IRA testimony - t104






MARCH 19, 1997

I am pleased to appear before you this morning to discuss issues of individual retirement accounts (IRAs) and alternative tax-qualified retirement saving plans. My name is Paul Yakoboski. I am a senior research associate at the Employee Benefit Research Institute (EBRI), a nonprofit, nonpartisan, public policy research organization based in Washington, DC.

EBRI has been committed, since its founding in 1978, to the accurate statistical analysis of economic security issues. Through our research we strive to contribute to the formulation of effective and responsible health and retirement policies. Consistent with our mission, we do not lobby or advocate specific policy solutions.

IRA Usage

Through enactment of the Employee Retirement Income Security Act of 1974 (ERISA), Congress established IRAs to provide workers who did not participate in employment-based retirement plans an opportunity to save for retirement on a tax-deferred basis. U.S. tax law has substantially changed the eligibility and deduction rules for IRAs since then. The Economic Recovery Tax Act of 1981 (ERTA) extended the availability of IRAs to all workers, including those with pension coverage. The Tax Reform Act of 1986 (TRA '86) retained tax-deductible IRAs for those who did not participate in an employment-based retirement plan (and if married, whose spouse did not participate in such a plan), but restricted the tax deduction among those with a retirement plan to individuals with incomes below specified levels. In addition, TRA '86 added two new categories of IRA contributions: nondeductible contributions, which accumulate tax free until distributed, and partially deductible contributions, which are deductible up to a maximum amount less than the $2,000 maximum otherwise allowable. The Small Business Job Protection Act of 1996 increased the amount that may be contributed on a deductible basis on behalf of a nonworking spouse (if the working spouse is eligible for a deductible contribution) from $250 to $2,000.

The overwhelming majority of those workers eligible to make a tax-deductible contribution to an IRA currently choose not to do so. This is true among single workers and among married couples (both one earner and two earner couples). And it is true across income groups, although those with higher incomes are more likely to contribute when eligible (table 1).

According to EBRI tabulations of the April 1993 Current Population Survey employee benefits supplement (CPS-ebs), in 1992, 89 percent of all single workers were eligible to make an IRA contribution that was at least partially tax deductible. All such workers earning less than $35,000 (86 percent of single workers) were eligible. In addition, 22 percent of single workers earning between $35,000 and $49,999 and 20 percent of those earning $50,000 or more were eligible for a deductible IRA contribution.

Among single workers, only 5 percent of those eligible for a deductible IRA contribution actually contributed to an IRA in 1992. The likelihood of making a contribution increased with worker earnings. Only 1 percent of those eligibles making less than $10,000 contributed, compared with 27 percent of those making $50,000 or more.

Fifty-six percent of married couples with both spouses working were eligible to make an IRA contribution that was at least partially tax deductible. All such couples with combined incomes of less than $50,000 were eligible, and 10 percent of those with combined incomes greater than $50,000 were eligible. Among eligible two earner couples, 10 percent made an IRA contribution in 1992. Among eligible two earner couples, the likelihood of making a contribution increased with the couples' income. Among couples with a combined income of less than $10,000, essentially none contributed, while 23 percent of couples making $50,000 or more made an IRA contribution.

Married couples with one earner are more likely than those with two earners to be eligible for a deductible IRA contribution. Seventy-two percent of single earner couples were eligible to make an IRA contribution that was at least partially tax deductible. This included 100 percent of those earning less than $35,000, 22 percent of those earning $35,000 to $49,999 and 16 percent of those earning $50,000 or more. Among eligible single earner couples, 9 percent made an IRA contribution in 1992. Six percent of those making less than $10,000 made a contribution, compared with 22 percent of those making $50,000 or more.

While IRAs were created to allow individuals without an employment-based retirement plan to save for retirement on a tax-deferred basis, the fact is that the vast majority of those eligible to make tax-deductible contributions to an IRA choose not to do so. It is often speculated that this is due to a lack of money, but even among higher earning workers, those who are eligible for a deductible IRA still do not, in general, participate. It is also often speculated that individuals are reluctant to tie up their savings in a vehicle where it is beyond their reach, without significant tax penalties, should they need the money before retirement.

Salary Reduction Plans

Alternatives to IRAs exist that allow workers to save money for retirement on the same tax-deferred basis enjoyed by fully deductible IRA contributions. These plans, referred to here as salary reduction plans, are offered through work at an employer’s discretion, and therefore are not available to all workers. However, when they are available to workers, they do have some advantages relative to IRAs as a retirement wealth accumulation tool. These are discussed shortly.

Salary reduction plans include 401(k) plans, 457 plans, 403(b) plans, and the federal Thrift Savings Plan (TSP). The Revenue Act of 1978 permitted employers to establish 401(k) arrangements, named after the Internal Revenue Code (IRC) section authorizing them. In 1981, the Internal Revenue Service (IRS) issued the first set of proposed regulations covering such plans. These proposed regulations provided some interpretive guidelines for sec. 401(k) and specifically sanctioned "salary reduction" plans. Through 401(k) arrangements, participants may contribute a portion of compensation (otherwise payable in cash) to a tax-qualified employment-based plan. Typically, the contribution is made as a pretax reduction in (or deferral of) salary that is paid into the plan by the employer on behalf of the employee. In many cases, an employer provides a "matching" contribution that is some portion of the amount contributed by the employee, generally up to a specified maximum. The employee pays no federal income tax on the contributions or on the investment earnings that accumulate until withdrawal. Some plans also permit employee after-tax contributions; the earnings on these contributions are also not taxed until withdrawal.

Public-sector employers can establish deferred compensation plans under IRC sec. 457; charitable organizations qualified under IRC sec. 501(c)(3) (for example, a tax-exempt hospital, church, school, or other such organization or foundation) and public school systems and public colleges and universities can establish tax-deferred annuity plans under IRC sec. 403(b). The 1983 Social Security Amendments required that a new civil service retirement system be established to cover federal employees hired after December 31, 1983. The Federal Employees Retirement System (FERS), which Congress adopted in 1986 and which went into effect in January 1987, combines Social Security, a defined benefit pension plan, and an optional tax-deferred thrift plan similar to a private-sector 401(k) arrangement. Employees hired before the end of 1983 were given the option of joining the new system or remaining in the old Civil Service Retirement System (CSRS) during a six-month period ending in December 1987.

Comparison with IRAs

Salary reduction plans offer an advantage over IRAs in that the amount that can be contributed on a tax-deductible basis is much higher. The maximum deductible IRA contribution is $2,000 annually, compared with $9,500 for 401(k), 403(b) plans, and the federal TSP, and $7,500 for 457 plans. Furthermore, the limits on the salary reduction plans are indexed for inflation, while the IRA maximum is not. However, nondiscrimination standards for salary reduction plans in the private sector may limit the amount that highly compensated employees can contribute. In some instances such highly compensated employees may not be allowed to contribute the dollar amount cited above, and in extreme cases they may not be allowed to contribute anything to the plan as a result. Since IRAs are not employment-based, they are not subject to such nondiscrimination standards.

Employers will often provide matching contributions on a certain percentage of the earnings that a worker chooses to contribute (e.g., an employer may match 50 percent of the first 6 percent of pay that participants in the plan choose to contribute). Such matching contributions are optional on the part of the employer, and thus do not constitute an inherent advantage for these plans over IRAs. They may, however, serve as a strong incentive to participate, as will be discussed later.

A second advantage of salary reduction plans over IRAs is that the plan sponsor serves as a fiduciary filter for the thousands of investment options available today. Salary reduction plans offer participants a limited menu of investment options from which to choose. The plan sponsor has a fiduciary duty to choose the options offered in a prudent manner. In essence, the sponsor has already done the first round of screening for the participant.

Sec. 401(k) and 403(b) plans can allow loans to participants. Whether a plan has a loan feature is at the discretion of the plan sponsor. The federal TSP does have a loan feature. Sec. 457 plans are not allowed to offer loans. IRAs do not have loan features. However, IRA money can be withdrawn at any time for any purpose (it is generally subject to a 10 percent penalty tax if withdrawn before age 59 ½, in addition to income taxation). Salary reduction plans may allow withdrawals in instances of "hardship," but they are not required to do so. If a plan does not allow loans or hardship withdrawals, a worker would not be able to access the funds in his or her account under any circumstances until the time he or she leaves that employer.


Salary reduction plans continue to grow as an important element of the employment-based retirement income system. According to EBRI tabulations of the April 1993 CPS-ebs, the percentage of civilian nonagricultural wage and salary workers with an employer who sponsors a salary reduction plan (the sponsorship rate) increased from 27 percent (27 million workers) in 1988 to 37 percent (39 million workers) in 1993 (table 2). Over the same time period, the fraction of all workers participating in such plans (the participation rate) rose from 15 percent (16 million workers) to 24 percent (25 million workers). The fraction of participating workers among those where a salary reduction plan was sponsored (the sponsored participation rate) also increased, rising from 57 percent to 65 percent. The growth in salary reduction plan sponsorship and participation has occurred across almost all worker and job-related characteristics, including firm size.

The likelihood of salary reduction plan sponsorship and participation increased with firm size (table 2). In 1993, 5 percent of those employed by a firm with fewer than 10 employees reported that their employer sponsored a salary reduction plan, as compared with 54 percent of those employed by firms with 1,000 or more employees. When a plan was sponsored, the participation rate did not vary systematically with firm size. In all but the smallest employer category, the participation rate among workers where a plan was sponsored was about two-thirds. In the smallest firms (fewer than 10 employees), almost three-quarters of workers where a plan was sponsored chose to participate. Therefore, the positive relationship between firm size and overall participation rates was solely a function of the positive relationship between firm size and sponsorship rates.

The higher a worker’s earnings, the more likely he or she was to have a plan available at work. Two-thirds of workers earning $50,000 or more had an employer that sponsored a salary reduction plan, compared with only 8 percent of workers earning less than $5,000 (table 2). Furthermore, when a plan was available, higher earning workers were more likely to participate than lower earners. Twenty percent of workers earning less than $5,000 contributed to a plan when one was offered, compared with 83 percent of workers earning $50,000 or more.


As seen above, participation rates among eligibles are much higher for employment-based salary reduction plans than for IRAs. Why?

Participation in a salary reduction plan is generally more convenient since it is offered through the workplace and involves automatic contributions from a worker's paycheck before he or she even sees the money. Plan sponsors will also market the plan to their employees and typically educate them as to the importance for their retirement income security of participating in the plan. With IRAs, on the other hand, an individual must make a conscious decision to seek out such information on his or her own (unless it is offered through work). Moreover, it has been speculated that some workers who are eligible for a tax-deducible IRA contribution may not be aware of their eligibility.

Another important reason is the availability of employer matching contributions with salary reduction plans. Among workers whose employer sponsored a salary reduction plan in 1993, 51.3 percent reported that their employer provided matching contributions to the plan. The actual percentage was likely higher because 30.2 percent did not know if their employer matched contributions. Among those responding that their employer did provide a matching contribution, the average reported match rate was 65 percent (i.e., for every $1 the employee contributed, the employer contributed 65 cents). Such employer matching contributions are not available with IRAs.

Studies have found evidence that the availability of an employer match does have an effect on participation. For example, a 1995 Hewitt Associates' study of 401(k) plans found an average participation rate of 76 percent in plans with an employer match as opposed to an average of 59 percent in plans with no employer match. Similarly, a 1996 Buck Consultants study of 401(k) plans found an average participation rate of 67 percent in plans with no employer match, compared with participation rates near 80 percent in plans with some form of employer matching contribution.

Finally, the other notable point from the data presented above is that, despite the rapid growth over recent years in the number of salary reduction arrangements in small firms, it is at the small plan level that a noticeable gap in plan sponsorship remains. The question naturally arises as to what, if anything, can be done to fill this void? SIMPLE IRAs and SIMPLE 401(k)s were created by the Small Business Job Protection Act of 1996 for this very reason. Time will tell how successful they will be.