Families process state, market and personal resources to negotiate their particular household mix of productive and reproductive activities, yet only recently have we begun to compare the household division of labour and its effect on family outcomes across societies. One body of work looks at policy effects on gender equity within the home (Cooke, Reference Cooke2007a; Stier and Lewin-Epstein, Reference Stier and Lewin-Epstein2007), while another assesses the impact of gender equity in the household division of labour on fertility (Cooke, Reference Cooke2004; Oláh, Reference Oláh2003; Torr and Short, Reference Torr and Short2004). Economists such as Gary Becker (Reference Becker1981, Reference Becker1985) argue these two are at cross-purposes, with any increase in women's employment predicted to decrease women's desired number of children. A growing group of theorists across disciplines, however, argue instead that equity is the only sustainable gender order in the post-industrial world of less sta employment and more diverse families (Fraser, Reference Fraser1997; McDonald, Reference McDonald2000; Oppenheimer, Reference Oppenheimer1997).

When analysed at the aggregate or household level, the evidence in support of either hypothesis of gender equity effects on fertility remains inconclusive. Until recently, aggregate fertility was higher in countries where women's aggregate employment was lower, but the sign of that relationship reversed in the 1990s (Sleebos, Reference Sleebos2003). At the individual level – where couples actually negotiate a division of labour and have children – effects vary in context. Among US couples, the most gendered and most egalitarian housework arrangements are both positively associated with fertility, although odds are slightly greater among the more egalitarian couples (Torr and Short, Reference Torr and Short2004). For German couples, male breadwinner families are more likely to have a second child, but father's greater participation in childcare also increases the odds (Cooke, Reference Cooke2004). In Sweden and Hungary, the likelihood of second birth is greater when couples share family responsibilities as compared with couples with more gendered divisions of paid and unpaid labour (Oláh, Reference Oláh2003).

Equity effects on fertility appear to vary by the degree to which policy or cultural elements reinforce women's responsibility for unpaid care work. (West) German policies historically reinforced women's responsibility for the private sphere and economic dependence on a male breadwinner (Cooke, Reference Cooke2004). Socialist Hungary and social-democratic Sweden historically provided a web of services and policies supporting full employment of all able-bodied adults (Oláh, Reference Oláh2003). In the liberal US regime, policy has supported equal education and employment opportunities for women through Title VII of the Civil Rights Act of 1964, but remains silent regarding how demands from the private sphere are met (Cooke, Reference Cooke2007a). Taken together, this evidence suggests that the more policy encourages greater equity in the division of paid and unpaid labour, the more gender equity enhances family fertility.

McDonald (Reference McDonald2000) offers a structural argument for this covariation between policy and equity effects on fertility. He claims fertility declines most sharply where there is conflict between gender equity in individual-level versus family-level institutions. To date, however, there are no suitable institutional measures with which to assess this claim. Instead, the analysis here focuses on examining whether individual gender equity effects on fertility differ across two countries with similar institutional structures vis-à-vis gender equity. Italy and Spain are Mediterranean countries that industrialised later than their northern neighbours; both rely on extended kinship networks in the provision of individual welfare (Ferrera, Reference Ferrera2005), as of 2000 had similar female labour force participation rates (OECD, 2007) and have experienced a dramatic decline in fertility rates since the 1960s (Kohler et al., Reference Kohler, Billari and Ortega2002). Yet the countries also differ slightly in terms of women's returns to education, the structure of female employment and the gender earnings gap (OECD, 2002, 2004, 2007). Using seven waves of the European Community Household Panel (ECHP), we explore whether the slight differences in gender equity across the 1990s in these two countries yielded differences in equity effects on fertility.

Family in the state

Individual-level theories of the gendered division of household labour model it as a negotiation based on each partner's relative resources, with earnings and education key sources of relative power (Blood and Wolfe, Reference Blood and Wolfe1960). These models predicted women's rising employment would lead to a revolution in domestic gender roles, a revolution that failed to occur. Breen and Cooke (Reference Breen and Cooke2005) suggest this is because a given woman's individual resources are not sufficient to shift the gendered division; what is necessary is that the proportion of economically autonomous women must be sufficiently high to change men's beliefs about what a partner would expect in the domestic sphere: in other words, greater aggregate equity begets greater individual-level equity. A broad range of policies – including education, labour and family law, social insurance and public and private transfers – shapes aggregate gender equity (Cooke, Reference Cooke2007a). Assessing these dynamics using US and German panel data, Cooke (Reference Cooke2007a, Reference Cooke2007b) found women's individual resources predicted very little shift in men's housework hours, although some structural factors predicted that men will increase their housework hours. Individual as well as structural factors predicted significant differences in women's housework hours. Where policies enhance women's aggregate resources, they spend fewer hours in domestic tasks. Together this evidence supports the importance of policy effects on gender equity in paid and unpaid labour, as argued by Breen and Cooke.

Yet specialisation and trading models of family (Oppenheimer, Reference Oppenheimer1997) posit that women's greater equity in employment incurs an ‘opportunity cost’ of children that predicts lower fertility (Becker, Reference Becker1981, Reference Becker1985). Such a negative relationship is worrisome for welfare states, as it forces them to choose between bolstering the current tax base via women's employment versus the future tax base via children, reflecting the Achilles heel of welfare state ‘familialism’ (Esping-Andersen, Reference Esping-Andersen1999: 70). This familialism stems from state reliance on the family as the welfare provider of first resort (Esping-Andersen, Reference Esping-Andersen1990), and women's unpaid care work under a male breadwinner model (Lewis, Reference Lewis1992).

Fraser (Reference Fraser1997) notes, however, that these policy assumptions about the gender order centred on the ideal of a family wage. Beginning in the nineteenth century, trade unions and the working class bargained for wages enabling a working man to support a dependent wife and several children, with first welfare state provisions in most countries providing some type of insurance for interruptions in the male wage (Pierson, Reference Pierson1998). These ideals experienced their last hurrah during the welfare state expansion and baby boom through the 1950s and 1960s. Fraser (Reference Fraser1997) argues a family wage system is no longer tenable within highly competitive, liberalised labour markets that denote post-industrial societies. As male employment becomes more precarious, particularly for new labour market entrants, women's employment enhances family economic flexibility (Oppenheimer, Reference Oppenheimer1997). US evidence supports this assertion: employed US women are now more likely to marry than unemployed women (Sweeney, Reference Sweeney2002), and US marriage rates have been declining at a faster rate among the least-educated women (Moffitt, Reference Moffitt and Waite2000).

During the transition from industrial to post-industrial societies, the observed relationship between aggregate female employment and fertility also shifted. While historically the correlation between the two across OECD countries had been negative, in the 1990s it became positive (McDonald, Reference McDonald2000; Sleebos, Reference Sleebos2003). Welfare state provisions in the Scandinavian countries and market services in the English-speaking countries supportive of maternal employment are credited for this trend reversal (Brewster and Rindfuss, Reference Brewster and Rindfuss2000; Sleebos, Reference Sleebos2003). During the same time period, however, very low fertility spread rapidly in southern Europe (Kohler et al., Reference Kohler, Billari and Ortega2002).

McDonald (Reference McDonald2000) argues these trends reflect the degree of institutional conflict between gender equity achieved in education and the labour market, and equity as related to the family – including the degree of flexibility in employment, extent of supportive services such as public day care and maternity leave − and within family members’ roles. Policies upholding a male breadwinner model block the evolution in gender equity that proves adaptive as nations move from industrial to post-industrial economies. Male breadwinner policies are prevalent within corporatist-conservative welfare regimes reinforcing other status hierarchies as well, such as between skilled and unskilled workers (Esping-Andersen, Reference Esping-Andersen1990). As male employment has become more precarious over the past few decades, corporatist-conservative regimes deepened insider–outsider cleavages by protecting secure positions of older, privileged male workers at the expense of job security among new labour market entrants (Esping-Andersen, Reference Esping-Andersen1999). In addition to facing a protracted youth-to-adult transition, today's young adults are also more likely to face fixed-term contracts and part-time rather than full-time employment (Taylor-Gooby, Reference Taylor-Gooby2004). So now, corporatist-conservative regimes can no longer offer the secure employment necessary to a male breadwinner model and have minimal supports in place for the female employment that enhances family economic flexibility. As a result, aggregate levels of female labour force participation rates and total fertility rates are lower in regimes with male breadwinner policies such as Germany, Belgium and the Netherlands, than in Scandinavia or the English-speaking countries (Brewster and Rindfuss, Reference Brewster and Rindfuss2000; Sleebos, Reference Sleebos2003).

The female employment/fertility trade-off under the male breadwinner model becomes more acute the greater the precariousness of male employment, which explains why female employment and fertility tend to be lower in the southern versus northern European countries (Adsera, Reference Adsera2003). In Italy and Spain, for example, traditional gender roles and women's responsibility for the domestic sphere were institutionalised in authoritarian state structures prior to the late industrialisation that began in the 1960s in Italy (Bernardi and Nazio, Reference Bernardi, Nazio, Blossfeld, Mills, Klijzing and Kurz2005) and 1970s in Spain (Jurado Guerrero and Naldini, Reference Jurado Guerrero and Naldini1996). As detailed in the next section, macro-economic factors since that time have made male employment extremely precarious. These simultaneous factors intensify the trade-off between female employment and fertility because traditional male breadwinner families are institutionally supported but no longer economically feasible, while dual-earner families are the most economically viable yet are not institutionally supported. Consequently, Italy and Spain have among the lowest female labour force participation and total fertility rates of all OECD countries (OECD, 2007; WHOSIS, 2002). At the same time, however, policy and cultural differences in gender equity also exist, which should predict some variation in equity effects on fertility across the two countries. The institutional factors affecting the family's division of labour and fertility in Italy and Spain are described next. Given the data window of the ECHP, the focus will be on the sociopolitical contexts in the second half of the twenty-first century.

Italy and Spain

Method

Discussion and conclusions

A growing body of evidence indicates that gender equity varies in context, with more egalitarian divisions of domestic tasks observed in countries with greater policy support for gender equity (Breen and Cooke, Reference Breen and Cooke2005; Cooke, Reference Cooke2007b; Stier and Lewin-Epstein, Reference Stier and Lewin-Epstein2007). Effects of gender equity on fertility also vary in context, with higher fertility associated with more egalitarian domestic divisions in countries such as Sweden where policies support maternal employment (Oláh, Reference Oláh2003) or remain silent regarding the domestic sphere such as in the United States (Torr and Short, Reference Torr and Short2004). In contrast, a gendered division of labour predicts higher fertility in countries with historical support for male breadwinner families such as the former West Germany (Cooke, Reference Cooke2004). Here we use the European Community Household Panel to assess whether differences in equity effects on fertility are evident across two countries with similar institutional reliance on extended kinship networks and women's unpaid care work, but distinct, albeit small, differences in gender employment equity during the 1990s.

Gender equity in employment is low in both Italy and Spain, with about half of the wives with one child out of the labour force at some time during the seven years of the panel. Wives who were employed tended to work full-time, with the gender earnings gap smaller in Italy than in Spain. Wives’ hours of work in both countries reduced the likelihood of a second birth, with the effect more acute in Italy. The presence of another adult in the household increased the likelihood of having a second child, although the effect only reached statistical significance in Italy. Dual-earner couples that paid for childcare were more likely to have another child, with the net effect appreciably greater in Spain than in Italy. Both of these sources of care were available to just a small number of couples in the countries, so an obvious policy solution to the low fertility in the two countries would be to expand infant childcare availability. The macro-economic factors that blunted employment, however, also reduced available tax dollars or disposable income for such public or private investment. So, ironically, the same contextual factors that undermined a male breadwinner model also prevented the alteration of it. Since the 1990s, childcare availability has been expanding, but primarily for children aged three to five (León, Reference León2007).

Despite the limited external care options, however, only in Italy did fathers' assumption of more caring responsibilities significantly increase the likelihood of a couple having another child, particularly within dual-earner couples. The magnitude of the effect was five times that found by Cooke (Reference Cooke2004) in her analysis of German couples. Yet when Italian fathers assumed more than one-quarter to one-third of childcare, the likelihood began to decline sharply. These findings suggest, first, that women's greater aggregate employment equity predicts not only modestly greater equity within households, but positive effects of equity on family outcomes. Consequently, although we have not realised the revolution in gender roles predicted by women's rising labour force participation, the evidence here suggests the evolution helps couples meet productive and reproductive challenges in post-industrial societies.

When Italian couples venture too far from traditional domestic roles, however, the likelihood of second birth declines sharply. Whether this reflects ‘maternal gatekeeping’ on the part of Italian mothers or Italian fathers' resistance to assuming more of the care tasks cannot be assessed with these data. Maternal gatekeeping is a term used by Allen and Hawkins (Reference Allen and Hawkins1999) referring to working US mothers who actively managed fathers' access to children. In a more recent analysis, Sayer (Reference Sayer2005) found that although US women still retain primary responsibility for childcare, US men's childcare hours have also increased since the 1960s, and they have taken on a significantly larger share of housework. The ECHP has no measures of housework with which to compare whether Italian or Spanish men are similarly performing more housework as compared with childcare. More qualitative cross-country comparisons are needed to unravel who in the family actively structures the household division of domestic labour.

On the other hand, no paternal care effects were found for Spanish couples. This might result from the lower levels of gender equity in employment in Spain as compared with Italy. More Spanish wives were out of the labour force and the gender wage gap is among the largest in OECD countries. To the extent aggregate gender equity in employment results in women's greater ability to negotiate more egalitarian divisions of domestic tasks as argued by Breen and Cooke (Reference Breen and Cooke2005), this ability is lower in Spain than in Italy. The findings here support the growing body of evidence that policy support for greater gender equity in the public sphere leads to greater gender equity within the domestic sphere, which in turn countervails some of the negative fertility effect associated with women's labour force participation (Cooke, Reference Cooke2004; Oláh, Reference Oláh2003; Torr and Short, Reference Torr and Short2004).

There are limitations to the study. First, defining childcare remains challenging. Is time spent playing with a child considered care, or is it equivalent to tending to a sick child in the middle of the night? Australian time diary data reveal that mothers are far more likely to be home alone with their children and responsible for their physical care, whereas Australian fathers spend more time with their children in play (Craig, Reference Craig2003). If this is the case, increases in paternal time may not be reducing mothers’ time in care, although undoubtedly children benefit from this extra parental time. Second, the ECHP limits analyses to the 1990s, so we are unable to see effects of policy changes since that time. As noted above, childcare is expanding in Spain, and more recent employment data suggest gender equity in that sphere is more similar now across the two countries (OECD, 2007). Future research will need to assess whether these policy changes have shifted the effects of individual equity factors on fertility within Italy and Spain.

Despite the limitations, the findings reported here support that a gendered division of labour appears increasingly inflexible and detrimental to modern families. It exacerbates the historical negative relationship between women's employment and fertility during a time when women's earnings are crucial to family economic flexibility. The findings also underscore the importance of fathers' childcare participation even in cultures with strong reinforcement of women's responsibility for care work. Finally, the results highlight the importance of assessing gender equity writ large, in terms of employment opportunities, structures and the variety of care solutions families use to fashion solutions for balancing work and family in post-industrial societies.