Social Influences in Couples: The Case of Early Retirement Decisions

Maria Eismann, Netherlands Interdisciplinary Demographic Institute (NIDI)
Kène Henkens, Netherlands Interdisciplinary Demographic Institute (NIDI-KNAW)
Matthijs Kalmijn, University of Amsterdam

Two people who share their lives as partners will inevitably influence one another. This interdependence raises considerable interest in the family literature. One life sphere in which the ‘linked lives’ of couples are particularly consequential is work. Each partner’s work decisions affect where the couple lives, both partners’ job level, the division of labor after childbirth, and the possibilities for joint leisure. With the pending retirement of a large group of older workers, spousal influence on the retirement decision is a particularly topical issue. However, few studies truly treat retirement as a couple-level decision and incorporate spouses as separate actors in their research design. Our study is among these few. Moreover, we aim to disentangle different mechanisms though which spouses’ influence workers’ retirement intentions and behavior. Spouses might prefer workers’ early retirement because of its effect on themselves (e.g., increased joint leisure), indicating self-interest. Spouses might prefer workers’ early retirement because of its effect on workers (e.g., improved well-being), indicating altruism. Spouses might also come to prefer for workers what workers prefer for themselves, indicating adaptation. We further hypothesize that spousal influence runs via persuasion, in which workers change their preferences according to spouses’ preferences and pressure, in which workers act according to their spouses’ preferences irrespective of their own preferences.

To test our hypotheses, we analyze data from the NIDI Pension Panel Survey (2015), a multi-actor study of 3,300 older workers (age 60-63) and their spouses. Administrative data about subsequent retirement behavior allow us to study the effect of spousal preferences on workers’ preferences as well as behavior. Results strongly support the self-interest, altruism, and adaptation hypothesis. Moreover, spouses influence workers’ retirement decision via both persuasion and pressure. This study contributes to life course theories by theoretically interpreting the concept of linked lives.


Introduction

When two people share their lives, either as a cohabiting or marital partners, they will inevitably influence one another. This interdependence between partners raises considerable interest in the in the family literature. One life sphere in which the ‘linked lives’ (Elder & Johnson, 2003) of couples are particularly consequential is work. Each partner’s work decisions affect where the couple lives (van der Velde, Jansen, Bal, & van Erp, 2017), both partners’ job level (Verbakel & de Graaf, 2009), the division of labor after childbirth (Stertz, Grether, & Wiese, 2017), and the possibilities for joint leisure (Carriero, Ghysels, & van Klaveren, 2009). With the pending retirement of a large group of older workers, spousal influences on the retirement decision are a particularly topical field of study.

Researchers increasingly acknowledge that retirement decision-making is a couple- rather than an individual-level process. Few studies however, incorporate the spouse as a separate actor in their research design. Our study is among these few and asks the following research questions: Why do spouses have an interest in influencing older workers’ retirement decision?, and How do spouses influence workers’ retirement preferences and subsequent retirement behavior?

The NIDI Pension Panel Survey

We analyzed data from the NIDI Pension Panel Survey (2015), a multi-actor study of 3,300 older workers (age 60-63) and their spouses. In contrast to most studies, we collected data of both spouses, allowing us to adequately investigate spouses’ preferences for workers’ early retirement. Moreover, the multi-actor data make the use of a two-stage least squares (2SLS) approach possible. This analytic method is particularly appropriate to estimate the mutual influence of preferences within a couple while accounting for interdependencies. By linking the survey data with administrative data about subsequent retirement behavior, the data also offer the possibility to study the effect of spousal preferences on workers’ preferences as well as behavior.

Spousal preferences for workers’ retirement

Spouses might aim to influence older workers’ retirement decision for various reasons. On the one hand, we hypothesize that spouses have self-interested reasons to prefer workers’ early retirement. Spouses who experience high relationship quality and prefer to be out of the labor force themselves are thought to look forward to spending more joint leisure time. Moreover, spouses who are in bad health are expected to prefer workers to retire in order to assume care tasks. On the other hand, we hypothesize that spouses also have altruistic reasons to prefer workers’ early retirement. Spouses who see that workers experience stress due to their work and spouses who worry about workers’ health are expected to prefer workers to retire in the hope that retirement will restore workers’ health and well-being. Lastly, we hypothesize that spouses will adapt their own preferences to those of the worker. Even though retirement is a couple decision, it is still likely to have a stronger impact on the workers than their spouses. Spouses who see that retirement is a central issue to workers might come to prefer for them what they prefer for themselves. The results support all three hypotheses.

Mechanisms of spousal influence

We propose two mechanisms by which spousal preferences might affect workers’ retirement process. One way for spouses to influence workers’ retirement process is by persuasion. Persuasion can take place by spouses providing workers with information that is congruent with spouses’ own preferences or simply trying to convince workers when discussing retirement plans. We hypothesize that workers have a stronger preference for early retirement to the extent that spouses also prefer workers to retire and that they will subsequently act upon this changed preference. Another way for spouses to influence workers’ retirement process is by applying pressure. Pressure can also take place in discussions of retirement plans. However, in these discussions spouses do not aim to change workers’ mind, but rather induce them to retire according to the spouses’ preferences irrespective of workers’ own preferences. We hypothesize that workers are more likely to retire early to the extent that spouses prefer workers to retire, even when controlling for workers’ own preferences. The results suggest that spousal preferences affect workers’ retirement process via persuasion as well as pressure. Interestingly, spouses seem to be more likely to pull workers out of the labor force than to push them to stay employed.

Conclusion

This study contributes to life course theories by theoretically interpreting the concept of linked lives. Spouses might have self-interested as well as altruistic reasons to seek influence in their partners’ decisions, but they might also simply adapt their own preferences to those of their partner. Moreover, spousal influence can take the form of persuasion as well as pressure. The results of the current study might help to explain spousal influence not only in the retirement decision, but also the work sphere more generally or even other life spheres.

Presented in Session 1115: Families and Households