a benefit based on half the
worker's basic benefit, regardless of the worker's benefit amount. 


--------------------
\18 In 1995, about 43 million people were receiving Social Security
benefits, of which 3.3 million were spouses, 5.5 million were widows
or widowers, and 3.7 million were children.  Half of children
beneficiaries were children of deceased workers and more than a third
were children of disabled workers.  About 27 million of the
beneficiaries were retired workers, and more than 4 million were
disabled workers. 


         INCREASING INCOME TAXES
         ON SOCIAL SECURITY
         BENEFITS
------------------------------------------------------ Chapter 3:1.1.5

In a sense, income taxes on Social Security benefits already provide
a limited means test.  Social Security benefits are subject to income
taxes if current income exceeds $25,000 for a single person or
$32,000 for a married couple.  Also, taxes currently apply to only a
portion of benefits, up to 85 percent of them, depending on adjusted
gross income.  Modifying these provisions further, by making 100
percent of benefits subject to income tax above certain income
levels, for example, would have an effect similar to a means-tested
benefit reduction. 


         INCREASING REVENUES FROM
         SOCIAL SECURITY PAYROLL
         TAXES
------------------------------------------------------ Chapter 3:1.1.6

The Social Security payroll tax provides about 90 percent of the
system's revenue, and increasing the payroll tax rate is the most
obvious way to increase revenues.  However, for many people, the
Social Security tax is already the largest tax they pay.  According
to SSA's intermediate actuarial assumptions, increasing the payroll
tax rate from 12.4 to 14.6 percent for 1996 and subsequent years
would have restored financial balance for the next 75 years.\19
Increasing the payroll tax rate would cause all covered workers and
their employers to pay more taxes, but workers would receive no
increase in benefits from paying these additional taxes.  Also, by
reducing workers' disposable income, such tax increases would make it
more difficult for some workers to save. 

Enlarging the Social Security tax base would also increase payroll
tax revenues.  For example, the maximum earnings subject to the
payroll tax are $65,400 per year in 1997.\20

Increasing the earnings subject to the payroll tax would increase
taxes for a relatively few high earners (about 6 percent of all
workers).  In addition, about 4 percent of the workforce remains
uncovered, which mostly includes some state and local government
employees and federal employees hired before 1984.  Bringing new
state and local workers into the Social Security system would expand
the payroll tax base.  However, expanding the tax base also increases
future benefit payments because the newly covered earnings would be
included in benefit computations. 


--------------------
\19 However, each year's actuarial valuation covers a new 75-year
period.  Therefore, long-term actuarial balance can change from year
to year. 

\20 The maximum earnings subject to Social Security payroll taxes is
updated automatically each year in proportion to increases in the
average annual wage. 


         INVESTING TRUST FUNDS IN
         THE STOCK MARKET
------------------------------------------------------ Chapter 3:1.1.7

Investing some portion of the Social Security trust funds in the
stock market rather than in government securities might increase
system revenues but has potential drawbacks as well.  Historically,
stock investments have earned higher returns on the average than
Treasury securities, but they also carry higher risks.  Nevertheless,
investing in stocks could increase program revenues without raising
taxes.  However, Social Security would no longer buy as many Treasury
bonds, so Treasury would have to sell more in the private capital
market. 

As a whole, the federal government would be selling lower-yielding
government bonds and buying higher-yielding private securities.  In
turn, such government activity could affect prices of stocks and
their expected returns and interest rates.  Also, as partial owners
of a company, stockholders can vote on company policies, which raises
the question of whether and how the government would exercise its
voting privileges.  While the funds could be invested in "passively
managed" portfolios, political pressures could also lead the
government to select or omit particular stocks for investment. 


      PROGRAM RESTRUCTURING SERVES
      OTHER OBJECTIVES AS WELL AS
      LONG-TERM BALANCE
-------------------------------------------------------- Chapter 3:1.2

A variety of proposals would address Social Security's long-term
funding problems by significantly restructuring the program, usually
by privatizing at least a portion of it.\21

Such proposals still essentially achieve financial balance by, in
effect, raising revenues and reducing costs, but they do so in ways
that pursue other objectives as well.  Some would reduce the role of
Social Security and the federal government in providing retirement
income and give individuals greater responsibility and control over
their own retirement.  Some proposals focus on trying to increase
national saving and funding future Social Security benefits in
advance. 

Generally, privatization proposals focus on setting up individual
retirement savings accounts and requiring workers to contribute to
them.  The accounts usually replace a portion of Social Security,
whose benefits would be reduced to compensate for revenues diverted
to the savings accounts.  Some combine new mandatory saving and
Social Security benefit cuts, hoping to produce a potential net gain
in retirement income.  The mandated savings deposits and revised
Social Security taxes combined would be greater than current Social
Security taxes, in most cases. 

Virtually all proposals addressing long-term financing issues would
increase the proportion of retirement assets invested in the stock
market or in other higher-risk, higher-return investments.  The
common objective is to finance a smaller share of retirement costs
with worker contributions and more of the costs with investment
returns. 

The federal government's Thrift Savings Plan (TSP) for federal
workers exemplifies how individually held accounts could be invested
in the private sector with a limited government role.  Under TSP,
vested federal workers own their savings accounts and choose how to
invest them in a range of passively managed funds, which include
funds invested in stocks, corporate bonds, and government bonds.  The
government administers these funds through contractors and delegates
proxy voting to them.\22

In the case of individually managed savings accounts, the risk of
economic and market performance would be shifted to the worker. 
Individuals with identical earning histories and retirement
contributions could have notably different retirement incomes because
of market fluctuations or individual investment choices.  Some
observers have expressed concern that many workers lack the knowledge
necessary to make the best investment decisions.  Many workers,
especially women, invest too large a portion of their savings in
relatively safe, fixed-income securities that earn a relatively low
return and shy away from more "risky" equity investments that
generally earn a higher average return.  Proponents of privatization
respond that investor education programs can teach workers how to
manage their investment portfolios better.  Limited research
indicates that investor education is effective at increasing both
workers' retirement contributions and their investment returns. 
Finally, care must be taken to ensure that retirement savings would
last until the retiree dies.  Some proposals would require retirees
to purchase a lifetime annuity with their retirement savings. 

To the extent that privatization requires workers to increase their
retirement saving, it could increase national saving.  However,
workers might reduce other saving either because they would have less
disposable income from which to save or because they would, in
effect, let the new retirement accounts substitute for their other
saving.  In cases in which privatization largely involves redirecting
the Social Security trust funds' balances into the stock market,
whether government directs the investment or not, national saving
would probably not increase significantly.  Any increase in private
sector investment could be offset, at least temporarily, by
Treasury's need to borrow money from the private sector to replace
the funds it currently borrows from the trust funds. 

Some of the privatization proposals raise the issue of how to make
the transition to a new system.  Currently, each year's Social
Security taxes are used for the most part to pay for that year's
benefits.  Financing retirement through individually owned savings
accounts requires "advance funding," or saving this year's
contributions to provide future retirement income.  Social Security
would still have to have revenues to pay benefits that retirees and
current workers have already earned.  The revenues needed to fund
both current and future liabilities would clearly be higher than
those currently collected. 

Privatization would also have a significant effect on the
distribution of retirement income between high and low earners,
although advocates claim that all would still be better off.  As
described above, the Social Security benefit formula redistributes
income from high to low earners and tries to balance the program's
goals of providing equitable benefits while also meeting basic income
needs.  To the extent that privatization involves workers'
contributing to their own retirement saving, their contributions are
not available for redistribution.  Some privatization proposals
retain some degree of Social Security coverage or at least a
means-tested safety net and therefore permit some redistribution to
continue.  In effect, such proposals separate the program's equity
and adequacy goals. 

Privatization proposals also tend to separate retirement benefits
from Social Security's survivors' and disability benefits.  In the
cases of death or disability before retirement, individual savings
may not have been building long enough to sufficiently replace lost
income.  Some privatization proposals, therefore, leave these social
insurance programs largely as they are now.  In the case of death
after retirement, the surviving spouse and dependents may still
depend on financial support from the retirement savings of the
deceased.  Therefore, some privatization proposals would require
retired workers to purchase a joint and survivor annuity with their
retirement savings or obtain spousal consent for the choice of a
single life annuity.  Joint and survivor annuities provide lower
annual benefits but continue to pay benefits to the surviving
spouses. 


--------------------
\21 Chile is often cited as an example of how privatization can work. 
Other countries also have planned or are planning to privatize their
social security systems.  However, their macroeconomic situations and
political institutions often differ substantially from ours.  For
example, Chile is a much less developed economy than the United
States, and when it privatized, it had limited capital markets and a
budget surplus. 

\22 However, the size of such funds under various reform proposals
would be far greater than under TSP and, thus, the voting issue would
be magnified. 


      CURRENT PROPOSALS MIX
      SEVERAL ELEMENTS
-------------------------------------------------------- Chapter 3:1.3

The 1994-96 Advisory Council on Social Security recently issued a
report in which its members offered three alternative reform
proposals.  While the council could not achieve majority support in
favor of any one proposal, it did appear to reach agreement on
certain points.  Some individual components appeared in more than one
of the three proposals: 

  -- extending mandatory coverage to all state and local government
     workers hired after 1997;

  -- assume that revisions to the CPI, which were announced by the
     Bureau of Labor Statistics in March 1996, will lower the COLAs
     for future benefits by 0.21 percentage points;

  -- increase the number of years of earnings used in computing
     benefits from 35 to 38;

  -- subject benefits to income taxes to the extent that workers'
     benefits exceed their contributions and deposit the proceeds in
     the Social Security trust funds;

  -- accelerate the increase in the retirement age, so that the NRA
     will reach 67 for persons born in 1949 instead of 1960 and
     increase it after 2011 according to increases in longevity. 

According to the National Academy of Social Insurance, these five
measures combined would eliminate more than two thirds of Social
Security's long-term financial deficit and postpone the exhaustion of
the trust funds from 2030 to 2052.\23

Of the three comprehensive reform proposals, the one called the
"maintain benefits" plan, generally associated with former Social
Security Commissioner Robert Ball, would make the fewest changes to
the current system.  The Ball proposal would make all the changes
above except for increasing the retirement age.  In addition, it
calls for studying the possibility of investing 40 percent of the
trust funds in the stock market by selecting a portfolio that would
track some broad market index such as the Wilshire 5000.  The
investments would be passively managed by an independent board.  In
addition to these and other changes, the Ball proposal would achieve
long-term balance by increasing Social Security taxes in 2045 by 1.6
percentage points (0.8 percentage points each for employers and
employees.)

A second proposal, called the "individual account" plan and generally
associated with Council Chairman Edward Gramlich, would make all five
of the changes above.  In addition, the Gramlich proposal would
gradually reduce benefits by modifying the benefit formula so that
eventually all benefits would be financed by the current Social
Security tax rate.  After completely phasing in this change, benefits
would be 17-percent lower for average earners, with larger reductions
for high earners and smaller ones for low earners.  Combined with
raising the NRA, the average earner's benefits would ultimately
decline about 30 percent.  To offset these benefit reductions, the
Gramlich proposal would add a mandatory individual savings account
financed by new contributions from workers of 1.6 percent of their
covered earnings, starting in 1998.  Workers would choose among
alternative investment options administered by a government board,
similar to the administration of the TSP.  At retirement, workers
would be required to purchase a lifetime annuity with the money in
these accounts. 

The third proposal, called the "personal security account" plan and
generally associated with benefits consultant Sylvester Schieber and
economist Carolyn Weaver, would make the most dramatic changes to
Social Security.  The Schieber-Weaver proposal would extend coverage,
assume cost reductions from corrections to the CPI, and accelerate
the increase in the retirement age.  In addition, it would transform
Social Security into a two-tier system.  In the first tier, a basic
flat benefit amount adjusted for years with covered earnings would be
paid to all retirees.  The amount would equal about two thirds the
current poverty level for a single elderly person, or $410 per month
in 1996.  The second tier would be a personal security account
financed by shifting 5 percentage points of the existing employee tax
into those accounts.  Workers could invest these funds in a wide
range of instruments but could not withdraw them before retirement. 
At the ERA, they could use the funds freely.  The transition to the
new system would require new taxes equaling 1.52 percent of covered
payroll, starting in 1998 and lasting 72 years.  Not insignificantly,
it would also require additional government borrowing of nearly $2
trillion in 1995 dollars during the first 40 years of the transition. 


--------------------
\23 National Academy of Social Insurance, "Advisory Council on Social
Security Plans," Social Insurance Update, Vol.  1, no.  3 (December
1996). 


   SOCIAL SECURITY TRUST FUNDS AND
   THE FEDERAL DEFICIT
---------------------------------------------------------- Chapter 3:2

As a result of previous financing reforms, Social Security collects
more in revenues than it pays out in benefits each year and builds up
substantial trust fund reserves.  This annual excess of revenues over
expenditures lowers the total federal deficit.  However, tapping
these reserves in the future to help pay benefits would pose a
substantial challenge for the overall federal budget. 

The Social Security trust funds are by law invested in federal
government securities and effectively lend money to Treasury.  By
investing in Treasury securities, Social Security reduces the amount
that the federal government must borrow in the private financial
markets.  Until 2009, Social Security is expected to collect roughly
$30 billion more in cash each year than is needed to pay benefits. 
In addition, the trust funds' interest earnings were nearly $40
billion last year and will grow as the reserves build up.  Treasury
credits this interest by issuing more securities.  If Treasury had to
replace borrowing from the trust funds with borrowing in the private
market, it would have to pay this interest in cash.  This cash could
be borrowed by selling bonds in the private capital market, but such
interest payments would further add to the overall federal deficit. 

Social Security's excess cash revenues are expected to start falling
rapidly in 2009 and to disappear in 2012, under the trustees' 1996
intermediate actuarial assumptions.  (See fig.  3.1.) Expenditures
will then exceed cash revenues, and the government's general fund
will have to make up the difference, in effect repaying Social
Security.  This will increase the budget deficit, unless offset by
revenue increases or spending reductions.  In 2028, this amount is
expected to reach about $183 billion in 1997 dollars.  In that year,
the budget deficit from Social Security alone would equal 1.4 percent
of GDP, which is the same share of GDP as last year's deficit for the
entire federal budget.  The trust funds are expected to be depleted
in 2029. 

   Figure 3.1:  Social Security's
   Revenues Exceed Expenditures
   Now but Fall Short Later

   (See figure in printed
   edition.)

Source:  1996 Annual Trustees' Report and unpublished SSA data. 

Solving Social Security's long-term financing problem will not
necessarily eliminate this budget challenge.  Some Social Security
reform proposals would diminish the trust fund build-up and
spend-down.  However, in any event, efforts to maintain a stable
fiscal policy will depend to some extent on how this spend-down is
addressed. 


PRIVATE PENSIONS AND RETIREMENT
SAVING CAN HELP WITH RETIREMENT
INCOME CHALLENGES
============================================================ Chapter 4

While Social Security provides a foundation for retirement income,
increasing private pension coverage and individual retirement saving
can contribute substantially toward meeting the income needs of the
elderly and help increase national saving at the same time.  To
fulfill their potential as major retirement income sources, both
pensions and savings must be secure and carefully managed. 

The two basic types of pension plans are defined benefit and defined
contribution plans.  A defined benefit (DB) pension plan promises the
worker a benefit based on a specific formula linked to the worker's
earnings and years of employment.  The employer, as the plan sponsor,
is responsible for funding the promised benefit, investing and
managing the funds, and bearing the investment risk.  In terms of
coverage, DB pensions were the predominant type of employer pension
for many years.  In 1975, three fourths of workers covered by a
private pension had DB plans. 

Under defined contribution (DC) plans, a percentage of pay is
contributed to an account for each worker.  While the employer
generally makes the contribution, the increasingly popular 401(k)
plans also allow contributions by workers.  Retirement income from DC
plans depends on how much money is deposited and how much the
invested funds earn.  In DC plans, the worker bears the investment
risk and often controls how the funds are invested.  Because nearly
all DC plan benefits are taken immediately as a lump sum, DC pensions
are considered to be more portable than DB plans, which frequently
pay a life annuity at retirement. 


   FEDERAL POLICIES PROMOTE
   PENSION COVERAGE AND BENEFIT
   SECURITY
---------------------------------------------------------- Chapter 4:1

The federal government has an interest in promoting pension coverage
to ensure greater retirement income security and to help increase
national saving.  Also, the government has assumed the role of
helping protect workers' legal rights to pension benefits.  The
Internal Revenue Code and ERISA define the government's role, and the
Department of Labor (DOL), the Internal Revenue Service (IRS), and
PBGC carry out the laws. 

The Internal Revenue Code provides preferential income tax treatment
for pension contributions and for the capital gains on the funds. 
These tax preferences have the greatest relevance to employees rather
than employers.  For employers, all wages are tax deductible just as
pension contributions are, but employees do benefit from the deferral
of income tax on their pension benefits until they receive them.  In
particular, the extension of preferential tax treatment to employee
contributions to 401(k) plans has resulted in the substantial growth
of these plans.  Even so, the effect of these tax preferences depends
significantly on the structure of income tax rates.  Marginal tax
rates on high and middle earners have dropped considerably since the
1970s.  The tax rate on capital gains also affects these tax
preferences because many DB and DC pension assets are invested in
equities and other financial instruments that would otherwise be
subject to the capital gains tax.  Also, tax preferences for pensions
result in substantial foregone federal revenues, and the Congress has
an interest in limiting these losses and has at times attempted to
target who benefits from them and how much.  Therefore, changes in
tax policy could have a significant effect on pension coverage. 

To qualify for preferential tax treatment, pension plans must adhere
to various standards.  For example, in 1942, the Congress enacted
"nondiscrimination" provisions to ensure that pension plans did not
receive favorable tax treatment if a disproportionate share of the
benefits accrued to company officers or other highly paid
individuals. 

In 1974, the Congress enacted ERISA to make pension promises more
explicit and workers' benefits more secure.  Among other things,
ERISA requires that private DB plan sponsors make contributions
according to actuarial standards to help ensure adequate funding.  In
addition, ERISA requires such sponsors to pay pension insurance
premiums to PBGC, which assumes the liability for terminated DB plans
and pays the retirement benefits, subject to certain limits.  ERISA
also requires that plan investments be diversified and that plan
fiduciaries adhere to prescribed standards of conduct. 


   MORE WILL RECEIVE PENSIONS
   ALTHOUGH COVERAGE HAS NOT GROWN
   RECENTLY
---------------------------------------------------------- Chapter 4:2

Pension coverage for workers in the private sector increased from 15
percent to 45 percent between 1940 and 1970.  This expansion resulted
from a variety of factors, including tax preferences, healthy
economic expansion, and the growth of labor unions and also from wage
and price controls during the 1940s and 1950s.  Despite wage
controls, employers were able to meet union requests for increased
compensation by offering pensions and other employee benefits. 
However, since 1970, the coverage rate has changed very little; it
stood at 47 percent in 1993. 

Nevertheless, in spite of stagnant pension coverage, more workers
covered under pension plans and their spouses will eventually receive
pensions.  ERISA established vesting rules in 1974 and set a 10-year
minimum before employees could vest--that is, earn a legal right to
their pension benefits.  In 1986, the rule was liberalized, allowing
either a 5-year "cliff" or a graded vesting schedule under which
participants are 20 percent vested after 3 years and receive an
additional 20 percent each subsequent year until they are fully
vested after 7 years.  As a result, more covered employees will meet
the vesting requirement, although benefit levels will reflect the
shorter job tenures of the employees affected. 

In addition, more women are working, so more retired couples will
have pension income.  Also, a larger share of widowed and divorced
retirees will receive pension income from their spouse's employers
because of spousal notification laws enacted in 1984.\24


--------------------
\24 The 1984 Retirement Equity Act made the joint-and-survivor
benefit the default option for married workers retiring on a pension. 
Under the joint-and-survivor option, a surviving spouse will continue
to receive pension benefits once the pension beneficiary dies.  To
waive this option, both the retiring worker and his or her spouse
must sign a waiver form.  Prior to this law, waivers from spouses
were not required, and few retiring workers selected the
joint-and-survivor option.  As a result, many spouses were left
without pension benefits when the retiree died first. 


      ECONOMIC CONSIDERATIONS AND
      REGULATION AFFECT PENSION
      COVERAGE
-------------------------------------------------------- Chapter 4:2.1

Employers offer pensions because they help attract and retain
valuable employees, encourage older workers to retire, and give
workers an additional incentive to perform well and promote the
ongoing success of the employer.  Workers accept pension arrangements
in place of higher wages and, in fact, unions often seek them because
workers benefit from the tax advantages and, with DB pensions, from
having employers bear the risk of long-term commitments (unless and
until benefits are distributed as a lump sum). 

However, some factors might make further increases in worker coverage
unlikely.  About 60 percent of all older (aged 40 to 60) workers are
already covered.  Moreover, employers are less likely to offer
pensions if their workforce tends to be less educated, lower-skilled,
or younger or exhibits high turnover.  Only 34 percent of those in
their twenties are covered by pensions.  Also, smaller employers are
less likely to offer pensions.  For example, only 28 percent of
workers at firms with 25-49 employees are covered, while at firms
employing more than 1,000, 67 percent are covered.  Finally, about 9
percent of the workforce is self-employed, and these workers must, of
course, provide for their own retirement income. 

Structural changes in the economy also influence employers' decisions
to provide pensions.  In 1945, 35 percent of the workforce, both
private and public, was unionized, compared with 16 percent in 1990,
and nearly 80 percent of union workers have pension coverage compared
with about 40 percent of nonunion workers.  In addition, the rising
cost and demand for health care has strained the resources that
employers have available for other employee benefits. 

Since ERISA was enacted, legislative and regulatory activity
regarding pensions has increased.  This activity provided improved
benefit security for workers but has affected the structure of
pensions.  Some of the changes have focused on limiting the federal
revenues foregone because of the favorable tax code provisions. 
However, federal regulations, especially when they change frequently,
tend to increase the cost of pensions and reduce employers' incentive
to sponsor pension plans.\25


--------------------
\25 Committee for Economic Development, Who Will Pay For Your
Retirement?:  The Looming Crisis (New York:  1995). 


      PENSION SIMPLIFICATION IS A
      FOCUS OF RECENT LEGISLATION
-------------------------------------------------------- Chapter 4:2.2

In response to concerns about regulatory burdens on employer pension
plans, recent federal initiatives have focused on pension
simplification.  These efforts seek to encourage expanded coverage by
reducing the regulatory requirements for small employers to start
pensions.  However, this regulatory relief is usually accompanied by
a required employer contribution for all employees that might not be
attractive to the small employer.  Past pension expansion efforts
along this line, such as simplified employee pensions (SEP), have not
been notably successful in raising pension plan sponsorship rates.\26

The Small Business Job Protection Act of 1996, created a new "savings
incentive match plan for employees" (SIMPLE) for firms with 100 or
fewer workers that do not already offer a pension plan.  The plan
could be based on IRA or 401(k) arrangements and, sponsors would be
exempt from certain "nondiscrimination" rules or, under certain
conditions, other rules that might otherwise deter plan sponsorship. 
An employee could contribute up to $6,000 yearly, and the employer
would have to meet a matching requirement.\27


--------------------
\26 See Private Pensions:  Changes Can Produce a Modest Increase in
Use of Simplified Employee Pensions (GAO/HRD-92-119, July 1, 1992). 

\27 For more detailed information on the provisions in the bill see
James R.  Storey, Pension Proposals for Simplification and Increased
Access (Washington, D.C.:  Congressional Research Service, July 16,
1996).  For another discussion on related proposals advanced by the
Clinton administration, see James R.  Storey and Ray Schmitt, Pension
Reform:  President Clinton's Proposed Retirement Savings and Security
Act (Washington, D.C.:  Congressional Research Service, July 10,
1996). 


      DC PENSION COVERAGE HAS
      GROWN
-------------------------------------------------------- Chapter 4:2.3

Although pension coverage on current jobs has been generally stable
overall, the number of workers participating in DC plans has grown
steadily over the past 20 years.  In 1975, 27 percent of pension plan
participants had DC plans; by 1990, 50 percent had them.\28

With this trend, more of the risk and responsibility for providing
pension income is shifting from the employer to the employee.  Recent
data and analysis suggest that, rather than representing replacement
of DB plans by DC plans, this trend mainly reflects that new or
additional pension benefits are often offered as DC plans.  Many
larger employers supplement their DB plans with DC plans.\29

In addition, the relatively faster growth of employment at smaller
firms may help explain the trend toward DC pensions, which are easier
for smaller employers to administer.  Between 1980 and 1993, the
number of employees in firms with fewer than 100 employees grew 30
percent, twice as fast as for firms with 500 or more employees.  In
1993, small firms accounted for 56 percent of all employees.  About
25 percent of full-time employees in small firms participate in DC
plans, compared with 12 percent participating in DB plans. 

The 401(k) plan is the fastest growing type of DC pension that allows
employees to make tax-deferred contributions that may be augmented by
the employer.  The growth of these accounts has contributed
substantially to the DC trend.  Between 1984 and 1990, 401(k) plans'
share of all private plans grew from 3 to 14 percent, and their share
of all active pension plan participants grew from 19 to 46
percent.\30 This growth appears to be continuing. 


--------------------
\28 Percentages are not adjusted for double counting of individuals
participating in more than one plan.  For a more recent analysis of
defined contribution plans see Private Pensions:  Most Employers That
Offer Pensions Use Defined Contribution Plans (GAO/GGD-96-181, Oct. 
3, 1996). 

\29 Employee Benefits Research Institute, "Pension Evolution in a
Changing Economy," Special Report and Issue Brief No.  141,
Washington, D.C., Sept.  1993. 

\30 401(k) participants may participate in one or more additional
plans. 


   EFFORTS TO INCREASE PERSONAL
   RETIREMENT SAVING HAVE NOT BEEN
   SUCCESSFUL
---------------------------------------------------------- Chapter 4:3

Personal savings outside pension arrangements can contribute
substantially to retirement income.  However, saving patterns vary
considerably among families.  While most families say they recognize
the need to save for retirement, many do not save in any systematic
way.  As with pension coverage, government efforts to encourage more
personal retirement saving seem to have had only marginal effects. 

According to our analysis of the University of Michigan's Health and
Retirement Survey, the average net worth of families whose heads are
nearing retirement age (from 55 to 61 years of age) is almost
$250,000.  However, the distribution of wealth is uneven, and many
families near retirement have relatively few retirement savings. 
According to our analysis, over half of these families have less than
$100,000 in assets, and 25 percent have less than $25,000 in assets. 
Over one quarter of families near retirement in this survey do not
own their own home, almost 60 percent have no IRA, over 70 percent do
not own stocks, 92 percent do not own bonds, and half have less than
$5,000 in bank accounts. 

Tax policy affects how people save for retirement, but its effect on
the overall level of personal saving is less clear.  Since the 1970s,
federal laws have granted favorable tax treatment to IRAs and to the
closely related 401(k) pension plans.  In 1981, most constraints on
IRAs were lifted so that nearly all workers and their spouses could
participate, making this form of tax-deferred saving very popular. 
After 1986, tax policy limited the group of workers who could make
tax-deductible IRA contributions.  Total contributions fell by more
than 60 percent, and the percentage of all tax returns claiming an
IRA deduction fell from 16 percent in 1985 to 6 percent in 1988. 
Recent polling evidence suggests that people would save more in IRAs
if the tax benefits were increased.  The 401(k) plan has many of the
same tax advantages as IRAs and is now the fastest growing type of
pension plan. 

Preferential tax treatment for IRAs and 401(k) accounts seems to
encourage saving in these vehicles, but they may not necessarily
represent totally new saving.  Some contributions may merely have
been shifted from other forms of saving.  Between 1970 and 1994,
personal saving rates declined from roughly 8 to 4 percent of
disposable personal income, even with these retirement saving
incentives.  Also, according to available data, most persons
responding to these tax preferences are middle- to upper-income, not
lower-income, who receive very little income from retirement savings. 
Low-income families may find it especially difficult to save, and
most public assistance programs penalize private saving by requiring
low levels of financial assets in order to qualify. 


   WILL PENSIONS AND SAVINGS BE
   THERE FOR RETIREMENT? 
---------------------------------------------------------- Chapter 4:4

To fulfill their potential as retirement income sources, pension
assets and savings must be preserved and carefully managed.  ERISA
sets out requirements regarding how pension assets must be managed,
invested, and preserved for retirement.  While tax rules discourage
using pension assets for nonretirement purposes, they do not prohibit
such use entirely.  Also, many pensions allow workers to take a lump
sum at retirement rather than receive an annuity.  If they elect to
take a lump sum, retirees must then carefully manage these assets
along with their other savings to make them last throughout their
retirement. 


      PROTECTING DB PLAN ASSETS
-------------------------------------------------------- Chapter 4:4.1

PBGC insures and guarantees DB pension benefits, and IRS monitors
plan funding status.  Funding requirements aim not only to protect
workers' pensions but also to limit PBGC's risk of assuming financial
responsibility for unfunded benefits when employers go out of
business.  Premiums for pension guaranty insurance are set so that
they increase as the extent to which the plan is underfunded
increases.  Underfunded plans may also be required to make
contributions that gradually reduce the plan's unfunded liability. 
Recently, the Retirement Protection Act of 1994 revised rules
relating to pension plan underfunding.  Along with the healthy
investment returns that pension funds have enjoyed, such reforms have
substantially improved funding levels and have reduced the financial
risk to the pension sponsors that pay the PBGC insurance premiums. 
Under current law, PBGC may borrow up to $100 million from Treasury,
but the federal government has no legal obligation to back up PBGC
otherwise. 

For well-funded plans, boosted by the strong stock market performance
of the 1980s and 1990s, the issue arises of how to deal with
overfunding.  Since firms own the pension assets, they sometimes want
to take some of the excess assets back and use them elsewhere.  The
courts have upheld plan sponsors' right to "revert" assets, and the
practice may be consistent with prudent plan funding.  Still, the
Congress has been concerned that such reversion may threaten benefit
security if the sponsors or their pension funds do poorly in the
future.  The Congress imposed substantial penalties in response to a
wave of reversions in the 1980s, largely stopping the practice.  As
recently as 1995, proposals were offered to loosen these
restrictions. 


      INVESTING AND PRESERVING DC
      PLAN ASSETS
-------------------------------------------------------- Chapter 4:4.2

With DC pensions, as well as with individual retirement saving,
workers generally have responsibility for directing how their assets
are invested.  Some invest conservatively and earn low returns even
though they still have many years before retirement and could
arguably bear the risks associated with higher-return investments. 
To help expand workers' investment choices, DOL recently issued an
interpretive bulletin regarding the 404(c) rules to clarify the kinds
of investment information firms can give their workers without being
professionally liable for giving investment advice. 

Workers also diminish their DC pension savings when they use the
funds for nonretirement purposes.  Most plans allow participants to
borrow from their accounts for nonretirement expenditures such as
purchasing a home or paying educational or emergency expenses. 
Overall, more than 75 percent of workers with DC pensions are able to
borrow from their retirement accounts.  Of these workers, almost 8
percent have outstanding loans, with an average balance of $3,000.  A
recent Clinton administration proposal would allow early,
penalty-free IRA withdrawals for selected expenses.  This added
liquidity may actually encourage more pension saving.  Still, if
workers do borrow, they may reduce the savings available for
retirement. 

At retirement, IRA and DC pensions do not have to be taken in the
form of an annuity and are generally taken as a lump-sum amount. 
Moreover, about 33 percent of DB plan sponsors, who covered more than
half of all participants, provided a lump-sum option for retiring
workers.  Annuities help insure retirees against living longer than
their retirement funds would last, but retirees pay insurers for
assuming this risk.  In contrast, without an annuity, retirees assume
much of the risk for their longevity as well as responsibility for
their wise use of the funds.  However, since the federal and state
governments provide public assistance benefits for the low-income
elderly, they also assume some of this risk. 


CONCLUDING OBSERVATIONS
============================================================ Chapter 5

Dramatic growth in our nation's elderly population raises issues not
just for the future of retirement income but also for the federal
budget and for the economy as a whole.  The elderly will consume an
increasing share of the nation's output, and unless retirement or
work patterns change, relatively fewer workers will be supporting
more retirees.  Our analysis leads us to five observations about the
future direction of federal retirement income policy. 

First, if the dominant purpose in examining Social Security is to
restore long-term financial balance to the Social Security system,
then some combination of traditional adjustments---revenue increases
or benefit reductions---could be sufficient.  However, some observers
believe that the structure of this program should be reevaluated in
light of the economic, political, and demographic changes of the past
60 years.  Even though Social Security has become the most important
single source of retirement income for many people, many believe that
now is an appropriate time to assess whether the current structure of
benefits best serves today's retirement needs. 

Second, giving people greater responsibility for their retirement
income and greater opportunity to invest their savings to earn
potentially higher rates of return might raise the general welfare of
the nation's elderly retirees.  But such potential gains must be
weighed against the potential problems associated with having
individuals bear much more of the risk in saving for retirement. 
Also, if the Social Security program is to continue to play an
important role in ensuring that the less-well-off have a basic
retirement income, then care must be taken to ensure that changes in
the Social Security program do not exacerbate the condition of
particular groups of elderly poor.  Unmarried women are currently
much more likely to have incomes below the poverty line than other
beneficiaries.  Changes such as increasing the number of years used
in computing benefits would disproportionately affect women because
they more often already have several years of zero earnings included
in the benefit calculation.  Such distributional effects will have to
be considered in any reform proposal. 

Third, while Social Security's long-term financing problem and the
federal budget deficit are different issues, the practice of using
the Social Security trust funds' surpluses to partially offset the
deficit elsewhere in government has substantially intertwined the two
issues.  The program's annual cash surplus is projected to start
falling by 2009 and to disappear entirely in 2012, under Social
Security's intermediate assumptions.  As payments to beneficiaries
begin to exceed cash receipts, Social Security will have to redeem
the Treasury certificates it now holds.  Redemptions will reach
nearly $200 billion annually in 2028 (in 1997 dollars) and will place
significant strains on the federal budget.  Moreover, some of the
proposals to privatize Social Security involve heavy additional
federal borrowing to finance the transition from a pay-as-you-go to a
partially or fully funded system.  All proposals will have to be
examined not only for their effect on Social Security but also for
their overall budgetary consequences. 

Fourth, to respond to the nation's retirement income challenges
effectively, we must also examine pensions and private savings and
the potential effect of Social Security reform proposals on them. 
For example, raising Social Security taxes could make it more
difficult for people to save for their retirement or to contribute to
their pensions.  Proposals to solve Social Security's long-term
financial problem will have to take into account the effect on
achieving our goals of encouraging greater employee participation in
pensions and increased rates of national saving. 

Finally, some of the Social Security privatization proposals focus on
increasing the national rate of saving.  If successful, these changes
could help raise our standard of living and help mitigate the strain
from dealing with Social Security's financing problem.  Increased
national saving could lead to higher rates of economic growth, which,
in turn, would make it easier to meet the financial challenges posed
by the shortfalls in Social Security.  However, raising the national
saving rate may prove to be difficult.  The proposals that aim to
increase national saving through creating individual retirement
savings accounts cannot guarantee that these savings will not simply
substitute for other forms of saving.  In addition, some of the
proposed changes to the nation's retirement system could have
significant effects on the nation's equity and bond markets, and this
will have to be evaluated before such changes are adopted. 

Ensuring that Americans have enough retirement income in the
twenty-first century to meet their needs will require that the nation
and the Congress make some difficult choices.  Social Security has
been an effective agent for ensuring a reliable source of income in
retirement and greatly reducing poverty among the elderly.  The
effect of changes to the system on other retirement income sources
and their effects on various groups within the aged population should
be well understood before decisions are made.  Further, the interplay
of budget and saving effects will have to be carefully considered
before any reform proposal is adopted. 


MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix I

Francis P.  Mulvey, Assistant Director, (202) 512-3592
Ken Stockbridge, Evaluator-in-Charge
Michael Packard, Economist
Ken Bombara, Economist
Thomas Hungerford, Economist
Alicia Cackley, Economist

RELATED GAO PRODUCTS

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

Private Pensions:  Most Employers That Offer Pensions Use Defined
Contribution Plans (GAO/GGD-97-1, Oct.  3, 1996). 

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

Public Pensions:  Section 457 Plans Pose Greater Risk Than Other
Supplemental Plans (GAO/HEHS-96-38, Apr.  30, 1996). 

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

Public Pensions:  State and Local Government Contributions to
Underfunded Plans (GAO/HEHS-96-56, Mar.  14, 1996). 

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

Private Pension Plans:  Efforts to Encourage Infrastructure
Investment (GAO/HEHS-95-173, Sept.  8, 1995). 

Private Pensions:  Funding Rule Change Needed to Reduce PBGC's
Multibillion Dollar Exposure (GAO/HEHS-95-5, Oct.  5, 1994). 

Pension Plans:  Stronger Labor ERISA Enforcement Should Better
Protect Plan Participants (GAO/HEHS-94-157, Aug.  8, 1994). 

Private Pensions:  Changes Can Produce a Modest Increase in Use of
Simplified Employee Pensions (GAO/HRD-92-119, July 1, 1992). 

Pension Plans:  Survivor Benefit Coverage for Wives Increased After
1984 Pension Law (GAO/HRD-92-49, Feb.  28, 1992). 

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

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


*** End of document. ***