Changes in retirement decisions: Determinants of plans and timing

Purpose. This study examined dynamic factors in a worker's background, resources, jobs or health which influence their expectations about future decisions to work or retire at ages 62 or 65, as well as how these influences and expected timing change as the individual approaches retirement.;Design and methods. Employing a conceptual framework based on Andersen's (1995) model of health care utilization, the first research question used Generalized Estimating Equations (GEE) to determine which predictors were associated with workers' expected retirement timing. The second question examined changes in expectations over time in response to changes in predictors, using a two-level approach modeled on Singer and Willett (2003). Utilizing the Core sample of the Health and Retirement Study (HRS); data on 5,989 individuals (48.3% female) were examined longitudinally across seven waves, 1992 and 2006.;Results. For research question one, female gender was associated with a lower expectations to continue working at both early (62), and normal (65) retirement ages. Non-married status had a similar effect for 62 but not for age 65. Differences emerged among sub-groups with non-married females reporting greater plans to continue working at both retirement ages; non-married blacks however were less likely to report plans to continue working at age 65 only. Non-whites with poorer subjective health or health that limited work were less likely to expect to work at age 62, but there was no impact on plans for age 65. Changes were also important. Leaving married status was linked with changes in plans to work at 65 but not 62. Gaining employer sponsored benefits (pension and health coverage) increased likelihood of continued work at both ages.;Implications. Findings confirmed that known predictors of retirement decisions are also associated with workers' expectations well before retirement age is reached. Andersen's model was useful for understanding the dynamic process leading up to retirement. Results also support the importance of employer sponsored benefits, such as pension and health insurance, on changes in plans as individuals approach retirement. Policy makers interested in prolonging work lives should look to modifiable factors (pensions and health insurance), which may impact retirement timing.