Family policy in comparative perspective: The concepts and measurement of familization and defamilization

Abstract

Family policy is not easily conceptualized or measured in comparative research. Previous approaches are highly diverse and have yielded mixed empirical results in terms of placing countries’ family policy profiles in the international landscape and mapping their trajectories. This article reviews the long-standing discussion of the familization and defamilization concepts popular in comparative research, derives a conceptual framework and provides an in-depth discussion of current empirical approaches. It tackles the lacking consensus on how familization and defamilization are measured, arguing that interventions in gender-specific and intergenerational dependencies are the key dimensions and that measurement at policy level is best suited to capture within- and cross-country variation in family policy. Using data on 21 European countries, the article proposes measures that acknowledge the different dimensions of familization and defamilization. The proposed indicators prove to be useful for mapping a range of European countries’ family policy constellations but are bound by data restrictions. Therefore, the article makes a strong claim for improving the availability of internationally comparable family policy data.

Keywords

Defamilization familization family policy individualization measurement

Introduction

Family policy has gained importance internationally as a policy field since the mid-1990s (Mätzke and Ostner, 2010), and research is increasingly concerned with explaining changes, outcomes and future directions in family policy. Family policies have conceptually been integrated less systematically into mainstream welfare state theories than other traditional fields such as unemployment and pension policies. One difficulty seems to be that, within the same welfare state setting, family policy combines different ideas and aims. An illustrative example for such tendencies is recent family policy reforms in Germany, which included the commitment to boosting childcare provision to relieve family care responsibilities as well as the introduction of a childcare premium for stay-at-home mothers. This article argues that, in order to describe and comprehensively map countries’ family policy profiles and changes over time, a fresh analytical perspective and new tools for the measurement of family policies are required.

The concepts of familization and defamilization have often been used in comparative welfare state research to discuss how countries address the issues of financial and care dependencies between family members, and attempts have been made to measure the concepts used to describe the differences between countries. This article congruently argues that the familization and defamilization concepts can be useful analytical tools for comparing family policy across welfare states. The multi-dimensional framework that the concepts offer seems suitable for analysing a policy field with opposing policy aims, such as supporting mothers’ labour market participation and boosting fertility. However, definitions of familization and defamilization in previous studies vary widely, and only a few approaches combine rigorous conceptualization with sound operationalization. Compromises are either made on the conceptual side, narrowing de-/familization to sub-dimensions, or on the measurement side by omitting central aspects of the concepts in their translation into empirical indicators. This has led to some confusion and arbitrariness around the terms and their use in empirical studies. The aim of this article is hence threefold: first, to derive a conceptual framework based on the comprehensive review of theoretical literature on de-/familization; second, to compare previous empirical approaches towards the measurement of de-/familization; and third, to provide a suggestion for the operationalization of the concepts, which allows for conducting empirical analyses of the complexities and dynamics of policy intervention in family responsibilities.

The concepts of familization and defamilization

Since the initial formulation of different ‘worlds of welfare’, a growing body of research has compared welfare states by the extent to which they support the ‘familization’ and ‘defamilization’ of their citizens, roughly defined as the degree of support of individuals’ independence from family relationships. While defamilization is primarily the focus in this field of research, familization is often treated as a sub-dimension or the counter-pole to defamilization. The academic debate around the defamilization concept can be described in three main phases, which may be labelled the emergence, consolidation and critical assessment phases. In the first phase, the term was introduced in two articles published in the same year by McLaughlin and Glendinning (1994) and Lister (1994) as part of the long-standing discussion within feminist literature on the position of women in welfare states (see, for example, Hernes, 1987; Lewis, 1992; Lister, 1990; Nelson, 1990; O’Connor, 1993; Orloff, 1993).¹ In the second phase, the concept was taken up by mainstream welfare state theorists in the reformulation of their earlier analyses of social policy regimes, most prominently by Esping-Andersen (1999) and Korpi (2000), who did not explicitly refer to the term ‘defamilization’. The third ongoing phase is characterized both by a critical conceptual assessment of defamilization, in which different sub-dimensions of the initial concept are highlighted, and by empirical application. The remainder of this section discusses the main aspects of each of the three phases in more detail.

Emergence phase

A focal point in the primarily feminist discussion that has led to the emergence of the concept of defamilization is the critical examination of Esping-Andersen’s (1990) analysis of welfare states and its central concept of decommodification. For example, Lister (1990), Lewis (1992), Orloff (1993) and O’Connor (1993) have argued, respectively, that gender-specific aspects such as dependencies between spouses ought to be included in the analysis of welfare states. With reference to the approach of Esping-Andersen and other welfare state theorists, Orloff (1993) writes, ‘Its concepts are explicitly gender-neutral – but the categories of workers, state–market relations, stratification, citizenship and decommodification are based on a male standard; moreover, gender relations and their effects are ignored’ (p. 307). As to the gendered implications of the decommodification concept, she argues that most women are not (fully) commodified and thus do not fulfil the precondition to benefit directly from a high level of decommodification through welfare states. While Esping-Andersen (1990) stresses the equalizing impact of decommodification, he ignores its impact on gender inequalities. Because women are predominantly the ones held responsible for care work in their families, they often face reconciliation issues with work in the labour market. As a consequence of their limited access to labour market income, women are often economically dependent on a male earner (e.g. Lister, 1990). Furthermore, non-standard labour market involvement can restrain eligibility to welfare benefits since these are often linked to continuous employment histories. The aspect of economic dependence is stressed in Lister’s (1994) definition of defamilization:

Arguably, the dimension of decommodification also needs to be complemented by what we might call ‘defamilization’, if it is to provide a rounded measure of economic independence. Welfare regimes might then also be characterized according to the degree to which individual adults can uphold a socially acceptable standard of living, independently of family relationships, either through paid work or through the social security system. (p. 37)

McLaughlin and Glendinning (1994) propose a more general concept of defamilization which does not as explicitly stress the aim of economic independence: ‘[D]e-familization is constituted by those provisions and practices which vary the extent to which well-being is dependent on “our” relation to the (patriarchal) family’ (p. 65). They furthermore stress that defamilization is not only about gendered but also about intergenerational dependencies:

The issue is not whether people are completely ‘de-familized’ but rather the extent to which packages of legal and social provisions have altered the balance of power between men and women, between dependents and non-dependents, and hence the terms and conditions under which people engage in familial or caring arrangements. (McLaughlin and Glendinning, 1994: 66)

The idea that care must be regarded as a reciprocal relationship, which is stressed again in more recent work on defamilization, is already spelled out in this conceptualization. Hence, at the time of its emergence, the concept of defamilization captures not only economic independence and the independence from care responsibility but also the freedom to choose who cares.

Consolidation phase

As argued above, the second phase is characterized by an integration of the concept of defamilization into the mainstream welfare state theory. However, looking at Esping-Andersen’s (1999)² definition of defamilization more closely reveals that only a partial integration of the concept was achieved:

The concept of de-familization parallels the concept of de-commodification; in fact for women de-familization is generally a precondition for their capacity to ‘commodify themselves’ (Orloff, 1993). Hence, de-familization would indicate the degree to which social policy (or perhaps markets) render women autonomous to become ‘commodified’, or to set up independent households, in the first place. (p. 51)

In this definition, enabling commodification (of women) appears to be the main, if not the only, aim of defamilization.³ Although not making use of the defamilization concept explicitly, Korpi (2000) takes up the ideas it conveys in his characterization of different family support models of the welfare state defined along the dimensions of general family support and dual-earner support. Korpi refers to the wider notion of gender inequalities in agency (and not exclusively to commodification), but his main focus is still on policies affecting the gendered patterns of labour market inclusion. Consequently, Korpi’s (2000) typology is based on ‘broadly conceived policy institutions likely to be of major relevance for gendered agency inequality in terms of labor force participation’ (p. 143). Hence, while integrating the ideas from feminist researchers into the core discussion of welfare states, both Korpi and Esping-Andersen focus on the gender aspect of labour market integration.

Critical assessment phase

In more recent years, which may be regarded as a third phase in the discussion of defamilization, various authors have criticised this narrowing of the perspective on defamilization to questions of employment. Two types of criticism can be distinguished in this phase. One perspective takes current policy changes and reform efforts as its starting point for scrutinizing the different facets of defamilization, and the other calls for a reconsideration of the defamilization concept from a more theoretical point of view. Both strands, however, reach similar conclusions, condemning that defamilization has become limited to the aspect of commodification (see Kröger, 2011; Leitner et al., 2004; Leitner and Lessenich, 2007; Lewis, 2001; Lewis and Giullari, 2005; Ostner, 2003).

Several authors have used the notions of de-/familization in their critical discussions of the policy reforms which have been taking place in many European countries since the 1990s. The centrepiece of the reforms was strategies for boosting female labour market participation. The overarching aim, as is also spelt out in the European Employment Strategy, is to strengthen the competitiveness of the European Union in a global perspective. Where such policy reforms are assessed against the categories implicit to the defamilization concept, it becomes apparent that the policies mainly address economic but not care dependencies (e.g. Leitner et al., 2004; Lewis, 2001; Lewis and Giullari, 2005). Conflicting reform aims and outcomes arise where support for the commodification of women does not include support for easing care responsibilities. Employment-oriented reforms have often implied that the social rights previously granted to women based on their status of dependency on a male earner were cut (e.g. financial aid for widows) without resolving care dependencies (Leitner et al., 2004). An underpinning issue is that, although increasing female employment is arguably inseparable from it, eradicating gender inequality is not usually considered to be an equal aim in European policy reform (Lewis and Giullari, 2005: 83). These aspects illustrate that policy strategies described as defamilizing can entail politically controversial ideas and conflicting outcomes for individuals.

The second strand of critical discussion in this ‘third phase’ of conceptual development is more concerned with the usefulness of the defamilization concept for comparing and typologizing welfare states. Kröger (2011), for example, claims that defamilization causes ‘conceptual confusion’ (p. 429) mainly because it mashes up distinct dimensions of economic and social/emotional independence. Unlike Leitner and Lessenich (2007), who define sub-dimensions of the defamilization concept in order to include social/emotional aspects, Kröger suggests reserving the term ‘defamilization’ for the economic dimension and to use the concept of ‘dedomestication’ for the social/emotional component. By contrast, other authors focus entirely on the economic dimension. Bambra’s (2004) approach stresses the origin of the defamilization concept in the feminist research tradition and defines it in terms of ‘factors concerned with female freedom from the family, rather than the freedom of the family’ (p. 204). Saxonberg (2013), too, highlights the concept’s roots in analysing gender equality but criticises the concept’s vagueness and proposes replacing it with the ‘degenderization’ concept. Daly (2011a, 2011b), on the other hand, argues that familization is a useful concept to describe policy effects, but that the defamilization concept is misleading. Arguing that ‘“familization” is the original orientation of family policy’ (Daly, 2011a: 88), she suggests that defamilization is inappropriately based on a static concept of the family and that it frames shifts in policy in negative terms with the ‘de’ prefix (Daly, 2011b: 7). Daly’s preferred contrasting notion to familization is ‘individualization’, which conveys policies’ focus on the individual rather than a collective, as is implied in the term ‘family’. Alongside the approaches suggesting alternative concepts to defamilization, other research has further developed the sub-dimensions included in the initial approach of McLaughlin and Glendinning. This has an advantage over newer suggestions which trade in the multi-dimensionality of the initial concept in attempting improved clarity. Leitner and Lessenich’s (2007) approach, for example, highlights the need to specify the roles of care giver and receiver within the conceptual framework for defamilization. Saraceno and Keck (2010) have also used the defamilization concept for describing state provision of individualized social rights that reduce family responsibility and dependence. However, they assume that the opposite pole to defamilization, ‘familialism’, can either be created by the lack of state provision of services and financial support (‘familialism by default’) or by active state support to individuals’ family care (‘supported familialism’). This categorization comes close to the one proposed by Leitner (2003), who applies the two-dimensional approach of familization and defamilization in describing state support and/or lack of care provision and financial incentives for family care. Both the studies by Leitner (2003) and Saraceno and Keck (2010) make the particular contribution of considering intergenerational dependencies, the rarely regarded dimension captured with the defamilization concept, alongside gender dependencies.

Patterns of familizing and defamilizing policies

As discussed in the previous section, the almost 20 years of theoretical debate around familization and defamilization entail different perspectives on the meanings of the concepts. Carving out common denominators in the literature, we distinguish the following sub-dimensions: (a) gender and/or generation, (b) reciprocity of care and (c) statutory policy and/or state and market provision. In general, we differentiate between the policy level and the level of expected outcomes (see also Leitner, 2003; Saraceno and Keck, 2010). At the policy level, we regard familizing and defamilizing policies which differ from country to country. These policy configurations are expected to produce ideal-typical patterns of outcomes with regard to gender-specific and intergenerational dependencies such as familialism and defamilialism or individualism, which will differ accordingly at the country level. We share Daly’s (2011b) preference for using the concept of individualism rather than that of defamilialism (see discussion below). We acknowledge that patterns of familialism and individualism result not only from the differences in policies but are also expressions of attitudinal and cultural differences (see, for example, Pfau-Effinger, 2004). Although we do not focus on such differences in the empirical part of this article, we suggest that, analytically, attitudinal and cultural differences should be considered alongside familizing and defamilizing policies for categorizing countries. Hence, in the next two sections, we present our definition of familizing and defamilizing policies before we discuss the expected patterns of outcomes. We assume that single policies can be characterized by their impact on familialism and individualism, but that, for the description of countries, the combination of the different policies has to be considered. We argue that policy-level analysis is central in the field of family policy for two reasons. First, only by taking into account single policies, the contradictions between and opposing directions of policies may be identified. Second, policy change and potential divergences from ‘path dependencies’ can easily go unnoticed by looking at the country level and may be addressed more readily with a policy-level perspective.

Defamilizing policies

We define defamilizing policies as welfare state provisions (social policies and regulations) that reduce care and financial responsibilities and dependencies between family members. It is hence a concept for describing policies in terms of their expected outcomes. We follow the approaches from above that define defamilization as capturing policy outcomes in terms of gender-specific as well as intergenerational relationships; describing the extent to which policies relieve family members from their gender- and generation-specific dependency roles. This definition goes beyond other approaches, such as Bambra’s (2004) reading of the defamilization concept and Saxonberg’s (2013) alternative concept of degenderization, which omit the intergenerational aspect. Defamilizing policies can, in principle, be provided by the state, the voluntary sector and ‘the market’ (i.e. commercial providers). In our definition, we focus on regulations defined by statutory bodies because we assume that these cover a substantial part of family policy provision and serve as a good basis for international comparisons.

In the present understanding, free public childcare is a prime example of a policy with a clear defamilizing effect because it reduces family dependencies for parents to care for their children and for children to rely on their parents as carers.⁴ We also consider parental leave policies as defamilizing as they may reduce family dependencies by allowing employed parents of young children to maintain high labour market attachment, guaranteeing the return to their jobs after a given period of full-time family care. This stands in contrast to Saraceno and Keck (2010) who stress that parental leave supports certain family members in keeping up their care responsibilities and categorize it as ‘supported familialism’. We argue that the defamilizing effect depends on the design of parental leave policies. Empirical research has found that long unpaid leave, in contrast to short periods of leave paid at high replacement rates, incentivize parents (mothers) to stay at home to care for their child (e.g. Akgunduz and Plantenga, 2013; Han et al., 2009). Therefore, we regard only policies granting rather short, well-paid absence from employment as defamilizing and other leave policies as familizing. Similar to parental leave policies, regulations that support employed individuals in taking periods of leave to care for their frail or sick relatives can be considered defamilizing. A further example of a defamilizing policy is free public care for older people because it implies that care responsibility does not remain with family members. In these cases of child and elderly care, the meaning of defamilizing policies resembles what Kröger (2011) defines as ‘dedomestication’ by social policies. Some scholars have noted the theoretical possibility of defamilization via the market in addition to statutory policies (see Esping-Andersen, 1999), for example, in the form of market-based care provision for children and older people. In fact, in most countries, care services are not exclusively provided by the public sector but also by non-profit and for-profit providers, which is sometimes referred to as ‘welfare mix’ (Bode, 2003). However, it is difficult to define the conceptual boundaries of market-based defamilizing policies, such as childcare provision, not least because often ‘markets’ provide state-subsidized services and contain third sector providers. Hence, in contrast to other studies (e.g. Leitner, 2003), we exclude market-based provision from our definition of defamilizing policies to avoid such conceptual difficulties.⁵ Instead, we argue that looking at publicly regulated care provision provides a good comparative picture of different welfare state strategies, that is, representing systems with different degrees of statutory involvement. This perspective also covers services organized on the basis of the subsidiarity principle, for example, in Germany (Bode, 2003), to the extent that they are regulated and funded by public bodies. If welfare states leave it up to individuals to arrange their family care, either privately or in the market, this means a lack of defamilizing policies. This is in line with the idea that some welfare states are more ‘hands off’ or ‘liberal’ than others in certain policy areas.

Familizing policies

Familizing policies are defined as social policies or regulations that foster dependencies among family members by actively lowering their negative social and economic consequences (also see Leitner, 2003; Leitner and Lessenich, 2007, Saraceno and Keck, 2010), such as women’s financial dependence on a breadwinner, children’s dependence on their parents’ care and elderly people’s dependence on their adult children. It is important to acknowledge that we do not understand defamilization and familization as polar opposites on a linear continuum. Many welfare states feature defamilizing and familizing policies at the same time (e.g. public childcare and generous family benefits). Likewise, countries with few defamilizing policies are not necessarily rich in familizing policies. Familizing policies comprise all cash payments that support family care at home. Among these are child benefits and family allowances, as well as home care allowances that are paid conditionally on parents’ (mothers’) home care for a dependent child, and payments that are conditional on home care for older people. Tax rebates granted for family members are considered to have a similar effect. Both sets of policies reduce the negative economic consequences associated with family dependency. Parental leave regulations are considered familizing when they are granted with no wage replacement or payment at a low level and for long periods. The familizing effect here is that the negative social consequences of family dependencies are reduced with the legal status supporting the acceptability of mothers staying at home for care (see discussion above). A further factor categorized as familizing policies are legal regulations for financial obligations between parents and their children. Where such regulations imply that parents’ financial obligation towards their children lasts for the entire dependency years (e.g. as child-related benefits) or beyond, strong familization can be assumed. Likewise, the statutory regulation of adult children’s obligations to care for their elderly parents (financially) strengthens dependencies among them.

Ideal-typical outcomes of policy configurations

Based on the definitions of defamilizing and familizing policies, we can derive expectations about ideal-typical patterns of policy outcomes for how individual family members tend to structure their family dependencies.⁶ Figure 1 illustrates the conceptual range of outcomes from constellations of defamilizing and familizing policies in a four-field matrix. In essence, similar to other approaches, we assume that combinations of strong/weak defamilizing policies with strong/weak familizing policies in welfare states produce typical patterns of family relationships.

Figure 1.

Conceptual dimensions for describing welfare state intervention in family responsibilities and dependencies.

Our proposal draws on existing typologies on familization and defamilization. Suggesting a similar matrix, Leitner (2003) categorizes four ideal types of outcomes as (a) defamilialism, (b) implicit familialism, (c) explicit familialism and (d) optional familialism based on the policy constellations of weak/strong defamilization and weak/strong familization. Saraceno and Keck (2010) and Keck and Saraceno (2012), on the other hand, distinguish three types: (a) defamilization, (b) familialism by default and (c) supported familialism as typical constellations. Across approaches, the category of familialism seems relatively undisputed. Likewise, defamilialism appears in both studies. Our model, however, makes important additions. First, we replace ‘defamilialism’ with ‘individualism’ following Daly’s (2011b) suggestion. In this reading, reducing gender-related and intergenerational dependencies through statutory support to individuals fosters individualism, but is not equal to a ‘family-hostile’ environment. Second, our approach goes further than previous ones in that it assumes two respective possible outcomes for individuals’ organization of family life where both familization and defamilization are weak or both are strong. Accordingly, the combination of strong familization and strong defamilization, labelled as ‘optional’ familialism by Leitner (2003), should at the same time allow for ‘optional individualism’. In this constellation of statutory family policies, both outcomes are possible: stronger reliance on family members on the one hand or weaker dependencies on the other hand. At the opposite pole, the provision of both weak familizing policies and weak defamilizing policies would create a pattern of implicit individualism alongside ‘implicit familialism’ (Leitner, 2003) or what is called ‘familialism by default’ by Saraceno and Keck (2010). Here, and in the case of optional familialism or individualism, the distinguishing criterion is not the degree of familizing and defamilizing policies but other factors such as attitudinal and cultural differences across welfare states (Bahle, 2008; Kangas and Rostgaard, 2007; Pfau-Effinger, 2004). Labelled as ‘cultural legacies’, the diagonal line in the graph illustrates that the organization of family dependencies at the household level is influenced by normative factors that lie beyond the mere effects of policies. This means that the direction in which countries pull on this dimension (more towards optional individualism or more towards optional familialism) is to a certain extent defined by predominant family ideals in terms of gender and intergenerational relations. Countries where no clear pattern of statutory familizing and defamilizing family policies prevails can differ in terms of the extent to which individuals and families fall back to familialist solutions for finances and care or to ‘individualist’ solutions, for example, in outsourcing family dependencies to market providers. However, it should also be noted that ‘cultural legacies’ do not constitute a factor that is independent from the design of family policies. Certainly, culture and policies are interrelated in a much more complex way than may be acknowledged in a simplifying graph like Figure 1. The aim of the presented model is to provide one possible framework for interpreting the differences in countries’ family policy constellations and for formulating expectations about differences in outcomes.

An evaluation of previous measures

Having developed a conceptual understanding of familization and defamilization at the policy level and of possible country-level outcomes, we now turn to evaluating previous empirical approaches. This section evaluates the studies’ choices of indicators against a set of criteria that define the desirable features of comparative indicators (for an earlier overview, see Lohmann 2009). The criteria are the measurement of institutional factors, the availability for many countries and repeated measurement. The following evaluation will demonstrate the variety of indicators measuring similar constructs and the problems associated with them. This is based on the assumption that the comparison of familizing and defamilizing policies across countries requires rigid operationalization of the concepts, which should ideally meet the following criteria. First, although it is usually more difficult to gather information on institutional factors (e.g. policies that support mother’s employment) than information on the assumed outcome of such factors (e.g. the female employment rate), the former are preferable because they do not require that a high degree of covariation is assumed between policies and outcomes. This assumption is not only critical because it may be wrong, but first and foremost because in many cases of international comparison the assumption equals the hypothesis to be tested (see, for example, Bambra, 2004). For instance, the question of whether defamilizing policies have an impact on women’s labour market participation cannot be answered using the female employment rate as a proxy for institutional factors because it is part of the concept to be explained. Furthermore, using outcomes as measures for institutional factors restricts the options for testing alternative hypotheses like the influence of attitudes towards female employment. Second, because country-level measures are often compiled for use in macro- and multi-level regression models to which the small-n-problem is inherent (Ebbinghaus, 2005), availability for many countries is a desirable characteristic of comparative indicators (usually at least all European Union (EU)–Organisation for Economic Co-operation and Development (OECD) and non-EU–OECD countries). Third, stability and change can only be captured if measures are surveyed on a regular basis (see, for example, Gauthier, 2002). Even if the defined criteria are ideal requirements that will be difficult to satisfy fully for any measure, they provide a framework against which the quality and the problems of the different indicators used to measure familizing and defamilizing policies can be evaluated. Not least, the criteria set a standard for the development of new measures of the concepts as well as for new attempts of compiling family policy data. In the following subsections, the choice of indicators in previous studies is discussed (section ‘Previously used indicators’) and their empirical approaches are assessed (section ‘Comparison of indices’). Studies considered in the evaluation (a) contain some theoretical discussion of familization and/or defamilization or closely related concepts, (b) suggest an operationalization and (c) use the measure(s) in the analysis of a larger set of countries.

Previously used indicators

Despite the long-standing discussion of the familization and defamilization concepts (see section ‘The concepts of familization and defamilization’), there is still only one proposal for a comprehensive ‘defamilization index’ (Bambra, 2004). A number of researchers have, however, proposed indices for related concepts which we also include in our discussion. Other approaches have opted to analyse single indicators to illustrate country differences in the sub-dimensions of familization and defamilization. Many of the studies considered here use indicators that come closer to the idea of ‘defamilization’ than to that of ‘familization’. However, some attempt is usually made to distinguish between different dimensions of de-/familization through welfare state activity. Table 1 provides an overview of the indicators used in the discussed studies. There are indicators in four different categories. As indicated, several studies provide measures of two different dimensions. While Esping-Andersen (1999) explicitly speaks of defamilizing and familizing factors, Korpi (2000) distinguishes the dimensions of dual-earner and general family support, combining different aspects of leave policies, provisions of childcare, home help for the elderly, child allowances and family tax benefits. The ‘dual-earner support’ dimension resembles the concept of defamilizing policies, while the ‘general family support’ one is closer to the idea of familizing policies.⁷ A similar operationalization is used by Ferrarini (2006) who refers to Korpi’s approach. Korpi et al. (2013) also draw on the initial idea of dual-earner/traditional family support, differentiating between ‘traditional family’, ‘market-oriented’ and ‘earner–carer’ models of welfare support, and also considering policies that support (temporary) ‘dual-carer’ arrangements (paternal care). The ‘traditional family’ dimension largely matches an idea of familizing policies, while the ‘earner–carer’ dimension adheres to the concept of defamilizing policies. Following our approach, Korpi et al.’s ‘market-oriented’ dimension could imply defamilizing or familizing policies. Gornick and Meyers (2003) provide several indices on policy measures (early childhood education and care, schooling schedules, family leave, working time regulation) that support an ‘earner–carer model’ in which parents will be able to equally share economic and care responsibilities.⁸ While this conceptualization comes close to the outcome that strong defamilizing policies would have, the contrasting male breadwinner–female carer model resembles one with dominating familizing policies. By contrast, using only measures of defamilization, Bambra’s (2004) approach is one-dimensional. Similarly, Mandel and Shalev (2009) draw on the defamilization concept, measured as both the enrolment rates of children in public childcare and weeks of maternity leave granted, which could, however, be said to convey aspects of both familizing and defamilizing policies. Only a few of the studies considered make the distinction between familization and defamilization explicit, constructing indicators for each dimension. Looking at the policies directed at the caring function of the family, Leitner (2003) defines two indicators for familization (paid parental leave and cash transfers to older people) and two for defamilization (childcare coverage and institutional care for older people). Saraceno and Keck (2010) use child allowances, child-related tax allowances and legal age thresholds ending parents’ financial obligations as measures of ‘familialism by default’. Effective maternity and parental leave in weeks and level of compensation were used as indicators of ‘supported familialism’ in their study and childcare coverage rates as a measure of ‘defamilization’. In addition, Saraceno and Keck use a range of indicators to consider responsibilities towards older family members: service provision for elder care (provided by the state or the market) for distinguishing defamilization and supporting familialism, and conditionality of financial support provided to the elderly and level of minimum pension for measuring defamilization versus familialism by default. Focusing on care dependencies, Kröger (2011) uses the time coverage of care services as hours per week, the legal right to childcare, the proportion of net costs for childcare in family income and the child–staff ratios for constructing a ‘dedomestication’ indicator. Given the wide variety of theoretical approaches, it is still noticeable how many similarities there are between the indicators used. The next step will show which are the most widely used indicators and how their country ranking outcomes compare across the studies.

Table 1.

Overview of de-/familization indicators used in previous studies.

	Esping-Andersen	Gornick et al.	Korpi	Bambra	Gornick and Meyers	Leitner	Ferrarini	Mandel and Shalev	Korpi et al.	Saraceno and Keck	Kröger	Frequency (total)
1. Maternity and parental leave policies
(Paid) maternity/ parental leave		D	D	D	D	F	D	D	D	D		9
(Paid) paternity leave		D	D		D		D	D	D	D		7
2. Care services
Guaranteed childcare coverage		D			D				D	D	D	5
Childcare coverage rate	D	D	D/F		D	D		D	D	D	D	9
(Low) cost of childcare	D				D						D	3
Pre-primary/primary school regulations		D			D							2
Home care coverage	D									D		2
Average hours of attendance in childcare											D	1
Childcare quality (child–staff-ratio)					D						D	2
Share of people 65+ years receiving home care						D				D		2
Share of people 65+ years in residential care										D		1
3. Generosity of family benefits (incl. tax benefits)
Generosity of child/family allowances	F		F			F	F		F	F		6
Householdmeans-testing (unemployment benefits)	F											1
Family/child/partner tax benefits	F	D	F		D		F		F			6
Payment of child allowance to mother/father										F		1
Legal age limit: financial obligation parent → child										D		1
Financial obligations child → parent										D		1
Minimum pension										D		1
4. Other indicators	4*			2*

For sources, see text; Bambra: D = defamilization; Esping-Andersen: D = defamilization, F = familization; Leitner: D = defamilization, F = familization; Ferrarini: D = dual-earner support, F = general family support; Gornick et al. : D = earner-carer model; Gornick and Meyers: D = earner-carer model; Korpi: D = dual-earner support, F = general family support; Korpi et al.: D = dual-earner/dual-carer support, F = traditional family support; Kröger: D = dedomestication; Mandel and Shalev: D = defamilization; Saraceno and Keck: D = defamilization, F = familialism.

Esping-Andersen considers the generosity of family expenditures as a defamilization indicator and indicators on family life and the division of labour as familization measures (share of elderly living with their children, share of young unemployed living with their parents, extent of female unpaid work); Bambra uses the female employment rate and the gender pay-gap as defamilization indicators.

Comparison of indices

Table 1 shows which indicators are used in the 11 studies considered. Information on maternity and parental leave regulations and on childcare coverage is used in more than half of the studies. Where used, indicators for the provision of childcare are assumed to relieve mothers from caring responsibilities. Assumptions around the impact of maternity/parental leave policies, on the other hand, are not unanimous. For example, while Gornick and Meyers (2003: 101) as well as Korpi (2000) interpret leave policies as supporting (female) labour market participation, Leitner (2003: 359f) argues that parental leave policies support the caring function of the family.⁹

As shown in Table 1, the bulk of indicators in the considered studies measures institutional factors rather than outcomes. Drawing on institutional indicators seems to have become common practice, and the outcome indicators describing family life and female employment used by Esping-Andersen (1999), on one hand, and Bambra (2004), on the other hand, have not been applied in any other study considered here. Only when it comes to measuring care provision for children and older people, given data limitations, can comparative scholars more often seem to revert to using outcome measures. Indicators which cover aspects of intergenerational dependencies can be found in studies by Esping-Andersen (1999), Leitner (2003) and Saraceno and Keck (2010). Esping-Andersen reports information on home care coverage and the share of elderly and young unemployed individuals who live with their families. The latter, however, are outcome indicators. In addition to the regulation of payment for elderly care, Leitner uses the percentage of home help services as a measure, which is also an outcome indicator. Saraceno and Keck use information on minimum pensions and cash for care payments as well as on the share of people over 65 years in residential and home-based care, the latter of which would also count as an outcome indicator. Finally, Kröger (2011) uses childcare take-up rates, average hours of attendance and child–staff ratios (by observation rather than by regulation) to describe state support for care dependencies. All of these indicators can be considered outcome measures.¹⁰

The comparison shows that there is a core group of seven countries which is covered in all of the studies considered: Belgium, Denmark, Finland, France, the Netherlands, Sweden and the United Kingdom. Eight further countries, Australia, Austria, Canada, Germany, Ireland, Italy, Norway and the United States, are included in nine of the studies.¹¹ It is likely that the limited availability of and access to data are severe obstacles to broadening the country sample. Issues of data availability will be considered in more detail in section ‘Discussion’.

In the discussion of the theoretical and methodological foundations of existing approaches to measure defamilization, familization and related concepts, we have pointed out differences besides some commonalities. A crucial question is how such differences are reflected at the level of substantial results, namely, if countries are ranked similarly despite using different measures. Given the wide use of country rankings or scores in league tables, or as input for macro or multi-level analyses on the outcomes of family policies across countries, evidence on the robustness of rankings using different approaches seems crucial. In our comparison, we focus on measures of defamilization and related concepts such as dual-earner support or dedomestication, because only a few studies provide measures of familization for a large sample of countries. Table 2 shows the country rankings of indicators from eight different studies (Bambra, 2004; Ferrarini, 2006; Gornick et al., 1997; Gornick and Meyers, 2003; Korpi, 2000; Korpi et al., 2013; Kröger, 2011; Mandel and Shalev, 2009). Some of the studies discussed above have been left out as they do not cover crucial countries or do not provide index values which can be brought into a rank order. In total, we used information on 13 countries (with the exception of two studies which only cover 11 of these countries). The table provides rank orders for all studies, information on the mean rank and the standard deviation (SD) of the rank across all studies. Looking at the mean rank and the SD, we can see that especially at the top or the bottom of the overall ranking countries are placed rather similarly in most studies. Sweden, Denmark or Finland is ranked at the top in most of the studies, while at the bottom we find either the United States or Australia. The similarity of country ranks across studies is reflected in a rather low SD. It is particularly low in the countries that we find consistently at the top or the bottom, but it is also rather low for Germany and the Netherlands. However, looking in more detail at where countries are ranked, there are even differences between the similar scores. For example, Kröger (2011) ranks the United States in ninth place, while it ranks 12 or 13 in all other studies. For children aged 3–5 years, Gornick et al. (1997) rank Sweden in fourth place, which contrasts with most other studies where it is ranked 1 or 2. This could reflect the different theoretical perspectives, but nonetheless is surprising given the general focus on policies for relieving individuals from family responsibilities. The most inconsistent overall results were found for Norway (SD = 2.9). According to Gornick et al. (1997), in particular with regard to policies that support the employment of mothers with children aged 3 years and above, Norway is among the least supportive countries (rank 11). In contrast, it performs second best in Bambra’s (2004) defamilization index and Ferrarini’s (2006) index. We cannot rule out that the different data sources or years of data collection explain these deviating results, but it seems more likely that differences in the basic conceptualization and, moreover, in the choice of indicators explain the variation across studies. The previously discussed Table 1 shows that Ferrarini’s indicator only reflects leave policies, while Bambra adds outcome indicators of female employment and gender pay gaps. In contrast, Gornick et al. include numerous indicators of childcare policies.

Table 2.

Country rankings using different family policy/defamilization indexes.

	G3	G35	G5	KO	GM5	GM6+	GM	BA	FE	MS	KR	KO+	Mean rank	Rank SD
SE	3	4	3	1	2	2	2	1	1	2	2	1	2.0	1.0
DK	2	3	2	2	1	1	1	3	5	1	1	2	2.0	1.2
FI	1	5	4	3	5	5	5	4	3	3	8	4	4.2	1.7
NO	6	11	10	4	4	4	4	2	2	4	7	3	5.1	2.9
FR	5	1	1	5	7	8	7	5	9	6	6	5	5.4	2.4
BE	4	6	5	6	6	9	6	8	11	5	5	6	6.4	2.0
IT	8	2	6	8	–	–	–	9	6	8	–	8	6.9	2.2
NL	10	9	9	9	8	7	8	6	7	7	4	7	7.6	1.6
DE	7	7	7	7	9	6	9	7	8	9	–	11	7.9	1.4
CA	9	8	8	11	12	12	12	10	4	10	12	12	10.0	2.4
GB	11	10	11	10	11	11	11	11	10	11	11	9	10.6	0.7
AU	12	13	12	12	–	–	–	12	12	13	13	10	12.1	0.9
US	13	12	13	13	13	13	13	13	13	12	9	13	12.5	1.2
Correlation of index ranking with mean ranking (over all indexes):
	0.92	0.72	0.84	0.98	0.98	0.93	0.98	0.94	0.72	0.98	0.83	0.92

SD: standard deviation.

G3: Gornick et al. (1997: 61) (children <3 years); G6: Gornick et al. (1997: 61) (children <6 years); G35: Gornick et al. (1997: 61) (children 3–5 years); KO: Korpi (2000: 33) (dual-earner support); GM5: Gornick and Meyers (2003: 320) (children <6 years); GM6+: Gornick and Meyers (2003: 320) (children ⩾6 years); GM: Gornick and Meyers (2003: 320) (all children); BA: Bambra (2004: 207); FE: Ferrarini (2006); MS: Mandel and Shalev (2009: 1885); KR: Kröger (2011: 432); KO+: Korpi et al. (2013; n = 13) (n = 11 in Gornick and Meyers, 2003 and Kröger, 2011, rank multiplied by 13/11).

Aside from the different country rankings, Table 2 provides information on the association of rankings in single studies with the mean ranking across all studies. Given the large differences in conceptual approaches and indicator choices and the wide variation in the ranking of single countries, the rather high correlation of most studies may come as a surprise. The majority of rankings are highly correlated with the overall mean ranking (r > 0.9). In only two cases (Gornick and Meyers, 2003 (3–5-year-olds); Ferrarini, 2006), correlation coefficients near 0.7 indicate a lower agreement with the rankings in other studies. The correlations should, however, not detract from the large differences observed between the studies’ country rankings for single countries, which point to crucial divergences between the conceptual and empirical approaches of comparing family policy constellations. Not least, this poses a major limitation on the systematic evaluation of developments in family policy configurations over time.

Construction of indices of defamilization and familization

After having evaluated previous studies’ approaches of measuring defamilization and familization across countries, we consider how the dimensions outlined in section ‘Patterns of familizing and defamilizing policies’ can be measured with the data currently available for a larger set of countries. We consulted several available data sources of international family policy indicators which provide indicators over several years, such as the Comparative Family Policy Database (Gauthier, 2011), the Multilinks Database (Keck and Saraceno, 2012), complementary databases to the Luxembourg Income Study (LIS) assembled on the LIS website,¹² the OECD Family Database (OECD, 2012a) and the OECD Social Expenditures Database (OECD, 2012b).¹³ Annual indicators with which some of the conceptual dimensions could be operationalized for a large number of countries and a relatively long period of time (starting in the 1980s–1990s) can be derived from OECD sources on one hand and the Comparative Family Policy Database on the other hand. The OECD Social Expenditure Database contains longitudinal information on spending for family services as a percentage of the country’s gross domestic product (GDP) from 1980, which may be used as an indicator for defamilizing policies. The Comparative Family Database includes an indicator for the level of child allowances paid to families covering a longer time span, which is suitable as a measure of familizing policy. Using these single indicators can, however, only give a very limited, one-dimensional view on countries’ patterns of familizing and defamilizing policies. We also found problematic irregularities in the OECD spending data series in the 1990s, which point to changes in the coding or underlying concepts. In conclusion, we found that the policy dimensions of defamilization and familization can currently be captured most adequately with indicators from the Multilinks Database (Keck and Saraceno, 2012). On the downside, this database is limited to 2 years (2004 and 2009), which led us to focus on constructing comprehensive, but cross-sectional indices. Although we had to relax some of our standards for the data, we are able to cover the theoretically crucial dimensions of intergenerational and gender dependencies between family members, extending the scope of previous empirical studies. Multilinks provides indicators on family policies and legal regulations, covering 27 EU member states (as of 2011) as well as Georgia, Norway and Russia. We restricted our selection to 2004 indicators to maximize the country sample because some of the indicators were not available for both years, which resulted in the range of measures listed in Table 3. We suggest nine policy categories for defamilizing and six for familizing policies.

Table 3.

Indicators used for de-/familization indices.

Defamilizing policies	Familizing policies
Early years care	Allowances and taxes
(1) Individual entitlement to childcare for children under 3 years (yes/no), (2) coverage rate for children under 3 years (share of children aged 0–2 years), (3) full-time childcare usage of children under 3 years (share of children aged 0–2 years), (4) coverage rate for children aged 3–5 years (share of children aged 3–5 years)	(1) Eligibility condition for child allowance: Universal benefit (yes/no), (2) child allowance for one child (share of net average income), (3) child allowance for three children (share of net average income), (4) tax deduction or tax credit for children (yes/no)
Elderly care	Parental leave
(1) Access to care services (yes/no), (2) home-based care service recipients (share of population aged 65 and older), (3) persons living in care institutions (share of population aged 65 years and older)	(1) Length of unpaid leave (months; constructed as the sum of maternity leave and parental leave minus the months of paid leave)
Parental leave	Support of the elderly
(1) Length of well-paid leave (months), (2) duration of paternity leave (days)	(1) Children’s legal obligation to support parents (yes/no)

These policy indicators were then used for the composite defamilization and familization indices. To standardize the indicator scales, we divided values by the observed maximum (Gornick and Meyers, 2003)¹⁴ and transformed dichotomous variables as 0 and 1 scores. We then added the single scores together to composite indices for defamilizing and familizing policies, respectively. Our final sample consists of the following 21 countries: Austria, Belgium, Czech Republic, Estonia, Finland, France, Germany, Great Britain, Greece, Hungary, Ireland, Italy, Latvia, Luxembourg, the Netherlands, Norway, Poland, Portugal, Slovenia, Spain and Sweden. In Figure 2, the countries are plotted on the dimensions of familizing and defamilizing policies.¹⁵ We also added the labels of the ideal-typical outcomes of family policy configurations illustrated in Figure 1. These are meant to provide a general impression of the orientation of the countries, not to define their exact location. We can see that in 2004, Sweden, Finland, Norway and Belgium, together with Estonia, come closest to a policy configuration supporting individualism (see also Daly, 2011b). There are no countries in our sample with strong policy support for both individualism and familialism, that is, leaving the option for either of the models. However, there are a few countries like France and Germany which, to some extent, combine familizing and defamilizing policies. Austria, Slovenia and Latvia can be said to feature the most familialist family policy landscape in our sample in 2004. Although scoring lower than Germany on the familization index, none of the three countries provides defamilizing policies. Countries such as Spain, Ireland, Great Britain and Greece score rather low on both policy dimensions. We argued that the provision of both weak familizing policies and weak defamilizing policies creates patterns of implicit individualism or implicit familialism depending on the predominant family ideals. Not showing a strong degree of familizing or defamilizing policies, other countries such as Poland still lean more towards either of the two directions. Hungary and Latvia score higher on familizing policies, while the Czech Republic and Portugal exhibit stronger defamilizing policies. Other countries, most prominently the Netherlands plotted in the middle of our graph, come out as ambiguous cases.

Figure 2.

Familization and defamilization across countries 2004, composite indices.

Discussion

The starting point for this article was the observation that the concepts of familization and defamilization are often used in comparative welfare state research, but that the translation of the concepts into empirical measures has had limited success in previous studies. Based on an extensive review of the theoretical and empirical literature, we have derived a conceptual framework and proposed indicators for operationalization. In the conceptual approach, we regarded dimensions of familizing and defamilizing policies in terms of intergenerational and gender-specific dependencies of individuals towards their families. The article proposes measuring familization and defamilization at the policy level, rather than at the country level. This makes it possible to consider both countries’ familizing and defamilizing policies, which often exist alongside each other and develop in different, but not always congruent directions over time. In addition, rather than defining the policy outcomes as familialism and defamilialism, we follow Daly (2011a; 2011b) in assuming that individualism is the outcome category of defamilizing and familizing policies that opposes familialism. Both of these, we argue, will vary across countries according to the cultural legacies in organizing dependency relationships in families. The article’s discussion of previous empirical studies, then, demonstrates the variety of approaches and divergence in how countries are ranked on a familization–defamilization continuum. In a final step, we propose indices of familizing and defamilizing policies which capture the multi-dimensionality of both concepts using current institutional data for a larger set of European countries. Using these indices, we can see that these countries exhibit family policy patterns that implicitly or explicitly support individualist or familialist family models. We do not find countries clearly falling into the category of support for optional individualism or optional familialism. However, the main focus of this exercise was directed less on substantial and more on practical issues. Operationalizing complex constructs such as defamilization is still hampered by a lack of data. Although data availability in the field of family policy has improved over the last few years, constraints prevail. Despite advances in recent years, there is still a lack of indicators for depicting change in family policy over time. In addition, in some areas, such as the quality and the costs of childcare or intergenerational dependencies, data availability is a bigger problem than in others. The situation clearly differs from other policy fields where discussion has moved from the question of the sheer availability of data to detailed issues about data quality (Danforth and Stephens, 2013). Valid and reliable data are the essential requirements for mapping the range and the trajectories of European countries’ family policy constellations.

Footnotes

Acknowledgements

We would like to thank two anonymous reviewers and Jochen Clasen for their helpful comments. We also thank the participants of the SFB 597 Workshop at the University of Bremen and especially Karin Gottschall for comments on an earlier draft.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Notes

References

Akgunduz

Y.E.

Plantenga

(2013) ‘Labour Market Effects of Parental Leave in Europe’, Cambridge Journal of Economics 37(4): 845–62.

Bahle

(2008) ‘Family Policy Patterns in the Enlarged EU’, in Alber

Fahey

Saraceno

(eds) Handbook of Quality of Life in the Enlarged European Union, pp. 100–25. London; New York: Routledge.

Bambra

(2004) ‘The Worlds of Welfare: Illusory and Gender Blind?’ Social Policy and Society 3(3): 201–11.

Bode

(2003) ‘The Organisational Evolution of the Childcare Regime in Germany: Issues and Dynamics of a Public–Private Partnership’, Annals of Public and Cooperative Economics 74(4): 631–58.

Crompton

(ed.) (1999) Restructuring Gender Relations and Employment: The Decline of the Male Breadwinner. Oxford: Oxford University Press.

Daly

(2011a) ‘Family Policy: Striving for Sustainability’, in Clasen

(ed.) Converging Worlds of Welfare? British and German Social Policy in the 21st Century, pp. 75–90. Oxford: Oxford University Press.

Daly

(2011b) ‘What Adult Worker Model? A Critical Look at Recent Social Policy Reform in Europe from a Gender and Family Perspective’, Social Politics 18(1): 1–23.

Danforth

Stephens

J.D.

(2013) ‘Measuring Social Citizenship: Achievements and Future Challenges’, Journal of European Public Policy 20(9): 1285–98.

Ebbinghaus

(2005) ‘When Less is More. Selection Problems in Large-N and Small-N Cross-National Comparisons’, International Sociology 20(2): 133–52.

10.

Esping-Andersen

(1990) The Three Worlds of Welfare Capitalism. Cambridge: Polity Press.

11.

Esping-Andersen

(1999) Social Foundations of Postindustrial Economies. Oxford: Oxford University Press.

12.

Ferrarini

(2006) Families, States and Labour Markets: Institutions, Causes and Consequences of Family Policy in Post-War Welfare States. Cheltenham: Edward Elgar.

13.

Ferrarini

Nelson

Höög

(2013) ‘From Universialism to Selectivity: Old Wine in New Bottles for Child Benefits in Europe and Other Countries’, in Marx

Nelson

(eds) Minimum Income in Flux, pp. 137–60. London: Palgrave Macmillan.

14.

Gauthier

A.H.

(2002) Family Policies in Industrialized Countries: Is There Convergence? Population 57(3): 447-474.

15.

Gauthier

A.H.

(2011) Comparative Family Policy Database, Version 3. Netherlands Interdisciplinary Demographic Institute and Max Planck Institute for Demographic Research (distributors), Available at: www.demogr.mpg.de.

16.

Gornick

Meyers

M.K.

(2003) Families that Work: Policies for Reconciling Parenthood and Employment. New York: Russell Sage Foundation.

17.

Gornick

Meyers

M.K.

Ross

K.E.

(1997) ‘Supporting the Employment of Mothers: Policy Applications across Fourteen Welfare States’, Journal of European Social Policy 7(1): 45–70.

18.

Han

W.-J.

Ruhm

Waldfogel

(2009) ‘Parental Leave Policies and Parents’ Employment and Leave-taking’, Journal of Policy Analysis and Management 28(1): 29–54.

19.

Hernes

H.M.

(1987) Welfare States and Woman Power: Essays in State Feminism. Oslo: Norwegian University Press.

20.

Kangas

Rostgaard

(2007) ‘Preferences or Institutions? Work–Family Life Opportunities in Seven European Countries’, Journal of European Social Policy 17(3): 240–56.

21.

Keck

Saraceno

(2012) ‘Multilinks Database on Intergenerational Policy Indicators’, Schmollers Jahrbuch 132: 453–61.

22.

Korpi

(2000) ‘Faces of Inequality: Gender, Class, and Patterns of Inequalities in Different Types of Welfare States’, Social Politics: International Studies in Gender, State and Society 7(2): 127–91.

23.

Korpi

Ferrarini

Englund

(2013) ‘Women’s Opportunities under Different Family Policy Constellations: Gender, Class, and Inequality Tradeoffs in Western Countries Re-examined’, Social Politics: International Studies in Gender, State and Society 20(1): 1–40.

24.

Kröger

(2011) ‘Defamilisation, Dedomestication and Care Policy: Comparing Childcare Service Provisions of Welfare States’, International Journal of Sociology and Social Policy 31(7/8): 424–40.

25.

Leitner

(2003) ‘Varieties of Familialism: The Caring Function of the Family in Comparative Perspective’, European Societies 5(4): 353–75.

26.

Leitner

Lessenich

(2007) ‘(In-)Dependence as Dependent Variable: Conceptualising and Measuring “De-familisation”’ in Clasen

Siegel

N.A.

(eds) Investigating Welfare State Change: the ‘Dependent Variable Problem’ in Comparative Analysis, pp. 244–60. Cheltenham: Edward Elgar.

27.

Leitner

Ostner

Schratzensteller

(2004) ‘Einleitung: Was kommt nach dem Ernährermodell? Sozialpolitik zwischen Re-Kommodifizieung und Re-Familisierung’, in Leitner

Ostner

Schratzensteller

(eds) Wohlfahrtsstaat und Geschlechterverhältnis im Umbruch. Was kommt nach dem Ernährermodell? pp. 9–27. Wiesbaden: Verlag für Sozialwissenschaften.

28.

Lewis

(1992) ‘Gender and the Development of Welfare Regimes’, Journal of European Social Policy 2(3): 159–73.

29.

Lewis

(2001) ‘The Decline of the Male Breadwinner Model: Implications for Work and Care’, Social Politics: International Studies in Gender, State and Society 8(2): 152–69.

30.

Lewis

Giullari

(2005) ‘The Adult Worker Model Family, Gender Equality and Care: The Search for New Policy Principles and the Possibilities and Problems of a Capabilities Approach’, Economy and Society 34(1): 76–104.

31.

Lister

(1990) ‘Women, Economic Dependency and Citizenship’, Journal of Social Policy 19(4): 445–67.

32.

Lister

(1994) ‘“She Has Other Duties”: Women, Citizenship and Social Security’, in Baldwin

Falkingham

(eds) Social Security and Social Change: New Challenges to the Beveridge Model, pp. 31–44. New York: Harvester Wheatsheaf.

33.

Lohmann

(2009) ‘Konzept und Messung von Defamilisierung in international vergleichender Perspektive’, in Pfau-Effinger

Magdalenić

Wolf

(eds) International vergleichende Sozialforschung, pp. 93–128. Wiesbaden: Verlag für Sozialwissenschaften.

34.

McLaughlin

Glendinning

(1994) ‘Paying for Care in Europe: Is There a Feminist Approach?’ in Hantrais

Mangen

(eds) Family Policy and the Welfare of Women, Cross-National Research Papers, pp. 52–69. Leicestershire: European Research Centre, Loughborough University of Technology.

35.

Mandel

Shalev

(2009) ‘How Welfare States Shape the Gender Pay Gap: A Theoretical and Comparative Analysis’, Social Forces 87(4): 1873–911.

36.

Mätzke

Ostner

(2010) ‘Change and Continuity in Recent Family Policies’, Journal of European Social Policy 20(5): 387–98.

37.

Nelson

(1990) ‘The Origins of the Two-Channel Welfare State: Workmen’s Compensation and Mother’s Aid’, in Gordon

(ed.) Women, the State, and Welfare, pp. 123–51. Madison, WI: University of Wisconsin Press.

38.

O’Connor

J.S.

(1993) ‘Gender, Class and Citizenship in the Comparative Analysis of Welfare State Regimes: Theoretical and Methodological Issue’, British Journal of Sociology 44(3): 501–18.

39.

Organisation for Economic Co-operation and Development (OECD) (1995) Employment Outlook. Paris: OECD.

40.

Organisation for Economic Co-operation and Development (OECD) (2012a) OECD Family Database. Paris: OECD, Available at: www.oecd.org/social/family/database (accessed 22 November 2013).

41.

Organisation for Economic Co-operation and Development (OECD) (2012b) OECD Social Expenditures Database. Paris: OECD, Available at: http://www.oecd.org/social/soc/socialexpendituredatabasesocx.htm (accessed 11 October 2013).

42.

Orloff

A.S.

(1993) ‘Gender and the Social Rights of Citizenship: The Comparative Analysis of Gender Relations and Welfare States’, American Sociological Review 58(3): 303–28.

43.

Ostner

(2003) ‘“Individualisation” – The Origins of the Concept and its Impact on German Social Policies’, Social Policy and Society 3(1): 47–56.

44.

Pfau-Effinger

(2004) Development of Culture, Welfare States and Women’s Employment in Europe. Aldershot: Ashgate.

45.

Saraceno

(1997) ‘Family Change, Family Policies and the Restructuring of Welfare’, in OECD (ed.) Family, Market and Community, pp. 81–100. Paris: OECD.

46.

Saraceno

Keck

(2010) ‘Can We Identify Intergenerational Policy Regimes in Europe?’ European Societies 12(5): 675–96.

47.

Saxonberg

(2013) ‘From Defamilialization to Degenderization: Toward a New Welfare Typology’, Social Policy and Administration 47(1): 26–49.