A Markov Process Model of Work-Life Expectancies for Ages 66-75 Based on Labor Force Activity in 1997-98

In a recent paper in this Journal, James Ciecka, Thomas Donley, and Jerry Goldman (2000) published work-life expectancies based on a Markov process model using labor force data for 1997-98. At the time work was done on that paper, work-lives were calculated to age 75; but the published paper contained work-life expectancies only to age 65. The rationale for not publishing results for ages 66-75 was that work-life tables become more sensitive to table closing assumptions at later ages. Closing assumptions entail (1) using participation rates instead of transition probabilities to measure labor market activity between ages 76-80, thereby abandoning the Markov process for those ages and (2) assuming no labor market activity after age 80. Since publishing the foregoing paper, the authors have received several requests for work-lives between ages 66-75. This paper responds to those requests. The methodology, data, and table closing procedures are exactly as specified in our earlier paper and will not be repeated here. There are 14 work-life tables in this paper; each is an extension of a corresponding table in the earlier paper. The final two tables in this paper give the reader information about the affect that labor force participation rates and methodological assumptions (for ages 76-80) have on work-lives between ages 60-75