Couples Approaching Retirement: The Organization of Pension Behavior within the Family

In this analysis I used Wave 1 of the Health and Retirement Study to explore how the utilization of pension opportunities intersects with family context. I examined how decisions about retirement savings are shaped by family structure and family dynamics. Structural explanations of pension opportunities take non-market factors such as family responsibilities into account only to the extent that they shape location in the labor market. However, like other decisions related to household resources and savings, many decisions regarding employer pensions occur within the context of the family and it is likely that they are influenced by the resources and behavior of other family members. This research is an initial attempt to incorporate spouse characteristics and household dynamics into explanations of pension behavior. The analyses show the importance of considering the interdependence of decision making among couple members when conducting analyses of financial behavior and pension participation. Results suggest that women, at least this cohort of women, are more influenced by their spouses than are men. Working women tended to coordinate pension decisions with their husbands in a manner that was consistent with or complimentary to the behavior of their husbands. This points to the importance of attending to the influence of the spouse on individual behavior, as well as gender differences in this influence, and the way in which power dynamics within the household filter the translation of individual preferences into individual behavior. The analyses also demonstrate the necessity of accounting for differences in individual preferences and "states of mind" in pension research, since preferences for planning and risk influenced pension participation. In addition, the analyses indicate that there are substantive reasons for making an analytic distinction between primary and supplemental participation. Results also show that the potential selection effects involved in modeling pension participation did not bias the estimates. Although the theoretical reasons for performing these corrections remain, little was gained analytically. In contrast, the analyses demonstrated the importance of acknowledging and correcting for the non-independence of observations when examining individual-level behavior with a sample that contains both members of a couple.

Middle Aged Adults

1238

Available from UMI, Ann Arbor, MI. Order No. DA3014359. (0419-4209)