Policy change and partisan politics: understanding family policy differentiation in two similar countries

. This paper looks at how different electoral competition dynamics can result in differentiated party positioning on childcare and family policy. Italy and Spain are compared using a most similar case design. The presence of women in politics, the socio-economic profiles of the voters of the two main left-wing and right-wing Italian and Spanish parties, and opinions on traditional norms of motherhood explain different policy trajectories and higher incentives for the conservative party in Spain to converge towards the social-democratic party in more progressive views of family policy.


Introduction
In this contribution we are interested in unveiling the socio-political dynamics behind changes in family policy.We engage with a growing body of research that looks at the interaction between changes in the political landscape of Western democracies and welfare reform (Beramendi et al. 2015;Häusermann, Picot and Geering 2012;Manow et al. 2018).We connect with a number of studies interested in understanding how changes in social demand and in politics can trigger changes in family policy from a more traditional to a more progressive conception by endorsing gender equality and supporting the employment of mothers (Ferragina and Seeleib-Kaiser 2015;Häusermann 2010;Mahon 2002;Morgan 2013).
We follow Daly and Ferragina's (2018) definition of family policy: public support to families with children as a package that includes money, services and time.Using the OECD typology (2009), we distinguish among: cash benefits (e.g.family allowances, parental leave, maternity and paternity leave), tax breaks, childcare benefits and other benefits in kind (mostly childcare services).If one broad issue is how generous and what specific type of support family policies provide, a more specific question is the overall goal of such policies We compare Italy and Spain, two countries that have been the focus of fruitful comparative analyses since Ferrera (1996) introduced the 'Southern Model' to the world of welfare regime typologies.Given shared socio-economic structures and strong institutional similarities (Castles 1995, Ferrera 1996, Guillén and León 2011), diverging outcomes between the two countries can be compared within a most similar case design.
A number of recent studies have shown how the economic recession heralded a new age of 'permanent strain' that was equally pervasive in both countries, causing welfare retrenchment (Pavolini et al. 2015;Petmesidou and Guillén 2014).Still, a number of studies have indicated that the two countries have not always responded in the same way to shared challenges before, during or after the economic crisis (León and Pavolini 2014;Naldini and Jurado 2013).
Italy and Spain are two familialistic countries that have long lacked policies oriented to socializing care, supporting families, or facilitating women's employment.
Over the last decades both countries have started to invest more in family policy but whilst Italy has kept its familialistic profile (León and Pavolini 2014;Naldini and Saraceno 2008), Spain has placed greater emphasis on de-familialization (Estévez-Abe and Naldini 2016), by investing in public childcare services, supporting the employment of women and encouraging fathers to share childbearing responsibilities.We want to understand the reasons behind this difference.
Recent research has provided different explanations for this absence of policy shift in Italy.Three relatively recent studies (Blome 2017;Morgan 2013;Schwander 2018) emphasize the greater traditionalism of Italian voters, who still hold conservative views regarding gender roles and the upbringing of children, to explain why political parties have not adopted more progressive stances with regards to family policy.We would go further, analyzing social demand factors and other less explored aspects that intervene between policy changes and partisan politics.
As Figure 1 below shows, our analysis considers one dependent variable, one intervening variable and three independent variables.Our dependent variable is the difference in family policy in Italy and Spain from 1990 to 2015 with regards to a) overall expansion and b) how important the de-familiarization approach has been.Our intervening variable is the difference in the way in which family policy figures in the party manifestos of the main center-left and centerright parties in the two countries and the proximity of left and right towards (progressive) family policy in their electoral programs.
Our analysis shows that whilst both countries have increased their spending on family policy during the period of study, they have done so by following different paths.
The Italian model is geared towards cash transfers and tax benefits.The Spanish model places greater emphasis on public childcare services and on increasing time off for fathers after childbirth.Our intervening variable reveals that in Spain, family policy is more visible in the party manifestos of both left-wing and right-wing parties.In addition, increasing convergence between the two largest parties toward more progressive views is observed.In Italy, however, family policy is less visible in the party manifestos of both the left and the right, while significant party differences remain.The Italian center-right party maintains a traditional approach to family policy while the left advocates for childcare services and leave schemes.
. In short, what our dependent and intervening variables show is that Italy is an outlier with respect to global trends in family policy in European countries, while Spain is more in line with the major change taking place in conservative or familialistic welfare systems towards support for female employment and a dual-earner model.How do we explain this divergence?
To understand these differences, and after considering a number of explanatory factors, ranging from social demand to power resources, we focus on three independent variables that we argue play a fundamental role: firstly, the presence of women in politics and how this is affected by rules of access; secondly, different socio-economic profiles of the voters of the two main left-wing and right-wing Italian and Spanish parties; thirdly, cultural preferences in the two countries and the way these have evolved over time.
Contrary to previous studies (Fleckenstein and Lee 2014;Morgan 2013;Schwander 2018) we did not find evidence of an effect of higher electoral volatility on the policy differences between the two countries.
For the dependent variable, we use various sources to collect data on expenditure and coverage as well as detailed accounts of policy reforms.For the intervening variable, we have carried out a content analysis of party manifestos of the two main left-wing and right-wing parties in our two countries from 1990 until 2015.For our three independent variables, we firstly look at the proportion of seats held by women in the Italian and Spanish national parliaments since the beginning of the 1990s, as well as the presence of women in cabinets and how this has been influenced by the adoption of gender quotas.
Secondly, we investigate the socio-economic characteristics of these parties' voters and changes made over time and, lastly, we compare the share of people in the two countries who agree with traditional norms on motherhood.
The paper is organized as follows: in the next section we present evidence on policy changes in family policy in Italy and Spain since the early 1990s.The next section looks at the main theories regarding family policy change with specific emphasis on politicalbased explanations.Section four describes our intervening variables.Section 5 develops our three independent variables.

The puzzle: diverging family policy expansion paths in Italy and Spain
Family policies have become key components in post-industrial societies' reorientation of welfare states towards New Social Risks (NSR) (Bonoli 2005).For Bismarckian and certain liberal welfare regimes this implies making the shift from a traditional view that preserves gender roles in family life to a progressive perspective more attentive to gender equality and the socialization of care (Klammer and Letablier 2005).For familialistic countries such as the ones under study here, the focus on NSRs requires intervening in a long-neglected policy field.
In this section, we look at childcare, parental leave and expenditure on cash transfers and tax benefits to families.Although we pay closer attention to the reforms that were enacted from the 2000s onwards, to provide context, our analysis covers policy trajectories since the beginning of the 1990s until the mid-2010s.
Table 1 presents data on public expenditure.In the early 2000s the whole of Southern Europe was still mostly familialistic.Spain, Portugal and Greece, were characterized by low total per capita expenditure concentrated in tax breaks and cash benefits.Italy was characterized by a light "supported familialism" model (Saraceno and Keck 2011) with a medium-low level of total per capita expenditure but also a moderate level of expenditure on services, higher than in Continental countries such as Germany.
Ten years later, the picture in Southern Europe has become more diverse.Greece is moving backwards in terms of expenditure.Portugal has partially expanded per capita expenditure but mostly through transfers and tax breaks, while reducing expenditure for services.Italy and Spain have both significantly expanded expenditure (+70% in Italy and +63% in Spain), but followed two paths.In Italy the growth of expenditure was concentrated on transfers and tax breaks, whereas expenditure for services decreased, with the overall effect of the whole model being based more on tax breaks and cash benefits than in 2003.In contrast, Spain has concentrated its expansionary efforts on services: in 2003, per capita expenditure on services was around one fourth of Italy's, whereas in 2013 it was more than Italy's.Spain is also the only Southern European country that now has a family policy model based on limited use of tax breaks and transfers.Also, the share of expenditure on the latter is almost in line with Germany and the traditional front-runners countries in family policy (the Nordic countries and France), showing continuous efforts towards defamilialization.Insufficient cash benefits to lowincome families with small children has repeatedly been identified as a main hindrance to combat Spain's high levels of child poverty (Ayala and Cantó 2018).
Table 1.Variation of total public expenditure for family policies for infants under three (2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013) (Saraceno 2003).If both parents make use of the leave, the total period is extended to ten months, or eleven when the father takes at least three months (Naldini and Jurado 2013).Naldini and Saraceno (2008: 740) argued that these incentives for fathers to take time off Spain introduced maternity provision much later than Italy.It was only in 1994 that maternity leave appeared as a specific section in Social Security payments, but there have been significant improvements since then (León and Salido 2016).It currently consists of sixteen weeks at 100% of earnings for workers with a minimum of 180 days of social security contributions.In addition, Law 39/1999 for the conciliation of family and working life introduced specific economic protection to pregnant workers who need to stop working for health reasons.Royal Decree 1251/2001 made further improvements to maternity and high-risk pregnancy protection, such as the possibility of suspending the job contract, reductions in working hours and the inclusion of self-employed workers.
Organic Law no.3/2007 for the effective equality between women and men increased the scope of the protection by extending maternity coverage to unemployed women and by eliminating the contributory period for young mothers.In 2014, 86% of working mothers benefited from maternity leave.Thanks to the shorter day due to time allowed for breastfeeding, the number of days are slightly over the statutory period (134.4 days).
Parental leave has been subject to few changes and consists of four unpaid weeks.Selfemployed workers are not eligible.Since 1994, time spent on leave is effectively considered as a contributory period but given the fact that it is unpaid, take up rates are very small.Comparing the two countries, total paid leave available to mothers at the equivalent of full rate is sixteen weeks for Spain and 25.2 weeks for Italy.
In contrast with the limited attention to paid parental leave, Spanish political parties' recent interest in paternity leave has been remarkable.Law no.3/2007 for the effective equality between men and women, approved by the Socialist government of Rodríguez Zapatero, increased paid paternity leave (for fathers only, non-compulsory) from thirteen days to four weeks at 100% of salary, provided that there had been 180 days of prior social security contributions (as in the case of maternity leave).The actual implementation of this policy was put on hold by the austerity measures during the economic recession until it was made effective a decade later by Rajoy's conservative government, which added one more fully paid week.
Moving from leave schemes to services for young children, Italy was an early starter with regards to pre-school.Universalization of early years' education for children aged 3 and older came about with the 1968 law regulating state pre-schools.Spain followed exactly the same path but two decades later.The 1990 National Organic Education Law (LOGSE) introduced three years of preschool education as part of the national education system.Since then, pre-school provision has achieved an almost universal coverage in both countries (at the beginning of 1990, around 84% in Spain and 98% in Italy).
In the early 1990s, however, both countries provided very few services for children under the age of three.It was not until the beginning of the 2000s that they experienced an expansionary trend.Figure 2 shows the situation only in relation to the percentage of infants under the age of three enrolled in public childcare services according to national registries and surveys.Enrolment data in public childcare collected by national institutional sources are more comprehensive than the EU-SILC childcare statistics.Moreover, EU SILC data on childcare coverage also include private childcare providers, thus making it complicated to differentiate between public and private provision (Keck and Saraceno 2011).Figure 2 shows that the differences between Italy  Figure 3 shows that the differences between Italy and Spain in relation to their public expenditure on childcare are coupled with a similar evolution in private provision.
At the beginning of the last decade Spain started from a level slightly above Italy's, but in twelve years, the gap between the two countries increased by almost seven points.
Overall  Overall, what the data on public expenditure, childcare coverage and leave policies show is that over the last two decades both countries tried to strengthen their family policies.Italy has kept spending more than Spain on a per-capita base in covering the needs of families with small children, as it did already at the turn of the century.
However, Spain has started to adopt a family policy model which is explicitly aimed at investing in childcare services and increasing time off for fathers after childbirth, while parental leave continues to be unpaid and cash transfers or tax benefits are kept at low levels.

What the literature tells us about how family policy changes can be explained
A large body of research has looked at the factors that trigger family policy change.Social demand analyses demonstrate how higher levels of education and participation in paid employment among women put pressure on governments to implement more egalitarian family policies, including policies that try to bridge the world of employment and family life.Furthermore, changes in societal values and norms also trigger policy changes (Pfau-Effinger 2010).Ideals and expectations of motherhood (and fatherhood) provide the social context in which family policies develop: if in a given country the idea that the best solution for children is that mothers give up their jobs to take care of their offspring, there may not be an explicit social demand for childcare services but for more traditional family policies (i.e.support for stay-at-home mums).
Social and cultural changes also lead changes in women's political preferences and interest in politics.In most Western countries there is no longer a substantial difference between men and women in terms of voting rates in elections, compared to a studies agree therefore with Gingrich and Häusermann (2015)'s point that women's dealignment from traditional constituencies is greater than men's and that this gives incentives to parties to re-position themselves with regards to policy issues that are of specific relevance to women.
Importantly, the likelihood of these shifts in policies will also depend on the configuration of the party and electoral systems and the type of competition these endorse.
Furthermore, where parties stand in relation to specific issues will depend on the position of parties seen as immediate rivals (Andersen and Beramendi 2012).
In sum, the power resources perspective seems to be moving from a focus on the presence of left-wing governments and women's advocacy groups in politics to greater attention being paid to the question of how and why right-wing parties change their position with regards to family policy, especially in terms of policies aiming at defamiliarization.Our paper engages with this debate.We want to have a better understanding of strategic policy adaptation (or the lack of it) by major political parties; we do so by concentrating on two main factors: the presence of women in politics and changes in the composition of the electorate.We also agree with the importance of a third socio-cultural variable, changes in societal values and norms related to motherhood.

The intervening variable: the positioning of political parties on family policy
To study the differences in program between parties in relation to family policy, we have carried out documentary and content analysis.In a "directed approach" (Hsieh and Shannon 2005) to content analysis, phrases on family policy (including childcare and gender equality) have been tagged, using a priori coding.The list of the 10 specific policies that form the basis for identifying and categorizing relevant party positions in these policy fields can be seen in Tables 2.1 and 2.2.The position of the party in each category is captured using a dichotomous measure that expresses either 'stated support' in the national elections between 1993 and 2015 were included in the analysis.
The results of the content analysis of party manifestos in both countries show two significant differences.Firstly, taking left-wing and right-wing parties together, there is a greater prominence of family/gender issues in Spanish parties than in Italian ones: the total number of positive policy proposals on family policy amounts to one hundred in the Spanish case, and sixty-six for the Italian parties.
Secondly, we observe different policy directions.Our analysis shows a steady divergence of the two main Italian parties on the priorities of different types of family policy.The center-right (FI) has been in favor of increasing cash transfers and tax benefits for families with children since the early 1990s.Support for public childcare services was rarely mentioned, 1 while the implementation of paternity and maternity leaves has never been endorsed in the party's manifestos.The Italian center-left, in contrast, has traditionally shown a slight preference towards strengthening leave schemes and services, although references to tax exemptions for families with children and cash transfer payments have been mentioned more in the most recent party manifestos.Gender equality is an issue touched on practically only by the Social Democratic party.Source: authors' analysis of Italian parties' manifestos In the Spanish case, by contrast, the positions on family policy of the two main parties have been converging since the early 1990s on measures supporting gender equality (Table 2.2).The content analysis of party manifestos shows how the evolution of the Spanish center-right with regards to childcare and gender equality policies has converged with the Spanish center-left since the early 1990s.The extension of childcare services (both public and private), the implementation and extension of maternity and parental leaves and the introduction of tax exemptions for working mothers and families with children under eighteen (up until 2011) have been three policy domains with broad consensus between the two parties.investment of around 780 million euros over a three-year period.
Our findings are in line with Schwander's (2018) analysis on party positions on family policy in electoral campaigns as reported in the media, where she concluded that whilst there was progressive consensus on family policy of the left and the right in Spain, political mobilization of this policy field was rather scant in Italy.

The independent variables: the socio-economic profile of voters, women in politics and changes in cultural preferences
The first of our independent variables -the presence of women in politicsshows a much higher presence of women in politics across the political system in Spain than in Italy.As shown in Table 3, the number of women in cabinet positions has been higher in Spain than in Italy independent of the color of the government.Since the early 1990s both PP and PSOE councils of ministers have had significantly higher percentages of female ministers than their Italian counterparts.With regards to the presence of women in the legislative, if in the 1990s center-right parties in both countries had a limited presence of female parliamentarians in their ranks (in 1995 10% for FI and 15% for the Popular Party), in 2015 this share reached a much higher level in the Spanish PP (36%) than in FI (25%) (Table 4).If we compare the socialdemocratic parties in the two countries, we see a more limited difference, although it is still significant: in 1995 the share of female parliamentarians was higher in Italy (21%) than in Spain (18%), whereas in 2015 this share was higher in Spain (45.6% vs. 38.5%).These different trajectories with regards to the presence of women in the respective parliaments and executives can be explained to a large extent by different paths towards the adoption of gender quota laws.Spain adopted a legislative quota in 2008.This was a consequence of the progressive adoption of gender quotas by left-wing parties from the late 1980s.The socialist party PSOE first adopted party quota of 25% in 1988.In 1997, both the PSOE and the post-communist United Left adopted quota systems that guaranteed parity (at a 40% minimum representation) in their electoral candidatures and party committees.According to Verge (2012), this triggered a contagion effect within the Spanish party system that led to a much higher representation of women already in the late 90s compared to many other European countries.In 1996 in Belgium, Portugal, Italy and France, between eleven percent and fourteen percent of those elected to the national parliament were women, while Spain already had twenty-four percent (Verge 2012).The Women's Political Power Index, which includes women's representation in ministries, parliaments, and regional assemblies, situated Spain well above the EU average as early as 2012 (EIGE 2012).This is line with previous research which has shown the positive impact of women's advocacy groups in the Spanish policy agenda and party system (León, Jun Choi and Jong-Soon 2016;Valiente 1996).
In contrast, Italy's path to quota reforms has been quite contradictory and convoluted (Palici di Suni 2012).In 1993, new electoral laws at national, regional and local levels adopted gender parity in electoral lists.Nevertheless, in 1995, the Italian Constitutional Court declared these measures unconstitutional.In 2001, the center-left majority introduced a constitutional reform which prompted the adoption a gender equality electoral mechanism at the regional and local levels such as the 2/3 same-sex candidates' threshold (Baltrunaite et. al. 2016;Guadagnini 2005).Only recently, the national Electoral Law no.167/2017, the so-called Rosatellum, established a mix of proportional and majoritarian system and was first applied at the national elections in 2018.On that occasion political parties had to present a 'zipper list' for the seats allocated through the proportional system while at least 40% of seats allocated through the majoritarian system had to be for women (Rodríguez-Ruiz and Rubio-Marín 2008).
However, Balduzzi and Voltolina (2018) argue that political parties use some elusive mechanisms hidden in the electoral law in order to bypass this gender quota, thus limiting the expected impact.
This explanation is also supported by the main characteristics and evolution of the different socio-economic profiles of the voters of the main Italian and Spanish centerright and center-left parties (respectively tables 5.1 and 5.2).Some potential variables do not show any particular difference among similar parties in the two countries.There are no relevant differences in the percentages of women among the voters of the two rightwing parties (in both cases women represent around 50 percent of their voters across the different elections), while the proportions of working women among their voters is also very similar.Two other variables, more potentially associated with preferences for conservative values (older age and being religious) do not work either: the Spanish PP has either a similar or a higher share of older voters and church-goers than the Italian FI.
The most striking difference is the much higher proportion (about three and four times higher) of highly educated women among the voting base of the Spanish conservative party compared to its Italian counterpart.This is a difference that already existed in the early 1990s and that persisted over the two decades under study.It is interesting to notice that women with an undergraduate degree represent around ten percent of the Spanish conservative party's total voters, whereas they only represent three percent among FI supporters.Compared to the center-right parties, the two center-left parties are more similar in all respects, including the education levels of their voters.Source: authors' elaboration on micro-data from ITANES (1994ITANES ( , 2013) ) for Italy and from the Centro de Investigaciones Sociológicas (1993, 2004 and 2015) for Spain Source: authors' elaboration on micro-data from ITANES (1994ITANES ( , 2013) ) for Italy and from the Centro de Investigaciones Sociológicas (1993, 2004 and 2015) for Spain These differences reflect the extent to which the two countries have diverged in relation to the access of women to the labor market and to higher education.As table 6 shows, both countries started at the beginning of the 1990s with largely similar situations: the labor market activity rate for women from twenty-five to forty-nine years old was similar; the shares of women with maximum a high school diploma or conversely, with a further education degree were also relatively similar (higher in Spain in both cases); the shares of women in the 25-49 age bracket who were active in the labor market with maximum a lower secondary degree or a tertiary degree were also similar.However, by 2010 female labor market participation and also the participation of highly educated women increased significantly in Spain but much less so in Italy.
Table 6 also shows the existence of significant differences in the diffusion of traditional norms of motherhood, our third independent variable between the two countries.These differences were already present in the 1990s and help to explain the subsequent transformations in family policies.As in Spain, the 1990s and the 2000s were decades in which these two political forces were able to gather a large majority of votes.It was only in the present decade that the decline in percentage of votes given to mainstream parties has been even sharper than in Spain: from around eighty-four percent in the elections held in 2000 to 59.4 percent in 2013.

Conclusions
This article aims to explain the divergent paths of Italy and Spain over the last two decades with regards to changes in family policy.Both countries have started to invest more in family policy, with an overall higher expenditure growth in Italy, but they have gone in very different directions.Italy has placed greater emphasis on cash transfers and tax-based family policy while Spain has been fostering the growth in coverage of childcare services.Our qualitative content analysis of the party manifestos has shown that in Spain, the positions on family policy of the two large parties have been converging to the point where views on gender roles, female employment or child upbringing do not appear to be significantly different.The extension of childcare services (both public and private), the implementation and extension of maternity and parental leaves and the introduction of tax exemptions for working mothers have constituted three broad lines of policy consensus within the Spanish party system since the late 1990s onwards.In contrast, family policy has less presence in the party manifestos of Italian parties during the same period, and, furthermore, there is a left/right divide with regards to the preferred family policy type (the left being more inclined to services, the right more in favor of cash transfers and tax benefits to families with children).In fact, if there is convergence, in more recent times this goes in the direction of the left also supporting monetary transfers over services.
We have identified the presence of women in politics, changes in the socio-economic profile of female supporters and changes in cultural attitudes as contributing to these important contrasting scenarios.During the last two decades, progress towards parity in politics has been far more pronounced in Spain than in Italy, mainly thanks to the stronger adoption of party and legal gender quotas; this has in fact placed the country above the EU average with regards to female representation in national parliaments.The differing pace of the implementation of gender quotas coincides with the different pace at which the mainstream parties have adopted gender equality and more progressive family policies in their political programs.
Regarding voters' socio-economic profiles, although right-wing parties in both countries have a similar level of female supporters (just over half of all voters across the different elections), the presence of highly educated women among the voting base of the Spanish PP is substantially higher than that of FI.This difference was already present two decades ago and reflects a deeper transformation from the social demand side (with regards to higher access of women to employment and further education and less attachment to traditional ideas of motherhood and childbearing).
In sum, since the 1990s Italy and Spain have developed increasingly different sociocultural and socio-institutional environments that have been changing the bases on which family policy has been crafted in the two countries.The result is a path of incremental institutional transformation that makes the two countries, once very similar to each other, increasingly different in relation to gender and family policy issues.

Notes
1. On most occasions, if the expansion of childcare services was discussed in the manifestos, it was in relation to the support for company crèches instead of the strengthening public supply.2. Since the 1990s these two parties have changed their names; for the sake of simplicity we refer to them as FI and the PD, given the stable composition of their internal elites.

Figure 1 .
Figure 1.A model for interpreting the differentiation in family policies in Italy and Spain in the for child rearing represented a major cultural and policy shift, although the change of government soon after, in 2001, left the measure with scarce impact.Between 2001 and 2004, 25% of male workers in the public sector benefited from parental leave, while only 11% in the private sector did(Gavio and Lelleri 2005).Law no.92/2012 (known as the Fornero law) introduced a three-day paternity leave for the first time in Italy as an experimental measure for children born between 2013 and 2015.Out of three days of paternity leave only one was compulsory.In 2015, Decree 80 extended the length of paternity leave to four fully paid days, two of which are compulsory and the other two only if the mother agrees to transfer her days (Sabatinelli 2017).

and
Spain in relation to their public childcare expenditure have become more evident over time: in 2003 public provision coverage in Spain was very low: half the level in Italy.Since then, Spain has increased its coverage steadily; this has also lasted during the austerity years, whereas coverage in Italy has grown at a very slow pace.The result is that in 2015 the public coverage rate in Spain reached 18%(from around 4% in 2003)    five percent points higher than in Italy, where the austerity years practically put a stop on public investment in childcare.

Figure 2 .
Figure 2. Public childcare coverage rate (infants under three) in Italy and Spain, 2003-2015

Figure 3 .
Figure 3. Share of children under three attending formal childcare facilities by type of provision past when men tended to vote more regularly.If in the past female voters were more conservative, since the early 1980s women in most countries appear to have a higher tendency to support the left(Abendschön and Steinmetz 2014;Inglehart and Norris 2003).AsEmmeneger and Manow (2014) have pointed out, the weakening of the religious cleavage has also increased party competition to attract women's votes, especially in Continental and Southern European countries.The realignment of the female vote is also forcing a revision of the main power resources theses.There is a large strand of power resources literature that considers that expansion of welfare provisions is strongly associated with partisanship factors -especially the presence of governments ruled by left-wing parties(Huber and Stephens 2000;Sainsbury 1999).In our two-country comparison, left-wing parties were in power for a similar number of years in Italy and Spain during the period of study (slightly longer in the Spanish case) 3 and therefore we have explored other potential explanatory factors.More recently, other authors have recognized the relevance of the strength of the left but in interaction with a social value dimension.By analyzing the variation of childcare expenditure in eighteen OECD countries,Hieda (2013) concluded that left-wing parties with a libertarian profile are those keenest on expanding childcare, compared with more conservative left-wing parties, thus emphasizing the relevance of value orientation.A complementary set of studies has considered the significant role played by women's capacity to mobilize politically, with a stronger capacity when center-left parties are in office(Stetson and Mazur 1995;O'Connor, Orloff and Shaver 1999; Kabeer   2010).A larger representation of women in politics is found to impact positively on resources allocated to policies which address gender issues more broadly due to a more effective interest representation (Atchison and Down 2009; Lovenduski and Norris 2003; Paxton, Hughes and Painter 2010; Wängnerud 2009).More recent studies (Fleckenstein and Lee 2014; Morgan 2013; Nauman 2012; Schwander 2018) have indicated that in changing political contexts, with much higher levels of electoral volatility and the emergence of new parties, the differences in the ideological positioning of right-wing and left-wing parties are blurring.A higher participation of women in paid employment and attitudinal shifts towards greater gender equality give parties incentives to shift positions towards more progressive family policy stances.This implies a fundamental break with party competition in the past when right-wing parties would endorse more traditional family policies whilst left-wing parties would defend more progressive views.According toMorgan (2013), this is what happened in Germany and the Netherlands when right-wing parties began to promote women's descriptive representation and to shift their family policy agendas.Fleckenstein and Lee (2014) make a similar point, noting that changes in the electoral behavior of womenwho used to traditionally support right-wing parties with a strong religious connotationled political parties in Germany, Sweden, UK and Korea to modernize their political stances towards family policy.Blome (2017) has also looked at the impact of changes in normative beliefs and voting behavior in party competition in Germany and Italy, concluding that while in Germany more egalitarian social norms have triggered policy change, the persistence of traditional normative assumptions in Italy has maintained long-established political constituencies.All these

(
+) or 'absence of reference to the category in question' (-) in the party's policies.A further category (*) has been used to indicate when the policy mentioned in the program as already achieved.All electoral programs presented by the Spanish center-left Partido Socialista Obrero Español (PSOE -Spanish Socialist Workers Party) and the Spanish center-right Partido Popular (PP -People's Party) and the Italian center-left Partito Democratico (PD -Democratic Party) and Italian center-right Forza Italia (FI -Go Italy) 1982 to 2015, the Spanish party system was dominated by two-way competition between the center-left PSOE and the conservative PP.The overall electoral strength of the two main parties steadily increased until the 2000s, reaching eighty-four percent of the total vote share in 2008.Party system stability was guaranteed by the specificities of Spanish parliamentary democracy: strong concentration of power around the executive and asymmetrical bicameralism, produced by a PR system with a majoritarian bias.It is only in more recent years (since the 2011 election) that this trend has been reversed(Bosco   and Verney 2016).However, most policy changes analyzed in this paper took place before 2011, in times of relative electoral stability.Italy shifted from a proportional voting system to a majoritarian-like system in 1993.The majority of Members of Parliament are elected with a majoritarian system but the remainder are elected on a proportional basis.Moreover, the competition has been between blocks/coalitions of parties.For around two decades, until the 2013 elections, the competing coalitions had two parties as pivotal players: Forza Italia on the center-right, and the Partito Democratico on the center-left.2 , if we consider public and private coverage together, in 2003 Spain's coverage incentivizing households to buy childcare on the private market) but we can conclude that this increase in expenditure in Italy did not translate into a strong indirect support for acquiring private childcare services Households in Spain have invested more than Italian households in private childcare services, even though on average they have fewer economic resources, if measured by GDP per capita.

Table 2 . 1 .
Positions of the two main Italian parties on family and gender policies

Table 2 . 2 .
Positions of the two large Spanish parties on family and gender policies in the expansion of family policy.In 1997 with Law 285 on childcare services; in 2000 and 2001 respectively with the Law 53 and Decree 151 on maternity and parental leave provision; in 2008 with an "Ad hoc public childcare expansion plan", with an just a year after the PSOE government expanded the existing provision from thirteen days to four weeks (see section 3).Meanwhile, in Italy only left-wing coalitions have shown interest

Table 4 .
Women in Parliament in the two main Italian and Spanish Centre-Right and Centre-Left parties Source: information retrieved from Spanish and Italian Parliaments' websites

Table 7 ,
Table7explores this socio-cultural dimension.The percentage of individuals who agree with the sentence 'the child suffers if the mother works' is not significantly different by gender in either country.Nor does age make a difference, although this cohort effect between the older generations and the rest of the population is much higher in Spain.When we look at preferences by party proximity, politics plays a role only in Italy but not in Spain.Voters of center-right PP and center-left PSOE have similar preferences with regards traditional attitudes whilst in Italy center-right and left voters have quite different preferences.As it can be seen in FI voters are almost totally against working mothers (over 86% for FI and 69% for PD).Overall, Italian voters hold more traditional views on motherhood compared to their Spanish counterparts.

Table 6 .
Socio-demand factors: a comparison between Italy and SpainNote: the most important differences are in bold; traditional norms concerning motherhood are measured by the share of individuals who agree with the statement "young children suffer if the mother works".* Data refer to women from 25 to 49 years ** Data refer respectively to years 1999 and 2008 Source: Eurostat online database for labor market statistics; authors' own elaboration from the European Values Survey microdata for norms on motherhood.

Table 7 .
(Blome 2017;Schwander 2018hat the incentives for the Spanish center-right to converge with the center-left on gender issues are stronger than those of the Italian center-right.They support findings from previous research(Blome 2017;Schwander 2018) that the social demand factors which trigger change in the political arena have remained relatively stable in Italy, whilst party competition to attract women's votes (especially highly educated ones) has been stronger in Spain, leading the main conservative party to support childcare expansion and engage in progressive family policy.One explanation that does not seem to hold in our research is the one related to electoral volatility in the two countries.In the last quarter-century, Italy and Spain have used electoral systems that are a blend of majoritarian and proportional systems (PR).From -demand factors: share of individuals who agree with traditional norms on motherhood in Italy and Spain by personal characteristics and political preferences (%)(2008)