Assessing the Impact of Retirement Resources on U.S. Older Female Workers’ Retirement Timing: A Theory of Planned Behavior Model

Retirement is an important life event for all workers. Most older workers look forward to
retirement and having a retirement plan is important for a successful transition. Those who plan
also demonstrate a more positive attitude and greater confidence in their retirement. Much previous
work on retirement and related policies or programs has focused on male workers. Consequently,
social awareness of the problems encountered by older women during retirement remains low.
Women have limited retirement resources (Vrdoljak & Rappaport, 2018) and are more likely to
live in poverty than older men. Many older women who lack financial security depend on Social
Security benefits. Moreover, the gender gap increases with age and is closely associated with
quality of life and health (James, Matz-Costa, & Smyer, 2016). Unequal work experience and
access to fewer retirement resources can postpone older women’s retirement, leading to a higher
retirement age for women.
Therefore, I examine the relationship between older women’s retirement resources and the
timing of their retirement. I also examine the relationship between older women’s marital status
and retirement timing. In this study, I investigated these questions with a sample of women aged
between 50 and 62 years old who worked either full - or part - time, using data from the RAND
Corporation and the University of Michigan Health and Retirement Study (2017). Guided by the
theory of planned behavior, multiple regression analysis was used to examine older adults’
expected retirement timing. Factors that might influence this timing included attitudes toward
retirement, subjective norms about retirement, and perceived behavioral control (retirement
security). My analyses also examined gender differences in predicting retirement timing.
Subsequent analyses were conducted with women only, primarily to examine a potential linear
relationship between retirement timing and marital status, one of the “background factors” in the
theory of planned behavior. Additionally, logistic regression analyzed the effects of respondents’
expectations of retirement (i.e., comparing respondents with an expected timing of their retirement
with those who did not).
The study findings indicated that theory of planned behavior factors are useful for
predicting retirement timing. The model works similarly for men and women, but there is a
difference according to marital status. Unmarried women are likely to anticipate a later retirement
than married women and are less likely to set an expected timing for retirement.