This is the accessible text file for GAO report number GAO-12-699 
entitled 'Retirement Security: Women Still Face Challenges' which was 
released on July 25, 2012. 

This text file was formatted by the U.S. Government Accountability 
Office (GAO) to be accessible to users with visual impairments, as 
part of a longer term project to improve GAO products' accessibility. 
Every attempt has been made to maintain the structural and data 
integrity of the original printed product. Accessibility features, 
such as text descriptions of tables, consecutively numbered footnotes 
placed at the end of the file, and the text of agency comment letters, 
are provided but may not exactly duplicate the presentation or format 
of the printed version. The portable document format (PDF) file is an 
exact electronic replica of the printed version. We welcome your 
feedback. Please E-mail your comments regarding the contents or 
accessibility features of this document to Webmaster@gao.gov.

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. 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: 
GAO: 

Report to the Chairman, Special Committee on Aging, U.S. Senate:

July 2012: 

Retirement Security:

Women Still Face Challenges:

GAO-12-699:

GAO Highlights:

Highlights of GAO-12-699, a report to the Chairman, Special Committee 
on Aging, U.S. Senate. 

Why GAO Did This Study:

Elderly women, who comprise a growing portion of the U.S. population, 
have historically been at greater risk of living in poverty than 
elderly men. Several factors contribute to the higher rate of poverty 
among elderly women including their tendency to have lower lifetime 
earnings, take time out of the workforce to care for family members, 
and outlive their spouses. Other factors could affect older women’s 
financial insecurity. These include the economic downturn and changing 
trends in pension plan offerings. In light of these circumstances, GAO 
was asked to examine (1) how women’s access to and participation in 
employer-sponsored retirement plans compare to men’s and how they have 
changed over time, (2) how women’s retirement income compares to men’s 
and how the composition of their income—the proportion of income 
coming from different sources—changed with economic conditions and 
trends in pension design, (3) how later-in-life events affect women’s 
retirement income security, and (4) what policy options are available 
to help increase women’s retirement income security. To answer these 
questions, GAO analyzed data from two nationally representative 
surveys, conducted a broad literature review, and interviewed a range 
of experts in the area of retirement security. 

What GAO Found:

Over the last decade, working women’s access to and participation in 
employer-sponsored retirement plans have improved relative to men. 
Indeed, from 1998 to 2009, women surpassed men in their likelihood of 
working for an employer that offered a pension plan, largely because 
the proportion of men covered by a plan declined. Furthermore, as 
employers have continued to terminate their defined benefit (DB) plans 
and have switched to defined contribution (DC) plans, the proportion 
of women who worked for employers that offered a DC plan increased. 
Correspondingly, women’s participation rates in DC plans increased 
slightly over this same period while men’s participation fell, thereby 
narrowing the participation difference between men and women to 1 
percentage point. At the same time, however, women contributed to 
their DC plans at lower levels than men. 

Although the composition of income for women age 65 and over did not 
vary greatly over the period—despite changes in the economy and 
pension system—-women continued to have less retirement income on 
average and live in higher rates of poverty than men in that age 
group. The composition of women’s income varied only slightly, in 
part, because their main income sources—-Social Security and DB 
benefits—were shielded from fluctuations in the market. Women, 
especially widows and those age 80 and over, depended on Social 
Security benefits for a larger percentage of their income than men. 
For example, in 2010, 16 percent of women age 65 and over depended 
solely on Social Security for income compared to 12 percent of men. At 
the same time, the share of household income women received from 
earnings increased over the period, but was consistently lower than 
for men. Moreover, women’s median income was approximately 25 percent 
lower than men’s over the last decade, and the poverty rate for women 
in this age group was nearly two times higher than men’s in 2010.   
For women approaching or in retirement, becoming divorced, widowed or 
unemployed had detrimental effects on their income security. Moreover, 
divorce and widowhood had more pronounced effects for women than for 
men. For example, women’s household income, on average, fell by 41 
percent with divorce, almost twice the size of the decline that men 
experienced. For widowhood, women’s household income fell by 37 
percent—while men’s declined by only 22 percent. Unemployment also had 
a detrimental effect on income security, though the effects were 
similar for women and men; household assets and income fell by 7 to 9 
percent. 

A range of existing policy options could address some of the income 
security challenges women face in retirement. For example, some would 
expand existing tax incentives to save for retirement while others 
would improve access to annuities. All of these options have 
advantages and disadvantages that would need to be evaluated prior to 
implementation. For example, increasing Social Security benefits for 
widows could provide additional income for women who have few options 
to increase their retirement savings. However, increasing benefits 
would also increase costs to the Social Security program and have 
implications for its long-term solvency. 

What GAO Recommends:

GAO is making no recommendations. GAO received technical comments on a 
draft of this report from the Department of Labor, the Department of 
the Treasury and the Social Security Administration, and incorporated 
them, as appropriate. 

View [hyperlink, http://www.gao.gov/products/GAO-12-699]. For more 
information, contact Charles Jeszeck at (202) 512-7215 or 
jeszeckc@gao.gov. 

[End of section] 

Contents:

Letter:

Background:

Working Women's Access to and Participation in Employer-Sponsored 
Pension Plans Have Improved Relative to Men:

While Income Composition Changed Only Slightly for Women Age 65 and 
Over, They Continue to Have Less Retirement Income Than Men:

Divorce, Widowhood, and Unemployment Had a Detrimental Effect on Older 
Women's Income Security:

Existing Policy Options Could Address Retirement Security Issues 
Facing Women:

Concluding Observations:

Agency Comments:

Appendix I: Objective, Scope, and Methods:

Section 1: Information Sources:

Section 2: Methods for Comparing Working Women's and Men's Access to 
and Participation in Employer-Sponsored Pension Plans:

Section 3: Methods for Comparing the Income of Women and Men Age 65 
and Over:

Section 4: Methods for Analyzing the Effects of Events Occurring Later 
in Life on Women's and Men's Household Income and Assets:

Appendix II: GAO Contact and Staff Acknowledgments:

Tables:

Table 1: Estimated Effects of Life Events on Household Assets and 
Income by Gender:

Table 2: Percent of Women and Men Reporting Their Health Is Poor Is 
Similar across Age Groups:

Table 3: Proposals to Expand Use of Existing Tax Incentives to Save 
for Retirement:

Table 4: Proposals to Expand Eligibility and Opportunities to 
Accumulate Social Security Credits:

Table 5: Proposals to Expand Access to Retirement Savings and 
Strengthen Spousal Protections:

Table 6: Proposals to Expand Opportunities for Saving Later in Life 
and Delay Social Security Benefit Receipt:

Table 7: Proposals to Ensure Lifetime Income:

Table 8: Proposals to Ensure Income Adequacy:

Table 9: Data Sources Used for Each Reporting Objective:

Table 10: SIPP Panels, Waves, and Questionnaires Used to Answer 
Objective 1 and Objective 2:

Table 11: Birth Years for the HRS Cohorts and the Year Data Collection 
Began for Each Cohort:

Table 12: Information Used from SIPP to Construct Key Variables:

Table 13: Characteristics of the Working Population over Time:

Table 14: Factors Associated with Working for an Employer That Offers 
a Plan, 2009:

Table 15: Factors Associated with Eligibility for Employer-Sponsored 
Pension Plan, 2009:

Table 16: Factors Associated with Participation in an Employer- 
Sponsored Pension Plan, 2009:

Table 17: Descriptive Statistics of Women and Men in the HRS by Age:

Table 18: Proportion of Individuals Changing Status between 
Observations:

Table 19: Divorce Effect on Household Assets and Income:

Table 20: Widowhood Effect on Household Assets and Income:

Table 21: Unemployment Effect:

Table 22: A Decline in Health's Effect on Household Assets and Income:

Table 23: Effects of Providing Financial Assistance or Physical Care 
on Household Assets and Income:

Figures:

Figure 1: Labor Force Participation Rates for Women, Ages 25 to 64:

Figure 2: In 2009, Working Women and Working Men Were Similar in Their 
Access to and Participation in Employer-Sponsored Pension Plans:

Figure 3: Proportion of Working Women and Men with Employers That 
Offered Any Type of Pension Plan and DC Plans Specifically:

Figure 4: The Proportion of Working Women and Working Men with 
Employers That Offered DC Pension Plans Varied, by Race:

Figure 5: The Proportion of Working Women and Men Who Were Eligible 
for Their Employer's Pension Plans (among the Population Whose 
Employers Offered a Plan):

Figure 6: The Proportion of Eligible Women and Men That Participated 
in Any Type of Employer-Sponsored Pension Plan or in DC Plans (among 
the Population That Was Eligible for a Plan):

Figure 7: The Proportion of Working Women and Working Men (among Those 
Who Were Eligible) Who Participated in Their Employer's Defined 
Contribution Pension Plans, by Race:

Figure 8: The Composition of Household Income for Women and Men Age 65 
and Over Did Not Fluctuate Greatly Over Time:

Figure 9: Differences in the Composition of Household Income for Women 
and Men Age 65 and Over, by Marital Status, 2010:

Figure 10: Differences in the Composition of Household Income for 
Women and Men Age 65 and Over, by Age Group, 2010:

Figure 11: Differences in the Composition of Household Income for 
Women and Men Age 65 and Over, by Race and Ethnicity, 2010:

Figure 12: Median Household Incomes in 2010 for Individuals 65 and 
Over by Age Group:

Figure 13: Poverty Rates by Demographic Categories in 2010 for Women 
and Men Age 65 and Over:

Figure 14: Estimated Effects of Divorce and Separation on Total 
Household Assets and Income:

Figure 15: Estimated Effects of Widowhood on Total Household Assets 
and Income:

Figure 16: Estimated Effects of Unemployment on Total Household Assets 
and Income:

Figure 17: Estimated Effects of a Decline in Health on Total Household 
Assets and Income:

Abbreviations:

AHEAD: Asset and Health Dynamics of the Oldest Old:

CODA: Children of the Depression Era:

DB: defined benefit:

DC: defined contribution:

ERISA: Employee Retirement Income Security Act of 1974:

HRS: Health and Retirement Study:

IRA: individual retirement account:

SCF: Survey of Consumer Finances:

SIPP: Survey of Income and Program Participation:

[End of section] 

United States Government Accountability Office: 
Washington, DC 20548: 

July 19, 2012:

The Honorable Herb Kohl:
Chairman:
Special Committee on Aging:
United States Senate:

Dear Mr. Chairman,

Historically, elderly women have been at greater risk of living in 
poverty than elderly men. Several economic and demographic factors 
contribute to their higher poverty rates in old age. First, women's 
average annual earnings are consistently lower than men's. Second, 
women are more likely to take time out of the workforce to care for 
children and elderly relatives. These employment patterns result in 
lower retirement savings, reduced Social Security benefits,[Footnote 
1] and smaller pension benefits for women in comparison to men. Third, 
women tend to live longer than men, increasing the risk of exhausting 
their retirement savings before death. Finally, women are more likely 
than men to live alone in old age,[Footnote 2] increasing their 
vulnerability to unexpected economic and health shocks due to the 
inability to pool resources with a partner or benefit from spousal 
care-giving in the event of an illness.

Recent economic events affecting both men and women have the potential 
to exacerbate older women's financial insecurity. The financial crisis 
and recent recession have resulted in depressed home values and high 
unemployment rates among younger and older Americans alike. At the 
same time, health care costs continue to rise. Efforts to address the 
financial challenges of Social Security and Medicare could lead to a 
reduction in benefits for retirees.[Footnote 3] In addition, the 
burden of saving for retirement and paying for old-age health care has 
been shifting from employers to employees in both the private and 
public sectors. In the private sector, for example, many employers 
continue to replace defined benefit (DB) pension plans with defined 
contribution (DC) plans and reduce or eliminate retiree health 
insurance benefits. At the same time, many employed in the public 
sector have seen a reduction in their pension benefits or an increase 
in employee contributions for those benefits.

In light of this unique confluence of circumstances, the Senate 
Special Committee on Aging requested that we explore the issue of 
women's retirement income security with a special focus on the effects 
of the recent financial crisis and subsequent recession, and the 
persistent trend of employers to replace DB with DC plans.[Footnote 4] 
Specifically, this report examines (1) how women's access to and 
participation in employer-sponsored retirement plans compare to men's 
and how they have changed over time, (2) how women's retirement income 
compares to men's and how the composition of their income changed with 
economic conditions and trends in pension design, (3) how events 
occurring later in life affect women's retirement income security, and 
(4) what policy options are available to help increase women's 
retirement income security.

To address these questions, we analyzed two nationally-representative 
datasets, conducted an extensive literature review, and consulted with 
numerous experts. Specifically, to analyze plan coverage and 
participation rates among the working-age population, we used data for 
the late 1990s through 2009 from the Survey of Income and Program 
Participation (SIPP), a nationally-representative survey.[Footnote 5] 
With these data, we computed descriptive statistics on plan coverage, 
eligibility, and participation rates and conducted an econometric 
analysis of each of these. To analyze median incomes and the income 
composition of the retirement-age population, we computed descriptive 
statistics using SIPP data from the late 1990s through 2010.[Footnote 
6] To understand the factors that affect women's income and assets, we 
developed a statistical model to estimate the effects of events 
occurring later in life, such as widowhood, using the Health and 
Retirement Study (HRS), a nationally representative dataset that 
tracks Americans 51 years or older over time.[Footnote 7] We conducted 
a data reliability assessment of selected SIPP and HRS data and found 
that, for the purposes of our analysis, the data that we analyzed were 
sufficiently reliable. Finally, to identify policy options that could 
increase retirement income security among women, we conducted an 
extensive literature review and interviewed a range of experts in the 
area of retirement income security.[Footnote 8]

We conducted this performance audit from March 2011 through July 2012 
in accordance with generally accepted government auditing standards. 
Those standards require that we plan and perform the audit to obtain 
sufficient, appropriate evidence to provide a reasonable basis for our 
findings and conclusions based on our audit objectives. We believe 
that the evidence we obtained provides a reasonable basis for our 
findings and conclusions based on our audit objectives. For more 
information on our scope and methodology, see appendix I.

Background:

Demographic and Labor Force Trends Affecting Women's Retirement Income 
Security:

Since the early 1900s, female life expectancy has exceeded male life 
expectancy, resulting in women outnumbering men in the older age 
groups. Although gender differences in life expectancy have been 
decreasing, women age 65 and over continue to outnumber men age 65 and 
over. This trend is projected to continue over the next 4 decades. 
Further, the population age 65 and over is expected to more than 
double from 2010 to 2050.[Footnote 9] The population of women among 
the "oldest-old"--those 85 and over--is also projected to grow. 
[Footnote 10] Today, of those age 65 and over, one-sixth of women and 
one-tenth of men are among the oldest-old and this is projected to 
grow to almost one-quarter of women and one-fifth of all men by 2050. 
[Footnote 11]

Women's workforce participation surged over the last half of the 20th 
century. Among women ages 25 to 54, the rate of labor force 
participation jumped from 42 percent by the end of the 1950s to about 
74 percent by the late 1980s. The rate continued to grow in the 1990s 
but at a slower pace. Over the last decade, the rate declined slightly 
from its peak of 76.8 percent in 1999, and was 74.7 percent in 2011. 
Labor force participation rates have varied by generation, with women 
born in the baby boom generation much more likely to be in the 
workforce than preceding generations.[Footnote 12] As baby boomers 
have aged, workforce participation rates have increased significantly 
for women ages 55 to 64 (see figure 1).

Figure 1: Labor Force Participation Rates for Women, Ages 25 to 64:

[Refer to PDF for image: multiple line graph] 

Year: 1950; 
Ages 25-54: 36.8%; 
Ages 55-64: 27%. 

Year: 1955; 
Ages 25-54: 39.8%; 
Ages 55-64: 32.5%. 

Year: 1960; 
Ages 25-54: 42.9%; 
Ages 55-64: 37.2%. 

Year: 1965; 
Ages 25-54: 45.2%; 
Ages 55-64: 41.1%. 

Year: 1970; 
Ages 25-54: 50.1%; 
Ages 55-64: 43%. 

Year: 1975; 
Ages 25-54: 55.1%; 
Ages 55-64: 40.9%. 

Year: 1980; 
Ages 25-54: 64%; 
Ages 55-64: 41.3%. 

Year: 1985; 
Ages 25-54: 69.6%; 
Ages 55-64: 42%. 

Year: 1990; 
Ages 25-54: 74%; 
Ages 55-64: 45.2%. 

Year: 1995; 
Ages 25-54: 75.6%; 
Ages 55-64: 49.2%. 

Year: 2000; 
Ages 25-54: 76.7%; 
Ages 55-64: 51.9%. 

Year: 2005; 
Ages 25-54: 75.3%; 
Ages 55-64: 57%. 

Year: 2011; 
Ages 25-54: 74.7%; 
Ages 55-64: 59.5%. 

Source: GAO analysis of BLS data. 

[End of figure]

Despite their economic gains, women continue to have lower annual 
earnings than men, on average, and much lower lifetime earnings. In 
2010, the median earnings of women working full-time were about 
$36,900, compared to $47,700 for men.[Footnote 13] One study reported 
that a 25-year-old woman with a college degree will make about 
$523,000 less in wages over her lifetime compared to a man with a 
college degree.[Footnote 14] Further, the study noted that of those 
retiring at age 62 in 2000, women were in the workforce for 12 years 
less than men, on average, primarily because they spent more time than 
men out of the workforce caring for family members.[Footnote 15]

Sources of Retirement Income:

Although the composition of retirement income--the proportion of 
income coming from different sources--varies greatly for individual 
households, Social Security benefits, pension income, and earnings 
make up the bulk of income for the U.S. population age 65 and over. 
Social Security provides retirement benefits to eligible workers, 
based on their work and earnings history. Social Security also 
provides benefits to eligible workers who become disabled before 
reaching retirement age, as well as spouses, widow(er)s, and children 
of eligible workers. Although all Social Security benefits are based 
upon a common formula, they are calculated in different ways for each 
beneficiary type.[Footnote 16] The level of the monthly benefit is 
adjusted for inflation and varies depending on the age at which the 
beneficiary chooses to begin receiving benefits. Generally, 
beneficiaries may begin receiving retirement benefits at age 62; 
however, the payments will be higher if they wait to receive benefits 
at their full retirement age, which varies from 65 to 67, depending on 
the beneficiary's birth year. The monthly retirement benefit continues 
to rise for workers who delay benefits beyond their full retirement 
age, up to age 70. Employees and employers pay payroll taxes that 
finance Social Security benefits. However, Social Security faces a 
long-term financing shortfall resulting largely from lower birth rates 
and longer life spans. According to the Social Security Trustees, the 
Social Security Trust Funds could be exhausted by 2033 and unable to 
pay full benefits.[Footnote 17]

Pension income from employer-provided retirement plans falls into two 
broad categories: DB and DC pension plans. DB plans typically provide 
retirement benefits to each retiree in the form of an annuity that 
provides a monthly payment for life, the value of which is typically 
determined by a formula based on particular factors specified by the 
plan, such as salary or years of service. Under DC plans, workers and 
employers may make contributions into individual accounts.[Footnote 
18] Workers can also save for retirement through an individual 
retirement account (IRA). IRAs allow workers to receive favorable tax 
treatment for making contributions to an individual account.[Footnote 
19]

At retirement, participants' distribution options vary depending on 
the type of pension plan. Private sector DB plans must offer 
participants a benefit in the form of a lifetime annuity (either 
immediately or deferred). An annuity can help to protect a retiree 
against risks, including the risk of outliving one's assets (longevity 
risk) and, when an inflation-adjusted annuity is provided, the risk of 
inflation diminishing one's purchasing power. Some DB plans also give 
participants a choice to take a lump sum cash settlement 
(distribution) or roll over funds to an IRA, instead of taking a 
lifetime annuity.[Footnote 20] In contrast, DC plan sponsors are not 
required to offer a lifetime annuity and more often provide 
participants with a lump sum distribution as the only option. Other 
options for DC participants may include leaving money in the plan, 
taking a partial distribution, rolling their plan savings into an IRA, 
or purchasing an annuity, which are typically only available outside 
of the plan.

In addition, whether a pension plan is a DB or DC has implications for 
whether a spouse is entitled to the pension's benefits. The Employee 
Retirement Income Security Act of 1974 (ERISA) requires that DB plans 
include a survivor's benefit, called a qualified joint and survivor 
annuity. Thus, after a worker with a DB plan dies, the surviving 
spouse continues to receive an annuity, but typically at a reduced 
level.[Footnote 21] A qualified joint and survivor annuity may only be 
waived through a written spousal consent. Under most DC plans, the 
plan is written so that the employee may, during his or her lifetime, 
make withdrawals from the account or roll over the balance into an IRA 
without spousal consent, provided that the employee's vested account 
balance is payable in full on death to the surviving spouse.

National Trends Affecting Retirement Security for Men and Women:

Over the past quarter-century, the percentage of private sector 
workers participating in employer-sponsored pension plans has held 
steady at about 50 percent. Although some workers choose not to 
participate in an employer-sponsored pension plan, the large majority 
of nonparticipating workers do not have access to one.[Footnote 22] In 
addition, over the last 3 decades, the U.S. retirement system has 
undergone a major transition from one based primarily on DB plans to 
one based on DC plans, increasing workers' exposure to economic 
volatility and usually shifting the burden of saving to the individual 
worker, which makes them more reliant on their own decision making. As 
we have previously reported, from 1990 to 2008, the number of active 
participants in private sector DB plans fell by 28 percent, from about 
26 million to about 19 million. Over the same period, the number of 
active participants in DC plans increased by 90 percent, from about 35 
million to about 67 million.[Footnote 23] DC plans generally do not 
offer annuities, so retirees are left with increasingly important 
decisions about managing their retirement savings to ensure they have 
income throughout retirement.[Footnote 24] These decisions may be more 
difficult to make in times of economic volatility. For example, two 
recent recessions--one beginning in March 2001 and ending in November 
2001 and the other beginning in December 2007 and ending in June 
2009[Footnote 25]--resulted in major stock indices falling 
dramatically. The long-term effects of financial market fluctuations 
on retirement income security are uncertain, but the effects may vary 
based on factors such as age, type of pension plan, and employment 
status.[Footnote 26] Employment status, in particular, can pose 
serious challenges for retirement security. As we recently reported, 
long-term unemployment can reduce an older worker's future monthly 
retirement income in numerous ways such as by reducing the number of 
years the worker can accumulate DB plan retirement benefits or DC plan 
savings, by motivating workers to claim Social Security at an earlier 
age, and by leading workers to draw down retirement savings to pay for 
expenses during unemployment.[Footnote 27]

Working Women's Access to and Participation in Employer-Sponsored 
Pension Plans Have Improved Relative to Men:

From 1998 to 2009, working women surpassed men in their likelihood of 
having an employer that offered a pension plan, but were slightly less 
likely to be eligible for and to participate in those plans.[Footnote 
28] However, this gap, narrowed over time. In fact, by 2009, the same 
proportion of working women and men ultimately participated in some 
type of plan (either a DB or a DC) as shown in figure 2. Nonetheless, 
women's contribution rates to DC plans remained lower than those of men.

Figure 2: In 2009, Working Women and Working Men Were Similar in Their 
Access to and Participation in Employer-Sponsored Pension Plans:

[Refer to PDF for image: illustration] 

Men: 
Not offered 42%; 
Employer offers a plan 58%; 
Not eligible 5%; 
Employee eligible 53%; 
Choose not to participate 7%; 
Participate 46%. 

Women: 
Not offered 39%; 
Employer offers a plan 61%; 
Not eligible 8%; 
Employee eligible 53%; 
Choose not to participate 7%; 
Participate 46%. 
 
Source: GAO analysis of SIPP data. 

Note: Percentage estimates in this figure have 95 percent confidence 
intervals that are within +/-1 percent of the estimate itself. 

[End of figure] 

Women Surpassed Men in Their Likelihood of Working for an Employer 
That Offers a Pension Plan:

While working men and women were just as likely to have employers that 
offered pension plans in 1998, by 2009, these women were more likely 
than men to work for employers that offered pension plans (see figure 
3). This may be due to the sectors and industries in which women 
worked. For example, a greater proportion of women than men worked in 
the public and nonprofit sectors--sectors that have higher proportions 
of workers with access to plans offered by employers--than the for-
profit sector. Women were also more likely to work in the education 
and health industries--industries that have higher proportions of 
workers with access to plans offered by employers.[Footnote 29] In 
contrast, men had higher rates of self-employment over this period, 
and self-employed individuals were much less likely to have retirement 
plans. In addition, from 1998 to 2009, the proportion of working women 
and men with employers that offered pension plans declined after 2003, 
possibly reflecting the decline in the number of employers offering DB 
plans.[Footnote 30] Furthermore, the proportion of women working for 
employers offering DC plans increased, rising from 41 to 49 percent 
(see figure 3). With the exception of 1998, women were more likely to 
work for employers that offered DC plans than were men.

Figure 3: Proportion of Working Women and Men with Employers That 
Offered Any Type of Pension Plan and DC Plans Specifically:

[Refer to PDF for image: horizontal bar graph] 

Percentage of workers with employer that offers a plan: 

All plans: 

1998: 
Men: 60%; 
Women: 60%. 

2003: 
Men: 61%; 
Women: 64%. 

2006: 
Men: 57%; 
Women: 61%. 

2009: 
Men: 58%; 
Women: 61%. 

DC plans only: 

1998: 
Men: 41%; 
Women: 41%. 

2003: 
Men: 42%; 
Women: 43%. 

2006: 
Men: 41%; 
Women: 45%. 

2009: 
Men: 46%; 
Women: 49%. 

Source: GAO analysis of SIPP data. 

Note: Percentage estimates in this figure have 95 percent confidence 
intervals that are within +/-1 percent of the estimate itself. 

[End of figure] 

Moreover, as shown in figure 4, while the proportion of working women 
with an employer that offered a DC plan increased through 2009--though 
not always steadily--it varied by racial and ethnic groups. White and 
Black women, for example, were the most likely to work for an employer 
that offered a plan, while Hispanic women were the least 
likely.[Footnote 31] Interestingly, with only a few exceptions (i.e., 
Whites in 1998 and Asians in 2003 and 2009), the proportion of women 
working for an employer offering a plan was equal to or higher than 
that of men of the same race.

Figure 4: The Proportion of Working Women and Working Men with 
Employers That Offered DC Pension Plans Varied, by Race:

[Refer to PDF for image: horizontal bar graph] 

Percentage of workers with employer that offers a plan: 

White: 

1998: 
Men: 44%; 
Women: 43%. 

2003: 
Men: 45%; 
Women: 45%. 

2006: 
Men: 44%; 
Women: 46%. 

2009: 
Men: 50%; 
Women: 51%. 

Black: 

1998: 
Men: 37%; 
Women: 38%. 

2003: 
Men: 38%; 
Women: 42%. 

2006: 
Men: 40%; 
Women: 45%. 

2009: 
Men: 43%; 
Women: 49%. 

Hispanic: 

1998: 
Men: 29%; 
Women: 30%. 

2003: 
Men: 27%; 
Women: 33%. 

2006: 
Men: 29%; 
Women: 36%. 

2009: 
Men: 30%; 
Women: 36%. 

Asian: 

1998: 
Men: 36%; 
Women: 40%. 

2003: 
Men: 42%; 
Women: 36%. 

2006: 
Men: 41%; 
Women: 44%. 

2009: 
Men: 48%; 
Women: 44%. 

Source: GAO analysis of SIPP data. 

Note: For Whites, percentage estimates in this figure have 95 percent 
confidence intervals that are within +/-2 percentage points or less of 
the estimate itself. For Blacks, Hispanics, and Asians, the 95 percent 
confidence intervals are within +/-3, 3 and 6 or fewer percentage 
points of the estimate itself respectively. For Asians, the variation 
by year may be due to their relatively small sample size. 

[End of figure] 

Women Working for Employers Offering Plans Made Gains in Plan 
Eligibility:

Among those working for an employer offering a pension, the vast 
majority of both women and men were eligible to participate in the 
plan, and their eligibility rates generally increased over time (see 
figure 5). Moreover, women's eligibility rates increased more than 
men's, thereby narrowing the gap between men and women from 7 to 4 
percentage points.

Figure 5: The Proportion of Working Women and Men Who Were Eligible 
for Their Employer's Pension Plans (among the Population Whose 
Employers Offered a Plan):

[Refer to PDF for image: horizontal bar graph] 

Percentage eligible: 

1998: 
Men: 78%; 
Women: 85%. 

2003: 
Men: 82%; 
Women: 88%. 

2006: 
Men: 83%; 
Women: 87%. 

2009: 
Men: 87%; 
Women: 91%. 

Source: GAO analysis of SIPP data. 

Note: Percentage estimates in this figure have 95 percent confidence 
intervals that are within +/-2 percent or less of the estimate itself. 

[End of figure] 

Of the women who were not eligible to participate in their employer's 
pension plan in 2009, the majority reported that they were not 
eligible because they did not work enough hours, weeks, or months per 
year at their place of employment. Correspondingly, women that worked 
part-time were significantly less likely to be eligible for their 
employer's plan, according to our analysis.[Footnote 32] In contrast, 
men most frequently cited insufficient tenure as the reason for 
ineligibility.

The Gender Gap in Plan Participation Narrowed:

Among those eligible to participate in their employer's pension plan, 
women had lower rates of participation than men, but this gap 
diminished over time as men's participation rates declined. 
Specifically, the participation rate for women in any type of plan (DB 
or DC) declined slightly from 87 percent in 1998 to 86 percent in 
2009, while the participation rate for men declined from 91 to 87 
percent (see figure 6).

Among those eligible for DC plans, women's participation rates 
increased by one percentage point over the years we analyzed, while 
men's declined by 2 percentage points. Taken together, the gender 
participation gap in DC take-up rates narrowed from 4 to 1 percentage 
points.

Figure 6: The Proportion of Eligible Women and Men That Participated 
in Any Type of Employer-Sponsored Pension Plan or in DC Plans (among 
the Population That Was Eligible for a Plan):

[Refer to PDF for image: horizontal bar graph] 

Percentage participating: 

All plans: 

1998: 
Men: 91%; 
Women: 87%. 

2003: 
Men: 90; 
Women: 88%. 

2006: 
Men: 88%; 
Women: 85%. 

2009: 
Men: 87%; 
Women: 86%. 

DC plans only: 

1998: 
Men: 81%; 
Women: 77%. 

2003: 
Men: 84%; 
Women: 79%. 

2006: 
Men: 81%; 
Women: 77%. 

2009: 
Men: 79%; 
Women: 78%. 

Source: GAO analysis of SIPP data. 

Note: Percentage estimates in this figure have 95 percent confidence 
intervals that are within +/-2 percent or less of the estimate itself. 

[End of figure] 

Women's participation rates in DC plans also varied by race. As shown 
in figure 7, White and Asian women had the highest participation rates 
in DC plans, ranging from 79 and 78 percent respectively in 1998 to 80 
and 85 percent in 2009. Black and Hispanic women had lower 
participation rates, but the rate for Black women increased over time 
from 66 to 70 percent. With some exceptions, across all racial and 
ethnic groups, eligible women tended to have lower participation rates 
than eligible men across all 4 years.

Figure 7: The Proportion of Working Women and Working Men (among Those 
Who Were Eligible) Who Participated in Their Employer's Defined 
Contribution Pension Plans, by Race:

[Refer to PDF for image: horizontal bar graph] 

Percentage participating: 

White: 

1998: 
Men: 83%; 
Women: 79%. 

2003: 
Men: 85%; 
Women: 80%. 

2006: 
Men: 84%; 
Women: 79%. 

2009: 
Men: 82%; 
Women: 80%. 

Black: 

1998: 
Men: 74%; 
Women: 66%. 

2003: 
Men: 76%; 
Women: 73%. 

2006: 
Men: 71%; 
Women: 68%. 

2009: 
Men: 64%; 
Women: 70%. 

Hispanic: 

1998: 
Men: 71%; 
Women: 70%. 

2003: 
Men: 77%; 
Women: 71%. 

2006: 
Men: 68%; 
Women: 72%. 

2009: 
Men: 69%; 
Women: 69%. 

Asian: 

1998: 
Men: 86%; 
Women: 78%. 

2003: 
Men: 84%; 
Women: 81%. 

2006: 
Men: 84%; 
Women: 79%. 

2009: 
Men: 85%; 
Women: 85%. 

Source: GAO analysis of SIPP data. 

Note: For Whites, percentage estimates in this figure have 95 percent 
confidence intervals that are within +/-2 percentage points or less of 
the estimate itself. For Blacks, Hispanics, and Asians, the 95 percent 
confidence intervals are within +/-5, +/-6, and +/-7 or fewer 
percentage points of the estimate respectively. 

[End of figure] 

Several characteristics of women help to explain their lower 
participation rates in DC plans. For one, women had significantly 
lower levels of household income than men across all 4 years. Our 
analysis, coupled with studies conducted by outside experts, indicates 
that higher incomes are associated with higher rates of plan 
participation.[Footnote 33] Further, despite women's increasing 
attachment to the labor force, they continue to be more likely than 
men to work part-time and to have less tenure--factors we and others 
have found to be associated with lower DC participation 
rates.[Footnote 34] At the same time, a higher proportion of women are 
single-parents--a factor that we found to be negatively associated 
with plan participation. After accounting for these differences (and 
differences in other factors) between men and women, women did not 
have significantly lower participation rates than men in 2009. 
[Footnote 35]

In addition to having lower rates of participation, women also 
contributed to their DC plans at lower levels than men. Among those 
reporting their contributions as shares of their salaries, women's 
contribution rates hovered around 6.7 percent of their salaries while 
men's contribution rates averaged around 7.2 percent over the years of 
our analysis.[Footnote 36] Among those reporting their contributions 
in dollar amounts, women's annual contributions were consistently 
around 30 percent lower than men's over the study period.

While Income Composition Changed Only Slightly for Women Age 65 and 
Over, They Continue to Have Less Retirement Income Than Men:

The composition of women's and men's retirement income did not vary 
greatly over the last decade despite changes in the economy and 
pension system, largely because their main income sources--Social 
Security and DB plans--were shielded from fluctuations in the 
financial market. However, women, especially widows and those 80 years 
and over, depended on Social Security benefits for a larger percentage 
of their income than men. In contrast, women received a lower share of 
their income from earnings than men. Women age 65 and over also had 
less retirement income on average and higher rates of poverty than men 
in that age group. Specifically, for the population age 65 and over, 
women's median income was approximately 25 percent lower than men in 
the same age group for all years.[Footnote 37] Moreover, women in this 
age group were nearly twice as likely to be living in poverty than men.

The Composition of Income for Women and Men Age 65 and Over Did Not 
Fluctuate Greatly Despite Changes in the Economy and Pension System:

The composition of household income for women and men age 65 and over 
fluctuated only slightly from 1998 to 2010, despite changes in the 
economy and the pension system (see figure 8). The composition of 
household income did not fluctuate drastically largely because Social 
Security and DB benefits comprised nearly three-quarters of household 
income for women and slightly less (around 70 percent) for men, 
providing them with guaranteed monthly income for life. Women tended 
to receive a higher proportion of household income from Social 
Security. In fact, in 2010, 16 percent of women age 65 and over 
depended solely on Social Security for income compared to 12 percent 
of men. At the same time, the share of income from earnings increased 
slightly for men and women, but was consistently lower for women than 
for men. Furthermore, the share of income from DC plans was very low 
(1 to 2 percent) across the entire period for both men and women. This 
is due to the fact that the lion's share of people age 65 and over did 
not report receiving any income from regular distributions from DC 
plans.[Footnote 38]

Figure 8: The Composition of Household Income for Women and Men Age 65 
and Over Did Not Fluctuate Greatly Over Time:

[Refer to PDF for image: stacked horizontal bar graph] 

Men: percentage of household income: 

1998: 
Social Security: 47%; 
Earnings: 17%; 
Defined benefit pensions: 21%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 13%. 

2003: 
Social Security: 49%; 
Earnings: 17%; 
Defined benefit pensions: 24%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 9%. 

2006: 
Social Security: 50%; 
Earnings: 18%; 
Defined benefit pensions: 21%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 8%. 

2010: 
Social Security: 50%; 
Earnings: 19%; 
Defined benefit pensions: 22%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 7%. 

Women: percentage of household income: 

1998: 
Social Security: 54%; 
Earnings: 14%; 
Defined benefit pensions: 18%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 14%. 

2003: 
Social Security: 54%; 
Earnings: 14%; 
Defined benefit pensions: 19%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 11%. 

2006: 
Social Security: 54%; 
Earnings: 16%; 
Defined benefit pensions: 19%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 10%. 

2010: 
Social Security: 54%; 
Earnings: 16%; 
Defined benefit pensions: 20%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 9%. 

Source: GAO analysis of SIPP data. 

Notes: Estimates for men and women include spousal income. The 
category for income from defined contribution pensions reflects total 
household distributions from IRAs, as well as 401(k) pension plans and 
similar defined contribution pension plans. Nonregular (lump sum) 
withdrawals from IRA and 401(k) plans are not included. The "other" 
category includes income from cash public assistance and property 
income including interest, dividends, rent and royalties. Percentages 
may not add to 100% due to rounding. Percentages are based on 
household incomes for each source including zero values. Percentage 
estimates of the income shares from Social Security, earnings, defined 
benefit pension, and defined contribution pensions have 95 percent 
confidence intervals that are within +/-2.5 percent of the estimate 
itself. For information on how these percentages were estimated, see 
appendix I. 

[End of figure] 

As shown in figures 9 to 11, in 2010, the composition of household 
income for individuals age 65 and over also varied by demographic 
group. Among marital-status categories, widowed women depended on 
Social Security benefits for a larger percentage of their income (58 
percent) than other women (see figure 9). In fact, about 21 percent of 
all widowed women depended on Social Security as their sole source of 
income. Separated women and men received higher shares of income from 
earnings, and married women and men received relatively higher shares 
of their income from DB plans.

Figure 9: Differences in the Composition of Household Income for Women 
and Men Age 65 and Over, by Marital Status, 2010:

[Refer to PDF for image: stacked horizontal bar graph] 

Men: percentage of household income: 

Married: 
Social Security: 49%; 
Earnings: 20%; 
Defined benefit pensions: 23%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 6%. 

Widowed: 
Social Security: 53%; 
Earnings: 13%; 
Defined benefit pensions: 23%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 8%. 

Divorced: 
Social Security: 54%; 
Earnings: 18%; 
Defined benefit pensions: 18%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 9%. 

Separated: 
Social Security: 49%; 
Earnings: 32%; 
Defined benefit pensions: 9%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 0%; 
Other: 11%. 

Never married: 
Social Security: 47%; 
Earnings: 20%; 
Defined benefit pensions: 19%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 12%. 

Women: percentage of household income: 

Married: 
Social Security: 52%; 
Earnings: 16%; 
Defined benefit pensions: 23%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 3%; 
Other: 7%. 

Widowed: 
Social Security: 58%; 
Earnings: 13%; 
Defined benefit pensions: 18%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 10%. 

Divorced: 
Social Security: 53%; 
Earnings: 21%; 
Defined benefit pensions: 16%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 9%. 

Separated: 
Social Security: 42%; 
Earnings: 30%; 
Defined benefit pensions: 9%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 0%; 
Other: 19%. 

Never married: 
Social Security: 48%; 
Earnings: 19%; 
Defined benefit pensions: 18%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 13%. 

Source: GAO analysis of SIPP data. 

Notes: In the category for married individuals, estimates for men and 
women include spousal income. The category for income from defined 
contribution pensions reflects total household distributions from 
IRAs, as well as 401(k) and similar defined contribution pension 
plans. Nonregular (lump sum) withdrawals are not included. The "other" 
category includes income cash public assistance and property income 
including interest, dividends, rent and royalties. Percentages may not 
add to 100% due to rounding. Percentage estimates of the income shares 
from Social Security have 95 percent confidence intervals that are 
within +/-2, +/-3, +/-4, +/-10 and +/-6 percent of the estimate itself 
for married, widowed, divorced, separated and never married 
individuals respectively. Percentage estimates of the income shares 
from earnings have 95 percent confidence intervals that are within +/-
2, +/-2, +/-3, +/-11 and +/-6 percent of the estimate itself for 
married, widowed, divorced, separated and never married individuals 
respectively. Percentage estimates of the income shares from defined 
benefit plans have 95 percent confidence intervals that are within +/-
2, +/-3, +/-3, +/-7 and +/-5 percent of the estimate itself for 
married, widowed, divorced, separated and never married individuals 
respectively. Percentage estimates of the income shares from defined 
contribution plans have 95 percent confidence intervals that are 
within +/-2 percent of the estimate itself for all marital status 
categories. 

[End of figure] 

As shown in figure 10, among different age groups, women age 80 and 
over received the highest share of their income from Social Security 
(61 percent). In fact, about 20 percent of them depended on Social 
Security for their sole source of income. Men in the youngest age 
category (65 to 69) received a higher share of their income from 
earnings (31 percent) relative to other groups, while individuals in 
the oldest age categories received the smallest share of income from 
earnings, likely reflecting the declining ability to work at older ages.

Figure 10: Differences in the Composition of Household Income for 
Women and Men Age 65 and Over, by Age Group, 2010:

[Refer to PDF for image: stacked horizontal bar graph] 

Men: percentage of household income: 

Age 65-69: 
Social Security: 41%; 
Earnings: 31%; 
Defined benefit pensions: 20%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 7%. 

Age 70-74: 
Social Security: 52%; 
Earnings: 19%; 
Defined benefit pensions: 21%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 5%. 

Age 75-79: 
Social Security: 55%; 
Earnings: 12%; 
Defined benefit pensions: 24%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 3%; 
Other: 6%. 

Age 80 or older: 
Social Security: 56%; 
Earnings: 8%; 
Defined benefit pensions: 23%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 11%. 

Women: percentage of household income: 

Age 65-69: 
Social Security: 46%; 
Earnings: 25%; 
Defined benefit pensions: 20%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 8%. 

Age 70-74: 
Social Security: 53%; 
Earnings: 16%; 
Defined benefit pensions: 21%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 9%. 

Age 75-79: 
Social Security: 58%; 
Earnings: 11%; 
Defined benefit pensions: 22%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 7%. 

Age 80 or older: 
Social Security: 61%; 
Earnings: 10%; 
Defined benefit pensions: 18%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 10%. 

Source: GAO analysis of SIPP data. 

Notes: Estimates for men and women include spousal income. The 
category for income from defined contribution pensions reflects total 
household distributions from IRAs, as well as 401(k) and similar 
defined contribution pension plans. Nonregular (lump sum) withdrawals 
are not included. The "other" category includes income from cash 
public assistance and property income including interest, dividends, 
rent and royalties. Percentages may not add to 100 percent due to 
rounding. Percentage estimates of the income shares from Social 
Security have 95 percent confidence intervals that are within +/-2, +/-
4, +/-3, and +/- 2 percent of the estimate itself for individuals in 
the 65-69, 70-74, 75-79, and 80+ age categories respectively. 
Percentage estimates of the income shares from earnings have 95 
percent confidence intervals that are within +/-2 percent of the 
estimate itself for individuals in all age categories respectively. 
Percentage estimates of the income shares from defined benefit pension 
plans have 95 percent confidence intervals that are within +/-2, +/-2, 
+/-2, and +/-4 percent of the estimate itself for individuals in the 
65-69, 70-74, 75-79, and 80+ age categories respectively. Percentage 
estimates of the income shares from defined contribution pension plans 
have 95 percent confidence intervals that are within +/-0.5, +/-3, +/-
1, and +/-1 percent of the estimate itself for individuals in the 65-
69, 70-74, 75-79, and 80+ age categories respectively. 

[End of figure] 

Finally, among racial and ethnic groups, White and Black women and men 
age 65 and over received the highest share of income from Social 
Security (see figure 11). In contrast, Asians and Hispanics tended to 
receive a lower share of their incomes from Social Security.[Footnote 
39] Asian men and women received a disproportionately higher share of 
income from earnings relative to other racial and ethnic categories. 
White and Black women and men received higher shares of income from DB 
plans, compared to Hispanics and Asians.

Figure 11: Differences in the Composition of Household Income for 
Women and Men Age 65 and Over, by Race and Ethnicity, 2010:

[Refer to PDF for image: stacked horizontal bar graph] 

Men: percentage of household income: 

White: 
Social Security: 50%; 
Earnings: 17%; 
Defined benefit pensions: 23%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 7%. 

Black: 
Social Security: 53%; 
Earnings: 19%; 
Defined benefit pensions: 21%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 6%. 

Hispanic: 
Social Security: 46%; 
Earnings: 30%; 
Defined benefit pensions: 13%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 0%; 
Other: 10%. 

Asian: 
Social Security: 33%; 
Earnings: 43%; 
Defined benefit pensions: 9%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 15%. 

Women: percentage of household income: 

White: 
Social Security: 56%; 
Earnings: 14%; 
Defined benefit pensions: 21%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 2%; 
Other: 8%. 

Black: 
Social Security: 52%; 
Earnings: 17%; 
Defined benefit pensions: 22%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 0%; 
Other: 9%. 

Hispanic: 
Social Security: 44%; 
Earnings: 31%; 
Defined benefit pensions: 11%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 14%. 

Asian: 
Social Security: 38%; 
Earnings: 34%; 
Defined benefit pensions: 14%; 
Defined contribution pensions (e.g. 401(k) plans, IRAs): 1%; 
Other: 13%. 

Source: GAO analysis of SIPP data. 

Notes: Estimates for men and women include spousal income. The 
category for income from defined contribution pensions reflects total 
household distributions from IRAs, as well as 401(k) and similar 
defined contribution pension plans. Nonregular (lump sum) withdrawals 
are not included. The "other" category includes income from cash 
public assistance and property income including interest, dividends, 
rent and royalties. Percentages may not add to 100 percent due to 
rounding. The size of the 95 percent confidence intervals for 
estimates presented in this figure varies by racial/ethnic category. 
Percentage estimates of the income shares from Social Security have 95 
percent confidence intervals that are within +/-2, +/-4, +/-5, and +/-
6 percent of the estimate itself for White, Black, Hispanic, and Asian 
individuals respectively. Percentage estimates of the income shares 
from earnings have 95 percent confidence intervals that are within +/-
1, +/-3, +/-5, and +/-7 percent of the estimate itself for White, 
Black, Hispanic, and Asian individuals respectively. Percentage 
estimates of the income shares from defined benefit plans have 95 
percent confidence intervals that are within +/-2, +/-3, +/-3, and +/-
7 percent of the estimate itself for White, Black, Hispanic, and Asian 
individuals respectively. Percentage estimates of the income shares 
from defined contribution plans have 95 percent confidence intervals 
that are within +/-1 percent for all the racial and ethnic categories. 

[End of figure] 

Median Household Income for Women Age 65 and Over Was 25 Percent Lower 
Than Men's:

Women age 65 and over had consistently lower median incomes than men 
across age and most race groups over time.[Footnote 40] Over the last 
decade, the median incomes of women age 65 and over were approximately 
25 percent lower than their male counterparts. Median incomes, did, 
however, vary by demographic category (see figure 12). Demographic 
groups with the lowest median incomes included women who were either 
unmarried--especially those who had been separated or never married-- 
over the age-of 80, or Black or Hispanic.

Figure 12: Median Household Incomes in 2010 for Individuals 65 and 
Over by Age Group:

[Refer to PDF for image: horizontal bar graph] 

Median household income: 

Median: 
Men: $44,390; 
Women: $33,150. 

Married: 
Men: $49,720; 
Women: $46,990. 

Widowed: 
Men: $30,700; 
Women: $23,190. 

Divorced: 
Men: $33,160; 
Women: $25,320. 

Separated: 
Men: $35,080; 
Women: $21,470. 

Never married: 
Men: $27,610; 
Women: $22,890. 

Age 65-69: 
Men: $53,460; 
Women: $44,070. 

Age 70-74: 
Men: $47,100; 
Women: $36,200. 

Age 75-79: 
Men: $40,390; 
Women: $30,610. 

Age 80 or older: 
Men: $37,130; 
Women: $25,020. 

White: 
Men: $45,800; 
Women: $33,590. 

Black: 
Men: $35,310; 
Women: $27,610. 

Hispanic: 
Men: $40,480; 
Women: $32,820. 

Asian: 
Men: $47,530; 
Women: $41,960. 

Source: GAO analysis of SIPP data. 

Notes: Estimates for men and women include spousal income. Estimates 
of median incomes have 95 percent confidence intervals that are within 
+/-$900 for women and +/-$1,200 for men in the entire U.S., +/-$1,600 
for married women, +/-$1,000 for widowed women, +/-$1,900 for divorced 
women, +/-$5,200 for separated women, +/-$4,000 for never married 
women, +/-$1,800 for married men, +/-$2,000 for widowed men, +/-$4,100 
for divorced men, +/-$8,000 for separated men, +/-$3,500 for never 
married men, +/-$1,800 for women ages 65-69, +/-1,700 for woman age 
ages 70-74, +/-$2,100 for women ages 75-79, +/-$900 for women 80 and 
older, +/-$2,700 for men ages 65-69, +/-$2,300 for men ages 70-74 and 
75-79, +/-$2,200 for men 80 and over, +/-$1,000 for White women, +/- 
$1,400 for Black women, +/-$4,300 for Hispanic women, +/-$5,000 for 
Asian women, +/-$1,400 for White men, +/-$2,900 for Black men, +/- 
$5,300 for Hispanic men, and +/-$7,100 for Asian men. 

[End of figure] 

In addition, a greater proportion of women age 65 and over lived in 
households with incomes below the poverty line than men in the same 
age group. Consistent with their relatively lower median incomes, the 
demographic groups with the highest poverty rates were women who were 
not married, over the age of 80, or non-White (see figure 
13).[Footnote 41] In contrast, married people and White men had the 
lowest poverty rates.

Figure 13: Poverty Rates by Demographic Categories in 2010 for Women 
and Men Age 65 and Over:

[Refer to PDF for image: horizontal bar graph] 

Percentage of population age 65 or over:  

Entire U.S.: 
Men: 5%; 
Women: 9%. 

Married: 
Men: 3%; 
Women: 3%. 

Widowed: 
Men: 6%; 
Women: 12%. 

Divorced: 
Men: 8%; 
Women: 13%. 

Separated: 
Men: 20%; 
Women: 22%. 

Never married: 
Men: 7%; 
Women: 20%. 

Age 65-69: 
Men: 5%; 
Women: 7%. 

Age 70-74: 
Men: 4%; 
Women: 9%. 

Age 75-79: 
Men: 4%; 
Women: 9%. 

Age 80 or older: 
Men: 4%; 
Women: 11%. 

White: 
Men: 3%; 
Women: 7%. 

Black: 
Men: 9%; 
Women: 16%. 

Hispanic: 
Men: 13%; 
Women: 16%. 

Asian: 
Men: 16%; 
Women: 17%. 

Source: GAO analysis of SIPP data. 

Note: Estimates for men and women include spousal income. Percentage 
estimates of poverty rates have 95 percent confidence intervals that 
are within +/-1 percent for the category for the entire U.S.; +/-1 
percent for married, +/-2 percent for widowed, +/-3 percent for 
divorced, +/-12 percent for separated, and +/-6 percent for never 
married individuals; +/-2 percent for all age-categories; +/-1 percent 
for Whites, +/-4 percent for Blacks, +/-5 percent for Hispanics, and +/
-8 percent for Asians. 

[End of figure] 

Divorce, Widowhood, and Unemployment Had a Detrimental Effect on Older 
Women's Income Security:

When women nearing or in retirement--women over age 50--became 
divorced, widowed or unemployed, the effects on their households' 
total assets and income were detrimental, according to our analysis 
(see table 1).[Footnote 42] Further, divorce and widowhood had more 
pronounced effects for women than for men. These effects may be 
contributing to elderly women's higher poverty rates and lower levels 
of income compared to men's. We also found, not surprisingly, that a 
decline in health after age 50 had a negative effect on household 
assets and income.[Footnote 43] Lastly, we also examined the effect of 
caring for elderly parents on income and assets, but we did not find 
statistically significant negative relationships. All of these effects 
may not be generalizable to younger cohorts as women's labor force 
participation and, correspondingly, their assets and income, have 
changed over the last several decades.[Footnote 44] 

Table 1: Estimated Effects of Life Events on Household Assets and 
Income by Gender:

Percent change:

Became divorced or separated; 
Total household assets: Women: -41%[A]; 
Total household assets: Men: -39%[A]; 
Total household income: Women: -41%[A,B]; 
Total household income: Men: -23%[A,B].

Became widowed; 
Total household assets: Women: -32%[A,B]; 
Total household assets: Men: -27%[A,B]; 
Total household income: Women: -37%[A,B]; 
Total household income: Men: -22%[A,B].

Became unemployed; 
Total household assets: Women: -7%[A]; 
Total household assets: Men: -7%[A]; 
Total household income: Women: -9%[A]; 
Total household income: Men: -7[A].

Health declined; 
Total household assets: Women: -8%[A]; 
Total household assets: Men: -10%[A]; 
Total household income: Women: -4%[A]; 
Total household income: Men: -3%[A].

Helped parents financially; 
Total household assets: Women: 3%[A]; 
Total household assets: Men: 3%[A]; 
Total household income: Women: 6%[A]; 
Total household income: Men: 7%[A].

Helped parents with daily activities; 
Total household assets: Women: 1%; 
Total household assets: Men: 1%; 
Total household income: Women: 2%[A]; 
Total household income: Men: 2%[A].

Source: GAO analysis of HRS data.

Notes: We used fixed-effects regression models to estimate the percent 
change in total household assets and income that occurs for an 
individual after a life event, relative to an individual that did not 
experience that life event. Total assets and income for the household 
were applied to each individual in the household. The estimated 
effects represent the average percent difference in household assets 
and income between all survey periods in which the household does 
experience an event and all survey periods in which the household does 
not experience an event. The event may have occurred in any year after 
the household entered the survey. For more details on the models, see 
appendix I.

[A] Estimate is significantly different from zero.

[B] Difference between women and men is statistically significant. 

[End of table] 

Became Divorced or Separated after Age 50:

As shown in figure 14, the effects of divorce or separation after age 
50 had substantial, negative effects on women's total household assets 
and income. For both women and men, assets fell by about 40 percent 
with a divorce or separation.[Footnote 45] The effects were less 
substantial for those living in households where at least one member 
was age 65 or over, but these women and men still lost about one-third 
of their total assets. The effects for income were more pronounced for 
women than for men. Women's income fell by 41 percent, nearly twice 
that of men's (23 percent). The effects were largest for women living 
in households where all members were age 64 or younger; for these 
women, income fell by 44 percent.[Footnote 46] However, while divorce 
had very detrimental effects, we found that, for women ages 51 and 
over, divorce or separation was less prevalent than widowhood. 
Specifically, for those age 85 and over in our sample, 4 percent of 
women and 2 percent of men had been divorced or separated.[Footnote 47]

Figure 14: Estimated Effects of Divorce and Separation on Total 
Household Assets and Income:

[Refer to PDF for image: vertical bar graph] 

Percentage change in total household assets: 

All households: 
Men: -41%; 
Women: -39%. 

Households where everyone is age 64 or younger: 
Men: -41%; 
Women: -39%. 

Households where at least one person is age 65 or older: 
Men: -32%; 
Women: -32%. 

Percentage change in total household income: 

All households: 
Men: -41%; 
Women: -23%. 

Households where everyone is age 64 or younger: 
Men: -44%; 
Women: -25%. 

Households where at least one person is age 65 or older: 
Men: -39%; 
Women: -23%. 

Source: GAO analysis of HRS data. 

Notes: All estimates in this figure have 95 percent confidence 
intervals within +/-8 percentage points of the estimate itself. For 
statistical comparisons of the estimates across different groups, see 
appendix I. We used fixed-effects regression models to estimate the 
percent change in total household assets and income that occurs for an 
individual after a life event, relative to an individual that did not 
experience that life event. Total assets and income for the household 
were applied to each individual in the household. The estimated 
effects represent the average percent difference in household assets 
and income between all survey periods in which the household does 
experience an event and all survey periods in which the household does 
not experience an event. The event may have occurred in any year after 
the household entered the survey. For more details on the models, see 
appendix I. 

[End of figure] 

Became Widowed after Age 50:

Not only did women's total household assets and income decline 
substantially with widowhood, but the effects were more pronounced for 
women than for men (see figure 15). For example, while men's income 
fell 22 percent after widowerhood, women's income fell by an even 
greater amount--37 percent. The effects were larger for women living 
in younger households than women living in older households. 
Specifically, women in households where all members were age 64 or 
younger experienced a 31 percent decrease in assets and a 47 percent 
decrease in income.[Footnote 48] Adding to these effects, widowhood 
was a much more common experience for women than men in our sample. In 
fact, women were at least twice as likely as men to become widowed 
between any two survey periods. Consequently, 70 percent of women age 
85 and over were widowed compared to only 24 percent of men age 85 and 
over.

Figure 15: Estimated Effects of Widowhood on Total Household Assets 
and Income:

[Refer to PDF for image: vertical bar graph] 

Percentage change in total household assets: 

All households: 
Men: -32%; 
Women: -27%. 

Households where everyone is age 64 or younger: 
Men: -31%; 
Women: -23%. 

Households where at least one person is age 65 or older: 
Men: -23%; 
Women: -18%. 

Percentage change in total household income: 

All households: 
Men: -37%; 
Women: -22%. 

Households where everyone is age 64 or younger: 
Men: -47%; 
Women: -30%. 

Households where at least one person is age 65 or older: 
Men: -35%; 
Women: -21%. 

Source: GAO analysis of HRS data. 

Notes: Because widows appear much more often in households where at 
least one person is over the age of 65 than in households where 
everyone is age 64 or younger, part of the overall effect is likely a 
comparison of the household's assets over time. This explains why the 
effect for the larger population is larger than the effect for each of 
the groups. Estimates for the "all households" and "households where 
at least one person is age 65 or older" categories have 95 percent 
confidence intervals within +/-5 percentage points of the estimate 
itself. Estimates for the "households where everyone is age 64 or 
younger" category have 95 percent confidence intervals within +/-10 
percentage points of the estimate itself. For statistical comparisons 
of the estimates across different groups, see appendix I. We used 
fixed-effects regression models to estimate the percent change in 
total household assets and income that occurs for an individual after 
a life event, relative to an individual that did not experience that 
life event. Total assets and income for the household were applied to 
each individual in the household. The estimated effects represent the 
average percent difference in household assets and income between all 
survey periods in which the household does experience an event and all 
survey periods in which the household does not experience an event. 
The event may have occurred in any year after the household entered 
the survey. For more details on the models, see appendix I. 

[End of figure] 

Became Unemployed after Age 50:

Similar to becoming widowed, unemployment had negative effects on 
total household assets and income, although the effects were similar 
for women and men (see figure 16).[Footnote 49] Women and men saw 
their assets and income decline by about 7 to 9 percent. The effects 
on income were most acute for households where at least one member of 
the household was age 65 or over. For these households, men's assets 
fell by 14 percent and their income fell by 12 percent. For women, 
there was not a significant decline in assets but their income fell by 
13 percent. In addition, older workers may have difficulty finding 
another job.[Footnote 50] However, unemployment was not very prevalent 
in the HRS sample, in part because many survey respondents were 
retired.[Footnote 51] On average, only 1 percent of men and women 
reported being out of work and actively looking for a job. For men and 
women ages 51 to 64, this percentage rose slightly to 2 percent.

Figure 16: Estimated Effects of Unemployment on Total Household Assets 
and Income:

[Refer to PDF for image: vertical bar graph] 

Percentage change in total household assets: 

All households: 
Men: -7%; 
Women: -7%. 

Households where everyone is age 64 or younger: 
Men: -9%; 
Women: -7%. 

Households where at least one person is age 65 or older: 
Men: -2%; 
Women: -14%. 

Percentage change in total household income: 

All households: 
Men: -9%; 
Women: -7%. 

Households where everyone is age 64 or younger: 
Men: -9%; 
Women: -6%. 

Households where at least one person is age 65 or older: 
Men: -13%; 
Women: -12%. 

Source: GAO analysis of HRS data. 

Notes: Estimates for the "all households" and "households where 
everyone is age 64 or younger" categories have 95 percent confidence 
intervals within +/-6 percentage points of the estimate itself. 
Estimates for the "households where at least one member is age 65 or 
older" category have 95 percent confidence intervals within +/-15 
percentage points of the estimate itself. For statistical comparisons 
of the estimates across different groups, see appendix I. We used 
fixed-effects regression models to estimate the percent change in 
total household assets and income that occurs for an individual after 
a life event, relative to an individual that did not experience that 
life event. Total assets and income for the household were applied to 
each individual in the household. The estimated effects represent the 
average percent difference in household assets and income between all 
survey periods in which the household does experience an event and all 
survey periods in which the household does not experience an event. 
The event may have occurred in any year after the household entered 
the survey. For more details on the models, see appendix I. 

[End of figure] 

Health Declined after Age 50:

As shown in figure 17, a decline in self-reported health status also 
had negative effects on total household income and assets, although to 
a lesser degree than widowhood, divorce, and unemployment. For all 
households in our sample, income fell by 4 percent for women and 3 
percent for men when self-reported health status changed from 
excellent, very good or good to fair or poor.[Footnote 52] The effects 
of a decline in health on assets varied by household type. The 
differences between women and men were the largest for younger 
households, where all members were age 64 or younger. For example, the 
loss of assets was greater for men (13 percent) compared to women (5 
percent).[Footnote 53]

Figure 17: Estimated Effects of a Decline in Health on Total Household 
Assets and Income:

[Refer to PDF for image: vertical bar graph] 

Percentage change in total household assets: 

All households: 
Men: -8%; 
Women: -10%. 

Households where everyone is age 64 or younger: 
Men: -5%; 
Women: -13%. 

Households where at least one person is age 65 or older: 
Men: -6%; 
Women: -4%. 

Percentage change in total household income: 

All households: 
Men: -4%; 
Women: -3%. 

Households where everyone is age 64 or younger: 
Men: -5%; 
Women: -3%. 

Households where at least one person is age 65 or older: 
Men: -3%; 
Women: -2%. 

Source: GAO analysis of HRS data. 

Notes: All estimates in this figure have 95 percent confidence 
intervals within +/-3 percentage points of the estimate itself. For 
statistical comparisons of the estimates across different groups, see 
appendix I. We used fixed-effects regression models to estimate the 
percent change in total household assets and income that occurs for an 
individual after a life event, relative to an individual that did not 
experience that life event. Total assets and income for the household 
were applied to each individual in the household. The estimated 
effects represent the average percent difference in household assets 
and income between all survey periods in which the household does 
experience an event and all survey periods in which the household does 
not experience an event. The event may have occurred in any year after 
the household entered the survey. For more details on the models, see 
appendix I. 

[End of figure] 

Although the effects of a decline in health were smaller than the 
effects of some of the other life events in our analysis, more 
individuals experienced this event than any other. Almost 30 percent 
of individuals ages 65 to 84 reported being in poor health (see table 
2). For individuals ages 85 and over, 40 percent reported being in 
poor health. Interestingly, as shown in table 2, women and men 
suffered from poor health at similar rates across age categories. 
Further, we found that, between any two HRS surveys, about 2 percent 
of both women and men reported entering a period of poor health.

Table 2: Percent of Women and Men Reporting Their Health Is Poor Is 
Similar across Age Groups:

Percent reporting their health is poor:

Ages 51-64; 
Women: 21%; 
Men: 20%.

Ages 65-84; 
Women: 28%; 
Men: 28%.

Ages 85 and over; 
Women: 40%; 
Men: 40%.

Source: GAO analysis of HRS data. 

[End of table] 

Lastly, we found that providing elderly parents with financial 
assistance or helping parents with basic activities of daily living 
(i.e., bathing, dressing, and eating) had a slightly positive effect 
on household assets and income. However, often these effects were not 
significantly different from zero, possibly because of limitations in 
our data and methods.[Footnote 54] In addition, we found that only a 
small percentage of the sample provided these types of assistance to 
their parents. Also, women and men age 51 through 64 were much more 
likely to provide assistance than women and men age 65 and over. But, 
as the baby boomers age, more children may be called upon to help 
their parents financially or with basic activities.

Existing Policy Options Could Address Retirement Security Issues 
Facing Women:

Through our interviews with experts and our literature review, we 
found that a range of existing policy options could help improve 
retirement income security for women.[Footnote 55] Our analysis 
focuses on how women would be affected by these policy options. While 
each of these options would be available for both women and men, they 
could help address some of the specific challenges women face in 
ensuring a secure retirement. For example, some options would expand 
the use of existing tax incentives, encouraging women to save more. 
Another set of options would expand access to and strengthen spousal 
protections for retirement savings. These options could increase 
women's retirement savings and preserve their retirement income if 
they become divorced or widowed. Other sets of options could motivate 
women nearing retirement to work longer and save more, ensure lifetime 
retirement income, or enhance benefit adequacy. These options could 
help shield women from the effects of divorce, widowhood, and 
unemployment and decrease their risk of living in poverty.

All of the options have cost implications that would need to be 
considered prior to implementation. Moreover, as with federal spending 
programs, any option that results in reduced or deferred federal tax 
revenue may require an offset, such as raising revenue elsewhere or 
cutting spending. While the federal government could bear some of 
these costs, workers and plan sponsors could be responsible for 
others. Also, although some of the options could have positive effects 
on women on their own, there could be an offsetting effect. If the 
plan sponsor, for example, is responsible for the increased cost of 
sponsorship and makes changes to the plan to offset those increased 
costs, women may not ultimately benefit from the policy option. 
Lastly, some of these changes may require legislative changes.

Proposals to Expand the Use of Existing Tax Incentives to Save for 
Retirement:

Some of the policy options we identified could expand the use of 
existing tax incentives for individuals to save for retirement during 
their working years (see table 3). These options could help lower-and 
moderate-income workers, as well as workers who take time out of the 
workforce to care for family members. Since women have lower earnings 
than men, on average, and are more likely to take time out of the 
workforce to care for family members, women may especially benefit 
from these options. However, pension experts are concerned that women 
may not be as financially literate as men, hindering them from taking 
full advantage of options for saving for retirement.[Footnote 56]

Table 3: Proposals to Expand Use of Existing Tax Incentives to Save 
for Retirement:

Policy option: Automatic IRA; 
Description of policy option: Employers who do not sponsor a pension 
plan would be required to automatically enroll employees in an IRA 
unless the employee opted out.[A] Automatic IRA proposals have been 
introduced before the four most recent Congresses.b However, this 
option would result in a loss of federal tax revenue.c Further, this 
kind of requirement could pose administrative burdens and costs for 
employers; 
Potential effects on women: According to one study, lower-and moderate-
income workers may be more likely to be eligible for automatic 
IRAs.[D] Women have lower incomes and retirement savings than men, but 
experts reported that automatic enrollment in IRAs could increase the 
number of women saving for retirement or increase their retirement 
savings. However, women from lower-income households may not be able 
to afford to make contributions to an IRA. 

Policy option: Expansion of Saver's Credit; 
Description of policy option: The Saver's Credit--a tax credit for 
retirement savings for low-and middle-income workers--could be 
expanded in a number of ways. For example, some experts have called 
for making the credit refundable.[E] This option would result in a 
reduction in tax revenue[F]; 
Potential effects on women: By enhancing the tax incentives to save 
for retirement, low-and middle-income workers may save more for 
retirement. However, women from lower-income households may choose not 
to take advantage of the credit because they may not be able to afford 
to contribute. Our previous work has shown that while the number of 
workers benefiting from an expansion of the Saver's Credit could be 
small, the increase in retirement savings could be sizable for those 
who do benefit.[G].

Policy option: Caregiver contributions to IRAs; 
Description of policy option: Allow all caregivers to contribute to 
IRAs up to the qualified contribution limit, which would be based on 
the individual's adjusted gross income in the year prior to becoming a 
qualified caregiver. Currently, a married caregiver who has no 
compensation or whose compensation is less than her spouse, and who 
files a joint return, can contribute to an IRA by using her spouse's 
compensation in determining her maximum contributions to an IRA. If 
implemented, tax revenue could fall; 
Potential effects on women: Women, who are more likely to take time 
out of the workforce to provide care for family members, could 
continue to save for retirement while providing care. However, women 
from lower-income households may not be able to afford to make 
contributions to an IRA while providing care to relatives. 

Policy option: Expand catch-up contributions; 
Description of policy option: Currently, workers age 50 and over are 
permitted to make additional, annual "catch-up" tax-deferred 
contributions of up to $5,500 to their DC plans. Under this option, 
workers ages 40-49 would become eligible to make such contributions, 
and the contribution limits would be raised. Simultaneously, a 
campaign could be launched to promote the catch-up contribution 
option. By expanding tax incentives, however, more tax revenue could 
be deferred; 
Potential effects on women: Women would be able to make larger 
contributions to DC plans for an additional decade, increasing their 
retirement savings. However, as we have previously reported, men are 
three times more likely than women to make catch-up contributions.[H] 
Because they have lower earnings than men, women may be constrained in 
their ability to save more. As a result, women may not choose to take 
advantage of extra years to make catch-up contributions. 

Source: GAO analysis of literature and expert interviews.

[A] It has been proposed that certain types of firms, such as those 
with fewer than 10 employees, would be exempt from the automatic IRA 
requirement. Our prior work has analyzed the automatic IRA proposal. 
See GAO, Retirement Savings: Automatic Enrollment Shows Promise for 
Some Workers, but Proposals to Broaden Retirement Savings for Other 
Workers Could Face Challenges, GAO-10-31 (Washington, D.C.: Oct. 23, 
2009) and Private Pensions: Low Defined Contribution Plan Savings May 
Pose Challenges to Retirement Security, Especially for Many Low-Income 
Workers, GAO-08-8 (Washington, D.C.: Nov. 29, 2007).

[B] See The Automatic IRA Act of 2012, H.R. 4049, 112th Cong. (2012) 
and the Automatic IRA Act of 2011, S. 1557, 112th Cong. (2011); the 
Automatic IRA Act of 2010, S. 3760 and H.R. 6099, 111th Cong. (2010); 
the Automatic IRA Act of 2007, S. 1141 and H.R. 2167, 110th Cong. 
(2007); and the Automatic IRA Act of 2006, S. 3952 and H.R. 6210, 
109th Cong. (2006).

[C] Treasury has estimated that if automatic enrollment in IRAs and 
doubling an existing employer tax credit for starting an employer- 
sponsored pension plan were implemented by the end of calendar year 
2013, then the revenue loss would be about $15 billion for fiscal 
years 2013-2022.

[D] Benjamin H. Harris and Ilana Fischer, The Population of Workers 
Covered by the Auto IRA: Trends and Characteristics, AARP Public 
Policy Institute (Washington, D.C.: Feb. 2012).

[E] Currently, the Saver's Credit is nonrefundable. A nonrefundable 
tax credit can reduce tax owed to zero, but it cannot be used to 
generate a refund payment to the filer in excess of taxes paid.

[F] The cost of expanding the Saver's Credit would depend on how the 
credit was expanded. For example, the President's fiscal year 2011 
budget proposed expanding the Saver's Credit by making the credit 
refundable and providing a 50 percent match on retirement 
contributions of up to $1,000 for families earning $85,000 or less. 
The estimated cost of this expansion was $29.8 billion for fiscal 
years 2011-2020. See Office of Management and Budget, Budget of the 
U.S. Government: Fiscal Year 2011 (Washington, D.C., Feb. 1, 2010).

[G] See GAO, Private Pensions: Some Key Features Lead to an Uneven 
Distribution of Benefits, GAO-11-333 (Washington, D.C.: Mar. 30, 2011).

[H] See GAO-11-333. 

[End of table] 

Proposals to Expand Opportunities to Accumulate Social Security Credits:

Experts also identified a set of policy options that would offer new 
opportunities to accumulate earnings credits for Social Security (see 
table 4). These options could enhance the retirement security of 
workers who experience a period of unemployment or who take time out 
of the workforce to care for family members. For example, counting 
unemployment insurance payments as creditable earnings under Social 
Security may be particularly helpful for women who become unemployed 
later in life and experience a notable decrease in their assets and 
income. However, because they would extend eligibility or increase 
benefits, these options would increase costs for Social Security and 
decrease solvency.

Table 4: Proposals to Expand Eligibility and Opportunities to 
Accumulate Social Security Credits:

Policy option: Count unemployment insurance payments as creditable 
earnings under Social Security; 
Description of policy option: Currently, workers do not receive 
earnings credits for unemployment compensation. However, two experts 
told us some countries consider unemployment compensation as 
creditable earnings under their social security systems. This allows 
workers to continue accruing earnings credits while unemployed. This 
option could increase costs and would decrease Social Security 
solvency; 
Potential effects on women: According to two of the experts we spoke 
with, women who experience bouts of unemployment would receive more 
earnings credits under Social Security, potentially increasing their 
benefits. This option may also help women become eligible for benefits.

Policy option: Allow care-giving credits for Social Security benefit 
calculations; 
Description of policy option: Under the current system, Social 
Security eligibility and benefit amounts depend on the amount of time 
a worker spends in covered employment. Under this option, workers who 
take time out of the workforce to provide care could have their Social 
Security benefits adjusted. For example, the benefits formula could 
impute earnings for years with zero or low earnings due to care- 
giving.[A] In addition, this option would increase Social Security 
costs and decrease solvency; 
Potential effects on women: Crediting time spent providing care could 
increase women's Social Security benefits or make them eligible for 
benefits. Our past work has shown that more women than men could 
benefit from care-giving credits.[B] However, as we have previously 
reported, care-giving credits may not reach the target population. For 
example, low-income people are less likely to be able to take time off 
from work. Therefore, people who have relatively higher incomes may 
benefit more from the creation of care-giving credits.[C] 

Source: GAO analysis of literature and expert interviews.

[A] SSA's Office of the Chief Actuary has estimated the effect of 
providing a care-giving credit to parents with a child under 6 for up 
to 5 years. In 2011, the Office of the Chief Actuary estimated these 
proposals would decrease solvency by 0.24 percent of payroll. See 
[hyperlink, http://www.ssa.gov/oact/solvency/provisions/index.html]. 

[B] See GAO-08-105.

[C] See GAO-10-101R. 

[End of table] 

Proposals to Expand Access to Retirement Savings and Strengthen 
Spousal Protections:

Other policy options could either expand access to retirement savings 
in DC plans and IRAs or strengthen spousal protections for retirement 
savings (see table 5). These options could address a variety of 
challenges women face, including their lower levels of income in 
retirement. In addition, they could preserve retirement income after a 
divorce or after becoming widowed. For example, requiring that a wife 
provides consent whenever a husband takes a distribution from his DC 
savings would protect the wife's access to household income in 
retirement. However, these options could increase costs for plan 
sponsors. For example, requiring notarized spousal consent whenever a 
husband takes a distribution could increase the administrative costs 
that must be paid by plan sponsors.

Table 5: Proposals to Expand Access to Retirement Savings and 
Strengthen Spousal Protections:

Policy option: Lower DC plan eligibility requirements; 
Description of policy option: Currently, employees are generally 
eligible for DC plans once they have at least 1,000 hours of service 
during a 12-month period. One proposal would require employers to 
offer DC plans to employees that have at least 500 hours of service 
per year for 3 years. This option could, in turn, increase costs for 
plan sponsors. It would also result in increased deferral of tax 
revenue if more workers made contributions to DC plans because DC 
contributions are typically tax- deferred; 
Potential effects on women: Women, who tend to move in and out of the 
workforce and/or work part-time, could become eligible to participate 
in DC plans. If they choose to participate, their retirement savings 
would increase. However, over 75 percent of women covered by a pension 
are eligible to participate, so the number of women affected by this 
option may be limited. Further, part-time workers have lower earnings 
than full-time workers and may not be able to make contributions to DC 
plans.

Policy option: Lower DC plan vesting requirements; 
Description of policy option: Currently, ERISA requires that employees 
become vested in DC plans in no more than 3 or 6 years, depending on 
whether the plan calls for graded or cliff vesting, respectively.[A] 
Experts have called for lowering these vesting requirements. For 
example, one proposal calls for lowering vesting requirements to 2 
years for plans with cliff vesting and 3 years for plans with graded 
vesting. Such options, however, could increase costs for plan sponsors 
and result in an increased deferral of tax revenue; 
Potential effects on women: Women, who tend to move in and out of the 
workforce and/or work part-time, would become more likely to vest more 
of their employer-sponsored pension plans, improving their access to 
pension benefits and retirement savings. In our 2008 report on women's 
retirement income security, we simulated lowering vesting 
requirements. We found that women in the lowest income quintile saw 
the largest change in benefits. Similarly, never married and divorced 
women saw a bigger increase in benefits than did married and widowed 
women.[B] 

Policy option: Provide spousal protection provisions for DC savings; 
Description of policy option: Currently, spousal consent is not 
required when married individuals take distributions from their IRA or 
DC savings. Under tax-qualified DB plans, the spouse must provide 
consent in order to elect a DB benefit that is not a qualified joint 
and survivor annuity. One proposal calls for requiring spousal consent 
for any distribution from an IRA or DC plan other than a joint and 
survivor annuity. This option could increase costs for plan sponsors 
and would defer tax revenue if requiring spousal consent results in 
individuals delaying withdrawals; 
Potential effects on women: Spousal protections for DB and DC plans 
would be similar. These changes would help to ensure that women were 
involved with decisions that would affect their retirement income and, 
in turn, would help improve the adequacy of their retirement income. 
However, officials and experts have noted that spouses often give 
consent to select a DB benefit other than a joint and survivor 
annuity, raising questions about the effectiveness of placing the same 
spousal consent requirements on DC plans.

Source: GAO analysis of literature and expert interviews.

[A] ERISA, as amended, governs vesting periods. Plans with cliff 
vesting have a specified point at which participants have a right to 
all benefits accrued to date and benefits accrued thereafter. Plans 
with graded vesting give participants a right to an increasing 
percentage of their total accrued benefit over time. For more 
information, see GAO, Answers to Key Questions about Private Pension 
Plans, GAO-02-745SP (Washington, D.C.: Sept. 18, 2002).

[B] See GAO-08-105. 

[End of table] 

Proposals to Expand Opportunities for Saving Later in Life and Delay 
Social Security Benefit Receipt:

Experts identified three policy options that could motivate women 
nearing retirement to remain in the workforce and delay claiming 
Social Security benefits, thereby giving them more time to save for 
retirement and increasing their Social Security benefits (see table 
6). Because women tend to have less income in retirement than men, and 
because elderly women face higher poverty rates than elderly men, 
these options for boosting retirement savings and benefits may improve 
women's overall retirement income security. For example, the full 
retirement age for Social Security could be increased, thus providing 
workers who are able to work with an incentive to keep doing so--
potentially saving more for retirement in the process. However, each 
of these options has disadvantages. In the case of increasing the full 
retirement age, this option may not prove to be effective because 
women may not be able to work longer or may choose to exit the 
workforce before the full retirement age. They would, in turn, suffer 
reductions in Social Security income.

Table 6: Proposals to Expand Opportunities for Saving Later in Life 
and Delay Social Security Benefit Receipt:

Policy option: Education on benefits of waiting to start collecting 
Social Security benefits; 
Description of policy option: According to experts, many people do not 
realize that waiting to claim Social Security benefits can 
significantly increase monthly benefit amounts for the rest of their 
lives. Better educational outreach could increase awareness. If 
workers delay claiming Social Security benefits, income and payroll 
tax revenues would be increased and solvency would be improved. 
Employer pension costs could be increased if workers continue 
participating in their pension plans; 
Potential effects on women: A larger monthly income could help many 
women avoid poverty in retirement and better protect against outliving 
their retirement assets. On the other hand, women may not have the 
savings they need or be able to keep working to have enough income to 
delay claiming. 

Policy option: Increase the early or full retirement ages; 
Description of policy option: Experts told us the Social Security 
early or full retirement ages could be increased. By increasing the 
Social Security retirement ages, workers may choose to work longer, 
resulting in additional payroll tax revenue, which would improve 
solvency.[A] However, employer pension costs could be increased if 
workers continue participating in their pension plans; 
Potential effects on women: Some experts told us that these changes 
could encourage people to delay retirement, potentially increasing 
their retirement savings. Others are concerned that these options 
would be harmful for women. For example, if the full retirement age is 
increased and women who planned to claim at the old full retirement 
age do not delay collecting Social Security benefits, they would 
receive a lower benefit. 

Policy option: Increase duration of unemployment benefits in lieu of 
applying for Social Security early; 
Description of policy option: According to one expert we spoke with, 
the eligibility period for unemployment compensation could be extended 
further for older workers. This could increase federal tax revenue 
because unemployment compensation is taxable. However, paying more in 
unemployment benefits would exacerbate the financial challenges state 
unemployment insurance programs face[B]; 
Potential effects on women: Unemployment can have a negative effect on 
women's income security. This option would provide additional income 
to unemployed older women, who may find it difficult to find another 
job. Instead of applying for early Social Security benefits, which 
results in a permanently lower benefit level, women could rely on 
unemployment compensation, thus preserving the value of their Social 
Security benefits. 

Source: GAO analysis of literature and expert interviews.

[A] SSA's Office of the Chief Actuary has estimated the effect various 
proposals to increase the full retirement age would have on solvency. 
In 2011, the Office of the Chief Actuary estimated these proposals 
would improve solvency by 0.32 to 0.98 percent of payroll. See 
[hyperlink, http://www.ssa.gov/oact/solvency/provisions/index.html].

[B] In April 2010, we reported that state unemployment insurance trust 
funds stood in historically poor financial condition. See GAO, 
Unemployment Insurance Trust Funds: Long-standing State Financing 
Policies Have Increased Risk of Insolvency, GAO-10-440 (Washington, 
D.C.: Apr. 14, 2010). 

[End of table] 

Proposals to Ensure Lifetime Income:

Experts also identified several policies that would ensure lifetime 
retirement income for women (see table 7). Women may especially 
benefit from these options, given that they (1) have lower levels of 
retirement income than men, (2) are more likely to live longer, and 
(3) are also more likely to become widowed. For example, Treasury 
recently proposed modifying the required minimum distribution rules so 
that individuals could use part of their retirement savings to 
purchase a longevity annuity.[Footnote 57] This option would provide 
older women with guaranteed additional income, which may be helpful if 
they live long lives or outlive a spouse. These options, however, 
often have cost implications for either federal tax revenue or plan 
sponsors. For example, if individuals purchased longevity annuities 
using tax-qualified retirement savings, the tax revenue generated from 
withdrawing these savings would be deferred until the annuity started 
paying out.

Table 7: Proposals to Ensure Lifetime Income:

Policy option: Encourage DC plan sponsors to offer annuities as a 
distribution option for a portion or the entire DC account balance; 
Description of policy option: Experts reported that steps could be 
taken to decrease the risks employers face when they offer an annuity 
as a distribution option for DC plans. For example, one expert told us 
the rules for using DC savings to purchase an annuity could be 
revised. These options could introduce greater costs and 
administrative burdens for plan sponsors; 
Potential effects on women: More DC plan participants could have the 
opportunity to secure guaranteed lifetime income. This could be 
especially beneficial for women given that they tend to live longer 
than men, have higher poverty rates, and are more likely to be widowed. 

Policy option: Modify required minimum distribution rules to allow for 
longevity annuities; 
Description of policy option: This option would modify the required 
minimum distribution rules so that it is easier to purchase longevity 
annuities with a portion of DC plan assets.[A] In February, Treasury 
proposed a regulation that would alter the required minimum 
distribution rules to make it easier for individuals to use a portion 
of their savings to purchase longevity annuities.[B] Tax revenue would 
be deferred until the annuity starts paying out; 
Potential effects on women: A longevity annuity would decrease the 
chances that a woman would outlive her retirement savings. Given 
women's tendency to live longer than men, as well as their higher 
poverty rates and likelihood of being widowed, this option could be 
especially beneficial for improving women's retirement income security. 

Policy option: Reduce eligibility requirements for divorced spousal 
benefits under Social Security; 
Description of policy option: Currently, a divorced spouse can receive 
benefits based on a retired worker's earnings record if the marriage 
lasted at least 10 years, and the spouse is unmarried and at least 62 
years old. Experts have recommended expanding eligibility for divorce 
benefits to require a minimum of 7 years of marriage. Additionally, 
some experts have suggested marriage years could also be accumulated 
across multiple marriages. This option would increase Social Security 
costs and the administrative burden for SSA, while decreasing solvency; 
Potential effects on women: More divorced women would qualify for 
spousal benefits. One study estimated that lowering the marriage-
duration requirement from 10 to 7 years would increase benefits for 
about 8 percent of all divorced women age 62 and over in the year 
2030.[C] However, as we have previously reported, this option could 
benefit higher-income women who are not economically vulnerable and it 
would not benefit women who were never married.[D] 

Source: GAO analysis of literature and expert interviews.

[A] A longevity annuity (sometimes referred to as "longevity 
insurance" or a "deeply deferred annuity") is an income stream that 
can be purchased at or near retirement but begins at an advanced age--
for example, age 85--and continues as long as the individual lives.

[B] Longevity Annuity Contracts, 77 Fed. Reg. 5443 (Feb. 3, 2012).

[C] Christopher R. Tamborini and Kevin Whitman, "Lowering Social 
Security's Duration-of-Marriage Requirement: Distributional Effects 
for Future Female Retirees," Journal of Women and Aging vol. 22 (2010).

[D] GAO, Social Security: Options to Protect Benefits for Vulnerable 
Groups When Addressing Program Solvency, GAO-10-101R (Washington, 
D.C.: Dec. 7, 2009). 

[End of table] 

Proposals to Ensure Income Adequacy:

There are also a number of policy options that could enhance Social 
Security benefits for vulnerable groups at risk of not having 
sufficient income or assets in retirement, including widows, divorced 
women, low-income women and women age 85 and over (see table 
8).[Footnote 58] For example, increasing the Social Security 
Survivor's benefit to 75 percent of the deceased worker's benefit 
would provide widows with more monthly income, helping to keep some 
women out of poverty. However, all of these options would increase 
existing costs or introduce new costs and, in turn, would decrease the 
solvency of the system.

Table 8: Proposals to Ensure Income Adequacy:

Policy option: Use consumer price index for the elderly (CPI-E) to 
calculate Social Security cost-of-living adjustments; 
Description of policy option: Currently, the Consumer Price Index for 
Urban Wage Earners and Clerical Workers (CPI-W) is used to calculate 
annual cost- of-living adjustments for Social Security benefits. 
However, some experts argue that the CPI-W does not accurately reflect 
expenses for the elderly. The CPI-E, an index designed to represent 
expenses of those age 62 and over,[A] could be used to calculate cost-
of-living adjustments for Social Security recipients. Experts say an 
advantage of the CPI-E is that it more accurately reflects the 
typically larger share of expenditures older Americans spend on 
medical care. This option would decrease Social Security solvency 
because it would generally increase benefit levels and, therefore, 
costs[B]; 
Potential effects on women: Advocates for the CPI-E reported that it 
more accurately reflect expenses for retirees, thereby improving 
income adequacy by providing more appropriate cost-of-living 
adjustments. While all Social Security recipients would benefit, women 
could benefit more than men as they tend to live longer. Moreover, 
benefit increases compound over time. However, some advocates believe 
benefits would still be insufficient under the CPI-E. 

Policy option: Update the Social Security Special Minimum Benefit; 
Description of policy option: Currently, Social Security includes a 
Special Primary Insurance Amount (also referred to as the Special 
Minimum Benefit) that is intended to reduce poverty among retired 
lifetime low-wage workers. However, very few people receive this 
benefit.[C] There are several options for increasing the minimum 
benefit. For example, one proposal would increase the minimum benefit 
and index it to wages.[D] While benefits would increase, decreasing 
poverty for some beneficiaries, this option would increase costs and 
decrease solvency; 
Potential effects on women: An increased Special Minimum Benefit could 
keep more elderly women out of poverty by increasing their monthly 
income. In addition, our past work found that while the share of women 
affected by the minimum benefit was fairly similar across marital 
statuses (never-married, divorced, married and widowed), never-married 
and divorced women had much larger percent changes in median 
benefits.[E] 

Policy option: Provide an additional Social Security benefit to the 
oldest old; 
Description of policy option: Social Security recipients over the age 
of 80 or 85 could receive an additional benefit, such as an extra 5 
percent on top of their regular benefit. While this option would 
increase benefits for the oldest old, it would also increase costs and 
decrease solvency[F]; 
Potential effects on women: Women, who tend to live longer than men, 
would be more likely to receive this extra benefit. Older women may 
need extra benefit as income and assets may have been used to care for 
a deceased spouse or to pay for increasing medical costs. An 
additional benefit may be particularly helpful for low-income women. 

Policy option: Increase Social Security Survivor's benefits to 75%; 
Description of policy option: Currently, when someone is widowed, 
total household income from Social Security decreases by one-third if 
the couple's benefits had been based on one spouse's work history and 
up to 50 percent if both spouses had been receiving retired worker 
benefits. Survivor's benefits could be increased to 75 percent of the 
couple's retired-workers benefits. Experts have proposed calculating 
this new benefit in different ways. For example, the surviving spouse 
could receive 75 percent of the couple's retired-workers benefit but 
the benefit would be capped at the maximum earner's benefit or at the 
benefit of the "lifelong average earner." However, increasing benefits 
would increase costs and decrease solvency; 
Potential effects on women: Increasing Survivor's benefits would 
increase income for widowed women. Widowhood can have a devastating 
effect on women's household assets and income. Further, women are more 
likely than men to be widowed so they would be more likely to benefit 
from an increase in the survivor's benefit. In fact, when we simulated 
the effects of this option in 2007, we found that three times the 
number of women as men were affected. However, the magnitude of the 
benefit increase was larger for men than for women.[G] 

Policy option: Increase Social Security spousal benefits for divorced 
spouses; 
Description of policy option: Currently, divorced spouses who qualify 
for spousal benefits receive a benefit equal up to 50 percent of the 
worker's benefits. This option would raise benefits for divorced 
spouses to 75 percent of the former spouse's benefit while the former 
spouse is still alive. Upon the death of the former spouse, the 
divorced spouse would receive the full widow's benefit of 100 percent. 
This benefit increase would decrease solvency because it would 
increase costs; 
Potential effects on women: Divorce can result in a substantial loss 
of assets and income for women. Some experts argue that a 50 percent 
benefit is not enough to keep divorced women from falling into 
poverty. It has been estimated that increasing the benefit rate for 
divorced spouses to 75 percent would lower the poverty rate among 
divorced spouses from 30 percent to 11 percent.[H] 

Policy option: Increase Social Security benefits for disabled 
surviving spouses; 
Description of policy option: Currently, to qualify for disabled 
surviving spouse benefits, disabled surviving spouses must be at least 
age 50 and have become disabled before or within 7 years of the 
spouse's death or before or within 7 years after last being eligible 
for benefits as a caretaking parent or eligible surviving child. In 
addition, disabled surviving spouses younger than the full retirement 
age generally receive lower benefits than those who wait to receive 
their benefits until the full retirement age. This option would raise 
benefits for disabled surviving spouses to 100 percent of the deceased 
spouse's benefit. It would also remove the 7 year limitation and the 
age 50 requirement. Lastly, it would make divorced spouses who are 
disabled eligible for benefits on the same basis as disabled surviving 
spouses. Although benefits would increase, Social Security solvency 
would decrease; 
Potential effects on women: Both divorce and widowhood can result in a 
decrease in retirement security. Further, disabled surviving spouses, 
including those who have been divorced, cannot work and may have no 
partner to depend on for support. In addition, disability issues 
affect a surprisingly high number of divorced spouses, making them 
more vulnerable to income insecurity. One study estimated that more 
than one-fifth of all divorced spouses had health problems that meet 
disability criteria established by SSA.[I] 

Policy option: Increase continuation percentage for qualified joint- 
and-survivor annuities; 
Description of policy option: Currently, if a worker receives a joint 
and survivor annuity, when the worker passes away, the spouse 
continues to receive the annuity, but at not less than 50 percent of 
the amount the worker received. This option would increase the minimum 
continuation percentage to 66 or 75 percent; 
Potential effects on women: It is about 40 percent more expensive to 
live as a single retiree than as a married retiree. After becoming 
widowed, household annuity income would be reduced by a smaller amount 
than it is currently. However, by increasing the continuation 
percentage, the cost of the joint-and-survivor annuity could increase. 

Source: GAO analysis of literature and expert interviews.

[A] The CPI-E is an experimental index developed by the Bureau of 
Labor Statistics. It takes into account increased utilization of 
medical care by the elderly. Officials from the Bureau of Labor 
Statistics have cautioned against using the CPI-E for pension and 
other adjustments because it is only an approximation of an index for 
older Americans. See GAO, Income Security: Older Adults and the 2007-
2009 Recession. GAO-12-76 (Washington, D.C.: Oct. 17, 2011).

[B] SSA's Office of the Chief Actuary has estimated the effect of 
using the CPI-E to calculate cost-of-living-adjustments would have on 
solvency. In 2011, the Office of the Chief Actuary estimated that 
solvency would be decreased by 0.35 percent of payroll. See 
[hyperlink, http://www.ssa.gov/oact/solvency/provisions/index.html].

[C] Currently, few people qualify for the special minimum benefit 
because the eligibility threshold has not kept pace with wage growth.

[D] SSA's Office of the Chief Actuary has estimated the effects 
various proposals to increase the Special Minimum Benefit would have 
on solvency. In 2011, the Office of the Chief Actuary estimated these 
proposals would decrease solvency by 0.10 to 0.28 percent of payroll. 
For these estimates, see [hyperlink, 
http://www.ssa.gov/oact/solvency/provisions/index.html].

[E] See GAO-08-105.

[F] SSA's Office of the Chief Actuary has estimated the effects 
various proposals to increase benefits for those age 85 and over would 
have on solvency. In 2011, the Office of the Chief Actuary estimated 
these proposals would decrease solvency by 0.10 to 0.13 percent of 
payroll. For these estimates, see [hyperlink, 
http://www.ssa.gov/oact/solvency/provisions/index.html].

[G] See GAO-08-105.

[H] Alison M. Shelton and Dawn Nuschler, Social Security: Revisiting 
Benefits for Spouses and Survivors, Congressional Research Service 
(Washington, D.C.: Nov. 5, 2010).

[I] David A. Weaver, "The Economic Well-Being of Social Security 
Beneficiaries, with an Emphasis on Divorced Beneficiaries," Social 
Security Bulletin, vol. 60, no. 4 (1997). 

[End of table] 

Concluding Observations:

To retirement security experts, our findings paint a familiar if 
disconcerting picture. Although increases in women's labor force and 
retirement plan participation have led to a marginal improvement in 
women's prospects for achieving a more secure retirement, our report 
also highlights the substantial risks women continue to face in 
accumulating adequate retirement income. Yet, despite the differential 
risks women face, retirement security in America continues to be a 
national dilemma that transcends gender differences. It is important 
to note that much of the relative improvement in women's retirement 
security has been a consequence of deterioration in men's retirement 
security. Recent economic volatility, coupled with the continued shift 
toward defined contribution plans, exposes all workers to more 
financial risk than previous generations. Further, older workers' 
financial security is increasingly dependent on individual choices 
regarding how much to save, how to invest those savings, at what age 
to retire, and how to make those savings last throughout retirement. 
Much of the total workforce continues to approach retirement age with 
no traditional pension. Unchecked, this problem will only grow in 
severity.

Nevertheless, women face a unique set of circumstances, which warrant 
special attention. In particular, our findings show that the 
disruptions that occur as a result of later-in-life events, such as 
divorce and widowhood, can be financially devastating for women. In 
addition, women's greater likelihood of being single, higher life 
expectancy, and lower average earnings continue to make saving for 
retirement and avoiding late-life poverty a challenge.

The challenges facing women's retirement income security do not lack 
for potential resolutions. In fact, our discussions with experts 
identified a number of policy options that would improve retirement 
income security for women. These options range from changes to Social 
Security to altering the private pension system. While these options 
involve tradeoffs and difficult choices, they have the potential to 
improve the retirement income security of men as well. Ultimately, 
such efforts provide opportunities to improve the retirement security 
of many Americans.

Agency Comments:

We provided a draft of this report to the Department of Labor, the 
Department of the Treasury, and the Social Security Administration for 
review and comment. While none of the agencies provided official 
comments, each provided technical comments, which we incorporated as 
appropriate.

As agreed with your office, unless you publicly announce its contents 
earlier, we plan no further distribution until 30 days after the date 
of this letter. At that time, we will send copies of this report to 
the Secretary of Labor, the Secretary of the Treasury, the 
Commissioner of Social Security, and other interested parties. In 
addition, the report will be available at no charge on the GAO website 
at [hyperlink, http://www.gao.gov].

If you or your staff have any questions about this report, please 
contact me at (202) 512-7215 or jeszeckc@gao.gov. Contact points for 
our Offices of Congressional Relations and Public Affairs may be found 
on the last page of this report. GAO staff who made contributions to 
this report are listed in appendix II.

Sincerely yours, 

Signed by: 

Charles A. Jeszeck: 
Director Education, Workforce,  and Income Security: 

[End of section] 

Appendix I: Objective, Scope, and Methods:

To analyze factors that affect women's retirement security, we 
examined (1) how women's access to and participation in employer-
sponsored retirement plans compare to men's and how they have changed 
over time; (2) how women's retirement income compares to men's and how 
the composition of their income has changed with economic conditions 
and trends in pension design; (3) how events occurring later in life 
affect women's retirement income; and (4) what policy options are 
available to help increase women's retirement income security. This 
appendix provides a detailed account of the information and methods we 
used to answer these questions. Section 1 describes the key 
information sources we used. Sections 2 through 4 describe the 
empirical methods we used to answer questions 1 through 3 respectively 
and the results of supplementary analyses.

Section 1: Information Sources:

To answer our questions, we obtained information from a variety of 
sources including two nationally representative surveys--the Survey of 
Income and Program Participation (SIPP) and the Health and Retirement 
Study (HRS)--the academic literature on retirement security, and a 
range of experts in the area of women's retirement security. Table 9 
summarizes the data sources used to answer each question. This section 
provides a description of our data sources and the steps we took to 
ensure their reliability.

Table 9: Data Sources Used for Each Reporting Objective:

Objective 1: Women and men's access to employer-sponsored pension 
plans; 
SIPP data: [Check]; 
HRS data: [Empty]; 
Academic literature: [Check]; 
Expert opinions[A]: [Check].

Objective 2: Women's and men's retirement income sources; 
SIPP data: [Check]; 
HRS data: [Empty]; 
Academic literature: [Check]; 
Expert opinions[A]: [Check].

Objective 3: Impact of late-in-life events on retirement income and 
assets; 
SIPP data: [Empty]; 
HRS data: [Check]; 
Academic literature: [Check]; 
Expert opinions[A]: [Check].

Objective 4: Policy options; 
SIPP data: [Empty]; 
HRS data: [Empty]; 
Academic literature: [Check]; 
Expert opinions[A]: [Check].

Source: GAO.

[A] Expert opinions were gathered from the literature and our 
interviews. We interviewed experts from government, academia, advocacy 
groups, and the private sector. For more information about our 
literature review and expert interviews, see below. 

[End of table] 

Survey of Income and Program Participation:

To answer Questions 1 and 2, we analyzed data collected through the 
SIPP, a nationally representative survey conducted by the U.S. Census 
Bureau that collects detailed information on income sources and 
pension plan coverage, among many other areas. The survey is conducted 
in a series of national panels, with sample sizes ranging from 
approximately 14,000 to 36,700 interviewed households. The duration of 
each panel ranges from 2 ½ years to 4 years. Within each panel, the 
data are collected in a series of "waves" which take place in 4-month 
cycles. Within each wave, Census administers a core survey consisting 
of questions that are asked at every interview, and several modules 
relating to a particular topic. We used data from the core survey and 
the topical module on retirement and pension coverage from the last 
four SIPP panels, which began in 1996, 2001, 2004, and 2008 
respectively. For all but the 2008 panel, the topical module on 
retirement and pension coverage was administered in Wave 7. For 
objective 1, we matched core data from Wave 3 of the 2008 panel with 
the topical module data, which was also administered in Wave 3. This 
ensured that the demographic data used in the analysis for that 
objective would match the time frame of the topical module data. 
However, to obtain the most up to date income data for objective 2, we 
used core data from Wave 7, which was the most recently available data 
as of October 2011. Table 10 shows the waves and questionnaires we 
used to answer each objective. It also shows the years that the data 
were collected during each panel and wave listed. The bolded years 
correspond to the years of data that are presented in the figures in 
objectives 1 and 2.

Table 10: SIPP Panels, Waves, and Questionnaires Used to Answer 
Objective 1 and Objective 2:

1996 Panel, Wave 7, Core questionnaire; 
Year data were collected: 1997, 1998[A]; 
Objective 1: [Check]; 
Objective 2: [Check].

1996 Panel, Wave 7, Topical Module on Retirement and Pension Plan 
Coverage; 
Year data were collected: 1997,1998[A]; 
Objective 1: [Check]; 
Objective 2: [Empty].

2001 Panel, Wave 7, Core questionnaire; 
Year data were collected: 2002, 2003[A]; 
Objective 1: [Check]; 
Objective 2: [Check].

2001 Panel, Wave 7, Topical Module on Retirement and Pension Plan 
Coverage; 
Year data were collected: 2002, 2003[A]; 
Objective 1: [Check]; 
Objective 2: [Empty].

2004 Panel, Wave 7, Core questionnaire; 
Year data were collected: 2005, 2006[A]; 
Objective 1: [Check]; 
Objective 2: [Check].

2004 Panel, Wave 7, Topical Module on Retirement and Pension Plan 
Coverage; 
Year data were collected: 2005, 2006[A]; 
Objective 1: [Check]; 
Objective 2: [Empty].

2008 Panel, Wave 3, Core questionnaire; 
Year data were collected: 2009; 
Objective 1: [Check]; 
Objective 2: [Check].

2008 Panel, Wave 3, Topical Module on Retirement and Pension Plan 
Coverage; 
Year data were collected: 2009; 
Objective 1: [Check]; 
Objective 2: [Empty].

2008 Panel, Wave 7, Core questionnaire; 
Year data were collected: 2010; 
Objective 1: [Empty]; 
Objective 2: [Check].

Source: GAO.

[A] In this report, the data are described by referring to the year 
from which the majority of the data was collected. For example, the 
2001 Wave 7 data is described as "2003 data" because the reference 
periods for 10 of the 16 rotation groups in this wave were in calendar 
year 2003. 

[End of table] 

In comparison to other nationally representative surveys, the SIPP had 
several main advantages. First, the SIPP collects separate information 
on defined benefit (DB) and defined contribution (DC) plans. Other 
surveys, such as the Current Population Survey, do not distinguish 
between income from and participation in DB and DC plans. Second, the 
SIPP sample is larger than comparable surveys, such as the Survey of 
Consumer Finances (SCF). Consequently, it is possible to produce point 
estimates for demographic subcategories with a higher degree of 
reliability. Further, in comparison to the SCF, which oversamples 
wealthy households, the SIPP oversamples lower-income households-- 
arguably an important component of an analysis of income security.

Despite its advantages, the SIPP has two limitations for our analysis. 
First, as with most survey data, SIPP data are self-reported. This can 
be problematic for the reporting of data on income sources and pension 
plan participation. For example, respondents might incorrectly report 
that they participate in a pension plan when they do not participate 
in one.[Footnote 59] Second, despite the fact that SIPP differentiates 
between participation in a DB or DC plan, it does not contain full 
information on whether an individual's employer offers a DB 
plan.[Footnote 60]

Health and Retirement Study:

To answer question 3--on the effects of events occurring later in life 
on women's retirement income security--we analyzed data collected 
through the HRS, a nationally representative survey primarily 
sponsored by the National Institute of Aging and conducted by the 
Institute for Social Research at the University of Michigan. This 
longitudinal survey collects data on individuals over age 50 and 
contains detailed information on health, marital status, assets, 
income, and care for elders. Respondents were first surveyed in 1992, 
when they were age 51 to 61 and continued to be surveyed every 2 
years. Additional cohorts were added in later years to maintain the 
representation of the older population. Table 11 presents the cohorts 
that are included in the HRS sample. Respondents are resurveyed every 
2 years. The data in our analysis span from the initial 1992 survey 
through the early release data for 2010, the most current data 
available. Our analysis follows over 30,000 individuals from the HRS 
sample.

Table 11: Birth Years for the HRS Cohorts and the Year Data Collection 
Began for Each Cohort:

Cohort: AHEAD[A]; 
Birth years: 1923 and earlier; 
Year data collection began: 1993.

Cohort: Children of the Depression Era (CODA)[A]; 
Birth years: 1924-1930; 
Year data collection began: 1998.

Cohort: Original HRS cohort; 
Birth years: 1931-1941; 
Year data collection began: 1992.

Cohort: War Babies[A]; 
Birth years: 1942-1947; 
Year data collection began: 1998.

Cohort: Early Baby Boomers; 
Birth years: 1948-1953; 
Year data collection began: 2004.

Source: RAND HRS Data Documentation, Version L.

[A] The Asset and Health Dynamics of the Oldest Old (AHEAD) survey 
began collecting data in 1993. Originally, the HRS and AHEAD were 
separate but related surveys. The AHEAD survey was initially funded as 
a supplement to the HRS. In 1998, the two surveys merged and the CODA 
and War Babies cohorts were added to the survey. 

[End of table] 

One of the main advantages of the HRS is that the same households are 
interviewed at different points of time, allowing us to examine the 
correlation of changes in life events to changes in household assets 
and income. Further, RAND, a research organization, cleans and 
processes the HRS data to create a user-friendly longitudinal dataset 
that has consistent and intuitive naming conventions, model-based 
imputations for missing wealth and income data, and spousal 
counterparts of most individual-level variables. We used these data 
for our analysis.

However, there are three limitations for our analysis. First, the 
women currently in the HRS survey may have very different retirement 
experiences from women in the workforce today due to changes in 
demographic trends and workforce participation. Second, as with the 
SIPP, data from the HRS are self-reported. Third, total household 
assets cannot be broken out at the individual level.

Data Reliability:

For each of the datasets described above, we conducted a data 
reliability assessment of selected variables by conducting electronic 
data tests for completeness and accuracy, reviewing documentation on 
the dataset, or interviewing knowledgeable officials about how the 
data are collected and maintained and their appropriate uses. When we 
learned that particular fields were not sufficiently reliable, we did 
not use them in our analysis. For example, we chose not to use data 
from the SIPP Topical Module on Annual Income and Retirement Accounts 
because many of the fields in that survey are not edited by the Census 
Bureau. For the purposes of our analysis, we found the variables that 
we ultimately reported on to be sufficiently reliable.

Literature Review and Interviews:

To gain an understanding of the challenges women face in attaining a 
secure retirement and policy options that could enhance women's 
retirement security, we conducted an extensive literature review and 
interviewed a range of experts. To identify existing studies, we 
conducted searches of various databases, such as EconLit, Electronic 
Collections Online, ProQuest, Academic OneFile, WorldCat, and Policy 
File. From these sources, we identified 128 articles that appeared in 
journals since 2007 and were relevant to our research objective on 
policy options that could enhance women's retirement security. From 
the articles identified in the preliminary search, we reviewed article 
abstracts, when available, to determine which articles contained 
information germane to our report and reviewed those articles. In 
addition, we reviewed articles that were collected during the previous 
GAO study on women's retirement security that contained information 
relevant to our empirical analyses, described below, and reviewed 
articles that were suggested to us by the experts we interviewed. We 
performed these searches and identified articles from May 2011 to 
October 2011.

To supplement the literature review, we conducted interviews with 
experts. To ensure that we obtained a balanced perspective, we 
interviewed experts with a range of perspectives and from different 
types of organizations including government, academia, advocacy 
groups, and the private sector. We also consulted several experts in 
government and academia on technical issues related to our analysis. 
Specifically, we interviewed agency officials at the departments of 
the Treasury and Labor, the Social Security Administration, and the 
Bureau of the Census; academic experts at the Employee Benefits 
Research Institute, Heritage Foundation, University of Pennsylvania, 
Stanford University, Urban Institute, and Wellesley College; and 
industry experts and advocates from the American Council on Life 
Insurers, Anna Rappaport Consulting, Financial Engines, the Institute 
for Women's Policy Research, the National Women's Law Center, AARP, 
the Pension Rights Center, the National Academy of Social Insurance, 
Social Security Works, and the Women's Institute for a Secure 
Retirement.

Section 2: Methods for Comparing Working Women's and Men's Access to 
and Participation in Employer-Sponsored Pension Plans:

To determine the proportion of men and women that (1) work for an 
employer that offers a plan, (2) are eligible for a plan, and (3) 
participate in a plan, we used data from the SIPP topical module on 
retirement and pension plan coverage. Specifically, we constructed 
five dummy variables using a combination of various questions in the 
SIPP. Table 12 shows the information we used to construct each 
variable. For each of these variables, we used SIPP individual-level 
weights to compute point estimates and, in conjunction with other 
factors, calculate the standard errors of those estimates so that we 
could accurately account for the complex survey design. We consulted 
statisticians from the U.S. Bureau of the Census on the appropriate 
use of these weights.

Table 12: Information Used from SIPP to Construct Key Variables:

Variable: Worker has employer that offers either a DB or a DC pension 
plan to some employees; 
Constructed with: A combination of two questions. One question asks 
whether the individual's job or business has any kind of pension or 
retirement plan for anyone in the company or organization, and a 
subsequent clarifying question asks if the individual's job or 
business offers a DC plan. 

Variable: Worker has employer that offers a DC pension plan to some 
employees; 
Constructed with: A combination of questions. If the respondent 
replied yes to the question listed above, a follow-up question is 
asked about whether the respondent participates in the plan, and if 
so, the type of plan. This series of questions enables us to identify, 
among those who participate, whether the individual's employer offers 
a DC plan. For those that said that their employer does not offer a 
pension or retirement plan, and those who said that their employer 
offers a plan but it does not include a DC-type component, SIPP asks a 
follow-up question about whether the employer offers a DC- type plan. 
By combining these two sets of information, we were able to construct 
a dummy variable to indicate whether the individual's employer offers 
a DC plan. 

Variable: Worker is eligible for employer-sponsored plan; 
Constructed with: A question in the SIPP topical module on retirement 
and pension plan coverage that asks the reason for not participating 
in the employer's plan. We defined individuals as not eligible if they 
listed one of the following reasons for not participating: no one in 
their type of job is eligible; they don't work enough hours, days, 
weeks or months; they don't have enough tenure in the job; they are 
too young; they started their job too close to retirement. We defined 
individuals as eligible if they participated in the plan or listed 
some other reason for not participating. 

Variable: Worker participates in employer-sponsored DB or DC plan; 
Constructed with: A combination of two questions. One question asks 
whether the individual participates in the employer-sponsored plan, 
and a subsequent clarifying question asks if the individual 
participates in an employer-sponsored DC plan. 

Variable: Worker participates in employer-sponsored DC plan; 
Constructed with:: A combination of questions. If the respondent 
replied yes to the question above and the respondent indicates that 
the type of plan in which he or she participated was a DC plan. 

Source: GAO analysis of SIPP questionnaire. 

[End of table] 

To better understand the factors that might explain gender differences 
in each of these variables, we developed a series of empirical models. 
Following the literature, we controlled for the following factors in 
our models: (1) demographic characteristics including gender, age, 
marital status, children present in the household, single parenthood, 
race and ethnicity, citizenship, immigrant status, and education 
level; and (2) occupational characteristics including part-time 
employment status, self-employment status, years of tenure, work 
experience, occupation, industry, sector, union status, and size of 
employing firm.[Footnote 61] To estimate these models, we used 
logistic regression--an appropriate technique when the dependent 
variable is binary, or has two categories such as participating in a 
plan or not participating in a plan. Logistic regression also allows 
for the coefficients to be converted into odds ratios, which are 
described below.

We conducted the modeling analyses in a series of steps whereby with 
each step, the sample of men and women that was included in the 
analysis was conditional on the previous step. Specifically, the first 
analysis involved analyzing the probability of working for an employer 
that offered a pension plan for all workers in the sample. The second 
analysis involved analyzing the probability of being eligible for a 
plan for those men and women that worked for an employer offering a 
plan. The third analysis involved analyzing the probability of 
participating in a plan for those that were eligible for their 
employer-sponsored plan.

Changes in the Working Population Over Time by Gender:

In conjunction with understanding the factors associated with each 
dependent variable in our models, it is essential to also understand 
how women and men differ in those factors. Taken together, the 
information from the model and information from a comparison of men's 
and women's characteristics enables us to understand what factors make 
women more or less likely to be employed by an employer that offers a 
plan, be eligible for the plan, and participate in the plan. For 
example, if we know that women are disproportionately more likely to 
work part-time and that part-time status is an important factor 
associated with plan participation, we can infer that women's higher 
rates of part-time status might contribute to their lower rates of 
plan participation. Table 13 compares the characteristics of men and 
women for each of the factors that we control for, across each year of 
the study period.

Generally, the characteristics of men and women in the working 
population did not change dramatically over the study period. 
Correspondingly, when we compare men and women in each year, several 
relationships between them were consistent across all of the study 
years. In terms of demographic characteristics, women were more likely 
than men to be widowed and divorced. Women were also more likely to 
have children present in the household, be single parents, and work 
part time. A higher proportion of men than women were Hispanic, and 
this proportion increased over the study period.[Footnote 62]

In terms of occupational characteristics, several gender differences 
persisted across the study years. Women consistently had higher levels 
of education and were more likely to work in the public or nonprofit 
sectors. Men were more likely to work in the private sector, be self- 
employed, have longer tenure at their current position, have more work 
experience, and to be in a union.

Although the occupational and industry categories in the SIPP data 
changed midway through the study periods, the distributions of men and 
women across occupations and industry were generally consistent for 
the last 2 study years. Specifically, the top three occupations for 
women were office and administrative support; sales and related 
services; and education, training, and library services, with 20, 10, 
and 10 percent of women working in these occupations respectively in 
2009. Men tended not to be as concentrated in just a few occupations. 
In 2009, the highest proportions of men were employed in management (9 
percent), sales and related occupations (8 percent), construction and 
extraction (8 percent), and transportation and material moving (8 
percent). Similarly, in 2009, the top three industries for women were 
health care and social assistance (21 percent), educational services 
(14 percent), and retail trade (10 percent). For men in this year, the 
top three industries in which men were employed were manufacturing (13 
percent), construction (9 percent), and retail trade (9 percent).

Table 13: Characteristics of the Working Population over Time:

Gender; 
1999: Percentage of: Men: 53; 
1999: Percentage of: Women: 47; 
2003: Percentage of: Men: 53; 
2003: Percentage of: Women: 47; 
2006: Percentage of: Men: 53; 
2006: Percentage of: Women: 47; 
2009: Percentage of: Men: 53; 
2009: Percentage of: Women: 47.

Age groups: 

18-24; 
1999: Percentage of: Men: 12; 
1999: Percentage of: Women: 13; 
2003: Percentage of: Men: 12; 
2003: Percentage of: Women: 13; 
2006: Percentage of: Men: 13; 
2006: Percentage of: Women: 13; 
2009: Percentage of: Men: 12; 
2009: Percentage of: Women: 13.

25-34; 
1999: Percentage of: Men: 26; 
1999: Percentage of: Women: 25; 
2003: Percentage of: Men: 24; 
2003: Percentage of: Women: 23; 
2006: Percentage of: Men: 23; 
2006: Percentage of: Women: 22; 
2009: Percentage of: Men: 23; 
2009: Percentage of: Women: 22.

35-44; 
1999: Percentage of: Men: 29; 
1999: Percentage of: Women: 29; 
2003: Percentage of: Men: 27; 
2003: Percentage of: Women: 27; 
2006: Percentage of: Men: 26; 
2006: Percentage of: Women: 25; 
2009: Percentage of: Men: 24; 
2009: Percentage of: Women: 23.

45-54; 
1999: Percentage of: Men: 22; 
1999: Percentage of: Women: 22; 
2003: Percentage of: Men: 24; 
2003: Percentage of: Women: 25; 
2006: Percentage of: Men: 24; 
2006: Percentage of: Women: 25; 
2009: Percentage of: Men: 25; 
2009: Percentage of: Women: 26.

55-64; 
1999: Percentage of: Men: 11; 
1999: Percentage of: Women: 11; 
2003: Percentage of: Men: 13; 
2003: Percentage of: Women: 13; 
2006: Percentage of: Men: 14; 
2006: Percentage of: Women: 14; 
2009: Percentage of: Men: 16; 
2009: Percentage of: Women: 16.

Marital status: 

Married; 
1999: Percentage of: Men: 62; 
1999: Percentage of: Women: 57; 
2003: Percentage of: Men: 61; 
2003: Percentage of: Women: 57; 
2006: Percentage of: Men: 59; 
2006: Percentage of: Women: 56; 
2009: Percentage of: Men: 59; 
2009: Percentage of: Women: 55.

Widowed; 
1999: Percentage of: Men: 1; 
1999: Percentage of: Women: 3; 
2003: Percentage of: Men: 1; 
2003: Percentage of: Women: 2; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 2.

Divorced; 
1999: Percentage of: Men: 9; 
1999: Percentage of: Women: 13; 
2003: Percentage of: Men: 10; 
2003: Percentage of: Women: 14; 
2006: Percentage of: Men: 10; 
2006: Percentage of: Women: 13; 
2009: Percentage of: Men: 9; 
2009: Percentage of: Women: 13.

Separated; 
1999: Percentage of: Men: 2; 
1999: Percentage of: Women: 3; 
2003: Percentage of: Men: 2; 
2003: Percentage of: Women: 3; 
2006: Percentage of: Men: 2; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 2; 
2009: Percentage of: Women: 2.

Never married; 
1999: Percentage of: Men: 26; 
1999: Percentage of: Women: 23; 
2003: Percentage of: Men: 27; 
2003: Percentage of: Women: 24; 
2006: Percentage of: Men: 29; 
2006: Percentage of: Women: 26; 
2009: Percentage of: Men: 29; 
2009: Percentage of: Women: 27.

Children in the household; 
1999: Percentage of: Men: 46; 
1999: Percentage of: Women: 49; 
2003: Percentage of: Men: 44; 
2003: Percentage of: Women: 46; 
2006: Percentage of: Men: 44; 
2006: Percentage of: Women: 47; 
2009: Percentage of: Men: 42; 
2009: Percentage of: Women: 44.

Single parent; 
1999: Percentage of: Men: 8; 
1999: Percentage of: Women: 16; 
2003: Percentage of: Men: 7; 
2003: Percentage of: Women: 16; 
2006: Percentage of: Men: 8; 
2006: Percentage of: Women: 16; 
2009: Percentage of: Men: 8; 
2009: Percentage of: Women: 16.

Race and ethnicity: 

White, Non-Hispanic; 
1999: Percentage of: Men: 76; 
1999: Percentage of: Women: 75; 
2003: Percentage of: Men: 73; 
2003: Percentage of: Women: 72; 
2006: Percentage of: Men: 69; 
2006: Percentage of: Women: 70; 
2009: Percentage of: Men: 69; 
2009: Percentage of: Women: 70.

Black, Non-Hispanic; 
1999: Percentage of: Men: 9; 
1999: Percentage of: Women: 12; 
2003: Percentage of: Men: 9; 
2003: Percentage of: Women: 12; 
2006: Percentage of: Men: 10; 
2006: Percentage of: Women: 12; 
2009: Percentage of: Men: 9; 
2009: Percentage of: Women: 12.

Hispanic; 
1999: Percentage of: Men: 11; 
1999: Percentage of: Women: 9; 
2003: Percentage of: Men: 14; 
2003: Percentage of: Women: 11; 
2006: Percentage of: Men: 15; 
2006: Percentage of: Women: 11; 
2009: Percentage of: Men: 16; 
2009: Percentage of: Women: 12.

Asian, Non-Hispanic; 
1999: Percentage of: Men: 3; 
1999: Percentage of: Women: 4; 
2003: Percentage of: Men: 4; 
2003: Percentage of: Women: 4; 
2006: Percentage of: Men: 4; 
2006: Percentage of: Women: 3; 
2009: Percentage of: Men: 4; 
2009: Percentage of: Women: 4.

Other, Non-Hispanic; 
1999: Percentage of: Men: 1; 
1999: Percentage of: Women: 1; 
2003: Percentage of: Men: 1; 
2003: Percentage of: Women: 1; 
2006: Percentage of: Men: 2; 
2006: Percentage of: Women: 3; 
2009: Percentage of: Men: 2; 
2009: Percentage of: Women: 3.

Citizenship: 

Noncitizen; 
1999: Percentage of: Men: 7; 
1999: Percentage of: Women: 6; 
2003: Percentage of: Men: 9; 
2003: Percentage of: Women: 6; 
2006: Percentage of: Men: 10; 
2006: Percentage of: Women: 7; 
2009: Percentage of: Men: 10; 
2009: Percentage of: Women: 6.

Immigrant status: 

Naturalized immigrant; 
1999: Percentage of: Men: 4; 
1999: Percentage of: Women: 4; 
2003: Percentage of: Men: 5; 
2003: Percentage of: Women: 5; 
2006: Percentage of: Men: 6; 
2006: Percentage of: Women: 6; 
2009: Percentage of: Men: 7; 
2009: Percentage of: Women: 7.

Education level: 

No high school diploma; 
1999: Percentage of: Men: 12; 
1999: Percentage of: Women: 9; 
2003: Percentage of: Men: 12; 
2003: Percentage of: Women: 8; 
2006: Percentage of: Men: 8; 
2006: Percentage of: Women: 5; 
2009: Percentage of: Men: 8; 
2009: Percentage of: Women: 5.

High school diploma; 
1999: Percentage of: Men: 32; 
1999: Percentage of: Women: 30; 
2003: Percentage of: Men: 29; 
2003: Percentage of: Women: 27; 
2006: Percentage of: Men: 30; 
2006: Percentage of: Women: 26; 
2009: Percentage of: Men: 27; 
2009: Percentage of: Women: 23.

Some college; 
1999: Percentage of: Men: 30; 
1999: Percentage of: Women: 34; 
2003: Percentage of: Men: 30; 
2003: Percentage of: Women: 35; 
2006: Percentage of: Men: 35; 
2006: Percentage of: Women: 39; 
2009: Percentage of: Men: 35; 
2009: Percentage of: Women: 39.

Bachelor's degree or higher; 
1999: Percentage of: Men: 26; 
1999: Percentage of: Women: 27; 
2003: Percentage of: Men: 29; 
2003: Percentage of: Women: 29; 
2006: Percentage of: Men: 28; 
2006: Percentage of: Women: 31; 
2009: Percentage of: Men: 30; 
2009: Percentage of: Women: 33.

Part-time status[A]: 

Part time; 
1999: Percentage of: Men: 22; 
1999: Percentage of: Women: 37; 
2003: Percentage of: Men: 23; 
2003: Percentage of: Women: 38; 
2006: Percentage of: Men: 22; 
2006: Percentage of: Women: 36; 
2009: Percentage of: Men: 26; 
2009: Percentage of: Women: 37.

Self-employment status: 

Self-employed; 
1999: Percentage of: Men: 16; 
1999: Percentage of: Women: 10; 
2003: Percentage of: Men: 15; 
2003: Percentage of: Women: 10; 
2006: Percentage of: Men: 15; 
2006: Percentage of: Women: 10; 
2009: Percentage of: Men: 16; 
2009: Percentage of: Women: 10.

Average years of tenure at current job; 
1999: Percentage of: Men: 8.0; 
1999: Percentage of: Women: 6.9; 
2003: Percentage of: Men: 8.0; 
2003: Percentage of: Women: 7.0; 
2006: Percentage of: Men: 7.8; 
2006: Percentage of: Women: 7.2; 
2009: Percentage of: Men: 8.2; 
2009: Percentage of: Women: 7.7.

Work experience: 

Less than 5 years; 
1999: Percentage of: Men: 26; 
1999: Percentage of: Women: 29; 
2003: Percentage of: Men: 26; 
2003: Percentage of: Women: 31; 
2006: Percentage of: Men: 25; 
2006: Percentage of: Women: 28; 
2009: Percentage of: Men: 26; 
2009: Percentage of: Women: 30.

5 to 9 years; 
1999: Percentage of: Men: 14; 
1999: Percentage of: Women: 17; 
2003: Percentage of: Men: 14; 
2003: Percentage of: Women: 15; 
2006: Percentage of: Men: 14; 
2006: Percentage of: Women: 16; 
2009: Percentage of: Men: 14; 
2009: Percentage of: Women: 16.

10 to 15 years; 
1999: Percentage of: Men: 11; 
1999: Percentage of: Women: 12; 
2003: Percentage of: Men: 11; 
2003: Percentage of: Women: 11; 
2006: Percentage of: Men: 10; 
2006: Percentage of: Women: 11; 
2009: Percentage of: Men: 12; 
2009: Percentage of: Women: 12.

More than 15 years; 
1999: Percentage of: Men: 49; 
1999: Percentage of: Women: 43; 
2003: Percentage of: Men: 49; 
2003: Percentage of: Women: 43; 
2006: Percentage of: Men: 50; 
2006: Percentage of: Women: 45; 
2009: Percentage of: Men: 48; 
2009: Percentage of: Women: 41.

Average years of total work experience; 
1999: Percentage of: Men: 10.8; 
1999: Percentage of: Women: 9.4; 
2003: Percentage of: Men: 11.1; 
2003: Percentage of: Women: 9.7; 
2006: Percentage of: Men: 11.4; 
2006: Percentage of: Women: 10.3; 
2009: Percentage of: Men: 11.9; 
2009: Percentage of: Women: 10.7.

Sector: 

Private for profit; 
1999: Percentage of: Men: 70; 
1999: Percentage of: Women: 63; 
2003: Percentage of: Men: 69; 
2003: Percentage of: Women: 62; 
2006: Percentage of: Men: 71; 
2006: Percentage of: Women: 62; 
2009: Percentage of: Men: 68; 
2009: Percentage of: Women: 61.

Private not for profit; 
1999: Percentage of: Men: 4; 
1999: Percentage of: Women: 10; 
2003: Percentage of: Men: 4; 
2003: Percentage of: Women: 10; 
2006: Percentage of: Men: 4; 
2006: Percentage of: Women: 10; 
2009: Percentage of: Men: 4; 
2009: Percentage of: Women: 10.

Government; 
1999: Percentage of: Men: 13; 
1999: Percentage of: Women: 18; 
2003: Percentage of: Men: 13; 
2003: Percentage of: Women: 19; 
2006: Percentage of: Men: 12; 
2006: Percentage of: Women: 18; 
2009: Percentage of: Men: 14; 
2009: Percentage of: Women: 19.

Family worker without pay; 
1999: Percentage of: Men: 1; 
1999: Percentage of: Women: 1; 
2003: Percentage of: Men: 0; 
2003: Percentage of: Women: 1; 
2006: Percentage of: Men: 0; 
2006: Percentage of: Women: 1; 
2009: Percentage of: Men: 0; 
2009: Percentage of: Women: 1.

Not in universe; 
1999: Percentage of: Men: 13; 
1999: Percentage of: Women: 7; 
2003: Percentage of: Men: 13; 
2003: Percentage of: Women: 8; 
2006: Percentage of: Men: 12; 
2006: Percentage of: Women: 8; 
2009: Percentage of: Men: 13; 
2009: Percentage of: Women: 8.

Union status: 

In union; 
1999: Percentage of: Men: 17; 
1999: Percentage of: Women: 12; 
2003: Percentage of: Men: 15; 
2003: Percentage of: Women: 12; 
2006: Percentage of: Men: 14; 
2006: Percentage of: Women: 12; 
2009: Percentage of: Men: 15; 
2009: Percentage of: Women: 13.

Occupation: 

Management; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 9; 
2006: Percentage of: Women: 7; 
2009: Percentage of: Men: 9; 
2009: Percentage of: Women: 7.

Business and Financial Operations; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 3; 
2006: Percentage of: Women: 4; 
2009: Percentage of: Men: 3; 
2009: Percentage of: Women: 5.

Computer and Mathematical; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 3; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 4; 
2009: Percentage of: Women: 1.

Architecture and Engineering; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 3; 
2006: Percentage of: Women: 1; 
2009: Percentage of: Men: 3; 
2009: Percentage of: Women: 1.

Life, Physical, and Social Services; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 1; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 1.

Community and Social Services; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 2.

Legal; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 1; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 1.

Education, Training, and Library; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 3; 
2006: Percentage of: Women: 10; 
2009: Percentage of: Men: 3; 
2009: Percentage of: Women: 10.

Arts, Design, Entertainment, Sports, and Media; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 1; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 1.

Healthcare Practitioners and Technical; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 2; 
2006: Percentage of: Women: 8; 
2009: Percentage of: Men: 2; 
2009: Percentage of: Women: 8.

Healthcare Support; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 0; 
2006: Percentage of: Women: 4; 
2009: Percentage of: Men: 0; 
2009: Percentage of: Women: 4.

Protective Service; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 3; 
2006: Percentage of: Women: 1; 
2009: Percentage of: Men: 3; 
2009: Percentage of: Women: 1.

Food Preparation and Serving Related; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 4; 
2006: Percentage of: Women: 6; 
2009: Percentage of: Men: 4; 
2009: Percentage of: Women: 6.

Building and Grounds Cleaning and Maintenance; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 4; 
2006: Percentage of: Women: 3; 
2009: Percentage of: Men: 4; 
2009: Percentage of: Women: 3.

Personal Care and Service; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 4; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 4.

Sales and Related; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 8; 
2006: Percentage of: Women: 10; 
2009: Percentage of: Men: 8; 
2009: Percentage of: Women: 10.

Office and Administrative Support; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 7; 
2006: Percentage of: Women: 22; 
2009: Percentage of: Men: 6; 
2009: Percentage of: Women: 20.

Farming, Forestry, and Fishing; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 0; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 0.

Construction and Extraction; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 9; 
2006: Percentage of: Women: 0; 
2009: Percentage of: Men: 8; 
2009: Percentage of: Women: 0.

Installation, Repair, and Maintenance; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 6; 
2006: Percentage of: Women: 0; 
2009: Percentage of: Men: 6; 
2009: Percentage of: Women: 0.

Production; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 9; 
2006: Percentage of: Women: 4; 
2009: Percentage of: Men: 7; 
2009: Percentage of: Women: 3.

Transportation and Material Moving; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 9; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 8; 
2009: Percentage of: Women: 2.

Not in universe[B]; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 12; 
2006: Percentage of: Women: 7; 
2009: Percentage of: Men: 14; 
2009: Percentage of: Women: 8.

Industry: 

Agriculture; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 0; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 1.

Mining; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 0; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 0.

Utilities; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 0; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 0.

Construction; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 10; 
2006: Percentage of: Women: 1; 
2009: Percentage of: Men: 9; 
2009: Percentage of: Women: 1.

Manufacturing; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 15; 
2006: Percentage of: Women: 7; 
2009: Percentage of: Men: 13; 
2009: Percentage of: Women: 6.

Wholesale Trade; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 4; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 3; 
2009: Percentage of: Women: 2.

Retail Trade; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 10; 
2006: Percentage of: Women: 10; 
2009: Percentage of: Men: 9; 
2009: Percentage of: Women: 10.

Transportation and Warehousing; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 5; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 5; 
2009: Percentage of: Women: 2.

Information; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 2; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 2; 
2009: Percentage of: Women: 2.

Finance and Insurance; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 3; 
2006: Percentage of: Women: 6; 
2009: Percentage of: Men: 3; 
2009: Percentage of: Women: 6.

Real Estate and Rental and Leasing; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 1.

Professional, Scientific, and Technical; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 5; 
2006: Percentage of: Women: 5; 
2009: Percentage of: Men: 5; 
2009: Percentage of: Women: 5.

Management, Administrative and Support; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 4; 
2006: Percentage of: Women: 3; 
2009: Percentage of: Men: 4; 
2009: Percentage of: Women: 3.

Educational Services; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 5; 
2006: Percentage of: Women: 14; 
2009: Percentage of: Men: 6; 
2009: Percentage of: Women: 14.

Health Care and Social Assistance; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 4; 
2006: Percentage of: Women: 19; 
2009: Percentage of: Men: 4; 
2009: Percentage of: Women: 21.

Arts, Entertainment, and Recreation; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 2; 
2009: Percentage of: Men: 2; 
2009: Percentage of: Women: 2.

Accommodations and Food Services; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 5; 
2006: Percentage of: Women: 7; 
2009: Percentage of: Men: 6; 
2009: Percentage of: Women: 8.

Other Services (Except Public Administration); 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 3; 
2006: Percentage of: Women: 4; 
2009: Percentage of: Men: 3; 
2009: Percentage of: Women: 4.

Public Administration; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 6; 
2006: Percentage of: Women: 5; 
2009: Percentage of: Men: 5; 
2009: Percentage of: Women: 5.

Active duty; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 1; 
2006: Percentage of: Women: 0; 
2009: Percentage of: Men: 1; 
2009: Percentage of: Women: 0.

Not in universe[B]; 
1999: Percentage of: Men: NA; 
1999: Percentage of: Women: NA; 
2003: Percentage of: Men: NA; 
2003: Percentage of: Women: NA; 
2006: Percentage of: Men: 12; 
2006: Percentage of: Women: 7; 
2009: Percentage of: Men: 13; 
2009: Percentage of: Women: 8.

Household income bracket: 

Less than $20,000; 
1999: Percentage of: Men: 7; 
1999: Percentage of: Women: 8; 
2003: Percentage of: Men: 6; 
2003: Percentage of: Women: 8; 
2006: Percentage of: Men: 7; 
2006: Percentage of: Women: 8; 
2009: Percentage of: Men: 8; 
2009: Percentage of: Women: 9.

$20,000-$40,000; 
1999: Percentage of: Men: 17; 
1999: Percentage of: Women: 18; 
2003: Percentage of: Men: 16; 
2003: Percentage of: Women: 18; 
2006: Percentage of: Men: 16; 
2006: Percentage of: Women: 17; 
2009: Percentage of: Men: 16; 
2009: Percentage of: Women: 16.

$40,000-$60,000; 
1999: Percentage of: Men: 20; 
1999: Percentage of: Women: 20; 
2003: Percentage of: Men: 19; 
2003: Percentage of: Women: 19; 
2006: Percentage of: Men: 19; 
2006: Percentage of: Women: 19; 
2009: Percentage of: Men: 18; 
2009: Percentage of: Women: 19.

$60,000-$80,000; 
1999: Percentage of: Men: 18; 
1999: Percentage of: Women: 17; 
2003: Percentage of: Men: 17; 
2003: Percentage of: Women: 17; 
2006: Percentage of: Men: 16; 
2006: Percentage of: Women: 16; 
2009: Percentage of: Men: 16; 
2009: Percentage of: Women: 15.

Greater than $80,000; 
1999: Percentage of: Men: 38; 
1999: Percentage of: Women: 36; 
2003: Percentage of: Men: 41; 
2003: Percentage of: Women: 39; 
2006: Percentage of: Men: 42; 
2006: Percentage of: Women: 41; 
2009: Percentage of: Men: 41; 
2009: Percentage of: Women: 40.

Firm size: 

Under 25 employees; 
1999: Percentage of: Men: 18; 
1999: Percentage of: Women: 19; 
2003: Percentage of: Men: 18; 
2003: Percentage of: Women: 19; 
2006: Percentage of: Men: 19; 
2006: Percentage of: Women: 19; 
2009: Percentage of: Men: 18; 
2009: Percentage of: Women: 18.

25 to 100 employees; 
1999: Percentage of: Men: 12; 
1999: Percentage of: Women: 11; 
2003: Percentage of: Men: 12; 
2003: Percentage of: Women: 11; 
2006: Percentage of: Men: 11; 
2006: Percentage of: Women: 11; 
2009: Percentage of: Men: 11; 
2009: Percentage of: Women: 11.

100+ employees; 
1999: Percentage of: Men: 58; 
1999: Percentage of: Women: 62; 
2003: Percentage of: Men: 58; 
2003: Percentage of: Women: 62; 
2006: Percentage of: Men: 57; 
2006: Percentage of: Women: 63; 
2009: Percentage of: Men: 58; 
2009: Percentage of: Women: 63.

Not in universe[B]; 
1999: Percentage of: Men: 13; 
1999: Percentage of: Women: 7; 
2003: Percentage of: Men: 13; 
2003: Percentage of: Women: 8; 
2006: Percentage of: Men: 12; 
2006: Percentage of: Women: 8; 
2009: Percentage of: Men: 13; 
2009: Percentage of: Women: 8.

Source: GAO analysis of SIPP data.

Note: The categories for occupation and industry changed between the 
2001 and 2003 SIPP panels. We present the categories for the two most 
recent panels.

[A] Part-time status is defined as working 35 hours or less per week 
during the reference period.

[B] The category "Not in universe" includes self-employed individuals. 

[End of table] 

Factors Associated with Working for an Employer That Offers a Plan:

Table 14 shows the results of two models that analyze factors 
associated with the probability of working for an employer that offers 
(1) any type of pension plan (DB or DC) or (2) a DC plan. The first 
column presents the variables that were included in each model. The 
third and fifth columns present odds ratios that are estimated for 
each variable in the model.5[Footnote 63] The interpretation of the 
odds ratio for a particular variable depends on whether the variable 
has only two or more than two categories.6[Footnote 64]For dichotomous 
(or dummy) variables, odds ratios that are statistically significant 
and greater than 1.00 indicate that individuals with that 
characteristic are more likely to work for an employer that offers a 
plan. For example, an odds ratio of 1.25 for women would mean that 
women are 1.25 times more likely to work for an employer that offers a 
plan. Odds ratios that are significantly lower than 1.00 indicate that 
individuals with that characteristic are less likely to work for an 
employer that offers a plan. For categorical variables with more than 
two categories, a statistically significant odds ratio that is 
greater/less than 1.00 indicates that individuals in that category are 
more/less likely to work for an employer that offers a plan than 
individuals in the category that is chosen as the referent or 
comparison category.

As shown in the body of the report, before controlling for differences 
between men and women in demographic and occupational characteristics, 
a greater proportion of women worked for employers that offered plans 
in 2009. Interestingly, table 14 shows that after accounting for 
demographic and occupational characteristics, women have slightly 
lower odds of working for an employer that offers a DC plan than men. 
In fact, the positive gender effect for women is eliminated when we 
control for occupational characteristics using a statistical model 
(results not shown below). In other words, women's higher likelihood 
of working for an employer that offers a plan is largely due to the 
types of occupations and industries in which women work. (The odds 
ratios for the specific occupations and industries, which are too 
numerous to discuss here, are listed in the table.):

We found that several other factors are associated with the likelihood 
of working for an employer that offers a plan. While the details are 
shown in the table, the factors that were positively associated with 
working for an employer that offers either a DB or DC plan (and that 
were statistically significant at the 95 percent confidence level) 
included age; being divorced (relative to married); education level; 
U.S. citizenship; working in the government or nonprofit sector (in 
comparison to the private sector); having 5 to 9 years of work 
experience (in comparison to having less than 5 years); union 
membership; job tenure; and firm size.

Factors that were negatively associated with working for an employer 
that offers a plan included being never married (in comparison to 
being married); being a single parent; being Black, Hispanic, or Asian 
(in comparison to White, non-Hispanics); being a naturalized 
immigrant; working part time; and being self-employed. While the 
results across both models were generally consistent, some results 
were significant in one model but not the other.

Table 14: Factors Associated with Working for an Employer That Offers 
a Plan, 2009:

Explanatory variables: 

Dependent variable: Gender (omitted category is men); 
Unadjusted proportion with employer that offers a DB or DC plan: 58%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 46%; 
Employer offers a DC plan: [Empty].

Dependent variable: Women; 
Unadjusted proportion with employer that offers a DB or DC plan: 61%; 
Employer offers a DB or DC plan: 0.948; 
Unadjusted proportion with employer that offers a DC plan: 49%; 
Employer offers a DC plan: 0.938**.

Dependent variable: Age groups (omitted category age 18-24); 
Unadjusted proportion with employer that offers a DB or DC plan: 42%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 33%; 
Employer offers a DC plan: [Empty].

Dependent variable: 25-34; 
Unadjusted proportion with employer that offers a DB or DC plan: 60%; 
Employer offers a DB or DC plan: 1.494***; 
Unadjusted proportion with employer that offers a DC plan: 49%; 
Employer offers a DC plan: 1.615***.

Dependent variable: 35-44; 
Unadjusted proportion with employer that offers a DB or DC plan: 62%; 
Employer offers a DB or DC plan: 1.499***; 
Unadjusted proportion with employer that offers a DC plan: 50%; 
Employer offers a DC plan: 1.608***.

Dependent variable: 45-54; 
Unadjusted proportion with employer that offers a DB or DC plan: 64%; 
Employer offers a DB or DC plan: 1.518***; 
Unadjusted proportion with employer that offers a DC plan: 51%; 
Employer offers a DC plan: 1.620***.

Dependent variable: 55-64; 
Unadjusted proportion with employer that offers a DB or DC plan: 63%; 
Employer offers a DB or DC plan: 1.229***; 
Unadjusted proportion with employer that offers a DC plan: 48%; 
Employer offers a DC plan: 1.300***.

Dependent variable: Marital status (omitted category married); 
Unadjusted proportion with employer that offers a DB or DC plan: 63%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 50%; 
Employer offers a DC plan: [Empty].

Dependent variable: Widowed; 
Unadjusted proportion with employer that offers a DB or DC plan: 59%; 
Employer offers a DB or DC plan: 1.059; 
Unadjusted proportion with employer that offers a DC plan: 46%; 
Employer offers a DC plan: 1.083.

Dependent variable: Divorced; 
Unadjusted proportion with employer that offers a DB or DC plan: 64%; 
Employer offers a DB or DC plan: 1.135**; 
Unadjusted proportion with employer that offers a DC plan: 51%; 
Employer offers a DC plan: 1.097**.

Dependent variable: Separated; 
Unadjusted proportion with employer that offers a DB or DC plan: 53%; 
Employer offers a DB or DC plan: 1.008; 
Unadjusted proportion with employer that offers a DC plan: 42%; 
Employer offers a DC plan: 1.000.

Dependent variable: Never married; 
Unadjusted proportion with employer that offers a DB or DC plan: 52%; 
Employer offers a DB or DC plan: 0.906**; 
Unadjusted proportion with employer that offers a DC plan: 41%; 
Employer offers a DC plan: 0.965.

Dependent variable: Children in the household; 
Unadjusted proportion with employer that offers a DB or DC plan: 59%; 
Employer offers a DB or DC plan: 1.101**; 
Unadjusted proportion with employer that offers a DC plan: 47%; 
Employer offers a DC plan: 1.027.

Dependent variable: Single parent; 
Unadjusted proportion with employer that offers a DB or DC plan: 49%; 
Employer offers a DB or DC plan: 0.793***; 
Unadjusted proportion with employer that offers a DC plan: 39%; 
Employer offers a DC plan: 0.869**.

Dependent variable: Race and ethnicity (omitted category White); 
Unadjusted proportion with employer that offers a DB or DC plan: 63%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 51%; 
Employer offers a DC plan: [Empty].

Dependent variable: Black, Non-Hispanic; 
Unadjusted proportion with employer that offers a DB or DC plan: 61%; 
Employer offers a DB or DC plan: 0.750***; 
Unadjusted proportion with employer that offers a DC plan: 46%; 
Employer offers a DC plan: 0.758***.

Dependent variable: Hispanic; 
Unadjusted proportion with employer that offers a DB or DC plan: 43%; 
Employer offers a DB or DC plan: 0.605***; 
Unadjusted proportion with employer that offers a DC plan: 32%; 
Employer offers a DC plan: 0.663***.

Dependent variable: Asian, Non-Hispanic; 
Unadjusted proportion with employer that offers a DB or DC plan: 56%; 
Employer offers a DB or DC plan: 0.761***; 
Unadjusted proportion with employer that offers a DC plan: 46%; 
Employer offers a DC plan: 0.882.

Dependent variable: Other, Non-Hispanic; 
Unadjusted proportion with employer that offers a DB or DC plan: 59%; 
Employer offers a DB or DC plan: 0.855*; 
Unadjusted proportion with employer that offers a DC plan: 47%; 
Employer offers a DC plan: 0.897.

Dependent variable: Education level (omitted category No high school 
diploma); 
Unadjusted proportion with employer that offers a DB or DC plan: 30%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 24%; 
Employer offers a DC plan: [Empty].

Dependent variable: High school diploma; 
Unadjusted proportion with employer that offers a DB or DC plan: 51%; 
Employer offers a DB or DC plan: 1.297***; 
Unadjusted proportion with employer that offers a DC plan: 39%; 
Employer offers a DC plan: 1.196***.

Dependent variable: Some college; 
Unadjusted proportion with employer that offers a DB or DC plan: 61%; 
Employer offers a DB or DC plan: 1.772***; 
Unadjusted proportion with employer that offers a DC plan: 48%; 
Employer offers a DC plan: 1.543***.

Dependent variable: Bachelor's degree or higher; 
Unadjusted proportion with employer that offers a DB or DC plan: 72%; 
Employer offers a DB or DC plan: 1.997***; 
Unadjusted proportion with employer that offers a DC plan: 57%; 
Employer offers a DC plan: 1.606***.

Dependent variable: Citizen; 
Unadjusted proportion with employer that offers a DB or DC plan: 62%; 
Employer offers a DB or DC plan: 1.577***; 
Unadjusted proportion with employer that offers a DC plan: 49%; 
Employer offers a DC plan: 1.499***.

Dependent variable: Naturalized immigrant; 
Unadjusted proportion with employer that offers a DB or DC plan: 54%; 
Employer offers a DB or DC plan: 0.737***; 
Unadjusted proportion with employer that offers a DC plan: 42%; 
Employer offers a DC plan: 0.787***.

Dependent variable: Part-time status (omitted category is full 
time)[A]; 
Unadjusted proportion with employer that offers a DB or DC plan: 66%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 52%; 
Employer offers a DC plan: [Empty].

Dependent variable: Part-time; 
Unadjusted proportion with employer that offers a DB or DC plan: 46%; 
Employer offers a DB or DC plan: 0.763***; 
Unadjusted proportion with employer that offers a DC plan: 38%; 
Employer offers a DC plan: 0.925***.

Dependent variable: Sector (omitted category private sector); 
Unadjusted proportion with employer that offers a DB or DC plan: 60%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 50%; 
Employer offers a DC plan: [Empty].

Dependent variable: Private not for profit; 
Unadjusted proportion with employer that offers a DB or DC plan: 73%; 
Employer offers a DB or DC plan: 1.430***; 
Unadjusted proportion with employer that offers a DC plan: 59%; 
Employer offers a DC plan: 1.243***.

Dependent variable: Government worker; 
Unadjusted proportion with employer that offers a DB or DC plan: 88%; 
Employer offers a DB or DC plan: 2.142***; 
Unadjusted proportion with employer that offers a DC plan: 61%; 
Employer offers a DC plan: 1.062.

Dependent variable: Occupation (omitted category Management); 
Unadjusted proportion with employer that offers a DB or DC plan: 76%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 64%; 
Employer offers a DC plan: [Empty].

Dependent variable: Business and Financial Operations; 
Unadjusted proportion with employer that offers a DB or DC plan: 82%; 
Employer offers a DB or DC plan: 1.133; 
Unadjusted proportion with employer that offers a DC plan: 70%; 
Employer offers a DC plan: 1.053.

Dependent variable: Computer and Mathematical; 
Unadjusted proportion with employer that offers a DB or DC plan: 85%; 
Employer offers a DB or DC plan: 1.222*; 
Unadjusted proportion with employer that offers a DC plan: 73%; 
Employer offers a DC plan: 1.036.

Dependent variable: Architecture and Engineering; 
Unadjusted proportion with employer that offers a DB or DC plan: 88%; 
Employer offers a DB or DC plan: 1.737***; 
Unadjusted proportion with employer that offers a DC plan: 73%; 
Employer offers a DC plan: 1.196.

Dependent variable: Life, Physical, and Social Services; 
Unadjusted proportion with employer that offers a DB or DC plan: 87%; 
Employer offers a DB or DC plan: 1.067; 
Unadjusted proportion with employer that offers a DC plan: 68%; 
Employer offers a DC plan: 0.779*.

Dependent variable: Community and Social Services; 
Unadjusted proportion with employer that offers a DB or DC plan: 74%; 
Employer offers a DB or DC plan: 0.895; 
Unadjusted proportion with employer that offers a DC plan: 54%; 
Employer offers a DC plan: 0.701***.

Dependent variable: Legal; 
Unadjusted proportion with employer that offers a DB or DC plan: 77%; 
Employer offers a DB or DC plan: 1.139; 
Unadjusted proportion with employer that offers a DC plan: 66%; 
Employer offers a DC plan: 1.190.

Dependent variable: Education, Training, and Library; 
Unadjusted proportion with employer that offers a DB or DC plan: 81%; 
Employer offers a DB or DC plan: 0.605***; 
Unadjusted proportion with employer that offers a DC plan: 57%; 
Employer offers a DC plan: 0.600***.

Dependent variable: Arts, Design, Entertainment, Sports, and Media; 
Unadjusted proportion with employer that offers a DB or DC plan: 64%; 
Employer offers a DB or DC plan: 0.729**; 
Unadjusted proportion with employer that offers a DC plan: 52%; 
Employer offers a DC plan: 0.745**.

Dependent variable: Healthcare Practitioners and Technical; 
Unadjusted proportion with employer that offers a DB or DC plan: 82%; 
Employer offers a DB or DC plan: 1.314***; 
Unadjusted proportion with employer that offers a DC plan: 68%; 
Employer offers a DC plan: 1.003.

Dependent variable: Healthcare Support; 
Unadjusted proportion with employer that offers a DB or DC plan: 57%; 
Employer offers a DB or DC plan: 0.635***; 
Unadjusted proportion with employer that offers a DC plan: 46%; 
Employer offers a DC plan: 0.617***.

Dependent variable: Protective Service; 
Unadjusted proportion with employer that offers a DB or DC plan: 77%; 
Employer offers a DB or DC plan: 0.684***; 
Unadjusted proportion with employer that offers a DC plan: 54%; 
Employer offers a DC plan: 0.631***.

Dependent variable: Food Preparation and Serving Related; 
Unadjusted proportion with employer that offers a DB or DC plan: 34%; 
Employer offers a DB or DC plan: 0.530***; 
Unadjusted proportion with employer that offers a DC plan: 27%; 
Employer offers a DC plan: 0.524***.

Dependent variable: Building and Grounds Cleaning and Maintenance; 
Unadjusted proportion with employer that offers a DB or DC plan: 42%; 
Employer offers a DB or DC plan: 0.619***; 
Unadjusted proportion with employer that offers a DC plan: 30%; 
Employer offers a DC plan: 0.545***.

Dependent variable: Personal Care and Service; 
Unadjusted proportion with employer that offers a DB or DC plan: 33%; 
Employer offers a DB or DC plan: 0.326***; 
Unadjusted proportion with employer that offers a DC plan: 25%; 
Employer offers a DC plan: 0.363***.

Dependent variable: Sales and Related; 
Unadjusted proportion with employer that offers a DB or DC plan: 60%; 
Employer offers a DB or DC plan: 0.634***; 
Unadjusted proportion with employer that offers a DC plan: 49%; 
Employer offers a DC plan: 0.620***.

Dependent variable: Office and Administrative Support; 
Unadjusted proportion with employer that offers a DB or DC plan: 69%; 
Employer offers a DB or DC plan: 0.864*; 
Unadjusted proportion with employer that offers a DC plan: 55%; 
Employer offers a DC plan: 0.785***.

Dependent variable: Farming, Forestry, and Fishing; 
Unadjusted proportion with employer that offers a DB or DC plan: 16%; 
Employer offers a DB or DC plan: 0.265***; 
Unadjusted proportion with employer that offers a DC plan: 12%; 
Employer offers a DC plan: 0.290***.

Dependent variable: Construction and Extraction; 
Unadjusted proportion with employer that offers a DB or DC plan: 44%; 
Employer offers a DB or DC plan: 0.690***; 
Unadjusted proportion with employer that offers a DC plan: 31%; 
Employer offers a DC plan: 0.595***.

Dependent variable: Installation, Repair, and Maintenance; 
Unadjusted proportion with employer that offers a DB or DC plan: 66%; 
Employer offers a DB or DC plan: 0.856; 
Unadjusted proportion with employer that offers a DC plan: 56%; 
Employer offers a DC plan: 0.929.

Dependent variable: Production; 
Unadjusted proportion with employer that offers a DB or DC plan: 66%; 
Employer offers a DB or DC plan: 0.607***; 
Unadjusted proportion with employer that offers a DC plan: 53%; 
Employer offers a DC plan: 0.628***.

Dependent variable: Transportation and Material Moving; 
Unadjusted proportion with employer that offers a DB or DC plan: 61%; 
Employer offers a DB or DC plan: 0.673***; 
Unadjusted proportion with employer that offers a DC plan: 48%; 
Employer offers a DC plan: 0.627***.

Dependent variable: Not in universe[B]; 
Unadjusted proportion with employer that offers a DB or DC plan: 15%; 
Employer offers a DB or DC plan: 0.894; 
Unadjusted proportion with employer that offers a DC plan: 10%; 
Employer offers a DC plan: 0.411***.

Dependent variable: Industry (omitted category Agriculture); 
Unadjusted proportion with employer that offers a DB or DC plan: 19%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 16%; 
Employer offers a DC plan: [Empty].

Dependent variable: Mining; 
Unadjusted proportion with employer that offers a DB or DC plan: 69%; 
Employer offers a DB or DC plan: 1.515; 
Unadjusted proportion with employer that offers a DC plan: 54%; 
Employer offers a DC plan: 1.235.

Dependent variable: Utilities; 
Unadjusted proportion with employer that offers a DB or DC plan: 89%; 
Employer offers a DB or DC plan: 3.134***; 
Unadjusted proportion with employer that offers a DC plan: 69%; 
Employer offers a DC plan: 1.785**.

Dependent variable: Construction; 
Unadjusted proportion with employer that offers a DB or DC plan: 45%; 
Employer offers a DB or DC plan: 1.379; 
Unadjusted proportion with employer that offers a DC plan: 33%; 
Employer offers a DC plan: 1.145.

Dependent variable: Manufacturing; 
Unadjusted proportion with employer that offers a DB or DC plan: 76%; 
Employer offers a DB or DC plan: 2.697***; 
Unadjusted proportion with employer that offers a DC plan: 63%; 
Employer offers a DC plan: 2.036***.

Dependent variable: Wholesale Trade; 
Unadjusted proportion with employer that offers a DB or DC plan: 69%; 
Employer offers a DB or DC plan: 2.725***; 
Unadjusted proportion with employer that offers a DC plan: 55%; 
Employer offers a DC plan: 1.923***.

Dependent variable: Retail Trade; 
Unadjusted proportion with employer that offers a DB or DC plan: 62%; 
Employer offers a DB or DC plan: 2.052***; 
Unadjusted proportion with employer that offers a DC plan: 51%; 
Employer offers a DC plan: 1.668**.

Dependent variable: Transportation and Warehousing; 
Unadjusted proportion with employer that offers a DB or DC plan: 71%; 
Employer offers a DB or DC plan: 1.763**; 
Unadjusted proportion with employer that offers a DC plan: 55%; 
Employer offers a DC plan: 1.525*.

Dependent variable: Information; 
Unadjusted proportion with employer that offers a DB or DC plan: 76%; 
Employer offers a DB or DC plan: 2.228***; 
Unadjusted proportion with employer that offers a DC plan: 63%; 
Employer offers a DC plan: 1.761**.

Dependent variable: Finance and Insurance; 
Unadjusted proportion with employer that offers a DB or DC plan: 84%; 
Employer offers a DB or DC plan: 3.571***; 
Unadjusted proportion with employer that offers a DC plan: 73%; 
Employer offers a DC plan: 2.614***.

Dependent variable: Real Estate and Rental and Leasing; 
Unadjusted proportion with employer that offers a DB or DC plan: 49%; 
Employer offers a DB or DC plan: 1.301; 
Unadjusted proportion with employer that offers a DC plan: 41%; 
Employer offers a DC plan: 1.255.

Dependent variable: Professional, Scientific, and Technical; 
Unadjusted proportion with employer that offers a DB or DC plan: 71%; 
Employer offers a DB or DC plan: 2.197***; 
Unadjusted proportion with employer that offers a DC plan: 61%; 
Employer offers a DC plan: 1.884***.

Dependent variable: Management, Administrative and Support; 
Unadjusted proportion with employer that offers a DB or DC plan: 42%; 
Employer offers a DB or DC plan: 1.059; 
Unadjusted proportion with employer that offers a DC plan: 34%; 
Employer offers a DC plan: 1.013.

Dependent variable: Educational Services; 
Unadjusted proportion with employer that offers a DB or DC plan: 84%; 
Employer offers a DB or DC plan: 2.120***; 
Unadjusted proportion with employer that offers a DC plan: 60%; 
Employer offers a DC plan: 1.461.

Dependent variable: Health Care and Social Assistance; 
Unadjusted proportion with employer that offers a DB or DC plan: 67%; 
Employer offers a DB or DC plan: 1.733**; 
Unadjusted proportion with employer that offers a DC plan: 55%; 
Employer offers a DC plan: 1.559*.

Dependent variable: Arts, Entertainment, and Recreation; 
Unadjusted proportion with employer that offers a DB or DC plan: 51%; 
Employer offers a DB or DC plan: 1.453; 
Unadjusted proportion with employer that offers a DC plan: 40%; 
Employer offers a DC plan: 1.268.

Dependent variable: Accommodations and Food Services; 
Unadjusted proportion with employer that offers a DB or DC plan: 34%; 
Employer offers a DB or DC plan: 0.996; 
Unadjusted proportion with employer that offers a DC plan: 28%; 
Employer offers a DC plan: 0.905.

Dependent variable: Other Services (Except Public Administration); 
Unadjusted proportion with employer that offers a DB or DC plan: 38%; 
Employer offers a DB or DC plan: 1.144; 
Unadjusted proportion with employer that offers a DC plan: 30%; 
Employer offers a DC plan: 0.972.

Dependent variable: Public Administration; 
Unadjusted proportion with employer that offers a DB or DC plan: 89%; 
Employer offers a DB or DC plan: 2.198***; 
Unadjusted proportion with employer that offers a DC plan: 64%; 
Employer offers a DC plan: 1.499.

Dependent variable: Work experience (omitted category Less than 5 
years); 
Unadjusted proportion with employer that offers a DB or DC plan: 56%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 45%; 
Employer offers a DC plan: [Empty].

Dependent variable: 5 to 9 years; 
Unadjusted proportion with employer that offers a DB or DC plan: 67%; 
Employer offers a DB or DC plan: 1.140***; 
Unadjusted proportion with employer that offers a DC plan: 53%; 
Employer offers a DC plan: 1.069*.

Dependent variable: 10 to 15 years; 
Unadjusted proportion with employer that offers a DB or DC plan: 71%; 
Employer offers a DB or DC plan: 1.033; 
Unadjusted proportion with employer that offers a DC plan: 56%; 
Employer offers a DC plan: 1.013.

Dependent variable: More than 15 years; 
Unadjusted proportion with employer that offers a DB or DC plan: 57%; 
Employer offers a DB or DC plan: 0.978; 
Unadjusted proportion with employer that offers a DC plan: 45%; 
Employer offers a DC plan: 0.989.

Dependent variable: Union status (omitted category not in a union); 
Unadjusted proportion with employer that offers a DB or DC plan: 63%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 51%; 
Employer offers a DC plan: [Empty].

Dependent variable: In a union; 
Unadjusted proportion with employer that offers a DB or DC plan: 87%; 
Employer offers a DB or DC plan: 1.903***; 
Unadjusted proportion with employer that offers a DC plan: 62%; 
Employer offers a DC plan: 1.094**.

Dependent variable: Self-employment status; 
Unadjusted proportion with employer that offers a DB or DC plan: 19%; 
Employer offers a DB or DC plan: 0.525***; 
Unadjusted proportion with employer that offers a DC plan: 14%; 
Employer offers a DC plan: 0.671***.

Dependent variable: Number of employees at current place of employment 
(omitted category Under 25 employees)); 
Unadjusted proportion with employer that offers a DB or DC plan: 25%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 19%; 
Employer offers a DC plan: [Empty].

Dependent variable: 25 to 100 employees; 
Unadjusted proportion with employer that offers a DB or DC plan: 57%; 
Employer offers a DB or DC plan: 3.291***; 
Unadjusted proportion with employer that offers a DC plan: 46%; 
Employer offers a DC plan: 3.021***.

Dependent variable: 100+ employees; 
Unadjusted proportion with employer that offers a DB or DC plan: 79%; 
Employer offers a DB or DC plan: 7.618***; 
Unadjusted proportion with employer that offers a DC plan: 63%; 
Employer offers a DC plan: 5.528***.

Dependent variable: Years of tenure at current job; 
Unadjusted proportion with employer that offers a DB or DC plan: 
[Empty]; 
Employer offers a DB or DC plan: 1.042***; 
Unadjusted proportion with employer that offers a DC plan: 9%; 
Employer offers a DC plan: 1.015***.

Tenure categories: 

Dependent variable: Less than 5 years; 
Unadjusted proportion with employer that offers a DB or DC plan: 56%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 45%; 
Employer offers a DC plan: [Empty].

Dependent variable: 5 to 9 years; 
Unadjusted proportion with employer that offers a DB or DC plan: 69%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 54%; 
Employer offers a DC plan: [Empty].

Dependent variable: 10 to 15 years; 
Unadjusted proportion with employer that offers a DB or DC plan: 76%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 60%; 
Employer offers a DC plan: [Empty].

Dependent variable: More than 15 years; 
Unadjusted proportion with employer that offers a DB or DC plan: 82%; 
Employer offers a DB or DC plan: [Empty]; 
Unadjusted proportion with employer that offers a DC plan: 63%; 
Employer offers a DC plan: [Empty].

Dependent variable: Number of observations; 
Employer offers a DB or DC plan: 37,038;  
Employer offers a DC plan: 37,038.

Source: GAO analysis of SIPP data.

* Indicates that the variable is statistically significant at the 90 
percent level.

** Indicates that the variable is statistically significant at the 95 
percent level.

*** Indicates that the variable is statistically significant at the 99 
percent level.

[A] Part-time status is defined as working 35 hours or less per week 
during the reference period.

[B] The category "Not in universe" includes self-employed individuals. 

[End of table] 

Factors Associated with Eligibility for Employer-Sponsored Pension Plan:

Table 15 shows the results of a model we estimated to analyze factors 
associated with whether an individual is eligible for their employer's 
plan. It is presented in the same format as table 14. As shown in the 
body of the report, women had lower rates of plan eligibility across 
all 4 study years. The results of the model show that, even after 
controlling for demographic and occupational differences between men 
and women, women had significantly lower rates of eligibility in 2009. 
Perhaps most interesting is the odds ratio for part-time status, which 
indicates that part-time workers are approximately one-third as likely 
to be eligible for their employer's plan as full-time workers. In 
addition, work experience and tenure are also significantly and 
positively related with eligibility. Union status is also positively 
associated with plan eligibility.

Table 15: Factors Associated with Eligibility for Employer-Sponsored 
Pension Plan, 2009:

Dependent variable: Gender (omitted category is men); 
Unadjusted proportion eligible for a DB or DC plan: 91%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Women; 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: 0.861**.

Dependent variable: Age groups (omitted category age 18-24); 
Unadjusted proportion eligible for a DB or DC plan: 55%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: 25-34; 
Unadjusted proportion eligible for a DB or DC plan: 88%; 
Individual is eligible for a DB or DC plan: 2.589***.

Dependent variable: 35-44; 
Unadjusted proportion eligible for a DB or DC plan: 93%; 
Individual is eligible for a DB or DC plan: 2.957***.

Dependent variable: 45-54; 
Unadjusted proportion eligible for a DB or DC plan: 94%; 
Individual is eligible for a DB or DC plan: 2.846***.

Dependent variable: 55-64; 
Unadjusted proportion eligible for a DB or DC plan: 93%; 
Individual is eligible for a DB or DC plan: 2.106***.

Dependent variable: Marital status (omitted category married); 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Widowed; 
Unadjusted proportion eligible for a DB or DC plan: 88%; 
Individual is eligible for a DB or DC plan: 0.637**.

Dependent variable: Divorced; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 1.021.

Dependent variable: Separated; 
Unadjusted proportion eligible for a DB or DC plan: 89%; 
Individual is eligible for a DB or DC plan: 0.998.

Dependent variable: Never married; 
Unadjusted proportion eligible for a DB or DC plan: 77%; 
Individual is eligible for a DB or DC plan: 0.795***.

Dependent variable: Children in the household; 
Unadjusted proportion eligible for a DB or DC plan: 89%; 
Individual is eligible for a DB or DC plan: 0.906.

Dependent variable: Single parent; 
Unadjusted proportion eligible for a DB or DC plan: 77%; 
Individual is eligible for a DB or DC plan: 0.963.

Dependent variable: Race and ethnicity (omitted category White); 
Unadjusted proportion eligible for a DB or DC plan: 89%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Black, Non-Hispanic; 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: 0.998.

Dependent variable: Hispanic; 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: 1.028.

Dependent variable: Asian, Non-Hispanic; 
Unadjusted proportion eligible for a DB or DC plan: 90%; 
Individual is eligible for a DB or DC plan: 1.011.

Dependent variable: Other, Non-Hispanic; 
Unadjusted proportion eligible for a DB or DC plan: 86%; 
Individual is eligible for a DB or DC plan: 1.023.

Dependent variable: Education level (omitted category No high school 
diploma); 
Unadjusted proportion eligible for a DB or DC plan: 82%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: High school diploma; 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: 0.881.

Dependent variable: Some college; 
Unadjusted proportion eligible for a DB or DC plan: 86%; 
Individual is eligible for a DB or DC plan: 0.872.

Dependent variable: Bachelor's degree or higher; 
Unadjusted proportion eligible for a DB or DC plan: 93%; 
Individual is eligible for a DB or DC plan: 1.128.

Dependent variable: Citizen; 
Unadjusted proportion eligible for a DB or DC plan: 89%; 
Individual is eligible for a DB or DC plan: 1.235.

Dependent variable: Naturalized immigrant; 
Unadjusted proportion eligible for a DB or DC plan: 91%; 
Individual is eligible for a DB or DC plan: 0.886.

Dependent variable: Part-time status (omitted category is full 
time)[A]; 
Unadjusted proportion eligible for a DB or DC plan: 94%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Part-time; 
Unadjusted proportion eligible for a DB or DC plan: 73%; 
Individual is eligible for a DB or DC plan: 0.315***.

Dependent variable: Sector (omitted category private sector); 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Private not for profit; 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: 0.867.

Dependent variable: Government worker; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 0.996.

Dependent variable: Occupation (omitted category Management); 
Unadjusted proportion eligible for a DB or DC plan: 96%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Business and Financial Operations; 
Unadjusted proportion eligible for a DB or DC plan: 95%; 
Individual is eligible for a DB or DC plan: 0.958.

Dependent variable: Computer and Mathematical; 
Unadjusted proportion eligible for a DB or DC plan: 95%; 
Individual is eligible for a DB or DC plan: 0.941.

Dependent variable: Architecture and Engineering; 
Unadjusted proportion eligible for a DB or DC plan: 96%; 
Individual is eligible for a DB or DC plan: 1.088.

Dependent variable: Life, Physical, and Social Services; 
Unadjusted proportion eligible for a DB or DC plan: 95%; 
Individual is eligible for a DB or DC plan: 0.924.

Dependent variable: Community and Social Services; 
Unadjusted proportion eligible for a DB or DC plan: 90%; 
Individual is eligible for a DB or DC plan: 0.579**.

Dependent variable: Legal; 
Unadjusted proportion eligible for a DB or DC plan: 94%; 
Individual is eligible for a DB or DC plan: 0.823.

Dependent variable: Education, Training, and Library; 
Unadjusted proportion eligible for a DB or DC plan: 88%; 
Individual is eligible for a DB or DC plan: 0.492***.

Dependent variable: Arts, Design, Entertainment, Sports, and Media; 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: 0.535**.

Dependent variable: Healthcare Practitioners and Technical; 
Unadjusted proportion eligible for a DB or DC plan: 90%; 
Individual is eligible for a DB or DC plan: 0.595***.

Dependent variable: Healthcare Support; 
Unadjusted proportion eligible for a DB or DC plan: 79%; 
Individual is eligible for a DB or DC plan: 0.346***.

Dependent variable: Protective Service; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 0.533***.

Dependent variable: Food Preparation and Serving Related; 
Unadjusted proportion eligible for a DB or DC plan: 65%; 
Individual is eligible for a DB or DC plan: 0.395***.

Dependent variable: Building and Grounds Cleaning and Maintenance; 
Unadjusted proportion eligible for a DB or DC plan: 83%; 
Individual is eligible for a DB or DC plan: 0.396***.

Dependent variable: Personal Care and Service; 
Unadjusted proportion eligible for a DB or DC plan: 69%; 
Individual is eligible for a DB or DC plan: 0.299***.

Dependent variable: Sales and Related; 
Unadjusted proportion eligible for a DB or DC plan: 81%; 
Individual is eligible for a DB or DC plan: 0.466***.

Dependent variable: Office and Administrative Support; 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: 0.474***.

Dependent variable: Farming, Forestry, and Fishing; 
Unadjusted proportion eligible for a DB or DC plan: 89%; 
Individual is eligible for a DB or DC plan: 0.598.

Dependent variable: Construction and Extraction; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 0.693.

Dependent variable: Installation, Repair, and Maintenance; 
Unadjusted proportion eligible for a DB or DC plan: 93%; 
Individual is eligible for a DB or DC plan: 0.638**.

Dependent variable: Production; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 0.482***.

Dependent variable: Transportation and Material Moving; 
Unadjusted proportion eligible for a DB or DC plan: 85%; 
Individual is eligible for a DB or DC plan: 0.417***.

Dependent variable: Not in universe[B]; 
Unadjusted proportion eligible for a DB or DC plan: 96%; 
Individual is eligible for a DB or DC plan: 1.461.

Dependent variable: Industry (omitted category Agriculture); 
Unadjusted proportion eligible for a DB or DC plan: 90%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Mining; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 0.765.

Dependent variable: Utilities; 
Unadjusted proportion eligible for a DB or DC plan: 98%; 
Individual is eligible for a DB or DC plan: 1.572.

Dependent variable: Construction; 
Unadjusted proportion eligible for a DB or DC plan: 91%; 
Individual is eligible for a DB or DC plan: 0.684.

Dependent variable: Manufacturing; 
Unadjusted proportion eligible for a DB or DC plan: 94%; 
Individual is eligible for a DB or DC plan: 1.055.

Dependent variable: Wholesale Trade; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 1.090.

Dependent variable: Retail Trade; 
Unadjusted proportion eligible for a DB or DC plan: 79%; 
Individual is eligible for a DB or DC plan: 0.682.

Dependent variable: Transportation and Warehousing; 
Unadjusted proportion eligible for a DB or DC plan: 90%; 
Individual is eligible for a DB or DC plan: 0.717.

Dependent variable: Information; 
Unadjusted proportion eligible for a DB or DC plan: 91%; 
Individual is eligible for a DB or DC plan: 0.952.

Dependent variable: Finance and Insurance; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 1.006.

Dependent variable: Real Estate and Rental and Leasing; 
Unadjusted proportion eligible for a DB or DC plan: 89%; 
Individual is eligible for a DB or DC plan: 0.907.

Dependent variable: Professional, Scientific, and Technical; 
Unadjusted proportion eligible for a DB or DC plan: 93%; 
Individual is eligible for a DB or DC plan: 0.854.

Dependent variable: Management, Administrative and Support; 
Unadjusted proportion eligible for a DB or DC plan: 85%; 
Individual is eligible for a DB or DC plan: 0.656.

Dependent variable: Educational Services; 
Unadjusted proportion eligible for a DB or DC plan: 88%; 
Individual is eligible for a DB or DC plan: 0.599.

Dependent variable: Health Care and Social Assistance; 
Unadjusted proportion eligible for a DB or DC plan: 88%; 
Individual is eligible for a DB or DC plan: 0.831.

Dependent variable: Arts, Entertainment, and Recreation; 
Unadjusted proportion eligible for a DB or DC plan: 75%; 
Individual is eligible for a DB or DC plan: 0.462.

Dependent variable: Accommodations and Food Services; 
Unadjusted proportion eligible for a DB or DC plan: 67%; 
Individual is eligible for a DB or DC plan: 0.538.

Dependent variable: Other Services (Except Public Administration); 
Unadjusted proportion eligible for a DB or DC plan: 86%; 
Individual is eligible for a DB or DC plan: 0.738.

Dependent variable: Public Administration; 
Unadjusted proportion eligible for a DB or DC plan: 95%; 
Individual is eligible for a DB or DC plan: 1.076.

Dependent variable: Work experience (omitted category Less than 5 
years); 
Unadjusted proportion eligible for a DB or DC plan: 78%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: 5 to 9 years; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 1.419***.

Dependent variable: 10 to 15 years; 
Unadjusted proportion eligible for a DB or DC plan: 95%; 
Individual is eligible for a DB or DC plan: 1.435***.

Dependent variable: More than 15 years; 
Unadjusted proportion eligible for a DB or DC plan: 92%; 
Individual is eligible for a DB or DC plan: 0.895*.

Dependent variable: Union status (omitted category not in a union); 
Unadjusted proportion eligible for a DB or DC plan: 87%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: In a union; 
Unadjusted proportion eligible for a DB or DC plan: 95%; 
Individual is eligible for a DB or DC plan: 2.070***.

Dependent variable: Self-employment status; 
Unadjusted proportion eligible for a DB or DC plan: 91%; 
Individual is eligible for a DB or DC plan: 0.864.

Dependent variable: Number of employees at current place of employment 
(omitted category Under 25 employees); 
Unadjusted proportion eligible for a DB or DC plan: 85%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: 25 to 100 employees; 
Unadjusted proportion eligible for a DB or DC plan: 88%; 
Individual is eligible for a DB or DC plan: 1.165*.

Dependent variable: 100+ employees; 
Unadjusted proportion eligible for a DB or DC plan: 89%; 
Individual is eligible for a DB or DC plan: 1.300***.

Dependent variable: Years of tenure at current job; 
Unadjusted proportion eligible for a DB or DC plan: [Empty]; 
Individual is eligible for a DB or DC plan: 1.169***.

Tenure categories: 

Dependent variable: Less than 5 years; 
Unadjusted proportion eligible for a DB or DC plan: 78%; 
Individual is eligible for a DB or DC plan: [Empty]. 

Dependent variable: 5 to 9 years; 
Unadjusted proportion eligible for a DB or DC plan: 93%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: 10 to 15 years; 
Unadjusted proportion eligible for a DB or DC plan: 97%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: More than 15 years; 
Unadjusted proportion eligible for a DB or DC plan: 99%; 
Individual is eligible for a DB or DC plan: [Empty].

Dependent variable: Number of observations; 
Individual is eligible for a DB or DC plan: 24,274.

Source: GAO analysis of SIPP data.

* Indicates that the variable is statistically significant at the 90 
percent level.

** Indicates that the variable is statistically significant at the 95 
percent level.

*** Indicates that the variable is statistically significant at the 99 
percent level.

[A] Part-time status is defined as working 35 hours or less per week 
during the reference period.

[B] The category "Not in universe" includes self-employed individuals. 

[End of table] 

Factors Associated with Participation in an Employer-Sponsored Pension 
Plan:

Table 16 shows the results of two models we estimated to analyze 
factors associated with the probability of participating in (1) any 
type of pension plan (DB or DC) or (2) a DC plan. Again, it is 
presented in the same format as tables 14 and 15.

As shown in the body of the report, before controlling for differences 
between men and women in demographic and occupational characteristics, 
a smaller proportion of women participated in an employer-sponsored 
pension plan. Our analysis shows that the gender differences in plan 
participation are largely accounted for by differences between men and 
women in demographic and occupational characteristics.

Similar to our other models, we identify a number of factors that are 
related to plan participation. The factors that were positively 
related to participating in either a DB or a DC (and that are 
statistically significant at the 95 percent level) include age; 
education-level; being Asian (relative to whites); U.S. citizenship; 
working in the nonprofit or government sector (relative to the private 
sector); work- experience; union membership; and tenure. Factors that 
were negatively related to participating in a plan included being a 
single parent; working part-time; and being Black or Hispanic. A 
number of industries and occupations, too numerous to list, were 
statistically significant as shown in the table below.

Table 16: Factors Associated with Participation in an Employer- 
Sponsored Pension Plan, 2009:

Dependent variable: Gender (omitted category is men); 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Women; 
Unadjusted proportion participating in a DB or DC plan: 86%; 
Individual participates in a DB or DC plan: 0.973; 
Unadjusted proportion participating in a DC plan: 78%; 
Individual participates in a DC plan: 1.099*.

Dependent variable: Age groups (omitted category age 18-24); 
Unadjusted proportion participating in a DB or DC plan: 63%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 54%; 
Individual participates in a DC plan: [Empty].

Dependent variable: 25-34; 
Unadjusted proportion participating in a DB or DC plan: 83%; 
Individual participates in a DB or DC plan: 1.547***; 
Unadjusted proportion participating in a DC plan: 75%; 
Individual participates in a DC plan: 1.659***.

Dependent variable: 35-44; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: 1.627***; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: 1.821***.

Dependent variable: 45-54; 
Unadjusted proportion participating in a DB or DC plan: 91%; 
Individual participates in a DB or DC plan: 1.843***; 
Unadjusted proportion participating in a DC plan: 83%; 
Individual participates in a DC plan: 1.924***.

Dependent variable: 55-64; 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: 1.691***; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 1.642***.

Dependent variable: Marital status (omitted category married); 
Unadjusted proportion participating in a DB or DC plan: 90%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Widowed; 
Unadjusted proportion participating in a DB or DC plan: 90%; 
Individual participates in a DB or DC plan: 1.173; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 1.129.

Dependent variable: Divorced; 
Unadjusted proportion participating in a DB or DC plan: 86%; 
Individual participates in a DB or DC plan: 0.867; 
Unadjusted proportion participating in a DC plan: 78%; 
Individual participates in a DC plan: 0.911.

Dependent variable: Separated; 
Unadjusted proportion participating in a DB or DC plan: 79%; 
Individual participates in a DB or DC plan: 0.871; 
Unadjusted proportion participating in a DC plan: 68%; 
Individual participates in a DC plan: 0.784.

Dependent variable: Never married; 
Unadjusted proportion participating in a DB or DC plan: 78%; 
Individual participates in a DB or DC plan: 0.888; 
Unadjusted proportion participating in a DC plan: 71%; 
Individual participates in a DC plan: 0.961.

Dependent variable: Children in the household; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: 1.125; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: 1.138*.

Dependent variable: Single parent; 
Unadjusted proportion participating in a DB or DC plan: 76%; 
Individual participates in a DB or DC plan: 0.805**; 
Unadjusted proportion participating in a DC plan: 68%; 
Individual participates in a DC plan: 0.844*.

Dependent variable: Race and ethnicity (omitted category White); 
Unadjusted proportion participating in a DB or DC plan: 88%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 81%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Black, Non-Hispanic; 
Unadjusted proportion participating in a DB or DC plan: 81%; 
Individual participates in a DB or DC plan: 0.705***; 
Unadjusted proportion participating in a DC plan: 68%; 
Individual participates in a DC plan: 0.579***.

Dependent variable: Hispanic; 
Unadjusted proportion participating in a DB or DC plan: 77%; 
Individual participates in a DB or DC plan: 0.684***; 
Unadjusted proportion participating in a DC plan: 69%; 
Individual participates in a DC plan: 0.737***.

Dependent variable: Asian, Non-Hispanic; 
Unadjusted proportion participating in a DB or DC plan: 90%; 
Individual participates in a DB or DC plan: 1.304*; 
Unadjusted proportion participating in a DC plan: 85%; 
Individual participates in a DC plan: 1.500***.

Dependent variable: Other, Non-Hispanic; 
Unadjusted proportion participating in a DB or DC plan: 86%; 
Individual participates in a DB or DC plan: 1.108; 
Unadjusted proportion participating in a DC plan: 77%; 
Individual participates in a DC plan: 0.950.

Dependent variable: Education level (omitted category No high school 
diploma); 
Unadjusted proportion participating in a DB or DC plan: 69%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 61%; 
Individual participates in a DC plan: [Empty].

Dependent variable: High school diploma; 
Unadjusted proportion participating in a DB or DC plan: 82%; 
Individual participates in a DB or DC plan: 1.275**; 
Unadjusted proportion participating in a DC plan: 72%; 
Individual participates in a DC plan: 1.190.

Dependent variable: Some college; 
Unadjusted proportion participating in a DB or DC plan: 85%; 
Individual participates in a DB or DC plan: 1.617***; 
Unadjusted proportion participating in a DC plan: 77%; 
Individual participates in a DC plan: 1.548***.

Dependent variable: Bachelor's degree or higher; 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: 2.318***; 
Unadjusted proportion participating in a DC plan: 84%; 
Individual participates in a DC plan: 1.871***.

Dependent variable: Citizen; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: 1.619***; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: 1.570***.

Dependent variable: Naturalized immigrant; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: 0.991; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: 1.055.

Dependent variable: Part-time status (omitted category is full 
time)[A]; 
Unadjusted proportion participating in a DB or DC plan: 88%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Part-time; 
Unadjusted proportion participating in a DB or DC plan: 81%; 
Individual participates in a DB or DC plan: 0.791***; 
Unadjusted proportion participating in a DC plan: 74%; 
Individual participates in a DC plan: 0.851***.

Dependent variable: Sector (omitted category private sector); 
Unadjusted proportion participating in a DB or DC plan: 83%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 77%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Private not for profit; 
Unadjusted proportion participating in a DB or DC plan: 88%; 
Individual participates in a DB or DC plan: 1.274***; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: 1.219***.

Dependent variable: Government worker; 
Unadjusted proportion participating in a DB or DC plan: 94%; 
Individual participates in a DB or DC plan: 1.902***; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 1.239**.

Dependent variable: Occupation (omitted category Management); 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 87%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Business and Financial Operations; 
Unadjusted proportion participating in a DB or DC plan: 91%; 
Individual participates in a DB or DC plan: 0.885; 
Unadjusted proportion participating in a DC plan: 85%; 
Individual participates in a DC plan: 0.847.

Dependent variable: Computer and Mathematical; 
Unadjusted proportion participating in a DB or DC plan: 91%; 
Individual participates in a DB or DC plan: 0.823; 
Unadjusted proportion participating in a DC plan: 86%; 
Individual participates in a DC plan: 0.790.

Dependent variable: Architecture and Engineering; 
Unadjusted proportion participating in a DB or DC plan: 94%; 
Individual participates in a DB or DC plan: 1.256; 
Unadjusted proportion participating in a DC plan: 89%; 
Individual participates in a DC plan: 0.969.

Dependent variable: Life, Physical, and Social Services; 
Unadjusted proportion participating in a DB or DC plan: 96%; 
Individual participates in a DB or DC plan: 1.387; 
Unadjusted proportion participating in a DC plan: 90%; 
Individual participates in a DC plan: 1.147.

Dependent variable: Community and Social Services; 
Unadjusted proportion participating in a DB or DC plan: 88%; 
Individual participates in a DB or DC plan: 0.636**; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 0.808.

Dependent variable: Legal; 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: 0.883; 
Unadjusted proportion participating in a DC plan: 86%; 
Individual participates in a DC plan: 0.856.

Dependent variable: Education, Training, and Library; 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: 0.606***; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: 0.520***.

Dependent variable: Arts, Design, Entertainment, Sports, and Media; 
Unadjusted proportion participating in a DB or DC plan: 86%; 
Individual participates in a DB or DC plan: 0.661*; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: 0.609**.

Dependent variable: Healthcare Practitioners and Technical; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: 0.712**; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: 0.694**.

Dependent variable: Healthcare Support; 
Unadjusted proportion participating in a DB or DC plan: 75%; 
Individual participates in a DB or DC plan: 0.545***; 
Unadjusted proportion participating in a DC plan: 64%; 
Individual participates in a DC plan: 0.457***.

Dependent variable: Protective Service; 
Unadjusted proportion participating in a DB or DC plan: 93%; 
Individual participates in a DB or DC plan: 0.727; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: 0.713**.

Dependent variable: Food Preparation and Serving Related; 
Unadjusted proportion participating in a DB or DC plan: 63%; 
Individual participates in a DB or DC plan: 0.556***; 
Unadjusted proportion participating in a DC plan: 53%; 
Individual participates in a DC plan: 0.488***.

Dependent variable: Building and Grounds Cleaning and Maintenance; 
Unadjusted proportion participating in a DB or DC plan: 76%; 
Individual participates in a DB or DC plan: 0.570***; 
Unadjusted proportion participating in a DC plan: 63%; 
Individual participates in a DC plan: 0.542***.

Dependent variable: Personal Care and Service; 
Unadjusted proportion participating in a DB or DC plan: 73%; 
Individual participates in a DB or DC plan: 0.516***; 
Unadjusted proportion participating in a DC plan: 63%; 
Individual participates in a DC plan: 0.483***.

Dependent variable: Sales and Related; 
Unadjusted proportion participating in a DB or DC plan: 82%; 
Individual participates in a DB or DC plan: 0.734**; 
Unadjusted proportion participating in a DC plan: 75%; 
Individual participates in a DC plan: 0.678***.

Dependent variable: Office and Administrative Support; 
Unadjusted proportion participating in a DB or DC plan: 84%; 
Individual participates in a DB or DC plan: 0.576***; 
Unadjusted proportion participating in a DC plan: 76%; 
Individual participates in a DC plan: 0.574***.

Dependent variable: Farming, Forestry, and Fishing; 
Unadjusted proportion participating in a DB or DC plan: 82%; 
Individual participates in a DB or DC plan: 0.913; 
Unadjusted proportion participating in a DC plan: 78%; 
Individual participates in a DC plan: 1.014.

Dependent variable: Construction and Extraction; 
Unadjusted proportion participating in a DB or DC plan: 89%; 
Individual participates in a DB or DC plan: 0.914; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: 0.681*.

Dependent variable: Installation, Repair, and Maintenance; 
Unadjusted proportion participating in a DB or DC plan: 85%; 
Individual participates in a DB or DC plan: 0.627***; 
Unadjusted proportion participating in a DC plan: 77%; 
Individual participates in a DC plan: 0.613***.

Dependent variable: Production; 
Unadjusted proportion participating in a DB or DC plan: 82%; 
Individual participates in a DB or DC plan: 0.476***; 
Unadjusted proportion participating in a DC plan: 75%; 
Individual participates in a DC plan: 0.523***.

Dependent variable: Transportation and Material Moving; 
Unadjusted proportion participating in a DB or DC plan: 83%; 
Individual participates in a DB or DC plan: 0.723**; 
Unadjusted proportion participating in a DC plan: 73%; 
Individual participates in a DC plan: 0.650***.

Dependent variable: Not in universe[B]; 
Unadjusted proportion participating in a DB or DC plan: 94%; 
Individual participates in a DB or DC plan: 2.604*; 
Unadjusted proportion participating in a DC plan: 88%; 
Individual participates in a DC plan: 0.765.

Dependent variable: Industry (omitted category Agriculture); 
Unadjusted proportion participating in a DB or DC plan: 76%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 72%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Mining; 
Unadjusted proportion participating in a DB or DC plan: 94%; 
Individual participates in a DB or DC plan: 5.874***; 
Unadjusted proportion participating in a DC plan: 91%; 
Individual participates in a DC plan: 5.571***.

Dependent variable: Utilities; 
Unadjusted proportion participating in a DB or DC plan: 94%; 
Individual participates in a DB or DC plan: 3.325***; 
Unadjusted proportion participating in a DC plan: 84%; 
Individual participates in a DC plan: 1.871*.

Dependent variable: Construction; 
Unadjusted proportion participating in a DB or DC plan: 89%; 
Individual participates in a DB or DC plan: 2.673***; 
Unadjusted proportion participating in a DC plan: 81%; 
Individual participates in a DC plan: 2.222**.

Dependent variable: Manufacturing; 
Unadjusted proportion participating in a DB or DC plan: 88%; 
Individual participates in a DB or DC plan: 2.907***; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 2.401***.

Dependent variable: Wholesale Trade; 
Unadjusted proportion participating in a DB or DC plan: 86%; 
Individual participates in a DB or DC plan: 2.567**; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 2.330**.

Dependent variable: Retail Trade; 
Unadjusted proportion participating in a DB or DC plan: 77%; 
Individual participates in a DB or DC plan: 1.770*; 
Unadjusted proportion participating in a DC plan: 69%; 
Individual participates in a DC plan: 1.435.

Dependent variable: Transportation and Warehousing; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: 1.937*; 
Unadjusted proportion participating in a DC plan: 77%; 
Individual participates in a DC plan: 1.683.

Dependent variable: Information; 
Unadjusted proportion participating in a DB or DC plan: 88%; 
Individual participates in a DB or DC plan: 2.520**; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 2.107**.

Dependent variable: Finance and Insurance; 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: 4.287***; 
Unadjusted proportion participating in a DC plan: 86%; 
Individual participates in a DC plan: 2.868***.

Dependent variable: Real Estate and Rental and Leasing; 
Unadjusted proportion participating in a DB or DC plan: 77%; 
Individual participates in a DB or DC plan: 1.346; 
Unadjusted proportion participating in a DC plan: 68%; 
Individual participates in a DC plan: 1.120.

Dependent variable: Professional, Scientific, and Technical; 
Unadjusted proportion participating in a DB or DC plan: 89%; 
Individual participates in a DB or DC plan: 2.555***; 
Unadjusted proportion participating in a DC plan: 84%; 
Individual participates in a DC plan: 2.059**.

Dependent variable: Management, Administrative and Support; 
Unadjusted proportion participating in a DB or DC plan: 74%; 
Individual participates in a DB or DC plan: 1.511; 
Unadjusted proportion participating in a DC plan: 68%; 
Individual participates in a DC plan: 1.291.

Dependent variable: Educational Services; 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: 2.231**; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: 1.612.

Dependent variable: Health Care and Social Assistance; 
Unadjusted proportion participating in a DB or DC plan: 83%; 
Individual participates in a DB or DC plan: 1.824*; 
Unadjusted proportion participating in a DC plan: 76%; 
Individual participates in a DC plan: 1.513.

Dependent variable: Arts, Entertainment, and Recreation; 
Unadjusted proportion participating in a DB or DC plan: 75%; 
Individual participates in a DB or DC plan: 1.253; 
Unadjusted proportion participating in a DC plan: 66%; 
Individual participates in a DC plan: 1.117.

Dependent variable: Accommodations and Food Services; 
Unadjusted proportion participating in a DB or DC plan: 60%; 
Individual participates in a DB or DC plan: 1.057; 
Unadjusted proportion participating in a DC plan: 55%; 
Individual participates in a DC plan: 1.017.

Dependent variable: Other Services (Except Public Administration); 
Unadjusted proportion participating in a DB or DC plan: 81%; 
Individual participates in a DB or DC plan: 1.845*; 
Unadjusted proportion participating in a DC plan: 71%; 
Individual participates in a DC plan: 1.183.

Dependent variable: Public Administration; 
Unadjusted proportion participating in a DB or DC plan: 96%; 
Individual participates in a DB or DC plan: 4.284***; 
Unadjusted proportion participating in a DC plan: 85%; 
Individual participates in a DC plan: 1.868*.

Dependent variable: Work experience (omitted category Less than 5 
years); 
Unadjusted proportion participating in a DB or DC plan: 77%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 69%; 
Individual participates in a DC plan: [Empty].

Dependent variable: 5 to 9 years; 
Unadjusted proportion participating in a DB or DC plan: 85%; 
Individual participates in a DB or DC plan: 1.150**; 
Unadjusted proportion participating in a DC plan: 77%; 
Individual participates in a DC plan: 1.185**.

Dependent variable: 10 to 15 years; 
Unadjusted proportion participating in a DB or DC plan: 89%; 
Individual participates in a DB or DC plan: 1.250***; 
Unadjusted proportion participating in a DC plan: 82%; 
Individual participates in a DC plan: 1.305***.

Dependent variable: More than 15 years; 
Unadjusted proportion participating in a DB or DC plan: 91%; 
Individual participates in a DB or DC plan: 1.195**; 
Unadjusted proportion participating in a DC plan: 84%; 
Individual participates in a DC plan: 1.250***.

Dependent variable: Union status (omitted category not in a union); 
Unadjusted proportion participating in a DB or DC plan: 85%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: [Empty]; 
Individual participates in a DC plan: [Empty].

Dependent variable: In a union; 
Unadjusted proportion participating in a DB or DC plan: 93%; 
Individual participates in a DB or DC plan: 1.579***; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: 0.980.

Dependent variable: Self-employment status; 
Unadjusted proportion participating in a DB or DC plan: 91%; 
Individual participates in a DB or DC plan: 1.020; 
Unadjusted proportion participating in a DC plan: 85%; 
Individual participates in a DC plan: 0.860.

Dependent variable: Number of employees at current place of employment 
(omitted category Under 25 employees); 
Unadjusted proportion participating in a DB or DC plan: 83%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: [Empty].

Dependent variable: 25 to 100 employees; 
Unadjusted proportion participating in a DB or DC plan: 82%; 
Individual participates in a DB or DC plan: 0.916; 
Unadjusted proportion participating in a DC plan: 77%; 
Individual participates in a DC plan: 0.878.

Dependent variable: 100+ employees; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: 1.172*; 
Unadjusted proportion participating in a DC plan: 79%; 
Individual participates in a DC plan: 0.936.

Dependent variable: Years of tenure at current job; 
Unadjusted proportion participating in a DB or DC plan: [Empty]; 
Individual participates in a DB or DC plan: 1.084***; 
Unadjusted proportion participating in a DC plan: [Empty]; 
Individual participates in a DC plan: 1.036***.

Tenure categories: 

Dependent variable: Less than 5 years; 
Unadjusted proportion participating in a DB or DC plan: 77%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 69%; 
Individual participates in a DC plan: [Empty].

Dependent variable: 5 to 9 years; 
Unadjusted proportion participating in a DB or DC plan: 87%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 80%; 
Individual participates in a DC plan: [Empty].

Dependent variable: 10 to 15 years; 
Unadjusted proportion participating in a DB or DC plan: 92%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 85%; 
Individual participates in a DC plan: [Empty].

Dependent variable: More than 15 years; 
Unadjusted proportion participating in a DB or DC plan: 96%; 
Individual participates in a DB or DC plan: [Empty]; 
Unadjusted proportion participating in a DC plan: 87%; 
Individual participates in a DC plan: [Empty].

Dependent variable: Number of observations; 
Individual participates in a DB or DC plan: 21,494; 
Individual participates in a DC plan: 17,067.

Source: GAO analysis of SIPP data.

* Indicates that the variable is statistically significant at the 90 
percent level.

** Indicates that the variable is statistically significant at the 95 
percent level.

*** Indicates that the variable is statistically significant at the 99 
percent level.

[A] Part-time status is defined as working 35 hours or less per week 
during the reference period.

[B] The category "Not in universe" includes self-employed individuals. 

[End of table] 

Section 3: Methods for Comparing the Income of Women and Men Age 65 
and Over:

To compute median incomes and income composition for men and women in 
different demographic groups, we used information from the core 
questionnaire of the SIPP data (as described above). We used the last 
month of the 4-month reporting period (within each "wave") with the 
assumption that individuals will more accurately recollect income from 
the most recent month than income from 4 months ago. To obtain an 
annual income estimate, we multiplied the monthly reported income by 
12.7[Footnote 65]

The poverty rate was computed using a SIPP variable that indicates the 
poverty threshold for an individual's household. The Census Bureau 
uses a set of money-income thresholds that vary by family size and 
composition to determine who is in poverty. If a family's total income 
is less than the family's threshold, then that family and every 
individual in it is considered in poverty. The official poverty 
thresholds do not vary geographically, but they are updated for 
inflation using Consumer Price Index (CPI-U). The official poverty 
definition uses money income before taxes and does not include capital 
gains or noncash benefits (such as public housing, Medicaid, and food 
stamps).

All of our income composition, median, and poverty level estimates 
were computed at the individual level, using household-level 
information. In other words, median incomes were computed by applying 
all household income to each individual in the household and taking 
the median across all individuals within a certain category (e.g., 
gender, or gender and race). For married individuals, this means that 
spousal income was included in these estimates. Correspondingly, we 
used SIPP individual- level weights to compute our point estimates 
and, in conjunction with other factors, calculate the standard errors 
of those estimates so that we could accurately account for the complex 
survey design.

The point estimates for household income for married men and married 
women may not be equal for a couple of reasons. First, the criteria 
for including an individual in the sample in our analysis was that he 
or she was 65 or above. While there are more women than men among all 
people over 65 in our sample, among married people over 65 there are 
more men than women.8[Footnote 66]One reason this might occur is due 
to demographic patterns of life-expectancy and the ages of marital 
partners.9[Footnote 67]Since women typically marry older men, and 
women typically have longer life-expectancies than men, it is not 
surprising that a sample of older individuals will include fewer 
married women than married men, as the spouses of older women are more 
likely to have died than the spouses of older men. For this reason, 
the sample of married older women could differ from the sample of 
married older men, so their household characteristics--including 
income--may not be the same. Further, the difference between the ages 
of the spouses of married men and married women could also result in 
different estimates of median income and income composition. For 
example, if women tended to be married to older men, the income 
composition of the household might be skewed away from earnings and 
towards Social Security. Conversely, if men tended to be married to 
younger women, a higher share of income might come from earnings.

Section 4: Methods for Analyzing the Effects of Events Occurring Later 
in Life on Women's and Men's Household Income and Assets:

We estimated the relationship between events that occur later in life 
and income and assets using fixed-effects panel regressions. The main 
advantage of fixed-effects models is that they are designed to isolate 
the effect of the event from all other permanent characteristics of 
the individual. We estimated our models using data from the HRS, which 
follows households over time. Our analysis focuses on life events that 
occur after age 50, as the HRS follows individuals age 51 and over.

Descriptive Analysis of the Frequency of Life Events by Gender:

Prior to analyzing the effect of the life events on assets and income, 
we first estimated the differences in the frequency of life events by 
gender. We estimated these differences in two ways. First, we 
estimated the proportion of women and men that had a life event across 
all the periods (e.g., proportion that were divorced). Second, we 
estimated the proportion of women and men that had that life event 
change between two periods of observation (e.g., proportion that 
became divorced between period 1 and period 2).

Table 19 uses the first method and presents some descriptive 
statistics on the women and men in our sample. Specifically, it shows 
the average values of characteristics for different ages for women and 
men.

* Real assets and real income. At ages 51 to 64 women and men have 
similar levels of assets. However, after age 65, men's average level 
of household assets becomes larger than the average level for women. 
Men's average levels of household income are higher than women's at 
every age level.

* Marital status. The rates of marriage and widowhood are relatively 
comparable between women and men before age 65. For example, 6 percent 
of women and 1 percent of men younger than age 65 were widowed. 
However, at older ages, more women were estimated to be widowed than 
men.

* Poor health. Individuals were classified as being in poor health 
based on a survey question of self-reported health, which asked the 
individual to rate his or her health on a scale from 1 to 5, where 1 
is excellent and 5 is poor. An answer of "fair" or "poor" was 
classified as being in poor health. As table 17 shows, rates of poor 
health were comparable between women and men at all age groups.

* Unemployment. This variable captures the percentage of individuals 
that responded to a labor force question as being "unemployed". It is 
important to note that this is not equivalent to an unemployment rate--
as individuals classified as not in the labor force were included in 
the denominator. Women and men were equally likely to report being 
unemployed.

* Helping parents financially or with daily activities. These 
variables capture the percentage of households that provided financial 
help or assistance with basic daily activities to either the parents 
of the respondent or spouse. Again, it appears that these rates were 
comparable for women and men.

Table 17: Descriptive Statistics of Women and Men in the HRS by Age:

Women: Mean real household assets; 

Age: Under age 51; 
Estimate: $446,436; 
Standard error: $18,911; 
Lower bound of 95 percent confidence interval: $409,372; 
Upper bound of 95 percent confidence interval: $483,500; 
Error over estimate: 4.24%.

Age: Ages 51-64; 
Estimate: $537,262; 
Standard error: $9,039; 
Lower bound of 95 percent confidence interval: $519,546; 
Upper bound of 95 percent confidence interval: $554,978; 
Error over estimate: 1.68%.

Age: Ages 65-84; 
Estimate: $522,190; 
Standard error: $6,020; 
Lower bound of 95 percent confidence interval: $510,391; 
Upper bound of 95 percent confidence interval: $533,989; 
Error over estimate: 1.15%.

Age: Ages 85-99; 
Estimate: $359,269; 
Standard error: $12,341; 
Lower bound of 95 percent confidence interval: $335,071; 
Upper bound of 95 percent confidence interval: $383,447; 
Error over estimate: 3.44%.

Women: Mean real household income; 

Age: Under age 51; 
Estimate: $128,325; 
Standard error: $6,899; 
Lower bound of 95 percent confidence interval: $114,803; 
Upper bound of 95 percent confidence interval: $141,847; 
Error over estimate: 5.38%.

Age: Ages 51-64; 
Estimate: $98,116; 
Standard error: $1,185; 
Lower bound of 95 percent confidence interval: $95,794; 
Upper bound of 95 percent confidence interval: $100,438; 
Error over estimate: 1.21%.

Age: Ages 65-84; 
Estimate: $55,014; 
Standard error: $348; 
Lower bound of 95 percent confidence interval: $54,332; 
Upper bound of 95 percent confidence interval: $55,696; 
Error over estimate: 0.63%.

Age: Ages 85-99; 
Estimate: $32,728; 
Standard error: $746; 
Lower bound of 95 percent confidence interval: $31,275; 
Upper bound of 95 percent confidence interval: $34,201; 
Error over estimate: 2.28%.

Women: Percent married; 

Age: Under age 51; 
Estimate: 87%; 
Standard error: 0.46%; 
Lower bound of 95 percent confidence interval: 86%; 
Upper bound of 95 percent confidence interval: 88%; 
Error over estimate: 0.52%.

Age: Ages 51-64; 
Estimate: 78%; 
Standard error: 0.19%; 
Lower bound of 95 percent confidence interval: 78%; 
Upper bound of 95 percent confidence interval: 78%; 
Error over estimate: 0.24%.

Age: Ages 65-84; 
Estimate: 64%; 
Standard error: 0.20%; 
Lower bound of 95 percent confidence interval: 63%; 
Upper bound of 95 percent confidence interval: 64%; 
Error over estimate: 0.32%.

Age: Ages 85-99; 
Estimate: 23%; 
Standard error: 0.48%; 
Lower bound of 95 percent confidence interval: 22%; 
Upper bound of 95 percent confidence interval: 24%; 
Error over estimate: 2.05%.

Women: Percent divorced or separated; 

Age: Under age 51; 
Estimate: 6%; 
Standard error: 0.33%; 
Lower bound of 95 percent confidence interval: 5%; 
Upper bound of 95 percent confidence interval: 6%; 
Error over estimate: 5.67%.

Age: Ages 51-64; 
Estimate: 12%; 
Standard error: 0.16%; 
Lower bound of 95 percent confidence interval: 12%; 
Upper bound of 95 percent confidence interval: 13%; 
Error over estimate: 1.27%.

Age: Ages 65-84; 
Estimate: 8%; 
Standard error: 0.12%; 
Lower bound of 95 percent confidence interval: 7%; 
Upper bound of 95 percent confidence interval: 8%; 
Error over estimate: 1.55%.

Age: Ages 85-99; 
Estimate: 4%; 
Standard error: 0.23%; 
Lower bound of 95 percent confidence interval: 4%; 
Upper bound of 95 percent confidence interval: %; 
Error over estimate: 5.71%.

Women: Percent widowed; 

Age: Under age 51; 
Estimate: 1%; 
Standard error: 0.11%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 1%; 
Error over estimate: 16.19%.

Age: Ages 51-64; 
Estimate: 6%; 
Standard error: 0.10%; 
Lower bound of 95 percent confidence interval: 6%; 
Upper bound of 95 percent confidence interval: 6%; 
Error over estimate: 1.74%.

Age: Ages 65-84; 
Estimate: 26%; 
Standard error: 0.19%; 
Lower bound of 95 percent confidence interval: 26%; 
Upper bound of 95 percent confidence interval: 26%; 
Error over estimate: 0.72%.

Age: Ages 85-99; 
Estimate: 70%; 
Standard error: 0.53%; 
Lower bound of 95 percent confidence interval: 69%; 
Upper bound of 95 percent confidence interval: 71%; 
Error over estimate: 0.76%.

Women: Percent in poor health; 

Age: Under age 51; 
Estimate: 15%; 
Standard error: 0.45%; 
Lower bound of 95 percent confidence interval: 14%; 
Upper bound of 95 percent confidence interval: 16%; 
Error over estimate: 3.01%.

Age: Ages 51-64; 
Estimate: 21%; 
Standard error: 0.18%; 
Lower bound of 95 percent confidence interval: 21%; 
Upper bound of 95 percent confidence interval: 21%; 
Error over estimate: 0.87%.

Age: Ages 65-84; 
Estimate: 28%; 
Standard error: 0.19%; 
Lower bound of 95 percent confidence interval: 28%; 
Upper bound of 95 percent confidence interval: 29%; 
Error over estimate: 0.66%.

Age: Ages 85-99; 
Estimate: 40%; 
Standard error: 0.57%; 
Lower bound of 95 percent confidence interval: 39%; 
Upper bound of 95 percent confidence interval: 41%; 
Error over estimate: 1.41%.

Women: Percent unemployed; 

Age: Under age 51; 
Estimate: 3%; 
Standard error: 0.22%; 
Lower bound of 95 percent confidence interval: 3%; 
Upper bound of 95 percent confidence interval: 3%; 
Error over estimate: 7.34%.

Age: Ages 51-64; 
Estimate: 2%; 
Standard error: 0.06%; 
Lower bound of 95 percent confidence interval: 2%; 
Upper bound of 95 percent confidence interval: 2%; 
Error over estimate: 3.48%.

Age: Ages 65-84; 
Estimate: 0.1%; 
Standard error: 0.01%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 0%; 
Error over estimate: 13.11%.

Age: Women: Ages 85-99; 
Estimate: Women: 0%; 
Standard error: Women: 0.01%; 
Lower bound of 95 percent confidence interval: Women: 0%; 
Upper bound of 95 percent confidence interval: Women: 0%; 
Error over estimate: 100.00%.

Women: Percent who helped their parents financially; 

Age: Under age 51; 
Estimate: 16%; 
Standard error: 0.48%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 17%; 
Error over estimate: 3.00%.

Age: Ages 51-64; 
Estimate: 11%; 
Standard error: 0.15%; 
Lower bound of 95 percent confidence interval: 11%; 
Upper bound of 95 percent confidence interval: 11%; 
Error over estimate: 1.33%.

Age: Ages 65-84; 
Estimate: 2%; 
Standard error: 0.06%; 
Lower bound of 95 percent confidence interval: 2%; 
Upper bound of 95 percent confidence interval: 2%; 
Error over estimate: 3.06%.

Age: Women: Ages 85-99; 
Estimate: 0.1%; 
Standard error: 0.03%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 0%; 
Error over estimate: 48.61%.

Women: Percent who helped their parents with daily activities; 

Age: Under age 51; 
Estimate: 8%; 
Standard error: 0.34%; 
Lower bound of 95 percent confidence interval: 7%; 
Upper bound of 95 percent confidence interval: 8%; 
Error over estimate: 4.56%.

Age: Ages 51-64; 
Estimate: 9%; 
Standard error: 0.13%; 
Lower bound of 95 percent confidence interval: 9%; 
Upper bound of 95 percent confidence interval: 9%; 
Error over estimate: 1.51%.

Age: Ages 65-84; 
Estimate: 3%; 
Standard error: 0.07%; 
Lower bound of 95 percent confidence interval: 2%; 
Upper bound of 95 percent confidence interval: 3%; 
Error over estimate: 2.70%.

Age: Ages 85-99; 
Estimate: 0.1%; 
Standard error: 0.03%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 0%; 
Error over estimate: 52.68%.

Men: Mean real household assets; 

Age: Under age 51; 
Estimate: $369,106; 
Standard error: $19,9401; 
Lower bound of 95 percent confidence interval: $330,023; 
Upper bound of 95 percent confidence interval: $408,189; 
Error over estimate: 5.40%.

Age: Ages 51-64; 
Estimate: $540,761; 
Standard error: $9,581; 
Lower bound of 95 percent confidence interval: $521,982; 
Upper bound of 95 percent confidence interval: $559,541; 
Error over estimate: 1.77%.

Age: Ages 65-84; 
Estimate: $638,166; 
Standard error: $8,352; 
Lower bound of 95 percent confidence interval: $621,796; 
Upper bound of 95 percent confidence interval: $654,537; 
Error over estimate: 1.31%.

Age: Ages 85-99; 
Estimate: $528,611; 
Standard error: $17,682; 
Lower bound of 95 percent confidence interval: $493,955; 
Upper bound of 95 percent confidence interval: $563,268; 
Error over estimate: 3.35%.

Men: 
Mean real household income; 
Age: Under age 51; 
Estimate: $107,801; 
Standard error: $3,371; 
Lower bound of 95 percent confidence interval: $101,194; 
Upper bound of 95 percent confidence interval: $114,409; 
Error over estimate: 3.13%.

Age: Ages 51-64; 
Estimate: $112,785; 
Standard error: $1,726; 
Lower bound of 95 percent confidence interval: $109,402; 
Upper bound of 95 percent confidence interval: $116,168; 
Error over estimate: 1.53%.

Age: Ages 65-84; 
Estimate: $72,767; 
Standard error: $1,718; 
Lower bound of 95 percent confidence interval: $69,400; 
Upper bound of 95 percent confidence interval: $76,133; 
Error over estimate: 2.36%.

Age: Ages 85-99; 
Estimate: $48,073; 
Standard error: $1,023; 
Lower bound of 95 percent confidence interval: $46,068; 
Upper bound of 95 percent confidence interval: $50,078; 
Error over estimate: 2.13%.

Men: Percent married; 

Age: Under age 51; 
Estimate: 76%; 
Standard error: 1.25%; 
Lower bound of 95 percent confidence interval: 74%; 
Upper bound of 95 percent confidence interval: 79%; 
Error over estimate: 1.63%.

Age: Ages 51-64; 
Estimate: 83%; 
Standard error: 0.20%; 
Lower bound of 95 percent confidence interval: 83%; 
Upper bound of 95 percent confidence interval: 84%; 
Error over estimate: 0.24%.

Age: Ages 65-84; 
Estimate: 85%; 
Standard error: 0.17%; 
Lower bound of 95 percent confidence interval: 85%; 
Upper bound of 95 percent confidence interval: 85%; 
Error over estimate: 0.19%.

Age: Ages 85-99; 
Estimate: 71%; 
Standard error: 0.62%; 
Lower bound of 95 percent confidence interval: 70%; 
Upper bound of 95 percent confidence interval: 72%; 
Error over estimate: 0.87%.

Men: Percent Divorced or Separated; 

Age: Under age 51; 
Estimate: 9%; 
Standard error: 0.85%; 
Lower bound of 95 percent confidence interval: 7%; 
Upper bound of 95 percent confidence interval: 11%; 
Error over estimate: 9.61%.

Age: Ages 51-64; 
Estimate: 10%; 
Standard error: 0.17%; 
Lower bound of 95 percent confidence interval: 10%; 
Upper bound of 95 percent confidence interval: 11%; 
Error over estimate: 1.60%.

Age: Ages 65-84; 
Estimate: 6%; 
Standard error: 0.11%; 
Lower bound of 95 percent confidence interval: 5%; 
Upper bound of 95 percent confidence interval: 6%; 
Error over estimate: 2.02%.

Age: Ages 85-99; 
Estimate: 2%; 
Standard error: 0.21%; 
Lower bound of 95 percent confidence interval: 2%; 
Upper bound of 95 percent confidence interval: 3%; 
Error over estimate: 8.62%.

Men: Percent widowed; 

Age: Under age 51; 
Estimate: 0.1%; 
Standard error: 0.10%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 0%; 
Error over estimate: 70.66%.

Age: Ages 51-64; 
Estimate: 1%; 
Standard error: 0.06%; 
Lower bound of 95 percent confidence interval: 1%; 
Upper bound of 95 percent confidence interval: 2%; 
Error over estimate: 4.24%.

Age: Ages 65-84; 
Estimate: 7%; 
Standard error: 0.11%; 
Lower bound of 95 percent confidence interval: 6%; 
Upper bound of 95 percent confidence interval: 7%; 
Error over estimate: 1.66%.

Age: Ages 85-99; 
Estimate: 24%; 
Standard error: 0.59%; 
Lower bound of 95 percent confidence interval: 23%; 
Upper bound of 95 percent confidence interval: 26%; 
Error over estimate: 2.41%.

Men: Percent in poor health; 

Age: Under age 51; 
Estimate: 18%; 
Standard error: 1.11%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 20%; 
Error over estimate: 6.34%.

Age: Ages 51-64; 
Estimate: 20%; 
Standard error: 0.20%; 
Lower bound of 95 percent confidence interval: 20%; 
Upper bound of 95 percent confidence interval: 21%; 
Error over estimate: 0.98%.

Age: Ages 65-84; 
Estimate: 28%; 
Standard error: 0.19%; 
Lower bound of 95 percent confidence interval: 28%; 
Upper bound of 95 percent confidence interval: 29%; 
Error over estimate: 0.68%.

Age: Ages 85-99; 
Estimate: 40%; 
Standard error: 0.66%; 
Lower bound of 95 percent confidence interval: 39%; 
Upper bound of 95 percent confidence interval: 41%; 
Error over estimate: 1.65%.

Men: Percent unemployed; 

Age: Under age 51; 
Estimate: 3%; 
Standard error: 0.49%; 
Lower bound of 95 percent confidence interval: 2%; 
Upper bound of 95 percent confidence interval: 4%; 
Error over estimate: 15.70%.

Age: Ages 51-64; 
Estimate: 2%; 
Standard error: 0.07%; 
Lower bound of 95 percent confidence interval: 2%; 
Upper bound of 95 percent confidence interval: 2%; 
Error over estimate: 3.58%.

Age: Ages 65-84; 
Estimate: 0.2%; 
Standard error: 0.02%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 0%; 
Error over estimate: 9.73%.

Age: Ages 85-99; 
Estimate: 0%; 
Standard error: 0.00%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 0%; 
Error over estimate: [Empty]. 

Men: Percent who helped their parents financially; 

Age: Under age 51; 
Estimate: 17%; 
Standard error: 1.14%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 20%; 
Error over estimate: 6.56%.

Age: Ages 51-64; 
Estimate: 13%; 
Standard error: 0.17%; 
Lower bound of 95 percent confidence interval: 13%; 
Upper bound of 95 percent confidence interval: 13%; 
Error over estimate: 1.34%.

Age: Ages 65-84; 
Estimate: 4%; 
Standard error: 0.09%; 
Lower bound of 95 percent confidence interval: 4%; 
Upper bound of 95 percent confidence interval: 4%; 
Error over estimate: 2.26%.

Age: Ages 85-99; 
Estimate: 0.2%; 
Standard error: 0.06%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 0%; 
Error over estimate: 28.54%.

Men: Percent who helped their parents with daily activities; 

Age: Under age 51; 
Estimate: 10%; 
Standard error: 0.91%; 
Lower bound of 95 percent confidence interval: 8%; 
Upper bound of 95 percent confidence interval: 12%; 
Error over estimate: 9.17%.

Age: Ages 51-64; 
Estimate: 9%; 
Standard error: 0.14%; 
Lower bound of 95 percent confidence interval: 8%; 
Upper bound of 95 percent confidence interval: 9%; 
Error over estimate: 1.67%.

Age: Ages 65-84; 
Estimate: 4%; 
Standard error: 0.09%; 
Lower bound of 95 percent confidence interval: 4%; 
Upper bound of 95 percent confidence interval: 4%; 
Error over estimate: 2.27%.

Age: Ages 85-99; 
Estimate: 0.3%; 
Standard error: 0.10%; 
Lower bound of 95 percent confidence interval: 0%; 
Upper bound of 95 percent confidence interval: 1%; 
Error over estimate: 28.65%.

Source: GAO analysis of HRS data. 

[End of table] 

Table 18 uses the second method to show the proportion of women and 
men that had a life event status change during the period of analysis. 
As table 18 shows:

* Divorce/separation. During the period in which both members of the 
household are less than 65, less than 1 percent of men experienced 
divorce or separation between any of the two waves. For women, the 
proportion was negative - indicating that more women went from 
divorced or separated to married than from married to divorced or 
separated.

* Widowhood. During the earlier period, about 1 percent of women 
became widowed between any of the two waves. This proportion increased 
to more than 2 percent as the household aged and was twice the rate 
for men.

* Decline into poor health. The rate of health decline was similar for 
women and men. On average, approximately 2 percent of women and men 
reported a decline in health from one period to another.

* Unemployment. Very few women and men reported a change to and from 
unemployment in our data.

* Helping parents financially or with daily activities. The proportion 
of women's and men's households providing personal or financial 
assistance fell as the household aged. This may be because older 
households were less likely to have living parents requiring assistance.

* Percent change in real assets. In the earlier period, assets for 
women and men increased at a rate of about 6 percent per 2-year 
period. Alternatively, the rate of asset growth became negative as the 
household aged.

* Percent change in real income. In both younger and older households, 
incomes fell at a rate of approximately 5 percent per 2-year period, 
on average.

Table 18: Proportion of Individuals Changing Status between 
Observations:

Women: Divorced or separated; 


Household type: Households where everyone is age 64 or younger; 
Estimate: -0.0022; 
Standard error: 0.0008; 
Lower bound of 95 percent confidence interval: -0.0038; 
Upper bound of 95 percent confidence interval: -0.0007; 
Error over estimate: (35.40%).

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0011; 
Standard error: 0.0005; 
Lower bound of 95 percent confidence interval: -0.002; 
Upper bound of 95 percent confidence interval: -0.0001; 
Error over estimate: (46.32%).

Household type: All households; 
Estimate: -0.0016; 
Standard error: 0.0005; 
Lower bound of 95 percent confidence interval: -0.0025; 
Upper bound of 95 percent confidence interval: -0.0007; 
Error over estimate: (28.13%).

Women: Widowhood; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0106; 
Standard error: 0.0007; 
Lower bound of 95 percent confidence interval: 0.0094; 
Upper bound of 95 percent confidence interval: 0.0119; 
Error over estimate: 6.13%.

Household type: Households where at least one person is 65 or over; 
Estimate: 0.0237; 
Standard error: 0.0007; 
Lower bound of 95 percent confidence interval: 0.0223; 
Upper bound of 95 percent confidence interval: 0.0251; 
Error over estimate: 3.05%.

Household type: All households; 
Estimate: 0.0177; 
Standard error: 0.0005; 
Lower bound of 95 percent confidence interval: 0.0168; 
Upper bound of 95 percent confidence interval: 0.0187; 
Error over estimate: 2.79%.

Women: Decline in health; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0128; 
Standard error: 0.0019; 
Lower bound of 95 percent confidence interval: 0.009; 
Upper bound of 95 percent confidence interval: 0.0165; 
Error over estimate: 14.93%.

Household type: Households where at least one person is 65 or over; 
Estimate: 0.026; 
Standard error: 0.0016; 
Lower bound of 95 percent confidence interval: 0.0229; 
Upper bound of 95 percent confidence interval: 0.0291; 
Error over estimate: 6.07%.

Household type: All households; 
Estimate: 0.02; 
Standard error: 0.0012; 
Lower bound of 95 percent confidence interval: 0.0176; 
Upper bound of 95 percent confidence interval: 0.0224; 
Error over estimate: 6.12%.

Women: Unemployment; 

Household type: Households where everyone is age 64 or younger; 
Estimate: -0.0026; 
Standard error: 0.0011; 
Lower bound of 95 percent confidence interval: -0.0047; 
Upper bound of 95 percent confidence interval: -0.0005; 
Error over estimate: (40.90%).

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0003; 
Standard error: 0.0003; 
Lower bound of 95 percent confidence interval: -0.001; 
Upper bound of 95 percent confidence interval: 0.0003; 
Error over estimate: (90.77%).

Household type: All households; 
Estimate: -0.0014; 
Standard error: 0.0005; 
Lower bound of 95 percent confidence interval: -0.0024; 
Upper bound of 95 percent confidence interval: -0.0004; 
Error over estimate: (37.43%).

Women: Helped parents financially; 

Household type: Households where everyone is age 64 or younger; 
Estimate: -0.0028; 
Standard error: 0.0018; 
Lower bound of 95 percent confidence interval: -0.0064; 
Upper bound of 95 percent confidence interval: 0.0008; 
Error over estimate: (65.46%).

Household type: Households where at least one person is 65 or over; 
Estimate: -0.006; 
Standard error: 0.0008; 
Lower bound of 95 percent confidence interval: -0.0074; 
Upper bound of 95 percent confidence interval: -0.0045; 
Error over estimate: (12.81%).

Household type: All households; 
Estimate: -0.0045; 
Standard error: 0.0009; 
Lower bound of 95 percent confidence interval: -0.0064; 
Upper bound of 95 percent confidence interval: -0.0027; 
Error over estimate: (20.76%).

Women: Helped parents with daily activities; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0069; 
Standard error: 0.0019; 
Lower bound of 95 percent confidence interval: 0.0031; 
Upper bound of 95 percent confidence interval: 0.0107; 
Error over estimate: 27.91%.

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0048; 
Standard error: 0.0009; 
Lower bound of 95 percent confidence interval: -0.0065; 
Upper bound of 95 percent confidence interval: -0.0031; 
Error over estimate: (18.32%).

Household type: All households; 
Estimate: 0.0006; 
Standard error: 0.001; 
Lower bound of 95 percent confidence interval: -0.0014; 
Upper bound of 95 percent confidence interval: 0.0025; 
Error over estimate: 177.50%.

Women: Real household assets; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0533; 
Standard error: 0.0058; 
Lower bound of 95 percent confidence interval: 0.0419; 
Upper bound of 95 percent confidence interval: 0.0646; 
Error over estimate: 10.87%.

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0361; 
Standard error: 0.004; 
Lower bound of 95 percent confidence interval: -0.0439; 
Upper bound of 95 percent confidence interval: -0.0283; 
Error over estimate: (11.09%).

Household type: All households; 
Estimate: 0.0041; 
Standard error: 0.0034; 
Lower bound of 95 percent confidence interval: -0.0026; 
Upper bound of 95 percent confidence interval: 0.0107; 
Error over estimate: Household type: 84.14%.

Women: Real household income; 

Household type: Households where everyone is age 64 or younger; 
Estimate: -0.0542; 
Standard error: 0.0051; 
Lower bound of 95 percent confidence interval: -0.0642; 
Upper bound of 95 percent confidence interval: -0.0441; 
Error over estimate: (9.448%).

Household type: Households where at least one person is 65 or over; 
Estimate: -0.054; 
Standard error: 0.0027; 
Lower bound of 95 percent confidence interval: -0.0593; 
Upper bound of 95 percent confidence interval: -0.0487; 
Error over estimate: (4.992%).

Household type: All households; 
Estimate: -0.0541; 
Standard error: 0.0028; 
Lower bound of 95 percent confidence interval: -0.0595; 
Upper bound of 95 percent confidence interval: -0.0487; 
Error over estimate: 5.085%).

Men: Divorced or separated; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0007; 
Standard error: 0.0009; 
Lower bound of 95 percent confidence interval: -0.0011; 
Upper bound of 95 percent confidence interval: 0.0025; 
Error over estimate: 126.10%.

Household type: Households where at least one person is 65 or over; 
Estimate: 0.0009; 
Standard error: 0.0005; 
Lower bound of 95 percent confidence interval: 0; 
Upper bound of 95 percent confidence interval: 0.0019; 
Error over estimate: 52.34%.

Household type: All households; 
Estimate: 0.0008; 
Standard error: 0.0005; 
Lower bound of 95 percent confidence interval: -0.0002; 
Upper bound of 95 percent confidence interval: 0.0019; 
Error over estimate: 62.40%.

Men: Widowhood; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.003; 
Standard error: 0.0005; 
Lower bound of 95 percent confidence interval: 0.002; 
Upper bound of 95 percent confidence interval: 0.004; 
Error over estimate: 17.07%.

Household type: Households where at least one person is 65 or over; 
Estimate: 0.0133; 
Standard error: 0.0006; 
Lower bound of 95 percent confidence interval: 0.0122; 
Upper bound of 95 percent confidence interval: 0.0145; 
Error over estimate: 4.46%.

Household type: All households; 
Estimate: 0.0082; 
Standard error: 0.0004; 
Lower bound of 95 percent confidence interval: 0.0074; 
Upper bound of 95 percent confidence interval: 0.009; 
Error over estimate: 4.79%.

Men: Decline in health; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0171; 
Standard error: 0.0021; 
Lower bound of 95 percent confidence interval: 0.013; 
Upper bound of 95 percent confidence interval: 0.0212; 
Error over estimate: 12.33%.

Household type: Households where at least one person is 65 or over; 
Estimate: 0.0348; 
Standard error: 0.0019; 
Lower bound of 95 percent confidence interval: 0.0311; 
Upper bound of 95 percent confidence interval: 0.0384; 
Error over estimate: 5.39%.

Household type: All households; 
Estimate: 0.026; 
Standard error: 0.0014; 
Lower bound of 95 percent confidence interval: 0.0232; 
Upper bound of 95 percent confidence interval: 0.0288; 
Error over estimate: 5.42%.

Men: Unemployment; 

Household type: Households where everyone is age 64 or younger; 
Estimate: -0.0002; 
Standard error: 0.0011; 
Lower bound of 95 percent confidence interval: -0.0024; 
Upper bound of 95 percent confidence interval: 0.002; 
Error over estimate: (513.2%).

Household type: Households where at least one person is 65 or over; 
Estimate: 0; 
Standard error: 0.0003; 
Lower bound of 95 percent confidence interval: -0.0007; 
Upper bound of 95 percent confidence interval: 0.0006; 
Error over estimate: (4975%).

Household type: All households; 
Estimate: -0.0001; 
Standard error: 0.0006; 
Lower bound of 95 percent confidence interval: -0.0012; 
Upper bound of 95 percent confidence interval: 0.001; 
Error over estimate: (520.0%).

Men: Helped parents financially; 

Household type: Households where everyone is age 64 or younger; 
Estimate: -0.0021; 
Standard error: 0.002; 
Lower bound of 95 percent confidence interval: -0.0061; 
Upper bound of 95 percent confidence interval: 0.0019; 
Error over estimate: (95.72%).

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0068; 
Standard error: 0.001; 
Lower bound of 95 percent confidence interval: -0.0087; 
Upper bound of 95 percent confidence interval: -0.0049; 
Error over estimate: (14.16%).

Household type: All households; 
Estimate: -0.0045; 
Standard error: 0.0011; 
Lower bound of 95 percent confidence interval: -0.0067; 
Upper bound of 95 percent confidence interval: -0.0023; 
Error over estimate: (24.88%).

Men: Helped parents with daily activities; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0061; 
Standard error: 0.002; 
Lower bound of 95 percent confidence interval: 0.0022; 
Upper bound of 95 percent confidence interval: 0.0101; 
Error over estimate: 32.59%.

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0042; 
Standard error: 0.0011; 
Lower bound of 95 percent confidence interval: -0.0063; 
Upper bound of 95 percent confidence interval: -0.0021; 
Error over estimate: (25.43%).

Household type: All households; 
Estimate: 0.0009; 
Standard error: 0.0011; 
Lower bound of 95 percent confidence interval: -0.0013; 
Upper bound of 95 percent confidence interval: 0.0031; 
Error over estimate: 120.90%.

Men: Real household assets; 

Household type: Households where everyone is age 64 or younger; 
Estimate: 0.0566; 
Standard error: 0.0061; 
Lower bound of 95 percent confidence interval: 0.0447; 
Upper bound of 95 percent confidence interval: 0.0684; 
Error over estimate: 10.70%.

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0268; 
Standard error: 0.0041; 
Lower bound of 95 percent confidence interval: -0.0349; 
Upper bound of 95 percent confidence interval: -0.0188; 
Error over estimate: (15.37%).

Household type: All households; 
Estimate: 0.0137; 
Standard error: 0.0036; 
Lower bound of 95 percent confidence interval: 0.0066; 
Upper bound of 95 percent confidence interval: 0.0209; 
Error over estimate: 26.41%.

Men: Real household income; 
Household type: Households where everyone is age 64 or younger; 
Estimate: -0.0486; 
Standard error: 0.0052; 
Lower bound of 95 percent confidence interval: -0.0587; 
Upper bound of 95 percent confidence interval: -0.0385; 
Error over estimate: (10.60%).

Household type: Households where at least one person is 65 or over; 
Estimate: -0.0536; 
Standard error: 0.0031; 
Lower bound of 95 percent confidence interval: -0.0596; 
Upper bound of 95 percent confidence interval: -0.0475; 
Error over estimate: (5.767%).

Household type: All households; 
Estimate: -0.0511; 
Standard error: 0.003; 
Lower bound of 95 percent confidence interval: -0.057; 
Upper bound of 95 percent confidence interval: -0.0453; 
Error over estimate: (5.835%).

Source: GAO analysis of HRS data. 

[End of table] 

Estimating the Effects of Events Occurring Later in Life on Assets and 
Income:

In order to examine whether the effects of certain events occurring 
later in life differ by gender, we used fixed-effects regression 
models. For example, we estimated how changes in health lead to 
changes in household assets and income. Researchers use the fixed-
effects method because much of the differences in income and wealth 
between households are consistent over time (as poorer households tend 
to stay poor and richer households tend to stay rich). The fixed-
effects method sweeps away these "time invariant" differences, thus 
better isolating the effect of health or other life events from other 
aspects of households that could explain differences.10[Footnote 68]

Specifically, we estimated variations of the following equation, 
separately by gender:

(1) Log (Household Assets or Income) = ai + at + *(poor health) + 
c*(marital status) + d*(other control variables):

Where, ai and at indicate fixed effects for the individual and wave. 
is the effect of poor health and d and c are the effect of other 
control variables and marital status.11B[Footnote 69]y including a 
dummy variable for each wave, we attempted to control for all national-
level changes that could have affected assets and income, and also 
have been associated with the life events. Therefore, can be 
interpreted as the effect of poor health, measured as the percent 
difference in average assets between periods where an individual 
reports poor versus not-poor health. Due to the additional controls, 
this average percent difference is measured relative to the changes 
over time, and also relative to the other time-variant measures 
captured, such as changes in marital status.[Footnote 70] 

However, while some of the life-events are likely associated with the 
passage of time, the regression does not assume that relationship. For 
example, if an individual switches from poor health to good health, 
the fixed-effects regression will also use those transitions to 
estimate the size of the effect. Similarly, the fixed- effects 
regression will also use transitions from married to widowed, as well 
as widowed to married, to estimate the effect of widowhood.

As is common among all regressions, a limitation of the fixed-effects 
method is that some important variable could be omitted from the 
model. While the fixed effect controls for all time-invariant 
attributes, there is still the possibility of endogenous 
relationships. For example, if an individual's health declined because 
income fell, and not the other way around, that bias could affect our 
findings.

Some of the life events we examined were likely correlated with 
changes in household structure, such as changes in marital status. 
However, if the income of a household falls when an individual leaves, 
the remaining individuals may not be worse off when it comes to 
resources because the household now requires fewer resources to meet 
its needs. To address this, we adjusted the estimated effects by 
household size; the household's income and assets were scaled by the 
square root of the individuals in the household. The rationale for 
using the square root is because the effect of reducing members is 
diminishing (changing from 1 to 2 has a larger effect than going from 
9 to 10). In addition, this analysis estimated the effect of an 
individual's life event on household assets or income. We did not 
attempt to determine to what extent a spouse's life event (for married 
individuals) may have affected household assets or income).

Divorce:

Table 19 contains the effects of the first event we analyzed: divorce. 
We analyzed the effect of divorce on household assets and income, both 
with and without controlling for the number of people in the 
household. Across almost all the groups and specifications, the effect 
of divorce is to reduce assets and income, with larger effects for 
women than for men. Adjusting for household size tended to reduce the 
magnitude of the effects.

* Effect on assets. Divorce tended to reduce assets for more women 
than men, with comparable sizes of effects for women and men. For 
example, among all households, the decline in assets associated with 
divorce was 41 percent for women and 39 percent for men. When the size 
of the household was adjusted for, the size of the effect declined, 
but was still statistically significant.

* Effect on income. Divorce reduced income for both women and men, 
with larger effects for women than men. For example, among all 
households, the decline in income associated with divorce was 41 
percent for women and 23 percent for men. When household size was 
adjusted for, the size of the effects were much smaller in magnitude.

Table 19: Divorce Effect on Household Assets and Income:

Effect on assets: 

Log point change; 
All households: Women: -0.53; 
All households: Men: -0.50; 
Households where everyone is age 64 or younger: Women: -0.54; 
Households where everyone is age 64 or younger: Men: -0.50; 
Households where at least one person is 65 or over: Women: -0.39; 
Households where at least one person is 65 or over: Men: -0.38.

Standard error; 
All households: Women: (0.022); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.03); 
Households where at least one person is 65 or over: Women: (0.04); 
Households where at least one person is 65 or over: Men: (0.040).

Percent change; 
All households: Women: -41%; 
All households: Men: -39%; 
Households where everyone is age 64 or younger: Women: -41%; 
Households where everyone is age 64 or younger: Men: -39%; 
Households where at least one person is 65 or over: Women: -32%; 
Households where at least one person is 65 or over: Men: -32%.

Effects on assets per household member: 

Log point change; 
All households: Women: -0.37; 
All households: Men: -0.32; 
Households where everyone is age 64 or younger: Women: -0.41; 
Households where everyone is age 64 or younger: Men: -0.32; 
Households where at least one person is 65 or over: Women: -0.18; 
Households where at least one person is 65 or over: Men: -0.24.

Standard error; 
All households: Women: (0.022); 
All households: Men: (0.022); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.03); 
Households where at least one person is 65 or over: Women: (0.04); 
Households where at least one person is 65 or over: Men: (0.04).

Percent change; 
All households: Women: -31%; 
All households: Men: -28%; 
Households where everyone is age 64 or younger: Women: -33%; 
Households where everyone is age 64 or younger: Men: -28%; 
Households where at least one person is 65 or over: Women: -17%; 
Households where at least one person is 65 or over: Men: -21%.

Effect on income: 

Log point change; 
All households: Women: -0.52; 
All households: Men: -0.26; 
Households where everyone is age 64 or younger: Women: -0.58; 
Households where everyone is age 64 or younger: Men: -0.29; 
Households where at least one person is 65 or over: Women: -0.49; 
Households where at least one person is 65 or over: Men: -0.27.

Standard error; 
All households: Women: (0.02); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.025); 
Households where at least one person is 65 or over: Women: (0.02); 
Households where at least one person is 65 or over: Men: (0.025).

Percent change; 
All households: Women: -41%; 
All households: Men: -23%; 
Households where everyone is age 64 or younger: Women: -44%; 
Households where everyone is age 64 or younger: Men: -25%; 
Households where at least one person is 65 or over: Women: -39%; 
Households where at least one person is 65 or over: Men: -23%.

Effect on income per household member: 

Log point change; 
All households: Women: -0.37; 
All households: Men: -0.09; 
Households where everyone is age 64 or younger: Women: -0.46; 
Households where everyone is age 64 or younger: Men: -0.12; 
Households where at least one person is 65 or over: Women: -0.29; 
Households where at least one person is 65 or over: Men: -0.13.

Standard error; 
All households: Women: (0.02); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.03); 
Households where at least one person is 65 or over: Women: (0.020); 
Households where at least one person is 65 or over: Men: (0.025).

Percent change; 
All households: Women: -31%; 
All households: Men: -9%; 
Households where everyone is age 64 or younger: Women: -37%; 
Households where everyone is age 64 or younger: Men: -11%; 
Households where at least one person is 65 or over: Women: -25%; 
Households where at least one person is 65 or over: Men: -12%.

Source: GAO analysis of HRS data. 

[End of table] 

Widowhood:

Table 20 contains the results for widowhood. As with divorce, we 
analyzed the effect of widowhood on household assets and income, both 
with and without controlling for the number of people in the 
household. Across almost all the groups and specifications, the effect 
of widowhood is to reduce assets and income, with larger effects for 
women than for men. Adjusting for household size tended to reduce the 
magnitude of the effects.

* Effect on assets. Widowhood reduced assets for both women and men, 
with larger effects for women than men. For example, among all 
households, the decline in assets associated with widowhood was 32 
percent for women and 27 percent for men. However, part of this effect 
seems to be associated with the size of the household. Among the 
households in which at least one member was 65 and over, the decline 
in assets was not significant when household size was adjusted for.

* Effect on income. Widowhood reduced income for both women and men, 
with larger effects for women than men. For example, among all 
households, the decline in income associated with widowhood was 37 
percent for women and 22 percent for men. Again, part of this effect 
seems to be associated with the size of the household. When household 
size was adjusted for, the size of the effects were much smaller in 
magnitude.

Table 20: Widowhood Effect on Household Assets and Income:

Effect on assets: 

Log point change; 
All households: Women: -0.39; 
All households: Men: - 0.31; 
Households where everyone is age 64 or younger: Women: - 0.37; 
Households where everyone is age 64 or younger: Men: -0.26; 
Households where at least one person is 65 or over: Women: - 0.26; 
Households where at least one person is 65 or over: Men: -0.20.

Standard error; 
All households: Women: (0.01); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.034); 
Households where everyone is age 64 or younger: Men: (0.051); 
Households where at least one person is 65 or over: Women: (0.02); 
Households where at least one person is 65 or over: Men: (0.02).

Percent change; 
All households: Women: -32%; 
All households: Men: -27%; 
Households where everyone is age 64 or younger: Women: -31%; 
Households where everyone is age 64 or younger: Men: -23%; 
Households where at least one person is 65 or over: Women: -23%; 
Households where at least one person is 65 or over: Men: -18%.

Effect on assets per household member: 

Log point change; 
All households: Women: -0.19; 
All households: Men: - 0.11; 
Households where everyone is age 64 or younger: Women: - 0.20; 
Households where everyone is age 64 or younger: Men: -0 .03; 
Households where at least one person is 65 or over: Women: - 0.02; 
Households where at least one person is 65 or over: Men: -0.00.

Standard error; 
All households: Women: (0.01); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.05); 
Households where at least one person is 65 or over: Women: (0.02); 
Households where at least one person is 65 or over: Men: (0.02).

Percent change; 
All households: Women: -17%; 
All households: Men: -10%; 
Households where everyone is age 64 or younger: Women: -18%; 
Households where everyone is age 64 or younger: Men: -3%; 
Households where at least one person is 65 or over: Women: -2%; 
Households where at least one person is 65 or over: Men: - .3%.

Effect on income: 

Log point change; 
All households: Women: -0.46; 
All households: Men: - 0.25; 
Households where everyone is age 64 or younger: Women: - 0.63; 
Households where everyone is age 64 or younger: Men: -0.36; 
Households where at least one person is 65 or over: Women: - 0.43; 
Households where at least one person is 65 or over: Men: -0.23.

Standard error; 
All households: Women: (0.01); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.04); 
Households where at least one person is 65 or over: Women: (0.01); 
Households where at least one person is 65 or over: Men: (0.01).

Percent change; 
All households: Women: -37%; 
All households: Men: -22%; 
Households where everyone is age 64 or younger: Women: -47%; 
Households where everyone is age 64 or younger: Men: -30%; 
Households where at least one person is 65 or over: Women: -35%; 
Households where at least one person is 65 or over: Men: -21%.

Effect on income per household member: 

Log point change; 
All households: Women: -0.27; 
All households: Men: - 0.06; 
Households where everyone is age 64 or younger: Women: - 0.48; 
Households where everyone is age 64 or younger: Men: -0.17; 
Households where at least one person is 65 or over: Women: - 0.20; 
Households where at least one person is 65 or over: Men: - 0.04.

Standard error; 
All households: Women: (0.01); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.04); 
Households where at least one person is 65 or over: Women: (0.01); 
Households where at least one person is 65 or over: Men: (0.02).

Percent change; 
All households: Women: -23%; 
All households: Men: -6%; 
Households where everyone is age 64 or younger: Women: -38%; 
Households where everyone is age 64 or younger: Men: -16%; 
Households where at least one person is 65 or over: Women: -18%; 
Households where at least one person is 65 or over: Men: -4%.

Source: GAO analysis of HRS data. 

[End of table] 

Unemployment:

As shown in table 21, unemployment tended to reduce assets and income, 
with comparable effects for women and men. The effects did not seem to 
dissipate when household size was adjusted for.

* Effect on assets. Unemployment reduced assets for both women and 
men, with comparable effects for women and men. For example, among all 
households, the decline in assets associated with unemployment was 7 
percent for women and 7 percent for men. An exception to this 
difference was in cases in which at least one member was 65 or over. 
For those individuals, the decline in household assets was only 2 
percent for women and 15 percent for men.

* Effect on income. Unemployment reduced income for both women and 
men, with comparable effects for women and men. For example, among all 
households, the decline in income associated with unemployment was 6 
percent for women and 8 percent for men.

Table 21: Unemployment Effect:

Effect on assets: 

Log point change; 
All households: Women: -0.07; 
All households: Men: - 0.07; 
Households where everyone is age 64 or younger: Women: - 0.09; 
Households where everyone is age 64 or younger: Men: -0.07; 
Households where at least one person is 65 or over: Women: - 0.02; 
Households where at least one person is 65 or over: Men: -0.15.

Standard error; 
All households: Women: (0.02); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.03); 
Households where at least one person is 65 or over: Women: (0.07); 
Households where at least one person is 65 or over: Men: (0.075).

Percent change; 
All households: Women: -7%; 
All households: Men: -7%; 
Households where everyone is age 64 or younger: Women: -9%; 
Households where everyone is age 64 or younger: Men: -7%; 
Households where at least one person is 65 or over: Women: -2%; 
Households where at least one person is 65 or over: Men: -14%.

Effects on assets per household member: 

Log point change; 
All households: Women: -0.06; 
All households: Men: -0.08; 
Households where everyone is age 64 or younger: Women: -0.08; 
Households where everyone is age 64 or younger: Men: -0.08; 
Households where at least one person is 65 or over: Women: -0.03; 
Households where at least one person is 65 or over: Men: -0 .16.

Standard error; 
All households: Women: (0.02); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.03); 
Households where everyone is age 64 or younger: Men: (0.03); 
Households where at least one person is 65 or over: Women: (0.07); 
Households where at least one person is 65 or over: Men: (0.076).

Percent change; 
All households: Women: -6%; 
All households: Men: -8%; 
Households where everyone is age 64 or younger: Women: -8%; 
Households where everyone is age 64 or younger: Men: -8%; 
Households where at least one person is 65 or over: Women: -3%; 
Households where at least one person is 65 or over: Men: -15%.

Effects on income: 

Log point change; 
All households: Women: -0.09; 
All households: Men: - 0.07; 
Households where everyone is age 64 or younger: Women: - 0.10; 
Households where everyone is age 64 or younger: Men: -0.06; 
Households where at least one person is 65 or over: Women: - 0.13; 
Households where at least one person is 65 or over: Men: -0.12.

Standard error; 
All households: Women: (0.02); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.02); 
Households where everyone is age 64 or younger: Men: (0.02); 
Households where at least one person is 65 or over: Women: (0.04); 
Households where at least one person is 65 or over: Men: (0.05).

Percent change; 
All households: Women: -9%; 
All households: Men: -7%; 
Households where everyone is age 64 or younger: Women: -9%; 
Households where everyone is age 64 or younger: Men: -6%; 
Households where at least one person is 65 or over: Women: -13%; 
Households where at least one person is 65 or over: Men: -12%.

Effects on income per household member: 

Log point change; 
All households: Women: -0.09; 
All households: Men: - 0.08; 
Households where everyone is age 64 or younger: Women: - 0.09; 
Households where everyone is age 64 or younger: Men: -0.07; 
Households where at least one person is 65 or over: Women: - 0.14; 
Households where at least one person is 65 or over: Men: -0.13.

Standard error; 
All households: Women: (0.02); 
All households: Men: (0.02); 
Households where everyone is age 64 or younger: Women: (0.02); 
Households where everyone is age 64 or younger: Men: (0.02); 
Households where at least one person is 65 or over: Women: (0.04); 
Households where at least one person is 65 or over: Men: (0.05).

Percent change; 
All households: Women: -8%; 
All households: Men: -7%; 
Households where everyone is age 64 or younger: Women: -8%; 
Households where everyone is age 64 or younger: Men: -7%; 
Households where at least one person is 65 or over: Women: -13%; 
Households where at least one person is 65 or over: Men: -12%.

Source: GAO analysis of HRS data. 

[End of table] 

A Decline in Health:

In general, across the specifications, the effect of a decline into 
poor health tended to reduce assets and income, with comparable 
effects for women and men (see table 22). One notable difference 
however, were the larger estimated effects of men's poor health on 
assets, but only in the case where both members of the household were 
less than 65 years of age. Specifically, we found that for individuals 
living in these households, poor health in men was associated with a 
drop in household assets of 13 percent, but 5 percent for women. 
[Footnote 71]

In general, the magnitude of the effect on assets was in the 10 
percent range for both women and men, and is statistically 
significant. The effects on income are about half that magnitude, but 
follow the same direction as the effects on assets. There is little 
difference in the effects when the level of assets and income are 
estimated with a correction for the size of the household.

Table 22: A Decline in Health's Effect on Household Assets and Income:

Effect on assets: 

Log point change; 
All households: Women: -0.09; 
All households: Men: - 0.10; 
Households where everyone is age 64 or younger: Women: - 0.05; 
Households where everyone is age 64 or younger: Men: -0.14; 
Households where at least one person is 65 or over: Women: - 0.06; 
Households where at least one person is 65 or over: Men: -0.04.

Standard error; 
All households: Women: (0.008); 
All households: Men: (0.008); 
Households where everyone is age 64 or younger: Women: (0.01); 
Households where everyone is age 64 or younger: Men: (0.01); 
Households where at least one person is 65 or over: Women: (0.01); 
Households where at least one person is 65 or over: Men: (0.01).

Percent change; 
All households: Women: -8%; 
All households: Men: -10%; 
Households where everyone is age 64 or younger: Women: -5%; 
Households where everyone is age 64 or younger: Men: -13%; 
Households where at least one person is 65 or over: Women: -6%; 
Households where at least one person is 65 or over: Men: -4%.

Effects on assets per household member: 

Log point change; 
All households: Women: -0.09; 
All households: Men: -0.11; 
Households where everyone is age 64 or younger: Women: - 0.06; 
Households where everyone is age 64 or younger: Men: -0.14; 
Households where at least one person is 65 or over: Women: - 0.06; 
Households where at least one person is 65 or over: Men: -0.05.

Standard error; 
All households: Women: (0.008); 
All households: Men: (0.008); 
Households where everyone is age 64 or younger: Women: (0.01); 
Households where everyone is age 64 or younger: Men: (0.01); 
Households where at least one person is 65 or over: Women: (0.01); 
Households where at least one person is 65 or over: Men: (0.01).

Percent change; 
All households: Women: -9%; 
All households: Men: -10%; 
Households where everyone is age 64 or younger: Women: -5%; 
Households where everyone is age 64 or younger: Men: -13%; 
Households where at least one person is 65 or over: Women: -6%; 
Households where at least one person is 65 or over: Men: -5%.

Effect on income: 

Log point change; 
All households: Women: -0.04; 
All households: Men: - 0.03; 
Households where everyone is age 64 or younger: Women: - 0.05; 
Households where everyone is age 64 or younger: Men: -0.03; 
Households where at least one person is 65 or over: Women: - 0.03; 
Households where at least one person is 65 or over: Men: -0.02.

Standard error; 
All households: Women: (0.006); 
All households: Men: (0.006); 
Households where everyone is age 64 or younger: Women: (0.01); 
Households where everyone is age 64 or younger: Men: (0.01); 
Households where at least one person is 65 or over: Women: (0.01); 
Households where at least one person is 65 or over: Men: (0.01).

Percent change; 
All households: Women: -4%; 
All households: Men: -3%; 
Households where everyone is age 64 or younger: Women: -5%; 
Households where everyone is age 64 or younger: Men: -3%; 
Households where at least one person is 65 or over: Women: -3%; 
Households where at least one person is 65 or over: Men: -2%.

Effect on income per household member: 

Log point change; 
All households: Women: -0.05; 
All households: Men: -0.04; 
Households where everyone is age 64 or younger: Women: - 0.05; 
Households where everyone is age 64 or younger: Men: -0.03; 
Households where at least one person is 65 or over: Women: - 0.03; 
Households where at least one person is 65 or over: Men: -0.02.

Standard error; 
All households: Women: (0.006); 
All households: Men: (0.006); 
Households where everyone is age 64 or younger: Women: (0.01); 
Households where everyone is age 64 or younger: Men: (0.01); 
Households where at least one person is 65 or over: Women: (0.01); 
Households where at least one person is 65 or over: Men: (0.01).

Percent change; 
All households: Women: -5%; 
All households: Men: -4%; 
Households where everyone is age 64 or younger: Women: -5%; 
Households where everyone is age 64 or younger: Men: -3%; 
Households where at least one person is 65 or over: Women: -3%; 
Households where at least one person is 65 or over: Men: -2%.

Source: GAO analysis of HRS data. 

[End of table] 

Helping Parents Financially or with Daily Activities:

As shown in table 23, the results for either helping parents 
financially or with basic daily activities--eating, dressing, and 
bathing--were not as consistently significantly negative as the other 
life events. In the fixed-effects regression, the effect of personal 
assistance did not appear to be statistically significant, while the 
effect of financial assistance tended to be significantly positive. It 
may be that when households have more assets or income they are more 
likely to provide assistance--which could explain these findings. 
There is little difference in the effects when the level of assets and 
income are estimated with a correction for the size of the household. 
To further understand these relationships, we explored the 
characteristics of those helping their parents with the basic daily 
activities of bathing, dressing, and eating. We found that only 2 
percent of the sample provided both financial help and help with basic 
daily activities. Further, those in the labor force (i.e., working or 
unemployed and looking for work) were more likely to help their 
parents with basic daily activities than those retired or not in the 
labor force.

Table 23: Effects of Providing Financial Assistance or Physical Care 
on Household Assets and Income:

Effect on assets: 

Log point change; 
Helped parents financially: Women: 0.028; 
Helped parents financially: Men: 0.034; 
Helped parents with basic daily activities: Women: 0.0; 
Helped parents with basic daily activities: Men: 0.01.

Standard error; 
Helped parents financially: Women: (0.01); 
Helped parents financially: Men: (0.01); 
Helped parents with basic daily activities: Women: (0.01); 
Helped parents with basic daily activities: Men: (0.01).

Percent change; 
Helped parents financially: Women: 3%; 
Helped parents financially: Men: 3%; 
Helped parents with basic daily activities: Women: 1%; 
Helped parents with basic daily activities: Men: 1%.

Effects on assets per household member: 

Log point change; 
Helped parents financially: Women: 0.032; 
Helped parents financially: Men: 0.038; 
Helped parents with basic daily activities: Women: 0.004; 
Helped parents with basic daily activities: Men: 0.01.

Standard error; 
Helped parents financially: Women: (0.01); 
Helped parents financially: Men: (0.01); 
Helped parents with basic daily activities: Women: (0.01); 
Helped parents with basic daily activities: Men: (0.02).

Percent change; 
Helped parents financially: Women: 3%; 
Helped parents financially: Men: 4%; 
Helped parents with basic daily activities: Women: 0.4%; 
Helped parents with basic daily activities: Men: 1%.

Effect on income; 
Helped parents financially: Women: [Empty]; 
Helped parents financially: Men: [Empty]; 
Helped parents with basic daily activities: Women: [Empty]; 
Helped parents with basic daily activities: Men: [Empty].

Log point change; 
Helped parents financially: Women: 0.056; 
Helped parents financially: Men: 0.071; 
Helped parents with basic daily activities: Women: 0.016; 
Helped parents with basic daily activities: Men: 0.020.

Standard error; 
Helped parents financially: Women: (0.008); 
Helped parents financially: Men: (0.008); 
Helped parents with basic daily activities: Women: (0.008); 
Helped parents with basic daily activities: Men: (0.008).

Percent change; 
Helped parents financially: Women: 6%; 
Helped parents financially: Men: 7%; 
Helped parents with basic daily activities: Women: 2%; 
Helped parents with basic daily activities: Men: 2%.

Effect on income per household member: 

Log point change; 
Helped parents financially: Women: 0.059; 
Helped parents financially: Men: 0.073; 
Helped parents with basic daily activities: Women: 0.013; 
Helped parents with basic daily activities: Men: 0.018.

Standard error; 
Helped parents financially: Women: (0.008); 
Helped parents financially: Men: (0.008); 
Helped parents with basic daily activities: Women: (0.008); 
Helped parents with basic daily activities: Men: (0.008).

Percent change; 
Helped parents financially: Women: 6%; 
Helped parents financially: Men: 8%; 
Helped parents with basic daily activities: Women: 1%; 
Helped parents with basic daily activities: Men: 2%.

Source: GAO analysis of HRS data. 

[End of table] 

[End of section] 

Appendix II: GAO Contact and Staff Acknowledgments:

GAO Contact:

Charles Jeszeck, (202) 512-7215 or jeszeckc@gao.gov:

Staff Acknowledgments:

Michael Collins, Assistant Director; Erin M. Godtland, Senior 
Economist, and Jennifer Gregory, Senior Analyst, led the engagement. 
In addition, James Bennett, Benjamin Bolitzer, David Chrisinger, 
Cynthia Grant, Jean Lee, Grant Mallie, Ashley McCall, Michael Morris, 
Rhiannon Patterson, Mark Ramage, James Rebbe, Douglas Sloane, Jeff 
Tessin, Shana Wallace, and Erin White made valuable contributions.

[End of section] 

Footnotes: 

[1] Generally, Social Security retirement benefits are based on up to 
35 years of a worker's indexed earnings. Average lower earnings over a 
lifetime and fewer years in the workforce lead to significantly lower 
average benefit amounts for women as compared to men. In 2009, the 
average annual Social Security income received by retired women was 
$12,155 compared to $15,620 for men, according to one study. See 
Carroll L. Estes, Terry O'Neill and Heidi Hartmann, Breaking the 
Social Security Glass Ceiling: A Proposal to Modernize Women's 
Benefits, Institute for Women's Policy Research, National Committee to 
Preserve Social Security & Medicare Foundation, and National 
Organization for Women Foundation (May 2012).

[2] This is due to at least two factors: women have longer life 
expectancies, and in marriages the husband is, on average, older than 
the wife by 3 years.

[3] In 2008, about 69 percent of single women 65 and over living alone 
would have been living in poverty if it were not for Social Security 
benefits they received, according to a study published by the 
Institute for Women's Policy Research. See Jeff Hayes, Heidi Hartmann, 
and Sunhwa Lee, Social Security: Vital to Retirement Security for 35 
Million Women and Men, Institute for Women's Policy Research Briefing 
Paper, IWPR Publication #D487 (March 2010).

[4] This report builds upon our past work for this committee. See GAO, 
Retirement Security: Women Face Challenges in Ensuring Financial 
Security in Retirement, GAO-08-105 (Washington, D.C.: Oct. 11, 2007).

[5] Specifically, we used data from the 1996, 2001, 2004, and 2008 
SIPP panel surveys, which are administered over a period of several 
years.

[6] Data on income were available through 2010, while data on 
retirement plan coverage and participation were only available through 
2009.

[7] Specifically, we used a statistical technique called "fixed-
effects regression." Because the HRS tracks individuals over time, we 
were able to estimate the percentage change in household income and 
assets that occurs for an individual after a life event, relative to 
an individual that did not experience that life event, but also became 
older. In this way, we were able to isolate the effect of the life 
event from other factors. We used all available years of HRS data, 
including early release data for 2010. For more information on 
methods, see appendix I. 


[8] To ensure that we obtained a balanced perspective, we interviewed 
experts with a range of viewpoints and from different types of 
organizations including government, academia, advocacy groups, and the 
private sector. For a list of organizations, see appendix I.

[9] Linda A. Jacobsen, Mary Kent, Marlene Lee, and Mark Mather, 
"America's Aging Population," Population Bulletin, Population 
Reference Bureau, vol. 66, no. 1 (2011).

[10] Ibid.

[11] Ibid. 

[12] The baby boom generation consists of individuals born from 1946 
to 1964.

[13] Carmen DeNavas-Walt, Bernadette D. Proctor, and Jessica C. Smith, 
"Income, Poverty, and Health Insurance Coverage in the United States: 
2010" Current Population Reports, Consumer Income, United States 
Census Bureau, P60-239 (September 2011).

[14] Cindy Hounsell, The Female Factor 2008: Why Women Are at Greater 
Financial Risk in Retirement and How Annuities Can Help (Washington, 
D.C.: Americans for Secure Retirement, 2008).

[15] Ibid.

[16] For example, wives may be eligible to receive a spousal benefit 
equal to 50 percent of their husbands' benefits. If a wife receiving 
this benefit becomes widowed, then the benefit becomes a survivor 
benefit, and may equal up to 100 percent of the husband's benefit. For 
more information on how the different types of benefits are 
calculated, see GAO, Social Security Reform: Issues for Disability and 
Dependent Benefits, GAO-08-26 (Washington, D.C.: Oct. 26, 2007).

[17] The Board of Trustees, Federal Old-Age and Survivors Insurance 
and Federal Disability Insurance Trust Funds, The 2012 Annual Report 
of the Board of Trustees of the Federal Old-Age and Survivors 
Insurance and Federal Disability Insurance Trust Funds (Washington, 
D.C.: Apr. 25, 2012).

[18] The most common type of DC plans are 401(k) plans, which 
typically allow workers to choose to contribute a portion of their 
pretax compensation to the plan. 

[19] The tax treatment differs depending on the type of IRA. For 
example, with traditional IRAs, workers who meet certain conditions 
can take an income tax deduction on some or all of the contributions 
they make to their IRA, but they must pay taxes on amounts they 
withdraw from the IRA. Workers below certain income limits may also 
contribute to Roth IRAs, which do not provide an income tax deduction 
on contributions, but permit tax-free withdrawals after a specified 
time period.

[20] Rolling funds over to an IRA allows participants to preserve the 
tax benefits enjoyed by the plan. 

[21] The qualified joint and survivor annuity must provide at least a 
50 percent benefit continuation to the surviving spouse.

[22] GAO, Retirement Savings: Automatic Enrollment Shows Promise for 
Some Workers, but Proposals to Broaden Retirement Savings for Other 
Workers Could Face Challenges, GAO-10-31 (Washington, D.C.: Oct. 23, 
2009).

[23] GAO, Retirement Income: Ensuring Income throughout Retirement 
Requires Difficult Choices, GAO-11-400 (Washington, D.C.: June 7, 
2011). 

[24] DC participants can purchase annuities on the retail market. 
However, retail annuities are typically more expensive for women than 
for men because of women's longer life expectancy, whereas in-plan 
annuity options, when they are offered, must be at gender-neutral 
rates. In addition, in-plan rates may be lower than retail rates 
because the in-plan rate may be able to take advantage of a lower, 
institutional price. Nonetheless, research shows that most people 
choose not to annuitize DC savings.

[25] These recession periods were identified by the National Bureau of 
Economic Research Business Cycle Dating Committee. 

[26] GAO, Private Pensions: Some Key Features Lead to an Uneven 
Distribution of Benefits, GAO-11-333 (Washington, D.C.: Mar. 30, 2011).

[27] See GAO, Unemployed Older Workers: Many Experience Challenges 
Regaining Employment and Face Reduced Retirement Security, GAO-12-445 
(Washington, D.C.: Apr. 25, 2012).

[28] The statistics we present in this section are unadjusted point 
estimates computed from the SIPP data without taking into account 
differences between men and women in demographic and occupational 
characteristics. To adjust these point estimates by taking into 
account different factors that might explain gender differences in 
these three outcomes--working for an employer that offers a plan, plan 
eligibility, and participation--we also conducted multivariate 
analysis. The detailed results of these analyses are presented in 
appendix I. 

[29] For more information on women's and men's occupations and 
industries and other factors associated with working for an employer 
that offers a plan, see appendix I.

[30] In prior work, GAO reported that net new plan formation between 
2003 and 2007 had been very small (about 1 percent) and that new plan 
formation was offset by plan terminations or mergers. In addition, 92 
percent of newly-formed plans were DC plans and were generally small, 
with 96 percent having fewer than 100 participants. See GAO, Private 
Pensions: Some Key Features Lead to an Uneven Distribution of 
Benefits, GAO-11-333 (Washington, D.C.: Mar. 30, 2011). GAO also 
reported that plan sponsors voluntarily terminated over 61,000 
sufficiently funded single-employer DB plans from 1990 to 2006 and 
that a number of large employers, representing a significant portion 
of participants in the DB pension system, had announced their 
intention to freeze one or more of their DB plans. See GAO, Defined 
Benefit Pensions: Plan Freezes Affect Millions of Participants and May 
Pose Retirement Income Challenges, GAO-08-817 (Washington, D.C.: July 
21, 2008).

[31] Individuals in the White, Black, and Asian racial and ethnic 
categories are non-Hispanic.

[32] For more information on other factors associated with employer- 
plan eligibility, see appendix I.

[33] See Alicia H. Munnell, Annika Sundén, and Catherine Taylor, "What 
Determines 401(k) Participation and Contributions?" Social Security 
Bulletin, vol. 64, no. 3 (2001/2002). See appendix I for additional 
information on our modeling analyses. 

[34] [4] See Lois Shaw and Catherine Hill, The Gender Gap in Pension 
Coverage: What Does the Future Hold?, Institute for Women's Policy 
Research, IWPR Publication #E507 (May 15, 2001). Shaw and Hill find 
that hours worked per week and job tenure are positively related with 
participating in a pension plan. 

[35] These results are consistent with those of outside researchers. 
For example, one study found that before controlling for differences 
between men and women in other factors that might affect 
participation, women had significantly lower participation rates than 
men. However, after controlling for differences between men and women, 
women were 6.5 percent more likely to participate in a DC plan. See 
Gur Huberman, Sheena S. Iyengar, and Wei Jiang, "Defined Contribution 
Pension Plans: Determinants of Participation and Contributions Rates," 
Journal of Financial Services Research (January 2007). For more 
information on other factors associated with employer-plan 
participation, see appendix I.

[36] These estimates of contribution levels are consistent with 
estimates (for both men and women combined) from other studies using 
recent SIPP data. See, for example, "Retirement Plan Participation: 
Survey of Income and Program Participation (SIPP) Data, 2009" Employee 
Benefit Research Institute Notes, vol. 31, no.11 (November 2010): 2.

[37] For the analysis in this section, we used SIPP data from 1998, 
2003, 2006, and 2010. See appendix I for more details on the data and 
our analyses.

[38] This may be due to the fact that retirees tend to withdraw funds 
from DC accounts irregularly, instead of annuitizing. To the extent 
that nonregular (lump sum) distributions comprise a significant 
portion of income, our estimates of income shares from other sources, 
such as Social Security, might be overstated. However, because of the 
irregularity of these lump sum distributions, it is difficult to 
observe them with household survey data because surveys generally 
measure income only at a particular point in time.

[39] In "Racial and Ethnic Differences in Women's Retirement 
Security," Journal of Women, Politics & Policy, 30 (2009): 141-171, 
Sunhwa Lee also notes that Social Security is the most common source 
of retirement income and that differences in immigrant status do not 
entirely account for the lower rates of Social Security receipt among 
Hispanics and Asians. Maya Rockeymoore and Meizhu Lui highlight that 
Hispanics are disproportionately represented in job categories that 
were previously excluded from the Social Security program, such as 
agricultural and household workers. They point out that, although 
these job categories are now covered, earnings in these categories 
might not be recorded accurately in Social Security tax payment 
records, which could lead to lower payments and therefore a lower 
share of income from Social Security. See Maya M. Rockeymoore and 
Meizhu Lui, Plan for a New Future: The Impact of Social Security 
Reform on People of Color (Washington, D.C.: Commission to Modernize 
Social Security, 2011).

[40] We used SIPP data to analyze household income among individuals 
65 and over from 1998 to 2010.

[41] Lee (2009) and Rockeymoore, et al. (2011) find similar results 
regarding higher poverty rates among unmarried and non-White women. 

[42] We estimated these effects using fixed-effects panel regressions. 
We used all available years of HRS data, from 1992 up through the 
early release data for 2010. Because the HRS tracks individuals over 
time, we were able to estimate the percent change in household assets 
and household income that occurs for an individual after a life event, 
relative to an individual that did not experience that life event, but 
also became older. In this way, we were able to isolate the effect of 
the life event from other factors. We analyzed the effect of each 
event individually. If a woman were to experience multiple events 
simultaneously, such as becoming divorced and unemployed, the effects 
on her household assets and income could be even larger. For more 
details on our methodology and results, including standard errors, see 
appendix I.

[43] For our analysis, we used a user-friendly longitudinal dataset 
created by RAND, a research organization. For total household assets, 
we used RAND's variable that includes all household assets except for 
secondary residences. For income, we used RAND's total household 
income variable. For more information on the RAND dataset, see 
appendix I.

[44] Respondents in our sample were born prior to 1954; the HRS 
grouped these individuals into five generational cohorts. In addition, 
these analyses did not adjust for the size of the household, but show 
the effect on total household income and assets for a person 
experiencing the event. When we adjusted our models for household 
size, we found smaller effects for divorce and widowhood, but these 
effects were still significant. See appendix I for more information.

[45] Our estimated effects represent the average percent difference in 
household assets and income between all survey periods in which the 
household does experience an event and all survey periods in which the 
household does not experience an event.

[46] Researchers have hypothesized that the drop in assets is due to 
households saving their assets for a rainy day and are primarily drawn 
down at the time of precipitating shocks, such as divorce. See James 
M. Poterba, Steven F. Venti, and David A. Wise, Family Status 
Transitions, Latent Health, and the Post-Retirement Evolution of 
Assets, NBER Working Paper 15789, issued in February 2010. Also, 
Wilmoth and Koso hypothesize that the mechanisms that systematically 
allocate wealth when a marriage ends are more effective at maintaining 
wealth for those who are widowed compared to those who are divorced. 
They conclude that divorce should be more detrimental to long-term 
wealth accumulation than widowhood. See Janet Wilmoth and Gregor Koso, 
"Does Marital History Matter? Marital Status and Wealth Outcomes Among 
Preretirement Adults," Journal of Marriage and Family, vol. 64, no. 1 
(2002).

[47] Further, some of these women and men could have been divorced 
prior to entering our sample.

[48] A widow's assets may be depleted by medical and other expenses 
incurred prior to the death of her spouse. See Kathleen McGarry and 
Robert F. Schoeni, "Medicare Gaps and Widow Poverty," Social Security 
Bulletin, vol. 66, no. 1 (2005). In addition, women's income may fall 
after widowhood if their husbands did not elect to take the husband's 
pension benefits in the form of a joint and survivor benefit. See 
Karen C. Holden and Angela Fontes, "Economic Security in Retirement: 
How Changes in Employment and Marriage Have Altered Retirement-Related 
Economic Risks for Women," Journal of Women, Politics & Policy, vol. 
30, no. 2-3 (2009).

[49] We defined unemployment as being out of work and actively looking 
for a job.

[50] We have previously reported that older workers generally have 
longer spells of unemployment than younger workers and that older 
workers report facing difficulties finding new jobs after being laid 
off. See GAO-12-445.

[51] When individuals enter the HRS sample, they are between the ages 
of 51 and 61. However, because this is a longitudinal study, all the 
survey members age over time. For example, someone who was age 61 at 
the time of the first HRS survey in 1992 was age 79 in 2010.

[52] This difference between women and men is not statistically 
significant. 

[53] Health care costs may deplete elderly individuals' resources. See 
McGarry and Schoeni (2005). Also see Richard W. Johnson, Gordon B.T. 
Mermin, and Cori E. Uccello, When the Nest Egg Cracks: Financial 
Consequences of Health Problems, Marital Status Changes, and Job 
Layoffs at Older Ages (Urban Institute: January 2006). 

[54] Although the fixed-effects method offers several advantages over 
other regression methods, it also has limitations that may affect our 
estimates. For example, while the fixed-effects method controls for 
all characteristics within a household that do not change over time, 
it is possible that the relationship between providing care for 
parents and household assets changes over time and works in multiple 
directions. For example, if a household sees an increase in the value 
of its assets, it may decide to use some of this new wealth to finance 
care for elderly parents. However, using these assets causes total 
household assets to fall. The fixed-effects method cannot control for 
these simultaneous effects and, thus, the two effects may cancel each 
other out. For more information on our analysis of the effects of 
providing help to elderly parents and an analysis describing the 
individuals who provided care to parents, see appendix I.

[55] To identify and analyze policy options that could enhance women's 
retirement security, we conducted an extensive literature review and 
interviewed a range of experts. To ensure that we obtained a balanced 
perspective, we interviewed experts with a range of perspectives and 
from different types of organizations including government, academia, 
advocacy groups, and the private sector. For more information on our 
literature review and expert interviews, see appendix I. Some of the 
options have been proposed in various permutations. Our analysis is 
not intended to describe any one proposal. Rather, we describe the 
basic features of the option; these features may be common across 
proposals. GAO did not independently evaluate the efficacy of these 
options, nor by including them in this report are we providing a 
position on or endorsing any of these options.

[56] We have previously reported that there is a need to improve 
individuals' financial literacy. Financial skills are increasingly 
important for ensuring a comfortable standard of living in retirement. 
GAO, Financial Literacy: Enhancing the Effectiveness of the Federal 
Government's Role, GAO-12-636T (Washington, D.C.: Apr. 26, 2012) and 
Financial Literacy: Strengthening Partnerships in Challenging Times, 
GAO-12-299SP (Washington, D.C.: Feb. 9, 2012).

[57] Certain provisions of the Internal Revenue Code set required 
minimum distributions from tax-deferred accounts, such as traditional 
IRAs and qualified plans, starting generally by April 1 in the year 
following the year in which the account holder reaches age 70 ½. These 
required minimum distributions help to ensure that account holders 
withdraw tax-deferred savings in retirement rather than accumulate 
savings for their estate.

[58] Experts we spoke with also identified women without long-term 
care insurance as a vulnerable population. Although the lack of long-
term care insurance does put women at risk of income insecurity, in 
general, we did not identify any long-term care policy options that 
addressed retirement income specifically.

[59] For more information regarding such misreporting, see Irena 
Dushi, Howard M. Iams, and Jules Lichtenstein, "Assessment of 
Retirement Plan Coverage by Firm Size, Using W-2 Tax Records," Social 
Security Bulletin, vol. 71, no. 2 (2011). 

[60] The survey contains catch-all questions for whether an 
individual's employer offers a DC plan, but it does not contain 
similar questions for DB plans. Specifically, those who are not 
included in their employer's plan are not asked whether their employer 
offers a DB plan.

[61] Note that in the models we present, we did not include income as 
a control variable. Income can be considered to be endogenously (or 
simultaneously) determined with an individual's decision to work for a 
particular employer that might be offering a plan and therefore have 
the potential to bias the model estimates. For example, one might 
deliberately choose to work in a lower-paid government position to 
ensure access to a DB plan. We did run versions of our model with 
income included as a control and found that it was significantly 
associated with the likelihood of working for an employer that offers 
a plan and of participating in a plan. 

[62] This result is consistent with Census findings, which note a 
higher male-to-female ratio among the Hispanic population in the 
United States than among the general population. 

[63] Odds (O) are mathematically related to but not the same as 
probabilities (P), that is O=P/[1-P]. 

[64] While dummy and categorical variables are both discrete 
variables, a dummy variable takes on a value of 0 or 1. A categorical 
variable takes a value that is one of several possible categories and 
there is no intrinsic ordering to the categories.

[65] This method might result in overstated estimates from earnings if 
workers do not work all 12 months of the year. 

[66] These patterns held across all the years we analyzed.

[67] It is also possible that the survey response rate was higher for 
married men than for married women.

[68] In addition to the fixed-effects analysis, we also developed 
"cross-section" regression models. In these models, we attempted to 
control for a set of demographic and other variables, such as 
education and age that could be correlated with life events, household 
assets, and household income. A challenge to this approach is that 
many factors that affect assets and income are unobserved, and lead to 
mistaken conclusions. For example, if an individual earns a low wage, 
that may be connected with poor health and the accumulation of assets. 
So, while the researcher is attempting to estimate the effect of 
health on income, what is actually measured is the effect of income on 
health. In general, in our cross-section models, we found that effects 
were larger in magnitude than in the fixed-effects models, but these 
models were not as good a fit to the data as the fixed-effects models.

[69] Other control variables that we included were age (measured as 
date of wave minus birth year), race and education (categorical), 
cohort of HRS survey, Census region, region of birth (12 categories, 
including non-U.S.). In general, in the cross-section models, we found 
that education was positively related to assets and income, while 
minority status was negatively related. With some slight variation, we 
based our choice of control variables on Coile and Milligan. (See 
Courtney Coile and Kevin Milligan, "How Household Portfolios Evolve 
After Retirement: The Effect of Aging and Health Shocks," The Review 
of Income and Wealth, vol. 55 no. 2 (Malden, MA: June 2009)).

[70] In order to estimate effects in terms of percents, we estimated 
the effects on the log of assets or income. In addition, we 
transformed the coefficients to more closely approximate percent 
changes by taking the exponent of the estimated coefficient and 
subtracting 1. Regression variables were weighted by household weights. 

[71] We tested this result by using an alternative measure of health: 
the extent to which there are challenges in daily living. In this 
case, we did not find that men's health had a larger effect.

[End of section] 

GAO’s Mission: 

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. 

Obtaining Copies of GAO Reports and Testimony: 

The fastest and easiest way to obtain copies of GAO documents at no 
cost is through GAO’s website [hyperlink, http://www.gao.gov]. Each 
weekday afternoon, GAO posts on its website newly released reports, 
testimony, and correspondence. To have GAO e-mail you a list of newly 
posted products, go to [hyperlink, http://www.gao.gov] and select 
“E-mail Updates.” 

Order by Phone: 

The price of each GAO publication reflects GAO’s actual cost of 
production and distribution and depends on the number of pages in the 
publication and whether the publication is printed in color or black 
and white. Pricing and ordering information is posted on GAO’s 
website, [hyperlink, http://www.gao.gov/ordering.htm]. 

Place orders by calling (202) 512-6000, toll free (866) 801-7077, or 
TDD (202) 512-2537. 

Orders may be paid for using American Express, Discover Card, 
MasterCard, Visa, check, or money order. Call for additional 
information. 

Connect with GAO: 

Connect with GAO on facebook, flickr, twitter, and YouTube.
Subscribe to our RSS Feeds or E mail Updates. Listen to our Podcasts.
Visit GAO on the web at [hyperlink, http://www.gao.gov]. 

To Report Fraud, Waste, and Abuse in Federal Programs: 

Contact: 
Website: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]; 
E-mail: fraudnet@gao.gov; 
Automated answering system: (800) 424-5454 or (202) 512-7470. 

Congressional Relations: 

Katherine Siggerud, Managing Director, siggerudk@gao.gov, (202) 512-4400
U.S. Government Accountability Office, 441 G Street NW, Room 7125
Washington, DC 20548. 

Public Affairs: 

Chuck Young, Managing Director, youngc1@gao.gov, (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149 
Washington, DC 20548.