Social Security Reform: Assessing Changes To Future Retirement Benefits

Order Code RL30380
CRS Report for Congress
Received through the CRS Web
Social Security Reform: Assessing Changes To
Future Retirement Benefits
December 14, 1999
David Stuart Koitz
Specialist in Social Legislation
Domestic Social Policy Division
Congressional Research Service ˜ The Library of Congress

ABSTRACT
The conventional means of illustrating possible changes to Social Security is to compare
estimated benefits resulting from a reform proposal to those of current law. However, many
people mistakenly perceive current law to mean what retirees receive today. Social Security
benefits for future retirees are scheduled to decline on a relative basis as a legislatively-
mandated increase in the age for receipt of full benefits from 65 to 67 is phased in over the
next 22 years, and may have to be further reduced to keep them within the system’s available
resources. Consequently, showing a change to future current law levels does not show how
different benefits would be compared to those of today’s retirees. Moreover, it would not
necessarily illustrate how a proposal would affect the role of Social Security in providing an
adequate base of retirement income. This report suggests that understanding of a reform
plan’s impact might be enhanced by comparing its estimated benefits to two different
baselines: one that holds Social Security “replacement rates” constant — that keeps its role
in replacing pre-retirement earnings the same in the future as it is for today’s retirees — and
the other, that holds the purchasing power of today’s benefits constant through time. This is
a one-time report and will not be updated.

Social Security Reform: Assessing Changes To
Future Retirement Benefits
Summary
For the past 15 years, the trustees of the Social Security trust funds have
repeatedly projected long-range financing problems for the system. They have
projected that its income will not keep up with its outgo and eventually it will be
unable to pay its promised benefits in full. Policymakers have reacted with a wide
range of ideas. Most attempt to show how they would restore the system to good
health or how they would phase in what they believe is a better system — the goal
being to show the public that a more secure national retirement system is attainable.
The Social Security debate, however, is as much about the insecurity people feel
about their retirement in general as it is about Social Security’s long-range finances.
People need to have some understanding of what their own benefits will be; how
much of their earnings they will replace, or how much purchasing power they will
have. Only then can they assess how much other income and savings they will need.
The prospect that Social Security will be changed can only heighten that uncertainty.
Those who attempt to illustrate the effects of their proposals on “individuals” do
so by showing how much their benefits would differ from those promised by current
law. However, many people mistakenly perceive current law to mean what retirees
receive today. Legislation enacted in 1983, mandating a gradual increase in the age
for receipt of full benefits from 65 to 67, will cause future retirement benefits to be
lower on a relative basis. A person, for instance, retiring at age 65 in 2035 will get
13a% less than if the age change had not been made. By the same token, people do
not realize that even with the rise in the full-benefit age, future benefits are projected
to be higher in real terms, meaning they would buy more than those of today’s
retirees. On the other hand, if current law benefits had to be delayed or reduced
because of the system’s financing shortfall, their buying power would drop below
current levels, and the earnings replacement that they would provide could be nearly
40% less than that afforded to a retiree today.
Simply comparing possible changes to current law does not adequately show
people how Social Security will help them meet their retirement goals. This report
suggests that public understanding might be enhanced by comparing a reform
measure’s benefits to baselines that hold Social Security “replacement rates” constant
— that keep its role in replacing pre-retirement earnings the same in the future as it
is for today’s retirees — or alternatively, that hold the purchasing power of today’s
benefits constant through time. These baselines pose the following question: if we
could choose an “ideal” policy, would we want future Social Security benefits to
reflect the rise in the nation’s standard of living (as reflected by wages) or,
alternatively, its cost of living?
Having them rise with the standard of living means
raising them above the levels prescribed by current law with higher commensurate
costs. Keeping them up with the cost of living means lowering retirement benefits
from current law levels with lower commensurate costs. Setting a future course for
Social Security is difficult because ultimately much of the decision involves a value
judgment, not an analytical one, about what an appropriate retirement income “floor”
would be.

Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
The Complexities of “Current Law” Benefit Baselines — Two Very
Different Approaches to Measure the Degree of a Benefit Change . . . 2
A Lay Person’s Perception of “Current Law” May Differ From What He
or She May Be Entitled To . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Other Baselines for Measuring Change — Projecting Today’s Benefit Levels
into the Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Measuring earnings replacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Measuring “real” growth in purchasing power . . . . . . . . . . . . . . . . . . 9
Policy Implications of the Two Baselines . . . . . . . . . . . . . . . . . . . . . . . . . . 9
How Much Would Benefits Have to Change From Today’s Levels To
Account For The System’s Future Financing Shortfall? . . . . . . . . . . 12
What Would It Mean to Keep Social Security Replacement Rates
“Constant” Through Time? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
What Would It Mean to Keep The Purchasing Power of Social Security
Benefits “Constant” Through Time? . . . . . . . . . . . . . . . . . . . . . . . . . 15
Conclusion — Why Consider More Baselines? . . . . . . . . . . . . . . . . . . . . . 16
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
List of Figures
Figure 1. Declining Future Replacement Rates . . . . . . . . . . . . . . . . . . . . . . . . 11
Figure 2. Rising Purchasing Power of Future Retirees’ Benefits . . . . . . . . . . . 11
Figure 3. Declining Future Replacement Rates Caused by System’s
Financing Shortfall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Figure 4. Drop in Purchasing Power of Future Benefits Caused by
System’s Financing Shortfall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Figure 5. Future Benefit Increases Needed to Achieve Constant Replacement
Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Figure 6. Future Benefit Changes Needed to Achieve Constant Purchasing
Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
List of Tables
Table 1. Projected Annual Social Security Retirement Benefits As a Percent
of Worker’s Final Year’s Earnings (Replacement Rate) . . . . . . . . . . . . . . . 6
Table 2. Change in Social Security Retirement Benefit Levels Caused By The
Social Security Amendments of 1983 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Table 3. Projected Increases in Purchasing Power of Social Security Benefits
for Future Retirees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Table 4. Illustrative Future Social Security Retirement Benefits and
Replacement Rates Computed Under Rules Prescribed by Current Law . . 19
Table 5. Illustrative Social Security Replacement Rates Computed Under
Rules Prescribed by Current Law, At Age 65 and the “Full Benefit” Age,
1940 - 2070 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Social Security Reform: Assessing Changes To
Future Retirement Benefits
Introduction
Over the past 15 years, the various Social Security boards of trustees serving
under three different Presidents have projected long-range financing problems for the
Social Security system. They have projected to varying degrees that its income will
not keep up with its outgo, its trust fund reserves will subsequently run out, and
eventually it will be unable to pay its promised benefits in full.1 Policymakers have
reacted with a wide range of ideas. Some would seek to increase the system’s
income, others would constrain its benefits, and still others would create an entirely
new system. Most attempt to illustrate the merits of their plans by showing how they
would restore long-range balance to the Social Security trust funds (75-years’ worth)
or how they would phase in what they perceive to be a better system for providing
retirement incomes — the goal being to show the public that a more secure national
retirement system is attainable.
Thus, for a reform advocate, it is a coveted “prize” to have the Social Security
Administration’s (SSA’s) actuaries estimate that a proposal would result in a
financially sound system. However, past legislative attempts to “rescue” Social
Security have shown that it is difficult to inspire public confidence in a reform effort
since ultimately a successful outcome does not come from projections. It comes from
the actual economic and demographic developments that drive the system’s income
and outgo.2 Moreover, talking about the system in the aggregate is a sometimes
obscure topic since actuarial science and the policy issues in financing retirement
programs are so complex, with so many uncertain variables, as to be beyond the
general comprehension of non-experts. The actuaries’ projections are strewn with
far-reaching assumptions and the debate is confusing, value-laden, and in many
respects arcane.
For the lay person, the issue is as much about his or her own benefits as it is
about the system’s viability. Illustrations of impacts on “individuals” are more
tangible (albeit no more certain). They shift the debate to a level that people can
understand by showing how much their potential benefits would change — talking
1 For more information, see: 1999 Annual Report of the Board of Trustees of the Old Age,
Survivors and Disability Insurance Trust Funds
, Washington, GPO, 1999.
2 During the past 25 years, there were two major attempts to shore up the system — in 1977
and 1983, P.L. 95-216 and 98-21, the Social Security Amendments of 1977 and 1983. In
both cases, favorable long-range projections made at the time of enactment proved to be too
optimistic within a few years as actuarial methods were refined and less favorable economic
trends emerged.

CRS-2
about how a $900 monthly benefit would be altered is a lot easier to comprehend than
system-wide changes involving trillions of dollars.
The conventional means of illustrating these effects is to compare estimated
benefit levels resulting from a proposal to those of current law. Comparisons to
current law are useful for policymakers as well since they show the degree of “policy
change” a particular proposal would make — how much benefits would be raised or
lowered, and who wins and loses. These comparisons are obviously important.
However, in and of themselves, comparisons to “current law” do not really illustrate
what a proposal will mean to someone attempting to plan for an adequate retirement
income. How well off can a 25-year old expect to be 40 years from now when
average yearly earnings are estimated to be $155,000, and his or her yearly Social
Security benefit would be $54,000? If shown these projections, how is this person to
view them, given that today’s average yearly earnings are less than $30,000 and the
yearly benefit arising from a comparable wage record is $12,000? Clearly, a $54,000
annual benefit would seem like a fortune to someone earning average earnings today.
Should this young worker be upset if a reform plan proposes to lower this amount by
25%? Or should this worker anticipate being so much richer that the cut will not
really matter?
Ultimately, for the public, the question is not about current law versus proposed
changes. To the extent that today’s workers have an expectation, it is likely to be that
their Social Security benefits will be comparable, relatively speaking, to what people
retiring today get.3 However, that would be inaccurate, since changes to constrain
the rise in the system’s future cost, as well as other benefit revisions yet to take effect,
are already imbedded in the law.
The Complexities of “Current Law” Benefit Baselines — Two Very
Different Approaches to Measure the Degree of a Benefit Change

Complicating the comparison issue is the fact that two very different portrayals
of current law now permeate the policy debate. Under the first, future Social Security
benefits are projected using the benefit computation rules of current law. Social
Security benefits are determined from a three-step formula that calculates benefits
3 Awareness by workers of their potential Social Security benefits should rise to some extent
in the near future with SSA’s recent implementation of a system to send people annual
“Personal Earnings and Benefit Estimate Statements” (PEBES). These are detailed,
individualized statements to be issued to most workers age 25 and older that reflect whether
a worker has met the requisite Social Security eligibility criteria and include estimated future
benefits based on the worker’s earnings record to date. This mailing was mandated in 1989
by legislation. Statements have been issued automatically to workers age 60 and older for a
number of years; younger workers could have received them upon request. Their required
issuance to workers age 25 and older became effective in October 1999. While these
statements should enhance workers’ understanding of their potential benefits, their greatest
utility will be for those in the latter portion of their careers or nearing retirement, for whom
most of the earnings record used to compute benefits will have been established. It will be less
useful for younger workers, who will still have much of their careers ahead of them, leaving
considerable uncertainty about the career earnings that will ultimately be used to determine
the level of their benefits.

CRS-3
based on a worker’s earnings record. The higher a worker’s career earnings, the
higher his or her benefits will be.4 These rules are specified in the Social Security Act.
By taking these rules and hypothesizing certain wage histories, dates of birth, and
ages at retirement, projections can be made to illustrate the potential benefits
promised to workers.
Alternatively, because the system’s financing is projected to become inadequate
— the latest projections of which suggest that only 71% of benefits would be payable
in 2034 — some reform proponents compare projected benefits resulting from their
plans, many of which include proposed benefit constraints, to those that would be
“payable” under current law given the system’s eventual income shortfall.5 By so
doing, they attempt to show how their plans actually increase benefits compared to
doing nothing to shore up the system. Simply stated, they contend that “current law”
will eventually result in benefit levels that will be 29% lower than prescribed by the
benefit computation rules in the law.
The first perspective reflects a “baseline” that assumes current law will be fully
funded for the next 75 years. It implicitly assumes that taxes will be raised or some
other income source found to meet the system’s ongoing costs. While useful in
showing the degree of “benefit policy” change a reform proposal would make, it does
not recognize that benefits eventually would not be payable at those levels. “Current
law” consists of more than prescribed benefit rules. Social Security is considered a
self-financed program, meaning that it also has prescribed financing provisions (i.e.,
its own taxes), and under those provisions, the system is projected to have fewer
resources than it will need. Hence, when “current law” is viewed from a perspective
that reflects both its benefit rules and financing provisions, the benefit levels
4 While benefits are larger the higher one’s earnings, lower earners’ benefits represent a larger
proportion of their average earnings. In other words, on a wage-replacement basis, the system
is “tilted” in favor of low-wage earners. Benefits are computed by applying a three-step
formula to a worker’s “average indexed monthly earnings” (AIME) calculated using as many
as 35 years’ worth of earnings. For workers reaching age 62 in 1999, monthly benefits are
the sum of 90% of the first $505 of AIME, 32% of the next $2,538, and 15% of the
remainder. Both the earnings used to compute the worker’s AIME and the so-called “bend
points” in the benefit formula (“$505" and “$3,043" in 1999) are indexed to reflect growth
in average earnings in the economy. For retirees, each year’s earnings are indexed from the
year they were earned to the year the worker reaches age 60. Earnings at age 60 and beyond
are included in the calculation at their nominal value.
5 Under the Social Security trustees’ 1999 intermediate (or “best guess”) forecast, the system
is projected to have an average 75-year deficit equal to 2.07% of taxable payroll under current
law (taxable payroll being the amount of earnings in the economy subject to Social Security
taxation). This amount is equal to about 15% of the average income of the system over the
75-year period as a whole. In terms of today’s taxable payroll, this shortfall would be
equivalent to $78 billion per year. In the year 2075, the deficit would be 6.53% of taxable
payroll, an amount equal to 49% of the system’s projected income for that year. (For
reference, see footnote 1).

CRS-4
prescribed only by the benefit computation rules overstate the baseline for measuring
the degree of change a reform proposal might make.6
These two baselines represent what might be considered broad boundaries for
measuring change. The first addresses the question of how much change in the
benefit rules is being proposed in a reform plan. While the benefits prescribed by
current law may not be sustainable given the system’s financing problems, those rules
do reflect the “benefit policy” that lawmakers established in 1983, when the last major
reforms were made. Lawmakers assumed then that the system’s financing would be
adequate to cover its commitments for the next 75 years.
The second measure addresses the question of how much change a proposal
would make given the system’s projected funding shortfall — it reflects the degree of
“taxation policy” that lawmakers established in 1983. As previously mentioned, to
say that a plan would cut benefits from current law levels, without regard to the
inability of the system to sustain those levels, may overstate the degree of reductions
being proposed. Indeed, in some circumstances, using a baseline that recognizes the
lower levels that can be financed under current law may change the portrayal of the
plan’s provisions from being benefit reductions to benefit increases.
A Lay Person’s Perception of “Current Law” May Differ From
What He or She May Be Entitled To

What is likely to surprise the lay observer is that neither of these measures of
future retirees’ benefits corresponds to what today’s retirees receive.
Workers who retired at age 65 over the past decade, who worked for at least 35
years and always earned average wages (almost $30,000 a year in 1999), on average
got about 42% of their pre-retirement earnings replaced by Social Security; low-wage
earners got 57% (defined here as those who always earned 45% of the average wage);
and above average-wage earners got 34% (defined here as those who always earned
160% of the average wage). These so-called replacement rates, however, would not
prevail in the future even if there were no financing problem. Under current law, they
will drop for comparable future retirees beginning with those who reach age 62 next
year. (See Table 1.)7
6 The reader should recognize that it would take legislation to raise Social Security taxes. In
the absence of such legislation, restraining benefit outlays would be the only option available
to the Secretary of the Treasury and Commissioner of Social Security. Although nothing in
the Social Security Act tells them what to do if resources become inadequate, it would be their
responsibility to constrain benefits in some fashion since under the law (Section 201 of the
Act) benefits can be paid only from available Social Security taxes or to the extent there are
“balances” recorded to the Social Security trust funds.
7 The literature of the Social Security Administration often shows that the system will provide
“constant” replacement rates over time. However, a close look at those portrayals shows that
the steady or constant replacement rate scenario occurs only when measured at the “full
retirement age.” Since the “full retirement age” will rise from 65 to 67, and workers retiring
at younger ages will incur reductions in monthly benefits due to their “early” retirement, the
(continued...)

CRS-5
This reduction will result from a provision enacted in 1983 under which the age
for receipt of full benefits will rise from 65 for today’s retirees to 67 for those born
after 1959. A person retiring at age 65 in 2025 will get a 13a% lower monthly
benefit than if the age change had not been made. This provision was enacted in 1983
and will be phased in from 2000 to 2022.8 When fully effective, benefits will be
payable beginning at age 62 as they are today, but workers retiring before age 67 will
have to take a permanent reduction in their benefits. (See Table 2.) Today,
retirement before age 65, i.e., taking “early” retirement, causes a reduction. As the
full-benefit age rises, the reduction will apply if a worker draws benefits before that
age, and the amount of the reduction at younger ages will be larger.9
7 (...continued)
notion that the current system provides “constant” replacement rates can be misleading. At
any given retirement age below 70, the system’s replacement rates will be lower for future
retirees than they are for workers choosing to retire at that age today.
8 This provision was intended to help shore up the system’s financial condition by reducing
its long-range benefit outgo. Actuarial projections made at that time showed that it would
eliminate about one-third of the system’s average 75-year deficit. While this paper focuses
on how the provision will affect future retirees, it will also affect the benefits of many future
spouses and survivors. They too will be lower than under the pre-1983 law, but by different
amounts. The reductions do not apply to the disabled or survivors under age 60. For further
discussion, see CRS Report 84-677, Various Effects of Raising the Normal Retirement Age
For Social Security Benefits
, by David Koitz.
9 There are additional scheduled changes in the law that will raise benefits for certain workers
in the future. Those who work after reaching the full-benefit age will be able to earn a larger
amount without forfeiting benefits than workers who retired in the past. The threshold for the
amount of earnings a person can have and still collect benefits will rise from its present level
of $15,500 to $30,000 in 2002 (indexed to wage growth thereafter). By 2002, the relative
level of this so-called “exempt amount” will be more than twice as large as it was through
much of the 1980s and 1990s. In addition, to the extent that workers choose to delay
retirement beyond the full-benefit age, or whose earnings are too high to permit the payment
of benefits, they will get larger “delayed retirement credits” (i.e., credits that cause larger
benefits to be paid when the worker does eventually retire) than were payable under the pre-
1983 law (24% at age 70 in 2029 versus 15% at that age under the old law). The increase in
the full-benefit age will be the more dominant of these pending changes, since most eligible
workers draw benefits at or before age 65. Data from SSA shows that 71% of newly-awarded
retirement benefits in 1997 went to people under age 65; 86% went to people under age 66.
Nonetheless, it would be inaccurate to infer that all future Social Security recipients will get
lower benefits as a result of this provision.

CRS-6
Table 1. Projected Annual Social Security Retirement Benefits As a
Percent of Worker’s Final Year’s Earnings (Replacement Rate)
Year of
Decrease in replacement rates (in
retirement
“rounded” percent) from those of
at age 65
Replacement rates (in percent) for
workers retiring in 1990-2000 period
workers retiring at age 65 with:
with:
Low wage
Average
High wage
Low wage
Average
High wage
record
wage
record
record
wage
record
record
record
1990-2000
57.1
42.4
34.0
—-
—-
—-
(average)
2010
53.2
39.6
32.7
- 7
- 7
- 4
2025
49.4
36.7
30.3
- 13
- 13
- 11
2050
49.5
36.7
30.3
- 13
- 13
- 11
Source: CRS benefit computation model based on the long-range “intermediate” economic
assumptions of the 1999 Social Security trustees’ report. Illustrations are for benefits computed
under current law for full-career workers.
Table 2. Change in Social Security Retirement Benefit Levels Caused
By The Social Security Amendments of 1983
Percent of basic benefit amounta payable at given age of
Benefit rules
retirement (for workers born after 1959)
Age 62
Age 65
Age 66
Age 67
Age 70
Pre-1983 law
80
100
103
106
115
After 1983 Amendments
70
86 2/3
93 1/3
100
124
Difference (as % of
-12.5%
-13.33%
-9.39%
-5.67%
+ 7.82%b
pre-1983 law level)
Source: derived from CRS Report 84-677, loc.cit.
a Under the law, this is referred to as a “Primary Insurance Amount” or PIA.
b Increase is due to change enacted in 1983 raising “delayed retirement credits” (see footnote 9).
Adding further complexity for the lay person is that, even with the scheduled
drop in replacement rates, under the trustees’ projections future retirees will receive
benefits that are more valuable in terms of their purchasing power than those of
comparable retirees today. Future Social Security benefits, adjusted for inflation,
would have more value. Under the projections, the real value of benefits is estimated
to rise over time — they will buy more goods and services as time passes — even
though the role of the system in replacing earnings will contract. The reason the

CRS-7
benefits will be larger is that earnings tend to grow faster than prices over time and
this is assumed in long-range Social Security projections.10 (See Table 3.)
Table 3. Projected Increases in Purchasing Power of Social Security
Benefits for Future Retirees
Year of
Projected purchasing power of benefits for workers retiring at age 65
retirement
at age 65
Annual benefit in constant 1999 dollars
Increase in purchasing power over
for retiree with:
benefits payable to retiree in 2000 (in
rounded percent)
Low wage
Average
High wage
Low wage
Average
High wage
record
wage
record
record
wage
record
record
record
2000
7,017
11,602
15,022
—-
—-
—-
2010
7,491
12,375
16,360
7%
7%
9%
2025
7,925
13,086
17,292
13%
13%
15%
2050
9,848
16,257
21,483
40%
40%
43%
Source: CRS benefit computation model based on the long-range “intermediate” economic
assumptions of the 1999 Social Security trustees’ report. Illustrations are for benefits computed
under current law for full-career workers.
Neither of these future trends is breaking news for Social Security and pension
experts, but they are not something the public readily understands. Thus, effectively
communicating the impact on individuals of potential Social Security reform plans is
not just a matter of comparing proposed changes to current law. What may be more
understandable to the public is to compare benefits under these proposals to what
people receive from the system today.
Other Baselines for Measuring Change — Projecting Today’s
Benefit Levels into the Future

Based on the premise that workers retiring in the future would want to know
how much more or less they would get than people retiring today, this report
describes and illustrates the effects of two other potential baselines for measuring
Social Security changes. Both have been employed in past debates and represent
measures frequently used by Social Security and pension experts. They start with the
approximate level of benefits paid to people retiring today.
While this may sound like a simple concept, because Social Security benefits are
based on a worker’s earnings history and earnings levels in the economy tend to rise
10 Under the trustees’ long-range “intermediate” assumptions, earnings are assumed to grow
by 4.2% annually; inflation as measured by the Bureau of Labor Statistics’ Consumer Price
Index for Wage Earners and Salaried Workers (CPI-W) by 3.3%. The difference, often
referred to as the “real wage” differential, is 0.9% per year.

CRS-8
over time, it is not a simple matter to express today’s benefit levels in values that
pertain to workers retiring in the future. For example, projected Social Security
benefits for someone retiring at age 65 in 2000, after a career of steady average
earnings, would be $987 a month. They would be calculated from cumulative
earnings over the worker’s highest 35 years of earnings ending with nearly $30,000
in 1999. Under assumptions about future wage growth currently used by the Social
Security trustees, a similarly situated worker retiring in 2025 would get benefits of
$2,442 a month based on a cumulative high-35 years of earnings ending with $80,000
in 2024.
Does this mean that the 2025 retiree will be 2.5 times better off than today’s
retirees simply because his or her monthly benefits are 2.5 times higher? No. In one
sense, this worker will be better off — his or her benefits will have more “buying
power” — but not by 2.5 times. The trustees’ projections assume that inflation in the
intervening years would average in excess of 3% annually, lowering the worth of
earnings and benefits paid in 2025. Converting the $2,442 monthly benefits payable
then into 1999 dollars would reduce them to $1,090. This is higher than the $987
monthly benefit payable to a comparable worker retiring next year, but not by 2.5
times. Average wage levels are projected to grow by 172% cumulatively from 2000
to 2025; prices would rise by 119%. Hence, the wage record used to compute the
2025 retiree’s benefits would have grown substantially more than if average earnings
had only kept up with inflation. In real terms, and even with the benefit constraints
of current law factored in (i.e., the rise in the full-benefit age), the 2025 retiree could
buy 13% more with his or her monthly benefits than the person retiring in 2000, as
shown in Table 3. The comparable retiree in 2050 could buy 40% more.
In another sense, however, these future retirees will be worse off. While the
their benefits will buy more, they would not represent a larger percentage replacement
of their pre-retirement earnings than that of current retirees because of the scheduled
increase in the full benefit age. In fact, they would be lower, as shown in Table 2.
The 2025 retiree’s annual benefits of $29,300 would equal 36.7% of his or her final
earnings of $80,000. Comparable workers retiring over the past decade would have
had a 42.4% replacement rate. Thus, over the next 25 years, the role of the Social
Security benefit in replacing a retiree’s pre-retirement standard of living would have
fallen by 13% (42.4 - 36.7 = 5.7; 5.7/42.4 = 13%).
The baselines discussed in this report attempt to provide a way of measuring how
much a proposal would change benefits compared to the levels prevailing today, while
recognizing that the economy, and wage levels in particular, are projected to grow in
“real” terms in the future.11
11 Obviously many different forecasts are possible, including ones that assume economic
stagnation or contraction. However, over the past 50 years, “real” or inflation-adjusted Gross
Domestic Product (GDP) has grown at average annual rates exceeding 3%, and it has been
conventional practice for most long-range government and private forecasts to assume
continued future growth of the U.S. economy. Real growth in both GDP and earnings is
assumed under all of the Social Security trustees’ 1999 financial projections for the Social
Security system, including their pessimistic forecast.

CRS-9
Measuring earnings replacement. One of these baselines starts with
replacement rates for people retiring today, and then uses those percentages to project
what future benefit levels would be if they were to prevail for future retirees. In
effect, Social Security benefits would grow as earnings grow. If earning levels rise
by 25%, Social Security benefits would do likewise. The “floor” of retirement
income, as reflected by this baseline, would be a constant share of pre-retirement
earnings from one generation to the next. With this approach, the underlying
assumption is that as the economy grows (and earnings rise), Social Security benefits
should keep pace. Because earnings tend to grow faster than prices over time, this
approach reflects a baseline under which future Social Security benefits would be
assumed to grow such that they would purchase more than today’s benefits can buy
— the idea being that the relative role of Social Security should grow in tandem with
the nation’s rising standard of living.
Measuring “real” growth in purchasing power. The other baseline assumes
that the Social Security benefits paid to people retiring today should rise for future
retirees at the same rate as inflation so that they will be able to “purchase” the same
amount of goods and services as the benefits of today’s recipients. Under this
approach, the underlying assumption is that as the economy grows and per capita
incomes rise, Social Security does not need to provide the same “earnings
replacement” as it does today because future workers will be able, or should be
encouraged, to develop supplementary sources of wealth to achieve reasonable
retirement living standards on their own. It implies that in the aggregate Social
Security benefits today provide a “sufficient” retirement base and that real increases
in the purchasing power of benefits are either (1) unnecessary or (2) to be left to
future generations to decide.12
Policy Implications of the Two Baselines
Projected current law benefits for future retirees — i.e., those based on the
benefit computation rules of current law — fall between these two baselines. They
are not so large that future retirees can count on getting the same replacement of their
earnings as today’s retirees, but they are higher than levels that merely match the
purchasing power of today’s initial benefit levels.
12 This issue should not be confused with the policy of providing cost-of-living adjustments
(COLAs) to retirees after they become recipients. Social Security COLAs are based on
inflation as measured by the CPI, the purpose being to maintain the purchasing power of the
benefits after they are awarded. Some policymakers have proposed that COLAs be cut to
adjust for a perceived “upward bias” in the CPI (i.e., that it overstates inflation) or simply to
help restore the system’s financial health. The issue posed in this report is not about COLAs.
It is about whether “initial” benefits of new retirees in the future should be allowed to go up
at the rate of wage growth or at the rate of inflation. In comparing different plans, it is
important not to consider only the value of initial benefits, because doing so can mask the
impact of a COLA constraint on a retiree’s ongoing income. Here, too, the issue is what role
the system should have in maintaining a worker’s standard of living once he or she becomes
a retiree — the question becomes “at what level should the value of post-entitlement benefits
be protected?”

CRS-10
For example, workers retiring at age 65 during the 1990s, after a career of
average earnings, would have had benefits in their first year of retirement equaling
42.4% of their final year’s earnings (on average). Comparable workers retiring in
2025 are projected to have benefits equaling 36.7% of their final earnings — that
equates to a 13% drop in relative benefits. (See Figure 1.)
However, the value of their initial benefits, after adjusting them for inflation
between now and 2025, would be 13% higher. The first year’s benefits for an
“average-earner” retiring in 2000 would be $11,602. The first year’s benefits for a
comparable worker retiring in 2025 would be $13,086 — 13% higher. (See Figure
2
.)
In sum, if one took the policy position that benefits for future retirees should
always replace the same proportion of workers’ pre-retirement standards of living
regardless of when they were born, Figure 1 indicates that the benefit levels promised
by current law will be inadequate. To cut them further as a means of restoring the
system’s solvency would exacerbate the inadequacy.
If, alternatively, one took the position that the benefits promised to future
retirees should always have the same purchasing power as the benefits of today’s
recipients, Figure 2 indicates that future benefits will be too generous and, therefore,
could be reduced to restore the system’s solvency without erosion of retirees’
purchasing power.
Both baselines have merit; there is no definitive right or wrong approach.
Choosing the “best” one to assess the worth of current law benefits and potential
changes is a value judgement for policymakers.

CRS-11
Figure 1. Declining Future Replacement Rates
(First Year's Benefits as % of Final Earnings)
46%
Average replacement rate
44%
for retirees, 1990-2000
42%
40%
38%
36%
Declining replacement rates for
34%
future retirees
32%
1995 2000 2005 2010 2020 2030 2040 2050
Year of Retirement
(Workers retiring at age 65 with "average-wage" record)
Figure 2. Rising Purchasing Power of Future
Retirees’ Benefits
$1,600
( M o n t h l y B e n e f i t s i n C o n s t a n t 1 9 9 9 D o l l a r s )
Rising value of future
$1,400
retirees' benefits
$1,200
$1,000
$800
Benefits of
$600
retiree in 2000
$400
2000
2010
2020
2030
2040
2050
Year of Retirement
( W o r k e r s r e t i r i n g a t a g e 6 5 w i t h " a v e r a g e - w a g e " r e c o r d )

CRS-12
How Much Would Benefits Have to Change From Today’s Levels To
Account For The System’s Future Financing Shortfall?

If one accepts the trustees’ projections that future Social Security benefits will
not be payable in full because of the system’s financing shortfall, Figures 1 and 2
overstate the worth of future current law benefits. If the system’s financing is not
increased legislatively between now and 2034, the Secretary of the Treasury and
Commissioner of Social Security will be forced to constrain benefit outlays in some
way to stay within the system’s future revenue stream — that is, under the trustees’
latest projections, in 2034 they would be forced to reduce benefits by 29% below the
promised levels.
Under this scenario, the replacement rate for an “average” earner would drop
from the 42.4% level experienced by recent retirees to about 26% for the comparable
worker retiring in 2034. (See Figure 3.) This reflects both the 29% reduction due
to the system’s financing shortfall, and the already scheduled 13a% reduction
stemming from the rise in the full-benefit age to 67. Together, they would cause
nearly a 40% reduction in replacement rates from those prevailing for recent retirees
(42.4 - 26.0 = 16.4; 16.4/42.4 = 38.5%).
Similarly, the purchasing power of those future benefits would be 29% less than
promised by current law. In fact, in contrast to Figure 2, where the purchasing
power of promised benefits is shown to rise over time, a 29% reduction in benefits
would result in 13% less purchasing power for the 2034 retiree than that afforded to
today’s retiree. The initial monthly benefit would be $837 (when adjusted for
inflation) for a worker retiring in 2034, compared to $967 for a worker retiring in
2000. (See Figure 4.)
Under this scenario, in which the benefits promised by the system cannot be
financed, the resulting “payable” benefits would fall below both baseline measures.
The initial benefits of future retirees, i.e., in 2034 and later, would provide only about
60% of the earnings replacement afforded to retirees in the 1990-2000 period. Said
another way, as a source of retirement income, the role of the system would shrink
by 40%. Similarly, the buying power of those future benefits would rise through 2033
and then, when the trust fund balances run out, decline sharply to a level below that
of today’s retirees. The benefits then would buy less goods and services than the
benefits of today’s retirees.

CRS-13
Figure 3. Declining Future Replacement Rates
Caused by System’s Financing Shortfall
45%
(First Year's Benefits as % of Final Earnings)
40%
35%
Decline in promised
replacement rates
30%
Further decline due to
25%
system's financing shortfall
20%
1995 2000 2005 2010
2020 2030 2040 2050
Year of Retirement
(Workers retiring at age 65 with "average-wage" record)
Figure 4. Drop in Purchasing Power of Future
Benefits Caused by System’s Financing Shortfall
$1,500
( M o n t h ly B e n e fits In C o n s t a n t 1 9 9 9 D o lla rs )
Rising purchasing power
$1,300
of initial benefits
$1,100
$900
$700
Purchasing power drops due to
system's financing shortfall
$500
2000
2010
2020
2030
2035
2040
2050
Year of Retirement
(Wo r k e r s re t irin g a t a g e 6 5 w i t h " a v e r a g e - w a g e " r e c o rd)

CRS-14
What Would It Mean to Keep Social Security Replacement Rates
“Constant” Through Time?

If policymakers desired a future course where Social Security’s role in replacing
pre-retirement income were to stay the same from one generation to the next,
replacement rates would have to be higher than both those that are “promised” — i.e.,
under the benefit rules written in the law — and those that are “payable” — i.e., what
can be funded under the financing provisions in the law. For example, to achieve this
for an “average earner,” the level of promised benefits would have to rise ultimately
by about 15% for future age-65 retirees, i.e., for persons born in 1960 and later. The
level of “payable” benefits — i.e., recognizing that the system’s financing shortfall
would have caused benefit reductions — would have to rise by about 63%. (See
Figure 5.) The implicit tax burden also would have to rise substantially.13
Figure 5. Future Benefit Increases Needed to Achieve
Constant Replacement Rates
60%
If baseline recognizes
63%
system's financing shortfall
40%
If baseline is
benefit rules in law
15%
20%
0%
1990/2000
2020
2040
(avg.)
Year of retirement
(Workers retiring at age 65 with "average-wage" record)
13 Under the trustees’ “intermediate” projections, the cost of the retirement and survivors’
portion of the Social Security system (i.e., ignoring the disability portion) is projected to be
17.3% of taxable payroll in 2075; the projected income, 11.51%. Thus, paying for the
promised benefits of current law assumes a relative tax rise of 50%. Raising future benefits
enough to produce constant replacement rates would only augment the necessary rise.

CRS-15
What Would It Mean to Keep The Purchasing Power of Social
Security Benefits “Constant” Through Time?


If, on the other hand, policymakers wanted to pursue a course where future
retirees could expect their Social Security benefits to buy the same amount of goods
and services as those of today’s retirees, future benefit levels would be set lower than
now promised in the law. For a 2025 “average earner” retiring at age 65, benefits
would be reduced by 11%; for a comparable retiree in 2050, they would be reduced
by 29%. (See Figure 6.) By 2075, the reduction would be 43%. The aggregate cost
of the system would drop accordingly.14
When compared to the baseline of benefits “payable” under current law (i.e.,
where benefits rise until 2034, but then fall by 29% because of the system’s limited
resources), benefits that provide constant purchasing power would be lower through
2033, but higher in 2034 and for some time thereafter. For comparable retirees, they
would be 15% higher in 2035, and about 1% higher in 2050.
Figure 6. Future Benefit Changes Needed to Achieve
Constant Purchasing Power
( M o n t h ly B e n e fits In C o n s t a n t 1 9 9 9 D o lla rs )
$ 1 , 3 0 0
Benefit reduction from levels
" p r o m ised" under current law
$ 1 , 1 0 0
$ 9 0 0
Benefit increase from levels
$ 7 0 0
"payable" under current law
$ 5 0 0
2 0 0 0
2 0 1 0
2 0 2 0
2 0 3 0
2 0 4 0
2 0 5 0
Y e a r o f R e t i r e m e n t
(Wo r k e r s re t irin g a t a g e 6 5 w i t h " a v e r a g e - w a g e " r e c o rd)
14 As described in the previous footnote, the system’s projected income of 11.51% of taxable
payroll is 33% lower than its projected 17.3% of payroll costs. Reductions in promised
retirement benefits of this magnitude would appear to be more than sufficient to close the
long-range gap between the system’s income and outgo.

CRS-16
Conclusion — Why Consider More Baselines?
When the Social Security system was created in 1935, its role was generally
viewed to be earnings replacement, not wealth accumulation. It emanated from the
dire economic conditions of the early 1930s, and its proponents saw it as an eventual
means of keeping a large segment of the workforce from slipping into poverty in old
age. It was to provide a “floor of protection” for society, a means of affording the
nation’s workers with at least a minimal income during retirement without having to
rely on welfare. While, as a class, the elderly are much better off today — in many
respects, they have achieved economic well-being equal to or greater than the non-
elderly — Social Security remains the dominant source of their retirement income.
All other things held constant, Social Security lifted some 41% of the elderly out of
poverty in 1996. Certainly for some, it simply adds to their wealth, an additional sum
to augment an already ample income from other sources. However, the elderly are
economically diverse. While only 4% of aged married couples had incomes below the
poverty line in 1996, the figure for single men was 13%, and for single women, 20%.
To the extent that Census Bureau surveys accurately reflect the elderly’s well-being,
Social Security accounted for 40% of the aggregate income of the population 65 and
older. As such, it was their largest single source of income, with earnings from work,
pensions, and asset income running a distant second, third, and fourth. For 66%,
Social Security represented more than half of their income; for one-third, it
represented 90%; and for 18%, it represented 100%.
Given its significance, the paramount question about the Social Security system’s
survival is how much individuals can count on getting. In the best of circumstances,
there is no certainty in retirement planning. The prospect that a major source of
retirement income may be changed by legislative reforms only heightens the
uncertainty.
Policymakers approach the issue by highlighting what changes they would make
to attain a financially sound system. Some, satisfied with the basic role that Social
Security now plays, want minimal changes. They seek to preserve its so-called social
safety net features and want its benefit provisions held constant to the maximum
extent possible. Others, seeing the public’s desire for a more secure system as an
opportunity for change, seek to create an entirely different way of providing for
retirement. They want something more — principally, to use “reform” as a potential
catalyst for economic expansion. They want a more individualized system where
potential wealth accumulation is a substantial or even primary catalyst for
participation.
While a number of these ideas are far-reaching, few really challenge the notion
that the nation needs a basic mandatory pension system. Financing issues aside, the
underlying debate is about how much of a role the system should have as a source of
future earnings replacement. However, in many instances, complex multi-faceted
proposals have emerged, often leaving a confusing array of images about what the
resulting benefits would be. Even when benefit examples are provided, they can be
confusing or misleading, depending in part on the basis for comparison. Lay
observers cannot be expected to see their way through the many varied distinctions.

CRS-17
Benefit examples are used as the means of communicating those distinctions, and
a great deal of attention is given to comparisons to current law. If a proposed reform
shows reductions, the “image” it purveys is “bad” (for future retirees); if it provides
increases, it’s “good.” However, to the extent that current law is mistakenly seen as
the level of benefits that retirees get today, any image of change is going to be
misleading.
If it matters that people know what amount of their aggregate retirement income
they can count on from a Social Security reform plan, then some measure of how
much its benefits will be worth is necessary. Comparisons to current law are not
sufficient. The level of benefits payable from a proposed reform has to have some
context — “how much of my pre-retirement earnings will Social Security replace?”
or “how much will my benefits buy?” And given those figures, “how much do I need
(or should I plan) to rely on other pensions, personal savings, or continued work?”
While current law, under whatever definition, provides a baseline by which the relative
amount of Social Security change can be observed, the ultimate measure is one
showing the extent to which a person’s retirement income goals are met.
How much, then, should future Social Security benefits be?
Retirement planners say that people should prepare for retirement by attempting
to determine the percent of pre-retirement earnings they will need to sustain their
desired standard of living. Thus, if Social Security were to provide the same relative
replacement of pre-retirement earnings from one cohort to the next, there would be
a “floor” for people to build on — a given base amount of retirement income would
always come from the program. Private pensions and personal savings could then be
used to supplement that stable base. If, on the other hand, Social Security’s role were
to decline as a replacement of pre-retirement earnings, dependency on other sources
would have to rise (all other things being equal). However, retirement planning is not
easy. Private pensions are not available to all workers, and many people, particularly
low-wage earners, do not always have discretionary income during their working
years to set aside in personal savings accounts. Even when savings are accumulated,
they may be depleted before retirement for other things (acquiring a home, high
medical costs, college tuition, job loss, etc). Hence, the questions arises, if Social
Security benefits are scheduled to decline as a percent of pre-retirement earnings, and
may have to fall further because of the system’s financing shortfall, to what extent will
future workers be unable to fill the growing gap between what Social Security
provides and the amount they will need to attain their desired retirement standard of
living?
From another perspective, the economic condition of the nation, and the elderly
in particular, is very different than when Social Security was conceived. In the
aggregate, the nation is much wealthier today. In 1935, the unemployment rate stood
at 20%, down from 25% in 1933; for much of 1999, it was under 4.5%. Disposable
per capita income (adjusted for inflation) is more than 4 times higher now. Average
earnings, similarly adjusted, are 2.3 times the 1937 level. By some estimates, more
than half of the elderly in the early 1930s were in a state of dependency. In 1996, the
overall poverty rate for people age 65 and older was under 11%. While much of the
drop can be attributed to Social Security, 63% of the elderly had income from
personal assets, and 41%, from other pensions. Almost half of employed workers

CRS-18
over age 15 today have employer-provided pension coverage; 62% of full-time, year-
round workers age 25-54 have it. Some 41% of households own stock; 43% have
IRAs, Keogh plans, or other tax-favored savings arrangements; 31% have life
insurance; 23% own savings bonds; and 65% own their own homes with an even a
higher percentage, between 70% and 80%, for the elderly. While these figures show
gaps as well as economic advancement, they raise the question of whether the
program established in the 1930s is the program needed for the 21st century, and,
specifically, whether the societal “floor” that Social Security affords today’s retirees
needs to be as large and extensive in the future. Social Security replacement rates are
higher now than in the system’s first 30 years and will continue to be so in the future,
even at the reduced levels scheduled in the law. Does the program’s role really need
to be held constant through time? Can individuals, rather than government, be given
a larger role? And if so, how much larger?
Setting an appropriate future role for the Social Security system is not simple,
because ultimately much of the decision involves a value judgment, not an analytical
one, about what an appropriate retirement income “floor” would be. The baselines
discussed in this report pose the following question: if we could choose an “ideal”
policy, would we want Social Security benefits for future retirees to rise with the
nation’s rising standard of living or, alternatively, with its cost of living?
Having
them rise with the standard of living means raising them above the levels prescribed
by current law with higher commensurate costs than projected by the trustees.
Keeping them up with the cost of living means lowering them from the levels
prescribed by current law with lower commensurate costs.
In reality, policymakers are going to have to balance resource questions against
what they ideally want benefit levels to be. Much of the debate should be about what
an appropriate “floor of protection” is, but people have to pay for whatever they
choose. Thus, the question and the ultimate response inevitably have to be a blend
of (1) seeking appropriate benefit levels and (2) tolerable levels of taxes and/or
mandatory savings.
Before that stage, however, policymakers have to understand what the level of
current law benefits represents to future retirees in order to make informed judgments
about what an appropriate level of Social Security benefits might be. Simply gauging
changes against current law will not render that understanding. The two baselines
described here attempt to measure economic outcomes instead of legal ones. They
are not presented as alternatives to current law comparisons as much as supplements.
Current law comparisons illustrate the degree of change from previously established
policy. The baselines suggested here can be used to re-examine past policy and, in a
fresh sense, what level of benefits we should provide.

CRS-19
Appendix
Table 4. Illustrative Future Social Security Retirement Benefits and
Replacement Rates Computed Under Rules
Prescribed by Current Law
Replacement rate
(first year’s benefits
Year of retirement at
First year’s benefits
First year’s benefits
as % of final year’s
age 65
in current $s
in constant 1999 $s
earnings)
1995
$10,296
$11,243
43.5%
2000
$11,844
$11,602
39.9%
2005
$14,508
$12,384
41.3%
2010
$17,028
$12,375
39.6%
2015
$20,904
$12,915
39.5%
2020
$25,380
$13,331
39.1%
2025
$29,304
$13,086
36.7%
2030
$35,988
$13,663
36.7%
2035
$44,220
$14,272
36.7%
2040
$54,312
$14,903
36.7%
2045
$66,732
$15,567
36.7%
2050
$81,972
$16,257
36.7%
2055
$100,692
$16,977
36.7%
2060
$123,696
$17,730
36.7%
2065
$151,944
$18,516
36.7%
2070
$186,660
$19,338
36.7%
2075
$229,296
$20,196
36.7%
Source: CRS benefit computation model based on the long-range “intermediate” economic
assumptions of the 1999 Social Security trustees’ report. Illustrations are for benefits determined
under the computation rules of current law for full-career workers who consistently earn average
wages and retire at age 65.

CRS-20
Table 5. Illustrative Social Security Replacement Rates Computed
Under Rules Prescribed by Current Law, At Age 65 and the “Full
Benefit” Age, 1940 - 2070
Replacement rate (first year’s
“Full benefit”
benefits as % of final year’s
age
Year attaining age 65
earnings)
Retirement at
Retirement at
age 65
“full benefit” age
1940
26.2%
26.2%
65
1950
19.7%
19.7%
65
1960
33.3%
33.3%
65
1970
34.3%
34.3%
65
1980
51.1%
51.1%
65
1990
43.2%
43.2%
65
2000
39.9%
39.9%
65
2005
41.3%
42.5%
65, 6 months
2010
39.6%
42.2%
66
2015
39.6%
42.2%
66
2020
39.1%
42.1%
66, 2 months
2025
36.7%
42.1%
67
2030
36.7%
42.1%
67
2040
36.7%
42.1%
67
2050
36.7%
42.1%
67
2060
36.7%
42.1%
67
2070
36.7%
42.1%
67
Source: The 1998 Green Book, House Committee on Ways and Means, GPO, May 19, 1998; and the
1999 Social Security trustees’ report (for reference, see footnote 1). Illustrations are for benefits
determined under the computation rules of current law for full-career workers who consistently earn
average wages and retire at age 65 or the full benefit age.