Labour and love: wives’ employment and divorce risk in its socio-political context

Dual-earning is now common in Western couple households, despite the fact that a wife’s employment was historically associated with greater marital instability. With support from the Leverhulme Trust, recent research led by Professor Lynn Prince Cooke, with an international team of researchers, sought to examine whether a country’s level of policy support for a wife’s employment affects its associated divorce risk.

The researchers argue that more extensive policy supports can ease competing time demands and financial pressures on families which, in turn, may ease the link between a wife’s employment and divorce risk. Using harmonized national data, the researchers found that only in the United States, where policy support is minimal, does a wife’s employment still significantly increase the risk of divorce.  In Australia, Flanders, France, Germany, Italy, the Netherlands and the United Kingdom, this is no longer the case.

What is more, in Finland, Norway and Sweden, countries with the most generous welfare state and policy supports, wives in paid work have a significantly lower risk of divorce, as compared with wives who remain outside the labour force.  The findings highlight the importance of policy in addressing gender equality in employment, as well as related family risks, with more extensive supports making dual-earning the most stable family form in post-industrial societies.

Key Findings

The research found that:

• Only in the United States, with its minimalist policy support for family, does a wife’s employment still significantly increase the risk of divorce, even though Americans have the most conservative attitudes towards marriage and divorce of the 11 countries included in the study
• In Australia, Flanders, France, Germany, Italy, the Netherlands and the United Kingdom, where male breadwinner couples were previously the norm, a wife’s employment no longer significantly increases the risk of divorce 
• In Finland, Norway and Sweden, countries with the most generous policy supports for women’s employment across the life course, wives who are employed have a significantly lower risk of divorce, as compared with wives who remain outside of the labour force
• The greater the policy support for gender employment equality, the more likely it is that dual-earner couples are more stable than male breadwinner couples
• This broad range of results highlights the importance of the socio-political context in influencing individual-level risks, notwithstanding the predictions of theories previously assumed to be generally valid
• The uniqueness of the US findings indicates the need to reconsider how gender- and family-related theories developed by US social scientists might differ in their applicability across countries and institutional settings

Background to the research and the findings in context

In Western industrialised countries, the increase in married women’s labour force participation in the post war era correlated with an increase in divorce rates. Evidence indicated this association was causal, with employed wives in many countries more likely to divorce than wives who remained outside of the labour force. Some theorists argued that the marital stability of male breadwinner couples was due to the interdependence fostered by a gendered division of paid and unpaid work within nuclear families. Other researchers instead suggested that the greater divorce risk of dual-earning couples stemmed from the competing work and family time demands, as working wives retain primary responsibility for housework and childcare. Thus policies such as public provision of child care or generous paid parental leaves, that support women’s employment and also encourage men’s greater participation in care activities, could reduce the work-family strains faced by dual-earning couples.

The limited comparative evidence on data into the 1990s revealed mixed results. Divorce rates tended to be higher in countries with greater female employment rates, but differences in individual-level effects were less pronounced. Exploring the impact of the socio-economic context on shaping individual-level divorce risk, now that dual-earners comprise the majority of households, has been hampered by the lack of suitable comparative data. With Leverhulme funding for an international network, researchers from 11 Western countries came together to harmonize the best-available national longitudinal data to explore the individual-level risk of marital breakdown among employed wives marrying for the first time in 1990 and into the new millennium. Also compiled was aggregate information on marriage, divorce and wives’ employment rates across the 1990s, along with attitudinal and policy information to ascertain whether there were distinct socio-political and demographic patterns.

At the aggregate level, factor analysis revealed two patterns of indicators that explained almost two-thirds of the context variation. A Policy Support for Equality factor reflected high social expenditure, generous maternity leave, public child care provision, and low income inequality. This pattern most clearly represents the Nordic countries of Finland, Norway and Sweden, but France and Belgium also loaded positively on this factor. A second factor bundled high wives’ employment rates, with yet more conservative attitudes regarding whether divorce is the best solution to persistent marital problems. The United States loads most highly on this factor, followed by the United Kingdom and Australia. Finland, Norway and Sweden also load positively on this factor, albeit more weakly than the United Kingdom. This factor therefore reflects high female employment, but without the impact of policy support.

At the individual level, event history analyses revealed that only in the United States does a wife’s employment still significantly increase the risk of divorce. In Australia, Flanders, France, Germany, Italy, the Netherlands and the United Kingdom, this is no longer the case. In Finland, Norway and Sweden, countries with the most generous policy support, employed wives have a significantly lower risk of divorce as compared with wives who remain outside of the labour force. Furthermore, the country loading on the Policy Support for Equality factor predicted the size of the individual-level impact of a wife’s employment on divorce: the greater the policy support for equality, the smaller the risk an employed wife will divorce.

Policy Implications

The findings suggest that a wives’ employment may indeed increase the risk of divorce, but the causal mechanism likely rests in the competing time demands between home and employment. In countries such as the United States where couples must purchase any supports such as private child care from the market, dual-earning places greater strains on family: either someone must work longer hours to earn the extra money, or the family must do without supports altogether. Relying on market solutions might therefore exacerbate work-family conflicts for dual-earning couples. In contrast, the evidence here suggests that social policies ameliorate the individual-level risk, possibly by easing the competing time demands and financial pressures that can destabilise families.

Methodology

Factor analysis was used to ascertain the degree of co-variation of aggregate measures of 1990 and 2000 crude marriage and divorce rates, wives’ employment rates, percentage of women working part-time, national income inequality, state expenditure on welfare, percentage of children in public child care place, weeks of paid maternity leave, attitudes toward marriage, divorce, and whether both partners should contribute to household income. The results yielded two primary factors that explained approximately 67 percent of the variation in these macro indicators.

To assess the individual-level divorce risk faced by employed wives, we located the best available national longitudinal data for each country, and then harmonized sampling, variable construction and modelling for women marrying for the first time in 1990 into the new millennium. Yearly observations for each wife throughout her marriage were then analysed with event history analysis with robust standard errors.