The Impact of Partners’ Relative Wages on Couples’ Gender Division of Paid Work after Parenthood across Origin Groups

Abstract

The transition to parenthood exacerbates gender inequality in couples’ division of paid work. While this is widely documented for general populations, in particular, potential underlying mechanisms of this phenomenon remain un(der)explored for couples with different migration backgrounds. Hence, this paper examines how women’s pre-birth relative wage potential affects the gender division of paid work after the transition to parenthood in Belgium among native, Southern-European and non-European origin couples. Our results show that, among all couples, the division of paid work is more gender-equal after childbirth when women’s wage potential is higher than or similar to that of their male partner. However, there is substantial variation by couples’ migration background and relative wage potential in partners’ gender division of paid work and the extent to which it changes after parenthood. These findings suggest that both normative and institutional factors moderate the impact of partners’ relative resources on couples’ division of paid work after parenthood, particularly among non-European origin couples.

1. Introduction

To this day, the transition to parenthood exacerbates gender inequality in couples’ division of (un)paid work (Kreyenfeld 2015; Kuhhirt 2011; Schober 2013; Wood et al. 2018). While men’s employment trajectories are relatively stable over their life course, many women reduce their working hours after the transition to parenthood. This decrease in female labour force participation after family formation and its determinants attract ample attention from both academics and policy makers since they have implications at both the societal level (e.g., covering welfare state costs in the context of accelerated population ageing and shrinking working-age populations) and the individual level (e.g., financial independence). In line with micro-economic theories (Becker 1991; Lundberg and Pollak 1996), research has shown that partners‘ relative labour market characteristics (e.g., in terms of earnings) prior to childbirth are key determinants in shaping (changes in) couples’ gender division of paid work after the birth of their first child: larger relative wages for women compared to their partners diminish the parenthood effect on gender inequality in paid work (Begall and Grunow 2015; Kanji 2011; Kuhhirt 2011; Vignoli et al. 2012; Wood et al. 2018). At the same time, prevailing gender norms and institutional contexts also influence couple-level gender dynamics around family formation and moderate whether and to what extent partners’ pre-birth relative resources shape (changes in) couples’ gender division of paid work after the transition to parenthood. However, this strand of research is also characterised by a major limitation: it predominantly studies general populations. As a result, gender dynamics in couples’ labour market participation around family formation and, in particular, potential underlying mechanisms, remain un(der)explored for couples with different migration backgrounds. This is remarkable given the increasing share of couples of migrant origin in European societies, who are or may be characterised by differential cultural settings and/or economic opportunity structures compared to couples without a migration background.

Therefore, this paper addresses how women’s pre-birth relative wages affect the gender division of paid work after the transition to parenthood in Belgium among native couples (i.e., both partners without a migration background), Southern European origin couples (i.e., at least one second-generation Southern European origin partner) and non-European origin couples (i.e., at least one second-generation Turkish or Maghreb origin partner). We build on and contribute to existing research in three ways. First and foremost, this paper addresses the lack of research on migrant populations by studying couples where at least one partner has a second-generation migration background. Examining to what extent the gender division of paid work after family formation of second-generation couples differs compared to native couples and how this is related to pre-birth labour market characteristics, such as earnings, may increase our understanding of the mechanisms behind the generally lower labour market participation of migrant origin women. This study focuses on Southern European, Maghreb and Turkish origin groups as their specific migration histories resulted in specific socio-economic and ideational contexts among these origin groups, which may in turn affect their gender dynamics in the division of paid work around family formation (Maes et al. 2022). Second, this paper adopts a much needed life course perspective. Life course research among general populations has substantially enhanced our understanding of gender dynamics in couples’ labour market participation around family formation by highlighting that couples’ pre-birth relative positions shape (changes in) their division of paid work after the transition to parenthood (i.e., principle of path dependency) and that these gender dynamics are also influenced by normative and institutional contexts (i.e., principles of time and place and bounded agency) (Bernardi et al. 2019; Elder et al. 2003). Such a perspective is missing in research on migrant populations: while recent research addresses (changes in) the division of paid work after family formation among migrant origin couples (Maes et al. 2022), it remains unexplored how partners’ pre-birth relative wages affect couples’ gender division of paid work after family formation among different origin groups despite considerable differences by origin group, migrant generation and gender in labour market opportunities and cultural settings. Third, this paper focuses on Belgium, which provides an interesting and well-suited context for three reasons: its institutional and labour market context, its sizeable migrant population and the availability of unique longitudinal microdata. First, alongside France and Nordic countries, Belgium is considered as a forerunner context in which work and family are relatively compatible and characterised by low employment gaps between mothers and childless women (Ciccia and Bleijenbergh 2014; Matysiak and Węziak-Białowolska 2016). As Belgium provides relatively large policy support to combine work and family, it is likely that partners’ relative resources have a stronger influence on couples’ gender division of paid work in comparison to other contexts (Wood et al. 2018). At the same time, Belgium displays larger socio-economic differences in the uptake of work–family reconciliation policies compared to other European countries, which has been related to the fact that access to formal childcare and parental leave is strongly conditioned on stable employment (Kil et al. 2017; Marynissen et al. 2021; Pavolini and Van Lancker 2018; Vandenbroeck et al. 2008). This differential access may diminish the role of partners’ relative wages in shaping couples’ gender division of paid work after family formation. Second, Belgium is an old immigration country with a substantial share of the population having a migration background, but also exhibits one of the largest employment gaps between natives and migrant origin groups (especially non-European origin groups) across Europe (Eurostat 2014). As a result of their long migration history, Southern European, Turkish and Maghreb origin groups constitute—apart from neighbouring countries—the largest foreign origin groups in Belgium, with a large second generation at working and childbearing age (FOD WASO and UNIA 2019). Third, addressing variation by couples’ migration background in the association between women’s pre-birth relative wages and couples’ gender division of paid work after family formation poses high data requirements. Therefore, this paper uses unique longitudinal microdata from Belgian Social Security registers, which provide detailed information at the couple level on both partners’ employment and family trajectories.

2. Theoretical Perspectives

In line with the life course principle of path dependency, long-standing micro-economic theories, such as New Home Economics (Becker 1991) and bargaining theories (Lundberg and Pollak 1996), argue that partners’ relative resources prior to childbirth are key determinants in shaping the gender division of paid work within couples after family formation. They suggest that couples aim to maximise household/individual utility and predict that, regardless of gender, the partner with the highest wage takes up more paid work while the other partner reduces working hours in order to take up more household work and childcare. Hence, paid work will be divided more equally after the transition to parenthood among couples where women have a pre-birth wage higher than or similar to that of their male partner in comparison to couples where the male partner has a higher pre-birth wage. In this respect, previous research indicates that higher relative wages for women compared to their partner before childbirth diminish the parenthood effect on gender inequality in paid work. A study for Belgium on dual earner couples shows that female main earner couples have a significantly larger probability of adopting female-oriented parental employment strategies (i.e., the female partner works more hours than the male partner after childbirth) than couples where the male partner has the highest income (Wood et al. 2018). Results for West Germany (Kuhhirt 2011) and the UK (Kanji 2011) also indicate that women who earned the same wage or more than their partner before childbirth were less likely to decrease their working hours following childbirth.

Despite the finding that partners’ pre-birth relative wages diminish the parenthood effect on gender inequality in paid work, empirical evidence for general populations indicates that male-oriented employment strategies remain dominant after childbirth, even in couples where women have a higher income than their male partner (Kuhhirt 2011; Schober 2013; Wood et al. 2018). In line with the life course principle of time and place, this suggests that the persistence of gendered parenting norms may hamper an economic cost–benefit rationale in couples’ division of paid work after family formation. Sociological theories highlight societal norms in shaping couple-level gender dynamics and argue that couples conform to and reproduce societal norms regarding the employment of mothers and fathers, even when this division does not maximise couples’ joint economic resources (Blumberg 1984; West and Zimmerman 1987). Deviating from the dominant gender role expectations may entail social penalties (e.g., negative reactions, social exclusion) (Heilman et al. 2004). Furthermore, life course research highlights that couples’ division of paid work after family formation is strongly influenced by countries’ work–family reconciliation policies, as policies such as parental leave1 and formal childcare2 shape the degree to which work and family life are (in)compatible (i.e., principle of bounded agency). Whereas economic theories highlight that the increasing availability of work–family reconciliation in most European countries limits opportunity costs of childrearing, sociological theories emphasise that the availability and designs of such policies reflect particular normative expectations regarding gender roles. Countries such as Belgium, with extensive (policy) support to combine work and family, enable gender egalitarian divisions among parents and provide a context in which partners’ pre-birth relative wages can influence couples’ decisions regarding their work–family strategies.

Based on these considerations and existing studies for general populations, where the findings predominantly reflect the patterns of majority groups, two hypotheses regarding the association between partners’ pre-birth relative wages and (changes in) the gender division in paid work after parenthood among native Belgian couples are put forward in this study:

H1a: 

Native couples’ division of paid work after the transition to parenthood is more gender-equal when women’s pre-birth wage is higher than or similar to that of their male partner compared to native couples where the female partner has a lower pre-birth wage.

H1b: 

Gender inequality in native couples’ division of paid work increases after the transition to parenthood, but to a lesser extent among couples where women have a higher relative wage before family formation.

Considering the aforementioned mechanisms, the role of economic mechanisms in shaping couples’ gender dynamics around family formation is likely to vary by couples’ migration background as migrant origin couples are or may be characterised by differential cultural settings and/or economic opportunity structures (which also affect access to work–family reconciliation policies) compared to native couples. With respect to Southern European origin couples (hereafter, South-EU origin couples), we expect similar patterns regarding the linkage between partners’ pre-birth relative wages and couples’ gender division in paid work after parenthood. Unfortunately, little is known about the gender role attitudes of Southern European origin groups in West European countries, but research for Belgium shows that the gender gap in employment among this population subgroup is relatively similar to the corresponding gender difference among native Belgians (FOD WASO and UNIA 2019). Moreover, a recent study indicated that South-EU origin couples display a gender division of paid work similar to that of native Belgian couples and also exhibit similar changes in this division around family formation (Maes et al. 2022), which suggests that gender role expectations of South-EU origin groups are similar to those of native Belgians and that a gender-equal division of paid work after family formation is not constrained by less access to work–family policies. Therefore, the hypotheses regarding South-EU origin couples are as follows:

H2a: 

Similar to the pattern among native couples, the gender division of paid work after the transition to parenthood is more gender-equal among South-EU origin couples where women’s pre-birth wage is higher than or similar to that of their male partner compared to South-EU origin couples where the female partner has a lower pre-birth wage.

H2b: 

Similar to the pattern among native couples, gender inequality in South-EU origin couples’ division of paid work increases after the transition to parenthood, but to a lesser extent among South-EU origin couples where women have a higher relative wage before family formation.

Regarding couples with at least one second-generation Turkish or Maghreb origin partner (hereafter, non-EU origin couples), we expect that the association between partners’ pre-birth relative wages and the gender dynamics around family formation differs compared to native couples as a result of two mechanisms: (i) alternative selection into parenthood and (ii) persistent traditional parenting norms and/or limited access to work-family policies. Hence, on the one hand, sociological theories of identity formation suggest that socialisation in different normative contexts and limited labour market opportunities (i.e., parent role as an alternative career to reduce labour market uncertainties) may foster more traditional work–family attitudes (Davis and Greenstein 2009; Friedman et al. 1994). In turn, it is likely that some couples do not postpone the transition to parenthood until both partners have established a stable position in the labour market (as is common in majority populations) and are more likely to have their first child in case only the partner with the highest wage potential is employed, as the other partner will specialise in unpaid work after the transition to parenthood. This alternative selection into parenthood may be more common among Turkish or Maghreb origin couples as a result of lower labour market opportunities compared to native Belgian couples, as well as different socialisation contexts. Empirical research for Belgium consistently shows that Turkish and Maghreb origin men and particularly women exhibit lower employment levels than natives and are also overrepresented in lower labour market segments3 (FOD WASO and UNIA 2019; Maes et al. 2019). Furthermore, non-EU origin couples may find themselves in an intermediate position between two normative contexts and may have to synthesise potentially opposing norms and expectations regarding gender roles. Accordingly, research for Belgium provides empirical evidence for differential preconditions for the transition to parenthood as native couples are most likely to have a first child in case both partners are employed, whereas Moroccan origin couples are most likely to have their first child when only the male partner is employed (Wood et al. 2017). This differential selection into parenthood implies that non-EU couples where women have lower relative wages before childbirth may be a selective group with more traditional work–family attitudes, which entails a highly unequal division after childbirth. Therefore, our first hypothesis regarding the role of partners’ pre-birth relative wages among non-EU couples is as follows:

H3a: 

There is a strong degree of gender inequality after the transition to parenthood among non-EU origin couples where women have a lower relative wage, which already existed prior to the birth of the first child.

On the other hand, the persistence of traditional norms regarding the employment of mothers and fathers and/or limited access to work–family policies may yield weaker associations between partners’ pre-birth relative wages and changes in the division of paid work among non-EU couples compared to native couples. Regarding the former, it is likely that higher relative wages for women prior to childbirth may not diminish the parenthood effect on gender inequality in paid work among non-EU origin couples due to (a stronger pressure to conform to) more traditional gender norms once there are children in the household. While previous research for Belgium suggests that marrying a partner from their origin country is for second-generation Turkish and Moroccan origin women a way to bend traditionally gendered power relations, and they typically have better labour market opportunities than their partner (Timmerman et al. 2009), it has been shown that the transition to parenthood still results in a strong increase in gender inequality among these couples (Maes et al. 2022). This suggests that it is no longer considered desirable to work within the Turkish or Moroccan community once women have children (Adam and Torrekens 2015) and that women will therefore reduce their working hours after family formation, regardless of their pre-birth relative wages. Furthermore, in case countries’ work–family policies primarily support parents who are firmly established in the labour market, such as in Belgium, variation by migration background in labour market opportunities also entails, by design, differential access to these policies. In this respect, research for Belgium indicates that non-EU origin parents with generally more precarious employment trajectories compared to natives have less access to formal childcare and parental leave (Kil et al. 2017; Marynissen et al. 2021; Vandenbroeck et al. 2008). When access to parental leave and formal childcare is limited, couples may turn to other strategies to combine work and family life, which are likely to entail reductions in working hours of one or both partners. To the extent that these alternative work–family strategies involve the persistence of gendered parenting norms, these constraints may diminish the role of partners’ pre-birth relative resources in shaping the gender division of paid work after family formation among non-EU couples.

As a result, gender inequality may increase to a stronger extent among non-EU origin couples compared to native couples where women have higher relative wages before family formation. Moreover, to the extent that non-EU couples where the female partner has a lower pre-birth relative wage already display a very unequal gender division of paid work before family formation (as a result of a differential selection into parenthood), gender inequality cannot increase to the same extent after the transition to parenthood compared to non-EU couples where women have a higher or similar relative wage, who are likely to exhibit a stronger degree of gender equality before first childbirth. Therefore, the second hypothesis related to non-EU origin couples is as follows:

H3b: 

Gender inequality in the division of paid work particularly increases after the transition to parenthood among non-EU couples where women have a higher or similar pre-birth relative wage, and this increase in gender inequality is stronger compared to native couples.

3. Data and Methods

3.1. Data

We used data from the Flemish administrative panel on Migration, Integration and Activation (MIA Panel), which provides longitudinal microdata from the Social Security Registers for the period 2005–2016 (CBSS)4. The MIA Panel provides information on a sample of individuals without a migration background (i.e., natives), individuals of Southern European origin5 and individuals of non-European origin6, aged 18–65 and legally residing in Flanders on 1 January 2005. Natives are defined as individuals who have the Belgian nationality at birth and with two parents that had the Belgian nationality at birth. An individual is considered to be of migrant origin when the person him/herself or one of their parents has a nationality at birth that is not Belgian. Individuals with a migration background who are not born in Belgium are defined as the first generation and individuals with a migration background who are born in Belgium are defined as the second generation7. Sampled individuals are followed until (i) the age of 65, (ii) emigration or death or (iii) the end of the observation period on the 31 December 2016. To maintain cross-sectional representativeness, supplementary annual samples of 18-year-olds were drawn to guarantee the presence of the youngest age group in the data throughout the observation period. For each observation year, household members of sampled individuals on the first of January are also included in the data. The MIA Panel is disproportionately stratified by age and migration background (i.e., over-representation of the younger age groups and individuals with a migration background8) and provides longitudinal information on individuals’ contractual working hours on a quarterly basis as well as the date of birth of all children, which allows us to analyse variation by migration background in couples’ gender division of paid work around the transition to parenthood.

For this study, we selected heterosexual couples who had their first child between 2006 and 2016 and restricted the analyses to couple quarters where both partners live in the same household and are not enrolled in education. Further, we focused on couples where at least one partner is employed and therefore excluded couples where both partners are not employed one year before first childbirth, as well as couple quarters in which both partners are not employed. In addition, we only included couples for which we have information on both partners’ contractual working hours in the fourth quarter before the birth of the first child (zero in case of unemployment or inactivity) and excluded couple quarters in case of missing information on contractual working hours for at least one partner during the observation period9. Couples are followed from one year before the birth of their first child until (i) two quarters before their second child is born10, (ii) the first child reaches the age of three, (iii) the couple is no longer in a co-residential union, (iv) reaching the end of the observation period or (v) the death or emigration of either partner.

Building on a recent study for Belgium (Maes et al. 2022), we distinguished three types of couples (

Table 1. Couples’ migration background considered in this study.

Native Couples
woman Belgium—man Belgium	100%
N couples	903
South-EU origin couples
woman 2G South-EU—man 2G South-EU	27%
woman Belgium—man 2G South-EU	34%
woman 2G South-EU—man Belgium	39%
N couples	549
Non-EU origin couples
woman 2G Turkey/Maghreb—man 2G Turkey/Maghreb	55%
woman 2G Turkey/Maghreb—man 1G Turkey/Maghreb	45%
N couples	722

Notes: 1G refers to first generation and 2G refers to second generation. Source: MIA Panel 2005–2016.

). First, we considered couples where both partners have no migration background (i.e., native couples). Second, Southern European origin couples (i.e., South-EU origin couples) are couples where (i) both partners have a second=generation Southern European origin or (ii) one partner has a native background and one partner a second-generation Southern European origin. The results of the study by Maes et al. (2022) indicated that these three types of South-EU origin couples, which are, in this study, clustered into one group, display the same gender dynamics in the division of paid work around parenthood. Third, non-European origin couples (i.e., non-EU origin couples) refer to couples where (i) both partners have a second-generation Turkish or Maghreb origin or (ii) the female partner has a second-generation Turkish or Maghreb origin and the male partner has a first-generation Turkish or Maghreb origin. The study of Maes et al. (2022) showed that these two types of non-EU origin couples considered in this study display similar gender dynamics around the transition to parenthood, whereas non-EU origin couples consisting of a first-generation woman and a second-generation man or mixed non-EU origin couples exhibit significantly different gender dynamics around family formation. These latter couples would need to be considered separately as a result of their distinct gender dynamics. However, we did not obtain robust estimations for these groups when distinguishing by women’s pre-birth relative wage potential (cf. Section 3.2.1) due to small sample sizes. Therefore, they are not included in this study. Descriptive statistics by couples’ migration background are provided in

Table A1. Descriptive statistics by couples’ migration background (in %).

	Native	South-EU	Non-EU
women’s duration of residence at first childbirth
2 years	0	0	0
3 years	0	0	0
4 years	0	0	0
5–10 years	0	0	0
10 years or more	0	0	0
missing	0	0	0
born in Belgium	100	100	100
men’s duration of residence at first childbirth
1 year	0	0	3.06
2 years	0	0	10.03
3 years	0	0	7.14
4 years	0	0	3.57
5–10 years	0	0	7.48
10 years or more	0	0	7.82
missing	0	0	2.55
born in Belgium	100	100	58.33
women’s age at migration
0–5	0	0	0
6–11	0	0	0
18–25	0	0	0
26–30	0	0	0
30+	0	0	0
born in Belgium	100	100	100
men’s age at migration
0–5	0	0	3.16
6–11	0	0	1.10
12–17	0	0	1.78
18–25	0	0	18.66
26–30	0	0	12.62
30+	0	0	3.84
born in Belgium	100	100	58.85
women’s age at first childbirth
younger than 22	0.24	1.16	3.40
22–24	6.71	7.56	27.21
25–27	29.23	22.88	38.26
28–30	37.25	36.04	21.25
31–35	22.63	27.52	8.67
36–40	3.36	3.88	0.85
40+	0.60	0.95	0.34
men’s age at first childbirth
younger than 22	0.00	0.38	0.17
22–24	2.40	3.30	5.61
25–27	16.17	14.92	29.07
28–30	34.00	27.91	29.58
31–35	35.58	38.56	28.39
36–40	9.95	11.63	5.61
40+	1.92	3.28	1.53

in Appendix A.

3.2. Methods

3.2.1. Women’s Pre-Birth Relative Wage Potential

The main independent variable of interest is women’s pre-birth relative wage potential. Since using partners’ observed wages in a specific quarter before first childbirth to calculate women’s pre-birth relative wage implies considering couples where both partners are employed, this may result in a highly selective subgroup, especially for migrant origin couples (with a non-EU background) with generally lower employment levels. Moreover, using partners’ observed wages at a specific moment before first childbirth may be problematic for (migrant origin) couples with less stable employment positions (Maes et al. 2021). Therefore, this study introduces the concept of partners’ pre-birth relative wage potential. This refers to the hourly wage that individuals would earn if they would have no children and would be employed, based on their socio-demographic profile. Women’s pre-birth relative wage potential is calculated in three stages. First, for all employed men and women in the MIA Panel who did not (yet) have children, we estimated in each quarter the hourly wage (i.e., the gross income divided by the work intensity) as a function of their (i) age (centred at age 18, quadratic specification), (ii) highest educational level (low, medium, high, unknown), (iii) origin group and generation (native 1G South-EU, 2G South-EU, 1G Turkey, 2G Turkey, 1G Maghreb, 2G Maghreb) and (iv) LIPRO household position11 (living in parental home, single, married, cohabiting, other), as well as (v) the interaction between origin group and age, (vi) the interaction between origin group and educational level, (vii) the interaction between origin group and LIPRO position, (viii) the interaction between educational level and age and (ix) the interaction between LIPRO position and age (model estimates12 for men and women are included in

Table A2. Regression model for estimating the hourly wage for women and men who have no children and are employed.

	Women		Men
	Coef.	Sig.	Coef.	Sig.
Age (centred at 18)	2.570	***	2.425	***
Age2 (centred at 18)	−0.038	***	−0.045	***
Educational level (ref. low)
medium	10.580	***	13.007	***
high	21.947	***	16.787	***
unknown	22.852	***	20.918	***
Origin group (ref. native)
Maghreb 1G	14.808	***	14.594	***
Maghreb 2G	6.967	***	7.747	***
Turkey 1G	14.699	***	15.151	***
Turkey 2G	8.482	***	10.463	***
South-EU 1G	7.462	***	8.941	***
South-EU 2G	7.055	***	9.062	***
LIPRO position (ref. living in parental home)
single	6.615	***	0.412	n.s.
married	2.606	***	0.494	n.s.
cohabiting	5.770	***	1.063	**
other	5.586	***	2.622	***
Origin group * age
Maghreb 1G	−0.551	***	−0.850	***
Maghreb 2G	−0.151	***	−0.346	***
Turkey 1G	−0.800	***	−1.063	***
Turkey 2G	−0.387	***	−0.612	***
South-EU 1G	−0.333	***	−0.647	***
South-EU 2G	−0.442	***	−0.614	***
Origin group * educational level
Maghreb 1G * medium	−5.411	***	−4.753	***
Maghreb 1G * high	−22.351	***	−30.509	***
Maghreb 1G * unknown	−22.971	***	−21.200	***
Maghreb 2G * medium	−4.177	***	−5.358	***
Maghreb 2G * high	−8.556	***	−9.602	***
Maghreb 2G * unknown	−10.679	***	−10.746	***
Turkey 1G * medium	−4.972	***	−4.916	***
Turkey 1G * high	−11.035	***	−26.215	***
Turkey 1G * unknown	−31.371	***	−23.063	***
Turkey 2G * medium	−4.612	***	−5.346	***
Turkey 2G * high	−9.111	***	−9.815	***
Turkey 2G * unknown	−14.159	***	−13.607	***
South-EU 1G * medium	1.215	n.s.	−3.257	***
South-EU 1G * high	−8.423	***	−8.816	***
South-EU 1G * unknown	−19.771	***	−20.901	***
South-EU 2G * medium	−3.751	***	−4.973	***
South-EU 2G * high	−2.526	***	−8.357	***
South-EU 2G * unknown	−12.022	***	−11.106	***
Origin group * LIPRO
Maghreb 1G * single	−7.325	***	−5.484	***
Maghreb 1G * married	−9.373	***	−8.438	***
Maghreb 1G * cohabiting	−7.247	***	−3.841	***
Maghreb 1G * other	−4.757	***	−4.531	***
Maghreb 2G * single	−2.856	***	−0.040	n.s.
Maghreb 2G * married	−1.454	**	−2.196	***
Maghreb 2G * cohabiting	−2.864	***	0.885	n.s.
Maghreb 2G * other	4.471	***	1.317	n.s.
Turkey 1G * single	−2.660	*	−2.412	*
Turkey 1G * married	−7.243	***	−7.121	***
Turkey 1G * cohabiting	−1.123	n.s.	−2.467	*
Turkey 1G * other	−5.312	***	−3.176	***
Turkey 2G * single	−0.712	n.s.	−1.724	n.s.
Turkey 2G * married	−4.550	***	−2.409	***
Turkey 2G * cohabiting	−1.545	n.s.	0.550	n.s.
Turkey 2G * other	−1.249	n.s.	−0.304	n.s.
South-EU 1G * single	−2.763	*	6.027	***
South-EU 1G * married	−3.764	***	−2.188	*
South-EU 1G * cohabiting	7.259	***	3.984	***
South-EU 1G * other	−5.425	***	0.419	n.s.
South-EU 2G * single	5.802	***	5.965	***
South-EU 2G * married	−1.467	**	−1.089	*
South-EU 2G * cohabiting	4.793	***	4.138	***
South-EU 2G * other	0.249	n.s.	2.284	***
Educational level * age
medium	−0.066	***	−0.098	***
high	0.798	***	1.353	***
unknown	0.539	***	0.628	***
LIPRO * age
single	−0.531	***	0.071	**
married	−0.455	***	0.382	***
cohabiting	−0.573	***	0.157	***
other	−0.767	***	−0.280	***
Constant	18.975	***	23.334	***
N	649,582		916,376
R2	0.18		0.22

Significance levels: * p < 0.05; ** p < 0.01; *** p < 0.001; n.s. not significant. Source: MIA Panel 2005–2016, calculations by authors.

in Appendix A)13. Second, in the fourth quarter before the birth of their first child, individuals in the analytical sample of interest in this paper are assigned the estimated wage potential of individuals who do not have children (yet) but who otherwise have an identical profile in terms of the aforementioned characteristics (i.e., matching for the covariates included in the model in stage 1). For employed men and women, the correlations between the observed and estimated hourly wages in the fourth quarter before the birth of their first child are r = 0.478 (p < 0.001) and r = 0.344 (p < 0.001), respectively. Third, we calculated women’s pre-birth relative wage potential as the ratio of the female partner’s wage potential to the sum of the male and female partner’s wage potential in the fourth quarter before first childbirth. We distinguished three categories: (i) couples where women have a lower pre-birth wage potential than their male partner (i.e., 0–44% of the total household wage potential), (ii) couples where women have a similar pre-birth wage potential (i.e., 45–48% of the total household wage potential) and (iii) couples where women have a higher pre-birth wage potential than their male partner (i.e., 49–100% of the total household wage potential). This estimation is subsequently used to address the extent to which changes in couples’ gender division of paid work around the transition to parenthood differ according to women’s pre-birth relative wage potential.

3.2.2. Modelling Strategy

To examine the link between women’s pre-birth relative wage potential and couples’ gender dynamics around family formation, we estimated couple-level fixed-effects models that address changes in couples’ gender division of paid work from one year before up to three years after the birth of their first child. Equation (1) shows the general equation for a fixed-effects model, where $Y_{it}$ denotes the dependent variable for couple $i$ at time $t$, $X_{it}$ is a vector of time-varying independent variables, ${\beta }_t$ are the parameter estimates for these time-varying independent variables, ${\alpha }_i$ is the time-invariant couple fixed effect and $u_{it}$ is the couple-level residual at time $t$. Hence, since the couple-level fixed-effects models only consider variation within couples over time, the analyses account for time-constant (un)observed heterogeneity between couples (Allison 2009; Stock and Watson 2015).

$$Y_{it}=\sum {\beta }_t{X}_{it}+{\alpha }_i+u_{it}$$

(1)

The dependent variable reflects women’s relative work intensity. In our dataset, work intensity reflects the percentage of contractual working hours compared to a full-time position in the employment sector considered. Although working hours differ between employment sectors14, work intensities provide an indication of variation in working hours around family formation. Unfortunately, the absolute number of working hours is not available in the data, which prevents sensitivity checks in this respect. The possible values of individuals’ work intensity range from 0% to 100%, where 0% reflects unemployment or inactivity and 100% reflects full-time employment15. Part-time jobs are combined to determine the total work intensity. Since our data do not provide the exact number of working hours, the work intensity of individuals that exceeds the standard number of working hours for a full-time position is considered 100%. The work intensity of women on maternity leave amounts to 0% and that of women on parental leave reflects their reduction in working hours. In turn, women’s relative work intensity is calculated as the ratio of the female partner’s work intensity to the sum of the male and female partner’s work intensity. The possible values of women’s relative work intensity range from 0% to 100%, where 0% refers to a situation in which only the male partner is employed, 50% reflects equal work intensity among both partners and 100% implies that only the female partner is employed.

We estimated two couple-level fixed-effects models to examine changes in women’s relative work intensity around the birth of their first child. Model 1 includes (i) time relative to the first birth (distinguishing quarters −4, −3, −2, −1, 0 1 2, 3, 4 to 7, 8 to 11), (ii) the interaction between time relative to the first birth and couples’ migration background (native, South-EU, non-EU) to assess whether the change in women’s relative work intensity around family formation differs by migration background and (iii) the interaction between time relative to the first birth and women’s pre-birth relative wage potential (lower, similar, higher) to address whether the change in women’s relative work intensity around the transition to parenthood differs by women’s pre-birth relative wage potential. The fourth quarter before the birth of the child is used as the reference category, implying that women’s relative work intensity in each quarter is compared to their relative work intensity one year before the birth of their first child. Besides partners’ pre-birth relative wages, the available literature indicates that also absolute wages are important determinants for couples’ gender dynamics around family formation as couples with a higher wage potential have higher opportunity costs of reducing working hours and this also affects couples’ ability to outsource childcare and domestic tasks (Schober 2013). Therefore, Model 1 additionally includes the interaction between time relative to the first birth and couples’ pre-birth income potential at the household level. This variable reflects what the household income would be if a couple would have no children and both partners would be full-time employed, based on the partners’ estimated pre-birth wage potential16. We distinguished couples with (i) a low pre-birth income potential (i.e., EUR 5000—EUR 12,499 per quarter), (ii) a medium pre-birth income potential (i.e., EUR 12,500—EUR 14,499 per quarter), and (iii) a high pre-birth income potential (i.e., EUR 14,500—EUR 19,500 per quarter).

In order to address whether the link between women’s pre-birth relative wage potential and couples’ gender division of paid work around the transition to parenthood differs by migration background, Model 2 additionally includes the three-way interaction between time relative to the first birth, women’s pre-birth relative wage potential and the couples’ migration background. Finally, given that changes in women’s relative work intensity may occur due to changes in both male and female partners’ work intensity, similar models with either women’s work intensity or men’s work intensity as dependent variable are estimated to examine the underlying gender dynamics.

4. Results

4.1. Variation by Migration Background in Women’s Pre-Birth Relative Wage Potential

Figure 1a shows the distribution of women’s pre-birth relative wage potential by couples’ migration background and shows clear differences between the three origin groups considered in this study. Whereas women have a pre-birth wage potential similar to their partner among half of native couples (49%), this amounts to 30% among both South-EU and non-EU origin couples. In turn, compared to native couples, there is a larger proportion of South-EU and non-EU origin couples where the female partner has either a higher or lower pre-birth relative wage potential. Furthermore, Figure 1b shows that the distribution of women’s pre-birth relative wage potential among South-EU and non-EU origin couples varies depending on the origin group and generation of both partners. Regarding South-EU origin couples, we find that a smaller proportion of mixed-origin couples where women have a similar pre-birth wage potential implies a larger proportion of couples where the native partner has a higher pre-birth wage potential. The distribution of South-EU origin couples with two second-generation partners is largely similar to that of native couples. With respect to non-EU origin couples, women have a lower pre-birth wage potential among half of non-EU couples where both partners are of the second generation (54%), while women have a higher pre-birth wage potential among 45% of non-EU origin couples with a second-generation woman and a first-generation man.

4.2. Changes in Couples’ Gender Division of Paid Work around the Transition to Parenthood by Women’s Pre-Birth Relative Wage Potential

Figure 2 displays the results of the couple-level fixed-effects analyses (Model 2), showing, for each origin group, women’s average relative work intensity one year before first childbirth and the average change in women’s relative work intensity around the transition to parenthood by women’s pre-birth relative wage potential. As changes in couples’ gender division of paid work may occur due to changes in both male and female partners’ work intensity, Figure A2, Figure A3 and Figure A4 in Appendix A show, for each origin group, the average development in work intensity for men and women separately by women’s pre-birth relative wage potential. The likelihood ratio test comparing Models 1 and 2 indicates that the impact of women’s pre-birth relative wage potential on the change in women’s relative work intensity around the transition to parenthood differs significantly by couples’ migration background. The following sections discuss the link between women’s pre-birth relative wage potential and couples’ gender dynamics around the transition to parenthood for native, South-EU origin and non-EU origin couples.

4.2.1. Native Couples

In line with Hypothesis 1a, Figure 2a shows that native couples where women have a lower pre-birth wage potential than their male partner display, on average, a slightly less equal division of paid work after the transition to parenthood compared to couples where women have a similar or higher wage potential. This is mainly due to the fact that women with a similar or higher pre-birth relative wage potential have, on average, a higher work intensity after family formation compared to women with a lower pre-birth relative wage potential, as differences in men’s work intensity depending on partners’ pre-birth relative wage potential are relatively limited (cf. Figure A2 in Appendix A).

Regarding changes in native couples’ gender division of paid work around family formation, we find that women’s relative work intensity drops to low values during the quarter of first childbirth (but also slightly during the quarter before and after childbirth) since women are on maternity leave, and recovers as the child becomes older17, but is slightly lower compared to one year before the transition to parenthood. This indicates that gender inequality in paid work increases after the birth of the first child among native couples. Figure A2 in Appendix A shows that this decrease in women’s relative work intensity after the transition to parenthood is primarily due to a decrease in women’s work intensity rather than an increase in men’s work intensity. However, we find no differential change in women’s relative work intensity depending on their pre-birth relative wage potential, as the same overall pattern can be observed. Hence, we cannot accept Hypothesis 1b, which suggests that gender inequality in native couples’ division of paid work would increase to a lesser extent among couples where women have a higher relative wage potential before family formation. In contrast, our results indicate that differences in native couples’ gender division of paid work by women’s pre-birth relative wage potential after the birth of their first child already existed prior to family formation and that these differences remain largely similar around the transition to parenthood.

4.2.2. South-EU Origin Couples

Similar to native couples, Figure 2b shows that South-EU origin couples where women have a lower pre-birth wage potential than their male partner display, on average, a less equal gender division of paid work after the transition to parenthood (women’s relative work intensity amounting, on average, to 42% at t + 3) compared to couples where women’s pre-birth relative wage potential is similar or higher (on average, 48% at t + 3). Figure A3 in Appendix A indicates that this is mainly due to the fact that women’s work intensity is, on average, lower when women have a lower wage potential than their male partners prior to childbirth, as men’s work intensity is largely similar regardless of partners’ pre-birth relative wage potential. This pattern is similar to native couples and we can therefore accept Hypothesis 2a.

When considering changes in South-EU origin couples’ gender division of paid work around family formation, we find that women’s relative work intensity slightly decreases after the birth of the first child, but there are no statistically significant differences by women’s pre-birth relative wage potential. Hypothesis 2b should therefore be rejected. Similar to native couples, the differential gender division of paid work by women’s pre-birth relative wage potential that can be observed among South-EU origin couples after the transition to parenthood already existed prior to family formation and is not the result of differential changes in couples’ gender division following family formation. In sum, we find overall no significant differences between native and South-EU origin couples in the link between women’s pre-birth relative wage potential and couples’ gender dynamics around family formation.

4.2.3. Non-EU Origin Couples

Figure 2c shows that non-EU origin couples’ gender division of paid work after family formation varies strongly depending on women’s pre-birth relative wage potential. Among non-EU origin couples where women have a higher pre-birth relative wage potential, women’s relative work intensity amounts, on average, to 44% three quarters after first childbirth, which is significantly higher compared to non-EU origin couples where women have a similar or lower pre-birth relative wage potential (on average, 32% and 28% at t + 3). This strong variation by women’s pre-birth relative wage potential is not only due to a higher work intensity for women who have a higher pre-birth relative wage potential (on average, 61% at t + 3) compared to women with a similar or higher wage potential (on average, 47% and 40%) (cf. Figure A4 in Appendix A). Men’s work intensity is also lower in cases where their partner has a higher pre-birth wage potential (on average, 78% at t + 3). Moreover, women’s relative work intensity after childbirth is significantly lower among non-EU origin couples than among native and South-EU origin couples where women have a lower pre-birth wage potential than their partner (cf. Figure A5 in Appendix A). Figure A1 in Appendix A shows that this is due to a large share of couples where only the male partner is employed after family formation (44% at t + 4). Our results furthermore indicate that this strong degree of gender inequality in paid work among non-EU origin couples where women have a lower pre-birth relative wage potential already existed prior to family formation, which confirms Hypothesis 3a.

With respect to changes in couples’ gender division after the transition to parenthood, non-EU origin couples where women have a higher or similar pre-birth relative wage potential display a stronger decrease in women’s relative work intensity after family formation compared to non-EU origin couples where women have a lower pre-birth relative wage potential (on average, 11 versus 7 percentage points at t + 2). As a result, although non-EU origin couples where women have a higher pre-birth relative wage potential still exhibit the highest degree of gender equality, differences in non-EU origin couples’ gender division of paid work by women’s relative wage potential become smaller after family formation. Furthermore, in line with Hypothesis 3b, we find that non-EU couples where women have a higher or similar pre-birth relative wage potential show a stronger increase in gender inequality following family formation compared to native couples (cf. Figure A5 in Appendix A). This results from the combination of a stronger decrease in women’s work intensity and a stronger increase in work intensity for men whose partner has a higher relative wage potential (cf. Figure A4 in Appendix A).

4.3. Sensitivity Analyses

In order to assess the robustness of the results, we performed two sensitivity analyses. First, we used alternative cut-off points for the categories of women’s pre-birth relative wage potential. In the main analyses, we distinguished couples where women’s pre-birth wage potential is lower, similar or higher than their male partner, with ranges of 0–44%, 45–48% and 49–100% of the total household wage potential, respectively. Narrowing the highest category by considering the categories 0–44%, 45–49% and 50–100% generates largely the same results among native, South-EU and non-EU couples. In addition, narrowing the lowest category by distinguishing the categories 0–43%, 44–48% and 49–100% also yields consistent results with the main analyses. Second, the main analyses include the interaction between time relative to the first birth and couples’ pre-birth income potential at the household level, distinguishing couples with a low, medium and high pre-birth income potential. As a robustness check, we replicated the models using a continuous variable for couples’ pre-birth income potential, centred around the mean household income potential one year before first childbirth (i.e., EUR 12,050). These alternative models yield the same results as the categorical approach for couples’ income potential.

5. Discussion

Although couples increasingly divide paid and unpaid work equally, this division is not stable over the life course and particularly the birth of the first child often entails a shift towards higher gender inequality (Kreyenfeld 2015; Kuhhirt 2011; Schober 2013; Wood et al. 2018). The (change in the) gender division of paid work within couples after the transition to parenthood, as well as the role of economic/ideational/institutional determinants of these couple-level gender dynamics, has been increasingly studied among general populations (Begall and Grunow 2015; Kanji 2011; Kuhhirt 2011; Vignoli et al. 2012; Wood et al. 2018). However, variation by couples’ migration background has hitherto only been examined to a limited extent (Maes et al. 2022), let alone potential underlying mechanisms. As a result, this paper adopts a life course perspective and explores variation by migration background in the linkage between women’s pre-birth relative wage potential and couples’ gender division of paid work from one year before up to three years after the transition to parenthood in Belgium. We focus on the origin groups that resulted from the large post-WWII waves of labour migration and distinguish native Belgian couples and couples with at least one second-generation Southern European and non-European origin partner.

In line with the life course principle of path dependency, our results indicate that couples’ gender division of paid work after the transition to parenthood is related to partners’ relative wages prior to childbirth. In accordance with expectations derived from economic theories (Becker 1991; Lundberg and Pollak 1996), we find, among all origin groups considered in this study, that the division of paid work is more gender-equal after family formation when women have a pre-birth wage potential higher than or similar to that of their male partner compared to couples where the female partner has a lower pre-birth wage potential. Hence, we can confirm H1a and H2a as native and South-EU origin couples where women have a lower pre-birth wage potential than their male partner display, on average, a less equal gender division of paid work after the transition to parenthood. The degree of gender inequality in paid work is particularly high among non-EU origin couples where the female partner has a lower relative wage potential, which is due to a large share of couples where only the male partner is employed. Moreover, in line with our expectations (H3a), our results indicate that this strong degree of gender inequality already existed prior to family formation. This may suggest that non-EU couples where women have a lower relative wage potential before childbirth may be a selective group with more traditional work–family attitudes (as a result of lower labour market opportunities and/or differential socialisation contexts), which entails an alternative selection into parenthood and, in turn, a highly unequal gender division of paid work after childbirth (Friedman et al. 1994; Wood and Neels 2017).

Consistent with previous research, we find that gender inequality in couples’ division of paid work increases after the transition to parenthood. However, in contrast to our expectation that gender inequality in native and South-EU origin couples’ division of paid work would increase to a lesser extent among couples where women have a higher pre-birth relative wage potential (H1b and H2b), the increase in gender inequality in couples’ division of paid work around family formation is similar for all native and South-EU origin couples, regardless of women’s pre-birth relative wage potential. This finding is consistent with the results of Schober (2013) for the UK, but contrasts with previous studies on general populations for West Germany (Kuhhirt 2011) and the UK (Kanji 2011) indicating that women who earned the same wage or more than their partner before childbirth were less likely to decrease their working hours after childbirth. Following the life course principles of time and place and bounded agency, the finding that, even among female main earner couples, women’s work intensity decreases to a stronger extent compared to men suggests that ideational (e.g., the persistence of gendered parenting norms) as well as institutional (e.g., gendered leave system) factors moderate the impact of women’s relative resources on couples’ division of paid work after family formation, and to the same extent among native and South-EU groups.

Furthermore, in line with Hypothesis 3b, we find a stronger increase in gender inequality among non-EU origin than native couples where women have a higher pre-birth relative wage potential. These non-EU couples mainly consist of a second-generation woman and a first-generation man and while previous research suggests that marrying a partner from their country of origin is, for second-generation Turkish and Moroccan origin women, a way to bend traditionally gendered relations (Timmerman et al. 2009), our findings indicate that the transition to parenthood remains a critical juncture for the development of a gendered division of paid work. Whereas women’s relative work intensity is slightly higher prior to family formation among non-EU couples than among native or South-EU origin couples where women have a higher pre-birth relative wage potential (as a result of a larger share of couples where only the female partner is employed), it decreases to a stronger extent following the transition to parenthood. This suggests that these non-EU couples experience (a strong pressure to conform to) more traditional parenting norms compared to native couples, which potentially moderates an economic cost–benefit rationale in their work–family organisation. In this respect, research among the Turkish and Moroccan second generation in Belgium suggests that it is no longer considered desirable to work within the Turkish or Moroccan community once women have children (Adam and Torrekens 2015). Based on the available literature, it is unfortunately unclear to what extent gender role expectations change around parenthood and whether this differs from natives (de Valk 2008). Furthermore, this stronger increase in gender inequality among non-EU couples than native couples where women have a higher pre-birth relative wage potential may also result from less access to parental leave and formal childcare, as these policies in Belgium primarily support parents who are firmly established in the labour market (Kil et al. 2017; Marynissen et al. 2021; Vandenbroeck et al. 2008). Given that particularly Turkish and Maghreb origin groups exhibit more precarious labour market positions in Belgium (FOD WASO and UNIA 2019; Maes et al. 2019), the options to combine work and family life might be more limited among these couples, more often yielding work–family reconciliation strategies that involve reduced labour market participation.

In sum, this paper sheds light on the role of economic mechanisms in the gender division of paid work after family formation among couples with different migration backgrounds. It highlights that partners’ relative wages prior to childbirth influence couples’ gender division of paid work after the transition to parenthood both among native and migrant origin couples. However, our results also suggest that normative as well as institutional factors moderate the impact of partners’ relative wages on couples’ gender division of paid work after family formation, particularly among non-EU origin groups. These findings imply that, besides differential employment opportunities, also differential parenting norms are important mechanisms behind varying gender dynamics around family formation between native and second-generation couples. As such, this paper contributes to a better understanding of (changes in) couples’ gender division of paid work after the transition to parenthood and how they vary by migration background, providing essential insights to develop specific policies that can enhance the labour market participation of migrant origin women. The latter is crucial for both societies as a whole and migrant women themselves: in the context of accelerated population ageing, increasing the labour market participation of the large and growing share of women with a migration background is considered as increasingly important for covering welfare state costs (e.g., pensions or health care) and labour shortages. Furthermore, increasing gender inequality in couples’ division of paid work following the transition to parenthood can jeopardise women’s financial independence, may increase poverty risks at the household level and has long-term implications for future labour market opportunities and social security protection (e.g., pensions) at later stages of the life course.

Next to these contributions, a number of limitations can be identified, which in turn offer fruitful paths for future research. First, this study only includes couples who have their first child during the observation period and does not compare their gender dynamics to those of couples who do not (yet) have children. As a result, potential selection into parenthood remains unaccounted for. While prior research suggests that the employment–fertility link varies by couples’ migration background (Wood et al. 2017), future studies should address how partners’ relative resources affect the link between couples’ gender division of paid work and their transition to parenthood among couples with different migration backgrounds. Second, in interpreting our findings, it is difficult to distinguish the role of economic and ideational mechanisms as they are strongly interrelated (differential labour market outcomes may occur as a result of differential gender role attitudes or vice versa). Complementing the rich register data used in this study with a longitudinal measurement of attitudes among native and migrant origin groups would be a particularly interesting avenue in that respect, as would a mixed-method approach. Finally, future studies could examine whether similar patterns could be observed in other European countries. The specific institutional context of Belgium shapes how couples with different migration backgrounds organise their work and family life and is likely to be crucial for the interpretation of our results. Since European countries vary in the extent to which policy designs challenge particular gender norms and imply subgroup differences in the access to these policies, comparing different countries could provide fruitful insights into the impact of policy designs on variation by migration background in couples’ gender dynamics around family formation (Huschek et al. 2011; Mussino and Duvander 2016; Sainsbury 2019). In Belgium, access to work–family reconciliation policies is strongly conditioned on stable employment, which contrasts with universal access in Nordic countries or policies in favour of the one-and-a-half-earner model (e.g., Germany) and may have influenced the results. For instance, studies for Sweden have found similar employment–fertility patterns for natives and migrants (Lundström and Andersson 2012), which have been associated with the universal and inclusive Swedish welfare regime. Moreover, comparative studies highlight the importance of policies in shaping the division of (un)paid work of second-generation migrants (Huschek et al. 2011). Yet, it remains unclear whether and to what extent changes in couples’ gender division of paid work around family formation vary by migration background across European countries, let alone how this is related to women’s pre-birth relative wage potential. In addition, variation by migration background in the link between women’s pre-birth relative wage potential and couples’ gender dynamics around family formation is likely to differ across European countries as a result of countries’ migration histories that may have induced specific origin groups, as well as specific socio-economic and ideational contexts among these groups.