GAO Says Congress Should Amend ERISA to Update 401(k) Plan Rules

GAO's nongeneralizable survey of 80 401(k) plans ranging in size from fewer than 100 participants to more than 5,000 and its review of industry data found that many plans have policies that affect workers' ability to (1) save in plans (eligibility policies), (2) receive employer contributions, and (3) keep those employer contributions if they leave their job (vesting policies). Thirty-three of 80 plans surveyed had policies that did not allow workers younger than age 21 to participate in the plan. In addition, 19 plans required participants to be employed on the last day of the year to receive any employer contribution for that year. Fifty-seven plans had vesting policies requiring employees to work for a certain period of time before employer contributions to their accounts are vested. Plan sponsors and plan professionals GAO surveyed identified lowering costs and reducing employee turnover as the primary reasons that plans use these policies.

The Employee Retirement Income Security Act of 1974 (ERISA) allows plan sponsors to set eligibility and vesting policies. Specifically, federal law permits 401(k) plan sponsors to require that workers be at least age 21 to be eligible to join the plan. The law also permits plans to use rules affecting 401(k) plan participants' receipt of employer contributions and the vesting of contributions already received. However, over time workers have come to rely less on traditional pensions and more on their 401(k) plan savings for retirement security. Further, while the rules were designed, in part, to help sponsors provide profit sharing contributions, today 401(k) plan sponsors are more likely to provide matching contributions and today's workers may be likely to change jobs frequently. GAO's projections for hypothetical scenarios suggest that these policies could potentially reduce workers' retirement savings. For example, assuming a minimum age policy of 21, GAO projections estimate that a medium-level earner who does not save in a plan or receive a 3 percent employer matching contribution from age 18 to 20 could have $134,456 less savings by their retirement at age 67 ($36,422 in 2016 dollars). Saving early for retirement is consistent with Department of Labor guidance as well as previous legislation and allows workers to benefit from compound interest, which can grow their savings over decades. In addition, the law permits plans to require that participants be employed on the last day of the year to receive employer contributions each year, which could reduce savings for today's mobile workforce. For example, GAO's projections suggest that if a medium-level earner did not meet a last day policy when leaving a job at age 30, the employer's 3 percent matching contribution not received for that year could have been worth $29,297 by the worker's retirement at age 67 ($8,150 in 2016 dollars). GAO's projections also suggest that vesting policies may also potentially reduce retirement savings. For example, if a worker leaves two jobs after 2 years, at ages 20 and 40, where the plan requires 3 years for full vesting, the employer contributions forfeited could be worth $81,743 at retirement ($22,143 in 2016 dollars).The Department of Treasury (Treasury) is responsible for evaluating and developing proposals for legislative changes for 401(k) plan policies, but has not recently done so for vesting policies. Vesting caps for employer matching contributions in 401(k) plans are 15 years old. A re-evaluation of these caps would help to assess whether they unduly reduce the retirement savings of today's mobile workers.

                                Contents

Employers frequently sponsor and contribute to defined contribution plans, such as 401(k) plans, in which workers can save for retirement.¹ Workers can contribute to these plans before taxes are applied, which can encourage them to establish and grow their retirement savings.² The amount of foregone federal tax revenue associated with defined contribution plans, estimated at $62.1 billion in fiscal year 2015, reflects the prevalence of workplace plans as a source of retirement security.³ Plan sponsors -- employers who offer 401(k) plans -- have the authority and flexibility to customize certain aspects of plans, but such customization can also affect workers' retirement savings. For example, some plan sponsors may use eligibility policies to limit enrollment in their plans and use vesting policies to limit participants' ability to keep employer contributions upon leaving their job.⁴

Federal policy encourages workers to save for retirement through employer plans, and has removed some barriers that prevent plan participants from saving, such as by facilitating automatic enrollment and automatic increases in savings rates in workplace plans. Researchers have also studied the efforts of different groups of workers in saving for retirement by comparing groups and outcomes by age, gender, race, and other factors. However, few studies have looked at the effects of eligibility and vesting policies on a participant's ability to save for retirement in a workplace plan and to keep employer contributions when changing jobs throughout their career. You asked us to look at plan sponsors' use of eligibility and vesting policies. This report examines: 1) what is known about the prevalence of 401(k) plans' eligibility and vesting policies and why plans use them, 2) the potential effect of eligibility and vesting policies on workers' retirement savings, and 3) participants' understanding of these policies.

To identify what is known about the prevalence of eligibility and vesting policies, we developed a non-generalizable survey of plan sponsors and plan professionals. This web-based questionnaire included questions on 401(k) plans' use of eligibility and vesting policies.⁵ Beginning in May 2015, we provided online access to the survey through the publications of three industry groups who agreed to help notify their members and readers about our survey. We also included a link to the survey in an online forum for American Society of Pension Professionals and Actuaries (ASPPA) members, who are owners or senior managers of plan administration firms.⁶ We received and accepted responses through August 31, 2015. We received 80 responses from plan sponsors and plan professionals regarding plans ranging in size from less than 100 participants to more than 5,000.⁷ In addition, we reviewed industry survey data, plan data from a large 401(k) plan record keeper, and U.S. Census Bureau (Census Bureau) data. We assessed the reliability of these data by interviewing the staff who maintain the data, reviewing related documentation, and testing the data for missing or erroneous values.⁸ We determined that the data we used were sufficiently reliable for our purposes. To examine why plans use eligibility and vesting policies, we included questions in our survey about what factors were important to plans in choosing specific policies. We based these questions on information we obtained from interviews and a discussion group with plan professionals. The survey also provided respondents with the opportunity to cite their own reasons for using eligibility and vesting policies. We also interviewed officials from the Department of Labor (DOL), Treasury, the Internal Revenue Service (IRS), and the Securities and Exchange Commission, as well as a total of 21 retirement professionals and academic researchers to obtain their perspectives on why plans use eligibility and vesting policies. The views of the individuals we interviewed are not generalizable and therefore do not represent all potential views on eligibility and vesting policies.

To examine the policies' potential effects on retirement savings over time, we developed hypothetical scenarios based on assumptions drawn from federal and industry data sources.⁹ Most of these projections assume contributions that are based on medium scaled lifetime earnings factors developed by the Social Security Office of the Chief Actuary (OCACT).

We also interviewed academic researchers and retirement professionals, as described above, on what the policies' effects on retirement savings were likely to be.

To examine workers' understanding of eligibility and vesting policies, we administered a nongeneralizable questionnaire to employees of four companies and tested the accuracy of their answers against the policies in their plans' documents. We asked plan sponsors who responded to our plan sponsor survey and provided contact information if they would help us by voluntarily inviting their 401(k) plan participants to complete a questionnaire regarding their eligibility and vesting status in the plan. Four companies agreed to help us facilitate this participant questionnaire. We analyzed the accuracy of responses from 46 401(k) plan participants by comparing their responses about plan policies against the summary plan descriptions for their actual plans. This allowed us to determine the accuracy of their knowledge rather than rely on their self-reported knowledge. To understand what sources of information on eligibility and vesting policies participants may have, we analyzed survey responses from the 80 plan sponsors and plan professionals which identified the methods used by plans to communicate these policies. In addition, we interviewed retirement professionals and academic researchers to get their perspectives on participants' understanding of their plan's eligibility and vesting policies. See Appendix I for additional information on our methodology.

We conducted this performance audit from January 2015 to October 2016 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

Types of Contributions in a 401(k) Plan

A 401(k) plan provides eligible plan participants the opportunity to choose to contribute a portion of their earnings, commonly called elective contributions, to their own individual account in a retirement plan.¹⁰ These contributions may be taken out of an employee's salary before taxes. Some employees affirmatively enroll in their plan and elect how much of their pay they want to contribute. These employee contributions can be a set dollar amount or a percentage of pay, within annual contribution limits set by the IRS.¹¹ Some employees are automatically enrolled in employers' plans and their contributions set at a default rate, though they can opt to adjust the contribution level later. Under federal law, an employee's own contributions and any returns on those contributions always belong to the employee and are not forfeitable to the plan if they leave their employer.¹²

An employer may also contribute to a participant's account, though not every plan includes an employer contribution. Generally, employers' contributions to participants' 401(k) accounts are voluntary, though once incorporated into plan documents contributions must be made as described.¹³ Unless the plan is a "Safe Harbor 401(k)", "SIMPLE 401(k)" plan, or a "SIMPLE IRA", plan sponsors have flexibility to decide whether to provide employer contributions and how quickly these contributions become vested, to the extent permitted by federal law.¹⁴ Formulas used to calculate employers' contributions vary across plans. One form of employer contribution is called a match, which is based on the amount that an employee contributes to the plan.¹⁵ Alternately, employers may provide non-matching contributions based on employer profits. Employer contributions may be made solely at the employer's discretion or may be required by plan documents. Unlike an employee's own contributions, employer contributions can be forfeited when an employee separates from their job if the employee is not vested in the plan.

Eligibility Policies and Relevant Laws

The Employee Retirement Income Security Act of 1974 (ERISA) provides the legal framework for eligibility policies used by workplace retirement plans, including minimum-age and minimum-service policies. The rules were designed, in part, to help sponsors provide profit sharing contributions, but after 401(k) plans were introduced, employer matching contributions became common. Under ERISA, the maximum age that plans may require as a condition of plan eligibility is 21.¹⁶ The maximum period of service -- or length of tenure with an employer -- a 401(k) plan may require for plan eligibility is 1 year.¹⁷ (See table 1.)

      Table 1: Summary of 401(k) Plan Eligibility Policies Allowable under

Vesting Policies and Relevant Laws

A 401(k) plan's vesting policy can require participants to work for a certain period of time with an employer before they can keep all or some of their employer's contributions to their account and investment returns on that money when they leave their job.¹⁹ Federal laws require that a minimum percentage of employer contributions are vested after a certain period of time; however, plans can choose to allow employer contributions to vest faster or even immediately.²⁰ The shorter the period of service required for 100 percent vesting, the "faster" the vesting. The minimum percentage that must be vested at a given time depends on the type of vesting policy used by the plan, either cliff or graduated (see table 2).

   Table 2: Descriptions of Effects on Employer Contributions to 401(k) Plans

 

Figure 1: Percentage of Employer Contributions to 401(k) Plans That

 

Are Nonforfeitable, Per Year of Employment, Under Maximum Cliff and

 

Graduated Vesting Policies Permitted by Federal Law

 

 

 

Similarly, in addition to a minimum-service policy required to receive employer contributions, plans may use a "last day policy," which can require up to an additional year of employment to earn the employer contribution for that year. When used, a last day policy applies to the employer contributions made each plan year, year after year (see fig. 2). According to recent industry data, plans offer matching contributions more commonly than non-matching contributions, which include profit sharing contributions.

 

Figure 2: A "Last Day" Policy for Employer Contributions

 

to 401(k) Plans Adds to Existing Service Requirements and

 

Continues Year after Year

 

 

 

Plan Policy Information Provided to Participants

Under federal law, plan sponsors must provide participants with a basic description of the plan, called a summary plan description (SPD), that explains participants' rights and responsibilities under the plan as well as the plan's key features, including eligibility and vesting policies.²⁵ Table 3 provides some key features of an SPD and other required plan documents.

        Table 3: Key Features of Required 401(k) Plan Communications to

Treasury and IRS, an agency within Treasury, share responsibility for overseeing provisions of ERISA applicable to eligibility and vesting in 401(k) plans. IRS has primary responsibility for overseeing eligibility and vesting policies and has promulgated regulations in these areas. For example, IRS regulations state that plans risk losing their tax qualified status if they impose policies that 1) do not specifically refer to service but have the effect of requiring service as a condition for plan participation, and 2) ultimately result in employees being excluded from participating in the plan for a period of time that exceeds the plan's stated minimum service policy.²⁶ IRS has also promulgated regulations for rules relating to "year of service." Additionally, IRS regulations and federal law address the timing of employer contributions -- when they must be deposited into participants' accounts in a defined contribution plan.²⁷ Treasury is responsible for developing proposals for legislative changes, which could include changes regarding eligibility and vesting requirements. In fulfilling this duty, Treasury prepares the "General Explanations of the Administration's Revenue Proposals," referred to as the "Greenbook," which accompanies the President's annual budget submission and outlines the President's tax-related legislative proposals.²⁸

DOL is responsible for prescribing regulations governing retirement plans in a variety of areas, including reporting, disclosure, and fiduciary requirements.²⁹ According to DOL officials, DOL could issue guidance on best practices to plan sponsors to help them better communicate plan policies in the summary plan description (SPD). With regard to eligibility, vesting, and related policies, DOL regulations impose certain requirements, such as the requirements that plans:

• Transmit employees' own contributions into their account no later than the 15th day of the month following the month in which the money comes out of their pay.³⁰

• Report and disclose certain information to participants regarding eligibility and vesting policies.³¹ ERISA specifically grants DOL the authority to prescribe the format and content of the SPD and other statements or documents which are required for plan participants and beneficiaries receiving benefits under a plan.³² In addition, DOL may prescribe regulations covering the format of these disclosures.³³

U.S. workers are likely to have multiple jobs throughout their careers. Each time an employee begins a new job where a 401(k) plan is offered, the plan's eligibility policy may affect the employee's ability to participate. Likewise, every time a plan participant leaves a job, the plan's vesting policy may affect the participant's ability to retain employer contributions to their account. According to workforce data collected by the federal government, from 1978 to 2012, the average number of total jobs held by men and women workers from age 18 to 48 was more than 11 (see table 4). The mobility of the U.S. workforce is also reflected by the median tenure, which was 4.1 years for private sector workers in January 2014, according to Bureau of Labor Statistics (BLS) data.³⁴

       Table 4: Average Total Jobs Held by Individuals from Age 18 to 48,

Many Plan Sponsors In Our Survey Use Policies that Determine Workers' Eligibility to Save in 401(k) Plans and to Retain Employer Contributions, Often to Reduce Costs and Employee Turnover

Many Plan Sponsors In Our Survey Use Minimum-Age and Minimum-Service Policies to Determine Plan Eligibility

Information from our non-generalizable survey of 80 plan sponsors ranging in size from fewer than 100 participants to more than 5,000, and our review of industry data, show that many 401(k) plans have minimum-age policies that do not allow workers to save in plans until they reach age 21 instead of immediately upon employment.³⁵ Our survey data found that 43 of 80 plans surveyed have minimum-age policies for plan eligibility, with 21 being the most frequently used minimum age, used by 33 plans.³⁶ Industry data from the Plan Sponsor Council of America (PSCA), which primarily cover large plans, also show that 376 out of 613 plans have minimum-age policies, with 21 the most frequently used minimum age.³⁷

According to our analysis of 2008 Survey of Income and Program Participation (SIPP) data from the Census Bureau, an estimated 405,000 workers whose employers offer 401(k)-type plans said they were ineligible to participate in the plan because of a minimum-age policy.³⁸ This is equal to about 2 percent of workers nationwide who were not participating in 401(k)-type plans offered by their employers that year.³⁹

Minimum-service eligibility policies require employees to work for an employer for a certain period of time before they can enroll and participate in a 401(k) plan. Fifty of the 80 plans we surveyed reported they have minimum-service eligibility policies. Twenty plans use a 1-year minimum. Industry data from Vanguard also show that about 40 percent of its plans have minimum-service eligibility policies, with a 1-year requirement the most prevalent policy among these plans.⁴⁰ The Census Bureau's SIPP data show that an estimated 16 percent of U.S. workers who do not participate in 401(k)-type plans offered by their employers said it is because they have not worked long enough.⁴¹ Projected to the total workforce nationwide, this would amount to about 3 million workers who are not able to save in their employers' 401(k) plans because of a minimum-service eligibility policy.

Certain types of 401(k) plans may be more likely than others to have policies that require employees to complete a minimum period of service before they can save in the plan. Vanguard found, among its plan clients, that small plans are more likely than large plans to use minimum service eligibility policies.⁴² PSCA also found that small plans are more likely to use minimum-service policies, as their data show that the percentage of plans with a minimum-service eligibility policy decreases as plan size increases.⁴³

Under federal law, plans have discretion to require a period of minimum service for plan eligibility, and some discretion on how long that period will be and how service will be measured. Our survey found that of 80 plans, 20 required employees to work a certain number of hours per year to be credited with a year of service for the purpose of meeting a plan's minimum-service eligibility policy. Of those 20, 12 required employees to work 1,000 hours during the year, which is the maximum number of hours that may be required under current law.⁴⁴ PSCA data covering a larger number of plans show that the 1,000 hour minimum is a common service requirement, used by about 30 percent of plans.⁴⁵ According to the most recently available SIPP data, an estimated 24 percent of U.S. workers who reported that they do not participate in their employer's plan said they do not do so because they do not work enough hours during the year.⁴⁶

Many Plans Use Policies that Affect the Receipt and Timing of Employer Contributions

Minimum-service policies that govern the receipt of employer contributions can prevent plan participants from receiving them. In our survey of 80 plans, 34 reported minimum-service policies for employees to receive employer contributions.⁴⁷ Vanguard data show that about 50 percent of plans have such policies, with a 1-year minimum-service policy being the most frequent.⁴⁸ Employers can use these policies in addition to minimum-service policies for 401(k) plan eligibility -- which determine a participant's ability to save his or her own earnings in a plan -- but ERISA caps the length of service needed to receive employer contributions at 2 years. Therefore, after a 1-year minimum-service policy for plan eligibility, a plan could only require 1 additional year of service prior to a participant receiving employer contributions.

The receipt of employer contributions can also be limited by "last day policies", which are additional policies that can, in some cases, prevent a participant's receipt of employer contributions. These policies require workers to be employed on the last day of the plan year to be eligible to receive an employer contribution for that year.⁴⁹ Plans can use these policies alone or in combination with minimum-service policies. Whether a worker leaves a job voluntarily or is laid off or fired, the worker can be affected by a last day policy even after satisfying up to a 2-year service policy for employer contributions. Our survey of plans found that 19 of 80 plans have last day policies.

Delaying employer contributions to employees is another policy that can affect the receipt of employer contributions. Twenty-four plans we surveyed provide contributions, including all types of employer contributions, on an annual basis instead of quarterly or per pay period.⁵⁰ With regard to matching contributions, PSCA's survey found that 18 percent of plans make those contributions on an annual basis rather than on a more frequent schedule (see fig. 3).⁵¹ Our survey data also show that plans often use an annual schedule for contributing to participant accounts in concert with a last day policy. In total, 12 of the 24 plans we surveyed that provide contributions on an annual basis said they also have last day policies.

 

Figure 3: Plans Use a Range of Schedules for

 

Providing Employer Matching Contributions

 

 

 

 

Figure 4: 401(k) Plan Policies Can Result in Employer Contributions

 

to Participant Accounts Being Delayed More Than a Year

 

 

 

Based on our survey of 80 plans, we found that 57 have vesting policies that require employees to work for a certain period of time before the employer contributions in their accounts are vested.⁵³ Among survey respondents, a 6-year graduated vesting schedule was used most frequently, meaning a worker must complete 6 years of employment before all of their employer contributions to their account would be retained if they left their job.⁵⁴ Industry data from Vanguard show that more than 55 percent of its plans also have vesting policies for matching contributions.⁵⁵ A 5-year graduated vesting policy was the most frequent. We also found that plans do not often change their vesting policies. For example, about 70 of the 80 plans we surveyed had not changed their vesting policies over the past 5 years.⁵⁶ A retirement professional we interviewed concurred that vesting policies have remained stable in recent years.

Plans Use Eligibility and Vesting Policies to Lower Costs and Reduce Employee Turnover

Based on our survey of plan sponsors and interviews with academic researchers, retirement professionals, and government officials, we found that plans cited lowering costs and reducing employee turnover as the most important reasons for using eligibility, vesting, and other related policies. Our survey asked respondents to identify the importance of various reasons for having eligibility and vesting policies (see table 5).⁵⁷

         Table 5: Most Frequently Cited Reasons for Using Eligibility,

Reducing the direct costs of employer contributions was also cited by plans that we surveyed as an important reason for using eligibility policies, although savings for some plans may be minimal.⁶² Minimum-age policies reduce costs because employers do not need to make contributions to otherwise eligible workers. If plans do not use minimum-age policies they can incur some added direct costs in the form of employer contributions to additional workers. However, our projections suggest that the direct costs associated with including workers under age 21 in existing plans may be small because younger workers earn less, on average, than older workers, so the same percentage match from the employer costs less for younger workers.⁶³ According to a retirement professional, plans may use a minimum-service policy for employer contributions to reduce costs for sponsors because the sponsor avoids making contributions to a participant account if that worker leaves their job before satisfying the policy.

Eligibility policies and policies that delay employers' contributions until the end of the year are also used by some plans for administrative convenience, according to government officials we interviewed and plans we surveyed, though delayed enrollment is possible without delayed eligibility and today's employer contribution formulas make delayed employer contributions often unnecessary. First, government officials said that some plans may use eligibility policies to delay workers' enrollment because it takes administrators time to determine eligibility and to process workers' enrollment in the plan. However, plans may delay a worker's enrollment up to 6 months after eligibility, which can help plans to mitigate administrative concerns.⁶⁴ According to PSCA's survey, about 25 percent of 401(k) type plans delay the enrollment of newly eligible workers and the rest enroll workers "anytime" after they are eligible.⁶⁵ Moreover, a retirement professional told us that, for medium and large-size plans, the payroll systems used to administer plan benefits can support the provision of employer contributions during the year as opposed to just at the end of the year. Second, some plans that we surveyed indicated they also delay employer contributions until the end of the year for administrative convenience. For example, one respondent reported that it is less work for the plan to make employer contributions once, at year's end. However, a government official told us that the reason most plans make employer contributions each pay period is that doing otherwise would have adverse effects on the plan sponsor by requiring a large outflow of cash at the end of the plan year. By making employer contributions more frequently, a company can spread that cost over the year. A retirement professional told us a likely reason that some plans make delayed employer contributions is employer inertia in keeping a policy that is a remnant of the past, when employer contributions were based on the employers' year-end profits. However, today, most plans are 401(k) plans and often provide matching employer contributions based on participants' own contributions, which are typically made each pay period.

Vesting policies can also reduce costs for plan sponsors by resulting in forfeited employer contributions when participants separate without fully satisfying the vesting policy. Those forfeitures can offset employer expenses and contributions. According to retirement professionals we interviewed and plans we surveyed, vesting policies reduce the direct cost of employer contributions for shorter-tenure employees who do not stay employed long enough to satisfy the vesting policy and keep employers' contributions.⁶⁶ However, tax benefits to employers from making employer contributions can partially offset direct costs to employers of making contributions to participant accounts, because they reduce an employer's taxable corporate income in proportion to the amount spent on contributions.⁶⁷

Finally, both eligibility and vesting policies were also used to reduce employee turnover by plans we surveyed. Retirement professionals and an academic researcher we interviewed explained that some plans see delayed eligibility in the 401(k) plan as an incentive that may convince new employees to stay longer in their job. Sponsors also use vesting policies to reduce turnover, according to government officials, retirement professionals, and an academic researcher we interviewed. For example, a retirement professional said that companies with a generous matching contribution and high employee turnover prefer to use a vesting schedule because it incentivizes employees to stay with the employer. However, the extent to which vesting policies are effective in changing behavior and reducing turnover depends in some measure on whether participants actually understand vesting policies. For example, if a participant incorrectly believes he or she is fully vested but is actually only partially vested, the vesting policy will not effectively incentivize the worker to extend tenure to fully vest.⁶⁸

GAO Hypothetical Projections Suggest Current Plan Eligibility, Employer Contribution, and Vesting Policies Can Result in Lower Retirement Savings

Plan sponsors use eligibility and vesting policies for a number of reasons. However, based on our projections, we found that although the effects on a worker's retirement savings can initially be minimal from policies affecting plan eligibility, eligibility to receive employer contributions and the timing of those contributions, and vesting of employer contributions, the cumulative effects can potentially result in significantly lower retirement savings, depending on the policies used (see table 6). (Also see Appendix II for detailed information about the assumptions used in these hypothetical projections.)

     Table 6: Based on the Results of GAO's Hypothetical Scenarios, Current

Currently, the law permits 401(k) plans to require a minimum age of 21 and at least 1,000 hours of service over 1 year for a worker to be eligible to join an employer's 401(k) plan. As discussed earlier, we found that plan sponsors may use these policies to reduce the costs and challenges incurred when short-term workers enroll in the plan and leave behind small accounts. While the law gives plan sponsors flexibility in establishing plan eligibility policies that meet their needs, federal caps on minimum age and minimum service policies serve to balance plan sponsors' needs with workers' interests in accessing and saving for retirement in workplace plans.⁶⁹ In addition, because of the potential for compound interest to grow savings over time, it is a widely accepted best practice for workers to start saving for retirement as early as possible, even if the amounts they save seem small.⁷⁰ Our projections suggest that these eligibility policies can potentially reduce workers' retirement savings.

Our Estimates Suggest that Plans' Use of Minimum-Age Policies Can Reduce Workers' Access to Retirement Plans and Accumulation of Savings

Our estimates, which project hypothetical retirement savings, suggest that minimum-age policies can potentially reduce young workers' future retirement savings. Minimum-age eligibility policies can lower workers' potential retirement savings through the loss of compound interest as well as employer contributions, and the policies disproportionately affect some groups. Because minimum-age policies can prevent young workers from saving for retirement in their workplace 401(k) plan early in their careers, they miss the opportunity to accrue compound interest and grow their initial contributions over the remaining decades of their working life.⁷¹ For example, an 18-year-old worker earning $15,822 per year who is ineligible to enroll in an employer's plan until age 21 may forego savings at retirement of $85,857⁷² ($23,258 in 2016 dollars) if they had been able to save and invest 5.3 percent (the average contribution level for non-highly compensated employees reported by plans in PSCA's 57th Annual Survey) of their salary over those 3 years.⁷³

Because of the effects of compound interest, saving for retirement at a young age is a one-time opportunity to optimize retirement savings despite making what is typically a low salary relative to lifetime earnings. An older worker has to save more to make the same gains over time, because the returns on savings made later have fewer years over which to compound. (See fig. 5.) (Also see Appendix II for detailed information about the assumptions used in these hypothetical projections.)

 

Figure 5: Compounding Means a Younger Worker Can Save Less than an

 

Older Worker for Same Savings Amount at Retirement

 

 

 

Minimum-age eligibility policies may mean foregone contributions from an employer. The amount of foregone retirement savings due to minimum-age policies can be higher when matching employer contributions are considered. For example, based on analysis of our hypothetical scenario, the savings of an otherwise eligible 18-year-old earning $15,822 per year could have been $134,456 at retirement, or $36,422 in 2016 dollars, if that individual had also received an employer match of their contribution up to 3 percent of salary from age 18 to 20.⁷⁹ An employer's match of employee contributions is typically, in effect, a 50 to 100 percent return on the employee's contributions; participants who receive an employer match can double the value of their contribution -- a 100 percent return -- if the employer makes a dollar for dollar match.

ERISA provides plan sponsors some flexibility to design plan policies that can restrict workers' eligibility to enroll and save in plans. ERISA specifically permits sponsors to limit enrollment in 401(k) plans to workers age 21 and older.⁸⁰ Over 8 million workers under the age of 21 are potentially subject to this policy.⁸¹ In passing ERISA, Congress supported the policy goal of increasing access to plans by allowing workers to save for retirement as early as possible.⁸² Increasing access to plans was also a policy goal supported by Congress more recently in passing the Pension Protection Act of 2006.⁸³ The current minimum-age policy does not further that goal. Extending eligibility to workers at an age earlier than 21 would also give young workers an opportunity to build their private sector savings at the same time they are earning credits toward future Social Security retirement benefits. In addition, increasing workers' access to workplace retirement plans is a current federal policy goal, reiterated by DOL guidance⁸⁴ and recently in DOL's fiscal year 2017 Budget Justification.⁸⁵

Employers may bear costs from enrolling additional workers in plans and administering their accounts after they leave their employ, but our prior work has shown that participants often bear the costs of administering their accounts and plans can use forced-transfers to eliminate small accounts left behind by separated employees.⁸⁶ IRS officials told us that the minimum-age policy is unnecessary because the minimum-service eligibility policy permits plans to exclude short-term employees. By extending eligibility to workers at an age earlier than 21, private retirement plan coverage could expand for young workers who research shows lack access to 401(k) plans.⁸⁷ We recently reported that when given the opportunity, young, low-income workers participate in workplace plans at high rates.⁸⁸

Allowing young people to contribute at the beginning of their careers would also help to mitigate the risk that potential unexpected events could reduce the length of their careers and the period to save for retirement and, thus, their retirement savings. For example, research shows that many workers retire sooner than they expect, due to physical limitations or the need to care for family members.⁸⁹ Such individuals will not save for retirement for as long as they had planned or during years in which their contributions may have been highest.⁹⁰ Extending eligibility to workers at an age earlier than 21 could help a significant number of workers to save at an earlier age and those who experience unforeseen absences from the workforce or premature retirement will be better positioned to maximize their retirement savings during their working years.

1-Year Minimum-Service Eligibility Policies Delay and May Prevent Plan Participation and Our Estimates Suggest they May Result in Lower Savings

Minimum-service eligibility policies of up to 1 year delay access to workplace 401(k) plans and can reduce potential retirement savings for workers of any age, according to our projections of retirement savings. For example, our projections suggest that for a 30-year-old worker earning a salary of $71,841 ($52,152 in 2016 dollars), a 1-year delay in plan eligibility could mean $51,758 less in savings at retirement ($14,021 less in 2016 dollars). That amounts to 3 percent of the worker's total projected retirement savings from their own savings alone. Including the employer match of 3 percent not received during the 1-year period of ineligibility, the worker could have $81,055 less at retirement ($21,957 in 2016 dollars).⁹¹ A minimum-service eligibility policy also means that any workers excluded from participating in a plan also will not receive any employer contribution to which they might otherwise be eligible.

Additionally, the more often a worker changes jobs the larger the potential effect of a minimum-service eligibility policy on the worker's retirement savings (see fig. 6). (See Appendix II, Table 5 for a similar table with values adjusted for inflation.) A longitudinal study by the BLS found that the average number of jobs for individuals born in the latter years of the baby boom was 11.7 jobs.⁹² The same study found that about half those jobs were held between ages 18 and 24. Being ineligible to save in a new employer's plan for 1 year on 11 occasions, especially occurring more frequently early in a worker's career, may result in $411,439 less retirement savings ($111,454 in 2016 dollars), based on our projections.⁹³

 

Figure 6: Potential Accumulated Value of Lost Savings

 

from Repeated Incidents of 1-Year Ineligibility for Participation

 

in a Workplace 401(k) Plan

 

 

 

In first establishing the rules for a minimum-service policy for plan eligibility, ERISA capped such policies at 1 year of service, defined as 1,000 or more hours worked over 1 year.⁹⁷ While plans can require fewer hours or no hours of service for plan eligibility, they cannot require more than 1,000 hours.⁹⁸ However, millions of part-time workers may never qualify for their employer's plan with a 1,000-hour requirement. Some members of Congress and the current administration have proposed amending the law to ensure that long-term part-time workers can become eligible to save in their employer's workplace plan. As part of its fiscal year 2017 budget submission, the current administration proposed amending federal law to require plans to expand eligibility to workers who have worked for their employer at least 500 hours per year for 3 consecutive years, allowing workers to contribute their own earnings, but not necessarily to receive employer contributions to their account.⁹⁹ A 2015 Senate Finance Committee bi-partisan working group endorsed this proposal and legislation was introduced that incorporated it.¹⁰⁰ Prior to the current administration's proposal, legislation was introduced that also would have required plans to cover long-term part-time employees.¹⁰¹

Given today's workforce, determining whether the current definition of "year of service" is consistent with the goal of expanding access to workplace retirement savings plans would be beneficial to all workers. Without revising the definition of "year of service", minimum-service eligibility policies may continue to reduce potential retirement savings for millions of workers who will remain ineligible to participate in a plan because their annual hours of service may fall below their plans' requirement. Some retirement professionals we interviewed said that operating without a 1,000 hour rule could result in more small accounts left behind by short-term workers, but as we previously noted, forced transfers of small balances can help plans manage any associated burden or cost, and often those costs are borne by participants themselves.

Our Projections Suggest that Current Employer Contribution Policies Can Also Reduce Workers' Retirement Savings in 401(k) Plans

Current law permits 401(k) plan sponsors to require participants to be employed on the last day of the plan year to be eligible to receive employer contributions to their account. Current law also permits 401(k) plan sponsors to delay the accrual of employer matching contributions until the end of the year. Information from our survey of 80 plan sponsors and plan professionals showed that plan sponsors often use these two requirements together. Based on our review of relevant statutes and interviews with retirement professionals and government officials, we found that these provisions were created decades ago when 401(k) plans did not exist and when profit sharing contributions were the norm. According to Treasury officials, the provisions met plan sponsors' need to wait until the end of the year to identify what their profits were for the year and thus what the employer's contribution would be for the year. However, today most employer-based plans are 401(k) plans, not traditional profit sharing plans, and based on our analysis of industry survey data, most 401(k) plans make matching contributions, which are based on participants' contributions throughout the year.¹⁰² Although plan sponsors may have previously found these two policies to be beneficial, our projections suggest they may also potentially reduce workers' retirement savings.

Last Day Policies Can Reduce Workers' Retirement Savings, Based on Our Projections

The law permits plans to apply policies that limit a participant's ability to receive employer contributions without additional service each year -- last day policies -- which can reduce participants' potential retirement savings. Although last day policies can provide financial benefits to plan sponsors and may also ease plan administration, these policies can reduce potential retirement savings for workers. Given a relatively mobile workforce, the requirement to be employed on the last day of the plan year to receive an employer's contributions for that year puts workers who separate from their job at risk of losing some of their potential retirement savings. For example, our projections suggest that for a 30-year-old earning $71,841 ($52,152 in 2016 dollars), the employer contribution not received due to an unmet "last day" policy is $2,155 in that year ($1,443 in 2016 dollars). But it could be worth $29,297 by retirement ($7,936 in 2016 dollars), which is 3 percent of the worker's total projected savings from employer contributions of $969,674 ($262,673 in 2016 dollars) at retirement.¹⁰³

Our projections also suggest that a last day policy can reduce potential retirement savings to the same extent as a 1-year minimum-service policy for employer contributions, except the potential savings are lost in the last year rather than in the first. Last day policies can reduce potential retirement savings for even long-tenure, full-time workers who separate from their employers before the official last day of the year. For example, a 67-year-old employee who has worked for the same employer during his or her entire career and retires when eligible for full Social Security benefits could lose the employer's $6,606 match for that last year ($1,837 in 2016 dollars) if it is before the end of the year.¹⁰⁴ Moreover, the last day policy can affect workers repeatedly throughout their careers. If our hypothetical worker leaves three jobs without satisfying a last day requirement, at ages 20, 30, and 40, the worker could lose a total of $6,542, which could be worth $69,583 at retirement ($18,849 in 2016 dollars).¹⁰⁵

ERISA caps at 2 years the length of service that a plan sponsor can require before a participant is eligible to receive employer contributions. Given the mobility of the workforce, this provision ensures that workers who change jobs are eligible, just like long-tenured workers, to benefit from employer contributions for which employers receive a tax benefit.

Given that job turnover is greater among younger workers, ERISA's cap on a service policy that delays employer contributions also helps to ensure that workers' savings in these tax-advantaged accounts are not excessively diminished at a time when, our projections suggest, contributions have the greatest ability for compound earnings over time to improve retirement savings. However, because a last day policy requires an additional year of service to receive any employer contributions, year after year, a worker who has already satisfied a 2-year minimum-service policy would have to wait up to another year to receive an employer contribution. (See Appendix II Table 8 for examples of how a last day policy could potentially reduce retirement savings.)

IRS oversees provisions of federal law applicable to eligibility and vesting and Treasury is responsible for developing proposals for legislative change in these areas. However, according to an IRS official, the law permits 401(k) plans to use a last day policy and, as long as the law's non-discrimination and coverage rules are satisfied, plans are generally free to structure their plans as they choose. In addition, according to that IRS official, the flexibility provided to plans by ERISA prevents IRS from prohibiting plans' use of a last day policy.¹⁰⁶ Considering whether an adjustment to the law's provisions regarding plans' use of a last day policy is needed could help to ensure that 401(k) plan policies reflect the current mobility and characteristics of today's workforce.

ERISA Allows Employers to Delay Employer Contributions, Which Our Projections Suggest Also May Reduce Potential Savings

Plan policies that delay employer contributions so that they are paid at the end of the year rather than being paid in tandem with employee contributions throughout the year affect participants' opportunities to earn compound interest on investment returns on the employer's contribution to their account over the course of the year. Our projections suggest it may also reduce participants' potential retirement savings. In contrast, regular employer contributions, such as those made bi-weekly, allow participants to potentially profit from the investment of that money and the reinvestment of those profits.¹⁰⁷ Delayed employer contributions may seem negligible at first, but, if left to compound over time, our projections suggest that the return on that employer contribution can amount to significant savings. Moreover, while the immediate value of savings lost in a single year can be relatively small, the potential value of that lost opportunity compounds year after year.¹⁰⁸ For example, our projections suggest that for a worker who remains with one employer throughout his or her career, the delay of the employer's contribution until the end of the year, each year, could mean $35,636 less in total savings at retirement ($9,653 in 2016 dollars) than if the employer's contribution was made on a per pay-period basis (bi-weekly), in concert with the worker's own contributions.¹⁰⁹ That is about 3.7 percent of our hypothetical worker's total 401(k) retirement savings based on employer contributions alone.

ERISA establishes rules for the accrual of retirement benefits in workplace retirement plans.¹¹⁰ Those rules permit flexibility as to when an employer's contributions go into -- or accrue to -- an individual's retirement account. ERISA permits plan sponsors to delay making the employer contribution to participants' accounts until as late as the end of the year or the date when the plan's tax return is filed, including extensions. However, the law permitting delayed employer contributions is from a time when profit sharing plans were the norm and plans typically waited until the end of the year to calculate and distribute employer contributions. Those plans predate 401(k) plans and the matching contributions that are now commonplace. Treasury officials told us that, because plan sponsors today generally make contributions that match a participant's own deferrals, delayed employer contributions are no longer necessary in most defined contribution plans. In addition, delayed employer contributions are inconsistent with the best practice of saving for retirement as early as possible.¹¹¹ While IRS has primary responsibility for overseeing eligibility and vesting policies and Treasury is responsible for developing proposals for legislative change in these areas, the provision permitting delayed employer contributions can only be changed by statute and not by either agency. Considering whether the law's provisions regarding the timing of matching employer contributions should be adjusted could be an opportunity to help ensure the provisions reflect the current mobility and characteristics of today's workforce.

Vesting Policies That Our Projections Suggest May Affect Participants' Retirement Savings when They Separate from a Job Have Not Been Recently Evaluated by Regulators

Our projections suggest that vesting policies can also reduce retirement savings when participants leave their job and the vesting policy is not satisfied in full.¹¹² (See fig. 7.) For example, our projections suggest that for a worker who twice separates from employment (at age 20 and 40) after 2 years without satisfying a 3-year cliff vesting policy, forfeiting the employer contributions already in their account, the lost savings could have grown to $81,743 by retirement ($22,143 in 2016 dollars).¹¹³

 

Figure 7: Potential Value of Lost Retirement Savings Due to 401(k)

 

Vesting Policies, Based on our Projections

 

 

 

Clearer Policy Descriptions by 401(k) Plans Could Help Participants to Better Understand Eligibility and Vesting Policies

GAO's Nongeneralizable Survey Found that While Knowledge Varied, Some Plan Participants Lacked Knowledge on Eligibility, Employer Contribution, and Vesting Policies

To examine plan participants' understanding of eligibility, employer contribution, and vesting policies, we analyzed 46 responses to an online survey administered to participants in four 401(k) plans in which they were asked about their own plans. Responses to the survey show that participants' knowledge varied.¹¹⁸ (See table 7.) The accuracy of responses was highest regarding the frequency of the employer's contribution, participants' initial eligibility to join the plan and the types of employer contributions made. However, based on our evaluation, some participants surveyed lacked knowledge of their plans' eligibility, employer contribution, and vesting policies that, as our projections discussed earlier suggest, can have important effects on retirement savings. Our survey results may reflect some degree of unwarranted confidence by plan participants regarding their understanding of plan policies. According to a behavioral finance expert, about one-quarter of people generally overestimate their understanding of finance-related terminology.¹¹⁹ An individual's lack of understanding of their employer's policies can result in suboptimal choices, such as choosing a job with a minimum-service eligibility policy over one offering immediate eligibility, when wages, working conditions, and other benefits are comparable.

           Table 7: GAO's Nongeneralizable Survey of 46 Participants'

Among those not immediately eligible to join the plan and asked to identify the reason for their initial ineligibility, more than half provided incorrect answers.¹²² That result is generally consistent with a national survey on self-reported financial literacy, which asked about 5,000 current defined contribution plan participants to assess their own knowledge of "eligibility requirements," among other topics.¹²³ That survey found that while 85 percent felt they had a working knowledge of and understood the term "eligibility requirements," nearly half (46 percent) were not confident enough in their understanding to teach others (see fig. 8).

 

Figure 8: Plan Participant Responses to a National Survey on

 

Financial Literacy Question About Their Understanding of the Term

 

"Eligibility Requirements"

 

 

 

Lastly, understanding of vesting requirements was mixed. More than half of participants gave accurate answers regarding their vesting status, given their plans' policies. Clearly written plan documents may also have helped those participants who understood their vesting status. Federal law requires that a plan's summary plan description (SPD) be written in a clear manner and uses a table to describe the maximum vesting schedule for 2- to 6-year vesting, which some plans use as a model in their SPD. At one of the companies where the majority of participants gave correct responses, the plan documents clearly stated the plan's 5-year graduated vesting policy and the calculation of vesting status based on different lengths of service by using a table (see fig. 9).

 

Figure 9: Vesting Schedule for Employer Contributions

 

in One 401(k) Plan

 

 

 

Unclear Plan Documents May Hinder Participants' Understanding of Their Employers' Eligibility and Vesting Policies

We found evidence that some eligibility and vesting policies in summary plan descriptions (SPD) can be unclear. Our review of five SPDs found that some eligibility and vesting policies were written using complex technical language which may make them less likely to be easily understood by the average plan participant.¹²⁸ (See text box.) Participants in two discussion groups comprised of plan sponsors and plan advisors which we convened in March 2015 also said that plan participants probably do not understand the eligibility and vesting requirements of their plan because the SPDs use complex legal terms.¹²⁹ Furthermore, one retirement professional we interviewed said that employers' own difficulty understanding their plans' eligibility and vesting policies contributes to employees' misunderstanding of these policies. According to a retirement professional, errors in interpretation can occur, making the clarity of plan documents more important.

Plans can also include information in their SPDs about provisions that are not used by the plan, a practice which is not explicitly prohibited by ERISA, making it difficult for participants to know which contribution is currently being offered and which eligibility and vesting policies apply. One plan sponsor explained that this is a common practice because plans want to avoid the time and expense of revising the plan documents later if they decide, for example, to change the type of employer contribution. One SPD we reviewed contained information about six employer contributions when only one contribution was offered (see fig. 10). This practice leaves participants with extraneous information and no clear way to tell what policies apply to them without additional information from their employer. Treasury officials said that it may be necessary for employers to describe contributions that they are not currently offering to employees, should they need to make these contributions at a later date in order to pass the Internal Revenue Code nondiscrimination tests. However, those officials agreed that the description of multiple employer contributions in SPDs can be confusing to participants as they may not be able to determine which contribution currently applies to them.

 

Figure 10: Comparison of Multiple Employer Contributions to a

 

401(k) Plan Included in a Summary Plan Description versus Only One

 

Actually Offered by the Employer

 

 

 

 

Figure 11: Sample 401(k) Plan Highlights Document

 

 

 

Conclusions

401(k) plans were created four years after ERISA was passed in 1974. Since that time, 401(k) plans have become the dominant employer-sponsored plan relied on by employees for retirement savings. Understanding the eligibility and vesting policies used by these plans has become increasingly important for workers, employers, and regulators. However, some of these policies were created to address issues when plans were more of a supplemental source of retirement income, in addition to a traditional pension, and reliant only on non-matching contributions from an employer rather than on a match of employee contributions. As 401(k) plans become the primary, and often sole, retirement savings vehicles for a large segment of the mobile workforce, there is a growing need to consider the effects of eligibility and vesting policies on workers, particularly those who are younger, less-educated, and with lower incomes. Moreover, one of the reported advantages of account-based plans like 401(k) plans is their enhanced portability over traditional pensions. Yet current rules and plan sponsor practices suggest some limitations on that portability, which can have potentially significant effects on retirement security.

Workers who must meet minimum-age eligibility policies to begin participating in a workplace plan miss out not only on contributing to 401(k) accounts, but also on the opportunity to receive employer matching contributions, which can significantly increase the amount contributed to their accounts. Saving early is particularly important for those who join the workforce out of high school and may never pursue higher education, with its associated higher wages. Our projections show that the inability to save in a 401(k) plan from ages 18 to 21 can result in tens of thousands of foregone retirement savings. While turnover among young workers can create administrative costs for an employer, these costs often are borne by participants in the form of fees. Extending plan eligibility to allow otherwise eligible workers to at least save their own contributions in their employers' 401(k) plans at an age earlier than 21 could help young workers improve their retirement security by saving through their workplace plan at a time when those savings have the most to gain.

Minimum-service eligibility policies can also affect employees' ability to save for retirement. Re-examining the legal definition of "year of service" used in minimum-service eligibility policies can help to ensure that the rules are consistent with today's mobile workforce and use of 401(k) plans. Opening 401(k) plans to more workers could result in additional small accounts, which are sometimes abandoned by their owner. But that challenge can be mitigated in ways that do not reduce savings. At the same time, opening 401(k) plans to more workers could help many who now lack coverage to access a workplace retirement plan and make tax-deferred contributions toward their retirement security.

Last day policies can also affect employees' retirement savings. Given the mobility of today's workforce, all workers are potentially affected by these policies. Further, policies that allow employers to delay making matching contributions to participants until the end of the year can also result in foregone savings. This type of policy can affect not only employees who separate from a job after a year or two, but also those who spend their entire career with the same employer. The policy is a remnant of the profit sharing plan era, when employers typically calculated benefits at the end of the year. Few employers rely on profit sharing contributions alone these days, so delayed employer contributions may be unnecessary. Considering whether these provisions should be adjusted could help ensure they reflect today's mobile workforce and use of matching contributions in 401(k) plans.

Vesting policies also present missed opportunities to improve savings for workers who are mobile. Given that the median length of stay with a private sector employer is currently about 4 years, the rule permitting a 6-year vesting policy may be outdated. Employees who forfeit employer contributions to their account when they leave a job prior to the end of the plan's vesting period lose the opportunity to have those funds grow in the plan or to transfer those contributions into their new employer's plan, reducing their retirement savings. A re-examination by Treasury of the appropriateness of current maximum vesting policies could help determine whether they unduly reduce the retirement savings of workers who change jobs.

Finally, having clear and concise information about their retirement plan's eligibility and vesting policies helps employees make informed decisions affecting their retirement savings. Guidance from DOL can help plan sponsors better inform participants about the plan policies that they must understand to make optimal decisions.

Matters for Congressional Consideration

To help increase plan participation and individuals' retirement savings, Congress should consider updating ERISA's 401(k) plan eligibility provisions to:

• extend plan eligibility to otherwise eligible workers at an age earlier than 21, and

• amend the definition of "year of service," given the prevalence of part-time workers in today's workforce.

Recommendations for Executive Action

To ensure that current vesting policies appropriately balance plans' needs and interests with the needs of workers to have employment mobility while also saving for retirement, Treasury should evaluate the appropriateness of existing maximum vesting policies for account-based plans, considering today's mobile labor force, and seek legislative action to revise vesting schedules, if deemed necessary. The Department of Labor could provide assistance with such an evaluation.

To help participants better understand eligibility and vesting policies, DOL should develop guidance for plan sponsors that identifies best practices for communicating information about eligibility and vesting policies in a clear manner in summary plan descriptions. For example, DOL could discourage plans from including in documents information about employer contributions or other provisions that are not actually being used by the plan sponsor.

Agency Comments and Our Evaluation

We provided a draft of this report to the Departments of the Treasury and Labor, and the Internal Revenue Service. Treasury provided technical comments, including those of IRS, which we have incorporated where appropriate, and oral comments, as discussed below. DOL provided written comments, which are summarized below and reproduced in Appendix III.

With respect to our recommendation that Treasury evaluate existing maximum vesting policies, Treasury had no formal comment. As we detail in our report on pages 44-46, given the effect that vesting policies can have on the retirement savings of mobile workers, we believe that it would be beneficial for Treasury to evaluate current vesting policies. Treasury may be able to incorporate an evaluation of these policies into the analysis it conducts in preparing the annual "Greenbook," highlighted in our report on page 12, and which accompanies the President's annual budget submission and outlines the President's tax-related legislative proposals. DOL, in its written comments, stated that substantive provisions of Title I of ERISA governing eligibility and vesting provisions in 401(k) plans are under the interpretive and regulatory jurisdiction of the Secretary of the Treasury. DOL also stated that Treasury and IRS generally consult with DOL on subjects of joint interest and it expects they will do so regarding our report.

With regard to our recommendation to develop guidance for plan sponsors that identifies best practices for communicating information about eligibility and vesting policies in a clear manner in summary plan descriptions, DOL agreed that disclosures explaining these policies are important to participants' ability to make informed choices about retirement savings. DOL also stated that under current law and regulations, this information must be written in a manner that can be understood by the average participant. DOL described planned actions that GAO believes are consistent with the intent of our recommendation. For example, DOL noted that an evaluation of best practices regarding eligibility and vesting should consider other disclosures provided to plan participants. DOL highlighted a long-term project on its current regulatory agenda relating to individual benefit statements, another type of disclosure provided to participants. DOL said that additional input from a broader range of plan sponsors and plan fiduciaries, possibly obtained through a Request for Information published in the Federal Register, could supplement the information we highlight in our report on pages 52-58 and contribute to an informed development of best practices guidance.

DOL stated that it did not agree that implementing the recommendation allows the best use of its limited resources. DOL stated it would not be appropriate at this time to reallocate resources away from its existing priority projects to a new best practices project focused on our recommendation, especially given the regulatory requirements that currently apply to summary plan discrimination disclosures on eligibility and vesting. However, DOL stated it would review its existing outreach materials on plan administration and compliance for opportunities to highlight the issues we raised in our report, as well as consider our recommendation in the ongoing development and prioritization of its agenda for regulations and sub-regulatory guidance. We agree with DOL that the efforts it plans to take in response to our recommendation, if fully implemented, will meet the intent of the recommendation and help plan sponsors more clearly communicate eligibility and vesting policies to plan participants without developing guidance.

As agreed with your office, unless you publicly announce the contents of this report earlier, we plan no further distribution until 30 days from the report date. At that time, we will send copies to the Secretaries of Labor and the Treasury, the Commissioner of Internal Revenue, and other interested parties. In addition, the report will be available at no charge on the GAO website at 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 key contributions to this report are listed in appendix IV.

Sincerely yours,

Charles Jeszeck

² Some plans also offer Roth accounts, in which participants can contribute earnings after taxes are paid. When taking eligible distributions later, withdrawals from those accounts are not taxed. 26 U.S.C. §§ 402A(a), (c)-(d), 408A.

³ Office of Management and Budget, Analytical Perspectives: Budget of the U.S. Government, Fiscal Year 2017 (Washington, D.C.: Feb. 9, 2016). 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 plan distributions to those who are currently receiving retirement benefits. Defined contribution plans are comprised largely of 401(k) plans.

⁴ For the purposes of this report, we use the term eligibility policies to refer to 401(k) plan policies that require employees to reach a minimum age (minimum-age policies) or work for a minimum length of time (minimum-service policies) before they can participate and save their earnings in a plan. In addition, in this report, we use the term vesting policies to refer to plan policies that require employees to work for a minimum length of time before their employer's contributions become fully nonforfeitable, meaning employees have an unconditional and legally enforceable right to keep their employer contributions if they separate from their job. To the extent permitted by federal law, some sponsors offer immediate eligibility and vesting, meaning their employees 1) do not have to reach a certain age or work for a minimum amount of time before they can save in a plan, and 2) do not have to work for a minimum length of time before their employer's contributions are nonforfeitable.

⁵ While plan sponsors and plan professionals of all defined contribution plans were invited to take the survey, because the majority of these plans are 401(k) plans according to Department of Labor data, we determined that it is reasonable to assume that the responses to our survey largely reflect the 401(k) plan experience. Therefore, we refer to the plans represented by survey respondents as 401(k) plans. See United States Department of Labor, Employee Benefits Security Administration, "Private Pension Plan Bulletin: Abstract of 2013 Form 5500 Annual Reports Data Extracted on 6/2/2015" (September 2015). The 2013 Private Pension Plan Bulletin was the most recent document available at the time of our review.

⁶ ASPPA is a non-profit professional organization with more than 7,000 members. Its goals are to educate retirement plan professionals and to create a framework of public policy that promotes retirement security.

⁷ Our web-based survey was an opt-in panel and open to anyone who received a link to the survey. Plan sponsors and other plan professionals who assist plan sponsors, such as human resource professionals, record keepers, plan administrators, and benefits consultants participated in the survey. The highest number of survey respondents identified themselves as plan sponsors. We asked for information regarding the plan sponsor's largest plan or the largest plan served by the plan professionals. We received 80 completed surveys. We evaluated the data by performing automated checks to identify inappropriate answers and checking for missing or ambiguous responses, following up with respondents when necessary to clarify their responses. On the basis of our application of recognized survey design practices and follow-up procedures, we determined that the data were of sufficient quality for our purposes.

⁸ See Appendix I for additional details on the steps we took to assess the reliability of the data.

⁹ Among other factors, we assumed an employee deferral rate of 5.3 percent of salary, an employer matching contribution of 3 percent of salary, employee and employer contributions made on a per pay period basis, and a return reflecting a mixed portfolio adjusted annually to decrease the allocation to equities over time. See Appendix II for more detailed information on our approach for developing hypothetical scenarios, including the assumptions we used.

¹⁰ 26 U.S.C. § 401(k)(2).These elective contributions are considered "employer contributions" under the Internal Revenue Code. However, for clarity in our report we refer to these contributions as employee contributions and only to additional contributions made by the employer as employer contributions.

¹¹ In 2016, employee contributions to 401(k) plans are limited to $18,000. Catch-up contributions, permitted for employees age 50 and over who participate in 401(k) plans, are limited to $6,000. Employer contributions are held to other limits. In 2016, the total annual additions to a participant's account by the participant and employer combined, cannot exceed the lesser of 100 percent of the participant's compensation, or $53,000 ($59,000 including catch-up contributions). 26 U.S.C. §§ 402(g)(1)(A)-(C), (4), 415(c), (d)(1)(C), (d)(2)-(4).

¹² 26 U.S.C. § 411(a)(1).

¹³ Plan sponsors must provide participants with a basic description of the plan that clearly explains participants' rights and responsibilities as well as the plan's key features. 29 U.S.C. § 1022.

¹⁴ A SIMPLE 401(k) plan is a type of plan that is intended to provide small businesses -- those with 100 or fewer employees who received at least $5000 in compensation in the preceding calendar year -- with a cost-efficient way to offer retirement benefits. 26 U.S.C. § 401(k)(11). In a SIMPLE 401(k) plan, the employer is required to make contributions that are fully vested, while employers that offer a traditional 401(k) plan can choose whether to provide employer contributions. A SIMPLE 401(k) plan is exempt from the Employee Retirement Income Security Act's "nondiscrimination requirements", which help to ensure that plan contributions do not discriminate in favor of highly-compensated employees. 26 U.S.C. §§ 401(a)(4), 410(b). Like SIMPLE 401(k) plans, Safe Harbor 401(k) plans are also exempt from nondiscrimination testing, and must provide for employer contributions. 26 U.S.C. § 401(k)(12)-(13). There are different safe harbor options. In one option, employer contributions may be matching contributions. Alternatively, contributions may be made on behalf of all eligible participants, regardless of whether they make elective contributions. The amount of employer matching contributions must comply with specific limits. 26 U.S.C. § 401(m). A SIMPLE IRA is a common account based retirement plan that is used by small employers.

¹⁵ For example, a plan could match 50 cents for each dollar a participant contributes.

¹⁶ 29 U.S.C. § 1052(a)(1)(A)(i).

¹⁷ 26 U.S.C. §§ 401(k)(2)(D), 410(a); 29 U.S.C. § 1052(a)(1)(A)(ii). 401(k) plans were created 4 years after ERISA was enacted.

¹⁸ The plan year is the calendar year, or an alternative 12-month period, which a retirement plan uses for plan administration. 26 U.S.C. § 410(a)(4).

¹⁹ Generally, all qualified years of service count toward satisfying a vesting period required by a plan, including when a worker was not yet participating in the plan due to ineligibility. 29 U.S.C § 1053(b)(1). ERISA also requires full vesting upon a participant reaching normal retirement age. For the purposes of this report, we discuss vesting in the context of individuals terminating employment prior to retirement.

²⁰ 26 U.S.C. § 411(a)(2)(B). Prior to the passage of ERISA, some retirement plans required an employee to work until normal retirement age before they would vest and keep their share of the employer's contribution when they left their job. At that time, plans were typically defined benefit plans, which generally provide a specified monthly lifetime retirement payment with the amount based on a formula that often reflects years of service, age at retirement, and salary. There has been a shift over time and, according to Department of Labor data, most private sector workplace plans are now defined contribution plans, which establish individual accounts for employees and provide benefits based on employer and participant contributions to individual accounts and investment returns on those contributions.

²¹ 29 U.S.C. § 1002(19).

²² Rev. Rul. 84-156, 1984-2 C.B. 97.

²³ 29 U.S.C. § 1053(a)(1). In some safe harbor 401(k) plans, non-discretionary employer contributions must also be immediately vested.

²⁴ 26 U.S.C. § 404(a)(6). Certain employer contributions to SIMPLE 401(k) plans must be made sooner.

²⁵ Information about the plan's policies can also be conveyed in other reports such as in annual reports and the summary of material modifications. Plans may provide participants with a summary of material modifications to inform them of plan policy changes or other changes to information included in the SPD instead of providing an updated SPD. 29 U.S.C. §§ 1021-1025.

²⁶ 26 C.F.R. § 1.410(a)-3(e)(2).

²⁷ 26 C.F.R. §§ 1.404(a)-3(a), 1.415(c)-1(b)(6); 26 U.S.C. § 404(a)(6).

²⁸ For the most recent version of this document, see Department of the Treasury, General Explanations of the Administration's Fiscal Year 2017 Revenue Proposals (Feb. 2016), accessible at http://www.treasury.gov/resource-center/tax-policy/Pages/general_explanation.aspx.

²⁹ DOL may intervene in any matters that materially affect the rights of participants. See Department of Labor, Employee Benefits Security Administration, History of EBSA and ERISA, accessible at www.dol.gov/ebsa/aboutebsa/history.html. DOL regulates the requirements for reporting and disclosure of financial information established by ERISA.

³⁰ 29 C.F.R. § 2510.3-102. DOL prescribes no standard for the timing of employer contributions, which is regulated by IRS.

³¹ 29 C.F.R. pt. 2520.

³² 29 U.S.C. § 1029(c).The statements or documents covered by this authority exclude the bargaining agreement, trust agreement, contract, or other instrument under which the plan is established or operated.

³³ 29 U.S.C. § 1135.

³⁴ BLS, Economic News Release, "Table 5: Median years of tenure with current employer for employed wage and salary workers by industry, selected years, 2004-14".

³⁵ Plan sponsors and other plan professionals who assist plan sponsors, such as human resource professionals, record keepers, plan administrators, and benefits consultants, participated in the survey. The highest number of survey respondents identified themselves as plan sponsors. Therefore, in this section we refer to the survey as a plan sponsor survey. See Appendix I for additional details on the survey. There are no current, comprehensive federal data that describe the prevalence of eligibility and vesting policies nationwide, and officials from the Department of Labor and the Internal Revenue Service said they do not have any current comprehensive data regarding the prevalence of eligibility and vesting policies. Prior to the enactment of the Taxpayer Relief Act of 1997, plans were required to submit summary plan descriptions -- documents that describe plan policies including those pertaining to eligibility and vesting -- to the Department of Labor. However, the Act eliminated this requirement. Pub. L. No. 105-34, § 1503, 111 Stat. 788, 1061-63.

³⁶ In this section and the sections that follow, we highlight the percentage or number of plans that have a particular type of eligibility or vesting policy. When we report that a certain percentage of plans have a particular policy that means that the remaining plans do not have the policy. Therefore, since 43 plans in our survey have minimum-age policies that require employees to reach a minimum age before they can participate in their plan, 37 plans do not require employees to reach a minimum age before they are eligible to participate.

³⁷ Of plans that reported having a minimum age policy (376), the highest number (207) used a minimum age of 21. Plan Sponsor Council of America, 57th Annual Survey: PSCA's Annual Survey of Profit Sharing and 401k Plans Reflecting 2013 Plan Experience (Chicago, IL: 2014). PSCA's survey, which is nongeneralizable, covers 613 plans, including both profit sharing and 401(k) plans. Because just 2 percent of plans surveyed are profit sharing plans, we determined the data are sufficiently representative of the 401(k) plan experience. PSCA's dataset includes a greater proportion of large plans based on participant numbers and assets and disproportionately represents the financial, insurance, and real estate industries when compared to the total population of plans, as measured by 2013 Form 5500 filings.

³⁸ U.S. Census Bureau, Survey of Income and Program Participation, 2008 Panel. These are the most recent SIPP data available that address workers' views on why they are not able to participate in workplace retirement plans. A 95 percent confidence interval of the population estimate is 1.6-2.5 percent.

³⁹ Because the SIPP data reflect workers' own assessment of why they are not participating in their workplace plans and workers can misunderstand plan policies, the data may therefore reflect some degree of inaccuracy. While our analysis was limited to workers responding that their employers offer a "401(k) type plan", that universe includes both public and private sector workers.

⁴⁰ See Vanguard, How America Saves 2015: A Report on Vanguard 2014 Defined Contribution Plan Data (Valley Forge, Pa., June 2015). This report is based on nongeneralizable data from the 1900 qualified plans for which Vanguard serves as record keeper. Survey data from the Plan Sponsor Council of America's 57th Annual Survey show that 36 percent of plans have minimum-service eligibility policies.

⁴¹ A 95 percent confidence interval of the population estimate is 15.2-17.8 percent. Few survey respondents reported that both minimum age and minimum service eligibility policies prevented them from participating in their employer's plan. Specifically, eight workers -- about 0.2 percent of SIPP respondents -- said they do not participate in their employer's plan because they have not satisfied a minimum-age policy and because they have not worked long enough. GAO analyzed data from SIPP's 2008 panel, the most recent data available on this topic. U.S. Census Bureau, Survey of Income and Program Participation, 2008 Panel.

⁴² Vanguard, How America Saves 2015. While it notes that small plans are more likely to use these policies, Vanguard's report does not define "small" or "large" and does not provide the number or percentage of small or large plans with minimum-service eligibility policies.

⁴³ Plan Sponsor Council of America, 57th Annual Survey. For example, among plans that have between 1 and 49 participants (the first category for plan size), about 57 percent of plans have minimum-service eligibility policies, while about 46 percent of plans that have between 50 and 199 participants (the second category for plan size) have these policies. This trend continues for all of the remaining plan size categories.

⁴⁴ Due to the small number of survey responses, the plans included in our survey may not be representative of all plans.

⁴⁵ Plan Sponsor Council of America, 57th Annual Survey. As previously noted, this survey covered 613 plans, while our survey covered 80 plans. Thirty-percent is roughly 184 plans.

⁴⁶ This would amount to about 4.8 million workers. A 95 percent confidence interval of the population estimate is 22.7-25.5 percent. U.S. Census Bureau, Survey of Income and Program Participation, 2008 Panel.

⁴⁷ In our survey, we did not differentiate between employer matching and non-matching contributions. Therefore, this figure includes matching and non-matching contributions.

⁴⁸ Vanguard, How America Saves 2015. Vanguard's data show that 53 percent of its plans have minimum-service eligibility policies for employees to receive employer matching contributions. PSCA's data show that 54 percent of plans have these policies. Plan Sponsor Council of America, 57th Annual Survey.

⁴⁹ As previously noted, a plan year is the calendar year, or an alternative 12-month period, that a retirement plan uses for plan administration.

⁵⁰ Workers make contributions through paycheck deductions, which occur throughout the year.

⁵¹ PSCA's percentage may be lower because PSCA's survey included a much larger number of plans (613) than our survey. Furthermore, our survey included a larger percentage of small plans (those with less than 100 participants) and, as noted by a retirement professional we interviewed, small plans may have cash flow concerns that lead them to make contributions at the end of the year to avoid contributing to the accounts of employees who leave during the year.

⁵² According to IRS officials, a corporation that operates on the basis of a calendar year normally has until March 15 to file an income tax return, but many get a 6 month extension to September 15.

⁵³ If employees are "vested", they can keep all or some of their employer's contributions and investment returns on that money when they leave their job. One plan did not provide a response to this question, therefore 79 of 80 plans surveyed provided a response. As previously noted, our survey did not differentiate between employer matching and non-matching contributions.

⁵⁴ Thirty of 80 plans surveyed used a 6-year graduated schedule.

⁵⁵ Vanguard, How America Saves 2015. PSCA's survey also found that more than half of surveyed plans (63 percent) have vesting policies for matching contributions. Plan Sponsor Council of America, 57th Annual Survey.

⁵⁶ Specifically, we asked respondents whether they have 1) added new vesting requirements, 2) decreased the total amount of time to become 100 percent vested, 3) eliminated vesting requirements, 4) increased the total amount of time to become 100 percent vested, 5) switched from cliff to graduated vesting, or 6) switched from graduated to cliff vesting. Between 72 and 75 respondents said they had not made a change in any one of these 6 categories.

⁵⁷ We asked survey respondents to provide reasons for using these policies for the largest plan they serve. We did not ask separately for the reasons for using minimum-age and minimum-service eligibility policies.

⁵⁸ IRS officials said that minimum-age policies are unnecessary because a minimum-service eligibility policy alone can prevent the enrollment of short-tenure workers into a plan. A minimum-service policy affects just those workers who are short-tenure, regardless of age.

⁵⁹ According to CPS data for January 2014, the median number of years with one's current employer among employed wage and salary workers was 0.8 years for 18 to 19-year-olds, 1.3 years for 20 to 24-year-olds, and 5.5 years for workers 25 and older. Tenure varies among workers age 21 and under. The average number of jobs started per year was 1.5 jobs for an 18-year-old, but nearly half that (0.8) for a 19-year-old, 0.7 for a 20-year-old and 0.6 for a 21-year-old. See U.S. Bureau of Labor Statistics, National Longitudinal Survey of Youth 1979 (NLSY79).

⁶⁰ GAO, 401(K) Plans: Increased Educational Outreach and Broader Oversight May Help Reduce Plan Fees, GAO-12-325 (Washington, D.C.: April 24, 2012). Here we use the term administrative fee to include record-keeping fees.

⁶¹ Absent instruction from the participant, a plan can opt to cash out a participant account with a balance of $1,000 or less and can transfer a vested balance up to $5,000, excluding rollovers, to a special individual retirement account . See GAO, Greater Protections Needed for Forced Transfers and Inactive Accounts, GAO-15-73 (Washington, D.C., Nov. 21, 2014). The "Retirement Savings Lost and Found Act of 2016" proposed in June 2016 would increase that maximum account balance to $6,000.

⁶² Of the 80 plans surveyed, direct cost was cited as an important reason for using minimum-service for employer contribution policies by 24 plans and for using vesting policies by 28 plans.

⁶³ Our analysis suggests that without paying a dollar more in employer contributions, a hypothetical employer currently making a 3.0 percent match to 1,000 employees ages 21 to 66 and earning, in 2016 dollars, an average salary of $65,387 may only have to reduce that match to 2.96 percent to include 50 (5 percent) more employees ages 18 to 20 who earn an average salary of $16,172. For the older workers, that amounts to an average annual employer contribution of $1,938, which is $24 less than they received before the adjustment, but younger workers could receive an annual average employer contribution of $479, instead of $0. Dollar values are rounded. See Appendix II for information about salary data used for hypothetical scenarios. Our analysis of Current Population Survey (CPS) data found that workers ages 16 to 20 are 5.2 percent of the overall workforce and just 2.3 percent of the full-time workforce, as of March 2016. CPS data are averages across all industries. However, some firms employ a higher percentage of workers ages 16 to 20, which would increase the cost of extending an employer match.

⁶⁴ Plans must enroll eligible workers no later than 6 months after becoming eligible or the first day of the new plan year, whichever is sooner. 26 U.S.C. § 410(a)(4).

⁶⁵ Plan Sponsor Council of America, 57th Annual Survey.

⁶⁶ For example, because a plan can require a 3-year cliff policy, there is no net cost in direct benefits for contributions made but later forfeited back to the employer. Returns on those contributions are also forfeited to the plan when vesting is unmet.

⁶⁷ Employers receive a tax deduction for their contributions to participant accounts, though Treasury officials said the deduction is the same as received for paying wages of the same amount. For tax year 2012, the most recent year for which data are available, IRS estimated that corporations claimed $145 billion in aggregate deductions for contributions to workplace retirement plans. This estimate is based on deductions claimed by companies who filed Form 1120, "U.S. Corporation Income Tax Return." See Department of the Treasury, Internal Revenue Service, 2012 Corporation Income Tax Returns Line Item Estimates (Washington, D.C). Because a tax deduction reduces the amount of an employer's income subject to income tax unlike a tax credit which directly offsets the tax amount owed, the federal tax revenue lost each year as a result of employer contributions does not equal the amount of deductions.

⁶⁸ A significant number of job separations are not voluntary. According to Bureau of Labor Statistics' data from the Job Openings and Labor Turnover Survey 2001-2015, on average, 40 percent of total annual job separations resulted from layoffs and discharges. Another 53 percent of job separations were due to workers voluntarily leaving their job and 7 percent was due to retirements and other reasons, on average. While the number of quits and layoffs may vary over the business cycle, over the last 15 years only in 2009 did the number of layoffs and discharges exceed voluntary separations.

⁶⁹ According to a statement from a House Committee on Ways and Means report issued in advance of ERISA: ". . . the committee bill sets minimum standards on the age and service requirements which can be used to exclude employees from participation in plans. . . . The Committee believes that these rules are reasonable. They provide a balance between the need to grant employees the right to participate in pension plans at a relatively early age so that they can begin to acquire pension rights and the need to avoid the administrative drawbacks that would be involved in granting coverage to immature and transient employees whose benefits would in any event be small." H.R. Rep. No. 93-807, at 15 (1974).

⁷⁰ See United States Department of Labor, Employee Benefits Security Administration, "Savings Fitness: A Guide to Your Money and Your Financial Future" (Washington, D.C.), accessible at http://www.dol.gov/ebsa/publications/savingsfitness.html, and "New Employee Savings Tips-Time Is On Your Side" (Washington, D.C.), accessible at http://www.dol.gov/ebsa/publications/main.html.

⁷¹ Compound interest refers to the exponential growth realized when returns are earned on returns.

⁷² Projected values "at retirement" are nominal, that is, they are not adjusted for inflation. Inflation is a general increase in the overall price level of the goods and services in the economy. The projected values provided in "2016 dollars" adjust the nominal value to show the current value of the projection, given anticipated increases in the price of goods and services. In other words, the 2016 value shows what the projected future value might be worth today.

⁷³ See Appendix II for a detailed description of our hypothetical projections and the assumptions we used. The Social Security Administration's (SSA) Office of the Chief Actuary (OCACT) publishes age-based earnings-adjustment factors that can be applied to SSA's average wage index (AWI) to produce an average pattern of wage progression over a worker's career, referred to as scaled factors. For our illustrations, we used the 2015 factors. SSA, Office of the Chief Actuary, SSA Bulletin: Actuarial Note, "Scaled Factors for Hypothetical Earnings Examples Under the 2015 Trustees Report Assumptions," No. 2015.3, July 2015. The amount of forgone savings is $23,258 in 2016 dollars -- more than the worker's entire starting salary in this scenario -- and amounts to 5 percent of their total potential retirement savings from their own contributions over their career. See Appendix II, table 3 for a comparison of projected savings from 3 years of employee contributions alone, occurring at different times during a worker's life.

⁷⁴ The data include those from ages 16 to 24 in the count of recent high school graduates. BLS, "College Enrollment and Work Activity of 2015 High School Graduates" April 28, 2016. According to BLS, information on school enrollment and work activity is collected monthly in the Current Population Survey (CPS), a nationwide survey of about 60,000 households that provides information on employment and unemployment. Each October, a supplement to the CPS gathers more detailed information about school enrollment, such as full- and part-time enrollment status.

⁷⁵ Ibid. As a percentage of those employed or looking for work, that amounts to 73 percent who are not enrolled in college and 36 percent who are enrolled.

⁷⁶ GAO, Retirement Security: Women Still Face Challenges, GAO-12-699 (Washington, D.C.: March 20, 2014).

⁷⁷ See Appendix II for a description of how we calculated flatter earnings growth for this comparison and table 2 for the projected value of savings not made from ages 18 to 20 for a low-wage earner with less wage growth than the SSA-scaled medium earner, from the SSA Bulletin, on which we based our other illustrations.

⁷⁸ Lost savings for a medium-level earner with constant wage growth of 5.6 percent (the average wage growth from salary data from the SSA Bulletin) would be $85,857, which is 5 percent of $1.7 million in total savings, not adjusted for inflation or $23,258 out of $464,056 in 2016 dollars. Lost savings for a low-level earner with less wage growth (4.6 percent, one point less than SSA's assumption for an average worker) would be $38,411, which is 11.5 percent of $334,783 in total savings at retirement (or $10,405 out of $90,689 in 2016 dollars). These values illustrate the potential effect of an eligibility policy on a worker's retirement savings over time and are not a prediction of a low-wage worker's likely retirement savings.

⁷⁹ In our hypothetical scenario, earnings for an 18-year-old are for 2016 so the inflation adjusted value is the same as the nominal value. PSCA survey data show that the most common employer contribution level among 401(k) plans is 2.9 percent, rounded in our standard projections to 3 percent. Similarly, the most commonly used match formula is 50 percent of employee deferrals up to 6 percent of salary, which is a maximum match of 3 percent, used by 26.2 percent of plans. See Appendix II for a detailed description of assumptions used in our hypothetical scenarios.

⁸⁰ 26 USC § 410(a)(1)(A)(i).

⁸¹ As noted, CPS data from March 2016 reflect that 5.2 percent of the total labor force (full-time and part-time) is ages 16 to 20, or 8.2 million workers. Just 2.3 percent of the full time workforce is younger than age 21.

⁸² According to a statement from a House Committee on Ways and Means report issued in advance of ERISA: ". . . it is desirable to have as many employees as possible covered by private pension plans and to begin such coverage as early as possible, since an employee's ultimate pension benefits usually depend to a considerable extent on the number of his years of participation in the plan." H.R. Rep. No. 93-807, at 43-44 (1974). ERISA set the minimum age cap at 25. The Retirement Equity Act of 1984 reduced the cap on minimum age to 21. According to a statement in the Congressional Record by an original cosponsor of the bill that later became the Retirement Equality Act of 1984, changing the cap on the minimum age to 21 better allowed women to save for retirement. Specifically, the statement in the Congressional Record states: "The permitted exclusion of younger women is particularly harmful to younger workers because female labor force participation is highest -- over 70 percent -- between the ages of 20 to 24. This change alone could result in the overall plan participation rate by women to increase up to 16 percent."

⁸³ The Pension Protection Act of 2006, Pub. L. No. 109-280, § 902, 120 Stat. 780, 1033-35, includes a provision that allows plans that automatically enroll eligible workers in the plan to be treated as meeting nondiscrimination testing requirements provided certain conditions, such as ensuring that workers' contribution amounts meet certain minimum levels, are met. A committee report by the House Committee on Education and the Workforce that preceded the enactment of PPA stated that "there is a growing consensus that the 401(k) plan rules need to be updated to encourage automatic enrollment (to get more workers in plans)." H.R. Rep. No. 109-232, pt. 1, at 280 (2005).

⁸⁴ DOL guidance highlights the importance of saving early for retirement. See United States Department of Labor, Employee Benefits Security Administration, "Savings Fitness: A Guide to Your Money and Your Financial Future" (Washington, D.C.), accessible at http://www.dol.gov/ebsa/publications/savingsfitness.html, and "New Employee Savings Tips -- Time Is On Your Side" (Washington, D.C.), accessible at http://www.dol.gov/ebsa/publications/main.html.

⁸⁵ The fiscal year 2017 Budget Justification of the Employee Benefits Security Administration, an agency within the Department of Labor, states a goal of expanding retirement plan coverage. See United States Department of Labor, Employee Benefits Security Administration, FY 2017 Congressional Budget Justification, accessible at http://www.dol.gov/sites/default/files/documents/general/budget/CBJ-2017-V2-01.pdf.

⁸⁶ GAO, Private Pensions: Changes Needed to Provide 401(k) Plan Participants and the Department of Labor Better Information on Fees, GAO-07-21 (Washington, D.C.: Nov.16, 2006) and GAO, 401(k) Plans: Greater Protections Needed for Forced Transfers and Inactive Accounts, GAO-15-73, (Washington, D.C.: Nov. 21, 2014).

⁸⁷ A worker could be ineligible for a reason other than a minimum-age policy, such as not working enough hours during the year.

⁸⁸ GAO, Retirement Security: Federal Action Could Help State Efforts to Expand Private Sector Coverage, GAO-15-556 (Washington, D.C., Sept.10, 2015).

⁸⁹ According to the 2015 Employee Benefit Research Institute's (EBRI) Retirement Confidence Survey, among workers 55 and older, 36 percent say they plan to retire at 66 or older, while 14 percent of current retirees report having done so. Also see GAO, Private Pensions: Low Defined Contribution Plan Savings May Pose Challenges to Retirement Security, Especially for Many Low-Income Workers, GAO-08-8 (Washington, D.C., Nov.29, 2007).

⁹⁰ Society of Actuaries. "Understanding and Managing the Risks of Retirement: 2013 Risks and Process of Retirement Survey Report", 2013, Online. Premature retirement also prevents workers from taking advantage of the additional "catch-up contributions" intended to help workers ensure adequate retirement savings. 26 U.S.C. § 414(v). In 2016, the maximum catch-up contribution permitted in a regular 401(k) plan was $6,000 above the general limit of $18,000 or the ADP test limit of section 401(k)(3) or the plan limit, if any. Catch-up provisions were created, in part, to help women save for retirement. See GAO, Private Pensions: Some Key Features Lead to an Uneven Distribution of Benefits, GAO-11-333 (Washington, D.C., March 30, 2011) and GAO, Private Pensions: Pension Tax Incentives Update, GAO-14-334R (Washington, D.C., March 20, 2014).

⁹¹ In our illustrative projections, total retirement savings from employee contributions alone would be $1.7 million ($464,056 in 2016 dollars). See Appendix II for a description of the assumptions used in our hypothetical projections.

⁹² Bureau of Labor Statistics' data from The National Longitudinal Survey of Youth 1979 (NLSY79). According to BLS, the survey's cohort consists of men and women born in the years 1957-64 who were age 14 to 22 when first interviewed in 1979. These individuals were ages 47 to 56 in 2012-13. Data are from the NLSY report "Number of Jobs Held, Labor Market Activity, and Earnings Growth Among the Youngest Baby Boomers: Results from a Longitudinal Survey."

⁹³ Lost savings of $411,439 amounts to 24 percent of our hypothetical worker's total retirement savings, based on nominal total employee retirement savings of $1,713,090 at age 67 ($464,056 in 2016 dollars) derived from employee contributions alone.

⁹⁴ The survey's categories for hours usually worked are 0-20, 21-34, 35-39, 40, 41-49, and 50 or more hours per week; 20 hours per week is the closest approximation to the average of 19 hours per week, which corresponds to 1,000 hours worked per year. BLS's Current Population Survey (CPS) data for March 2016.

⁹⁵ GAO analysis of BLS's CPS data issued in March 2016 found that 9.2 million women worked 20 or fewer hours per week while 5.2 million men did so. Prior GAO work has shown that women have less retirement income than men, on average. See GAO, Retirement Security: Women Still Face Challenges, GAO-12-699 (Washington, D.C., July 19, 2012).

⁹⁶ In 2015, nearly 5 percent of the workforce worked multiple jobs and women were somewhat more likely than men to do so (5.3 percent of women in the workforce worked multiple jobs versus 4.6 percent of men). BLS data from the Current Population Survey, 2015.

⁹⁷ According to the Congressional Record dated Feb. 26, 1974, the 1-year service cap was aimed to address concerns about pension plan coverage stemming, in part, from plans using overly restrictive service requirements to exclude workers from participating in plans.

⁹⁸ Under federal law, plans may credit employee service on the basis of actual hours worked or elapsed time. Using the actual number of hours an employee has worked, a plan may require an employee to work up to 1,000 hours during a 12-month period to be credited with a year of service. Employees are not eligible to save in the plan and receive employer contributions, if they are offered by the plan, until they have worked the requisite number of hours set by the plan. However, under the elapsed time approach, a plan credits service based on the total period of time that has elapsed since the employee began working, irrespective of the actual number of hours the employee has worked. 26 U.S.C. § 410(a)(3)(A); 26 C.F.R. § 1.410(a)-7(a)(1)(i)-(ii); and 29 U.S.C. § 1052(a)(3)(A).

⁹⁹ Department of the Treasury, General Explanations of the Administration's Fiscal Year 2017 Revenue Proposals (Washington, D.C.: February. 2016).

¹⁰⁰ Women's Pension Protection Act of 2015, H.R. 4235, 114th Cong. § 102 and Women's Pension Protection Act of 2015, S. 2110, 114th Cong. § 102.

¹⁰¹ Retirement Plan Simplification and Enhancement Act of 2013, H.R. 2117, 113th Cong. § 103.

¹⁰² Plan Sponsor Council of America's 57th Annual Survey of 401(k) and profit sharing plans covers the 2013 plan year. These were the most recent PSCA survey data available when we developed our report.

¹⁰³ See Appendix II for details of our projection assumptions and Appendix II Table 6 for examples of potential lost savings from a "last day" policy applied at different ages.

¹⁰⁴ A participant would lose whatever portion of the annual contribution they would otherwise have earned by the point in the year that they leave their job. See Appendix II for details about the assumptions used. Many older workers also leave their job due to health issues. For more on this phenomenon and the effects on retirement savings see GAO, Retirement Security: Most Households Approaching Retirement Have Low Savings, GAO-15-419, (Washington, D.C., May 12, 2015).

¹⁰⁵ Those lost savings amount to about 7 percent of total retirement savings from employer contributions. See Appendix II for the assumptions we used.

¹⁰⁶ 26 U.S.C. §§ 401(a)(4), 410(b).

¹⁰⁷ Investments typically have no guarantee of returns and can lose value.

¹⁰⁸ For example, under our hypothetical scenario, for a 20-year-old, the foregone investment returns over the year resulting from an employer contribution made at the end of the year instead of bi-weekly could be just $22 for that year and $599 at the worker's retirement ($162 in 2016 dollars). See Appendix II Table 7 for the projected value of retirement savings lost from delayed employer contributions over multiple years of employment occurring at different ages.

¹⁰⁹ See Appendix II for details of our projection assumptions and Appendix II Table 7 for examples of potential lost savings from delayed employer contributions over multiple years of employment occurring at different ages.

¹¹⁰ 26 U.S.C. § 404(a)(6).

¹¹¹ See United States Department of Labor, Employee Benefits Security Administration, "Savings Fitness: A Guide to Your Money and Your Financial Future" (Washington, D.C.), accessible at http://www.dol.gov/ebsa/publications/savingsfitness.html, and "New Employee Savings Tips -- Time Is On Your Side" (Washington, D.C.), accessible at http://www.dol.gov/ebsa/publications/main.html.

¹¹² Vesting policies differ from eligibility and other polices in the way in which they may reduce potential retirement savings because participants lose the employer contribution after it has been deposited and invested in participants' accounts. Consequently, the employer contributions that participants see on their account statements can be taken out of their account if they leave before they are vested. Initial savings forfeited is the amount of employer contributions made up to that time, and investment returns on those contributions, less any portion that is considered vested.

¹¹³ See Appendix II for a full description of the assumptions used in this hypothetical scenario.

¹¹⁴ The Economic Growth and Tax Relief Reconciliation Act of 2001 (EGTRRA) reduced the cap on vesting policies in defined contribution plans for employer matching contributions to 3 years for cliff vesting and 6 years for graduated vesting policies. More recently, the Pension Protection Act of 2006 ensured that those caps would apply to employer non-elective contributions made after 2006.

¹¹⁵ DOL officials told us that the reorganization of duties under ERISA that occurred in 1980 made Treasury responsible for the regulation of vesting policies.

¹¹⁶ Employer contributions that are forfeited when a participate leaves their job before satisfying the vesting requirement can offset plan expenses and reduce costs for the remaining participants.

¹¹⁷ The Pension Protection Act of 2006 revised rules affecting portability in some workplace plans. In addition, there have also been some recent policy proposals in this area. As previously noted, the current administration's FY 2017 Budget Proposal includes a package of proposals aimed at increasing access to retirement plans and increasing the portability of retirement savings and benefits. The proposals are intended to ensure near-universal access to workplace retirement savings accounts and test new approaches to making retirement benefits more portable across jobs. See Office of Management and Budget, Fiscal Year 2017 Budget Overview, accessible at https://www.whitehouse.gov/omb/overview.

¹¹⁸ Four plan sponsors agreed to let us survey their plan participants about their own eligibility to join the plan and to receive employer contributions, their vesting status, and other related plan policies. We examined the accuracy of 46 respondents' answers by identifying their actual plan policies through our analysis of summary plan descriptions (SPD), summaries of material modification, and exchanges with the plan administrators. While participating employers were asked to send the survey to both employee and participant groups, of the 46 survey responses we analyzed, two were from individuals without an active 401(k) plan account with their current employer. Therefore, we refer to respondents generally as plan participants.

¹¹⁹ Overconfidence among individuals assessing their own financial literacy is demonstrated in a study: Annamaria Lusardi and Olivia Mitchell, "Financial literacy and retirement planning in the United States" PEF, 10(4):509-525, October 2011. Cambridge University Press.

¹²⁰ For example, the six of eight respondents from a law firm we surveyed who provided correct answers regarding their immediate eligibility all have some college education and four have a doctor of philosophy or professional degree, such as a Juris Doctor (law degree). Of the 29 respondents who answered correctly, about 21 had tenure of 4 or more years. One retirement plan professional we interviewed also observed that longer tenure may be correlated with better knowledge of plan policies.

¹²¹ Several participants said they had been immediately eligible to join the plan when they started their job although their plan required an initial minimum-service period for eligibility.

¹²² For example, several participants said that they were not able to join the plan immediately after hiring because the plan did not yet exist; however, we found that the plan did exist and their ineligibility was instead due to the plan's minimum-service eligibility policy.

¹²³ Boston Research Technologies and the National Association of Retirement Plan Participants, "Participant Survey: Study of Financial Literacy, Financial Behavior and Trust in Financial Institutions," 2015. These two groups jointly conduct the annual Financial Empowerment Literacy and Trust (FELT) survey. We estimate that the sampling error is well within an acceptable range due to the large number of survey responses (about 5,000), and thus conclude that these survey responses are sufficiently reliable for our purposes.

¹²⁴ Aon Hewitt, "Workforce Mindset Study: Key Findings on What Differentiates, What Rewards, and What Communicates," 2015. The survey was administered online within the United States in August 2014 to employees of a diverse range of companies and across different geographical regions. Of 2,539 employees surveyed, 86 percent agreed that they had a good understanding of their employer's company match and/or basic contributions to their 401(k)-type plan.

¹²⁵ Some companies require plan participants to be employed on the last day of the year in order to receive the employer's matching contributions for that year. Vesting requirements may still apply.

¹²⁶ 2015 Financial Empowerment Literacy and Trust survey, conducted by Boston Research Technologies. In the survey, those who said they had a working knowledge of the term said that they could not teach it to others, while those who said that they understood the term well said that they could explain it to others.

¹²⁷ For example, one participant who started his job in 2015 said that he would be able to keep all of his employer contributions if he were to leave his job even though he would not satisfy the company's 6-year graduated vesting policy until 2021. As described in Appendix I, the participant survey was administered over several weeks from December 2015 through January 2016.

¹²⁸ The summary plan descriptions (SPD) were provided by the four companies whose participants we surveyed regarding participant understanding, and by a third-party plan administrator who we interviewed.

¹²⁹ We convened the discussion groups at a regional conference of defined contribution plan sponsors and plan advisors in March 2015, issuing an open invitation to members of these groups to share their views and experiences regarding participant understanding of eligibility and vesting policies as well as what factors determine how plans use these policies.

¹³⁰ ERISA requires plans to present the advantages and disadvantages of the plan without either exaggerating the benefits or minimizing the limitations. 29 C.F.R. § 2520.102-2(b).

¹³¹ See Appendix I for a description of our plan sponsor survey.

This report examines: (1) what is known about the prevalence of 401(k) plans' eligibility and vesting policies and why plans use them, (2) the potential effects of eligibility and vesting policies on workers' retirement savings, and (3) participants' understanding of these policies. We did not examine the use of these policies by defined benefit plans or by public retirement plans.

We reviewed survey data, industry data, and conducted interviews to determine what is known about the prevalence of eligibility and vesting policies and why plans use them. To identify what is known about the prevalence of eligibility and vesting policies, we conducted a non-generalizable survey of plan sponsors and plan professionals with plans ranging in size from less than 100 participants to more than 5,000.¹ We received 80 responses to this survey. We also reviewed nongeneralizable data from Plan Sponsor Council of America's (PSCA) annual survey of defined contribution plans, which covered 613 plans with 8 million participants and $832 billion in plan assets.² PSCA's dataset represents plans of varying sizes and industries, but includes a greater proportion of large plans based on participant numbers and assets and disproportionately represents the financial, insurance, and real estate industries when compared to the total population of plans, as measured by 2013 Form 5500 filings. In addition, we reviewed non-generalizable data from Vanguard's 2015 report on the 1,900 qualified plans for which it serves as record keeper.³ These two were the only datasets we identified with detailed information on eligibility and vesting policies. We evaluated another large dataset and determined that it was not reliable enough for our purposes. We also analyzed generalizable Census Bureau survey data to identify which eligibility policies affect workers, as reported by the workers themselves.⁴

To assess the reliability of these data, we interviewed staff from Vanguard and the Census Bureau who are responsible for producing the data. We also reviewed documentation related to these data sources, such as descriptions of their methodology. We also performed various tests on the data. For the PSCA data, we compared selected data from PSCA's 2013 survey -- the survey we used -- to data from PSCA's 2012 survey to assess the consistency of the data. Specifically, we reviewed basic survey respondent demographic data (respondents by plan type, amount of plan assets, and industry) and key data on eligibility and vesting (the percentage of plans with service restrictions for plan eligibility, the percentage of plans with age restrictions for plan eligibility, and the percentage of plans that have vesting policies that require a service period for the vesting of matching employer contributions). We also evaluated the makeup of the plans included in PSCA's survey by comparing selected demographic data from PSCA's survey to Department of Labor Form 5500 plan demographic data.⁵ For the Vanguard data, we asked a Vanguard official who is responsible for the data questions about the scope of the data, how the data are collected and maintained, and whether they have any limitations. We also compared the eligibility and vesting data to the data from PSCA's survey and our own survey of plan sponsors. For the Census Bureau data, we performed electronic testing for missing data, outliers, and obvious errors. Based on these steps, we determined that the PSCA, Vanguard, and Census Bureau data were sufficiently reliable for our purposes. We also interviewed government officials from the Securities and Exchange Commission, the Internal Revenue Service, and the Department of Labor, as well as retirement professionals, to determine whether they were aware of any data sources regarding the prevalence of eligibility and vesting policies and to obtain their perspectives on the policies' prevalence. See "Selection and Categorization of Interviewees" later in this section for more information on our interviews.

To determine why plans use eligibility and vesting policies, we surveyed plan sponsors and plan professionals by including a link to our survey in four industry publications (80 responded). The questions asked respondents to identify and rank the importance of multiple factors in their decision to use specific eligibility and vesting policies. To supplement our survey, we interviewed government officials, retirement professionals, and academic researchers to obtain their perspectives on why plans use eligibility and vesting policies. For additional context and perspectives, we also held two structured group interviews at a regional defined contribution plan conference with an open invitation to plan sponsors and plan service providers to discuss the reasons why plans use these policies.

To examine the policies' potential effects on retirement savings over time, we developed hypothetical scenarios to illustrate what the effect could be for an individual based on a number of assumptions. For the hypothetical projections, we made a number of assumptions regarding salary levels, employee deferrals, employer matches, and investment returns drawn from federal and industry data sources, including Bureau of Labor Statistics data, PSCA's annual defined contribution plan sponsor survey on plan policies used, and the Social Security Trustees' report (2015 Trustees Report, long range projections, intermediate assumptions). (See Appendix II for a detailed explanation of the assumptions used and additional tables providing more projections for comparison as well as inflation-adjusted values.) We also interviewed officials from the Department of Labor, Treasury, the Internal Revenue Service, and the Securities and Exchange Commission, as well as a total of 21 retirement professionals and academic researchers to discuss what is known about these policies' effects on savings over time.

To assess participants' understanding of eligibility and vesting policies, we used data from the National Association of Retirement Plan Participants' (NARPP) Participant Survey: Study of Financial Empowerment Literacy and Trust (FELT survey) to identify participants' understanding of financial terms, and generalizable data from a Defined Contribution Plan 2015 Study of Participant Satisfaction and Loyalty (DCP) to identify the decisions that participants would make about their employment based on their understanding of their companies' 401(k) policies. NARPP's FELT surveys approximately 5,000 current 401(k) and 403(b) participants yearly, and was administered in April 2015. To assist us with our work, NARPP agreed to include the terms "eligibility requirement" and "vesting period" to the list of terms already included in their financial literacy survey. The DCP surveys an internet panel of over 5,000 respondents and was administered during the month of April 2015. NARPP provided the surveys' summary responses to us for our analysis. To assess the reliability of the data from the DCP and FELT surveys, we interviewed an official from Boston Research Technologies who is responsible for managing the two surveys. We also independently conducted manual tests of the DCP survey data and compared selected survey responses to similar data from other samples of defined contribution plan participants. We found the data to be sufficiently reliable for our purposes. We did not perform separate tests on the FELT survey, which uses the same sampling frame and procedures as the DCP survey.

We further assessed participants' understanding of eligibility and vesting policies by administering a questionnaire to individuals testing their knowledge of the eligibility and vesting policies used by their 401(k) plans and comparing their answers to the actual policies in their plans' summary plan descriptions (SPD), summaries of material modification, and exchanges with plan administrators . This allowed us to determine the accuracy of their knowledge rather than rely on their self-reported knowledge. We analyzed the accuracy of the 46 responses by reviewing plan documents provided by the sponsors or their service providers and emailed the sponsors directly to confirm that we understood the plan policies correctly, as necessary. See Survey Data Sources below for details on the survey design and implementation. To further assess the clarity of plans' eligibility and vesting policies, we reviewed the SPDs of five companies, including those we surveyed regarding participant understanding. One SPD was provided by a third-party plan administrator who we interviewed. Additionally, we conducted online searches for relevant literature and asked interviewees to advise us regarding relevant studies and papers. We then assessed the relevance of the studies to our research questions to include those findings, as appropriate. Also, to understand what sources of information on eligibility and vesting policies participants typically have, we held two discussion groups with plan sponsors and plan advisors and asked them to identify the methods plans use to communicate their policies. Results of our review of the SPDs and discussion groups are not generalizable, but provide additional context and perspective. Finally, we interviewed retirement professionals and academic researchers, as discussed below, to get their perspectives on participants' understanding of eligibility and vesting policies and to invite their suggestions on how this information can be more effectively presented to improve participant understanding. For example, we spoke to several financial literacy experts who made observations about financial decision making more generally and what that could tell us about eligibility and vesting policies specifically.

Survey Data Sources

Plan Sponsor and Plan Professional Questionnaire

Design and implementation. We developed a web-based questionnaire for plan sponsors and plan professionals to collect information on eligibility and vesting policies. The questionnaire included questions on the types of eligibility and vesting policies plans use and the reasons they use these policies. In addition, the questionnaire included questions on the timing of employer contributions, including whether plan participants have to be employed on the last day of the year to receive employer contributions and the schedule for making employer contributions. Throughout the questionnaire, our questions focused on employer contributions and did not distinguish between matching and non-matching contributions. To inform our understanding of participants' understanding of eligibility and vesting policies, the questionnaire also included questions regarding their views on the degree to which participants understand these policies and how they are informed of the policies.

To minimize errors arising from differences in how questions might be interpreted and to reduce variability in responses that should be qualitatively the same, we conducted pretests with three individuals external to GAO. The individuals included: an academic researcher, a retirement professional, and a benefits coordinator for a manufacturing company. Further, two GAO staff with expertise in the retirement area and two staff with expertise in survey design reviewed the survey for content and consistency. Based on feedback from these pretests, we revised the questionnaire in order to improve question clarity. For instance, in response to the benefits coordinator's suggestion to clarify the language used in a question focused on whether certain types of employees are excluded from the plan, we modified the question to clarify that we were asking whether employees in certain job classifications are excluded from the plan.

After completing the pretests, we administered the survey. Starting in May 2015, we asked three industry groups to announce -- through their publications -- an invitation for plan sponsors to complete our survey. These groups included a link to our survey in their publications. Those publications were: an email newsletter published by PLANSPONSOR,⁶ Plan Sponsor Council of America's annual survey of plans,⁷ and Pensions & Investments'⁸ Plan Sponsor Digest and Pensions & Investments Daily, which are two publications directed at plan sponsors. We also included a link to the survey in an online forum for American Society of Pension Professionals and Actuaries (ASPPA)⁹ members who are owners or senior managers of plan administration firms. We directly received responses through August 31, 2015. We received 80 completed surveys. We cannot report a response rate as it is possible that respondents submitted multiple surveys or individuals responded who were not plan sponsors. Sponsors and plan professionals could respond anonymously, and some respondents did not provide contact information.

Analysis of responses and data quality. We used standard descriptive statistics to analyze responses to the questionnaire. Because this was not a sample survey, there are no sampling errors. To minimize other types of errors, commonly referred to as non-sampling errors, and to enhance data quality, we employed recognized survey design practices in the development of the questionnaire and in the collection, processing, and analysis of the survey data. For instance, as previously mentioned, we pretested and reviewed the questionnaire with individuals internal and external to GAO to enhance the clarity of our questions, which minimizes the likelihood of errors arising from differences in how questions might be interpreted and helps to reduce the variability in responses that should be qualitatively the same. To help reduce nonresponse, another source of non-sampling error, we asked the industry groups that publicized the survey to re-publicize it to further encourage respondents to complete the survey. In reviewing the survey data, we performed automated checks to identify inappropriate answers. We further reviewed the data for missing or ambiguous responses and followed up with respondents when necessary to clarify their responses. On the basis of our application of recognized survey design practices and follow-up procedures, we determined that the data were of sufficient quality for our purposes.

Plan Participant Questionnaire

Design and implementation. We developed a web-based questionnaire directed at plan participants to collect information on their understanding of their companies' 401(k) eligibility and vesting policies. The questionnaire included questions about the individual's own eligibility and vesting status. In addition, the questionnaire included questions on the timing of employer contributions, including whether plan participants have to be employed on the last day of the year to receive employer contributions. The participants' responses and our analysis of their accuracy are not generalizable.

To minimize errors arising from differences in how questions might be interpreted and to reduce variability in responses that should be qualitatively the same, we conducted pretests with two subject matter experts and conducted pretests with three plan participants from external companies. We also reviewed the survey with internal survey experts. Based on feedback from these pretests, we revised the questionnaire to improve question clarity. After completing the pretests, we administered the survey using an online platform. To identify participants to test, we first invited every plan sponsor who had provided contact information to us when completing a separate plan sponsor survey to distribute a separate questionnaire to their plan participants. Four plans agreed to distribute a hyperlink to our participant questionnaire. We sent a link to the survey to the participating sponsors and asked them to disseminate the link and invite their plan participants to complete the questionnaire. Starting on December 8, 2015, we made the survey available for the participants to complete. While we asked the plans to send periodic reminders, we did not have direct access to the participant groups and we were not included on those communications. We did not select the participants in any way or have input into who responded to the survey. The survey closed on January 27, 2016. We received 50 completed questionnaires and analyzed 46 of them. In the process of comparing the survey responses to the plan descriptions, we manually reviewed each survey response for outliers or errors in skip patterns. We excluded two surveys because the respondents had no 401(k) account but said they were eligible so they were not asked any further substantive questions. Two others were excluded because they began employment on a date before which we had a record of plan policies, so we could not analyze the accuracy of their responses. Some respondents did not click the "completed" button at the end of the survey and were not included in our analysis. We cannot report a response rate as it is possible that respondents submitted multiple surveys. In addition, most respondents did not provide contact information. Evaluating the accuracy of participant responses was the purpose of this survey, so our findings in that regard are an assessment of data quality.

Analysis of participant responses and data quality. To describe what is known about participants' understanding of their companies' 401(k) plan eligibility and vesting requirements, we reviewed participant responses at the four companies we surveyed. We compared the employee responses regarding the eligibility and vesting requirements that applied to them when they began working to information in the respective summary plan description (SPD), and obtained clarification from plan officials, as needed. Plans may issue new plan documents and change policies over time, so to test individuals' knowledge of the eligibility policies that affected them when they were hired, we used the policies in place at that time rather than the eligibility policies in place most recently, when they differed. For example, to determine the employee's eligibility to join the plan, we identified the employee's start date and determined if there were any factors that would have prevented them from immediately enrolling in the plan.¹⁰ In cases where we did not have the plan document for the year the participant was hired, we reviewed the plan document published before and after their hire date and reviewed summaries of material modifications. If the information about the plan policies in the two plan documents from before and after their date of hire were the same, we used it to determine the accuracy of the participant's answer. However, when we could not determine the plan policy for the relevant time period, we excluded the participant response from our analysis of that question. For policy questions regarding current plan policies (like frequency of employer contributions, and if the participant is required to work on a particular day to receive employer contributions) we used the current plan policy to determine the accuracy of responses. We reviewed the plan details to identify information about the vesting schedule. Three of the four companies we surveyed currently offer automatic enrollment to eligible workers. Automatic enrollment is a plan feature by which eligible workers are enrolled in the plan by default and can opt-out if they do not wish to participate in the plan.

Survey of Income and Program Participation (SIPP)

Administered by the Census Bureau, the SIPP is a household-based survey designed as a continuous series of national panels. The Census Bureau uses a two-stage stratified design to produce a nationally representative panel of respondents who are interviewed over a period of approximately 3 to 4 years. Within a SIPP panel, the entire sample is interviewed at various intervals called waves (from 1983 through 2013, generally 4-month intervals). In addition to income and public program participation, the SIPP includes data on other factors of economic well-being, demographics, and household characteristics. We used data from the most recent relevant data set, the 2008 SIPP.

Plan Sponsor Council of America's (PSCA) 57th Annual Survey of Plans

PSCA's survey, which includes data for plan year 2013, covers a total of 613 profit sharing, 401(k), and combination 401(k)/profit sharing plans. Only 2 percent of plans included in the survey are profit sharing plans and therefore we determined the data are sufficiently representative of the 401(k) plan experience. PSCA, established in 1947, is a national, non-profit trade association of 1,200 companies and over six million plan participants. PSCA has conducted an annual survey of plans for nearly 60 years.

Selection and Categorization of Interviewees

As part of our approach for obtaining information on why plans use eligibility and vesting policies and participant understanding of these policies, we interviewed government officials and a total of 21 retirement professionals and academic researchers. We identified interviewees based on our prior work examining 401(k) plans and recommendations from initial interviewees. We selected interviewees who reflect a range of perspectives, from those with a focus on plan participants to those with a focus on plan sponsors. We selected federal government officials with a role in overseeing eligibility and vesting policies: officials of the Department of Labor, the Department of the Treasury, and the Internal Revenue Service. We also interviewed an official from the Securities and Exchange Commission, an agency that has a role in regulating the investment options into which plan participants direct their contributions. We categorized interviewees as retirement professionals if they provide retirement plan-related services (such as those who serve as consultants to plans), represent the interests of retirement plans or plan participants, or otherwise perform work in the retirement area. We categorized interviewees as academic researchers if they teach at an institution of higher education and focus on conducting scholarly research relevant to the retirement area. The views of those interviewed are not generalizable.

² Plan Sponsor Council of America, 57th Annual Survey: PSCA's Annual Survey of Profit Sharing and 401k Plans Reflecting 2013 Plan Experience (Chicago, IL: 2014). This was the most recently issued PSCA survey data available at the time of our review. PSCA's report shows the percentage of plans with eligibility, vesting, and related policies. However, since the report indicates that these data cover a total of 613 plans, we converted these percentages to numbers for reporting purposes, rounding to the nearest number.

³ Vanguard, How America Saves 2015: A Report on Vanguard 2014 Defined Contribution Plan Data (Valley Forge, PA.: June 2015). Vanguard is one of the nation's largest record keepers based on total assets managed and the number of participants served according to a 2015 record keeping survey conducted by PLANSPONSOR, an industry organization. The survey results can be accessed at www.plansponsor.com/2015-Recordkeeping-Survey/?type=top.

⁴ U.S. Census Bureau, Survey of Income and Program Participation, 2008 Panel. Specifically, the survey asked participants to indicate why they do not participate in their employer's retirement plan by selecting from a list of several different reasons. Failing to meet the plan's eligibility policies was among the answer choices provided.

⁵ United States Department of Labor, Employee Benefits Security Administration, Private Pension Plan Bulletin: Abstract of 2013 Form 5500 Annual Reports Data Extracted on 6/2/2015 (September 2015).

⁶ PLANSPONSOR, a member organization, provides information and solutions for America's retirement benefits decision makers. For more information, see http://www.plansponsor.com/about_plansponsor/

⁷ Plan Sponsor Council of America, 58th Annual Survey: PSCA's Annual Survey of Profit Sharing and 401k Plans Reflecting 2014 Plan Experience (Chicago, IL). PSCA is a non-profit trade association that supports employer-sponsored retirement plans. PSCA has conducted an annual survey of plans for nearly 60 years.

⁸Pensions & Investments is an international newspaper of money management that is written for pension, portfolio, and investment management. Its coverage includes business and financial news, legislative reports, global investments, product development, and other topics.

⁹ ASPPA is a non-profit professional member organization. Its goals are to educate retirement plan professionals and to create a framework of public policy that promotes retirement security. ASPPA has more than 7,000 members.

¹⁰ Some plans have age and service requirements that limit employees' ability to join the plan.

Employee salary and retirement age. We assume that the worker starts working at age 18 in 2016, and is continuously employed through retirement at age 67 in 2065. We modeled lifetime earnings using medium scaled earnings factors developed by the Social Security Administration's Office of the Chief Actuary (OCACT) (see table 8).¹ These factors express hypothetical earnings at each age as a percent of the Social Security Administration's national average wage index.² The scaled factors are based on average work and earnings of actual insured workers over their careers. This approach has the advantage of reflecting actual earnings histories, with steeper wage growth earlier in early- and mid-career work years, and flatter wage growth in late-career work years. Nominal wage increases an average of 5.6 percent per year. However, this approach does not reflect the possibility that less-skilled workers and lower earners may have flatter wage growth over their lifetime than higher-skilled workers. For this reason, we ran an alternative scenario for low earners featuring a constant nominal wage growth of 4.6 percent (see table 9). This alternative scenario demonstrates that the loss of early savings has a bigger effect on total savings at retirement for workers with flatter earnings growth than for workers with steeper earnings growth. We assume a retirement in 2065 at age of 67 because that is the Social Security full retirement age for workers starting their careers in 2016.

Inflation. To report adjusted salaries and other figures we indexed to 2016 dollars using the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W) projected under intermediate assumptions from the 2015 Old-Age, Survivors, and Disability Insurance (OASDI) Trustee's report.³

       Table 8: Annual Average Salaries for a Medium-Level Earner

Employer matching contributions. We assume that the individual in our hypothetical projections receives employer contributions through their continuous employment from age 18 to 66, except where employer contributions are delayed or forfeited to illustrate the effect of an eligibility or vesting policy. We again referred to the PSCA survey data from the 2013 plan year. The average employer contribution among 401(k) plans is 2.90 percent, which we rounded to 3 percent for our hypothetical calculations. Similarly, the most commonly used match formula is 50 percent of employee deferrals up to 6 percent of salary (a max of 3 percent) used by 26.2 percent of plans. The next most common levels of employer contribution are a 100 percent match up to 4 percent of employee pay and a 100 percent employer match up to 5 percent, used by 10.9 percent and 9.9 percent of plans, respectively. The hypothetical projected amounts of retirement savings from employer contributions that is foregone from delayed eligibility and forfeited contributions from unmet vesting policies could be even larger if one assumed a higher, though not uncommon, level of employer matching contributions. We also assume that employer contributions are made on a per-pay-period basis, unless calculated otherwise for comparison.

Returns. To set an annual return for each of the years of the worker's career for our hypothetical scenarios we formulated a composite return based on Social Security Trustee projections. For the fixed income portion of the return, we used the Social Security Trustees' projected annual trust fund real interest rate of 2.90 percent, and the projected consumer-price index for urban wage earners and clerical workers (CPI-W) of 2.70 percent, both published in the 2015 OASDI Trustee's Report intermediate long range economic assumptions⁴, for a nominal interest rate of 5.6 percent. For the equity portion of the return, we added an estimated long-term equity risk premium of 3.5 percentage points⁵ to the annual trust fund nominal interest rate, for a nominal rate of return on stocks of 9.1 percent. We changed the ratio slightly each year to reduce equities exposure and investment risk as the worker approaches retirement. The ratio corresponds to a "100 minus age" rule, which means that the percentage of assets invested in equities is set at 100 minus the worker's current age. For example, when the worker is 50 years old, we assume that just half of their assets are invested in equities and assign the nominal return on stocks to just 50 percent of their retirement savings, while the remaining portion is calculated to earn the nominal interest rate on bonds. The "100 minus age" rule for portfolio diversification is relatively conservative, thus the projected value at retirement that we report is less than it would be if we had assumed a rule using a higher equities allocation. For example, another rule used for portfolio diversification is "120 minus age," which would mean a higher equities allocation and higher projected returns, because over time equities tend to produce a higher average annual return than corporate bonds. A lower return assumption would result in lower projected savings lost or forfeited from the policies discussed in this report.

Leakage and fees. We assumed no leakage from the worker's account over time and did not apply any plan fees to the account balance. Both of these factors could decrease the rate of the account balance's growth over time.

Tables of Hypothetical Retirement Savings Using Alternate Assumptions

    Table 9: Comparison of the Value at Retirement of Potential Lost Savings

An employer's use of a minimum-service policy for employer contributions can also reduce potential retirement savings significantly. For example, our projections suggest that a 40-year-old worker who saves 5.3 percent of their own earnings (the average employee deferral percentage among non-highly compensated workers, according to the PSCA survey) could earn $526 over the year from an investment return of 7.7 percent, which is the net portfolio return for this age in our projections. However, if the worker receives an employer match of 3 percent of pay, that contribution amounts to a 57 percent annual return on their own contribution and an 8 percent investment return on the employer match, and adds another 4 percent annual return on their own contribution.⁶ Projected to retirement, the value of the employer contribution not received by that 40-year-old worker for 1 year could be $24,614 ($13,685 in 2016 dollars), or 1.4 percent of what they could save on their own by retirement.

A second, additional year of service before eligibility for the employer contributions results in more potential lost savings for the worker. If subject to a second year minimum-service policy three times over a career,⁷ a worker could miss out on 6 years of employer contributions. As shown in table 15 below, three employers applying a 2-year minimum-service policy for employer contributions (at ages 20, 30, and 40) could reduce potential 401(k) savings by $137,912 at retirement, which is 14 percent of the total savings at retirement that a worker could potentially have from employer contributions alone.⁸

     Table 15: Total Value of Potential Lost Savings from Five Episodes of

² In the OCACT actuarial note, scaled earnings factors begin at age 21 and end at age 64. To construct a hypothetical earnings history for workers who are 18 to 20, we assume that the scaled earnings factor is constant from age 18 to 21; because the average wage index (AWI) increases each year, scaled earnings (which equal the scaled earnings factor multiplied by the AWI) therefore increase slightly from age 18 to 21. We assume that scaled earnings are constant from age 65 to 66.

³ "2015 OASDI Trustee's Report: Estimates in Dollars," Sept. 1, 2015. Online: http://www.ssa.gov/OACT/TR/2015/VI_G3_OASDHI_dollars.html#182913.

⁴ 2015 OASDI Report, Section 3: Estimates in Dollars, Sept.1, 2015, online: http://www.ssa.gov/OACT/TR/2015/VI_G3_OASDHI_dollars.html#182913.

⁵ Peter A. Diamond, "What Stock Market Returns to Expect for the Future?" Social Security Bulletin, Vol.63, No.2, 2000.

⁶ This projection is based on a 40-year-old saving $6,827 ($3,796 in 2016 dollars) or 5.3 percent of their pay. The employer's contribution is a match of 3 percent of pay, which is $3,864 ($2,148 in 2016 dollars). The value of the investment return on the employer contribution is $298 ($166 in 2016 dollars). For details about assumptions used see above in Appendix II.

⁷ According to BLS data, the average workers starts about 11 new jobs, but not every workplace plan will have a second year service policy, so for illustrative purposes we assume only about a quarter (3) of those plans have this policy.

⁸ That amount is $8,440 in 2016 dollars. Values in this calculation are for jobs started at age 20, 30 and 40, which reflects the fact that turnover is more common at younger ages. Percentages are calculated based on projected total retirement savings of $969,674 ($262,673 in 2016 dollars) and based on employer contributions alone.

Thank you for the opportunity to review the Government Accountability Office's (GAO) draft report entitled "401(k) Plans: Effects of Eligibility and Vesting Policies on Workers' Retirement Savings." The draft contains two recommendations for the Secretary of Labor: (1) provide assistance to the Department of the Treasury in evaluating the appropriateness of existing maximum vesting policies for individual account plans and in seeking legislative action to revise vesting schedules, if deemed necessary; and (2) develop guidance for plan sponsors that identifies best practices for communicating information about eligibility and vesting policies in a clear manner in summary plan descriptions (SPDs).

The first recommendation, as noted in your draft report, involves matters within the jurisdiction of the Department of the Treasury/Internal Revenue Service (IRS). Specifically, the substantive provisions of Title I of ERISA governing eligibility and vesting provisions in 401(k) plans are under the interpretive and regulatory jurisdiction of the Secretary of the Treasury pursuant to Section 3002(c) of ERISA and section 101 of Reorganization Plan No. 4 of 1978. We understand that your draft report is being submitted to the Treasury Department for it to evaluate your recommendation and consider possible legislative proposals. The Treasury Department and IRS generally consult with the Department of Labor on subjects of joint interest and we expect they will do so in the case of your report.

With respect to the second recommendation, although the Department of Labor has general interpretive and regulatory authority over the participant disclosure and fiduciary responsibility provisions in Parts 1 and 4 of Title I of ERISA, we do not agree that implementing the recommendation allows best use of our limited resources. We agree that disclosures to participants that explain the eligibility and vesting provisions in their 401(k) plans are important to their ability to make informed choices about retirement savings. As your report notes, Title I of ERISA and the Department's regulations already require the SPD for a 401(k) plan to be written in a manner calculated to be understood by the average plan participant. The SPD must be sufficiently comprehensive to inform participants regarding their rights and obligations under the plan, including the plan's requirements respecting eligibility for participation and vesting of benefits. The information in the SPD cannot be presented in a way that would have the effect of misleading, misinforming or failing to inform participants and beneficiaries. Any exceptions, limitations, reductions and other restrictions on plan benefits cannot be minimized, rendered obscure, or otherwise be made to appear unimportant. The SPD also must clearly identify circumstances that may result in disqualification, ineligibility, or denial, loss, forfeiture, suspension, offset, or reduction of any benefits that a participant or beneficiary might otherwise reasonably expect the plan to provide based on the SPD's description of benefits under the plan.

An evaluation of "best practices" regarding eligibility, benefit accrual, and vesting disclosures should consider other disclosures provided to plan participants. For example, your report references the benefit statements that 401(k) plans must provide to participants regarding their individual accounts, and notes other tools that plan sponsors and plan fiduciaries use to communicate information to plan participants. We have a long-term project on our current regulatory agenda relating to the individual benefit statement requirements. Additional input from a broader range of plan sponsors and plan fiduciaries, possibly in the form of a Request for Information published in the Federal Register, could supplement the information GAO gathered and contribute to an informed development of "best practices" guidance.

We do not believe it would be appropriate at this time to reallocate resources away from our existing priority projects to a new "best practices" project focused on the recommendations in your draft report, especially given the regulatory requirements that currently apply to SPD disclosures on eligibility and vesting. We will, however, review our existing outreach materials on plan administration and compliance for opportunities to highlight the issues and recommendations in your draft report. We will also consider your recommendations in our ongoing development and prioritization of EBSA's agenda for regulations and sub-regulatory guidance.

We appreciate having had the opportunity to review and comment on your draft report. Please do not hesitate to contact us if you have questions concerning this response or if we can be of further assistance.

Sincerely,

Phyllis C. Borzi

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

Staff Acknowledgments

In addition to the contact named above, Tamara Cross (Assistant Director), Angie Jacobs (Analyst in Charge), Sherwin Chapman, Katherine D. Morris, Rhiannon Patterson, and Stacy Spence made significant contributions to this report. Additional support was provided by Jessica Artis, Deborah Bland, Julianne Cutts, Laura Hoffrey, Saida Hussain, Gene Kuehneman, Jill Lacey, Sheila McCoy, Mimi Nguyen, Dae Park, Joe Silvestri, Frank Todisco, Walter Vance, Kate Van Gelder, Adam Wendel, Jill Yost, and Chris Zbrozek.

Related GAO Products

Retirement Security: Federal Action Could Help State Efforts to Expand Private Sector Coverage. GAO-15-556. Washington, D.C.: September 10, 2015.

Retirement Security: Most Households Approaching Retirement Have Low Savings. GAO-15-419. Washington, D.C.: May 12, 2015.

401(k) Plans: Greater Protections Needed for Forced Transfers and Inactive Accounts. GAO-15-73. Washington, D.C.: November 21, 2014.

Private Pensions: Pension Tax Incentives Update. GAO-14-334R. Washington, D.C.: March 20, 2014.

Retirement Security: Women Still Face Challenges. GAO-12-699. Washington, D.C.: July 19, 2012.

401(K) Plans: Increased Educational Outreach and Broader Oversight May Help Reduce Plan Fees. GAO-12-325. Washington, D.C.: April 24, 2012.

Private Pensions: Some Key Features Lead to an Uneven Distribution of Benefits. GAO-11-333. Washington, D.C.: March 30, 2011.

Private Pensions: Low Defined Contribution Plan Savings May Pose Challenges to Retirement Security, Especially for Many Low-Income Workers. GAO-08-8. Washington, D.C.: November 29, 2007.

Private Pensions: Changes Needed to Provide 401(k) Plan Participants and the Department of Labor Better Information on Fees. GAO-07-21. Washington, D.C.: November 16, 2006.

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 (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 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, 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 www.gao.gov.

To Report Fraud, Waste, and Abuse in Federal Programs

Contact: