ERISA at 50: A New Model for the Private Pension System Pamela Perun and C. Eugene Steuerle T H E R E T I R E M E N T P R O J E C Occasional Paper Number 4 URBAN INSTITUTE T ERISA at 50: A New Model for the Private Pension System Pamela Perun and C. Eugene Steuerle T H E R E T I R E M E N T P R O J E C Occasional Paper Number 4 T The Retirement Project ABOUT THE SERIES THE RETIREMENT PROJECT IS A MULTIYEAR research effort that will address the challenges and opportunities facing private and public retirement policies in the twenty-first century. As the number of elderly Americans grows more rapidly, Urban Institute researchers will examine this population’s needs. The project will assess how current retirement policies, demographic trends, and private-sector practices influence the well-being of older individuals, the economy, and government budgets. Analysis will focus on both the public and private sectors and will integrate income and health needs. Researchers will also evaluate the advantages and disadvantages of proposed policy options. Drawing on the Urban Institute’s expertise in health and retirement policy, the project will provide objective, nonpartisan information for policymakers and the public as they face the challenges of an aging population. All Retirement Project publications can be found on the Urban Institute’s Web site, http://www.urban.org. The project is made possible by a generous grant from the Andrew W. Mellon Foundation. This study was made possible by a generous grant from the J.M. Kaplan Foundation. Copyright © March 2000. The Urban Institute. All rights reserved. Permission is granted for reproduction of this document, with attribution to the Urban Institute. The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Table of Contents About the Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 ERISA at Zero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 ERISA at 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 ERISA at 50 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Coda: Retire ERISA at 65 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Endnotes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Appendix. Summary of Selected Rules circa 2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 FIGURES AND TABLES Figure 1. ERISA at Zero . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Figure 2. ERISA at 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Figure 3. ERISA at 50 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Figure 4. ERISA at 65 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Appendix. Summary of Selected Rules circa 2000: Part 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Appendix. Summary of Selected Rules circa 2000: Part 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 ABOUT THE AUTHORS PAMELA PERUN IS A CONSULTANT TO THE Retirement Project at the Urban Institute. She holds a law degree from the University of California at Berkeley and a Ph.D. in developmental psychology (adult development and aging) from the University of Chicago. She has practiced as a benefits lawyer for 10 years in Boston and Washington, D.C., and also has held research appointments at Duke University, Wellesley College, and Harvard Medical School. C. Eugene Steuerle is a senior fellow at the Urban Institute and author of a weekly column, “Economic Perspective,” for Tax Notes magazine. He has worked under four different U.S. presidents on a wide variety of social security, budget, tax, health, and other major reforms, including service both as the deputy assistant secretary of the Treasury for tax analysis (1987–89) and as the original organizer and economic coordinator of the Treasury’s 1984–86 tax reform effort. He is the author or coauthor of over 150 books, articles, reports, and testimonies, including the recent Urban Institute Press books The Government We Deserve: Responsive Democracy and Changing Expectations, Retooling Social Security for the 21st Century, The New World Fiscal Order, Serving Children with Disabilities, and The Tax Decade. ERISA at 50: A New Model for the Private Pension System INTRODUCTION IN 1999, THE EMPLOYEE RETIREMENT Income Security Act of 1974 (ERISA), the primary law regulating the private pension system, turned 25.1 For the most part, people agree that ERISA is showing its age and is in need of substantial reform. However, few agree on the shape such reform should take.2 Since its enactment, ERISA has expanded in often unanticipated and irrational ways. As a result, the private pension system is now burdened with overly complex rules, regulations, and plan types that inhibit its ability to generate adequate retirement income to millions of Americans. This paper proposes a new model for ERISA of vastly simplified plans and rules intended to make the private pension system more accessible by employers and employees alike. ERISA AT ZERO Twenty-five years ago, the pension world was a much simpler place. This was the basic structure of the U.S. pension system circa 1974 (figure 1). ERISA, as originally conceived, was a sensible and rational approach to pension regulation. Its primary goal was to create a basic regulatory framework for the U.S. pension system that would remedy perceived inadequacies in participant rights and enforcement, benefit security, funding adequacy, and government oversight.3 The expectation was that the reforms embodied in ERISA would, in the long run, foster the expansion of the pension system. In some respects, it was a revolutionary statute. It established requirements for reporting and disclosure and standards for fiduciary responsibility, administration, and enforcement where previously none existed. In other respects, it was largely evolutionary. It reformed tax code rules for retirement plans that had been evolving over the previous 50 years for eligibility, participation, coverage, vesting, benefit accrual, and funding. ERISA, at birth, gave us a comprehensive set of rules designed to promote uniformity, consistency, and predictability in the pension system. It was a difficult and complex statute, but it was also a practical and workable statute whose reforms were well reasoned and well intentioned. FIGURE 1. ERISA at Zero ERISA at Zero § 401(a) Qualified Plans Defined Benefit Plans Money Purchase Stock Bonus Defined Contribution Plans Profit Sharing § 403 TaxSheltered Annuities § 403(a) Annuities Thrift Executive Compensation Arrangements § 403(b) Arrangements Employee Contributions Only Employer Contributions ERISA was a product of its time. It was based on the assumption that defined benefit plans would be the mainstay of the pension system. It emphasized replacement ratios, rather than account balances, as a measure of pension adequacy. The long-term viability of such plans, which contain an inherent age bias, was not questioned, even though the baby boom was just then entering adulthood and the demographic significance of such a large cohort was well recognized. Nor did ERISA’s proponents consider how plans should evolve if life expectancies increased and the demand for labor changed. ERISA may have been appropriate for the benefits world of its day, but many of the design decisions made 25 years ago have since proven problematic. ERISA AT 25 Life has gotten more complicated in the last 25 years, and the U.S. pension system is no exception (figure 2). The number and types of plans continue to multiply. Special purpose plans, such as employee stock ownership plans (ESOPs), have been invented or expanded. Families of plans, such as simplified employee pension plans (SEPs) and individual retirement account–based savings incentive match plans for employees (SIMPLE IRAs), have been created for special employers, such as small businesses. Plans with special tax rules, such as Roth IRAs, have also been made available. New plan designs, such as cash balance plans and 401(k) plans, have transformed older plan formats. ERISA’s regulatory framework has also become more elaborate. In the past 25 years, successive pieces of legislation have modified the operation of the pension system, almost always in a patchwork fashion. While the basic framework set in place by ERISA remains, critical elements of its infrastructure—funding, participation, coverage, vesting, and nondiscrimination—have been continually amended.4 Some of these changes had laudable outcomes. For example, the acceleration of vesting requirements improved worker ownership of benefits. Spousal rights to survivor benefits have been strengthened. Others had positive intentions but more mixed results. For example, the development of quantitative nondiscrimination rules and similar measures had the positive intention of promoting pension equity for lower-paid workers but may have induced the negative outcome of reducing plan sponsorship by small businesses. Despite their potential merits, the sheer accumulation of changes has added multiple layers of complexity to ERISA. The appendix illustrates some of the basic rules that apply to the more common types of plans in 2000. This modest summary requires four pages of charts, and 2 T H E R E T I R E M E N T P R O J E C T it does not attempt to discuss multiple plan interactions, different rules for different employers, or instances in which ERISA does and does not apply. Over the past 25 years, the tax code has spawned an extraordinary number of incentives for saving, each with its special plan design and its own separate limits, exceptions, and requirements. At the same time, the pension system is also vulnerable to changes in the fiscal health of the federal budget. When times are tight, the pension system is an obvious candidate for raising additional tax revenues. During these times, pension tax provisions are amended to restrict contributions and tighten other limits. For example, the reimposition of limits on deductible contributions to IRAs in the 1980s was in part designed to finance tax reform. When times are flush, Congress becomes generous in offering new ways to save. For example, previously imposed limits are often raised or there is a political desire to do something new so even more savings incentives are created. By the time the cycle is complete, pension law has become more complex and less rational. Current proposals to change pension law illustrate this process well. With budget surpluses available, the pendulum has swung away from the contribution restrictions and cutbacks prevalent during the past 15 years. Many legislators now advocate adding new incentives for saving, such as more Roth IRA−like contributions, to the private pension system. Others would like to expand the current system with an upward adjustment to almost all existing limits on contributions. A number of such proposals were incorporated into the Taxpayer Refund and Relief Act of 1999, vetoed by President Clinton, and may be added to other bills in the near future. If enacted, these changes would substantially reverse the direction that pension policy has taken in recent years. Whatever their economic merits, these changes would inevitably add more regulatory complexity to plan administration. POPULAR PROPOSALS INCLUDE: · · · · · · Increasing the amount of allowed annual contributions by employers and employees. Counting higher amounts of compensation in benefit calculations. Permitting Roth IRA−type contributions in employee savings plans. Creating “catch-up” elections for more savings by older employees. Relaxing nondiscrimination, top-heavy, and coverage rules for qualified plans. Enhancing portability of benefits between plans and IRAs. FIGURE 2. ERISA at 25 ERISA at 25 § 403 TaxSheltered Annuities § 401(a) Qualified Plans Defined Benefit Plans ERISA at 50: A New Model for the Private Pension System Traditional Defined Contribution Plans Money Cash Balance Purchase ESOPs With 401(k) Feature Stock Bonus § 403(a) Annuities Thrift Without 401(k) Feature § 403(b) Arrangements Individual IRAs Employee Employer Profit Contributions Contributions Sharing Only With 401(k) Feature SIMPLE 401(k) Without 401(k) Feature Standard 401(k) Nonqualified Deferred Compensation § 408, § 408A IRAs Regular IRAs Employer IRAs Roth IRAs SEP IRAs § 457(b) Plans SIMPLE IRAs Executive Compensation Arrangements § 457(f) Plans for Governmental Tax-Exempt Employers Plans of Other Employers 3 Pension law has also grown more complex because of the tax code’s carrot-and-stick treatment of employers. On the one hand, Congress provides tax incentives to encourage employers to sponsor plans. On the other hand, it developed layers of complex rules to inhibit employers whom it fears will abuse those tax incentives by providing substantial benefits only to high-paid workers. For example, top-heavy rules, permitted disparity rules, controlled group rules, leased employee rules, coverage rules, and affiliated service group rules ensure that employers provide at least some benefits to low-paid workers as well. In addition, the nondiscrimination rules themselves (which measure whether employers have been too generous to high-paid workers) further control employer discretion but increase the costs of compliance. Of course, not all blame for the complexity of pension law belongs to Congress. Lobbyists often seek special benefits provisions or fight simplification efforts on behalf of their clients. Benefits practitioners, who rarely take the lead in promoting simplification proposals, also share some of the responsibility. Practitioners are not likely to question a proposed rule’s underlying rationale; instead, they focus on the legal minutiae required to keep their clients in compliance. Perhaps simplification seems too difficult to achieve, or the status quo appears reasonable enough. As a result, ERISA is no longer a sensible or practical statute. The pension system it governs is an uncoordinated jumble of plans, with a labyrinth of rules and regulations. The following are a few of its most frustrating characteristics from a legal perspective: Complexity: There are too many rules. Most plans require, in addition to in-house benefits staff, an army of outside experts—attorneys, accountants, actuaries, consultants, record-keepers, investment professionals, and communication specialists—just to keep plans in compliance with existing laws. The rules also change constantly, usually in the direction of greater complexity. Prior to ERISA, IRC § 401(a), the major tax statute for pension plans, ended with IRC § 401(a)(10). In 1999, it ended with IRC § 401(a)(34). This averages out to one fundamental legal requirement added each year since ERISA was enacted. Uncertainty: The rules are almost never clear. Even when statutes appear straightforward, voluminous pages of regulations are required to explain their application to various situations. Because benefits law questions are heavily fact-specific, even those regulations cannot address all issues. When the law changes, as it often does, regulatory guidance follows slowly, leading to further uncertainty. For example, the nondiscrimination rules for qualified plans were changed by the Tax Reform Act of 1986 to more quantitative standards. Those changes 4 T H E R E T I R E M E N T P R O J E C T affected many other sections of pension law. The new rules were to become effective as of January 1, 1989. But because the changes were so extensive, regulatory interpretation evolved only over a number of years. The rules were not actually finalized until the early 1990s and finally became effective only as of January 1, 1994. In the interim, plans lived in compliance limbo. The pension system is supposed to be a self-regulating system, but employers often find even substantial compliance to be difficult given the legal uncertainties created by the sheer volume of rules. Inconsistency: There are more exceptions than rules. Not all rules apply to all employers, contributions, or plans. Some, though not all, types of contributions are subject to Social Security taxes. A defined benefit plan sponsored by a for-profit employer is heavily regulated, but the identical plan sponsored by a state or local government employer escapes most regulation. The same is true of employee savings plans. A 401(k) plan, used primarily by corporate employers, must comply with difficult regulations that require complicated annual testing procedures. This is less true of 403(b) arrangements for tax-exempt employees and 457 plans for governmental employees, which are minimally regulated at the employer level. Volatility: The rules often flip-flop. For example, tax-exempt employers could choose to sponsor new 401(k) plans before July 2, 1986, and after December 31, 1996, but not in the interim. Before 1989, taxpayers who continued to work after age 70½ were not required to begin taking plan distributions until they actually retired; however, between 1990 and 1996, they were. Since 1997, the old rule that no distributions are required until actual retirement has been reinstated. The Taxpayer Refund and Relief Act of 1999 would have largely repealed the complicated nondiscrimination regulations authorized by the Tax Reform Act of 1986 and finalized just five years ago. Any change, even if relatively insignificant legally, imposes new costs on plans that must be paid either by plan sponsors or participants. These costs add to the already considerable expenses required to keep a plan in compliance with existing rules every year. Every time a rule is rewritten, a special exception is provided, or a new incentive is created, plan documents must be drafted or redrafted in a form approved by the Internal Revenue Service. Participant disclosure materials and plan forms must also be rewritten. Record-keeping, tax reporting, investment data, and other information processing systems often must be modified. There are additional costs, not related to taxes, generated by this complexity. Employers, and indirectly employees, bear the costs of the in-house benefits per- sonnel required to administer each plan. Other costs are usually borne more directly by the plan and, ultimately, plan participants. These costs include Pension Benefit Guaranty Corporation (PBGC) premiums, in the case of a defined benefit plan, and the fees of the outside consultants and advisers required to keep the plan in compliance with applicable law. There are, in addition, custody and trustee costs related to the fiduciary requirements for holding and investing plan assets. The investments themselves involve investment management fees, transaction costs, operating expense fees, account maintenance fees in the case of mutual fund investments, sales charges, redemption fees, premium taxes, and so on. The cumulative effect of all these separate costs, charges, and expenses has been masked in recent years by the performance of the stock market. However, they inevitably reduce the rate of return obtained by pension plan assets and, in the case of defined contribution plans, the amount of retirement income ultimately payable to plan participants. The cost of this complexity extends beyond the pension system to society as a whole. Its true cost is an expensive and inefficient pension system that provides insufficient benefits to many workers and no benefits to many more. The fact that ERISA is now a partially dysfunctional statute would generally be of interest only to benefits practitioners. But ERISA is not an ordinary statute. It affects virtually every American because it is the prime mechanism, other than Social Security, through which people accumulate private resources for retirement. For this reason, we have great need for it to perform better. From a societal perspective, one of the most important issues for the next 25 years will be the aging of the U.S. population. Due to increases in longevity in the twentieth century, more people are living long enough to retire and are living longer in retirement. In addition, the number of retirees is scheduled to increase rapidly once the baby boomers begin to retire in large numbers. Already over half of the civilian programmatic spending in the federal budget goes to the elderly. The long-term ability of Social Security and Medicare to continue paying the current level of benefits is in question. Government spending on retirees is now growing faster than the economy.5 Close to one-third of the adult population will soon be either retired or disabled if current patterns hold. Currently, the majority of elderly Americans rely on Social Security as their primary source of retirement income and about 40 percent rely on it almost exclusively. Yet Social Security benefits were never intended to serve as the sole source of retirement income; they were meant to be supplemented with personal savings and pension income. However, personal savings rates have been declining for decades and today many workers are not preparing for retirement by saving on their own. Those who do save may not be able to accumulate sufficient assets to maintain their preretirement standard of living.6 Although the government is now promoting educational campaigns on the importance of saving for retirement, it is unlikely that they will have more than marginal success. Without additional private resources, the financial demands created by an aging society will be difficult to meet, and generating those additional resources will also be difficult. Through a combination of Social Security, pension income, and private savings, those in the top half by income have the capacity to prepare for the 20 or so years of retirement that is now the norm. Providing those in the bottom half with the means to accumulate private resources remains a difficult societal problem. The pension system has a critical role to play in the next 25 years. It could mitigate the financial burdens of an aging society in several important respects. For example, it could provide more incentives for people to retire later and save more. It could also provide flexible benefits options for older workers. It could promote more plan funding by employers and expand plan participation by employees. However, 1999 pension coverage rates are not much different from those in 1974. At any point in time, fewer than 50 percent of workers are participating in a retirement or savings plan, a figure that has not changed since the 1970s (U.S. General Accounting Office 1997). Although many of these workers will at some point be covered by a plan, overall participation rates have been falling in recent years.7 Moreover, aggregate pension benefits are unevenly distributed. Bettereducated, longer-service, full-time, higher-paid employees at larger companies, especially federal and state government employees, get the majority of pension benefits. In one important respect, the pension system in 1999 is unlike its 1974 version and produces a very different type of retirement income. The number of defined benefit plans that usually provide a guaranteed lifetime income in retirement has fallen by 60 percent since 1984. Although the actual number of workers covered by defined benefit plans has not decreased significantly, the ratio of active to retired employees covered continues to decline (Private Pension Plan Bulletin 1999: 2). Few new defined benefit plans are being established, with the possible exception of plans for professional corporations or sole proprietorships covering only a small number of workers. However, defined contribution plans, which provide retirement income based on individual account contributions and their investment earnings, are thriving. Today, defined benefit plans are commonly found only ERISA at 50: A New Model for the Private Pension System 5 in large companies while defined contribution plans are prevalent among companies of all sizes. In 1975, 32 percent of workers with employer plans participated in a defined contribution plan and 87 percent participated in a defined benefit plan. By 1995, 81 percent of similar workers were covered by a defined contribution plan and 51 percent by a defined benefit plan (Private Pension Plan Bulletin 1999: 3). Defined benefit and defined contribution plans held roughly the same amount of assets in the aggregate in 1995, but defined contribution plans received more in contributions ($117 billion for defined contribution plans and $41 billion for defined benefit plans) and paid out more in benefits ($98 billion for defined contribution plans and $85 billion for defined benefit plans) (Private Pension Plan Bulletin 1999: 3, 8). While employers now favor defined contribution plans, the stunning growth of 401(k)-type plans is employee driven. Although a 401(k) plan is employer sponsored, its primary attraction is that it allows employees to decide whether, and how much, to save on a pretax basis. In 1984, there were only 17,000 of these plans. By 1995, they had increased 10-fold to about 200,000, and there were an estimated 320,000 in 1998. The number of workers participating in them has also increased exponentially from about 7.5 million employees in 1984 to 34 million in 1998 (Private Pension Plan Bulletin 1999: 86).8 Despite the additional cost of Social Security taxes on contributions by employees to 401(k)-type plans, employers who sponsor such plans increasingly use them as their only plans.9 Although employers can and do contribute to these plans, employees contribute more in the aggregate than employers to 401(k)-type and other defined contribution plans (Private Pension Plan Bulletin 1999: 3).10 The popularity of defined contribution−type formats has had an effect on the structure of defined benefit plans themselves. Many employers are adopting the increasingly common but controversial practice of converting their traditional defined benefit plan into a cash balance plan, a hybrid between a defined benefit plan and a defined contribution plan. Employers fund cash balance plans under traditional defined benefit plan rules, but employees accrue benefits under defined contribution plan−like rules. In a traditional defined benefit plan, employees accrue benefits based on years of service and compensation earned over a specified period. Employees accrue benefits under a cash balance plan through hypothetical allocations to their accounts each year based on their compensation and a specified interest credit. Account accumulations determine the pension benefit ultimately received. Conversions of traditional defined benefit plans to cash balance plans have recently become controversial because many older employees have found 6 T H E R E T I R E M E N T their expected benefits in the converted plan to be much less than they would have earned for additional work under their prior plan. This issue aside, cash balance plans can be popular with employees because the format is more appealing, younger workers earn larger benefits, and benefits are more portable than under the traditional defined benefit plan. The popularity of defined contribution plans, particularly 401(k)-type plans, and the decline in traditional defined benefit plans have changed the U.S. pension system. In 1974, employers assumed responsibility for sponsoring and funding the retirement income of their employees. Today, individual workers take on much more of the risk and responsibility for funding their own retirements. There are several possible reasons for this shift. For example, employers recognize the costs of providing benefits that replace a high ratio of preretirement income for extended retirements. In addition, fewer employees remain with one employer for substantial portions of their work lives, and both employers and employees are attracted to the flexibility of defined contribution plans. The trend to defined contribution plans increasingly funded by employees is expected to continue for the foreseeable future. ERISA AT 50 About 50 bills proposing changes to one or more aspects of ERISA were introduced in Congress last year, and many were included in the vetoed Taxpayer Refund and Relief Act of 1999.11 Most proposals were well intentioned, but they all suffered from one common defect: the assumption that the current system is worth preserving in more or less its current form. They conserved the status quo while tinkering at the edges, changing this rule here and that limit there. They also sought to add new types of plans and features to a system already in regulatory gridlock. In short, they failed to demonstrate any broad vision about what ERISA could be. The vision of ERISA at 50 presented here is predicated on the following observations: There is a widely held belief that the key to fixing ERISA is to get more employers to provide a plan, especially a defined benefit plan, for their employees. This is wishful thinking.12 It has not happened in the past 25 years and there is no indication that it will happen in the future in the absence of, and possibly even with, new or different incentives or true regulatory simplification. The primary growth in the pension system will continue to be in defined contribution and defined contribution−like plans, especially those permitting employee savings but also with employer contributions. · · P R O J E C T defined benefit plans dominated the pension · When system, the rationale for structuring it through employer-based plans and rules was clear. The growth in plans where employees save their own money for retirement, or employers act more as contributors than managers, calls this structure into question, although employers can play an important role in encouraging employees to participate in these plans. As employee savings plans become the primary engine for generating retirement income, the adequacy of such plans for moderate-income workers must be addressed. These observations lead to the following propositions. First, the pension system should be devoting more of its resources to increasing employee savings. This means concentrating more of its efforts on defined contribution and defined contribution−like plans. It also means rationalizing the means by which employees save by standardizing the different types of plans now available. Second, while defined benefit plans will continue to play an important role in generating retirement income for millions of Americans, it is unlikely that sponsorship of new defined benefit plans will increase substantially in the coming years. However, existing plans could be reformed to remove their disincentives for employees to work longer. The pension system should be exploring ways to make these plans more age-neutral. Third, relying on nondiscrimination rules to provide adequate benefits for moderate-income workers is insufficient; new approaches will be · required, not only by themselves but in conjunction with any simplification efforts as well. Figure 3 illustrates one relatively modest proposal for how ERISA might look at 50, based on these propositions. At the heart of the proposal is the creation of a single, standard form of defined contribution plan. It would replace the multiple plan types now available as well as the separate, “simplified” set of plans for small employers. Such a plan is feasible because contributions to profitsharing plans no longer depend on employer profits, which was the historical rationale for maintaining separate money purchase, stock bonus, and profit-sharing plans. The plans continue to differ in terms of formula (fixed or discretionary), spousal rights, limits on deductible contributions, and benefits distribution (cash or stock). Because these are not difficult design issues, it should be relatively easy to compromise on a standard form with sufficient flexibility to be attractive to small and large employers alike. The standard plan would also provide a simplified option for employee savings. The current plans—401(k) for corporate employers, 403(b) for certain tax-exempt employers, and 457 for governmental employers—are very similar in terms of the amounts employees are permitted to save. However, they operate under very different regulatory regimes designed for different types of employers. Linking employees’ opportunities to save with the tax attributes of their employer no longer seems reasonable, in light of the need to increase personal savings for retirement. FIGURE 3. ERISA at 50 ERISA at 50 Defined Contribution Plans Defined Benefit Plans · ·· · ·· · Coordinate normal retirement age (NRA) with Social Security retirement age (SSRA) Reduce early retirement subsidies Reduce late retirement penalties Provide incentives for part-time work and phased retirement for older workers Reduce disincentives for post-NRA work Provide a minimum portable benefit for short-term workers Permit employees to purchase a larger benefit or additional years of service ·· ·· · · Single plan for all types of employers Single eligibility standard Same rules on withdrawals Same rules on portability Same deductible limit on contributions by employees and employers Same FICA-tax treatment for employee and employer contributions Government Matching Contributions for Low- and Moderate-Income Workers Individual Retirement Accounts Coordinated Limit on Employee Contributions ERISA at 50: A New Model for the Private Pension System 7 The proposal also enhances the role played by IRAs. Although IRAs were initially created to provide workers without an employer plan with some means of taxdeferred savings for retirement, they have never really developed their potential. While they are a popular vehicle for rollover contributions, they remain a marginal, underperforming producer of new defined contribution savings. Under this proposal, IRAs achieve parity with employer plans through a coordinated individual savings limit. All workers, except perhaps the highestincome workers, without an employer plan would be permitted the same level of deductible contributions as workers covered by a standard plan. Thus, we retain a simplified “discrimination” rule, but one based on total employee plus employer contributions. Contributions made to the standard plan would reduce the deductible limit under the IRA and vice versa. Canada, for example, has had a similar arrangement for many years with its IRA-equivalent Registered Retirement Savings Program (RRSP). Canadian tax law establishes an individual taxdeferred savings limit similar to the IRC § 415(c) limit. This limit is coordinated between contributions to an RRSP and any employer plan. Using this model in the United States would enable millions of Americans without an employer plan to increase substantially their savings for retirement. Defined benefit plans raise a different set of issues. Although it is unlikely that defined benefit plans will achieve renewed popularity, they will continue to exist, largely within larger corporations and the public retirement system. It is important to preserve as many of these plans as possible because they can provide important incentives to encourage people to work longer. The task here is not so much to simplify the law but to make it more flexible. The following changes are suggested: (1) coordinate normal retirement age under tax law with Social Security retirement age to gradually increase the age of normal retirement; and (2) reduce or eliminate early retirement incentives. It is also possible to adopt flexible or phased retirement rules that would enable workers to make a gradual transition into retirement with wages from part-time work supplemented by pension payments. This would require liberalizing the in-service distribution rules for pension plans, which is not a difficult task. At the same time, disincentives to work past normal retirement age could be reduced by eliminating suspension-of-benefits provisions and restructuring benefit formulas for more actuarial fairness by age. The public sector is experimenting with new arrangements such as delayed/deferred retirement option plans (DROPs), which, if successful in providing additional flexibility for employees without increasing costs for employers, could serve as models for the private sector as well. 8 T H E R E T I R E M E N T P R O J E C T Defined benefit plans could also be made more attractive to mobile workers by adding portability features such as a minimum benefit provision with cash-out features for shorter-term employees. It might also be worth exploring whether defined benefit plans could induce additional worker savings by providing an opportunity for older but shorter-service employees to purchase additional years of service for a larger benefit. The United Kingdom, for example, permits employees to make additional voluntary contributions to defined benefit plans with a formula based on final salary to purchase a larger benefit. Many government defined benefit plans in the United States currently have similar provisions, but private employer plans do not. Two additional and interrelated issues merit discussion: nondiscrimination rules and the adequacy of retirement benefits available to moderate-income workers. The nondiscrimination rules attempt to prevent employers from skewing plan benefits to high-income employees. These rules even apply to employee savings in, for example, a 401(k) plan where the amount high-income employees can save depends on how much low-income employees, on average, have saved. While serving a noble purpose, these rules are responsible for much of the complexity of contemporary benefits law. They are mostly effective in limiting the benefits of high-income workers in qualified plans who generally have access to nonqualified deferred compensation plans that replace benefits lost due to these rules. Perhaps other direct approaches to providing an adequate retirement income for moderate-income workers should be considered, even if they generate additional complexity. Despite the nondiscrimination rules, these workers do not thrive in a pension system dominated by defined contribution plans. For many, employer contributions based on a percentage of pay do not create account accumulations sufficient to fund adequate retirement income. In addition, many cannot or do not contribute to employee savings plans on their own. Defined benefit−type arrangements often provide more valuable benefits for these workers, although few get those benefits either. Creating incentives for employers to provide additional benefits in either type of plan for these workers may be a useful trade-off for relaxing or eliminating the nondiscrimination rules. In the end, these workers may need further help from the federal government to save for retirement. Several proposals to provide government contributions to stimulate retirement savings by moderate-income workers are currently being discussed in Congress. For example, the Clinton administration once advocated creating USA Accounts for this purpose, and two Social Security reform proposals, the Bipartisan Social Security Reform Act of 1999 (S. 1383) and the 21st Century Retirement Act (H.R. 1793), have similar features. Each of these plans provides federal dollars through a basic saving contribution as well as matching contributions for low- and moderate-income workers. Currently, none of these plans work well administratively (Perun 1999). In addition, it is questionable whether a new federal entitlement program on the magnitude of the USA Accounts proposal is necessary or how much federal funds should be spent in any of these programs to encourage retirement savings by young people or those with only temporarily low incomes. However, integrating some federal contribution into IRAs and employee savings plans could provide valuable incentives for additional retirement savings by moderateincome workers. For example, federal dollars could be used to provide matching grants for savings within plans that meet some minimum standard of portable benefits. For low-income workers, some defined benefit features in Social Security to provide a minimum retirement income will still be required. CODA: RETIRE ERISA AT 65 Of course, there’s no need to stop with ERISA at 50. Let’s look just another 15 years into the future. By that time, there will be just a few defined benefit plans left in America, and they will largely cover retirees. The plan to simplify defined contribution plans and to give defined benefit plans more defined contribution−like features has been such a success that even more drastic simplification is possible. It is now time to retire the ERISA-at-50 framework and think about what ERISA at 65 could look like (figure 4). In just 15 short years, it’s finally become apparent that a defined contribution plan is just a collection of individual retirement accounts. It’s also become obvious that individual defined contribution accounts are the perfectly simple retirement savings vehicle for the post-ERISA era. Defined contribution plans have become obsolete. Individual retirement savings vehicles have become a garden variety financial services industry product. Even today, the financial services industry provides almost all the services required by most employer-based defined contribution plans: off-the-shelf prototype plans, custodial services, investment options, employee communication, distribution functions, accounting, and tax reporting. Practically speaking, the employer is often just the conduit of contributions to the individual accounts held in the plan. Removing the superstructure of a plan has simplified benefits law. Individual defined contribution accounts work as well, if not better, in delivering retirement benefits to workers as the old employer plan. Of course, the tax code continues to have rules for how much employers and employees can and must contribute to these individual accounts. Employers will still have incentives to contribute on behalf of their employees and to encourage their employees to save for retirement. Some nondiscrimination rules will still be satisfied. In addition, there will be rules about appropriate investments and permissible distributions. The main distinction is that employers are happy. They no longer have to supervise or assume fiduciary responsibility for their employees’ retirement funds. The administrative burden and expense of maintaining a plan no longer exist. Employers spend more of their employee benefit dollars directly on their employees and less on the plan compliance industry. They need only send their contributions to workers’ accounts through their payroll system. FIGURE 4. ERISA at 65 ERISA at 65 The Last Defined Benefit Plans Employer Contributions Individual Defined Contribution Accounts Employee Contributions Government Contributions for Low- and ModerateIncome Workers ERISA at 50: A New Model for the Private Pension System 9 Employees are happy, too. Their own contributions go directly to a centralized retirement account that they manage as they wish. They have administrative and investment control over their retirement assets, no matter who their employer is or how many employers they have—although these managed accounts will still have restrictions on withdrawals and maximum contributions eligible for various tax benefits. Saving for retirement has become almost as easy as opening up a bank account. This is how it should be. Retirement Savings, 1999, Washington, D.C.: U.S. Department of Labor, available at http://www.dol.gov/dol/pwba. 7. For example, the Employee Benefit Research Institute reports, based on Bureau of Labor Statistics data, that 91 percent of all fulltime workers in medium and large establishments were covered by a plan (either defined benefit or defined contribution) in 1986 but only 78 percent in 1993. Parallel figures in small private establishments were 47 percent in 1992 and 42 percent in 1994. Participation rates for state and local government employees have, however, remained relatively stable. See EBRI Databook on Employee Benefits, 1997, Washington, D.C.: Employee Benefit Research Institute. 8. See also the Profit Sharing Council of America at www.psca.org. It should be noted that the definition of 401(k)-type plans used in the Department of Labor report includes more plans than the traditional 401(k)-type plans such as 401(k), 403(b), and 457(b) plans. ENDNOTES 1. This paper was prepared for the ALI-ABA pension policy conference, “ERISA after 25 Years,” Washington, D.C., October 1, 1999. We are grateful to Rudolph G. Penner for his comments. 2. For example, see Theodore R. Groom and John B. Shoven, 1999, “How the Pension System Should Be Reformed,” and Daniel I. Halperin and Alicia H. Munnell, 1999, “How the Pension System Should Be Reformed,” papers presented at the Brookings Institution/SIEPR/TIAA-CREF conference, “ERISA after 25 Years: A Framework for Evaluating Pension Reform,” Washington, D.C., September 17. The papers present almost diametrically opposed reform proposals. 3. ERISA refers generically to the U.S. pension system, both public and private. The authors recognize that the statute does not cover all plans and employers in the system. ERISA also covers certain nonretirement plans, such as health and welfare plans, which will not be discussed here. 9. About 87 percent of 401(k)-type plans are the only plan sponsored by the employer. This statistic, however, includes more plans than the traditional 401(k)-type plan under the definition used in this report. 10. Contributions by participants to defined contribution plans have been increasing over the past 10 years and in 1995 were 6 percent higher than employer contributions. 11. Information about these bills can be found at http://thomas.loc.gov. 12. See The 1999 Small Employer Retirement Survey: Building a Better Mousetrap Is Not Enough, 1999, Issue Brief No. 212, Washington, D.C.: Employee Benefit Research Institute. REFERENCES 4. A full discussion of past amendments to ERISA is beyond the scope of this paper. For a good history of pension legislation since 1974, see EBRI Databook on Employee Benefits, 1997, Washington, D.C.: Employee Benefit Research Institute, Appendix E. 5. See, for example Rudolph G. Penner, 1999, “The Coming Collapse of the U.S. Economy?” The Retirement Project, Brief No. 4, Washington, D.C.: The Urban Institute; and C. Eugene Steuerle and Christopher Spiro, 1999, “Can Spending on the Elderly Always Grow Faster Than the Economy? Should It?” The Retirement Project, Straight Talk on Social Security and Retirement Policy, No. 3, Washington, D.C.: The Urban Institute. 6. A more detailed discussion of these issues can be found in a number of recently issued reports and white papers. See, for example, Policy Challenges Posed by the Aging of America, 1998, a discussion briefing dated May 29, Washington, D.C.: The Urban Institute, available at www.urban.org; Who Will Pay for Your Retirement? The Looming Crisis, 1995, Washington, D.C.: Committee for Economic Development; Financing the Retirement of Future Generations: The Problem and Options for Change, 1998, Public Policy Monograph No. 1, Washington, D.C.: American Academy of Actuaries; The 21st Century Retirement Plan, 1999, Washington, D.C.: The Center for Strategic and International Studies; William G. Gale, 1998, Are Americans Saving Enough for Retirement? New York: The Twentieth Century Fund Foundation; and the Final Report on the National Summit on 10 T H E R E T I R E M E N T P R O J E C T Perun, Pamela. 1999. “Matching Private Savings with Federal Dollars: USA Account and Other Subsidies for Saving.” Washington, D.C.: The Urban Institute. The Retirement Project, Brief No. 8. Private Pension Plan Bulletin. 1999. Abstract of 1995 Form 5500 Annual Reports, Number 8. Washington, D.C.: U.S. Department of Labor, Pension and Welfare Benefits Administration. U.S. General Accounting Office (GAO). 1997. Retirement Income: Implications of Demographic Trends for Social Security and Pension Reform. GAO/HEHS-97-81. Washington, D.C.: GAO. Appendix SUMMARY OF SELECTED RULES CIRCA 2000 12 Appendix. Summary of Selected Rules circa 2000: Part 1 T H E IRC § 403 Arrangements IRC § 401(a) Plans R E T I R E M E N T Defined Benefit Eligible employer Money Purchase Any employer Profit Sharing or Stock Bonus & Standard 401(k) Profit Sharing or Stock Bonus & SIMPLE 401(k) Other Profit Sharing or Stock Bonus without 401(k) Employee Stock Ownership Plan IRC § 403(b) Any employer except state and local governments 401(k)-eligible employer with <100 employees and no other plan Any employer Corporate employer IRC § 501(c)(3) organizations and public schools P R O J E C T Lesser of $30,000 or Annual benefit 25% of pay or limit = lesser of exclusion allowance $135,000 or 100% Per person annual limit = lesser of Per person annual Per person annual limit = lesser of $30,000 Overall limits 1 1 $30,000 or 25% of pay limit = 401(k) + match or 25% of pay (20% of pay × years of × high 3 years’ pay service – employer (not governmental contribution)1 plans) Pay limit $170,000 Normal cost + past service liability Match up to 3% or Amount required Annual funding amortization over NA fixed 2% of pay NA, usually by plan formula 10 years or 155% contribution required of current liability Employee limits NA NA $10,500 $6,000 Employer deduction limits Based on funding requirement Based on funding requirement 15% of taxable pay Greater of 15% of taxable pay or required contribution Exclusion from SS tax Yes, Yes, contributions contributions and distributions and distributions 10% early withdrawal tax Sometimes Yes No 401(k), yes No 401(k), yes other other contributions and contributions distributions and distributions Yes Yes NA NA 25% leveraged/15% 15% of taxable nonleveraged of taxable pay pay + dividends paid out + interest on loan $10,500 NA Yes Yes No deferrals, yes employer contributions and distributions Yes Yes Yes IRC § 403 Arrangements IRC § 401(a) Plans Defined Benefit In-service withdrawals ERISA at 50: A New Model for the Private Pension System Nondiscrimination rules (not governmental plans) Integrated with Social Security Money Purchase 3 General rule May be Spousal protection Survival annuity, consent and death benefit rights Vesting Deferred Special requirements Profit Sharing or Stock Bonus & SIMPLE 401(k) Financial hardship,2 minimum 2-year holding period (employer), loans Not allowed May be Profit Sharing or Stock Bonus & Standard 401(k) PBGC guarantee and Minimum funding premium of $19 per required in full participant each year ADP, ACP, safe harbors and general rule4 NA May be (employer) No Other Profit Sharing or Stock Bonus without 401(k) Employee Stock Ownership Plan Minimum 2-year holding period, loans Financial hardship,2 loans ACP for match, availability test for deferrals and general rule4 3 General rule May be IRC § 403(b) No May be Only death benefit usually5 401(k) immediate; others deferred Immediate None Deferred Deferrals immediate; others deferred Forfeitures/interest raise annual limit if ≤ 1/3 Catch-up contributions contributions for HCEs. permitted with 15+ 100% employer securities years of service. May allowed. Diversification also be a DB plan. optional at 55. Put option/voting rights. 1. The $30,000 overall dollar limit is a cumulative limit for employers and employees across all qualified plans. 2. Financial hardship is an immediate and heavy financial need, even if foreseeable or voluntarily incurred, not satisfiable by other resources. 3. No HCE may receive a contribution or benefit of a higher percentage of pay than any NHCE. 4. Both the Actual Deferral Percentage (ADP) test for 401(k) contributions and the Average Contribution Percentage (ACP) test for matching and after-tax contributions are designed to limit contributions for HCEs (highly compensated employees) based on the average contributions for NHCEs (non–highly compensated employees). 5. The surviving spouse receives the account balance as a death benefit, usually in a lump sum. 13 14 T H E Appendix. Summary of Selected Rules circa 2000: Part 2 R E T I R E M E N T IRC § 408, 408A IRAs Traditional IRA Roth IRA Nonqualified Deferred Compensation Plans SEP-IRA Earnings less than $110,000 for Employees of all individuals and employers $160,000 for couples SIMPLE IRA Eligible 457(b) plans Executive Arrangements Employees of Employees of state and local Select group of officers employers with no government and officers of tax-exempt or highly compensated other plan and organizations employees <100 employees P R O J E C T Eligibility Anyone Dollar limit $2,000 for all IRAs, deductible if no employer plan or income less than $42,000 for individuals and $62,000 for couples $2,000 for all IRAs $30,000 NA $8,000 − contributions to 403(b), SIMPLE, and 401(k) plans None Maximum % of pay 100% 100% 15% NA 33 1/3% NA Employer limits NA NA Lesser of $30,000 or 15% of pay Match of up to 3% or fixed 2% of pay Lesser of $8,000 as reduced above or 33 1/3% None $2,000 $2,000 NA $6,000 Lesser of $8,000 as reduced above or 33 1/3% None NA NA 15% of aggregate pay None NA None Yes on contribution and distribution No on employee; yes on employer contribution and distribution No on employee contribution; yes on distribution No (except after vesting) Yes, increased to 25% in first 2 years No No (unless annuity purchased) Employee limits Employer deduction limits Exclusion from SS tax 10% early withdrawal tax No on contribution; yes on distribution Yes Maybe Yes IRC § 408, 408A IRAs Traditional IRA Penalty tax exceptions Withdrawals Roth IRA Medical, first home purchase and higher education expenses, health insurance payments for unemployed Yes SEP-IRA First home purchase Same as (up to $10,000) traditional IRA 5-year waiting period Loans available ERISA at 50: A New Model for the Private Pension System 15 Nondiscrimination rules Nonqualified Deferred Compensation Plans Yes SIMPLE IRA Eligible 457(b) plans Same as traditional IRA NA Yes Unforeseeable emergency only while employed Yes Unclear Yes Can favor HCEs None No None None Uniform percent of pay contribution Required match or contribution None (unless annuity purchased) Pay limit See above $170,000 $170,000 for 2% contribution NA Integrated with Social Security NA May be No NA Spousal protection None Vesting Special restrictions and benefits Executive Arrangements Immediate None After-tax contributions only Employer does not have to contribute every year Employees generally responsible for investments Immediate for employee; deferred Immediate usually for employer $15,000 catch-up limit for 3 years Taxed when paid or before retirement. Unfunded plan made available (or but trust requirement for public- when vested for taxsector plans. May also be a DB exempts). plan. May be DC or DB. Note: All dollar values are indexed, and the given figures are as of 2000. Sources: Harry Conaway, William M. Mercer and C. Eugene Steuerle, Andrea Barnett and Pamela Perun, the Urban Institute. T H E R E T I R E M E N T P R O J E C T Occasional Paper Number 4 URBAN INSTITUTE 2100 M Street, N.W. Washington, DC 20037 Phone: 202.833.7200 Fax: 202.429.0687 e-mail: paffairs@ui.urban.org http://www.urban.org Nonprofit Org. U.S. Postage PAID Permit No. 8098 Washington, DC