Social Security Statements: Social Security Administration Should
Better Evaluate Whether Workers Understand Their Statements	 
(01-APR-05, GAO-05-192).					 
                                                                 
The Social Security Statement is the federal government's main	 
document for communicating with more than 140 million workers	 
about their Social Security benefits. By law, the statement must 
show an individual's annual earnings, payments into Social	 
Security and Medicare, and projected benefits. The Social	 
Security Administration also uses the statement to explain the	 
various types of Social Security benefits and to encourage	 
greater financial planning for retirement. GAO conducted a review
to examine (1) how well recipients understand the current	 
statement, (2) how the Social Security Administration is	 
evaluating the statement's understandability, and (3) the	 
promising practices used by private sector companies and other	 
industrial countries. GAO's information was obtained from its	 
national survey and focus groups of statement recipients, a firm 
that evaluates benefit statements, officials from three other	 
countries (Canada, Sweden, and the United Kingdom), and other	 
experts from the private sector.				 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-192 					        
    ACCNO:   A20619						        
  TITLE:     Social Security Statements: Social Security	      
Administration Should Better Evaluate Whether Workers Understand 
Their Statements						 
     DATE:   04/01/2005 
  SUBJECT:   Data collection					 
	     Federal records management 			 
	     Federal social security programs			 
	     Government information dissemination		 
	     Reporting requirements				 
	     Social security benefits				 
	     Surveys						 
	     SSA Personal Earnings and Benefits 		 
	     Estimate Statement 				 
                                                                 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO 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.                                         **
**                                                              **
******************************************************************
GAO-05-192

United States Government Accountability Office

GAO

                       Report to Congressional Requesters

April 2005

SOCIAL SECURITY STATEMENTS

Social Security Administration Should Better Evaluate Whether Workers Understand
                                Their Statements

GAO-05-192

[IMG]

April 2005

SOCIAL SECURITY STATEMENTS

Social Security Administration Should Better Evaluate Whether Workers Understand
Their Statements

                                 What GAO Found

Participants in our review identified both strengths and weaknesses in the
current statement's understandability. Many respondents to GAO's national
survey recalled receiving a statement, but had little recollection of some
components, for example, the information on Social Security's future.
Focus group participants provided more detailed information; they found
the statement to be comprehensive but less well presented than a
comparison statement they also reviewed. A firm that evaluates benefits
statements had similar conclusions. The firm rated the quality of primary
content and said the general understandability of the statement favorably
compared with that of other statements, but use of design to help convey
information and quality of secondary content fared less well.

The Social Security Statement's Evaluation Scores Compared with Industry
Average Scores

Criteria

Quality of primary content

Understandability

Absence of undesirable design features 93

Presence of desirable design features

Quality of secondary content

0 20 40 60 80 100

Percentage of maximum total points achievable

Social Security Statement Industry average

Source: GAO presentation of Dalbar's data.

The Social Security Administration's current evaluation of the statement's
understandability is limited because it does not include focus groups or
data from the agency's many public contacts. For feedback, the agency
relies almost exclusively on an annual survey covering many aspects of the
Social Security program. Its questions about the statement are general and
change each year, limiting their effectiveness. The agency also does not
routinely use data collected from such sources as its telephone call
centers, walk-in traffic, or Web site to help determine whether the
statement is meeting its goals.

Private sector experts and countries GAO studied use several practices the
Social Security Administration may find helpful, such as regularly
gathering feedback from statement recipients, customizing messages for
different age groups, and changing statements every few years to keep
readers interested. They also tended to design their statements in ways
GAO's focus groups preferred-for example, putting the most important
information at the start.

                 United States Government Accountability Office

Contents

Letter

Results in Brief
Background
Statement Found to Have Both Strengths and Weaknesses in Its

Understandability
SSA's Current Evaluation of the Statement's Understandability Is
Limited
Other Countries Studied and Private Sector Benefits Consultants

Use Several Promising Practices That May Be Helpful to SSA
Conclusion
Recommendations for Executive Action
Agency Comments

                                       1

                                      3 5

                                       8

22

25 29 29 30

Appendix I Court Rulings on Rights to Social Security Benefits 31

The Precedent-Setting Decision 31
Other Challenges under the Due Process Clause 33
Gender-Based Discrimination Decisions 34
Changes to Benefits as a Result of Social Security Reform 35
Summary 37

Appendix II Scope and Methodology

Appendix III 2004 Social Security Statement

Appendix IV Examples of Private Sector Retirement Statements

Appendix V Example of a Statement from Canada

Appendix VI Comments from the Social Security Administration 55

Appendix VII GAO Contacts and Staff Acknowledgments 58

Contacts 58 Staff Acknowledgments 58

Table

Table 1: Percentage of Survey Respondents Who Remembered Elements of the
Social Security Statement

Figures

Figure 1: Examples of Technical Terms That Focus Group Participants Found
Confusing 11 Figure 2: Excerpts of Explanations That Focus Group
Participants Found Confusing 13 Figure 3: Examples of Explanations That
Focus Group Participants Found Contradictory 15 Figure 4: Statement's
Evaluation Scores Compared with Industry Average Scores 16 Figure 5:
Example of Related Information Located in Different

Parts of the Statement 18 Figure 6: Example of Chart in the Private Sector
Statement 19 Figure 7: Potential Illustration Showing Monthly Social
Security

Retirement Benefit by Age, from 62 to 70 20 Figure 8: Graphic Used by SSA
to Illustrate Demographic Changes Affecting Social Security 21 Figure 9:
Example of Supplemental Information Included by Canada in an Insert 26

Abbreviations

PEBES Personal Earnings and Benefit Estimate Statement
PUMS Public Understanding Measurement System
SSA Social Security Administration

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

United States Government Accountability Office Washington, DC 20548

April 1, 2005

The Honorable Tom Delay
Majority Leader
House of Representatives

The Honorable Jim McCrery
Chair, Subcommittee on Social Security
Committee on Ways and Means
House of Representatives

The Honorable E. Clay Shaw, Jr
House of Representatives

Social Security, the nation's largest federal program, touches the lives
of
virtually all the nation's citizens. It provides a measure of economic
security and financial stability by providing retirement, disability, and
survivorship benefits. The Social Security Administration (SSA) provides
benefits information as well as other related information on an annual
statement to virtually all American workers. The Social Security Statement
(the statement) represents one of the federal government's largest efforts
to communicate directly with workers about their Social Security benefits.
In 2003, SSA mailed statements to more than 140 million people at a cost
of approximately $45 million. Because the statement is mailed to so many
homes, the Social Security Advisory Board recommended in 1997 that it
should be one of the highest priorities of the agency and receive careful
and high-level attention with respect to content and design.1 Since then,
SSA has made some significant revisions, based in part on our work, to
improve the statement's clarity and understandability.2 In 2000, the
Social
Security Commissioner testified that the statement is the most significant
vehicle that SSA has to increase the public's understanding of the basic

1The board is independent and bipartisan and advises the Commissioner of
Social Security on policies related to the Social Security and the
Supplemental Security Income programs.

2GAO, SSA Benefit Statements: Well Received by the Public but Difficult to
Comprehend, GAO/HEHS-97-19 (Washington, DC: December 5, 1996), and GAO,
Social Security: Providing Useful Information to the Public,
GAO/T-HEHS-00-101 (Washington, DC: April 11, 2000).

features of Social Security and enable Americans to prepare for long-term
financial security.

The individualized nature of the statement and the diverse population it
serves make developing a clear and understandable document a complex
process. Statements must correctly list each individual's annual earnings
history as well as future projections of retirement, disability, and
survivorship benefits. The statement must also be understandable to an
audience that varies considerably in age, educational level, and cultural
background. In addition, the production and evaluation of the statement
involve several SSA offices and contractors. Further, the statement may be
taking on added importance as a tool for communicating to the public
potential changes in Social Security. With the baby boom generation's
forthcoming retirement, longer life spans, and lower birthrates, Social
Security's financing shortfall will grow. The President has identified the
reform of the Social Security system as a major agenda item for his
administration, and if reform is enacted, educating the public about any
program changes and how they will affect benefits will likely be a high
priority for SSA.

Given the statement's importance, you asked us to review SSA's efforts to
ensure its understandability and to do so in the context of what
information private sector employers are providing to their employees and
what other industrialized countries provide to citizens about their public
pension plans.3 Specifically, our objectives were to address (1) how well
recipients understand the current statement, (2) how SSA is evaluating the
statement's understandability, and (3) the promising practices used by
private sector companies and other industrialized countries that could be
used by the Social Security Administration. In addition, you asked us to
review a person's legal rights to benefits. Appendix I summarizes the
applicable caselaw regarding an individual's rights to Social Security
benefits.

To answer the above questions, we collected data from a wide variety of
sources. We assessed the statement's understandability and usefulness in
conjunction with a GAO national random telephone survey on financial

3Private sector defined benefit plans (plans that promise to provide a
benefit that is generally based on an employee's salary and years of
service) are required to furnish plan participants, upon request, a
statement of their accrued benefits and total nonforfeitable pension
benefits, if any, which have accrued, or the earliest date on which
benefits become nonforfeitable.

literacy. We also contracted with a private market research firm, OneWorld
Communications, to conduct six focus groups that reviewed the statement's
content and design and compared them with those of a private sector
benefit statement. These focus groups were held in Seattle, Chicago, and
Atlanta and were balanced for age, gender, and ethnicity. We also
contracted with a nationally recognized benefit statement evaluation firm,
Dalbar Inc., to evaluate and score the statement based on the firm's
criteria for private sector benefit statements. We interviewed SSA
officials to discuss their procedures for evaluating the statement. To
gather data about private sector benefit statements and statements used by
other industrialized countries, we interviewed private companies that
specialize in benefit statement preparation and distribution as well as
public pension officials in Canada, Sweden, and the United Kingdom. To
determine the legally required information in both the Social Security
Statement and private pension benefit statements, we reviewed applicable
laws. We also analyzed caselaw on legal rights to benefits under the
Social Security program. Appendix II explains the scope and methodology of
our work in greater detail. We conducted our work between March 2004 and
March 2005 in accordance with generally accepted government auditing
standards.

The respondents to our survey, as well as the focus group participants and
the benefits consulting firm that we asked to review the statement,
identified both strengths and weaknesses in its understandability. About
66 percent of people responding to our survey remembered receiving the
statement, though some parts-for example, the benefit estimates and
earnings history-tended to be remembered much more than the information
about Social Security's future. Focus group participants generally found
their earnings history and benefit estimates understandable and useful.
However, they found some of the other information confusing or
contradictory-for example, the role credits play in determining benefit
eligibility and the role earnings play in calculating benefits. To be
eligible for benefits, workers must have earned enough income over their
working lifetimes to qualify for at least 40 credits.4 However, once
eligible, the worker's actual benefit is not determined by credits but by
a measure of lifetime earnings.5 Many focus group

4A total of 4 credits can be earned in each year. In 2004, 1 credit was
awarded for each $900 of wages or self-employment income.

5In calculating retirement benefits, SSA averages an individual's 35
highest years of earnings.

Results in Brief

participants were still confused about these complex concepts after
reading their statements. They were also confused about the statement's
explanation of Social Security's solvency challenges. When participants
were asked to compare the Social Security Statement with a private sector
benefit statement, many found the Social Security Statement more
comprehensive in scope but lacking in helpful design features, such as the
use of color and graphics. Some participants also preferred the private
sector statement's presentation of the need for supplemental income in
retirement. The benefits consulting firm that reviewed the statement gave
it a communication effectiveness score of 65 out of 100, which fell below
the industry average score of 73. The firm found that the statement was
strong in its presentation of benefits and use of plain language but could
be improved through better design, targeted messaging, and more
information on income needs in retirement.

SSA's current methods of evaluating the statement are limited. To date,
SSA has evaluated the statement primarily through an annual survey of the
general public. This survey is conducted for many other purposes besides
evaluating the statement, and the survey's questions specifically about
the statement have been general and have changed from year to year.
Assessing the statement through this limited approach may not provide the
depth and detail needed to comprehensively assess the statement's
understandability. Also, while SSA collects and analyzes a variety of data
about its services, it is not systematically collecting any data from
internal sources, such as the customer call centers, walk-in traffic, and
its Web site, about whether the statement is understandable or meeting its
goals. For example, although correction of inaccurate earnings records is
a goal of the statement and SSA officials reported that they corrected
360,000 earnings records in 2003, no data were collected to determine how
many corrections resulted from receiving a statement.

Private sector benefits consultants and other industrialized countries
utilize several promising practices that may be helpful to SSA, such as
regularly gathering feedback from statement recipients and utilizing
customized messages. All of the consultants and officials from other
countries that we spoke with regularly and systematically collect feedback
from recipients regarding the usefulness and understandability of their
statements. For example, officials from Sweden told us that they utilized
what they called "hot surveys," which are surveys that are sent out within
a month of when the statement is sent to specifically assess the
usefulness and understandability of the statement. This survey technique
allows the recipients to give feedback about the statement while their
impressions regarding the statement are relatively current. Benefits
consultants and

other countries also utilize customized messages that differ depending on
the age of the recipient. For example, officials in the United Kingdom
told us that younger workers receive messages emphasizing the importance
of starting to save for their retirement early in their careers and
workers closer to retirement receive messages regarding how to maintain
their current lifestyle, including increasing personal savings and working
longer. Officials at both the consulting firms and in other countries told
us that workers are more likely to read and remember information that is
relevant to them. Finally, consulting firm officials told us that
regularly modifying the statement's appearance could help make the
statement more effective, since recipients would be less likely to assume
that they are receiving the same old statement year after year.

This report contains recommendations for executive action. SSA should
develop a plan for regularly evaluating the statement through the
collection of data from multiple sources, including information from
surveys, focus groups, call centers, walk-in traffic, and the Web site. On
the basis of data it receives and a review of promising practices in the
private sector and other countries, SSA should consider revising the
statement to include showing the personalized benefit information first
and using more graphics. In its response to our draft report, SSA agreed
to develop a plan for regularly evaluating the statement and will consider
revising the statement as appropriate and necessary. (SSA's comments are
reproduced in app. VI.)

SSA is required by law to provide annual statements with benefits and
earnings information to individuals over the age of 25 who have a Social
Security number and have wages or earnings from self-employment.6 The law
requires each statement to contain the following:

o  	an estimate of the potential monthly Social Security retirement,
disability, survivor, and auxiliary benefits and a description of the
benefits under Medicare;7

o  	the amount of wages paid to the employee or income from
selfemployment;

Background

642 U.S.C. 1320b-13.

7The law requires that only the statements sent to people aged 50 and
older contain actual benefit estimates, but SSA provides benefit estimates
regardless of age.

o  an estimate of the individual's aggregate taxes paid to Social
Security; and

o  an estimate of the individual's aggregate taxes paid to Medicare.

Appendix III contains an example of the current statement with the legally
required information highlighted.

The requirement to mail the statements was phased in beginning in fiscal
year 1995, when SSA was required to mail the annual statement-then named
the Personal Earnings and Benefit Estimate Statement (PEBES)- to all
workers age 60 and older. In addition to the informational requirements,
Congress required SSA to send the annual statement to selected age groups
of workers not receiving benefits and for whom a mailing address can be
determined. Starting in fiscal year 2000, SSA was required to mail the
annual statement, now called the Social Security Statement, to workers age
25 and older. However, SSA officials decided to mail a larger number of
statements earlier than fiscal year 2000 to phase in the added workload
over several years. The accelerated schedule began in fiscal year 1996,
when SSA added individuals turning 58 and 59. In subsequent years, SSA
added increasingly younger people until people as young as 25 began
receiving the statement in fiscal year 2000.8 These statements are
generally mailed about 3 months before the worker's birthday.

The current Social Security Statement has evolved over several years. The
initial PEBES was a six-page document and contained information such as
the worker's earnings record, benefits estimates and a question-andanswer
section about Social Security. However, in a previous report, we found
that PEBES did not clearly communicate the complex information that
workers needed to understand about SSA's programs and benefits.9 In
response, SSA made significant changes to the format and presentation of
the PEBES, tested a four-page Social Security Statement in 1997, and began
mailing it to the public in October 1999. The newer statement was shorter,
better organized, and easier to read than the PEBES, but our follow-up
review in 2000 identified some remaining rough spots. The January 2005
statement retains the four-page organization of the 1999 statement and
generally used green underlining to highlight the page headings for each
page of the statement. Although the statement has

8The Social Security Statement replaced PEBES in fiscal year 2000. 9See
GAO/HEHS-97-19.

evolved over the years, some information in the statement has been
included since it was first mailed to the public, such as the description
of Social Security programs and the financial future of Social Security,
including the long-term financial challenges that the system faces. To
complement the statement, SSA created a special one-page insert for
workers age 55 and older that was first mailed in October 2000. This
insert is mailed with the statement and provides information to assist
workers in making decisions about when to retire.

In addition to complying with the legal requirements, SSA has established
three agency goals for the statement-to educate the public about benefits
under SSA programs, to aid in financial planning, and to ensure the
worker's earnings records are complete and accurate. SSA believes that its
statement is providing useful information to the general public. The
following goals have remained constant since 1995, when SSA began
designing earlier versions of the statement:

o  	To educate the public about SSA programs, the statement contains
information about Social Security benefits. The statement provides
information about various benefits, including retirement, disability,
family, and survivor benefits.

o  	To aid financial planning, the statement contains information about
planning for retirement. For example, the statement says that Social
Security was never intended to be the only source of retirement income and
that an individual needs other savings, investments, pensions, or
retirement accounts to ensure having enough money to live comfortably. The
statement also contains information about working and still receiving
Social Security retirement benefits and makes reference to an SSA booklet
to help determine the best time to retire.

o  	To verify earnings, the statement asks the worker to review the
earnings information and make certain that the information is complete
because Social Security benefits are based on earnings.

Statement Found to Have Both Strengths and Weaknesses in Its
Understandability

The respondents to our survey, as well as the focus groups and the
benefits consulting firm we asked to review the statement, identified both
strengths and weaknesses in its understandability. Our survey provided
general information about the statement in terms of how many people
recalled receiving it and what types of information they remembered.
Specifically, about 66 percent of survey respondents remembered receiving
their statement.10 However, more respondents remembered seeing the
statement's personalized information, such as benefit estimates and
earnings history, than remembered seeing the more generic information.
Feedback from the focus group participants provided detailed insight about
the areas of the statement they understood and how confusing information
might be improved. For example, focus group participants generally found
their earnings records and benefit estimates understandable and useful but
had difficulty with information about how the program operated and what
its long-term prospects were. When participants were asked to compare the
Social Security Statement with a private sector benefit statement, many
found the Social Security Statement more comprehensive, but they preferred
many aspects of the private sector statement. In particular, participants
liked the private sector statement's use of color and graphics. The
benefits consulting firm, which evaluated the statement against the
industry's best practices, scored it lower than the industry average for
defined benefit statements. Officials from the firm said the statement was
strong in its presentation of benefits and use of plain language but could
be improved through better design, targeted messaging, and more
information on income needs in retirement.

What Works Well in the Statement?

A majority of respondents in our national survey indicated a familiarity
with the statement and could recall the personal information it contained.
Our national telephone survey asked respondents if they remembered
receiving a statement, whether they recalled specific sections of the
statement, and how understandable and useful they found each section.
Survey results indicated that approximately 66 percent of respondents who
were eligible to receive a statement remembered receiving it.11 Of the

10Percentage estimates for respondents in our survey are based on a sample
and are subject to sampling error. Unless otherwise noted, estimates from
our survey have 95 percent confidence intervals of plus or minus 5
percentage points of the estimate. See appendix II for more information.

11Sixty-four percent of respondents remembered receiving their statement.
Those who did not remember or who did not know or were unsure were
prompted with a further description of the statement. Fourteen percent of
those who were prompted then remembered receiving the statement.

five sections of the statement-explanation of Social Security benefits,
estimated monthly benefits upon retirement, earnings per year, general
facts about Social Security benefits, and the statement on the future of
Social Security-the section on earnings per year was, we believe, the
easiest for most people to understand.12

The six focus groups conducted by our focus group moderator generally
reported that except for several key items that were confusing or hard to
understand, the information presented in the statement was understandable.
These groups, composed of people of various ages and educational
backgrounds, were instructed by the moderator to carefully read each page
of the statement, which participants admitted is less likely to happen
when they receive the statement in the mail. Participants found the
benefit estimates and earnings history understandable and frequently
commented that it was the information they were most likely to look at
when they received the statement in the mail.13 Some of the older
participants found the benefit estimates particularly useful, since their
estimates were based on a longer earnings history and were likely to be
more accurate than those for younger participants with fewer work years.
Participants appreciated the opportunity to review their earnings history,
and many said they would follow the directions in the statement and call
SSA's 800 number if they discovered errors. Many participants found the
total amount of Social Security and Medicare taxes they paid to be
understandable.

Several of these themes were echoed by the benefits consulting firm we
asked to evaluate the statement. This firm, which specializes in
evaluating the effectiveness and consistency of written materials from
financial services institutions to their customers, reviewed and scored
the statement based on the firm's standardized criteria. The firm gave the
statement an overall score of 65 out of 100 based on these criteria. By
comparison, the average score of private sector statements in the firm's
database was 73. The firm identified the statement's strengths as

o  being written in easy-to-understand language;

12This result is statistically significant even after accounting for the
multiple comparisons that can be made between proportions across the five
sections. For more information see appendix II.

13Workers who do not have 40 credits do not receive benefit estimates but
are told how many credits they have and how many they will need before
they become eligible for benefits.

o  	providing benefit estimates for different retirement ages as well as
for disability and survivorship;

o  identifying the statement's source and purpose;

o  	providing the personal data used to calculate benefits, such as date
of birth and earnings record, so recipients can check information for
accuracy

What Does Not Work Well? 	The survey respondents who remembered receiving
a statement varied considerably in the extent to which they remembered the
different types of information that SSA was trying to communicate. For
example, of the people who said they remembered receiving a statement,
about one-fifth remembered seeing all the major areas of information on
the statement. As table 1 shows, of all the statement's information,
respondents least remembered the information on the future of Social
Security.

Table 1: Percentage of Survey Respondents Who Remembered Elements of the
Social Security Statement

Percentage of respondents Element of the statement who recalled seeing the
element

                    Explanation of Social Security benefits

Estimated monthly benefits upon retirement

Earnings per year

General facts about Social Security benefits

The statement on the future of Social Security

Source: GAO survey data.

Many focus group participants identified some of the statement's
information about the Social Security program as confusing. In some cases,
participants did not understand the actual meaning of a word or phrase,
such as "actuary" and "intermediate assumption" (see fig. 1 for the
sentence in which the terms appear). Also, the meaning of the phrase

"compact between generations," used to describe the pay-as-you-go nature
of Social Security, was unclear to many.14

Figure 1: Examples of Technical Terms That Focus Group Participants Found
Confusing

Source: GAO representation of information from the 2004 Social Security
Statement and focus group participants.

Besides not understanding certain terms, participants across the focus
groups also did not understand explanations of certain concepts discussed
in the statement. For example, many participants were confused about the
role of credits in determining their eligibility for retirement benefits
versus the role of earnings in determining their actual retirement
benefit. As figure 2 shows, these are very complex issues to explain in
brief and straightforward language. While SSA has made changes to improve
explanations of concepts, our focus group results suggest there is a need
for further refinements. Finally, information that applies only to certain
groups of workers confused some participants. For example, a couple of
participants questioned whether they needed to request the free
publication about the windfall elimination provision because the
statement's description of this publication is listed with other
publication

14Pay-as-you-go means the taxes paid by today's workers and their
employers are used to pay the benefits for today's retirees and other
beneficiaries.

titles that pertain to all workers and the statement does not make it
clear that this provision primarily affects government workers.15

15The windfall elimination provision affects the calculation of retirement
or disability benefits if the worker received a pension from work where
Social Security taxes were not taken out of his or her pay. A modified
formula is used to calculate the benefit amount, resulting in a lower
Social Security benefit.

Figure 2: Excerpts of Explanations That Focus Group Participants Found Confusing

Source: GAO representation of information from the 2004 Social Security
Statement and focus group participants.

The statement's information about the financial stability of the Social
Security system was confusing and a source of concern for most focus group
participants. Virtually all participants said they were at least generally
aware that Social Security is facing fiscal challenges. Several voiced
concerns that their estimated benefits might not be available when they
are ready to retire, given the statement's information about the
exhaustion of the trust funds and other caveats on future benefits, such
as

being estimated in today's dollars and based on current law. Many also
found the reassurances on page 1 about Social Security being available
upon retirement followed by the information about it facing serious
financial problems to be contradictory (see fig. 3). The statement's
explanation that Social Security will still be able to pay 73 cents for
each dollar of scheduled retirement benefits even though the trust funds
will be exhausted in 2042 confused many participants.16 While most
participants felt it was important to have information in the statement
about Social Security's future insolvency, phrases such as "unless action
is taken soon" and "we will need to resolve these issues soon" did not
provide the information many felt they needed to understand the problem
and what personal action, if any, they were expected to take.

16Social Security is now taking in more money than it is paying out in
benefits. The surplus taxes are going into a trust fund, which will begin
to be used around 2018 when benefits will begin to exceed the amount of
Social Security taxes collected. The trust funds will then be used to help
pay benefits until around 2042, at which time the trust funds will be
depleted. After that, Social Security taxes will only cover about 73
percent of beneficiaries full benefits under the current system.

Figure 3: Examples of Explanations That Focus Group Participants Found
Contradictory

Source: GAO representation of information from the 2004 Social Security
Statement and focus group participants.

The benefits consulting firm that evaluated the statement, like the focus
group participants, found the statement was weak in design and conveying
financial planning information. The firm designated the statement as good
and awarded it points in five standardized evaluation categories.17 Figure
4 shows more specifically how this score was divided among the five. The
statement scored above or near the industry average in quality of primary
content and understandability. However, the statement scored below the
industry average in quality of secondary content and the two categories
related to design features.

17Dalbar designates statements that score 80 to 100 points as excellent,
statements that score 70 to 79 points as very good, and statements that
score 60 to 69 points as good.

Figure 4: Statement's Evaluation Scores Compared with Industry Average
Scores

Criteria

Quality of primary content

Understandability

Absence of undesirable design features 93

Presence of desirable design features

Quality of secondary content

0 20406080

Percentage of maximum points achievable

                           Social Security Statement

                                Industry average

Source: GAO presentation of Dalbar's data.

Specific reasons the statement lost points when compared with industry
best practices include the following:

o  	The statement does not compare the retirement benefit with how much
income a person may need in retirement or offer suggestions and strategies
for meeting income goals through other sources of retirement income.

o  	The design does not clearly identify the most important information or
easily lead the reader through the document. For example, the retirement
benefit information is not prominent, and the statement uses multiple
styles of presentation within a single page.

o  	Much of the statement is relevant to only certain groups of
recipients, or if relevant to all recipients, the information is not
customized appropriately for differing audiences. Irrelevant content can
create confusion and reduce readership.

o  	The statement has some repetitive content and uses text instead of
graphics, such as charts and tables. It also contains numerical data,
other

than a citizen's actual benefit, that may be confusing or result in a
failure to communicate critical information.

How Much Information Do People Want?

The focus group participants had mixed opinions about the amount of
information the statement should contain. On the one hand, some
participants felt that the statement contained so much information that it
reduced their interest in reading it. These people claimed they generally
reviewed the information on pages 2 and 3-their benefit estimates and
earnings history-and filed the statement away. On the other hand, some
participants believed that the statement introduced certain issues but
then did not provide an adequate explanation. For example, some
participants wanted more information on survivorship benefits,
specifically children's eligibility ages and whether statement estimates
were per child or a total amount for all children. Many participants
wanted retirement benefit estimates for additional ages, specifically 63,
64, and 66.18 Also, while participants generally understood that the
retirement age was rising, they felt the information about the full
retirement age for people born between 1938 and 1960 was missing. Finally,
many participants had questions about information presented early in the
statement (for example, disability benefits) that information on the
statement's last page answered. When SSA redesigned the statement in 1999,
it reduced the need for the reader to flip to other pages to find related
information. However, our focus group responses suggest that more could be
done in this area (see fig. 5).

18The Social Security Statement currently provides retirement benefit
estimates for 62, the individual's full retirement age, and 70.

Figure 5: Example of Related Information Located in Different Parts of the
Statement

Source: GAO representation of information from the 2004 Social Security
Statement and focus group participants.

To make the statement be most effective, the benefits consulting firm
recommended keeping the information on the statement brief and highly
relevant to the individual, with clear directions on how to get more
information. The firm recommended avoiding information attempting to
encompass all recipients, since the readers generally ignore such
information. For example, the message on page 1 that "Social Security is
for people of all ages" does not provide any specific information for the
individual. The firm defined the most important content as the benefit
estimates, factors that may affect benefit amounts, and points of contact
that can provide more information.

How Do People Want the When asked to compare the Social Security Statement
with a private

Information Presented? 	sector statement, many focus group participants
recognized that the Social Security Statement had different and more
complex information to

convey, but they preferred the design of the private sector statement.19
Our focus group moderator asked participants to compare the Social
Security Statement with a private sector statement that had been rated as
excellent by the benefits consulting firm with which we contracted. The
focus group moderator asked participants to compare the two statements in
terms of length and presentation of information. Many participants favored
the way the private sector statement grouped information and used color
and graphics. Specifically, many liked the private sector statement's use
of a colored pie chart illustrating what percentage Social Security and
other retirement savings will be required to replace current monthly
salary. (This chart is shown in fig. 6). Although the focus group
participants preferred the use of color in the private sector statement,
they did not discuss the likely higher costs associated with producing
such statements. Some participants may not have preferred the use of color
in the statement if the cost of producing such statements was
significantly higher than statements with little or no color.

Figure 6: Example of Chart in the Private Sector Statement

What Specific Changes Were Suggested?

One common theme that emerged in the suggestions made by the focus groups
and the benefits consulting firm was to use graphics to replace text,
making some information more quickly and easily understandable.
Specifically, the benefits consulting firm said that people are more
likely to scan numerical data than to read large amounts of text. Figure 6
above shows a pie chart that is central to one of SSA's goals for the
statement- aiding in financial planning. This chart was something that
focus group participants said they liked about the private sector
statement, and the

19The private sector statement that focus group participants reviewed for
comparison is contained in appendix IV.

benefits consulting firm recommended it as a way to provide more context
on income needs in retirement.

Some focus group participants also said they would like to see benefit
estimates for the ages of 63, 64, and 66 in addition to the ages in the
statement. Figure 7 shows how this information might be depicted. It helps
quickly communicate that working longer may result in a higher benefit.

Figure 7: Potential Illustration Showing Monthly Social Security
Retirement Benefit by Age, from 62 to 70

Additionally, SSA is currently using graphics in its other publications to
illustrate many of the concepts explained with text in the statement. For
example, in its publication "The Future of Social Security," SSA uses a
graphic to illustrate the demographic changes contributing to Social
Security's financial insolvency (see fig. 8).

Figure 8: Graphic Used by SSA to Illustrate Demographic Changes Affecting
Social Security

Source: SSA, "The Future of Social Security" January 2004.

In addition to suggesting replacing some of the statement's text with
graphics, the focus group participants and benefits consulting firm made a
number of other specific recommendations about how to improve the
statement. On the basis of the participants' comments, the focus group
moderator recommended design improvements to help readers more quickly
find and absorb information; the suggestions included combining the
information on pages 1 and 4 and moving the personalized information on
pages 2 and 3 to earlier in the statement because it is the information
most useful to participants. The benefits consulting firm recommended
improvements that focus on meeting recipients' information needs,
increasing readership, minimizing unnecessary inquiries at SSA, and
increasing the savings rate among U.S. citizens. For example:

o  	Provide contextual information about income needs in retirement. For
example, give recipients a retirement goal, such as 80 percent of current
income. Then show how much of that goal Social Security provides and how
much must come from other income sources.

o  	Align the order of information with recipients' priorities. For
example, move the retirement benefit information from page 2 to page 1 and
add an "Important Changes to Your Benefits" section to the beginning of
the statement that documents any new changes to the recipient's
calculation of benefits.

o  Target messages to keep information highly relevant to the

SSA's Current Evaluation of the Statement's Understandability Is Limited

individual recipient. Specific messages could be based on whether the
person is receiving the statement for the first time, his or her age, or
potential changes to benefit calculations. For example, a 25-year-old
recipient may need information about how to begin saving instead of
information on how to apply for Medicare.

SSA has not identified any significant problems with the statement, but
its evaluation of the statement's understandability is limited and may not
be identifying problems that exist. SSA has relied almost exclusively on
questions in its annual Public Understanding Measurement System (PUMS)
survey, which assesses how much adults know about Social Security
programs, guides SSA's education and communication activities, and gauges
the awareness and effectiveness of the statement. Although these survey
results are generally useful, if they are used as the only evaluation,
they may not provide the depth and detail needed to assess the
understandability of each individual section of the statement. Past survey
questions about the statement have been very general, and the number and
types of questions have changed from year to year. Since the current
format was adopted in 1999, SSA has not used focus groups or other means
to obtain detailed participant views to evaluate the complete statement.
Additionally, while SSA collects a variety of data about its services, it
does not systematically collect data about the statement's
understandability and how well the statement is achieving its goals from
internal sources, such as its customer call centers, walk-in traffic, and
its Web site.

SSA Uses Survey to Evaluate Statement, but Its Survey Does Not Effectively
Identify Problem Areas

Although the PUMS survey has been SSA's primary tool for evaluating the
statement in 6 of the last 7 years, for several reasons it is difficult to
evaluate the statement's understandability from the data SSA receives.20
First, the questions about the statement are very general, making it
difficult to obtain any detailed information. In its 2000 report, SSA's
contractor for the PUMS survey recommended that SSA consider a separate
project that would address all aspects of the statement and track customer
satisfaction and use. However, while the 2004 survey asks respondents if
they remember certain parts of the statement, it does not ask them if the
information was understandable or useful.21 Second, it is difficult to
identify trends in the survey responses because the number and type of
questions specifically about the statement have changed each year. For
example, from 1999 through 2001, the PUMS survey asked respondents to rate
the overall understandability of the statement, but neither the 2003 nor
the 2004 survey asked the same question. Similarly, a question was added
about satisfaction with the statement in the surveys for 2000 and 2001,
but this question did not appear in the 2003 or 2004 survey. SSA officials
said they eliminated these questions because there was a high percentage
of positive responses, and because the statement had not been markedly
altered in recent years, they did not expect the percentages would have
changed significantly. Third, since the survey is not conducted in
conjunction with any particular mailing, respondents may have received the
statement months before being surveyed about its contents and may not
recall certain portions of the statement or even remember receiving it. In
2003, 35 percent of survey respondents said they did not receive a
statement even though every eligible worker 25 and older should have been
mailed an annual statement since 1999. With over onethird of survey
respondents not receiving or not remembering they received a statement, it
is difficult to use survey results to fully assess the statement.

20The PUMS survey was not conducted in 2002. 21Results of the 2004 survey
were not available at this writing.

SSA's Use of Focus Groups for the Statement Has Been Limited

Since the statement was first mailed to workers in 1999, SSA has not
conducted focus groups on the entire statement or its preferred format.22
The current four-page statement was originally developed with the aid of
focus groups and a national public opinion survey. The focus groups and
survey respondents were asked to review a (1) one-page prototype, (2) a
one-page prototype with a pamphlet, (3) a four-page prototype, and (4) a
self-mailer prototype.23 SSA's survey report indicates that all four
prototypes were an overwhelming improvement over the six-page PEBES.
However, no clear preference for a particular format emerged. The fourpage
format was ultimately selected, as it had a higher clarity rating for
explanations of technical issues and the percentage of respondents with
questions was somewhat lower than for any other prototype. In addition,
respondents tended to indicate a preference for the four-page length.
Since the selection of the four-page format, SSA has not had any focus
groups reevaluate the entire statement. Additionally, SSA officials said
they do not have plans to use focus groups unless they are going to make
major changes to the statement.

SSA Is Not Examining Data from Internal Sources

SSA is not systematically examining data from internal sources, such as
its customer call centers, walk-in traffic, or its Web site, about the
statement's understandability or how well the statement is meeting its
goals. SSA officials said that informal feedback comes from its 800
number, the employee suggestion program, and links on its Web site.
However, SSA does not track specific data about the statement from these
sources. For example, although correction of inaccurate earnings is one of
SSA's primary goals for the statement, the agency does not have any way of
knowing or reliably estimating how many of the reported 360,000 earnings
records that were corrected in 2003 were corrected as the result of
receiving a statement. Additionally, SSA's 2003 service quality report
indicated that the third most common reason people called the 800 number
was because of the statement, but no data are available about the specific
reasons for the calls. Currently, SSA's 800 number telephone menu includes
a selection option for people who are calling because they recently
received their statement. However, if callers choose the selection

22SSA has conducted focus groups on other specialized topics relating to
the statement, such as creating an Internet PEBES (1998), revising the
insert to the statement (2000), creating age-specific messages for the
statement (2001), and revising PUMS questions (2004).

23A self-mailer is a mailing that is self-contained and does not require a
separate envelope.

Other Countries Studied and Private Sector Benefits Consultants Use
Several Promising Practices That May Be Helpful to SSA

option for questions about their statement and then mention that they
found a mistake on their earnings history, SSA codes the call in the
earnings category without any documentation that it was originally
prompted by receipt of a statement. Similarly, while the SSA Web site
gives users the option of selecting the statement as the subject of an
inquiry, SSA does not formally review questions from the Web site about
the statement for trends.

Other countries we studied, as well as the private sector benefits
consultants we spoke with in this country, use several promising practices
that may be helpful to SSA to ensure that the statement is useful. All of
them regularly and systematically collect feedback from statement
recipients regarding the usefulness and understandability of their
statements. Many of them also utilize customized messages that differ
depending on the age of the recipient. Finally, many of the benefits
consultants recommended significantly modifying the statement every few
years so that recipients would be less likely to assume that it is exactly
the same statement that they received the prior year.

Other Countries We Studied and Private Sector Benefits Consultants
Regularly and Systematically Gather Feedback Regarding Their Statements

Officials of the entities we studied use various systematic methods of
collecting recipient feedback to assess the understandability and
effectiveness of their benefit statements. For example, officials from
Sweden told us they use hot surveys, which are surveys that are sent out
within a month of when the statement is mailed, and ask questions
regarding the statement's understandability and usefulness. They said the
results of these surveys have resulted in slight changes to their
statements each year and that before changes are made, they are tested in
a focus group. Swedish officials told us that this combination of surveys
and focus groups helps to improve understandability. Sweden also uses
design consultants from the private sector to make sure that its
statements conform to industry best practices.

The benefits consulting firms that we spoke with also attempt to assess
the understandability of their statements through participant feedback.
For example, one consulting firm includes a detachable postage-paid survey
card that asks a few questions regarding the participant's perception of
the statement, including whether the participant found it to be
understandable. Although even good response rates for this type of
feedback mechanism are often no higher than 7 to 12 percent, the feedback
received served as an indicator as to whether or not the statements were
effective in meeting their objectives and provided insight

into what changes could be made to improve their understandability and
usefulness.

Other Countries We Studied and Private Sector Benefits Consultants Use
Design Enhancements, Customized Messages, and Innovative Delivery Options

Officials from benefits consulting firms told us that they increase the
usefulness of their statements by using white space, graphics, and colors,
as well as by ensuring that their statements are not too text-intensive.
The consultants told us their focus group participants found that the use
of white space and graphics made the statement easy to read. Officials
from Canada also told us about the importance of using white space as a
way to make the statement more effective, and they told us that one way
they maintain an ample amount of white space is by placing only the most
important information, such as earnings history and benefit estimates, on
the statement itself. Other, more supplemental information, such as
detailed information on the importance of saving for retirement, is
presented on inserts that are included with the statement. In addition to
including inserts, they also limit the amount of text on their statements
by providing the Web address and phone numbers that the recipient can use
to obtain additional information instead of trying to put this
supplemental information on the statement. Figure 9 shows an example of
the type of supplemental information that Canada includes in an insert to
its statement.

Figure 9: Example of Supplemental Information Included by Canada in an
Insert

Taxes and your pension

Your CPP retirement pension is taxable. We will not deduct income tax from
your monthly payment unless you request it. Once we have approved your
pension, we will send you a letter stating the amount of your pension. It
will give you information on how to request tax deductions.

Source: Canada Pension Plan (Social Development Canada).

Benefit consultants also recommended keeping the statement simple and
providing supplemental information through other means. For example, one
benefits consultant limits the information on the statement itself only to
personal information, such as projected benefit amounts and vesting
status, and provides supplemental information through a Web site, which is
prominently advertised on each page of the statement. The benefits
consultants told us that their surveys and focus groups have indicated
that participants prefer a simple statement over a statement that is
textintensive, as long as there are easily accessible resources available
for

further information. For SSA, using these design elements would probably
mean (1) less text and (2) greater reliance than at present on directing
recipients to the SSA Web site for additional information. However, SSA
recognizes that not all workers have regular and consistent access to the
Internet.

In addition to keeping the statement as simple as possible, most of the
benefits consulting firms that we spoke with echoed some of the
recommendations that our focus groups and contracted benefits consulting
firm made. For example, many of them said placing all of the personal
information, such as benefit projections and earnings history, at the
front of the statement is preferable to spreading it throughout the
statement. They said recipients are more likely to become interested in
the statement if the most personalized information is shown on the first
page. Consultants at one firm acknowledged that some recipients would
simply read the first page and disregard the rest, but they said they have
found that more people are likely to keep reading the information beyond
the first page once the personal information has gotten their interest.
According to the benefits consultants we spoke with, SSA could increase
the effectiveness of the statement by putting some, if not all, of the
personalized information toward the front of the statement and providing
additional resources for more information.

Benefits consultants in this country and agencies in other countries also
use customized messages that tailor their statements for each recipient.
For example, the United Kingdom provides different messages to workers
based on their age. Younger workers receive messages emphasizing the
importance of starting to save for their retirement early in their
careers. In contrast, workers closer to retirement receive messages
regarding how to maintain their current lifestyle, including increasing
personal savings and working longer. Officials in the United Kingdom told
us that using these customized messages increases the likelihood that the
recipients will read and understand the information contained on the
statement since it is more applicable to them. Providing information that
is customized for each recipient would represent a change for SSA, which
currently provides all recipients with information, such as the windfall
elimination provision, that is applicable to only a small number of
individuals. The benefits consultants that we spoke with told us that
including information on a statement that is not applicable to certain
recipients could confuse the recipients and cause them not to finish
reading the statement or to wonder if something is applicable to them when
it is not. The benefits consultants also told us that statement recipients
are more likely to take action, such

as saving more for retirement, if they can easily understand the statement
they receive.

Finally, these other producers of retirement benefit statements use
various innovative delivery methods. One involves customizing the length
of time between statements. For example, Canadian officials said that they
formerly provided their statements every year to all recipients but now
distribute their statements less often through what they call "smart
mailing" and have made their statement available for viewing at any time
over the Internet. Canada provides a statement only to individuals who
meet certain criteria, such as joining the workforce or turning a certain
age. For example, individuals who have just joined the workforce and have
made contributions to the public pension system receive a statement
welcoming them to the public pension system, providing them with limited
information about the system, and informing them about where to go for
further information. After this initial statement, workers would not
receive another one for 3 to 4 years, at which point they would receive a
more detailed statement that does provide a benefit estimate. Canadian
officials told us that their approach saved money, since postage is the
majority of each statement's cost, and also was more effective in that
when a statement did arrive, recipients were more likely to read it.

Another idea for consideration came from several benefits consultants that
we spoke with, who said that when a statement is distributed annually,
periodically changing the statement's appearance could help make it more
effective. One benefit consultant said annual distribution was good in
that people think about retirement savings so infrequently that at least
they would think about it once a year when the statement arrived. However,
another consultant told us that if the statement is provided every year,
its appearance should change so that recipients do not assume that it is
the same old statement and disregard it. Another consultant suggested
changing the appearance once every 3 years to keep the statement looking
current, and another consultant said that if the statement is to be
provided every year, SSA should include a "what's new" section on the
first page that highlights any changes or additions to the current year's
statement. The use of these practices would represent a change for SSA,
which has been sending relatively the same statement to all eligible
workers for the past 6 years. Subtle changes have been made, such as the
addition of another year of earnings history and slight changes to the
Commissioner's message. However, the statement has looked virtually the
same from the front since 1999. According to the consultants, recipients
could get in the habit of expecting nothing new or fresh on the statement
and choose to disregard it.

Conclusion

Recommendations for Executive Action

Our work suggests that in using the annual survey as its only evaluation
method, SSA is limiting its ability to make well-informed and effective
improvements to the statement. Surveys do have value; our survey, for
example, did identify the general areas of the statement that were either
not recalled or were less understandable than other areas. It was people
in our focus groups, however, who provided a much more detailed assessment
of what specific information was confusing and how the statement might be
improved. Likewise, the assessment by an outside specialist helped provide
information about how the statement is succeeding relative to statements
from financial services institutions.

While SSA faces many challenges that private sector benefit firms do not
face in communicating about their programs, there are some practices in
both the private sector and other countries that, if adopted by SSA, could
improve the current statement. SSA has made improvements to the statement
in the past but has not comprehensively reevaluated it in approximately 6
years, and key content areas remain confusing or are not presented
effectively. If key content areas of the statement confuse workers,
workers might not fully understand their benefits and the role Social
Security should play in their retirement planning. For instance, they
might rely too heavily on their Social Security retirement income and not
have enough personal savings to maintain their standard of living
throughout their retirement. Additionally, because the statement does not
utilize certain design features, such as placing personalized information
on the first page or presenting information graphically, some recipients
may not read the statement.

With the current debate on the long-term fiscal challenges facing Social
Security and the need to address them, the statement will likely become an
increasingly important tool for communicating how enacted changes to the
program will affect benefits. Regularly evaluating the statement through
multiple methods and making revisions to it will involve several SSA
offices. Therefore, our work suggests that developing and implementing a
plan for evaluating the statement and revising it based on that plan will
require attention from SSA's senior leadership.

SSA should develop a plan for regularly evaluating the statement through
the collection of data from multiple sources, including information from
surveys, focus groups, call centers, walk-in traffic, and its Web site. On
the basis of the data it receives and a review of promising practices in
the private sector and other countries, SSA should consider revising the
statement. These revisions could include, for example:

o  showing the personalized benefit information first,

o  using graphics to aid readers in quickly comprehending information, and

o  	providing information to help recipients understand Social Security's
contribution to their total retirement income.

In commenting on a draft of this report, SSA agreed with our conclusions
and recommendations. By the end of fiscal year 2005, SSA anticipates

                                Agency Comments

having a draft plan for regularly evaluating the statement. In addition
SSA will revise the statement when appropriate and necessary, and as a
first step, will utilize information from this report. We also received
technical comments from the agency that were incorporated as appropriate.
Appendix VI shows the agency's comments.

We are sending copies of this report to the Commissioner of the Social
Security Administration, appropriate congressional committees, and other
interested parties. We will also make copies available to others on
request. In addition, the report will be available at no charge on GAO's
Web site at http://www.gao.gov.

If you have any questions concerning this report please contact me at
(202) 512-7215 or George Scott at (202) 512-5932. Other contacts and
acknowledgments are listed in appendix VII.

Sincerely yours,

Barbara D. Bovbjerg Director, Education, Workforce, and Income Security
Issues

Appendix I: Court Rulings on Rights to Social Security Benefits

Many people believe that after they have worked a lifetime and paid Social
Security taxes, they have earned a legal right to Social Security benefits
and that these benefits cannot be reduced or completely taken away from
them. They believe that they have a contractual or property right to these
benefits by virtue of their contributions to the system through the years.
However, as the cases discussed below show, the courts that have examined
this issue have ruled that people do not have a legal right to these
benefits and that Congress may change the rules regarding eligibility. The
courts have ruled that Congress has broad power to reduce, and even
terminate, an individual's right to benefits.

After the passage of the Social Security Act in 1935, Congress has
occasionally amended the law to deny or reduce benefits to people who had
previously qualified for benefits. People adversely affected by these laws
brought cases in federal court challenging the constitutionality of these
laws. In deciding these cases, the federal courts have established a
framework for analyzing the constitutionality of the laws and have
discussed the extent to which individuals have legal rights to Social
Security benefits.

The first time the Supreme Court addressed this issue was in 1960 in the
case of Flemming v. Nestor, 363 U.S. 603. Nestor, an alien, immigrated to
the United States in 1913 and after working for a number of years became
eligible for old age benefits, which he began receiving in 1955. In 1956,
he was deported for being a member of the Communist Party from 1933 to
1939. Membership in the Communist Party became grounds for deportation in
1940. The Social Security Act was amended in 1954 and provided for the
termination of old age benefits to an alien who is deported after
September 1, 1954.1

Nestor challenged the law on two separate constitutional grounds. First,
he argued that he had an accrued property right to his Social Security
benefits and that the law violated the due process clause of the Fifth
Amendment. Second, he argued the termination of the benefits amounts to an
ex post facto law or a punishment without a judicial trial in violation of
Article I and Article III and the Sixth Amendment to the Constitution.

The Precedent-Setting Decision

142 U.S.C. S: 402(n).

Appendix I: Court Rulings on Rights to Social Security Benefits

The Fifth Amendment states that no person shall be deprived of life,
liberty, or property without due process of law, nor shall private
property be taken for public use without just compensation. In this case,
the Supreme Court held that Nestor had not been deprived of an "accrued
property right" under the due process clause of the Fifth Amendment. The
Court noted that entitlement to Social Security benefits is based on a
worker's earnings record and not on taxes paid. The Court stated that a
worker's non-contractual interest in benefits cannot be analogized to the
interest of the holder of an annuity, where the right to benefits is based
on contractual payments made. The Court also stated that to give
individuals an accrued property right would deprive the Social Security
Act of its flexibility to adjust to ever changing conditions. The Court
relied heavily on section 1104 of the original Social Security
legislation, still in effect today, in which Congress expressly reserves
the "right to alter, amend, or repeal any provision." See 42 U.S.C. S:
1304.

The Court went on to state that this does not mean that Congress may
modify the Social Security Act free of all constitutional restraint. The
due process clause bars the withholding of non-contractual benefits if the
statute manifests a patently arbitrary classification, utterly lacking in
rational justification. The Court analyzed the statute that terminated old
age benefits to deported aliens under this standard and concluded that it
was rational for Congress to conclude that the public purse should not be
utilized to contribute to the support of those deported. The Court
therefore upheld the constitutionality of the statute under the due
process clause.

The second issue raised was whether the termination of the benefits
amounted to an ex post facto law or a punishment without a judicial trial
in violation of Article I and Article III and the Sixth Amendment to the
Constitution. An ex post facto law punishes an individual for past conduct
that was not illegal when the conduct was engaged in. In this case, since
membership in the Communist Party was not grounds for deportation when
Nestor was a member, he argued that he could not be punished for his
membership. The Court held that it did not consider the termination of
Social Security benefits under these circumstances to have been intended
as punishment and set the standard that there must be unmistakable
evidence of punitive purpose before a congressional enactment of this kind
is struck down.

It should be noted that the case was decided by a five-four margin. Three
of the dissenters decided that the statute was unconstitutional because it
violated the due process clause of the Fifth Amendment by taking away

Appendix I: Court Rulings on Rights to Social Security Benefits

Other Challenges under the Due Process Clause

earned insurance without just compensation. They stated that Social
Security benefits are an earned right, and people earning this benefit
should be able to rely on getting the benefits they paid for and have been
promised. Four justices concluded that the 1954 amendment was
unconstitutional because it imposes a punishment in violation of the
prohibition against ex post facto laws and without a judicial trial.
Despite this vigorous dissent, the majority decision is still the law
today and serves as precedent for challenges to changes in the Social
Security Act.2

In the 1970s the Supreme Court decided two challenges to provisions in the
Social Security Act under the due process clause. In Richardson v.
Belcher, 404 U.S. 78 (1971), Belcher challenged a section of the Social
Security Act, which requires a reduction, or offset, in Social Security
disability benefits if the individual is also receiving state worker
compensation benefits.3 He claimed that he was denied due process of law
under the Fifth Amendment because his benefits would not have been reduced
if he had received private insurance instead of worker compensation.
Citing the framework established under Nestor, the Court held that there
was a rational basis for the reduction based on receiving worker
compensation benefits, and it did not violate the due process clause. The
Court specifically stated that an expectation of public benefits does not
confer a contractual right to receive the expected amounts.

In the second case, Weinberger v. Salfi, 422 U.S. 749 (1975), Mrs. Salfi's
husband of less than 6 months died and she filed for widow's and child's
benefits. She was denied benefits on the basis of the duration of marriage
requirements of the act that define widow and child so as to exclude
surviving widows and stepchildren who had relationships to the deceased
wage earner for less than 9 months.4 Mrs. Salfi argued that since the
duration of relationship provision was intended to prevent the use of sham

2The Court's decision was predicated, in large measure, on an examination
of the purposes underlying the Social Security Act as well as an
examination of the way the program was structured to achieve these
purposes. Thus, a fundamental change to the purpose and the structure of
the program, such as the creation of individual accounts, could lead the
courts to revisit the issue of whether individuals have accrued property
rights or contractual rights.

342 U.S.C. S: 424a.

442 U.S.C. S: 416(c).

Appendix I: Court Rulings on Rights to Social Security Benefits

Gender-Based Discrimination Decisions

marriages to secure Social Security benefits, she should be entitled to
present proof that her marriage was not a sham.

Using the analysis set forth in Nestor and Belcher, the Court held that
under the due process clause, as long as the statutory goals are
legitimate and the classification is rationally related to the achievement
of these goals, then the action of Congress is not so arbitrary as to
violate the due process clause. Since the duration-of-relationship
requirement operates to lessen the likelihood of sham marriages, it is
related to insuring the integrity of the Social Security system and has a
rational basis. The Court held that the fact that the duration of
relationship provision may exclude legitimate claimants did not make it
unconstitutional.

The Supreme Court did overturn several provisions of the Social Security
Act in the 1970s on the basis of gender-based discrimination. In
Weinberger v. Wiesenfeld, 420 U.S. 636 (1975), Ms. Wiesenfeld, the
principal support of her family, died in childbirth survived by her baby
and husband. Her husband applied for survivor's benefits for himself and
the baby. Benefits were granted to the baby but denied the father on the
grounds that the Social Security Act allowed for "mother benefit's" only
to women.5 The Court found that this provision deprived women of
protection of their families that men receive as a result of their
employment. The Court rejected the argument that since Social Security
benefits are non-contractual, Congress is not required to provide a
covered female employee with the same benefits as it provides to a man.
The Court held that the due process clause forbids gender-based
differentiation that results in the efforts of female workers covered by
Social Security to receive less protection for their families than men.

In Califano v. Goldfarb, 430 U.S. 199 (1977), Goldfarb applied for Social
Security survivor benefits after the death of his wife, but benefits were
denied on the grounds that he had not been receiving at least one-half
support from his wife when she died. At that time, survivor benefits based
on the earnings of a deceased husband covered by the Social Security Act
were payable to his widow, but survivor benefits based on the earnings of

542 U.S.C. S: 402(g).

Appendix I: Court Rulings on Rights to Social Security Benefits

a deceased wife were payable to the widower only if he was receiving at
least one-half of his support from his deceased wife.6

The Court held that this gender-based distinction violated the due process
clause of the Fifth Amendment, following the rationale of the Wiesenfeld
case. The Court stated that if a classification by gender is to survive
constitutional challenge, it must serve important governmental objectives
and must be substantially related to the achievement of these objectives.
This test is harder to meet than the rational basis test, and in this
case, the Court held that this gender-based distinction was not justified
but was based on an unverified assumption that wives are usually
dependent.

The Supreme Court did uphold a gender-based classification in Califano v.
Webster, 430 U.S. 313 (1977). The Court upheld a provision of the act that
allowed women to eliminate more low-earning years than men from the
calculation of their retirement benefit.7 The court found that redressing
the disparity in economic condition between men and women caused by
discrimination is an important governmental objective. The Court held that
this provision worked directly to remedy some part of the effect of this
past discrimination and was therefore constitutional.

Changes made to the Social Security Act in the early 1980s as a result of
Social Security reform led to several important challenges and subsequent
decisions upholding the changes. An important case that involved
litigation between the federal government and the states was Bowen v.
Public Agencies Opposed to Social Security Entrapment, 477 U.S. 41 (1986).
Enacted in 1950, Section 418 of Title 42, United States Code, authorizes
voluntary participation by states in the Social Security system. Until
1983, this section allowed states entering into a Section 418 agreement to
provide Social Security coverage to state employees to terminate these
agreements after giving at least 2 years notice. In 1983, Congress
repealed the termination provision, even expressly preventing termination
by a state that had already provided notice, because of concerns that an
increase in terminations had threatened the integrity of the system.8 The
state of California challenged the law, arguing it had a

Changes to Benefits as a Result of Social Security Reform

642 U.S.C. S: 402(f)(1).

742 U.S.C. S: 415(b)(3). Prior to the Court's decision, Congress had
amended the law and eliminated the gender-based distinction. Pub. L. No.
92-603 S: 104(b).

8Pub. L. No. 98-21.

Appendix I: Court Rulings on Rights to Social Security Benefits

contractual right to terminate its Section 418 agreement and that Congress
had violated the due process clause of the Fifth Amendment by denying the
state its contractual right without just compensation.

The Supreme Court rejected California's argument that the contract had
established a property right within the meaning of the Fifth Amendment.
The Court again relied heavily on Section 1104 of the Social Security Act,
which reserves to Congress the right to alter, amend or repeal any
provision. The Court stated that this provision was necessary to protect
future generations of workers by giving Congress the flexibility to meet
ever-changing conditions. The Court concluded that even though the federal
government had entered into a contract with the state, Congress had
reserved the power to amend the contract through legislation.

Another statutory change made in 1983 was a provision denying Social
Security benefits to incarcerated felons.9 Incarcerated felons challenged
that provision, but decisions of various Appellate Courts in separate
cases all upheld this provision. 10 The courts rejected due process and ex
post facto law claims. As to the ex post facto law claim, the courts held
that the suspension of the benefits was not considered a punishment under
the strict standard set forth in Nestor. The D.C. Court of Appeals stated
that Congress came as close to the line of unconstitutionality as possible
but accepted the government's contention that the statute was primarily a
cost-saving measure that eliminates retirement benefits to individuals
whose basic needs are already provided at public expense.

One specific change to the Social Security Act pertaining to federal
judges resulted in a court challenge in Robinson v. Sullivan, 905 F.2d
1199 (8th Cir. 1990). Prior to 1984, federal judges were not covered by
the Social Security Act. Congress amended the law to extend Social
Security coverage to the judiciary, including senior judges performing
judicial services but not those senior judges who no longer performed
duties.11 Senior judges performing duties and senior judges not performing
duties both receive their full salary. Under the legislation as enacted in
1984, however, senior judges performing duties would have their pay
reduced by

942 U.S.C. S: 402(x).

10See for example, Jones v. Heckler, 774 F.2d 997 (10th Cir. 1985), Peeler
v. Heckler, 781 F.2d 649 (8th Cir. 1986), and Wiley v. Bowen, 824 F.2d
1120 (D.C. Cir. 1987).

11Pub. L. No. 98-21.

Appendix I: Court Rulings on Rights to Social Security Benefits

FICA contributions, but senior judges no longer performing duties would
not. Congress, realizing the inequity in that distinction, delayed
effective date of that provision of law so that it could enact corrective
legislation to exempt senior judges performing judicial duties from Social
Security coverage. However, Congress did not act before the provision went
into effect on January 1, 1986. Legislation was not signed into law until
April 8, 1986, which made the exemption for senior judges performing
duties permanent and retroactive to 1983.

Judge Robinson was a senior judge who was still performing duties and had
worked in covered employment prior to becoming a judge. He needed one more
quarter to qualify for Social Security benefits, and he claimed, after
working the first quarter of 1986, that quarter. He was determined
ineligible for Social Security benefits and challenged that determination,
arguing that the retroactive application of the legislation passed on
April 8, 1986, violated his due process rights.

The court held that it must give retroactively applied legislation the
same deference as legislation applied prospectively. Legislation
readjusting rights and burdens is not unlawful solely because it upsets
otherwise settled expectations. The court held that the due process clause
does not prevent Congress from repealing the right to receive Social
Security benefits at any time, however settled the beneficiary's
expectations might be, for any reason that is not arbitrary or utterly
lacking in rational justification.12

In summary, courts have ruled that the right to Social Security benefits
is neither contractual nor a vested property right but is based entirely
on statutory provisions that Congress reserves to itself to alter, amend,
or repeal. A strong presumption of constitutionality attaches to decisions
made by Congress in allocating scarce Social Security resources, and
courts generally have upheld congressional amendments to the Social
Security Act in pursuit of that goal, even where doing so has resulted in
some portion of the population being excluded from receiving benefits.

12In the early 1990s, the courts rejected challenges to the reduction of
benefits under the Social Security windfall elimination provision, Das v.
Department of HHS, 17 F.3d 1250 (9th Cir. 1994) and the excess earnings
provision, Brown v. Bowen, 905 F.2d 632 (2nd Cir. 1990).

Summary

                       Appendix II: Scope and Methodology

To determine the legal requirements for the statement, we reviewed the
applicable laws. We determined additional agency goals for the statement
by interviewing Social Security Administration (SSA) officials. We also
analyzed caselaw on legal rights to benefits under the Social Security
program.

To determine the strengths and weaknesses of the current statement, we
conducted focus groups to gather qualitative information on recipients'
attitudes, beliefs, and perceptions. We contracted with a marketing and
advertising research firm, OneWorld Communications, to coordinate and
moderate six focus groups. The focus groups were conducted in Seattle,
Chicago, and Atlanta to ensure geographic diversity. In each city we
conducted two focus groups-one for participants with some college
education and below and one for participants with a college degree or
higher. Each focus group had between 8 and 10 participants. Participants
were required to be 25 or older, currently not receiving Social Security
benefits, and not employed by the Social Security Administration or in the
private pension or financial services fields. To encourage focus group
participants to speak freely, we wanted to avoid having participants from
widely divergent income levels within the same group. However, given the
sensitivity of directly asking participants to disclose their income, we
instead divided the focus groups based on education because of education's
close correlation with income. To the extent possible, we also balanced
the groups for ethnicity, gender, and age. Following a GAOdeveloped
discussion guide, the moderator asked focus group participants to
carefully read and then comment on each page of the statement (see app.
III). Additionally, participants were asked to review a private sector
statement from the Principal Financial Group (see app. IV). We chose the
Principal statement as the private sector comparison because it is a
defined benefit statement that represents a retirement program similar to
Social Security. The Principal statement was also rated excellent by a
consulting firm that compares and scores benefit statements based on
standardized criteria. Only one private sector statement was chosen for
comparison, given the time constraints of the focus groups. Focus group
participants were asked to compare the Social Security Statement and the
Principal statement on length, content, and presentation of information.

We also contracted for a national random telephone survey on the
understandability and usefulness of the statement in conjunction with
another GAO study on financial literacy. We developed our questions by
reviewing the statement and reviewing reports from SSA's previous surveys
and focus groups, as well as by consulting with benefit consultants and
experts at GAO. We pretested successive drafts of the questionnaire

Appendix II: Scope and Methodology

with members of various demographic groups and translated the survey into
Spanish. Overall, a total of 17 pretests were conducted, including 5 in
Spanish. Additionally, a pilot test of over 20 random calls resulted in
final modifications to the survey. Analytica Research, in affiliation with
Dove Consulting Group, obtained and dialed 9,784 randomly generated phone
numbers between July and October 2004. The survey oversampled for African
American and Hispanic respondents, and non-minorities with family incomes
under $25,000. We received a total of 1,578 usable responses. The
contractor took various steps to increase the response rate, including
scheduling calls throughout the week, offering cash incentives, and the
mailing of postcards to non-responding numbers. The survey's response rate
was 48 percent with a 59 percent cooperation rate.1 The difference between
the response rate and the cooperation rate suggests that the difficulty in
contacting anyone on the generated phone numbers was a significant factor
in the low response rate, because interviewers were able to convince
nearly 60 percent of those they contacted to complete the survey.

Results from sample surveys are subject to error introduced by
interviewing only a sample of the entire U.S. population, and the
practical difficulties of conducting any survey may introduce other errors
into estimates made from surveys. Because we followed a probability
procedure based on random selections, our sample is only one of a large
number of samples that we might have drawn, and thus our results are
subject to sampling error. Since each sample could have provided different
estimates, we express our confidence in the precision of our particular
sample's results as a 95 percent confidence interval around each
estimate-for example, plus or minus 10 percentage points. That is, we are
95 percent confident that each of the confidence intervals for estimates
in this report will include the true value that we would have obtained had
the entire population been surveyed. All percentage estimates from this
survey have confidence intervals of plus or minus 5 percentage points or
less, unless otherwise noted.

1We calculated the response rate as the total number of usable surveys
divided by the total eligible sample called. This is in accordance with
the American Association for Public Opinion Research response rate 3. See
www.aapor.org for the formula used. The cooperation rate measures the
success at completing interviews with contacted people who are determined
to be eligible. It is defined as the number of complete interviews divided
by the total number of complete and partial interviews, refusals, and
callbacks that were arranged but not completed.

Appendix II: Scope and Methodology

In addition to the reported sampling errors, the practical difficulties of
conducting any survey may introduce other types of errors, commonly
referred to as nonsampling errors. For example, differences in how a
particular question is interpreted, the sources of information available
to respondents, or the types of people who do not respond can introduce
unwanted variability into the survey results. We included steps in both
the data collection and data analysis stages for the purpose of minimizing
such nonsampling errors. Some of these steps included pretesting
questionnaires with members of the population, reviewing answers for
inconsistencies, and checking computer analyses with a second analyst.

In assessing whether certain sections of the statement were more easily
understood than others, we used a statistical test based on the Bonferroni
inequality that accounted for all comparisons that could be made across
sections. The use of this multiple comparison test assures us that the
confidence level for any statistically significant difference between
sections is more than 95 percent.

Additionally, we contracted with Dalbar, a consulting firm that evaluates
written communications from financial service firms to their customers and
the same company that rated the Principal statement as excellent, to
review the statement. Dalbar evaluated the statement against its
standardized rating criteria, which are based on customer needs and
preferences. Dalbar's report documented the scores achieved by the
statement, maximum obtainable scores, explanations for lost points and
recommendations for improvements, as well as how the statement compares
with industry averages and benchmarks.

We determined how SSA is evaluating the statement through interviews and
review of documents. We interviewed SSA officials about their current
evaluation processes, including the use of surveys, focus groups, and
information from internal sources such as their call centers, field
offices, and Web site. We also reviewed past survey and focus group
reports including the Public Understanding Measurement System (PUMS)
survey questionnaires from 1998 to 2004, the contractor's survey summary
reports for 1998 to 2003, as well as the survey and focus group reports
about (1) the initial Personal Earnings and Benefit Estimate Statements
(PEBES), (2) the four prototypes of the new statement, (3) an Internet
statement, (4) a revised insert, and (5) the inclusion of age-specific
messages. We also reviewed a 2002 SSA Inspector General report on the
reliability of the data used to measure public knowledge of SSA.

Appendix II: Scope and Methodology

To describe how private sector benefits consultants develop and assess the
understandability of their benefit statements, we visited several benefits
consultants, conducted interviews, and observed examples of their
statements. We spoke to these benefits consultants about industry best
practices for preparing an effective and understandable statement and in
some cases asked the consultants to assess the strengths and weaknesses of
the current Social Security Statement. In some cases, we obtained sample
statements from these benefits consultants, which are included in appendix
IV.

To describe how other industrialized countries develop and assess the
understandability and effectiveness of their public pension system
statements, we interviewed officials from the respective countries and
observed examples of their statements. We also gathered information from
officials in the countries we spoke with regarding their experiences with
statement distribution, which included discussions on frequency of
delivery and in some cases the intended recipients of their statements. In
some cases, we obtained sample statements from these countries, one of
which is included in appendix V.

To select the benefits consultants that we interviewed, we first selected
the top three statement preparers as ranked by Dalbar. We used Dalbar's
2003 ranking for defined benefit plan statements, which ranked CIGNA,
Principal, and MassMutual as the top three preparers of defined benefit
pension plan statements. We also selected three other benefit consulting
firms that have practices dedicated to benefit statement design. Finally,
we discussed industry best practices in statement preparation with
officials from a nonprofit retirement organization and academia.

To select the industrialized countries whose officials we interviewed, we
performed literature searches using the Internet, past GAO reports, and
public pension-related journals and articles to determine which countries
provide their citizens with a public pension statement. We determined
through our search that Canada, Sweden, and the United Kingdom all
provided statements to their citizens and all of them had been doing so
for several years, increasing the possibility that they may have developed
some best practices. We performed our interviews and other correspondence
with officials in these countries via telephone conferences and e-mail.

                  Appendix III: 2004 Social Security Statement

Note: The highlighting on pages 2 through 4 shows the information legally
required to be included in the statement. Also, the original statement is
a two-color document.

Appendix III: 2004 Social Security Statement Appendix III: 2004 Social Security
                                   Statement

                    Source: Social Security Administration.

                  Appendix III: 2004 Social Security Statement

                    Source: Social Security Administration.

Appendix IV: Examples of Private Sector Retirement Statements

Note: This statement was used in our focus groups and is a multicolor
document.

Appendix IV: Examples of Private Sector Retirement Statements

                       Source: Principal Financial Group.

Appendix IV: Examples of Private Sector Retirement Statements

             Note: The original statement is a multicolor document.

Appendix IV: Examples of Private Sector Retirement Statements

                         Source: Prudential Financial.

Appendix IV: Examples of Private Sector Retirement Statements

Appendix IV: Examples of Private Sector Retirement Statements

Source: Massachusetts Mutual Life Insurance Company. Note: The original
statement is a multicolor document.

Appendix IV: Examples of Private Sector Retirement Statements

Source: Massachusetts Mutual Life Insurance Company.

            Source: Canada Pension Plan (Social Development Canada).

Note: The original statement is a tri-fold and a multicolor document.

Page 53 GAO-05-192 Social Security Statements

            Source: Canada Pension Plan (Social Development Canada).

                 Page 54 GAO-05-192 Social Security Statements

Appendix VI: Comments from the Social Security Administration

Appendix VI: Comments from the Social Security Administration

Appendix VI: Comments from the Social Security Administration

Appendix VII: GAO Contacts and Staff Acknowledgments

Contacts 	George A. Scott, Assistant Director (202) 512-5932 Richard L.
Harada, Analyst-in-Charge (206) 287-4841

Staff 	In addition to those named above, Allison Abrams, Amy Anderson,
Debra Johnson, Michael Morris, Luann Moy, Jennifer Popovic,

Acknowledgments 	Daniel Schwimer, Andrew Stichman, and Wendy Wierzbicki
made important contributions to this report.

             (130361) Page 58 GAO-05-192 Social Security Statements

GAO's Mission

Obtaining Copies of GAO Reports and Testimony

The Government Accountability Office, the audit, evaluation and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site (www.gao.gov). Each weekday, GAO posts newly
released reports, testimony, and correspondence on its Web site. To have
GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone 	The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out
to the Superintendent of Documents. GAO also accepts VISA and Mastercard.
Orders for 100 or more copies mailed to a single address are discounted 25
percent. Orders should be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: 	Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Contact:

Waste, and Abuse in Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Gloria Jarmon, Managing Director, [email protected] (202)
512-4400Congressional U.S. Government Accountability Office, 441 G Street
NW, Room 7125 Relations Washington, D.C. 20548

Public Affairs 	Paul Anderson, Managing Director, [email protected] (202)
512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

                           PRINTED ON RECYCLED PAPER
*** End of document. ***