Gendered division of housework and childcare and women’s intention to have a second child in Spain

Abstract

The Gender Revolution Theory forecasts a rise in fertility as gender equality increases and permeates the private sphere. However, empirical evidence across different societies has not always been conclusive, which suggests further research on the relationship between gender equality and fertility is needed. This research aims to assess the impact of the distribution of housework and childcare within the couple on one-child mothers’ fertility intentions in Spain. The educational expansion in recent decades has facilitated women’s mass entry into the workforce in this country. Nevertheless, the distribution of unpaid labour remains quite gendered and public support for families is scarce, making family/work balance challenging, especially for women. Using data from the 2018 Spanish Fertility Survey, the results suggest that when the distribution of housework is highly unbalanced between mothers and fathers, women are less likely to intend to have a second child. In contrast, unequal childcare workload does not substantially impact their intention to have a second child. Furthermore, whereas women’s satisfaction with the distribution of housework and childcare does not seem to moderate the effect of distribution, their education level slightly does. University-educated women are more likely to revise their fertility intentions downwards when the distribution is highly unequal.

Introduction

Over the last decade, the relationship between gender equality and fertility has been largely explained by the Gender Revolution Theory (Goldscheider et al., 2015). This theory predicts an increase in fertility when gender equality permeates the private sphere and men take a more active role in housework and childcare responsibilities (Hochschild & Machung, 1989; McDonald, 2000, 2013). The Gender Revolution Theory relied on the experience of the Nordic countries to illustrate the positive association between gender equality within the family sphere and fertility (Duvander et al., 2019). These countries recorded the highest levels of women’s participation in the labour force for decades, characterised by a more equal sharing of household responsibilities than other countries, and policies that promoted active fatherhood (Andersson et al., 2009; Duvander et al., 2019; Esping-Andersen, 2009). Moreover, they did not experience the very low fertility rates prevailing in other European countries (Anderson & Kohler, 2015). Nevertheless, around 2010, fertility levels in the Nordic countries began to decline without undergoing a parallel reduction in gender equality (Hellstrand et al., 2020, 2021; Jalovaara et al., 2019). This shift in trends poses a challenge to gender theories of family change and calls for further research on the relationship between gender equality and fertility (Hellstrand et al., 2021), particularly in other countries that are at different stages of the Gender Revolution.

The Spanish context is of particular interest in assessing this relationship. It is a lowest-low fertility context (Kohler et al., 2002) situated in an intermediate stage of the Gender Revolution: recent decades have seen the proportion of women with tertiary education increase considerably (Ortiz & Rodriguez-Menés, 2016), with an ensuing mass entry into the labour force (Guinea-Martin et al., 2018). However, the distribution of unpaid labour within couples continues to rest largely on the shoulders of women (Abril et al., 2015; Moreno-Colom, 2017).

Previous quantitative research conducted in the 2000s on gender equality and fertility in Spain found that the division of childcare between women and men had no effect on fertility (Brodmann et al., 2007; Cooke, 2009), but Spanish society has undergone profound transformations since then, including rapid changes in family dynamics (Borràs et al., 2021; Bueno, 2020; Domínguez-Folgueras & Castro-Martín, 2013; García-Román, 2021). However, despite the relevance of the Spanish case, there have been no quantitative data in the last two decades to appropriately assess the relationship between gender equality within the family and fertility. Using data from the 2018 Spanish Fertility Survey (SFS), this study aims to fill this gap in the literature by focusing on the relationship between the distribution of housework and childcare within couples and one-child mothers’ fertility intentions. The analysis also explores the interaction between the allocation of household chores (objective component) and the satisfaction with that distribution (subjective component), as well as the moderating role of women’s educational level.

Theoretical framework and hypotheses

Theoretical approaches to the link between the gendered distribution of unpaid labour and fertility

In most high-income societies, gender equality in the public sphere has notably increased. Recent decades have seen the proportion of women with tertiary education grow substantially, surpassing men’s educational attainment. Consequently, the proportion of hypogamous couples is also increasing (Esteve et al., 2016b; Klesment & van Bavel, 2017; van Bavel et al., 2018). As the proportion of women participating in the labour force has also markedly increased, dual-earner couples have become the most common form of family (García Román, 2013).

Traditional approaches, such as the New Home Economics (NHE), have argued that fertility is relatively high when there is a division of paid and unpaid work between men and women. When women enter the workforce, fertility falls due to the difficulties of balancing household and paid work, and the opportunity costs that childbearing places on them (Becker, 1993). The Second Demographic Transition (SDT) theory (Lesthaeghe, 1995) also anticipates a negative relationship between gender equality and fertility. According to the SDT, a long-lasting shift in family values has taken place in Europe since the late 1960s, resulting in the weakening of the family as an institution (Sobotka, 2008; Van de Kaa, 2002). One of the consequences of this is a decline in fertility (Lesthaeghe, 2010).

In contrast, the Gender Revolution Theory (GRT) (Goldscheider et al., 2015) and the Multiple Equilibrium Framework (MEF) (Esping-Andersen & Billari, 2015) forecast a reversal of fertility trends as gender equality progresses. The GRT argues that fertility initially declines in the first part of the Gender Revolution as gender equality increases in society. Women attain higher levels of education, enter the workforce, and begin to face the double burden of domestic and paid work responsibilities (Hochschild & Machung, 1989). However, in the second part of the Gender Revolution, fertility is expected to rise again as gender equality increases in the family sphere and men shoulder a greater share of domestic responsibilities, relieving women of this the double burden (Goldscheider et al., 2015).

The MEF also argues that fertility is likely to increase as gender attitudes and the division of labour become more egalitarian. According to this framework, fertility declines in the transitional stage going from a family equilibrium characterised by gender specialisation to a new equilibrium where both men and women share responsibilities inside and outside the home (Esping-Andersen & Billari, 2015).

All the approaches mentioned above are consistent with the initial stage of the Gender Revolution: fertility declines as women enter the public spheres of education and employment. They disagree, however, on what comes next: a fertility reversal, a stabilisation at very low level or a further decline (Lesthaeghe, 2020). Given that no society has achieved a very high level of gender equality in the domestic sphere so far, the ultimate implications of the changing roles of women and men for fertility are still uncertain (Lappegård, 2020).

Additional factors, therefore, need to be determined in the study of the fertility decision-making process, aside from gender (in)equality. Firstly, male fertility intentions should be taken into consideration (Testa & Bolano, 2021). The couple usually make a joint decision to have a child (Hollerbach, 1980). Although there may be disagreements (Duvander et al., 2020), men’s and women’s intentions influence each other (Thomson, 1997). If men take on more household responsibilities, they may also endure the ‘double shift’ and reduce their fertility intentions (Okun & Raz-Yurovich, 2019), influencing women’s intentions. Secondly, distinct socioeconomic groups react differently to the institutional context and may obtain unique benefits from social policies. Low- and high-educated individuals may face different barriers when forming a family (van Bavel, 2012). Therefore, these differences need to be considered to understand fertility patterns and trends (Lappegård, 2020). Lastly, the intersection of the context and the micro–macro level are important here (Lappegård & Kornstad, 2019). The level of gender equality within a society may affect the relationship between gender equality and fertility in the private sphere and vice versa. In this respect, public policies and social norms play an essential role as instruments that may influence family dynamics and gender-role transitions (Oláh et al., 2021).

Research on gender equality and fertility at the micro-level

A growing number of studies have assessed whether couples in fairer arrangements where men are more involved in family responsibilities have different fertility preferences and behaviours than less egalitarian couples. Empirical results have not been entirely conclusive and vary to a large extent according to the context.

In the Nordic countries, most studies have shown a positive relationship between gender-egalitarian behaviours within the family and fertility. In Norway, a longitudinal study found that an unbalanced distribution of housework between men and women (either when men or women do more) was associated with a lower probability of having a(nother) child. For the transitions to the first and second child, that probability declined when men did a higher share of housework than their partners. In contrast, for the transition to the third child, couples where women took on a greater burden of care had a lower likelihood of enlarging their families. Moreover, among couples with one child, women satisfied with the division of childcare were more prone to have a second child (Dommermuth et al., 2017). Positive effects of fathers’ involvement in childcare on fertility were also found in Sweden (Duvander et al., 2010; Oláh, 2003). In Finland, men’s participation in housework did not seem to increase the likelihood of a couple having a second child, whereas involvement in childcare did (Miettinen et al., 2015).

Studies focused on Southern Europe, conversely, have shown mixed results. A comparison between Italy and Spain found that, while the degree of fathers’ involvement in childcare had no effect on the probability of having a second child in Spain, in Italy there was an inverted U-shaped relationship among dual-earner couples (Cooke, 2009). Comparing Denmark and Spain, Brodmann et al. (2007) found that fathers’ involvement in childcare had no effect on the likelihood of having a second child in Spain. In Italy, an analysis of second and third birth transitions showed that the contribution of fathers in childcare increased the probability of a second birth, but their participation in housework did not (Mencarini & Tanturri, 2004). When analysing women’s fertility intentions to have a second child and the realisation of those intentions in Italy, Rinesi and colleagues showed that there was no positive effect of fathers’ involvement in housework (Rinesi et al., 2011). For the same context, Fiori (2011) indicated a weak positive impact of paternal involvement in childcare and housework on women’s intention to have a second child. The effect of men’s participation in household responsibilities may also depend on women’s employment status. A previous study found that it increased the intention to have a second child only among working mothers (Pinnelli & Fiori, 2008).

In other countries in Europe, the results have also been inconsistent. In Germany, a U-shaped relationship was identified between fathers’ involvement in childcare and the transition to a second child. In contrast, no effect was observed regarding male participation in housework (Cooke, 2004). A recent study in this country showed a positive association between women’s satisfaction with the division of housework and the probability of having first and second births (Köppen & Trappe, 2019). In the United Kingdom, an inverted U-shaped relationship was observed between the share of housework among one-child mothers and the transition to the second child (Schober, 2013).

Looking outside of Europe, a U-shaped relationship was found in the United States, as the most traditional and egalitarian couples were the most likely to have a second child (Torr & Short, 2004). Similar results were reported for Australia, considering the actual sharing of housework and satisfaction with the division of unpaid labour (Luppi, 2016).

The Spanish case

The Spanish context represents an insightful case study to examine fertility trends in general, and the relationship between gender equality and fertility in particular. Spanish fertility has been at very low levels for a long time (Billari & Kohler, 2004; Castro-Martín & Martín-García, 2013; Kohler et al., 2002), with total fertility rates below 1.4 children per woman for the last three decades, driven mainly by relatively high rates of childlessness and a low rate of parity progression to two children (Esteve et al., 2016a). Cohort fertility rates have also been below 1.4 for all cohorts born after 1965 (Esteve et al., 2021). Moreover, the age at first motherhood is one of the highest in the world after Italy. In 2018, the average age at first birth among women was 31, while the average age in the European Union (EU-28) was 29.2 (Eurostat, 2018).¹ Unfavourable economic and labour conditions including high unemployment rates, precarious employment, difficulties in balancing work and family and inadequate access to affordable housing have influenced Spanish fertility decisions for decades (Adsera, 2011; Bernardi & Requena y Díez de Revenga, 2003; Castro-Martín & Martín-García, 2016).

Spain is a familistic country where childcare support from the State is limited (Esping-Andersen, 2009). The public expenditure on family benefits is among the lowest of the OECD countries: 1.3% of the GDP, while the average of the OECD is 2.3% (OECD, 2017). The role of grandparents as informal caregivers has therefore become essential (Rutigliano & Lozano, 2022). Pre-school enrolment rates in 2018 were 12% among children younger than 1 year old, 40% among 1-year-olds and 60% among 2-year-olds (OECD Publishing, 2020). However, public pre-school provision for children under the age of 3 does not meet the demand. Despite these adverse childrearing conditions, fertility ideals have not decreased overall; they have remained around two children per woman, leading to a wide gap between desired and actual fertility (Castro-Martín & Martín-García, 2013).

Women’s education levels in Spain have increased considerably in recent decades (Ortiz & Rodríguez-Menés, 2016). The percentage of individuals with tertiary education in 2018 was 50.1% among women and 38.4% among men (Eurostat, 2018). In parallel, gender-egalitarian values have become widespread (Arpino et al., 2015; Knight & Brinton, 2017). According to the European Values Study (2017), 58% of the European population aged between 18 and 45 disagree or strongly disagree with the statement ‘family life suffers when a woman has a full-time job’. In Spain, this figure increases to 74.2%.

Despite this apparently egalitarian context for women, there is an unequal division of labour among Spanish couples (Abril et al., 2015; Borràs et al., 2021; Botía-Morillas, 2019), particularly after the birth of the first child (Seiz et al., 2019). According to the European Quality of Life survey, 2017, women spend around 38 h caring for their children per week, while men only spend 23 (Eurofund, 2017). Not surprisingly, women work part-time far more often than men. According to data from the Spanish Labour Force Survey, 2018, around 24% of working women were in part-time employment, compared to 7% of men (INE, 2018).

Hypotheses

Fertility intentions ‘involve a specific decision to pursue an actionable goal, with an associated commitment and, commonly, a plan for implementing the decision’ (Miller, 2011, p. 78). Since the decision to have a child depends largely on individuals’ previous experiences (Presser, 2001), one-child mothers’ parenting experience is likely to influence their desire to increase their families. If women experience a ‘double shift’ (Hochschild & Machung, 1989) and do not receive the appropriate support from their partners (Cheng & Hsu, 2020), they may feel that an additional child would increase their burden and, thus, prefer not to expand their families. Therefore, Hypothesis 1 is as follows:

H1. Women in couples with an unequal distribution of unpaid labour have a lower probability of intending to have a second child.

It is necessary to distinguish between the distribution of housework and the distribution of childcare (Coltrane, 2000). With the arrival of a new child, the amount of housework and childcare tasks increases considerably. Although the latter increases more, childcare is often considered more pleasant and rewarding. In contrast, housework is regarded as dull and tedious (Sullivan, 2013). Thus, Hypothesis 2 is as follows:

H2. The distribution of housework has a more decisive effect on a mother’s probability of intending to have a second child than the distribution of childcare.

Neyer and colleagues also reported differences in fertility intentions depending on whether the analysis focuses on equality (the actual division of chores) or equity (the personal perception of that division) (Neyer et al., 2013). Given that the subjective perception of fairness shapes satisfaction with the division of domestic labour (Fraser, 1994; Neyer et al., 2013), and previous studies have found that the level of satisfaction with the workload distribution moderates the relationship between the gender distribution of unpaid labour and fertility (Riederer et al., 2019), Hypothesis 3 is as follows:

H3. Women’s satisfaction with the distribution of unpaid labour moderates the effect of the actual workload distribution on fertility intentions.

Lastly, previous research has shown how having a child has different implications for women with low and high levels of education (Evertsson, 2016; Jalovaara et al., 2019). Since highly educated women tend to choose gender-egalitarian partners (Van Bavel, 2012), have higher bargaining power (Brodmann et al., 2007), and may face different constraints in balancing work and family life (Jalovaara et al., 2019), it seems reasonable to consider how the relationship between gender equality and fertility varies at different educational attainment levels (Lappegård, 2020). Therefore, Hypothesis 4 is as follows:

H4. A woman’s education level moderates the effect of the distribution of unpaid labour on fertility intentions.

Data and analytical strategy

Data

This study uses data from the 2018 Spanish Fertility Survey (SFS), a cross-sectional survey conducted by the Spanish Statistical Office using a representative sample of 14,556 women and 1619 men aged 18 to 55. It contains information on the respondents’ fertility intentions and the gendered distribution of household responsibilities, among other relevant characteristics.² The analytical sample is confined to women aged 18 to 44 with one biological or adopted child, who lived with their male partners and were neither pregnant nor sterilised. In order to assess the effect of the distribution of childcare, the sample is restricted to mothers of children up to 13 years old, and excludes some cases because of missing information, giving an analytical sample of 1164 women.

The dependent variable is women’s intention to have a second child within the following 3 years, a more accurate measure than long-term intentions (Schoen et al., 1999; Testa & Basten, 2014). As shown in Table 1 in the Appendix, 44.8% of one-child mothers in the sample intend to have a second child in the following 3 years. The main explanatory variables to be evaluated are the distribution of housework and the distribution of childcare within the couple, measured as follows.

Table 1 Descriptive statistics of the variables included in the models

The distribution of housework data is collected from a question that asks women to estimate the approximate percentage of housework that they, their partners, and other persons usually do. The share of housework done by the couple is re-scaled to exclude the housework performed by other persons:

$$\text{Percentage \, of \, housework}=\frac{{P}_{\text{w}}}{{P}_{\text{w}}+{P}_{\text{M}} } \times 100,$$

where P_W represents the percentage of housework performed by the woman, and P_M the percentage of housework done by the man. I identified four categories, which record the percentage of housework done by the women out of the total housework performed by couples, namely: 0–40%, 41–60%, 61–80%, and more than 80%.

The distribution of childcare data is obtained from a question including a battery of ten childcare activities, where women indicated who mainly performs each of them.³ The proportion of activities done by the woman out of the total number of activities performed by the couple is calculated:

$$\mathrm{Percentage \, of \, childcare}=\frac{{A}_{\mathrm{w}}+{A}_{\mathrm{c}} \times 0.5}{{A}_{\mathrm{w}}+{A}_{\mathrm{c}}+{A}_{\mathrm{M}} }\times100,$$

where A_W refers to the activities done by the woman, A_c to the activities done by the couple (shared), and A_M to the activities done by the man. This variable is later transformed into the same four categories as the distribution of housework.

Women were also asked about their satisfaction with the distribution of housework and childcare on a scale from 0 to 10. Given the skewed distribution of their responses (mean satisfaction with childcare = 8.04; mean satisfaction with housework = 7.25), women’s satisfaction is classified into two categories: 9–10 (highly satisfied) and 0–8 (less satisfied). Their education level is recorded using a binary indicator which identifies the women who have completed university education.

Finally, several variables that have been shown to influence fertility intentions are included in the analysis as controls. Firstly, the models control for the proportion of housework and childcare performed by other individuals, as the ability to rely on external help may help women’s life–work balance and affect their fertility intentions (Baizán, 2009; Rutigliano, 2020). Moreover, the child’s age is relevant, as it affects both the type of childcare activities the couple must address and their fertility intentions (Berrington, 2004; Quesnel-Vallée & Morgan, 2003).⁴ An additional variable indicates whether the partner had other children from previous relationships.⁵

Secondly, following previous research, a block of variables is included that controls for those characteristics which may influence fertility intentions and the distribution of household responsibilities: women’s age (Sobotka & Beaujouan, 2018); the migrant background of either partner (Brini et al., 2022; Kraus & Castro-Martín, 2018); women’s educational attainment (McDonald, 2006), and whether their partner has a higher, similar, or lower educational level (Brodmann et al., 2007).⁶

Thirdly, since job conditions play a considerable role in fertility decisions and the distribution of chores (Bueno & Brinton, 2019; Busetta et al., 2019), the models control for women’s attachment to the labour force (full-time, part-time, unemployed and inactive) and men’s labour force status (self-employed, employed, unemployed/inactive).

Finally, all models include overall satisfaction with the relationship to account for the possibility that their dissatisfaction with the distribution of unpaid work might reflect more profound relationship issues (Berninger et al., 2011). Descriptive information about all these variables can be found in Table 1 in the Appendix.⁷

Analytical strategy

Logistic regression models are used to analyse the relationship between the gender distribution of housework and childcare and women’s intention to have a second child in the following 3 years. Each distribution is assessed separately in order to avoid possible interferences between them. The first step involves carrying out a regression analysis where fertility intentions (Y) is predicted by the distribution of unpaid labour (D) and a vector of controls (Z) (Model 1). Then, the interaction between the distribution of household responsibilities and the satisfaction with that distribution (S) is added to assess whether satisfaction moderates the effect of the distribution (Model 2). Finally, the interaction of the distribution of chores with women’s educational level (E) (Model 3) is evaluated:

$$M1: \mathrm{logit}\left({Y}_{i}\right)=\alpha +\beta {D}_{i}+\phi {{\varvec{Z}}}_{{\varvec{i}}}+{\varepsilon }_{i},$$

$$M2: \mathrm{logit}\left({Y}_{i}\right)=\alpha +\beta {D}_{i}+\gamma {S}_{i}+\varphi {D}_{i}\times {S}_{i}+\phi {{\varvec{Z}}}_{{\varvec{i}}}+{\varepsilon }_{i},$$

$$M3: \mathrm{logit}\left({Y}_{i}\right)=\alpha +\beta {D}_{i}+\varphi {D}_{i}\times {E}_{i}+\phi {{\varvec{Z}}}_{{\varvec{i}}}+{\varepsilon }_{i}.$$

For clarity, the main results are shown graphically as predicted probabilities, although complete models can be found in Appendix B (Tables 2, 3, 4, 5, 6). To analyse the interactions, I assess whether the Average Partial Effect (APE) of the distribution of chores varies over the categories of the two moderating variables (Mize, 2019). A second difference test is carried out to test the statistical significance of that difference (Long & Freese, 2014).

Table 2 Logit models: housework

Table 3 Logit models: childcare

Table 4 Predicted probabilities and average partial effects of the distribution of housework and childcare

Table 5 Predicted probabilities and average partial effects of the interaction in Model 2

Table 6 Predicted probabilities and average partial effects of the interaction in Model 3

Results

Women’s fertility intentions by distribution of housework and childcare: a bivariate analysis

Figure 1 describes one-child mothers’ share of housework and childcare and their satisfaction with that distribution.⁸ As previous literature has documented (Abril et al., 2015; Borràs et al., 2021; Botía-Morillas, 2019), Spanish women perform the greatest share of chores in Europe. Over two-thirds of women declare that they do more than 60% of housework or childcare tasks. Only about 2% of one-child mothers report that their partners do more than 60% of housework or childcare, and around 30% state that they have a balanced distribution (as they perform between 41 and 60% of the total workload).

Fig. 1

Percentage of housework and childcare duties performed by women and reported satisfaction levels. Percentages are weighted

Regarding satisfaction with the allocation of household responsibilities, women are more satisfied with the distribution of childcare than of housework tasks. However, they stated that they were quite satisfied with both distributions. On a scale from 0 to 10, only 32% and 19% of women reported a satisfaction level below 7 with the allocation of housework and childcare duties, respectively. In this regard, it is important to note that more women receive childcare support than housework support from their partners and from other relatives.

Figure 2 depicts the percentage of women who intend to have a second child, by share in the couple’s division of housework and childcare. Fertility intentions peaked among women performing between 61 and 80% of chores: 53.9% intend to have a second child doing that amount of housework, and 48.6% intend to have a second child performing that amount of childcare. In comparison, women in very unfavourable distributions (where they are responsible for more than 80% of chores) have lower fertility intentions. Additionally, only 18% of women in favourable distributions of housework (0–40%) intend to have a second child, and just 36% of women in favourable distributions of childcare (0–40%) plan to increase their family size. Therefore, a slight inverted U-shaped relationship is observed between fertility intentions and women’s share of both housework and childcare.

Fig. 2

Percentage of women intending to have another child, by housework and childcare distribution and satisfaction. Percentages are weighted

The effect of the distribution of housework and childcare on fertility intentions: a multivariate analysis

Figure 3 shows the findings from the multivariate analyses. The results barely change after including the controls. The probability of intending to have a second child peaks when women do between 61 and 80% of the housework (predicted probability = 0.498). Among women with a heavy load (81–100%), that probability decreases by 6.6 percentage points (p-value = 0.039), a statistically significant negative effect. This finding is in line with previous research conducted in Italy and the Netherlands (Mills et al., 2008).

Fig. 3

Predicted probabilities of one-child mothers’ fertility intentions by housework and childcare distribution. Error bars at 90% (thin) and 95% (thick). Models control for housework/childcare duties done by other individuals, woman’s age, child’s age, migrant background, women’s attachment to the labour force, men’s labour force status, differences in partners’ educational attainment, partner’s children and women’s partnership satisfaction

Women with egalitarian distributions of housework (41–60%) are also less likely to intend to have a second child than women who perform 61–80% of the total housework, although in this case, the differences are not statistically significant (APE = − 0.046; p-value = 0.145). However, women with favourable distributions of housework (0–40%) had the lowest probability of intending to have a second child (predicted probability = 0.278). This result is intriguing, although it has already been observed in other contexts (Dommermuth et al., 2017). The additional analysis included in Appendix C shows the specific characteristics of this particular group of women. Apart from being a small group (2.2% of the sample), it encompasses a larger proportion of women that work full-time (79% versus 52%) and a larger proportion of women whose partners work part-time (21 versus 6%) or are unemployed or inactive (23% versus 8%) than the rest of the sample. Although all models control for both variables, the descriptive statistics show that this group of women is somewhat different from the rest, so that the possibility of other unobserved factors affecting this group of women cannot be excluded.

In turn, the distribution of childcare does not seem to significantly influence fertility intentions. The estimated probability of intending to have a second child remains around 45%, irrespectively of the share of childcare done by women. This finding endorses previously observed differences between the effect of the distribution of housework and childcare (Sullivan, 2021). It is also consistent with previous studies that have documented the fact that the distribution of childcare in Spain does not affect fertility intentions (Brodmann et al., 2007; Cooke, 2009).

These results only partly confirm Hypothesis 1. The results corroborate that when the distribution of housework is highly unbalanced, women’s fertility intentions are lower. However, the relationship between women’s share of housework and fertility intentions is not linear, as anticipated, but resembles an inverted-U pattern. Women who do much more or much less work than their partners have lower fertility intentions. There are no significant differences between women who perform 61–80% of the housework, and those in couples with egalitarian distributions. It is not the distribution of all household responsibilities that affects the respondents’ fertility intentions, but the distribution of housework. This result is in line with Hypothesis 2, which posited that the distribution of housework is more important than the distribution of childcare for one-child mothers’ fertility intentions.

The effect of the variables included in the models as controls is as expected. University-educated women and women who receive childcare support from other individuals (not their partners), are more likely to intend to have a second child. Conversely, there is a negative relationship between a woman’s age and her fertility intentions and between the first child’s age and fertility intentions. When women are in a hypogamous partnership in education terms, the probability of intending to have a second child is also lower; so it is when their partner has children from other relationships or when women are unsatisfied with their relationship. The rest of the variables do not substantially impact women’s fertility intentions.

The moderation effect of women’s satisfaction with the distribution of unpaid labour

In order to assess whether women’s satisfaction with the distribution moderates the effect of the distribution of unpaid labour an interaction effect between both variables was included. The estimated probabilities of intending a second child are depicted graphically in Fig. 4. Light lines represent high satisfaction levels (9–10), and dark curves represent lower satisfaction (0–8).

Fig. 4

Predicted probabilities of one-child mothers’ fertility intentions. Interaction between housework/childcare distribution and mothers’ satisfaction. Error bars at 90% (thin) and 95% (thick). Models control for housework/childcare duties done by other individuals, woman’s age, child’s age, migrant background, women’s attachment to the labour force, men’s labour force status, differences in partners’ educational attainment, partner’s children and women’s partnership satisfaction

Contrary to the expectations posited in Hypothesis 3, satisfaction with the distribution of unpaid labour does not seem to play an important moderating role. There is no significant effect of the gendered distribution of housework among women with high levels of satisfaction, except when they have a favourable distribution (0–40%), in which case they have the lowest fertility intentions (Predicted probability = 0.282). Women with lower levels of satisfaction reduce their intentions when they have a very unfavourable division of labour (> 80%) compared to those with unfavourable distributions (61–80%). In that case, the probability of intending to have a second child declines from 0.484 to 0.423 (AME = − 0.061; p-value = 0.081). In turn, the satisfaction with the distribution of childcare does not influence women’s fertility intentions, regardless of whether they are highly satisfied or less satisfied with the distribution of labour.

Heterogeneity in the effect of the distribution of chores by women’s educational attainment

Finally, an analysis was made of whether the effect of the distribution of housework and childcare varies across women’s educational attainment. Figure 5 distinguishes between the fertility intentions of university-educated women (light line) and less-educated women (dark line) according to their share of housework and childcare. Interestingly, the effect of the distribution of housework only influences the fertility intentions of university-educated women. Again, an inverted-U pattern is observed for this group. Fertility intentions peak when highly educated women do 61–80% of the housework (58.6% intend to have a second child) and decline when they do most of the housework. The probability of highly educated women in favourable distributions (0–40 having a second child is much lower (17.8%). In contrast, around 43% of less-educated women intend to have a second child regardless of the distribution of housework. As for childcare, there are no significant differences between women with high and low education attainment.

Fig. 5

Predicted probabilities of one-child mothers’ fertility intentions. Interaction between housework/childcare distribution and mothers’ educational attainment. Error bars at 90% (thin) and 95% (thick). Models control for housework/childcare duties done by other individuals, woman’s age, child’s age, migrant background, women’s attachment to the labour force, men’s labour force status, differences in partners’ educational attainment, partner’s children and women’s partnership satisfaction

The educational attainment of women moderates the effect of the distribution of chores on the intention to have a second child (Hypothesis 4), but only for housework. Highly educated women reduce their intentions compared to their less-educated counterparts when the distribution of labour is highly unequal, regardless of whether it is the woman or the man who performs the largest part of housework. The bargaining power of highly educated women with their partners (Zhou & Kan, 2019) may be behind this result, as they may have considered the degree of involvement of their partners in household responsibilities (among other factors) when making reproductive decisions.

Additional analyses and robustness checks

In order to ensure the robustness and consistency of the results, several robustness checks are conducted. The results are shown in Appendix C. The first type of analysis assesses the internal validity of the main explanatory variables: the distribution of housework and childcare. The analyses are rerun using a different indicator: the proportion of housework and childcare women performs over the total workload (including what other individuals do), instead of the proportion the woman performs out of the total workload. Figure 6 in the Appendix shows that the results do not change substantively under the two alternative definitions of the distribution of housework and childcare.

Fig. 6

Predicted probabilities of one-child mothers’ fertility intentions by woman’s proportion of total burden. Error bars at 90% (thin) and 95% (thick). Models control for housework/childcare duties done by other individuals, woman’s age, child’s age, migrant background, women’s attachment to the labour force, men’s labour force status, partners’ educational attainment, differences in educational attainment, partner’s children and partnership satisfaction

The second type of analysis focussed on a selected subgroup of the analytical sample to further examine the results. Firstly, women in partnerships where both partners work full-time are examined, since these women may bear the ‘double shift’ more directly and reflect more clearly the transitional stage of the Gender Revolution. Figure 7 shows the results of this analysis. The outcomes are similar to the findings obtained among the whole sample. Nevertheless, among this group of women, those with a highly unequal distribution of childcare (81–100%) have a lower probability of intending to have a second child, and the difference is statistically significant (p-value: 0.08).

Fig. 7

Predicted probabilities of one-child mothers’ fertility intentions by housework/ childcare distribution where both partners work full-time. Error bars at 90% (thin) and 95% (thick). Models control for housework/childcare duties done by other individuals, woman’s age, child’s age, migrant background, women’s attachment to the labour force, men’s labour force status, partners’ educational attainment, differences in educational attainment, partner’s children and partnership satisfaction

Secondly, only mothers with children under 6 years old were considered. Women with a strong desire for two children may have already moved to the second by the time of the interview and may not be included in the analytical sample. To avoid this possible bias, and assuming that women with a solid intention to have a second child will have it shortly after the first, the analyses are rerun among mothers of younger children. Figure 8 in Appendix 3 reports the results of the analysis. The findings do not differ broadly from the initial analyses. Although the effects are smaller and not statistically significant, there is still a decline in fertility intentions among women doing more than 80% of housework and among women in favourable distributions (0–40%).

Fig. 8

Predicted probabilities of one-child mothers’ fertility intentions by housework/childcare distribution among mothers of children under 6. Error bars at 90% (thin) and 95% (thick). Models control for housework/childcare duties done by other individuals, woman’s age, child’s age, migrant background, women’s attachment to the labour force, men’s labour force status, partners’ educational attainment, differences in educational attainment, partner’s children and partnership satisfaction

Thirdly, the socio-demographic profile of women whose share of housework is 0–40% is examined. This is an atypical group where men perform significantly more housework than women. In order to draw a profile showing their particular characteristics, Table 7 in Appendix C shows the descriptive statistics of this group of women. This group contains a larger proportion of women working full-time and with partners working part-time or are unemployed or inactive.

Table 7 Descriptive statistics of women experiencing favourable distributions of housework and the rest

Conclusions

Although the Gender Revolution Theory has prevailed in recent years underscoring the importance of gender equality in understanding contemporary fertility patterns and trends, mixed empirical evidence across different societal contexts has highlighted the need to continue evaluating empirical evidence for the Gender Revolution arguments and predictions. This research aimed to provide further evidence on the relationship between gender equality and fertility in a lowest-low fertility society (Billari & Kohler, 2004) where the expansion of women’s levels of education (Ortiz & Rodríguez-Menés, 2016) was the driver for a rapid spread of egalitarian attitudes among society (Arpino et al., 2015). However, the division of unpaid labour has not adapted to the new circumstances and family-friendly policies remain scarce (Castro-Martín & Martín-García, 2016; Esping-Andersen, 2009).

Inspired by the GRT (Goldscheider et al., 2015), it was expected that a gendered allocation of housework and childcare within the couple would reduce one-child mothers’ fertility intentions. Although the overall results partially confirmed this expectation, the conclusions are more nuanced. First, although the distribution of housework between partners impacts women’s fertility intentions, the distribution of childcare does not. One may speculate that, in the Spanish context, given the late age at first birth and the large gap between desired and actual fertility (Castro-Martín & Martín-García, 2013), devoting time to childcare could be perceived as a rewarding activity (Sullivan, 2021), and may not negatively influence second-child intentions.

Second, the results suggest an inverted-U pattern relationship between women’s share of housework and fertility intentions rather than a linear relationship. On the one hand, women who perform more than 80% of the housework load reduce their fertility intentions compared to women in fairer distributions. This result aligns with the postulates of the GRT. When women carry the bulk of housework and cannot rely on their partners’ participation, they may prefer not to enlarge their families (Goldscheider et al., 2015; McDonald, 2000). Nevertheless, women in unfavourable distributions (60–80%) have the highest probability of intending to have a second child. This finding suggests that, despite their generally egalitarian gender attitudes, Spanish women still tolerate a certain level of inequality in the distribution of chores (Abril et al., 2015). The shift from an inegalitarian to a more egalitarian society may have taken place so fast in Spain that it has not yet permeated the population’s everyday life. Therefore, some women may feel comfortable with an unequal distribution of household responsibilities when it is not heavily unbalanced. In fact, that is in line with women’s declared satisfaction with the allocation of housework and childcare: in spite of unequal distributions, women report a high level of satisfaction, revealing that the perception of fairness in the division of unpaid work does not necessarily align with the (in)equality of that division (Gonzalez et al., 2018; Koster et al., 2022).

On the other hand, women whose partners perform a higher percentage of housework than them have a lower probability of intending to have a second child. Although this result may seem counter-intuitive, in a context where balancing work and family is challenging, men involved in housework may also face this ‘double shift’, which may reduce their fertility intentions, and ultimately downgrade their partners’ fertility intentions (Okun & Raz-Yurovich, 2019). Moreover, some other factors affecting this uncommon distribution of housework may also impact women’s low fertility intentions, such as material reasons, which are among the main causes of the low fertility in the country (Adsera, 2011; Castro-Martín & Martín-García, 2016). Similar results were found in a very different context, in Norway, where the probability of having a second child was lower when men took a leading role in housework (Dommermuth et al., 2017). Even so, one should be cautious with this result due to the small sample size (2%) and the possible selectivity of this group of women. For instance, the 2018 SFS does not provide information on health issues or other circumstances that may explain this atypical allocation of chores.

Regarding the moderating effect of satisfaction with the distribution, only a weak effect is found. Women who perform more than 80% of housework reduce their fertility intentions compared to women who do a lower amount of housework only when they are less satisfied, but not when they are highly satisfied with their share. In turn, the decline in fertility intentions among highly satisfied women is only observed when their partners perform more housework than them. One may speculate that this group of women may be afraid that a new child would alter the state of things and would therefore be less willing to increase their family size. However, as already noted, the limited size of this group of women in the sample does not permit firm conclusions to be drawn.

As for the moderating role of women’s educational attainment, the slight inverted U-shaped relationship between housework and fertility intentions only holds true for highly educated women. They may have higher bargaining power in their partnerships (Kühhirt, 2012) and, therefore, make fertility decisions more directly based on their partners’ support. Surprisingly, among their less educated counterparts, who presumably face more barriers to work–family reconciliation, their partners’ share of housework does not seem to influence their fertility intentions. This shows how different socioeconomic groups have different needs and behave differently in the process of family formation (van Bavel, 2012), as well as highlighting the importance of considering the level of education to understand fertility trends (Lappegård, 2020).

Finally, it is important to acknowledge some limitations in this research. Firstly, the analysis only includes women. The results would be more insightful if both women’s and men’s intentions were included (Trimarchi & Van Bavel, 2017). However, the SFS did not collect partners’ fertility intentions, and the sample of one-child fathers is too small to conduct a parallel analysis (n = 414). Secondly, the distribution of household workload was self-reported by women, who may overstate their contribution and downplay their partners’ share. Nevertheless, previous studies using the Time Use Survey have shown a very similar pattern of domestic work distribution in Spain (Borràs et al., 2021; González & Jurado-Guerrero, 2009). Thirdly, the questions in the SFS for the distribution of housework and childcare were not in a similar format, but were re-coded in a similar fashion to facilitate comparison.

Notwithstanding its shortcomings, this piece of research adds new empirical evidence to the study of the relationship between gender equality and fertility by using recent data to examine a lowest-low fertility country in an intermediate stage of the Gender Revolution. If the perception of a ‘double shift’ by women—when they do more than their partners at home—or by men—when they do more than women—is one of the reasons of lowest-low fertility levels in Spain, the availability of public policies that contribute to alleviating this double shift will be of particular importance. This research suggests that these public policies should be designed to lessen the burden of both women and men, something particularly important in Spain, where public support for families is scarce (Esping-Andersen, 2009). The results show that women who receive childcare support from other individuals (not their partners) are more likely to intend to have a second child, which seems to point to a need for care alternatives that relieve the burden. Some new public policies aimed at promoting men’s participation in childcare have been recently implemented in Spain. The equalisation of paternal and maternity leave (16 weeks) in 2021 might foster fathers’ involvement in family responsibilities and fertility intentions in the future (Meil et al., 2021). More policies in this direction are needed in order to promote a new gender equilibrium (Esping-Andersen & Billari, 2015), release the excessive burden on women, facilitate the balance between work and family life, and contribute to closing the gap between desired and actual fertility in Spain (Castro-Martín & Martín-García, 2013).

To conclude, this research largely supports the relevance of the Gender Revolution Theory in understanding fertility patterns in a country situated in an intermediate stage of the Revolution—relatively high gender equality in the public sphere but still low in the private sphere. Although there is no linear relationship between gender equality within the family and fertility intentions currently in Spain, a highly unequal distribution of housework reduces women’s likelihood to intend to have a second child.

Appendix A. Descriptive information

See Table 1.

Appendix B. Full results from the multivariate analysis

See Tables 2, 3, 4, 5, and 6.

Appendix C. Additional analyses and sensitivity analyses

See Table 7 and Figs. 6, 7, 8.