Turning workers into savers? Incentives, liquidity, and choice in 401(k) plan design.

INTRODUCTION

Company-sponsored defined contribution (DC) plans are now at the heart of the US private-sector retirement system, with nearly 60 million private sector employees holding $2 trillion in 401(k) pension assets. (1) Those who contribute to a 401(k) plan throughout their working careers can anticipate replacement incomes of nearly 60 percent from these plans; indeed, for Baby Boomers, company-based DC plans will provide larger retirement benefits than Social Security. (2) Against this optimistic backdrop, however, there is concern that some employees may fail to properly exercise the new responsibilities imposed on them by participant-directed DC plans, particularly if they fail to join their 401(k) plans, save inadequately, overly concentrate their portfolios in a single asset, or spend their retirement savings too quickly. (3)

This paper seeks to explain how the incentives embedded in 401(k) plans influence workers' propensity to save on a tax-deferred basis. Specifically we evaluate how saving behavior is influenced by 401(k) plan design features, including the structure of the promised employer match, the presence of loans, and the investment menu design, and to differentiate these effects from saving patterns attributable to workers' own preferences or firm-specific characteristics. We also evaluate the determinants of employer actual costs for matching contributions as well as how these costs vary across workforce and firm characteristics. We focus on the 401(k) plan as the unit of analysis, rather than on individual employees, as this is critical in understanding how firm-specific rules shape the incentives for retirement saving.

Prior research has focused mainly on how employer matching contributions in a DC plan might shape employee retirement saving behavior and, to a lesser extent, why employers might offer such incentives to workers. We develop an empirical model of these incentives using a rich new dataset of more than 500 defined contribution pensions covering nearly 740,000 employees, with unique detail on plan design, employer, and workforce characteristics. In particular, the dataset includes high-quality information on the design of matching contributions as well as investment and liquidity features of each plan. Only with this sort of highly nuanced plan-specific information is it possible to properly quantify the complex set of incentives and constraints shaping employee saving.

Our empirical results using this large and rich cross-section of pension plans point to two main conclusions. First, and most importantly, the incentive effects of employer matching contributions are quite small. For instance, in a typical plan, 65 percent of non-highly compensated workers participate in the retirement plan independent of the employer match, and 25 percent fail to participate at all. Only ten percent of workers respond to the match and join the plan as a result. Importantly, these results are after controlling for the unique liquidity and investment constraints of 401(k) plans. Second, the typical employer promises a three percent match but pays much less due to low employee saving rates. Overall, workers forfeit about half of the promised employer saving incentive (after controlling for workforce and firm differences).

These results underscore the limited efficacy of using contribution matches to boost retirement saving. More broadly, matching contributions appear to do little to turn workers into savers, with 90 percent of workers ignoring these employer incentives. To foster broader participation in 401(k) plans, consideration should be given to other approaches such as automatic enrollment, non-elective employer contributions, and mandatory employee contributions. Additionally, our research substantiates the importance of provisions such as those in the 2006 US Pension Protection Act (PPA) encouraging automatic enrollment, and the UK National Pension Savings Scheme, which combines automatic enrollment with employer matching contributions for workers not covered by a private-sector pension. (4) Our findings on matching incentives, liquidity, and investment choice are also relevant when considering possible models for a reformed Social Security system or other public defined contribution systems with individual accounts, as well as the design of private-sector 401(k) plans.

In what follows, we first briefly review the characteristics of 401(k) plans as well as prior research. Next we describe our data and methodology. We then report on an analysis of the determinants of employee savings behavior, followed by an assessment of employer matching costs. A final section offers conclusions and implications.

CHARACTERISTICS OF 401(K) PLANS

US employers may choose to voluntarily provide tax-qualified retirement plans to their workers; today most private-sector firms with a company-based program offer a 401(k) plan. (5) 401(k) plans are funded principally by employee contributions from wages, know as "elective deferrals." Employer contributions are typically in the form of matching contributions, which are contingent on employee contributions, although some employers may offer a non-matching contribution for all eligible employees, whether or not they participate in the plan.

Workers receive meaningful tax incentives for 401(k) contributions. In 2001, the year of our study, employee 401(k) contributions were tax-deductible up to the section 402(g) limit of $10,500. All employer contributions (up to specified limits) are also non-taxable to the employee at the time they are made. All employer and employee contributions, along with related earnings, compound tax-free until withdrawn. Plan withdrawals are subject to tax only when they are taken (typically later in life when tax rates may be lower). (6)

Besides the annual 402(g) contribution limit, retirement plans are also subject to nondiscrimination testing (NDT) rules for employee pre-tax elective deferrals. (7) To implement these rules, a plan sponsor must divide eligible workers into two groups: highly compensated employees (HCEs, earning $85,000 or more in 2001) and non-highly compensated employees (NHCEs, or those earning below $85,000 in 2001). In the most common situation covering employee contributions (the so-called ADP or "actual deferral percentage" test), the HCEs' plan contribution rate may not exceed that of the NHCEs by more than two percent. (8) Moreover, when calculating plan saving rates under the NDT rules, the employer can only count pay subject to the section 401(a) definition of compensation, which caps the maximum level of annual pay considered for retirement plan purposes ($170,000 in 2001). Consequently an employee earning $1 million and contributing $10,500 would have an actual plan contribution rate of just over one percent ($10,500 divided by $1 million), but his plan saving rate for nondiscrimination testing purposes would be 6.18 percent ($10,500 divided by $170,000). Figure I demonstrates this effect generally for HCEs making the maximum 402(g) contribution at various income levels.

[FIGURE 1 OMITTED]

Under the NDT rules, as the number of HCEs contributing the maximum (and the number earning more than $170,000) rises, plan contribution rates for the HCE group converge toward 6.18 percent for NDT testing purposes (in 2001). As a result, employers generally must encourage NHCEs to save at least 4.18 percent of earnings (two percent less than the HCE threshold of 6.18 percent) in order to comply with nondiscrimination testing. NHCE savings rates will actually have to be higher than 4.18 percent if not all HCEs contribute the maximum (or more of the HCE population earns between $85,000 and $170,000 per year). When many NHCEs fail to join the plan (i.e., their deferral rate is zero), it becomes difficult to satisfy this criterion, and the plan will "fail testing." In this case, employers will typically limit contributions by the highly paid (to an amount below $10,500) in order to comply with the rules. (9)

Beyond these tax considerations, there are two structural constraints that employees face when deciding whether to participate in an employer's plan. First, by law 401(k) saving is relatively illiquid. As a rule, the participant's savings must remain in the plan until the worker changes jobs or retire, and he faces a ten percent tax penalty if the withdrawal occurs before age 59 1/2. There are exceptions to this: many 401(k) plans offer limited access via a loan program, for instance, allowing participants to borrow up to half their balance with no tax consequences to a maximum of $50,000. (10) Some plans permit after-tax contributions, which are typically liquid and available for withdrawal at any time. (11) And in other cases, participants facing financial hardships may request in-service withdrawals. (12) But generally, the money is intended for retirement purposes and there are limits to pre-retirement access. Second, 401(k) saving is restricted to a rather limited investment menu. For instance, the employer may include in the menu a dozen or so funds and would not permit access to all retail mutual fund and brokerage account offerings.

To some extent, these limitations may be offset by the fact that many employers offer incremental compensation, the matching contribution, as a reward for plan saving. In practice, as we noted below, promised matching contributions vary from zero to over six percent of pay. In the modal case, an employee who contributes six percent of pay to a 401(k) plan receives a match of three percent (the equivalent of an instantaneous return of 50 percent on plan saving), for a total plan savings rate of nine percent of income.

PRIOR RESEARCH