No. 16 in Series, "The Retirement Project"
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.
Note: This report is available in its entirety in the Portable Document Format (PDF).
Within 10 years, baby boomers will begin retiring in large numbers. The United States will lose the services of millions of highly skilled, experienced workers. Because of the baby dearth that followed the baby boom, there will not be many new workers to replace the seniors leaving the labor force. Labor force growth is expected to fall from 1.1 percent per year in the 1990s to 0.36 percent per year in the period 2010 to 2020.
The problem afflicts some professions more than others. The population of nurses and teachers is aging at a particularly rapid rate. In contrast, the supply of blue-collar workers will be much less affected, partly because people tend to work their way into white-collar jobs as they age.
The situation could obviously be improved if older workers could be induced to work longer. This should be feasible as improved health has accompanied longer longevity and as jobs that require hard physical labor have declined in relative importance. Allowing older workers more flexible work arrangements, such as shorter hours and longer vacations, could enhance the attractiveness of working longer.
Unfortunately, there are a large number of economic, legal, and institutional barriers to more flexible employment arrangements for older workers. These were not considered a problem when there was a positive desire to move people out of the labor force early to make room for the horde of baby boomers working its way up the career ladder. While it was usually not appropriate to erect these barriers then, the imminent, drastic change in demographic conditions now creates a much more urgent need for reform.
It would be foolish to believe that any reforms could completely counter the projected decline in labor force growth. Our own projections suggest that to eliminate the decline by 2030, the labor force participation rate among the 55 to 64 age group would have to rise to a level only slightly lower than that for the 25 to 49 age group (Penner, Perun, and Steuerle 2002). Large increases for workers over age 65 are also implausible, because many older workers very much want to retire and some are forced to retire for health reasons. Of all older workers leaving their jobs between 1992 and 2000, only about 13 percent, or 1.5 million, say that they would have stayed on their jobs longer if the employer had offered them fewer hours with commensurately less pay. However, others, now working part-time might increase their hours if it were easier legally and institutionally. Legal barriers are particularly difficult to overcome for individuals working more than 1,000 hours per year.
Although labor force growth will slow regardless of changes in public and private policies, reforms could noticeably ease those macroeconomic problems associated with an aging labor force: a slower rate of economic and labor force growth, more transfer payments from government, and fewer taxes collected. But this is not the only reason for reform. The welfare of individual older workers can be increased significantly by giving them a greater variety of employment opportunities. More work early on also leaves more resources for later retirement years. These are sufficient reasons for reform, even if the macroeconomic effects turn out to be modest.
Institutional Barriers to Encouraging Longer Work
Because the number of those age 55 and over will grow remarkably as the baby boomers age, healthy retirees with 10 to 30 years of life expectancy are fast becoming the largest untapped source of potential labor in the economy. Yet they continue to face disincentives to work, due to an archaic employee compensation system designed when the goal was to get rid of long- term, high-cost workers and replace them with relatively less expensive younger workers.
An employer has an incentive to let a worker go if the worker's total compensation is above the average paid to other workers for the same services. When a worker is paid less than the going rate of compensation, on the other hand, then he or she has an incentive to leave the firm. Such compensation differentials often arise from the ways that employee benefit plans are structured.
One source of this disparity comes from traditional defined benefit plans. These plans typically provide benefits based on a formula consisting of number of years of service times some fixed percentage times salary for the highest years (often five or three), up to some maximum percentage of that salary. For example, a worker might be granted 1 percent of pay times years of service times salary in the highest five years, up to a maximum of 30 percent. However, payments usually won't commence until either age 65 or the attainment of many years of service (such as 30 years) with the employer.
The economic value of the annual benefit accrual in these pension plans can be calculated as the change in the present value of all future pension benefits for staying on the job one more year. During the accrual stage, an additional year on the job increases pension benefits not only because it adds an additional year of pay but also because the pay on which the benefits are based is likely to grow with inflation and real wage increases. This creates a compounding effect. On the other hand, at certain points (e.g., normal retirement age or early eligibility age due to attaining some maximum number of years of creditable service), one more year of work implies that the pension is received one year less. Furthermore, after reaching some maximum percentage of salary to be replaced, future annual benefit increases are limited. Consequently, accruals often turn negative.
Our examination of some 340 private pension plans (including most of the largest employers in the country) finds that on average the benefits they provide by age resemble a hill.1 The present value of the pension accrual rises along a slope that becomes increasingly steep as the worker gains in age and time on the job and then, after peaking, suddenly falls dramatically. Peak accrual typically occurs after about 30 years; for example, at age 55 for a worker who starts at age 25. For this worker, staying on the job for additional years beyond the peak can result in negative pension accruals—for example, an average of –13.9 percent from ages 60 to 65.
Partly to get around these problems, the private sector has been moving steadily away from a traditional defined benefit structure—either by adopting defined contribution plans or converting traditional defined benefit plans to so-called cash balance plans. Both defined contribution and cash balance plans usually base retirement plan contributions on a relatively constant percentage of salary, regardless of age. However, many traditional defined benefit plans remain, especially in the public sector. In that sector, the incentive to retire early can be even stronger than in the private sector. We found that pension accruals first rise especially quickly with length of service and then fall more rapidly for the federal employee retirement system (FERS) and for California, Illinois, and New York teachers. For instance, the New York teacher who starts at age 25 gets pension accruals of +6.4 percent between ages 36 and 40, +62.9 percent between ages 51 and 55, and –39.7 percent between ages 61 and 65.
Health benefits also vary in value by age. A 1996 panel study revealed that private health costs rose from between $500 to $1,000 for workers age 20–40 to close to $2,000 for most workers over age 55. For employees who stay with a job, health costs rise as a proportion of pay mainly when they come close to old age rather than in middle age when both pay and health costs are going up together. For longer-term workers, the age disparities in employee benefits as a proportion of pay are often less severe for health insurance than for many traditional defined benefit pension plans.
When people try to take a new job (as opposed to staying on the same job), age disparities in pay are again present—making it more difficult for older workers to find new jobs and stay in the labor force. Take a set of workers who start a new job and stay for five years. In a typical defined benefit plan, on average someone who works from ages 25 to 30 accrues about 2.1 percent of pay in pension benefits, but the accrual rate rises to 8 to 10 percent of pay for those between the ages of 50 and 65. When it comes to health insurance, the difference in health costs by age now comes into play in a more pronounced way than when older workers were assumed to have increased productivity because of more time on the job. Here the large rise in health costs from younger to older workers makes the older employee much more expensive in terms of health insurance benefits without any necessary productivity offset.
One clear-cut tax on older workers is a result of the federal government's requirement that Medicare be a secondary payer in cases where the employer offers health insurance to other employees. The older employee in such a firm has to give up several thousand dollars worth of Medicare benefits every year just to work past age 65. An employee might get around the tax by taking a job with an employer offering no health insurance.
These various disincentives to work longer, of course, are not entirely independent of those in the Social Security system. Now that some Social Security changes are under way—such as the increase in the normal retirement age—and others are inevitable given its funding crisis, much more attention must be given to the combined disincentives to work created by all these systems.
Making Phased Retirement a Routine Employee Benefit
Inducing older workers to work longer would be more feasible if phased retirement programs were a routine employee benefit. Such programs would permit workers to make a gradual— rather than an abrupt—transition from work to retirement, and provide them with an opportunity to work longer while working less.
Phased retirement programs are frequently available to state and local government workers and tenured faculty in higher education. But they are rare today in the private sector. Many employers express interest in phased retirement but only a small minority try to implement it. There is no uniform model. Many employers prefer to make individual arrangements to retain employees with specialized skills and expertise. Others offer reduced hours or work schedules to a larger group. But most private employers do not have any phased retirement option for current employees. Instead, if they have any program at all, they prefer to rehire previously retired workers for part-time and temporary work whether or not they once worked for the same firm.
To expand the reach of phased retirement programs will require a 180 degree shift in traditional benefits thinking. For decades, employers have looked for benefits packages to ease older workers out of the workforce. Phased retirement programs, however, have very different design needs. In order to facilitate a gradual transition to full retirement through adjusted work hours and responsibilities, they should permit flexible compensation and benefits structures while providing employers with reasonable and predictable costs. They should also impose minimal administrative responsibilities, and provide legal protection from age discrimination claims. Sufficient information on the details of a plan should be provided to enable workers to make informed decisions about participation and the plan must maintain current law protections for older workers, especially for those who work out of financial necessity.
Legal Restrictions on Phased Retirement Plans
These objectives are difficult to achieve today. Employers who offer benefit programs must comply with the rules of three complex statutes—the tax code, the Employee Retirement Income Security Act of 1974 (ERISA) and the Age Discrimination in Employment Act (ADEA). Each has features that create a design nightmare and make phased retirement programs unappealing. Both ERISA and the tax code are complex and inflexible statutes. They set rules on who can and must participate in a plan, on the amount and types of benefits that can be paid, on when benefits can be paid, and on how those benefits will be taxed.
For example, hiring retirees for part-time and temporary work seems like an obvious option because it avoids many benefit complications, but it is straightforward only if the retirees are from other companies. In addition, employers cannot pay pensions from defined benefit plans to their workers before they terminate employment or attain normal retirement age. But many workers would count on these payments to supplement pay from part-time work in a phased retirement arrangement. So many employers use a strategy of rehiring workers shortly after they "retire" in an attempt to satisfy this rule. This is a risky strategy because tax regulations have never specified how long an employee must separate to make his or her "retirement" legitimate. But severe penalties are imposed for violating the law. The tax-qualified status of the plan could be jeopardized for everyone. Hiring retirees as consultants is another risky strategy. Many of these retirees do not fit the tax code definition for an independent contractor, and misclassification of employees as independent contractors has recently become a high-profile legal issue. Companies that do misclassify employees risk not only regulatory action by the IRS but law suits by their own employees as well.
The ADEA, which forbids employers from discriminating against workers age 40 and older, is even more problematic. Its impact on employee benefit plans is just beginning to be fleshed out by the courts. Until there is more guidance on the extent to which benefit plans that satisfy the tax code and ERISA must be changed to comply with ADEA, employers will be reluctant to adopt phased retirement plans, largely because of their legal exposure.
Phased retirement programs today may pose some dilemmas for workers too. Because a phased retirement option is usually an ad hoc arrangement, many will find that working parttime has significant drawbacks. They may find that part-time work significantly reduces their pension benefits. They may also lose all or part of other employee benefits, such as health insurance, life insurance, and disability insurance. Age-based restrictions on their ability to receive pension payments plus extra tax penalties make a phased retirement program with their current employer an unattractive option. Many who might otherwise prefer to remain at the same job find it easier to negotiate flexible work arrangements with a new employer.
Several regulatory changes and statutory amendments might make phased retirement programs more appealing to both employers and workers. For example, the IRS could issue rules about when a bona fide termination of employment occurs so employers would know when pension payments could safely continue to returning workers. In addition, the IRS could also clarify whether workers who switch to parttime work as they near retirement lose a portion of their pension benefits that are attributable to their earlier full-time pay.
Many have argued for new laws to lower the age at which workers could start their pension payments while continuing to work for their current employer. Others have suggested giving workers with 401(k) plans the ability to access their own funds before the current threshold age of 59 1/2. These changes would provide part-time workers with supplemental pay from part–time work but they are controversial and not necessarily good pension policy. The pension system already has a problem with "leakage," that is, with retirement assets being consumed prior to actual retirement, and these changes would only exacerbate that problem. One less controversial change is to give workers who work after normal retirement age more control over when their pension benefits begin. Many defined benefit plans don't permit payments until workers actually retire and some even require workers to forfeit their earned benefits while they continue to work. Giving workers who have reached normal retirement age, rather than employers, control over the timing of their benefits and ensuring they are not penalized financially for continuing to work seems a reasonable trade off for longer work. Finally, giving employees more information about the value of working additional years as they draw near retirement age and the effect of entering a phased retirement program on their benefits seems fair and noncontroversial.
Such changes would eliminate some of the current barriers to phased retirement programs but it is difficult to argue that they would have more than a marginal effect. Most of the legal complexities and ambiguities that plague such programs today would still exist. The reality is that a special statute amending the tax code, ERISA and the ADEA to authorize phased retirement programs will probably be required before they can become a routine employee benefit program. This is not that difficult to do, and there is a precedent for such a statute. About a decade ago, early retirement programs faced similar legal obstacles. A compromise solution was reached, all three statutes were amended, and now these programs are relatively routine. In terms of phased retirement programs, the goal of such a statute would be to
- protect employers against age discrimination claims under the ADEA,
- minimize the costs and administrative burden of phased retirement programs,
- set standards for model benefit packages in return for some relief from the plan compliance rules,
- permit short-term plans and flexible compensation arrangements,
- provide flexibility on eligibility criteria, including age and service standards, and
- set full disclosure and informed consent standards to protect employees.
There are many ways in which such a statute could be structured, and no single design or feature is pivotal. The most important consideration to be kept in mind is that this is an opportunity to import some flexibility and creativity into the rigid and overly complex world of benefits law. Any statute should be capable of stimulating the creation of some innovative plan designs for private sector employers. The deferred option retirement plans (DROPs) now common among state and local government employers might serve as a model. It should also establish some parameters for safe harbor plans that employers could adopt without assuming the burden of expensive administrative requirements. In addition, such a statute might allow late retirement benefits to be added to defined benefit plans to provide incentives for continued work. Another important contribution would be to authorize special benefits packages just for phased retirees that would not be subject to the current cumbersome rules for nondiscrimination and coverage testing.
In many respects, the legal and regulatory problems facing workers in phased retirement programs reflect more their status as part-time workers than as older workers. Phased retirement programs raise benefits issues that are common to all types of flexible work arrangements. By working through these issues for phased retirees, the development of phased retirement programs—whether by enacting a special statute or adding flexibility to existing law—could serve as a model for adapting various laws to meet the needs of the twenty-first century workforce.
In addition to enacting legislation modifying tax laws, ERISA, and ADEA, private employment practices must be reexamined. More flexible compensation structures could remove some of the disincentives for work at older ages. While defined contribution and cash balance types of plans—essentially plans with retirement benefits usually equal to a percent of cash compensation each year—get around some problems, many traditional defined benefit plans (especially in the public sector) are likely to be around for a long time. In the future, these plans might be redesigned to add features letting workers at older ages shift into a structure where benefits are in line with compensation—not so high that the employer won't hire them or so low that older workers suffer economic discrimination relative to other workers. Similarly, employers should be given greater flexibility in limiting any rise in health costs simply because they hire older workers. Also important is removal of the requirement that Medicare serve as a secondary payer—in many firms, a significant barrier to retaining or hiring workers over age 65.
Notes
The authors are grateful to the Alfred P. Sloan Foundation for financial assistance. This brief is a summary of a longer report (Penner, Perun, and Steuerle, 2002) available at http://www.urban.org/url.cfm?ID=410587. The views expressed are those of the authors and do not necessarily reflect those of the Urban Institute, its board, its sponsors, or other authors in the series.
1. The data base was salary-based traditional defined benefit plans in the Pension Insurance Modeling System (PIMS) developed by the Pension Benefit Guaranty Corporation.
About the Authors
Rudolph G. Penner is a senior fellow at the Urban Institute, where he holds the Arjay and Frances Miller chair in public policy.
Pamela Perun is a consultant to the Retirement Project at the Urban Institute.
Eugene Steuerle is a senior fellow at the Urban Institute and former chair of the Social Security Administration's technical panel. His work includes several books on Social Security, as well as a series, Straight Talk on Social Security and Retirement Policy.
The Retirement Project
The Retirement Project is a research effort that addresses how current and proposed retirement policies, demographic trends, and private-sector practices affect the well-being of older individuals, the economy, and government budgets.
Note: This report is available in its entirety in the Portable Document Format (PDF).
Latest Research
PDF
Print this page