Time to Reconsider Work: Dual-Earner Couples’ Work-Related Adaptive Strategies and Preferences for Reduced Work Hours

Abstract

In this study, we use the life course perspective and the paradigm of the social construction of gender to examine the relationships between dual-earner couples’ adaptive strategies, such as their work-hour arrangements, conjoint occupational status, and relative earnings, and men’s and women’s own preferences for reduced work hours as well as their desire for reduced work hours of their spouses. Using the 2010 European Social Survey, we document a pervasive preference for reduced work hours in European countries, which is common to both men and women. Our findings indicate that, regardless of their actual work hour arrangements, conjoint occupational status, and relative earnings, couples generally report preferences for working hours for themselves and their spouses that conform to a modified male breadwinner/female homemaker template. More specifically, the ideal couple-level working time arrangement combines a husband who has a full-time job but does not work long hours and a wife working shorter hours, either in a part-time or full-time job. We discuss the impact of these findings on gender inequality, individuals, and organizations.

Keywords:

• Dual-earner couples
• gender
• preferences
• work hours

Notes

For more information, see: https://www.europeansocialsurvey.org/docs/methodology/ESS_weighting_data_1.pdf.

Cross-national research documents differences across countries and welfare regimes in working time regulations. Hence, we control for membership in specific welfare regime clusters (see Gornick and Meyers 2004).

We examined whether age has a non-linear association with working time preferences and the gap between actual and preferred working time. Age had a non-linear association with the gap only for men. However, inclusion of a squared age term did not impact the association of the three main independent variables and the dependent variable. Hence, for simplicity, we report results from the models excluding the squared term.

A count of the number of children in each of these two categories would have enabled a more accurate estimation of the impact of children on desired work hours. Unfortunately, such information was not available.

A factor analysis (using principal factors) produced factor loadings of 0.75, 0.58, and 0.67 and Cronbach’s alpha of 0.77 for the work autonomy construct for men and factor loadings of 0.76, 0.6, and 0.7 and Cronbach’s alpha of 0.78 for women.

The correlation between these two variables was r = 0.4, p < 0.05 for men and r = 0.32, p < 0.05 for women.

Specifically, Great Britain and Ireland were defined as having a liberal regime; Belgium, France, Germany, Netherlands, and Switzerland as having a conservative regime; Denmark, Finland, Norway, and Sweden as having a Social Democratic regime; and Israel, Portugal, and Spain as belonging to the Southern European/Mediterranean regime. This classification is commonly used in cross-national research focusing on the interaction between welfare policy and gender gaps in employment outcomes (e.g., Gornick and Meyers 2004).

Only two differences between the models emerged. Both of these differences were in the models estimated for men. First, men in couples where only the husband is a manager or a professional are more likely to want to reduce the work hours for both themselves and their partners in the models excluding the controls, but the impact is not significant in the models with the control variables. Second, men in couples where both partners are managers or professionals are more likely to want to reduce the work hours for both themselves and their partners in the models excluding the controls, but the association is not significant in the models with the control variables.

We found that both the variable recording the gap in working hours (i.e., the gap between the actual and preferred number of hours) and the residual from a regression predicting the gap between the variables described below were normally distributed.