While increases in life expectancy are the hoped-for answer to the ancient prayer for a long life, they pose a challenge to retirement income policy. How should retirement income policy adjust to longer life expectancy? Over the long term, with fertility rates stabilized at low levels, this is the primary question in retirement income policy. The papers by Sharon Hermes on phased retirement and by Kevin Neuman and Daniel Lawson on the distribution of years of retirement leisure both have implications for a possible policy response to increased longevity.

With increasing life expectancy placing financial pressure on the Social Security system, but also providing more years in retirement when retirement age is fixed, some countries have raised the social security early retirement age (ERA), which is now legislated to be age sixty-five or higher in the majority of Organization for Economic Co-operation and Development (OECD) countries (Turner 2005). If the Social Security ERA were raised from its current level of age sixty-two in the United States, policy could be developed to make work at older ages more flxible, such as thr more flexible, such as through phased retirement. If the ERA were raised, determine the characteristics of workers adversely affected when assessing the effects of such a possible change. These two papers address these issues.

Hermes

The data collection and paper by Hermes greatly expand our knowledge concerning the prevalence and characteristics of phased retirement programs. The key finding is that many firms claim to offer phased retirement on an informal but selective basis. Thus, the possibilities for phased retirement may be more prevalent than it would appear based on the limited availability of formal phased retirement plans. Within workplaces, however, the availability of phased retirement is limited to workers with certain character-istics—such as high performers and nonsupervisors—rather than being generally available.

A surprising finding is that the presence of a defined benefit (DB) pension plan did not make the offer of phased retirement less likely. However, the paper does not assess the attractiveness to workers of the phased retirement offer. The offer may not be attractive if it means that workers would take a penalty on their DB plan benefits. This penalty would occur in many DB plans because benefits are based on the average of the highest three or five years of pay, which are generally the worker's last years unless the worker took phased retirement, in which case it would be earlier years. Those earlier years, and the pension benefits derived from them, would be worth less in real terms at retirement because they would enter in the benefit calculation at their nominal value rather than being indexed for inflation to the point of retirement.

Even taking into account the effect on pension benefits of increased service, workers with twenty-five years or more of service could find that the real value of their annual pension benefit declined by taking phased retirement and postponing initial benefit receipt. The lifetime value of their benefits would decline even more than the annual value because of the postponement of retirement. That problem could be dealt with by using the full-time equivalent salary, rather than the part-time salary, to determine which years would be the high five years, but that is not done by firms.

In 2005 the Treasury Department proposed regulations that would permit employees to receive a prorated share of a pension if they had attained a certain age and reduced their hours from full- to part-time. Current law permits employees to receive a defined benefit pension while still working for the employer providing the pension if they have attained the normal retirement age in the plan, which typically is age sixty-five. The Treasury Department regulation would allow pension receipt while working to occur at younger ages in the context of phased retirement. The Treasury Department has not finalized these regulations. The results of Hermes's paper suggest these regulations may have little effect because of the limited availability of formal phased retirement and the selective availability within firms of informal phased retirement. To the extent that DB plans have decreased the desirability of phased retirement programs, that effect is declining over time with the decrease in the number of these plans.

This paper suggests that if Congress raised the ERA for Social Security, some people would be able to adjust by working reduced hours before Social Security benefits were available through phased retirement offered by their employer. The paper indicates job characteristics of such workers. It would be useful if a follow-up paper could assess what percentage of the workforce likely would have the option of phased retirement with their current employer.

Neuman and Lawson

Neuman and Lawson investigate the number of years people of different economic and demographic characteristics spend in retirement. Their study is based on the insight that raising the ERA would most adversely affect those people who spend the fewest years in retirement. That insight holds for people who have fewer years in retirement because of short life expectancy. Previous discussions of the issue have tended to focus on those who retire early because they would be directly affected by an increase in the ERA.

They note that though people who retire early would be affected by raising the ERA, for some of them the concern would be limited because they spend more years in retirement than is typical. However, people who retire early do not necessarily spend the most years in retirement because they may die earlier than other people. If people with shorter life expectancies tended to retire earlier, they would be the group most adversely affected by a policy that raised the ERA in Social Security. However, the opposite would be the case if wealthier people retired earlier and had longer life expectancies than other economic groups. Studies that examine retirement age and mortality separately do not provide the information that is needed to adequately assess the effects of a policy to raise the ERA in Social Security.

These authors find that workers taking early retirement include people with different life expectancies. Women tend to retire earlier and live longer than men. Less healthy people retire earlier, but their expected length of time in retirement is the same as healthy people. Thus, they would be adversely affected by raising the ERA. Individuals who are Hispanic, in good health, or have a defined contribution (DC) plan retire later but do not live longer than others.

In considering policy to raise the ERA, interpersonal comparisons can be made at a point in time or across time. This paper provides evidence concerning the distribution of the consumption of retirement leisure in the cross section at a point in time. However, over time there have been major increases in life expectancy at older ages, and further increases are projected. The average age at retirement has been constant, or perhaps risen slightly, since the early 1980s. Thus, over time, the number of years of retirement leisure has increased. If the ERA were raised, it would not reduce retirement leisure on average relative to people retiring one or two decades earlier, but some analysts may question whether that intergenerational comparison is relevant because of increasing lifetime wealth raising the demand for retirement leisure.

The Neuman and Lawson study finds that the groups who consume less retirement leisure than others are men, Hispanics, people in good health, and people participating in DC plans. These groups are not generally considered to be particularly vulnerable to the costs of a policy change. This is not the case, however, for low-skilled white collar workers, who also have shorter expected years of retirement. This group, therefore, would be adversely affected by a policy to raise retirement ages.

Author's address: 601 E. St. NW, Washington, DC 20049

Note

This material is the responsibility of the author and does not necessarily represent the positions of the AARP.