The Labour Market Preconditions to Fertility in Belgium: The Importance of Gender
Leen Marynissen, University of Antwerp
Micro-economic theories predict that the fulfilment of the aforementioned economic preconditions has a positive effect on first birth hazards, but do not differentiate in terms of partners’ relative labour force positions: couples are assumed to divide paid and unpaid work in the economically most efficient manner, regardless of gender.
The doing gender hypothesis on the other hand argues that couples conform to and reproduce gender norms, suggesting that traditional gender roles may persist, resulting in differential effects of women’s versus men’s labour market positions on first birth hazards.
Empirically, previous studies on couples’ labour market positions and fertility routinely lacked the necessary data required to look into different dimensions of labour market positions rather than mere activity or employment. Using data from a Belgian Administrative Socio-Demographic Panel (1999-2010) which contains detailed quarterly data on labour market and income positions for all household members, this paper examines the effect of women’s labour force position relative to that of their partners on first birth decisions in Belgium between 2000 and 2010.
The gender revolution that started aroundthe second half of the twentieth century gave rise to increasing femaleeducational attainment and female labour market participation. Whereas, genderdifferences are becoming smaller in the public sphere, gender role differencespersist in the private domain. This raises questions on whether labour marketpreconditions to parenthood have changed or remain traditionally gendered.Although available research widely supports that financial resources, time andcertainty about future labour market positions are preconditions for thetransition to parenthood, the degree to which it matters whether theserequirements are fulfilled through mens or womens economic position dependsstrongly on the theory considered.
Micro-economic theories predict that thefulfilment of the aforementioned economic preconditions has a positive effecton first birth hazards, but do not differentiate in terms of partners relativelabour force positions: couples are assumed to divide paid and unpaid work inthe economically most efficient manner, regardless of gender.
The doing gender hypothesis on the otherhand argues that couples conform to and reproduce gender norms, suggesting thattraditional gender roles may persist, resulting in differential effects ofwomens versus mens labour market positions on first birth hazards.
Empirically, previous studies on coupleslabour market positions and fertility routinely lacked the necessary data requiredto look into different dimensions of labour market positions rather than mereactivity or employment. Using data from a Belgian AdministrativeSocio-Demographic Panel (1999-2010) which contains detailed quarterly data onlabour market and income positions for all household members, this paperexamines the effect of womens labour force position relative to that of theirpartners on first birth decisions in Belgium between 2000 and 2010.
2 Data and methods
We use data from the Belgian AdministrativeSocio-Demographic Panel that was constructed using microdata from the NationalRegister and the Crossroads Bank for Social Security, providing detailed longitudinalinformation on a representative sample of women aged 15-50 years, legallyresiding in Belgium for the period of January 1st 1999 to December 31st 2010, includingtheir household members. To maintain the cross-sectional representativeness ofthe panel throughout the observation period, annual top-up samples of 15-yearolds were drawn to guarantee the presence of the youngest age group in thepanel, as well as annual samples of women aged 16-50 years who settled inBelgium in the preceding year.
Our selection contains all nulliparous womenaged 18 and older, who have a co-resident partner and are no longer enrolled ineducation. We observe these women until their first child is born or until theyare censored as a result of mortality, emigration or reaching the end of theobservation period.
In our analyses, we estimate discrete-timehazard models of first birth decisions. First birth decisions are measured by laggingfirst births by 4 quarters in order to avoid reverse causality. We examine theextent to which first birth decisions depend on mens and womens employment(unemployed/inactive/employed/self-employed) (model1), controlling forcomposition effects such as origin of both partners (model2). In the last twomodels (models 3 and 4), we examine whether first birth decisions depend onwomens working hours in paid employment relative to those of their partnersand on womens relative salary, for couples where both partners are employed.
3 Results and discussion
Results of model 1 (Table 1) show that the odds of having a firstbirth are higher for unemployed and inactive women than for women that areemployed on the labour market. Employment of the male partner has nosignificant effect on the first birth decision. After controlling for origin ofboth partners, region, age of the partner and age difference between thepartners, only the positive effect of being inactive versus employed remainsfor women. Furthermore, for men, being inactive versus employed decreases the oddsof having a first birth.
Results of model 3 (Table 2) show no significant effects ofwomens working hours relative to those of their partners. With respect to income,the odds of having a first child are higher when women earn substantially lessthan their partner. This effect remains when controlling for compositioneffects in model 4. Finally, total household income, which we included in theanalyses to control for a couples socio-economic position, has no significanteffect.
In line with our expectations, the results indicatethat it does matter whether the economic preconditions to parenthood arefulfilled by men or women, as illustrated by the positive effect of beinginactive versus employed on first birth decisions for women whereas this effectis negative for men. Also the results of the 3rd and 4th model indicate agendered effect of income. Overall, our findings provide more support fortheories drawing attention to gender differences in labour market preconditionsto parenthood than for micro-economic theories neglecting the role of gender.
Presented in Session 1234: Posters