EUROPEAN PAPERS ON THE NEW WELFARE

Phased Retirement: Who Opts for It and Toward What End?

Gustman and Steinmeier conducted one of the first empirical investigations into what they termed ‘partial retirement’. Using the first four waves of the Social Security Administration’s Retirement History Study (RHS), a longitudinal survey of men aged 58 to 63 when initially surveyed in 1969, they showed that a dichotomous outcome (retired, not retired) was not appropriate for predicting retirement behavior. The sample was limited to white males who were not self-employed in their main jobs, and for data purposes, ‘main job’ was the full-time job held at age 55. The study showed that about 3% of workers not facing mandatory retirement were partially retired in their main jobs while 11% were partially retired outside of their main jobs. Moreover, partial retirement increased with age. For example, in the 1975 wave, the percentage of the sample reporting partial retirement increased monotonically from 23.5 for those aged 64 to 38 at age 69 (Gustman and Steinmeier, 1984).
In 2000, Gustman and Steinmeier reexamined partial retirement using the self-reported definition of partial retirement, but this time using a different dataset, the Health and Retirement Study (HRS). In 1992, 6.3% of respondents reported themselves as partially retired, but that number rose to 12.7% by 1998.
Examining earnings is another method of defining partial retirement. Honig and Hanoch (1985), using the first three waves of the RHS, based ‘partial retirement’ on the ratio of an individual’s current earnings to maximum earnings over the entire career. If the ratio was 0.5 or less, the person was partially retired. Under this definition, nearly 20% of the sample was partially retired. Of those classified by the authors as partially retired, 39% considered themselves to be fully retired, 43% reported their status as partially retired, and the remaining 18% did not consider themselves retired at all.
Some studies have defined partial retirement by the number of hours worked. For example, Haider and Loughran (2001), using data from the Current Population Survey, found that 22% of those aged 50 to 58 and 31% of those aged 59 to 61 worked part time (less than 1,750 hours annually) from 1996 through 1998. Gustman and Steinmeier (2000) used two hours-based measures of partial retirement: usual hours worked per week (1-24 hours per week indicating partial retirement) and usual hours worked per year (1-1,199 hours per year indicating partial retirement). The usual hours per week measure found that 7% of respondents were partially retired in 1992 with an increase to 9.3% by 1998, and the use of an annual hours measure resulted in slightly higher percentages, ranging from 8.1% in 1992 to 10.6% in 1998.

Gustman and Steinmeier’s 2000 study also used job tenure to measure partial retirement. Partial retirement was defined as leaving a long-term job (long term meaning at least 10 years of tenure) for a new job. The measure of leaving a job of 10+ years resulted in 24% being classified as partially retired over the four waves of the HRS, and the definition of leaving a job of 20+ years found an average of 21% partially retired.
Combining earnings and self-reported retirement status in his study of the RHS, Ruhm (1990) was concerned that involuntary reductions in hours or wages might cause an erroneous classification of partial or full retirement. Approximately half of all workers under this definition were partially retired at some point in their lifetimes, but only about 6.2% partially retired from their career job. Ruhm also focused on the duration of partial retirement, finding that the average duration (from onset of partial retirement to full retirement) exceeded five years.

Table 1: Comparison of studies of partial retirement: definitions and findings
bing-tab-1.gif
Source: Authors’ compilation, 2005.


Pages: 1 2 3 4 5 6 7


Tags: , , , ,