Retirement and Work, Work and Retirement: How Do We Know?

John Scott, University of Maryland

Like most other countries in the world, the population of the United States is aging due to the trends of declining fertility and increasing longevity. As a result of the aging of the population, the workforce is also aging. While in 1990 the percentage of those 65 and older was 11.8 percent, by 2015 nearly 15 percent of the American labor force will consist of workers aged 65 and older. Labor force participation rates for older workers have apparently stopped declining, and there may be a new trend of increasing participation by older workers. While continued employment by older employees appears to be on the increase, such workers do not necessarily wish to work full-time. In fact, there is evidence that alternative work arrangements are being utilized by employers in order to meet the needs of older workers. Some employees are able to modify full-time status in some fashion in order to “phase down” the employment as they approach full retirement. Workers who cannot participate in such arrangements through their career employer often “retire” and find part-time work with a different employer. There is not one definition of phased retirement (also known as partial or gradual retirement), but instead the term indicates a process whereby one reduces their active working life in favor of increased amount of time spent in leisure or retirement. Phased retirement implies a blending of work and leisure as well as a transition to retirement that stands in contrast to the abrupt and complete retirement at a specified age. Such phased, partial, or gradual retirement arrangements are today relatively infrequent as formal or broad-based programs, but there is some evidence that more employers may implement such programs in the future. The difficulty is that phased retirement could be expressed in terms of one’s own perception, hours of work, earnings, and change in employers or reductions in responsibilities. As these definitions change, the number of individuals meeting the definition could also change in a meaningful way. This paper will examine the alternative conceptions of phased retirement. This paper will apply alternative definitions of phased retirement to a sample of older workers to illustrate changes in phased retirees. In addition, this paper will attempt to blend objective measures of work/retirement status with qualitative measures of retirement such as self-reported status. Finally, this composite definition will be refined to include changes in job title in order to reflect possible de-skilling as workers prepare to make the transition to full retirement. This study is based on two waves – 1992 and 1996 – of the Health and Retirement Survey (HRS). The HRS is a nationally representative sample of persons aged 51 to 61 in 1992 and their spouses or partners. The survey is longitudinal in nature, with the baseline interview conducted in 1992 and subsequent waves every two years, and it collects extensive information regarding employment, pension, health, family structure, and income and wealth characteristics of age-eligible respondents and their spouses or partners. Because this study will examine the retirement patterns for those working at the survey's onset, the study restricts the analysis to age-eligible respondents who are full-time employees working more than 1,599 hours per year in 1992. Of the total respondents, 8,003 responded that they were currently working. Both self-employed persons and respondents working for someone else are included. The definition of full-time status is based on the report of hours worked per week in the respondent’s current job multiplied times the number of weeks of work in the prior year. With these definitions, the dataset consists of 6,371 observations. This paper first will apply an objective measure of phased retirement in terms of the change in annual hours worked from the first wave of the survey in 1992 to 1996. As the literature review makes clear, there could be other objective measures of changes in work. Earnings is certainly one such indicator. However, it is this paper's contention that hours are more important than earnings in terms of observing a change to phased retirement status. Earnings do not adequately capture the interplay between work and leisure, and earnings can be subject to variation that has nothing to do with one’s desire for blending work and leisure in old age. In this paper, then, the objective measure will be based on annual hours worked. Classification will be based on the percentage change in annual hours worked between the 1992 and 1996 waves. If there is no change in the number of hours worked annually or even an increase, the respondent will be classified as ‘not retired.’ If there is a decline of at least 20 percent in annual hours worked but the respondent is still reporting some hours of work in 1996, then he or she will be classified as ‘partially retired.’ Finally, if the decline is down to 0 hours of work in 1996, that person is ‘completely retired.’ The subjective measure of self-reported status of the sample also will be examined. Results of retirement outcomes will be shown purely from the respondents’ classification of themselves as ‘completely retired,’ ‘partially retired’ or ‘not retired.’ These two measures will then be combined to form a composite definition of phased retirement in order to blend outward changes in workforce behavior with the individual’s intention and perception (‘Composite 1’). Finally, this paper will see whether persons who are classified as not retired under the composite definition could be reclassified as phased retirees given any changes in their job titles over the four year period (‘Composite 2’). Occupations are classified according to status (white collar and blue collar) and skill level (highly skilled and other). If a person is considered to be Not Retired but their job title has changed categories in a way that evidences a loss of skill or status level between 1992 and 1996, they were reclassified as a phased retiree. This paper finds that different definitions can create a great deal of variability as to who is considered to be partially retired. This variability is due to the conflict between people’s perceptions and/or desires and the outward indicators of their work/retirement status. A narrow definitions that focus on one indicator are likely to miss some activity or a perspective on retirement status. An additional conclusion is that a composite definition providing for both objective measures of retirement status and subjective measures of the respondent’s perceptions or intentions is preferred. Moreover, such composite definitions should include aspects of deskilling that could occur in a phased retirement state. Finally, even a composite definition of retirement status is likely to be incomplete. This analysis calls for more categories beyond the ones considered here in order to reflect desired states and temporal trends such as ‘involuntary full retirement’ or ‘steadily partially retired.’ While perhaps impractical or awkward, such constructions would help identify underlying social issues and trends and perhaps some of the variables that shape such issues and trends. Such new measures are important because retirement is likely to become more heterogeneous as constraints and institutions change. More aging people and a more varied population will aid such heterogeneity in outcomes, and our measures of outcomes have to maintain pace with these changes in order for social research to remain relevant.