Are Female-Breadwinner Couples Always Less Stable? Evidence from French Administrative Data

Giulia Ferrari , Institut National d'Études Démographiques (INED)
Anne Solaz , Institut National d'Études Démographiques (INED)
Agnese Vitali , University of Trento

With the increase of dual-earner couples and women’s educational level, couples in which the woman earns more than her partner are structurally becoming more widespread. Because female breadwinning challenges long-lived social norms regarding traditional gender roles, scholars have theorized that female-breadwinner couples would have a higher dissolution risk compared to couples in other earning arrangements. Using a French unique administrative panel database with an unconventionally high number of couples and separations events, this article aims at documenting whether female breadwinner couples are more likely to separate. Our results confirm that female-breadwinning couples are more likely to separate in France. All other things being equal, couples in which the woman’s share of the couple’s total income is higher than 55% are significantly more unstable than all other couples. This holds whatever the counterfactual couples considered, either equal-earners couples or male-breadwinner couples. We do not find any sign of a fading effect among the younger cohorts, as suggested by previous research. The higher risk of union dissolution for female-breadwinner couples hold across various marital arrangements (marriage, non-marital cohabitation, registered partnership). However, some interesting couple-stabilizing effects for dual-earner couples who have roughly similar individual earnings are observed for cohabitants and couples with a civil partnership arrangement. The female-breadwinner penalty is always present and robust to the inclusion of different control variables, interaction effects and model specifications. Our results provide an interesting sign of evolving behavior and norms that has not yet been emphasized by previous literature, usually considering restricted number of counterfactual couples.

See extended abstract

 Presented in Session 37. Divorce determinants