Research Report. The Population of Workers Covered by the Auto IRA: Trends and Characteristics. AARP Public Policy Institute.

AARP Public Policy Institute C E L E B R A T I N G years The Population of Workers Covered by the Auto IRA: Trends and Characteristics Benjamin H. Harris 1 Ilana Fischer The Brookings Institution 1 Harris is on leave from the Brookings Institution while serving as a senior economist with the Council of Economic Advisers. He was a research economist with the Brookings Institution when this paper was written. Research Report

The Population of Workers Covered by the Auto IRA: Trends and Characteristics Benjamin H. Harris Ilana Fischer The Brookings Institution AARP s Public Policy Institute informs and stimulates public debate on the issues we face as we age. Through research, analysis and dialogue with the nation s leading experts, PPI promotes development of sound, creative policies to address our common need for economic security, health care, and quality of life. The views expressed herein are for information, debate, and discussion, and do not necessarily represent official policies of AARP. #2012-03 February 2012 2012, AARP Reprinting with permission only AARP Public Policy Institute 601 E Street, NW, Washington, DC 20049 http://www.aarp.org/ppi

Acknowledgment The authors thank David John, Surachai Khitatrakun, and Gary Koenig for helpful comments. iv

Table of Contents Acknowledgment...iv Abstract...vii Executive Summary...1 Results...1 Introduction...3 Previous Research...4 Methodology and Data...5 Results...7 Conclusion...14 List of Tables Summary Table. Workers Eligible for Automatic Enrollment under Various Restrictions... 2 Table 1. Table 2. Workers Eligible for Automatic Enrollment Under Various Restrictions 2007 Survey of Consumer Finances... 8 Workers Eligible for Automatic Enrollment Under Various Restrictions 2009 Current Population Survey (2008 data)... 9 Table 3. Distribution of Workers Eligible for Auto IRA... 11 Table 4. Workers Eligible for Automatic Enrollment under Various Restrictions 1998 Current Population Survey (1997 data)... 12 v

AbstrAct This study investigates the claim that the Auto IRA proposal can potentially boost the retirement adequacy of tens of millions of workers. We estimate that between 24 million and 43 million workers approximately one-quarter of the workforce would be eligible for automatic enrollment in the proposals under consideration in Congress. Moreover, we find that the eligible population is heavily skewed toward workers with low and moderate wage levels and that exempting very low-wage workers from the automatic enrollment mandate would have only a modest impact on the number of eligible workers. Lastly, we find that the proportion of Auto IRA eligible workers is increasing over time owing to the shift away from employer-sponsored retirement plans. vii

Executive Summary Automatic enrollment in individual retirement accounts (IRAs) is an innovative proposal designed to increase retirement saving among millions of Americans. The plan which we refer to as the Auto IRA aims to improve retirement security for workers without employer-provided retirement plans. Under the Auto IRA, eligible workers are automatically enrolled in a private retirement account unless they explicitly elect not to participate ( opt out ). Using employer payroll systems, contributions are automatically withdrawn from workers paychecks as a set percentage of pay, deposited in worker-owned retirement accounts, and invested in low-cost diversified investment vehicles (such as life-cycle funds). Workers have ultimate control over their participation in Auto IRAs and can opt out of any default option, such as changing the amount contributed or the default investment. Since the concept was introduced in 2006, the Auto IRA has gained substantial momentum among policy makers and sparked the introduction of two bills in Congress. The goal of this report is to estimate the population of workers who would be automatically enrolled in an IRA after accounting for the bills various restrictions and exemptions. Results We estimate the population of workers eligible for automatic enrollment given the restrictions in Congress s Auto IRA proposals; this forms our baseline scenario. We also estimate the population of workers subject to the Auto IRA under a scenario where very low-wage workers are exempt from the program, and a scenario where half of small businesses, otherwise exempt because of their size, voluntarily enroll employees. Our estimates of the population of workers eligible for automatic enrollment are presented in the summary table below. Under the baseline scenario, we estimate that between 24.4 million (19.7 percent) and 35.8 million (26.6 percent) workers would be eligible for the Auto IRA. Most of these workers are employed full time. Between 19.3 million and 27.2 million full-time workers and between 5.0 million and 8.5 million part-time workers would be eligible. The population of workers subject to automatic enrollment would shrink if Congress elected to exclude very low-wage workers from the mandate. Under a scenario where workers with low annual wages less than $5,000 were excluded from the Auto IRA program, the number of workers subject to automatic enrollment would drop modestly to between 23.6 million (19.1 percent) and 32.7 million (24.3 percent). The baseline scenario might underestimate the number of eligible workers. The baseline estimates assume that firms will enroll workers only if required to do so, excluding the possibility of voluntary enrollment. We estimate the number of workers subject to the Auto IRA under the condition that half of workers at small firms are automatically enrolled voluntarily by their employer. Under this scenario, the number of workers subject to automatic enrollment increases to between 29.3 million (23.7 percent) and 43.1 million (32.0 percent). Given the experience with automatic enrollment in 401(k)s, the assumption that small firms would elect to voluntarily enroll workers in an IRA is plausible. 1

Summary Table Workers Eligible for Automatic Enrollment under Various Restrictions (1) Baseline Eligible Workers Number Worker Characteristic of Workers 2007 Survey of Consumer Finances Percentage of Total Eligible Full-Time Workers 19,358,685 18.0% Eligible Part-Time Workers 5,049,241 31.5% (2) Eligible Workers with Wage Restriction Total Eligible Workers 24,407,926 19.7% Eligible Full-Time Workers 19,214,842 17.8% Eligible Part-Time Workers 4,363,853 27.2% (3) Eligible Workers with Voluntary Small Business Participation Total Eligible Workers 23,578,695 19.1% Eligible Full-Time Workers 23,076,540 21.4% Eligible Part-Time Workers 6,272,438 39.1% (1) Baseline Eligible Workers Total Eligible Workers 29,348,978 23.7% 2009 Current Population Survey (2008 data) Eligible Full-Time Workers 27,239,255 24.6% Eligible Part-Time Workers 8,517,329 35.6% (2) Eligible Workers with Wage Restriction Total Eligible Workers 35,756,584 26.6% Eligible Full-Time Workers 26,281,578 23.8% Eligible Part-Time Workers 6,387,598 26.7% (3) Eligible Workers with Voluntary Small Business Participation Total Eligible Workers 32,669,176 24.3% Eligible Full-Time Workers 32,534,775 29.4% Eligible Part-Time Workers 10,542,792 44.1% Total Eligible Workers 43,077,567 32.0% We further estimate the distribution of eligible workers by wage level and calculate the changes in the proportion of eligible workers over the past decade. We find that automatic enrollment is decidedly progressive in terms of eligible workers, in large part due to the correlation between lower wages and lack of employer-sponsored retirement plans. Survey of Consumer Finance data show that more than 80 percent of the population eligible for automatic enrollment earn less than $50,000 in wages, and more than one-third earn less than $20,000 in wages. Fewer than 5 percent of eligible workers have wages in excess of $100,000. Current Population Survey data show a nearly identical trend. We also find that the proportion of Auto IRA-eligible workers is increasing over time owing to the shift away from employer-sponsored retirement plans. 2

Introduction Automatic enrollment in individual retirement accounts (IRAs) is an innovative proposal designed to increase retirement saving among millions of Americans. The plan which we refer to as the Auto IRA aims to improve retirement security for workers without company retirement plans. Under the Auto IRA, eligible workers are automatically enrolled in a private retirement account unless they explicitly elect not to participate ( opt out ). Using employer payroll systems, contributions are automatically withdrawn from workers paychecks as a set percentage of pay, deposited in workerowned retirement accounts, and invested in low-cost diversified investment vehicles (such as life-cycle funds). Workers have ultimate control over their participation in Auto IRAs and can opt out of any default option, such as changing the amount contributed or the default investment. Since the concept was introduced in 2006, the Auto IRA has gained substantial momentum among policy makers. Automatic enrollment in IRAs was included in the Obama administration s fiscal year 2012 budget proposal; 2 the President s Economic Recovery Advisory Board 3 endorsed automatic enrollment in IRAs in its recent report on tax policy. In addition, two Auto IRA bills were recently introduced in Congress, one in each chamber, (S. 3760 4 and H.R. 6099) aimed at automatically enrolling a substantial proportion of private sector employees in retirement accounts. Slightly differing legislation was introduced in earlier sessions of Congress. The Senate and House bills are similar but not identical. Both bills would require that firms meeting certain criteria automatically enroll their employees in IRAs; the bills would provide a modest tax credit for the administrative costs of automatic enrollment while using tax enforcement mechanisms to ensure firms compliance. The bills differ in other respects, but differ only slightly in their restrictions on workers who are exempt from automatic enrollment. Under both bills, several classes of workers would be exempt from Auto IRAs. Public sector workers, workers with access to employer-sponsored pensions, employees of church and religious organizations, workers at firms with fewer than 10 employees, workers with fewer than three months tenure, workers employed by firms established for fewer than two years, 5 and workers younger than 18 would all be exempt from the automatic enrollment mandate. 2 U.S. Treasury, General Explanation of the Administration s Fiscal Year 2012 Revenue Proposals (2011 Green Book) (Washington, DC: Government Printing Office, 2010). http://www.treasury.gov/resourcecenter/tax-policy/documents/final%20greenbook%20feb%202012.pdf. 3 President s Economic Recovery Advisory Board, The Report on Tax Reform Options: Simplification, Compliance, and Corporate Taxation (2010), http://www.whitehouse.gov/sites/default/files/microsites/ PERAB_Tax_Reform_Report.pdf. 4 On September 14, 2011, Senator Jeff Bingaman reintroduced the Auto IRA bill as S. 1557 the Automatic IRA Act of 2011. 5 Both bills exclude new businesses. New businesses are defined as those not in existence at all times during the calendar year and the preceding calendar year. Thus, it is possible for a firm established for nearly two years to be exempt from the Auto IRA. 3

The goal of this paper is to estimate the population of workers who would be automatically enrolled in an IRA after accounting for the bills various restrictions and exemptions. We further estimate the distribution of eligible workers by wage level and calculate the changes in the proportion of workers who would have been eligible over the past decade. We find that between 24 million and 43 million workers approximately one-quarter of the workforce would be eligible for automatic enrollment under the proposals recently introduced in Congress. Moreover, we find that the eligible population is heavily skewed toward workers with low and moderate wage levels, and that the proportion of Auto IRA eligible workers is increasing over time owing to the shift away from employer-sponsored retirement plans. We present a review of the existing literature on automatic enrollment in the next section, followed by a description of our methodology. We then present our findings and conclude with a brief summary and discussion. Previous Research There is a growing literature on the effects of automatic enrollment broadly defined, but little is known about the potential effects of automatic enrollment in IRAs. Perhaps the most well-known study of the effects of automatic enrollment was conducted by Madrian and Shea, 6 who studied the effects of automatic enrollment in the 401(k) plan of a large U.S. corporation. Madrian and Shea found that automatic enrollment substantially boosted employee participation in the company s retirement saving plan. Under automatic enrollment, the average participation rate in the company s 401(k) jumped from 37.4 percent to 85.9 percent. 7 Madrian and Shea also found that employees who were automatically enrolled in a retirement plan tended to maintain the default contribution rate and investment allocation. Other studies report similarly high rates of participation under automatic enrollment. 8 In an important study, Beshears et al. 9 examined the impact of the employer match on opt-out rates under automatic enrollment. They found that the existence of an employer match creates only a modest reduction 5 to 11 percentage points in the proportion of employees who opt out. Such a finding is relevant to automatic enrollment in IRAs, since IRAs usually offer saving incentives similar to an 6 Brigitte C. Madrian and Dennis F. Shea, The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior, The Quarterly Journal of Economics 116, no. 4 (2001): 1149 87. 7 Madrian and Shea control for demographic and economic factors, including gender, race, age, and compensation, and find the regression-adjusted effect of automatic enrollment to be even larger. After controlling for these factors, the difference is 50.4 percentage points, compared to the raw difference of 48.5 percentage points noted above. 8 John Beshears, James J. Choi, David Laibson, and Brigitte C. Madrian, The Importance of Default Options for Retirement Savings Outcomes: Evidence from the United States, in Lessons from Pension Reform in the Americas, ed. Stephen J. Kay and Tapen Sinha (New York: Oxford University Press, 2008); James J. Choi, David Laibson, Brigitte C. Madrian, and Andrew Metrick, For Better or For Worse: Default Effects and 401(k) Savings Behavior, in Perspectives in the Economics of Aging, ed. David Wise (Chicago: University of Chicago Press, 2004). 9 John Beshears, James J. Choi, David Laibson, and Brigitte C. Madrian, The Impact of Employer Matching on Savings Plan Participation under Automatic Enrollment, in Research Findings in the Economics of Aging, ed. David A. Wise (Chicago: University of Chicago Press, 2009). 4

unmatched contribution to a 401(k). Soto and Butrica 10 found a correlation between automatic enrollment and lower employer match rates in retirement plans, but were unable to provide causal evidence that automatic enrollment led to lower match rates. Using a different methodology and alternative sample, VanDerhei 11 found the opposite effect. Geissler and Harris 12 estimated the revenue and distributional effects of automatic enrollment in 401(k)s, finding revenue costs to be modest and the reform to be generally progressive. While the growing literature on automatic enrollment in 401(k)s can inform policy makers about the Auto IRA, the literature s lessons may not translate perfectly owing to differences in populations eligible for the respective programs. Workers who are eligible for employer-sponsored retirement plans tend to be older, better educated, and more highly compensated than workers eligible for the Auto IRA. 13 Since the Auto IRA has yet to be implemented, the few studies that focus on the Auto IRA are speculative. Iwry and John 14 discussed the merits of Automatic IRAs, while also describing some of the practical concerns with the proposal. Harris and Johnson 15 found that Auto IRA benefits would be primarily spread across the middle three quintiles of the income distribution, with small average changes in benefits for those at either end of the income distribution. The study also found that if automatic enrollment is coupled with an expansion of the Saver s Credit, benefits are more concentrated toward lower-income taxpayers, with the largest gains in after-tax income accrued by those in the second and third income quintiles. Methodology and Data To measure the potential universe of employees who might be affected by the Auto IRA, we begin with two related bills introduced in the U.S. Congress. In the Senate, Senator Jeff Bingaman (D-NM) introduced the Automatic IRA Act of 2010 (S. 3760), which would require most private firms without employer-sponsored retirement plans to automatically enroll employees in a Roth IRA, to automatically direct a preset proportion 10 Mauricio Soto and Barbara A. Butrica, Will Automatic Enrollment Reduce Employer Contributions to 401(k) Plans? Urban Institute Retirement Policy Program Discussion Paper 09-04 (Washington, DC: The Urban Institute, 2009). 11 Jack VanDerhei, The Impact of Automatic Enrollment in 401(k) Plans on Future Retirement Accumulations: A Simulation Study Based on Plan Design Modifications of Large Plan Sponsors, Employee Benefit Research Institute Issue Brief 341 (Washington, DC: Employee Benefit Research Institute, April 2010). 12 Christopher Geissler and Benjamin H. Harris, The Automatic 401(K): Revenue and Distributional Estimates, in Automatic: Changing the Way America Saves, ed. William G. Gale, J. Mark Iwry, David John, and Lina Walker (Washington, DC: Brookings Institution Press, 2009). 13 Investment Company Institute, Who Gets Retirement Plans and Why, Research Perspective 14, no. 2 (2008). 14 J. Mark Iwry and David C. John, Pursuing Universal Retirement Security Through Automatic IRAs, Retirement Security Project Discussion Paper 2009-3 (Washington, DC: Brookings Institution, 2009). 15 Benjamin H. Harris and Rachel M. Johnson, Automatic Enrollment in IRAs: Costs and Benefits, Tax Notes 124, no. 9 (2009): 903 14. 5

of the employees wages to the IRA on behalf of the employees, and to automatically invest these contributions in a designated low-cost investment. The Senate version of the legislation does not require all workers to be automatically enrolled in an IRA. Firms that offer retirement plans are generally exempt from the requirement, but firms that exempt particular divisions from the retirement plan are required to automatically enroll ineligible workers in an IRA. Other classes of exempt workers include those who have not satisfied the minimum age and job tenure requirements, government workers and church employees, employees with fewer than three months of service at a company, employees of newly established firms, and workers under 18 years of age. In the first year of enactment, firms with fewer than 100 employees are exempt from the requirement; this exemption is gradually phased down over four years to firms with fewer than 10 employees. For these purposes, an employee is counted as a worker earning more than $5,000 in annual wages. A nearly identical companion bill with the same title (H.R. 6099) was introduced in the House by Rep. Richard Neal (D-MA). With respect to automatic enrollment, the primary difference between the bills is the threshold for the size of company subject to the provisions. While the Senate bill gradually reduces the minimum size of the firm subject to the requirement from 100 to 10, the House bill immediately requires firms with 10 or more employees to enroll workers in an Auto IRA. 16 We use data from two large-scale government surveys to estimate the number of workers who would be eligible for the Auto IRA under the congressional proposals. One of the surveys is the 2007 Survey of Consumer Finances (SCF). The SCF is a triennial cross-sectional survey conducted by the Federal Reserve Board. The SCF contains a wealth of data on household finances, as well as substantial data on demographics, employment, and participation in employer-sponsored retirement accounts. The 2007 SCF contains 4,422 family-level observations. The other data source is the Current Population Survey (CPS). The CPS is a monthly survey of labor force participation conducted by the Bureau of Labor Statistics within the U.S. Census Bureau; it is a primary data source for U.S. labor market data. The CPS also contains an Annual Social and Economic Supplement administered in March. The March Supplement collects additional data on income, demographics, and labor market activity not captured in the monthly survey. In addition, the CPS administers a Displaced Worker, Employee Tenure, and Occupational Mobility Supplement in January. We primarily use labor force, demographic, and pension data from the 1998 and 2009 March Supplements in our estimates, but also use the job tenure data in the January Supplement to impute data on job tenure. Since the March CPS Supplements ask respondents about their behavior in the prior year, our data correspond to 1997 and 2008, respectively. Neither data source can perfectly estimate the population of workers who would be subject to automatic enrollment under the congressional bills. From the SCF, we can identify workers who are employed in the private sector, work at firms with 10 or more employees, and do not have a pension plan and are not eligible to receive one; these 16 For the purposes of measuring the proportion of workers exempt from automatic enrollment, we ignore the phase-in provision in the Senate bill. 6

workers are subject to the Auto IRA. We exclude workers who are currently ineligible to receive a pension but expect to be eligible in the future. 17 The SCF allows the identification of employees with less than one year of job tenure. To approximate private sector workers employed fewer than three months, we impute short-term job tenure based on data from the Displaced Worker/Job Tenure Supplement to the CPS. Among private sector workers employed less than a year, 27.6 percent were employed fewer than three months in 2008; a similar proportion was employed fewer than three months in 1998. Based on these estimates, we randomly assign short-term job tenure among private sector employees reporting job tenure of less than one year. 18 Due to survey design, our CPS estimates are slightly less precise. We use the same approach applied to SCF data, estimating the number of private sector workers, workers without a pension who are ineligible to receive one, and employees of firms with more than 10 workers. Unlike the SCF, the CPS does not report employees who expect to eventually become eligible for a pension. We also cannot determine job tenure using CPS data and consequently do not exempt workers under this criterion. 19 We make two adjustments to the baseline calculations. First, we calculate the number of workers who would be covered under the automatic IRA if the proposal exempts workers with very low wages to increase administrative ease. For purposes of this calculation, we define very low-wage workers as those with earnings less than $5,000 annually. While neither congressional bill includes this exemption, it might be considered in response to criticism that the administrative costs associated with the Auto IRA exceed the benefits for very low-wage workers. Second, we estimate the number of workers eligible for the Auto IRA if 50 percent of small business employees are automatically enrolled despite their exemption. To achieve this estimate, we randomly assign participation to half of the small business employees who would otherwise be exempt. Results The population of workers eligible for automatic enrollment are presented in table 1 (SCF) and table 2 (CPS). Under the baseline scenario, we estimate that between 24.4 million (19.7 percent) and 35.8 million (26.6 percent) workers would be eligible for the Auto IRA. Most of these workers are employed full time. Between 19.3 million and 27.2 million full-time workers and between 5.0 million and 8.5 million part-time workers would be eligible. 17 These workers most likely do not meet the minimum age and service requirements under the employer s plan guidelines. 18 Firms without employer-sponsored retirement plans are more likely to have short-term workers. A survey of small employers found that in 2000, 34 percent of employees in small businesses without retirement plans had fewer than three years job tenure, compared to just 13 percent among small firms with retirement plans. Dallas L. Salisbury, Hearing on H.R. 10, The Comprehensive Retirement Security & Pension Reform (2001), http://www.ebri.org/publications/testimony/index.cfm?fa=t127. 19 Both surveys also allow us to identify self-employed workers. While the Auto IRA proposals do not exempt self-employed workers, it seems unlikely that the financial inertia that makes automatic enrollment effective would apply to self-employed workers; establishing an IRA would likely require more administrative effort than would opting out. Consequently, we assume that self-employed workers will not enroll in retirement plans under the Auto IRA. 7

Table 1 Workers Eligible for Automatic Enrollment Under Various Restrictions 2007 Survey of Consumer Finances (1) Baseline Eligible Workers Worker Characteristic Number of Workers Percent of Total Eligible Full-Time Workers 19,358,685 18.0% Eligible Part-Time Workers 5,049,241 31.5% (2) Eligible Workers with Wage Restriction Total Eligible Workers 24,407,926 19.7% Eligible Full-Time Workers 19,214,842 17.8% Eligible Part-Time Workers 4,363,853 27.2% (3) Eligible Workers with Voluntary Small Business Participation Total Eligible Workers 23,578,695 19.1% Eligible Full-Time Workers 23,076,540 21.4% Eligible Part-Time Workers 6,272,438 39.1% Incremental Calculation of Total Baseline Eligible Workers Total Eligible Workers 29,348,978 23.7% Full-Time Workers 107,692,958 87.0% Part-Time Workers 16,052,254 13.0% Total Workers 123,745,212 100.0% Public-Sector Workers 7,797,111 6.3% Self-Employed Private Sector Workers 17,031,665 13.8% Private-Sector Workers Employed by a Firm 98,916,435 79.9% Workers with a Pension 56,295,802 45.5% Eligible Workers without a Pension 35,807,512 28.9% Workers without Pension Eligibility 14,610,231 11.8% Self-Employed Private Sector Workers 17,031,665 13.8% Subtotal: Private Sector Workers (non self-employed) without Pension Eligibility 35,271,349 28.5% Workers at a Firm with 10 or More Employees 97,880,206 79.1% Workers at a Firm with Less than 10 Employees 25,865,006 20.9% Subtotal: Private Sector Workers (non self-employed) without Pension Eligibility at Firms with 10 or More Employees 25,124,610 20.3% Employees with Three Months or Less of Job Tenure 2,813,706 2.3% Employees with Greater than Three Months of Job Tenure 120,931,506 97.7% Employees with Less than $5,000 in Wages 11,804,017 9.5% Employees with At Least $5,000 in Wages 111,941,196 90.5% Total Eligible Workers 24,407,926 19.7% 8

Table 2 Workers Eligible for Automatic Enrollment Under Various Restrictions 2009 Current Population Survey (2008 data) (1) Baseline Eligible Workers Worker Characteristic Number of Workers Percent of Total Eligible Full-Time Workers 27,239,255 24.6% Eligible Part-Time Workers 8,517,329 35.6% (2) Eligible Workers with Wage Restriction Total Eligible Workers 35,756,584 26.6% Eligible Full-Time Workers 26,281,578 23.8% Eligible Part-Time Workers 6,387,598 26.7% (3) Eligible Workers with Voluntary Small Business Participation Total Eligible Workers 32,669,176 24.3% Eligible Full-Time Workers 32,534,775 29.4% Eligible Part-Time Workers 10,542,792 44.1% Incremental Calculation of Total Baseline Eligible Workers Total Eligible Workers 43,077,567 32.0% Full-Time Workers 110,606,954 82.2% Part-Time Workers 23,918,236 17.8% Total Workers 134,525,190 100.0% Public-Sector Workers 20,648,844 15.3% Self-Employed Private Sector Workers 9,060,147 6.7% Private-Sector Workers Employed by a Firm 104,816,199 77.9% Workers with a Pension 59,598,722 44.3% Workers without a Pension, Employer Offers Pension 12,756,854 9.5% Workers without a Pension, No Pension Offered 62,169,614 46.2% Subtotal: Private Sector Workers (non self-employed) without Pension Eligibility 61,386,576 45.6% Workers at a Firm with 10 or More Employees 107,294,274 79.8% Workers at a Firm with Less than 10 Employees 27,230,916 20.2% Subtotal: Private Sector Workers (non self-employed) without Pension Eligibility at Firms with 10 or More Employees 46,039,825 34.2% Employees with Less than $5,000 in Wages 14,981,704 11.1% Employees with At Least $5,000 in Wages 119,543,486 88.9% Total Eligible Workers 35,756,584 26.6% 9

The population of workers subject to automatic enrollment would shrink if Congress elected to exclude very low-wage workers from the mandate. Under a scenario where workers with annual wages less than $5,000 were excluded from the Auto IRA program, the number of workers subject to automatic enrollment would drop modestly to between 23.6 million (19.1 percent) and 32.7 million (24.3 percent). In table 1 and table 2, this scenario is referred to as Eligible Workers with Wage Restriction. The baseline scenario might underestimate the number of eligible workers. The baseline estimates assume that firms will enroll workers only if required to do so, excluding the possibility of voluntary enrollment. The growing trend of automatic enrollment in 401(k)s 20 has shown that many employers will automatically enroll workers in retirement plans even though not required to do; such a trend would likely be stronger for the Auto IRA, which will generally cost employers less than 401(k)s. We estimate the number of workers subject to the Auto IRA under the condition that half of workers at small firms are automatically enrolled voluntarily by their employer. Under this scenario, the number of workers subject to automatic enrollment increases to between 29.3 million (23.7 percent) and 43.1 million (32.0 percent). In table 1 and table 2, this scenario is referred to as Eligible Workers with Voluntary Small Business Participation. Tables 1 and 2 also show the effects of individual restrictions on the potential pool of Auto IRA eligible workers. The number of U.S. workers in our data amounts to between 123.7 million and 134.5 million; the higher estimate corresponds to the CPS data. Of these, between 35.2 million (28.5 percent) and 61.4 million (45.6 percent) are private sector workers employed by a firm (i.e., not self-employed) who are ineligible to receive retirement benefits through their employer. The small business exemption reduces this pool of eligible workers to between 25.1 million (20.3 percent) and 46.0 million (34.2 percent). Automatic enrollment is decidedly progressive in terms of eligible workers, in large part owing to the correlation between lower wages and lack of an employer-sponsored retirement plan. SCF data show that more than 80 percent of the population eligible for automatic enrollment earn less than $50,000 in wages, and more than one-third earn less than $20,000 (table 3). Fewer than 5 percent of eligible workers have wages in excess of $100,000. CPS data show a nearly identical trend. The population of workers eligible for automatic enrollment is gradually increasing over time (table 4). Using the CPS, we find that in 1997 29.1 million workers (23.4 percent) were eligible under the baseline scenario, 24.8 million (19.9 percent) were eligible under a wage restriction, and 35.3 million (28.3 percent) were eligible assuming some small businesses voluntarily enroll workers in IRAs. 20 The percentage of firms with 401(k)s offering automatic enrollment has steadily risen from 4.2 percent in 1999 to 23.6 percent in 2006. Large firms are much more likely than smaller firms to offer automatic enrollment. In 2006, 41.3 percent of plans offered by large firms (more than 5,000 employees) included automatic enrollment, compared to just 6.8 percent of those offered by small firms (49 or fewer employees); however, the proportion of small firms offering automatic enrollment is growing over time. Profit Sharing/401(k) Council of America, Annual Survey of Profit Sharing and 401(k) Plans (Chicago: Profit Sharing/401(k) Council of America, 1999); Profit Sharing/401(k) Council of America, Annual Survey of Profit Sharing and 401(k) Plans (Chicago: Profit Sharing/401(k) Council of America, 2006). 10

Table 3 Distribution of Workers Eligible for Auto IRA Baseline Eligible Workers Millions of Eligible Workers Percent of Eligible Workers Eligible Workers with Wage Restriction Eligible Workers with Voluntary Small Business Participation SCF Baseline Eligible Workers Eligible Workers with Wage Restriction Eligible Workers with Voluntary Small Business Participation Millions of Eligible Workers Percent of Eligible Workers Wages Less than $20,000 9,141,101 8,311,870 11,355,338 37.5% 35.3% 38.7% $20,001 - $50,000 10,833,941 10,833,941 13,140,882 44.4% 45.9% 44.8% $50,001-$100,000 3,479,545 3,479,545 3,890,735 14.3% 14.8% 13.3% $100,000+ 953,338 953,338 962,023 3.9% 4.0% 3.3% Total 24,407,925 23,578,694 29,348,978 100.0% 100.0% 100.0% CPS Millions of Eligible Workers Percent of Eligible Workers Wages Less than $20,000 14,037,195 10,949,787 16,985,909 39.3% 33.5% 39.4% $20,001 - $50,000 14,692,374 14,692,374 17,682,743 41.1% 45.0% 41.0% $50,001-$100,000 5,379,996 5,379,996 6,351,188 15.0% 16.5% 14.7% $100,000+ 1,647,019 1,647,019 2,057,727 4.6% 5.0% 4.8% Total 35,756,584 32,669,176 43,077,567 100.0% 100.0% 100.0% 11

Table 4 Workers Eligible for Automatic Enrollment under Various Restrictions 1998 Current Population Survey (1997 data) Baseline Eligible Workers Worker Characteristic Number of Workers Percent of Total Change from 1997 to 2008 Number of Workers Percent of Total Percentage Point Change Eligible Full-Time Workers 22,938,469 21.9% 4,300,786 2.7% Eligible Part-Time Workers 6,197,871 31.2% 2,319,458 4.4% Eligible Workers with Wage Restriction Total Eligible Workers 29,136,340 23.4% 6,620,244 3.2% Eligible Full-Time Workers 21,139,931 20.2% 5,141,647 3.6% Eligible Part-Time Workers 3,640,410 18.3% 2,747,187 8.4% Eligible Workers with Voluntary Small Business Participation Total Eligible Workers 24,780,341 19.9% 7,888,835 4.4% Eligible Full-Time Workers 27,605,000 26.3% 4,929,774 3.1% Eligible Part-Time Workers 7,727,887 38.9% 2,814,905 5.1% Incremental Calculation of Total Baseline Eligible Workers Total Eligible Workers 35,332,887 28.3% 7,744,679 3.7% Full-Time Workers 104,831,496 84.1% 5,775,458-1.9% Part-Time Workers 19,843,991 15.9% 4,074,245 1.9% Total Workers 124,675,487 100.0% 9,849,703 0.0% Public-Sector Workers 17,975,247 14.4% 2,673,597 0.9% Self-Employed Private Sector Workers 9,248,507 7.4% -188,361-0.7% Private-Sector Workers Employed by a Firm 97,451,733 78.2% 7,364,467-0.2% Workers with a Pension 57,345,226 46.0% 2,253,497-1.7% Workers without a Pension, Employer Offers Pension 14,651,885 11.8% -1,895,030-2.3% Workers without a Pension, No Pension Offered 52,678,377 42.3% 9,491,237 4.0% Subtotal: Private Sector Workers (non self-employed) without Pension Eligibility 54,555,574 43.8% 6,831,002 1.9% Workers at a Firm with 10 or More Employees 99,868,626 80.1% 7,425,648-0.3% Workers at a Firm with Less than 10 Employees 24,806,860 19.9% 2,424,055 0.3% Subtotal: Private Sector Workers (non self-employed) without Pension Eligibility at Firms with 10 or More Employees 40,956,783 32.9% 5,083,043 1.4% Employees with Less than $5,000 in Wages 18,558,483 14.9% -3,576,779-3.7% Employees with At Least $5,000 in Wages 106,117,004 85.1% 13,426,482 3.7% Total Eligible Workers 29,136,340 23.4% 6,620,244 3.2% 12

Comparing CPS data between 1997 and 2008, the number of eligible workers increases by 6.6 million (or from 23.4 to 26.6 percent of total workers a 3.2 percentage point increase) under the baseline scenario, 7.9 million (or from 19.9 to 24.3 percent of total workers a 4.4 percentage point increase) under the wage restricted scenario, and 7.7 million (or from 28.3 to 32 percent of total workers a 3.7 percentage point increase) under the voluntary small business participation scenario. The proportion of part- and full-time workers increased in each scenario. The increase in the number of workers eligible for the Automatic IRA is driven by both a growing labor force and the growing proportion of workers without access to employer-sponsored pension coverage. The size of the workforce increased by 9.8 million between 1997 and 2008; the increase boosted the number of workers eligible for the Auto IRA. The shift away from employersponsored retirement plans also increased the proportion of workers eligible for the Auto IRA. Between 1997 and 2008, the percentage of workers ineligible for an employer retirement plan rose from 42.3 percent to 46.2 percent, an increase of 4.0 percentage points (table 4). The difference in estimates between the CPS and SCF is substantial. We attribute these differences to three primary factors. First, we attribute differences in estimated number of eligible workers (as opposed to proportion of eligible workers) to differences in total population of U.S. workers indicated in the two data sets; the SCF estimates U.S. workers at 123.7 million, compared to 134.5 million in the CPS. Second, as discussed above, the SCF offers more extensive data on worker characteristics relevant to this study, which allows for a more comprehensive measurement of eligible workers. Third, the survey questionnaires differ in their wording among comparable questions, and can thus produce different estimates of a given population. For example, the CPS asks respondents if they have worked at all in the previous year, while the SCF asks respondents whether they are working at the present time. Differences in questionnaires can lead to moderate differences in estimates of the same population. 21 One noteworthy comparison is that the number of Auto IRA eligible workers is substantially larger than the potential pool of Auto 401(k) eligible workers. Using the 2007 SCF, we estimate that approximately 10.6 million workers (8.5 percent of workers) would be eligible for automatic enrollment in 401(k)s if it were universally applied to all existing 401(k) plans, 22 less than half the number eligible for the Auto IRA. While opt-out rates would likely be higher for the Auto IRA than for automatic enrollment in 401(k)s, the significantly higher proportion of Auto IRA eligible workers illustrates the potential for the program to affect saving behavior for a large number of workers. Moreover, a combined policy of automatic enrollment in IRAs and 401(k)s has the potential to affect nearly one-third of American workers. We conclude this section with two important points. First, we emphasize that our estimates calculate the number of American workers who will be eligible for the Auto IRA, not those who will actually enroll in an IRA after being given the option to opt 21 For example, Sanzenbacher shows that pension coverage estimates vary across major surveys. Geoffrey Sanzenbacher, Estimating Pension Coverage Using Different Data Sets, Center for Retirement Research Issue Brief, No. 51 (Chestnut Hill, MA: Center for Retirement Research, 2006). 22 This estimate is not shown in the tables. The calculation is based on the number of employees eligible for, but not enrolled in, an employer-sponsored plan. Several of the Auto 401(k) proposals in Congress would exempt small businesses and particular classes of employees from the automatic 401(k) mandate; these restrictions would reduce the number and proportion of eligible workers. 13

out. Second, we note that the data are based on snapshots of employee characteristics at a given point in time. Employee turnover, which may be higher for the population of workers most likely to be subject to the Auto IRA, will increase the number of workers eligible for the program over the course of a year. Data limitations preclude us from estimating the impact of employee turnover on Auto IRA eligibility. Conclusion Auto IRA proposals continue to gain momentum in Congress. The bills sponsors and other supporters of automatic enrollment claim the Auto IRA will put millions of workers on the path toward better saving and will particularly help low- and moderate-income workers. The limited number of studies focusing on the Auto IRA s potential has made it difficult to either support or reject these claims. This study reinforces the claim that the Auto IRA can potentially boost the retirement adequacy of tens of millions of workers. We estimate that between 24 million and 43 million workers approximately one-quarter of the workforce would be eligible for automatic enrollment in the proposals recently introduced in Congress. Moreover, we find that the eligible population is heavily skewed toward workers with low and moderate wage levels, and that exempting very low-wage workers from the automatic enrollment mandate would have only a modest impact on the number of eligible workers, while potentially reducing administrative burdens. Lastly, we find that the proportion of eligible workers is increasing over time and that this trend is due, in large part, to the shift away from employer-sponsored retirement plans. These results reflect some uncertainty. Due to data limitations, it is difficult to determine precisely which workers would be exempt from automatic enrollment on the basis of job tenure, although the effect of this requirement on firms behavior is unclear. Firms might, for example, opt to automatically enroll workers on or near their hire date, rather than wait three months. In addition, small firms that are exempt from the automatic enrollment mandate may opt to participate anyway once the infrastructure is established; this scenario would be consistent with the recent trend of large corporations adopting automatic enrollment voluntarily. Data limitations also make it impossible to identify workers at new firms; these firms would also be exempt from automatic enrollment. Like the exemption for small firms, it is unclear whether this exemption would be binding. Despite these limitations, the estimates presented here indicate that the Auto IRA can have a widespread impact on worker participation in retirement saving accounts, especially for workers with low or moderate wage levels. The economic effects of automatic enrollment remain somewhat of a mystery: More research is needed to determine the effects of automatic enrollment generally, and the Auto IRA in particular, on net saving, labor supply, and retirement adequacy. 14