Article Journal Of European Social Policy Measuring state de-familialism: Contesting post-socialist exceptionalism Jana Javornik Umea University, Sweden; University of Leeds, UK Abstract This article offers a methodology to reveal the latent constructs that underlie policies on parental leave and childcare services. It is constructed to uncover state assumptions about the social organization of childcare and gender roles in a country-comparative perspective. Legislation is central to this analysis, and combinations of policy components take centre stage. An index of state de-familialism is proposed and its analytical potential tested on eight post-socialist European Union (EU) states. Grounded in Leitner's conceptualization of "varieties of familialism", it gauges three policy types: (I) Slovenian and Lithuanian supported de-familialism incentivizes women's continuous employment and active fatherhood; (2) explicit familialism in Hungary, the Czech Republic and Estonia supports familial childcare and reinforces gendered parenting; and (3) implicit familialism in Poland, Slovakia and Latvia leaves parents without public support. These groups share core characteristics with developed welfare state regimes. This methodology has the potential to discredit claims of post-socialist exceptionalism and allows researchers to test new hypotheses. Keywords Central and Eastern Europe, childcare, comparative analysis, policy index, state de-familialism Introduction Parents increasingly have careers, not just jobs, and ever more women and men expect to 'have it all', but it remains embedded in the 'culture of social obligation' that women are the primary caregivers (for example, Daly, 2002: 262). This 'gender contract' (Gottfried, 2000: 253) affects mothers' employment opportunities and limits their possibility to compete with men for the best paying jobs. Moreover, it constrains fathers' parenting possibilities. This division of unpaid work is not trivial, especially for women, who remain most likely to accommodate their economic Journal of European Social Policy 2014, Vol. 24(3) 240-257 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0958928714525815 esp.sagepub.com activities in order to carry out family responsibilities; Land and Rose (1985; in Orloff, 1993: 313) call this 'women's compulsory altruism'. For various reasons, welfare states have designed policies aimed at solving the problems that arise from this inherent conflict. These policies set 'the Corresponding author: Jana Javornik, School of Sociology and Social Policy, University of Leeds, LS2 9JT Leeds, UK. Email: j.javornik@leeds.ac.uk Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 Javorník 241 rules of the game' (North, 1990). They endorse, legitimize and reproduce the 'ethos' of social norms and practices (Hobson et al., 2011: 173; Misra and Akins, 1998). Following the feminist critique of welfare state regimes, normative assumptions about the social organization of care and gender roles most clearly underpin regulations on parental leave and childcare services (for example, Ferrarini, 2006; Knijn and Kremer, 1997; Leitner, 2003; Rubery et al., 1998: 223-34; Ruhm, 1998). These constrain parents' choices, and hence their 'opportunities to be and do' following childbirth (Hobson et al., 2011: 169). I call this policy conceptual logic. Welfare states choose different policy combinations. These have various implications, one being the disruption of female employment following childbirth (for example, Gornick and Meyers, 2003; Misra et al., 2010). To illustrate, parental leave may come with penalties, including job loss or discriminatory treatment in pay and promotion. Thus, for some parents, the decision to take leave could be an option on which to fall back instead of being a first choice, especially when childcare services are limited. Alternatively, welfare states might use these policies to challenge the normative parenthood ideals (for example, Ciccia and Verloo, 2012; Leitner, 2003). A good example of this is Sweden. Since the 1970s, the state has promoted active fatherhood, and shared parenting has increased as a result. Since more men use longer parental leave, the problems associated with taking leave have decreased. This suggests that government initiatives could also transform gendered roles. I call this policy transformative potential, to refer to the degree to which the states support a dual-earner/dual-carer family model (Gornick and Meyers, 2003). Policies on parental leave and childcare services are two sides of the same coin (Misra et al., 2010). Parents can view and experience them as a 'package' that constrains their options following childbirth. Therefore, to fully understand their implications, the underlying logics of these policies must be considered, although previous comparative analysis is limited in its ability to explain their various configurations and implications. There are two main reasons for these limitations. First, previous research on the implication of welfare state regimes tends to rely on welfare state output which is problematic if the data are less than ideal. There is a limited amount of cross-national data on comprehensive family policies. Time-series in international databases is short, and available indicators on policy delivery and use are scarce or flawed. Administrative data, which often are not translated, are inconsistent across countries. Therefore, we see a tradition where policies are measured with subpar data such as the generosity of leave policies (the correlation between leave duration and payment), public spending on childcare, participation rates and number of hours in public childcare and staff-to-child ratios. Serious methodological issues underlie these data (for example, Fagan and Hebson, 2005; Gilbert, 2008: 129, 145; Lambert, 2008: 316; Plantenga and Remery, 2005). For instance, states may use different funding streams to finance childcare services, yet this is not reflected in national statistics of state investment in childcare if municipalities do not report their spending (for example, Organisation of Economic Co-operation and Development (OECD), 2013: 3). Moreover, international databases on public expenditure cover information on education-related programmes while often excluding care-related services. This is a critical omission because public support for parents with children aged 0-3 years is concealed. Second, research tends to combine independent and dependent variables. Previous studies typically use indicators of policy provision and uptake; however, these are framed by a range of factors such as rules and regulations, national funding streams (for example, OECD, 2013), social norms, preferences and household income (for example, Gornick and Meyers, 2003; Lambert, 2008). For example, parents' use of entitlements reflects social norms and values that influence whether they feel entitled to claim them. To illustrate, Swedish parents do not use all the public childcare that is available to them (Gronlund and Javornik, 2013). This pattern appears to be informed by their notion of a 'good worker/good parent' and the need to put children first (Hobson et al., 2011). Therefore, by using the number of hours in childcare to discern policy logics, we could draw erroneous conclusions about the availability of Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 242 Journal of European Social Policy 24(3) public childcare. Notwithstanding their weaknesses, these measures are often presented as an intellectual force with an autonomous quality of their own while critical concerns are often missing.1 This is where the contribution of this article lies. I will present a critical perspective on the power of 'standard' policy measures to explain policy logics and will offer an approach that puts legal formulations on parental leave and childcare services in the centre of analysis. To capture state assumptions about social organization of care and gendered roles, I will expand upon Weber's (1949) use of 'ideal types' and evaluate policies across multiple dimensions. Then, I will examine their combinations and classify policies using the 'varieties of familialism' framework. Thus, this article will be structured in four sections. First, I will consider the issue of how to think about policies in the area of childcare in terms of social organization of care, mothers' employment and active fatherhood. Then, I will briefly review earlier approaches to comparative policy analysis. Drawing on these, I will propose a new approach and investigate its analytical potential on eight post-socialist European Union (EU) countries. Finally, I will conclude by expressing critical reflections and possible directions for future research. Theoretical perspectives Research on the gendered welfare state views the state as 'not just a set of services; it is also a set of ideas about society, about the family, and - not least importantly, about women who have a centrally important role within the family, as its linchpin' (Wilson, 1977: 9). The feminist response to Esping-Andersen's (1990) welfare regimes has contributed through providing important accounts of crosscountry variation in women's experiences (for example, Fraser, 1994; Leitner, 2003; Lewis, 1992; Orloff, 1993; Sainsbury, 1996; Saraceno and Keck, 2008). This literature maintains that welfare states vary in the extent to which governments reinforce or challenge gendered roles. Orloff (1993) describes this by saying, ' [T]he character of public social provision affects women's material situations, shapes gender relationships, structures political conflict and participation, and contributes to the formation and mobilization of specific identities and interests' (pp. 303-4). Considerable bodies of research demonstrate that parental leave and childcare services have the highest explanatory power for cross-country variation in female employment (for example, Eliason et al., 2008; Misra et al., 2010; Pettit and Hook, 2005; Rubery et al., 1998: 223-34; Ruhm, 1998; Uunk et al., 2005). These not only frame the ways in which women engage in employment but also how men engage in parenting (for example, Leitner, 2003). For example, Haas and Hwang (2008) argue that state incentives for father's use of leave affect active fatherhood throughout the family cycle, while Pylkkiinen and Smith (2003) report a 'substitution effect': when fathers use leave, mothers resume employment earlier which positively affects their job prospects. Therefore, welfare states use policies to determine 'proper' parenthood ideals, and these create both incentives and disincentives for shared parenting (for example, Ciccia and Verloo, 2012; Gornick and Meyers, 2008; Hobson et al., 2011; Leitner, 2003; Smith and Williams, 2007). Leitner (2003), however, maintains that policy conceptual logics reflect how welfare states choose to allocate childcare responsibilities between the state, the family and the market, and within the family between men and women (see also Saraceno and Keck, 2008). Leitner (2003) organizes policy combinations in a four-type framework of state de-familialism: explicit and implicit familialism, de-familialism and optional de-familialism. This framework is concerned with the degree to which the welfare states assume and support family interdependencies, and reinforce, or challenge, gendered care giving. Familialism promotes and supports childcare by the family (familial childcare). States do this either implicitly by leaving parents without publicly financed support or explicitly by rewarding them with public money to provide childcare themselves (that is, 'caregiver parity model' by Fraser, 1994). In contrast, de-familialism promotes the dual-earner family model (i.e. the 'adult worker model' by Lewis, 2001: 154), with the state investing in publicly funded childcare, whereas optional de-familialism allows parents to choose between preferable childcare options. Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 Javorník 243 Such conceptualization also considers the transformative potential of leave (the familialistic policy element) which can 'provide incentives to ensure that care provision is shared on equal terms among male and female family members' (p. 367). This opens up the analytical space for comparative research. First, Leitner's (2003) 'varieties of famil-ialism' distinguishes between policy areas, such as childcare and old age, relative to the broad welfare state regimes. Second, legal regulations on parental leave and childcare services become central analytical dimensions relevant to parents with preschool children. Third, it questions assumptions in which policy dimensions combine to reflect different policy logics. Leitner (2003) argues that different combinations frame different opportunities or constraints. Hence, this concept captures policy transformative potential, such as the extent to which the state challenges gendered roles. Moreover, it recognizes that de-familialism may occur through both the state and the market, although the two are assigned different conceptual status, acknowledging that recourse to the consumer market is mediated both by family resources and alternatives available through parental leave. In summary, this conceptual framework offers fruitful lines for analytical developments about the gendered implications of the welfare state in a country-comparative perspective. Country-comparative family policy research Comparisons play an important role in social policy analysis and welfare state research. Applications of various methods in this research have grown, including comparisons of welfare state regimes and policy typologies, while also studying changes over time. On the one hand, this literature provides invaluable information about policies across countries. On the other hand, quantitative analyses provide important information about policy implications. Country-comparative research includes increasingly national data. The selection of cases and appropriate research strategy is thus a challenge; important factors of comparative policy analysis are thus the ways in which we source, link and present policy information. Parental leave and childcare services are 'complex objects of investigation' (Ciccia and Verloo, 2012: 507). Their complexity derives from the fact that they are made from a multiplicity of relevant components. These include the duration and payment of leave, flexibility, eligibility, accessibility, affordability and quality of services, while each element carries implications for the policy as a whole (Javornik, 2010; Leitner, 2003). This study will assess multiple policy components as structural elements that can combine into a 'package' which is available to working parents following childbirth. Considering their impact on gender roles and parental employment, it will, in essence, offer a new typology for understanding this policy area. Policy indexing Sainz (1989) argues that the heterogeneity of countries necessitates flexible analytical approaches that combine various types of information. Composite measures are increasingly recognized as useful analytical instruments that can capture multifaceted concepts (for example, Booysen, 2002: 145; Ragin, 2009: 5). This type of research, in which child-related policies are included, has developed a series of indices that can measure gender equality (for example, Kershaw, 2010; Plantenga et al., 2009), active fatherhood (for example, Smith and Williams, 2007), work-family reconciliation (for example, Gornick et al., 1997; Gornick and Meyers, 2003; Lambert, 2008; Plantenga and Hansen, 1999; Ray et al., 2010), child-related benefit programmes (for example, Bradshaw et al., 2007; Mandel and Semyonov, 2005) and family poverty (Misra et al., 2007). In this field, the approach by Gornick et al. (1997) and Gornick and Meyers (2003) represents a seminal attempt to include multiple policy dimensions. The authors offer three indices on leave, childcare and early-school services, which are considered gendered policy incentives for working parents. These comprise 22 indicators, of which 20 are statistical data, such as participation rates in day care, and two are textual, such as service quality and tax relief. Previous efforts offer tremendous opportunities to analytically systemize policy data. They open up the possibility for new types of scholarly investigations Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 244 Journal of European Social Policy 24(3) and activities in comparative research. First, they address the need to investigate policies with multiple dimensions and advance legal regulations as principle ingredients of comparative analysis (largely used in studies on leave, for example, Bruning and Plantenga, 1999; Ciccia and Verloo, 2012; Ferrarini, 2006; Ray et al., 2010). Second, they demonstrate that effective composite measures depend on comparable and valid metrics. No straightforward rule yet exists to decide which and how many indicators to use, however. In earlier work, selection depends on research focus, theory, pragmatism and intuitive appeal (Javornik, 2012a). As a consequence, existing indices neither focus on policy models nor on policy conceptual logics (Ciccia and Verloo, 2012; Hudson and Kuhner, 2009). Instead, they often combine conceptually different indicators (family income, employment rates) across multiple policy areas (childcare, employment, social security). Although comprehensive, these all-inclusive indices may conceal the latent constructs that underlie these data, rendering the complexities invisible. Given the lack of theoretical assumptions in which policy dimensions combine to reflect different policy logics, several questions remain open. What are the central aspects of parental leave and childcare services? How do they combine to reflect gender and parenting norms? How are these ideal types operationalized? Measuring state de-familialism I propose a different approach to evaluating national legislation on parental leave and childcare services and how they conform to Leitner's (2003) ideal policy types. Using Weber's ideal types, I have developed an index of de-familialism; this measures the degree to which the state supports women's continuous employment and promotes active fatherhood. This approach aims to bridge area-specific and country-comparative policy analyses, to connect analysis with theoretical models of varieties of familialism, to map the differences across countries and to yield a grouped variable with enhanced explanatory power that can be used in quantitative analysis. This approach is based on the premise that policies are conceptually rooted, and hence their logics can be best understood in terms of legal formulations. Through laws, the state regulates who should be the primary caregiver and who should bear the costs of childcare. The central concerns of this analysis will be related to the extent to which the states: (1) support public childcare, (2) financially compensate family care and (c) expect women and men to engage in childcare. To assess parents' opportunities and constraints, the policies I will be considering will be for children from birth until the mandatory school age. This approach builds on the assumption that stages in people's lifecycles are causally linked; the choice of motherhood is a key element in family decisionmaking due to the costs associated with employment interruptions (for example, Esping-Andersen, 2009: 24). I will focus on working parents with preschool children because of the volume of their childcare responsibilities. Younger children demand more attention and time as compared with older children who are more independent. Moreover, older children are obliged to go to school; hence, access is granted without delay, and childcare needs subside (for example, Gilbert, 2008: 6; Gornick et al., 1997: 54-5). There are a number of reasons why such an approach to comparative analysis is appropriate for the purposes of my investigation. First, multidimensional-ity is its key advantage. This is important because policies comprise multiple components; each of these defines the policy as a whole (Ciccia and Verloo, 2012). Second, this approach can integrate various types of policy components - be it a quantitative or qualitative attribute (Booysen, 2002). This is important because of how policy logics are reflected in legislation (for example, Gilbert, 2008; Gornick and Meyers, 2003; Ragin, 2009: 5). To my knowledge, no study has employed this information exclusively, at least not beyond policies on leave (for example, Ciccia and Verloo, 2012; Smith and Williams, 2007). Furthermore, legal regulations are largely textual data; the possibility to use these types of data enables us to go beneath the surface, striking a better balance between the simplification and complication of policies (for example, Ragin, 2009: 5). Third, I will use graphs to summarize and visually present policy information in a comprehensible way. In summary, this Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 Javorník 245 approach aims to capture policy combinations and highlight inconsistencies therein. Hence, it will facilitate country-comparative analysis. Constructing an index with benchmarking The proposed approach entails four steps: selection of policy components, their assessment and aggregation, and validation (Booysen, 2002). I will assess and score each policy component against a set of theoretically guided standards. Then, I will aggregate the scores into an index of state de-familialism that measures policy logic: the higher the score, the higher its transformative potential. Finally, I will systemize this information using a graphical method of spider charts that will provide a synoptic overview of policies. Selection of policy components In this study, I analyse policies on parental leave and childcare services as a single policy programme related to childcare. These are distinct from other work-family measures in two ways. First, they frame parents' opportunities and constraints following childbirth. With reference to empirical studies of female employment, they critically frame women's return to employment following childbirth, whereas other work-family policy measures (for example, working time, workplace flexibility) frame the ways in which women return to work, and parents combine multiple roles. Second, they reflect a cultural script for socially acceptable allocation of childcare, and hence for 'proper' parenting.2 Central analytical categories are legal formulations on parental leave and childcare services. These provide the right to time off from work for parental childcare and the right to external childcare. I evaluate them across 11 attributes from which assumptions about their implications can be drawn. I examine them as single objects of inquiry, and then through their combinations. I draw the defamilialistic marker between parental leave (familialism) and childcare services for children aged from birth to school age (de-familialism). Policies on leave - maternity, paternity, parental and extended childcare leave - support family care while allowing parents to stay connected to employment (Misra et al., 2010). Maternity, paternity and parental leaves are childbirth-related and accompanied by earnings-related benefits, whereas extended childcare leave refers to longer leave that enables family care; the latter is either paid or not, and available to wider groups of parents. Henceforth, I use 'parental leave' when I refer to all these types. When more than one type is granted, I follow the rule of the most 'familialistic' option and consider the longest period available. Given that in some countries differential arrangements are possible, such as in the case of multiple births or the birth of a sick child, I follow Smith and Williams (2007) and consider the least generous regulation. Leave consists of six components (Table 1). Limited theoretical knowledge about the implications of some policy components prevents me from considering certain legal aspects such as eligibility according to parents' employment histories and benefit payment caps. Since there is no standard way to determine what is optimal, I could draw erroneous conclusions (Bradshaw et al., 2007; Plantenga et al., 2009: 25). Exclusion of these regulations is, however, acceptable (Ciccia and Verloo, 2012; Smith and Williams, 2007), considering that the focus is on norms and not on actual outcomes of policy regulations. Childcare services support children's early education and help parents combine family with employment (for example, Misra et al., 2010). I include regulations on publicly funded services for children aged from birth to the compulsory school age in centre-based day care and focus on five aspects that pertain to the availability, affordability, quality and intra-country disparity in service provision. When municipalities are responsible for childcare services, I check whether national regulations ensure that these are comparable across the state. Optimal policies There is no simple formula to determine the optimal regulation, and the preferences, social norms and workplace cultures will often shape what is considered optimal. In this study, assessment criteria are Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 246 Journal of European Social Policy 24(3) Table I. Policy dimensions, assessment criteria, scores. Score Policy dimensions and assessment criteria Total leave time 8 Minimum 6 months, maximum I year 4 More than I year up to 2 years 2 More than 2 years up to 3 years 1 Less than 6 months or more than 3 years Leave financial sustainability (maternity + parental) 8 Paid at 85-100 percent of previous earnings for minimum 6 months 4 65-84 percent of previous earnings for minimum 6 months 2 50-64 percent of previous earnings for minimum 6 months 1 70-100 percent for 3 months or less, then means-tested flat rate or unpaid Job security 8 Leave users retain a contract and associated benefits to previous job or equivalent during parental leave 4 Job-secured leave shorter than earnings-related leave 2 Leave users retain a contract, but leave does not guarantee a full set of rights for persons returning from leave 1 Job not protected Parental entitlement (excluding maternity and paternity) 8 Individual right of each parent, joint decision, change in arrangement possible and not limited 4 Individual right of each parent, joint decision, fixed arrangement 2 Fathers entitled to shorter portion of parental leave 1 Individual right of mothers Flexibility (including parental and extended childcare leave) 8 Leave can be used in sections over a longer period (minimum 6 years), different combinations possible 4 In sections over 3—4 years or a portion of leave can be reserved and used flexibly (in sections) before child's school age 2 In one block, limited gainful activity allowed with proportional reduction in benefits and leave time 1 Full-time familial care, no gainful activity allowed/gainful activity allowed but familial care required (not in day care) Father's non-transferable entitlement 8 Minimum I month at 100 percent of previous earnings 4 Two weeks at 100 percent following childbirth + extra days at lower payment spread over a longer period of time 2 Less than 2 weeks, at 80-100 percent 1 No individual provision or less than 80 percent Availability of childcare services Allocation of places - access criteria 8 Legally set access criteria, central capacity/demand planning (including creches when ran separately) 4 Legal guidelines on access criteria (including creches), providers allowed to add criteria when demand exceeds supply, no central capacity/demand planning 2 Conditional/various access to creches, legally set nationwide access criteria to kindergartens, no central planning 1 Providers autonomous in setting access criteria, no central planning Admission age 8 No lower age limit or child can be admitted before the end of paid leave 4 Public childcare and paid leave are congruous 2 Time gap between paid leave and childcare services 1 Intra-country variation in admission age Compatibility of service hours with working hours of parents 8 Prescribed full-time with flexible provision to accommodate parents' care needs, around the year 4 Prescribed to cover typical day/week/year, limited flexibility 2 Prescribed to cover typical day/week, spells of shorter breaks allowed (limited alternatives available) I Variant opening hours across municipalities (Continued) Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 Javorník 247 Table I. (Continued) Score Policy dimensions and assessment criteria Affordability Parental fees 8 Sliding-fee scale based on family income for both creches and kindergartens 4 Sliding-fee scale based on criteria other than income for both creches and kindergartens 2 Providers autonomous in setting rules for creches, legal guidelines for kindergartens 1 Providers autonomous in setting rules, no legal guidelines/ceiling Quality National co-ordination of service delivery 8 Services joint responsibility of state and municipalities, legally set operating standards 4 Services legal duty of municipalities (to establish and administer both creches and kindergartens), regulated standards and rules of operation 2 Legal duty of municipalities (both creches and kindergartens), providers autonomous in some elements of provision I Creches at discretion of municipalities (legal right), kindergartens prescribed (legal duty), autonomy in provision sourced from the predominant theories and empirical evidence on female employment and gender equity. This approach is based on endorsing maternal employment and active fatherhood and I consider the optimal policy type to be 'supported de-familialism'. This is characterized by gender-neutral, well-paid and flexible parental leave with incentives for active fatherhood. After 1 year, publicly financed childcare becomes a pronounced alternative to familial child-care. States accommodate the childcare needs of as many parents as possible by prescribing accessible, affordable and high-quality public services across the state. Each dimension is scored against a set of theoretically guided standards (see Table 1). Parental leave Leave time. Parental time following childbirth is a valued commodity, and leave that is either too short or too long can have detrimental effects on women's likelihood to withdraw from employment. Earlier studies show that a 1 -year leave has the fewest negative implications (Bruning and Plantenga, 1999: 207; Gornick and Meyers, 2003: 122; Pettit and Hook, 2005; Wall et al., 2009: 36), but only when the job is safeguarded (Fagan and Hebson, 2005; Jaumotte, 2003: 17). The downturn effect is identified at 6 months and 2 years, with both shorter and longer leave as having the most adverse effects on female employment and gender equity (Leitner, 2003: 370; OECD, 2007: 118; Pettit and Hook, 2005). In this study, both receive equally low scores because they could lead to women's complete withdrawal from employment, especially for women in less protected jobs (Fagan and Hebson, 2005: 90). Monetary value of family childcare. Income-replacement levels determine who claims the entitlement. When payment is low, the magnitude of household income shock is lower if used by women, who on average earn less than men (Leitner, 2003: 372). Entitlement disadvantages carers and reinforces gendered parenting (Fagan and Hebson, 2005: 8, 89). In this study, financial compensation is optimal when leave is paid at a minimum of 85 percent of previous earnings because fathers increasingly use only fully paid time off. Gornick and Meyers (2003: 122) set the optimal replacement level at 80 percent and Wall et al. (2009: 36) at 70 percent, which I consider a crossover point. The policy facilitates shared parenting by granting leave as an individual right to both the time and the payment (Bruning and Plantenga, 1999: 196; Fagan and Hebson, 2005: 95; Leitner, 2003: 368). The state may also use the force of fully paid nontransferable 'daddy-quotas' (for example, Pylkkiinen Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 248 Journal of European Social Policy 24(3) and Smith, 2003). These provide normative guidelines for 'proper' fatherhood, increase fathers' sense of entitlement and their opportunities for making claims at the workplace (Hobson et al., 2011). Finally, when flexibility in uptake is allowed, parents can stay connected to the workplace; this incen-tivizes fathers' use, speeds up mothers' return to employment and has less detrimental effects on their job progression and lifetime earnings (for example, Plantenga and Remery, 2005: 48). Childcare services. Optimal childcare services are regulated and overseen by the state. A guaranteed childcare system is provided when services are affordable, of high-quality and available across the state during a typical working day/week/year. Ideally, the state would give parents a choice between parental or public childcare (that is, optional de-familialism). This is reflected in the social organization of childcare, especially over the first 3 years of a child's life. In this study, the state allows for a realistic choice when it opens access to childcare services before the earnings-related parental leave ends (Fagan and Hebson, 2005: 105; Plantenga and Remery, 2005: 38). A second variation is also possible, in which mothers' employment is given prominence and earnings-related leave and services are contiguous. Furthermore, by guaranteeing a place in day care without delay, the state sends a clear message about social acceptability of public childcare (Gornick and Meyers, 2003: 197-206). Care is continuous, and parents can resume working, but when a child can be refused a place based on preferential criteria, or local governments are autonomous in setting access priorities, policy can create tensions and the state contributes to varied availability (for example, Fagan and Hebson, 2005: 108-9). Opening hours determine the extent to which services can serve as a supervised form of childcare (Gornick and Meyers, 2003: 227). For parents to have realistic opportunities to use the services, these should be available on a full-time basis and allow for a flexible provision to accommodate the different childcare needs around the year (OECD, 2007; Plantenga and Remery, 2005: 38—42). When opening hours are incompatible with parents' working hours, or vary across providers, then policy creates tensions and affects mothers' employment opportunities (Kreyenfeld and Hank, 2000: 318-19). Childcare costs are the equivalent of a regressive tax on mothers' labour supply and reduce financial returns from employment (Esping-Andersen, 2009: 91). Affordable childcare services generate incentives for maternal employment, especially for women with limited earning prospects (for example, Fagan and Hebson, 2005: 12). An optimal policy, then, provides services free of charge. Almost no country, however, provides such entitlement. Thus, an optimal entitlement is a legally prescribed parental fee on a sliding-fee scale (Gornick and Meyers, 2003: 206). When providers are autonomous in setting the price and operating mechanisms on who pays how much, the state allows for variants of affordability, which hinders access to public services (for example, Fagan and Hebson, 2005: 106-9). High-quality service generates incentives for using public childcare (Plantenga and Remery, 2005: 35). When the state prescribes and regulates minimal standards and oversees provision, services are considered trustworthy and of comparable quality across the state (for example, Gornick and Meyers, 2003: 195; 218-26). When providers are autonomous in service provision, the state contributes to variant service quality, and this hinders access to public childcare (Kamerman, 2006). Scoring Adapting Smith and Williams' (2007) approach, each policy component is scored repeatedly across countries using a scale with a four-value set of 1-2-4-8 (Table 1): • 8 indicates that the component is close to the optimal regulation • 4 that it is moderately close • 2 that it is far from optimal • 1 that the component is the furthest from the optimum The idea behind this scale is a geometric sequence, so that logarithms yield an arithmetic sequence; 8 is a maximum because 10 or 100 (commonly used in Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 Javorník 249 earlier studies) have too strong an influence on the results.3 Since the index is computed using the multiplicative method (explained below), there would be problems if figures were 0 (multiplying by 0 means losing information). Aggregation Adapting Smith and Williams' (2007) approach, the index score is derived through a standard mathematical procedure using the following formulae: Leave index = %jsumproduct Day-care index = %Jsumproduct Composite index = \y sumproduct (sumproduct = sum of products of individual scores) The final product score can be interpreted as the mean value of a binominal distribution (Argasinski and Kozlowski, 2008: 253), and the method may be used for each policy separately. The scores range between 1 and 8 and reveal how far the policy is from the optimum: only policies with consistently high scores on components obtain a high overall score, and vice versa. The advantage of this method over the additive (used in earlier studies) refers to 'the inequality of arithmetic and geometric means' (Beckenbach and Bellman, 1961: 4). Theorem 1 explains that when we take an average over a set of n positive numbers by dividing their total sum by n, the value we get is never smaller than the one we get by taking the nth root of their product. Thereby, the additive method always yields a more favourable picture, but the multiplicative increases the difference in fitness and is more neutral (Smith and Williams, 2007: 183).4 In the absence of any theoretical or empirical justification, I employ no explicit weighting other than the weights implicitly introduced during the scoring (following Booysen, 2002: 131—42; Bradshaw et al., 2007: 24-5; Plantenga et al, 2009: 25; Schütz et al, 1998: 41-3). Thus, all dimensions are considered relatively equal in terms of equal intrinsic importance as a component (Ura, 2008). This also allows decomposing the index by policies and/or components. Validation: Application to post-socialist countries The proposed approach invites policy analysis across countries and over time. To test its applicability empirically, I extend it to post-socialist EU states that include the Czech Republic, Hungary, Poland, Slovakia and Slovenia, Estonia, Latvia and Lithuania. These countries are a good choice for this study, both empirically and theoretically. First, they share a distinctive history with a common background in terms of state socialist legacies, transition and EU accession processes. These contexts shaped their institutional legacies, not only in terms of female employment but also in the state's role in the family (Javornik, 2010, 2012b, 2014). Second, welfare regime literature often views them as a distinct type. Conventionally, however, only Hungary, the Czech Republic and Poland are included (with few exceptions, for example, Ciccia and Verloo, 2012; Szelewa and Polakowski, 2008). This opens the question of 'representation' of 'former communist' regime (for example, Esping-Andersen, 1990; Hantrais, 2004; Rostgaard, 2004; Saraceno and Keck, 2008: 22). I find their 'typicality' a litmus test to determine whether this method can uncover state de-familialism. To date, only a few studies have challenged their homogeneity, suggesting that these countries provide conceptually different policies (for example, Ciccia and Verloo, 2012; Szelewa and Polakowski, 2008; Thévenon, 2011). In this article, Szelewa and Polakowski's (2008) study represents a touchstone for my approach because it includes the same eight countries, is framed in Leitner's (2003) 'varieties of familialism' and includes both parental leave and childcare services. The authors, however, employ a different method (fuzzy sets), operationalize policies with more 'standard measures' and omit childcare services for children aged 0-3 years.5 Since the purpose of this exercise goes beyond classifying or tracking institutional change, I report Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 250 Journal of European Social Policy 24(3) policies for a single year (2008), in order to avoid changes caused by the recent economic recession -an important empirical question but one that remains outside the scope of this analysis. Data come from the Mutual Information System on Social Protection in EU, Eurydice, OECD Education at a Glance and Family Database, Council of Europe, United Nations Children's Fund (UNICEF), national administrative and other relevant agencies and ministries. I also contacted national experts and agencies, who reviewed policy information. Varieties of familialism: Index Mandel and Semyonov (2005) maintain that the index is 'a composite phenomenon with consequences that go beyond the unique effects of each of its components' (p. 964). I agree that the index is a grouped variable with enhanced explanatory potential. As such, it could be used in analysis of policy implications, for example, in order to investigate variance in maternal employment6 or to monitor policy change and progress over time. Figure 1 reflects the degree of state de-familialism and a transformative potential of national policies on parental leave and childcare services: the higher the score, the higher the support for dual roles. It shows how close to the optimum each country was in 2008: with 73 percent of the achievable score, Slovenia comes closest to supporting dual roles of women and men, while Poland, Slovakia and the Czech Republic are the farthest. An index score, however, is not sufficient if one hopes to fully understand the policy design. Therefore, to determine policy configurations and identify critical policy components and any inconsistencies, I will employ a graphical analysis. Varieties of familialism: Graphical analysis To provide a synoptic overview of policies and their distance from the optimum, I use spider charts (for example, Plantenga and Hansen, 1999; Schütz et al., 1998). Each chart represents one country and comprises 11 equiangular spokes, one for each policy component. Their length is proportional to the score, and ranges between 1 and 8: the higher the score, the closer to the optimum. The line connects them into a radial figure that gives each policy a spider web appearance: the larger the area, the more optimal the policy. The advantage of this method is that policies can be easily read from the chart: the right-hand side illustrates leave, with childcare services on the left, and the synoptic overview reflects their overall policy configuration. Figure 2 shows how diversified national policies are, with only Slovenia and Lithuania resembling spider webs. In the other six countries, familial childcare is well-entrenched with policies cast in traditionally gendered roles. These countries largely grant leave as a family entitlement, with limited flexibility in uptake and paternity quota. Then again, countries significantly differ in the monetary value of familial care: while Hungary, Estonia and the Czech Republic financially invest (more) in familial care, Poland and Slovakia do not. Maternity leave is too short in Poland and poorly paid in Slovakia. Their parental and extended leaves are means-tested, and Slovakia and Hungary do not guarantee the full set of rights for workers returning from leave.7 In contrast, the Slovenian, Lithuanian and Latvian leave regulations carry a higher transformative potential. Their 1-year-long leaves are open to both parents and paid at a 100 percent income-replacement rate. States allow some flexibility in uptake and use the force of paternal quotas: Lithuania grants 1-month fully paid paternity leave, and Slovenia combines a 15-day leave at 100 percent pay with 75 practically unpaid days that last until the child's third birthday. Slovenian leave, however, is cast in terms of employment - compared to the Lithuanian, extended full-time familial childcare is not supported. The eight countries also vary in childcare services, most distinctly for children aged 0-3 years. Services are a legal responsibility of municipalities in all countries. In Slovenia and Lithuania, however, they are centrally regulated, coordinated and supervised. Their creches are available to children aged 1-3 years and are part of unitary childcare centres. Their legislation reflects a strong emphasis on high-quality services for different groups of parents across the state: they are prescribed on a full-time basis Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 Javorník 251 100 -i 90 -SO -70 -60 ■ 50 -40 -30 -20 io H o so ■■ 33 í 33 í 18.2 132 20.7 20.; 20.7 I I ~l , I I . I—L—I—_ 73 0 Poland Slovakia Czech Hungary Latvia Estonia Lithuania Slovenia Figure I. Degree of state de-familialism, 2008 (in percent of maximum achievable score). Source: Author calculations using MISSOC/MISSEC, Eurydice, OECD and administrative data (see Figure 2). around the year with some flexibility in provision and a centrally regulated sliding-fee scale. In the other six countries, public childcare for children aged 0-3 years is limited and varies across the state, with a strong urban/rural divide. This applies to services for older children as well, but Hungary and Poland stand out for prescribing free childcare without age limits. That notwithstanding, their regulations give providers permission to set access criteria and charge for any extracurricular activities, thus hindering access to publicly funded childcare. Finally, I have used spider charts to compare policies against Leitner's (2003) ideal types. Familialism appears particularly strong in Hungary, the Czech Republic and Estonia. There, states rely upon and support family childcare. Access to longer paid parental leave is a key feature, while public childcare services are limited, especially for children aged 0-3 years. Following Leitner (2003), these are characteristics of explicit familialism since the states equate the right to family childcare with the family's obligation to provide it. This is reflected in wide gaps in the employment rates for women with preschoolers as compared to those without, with a range of up to 50 percentage points in the Czech Republic (Javornik, 2010). In contrast, Poland, Slovakia and Latvia practically leave parents without any public support. This resembles implicit familialism. While such regulations do not explicitly promote traditional gender roles, the lack of public support implicitly puts the responsibility for childcare primarily within families. Latvia, however, stands out with a high score on parental leave. While its leave policy is closer to the optimal, the state does not provide public alternatives thereafter. Given a high reliance on private childcare and limited regulation in the consumer markets, access to high-quality childcare services is limited in all three countries, particularly for low-income families, and the employment of women with preschoolers is negatively associated with their education - given that education is related to one's potential wage (for example, Steiber and Haas, 2009: 646). Slovenia and Lithuania focus on women's continuous employment which is reflected in the policy logic of supported de-familialism. Parental leave generates incentives for mothers' continuous employment and a more active fatherhood following childbirth, and leave and childcare services are contiguous. In Slovenia, public responsibility for childcare has been a long-standing principle and parents' right to family time has received limited recognition: while it remains a cultural norm to spend a year with the newborn, parents have been incentivized to use public childcare thereafter (Javornik, 2012b, 2014). This is shown in their maternal employment rates, which are among the highest in the EU and the OECD: the gaps between different groups of women are narrow and mothers prevailingly work on a full-time basis (Javornik, 2010). Optional de-familialism remains as an ideal since none of the analysed countries equally distributes childcare responsibilities between the state and the family, and between the mother and the father, although the Lithuanian, Hungarian and Estonian policies come Downloaded from esp.sagepub.com at Masarykova Univerzita on January 3, 2015 252 Journal of European Social Policy 24(3) SUPPORTED DE-FAMILIALISM Slovenia LEftVE lime duality Affotdrtiiity Lithuania Jobsecurity AffordaMity LEAVE time 1\ Parental Opening ff entitlement r""1" Doddy-quota AdBliffí ■ Icxib lity EXPLICIT FAMILIALISM Czech R. Hungary LEAVE time Quality. jot>si«""ty Affoidarjiltv. legibility EXPLICIT FAMILIALISM Estonia Mfůrdability Job security IMPLICIT FAMILIALISM Poland AFIg ;:,il. I [y Parental enii|l((nent Opífliňghíuli Admission Afp Daddy-quoN Fk.-