Spousal resources and relationship quality in eight European countries

ABSTRACT

We relate relationship satisfaction and thoughts about leaving a romantic relationship to a couple’s relative and absolute resources and check for context-dependency of those associations. Our theoretical point of departure is that the more resources women have compared to their spouses, the higher their intra-household bargaining power to negotiate themselves out of unpleasant tasks, particularly in gender-egalitarian and very income equal and unequal societies. In traditional societies (which score low on the Gender Empowerment Measure (GEM)), the inflexible role of men within the household presumably prevents women from bargaining a better position, which in turn negatively affects relationship quality. Income equality (low GINI coefficient) may be a prerequisite for women’s bargaining position, where more inequality (mid-GINI) may be detrimental for it. Nevertheless, extreme income inequality (high GINI) may again be favorable for women’s relationship power. Using country fixed effects models on data from the Generations and Gender Surveys (GGS), we compare men and women who are in a couple (formed after 1995) for eight European countries. We find that absolute resources matter more than relative resources, at least for relationship satisfaction: Higher educated couples are more satisfied with their relationships, which could suggest lower stress levels in those couples (in more traditional contexts). Second, we observe GINI context-dependency of the association between relative education and relationship satisfaction for women and relative education and exit thoughts for men, although opposite to what we expected. Perhaps reference group theory or gender display theory can explain these unexpected results. Finally, we find that women have more break-up plans in societies with a lower score on GEM. This last result is consistent with the notion that bargaining only works in egalitarian contexts.

RESUMEN

Relacionamos el nivel de satisfacción de una relación romántica y los pensamientos sobre dejar la relación con los recursos relativos y absolutos de la pareja, y verificamos si estas asociaciones dependen del contexto. Nuestro punto de partida teórico es que mientras más recursos tengan las mujeres con respecto a sus esposos, mayor será su poder de negociación al interior del hogar para dejar de realizar tareas poco placenteras, en particular en contextos igualitarios de género y en sociedades que tienen desigualdades extremas en los ingresos (tres igual o tres desigual). En sociedades tradicionales (las cuales tienen un bajo puntaje en el Gender Empowerment Measure (GEM)) el rol inflexible de los hombres dentro del hogar previene, presumiblemente, que las mujeres tengan una mejor posición de negociación, lo cual a su vez afecta negativamente la calidad de la relación. La igualdad en los ingresos (un bajo coeficiente de Gini) podría ser un requisito para una mejor posición de negociación de las mujeres, mientras que más desigualdad (un mediano coeficiente de Gini) puede ser contraproducente. Sin embargo, una desigualdad extrema en los ingresos (alto coeficiente de Gini) podría ser favorable para el poder de negociación de las mujeres. Usamos modelos de efectos fijos de países y datos de Generations and Gender Surveys (GGS) para comparar parejas de hombres y mujeres (siendo pareja desde 1995) para ocho países Europeos. Encontramos, primero, que los recursos absolutos son más importantes que los relativos, al menos para el nivel de satisfacción con la relación: parejas con más educación están más satisfechas con sus relaciones, lo cual sugiere que hay menos estrés en estas parejas (en contextos más tradicionales). Segundo, observamos dependencia del contexto (coeficiente de Gini) en la asociación entre educación relativa y la satisfacción de la relación para el caso de las mujeres, y en la educación relativa y pensamientos de dejar la relación en el caso de los hombres, aunque esto es opuesto a lo esperado. Tal vez, la teoría de los grupos de referencia o del gender display puede ayudar a explicar estos resultados no esperados. Finalmente, encontramos que las mujeres tienen más planes para romper la relación en sociedades con un menor puntaje en GEM. Este último resultado es consistente con la noción que la negociación solo funciona en contextos igualitarios.

KEYWORDS:

• Relationship satisfaction
• break-up plans
• relative resources
• absolute resources
• bargaining
• gender equality

PALABRAS CLAVE:

• Satisfacción de una relación
• planes para romper la relación
• recursos relativos
• recursos absolutos
• negociación
• igualdad de género

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Maike van Damme is a postdoc in sociology at the University of Cologne. Her research interests include life course events and social inequalities in comparative perspective. One of her publications is on the cross-national comparison of the dynamic relationship between women’s employment and separation in 16 countries, published in Social Science Research (2014; vol 48).

Pearl Dykstra is a professor in Empirical Sociology at Erasmus University Rotterdam. She is an internationally regarded specialist on intergenerational solidarity, aging societies, family change, aging and the life course, and loneliness. Large scale projects for which she has been awarded grants include the Netherlands Kinship Panel Study (NKPS) and the EU 7th Framework program MULTILINKS. She also received an ERC Advanced Investigator Grant for the research project ‘Families in context’, which focuses on the ways in which policy, economic, and cultural contexts structure intergenerational and gendered dependencies in families.

ORCID

Maike van Damme http://orcid.org/0000-0003-1080-9789

Pearl Dykstra http://orcid.org/0000-0003-1518-1476

Notes

1 Referred to as external ‘threat point’ in the economic literature.

2 Note that also individual level gender role beliefs could influence partner’s negotiations. We checked interactions of relative resources with individual level gender role values, but found no significant effects.

3 Cross-national comparability of both spouses’ earnings was done by Joanne Muller (Earnings File (W1) provided on http://www.ggp-i.org/data, retrieved April 13, 2017).

4 We take the natural logarithm of wife’s relative earnings to transform the right skewed distribution into a normally distributed one.

5 To check whether this sample selection affects our results, we also included a human capital measure instead of earnings (Sullivan & Gershuny, 2016).This did not substantially change the results so we present here only the results using earnings.

6 We do not include the division of household labor itself, since this is a collider variable.

7 In many countries, women working more hours have been found to be more likely to break up (van Damme & Kalmijn, 2014). We do not expect to find a relationship for men though but include the variable anyway for consistency among the models.

8 As a sensitivity check, we ran models with aggregated gender role values (from GGS) instead of GEM as a macro-indicator of the degree of gender egalitarianism in a country. We did not find any significant effect of this aggregated measure as a main effect. With respect to the interactions with resources, we find weaker effects than when including GEM. Note that a macro-indicator from another source would be a more reliable measure than this aggregated measure of GGS on values. Unfortunately, in other data sources there is no other time-varying measure of gender egalitarianism available.

9 Note that there can be gender bias in men’s and women’s reporting of the same variable (Doorten, 2008).

10 Notice the gender bias in reporting. Men report more often that they are higher educated than women.

11 Note that the correlation between relative earnings and absolute individual earnings for women is rather high (around 0.75 for women). Such high correlations make it difficult to disentangle relative from absolute resources effects for earnings. Therefore, we also ran all models with relative earnings only included (minor changes in the coefficients). Associations between absolute and relative levels of education and occupational status are very low and thus can be distinguished from one another.

12 Between brackets is the range of the variable.