This is the accessible text file for GAO report number GAO-13-699 
entitled 'Automatic IRAS: Lower-Earning Households Could Realize 
Increases in Retirement Income' which was released on Septembver 23, 
2013. 

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 Ranking Member, Subcommittee on Select Revenue Measures, 
Committee on Ways and Means, House of Representatives: 

August 2013: 

Automatic IRAS: 

Lower-Earning Households Could Realize Increases in Retirement Income: 

GAO-13-699: 

GAO Highlights: 

Highlights of GAO-13-699, a report to the Ranking Member, Subcommittee 
on Select Revenue Measures, Committee on Ways and Means, House of 
Representatives. 

Why GAO Did This Study: 

receive tax incentives for their contributions and lower-earning 
households may qualify for the Saver’s Credit, an additional tax 
incentive for their contributions. However, less than half of the 
workforce participates in an employer-sponsored plan and upper-income 
workers have been more likely to take advantage of associated tax 
incentives. In recent years, proposals have been put forth to modify 
the Saver’s Credit and create automatic IRAs, under which employers 
who do not sponsor a plan would generally be required to offer their 
employees the opportunity to save in an IRA through payroll deduction. 
These proposals would have fiscal impacts for the federal government. 

GAO was asked to review tax incentives for contributions to DC plans 
and automatic IRAs. GAO examined (1) the earnings and tax rates of 
households that do not have DC plans or IRAs, (2) the effects of the 
Saver’s Credit on retirement income, and (3) the effects of automatic 
IRAs on retirement income, especially for low- and middle-income 
workers. GAO examined the characteristics of households that do not 
take advantage of these tax incentives using data from the 2010 Survey 
of Consumer Finances, simulated the effects of the Saver’s Credit and 
automatic IRAs, and reviewed related proposals. 

What GAO Found: 

Households without employer-sponsored defined contribution (DC) 
pension plans or individual retirement accounts (IRA) had lower 
incomes and tax rates than households with those plans, and are also 
likely to have limited additional resources to draw upon in 
retirement, according to GAO estimates. The median adjusted gross 
income for households without DC plans or IRAs was $32,000, compared 
to $75,000 for those that did have them. The median marginal tax rate 
for households without DC plans or IRAs was 15 percent, compared to 25 
percent for households with those savings vehicles. A defined benefit 
(DB) pension plan could provide a monthly benefit during retirement 
years for those without a DC plan or IRA; however, in 2010 only 15 
percent of married households and 11 percent of single households 
without a DC plan or IRA had a DB plan. 

The existing Saver’s Credit tax incentive could result in small 
increases in a household’s retirement annuity—-that is, the 
household’s annual retirement income received from DC or DB plans. GAO 
estimates that, on account of this credit, the median annuity increase 
for households in the lowest earnings quartile ($929-34,377) would be 
$155. If, however, the Saver’s Credit was refundable (i.e., could 
generate a tax refund in excess of tax paid), it could result in 
larger increases in households’ annuities across all earnings levels, 
and the median increase for households in the lowest earnings quartile 
would be $876 per year. 

Implementing automatic IRAs, unless waived by participants, could 
expand retirement coverage and modestly increase retirement annuities 
for households at all earnings levels. Specifically, 7 percent of all 
households could receive retirement annuities from automatic IRAs even 
though these households had no DB or DC plans, according to GAO’s 
projections. Workers with DB or DC plans could also benefit from 
automatic IRAs at certain points in their lifetime if their jobs do 
not offer such plans. Moreover, low-income workers could see a sizable 
increase in their annuities under automatic IRAs and the existing Saver’
s Credit—the projected median dollar increase for these households’ 
annual retirement annuity would be $479. 

Figure: Projected Annual Changes in Households’ Retirement Annuity 
under Automatic IRAs: 

[Refer to PDF for image: horizontal bar graph] 

Median dollar change in annual household retirement annuity: 

Total: $1,046; 
1st earnings quartile ($929-$34,377): $479; 
2nd earnings quartile ($34,379-$60,682): $1,043; 
3rd earnings quartile ($60,683-$104,440): $1,388; 
4th earnings quartile ($104,450-$3,275,800): $1,721. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

[End of figure] 

What GAO Recommends: 

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

Contents: 

Letter: 

Background: 

Households Without DC Plans or IRAs Had Lower Adjusted Gross Incomes: 

Modifying the Existing Saver's Credit Could Result in Larger Increases 
in Retirement Income: 

Automatic IRAs Could Increase Retirement Income for Households at All 
Earnings Levels: 

Concluding Observations: 

Agency Comments and Our Evaluation: 

Appendix I: Objectives, Scope, and Methodology: 

Appendix II: Sensitivity Analysis and Summary Statistics for Saver's 
Credit and Automatic IRA Modeling Scenarios: 

Appendix III: GAO Contact and Staff Acknowledgments: 

Tables: 

Table 1: Saver's Credit Rates, AGI Limits, and Maximum Credits by Tax 
Filing Status, 2012: 

Table 2: Median AGI for Households With and Without DC Plans or IRAs, 
by Age, 2010: 

Table 3: Median AGI for Households With and Without DC Plans or IRAs, 
by Tax Filing Status, 2010: 

Table 4: Median Marginal Tax Rate for Households With and Without DC 
Plans or IRAs, by Age, 2010: 

Table 5: Median Marginal Tax Rate for Households With and Without DC 
Plans or IRAs, by Tax Filing Status, 2010: 

Table 6: Percent of Households Without a DC Plan or IRA, by Whether 
They Have a DB Plan, 2010: 

Table 7: Results for Saver's Credit Scenarios With Immediate Tax 
Increase Social Security Benchmark: 

Table 8: Results for Saver's Credit Scenarios With Standard Fees, 100 
Percent Utilization Rate, and Social Security Benefit Reduction 
Benchmark: 

Table 9: Results for Automatic IRA and Refundable Saver's Credit 
Scenarios With Standard Fees, 69 Percent Automatic IRA Participation 
Rate, and Immediate Tax Increase Social Security Benchmark: 

Table 10: Results for Automatic IRA and Refundable Saver's Credit 
Scenarios With High Fees, 48 Percent Automatic IRA Participation Rate, 
and Immediate Tax Increase Social Security Benchmark: 

Table 11: Results for Automatic IRA and Refundable Saver's Credit 
Scenarios With Standard Fees, 69 Percent Automatic IRA Participation 
Rate, and Social Security Benefit Reduction Benchmark: 

Table 12: Sample Summary Statistics at Age 67, 1995 PENSIM cohort: 

Table 13: Means at Age 67, 1995 PENSIM cohort: 

Figures: 

Figure 1: Estimated Median AGI by Demographic Group for Households 
With and Without Defined Contribution (DC) Plans or Individual 
Retirement Accounts (IRAs), 2010: 

Figure 2: Estimated Median Marginal Tax Rate by Demographic Group for 
Households With and Without Defined Contribution (DC) Plans or 
Individual Retirement Accounts (IRAs), 2010: 

Figure 3: Percent of Households Without a Defined Contribution (DC) 
Plan or Individual Retirement Account (IRA) by Whether They Have a 
Defined Benefit (DB) Plan, 2010: 

Figure 4: Projected Percent of Individuals Who Received the Saver's 
Credit During their Working Years: 

Figure 5: Projected Annual Changes in Households' Retirement Annuity 
under Saver's Credit: 

Figure 6: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs: 

Figure 7: Projected Percent of Households That Could See an Increase 
in Retirement Annuity under Automatic IRAs and a Refundable Saver's 
Credit: 

Figure 8: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs and Refundable Saver's Credit: 

Figure 9: Projected Median Percent of Household Income Coming from 
Retirement Annuities and Social Security: 

Figure 10: Projected Annual Changes in Households' Retirement Annuity 
under Saver's Credit: 

Figure 11: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs: 

Figure 12: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs and Refundable Saver's Credit: 

Figure 13: Projected Mean Percent of Household Income Coming from 
Retirement Annuities and Social Security: 

Abbreviations: 

AGI: adjusted gross income: 

DB: defined benefit: 

DC: defined contribution: 

EBSA: Employee Benefits Security Administration: 

ERISA: Employee Retirement and Income Security Act of 1974: 

IRA: individual retirement account: 

IRC: Internal Revenue Code: 

IRS: Internal Revenue Service: 

OASDI: Old-Age, Survivors, and Disability Insurance: 

OLC: overlapping cohorts: 

PSG: Policy Simulation Group: 

SCF: Survey of Consumer Finances: 

RCS: representative cohort sample: 

[End of section] 

GAO:
United States Government Accountability Office: 
441 G St. N.W. 
Washington, DC 20548: 

August 23, 2013: 

The Honorable Richard Neal: 
Ranking Member: 
Subcommittee on Select Revenue Measures: 
Committee on Ways and Means: 
House of Representatives: 

Dear Mr. Neal: 

Workers who make contributions to defined contribution (DC) pension 
plans and individual retirement accounts (IRAs) meeting certain 
qualifications receive favorable federal tax treatment, such as tax 
deferred savings. These tax preferences are one of the largest tax 
expenditures in the federal government. Under DC plans and certain 
IRAs, taxes on contributions and investment earnings are deferred 
until benefits are received in retirement. In fiscal year 2012, the 
estimated revenue loss associated with these accounts included $51.8 
billion for DC plans and $16.2 billion for IRAs.[Footnote 1] In 
addition, an estimated $1.1 billion in federal tax revenue was forgone 
in fiscal year 2012 due to the Saver's Credit, a tax credit to 
encourage workers with low earnings to contribute to a pension plan or 
an IRA. The purpose of giving such plans and IRAs preferential tax 
treatment is to encourage employers to form new plans or maintain 
existing ones and to encourage workers to save for retirement. 
[Footnote 2] 

However, in 2011, only about 45 percent of wage and salary workers 
aged 21-64 participated in employer-sponsored pension plans.[Footnote 
3] Studies by the Center for Retirement Research at Boston College and 
the Employee Benefit Research Institute show that low-and middle-wage 
workers are not prepared for retirement, even with these tax 
incentives.[Footnote 4] Further, as we have previously reported, there 
is concern that much of the tax benefits flow to workers with higher 
earnings who have a greater incentive to save, and the financial 
constraints on lower-wage workers limit their ability to contribute to 
tax-qualified plans.[Footnote 5] 

In recent years, both members of Congress and the President have 
proposed requiring employers who do not sponsor a plan to offer their 
employees the opportunity to make contributions to an IRA through 
automatic payroll deduction.[Footnote 6] In addition, in his 2010 and 
2011 budgets, the President proposed modifying the existing Saver's 
Credit by making it refundable, expanding the number of households 
eligible for the credit by increasing the adjusted gross income (AGI) 
limit, and providing the same credit rate to all recipients.[Footnote 
7] Bills pending before Congress in 2012 and 2013 would make various 
changes to the Saver's Credit as well, including making it refundable. 
[Footnote 8] 

In light of your interest in understanding the extent to which workers 
with lower earnings could benefit from tax-deferred retirement savings 
accounts and automatic IRAs, this report addresses the following 
questions: 

1. What are the earnings and tax rates of households that do not have 
DC plans or IRAs? 

2. What are the effects of the Saver's Credit on retirement income? 

3. How would automatic IRAs affect retirement income, especially for 
low-and middle-income workers? 

To answer the first question, we used data from the 2010 Survey of 
Consumer Finances (SCF) to analyze the earnings and tax rates of 
households that did not take advantage of the existing tax incentives 
for retirement saving. To answer the two subsequent questions, we used 
the Policy Simulation Group's (PSG) microsimulation models to assess 
the distributional effects of the existing Saver's Credit, the 
President's 2011 budget proposal to modify the Saver's Credit, and 
H.R. 4049, which would have established automatic IRAs and which was 
pending before the 112th Congress.[Footnote 9] We also reviewed 
relevant federal laws and regulations, as well as, presidential and 
legislative proposals related to the Saver's Credit and automatic 
IRAs. We conducted a data reliability assessment of the PSG models and 
selected variables from the SCF 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. We found that for 
the purposes of our analysis, the data that we analyzed were 
sufficiently reliable. Finally, we interviewed experts in the area of 
retirement income security. 

We conducted this performance audit from May 2012 to August 2013 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 obtained provides a reasonable basis for our 
findings and conclusions based on our audit objectives. 

Background: 

U.S. Retirement Saving System: 

According to one study, in 2011, 54 percent of wage and salary workers 
aged 21-64 worked for an employer that sponsored a pension plan, such 
as a defined benefit (DB) plan or a DC plan, but only about 45 percent 
of wage and salary workers aged 21-64 actually participated in the 
plan.[Footnote 10] DB plans provide periodic benefits in retirement 
that are generally based on employees' salaries and years of service. 
Employers may also choose to sponsor DC plans, under which both 
employers and employees can make contributions to the plan. 
Distributions in retirement are, in turn, based on contributions and 
investment returns in these accounts. 

Private sector employer-sponsored DB and DC plans are generally 
subject to the Employee Retirement Income Security Act of 1974 
(ERISA), as amended, which establishes standards for private sector 
pension plans and sets forth protections for participants in these 
plans. The Department of Labor's Employee Benefits Security 
Administration (EBSA) generally administers and enforces the Title I 
provisions of ERISA. These employer-sponsored plans must also meet 
certain requirements in the Internal Revenue Code (IRC), which are 
enforced by the Internal Revenue Service (IRS). 

Individuals can also save for retirement through IRAs, which allow 
individuals to make contributions for retirement regardless of whether 
they are covered by an employer-sponsored plan.[Footnote 11] The IRS 
has primary responsibility for ensuring that IRAs meet IRC 
requirements necessary to qualify for preferential tax treatment. 

Employers are continuing to shift away from sponsoring DB plans toward 
sponsoring DC plans. Data from the Department of Labor show that over 
the past few decades, DC plans have become the predominant plan type 
offered by private sector employers. Indeed, in 2010, over 90 percent 
of all employer-sponsored plans were DC plans.[Footnote 12] IRAs have 
also grown in importance in recent years and are a key retirement 
savings vehicle for many individuals. According to data from the 
Investment Company Institute, in the first quarter of 2013, IRA assets 
represented a larger portion of total U.S. retirement assets than 
401(k) plans, the main type of DC plan. Specifically, IRA assets 
totaled almost $5.7 trillion, which represents about 27 percent of U.S 
retirement assets. In comparison, 401(k) assets accounted for about 
$3.8 trillion, or 18 percent, of U.S. retirement assets.[Footnote 13] 
Rollovers from 401(k) plans and other employer-sponsored plans are the 
predominant source of contributions to IRAs. Approximately 95 percent 
of money contributed to traditional IRAs in 2008 was attributable to 
rollovers, primarily from employer-sponsored plans.[Footnote 14] The 
greater reliance on DC plans and IRAs in the current retirement 
landscape indicates that the tax incentives are increasingly relevant 
for promoting retirement saving. 

Tax Incentives for Retirement Saving: 

Tax preferences for pension plans are structured to encourage 
individuals to save for retirement. Contributions to DC plans that 
fall within certain limits and investment earnings on assets are not 
taxed until distributions are paid to participants.[Footnote 15] 
Further, there are tax incentives for contributions to IRAs. There are 
two main types of IRAs: traditional and Roth. 

* A traditional IRA allows individuals to make tax deductible 
contributions to their accounts, and distributions are generally 
subject to income tax. For traditional IRAs, deductions for 
contributions are subject to limits based on income and pension 
coverage. Distributions made prior to age 59 ½, other than under 
specific exceptions, are generally subject to additional 10 percent 
tax. 

* A Roth IRA allows individuals to make after-tax contributions to 
their accounts, and generally distributions after age 59 ½ for 
accounts at least 5 years old are not subject to income tax. For Roth 
IRAs, distributions prior to age 59 ½ are taxable on portions 
attributable to earnings on contributions with an additional 10 
percent tax on distributions other than for specified purposes. 

Saver's Credit: 

To further encourage low-and middle-income individuals and families to 
save for retirement, the Economic Growth and Tax Relief Reconciliation 
Act of 2001 authorized a nonrefundable tax credit (the Saver's Credit) 
of up to $1,000 against federal income tax.[Footnote 16] Eligibility 
is based on a worker's adjusted gross income (AGI) and contributions 
made to qualified pension plans and IRAs.[Footnote 17] The credit rate 
phases out as AGI increases, and the rate is applied to qualified 
contributions up to $2,000 for individuals and $4,000 for 
households.[Footnote 18] The total Saver's Credit amount is equal to 
the amount of contributions multiplied by the credit rate (see table 
1). Because the credit is nonrefundable, the ability to receive the 
full amount of credit is limited not only by income eligibility, but 
also by whether the taxpayer's tax liability is large enough. In 
fiscal year 2012, the Saver's Credit cost the federal government about 
$1.1 billion in revenue forgone, and revenue losses are estimated to 
amount to $6.2 billion for fiscal years 2013-2017, according to the 
Department of the Treasury.[Footnote 19] 

Table 1: Saver's Credit Rates, AGI Limits, and Maximum Credits by Tax 
Filing Status, 2012: 

Single, married filing separately, or qualifying widow/widower: 

Credit Rate: 50%; 
AGI limit: $17,250; 
Maximum Available Credit: $1,000. 

Credit Rate: 20%; 
AGI limit: $18,750; 
Maximum Available Credit: $400. 

Credit Rate: 10%; 
AGI limit: $28,750; 
Maximum Available Credit: $200. 

Head of Household: 

Credit Rate: 50%; 
AGI limit: $25,875; 
Maximum Available Credit: $1,000. 

Credit Rate: 20%; 
AGI limit: $28,125; 
Maximum Available Credit: $400. 

Credit Rate: 10%; 
AGI limit: $43,125; 
Maximum Available Credit: $200. 

Married filing jointly: 

Credit Rate: 50%; 
AGI limit: $34,500; 
Maximum Available Credit: $2,000. 

Credit Rate: 20%; 
AGI limit: $37,500; 
Maximum Available Credit: $800. 

Credit Rate: 10%; 
AGI limit: $57,500; 
Maximum Available Credit: $400. 

Source: Department of the Treasury, IRS, Form 8880. 

[End of table] 

To expand retirement savings incentives for working families, the 
President's fiscal year 2010 and 2011 budgets proposed modifying the 
Saver's Credit. These modifications included: (1) making the credit 
refundable, (2) eliminating phase-out of the credit rate so that all 
eligible households would receive a 50 percent credit rate, and (3) 
extending the AGI limit to $85,000 for married couples filing their 
federal income taxes jointly.[Footnote 20] The estimated cost for this 
expansion of the Saver's Credit was $323 million for fiscal year 2011 
and $29.8 billion for fiscal years 2011-2020, according to the 
President's 2011 budget proposal.[Footnote 21] The Joint Committee on 
Taxation estimated that the Administration's 2011 proposal would have 
cost about $27.5 billion for the same period.[Footnote 22] These 
modifications were not adopted and were not included in the 
President's budget proposals for fiscal years 2012 through 2014. Bills 
introduced during the current and immediately preceding legislative 
sessions include provisions for similar modifications to the Saver's 
Credit, including making the credit refundable and increasing the AGI 
range for taxpayers eligible for the 50 percent credit rate.[Footnote 
23] 

Automatic IRA Proposals: 

Automatic enrollment has been advocated as a way to encourage greater 
participation in 401(k) plans. Our past work has shown that automatic 
enrollment policies considerably increased 401(k) participation rates 
for plans that adopted them.[Footnote 24] Automatic IRAs would 
similarly extend the benefits of payroll-deduction savings and 
automatic enrollment. Legislative proposals to establish automatic 
IRAs have been introduced in both the House of Representatives and the 
Senate in recent legislative sessions.[Footnote 25] Additionally, the 
President's budgets for fiscal years 2010 through 2014 included a 
proposal to establish automatic IRAs. 

Under the 2012 automatic IRA legislative proposal, certain employers 
who do not maintain a qualified pension plan would be required to make 
available an automatic IRA arrangement to eligible employees. 
Typically, under this proposal, 3 percent of the employee's salary 
would be automatically contributed to an IRA through payroll 
deduction, unless the employee elects to terminate his or her 
participation.[Footnote 26] Investment options for automatic IRAs 
would be limited to certain types of funds, such as principal 
preservation and target-date or life cycle funds, and target-date 
funds would be the default investment option. If the employer elects, 
employees would also have the option of using their contributions to 
purchase a retirement bond, which would provide a low-cost investment 
option where small account balances are pooled together until they are 
large enough to be profitable in the private market.[Footnote 27] This 
proposal also includes a provision for a tax credit for small 
employers to recoup start-up costs of establishing and maintaining 
automatic IRAs, such as setting up an automatic payroll deduction for 
employees. Automatic IRAs would also introduce costs to the federal 
government through the loss of federal income tax revenue.[Footnote 
28] According to the President's 2014 budget proposal, losses for 
automatic IRAs are estimated to amount to $1.1 billion for fiscal year 
2015 and $17.6 billion for fiscal years 2014 through 2023, if the 
proposal were to be enacted.[Footnote 29] 

Households Without DC Plans or IRAs Had Lower Adjusted Gross Incomes: 

Households That Do Not Save for Retirement Have Lower Adjusted Gross 
Incomes and Marginal Tax Rates Than Those That Do Save: 

We found that households that do not save for retirement had lower AGI 
than those that do, regardless of age group or tax filing status (see 
figure 1). Based on our analysis of the SCF data, we estimate that in 
2010 approximately 43 percent of households working in the private 
sector did not have a DC plan or IRA.[Footnote 30] We also estimate 
that the median AGI of all households that did not have a DC plan or 
IRA was $32,000, compared to $75,000 for those that did--a difference 
of about 43 percent. These trends hold across age groups and tax 
filing statuses in 2010. In addition, we estimate that 56 percent of 
single households did not have a DC plan or IRA, while 36 percent of 
married households did not and married household AGI was consistently 
higher than single household AGI. 

Figure 1: Estimated Median AGI by Demographic Group for Households 
With and Without Defined Contribution (DC) Plans or Individual 
Retirement Accounts (IRAs), 2010: 

[Refer to PDF for image: horizontal bar graph] 

Median household adjusted gross income: 

Age group: Age less than 35; 
Married households: 
Household has DC or IRA: $68,000; 
Household does not have DC or IRA: $38,000; 
Single households: 
Household has DC or IRA: $42,000; 
Household does not have DC or IRA: $18,200. 

Age group: Age 35-44; 
Married households: 
Household has DC or IRA: $90,000; 
Household does not have DC or IRA: $44,000; 
Single households: 
Household has DC or IRA: $50,000; 
Household does not have DC or IRA: $27,000. 

Age group: Age 45-54; 
Married households: 
Household has DC or IRA: $100,000; 
Household does not have DC or IRA: $51,000; 
Single households: 
Household has DC or IRA: $48,000; 
Household does not have DC or IRA: $28,000. 

Age group: Age 55-64; 
Married households: 
Household has DC or IRA: $102,000; 
Household does not have DC or IRA: $44,000; 
Single households: 
Household has DC or IRA: $53,000; 
Household does not have DC or IRA: $31,000. 

Filing status: Married filing jointly; 
Married households: 
Household has DC or IRA: $90,000; 
Household does not have DC or IRA: $42,000; 
Single households: 
Household has DC or IRA: N/A; 
Household does not have DC or IRA: N/A. 

Filing status: Single; 
Married households: 
Household has DC or IRA: $48,000; 
Household does not have DC or IRA: $48,000; 
Single households: 
Household has DC or IRA: $25,000; 
Household does not have DC or IRA: 25,000. 

Filing status: Head of household; 
Married households: 
Household has DC or IRA: N/A; 
Household does not have DC or IRA: N/A; 
Single households: 
Household has DC or IRA: $46,000; 
Household does not have DC or IRA: $20,000. 

Source: GAO analysis of the 2010 Survey of Consumer Finances. 

Notes: Results are for households where the household head was under 
the age of 65 and either the respondent or spouse worked in the 
private sector. The Federal Reserve provides a tax filing status 
variable, which we used in our analysis. For the married filing 
jointly category, the Federal Reserve included married households that 
filed income taxes jointly and separately. For married households, the 
single tax filing status category included households where only the 
respondent or their spouse filed taxes and the filer claimed one 
personal exemption. There are about 48 million married households in 
the married filing jointly category and about one million married 
households in the single tax filing status category. 

[End of figure] 

In addition, households without DC plans or IRAs have lower median 
marginal tax rates than households with DC plans or IRAs. A marginal 
tax rate is the rate of tax paid on the next dollar of income that a 
taxpayer earns.[Footnote 31] Generally households with higher earnings 
have higher marginal tax rates, and tax incentives are worth more to 
these households. Based on our analysis of the 2010 SCF data, we 
estimate the median marginal tax rate for households with DC plans or 
IRAs to be 25 percent and 15 percent for households without such 
savings vehicles. Across age groups and tax filing statuses, more 
households with higher median marginal tax rates took advantage of tax 
incentives for retirement saving than households with lower median 
marginal tax rates (see figure 2).[Footnote 32] For example, married 
households age 45 to 54 that did not have a DC plan or IRA had a 
median marginal tax rate of 15 percent, and those that did save had a 
median marginal tax rate of 25 percent. Single households age 45 to 54 
that did not have a DC plan or IRA had a median marginal tax rate of 
15 percent, compared to 22 percent for those that did save.[Footnote 
33] 

Figure 2: Estimated Median Marginal Tax Rate by Demographic Group for 
Households With and Without Defined Contribution (DC) Plans or 
Individual Retirement Accounts (IRAs), 2010: 

[Refer to PDF for image: horizontal bar graph] 

Median marginal tax rate: 

Age group: Age less than 35; 
Married households: 
Household has DC or IRA: 15%; 
Household does not have DC or IRA: 15%; 
Single households: 
Household has DC or IRA: 15%; 
Household does not have DC or IRA: 15%. 

Age group: Age 35-44; 
Married households: 
Household has DC or IRA: 25%; 
Household does not have DC or IRA: 15%; 
Single households: 
Household has DC or IRA: 25%; 
Household does not have DC or IRA: 15%. 

Age group: Age 45-54; 
Married households: 
Household has DC or IRA: 25%; 
Household does not have DC or IRA: 15%; 
Single households: 
Household has DC or IRA: 22%; 
Household does not have DC or IRA: 15%. 

Age group: Age 55-64; 
Married households: 
Household has DC or IRA: 25%; 
Household does not have DC or IRA: 15%; 
Single households: 
Household has DC or IRA: 22%; 
Household does not have DC or IRA: 15%. 

Filing status: Married filing jointly; 
Married households: 
Household has DC or IRA: 25%; 
Household does not have DC or IRA: 15%; 
Single households: 
Household has DC or IRA: N/A; 
Household does not have DC or IRA: N/A. 
        
Filing status: Single; 
Married households: 
Household has DC or IRA: 25%; 
Household does not have DC or IRA: 15%; 
Single households: 
Household has DC or IRA: 25%; 
Household does not have DC or IRA: 15%. 

Filing status: Head of household; 
Married households: 
Household has DC or IRA: N/A; 
Household does not have DC or IRA: N/A; 
Single households: 
Household has DC or IRA: 15%; 
Household does not have DC or IRA: 9%. 

Source: GAO analysis of the 2010 Survey of Consumer Finances. 

Notes: Results are for households where the household head was under 
the age of 65 and either the respondent or spouse worked in the 
private sector. The Federal Reserve provides a tax filing status 
variable, which we used in our analysis. For the married filing 
jointly category, the Federal Reserve included married households that 
filed income taxes jointly and separately. For married households, the 
single tax filing status category included households where only the 
respondent or their spouse filed taxes and the filer claimed one 
personal exemption. There are about 48 million married households in 
the married filing jointly category and about one million married 
households in the single tax filing status category. A marginal tax 
rate is the rate of tax paid on the next dollar of income that a 
taxpayer earns. 

[End of figure] 

While DB Plans Could Provide Additional Retirement Income, Few 
Households Without DC Plans or IRAs Have Them: 

As mentioned previously, DB plans typically provide a monthly benefit 
once an individual reaches retirement age and could be a source of 
retirement income for households without DC plans or IRAs. However, 
the prevalence of DB plans has declined over the past few decades. 
Consistent with overall trends, we found that few households without 
DC plans or IRAs have a DB plan (see figure 3). Only 15 percent of 
married households and 11 percent of single households without a DC 
plan or IRA had a DB plan. Although most households without a DC plan 
or IRA also do not have a DB plan, the first earnings quartile is 
worse off. For example, an estimated 5 percent of married households 
in the first earnings quartile had a DB plan, compared to an estimated 
27 percent and 31 percent of married households in the third and 
fourth earnings quartiles, respectively.[Footnote 34] The shift away 
from DB plans and the limited number of households that have them 
highlights the importance of other retirement savings vehicles such as 
DC plans or IRAs. 

Figure 3: Percent of Households Without a Defined Contribution (DC) 
Plan or Individual Retirement Account (IRA) by Whether They Have a 
Defined Benefit (DB) Plan, 2010: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Married households: 
Has a DB: 15%; 
Does not have a DB: 85%; 
Single households: 
Has a DB: 11%; 
Does not have a DB: 89%. 

1st earnings quartile: 
Married households: 
Has a DB: 5%; 
Does not have a DB: 95%; 
Single households: 
Has a DB: 4%; 
Does not have a DB: 96%. 

2nd earnings quartile: 
Married households: 
Has a DB: 21%; 
Does not have a DB: 79%; 
Single households: 
Has a DB: 9%; 
Does not have a DB: 91%. 

3rd earnings quartile: 
Married households: 
Has a DB: 27%; 
Does not have a DB: 73%; 
Single households: 
Has a DB: 25%; 
Does not have a DB: 75%. 

4th earnings quartile: 
Married households: 
Has a DB: 31%; 
Does not have a DB: 69%; 
Single households: 
Has a DB: 25%; 
Does not have a DB: 75%. 

Source: GAO analysis of the 2010 Survey of Consumer Finances. 

Notes: Results are for households where the household head was under 
the age of 65 and either the respondent or spouse worked in the 
private sector and do not save for retirement through a DC plan or 
IRA. We analyzed our results by earnings level using earnings 
quartiles. Households with adjusted gross incomes (adjusted for family 
size) of less than $18,960 were in the first earnings quartile, 
households with incomes of $18,961-33,519 were in the second earnings 
quartile, households with incomes of $33,520-56,559 were in the third 
earnings quartile, and households with incomes of $56,560-25,583,473 
were in the fourth earnings quartile. 

[End of figure] 

We also analyzed the net worth--the difference between gross assets 
and liabilities--of single and married households without DC plans or 
IRAs by income quartile. Our evidence suggests that in addition to 
some households not having a DC plan or IRA, these households also may 
have limited additional assets to draw upon in their retirement years. 
[Footnote 35] For example, single households in the first income 
quartile without a DB plan had an estimated average net worth of 
$17,164.[Footnote 36] Married households in the first income quartile 
without a DB plan had an estimated average net worth of $70,485. 

Modifying the Existing Saver's Credit Could Result in Larger Increases 
in Retirement Income: 

The Saver's Credit could increase retirement income for a sizable 
portion of the population, especially if modifications, such as making 
it refundable and expanding eligibility, are included in its design. 
[Footnote 37] To analyze the effect of the credit on retirement 
income, we modeled two different Saver's Credit scenarios: the 
existing credit and a modified refundable credit.[Footnote 38] (See 
"Saver's Credit Modeling Scenarios" text box and appendix II for 
detailed descriptions of these scenarios and the underlying 
assumptions).[Footnote 39] Specifically, we projected the effects of 
the two different Saver's Credits for a cohort of individuals born in 
1995 (see figure 4).[Footnote 40] We assumed that all eligible 
taxpayers claimed the credit, and found that under the existing 
Saver's Credit, just over a third of all individuals would be 
projected to receive the credit at some point during their working 
years. In contrast, we projected that about half of all individuals 
would receive the credit at some point during their working years if 
it were refundable. According to experts, the nonrefundable nature of 
the existing credit may limit its utility for low-income households. 
Specifically, many low-income households do not receive the full 
amount of the credit because their tax liability is not high enough. 
[Footnote 41] Our estimates show that changing the current design of 
the credit to make it refundable could substantially increase the 
percentage of individuals who would receive the credit. However, such 
modifications to the credit would also result in increased costs to 
the federal government through the loss of tax revenue and increased 
outlays. 

Under both scenarios, the lowest two earnings quartiles have the 
largest percentage of individuals who could receive the credit. 
[Footnote 42] For example, under the existing Saver's Credit, about 45 
percent of individuals in each of the lowest two earnings quartiles 
could receive it. Alternatively, a refundable Saver's Credit could 
result in more middle-income workers receiving the credit, with an 
estimated 60 percent of individuals in the second earnings quartile 
potentially receiving it. Because we present our results across 
earnings quartiles that were based on lifetime earnings, these results 
reflect individuals who could have received the credit while they were 
lower earners but subsequently became higher earners. 

Figure 4: Projected Percent of Individuals Who Received the Saver's 
Credit During their Working Years: 

[Refer to PDF for image: horizontal bar graph] 

Percentage of individuals who received the Saver’s Credit: 

Total: 
Existing Saver’s Credit: 36%; 
Refundable Saver’s Credit: 50%. 

1st earnings quartile: 
Existing Saver’s Credit: 45%; 
Refundable Saver’s Credit: 54%. 

2nd earnings quartile: 
Existing Saver’s Credit: 45%; 
Refundable Saver’s Credit: 60%. 

3rd earnings quartile: 
Existing Saver’s Credit: 33%; 
Refundable Saver’s Credit: 51%. 

4th earnings quartile: 
Existing Saver’s Credit: 20%; 
Refundable Saver’s Credit: 36%. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: We projected the effects of the Saver's Credit for a cohort of 
individuals born in 1995. The effects were compared to a situation in 
which the credit did not exist. The results shown are projections for 
the year 2062 when the cohort is age 67. We assumed that these 
individuals retired between ages 62 and 67. We analyzed our results by 
earnings level using earnings quartiles, which are calculated based on 
a measure of lifetime household earnings. Households with lifetime 
earnings of $929-34,377 on an annual basis were in the first quartile, 
households with earnings of $34,379-60,682 were in the second 
quartile, households with earnings of $60,683-104,440 were in the 
third quartile, and households with earnings of $104,450-3,275,800 
were in the fourth quartile. The PENSIM model does not report whether 
the spouses of individuals born in 1995 received the Saver's Credit. 
As a result, this figure does not represent the percent of households 
that benefit from the credit. 

[End of figure] 

Text box: 

Saver’s Credit Modeling Scenarios: 

Scenario I: Existing Saver’s Credit. Individuals received a 
nonrefundable credit of up to $1,000 if filing singly or $2,000 if 
married filing jointly. The credit rate is 10, 20, or 50 percent 
depending on AGI and tax filing status. For 2013, the AGI limit is 
$59,000 for individuals with a filing status of married filing 
jointly, $44,250 for individuals with a filing status of head of 
household, and $29,500 for individuals with a filing status of single, 
married filing separately, or widow(er). Limits in subsequent years 
were indexed to inflation. 

Scenario II: Refundable Saver’s Credit. This scenario was drawn from 
the President’s fiscal year 2011 budget proposal because its 
provisions align with possible modifications identified by experts. In 
this scenario, the credit would become refundable in 2014. The credit 
rate would be 50 percent for all tax filers. In 2014, AGI limits would 
be extended to $85,000 for individuals with a filing status of married 
filing jointly, $63,750 for individuals with filing status of head of 
household, and $42,500 for individuals with a filing status of single, 
married filing separately, or widow(er). Limits in subsequent years 
would be indexed to inflation. Individuals could receive a refundable 
credit of up to $1,000 if filing singly or $2,000 if married filing 
jointly. 

Assumptions Applied to Each Scenario: 

We projected the effects of both Saver’s Credit scenarios for a cohort 
of individuals born in 1995. Cohort members could have worked in 
either the private or public sectors. The effects were compared to a 
situation in which the credit did not exist. Our projections show 
household income for retired individuals in the year 2062, when the 
cohort is 67 years old.[A] In our projections, individuals could 
retire at any time after they turned 62. Projections were adjusted for 
household size and reported in 2013 dollars. We analyzed our results 
by earnings level using earnings quartiles, which were calculated 
based on a measure of lifetime household earnings. Households with 
lifetime earnings of $929-34,377 on an annual basis were in the first 
quartile, households with earnings of $34,379-60,682 were in the 
second quartile, households with earnings of $60,683-104,440 were in 
the third quartile, and households with earnings of $104,450-3,275,800 
were in the fourth quartile.[B] 

Our projections assumed a 100 percent utilization rate for the Saver’s 
Credit under both the existing and refundable scenarios, meaning that 
all eligible taxpayers claimed the credit. This utilization rate shows 
the maximum potential impact of the Saver’s Credit on retirement 
income.[C] We assessed the Saver’s Credit as a mechanism for 
accumulating retirement income and did not analyze behavioral 
responses associated with the credit.[D] Research suggests that early 
utilization rates for the existing Saver’s Credit, when it was first 
enacted, ranged from 60-67 percent. Account fees varied by investment 
fund type. We used annual non-stochastic nominal rates of return of 
9.2 percent for stocks and 5.7 percent for government bonds.[E] At 
retirement, workers used their entire defined contribution (DC) 
account balance to purchase a single-life, non-inflation-adjusted 
annuity. This annuity uses prices that are based on projected 
mortality rates for the 1995 birth cohort and annuity price loading 
factors that ensure that the cost of providing these annuities equals 
the revenue generated by selling them at those prices. Benefits from 
defined benefit (DB) plans are also provided in the form of an 
annuity. For the purposes of this report, we defined the household 
retirement annuity as including the annual retirement income received 
from all DC or DB plans. Individuals also received Social Security 
benefits promised under current law. 

Because these scenarios are sensitive to the assumptions used, we ran 
alternate simulations. These simulations used a lower Saver’s Credit 
utilization rate and higher investment fund account fees. 

See appendix II for a detailed explanation of the underlying 
assumptions and projections under our alternate simulations. 

[A] In our simulations, retirement occurs when an individual leaves 
the labor force and starts claiming a Social Security retirement 
benefit. 

[B] For example, a household in the first quartile could have had 
lifetime household earnings of $25,000 on an annual basis over the 
course of its working years. 

[C] PENSIM assumes that the Saver's Credit amount is automatically 
deposited into a recipient's DC account. However, under the existing 
credit, a tax filer would receive the credit amount as part of his 
refund and would be free to use it for any purpose. 

[D] We did not account for any behavioral response that changes in the 
Saver's Credit may have created. In reality, some individuals may 
choose to contribute more to their pension plans in response to 
receiving the Saver's Credit or may choose to start saving in a 
pension plan or IRA. 

[E] Using different rates of return would result in different DC 
account balances at retirement and, as a result, the size of the 
household's retirement annuity could be different. Also, because our 
projections do not stochastically model stock returns, assuming a rate 
of return on assets equal to the historical return on stocks does not 
capture the risks associated with stock returns. 

[End of text box] 

Households could experience larger retirement annuity increases from a 
refundable Saver's Credit than the existing credit, according to our 
projections (see figure 5). For the purposes of this report, household 
retirement annuities include all annual retirement income received 
from DC or DB plans and represent income that the household receives 
in retirement. DB plans typically provide benefits in the form of 
periodic payments which can be summed together into one annual 
payment. DC account balances represent a worker's DC or IRA savings 
from past jobs, and we assumed that at retirement, workers used their 
entire DC account balance to purchase an annuity. We estimate that the 
median increase in households' retirement annuity under the existing 
credit could be $103 per year.[Footnote 43] In contrast, under a 
refundable Saver's Credit, the median increase could be $655 per year, 
according to our simulations. For both Saver's Credit scenarios, low-
and middle-earning households could receive the largest benefit from 
the credit.[Footnote 44] For households in the lowest earnings 
quartile, we projected that the median increase in annuity under the 
existing Saver's Credit would be $155 per year, while the median 
increase in annuity under a refundable credit would be $876 per year. 
In addition, about 21 percent of households in the lowest earnings 
quartile could see an additional $1,000 to $1,999 per year in their 
retirement annuities under a refundable credit. Under the existing 
credit, only 6 percent of households could potentially receive an 
increase in annuity of this size. 

Figure 5: Projected Annual Changes in Households' Retirement Annuity 
under Saver's Credit: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Percent change in median annual household retirement annuity: 
Existing Saver’s Credit: 1%; 
Refundable Saver’s Credit: 6%; 
Median dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $103; 
Refundable Saver’s Credit: $655. 

1st earnings quartile: 
Percent change in median annual household retirement annuity: 
Existing Saver’s Credit: 4%; 
Refundable Saver’s Credit: 15%; 
Median dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $155; 
Refundable Saver’s Credit: $876. 

2nd earnings quartile: 
Percent change in median annual household retirement annuity: 
Existing Saver’s Credit: 1%; 
Refundable Saver’s Credit: 6%; 
Median dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $111; 
Refundable Saver’s Credit: $884. 

3rd earnings quartile: 
Percent change in median annual household retirement annuity: 
Existing Saver’s Credit: 0.3%; 
Refundable Saver’s Credit: 2%; 
Median dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $71; 
Refundable Saver’s Credit: $499. 

4th earnings quartile: 
Percent change in median annual household retirement annuity: 
Existing Saver’s Credit: 0.1%; 
Refundable Saver’s Credit: 1%; 
Median dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $68; 
Refundable Saver’s Credit: $342. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: These projections show the results for only households that 
benefited from the Saver's Credit. We projected the effects of the 
Saver's Credit for a cohort of individuals born in 1995. The effects 
were compared to a situation in which the credit did not exist. The 
results shown are projections for the year 2062 when the cohort is age 
67. We assumed that these individuals retired between ages 62 and 67. 
For purposes of this report, we defined the household retirement 
annuity as including the annual retirement income received from all 
defined contribution or defined benefit plans. This annuity does not 
include Social Security benefits. The annuity values are adjusted for 
family size and are presented in 2013 dollars. We analyzed our results 
by earnings level using earnings quartiles, which are calculated based 
on a measure of lifetime household earnings. Households with lifetime 
earnings of $929-34,377 on annual basis were in the first quartile, 
households with earnings of $34,379-60,682 were in the second 
quartile, households with earnings of $60,683-104,440 were in the 
third quartile, and households with earnings of $104,450-3,275,800 
were in the fourth quartile. 

[End of figure] 

Our estimates of the percent change in median household retirement 
annuities show similar trends. The refundable Saver's Credit results 
in a larger percent increase in the median household retirement 
annuity at all earnings levels. In both scenarios, the projected 
percent increase in the median household retirement annuity for the 
lowest earnings quartile was larger than the increase for other 
earnings quartiles. According to our projections, the percent increase 
in the median retirement annuity for low-earning households would be 4 
percent under the existing Saver's Credit and 15 percent under a 
refundable credit. 

Automatic IRAs Could Increase Retirement Income for Households at All 
Earnings Levels: 

Automatic IRAs Could Increase the Number of Lower-Earning Households 
with Retirement Savings and Modestly Increase Retirement Income: 

Automatic IRAs provide a new opportunity to increase the number of 
households saving for retirement at all earnings levels. We project 
that 7 percent of all households had no retirement annuities from DB 
or DC plans but could receive annuity income from automatic IRAs. (See 
"Automatic IRA and Saver's Credit Modeling Scenarios" text box and 
appendix II for detailed descriptions of the scenarios we modeled and 
their underlying assumptions.[Footnote 45]) Based on our projections, 
more households in the lowest earnings quartile would benefit than in 
any other earnings quartile.[Footnote 46] An estimated 12 percent of 
households that do not have a DB or DC plan in the lowest earnings 
quartile could receive annuity income from automatic IRAs, according 
to our analysis. 

[Text box] 

Automatic IRA and Saver’s Credit Modeling Scenarios: 

Scenario I: Automatic IRAs. This scenario was drawn from the 
provisions of H.R. 4049, the Automatic IRA Act of 2012 and assumed 
automatic IRAs became effective in 2014.[A] We assumed an aggregate 
participation rate of 69 percent, a default contribution rate of 3 
percent, and a target-date fund as the default investment. Consistent 
with the provisions of the bill, individuals may choose to terminate 
their participation after being enrolled in the plan. Provisions for 
the Saver's Credit matched those of our existing Saver's Credit 
scenario. 

Scenario II: Automatic IRAs plus refundable Saver’s Credit. This 
scenario used the automatic IRAs scenario described above and a 
refundable Saver’s Credit. The Saver’s Credit was drawn from the 
President’s fiscal year 2011 budget proposal. We assumed that the Saver’
s Credit would be refundable; all eligible tax filers would receive a 
50 percent credit rate; and the AGI limits would be increased to 
$85,000 for households filing joint income tax returns, $63,750 for 
those filing as head of household, and $42,500 for those filing 
singly. AGI limits in subsequent years would be indexed to inflation. 
Individuals could receive a refundable credit of up to $1,000 if 
filing singly or $2,000 if filing jointly. The credit would be 
automatically deposited into the recipient’s retirement account and 
the utilization rate was 100 percent. 

Assumptions Applied to Each Scenario: 

We projected the effects of both automatic IRA scenarios for a cohort 
of individuals born in 1995. Cohort members could have worked in 
either the private or public sectors. Our projections show household 
income for retired individuals in the year 2062,b when the cohort is 
67 years old. In our projections, individuals could retire at any time 
after they turned 62. Projections were adjusted for household size and 
reported in 2013 dollars. We analyzed our results by earnings level 
using earnings quartile, which were calculated based on a measure of 
lifetime household earnings. Households with lifetime earnings of $929-
34,377 on an annual basis were in the first quartile,c households with 
earnings of $34,379-60,682 were in the second quartile, households 
with earnings of $60,683-104,440 were in the third quartile, and 
households with earnings of $104,450-3,275,800 were in the fourth 
quartile. 

Account fees varied by investment fund type. We used annual 
nonstochastic nominal rates of return of 9.2 percent for stocks and 
5.7 percent for government bonds.d At retirement, workers used their 
entire defined contribution (DC) and automatic IRA account balances to 
purchase a single-life, non-inflation-adjusted annuity. This annuity 
uses annuity prices that are based on projected mortality rates for 
the 1995 birth cohort and annuity price loading factors that ensure 
that the cost of providing these annuities equals the revenue 
generated by selling them at those prices. Benefits from defined 
benefit (DB) plans are also provided in the form of an annuity. For 
the purposes of this report, we defined the household retirement 
annuity as including the annual income received from all DB and DC 
plans and automatic IRAs. Individuals also received Social Security 
benefits promised under current law. 

Because these scenarios are sensitive to the assumptions used, we ran 
alternate simulations. These simulations used lower aggregate 
participation rates, higher investment fund account fees, or different 
measures for calculating Social Security benefits. 

Our simulations do not include behavioral responses that the 
introduction of automatic IRAs or changes to the Saver’s Credit may 
have created. In reality, some individuals may choose to contribute 
more to their pension plans or may choose to start saving in a pension 
plan or automatic IRA. Further, individuals automatically enrolled in 
automatic IRAs may decide to increase their contributions in the 
future. However, automatic IRA participants may need to choose between 
participating in an automatic IRA or paying down debt. In addition, 
some households may decide to participate in automatic IRAs but borrow 
money to make up for the loss in take-home pay. Also, some of the 
money invested in the automatic IRA may have been invested in other 
vehicles, had the IRA not been available.e 

See appendix II for a detailed explanation of the underlying 
assumptions and projections under our alternate simulations. 

[A] We analyzed H.R. 4049, the Automatic IRA Act of 2012 because it was 
the most recent legislative proposal at the time we conducted our 
analysis. On May 16, 2013, H.R. 2035, the Automatic IRA Act of 2013 was 
introduced before the House of Representatives. The differences between 
H.R. 2035 and H.R. 4049 are minimal and we determined that they would 
not have a significant effect on our simulation results. 

[B] In our simulations, retirement occurs when an individual leaves the 
labor force and starts claiming a Social Security retirement benefit. 

[C] For example, a household in the first quartile could have had 
lifetime household earnings of $25,000 per year over the course of its 
working years. 

[D] Using different rates of return would result in different DC and 
automatic IRA account balances at retirement and, as a result, the size 
of the household's retirement annuity. Also, because our projections do 
not stochastically model stock returns, assuming a rate of return on 
assets equal to the historical return on stocks does not capture the 
risks associated with stock returns. 

[E] These other savings vehicles may or may not have been tax 
advantaged. 

[End of text box] 

Under automatic IRAs, 36 percent of households could see modest 
increases in their retirement annuities, according to our projections. 
These households include workers who have not had access to a DB or DC 
plan as well as workers who had access to DB or DC plans at some jobs 
but not at others. Specifically, we projected that the median dollar 
increase in a household's annuity was $1,046 (see figure 6).[Footnote 
47] While our projections show that those in the lowest earnings 
quartile saw the smallest median dollar increase in the household's 
annuity ($479), the percent change in the median annuity was 66 
percent. In addition, we estimate that 30 percent of households in 
this quartile saw an increase in their annuity of $1,000 or more. In 
comparison, the percent change in the median retirement annuity for 
households in the second earnings quartile was 16 percent and the 
median dollar increase in the household's retirement annuity was 
$1,043. 

Figure 6: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Percent change in median annual household retirement annuity: 13%; 
Median dollar change in annual household retirement annuity: $1,046. 

1st earnings quartile: 
Percent change in median annual household retirement annuity: 66%; 
Median dollar change in annual household retirement annuity: $479. 

2nd earnings quartile: 
Percent change in median annual household retirement annuity: 16%; 
Median dollar change in annual household retirement annuity: $1,043. 

3rd earnings quartile: 
Percent change in median annual household retirement annuity: 8%; 
Median dollar change in annual household retirement annuity: $1,388. 

4th earnings quartile: 
Percent change in median annual household retirement annuity: 4%; 
Median dollar change in annual household retirement annuity: $1,721. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: These projections show the results for only households that 
benefited from automatic IRAs. We projected the effects of automatic 
IRAs for a cohort of individuals born in 1995. The results shown are 
projections for the year 2062 when the cohort is age 67. We assumed 
that these individuals retired between ages 62 and 67. For the 
purposes of this report, we defined the household retirement annuity 
as including the annual retirement income received from all defined 
contribution or defined benefit plans and automatic IRAs. This annuity 
does not include Social Security benefits. The annuity values are 
adjusted for family size and are presented in 2013 dollars. We 
analyzed our results by earnings level using earnings quartiles, which 
are calculated based on a measure of lifetime household earnings. 
Households with lifetime earnings of $929-34,377 on an annual basis 
were in the first quartile, households with earnings of $34,379-60,682 
were in the second quartile, households with earnings of $60,683-
104,440 were in the third quartile, and households with earnings of 
$104,450-3,275,800 were in the fourth quartile. For the first earnings 
quartile, while the percent change in the median household annuity is 
high, the dollar amounts are small. The median household retirement 
annuity is $1,998 under current law and $3,317 under automatic IRAs. 

[End of figure] 

Almost 75 Percent of Households Could Experience an Increase in 
Retirement Income under Automatic IRAs Combined with a Refundable 
Saver's Credit: 

Almost 75 percent of households could experience an increase in their 
retirement annuity if the Saver's Credit were made refundable at the 
same time automatic IRAs were implemented (see figure 7).[Footnote 48] 
Further, according to our projections, making the Saver's Credit 
refundable in addition to implementing automatic IRAs could double the 
number of households experiencing an increase in retirement income. 
[Footnote 49] Under this scenario, many more households could benefit 
than under automatic IRAs alone. Retirement annuities could increase 
if the household participated in automatic IRAs, was eligible to 
receive the refundable Saver's Credit but not the existing Saver's 
Credit, received a larger credit under the refundable Saver's Credit, 
or some combination of these events. Further, under the proposals we 
used for our modeling, the refundable Saver's Credit would be 
available for certain households with AGI of up to $85,000 a year, 
resulting in all lower and many middle-income households seeing an 
increase in their retirement savings. Households in the two lowest 
earnings quartiles could be most likely to benefit, with about 80 
percent of households seeing an increase in their retirement 
annuities, according to our projections. 

Figure 7: Projected Percent of Households That Could See an Increase 
in Retirement Annuity under Automatic IRAs and a Refundable Saver's 
Credit: 

[Refer to PDF for image: horizontal bar graph] 

Percentage of households: 

Total: 
Automatic IRAs: 36%; 
Automatic IRAs and Refundable Saver’s Credit: 74%. 

1st earnings quartile: 
Automatic IRAs: 37%; 
Automatic IRAs and Refundable Saver’s Credit: 78%. 

2nd earnings quartile: 
Automatic IRAs: 38%; 
Automatic IRAs and Refundable Saver’s Credit: 82%. 

3rd earnings quartile: 
Automatic IRAs: 37%; 
Automatic IRAs and Refundable Saver’s Credit: 75%. 

4th earnings quartile: 
Automatic IRAs: 31%; 
Automatic IRAs and Refundable Saver’s Credit: 61%. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: We projected the effects of automatic IRAs and a refundable 
Saver's Credit for a cohort of individuals born in 1995. Results shown 
are projections for the year 2062 when the cohort is age 67. We 
assumed that these individuals retired between ages 62 and 67. For the 
purposes of this report, we defined the household retirement annuity 
as including the annual retirement income received from all defined 
contribution or defined benefit plans and automatic IRAs. This annuity 
does not include Social Security benefits. We analyzed our results by 
earnings level using earnings quartiles, which are calculated based on 
a measure of lifetime household earnings. Households with lifetime 
earnings of $929-34,377 on an annual basis were in the first quartile, 
households with earnings of $34,379-60,682 were in the second 
quartile, households with earnings of $60,683-104,440 were in the 
third quartile, and households with earnings of $104,450-3,275,800 
were in the fourth quartile. 

[End of figure] 

Our projections show that households in the lowest earnings quartile 
could gain the most from the combination of automatic IRAs and a 
refundable Saver's Credit. For these households, the percent increase 
in the median retirement annuity could be 21 percent, compared to a 
potential 2 percent increase for households in the highest quartile 
(see figure 8). Furthermore, 48 percent of households in the lowest 
earnings quartile could see an increase in their retirement annuities 
of $1,000 or more. 

Figure 8: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs and Refundable Saver's Credit: 

[Refer to PDF for image: horizontal bar graph] 

Percentage of households: 

Total: 
Percent change in median annual household retirement annuity: 8%; 
Median dollar change in annual household retirement annuity: $1,060. 

1st earnings quartile: 
Percent change in median annual household retirement annuity: 21%; 
Median dollar change in annual household retirement annuity: $922. 

2nd earnings quartile: 
Percent change in median annual household retirement annuity: 9%; 
Median dollar change in annual household retirement annuity: $1,300. 

3rd earnings quartile: 
Percent change in median annual household retirement annuity: 4%; 
Median dollar change in annual household retirement annuity: $1,060. 

4th earnings quartile: 
Percent change in median annual household retirement annuity: 2%; 
Median dollar change in annual household retirement annuity: $953. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: These projections show the results for only households that 
benefited from automatic IRAs. We projected the effects of automatic 
IRAs and a refundable Saver's Credit for a cohort of individuals born 
in 1995. The results shown are projections for the year 2062 when the 
cohort is age 67. We assumed that these individuals retired between 
ages 62 and 67. For the purposes of this report, we defined the 
household retirement annuity as including the annual retirement income 
received from all defined contribution or defined benefit plans and 
automatic IRAs. This annuity does not include Social Security 
benefits. The annuity values are adjusted for family size and are 
presented in 2013 dollars. We analyzed our results by earnings level 
using earnings quartiles, which are calculated based on a measure of 
lifetime household earnings. Households with lifetime earnings of $929-
34,377 on an annual basis were in the first quartile, households with 
earnings of $34,379-60,682 were in the second quartile, households 
with earnings of $60,683-104,440 were in the third quartile, and 
households with earnings of $104,450-3,275,800 were in the fourth 
quartile. 

[End of figure] 

In general, under our projections, lower-earning households were more 
reliant on Social Security for income in retirement than higher-
earning quartiles (see figure 9). This may be because higher-earning 
households have the ability to save more for retirement, generating a 
larger retirement savings account at retirement. Further, Social 
Security benefits are progressive and replace a larger portion of 
lifetime earnings for people with low earnings than for people with 
high earnings. Thus, higher-earning households may receive a greater 
share of retirement income from annuitizing their retirement savings 
than from their Social Security benefits. The combined effects of 
automatic IRAs and a refundable Saver's Credit could decrease lower-
and middle-income household's reliance on Social Security in 
retirement. Under our projections of automatic IRAs, the two lowest 
earnings quartiles relied on Social Security for 80 and 62 percent of 
their total household income, respectively. However, the percent of 
income derived from a household's retirement annuity increased when we 
added a refundable Saver's Credit. The two lowest earnings quartiles 
relied on retirement annuities for 31 and 46 percent, respectively, of 
their household's total income. 

Figure 9: Projected Median Percent of Household Income Coming from 
Retirement Annuities and Social Security: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Annual total household income under automatic IRAs: $46,667; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 39%; 
Annual household income coming from Social Security: 57%; 
Annual household income coming from earnings: 0%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $54,463; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 50%; 
Annual household income coming from Social Security: 46%; 
Annual household income coming from earnings: 0%. 

1st quartile: 
Annual total household income under automatic IRAs: $20,939; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 18%; 
Annual household income coming from Social Security: 80%; 
Annual household income coming from earnings: 0%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $24,587; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 31%; 
Annual household income coming from Social Security: 66%; 
Annual household income coming from earnings: 0%. 

2nd quartile: 
Annual total household income under automatic IRAs: $39,272; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 35%; 
Annual household income coming from Social Security: 62%; 
Annual household income coming from earnings: 0%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $46,692; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 46%; 
Annual household income coming from Social Security: 51%; 
Annual household income coming from earnings: 0%. 

3rd quartile: 
Annual total household income under automatic IRAs: $65,156; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 47%; 
Annual household income coming from Social Security: 49%; 
Annual household income coming from earnings: 0%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $78,762; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 57%; 
Annual household income coming from Social Security: 40%; 
Annual household income coming from earnings: 0%. 

4th quartile: 
Annual total household income under automatic IRAs: $107,069; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 60%; 
Annual household income coming from Social Security: 36%; 
Annual household income coming from earnings: 0%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $124,906; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 66%; 
Annual household income coming from Social Security: 30%; 
Annual household income coming from earnings: 0%. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: These projections show the results for only households that 
benefited from automatic IRAs. We projected the effects of automatic 
IRAs for a cohort of individuals born in 1995. The results shown are 
projections for the year 2062 when the cohort is age 67. We assumed 
that these individuals retired between ages 62 and 67. For the 
purposes of this report, we defined the household retirement annuity 
as including the annual retirement income received from all defined 
contribution or defined benefit plans and automatic IRAs. This annuity 
does not include Social Security benefits. The annuity values and 
total household incomes are adjusted for family size and are presented 
in 2013 dollars. Median annual household income is the household's 
income in retirement, at age 67. Median annual household income coming 
from earnings was 0 percent across all categories. We analyzed our 
results by earnings level using earnings quartiles, which are 
calculated based on a measure of lifetime household earnings. 
Households with lifetime earnings of $929-34,377 on an annual basis 
were in the first quartile, households with earnings of $34,379-60,682 
were in the second quartile, households with earnings of $60,683-
104,440 were in the third quartile, and households with earnings of 
$104,450-3,275,800 were in the fourth quartile. 

[End of figure] 

Concluding Observations: 

Concerns about retirement security have grown as private sector 
employers have shifted to predominantly sponsoring DC plans. First, 
despite existing tax incentives aimed at fostering plan formation and 
coverage, less than half of workers aged 21-64 were covered by an 
employer-sponsored plan in 2011 and the participation rate has barely 
changed over the last several decades. Second, despite significant tax 
incentives aimed at increasing retirement savings, many households 
enter retirement without adequate financial resources. 

Recent proposals, such as those that would establish the automatic 
IRA, could represent a significant step toward increasing 
participation to over 50 percent within the current framework of a 
voluntary system. Further, redesigning the Saver's Credit could 
improve the retirement income and savings imbalance between lower and 
higher earners by giving lower-and middle-earning households an 
additional incentive to accumulate more in their retirement accounts 
or to start saving for retirement. By itself, increasing coverage 
through automatic IRAs will likely result in, at best, moderate 
increases in retirement income. However, even these increases could 
bolster the financial prospects of many future retirees. These options 
do, of course, pose important trade-offs for individuals, employers, 
and the government that will need to be carefully weighed. 
Individuals, especially those in lower-income households, will still 
have to make difficult choices between spending now and saving for 
later. In addition, the government would forego tax revenue in the 
near-term because contributions to automatic IRAs would be tax-
deferred. Moreover, the expansion of the Saver's Credit would result 
in both a loss of revenue and increased federal spending. Such costs 
will have to be considered along with the proposals' potential effects 
on labor force participation and dependence on other government 
programs, such as Social Security, in retirement. 

Agency Comments and Our Evaluation: 

We provided a draft of this report to the Department of Labor and the 
Department of the Treasury for review and comment. While neither 
agency 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, 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 III. 

Sincerely yours, 

Signed by: 

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

[End of section] 

Appendix I: Objectives, Scope, and Methodology: 

To analyze the extent to which workers with lower earnings can benefit 
from tax incentives for retirement savings and automatic IRAs, we 
examined (1) the earnings and tax rates of households that do not have 
DC plans or IRAs, (2) the effects of the Saver's Credit on retirement 
income, and (3) how automatic IRAs could affect retirement income, 
especially for low-and middle-income workers. This appendix and 
appendix II provide a detailed account of the information and methods 
we used to answer these objectives. Section 1 describes the key 
information sources we used. Section 2 describes the empirical methods 
we used to answer objective 1 and presents information on standard 
errors and confidence intervals for our estimates. 

Section 1: Information Sources: 

To answer our objectives, we obtained information from a variety of 
sources including the Survey of Consumer Finances (SCF); the Policy 
Simulation Group's (PSG) microsimulation models; relevant literature; 
interviews with a range of experts in the area of retirement security; 
presidential and legislative proposals to modify the Saver's Credit 
and create automatic IRAs; and relevant federal laws and regulations. 

SCF: 

To answer the first objective, we used data from the 2010 SCF to 
analyze incomes and tax rates of households that did not take 
advantage of the existing tax incentives for retirement savings. The 
SCF is a triennial, nationally representative survey from the Board of 
Governors of the Federal Reserve. The 2010 SCF surveyed 6,482 
households about their pensions, incomes, labor force participation, 
asset holdings and debts, use of financial services, and demographic 
information. The SCF is conducted using a dual-frame sample design. 
One part of the design is a standard, multistage area-probability 
design, while the second part is a special over-sample of relatively 
wealthy households. This is done in order to accurately capture 
financial information about the population at large as well as 
characteristics specific to the relatively wealthy. The two parts of 
the sample are adjusted for sample nonresponse and combined using 
weights to make estimates from the survey data representative of 
households overall. In addition, the SCF excludes people included in 
the Forbes Magazine list of the 400 wealthiest people in the United 
States. Furthermore, the 2010 SCF dropped 10 observations from the 
public data set that had net worth at least equal to the minimum level 
needed to qualify for the Forbes list. 

For the purposes of this report, a household refers to the primary 
economic unit within a household, which the SCF refers to as a family. 
To estimate the age, marital status, net worth, tax filing status, and 
whether the household had a member with a DC or DB plan, we relied on 
variable definitions used for Federal Reserve publications using the 
SCF.[Footnote 50] To estimate AGI and tax rates, we used TAXSIM, an 
application provided by the National Bureau of Economic Research that 
estimates tax information for households using survey data. To prepare 
the data for TAXSIM, we relied on a program provided by the Federal 
Reserve. We followed SCF documentation and estimated the standard 
errors that incorporate both implicates and replicate weighting. The 
estimated populations we used for our analysis were derived with the 
first implicate. 

The SCF and other surveys that are based on self-reported data are 
subject to several other sources of nonsampling error, including the 
inability to get information about all sample cases; difficulties of 
definition; differences in the interpretation of questions; 
respondents' inability or unwillingness to provide correct 
information; and errors made in collecting, recording, coding, and 
processing data. These nonsampling errors can influence the accuracy 
of information presented in the report, although the magnitude of 
their effect is not known. 

PSG Microsimulation Models: 

To answer objectives 2 and 3, we used PSG's SSASIM, GEMINI, and PENSIM 
simulation models. GEMINI simulates Social Security benefits and taxes 
for large representative samples of people born in the same year. 
GEMINI simulates all types of Social Security benefits including 
retired workers', spouses', survivors', and disability benefits. It 
can be used to model a variety of changes to Social Security. GEMINI 
uses inputs from SSASIM, which has been used in numerous prior GAO 
reports, and PENSIM, which was developed for the Department of Labor. 
GEMINI relies on SSASIM for economic and demographic projections and 
relies on PENSIM for simulated life histories of large representative 
samples of people born in the same year and their spouses.[Footnote 
51] Life histories include educational attainment, labor force 
participation, earnings, job mobility, marriage, disability, 
childbirth, retirement, and death. Life histories are validated 
against data from the Survey of Income and Program Participation, the 
Current Population Survey, Modeling Income in the Near Term 
(MINT),[Footnote 52] and the Panel Study of Income Dynamics. 
Additionally, any projected statistics (such as life expectancy, 
employment patterns, and marital status at age 60) are, where 
possible, consistent with intermediate cost projections from the 
Social Security Administration's Office of the Chief Actuary (OCACT). 
At their best, such models can provide only very rough estimates of 
future incomes. However, these estimates may be useful for comparing 
future incomes across alternative policy scenarios and over time. 

GEMINI can be operated as a free-standing model or it can operate as a 
SSASIM add-on. When operating as an add-on, GEMINI is started 
automatically by SSASIM for one of two purposes. GEMINI can enable the 
SSASIM macro model to operate in the Overlapping Cohorts (OLC) mode or 
it can enable the SSASIM micro model to operate in the Representative 
Cohort Sample (RCS) mode. The SSASIM OLC mode requests GEMINI to 
produce samples for each cohort born after 1934 in order to build up 
aggregate payroll tax revenues and Old-Age, Survivors, and Disability 
Insurance (OASDI) benefit expenditures for each calendar year, which 
are used by SSASIM to calculate standard trust fund financial 
statistics. In either mode, GEMINI operates with the same logic, but 
typically with smaller cohort sample sizes in OLC mode than in the RCS 
or stand-alone-model mode. 

PENSIM simulates the timing for each life event by using data from 
various longitudinal data sets to estimate a waiting-time model (often 
called a hazard function model) using standard survival analysis 
methods. PENSIM incorporates many such estimated waiting-time models 
into a single dynamic simulation model. This model can be used to 
simulate a synthetic sample of complete life histories. PENSIM employs 
continuous-time, discrete-event simulation techniques, such that life 
events do not have to occur at discrete intervals, such as annually on 
a person's birthday. PENSIM also uses macro-demographic and 
macroeconomic variables generated by SSASIM. 

PENSIM imputes pension characteristics using a model estimated with 
1996--1998 establishment data from the Bureau of Labor Statistics 
Employee Benefits Survey (now known as the National Compensation 
Survey). Pension offerings are calibrated to historical trends in 
pension offerings from 1975 to 2005, including plan mix, types of 
plans, and employer matching. Further, PENSIM incorporates data from 
the 1996-1998 Employee Benefits Survey to impute access to and 
participation rates in DC plans in which the employer makes no 
contribution, which the Bureau of Labor Statistics does not report as 
pension plans in the National Compensation Survey. The inclusion of 
these "zero-matching" plans enhances PENSIM's ability to accurately 
reflect the universe of pension plans offered by employers. The 
baseline PENSIM assumption, which we adopted in our analysis, is that 
2005 pension offerings, including the imputed zero-matching plans, are 
projected forward in time. PENSIM also simulates federal income taxes. 

PSG has conducted validation checks of PENSIM's simulated life 
histories against both historical life history statistics and other 
projections. Different life history statistics have been validated 
against data from the Survey of Income and Program Participation, the 
Current Population Survey, MINT, the Panel Study of Income Dynamics, 
and the Social Security Administration's Trustees Report. PSG reports 
that PENSIM life histories have produced similar annual population, 
taxable earnings, and disability benefits for the years 2000 to 2080 
as those produced by the Congressional Budget Office's long-term 
Social Security model and as shown in the Social Security 
Administration's 2004 Trustees Report. According to PSG, PENSIM 
generates simulated DC plan participation rates and account balances 
that are similar to those observed in a variety of data sets. For 
example, measures of central tendency in the simulated distribution of 
DC account balances among employed individuals are similar to those 
produced by an analysis of the Employee Benefit Research Institute-
Investment Company Institute 401(k) database and of the 2004 SCF. We 
performed no independent validation checks of PENSIM's life histories 
or pension characteristics. 

In 2006, the Employee Benefits Security Administration (EBSA) 
submitted PENSIM to a peer review by three economists. The economists' 
overall reviews ranged from highly favorable to highly critical. While 
the economist who gave PENSIM a favorable review expressed a "high 
degree of confidence" in the model, the one who criticized it focused 
on PENSIM's reduced form modeling. This means that the model is 
grounded in previously observed statistical relationships among 
individuals' characteristics, circumstances, and behaviors, rather 
than on any underlying theory of the determinants of behaviors, such 
as the common economic theory that individuals make rational choices 
as their preferences dictate and thereby maximize their own welfare. 
The reduced form modeling approach is used in pension microsimulation 
models and the feasibility of using a nonreduced form approach to 
build such a model may be questionable given the current state of 
economic research. The third economist raised questions about specific 
modeling assumptions and possible overlooked indirect effects. 

Data Reliability: 

We conducted a data reliability assessment of the PSG models and 
selected variables from the SCF by conducting electronic data tests 
for completeness and accuracy, reviewing documentation on the data 
set, and 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 the purposes of our analysis, we found the 
variables that we ultimately reported on to be sufficiently reliable. 

Literature Review and Interviews: 

We conducted an extensive literature review and interviewed a range of 
experts. To identify existing studies, we conducted searches of 
various databases, such as ECO, ArticleFirst, WorldCat, Social 
SciSearch, Harvard Business Review, EconLit, ProQuest, PolicyFile and 
CQ.com. From these sources, we reviewed article abstracts, when 
available, to determine which articles contained information germane 
to our report and reviewed those articles. In addition, we collected 
articles posted on the websites of organizations such as Brookings, 
the Heritage Foundation, and AARP. We performed these searches and 
identified articles from June 2012 through October 2012. 

We also interviewed experts. To ensure we obtained a balanced 
perspective, we interviewed experts with a range of perspectives and 
from different types of organizations including government, research 
organizations, advocacy groups, and the private sector. We also 
conducted interviews with several experts in government and the 
private sector on technical issues related to our analysis. 
Specifically, we interviewed agency officials at the departments of 
the Treasury and Labor; researchers from the Urban Institute, the 
Heritage Foundation, and the Economic Policy Institute; experts and 
advocates from the Pension Rights Center, Demos, Aspen Institute, 
American Society of Pension Professionals and Actuaries, and AARP; and 
private sector professionals from PAi, Prudential Financial, Inc., and 
Putnam Investments. We consulted with officials at the Social Security 
Administration (SSA) and an expert from the PSG on technical issues. 

Section 2: Analysis of Incomes and Tax Rates of Households That Do Not 
Take Advantage of Incentives to Save for Retirement: 

Methodology: 

To analyze the earnings and tax rates of households that do not take 
advantage of the tax incentives for retirement saving, we used the 
2010 SCF. We limited our sample to households where the household head 
was under age 65 and either the respondent or spouse worked in the 
private sector. Our estimates for households only include retirement 
benefits and savings of the survey respondent and a spouse or partner 
and do not include retirement benefits or savings held by additional 
family members. As a result, our estimates may understate retirement 
assets held by a household. We estimated whether the household had a 
DB plan, and AGI and marginal tax rates for households with and 
without DC plans or IRAs. We also estimated net worth using variable 
definitions outlined by the Federal Reserve.[Footnote 53] We estimated 
AGI and marginal tax rates by inputting the SCF data into the National 
Bureau of Economic Research's (NBER) TAXSIM Model, a microsimulation 
model of U.S. federal and state income tax systems.[Footnote 54] 
TAXSIM calculates estimated liabilities under U.S. federal and state 
income tax laws from actual tax returns that have been prepared for 
public use by the Statistics of Income Division of the IRS.[Footnote 
55] 

Confidence Intervals for Our Estimates: 

The SCF is a probability sample based on random selection, so the 2010 
SCF sample is only one of a large number of samples that might have 
been drawn. Since each sample could have provided different estimates, 
we express our confidence in the precision of our particular sample's 
results as a 95 percent confidence interval (e.g., plus or minus 4 
percentage points). This is the interval that would contain the actual 
population value for 95 percent of the samples we could have drawn. As 
a result, we are 95 percent confident that each of the confidence 
intervals in this report will include the true values in the study 
population. Tables 2 through 6 show the confidence intervals for our 
estimates of AGI, marginal tax rates, and retirement plan ownership. 

Table 2: Median AGI for Households With and Without DC Plans or IRAs, 
by Age, 2010: 

Married with DC plan or IRA: 

Age: Less than 35; 
Estimate: $68,000; 
Standard error: $1,969; 
Lower bound of 95 percent confidence interval: $64,141; 
Upper bound of 95 percent confidence interval: $71,859. 

Age: 35-44; 
Estimate: $90,000; 
Standard error: $3,759; 
Lower bound of 95 percent confidence interval: $82,632; 
Upper bound of 95 percent confidence interval: $97,368. 

Age: 45-54; 
Estimate: $100,000; 
Standard error: $2,507; 
Lower bound of 95 percent confidence interval: $95,086; 
Upper bound of 95 percent confidence interval: $104,914. 

Age: 55-64; 
Estimate: $102,000; 
Standard error: $5,181; 
Lower bound of 95 percent confidence interval: $91,845; 
Upper bound of 95 percent confidence interval: $112,155. 

Married without DC plan or IRA: 

Age: Less than 35; 
Estimate: $33,000; 
Standard error: $1,824; 
Lower bound of 95 percent confidence interval: $29,424; 
Upper bound of 95 percent confidence interval: $36,576. 

Age: 35-44; 
Estimate: $44,000; 
Standard error: $2,027; 
Lower bound of 95 percent confidence interval: $40,027; 
Upper bound of 95 percent confidence interval: $47,973. 

Age: 45-54; 
Estimate: $51,000; 
Standard error: $2,901; 
Lower bound of 95 percent confidence interval: $45,315; 
Upper bound of 95 percent confidence interval: $56,685. 

Age: 55-64; 
Estimate: $44,000; 
Standard error: $3,736; 
Lower bound of 95 percent confidence interval: $36,677; 
Upper bound of 95 percent confidence interval: $51,323. 

Not married with DC plan or IRA: 

Age: Less than 35; 
Estimate: $42,000; 
Standard error: $2,447; 
Lower bound of 95 percent confidence interval: $37,203; 
Upper bound of 95 percent confidence interval: $46,797. 

Age: 35-44; 
Estimate: $50,000; 
Standard error: $2,789; 
Lower bound of 95 percent confidence interval: $44,533; 
Upper bound of 95 percent confidence interval: $55,467. 

Age: 45-54; 
Estimate: $48,000; 
Standard error: $1,877; 
Lower bound of 95 percent confidence interval: $44,321; 
Upper bound of 95 percent confidence interval: $51,679. 

Age: 55-64; 
Estimate: $53,000; 
Standard error: $3,237; 
Lower bound of 95 percent confidence interval: $46,655; 
Upper bound of 95 percent confidence interval: $59,345. 

Not married without DC plan or IRA: 

Age: Less than 35; 
Estimate: $18,200; 
Standard error: $1,034; 
Lower bound of 95 percent confidence interval: $16,173; 
Upper bound of 95 percent confidence interval: $20,227. 

Age: 35-44; 
Estimate: $27,000; 
Standard error: $1,451; 
Lower bound of 95 percent confidence interval: $24,157; 
Upper bound of 95 percent confidence interval: $29,843. 

Age: 45-54; 
Estimate: $28,000; 
Standard error: $1,563; 
Lower bound of 95 percent confidence interval: $24,937; 
Upper bound of 95 percent confidence interval: $31,063. 

Age: 55-64; 
Estimate: $31,000; 
Standard error: $2,817; 
Lower bound of 95 percent confidence interval: $25,478; 
Upper bound of 95 percent confidence interval: $36,522. 

Source: GAO analysis of the 2010 SCF. 

Note: Results are for households where the household head was under 
the age of 65 and either the respondent or spouse worked in the 
private sector. 

[End of table] 

Table 3: Median AGI for Households With and Without DC Plans or IRAs, 
by Tax Filing Status, 2010: 

Married with DC plan or IRA: 

Tax filing status: Single; 
Estimate: $48,000; 
Standard error: $4,661; 
Lower bound of 95 percent confidence interval: $38,864; 
Upper bound of 95 percent confidence interval: $57,136. 

Tax filing status: Married filing jointly; 
Estimate: $90,000; 
Standard error: $1,360; 
Lower bound of 95 percent confidence interval: $87,335; 
Upper bound of 95 percent confidence interval: $92,665. 

Married without DC plan or IRA: 

Tax filing status: Single; 
Estimate: $25,000; 
Standard error: $3,368; 
Lower bound of 95 percent confidence interval: $18,398; 
Upper bound of 95 percent confidence interval: $31,602. 

Tax filing status:Married filing jointly; 
Estimate: $42,000; 
Standard error: $1,768; 
Lower bound of 95 percent confidence interval: $38,534; 
Upper bound of 95 percent confidence interval: $45,466. 

Not married with DC plan or IRA: 

Tax filing status: Single; 
Estimate: $48,000; 
Standard error: $1,821; 
Lower bound of 95 percent confidence interval: $44,432; 
Upper bound of 95 percent confidence interval: $51,568. 

Tax filing status: Head of household; 
Estimate: $46,000; 
Standard error: $1,613; 
Lower bound of 95 percent confidence interval: $42,838; 
Upper bound of 95 percent confidence interval: $49,162. 

Not married without DC plan or IRA: 

Tax filing status: Single; 
Estimate: $25,000; 
Standard error: $415; 
Lower bound of 95 percent confidence interval: $24,187; 
Upper bound of 95 percent confidence interval: $25,813. 

Tax filing status: Head of household; 
Estimate: $20,000; 
Standard error: $1,050; 
Lower bound of 95 percent confidence interval: $17,942; 
Upper bound of 95 percent confidence interval: $22,058. 

Source: GAO analysis of the 2010 SCF. 

Notes: Results are for households where the household head was under 
the age of 65 and either the respondent or spouse worked in the 
private sector. The Federal Reserve provides a tax filing status 
variable, which we used in our analysis. For the married filing 
jointly category, the Federal Reserve included married households that 
filed income taxes jointly and separately. For married households, the 
single tax filing status category included households where only the 
respondent or their spouse filed taxes and the filer claimed one 
personal exemption. There are about 48 million married households in 
the married filing jointly category and about one million married 
households in the single tax filing status category. 

[End of table] 

Table 4: Median Marginal Tax Rate for Households With and Without DC 
Plans or IRAs, by Age, 2010: 

Married with DC plan or IRA: 

Age: Less than 35; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Age: 35-44; 
Estimate: 25%; 
Standard error: 3%; 
Lower bound of 95 percent confidence interval: 19%; 
Upper bound of 95 percent confidence interval: 31%. 

Age: 45-54; 
Estimate: 25%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 25%; 
Upper bound of 95 percent confidence interval: 25%. 

Age: 55-64; 
Estimate: 25%; 
Standard error: 0.1%; 
Lower bound of 95 percent confidence interval: 25%; 
Upper bound of 95 percent confidence interval: 25%. 

Married without DC plan or IRA: 

Age: Less than 35; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Age: 35-44; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Age: 45-54; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Age: 55-64; 
Estimate: 15%; 
Standard error: 0.2%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Not married with DC plan or IRA: 

Age: Less than 35; 
Estimate: 15%; 
Standard error: 2%; 
Lower bound of 95 percent confidence interval: 12%; 
Upper bound of 95 percent confidence interval: 18%. 

Age: 35-44; 
Estimate: 25%; 
Standard error: 4%; 
Lower bound of 95 percent confidence interval: 17%; 
Upper bound of 95 percent confidence interval: 33%. 

Age: 45-54; 
Estimate: 22%; 
Standard error: 5%; 
Lower bound of 95 percent confidence interval: 13%; 
Upper bound of 95 percent confidence interval: 31%. 

Age: 55-64; 
Estimate: 25%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 23%; 
Upper bound of 95 percent confidence interval: 27%. 

Not married without DC plan or IRA: 

Age: Less than 35; 
Estimate: 15%; 
Standard error: 2%; 
Lower bound of 95 percent confidence interval: 11%; 
Upper bound of 95 percent confidence interval: 19%. 

Age: 35-44; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Age: 45-54; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Age: 55-64; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Source: GAO analysis of the 2010 SCF. 

Notes: Results are for households where the household head was under 
the age of 65 and either the respondent or spouse worked in the 
private sector. The standard error of our estimate of the midpoint was 
larger when there were multiple modes of the data--when two different 
marginal tax rates were close to being equally likely. A marginal tax 
rate is the rate of tax paid on the next dollar of income that a 
taxpayer earns. 

[End of table] 

Table 5: Median Marginal Tax Rate for Households With and Without DC 
Plans or IRAs, by Tax Filing Status, 2010: 

Married with DC plan or IRA: 

Tax filing status: Single; 
Estimate: 25%; 
Standard error: 4%; 
Lower bound of 95 percent confidence interval: 16%; 
Upper bound of 95 percent confidence interval: 34%. 

Tax filing status: Married filing jointly; 
Estimate: 25%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 22%; 
Upper bound of 95 percent confidence interval: 28%. 

Married without DC plan or IRA: 

Tax filing status: Single; 
Estimate: 15%; 
Standard error: 0.2%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Tax filing status: Married filing jointly; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Not married with DC plan or IRA: 

Tax filing status: Single; 
Estimate: 25%; 
Standard error: 2%; 
Lower bound of 95 percent confidence interval: 22%; 
Upper bound of 95 percent confidence interval: 28%. 

Tax filing status: Head of household; 
Estimate: 15%; 
Standard error: 0.2%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Not married without DC plan or IRA: 

Tax filing status: Single; 
Estimate: 15%; 
Standard error: 0.0%; 
Lower bound of 95 percent confidence interval: 15%; 
Upper bound of 95 percent confidence interval: 15%. 

Tax filing status: Head of household; 
Estimate: 9%; 
Standard error: 3%; 
Lower bound of 95 percent confidence interval: 3%; 
Upper bound of 95 percent confidence interval: 15%. 

Source: GAO analysis of the 2010 SCF. 

Notes: Results are for households where the household head was under 
the age of 65 and either the respondent or spouse worked in the 
private sector. The standard error of our estimate of the midpoint was 
larger when there were multiple modes of the data--when two different 
marginal tax rates were close to being equally likely. The Federal 
Reserve provides a tax filing status variable, which we used in our 
analysis. For the married filing jointly category, the Federal Reserve 
included married households that filed income taxes jointly and 
separately. For married households, the single tax filing status 
category included households where only the respondent or their spouse 
filed taxes and the filer claimed one personal exemption. There are 
about 48 million married households in the married filing jointly 
category and about one million married households in the single tax 
filing status category. A marginal tax rate is the rate of tax paid on 
the next dollar of income that a taxpayer earns. 

[End of table] 

Table 6: Percent of Households Without a DC Plan or IRA, by Whether 
They Have a DB Plan, 2010: 

Married, has DB: 

Income quartile: 1st quartile; 
Estimate: 5%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 3%; 
Upper bound of 95 percent confidence interval: 7%. 

Income quartile: 2nd quartile; 
Estimate: 21%; 
Standard error: 3%; 
Lower bound of 95 percent confidence interval: 16%; 
Upper bound of 95 percent confidence interval: 26%. 

Income quartile: 3rd quartile; 
Estimate: 27%; 
Standard error: 3%; 
Lower bound of 95 percent confidence interval: 20%; 
Upper bound of 95 percent confidence interval: 33%. 

Income quartile: 4th quartile; 
Estimate: 31%; 
Standard error: 6%; 
Lower bound of 95 percent confidence interval: 20%; 
Upper bound of 95 percent confidence interval: 42%. 

Total: 
Estimate: 15%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 13%; 
Upper bound of 95 percent confidence interval: 18%. 

Married, does not have a DB: 

Income quartile: 1st quartile; 
Estimate: 95%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 93%; 
Upper bound of 95 percent confidence interval: 97%. 

Income quartile: 2nd quartile; 
Estimate: 79%; 
Standard error: 3%; 
Lower bound of 95 percent confidence interval: 74%; 
Upper bound of 95 percent confidence interval: 84%. 

Income quartile: 3rd quartile; 
Estimate: 73%; 
Standard error: 3%; 
Lower bound of 95 percent confidence interval: 67%; 
Upper bound of 95 percent confidence interval: 80%. 

Income quartile: 4th quartile; 
Estimate: 69%; 
Standard error: 6%; 
Lower bound of 95 percent confidence interval: 58%; 
Upper bound of 95 percent confidence interval: 80%. 

Total: 
Estimate: 85%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 82%; 
Upper bound of 95 percent confidence interval: 87%. 

Not married, has DB: 

Income quartile: 1st quartile; 
Estimate: 4%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 3%; 
Upper bound of 95 percent confidence interval: 6%. 

Income quartile: 2nd quartile; 
Estimate: 9%; 
Standard error: 2%; 
Lower bound of 95 percent confidence interval: 6%; 
Upper bound of 95 percent confidence interval: 12%. 

Income quartile: 3rd quartile; 
Estimate: 25%; 
Standard error: 4%; 
Lower bound of 95 percent confidence interval: 17%; 
Upper bound of 95 percent confidence interval: 32%. 

Income quartile: 4th quartile; 
Estimate: 25%; 
Standard error: 4%; 
Lower bound of 95 percent confidence interval: 17%; 
Upper bound of 95 percent confidence interval: 33%. 

Total: 
Estimate: 11%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 9%; 
Upper bound of 95 percent confidence interval: 13%. 

Not married, does not have a DB: 

Income quartile: 1st quartile; 
Estimate: 96%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 94%; 
Upper bound of 95 percent confidence interval: 97%. 

Income quartile: 2nd quartile; 
Estimate: 91%; 
Standard error: 2%; 
Lower bound of 95 percent confidence interval: 88%; 
Upper bound of 95 percent confidence interval: 94%. 

Income quartile: 3rd quartile; 
Estimate: 75%; 
Standard error: 4%; 
Lower bound of 95 percent confidence interval: 68%; 
Upper bound of 95 percent confidence interval: 83%. 

Income quartile: 4th quartile; 
Estimate: 75%; 
Standard error: 4%; 
Lower bound of 95 percent confidence interval: 67%; 
Upper bound of 95 percent confidence interval: 83%. 

Total: 
Estimate: 89%; 
Standard error: 1%; 
Lower bound of 95 percent confidence interval: 87%; 
Upper bound of 95 percent confidence interval: 91%. 

Source: GAO analysis of the 2010 SCF. 

Notes: Results are for households where the household head was under 
the age of 65 where either the respondent or spouse work in the 
private sector and do not save for retirement through a DC or IRA 
plan. We analyzed our results by earnings level using earnings 
quartiles. Households with adjusted gross incomes (adjusted for family 
size) of less than $18,960 were in the first earnings quartile, 
households with incomes of $18,961-$33,519 were in the second earnings 
quartile, households with incomes of $33,52 -$56,559 were in the third 
earnings quartile, and households with incomes of $56,560-$25,583,473 
were in the fourth earnings quartile. 

[End of table] 

[End of section] 

Appendix II: Sensitivity Analysis and Summary Statistics for Saver's 
Credit and Automatic IRA Modeling Scenarios: 

This appendix describes the modeling scenarios and assumptions for 
objectives 2 and 3. It also presents the results from our sensitivity 
analyses and describes cohort summary statistics resulting from our 
simulations. 

Modeling Scenarios and Assumptions: 

To analyze the effects of the Saver's Credit and automatic IRAs on 
retirement income, we used the PSG models. We started with a 2 percent 
sample of a 1995 cohort, totaling 118,142 people at birth. We 
projected income in retirement when the cohort is age 67. 

Our simulations included some of the following key assumptions and 
features: 

* Individuals who died before they retire, died before age 67, retire 
after age 67, immigrated into the cohort at an age older than 25, 
emigrated before age 67, or became permanently disabled before age 62 
were omitted from the sample. We did not include individuals who 
became permanently disabled before age 62 because we could not account 
for missed career growth and opportunities to participate in employer-
sponsored pension plans. 

* Retirement occurred as early as age 62. 

* Anyone who became disabled at age 62 or older was considered retired. 

* Workers could be covered by DB plans. We relied on PENSIM's defaults 
to determine DB plan coverage and benefit amounts. 

* Rates of return were fixed. The annual nonstochastic nominal rate of 
return was 9.2 percent for stocks and 5.7 percent for government 
bonds. Using different rates of return would result in different 
effects on DC and automatic IRA account balances at retirement and, as 
a result, the size of the household's retirement annuity. Also, 
because our projections did not stochastically model stock returns, 
assuming a rate of return on assets equal to the historical return on 
stocks did not capture the risks associated with stock returns. 
Further, the nominal rate of return for stocks is based in part on a 
long-term equity risk premium calculated in 2000.[Footnote 56] 

* PENSIM assigned each individual one of four "lifetime asset 
allocation styles." DC assets were invested according to the 
individual's asset allocation style. These styles included (1) all 
assets are invested in a diversified-equity fund, (2) all assets are 
invested in a government bond fund, (3) 15 percent of assets are 
invested in a collection of individuals stocks and the remaining 85 
percent are invested in an age-specific mixture of a diversified-
equity fund and a government-bond fund, and (4) 15 percent of assets 
are invested in a collection of individual stocks and the remaining 85 
percent are invested in a target-date fund. 

* Account fees for investment funds were 75 basis points for target-
date funds; 100 basis points for diversified-equity funds; 45 basis 
points for money-market funds, equity-index, and government bond 
funds; and 0 basis points for stable-value and guaranteed-return 
funds. We considered these account fees to be our "standard fees" 
scenario. 

* When workers change jobs, they either roll over their DC account 
balances or cash them out. Whether workers roll over their DC account 
balances or cash them out depends on the relative size of their 
account balances. We relied on PENSIM's defaults to determine whether 
workers rolled over their account balances. PENSIM does not allow for 
hardship withdrawals. Workers accumulated DC and IRA savings from past 
jobs in one rollover account, which continued to receive investment 
returns. At retirement, benefits were consolidated into one account. 

* Workers used their entire DC account balance at retirement to 
purchase a single-life annuity at retirement that was not adjusted for 
inflation. DB benefits were also provided in the form of an annuity. 
DB and DC annuities were combined into one retirement annuity. 

* Annuity prices were based on projected mortality rates for the 1995 
birth cohort and on loading factors that ensured that the cost of 
providing annuities in PENSIM equaled the revenue generated by selling 
them at those prices. We assume that the annuity provider had no 
administrative or marketing costs, no costs in acquiring the capital 
it needs to hold in reserve, and earns no profits. 

* Eligible individuals received Social Security benefits in 
retirement. According to 2012 projections of the Social Security 
Trustees for the next 75 years (2012-2087), revenues will not be 
adequate to pay full benefits as defined by the current benefit 
formula. Therefore, estimating future Social Security benefits should 
reflect that actuarial deficit and account for the fact that some 
combination of benefit reductions and revenue increases will be 
necessary to restore long-term solvency. Our tax-increase-only 
benchmark simulated "promised benefits," or those benefits promised by 
the current benefit formula. We also developed an alternative benefit-
reduction-only benchmark which simulated "funded benefits," or those 
benefits for which currently scheduled revenues are projected to be 
sufficient (see Sensitivity Analysis below). Our tax-increase-only 
benchmark raised payroll taxes once and immediately by the amount of 
Social Security's actuarial deficit as a percentage of payroll. It 
resulted in the smallest ultimate tax rate of those we considered and 
spread the tax burden most evenly across generations; this was the 
primary basis for our selection. The later that taxes are increased, 
the higher the ultimate tax rate needed to achieve solvency, and, in 
turn, the higher the tax burden on later taxpayers and lower on 
earlier taxpayers. Still, any policy scenario that achieves 75-year 
solvency only by increasing revenues would have the same effect on the 
adequacy of future benefits in that promised benefits would not be 
reduced. Nevertheless, alternative approaches to increasing revenues 
could have very different effects on individual equity. All estimates 
related to this benchmark were simulated using the SSASIM OLC mode. 
Starting in 2013, we increased the tax rate by 2.61 percent in order 
to achieve 75-year solvency. As reported in the Social Security 
Administration's 2012 Trustees Report, immediately raising taxes by 
2.61 percent would achieve solvency.[Footnote 57] 

In our policy scenarios, we varied other assumptions to see how these 
variations affect retirement income at age 67. We did not account for 
any behavioral responses that changes in policy may have created. In 
reality, some individuals may choose to contribute more to their 
pension plans or may choose to start saving in a pension plan or IRA 
in response to policy changes. Policy scenarios we analyzed include: 

* No Saver's Credit. In addition to the assumptions described above, 
this scenario assumed no Saver's Credit's was available. This scenario 
served as the baseline against which we compared the results for the 
existing and refundable Saver's Credit scenarios. 

* Existing Saver's Credit. This scenario was drawn from the existing 
design of the Saver's Credit as established by the Economic Growth and 
Tax Relief Reconciliation Act of 2001. Under the credit's current 
design, taxpayers may receive a non-refundable credit of up to $1,000 
($2,000 if married filing jointly) for contributing to a pension plan 
or IRA. The credit amount is equal to the credit rate multiplied by 
the amount of qualified contribution (of up to $2,000 per individual). 
Taxpayers receive a credit rate of 50, 20, or 10 percent depending on 
AGI and filing status. Certain students, individuals under age 18, and 
those who are claimed as a dependent on another taxpayer's return are 
not eligible to receive the credit. If a worker receives a pre-
retirement distribution from a pension plan, any credit received in 
that year and the subsequent two years will be reduced by the amount 
of the distribution. We used information from the Internal Revenue 
Service to determine current AGI limits. 2013 AGI limits are: $59,000 
for individuals with a filing status of married filing jointly, 
$44,250 for individuals with a filing status of head of household, and 
$29,500 for individuals with a filing status of single, married filing 
separately, or widow(er). Limits in subsequent years were indexed to 
inflation. PENSIM assumed the credit was deposited directly into the 
taxpayer's DC account. This is not a requirement under law, and in 
reality the credit would be part of the household's tax refund. We 
assumed a 100 percent utilization rate, meaning that all eligible 
taxpayers received the credit, to show the maximum potential 
accumulation for retirement income from the Saver's Credit. This 
utilization rate also reflects PENSIM's default values for 2012 and 
beyond. We did not assess behavioral responses related to the 
utilization of the Saver's Credit. 

* Refundable Saver's Credit. This scenario had the same eligibility 
requirements and 2013 AGI limits as the existing Saver's Credit 
scenario. Starting in 2014, the following changes were implemented, 
drawn from our understanding of modifications included in the 
President's budget proposal of fiscal year 2011: (1) the credit became 
fully refundable, (2) all taxpayers received a 50 percent credit rate 
(there is no phase-out), and (3) AGI limits were increased. The 2014 
AGI limits were: $85,000 for individuals with a filing status of 
married filing jointly, $63,750 for individuals with a filing status 
of head of household, and $42,500 for individuals with a filing status 
of single, married filing separately, or widow(er). Limits in 
subsequent years were indexed to inflation. Individuals could receive 
a refundable credit of up to $1,000 if filing singly or $2,000 if 
filing jointly. PENSIM assumed that the credit is deposited directly 
into the taxpayer's DC account. This change was not proposed in the 
President's budget. We assumed a 100 percent utilization rate, meaning 
that all eligible taxpayers claimed the credit, to show the maximum 
potential accumulation for retirement income from the Saver's Credit. 
We did not assess behavioral responses related to the utilization of 
the Saver's Credit. 

* Automatic IRAs. This scenario is drawn from our understanding of 
H.R. 4049 The Automatic IRA Act of 2012.[Footnote 58] For our 
scenario, we assumed that private sector employers began offering 
automatic IRAs in 2014. Employers who employed more than 10 employees 
and did not offer a DB or DC plan were required to offer automatic 
IRAs to employee ages 18 and over after they have been employed for 90 
days. Contributions were made to traditional IRAs.[Footnote 59] If the 
employee did not select a contribution rate or investment fund, 3 
percent of the employee's salary was contributed to a target-date 
fund. The aggregate participation rate is centered around a target 
participation rate of 69 percent in the year 2035, when the cohort was 
age 40 and in the midst of its prime working years. We selected this 
rate by estimating a hypothetical average participation rate using 
parameters from economic literature and SCF data on households that 
did not currently participate in a DC or IRA plan. Our model 
considered the effects of automatic enrollment in 401(k) plans, the 
effects of an employer match (or a lack there-of), income, and age. 
Individuals may choose to terminate their participation in the 
automatic IRA after being enrolled in the plan. The Saver's Credit 
provisions matched the Existing Saver's Credit scenario described 
above. We assumed that the utilization rate for the credit was 100 
percent to model the full savings potential of the credit. 

* Automatic IRAs and a refundable Saver's Credit. In addition to 
modeling automatic IRAs as described above, we modeled the Saver's 
Credit with the same provisions as in the Refundable Saver's Credit 
scenario, also described above. The refundable Saver's Credit was 
fully refundable, there was no credit rate phase-out, and AGI limits 
were extended starting in 2014. Individuals could receive a refundable 
credit of up to $1,000 if filing singly or $2,000 if filing jointly. 
We assumed that the utilization rate for the refundable Saver's Credit 
was 100 percent. 

Sensitivity Analysis: 

Results Using Means: 

Figures 10 through 13 show the results of our modeling scenarios 
presenting means instead of medians. In general, for dollar changes, 
the means were higher than the medians. 

Figure 10: Projected Annual Changes in Households' Retirement Annuity 
under Saver's Credit: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Percent change in average annual household retirement annuity: 
Existing Saver’s Credit: 1%; 
Refundable Saver’s Credit: 3%; 
Average dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $290; 
Refundable Saver’s Credit: $1,386. 

1st quartile: 
Percent change in average annual household retirement annuity: 
Existing Saver’s Credit: 3%; 
Refundable Saver’s Credit: 12%; 
Average dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $363; 
Refundable Saver’s Credit: $1,552. 

2nd quartile: 
Percent change in average annual household retirement annuity: 
Existing Saver’s Credit: 1%; 
Refundable Saver’s Credit: 6%; 
Average dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $306; 
Refundable Saver’s Credit: $1,687. 

3rd quartile: 
Percent change in average annual household retirement annuity: 
Existing Saver’s Credit: 0.4%; 
Refundable Saver’s Credit: 2%; 
Average dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $219; 
Refundable Saver’s Credit: $1,207. 

4th quartile: 
Percent change in average annual household retirement annuity: 
Existing Saver’s Credit: 0.2%; 
Refundable Saver’s Credit: 1%; 
Average dollar change in annual household retirement annuity: 
Existing Saver’s Credit: $220; 
Refundable Saver’s Credit: $922. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: These projections show the results for only households that 
benefited from the Saver's Credit. We projected the effects of the 
Saver's Credit for a cohort of individuals born in 1995. The results 
shown are projections for the year 2062 when the cohort is age 67. The 
effects were compared to a situation in which the credit did not 
exist. We assumed that these individuals retired between ages 62 and 
67. For the purposes of this report, we defined the household 
retirement annuity as including the annual retirement income received 
from all defined contribution or defined benefit plans. This annuity 
does not include Social Security benefits. The annuity values are 
adjusted for family size and are presented in 2013 dollars. We 
analyzed our results by earnings level using earnings quartiles, which 
are calculated based on a measure of lifetime household earnings. 
Households with lifetime earnings of $929-34,377 on an annual basis 
were in the first quartile, households with earnings of $34,379-60,682 
were in the second quartile, households with earnings of $60,683-
104,440 were in the third quartile, and households with earnings of 
$104,450-3,275,800 were in the fourth quartile. 

[End of figure] 

Figure 11: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Percent change in average annual household retirement annuity: 6%; 
Average dollar change in annual household retirement annuity: $1,842. 

1st quartile: 
Percent change in average annual household retirement annuity: 13%; 
Average dollar change in annual household retirement annuity: $899. 

2nd quartile: 
Percent change in average annual household retirement annuity: 9%; 
Average dollar change in annual household retirement annuity: $1,680. 

3rd quartile: 
Percent change in average annual household retirement annuity: 6%; 
Average dollar change in annual household retirement annuity: $2,249. 

4th quartile: 
Percent change in average annual household retirement annuity: 4%; 
Average dollar change in annual household retirement annuity: $2,688. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: These projections show the results for only households that 
benefited from automatic IRAs. We projected the effects of automatic 
IRAs for a cohort of individuals born in 1995. The results shown are 
projections for the year 2062 when the cohort is age 67. We assumed 
that these individuals retired between ages 62 and 67. For the 
purposes of this report, we defined the household retirement annuity 
as including the annual retirement income received from all defined 
contribution or defined benefit plans and automatic IRAs. This annuity 
does not include Social Security benefits. The annuity values are 
adjusted for family size and are presented in 2013 dollars. We 
analyzed our results by earnings levels using earnings quartiles, 
which are calculated based on a measure of lifetime household 
earnings. Households with lifetime earnings of $929-34,377 on annual 
basis were in the first quartile, households with earnings of $34,379-
60,682 were in the second quartile, households with earnings of 
$60,683-104,440 were in the third quartile, and households with 
earnings of $104,450-3,275,800 were in the fourth quartile. 

[End of figure] 

Figure 12: Projected Annual Changes in Households' Retirement Annuity 
under Automatic IRAs and Refundable Saver's Credit: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Percent change in average annual household retirement annuity: 5%; 
Average dollar change in annual household retirement annuity: $1,880. 

1st quartile: 
Percent change in average annual household retirement annuity: 14%; 
Average dollar change in annual household retirement annuity: $1,497. 

2nd quartile: 
Percent change in average annual household retirement annuity: 8%; 
Average dollar change in annual household retirement annuity: $2,075. 

3rd quartile: 
Percent change in average annual household retirement annuity: 4%; 
Average dollar change in annual household retirement annuity: $1,983. 

4th quartile: 
Percent change in average annual household retirement annuity: 2%; 
Average dollar change in annual household retirement annuity: $1,981. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: These projections show the results for only households that 
benefited from automatic IRAs or the refundable Saver's Credit. We 
projected the effects of automatic IRAs for a cohort of individuals 
born in 1995. The results shown are projections for the year 2062 when 
the cohort is age 67. We assumed that these individuals retired 
between ages 62 and 67. For the purposes of this report, we defined 
the household retirement annuity as including the annual retirement 
income received from all defined contribution or defined benefit plans 
and automatic IRAs. This annuity does not include Social Security 
benefits. The annuity values are adjusted for family size and are 
presented in 2013 dollars. We analyzed our results by earnings level 
using earnings quartiles, which are calculated based on a measure of 
lifetime household earnings. Households with lifetime earnings of $929-
34,377 on an annual basis were in the first quartile, households with 
earnings of $34,379-60,682 were in the second quartile, households 
with earnings of $60,683-104,440 were in the third quartile, and 
households with earnings of $104,450-3,275,800 were in the fourth 
quartile. 

[End of figure] 

Figure 13: Projected Mean Percent of Household Income Coming from 
Retirement Annuities and Social Security: 

[Refer to PDF for image: horizontal bar graph] 

Total: 
Annual total household income under automatic IRAs: $65,819; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 39%; 
Annual household income coming from Social Security: 57%; 
Annual household income coming from earnings: 4%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $75,910; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 46%; 
Annual household income coming from Social Security: 50%; 
Annual household income coming from earnings: 4%. 

1st quartile: 
Annual total household income under automatic IRAs: $24,254; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 25%; 
Annual household income coming from Social Security: 73%; 
Annual household income coming from earnings: 2%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $28,802; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 33%; 
Annual household income coming from Social Security: 64%; 
Annual household income coming from earnings: 2%. 

2nd quartile: 
Annual total household income under automatic IRAs: $46,121; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 35%; 
Annual household income coming from Social Security: 61%; 
Annual household income coming from earnings: 3%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $54,148; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 44%; 
Annual household income coming from Social Security: 53%; 
Annual household income coming from earnings: 3%. 

3rd quartile: 
Annual total household income under automatic IRAs: $76,636; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 44%; 
Annual household income coming from Social Security: 52%; 
Annual household income coming from earnings: 4%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $90,080; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 52%; 
Annual household income coming from Social Security: 44%; 
Annual household income coming from earnings: 4%. 

4th quartile: 
Annual total household income under automatic IRAs: $126,873; 
Automatic IRAs: 
Annual household income coming from retirement annuity: 54%; 
Annual household income coming from Social Security: 41%; 
Annual household income coming from earnings: 5%; 
Annual total household income under automatic IRAs and a refundable 
Saver’s Credit: $148,597; 
Automatic IRAs and refundable Saver’s Credit: 
Annual household income coming from retirement annuity: 60%; 
Annual household income coming from Social Security: 35%; 
Annual household income coming from earnings: 5%. 

Source: GAO analysis using the Policy Simulation Group's 
microsimulation models. 

Notes: Sum totals may not equal 100 percent due to rounding. These 
projections show the results for only households that benefited from 
automatic IRAs. We projected the effects of automatic IRAs for a 
cohort of individuals born in 1995. The results shown are projections 
for the year 2062 when the cohort is age 67. We assumed that these 
individuals retired between ages 62 and 67. For the purposes of this 
report, we defined the household retirement annuity as including the 
annual retirement income received from all defined contribution or 
defined contribution plans and automatic IRAs. This annuity does not 
include Social Security benefits. The annuity values and total 
household incomes are adjusted for family size and are presented in 
2013 dollars. Median annual household income is the household's income 
in retirement, at age 67. We analyzed our results by earnings level 
using earnings quartiles, which are calculated based on a measure of 
lifetime household earnings. Households with lifetime earnings of $929-
34,377 on an annual basis were in the first quartile, households with 
earnings of $34,379-60,682 were in the second quartile, households 
with earnings of $60,683-104,440 were in the third quartile, and 
households with earnings of $104,450-3,275,800 were in the fourth 
quartile. 

[End of figure] 

Results Using Alternative Assumptions: 

In addition, we modeled alternative scenarios to examine how sensitive 
our results were to our assumptions. Specifically, we ran the PSG 
models using the following alternative assumptions: 

* Alternative assumptions for investment account fees. Should 
investment account fees be higher than those in our standard fees 
assumption, we established a "high fees" assumption. Account fees for 
target-date funds were 200 basis points; account fees for diversified-
equity funds were 225 basis points; account fees for money-market 
funds, equity-index, and government bond funds remained at 45 basis 
points; and account fees for stable-value and guaranteed-return funds 
remained at 0 basis points. 

* Alternative assumptions for Social Security benefits. Our benefit-
reduction-only benchmark simulated "funded benefits," or those 
benefits for which currently scheduled revenues are projected to be 
sufficient. Under this scenario, the benefit reduction did not begin 
until 2018 to give workers and Social Security beneficiaries time to 
prepare for the reduction. In 2018, benefits were reduced by 18.2 
percent for new and existing beneficiaries. SSA actuaries scored this 
benchmark and determined it would achieve 75-year solvency. 

* Alternative assumptions for Saver's Credit utilization rate. 
Research on Saver's Credit utilization rate for the first couple of 
years it was available indicates that between 60-67 percent of 
eligible taxpayers claimed the credit.[Footnote 60] We modeled a 
utilization rate of 60 percent to reflect the lower bound of this 
range. This scenario demonstrates the effects of the Saver's Credit if 
utilization is as low as the initial rate at which taxpayers claimed 
the credit. 

* Alternative assumptions for automatic IRA participation rate. We 
modeled an aggregate automatic IRA participation rate of 48 percent to 
demonstrate the effects of automatic IRAs should overall participation 
be lower than expected. This alternative participation rate is similar 
to one study's estimate for the participation rate for new employees 
in retirement plans where participation is voluntary.[Footnote 61] 

The results for all of our modeling scenarios, including those using 
our alternative assumptions, are presented in tables 7 through 11 
below. 

Table 7: Results for Saver's Credit Scenarios With Immediate Tax 
Increase Social Security Benchmark: 

Existing Saver's Credit, standard fees, 100% utilization rate, 
immediate tax increase Social Security benchmark: 

Percent of Individuals who received the Saver's Credit[A]; 
Total: 36%; 
By earnings quartile: 
$929-34,377: 45%; 
$34,379-60,682: 45%; 
$60,683-104,440: 33%; 
$104,450-3,275,800: 20%. 

Median dollar change in household's retirement annuity[B]; 
Total: $103; 
By earnings quartile: 
$929-34,377: $155; 
$34,379-60,682: $111; 
$60,683-104,440: $71; 
$104,450-3,275,800: $68. 

Percent change in median household retirement annuity[B]; 
Total: 1%; 
By earnings quartile: 
$929-34,377: 4%; 
$34,379-60,682: 1%; 
$60,683-104,440: 0.3%; 
$104,450-3,275,800: 0.1%. 

Refundable Saver's Credit, standard fees, 100% utilization rate, 
immediate tax increase Social Security benchmark: 

Percent of Individuals who received the Saver's Credit[A]; 
Total: 50%; 
By earnings quartile: 
$929-34,377: 54%; 
$34,379-60,682: 60%; 
$60,683-104,440: 51%; 
$104,450-3,275,800: 36%. 

Median dollar change in household's retirement annuity[B]; 
Total: $655; 
By earnings quartile: 
$929-34,377: $876; 
$34,379-60,682: $884; 
$60,683-104,440: $499; 
$104,450-3,275,800: $342. 

Percent change in median household retirement annuity[B]; 
Total: 6%; 
By earnings quartile: 
$929-34,377: 15%; 
$34,379-60,682: 6%; 
$60,683-104,440: 2%; 
$104,450-3,275,800: 1%. 

Existing Saver's Credit, high fees, 60% utilization rate, immediate tax 
increase Social Security benchmark: 

Percent of Individuals who received the Saver's Credit[A]; 
Total: 31%; 
By earnings quartile: 
$929-34,377: 41%; 
$34,379-60,682: 40%; 
$60,683-104,440: 26%; 
$104,450-3,275,800: 15%. 

Median dollar change in household's retirement annuity[B]; 
Total: $62; 
By earnings quartile: 
$929-34,377: $85; 
$34,379-60,682: $63; 
$60,683-104,440: $46; 
$104,450-3,275,800: $49. 

Percent change in median household retirement annuity[B]; 
Total: 1%; 
By earnings quartile: 
$929-34,377: 2%; 
$34,379-60,682: 1%; 
$60,683-104,440: 0.2%; 
$104,450-3,275,800: 0.2%. 

Refundable Saver's Credit, high fees, 60% utilization rate, immediate 
tax increase Social Security benchmark: 

Percent of Individuals who received the Saver's Credit[A]; 
Total: 45%; 
By earnings quartile: 
$929-34,377: 51%; 
$34,379-60,682: 56%; 
$60,683-104,440: 44%; 
$104,450-3,275,800: 28%. 

Median dollar change in household's retirement annuity[B]; 
Total: $374; 
By earnings quartile: 
$929-34,377: $460; 
$34,379-60,682: $469; 
$60,683-104,440: $296; 
$104,450-3,275,800: $224. 

Percent change in median household retirement annuity[B]; 
Total: 4%; 
By earnings quartile: 
$929-34,377: 9%; 
$34,379-60,682: 4%; 
$60,683-104,440: 1%; 
$104,450-3,275,800: 1%. 

Source: GAO analysis using the PSG microsimulation models. 

Notes: We projected the effects of the Saver's Credit for a cohort of 
individuals born in 1995. The effects were compared to a situation in 
which the credit did not exist. Results shown are projections for the 
year 2062 when the cohort is age 67. We assumed that these individuals 
retired between ages 62 and 67. For the purposes of this report, we 
defined the household retirement annuity as including the annual 
retirement income received from all DC or DB plans. This annuity does 
not include Social Security benefits. Earnings and annuity values are 
adjusted for family size and are presented in 2013 dollars. We 
analyzed our results by earnings level using earnings quartiles, which 
are calculated based on a measure of lifetime household earnings. 

[A] The PENSIM model does not report whether the spouses of 
individuals born in 1995 received the Saver's Credit. As a result, 
this figure does not represent the percent of households that benefit 
from the credit. 

[B] Calculated for households that received the Saver's Credit only. 

[End of table] 

Table 8: Results for Saver's Credit Scenarios With Standard Fees, 100 
Percent Utilization Rate, and Social Security Benefit Reduction 
Benchmark: 

Existing Saver's Credit, standard fees, 100% utilization rate, benefit 
reduction Social Security benchmark: 

Percent of Individuals who received the Saver's Credit[A]; 
Total: 36%; 
By earnings quartile: 
$929-34,377: 45%; 
$34,379-60,682: 45%; 
$60,683-104,440: 33%; 
$104,450-3,275,800: 20%. 

Median dollar change in household's retirement annuity[B]; 
Total: $103; 
By earnings quartile: 
$929-34,377: $155; 
$34,379-60,682: $111; 
$60,683-104,440: $71; 
$104,450-3,275,800: $68. 

Percent change in median household retirement annuity[B]; 
Total: 1%; 
By earnings quartile: 
$929-34,377: 4%; 
$34,379-60,682: 1%; 
$60,683-104,440: 0.3%; 
$104,450-3,275,800: 0.1%. 

Refundable Saver's Credit, standard fees, 100% utilization rate, 
benefit reduction Social Security benchmark: 

Percent of Individuals who received the Saver's Credit[A]; 
Total: 50%; 
By earnings quartile: 
$929-34,377: 54%; 
$34,379-60,682: 60%; 
$60,683-104,440: 51%; 
$104,450-3,275,800: 36%. 

Median dollar change in household's retirement annuity[B]; 
Total: $655; 
By earnings quartile: 
$929-34,377: $876; 
$34,379-60,682: $884; 
$60,683-104,440: $499; 
$104,450-3,275,800: $342. 

Percent change in median household retirement annuity[B]; 
Total: 6%; 
By earnings quartile: 
$929-34,377: 15%; 
$34,379-60,682: 6%; 
$60,683-104,440: 2%; 
$104,450-3,275,800: 1%. 

Source: GAO analysis using the PSG microsimulation models. 

Notes: We projected the effects of the Saver's Credit for a cohort of 
individuals born in 1995. The results were compared to a situation in 
which the credit did not exist. Results shown are projections for the 
year 2062 when the cohort is age 67. We assumed that these individuals 
retired between ages 62 and 67. For the purposes of this report, we 
defined the household retirement annuity as including the annual 
retirement income received from all DC or DB plans. This annuity does 
not include Social Security benefits. Earnings and annuity values are 
adjusted for family size and are presented in 2013 dollars. We 
analyzed our results by earnings level using earnings quartiles, which 
are calculated based on a measure of lifetime household earnings. 

[A] The PENSIM model does not report whether the spouses of 
individuals born in 1995 received the Saver's Credit. As a result, 
this figure does not represent the percent of households that benefit 
from the credit. 

[B] Calculated for households that received the Saver's Credit only. 

[End of table] 

Table 9: Results for Automatic IRA and Refundable Saver's Credit 
Scenarios With Standard Fees, 69 Percent Automatic IRA Participation 
Rate, and Immediate Tax Increase Social Security Benchmark: 

Automatic IRAs, standard fees, 69% participation rate, immediate tax 
increase Social Security benchmark. 

Median dollar change in household retirement annuity[A]; 
Total: $1,046; 
By earnings quartile: 
$929-34,377: $479; 
$34,379-60,682: $1,043; 
$60,683-104,440: $1,388; 
$104,450-3,275,800: $1,721. 

Percent change in median household retirement annuity[A]; 
Total: 13%; 
By earnings quartile: 
$929-34,377: 66%; 
$34,379-60,682: 16%; 
$60,683-104,440: 8%; 
$104,450-3,275,800: 4%. 

Percent of households that see an increase in retirement annuity; 
Total: 36%; 
By earnings quartile: 
$929-34,377: 37%; 
$34,379-60,682: 38%; 
$60,683-104,440: 37%; 
$104,450-3,275,800: 31%. 

Household's total annual income (median)[A]; 
Total: $46,667; 
By earnings quartile: 
$929-34,377: $20,939; 
$34,379-60,682: $39,272; 
$60,683-104,440: $65,156; 
$104,450-3,275,800: $107,069. 

Percent of household's annual income coming from retirement annuity 
(median)[A]; 
Total: 39%; 
By earnings quartile: 
$929-34,377: 18%; 
$34,379-60,682: 35%; 
$60,683-104,440: 47%; 
$104,450-3,275,800: 60%. 

Percent of household's annual income coming from Social Security 
benefits (median)[A]; 
Total: 57%; 
By earnings quartile: 
$929-34,377: 80%; 
$34,379-60,682: 62%; 
$60,683-104,440: 49%; 
$104,450-3,275,800: 36%. 

Percent of household's annual income coming from earnings (median)[A]; 
Total: 0%; 
By earnings quartile: 
$929-34,377: 0%; 
$34,379-60,682: 0%; 
$60,683-104,440: 0%; 
$104,450-3,275,800: 0%. 

Automatic IRAs, refundable Saver's Credit, standard fees, 69% 
participation rate, immediate tax increase Social Security benchmark: 

Median dollar change in household retirement annuity[B]; 
Total: $1,060; 
By earnings quartile: 
$929-34,377: $922; 
$34,379-60,682: $1,300; 
$60,683-104,440: $1,060; 
$104,450-3,275,800: $953. 

Percent change in median household retirement annuity[B]; 
Total: 8%; 
By earnings quartile: 
$929-34,377: 21%; 
$34,379-60,682: 9%; 
$60,683-104,440: 4%; 
$104,450-3,275,800: 2%. 

Percent of households that see an increase in retirement annuity; 
Total: 74%; 
By earnings quartile: 
$929-34,377: 78%; 
$34,379-60,682: 82%; 
$60,683-104,440: 75%; 
$104,450-3,275,800: 6%1. 

Household's total annual income (median)[B]; 
Total: $54,463; 
By earnings quartile: 
$929-34,377: $24,587; 
$34,379-60,682: $46,692; 
$60,683-104,440: $78,762; 
$104,450-3,275,800: $124,906. 

Percent of household's annual income coming from retirement annuity 
(median)[B]; 
Total: 50%; 
By earnings quartile: 
$929-34,377: 31%; 
$34,379-60,682: 46%; 
$60,683-104,440: 57%; 
$104,450-3,275,800: 66%. 

Percent of household's annual income coming from Social Security 
benefits (median)[B]; 
Total: 46%; 
By earnings quartile: 
$929-34,377: 66%; 
$34,379-60,682: 51%; 
$60,683-104,440: 40%; 
$104,450-3,275,800: 30%. 

Percent of household's annual income coming from earnings (median)[B]; 
Total: 0%; 
By earnings quartile: 
$929-34,377: 0%; 
$34,379-60,682: 0%; 
$60,683-104,440: 0%; 
$104,450-3,275,800: 0%. 

Source: GAO analysis using the PSG microsimulation models. 

Notes: We projected the effects of automatic IRAs and a refundable 
Saver's Credit for a cohort of individuals born in 1995. Results shown 
are projections for the year 2062 when the cohort is age 67. We 
assumed that these individuals retired between ages 62 and 67. For the 
purposes of this report, we defined the household retirement annuity 
as including the annual retirement income received from all DC or DB 
plans and automatic IRAs. This annuity does not include Social 
Security benefits. Earnings, annuity values, and total household 
incomes are adjusted for family size and are presented in 2013 
dollars. We analyzed our results by earnings level using earnings 
quartiles, which are calculated based on a measure of lifetime 
household earnings. 

[A] Calculated for households that participated in automatic IRAs only. 

[B] Calculated for households that participated in automatic IRAs or 
received a refundable Saver's Credit. 

[End of table] 

Table 10: Results for Automatic IRA and Refundable Saver's Credit 
Scenarios With High Fees, 48 Percent Automatic IRA Participation Rate, 
and Immediate Tax Increase Social Security Benchmark: 

Automatic IRAs, high fees, 48% participation rate, immediate tax 
increase Social Security benchmark: 

Median dollar change in household retirement annuity[A]; 
Total: $1,019; 
By earnings quartile: 
$929-34,377: $451; 
$34,379-60,682: $920; 
$60,683-104,440: $1,269; 
$104,450-3,275,800: $1,488. 

Percent change in median household retirement annuity[A]; 
Total: 10%; 
By earnings quartile: 
$929-34,377: 85%; 
$34,379-60,682: 18%; 
$60,683-104,440: 7%; 
$104,450-3,275,800: 3%. 

Percent of households that see an increase in retirement annuity; 
Total: 24%; 
By earnings quartile: 
$929-34,377: 19%; 
$34,379-60,682: 24%; 
$60,683-104,440: 26%; 
$104,450-3,275,800: 25%. 

Household's total annual income (median)[A]; 
Total: $47,568; 
By earnings quartile: 
$929-34,377: $20,564; 
$34,379-60,682: $36,165; 
$60,683-104,440: $59,982; 
$104,450-3,275,800: $96,642. 

Percent of household's annual income coming from retirement annuity 
(median)[A]; 
Total: 35%; 
By earnings quartile: 
$929-34,377: 14%; 
$34,379-60,682: 28%; 
$60,683-104,440: 41%; 
$104,450-3,275,800: 56%. 

Percent of household's annual income coming from Social Security 
benefits (median)[A]; 
Total: 60%% 
By earnings quartile: 
$929-34,377: 84%; 
$34,379-60,682: 69%; 
$60,683-104,440: 55%; 
$104,450-3,275,800: 40%. 

Percent of household's annual income coming from earnings (median)[A]; 
Total: 0%; 
By earnings quartile: 
$929-34,377: 0%; 
$34,379-60,682: 0%; 
$60,683-104,440: 0%; 
$104,450-3,275,800: 0%. 

Automatic IRAs, refundable Saver's Credit, high fees, 48% participation 
rate, immediate tax increase Social Security benchmark: 

Median dollar change in household retirement annuity[B]; 
Total: $757; 
By earnings quartile: 
$929-34,377: $692; 
$34,379-60,682: $914; 
$60,683-104,440: $732; 
$104,450-3,275,800: $666. 

Percent change in median household retirement annuity[B]; 
Total: 7%; 
By earnings quartile: 
$929-34,377: 15%; 
$34,379-60,682: 7%; 
$60,683-104,440: 3%; 
$104,450-3,275,800: 2%. 

Percent of households that see an increase in retirement annuity; 
Total: 69%; 
By earnings quartile: 
$929-34,377: 71%; 
$34,379-60,682: 77%; 
$60,683-104,440: 71%; 
$104,450-3,275,800: 57%. 

Household's total annual income (median)[B]; 
Total: $52,475; 
By earnings quartile: 
$929-34,377: $24,589; 
$34,379-60,682: $44,736; 
$60,683-104,440: $74,454; 
$104,450-3,275,800: $116,181. 

Percent of household's annual income coming from retirement annuity 
(median)[B]; 
Total: 48%; 
By earnings quartile: 
$929-34,377: 31%; 
$34,379-60,682: 44%; 
$60,683-104,440: 54%; 
$104,450-3,275,800: 64%. 

Percent of household's annual income coming from Social Security 
benefits (median)[B]; 
Total: 49%; 
By earnings quartile: 
$929-34,377: 67%; 
$34,379-60,682: 53%; 
$60,683-104,440: 43%; 
$104,450-3,275,800: 33%. 

Percent of household's annual income coming from earnings (median)[B]; 
Total: 0%; 
By earnings quartile: 
$929-34,377: 0%; 
$34,379-60,682: 0%; 
$60,683-104,440: 0%; 
$104,450-3,275,800: 0%. 

Source: GAO analysis using the PSG microsimulation models. 

Notes: We projected the effects of automatic IRAs and a refundable 
Saver's Credit for a cohort of individuals born in 1995. Results shown 
are projections for the year 2062 when the cohort is age 67. We 
assumed that these individuals retired between ages 62 and 67. For the 
purposes of this report, we defined the household retirement annuity 
as including the annual retirement income received from all DC or DB 
plans and automatic IRAs. This annuity does not include Social 
Security benefits. Earnings, annuity values and total household 
incomes are adjusted for family size and are presented in 2013 
dollars. We analyzed our results by earnings level using earnings 
quartiles, which are calculated based on a measure of lifetime 
household earnings. 

[A] Calculated for households that participated in automatic IRAs only. 

[B] Calculated for households that participated in automatic IRAs or 
received a refundable Saver's Credit. 

[End of table] 

Table 11: Results for Automatic IRA and Refundable Saver's Credit 
Scenarios With Standard Fees, 69 Percent Automatic IRA Participation 
Rate, and Social Security Benefit Reduction Benchmark: 

Automatic IRAs, standard fees, 69% participation rate, benefit 
reduction Social Security benchmark: 

Median dollar change in household retirement annuity[A]; 
Total: $1,046; 
By earnings quartile: 
$929-34,377: $479; 
$34,379-60,682: $1,043; 
$60,683-104,440: $1,388; 
$104,450-3,275,800: $1,721. 

Percent change in median household retirement annuity[A]; 
Total: 13%; 
By earnings quartile: 
$929-34,377: 66%; 
$34,379-60,682: 16%; 
$60,683-104,440: 8%; 
$104,450-3,275,800: 4%. 

Percent of households that see an increase in retirement annuity; 
Total: 36%; 
By earnings quartile: 
$929-34,377: 37%; 
$34,379-60,682: 38%; 
$60,683-104,440: 37%; 
$104,450-3,275,800: 31%. 

Household's total annual income (median)[A]; 
Total: $41,462; 
By earnings quartile: 
$929-34,377: $17,859; 
$34,379-60,682: $34,668; 
$60,683-104,440: $59,323; 
$104,450-3,275,800: $99,654. 

Percent of household's annual income coming from retirement annuity 
(median)[A]; 
Total: 44%; 
By earnings quartile: 
$929-34,377: 21%; 
$34,379-60,682: 39%; 
$60,683-104,440: 52%; 
$104,450-3,275,800: 65%. 

Percent of household's annual income coming from Social Security 
benefits (median)[A]; 
Total: 52%; 
By earnings quartile: 
$929-34,377: 77%; 
$34,379-60,682: 57%; 
$60,683-104,440: 44%; 
$104,450-3,275,800: 31%. 

Percent of household's annual income coming from earnings (median)[A]; 
Total: 0%; 
By earnings quartile: 
$929-34,377: 0%; 
$34,379-60,682: 0%; 
$60,683-104,440: 0%; 
$104,450-3,275,800: 0%. 

Automatic IRAs, refundable Saver's Credit, standard fees, 69% 
participation rate, benefit reduction Social Security benchmark: 

Median dollar change in household retirement annuity[B]; 
Total: $1,060; 
By earnings quartile: 
$929-34,377: $922; 
$34,379-60,682: $1,300; 
$60,683-104,440: $1,060; 
$104,450-3,275,800: $953. 

Percent change in median household retirement annuity[B]; 
Total: 8%; 
By earnings quartile: 
$929-34,377: 21%; 
$34,379-60,682: 9%; 
$60,683-104,440: 4%; 
$104,450-3,275,800: 2%. 

Percent of households that see an increase in retirement annuity; 
Total: 74%; 
By earnings quartile: 
$929-34,377: 78%; 
$34,379-60,682: 82%; 
$60,683-104,440: 75%; 
$104,450-3,275,800: 61%. 

Household's total annual income (median)[B]; 
Total: $49,478; 
By earnings quartile: 
$929-34,377: $21,431; 
$34,379-60,682: $42,287; 
$60,683-104,440: $73,037; 
$104,450-3,275,800: $118,051. 

Percent of household's annual income coming from retirement annuity 
(median)[B]; 
Total: 55%; 
By earnings quartile: 
$929-34,377: 36%; 
$34,379-60,682: 51%; 
$60,683-104,440: 61%; 
$104,450-3,275,800: 71%. 

Percent of household's annual income coming from Social Security 
benefits (median)[B]; 
Total: 41%;
By earnings quartile: 
$929-34,377: 61%; 
$34,379-60,682: 46%; 
$60,683-104,440: 35%; 
$104,450-3,275,800: 26%. 

Percent of household's annual income coming from earnings (median)[B]; 
Total: 0%; 
By earnings quartile: 
$929-34,377: 0%; 
$34,379-60,682: 0%; 
$60,683-104,440: 0%; 
$104,450-3,275,800: 0%. 

Source: GAO analysis using the PSG microsimulation models. 

Notes: We projected the effects of automatic IRAs and a refundable 
Saver's Credit for a cohort of individuals born in 1995. Results shown 
are projections for the year 2062 when the cohort is age 67. We 
assumed that these individuals retired between ages 62 and 67. For the 
purposes of this report, we defined the household retirement annuity 
as including the annual retirement income received from all DC or DB 
plans and automatic IRAs. This annuity does not include Social 
Security benefits. Earnings, annuity values and total household 
incomes are adjusted for family size and are presented in 2013 
dollars. We analyzed our results by earnings level using earnings 
quartiles, which are calculated based on a measure of lifetime 
household earnings. 

[A] Calculated for households that participated in automatic IRAs only. 

[B] Calculated for households that participated in automatic IRAs or 
received a refundable Saver's Credit. 

[End of table] 

Cohort Summary Statistics: 

Lifetime summary statistics of the simulated 1995 cohort's workforce 
and demographic variables give some insight into the PSG model's 
projections of income in retirement in our report (see tables 12 and 
13). By restricting the sample to retirees who do not immigrate into 
the cohort after age 25, do not emigrate or die before age 67, and do 
not become permanently disabled before age 62, we reduce the full 
sample of 118,142 individuals to a sample of 60,813 individuals. 

Table 12: Sample Summary Statistics at Age 67, 1995 PENSIM cohort: 

Demographic variables: Total number of cohort members at age 67; 
Full sample: 60,813; 
By earnings quartile: 
$929-34,377: 15,203; 
$34,379-60,682: 15,203; 
$60,683-104,440: 15,204; 
$104,450-3,275,800: 15,203. 

Demographic variables: Percent male; 
Full sample: 48%; 
By earnings quartile: 
$929-34,377: 37%; 
$34,379-60,682: 46%; 
$60,683-104,440: 51%; 
$104,450-3,275,800: 59%. 

Demographic variables: Highest level of schooling (median); 
Full sample: Some college; 
By earnings quartile: 
$929-34,377: High school graduate; 
$34,379-60,682: High school graduate; 
$60,683-104,440: Some college; 
$104,450-3,275,800: College or graduate degree. 

Demographic variables: Percent whose longest-held job offers a pension; 
Full sample: 74%; 
By earnings quartile: 
$929-34,377: 60%; 
$34,379-60,682: 71%; 
$60,683-104,440: 78%; 
$104,450-3,275,800: 85%. 

Demographic variables: Percent who were ever eligible to participate in 
a DC plan; 
Full sample: 91%; 
By earnings quartile: 
$929-34,377: 85%; 
$34,379-60,682: 90%; 
$60,683-104,440: 93%; 
$104,450-3,275,800: 95%. 

Source: GAO analysis using the PSG microsimulation models. 

Notes: The sample excludes cohort members who immigrated into the 
cohort after age 25, emigrated or died before age 67, or became 
permanently disabled before age 62. Earnings are adjusted for family 
size and are presented in 2013 dollars. We analyzed our results by 
earnings level using earnings quartiles, which are calculated based on 
a measure of lifetime household earnings. 

[End of table] 

Table 13: Means at Age 67, 1995 PENSIM cohort: 

Demographic variables: Annual lifetime household earnings; 
Full sample: $82,485; 
By earnings quartile: 
$929-34,377: $22,600; 
$34,379-60,682: $46,670; 
$60,683-104,440: $80,105; 
$104,450-3,275,800: $180,565. 

Demographic variables: Number of years working full time; 
Full sample: 27; 
By earnings quartile: 
$929-34,377: 20; 
$34,379-60,682: 27; 
$60,683-104,440: 29; 
$104,450-3,275,800: 32. 

Demographic variables: Number of years working part time; 
Full sample: 6; 
By earnings quartile: 
$929-34,377: 10; 
$34,379-60,682: 6; 
$60,683-104,440: 5; 
$104,450-3,275,800: 4. 

Demographic variables: Number of jobs held over lifetime; 
Full sample: 5; 
By earnings quartile: 
$929-34,377: 5; 
$34,379-60,682: 5; 
$60,683-104,440: 5; 
$104,450-3,275,800: 5. 

Demographic variables: Duration of longest job (years); 
Full sample: 18; 
By earnings quartile: 
$929-34,377: 17; 
$34,379-60,682: 19; 
$60,683-104,440: 19; 
$104,450-3,275,800: 20. 

Demographic variables: Retirement age; 
Full sample: 63; 
By earnings quartile: 
$929-34,377: 63; 
$34,379-60,682: 63; 
$60,683-104,440: 63; 
$104,450-3,275,800: 63. 

Demographic variables: Number of years vested in a DC plan; 
Full sample: 15; 
By earnings quartile: 
$929-34,377: 11; 
$34,379-60,682: 14; 
$60,683-104,440: 16; 
$104,450-3,275,800: 18. 

Demographic variables: Number of jobs on which eligible for a DC plan; 
Full sample: 2; 
By earnings quartile: 
$929-34,377: 2; 
$34,379-60,682: 2; 
$60,683-104,440: 2; 
$104,450-3,275,800: 3. 

Source: GAO analysis using the PSG microsimulation models. 

Notes: The sample excludes cohort members who immigrated into the 
cohort after age 25, emigrated or died before age 67, or became 
permanently disabled before age 62. Earnings are adjusted for family 
size and are presented in 2013 dollars. We analyzed our results by 
earnings level using earnings quartiles, which are calculated based on 
a measure of lifetime household earnings. 

[End of table] 

[End of section] 

Appendix III: GAO Contact and Staff Acknowledgments: 

GAO Contact: 

Charles A. Jeszeck, (202) 512-7215 or jeszeckc@gao.gov: 

Staff Acknowledgments: 

In addition to the contact named above, Michael Collins, Assistant 
Director; Andrea Dawson, Jennifer Gregory, and Amrita Sen made key 
contributions. Also contributing to this report were Alicia Atkinson, 
Benjamin Bolitzer, Alicia Cackley, Colleen Candrl, David Chrisinger, 
Laura Henry, Sharon Hermes, Gene Kuehneman, Kathy Leslie, Thomas 
McCool, Mimi Nguyen, Marylynn Sergent, Kenneth Stockbridge, Nyree 
Ryder Tee, Roger Thomas, Frank Todisco, and Walter Vance. 

[End of section] 

Footnotes: 

[1] Office of Management and Budget, Fiscal Year 2014 Budget of the 
U.S. Government (Washington, D.C.: April 2013). The tax expenditure is 
measured as the tax revenue that the government does not currently 
collect on contributions and earnings amounts, offset by the taxes 
paid on pensions by those who are currently receiving retirement 
benefits. 

[2] Employer contributions to qualified pension plans are a tax 
deductible business expense. In addition, offering plans allows 
employers to retain employees. 

[3] Craig Copeland, "Employment-Based Retirement Plan Participation: 
Geographic Differences and Trends, 2011," EBRI Issue Brief, no. 378 
(November 2012). 

[4] Alicia Munnell, Anthony Webb, and Francesca Golub-Sass, Is There 
Really a Retirement Savings Crisis? An NRRI Analysis (Boston, MA: 
Center for Retirement Research at Boston College, August 2007); Nadia 
Karamcheva and Geoffrey Sanzenbacher, Is Pension Inequality Growing? 
(Boston, MA: Center for Retirement Research at Boston College, January 
2010). Jack VanDerhei, Retirement Income Adequacy for Boomers and Gen 
Xers: Evidence from the 2012 EBRI Retirement Security Projection 
Model®, Employee Benefit Research Institute (Washington DC, May 2012). 

[5] GAO, Private Pensions: Some Key Features Lead to an Uneven 
Distribution of Benefits, [hyperlink, 
http://www.gao.gov/products/GAO-11-333] (Washington, D.C.: Mar. 30, 
2011). 

[6] Under these proposals, participants may choose to opt out of or 
terminate their participation in automatic IRAs. 

[7] A nonrefundable tax credit can only reduce tax owed to zero. In 
contrast, a refundable tax credit can be used to generate a refund 
payment to the tax filer in excess of taxes paid. 

[8] See the Savings for American Families' Future Act of 2012, H.R. 
6472, 112TH Cong. (2012) and the Savings for American Families' Future 
Act of 2013, H.R. 837, 113TH Cong. (2013). 

[9] We used the PSG's SSASIM, PENSIM, and GEMINI microsimulation 
models. These models analyze lifetime coverage and adequacy issues 
related to employer-sponsored pensions and Social Security in the 
United States. These models can estimate individual effects of policy 
scenarios for a representative sample of future beneficiaries. They 
can also simulate different reform features for their effects on the 
level and distribution of benefits. 

[10] Copeland, "Employment-Based Retirement Plan Participation: 
Geographic Differences and Trends, 2011." 

[11] While the vast majority of IRAs are individual retirement 
accounts, section 408(b) of the Code provides for an "individual 
retirement annuity" which represents an annuity contract or endowment 
issued by an insurance company and meeting certain regulatory 
requirements under 26 C.F.R. § 1.408-3. 

[12] Department of Labor, Employee Benefits Security Administration, 
Private Pension Plan Bulletin Historical Tables and Graphs 
(Washington, D.C.: 2012). Data exclude plans with only one participant. 

[13] Investment Company Institute, The U.S. Retirement Market, First 
Quarter 2013, (June 2013), accessed August 6, 2013, [hyperlink, 
http://www.ici.org/info/ret_13_q1_data.xls]. 

[14] GAO, 401(K) Plans: Labor and IRS Could Improve the Rollover 
Process for Participants, [hyperlink, 
http://www.gao.gov/products/GAO-13-30] (Washington, D.C.: Mar. 7, 
2013). 

[15] Some defined contribution plans may have Roth accounts, where 
participants can make after-tax contributions. 

[16] Pub. L. No. 107-16, 115 Stat. 38. Begun in 2002 as a temporary 
provision, the Saver's Credit was made a permanent part of the tax 
code by the Pension Protection Act of 2006, Pub. L. No. 109-280, § 
812, 120 Stat. 780, 997. 

[17] AGI is defined as gross income minus adjustments to income. 

[18] Taxpayers under 18 years of age, certain students, and those 
claimed as a dependent on another taxpayer's return are not eligible 
for the credit. 

[19] Department of the Treasury estimates of revenue loss for 2013 and 
beyond do not reflect current law tax rates. In contrast to the 
Department of the Treasury's estimate of five-year revenue losses of 
$6.2 billion, the Joint Committee on Taxation estimated, based on 
current law tax rates, that the Saver's Credit will cost $5.3 billion 
for the same period (fiscal years 2013-2017). Joint Committee on 
Taxation, Estimates of Federal Tax Expenditures for Fiscal Years 2012-
2017, JCS-1-13 (Washington, D.C.: Feb. 1, 2013). 

[20] The President's 2011 proposal called for increasing the AGI limit 
to $85,000, but the 2010 proposal would have increased it to $65,000. 
The President's 2011 proposal would also limit the maximum available 
credit to $500 per family. Office of Management and Budget, Executive 
Office of the President of the United States, Budget of the U.S. 
Government, Fiscal Year 2011 (Washington, D.C.: Feb. 1, 2010). For 
this report, we modeled modifications to the Saver's Credit that were 
drawn from the President's 2011 budget because its provisions align 
with possible modifications suggested by experts we identified--
including making the Saver's Credit refundable--that could improve 
retirement savings for low-income households. 

[21] Budget of the U.S. Government, Fiscal Year 2011. The total 
estimated cost includes both reduced revenue as well as outlays for 
the refundable portion of the credit. 

[22] Joint Committee on Taxation, Estimated Budget Effects of the 
Revenue Provisions Contained in the President's Fiscal Year 2011 
Budget Proposal, JCX-7-10R (Washington, D.C.: Mar. 15, 2010). 

[23] See e.g., the Savings for American Families' Future Act of 2012, 
H.R. 6472, 112th Cong (2012), and the Savings for American Families' 
Future Act of 2013, H.R. 837, 113th Cong. (2013). 

[24] GAO, Retirement Savings: Automatic Enrollment Shows Promise for 
Some Workers, but Proposals to Broaden Retirement Savings for Other 
Workers Could Face Challenges, [hyperlink, 
http://www.gao.gov/products/GAO-10-31] (Washington, D.C.: Oct. 23, 
2009). 

[25] Automatic IRA legislative proposals have been introduced in the 
House of Representatives and the Senate in many recent years. The most 
recent legislative proposal for automatic IRAs is The Automatic IRA 
Act of 2013, H.R. 2035, 113th Cong. (2013). However, our analysis is 
based on the provisions of The Automatic IRA Act of 2012, H.R. 4049, 
112th Cong. (2012) because it was the most recent proposal available 
when we conducted our analysis. We determined that the differences 
between H.R. 2035 and H.R. 4049 are minimal and would not have a 
significant effect on our simulation results. 

[26] Employers that do not have more than 10 employees who received at 
least $5,000 of compensation from the employer in the preceding 
calendar year and employers that were not in existence at all times 
during the calendar year and the preceding calendar year would be 
exempt. Under the proposal, employers would be permitted to exclude 
certain employees, including those under the age of 18 and employees 
who have not completed at least 3 months of service with the employer. 

[27] Under this proposal, the Department of the Treasury would issue 
retirement bonds. 

[28] Our prior work has examined the effects of automatic IRAs on 
employees, employers, the federal government, and 401(k) plans, 
including concerns that some employers may have about the 
administrative burden and costs of implementing automatic IRAs. See 
[hyperlink, http://www.gao.gov/products/GAO-10-31]. 

[29] At the time of our analysis, there were no publicly available 
cost estimates for the 2012 legislative proposal for automatic IRAs. 
As a result, the estimates presented here pertain to the President's 
budget proposal regarding automatic IRAs. These estimates include 
doubling of the tax credit for small employer plan start-up costs. See 
Fiscal Year 2014 Budget of the U.S. Government. The Joint Committee on 
Taxation estimated the President's 2014 budget proposal for automatic 
IRAs would cost about $10.6 billion for fiscal years 2014 through 
2023. Joint Committee on Taxation, Estimated Budget Effects of the 
Revenue Provisions Contained in the President's Fiscal Year 2014 
Budget Proposal, JCX-11-13 (Washington, D.C.: May 10, 2013). 

[30] We analyzed households where the household head was under the age 
of 65 and either the respondent or spouse worked in the private sector 
in 2010. We selected households as the unit of analysis because the 
tax incentives for retirement saving we examined and taxes are filed 
at the household level. All estimates from the SCF are subject to 
sampling error. See appendix I for more information on standard errors 
and confidence intervals for SCF estimates. 

[31] Federal marginal tax rates reflect effective tax rates taking 
into account phase-outs and use of other tax benefits, such as the 
earned income tax credit, rather than the scheduled or statutory tax 
rates. 

[32] The median marginal tax rate for married and single households 
less than age 35 is 15 percent, regardless of whether or not the 
household had a DC plan or IRA. However, the average marginal tax rate 
varies. Married households age less than 35 with a DC plan or IRA have 
an average marginal tax rate of 18 percent, compared to 8 percent for 
those without. Single households age less than 35 with a DC plan or 
IRA have an average marginal tax rate of 18 percent, compared to 5 
percent for those without. 

[33] Households with lower earnings have lower tax rates and generally 
have less retirement savings. Our analysis did not test for and does 
not imply causality between retirement savings and tax rates. 

[34] We analyzed our results by earnings level using earnings 
quartiles. Households with adjusted gross incomes (adjusted for family 
size) of less than $18,960 were in the first earnings quartile, 
households with incomes of $18,961-33,519 were in the second earnings 
quartile, households with incomes of $33,520-56,559 were in the third 
earnings quartile, and households with incomes of $56,560-25,583,473 
were in the fourth earnings quartile. 

[35] Households without DC and DB plans or IRAs may have other assets, 
but there is no way to determine if these assets will be used for 
retirement or if they will even be available at retirement. 

[36] We used the SCF's definition of net worth-all of the household's 
assets, including houses and vehicles, minus its debts. 

[37] The refundable portion paid to a taxpayer increases federal 
outlays. 

[38] For the purposes of this report, the "refundable Saver's Credit" 
scenario includes the following modifications: making the credit 
refundable, having a 50 percent credit rate for all tax filers, and 
increasing AGI limits. See "Saver's Credit Modeling Scenarios" text 
box and appendix II for more information on the assumptions for this 
scenario. 

[39] Both our existing Saver's Credit and refundable Saver's Credit 
scenarios assume that all workers eligible for the credit take it. 
Thus, our scenarios show the maximum potential of the Saver's Credit 
for accumulating retirement income. We compared the effects of the 
Saver's Credit scenarios to a simulation where there is no Saver's 
Credit available. The results of the modeling scenarios are sensitive 
to our assumptions. We also modeled alternate scenarios using 
different assumptions for utilization, investment fund fees, and 
Social Security benefits. See appendix II for our projections under 
our alternate simulations as well as reproductions of the figures in 
this section using means instead of medians. In general, mean values 
for dollar changes were higher than median values. 

[40] Our simulation model only included this information for 
individuals born in 1995. Spouses may also have received the Saver's 
Credit. As a result, these estimates do not reflect the percentage of 
households that could receive the credit. 

[41] One analysis shows that a majority of taxpayers who qualify for 
the Saver's Credit based on their incomes would not be in a position 
to take advantage of the credit because they have no tax liability. 
According to this analysis, 69 million taxpayers qualified for the 
Saver's Credit based on their income in 2007, but about 45 million of 
these taxpayers had no federal tax liability. See Lisa Mensah, Raymond 
O'Mara III, Colby Farber, and Robert Weinberger, The Freedom Savings 
Credit: A Practical Step to Build Americans' Household Balance Sheets 
(Washington, D.C.: the Aspen Institute Initiative on Financial 
Security, February 2012). 

[42] We analyzed our results by earnings levels using earnings 
quartiles, which were calculated based on a measure of lifetime 
household earnings. Households with lifetime earnings of $929-34,377 
on an annual basis were in the first quartile, households with 
earnings of $34,379-60,682 were in the second quartile, households 
with earnings of $60,683-104,440 were in the third quartile, and 
households with earnings of $104,450-3,275,800 were in the fourth 
quartile. 

[43] For purposes of this report, we defined the household retirement 
annuity as including the annual retirement income received from all DC 
or DB plans. Social Security benefits are not included in this 
annuity. Dollar amounts from our simulations are adjusted for family 
size and are presented in 2013 dollars. 

[44] Under the existing Saver's Credit, households in the lowest 
earnings quartile receive the largest median dollar increase. Under 
the refundable Saver's Credit, households in the second-lowest 
earnings quartile receive the largest median dollar increase ($884). 

[45] We also modeled alternate scenarios using different assumptions 
for the aggregate participation rate, investment fund fees, and Social 
Security benefits. See appendix II for our projections under our 
alternate scenarios as well as reproductions of the figures in this 
section using means instead of medians. In general mean values for 
dollar changes were higher than median values. 

[46] We analyzed our results by earnings level using earnings 
quartiles, which were calculated based on a measure of lifetime 
household earnings. Households with lifetime earnings of $929-34,377 
on an annual basis were in the first quartile, households with 
earnings of $34,379-60,682 were in the second quartile, households 
with earnings of $60,683-104,440 were in the third quartile, and 
households with earnings of $104,450-3,275,800 were in the fourth 
quartile. 

[47] Dollar amounts from our simulations are adjusted for family size 
and are presented in 2013 dollars. 

[48] See text box above, "Automatic IRA and Saver's Credit Modeling 
Scenarios", for a description of this scenario and its assumptions. 

[49] As previously discussed, making the Saver's Credit refundable 
would result in a loss of revenue and additional outlays for the 
federal government. 

[50] See Jesse Bricker, et al. "Changes in U.S. Family Finances From 
2007 to 2010: Evidence from the Survey of Consumer Finances," Federal 
Reserve Bulletin, vol. 98, no. 2 (June 2012). 

[51] While these models use sample data, our report, like others using 
these models, does not address the issue of sampling errors. The 
results of the analysis reflect outcomes for individuals in the 
simulated populations and do not attempt to estimate outcomes for an 
actual population. 

[52] MINT is a detailed microsimulation model developed jointly by the 
Social Security Administration, the Brookings Institution, RAND, and 
the Urban Institute to project the distribution of income in 
retirement. MINT6 is the latest version and makes projections for the 
1926 to 1975 birth cohorts. 

[53] Bricker et. al., "Changes in U.S. Family finances from 2007 to 
2010." 

[54] AGI is defined as gross income minus adjustments to income. A 
marginal tax rate is the rate of tax paid on the next dollar of income 
that a taxpayer earns. 

[55] See Daniel Feenberg and Elisabeth Coutts, "An Introduction to the 
TAXSIM Model," Journal of Policy Analysis and Management, vol. 12, no. 
1 (1993), 189-194. 

[56] Peter A. Diamond, "What Stock Market Returns to Expect for the 
Future?" Social Security Bulletin, vol. 63, no. 2 (2000). 

[57] At the time of our analysis, the Social Security Administration's 
2013 Trustees Report had not been released. This report was released 
on May 31, 2013. As reported in 2012, the 2013 report found that 
revenues will not be adequate to pay full Social Security benefits for 
the next 75 years. Also, as in 2012, the asset reserves in the Social 
Security Trust Funds are projected to deplete in 2033. However, even 
though the depletion date for the Trust Funds has not changed, we are 
one year closer to asset reserve depletion. Our Social Security 
benchmarks would require minor adjustments to achieve solvency for the 
2013-2087 period. See The Board of Trustees, Federal Old-Age and 
Survivors Insurance and Federal Disability Insurance Trust Funds, The 
2013 Annual Report of the Board of Trustees of the Federal Old-Age and 
Survivors Insurance and Federal Disability Insurance Trust Funds 
(Washington, D.C.: May 31, 2013) and 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). 

[58] We analyzed H.R. 4049, The Automatic IRA Act of 2012 because it 
was the most recent legislative proposal at the time we conducted our 
analysis. On May 16, 2013, H.R. 2035, The Automatic IRA Act of 2013, 
was introduced before the House of Representatives. We determined that 
the differences between H.R. 2035 and H.R. 4049 are minimal and would 
not have a significant effect on our simulation results. For a small 
group of individuals in our simulations, i.e. individuals who 
participated in automatic IRAs in 2014, income in retirement would be 
slightly less because they would miss out on making contributions in 
2014. 

[59] While H.R. 4049 calls for depositing automatic IRA contributions 
into Roth IRAs as the default option, PENSIM models traditional IRAs. 

[60] We reviewed the following three studies with estimates of the 
initial utilization rate for the Saver's Credit. Lisa Southworth and 
John Gist, The Saver's Credit: What Does it Do for Saving? 
(Washington, D.C.: AARP Public Policy Institute, February 2008). The 
Aspen Institute Initiative on Financial Security, The Freedom Savings 
Credit: A Practical Step to Build Americans' Household Balance Sheets 
(2012); and Gary Koenig and Robert Harvey, "Utilization of the Saver's 
Credit: An Analysis of the First Year," National Tax Journal, vol. 
LVIII, no. 4 (2005). 

[61] William E. Nessmith, Stephen P. Utkus, and Jean A. Young, 
Measuring the Effectiveness of Automatic Enrollment (Vanguard Center 
for Retirement Research, December 2007). 

[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. 

[End of document]