„4 The current issue and full text archive of this journal is available at *&£2)~ www.emeraldinsight.com/0144-333X.htm IJSSP 31,7/8 424 Received 16 October 2010 Revised 12 December 2010 Accepted 11 January 2011 Def amilisation, dedomestication and care policy Comparing childcare service provisions of welfare states Teppo Kröger Department of Social Sciences and Philosophy, University ofjyväskylä, Jyväskylä, Finland Abstract Purpose - This paper aims to use perspectives from both mainstream and feminist welfare state research in drafting a conceptual approach for social care research. This approach is then applied empirically to a comparative analysis of childcare provisions of 15 OECD countries. Design/methodology/approach - The concept of dedomestication is developed from a discussion on the notions of decommodification and defamilisation, and it is defined as the degree to which social care policies make it possible for people to participate in society and social life outside their homes and families. In the empirical part of the paper, dedomestication of childcare service provisions of 15 welfare states is measured by an index that is constructed on the basis of time replacement rate, availability, affordability, quality, and take-up rates of care services. Findings - Denmark offers the highest degree of dedomestication to parents of young children, followed by a group of Nordic and Western European countries. In English-speaking "liberal regime" nations, dedomestication remains more limited but it is lowest in the Central European countries of Hungary and Austria. The findings only partly follow earlier welfare regime categorisations. Originality/value - The paper develops a new original conceptual framework for comparative study of care services that is then applied to an empirical analysis of childcare provisions in 15 welfare states, bringing out new results on the breadth of welfare state services. Keywords Defamilization, Dedomestication, Care services, Child care, Comparative research, Welfare regimes, Social welfare policy, Social care facilities Paper type Conceptual paper International Journal of Sociology and Social Policy Vol. 31 No. 7/8, 2011 pp. 424-440 © Emerald Group Publishing Limited 0144-333X DOI 10.1108/01443331111149860 1. Introduction Comparative research on social care systems has on the one hand followed the same regime-centred lines of discussion as the mainstream of comparative welfare state research; on the other hand it has developed in novel directions, due mainly to its closeness to feminist theoretisation (Kroger, 2001). This paper uses concepts and perspectives from both mainstream and feminist welfare state research in drafting a conceptual approach for The author has used the notion of dedomestication since 1995 and has always thought that he is the first and so far the only person to use the concept. However, in January 2010, he heard from Anne Skevik that she had actually used exactly the same term with an almost identical meaning in her unpublished thesis already in 1994. So, the author wishes to acknowledge her as co-creator of the concept. He is grateful for comments on earlier versions of this paper to Anneli Anttonen, Silke Bothfeld, Yueh-Ching Chou, Trudie Knijn, Jorma Sipila, and anonymous reviewers. The writing of this article was financed by two research grants from the Academy of Finland (under the titles "Care in context" no. 113340 and "Working carers and caring workers" no. 124450). comparative social care research. This approach is then applied empirically to a Defamilisation, comparative analysis of childcare provisions of 15 OECD countries. dedomestication Comparative welfare state research focused for a long time mostly on social j expenditure and social security systems, in particular on the coverage and replacement levels of pension systems and sickness and unemployment benefits. It was exactly these three systems of welfare benefits that Esping-Andersen (1990) chose to be included in his famous decommodification score, which laid the foundation for his three regimes of 425 welfare capitalism. Ever since, welfare regime thinking has characterised almost all comparative social policy research. Comparative studies made during the last two decades refer almost without an exception to Esping-Andersen's original categorisation of welfare states, often in a critical tone but nonetheless starting from his regime conceptualisations, adding new regimes, deconstructing old ones, or developing alternative criteria for clustering welfare states. Comparative studies on care developed considerably later than international research on social benefit systems. Whereas in social security research, Esping-Andersen's book represented a synthesis and culmination of a long tradition of comparative research, in social care research it sparked off a new strand of comparative literature - as a reaction from care researchers to Esping-Andersen's original omission of care (Alber, 1995; Anttonen and Sipila, 1996; Kautto, 2002; Rauch, 2007). Another major source of influence for international comparisons on care has been feminist scholarship, focusing in particular on childcare. Even before Esping-Andersen's book, there were comparative studies on childcare issues. These particularly emphasised the importance of formal childcare provisions to lone mothers (Kamerman and Kahn, 1981; Borchorst, 1990). The launch of The Three Worlds of Welfare Capitalism was followed by intensive counter argumentation from a number of feminist scholars (Langan and Ostner, 1991; Lewis, 1992; O'Connor, 1993; Orloff, 1993; Daly, 1994) as Esping-Andersen's approach was claimed to be gender-blind and to disregard those welfare benefits and services that are used and needed by women. A number of new concepts were soon developed for a more gender-sensitive comparative analysis of welfare states. Lewis (1992) created her typology of breadwinner models; Sainsbury (1996) spoke about "the maximum private responsibility model" and "the maximum public responsibility model" in reference to childcare policies. In connection to labour market participation of lone mothers, Lewis and Hobson (1997) drafted two ideal types of care regimes ("parent/worker model" vs "care-giving model"). Other writers took the concepts of personal autonomy and independence (of women) as the primary yardsticks in comparing social policies of different nations (O'Connor, 1993; Orloff, 1993). However, the feminist concept that has probably received the widest use in comparative welfare research is defamilisation. This was originally developed by Lister (1994,1995) expressly as a counter concept to Esping-Andersen's decommodification. Based on a discussion on the notions of defamilisation and decommodification, this article develops another concept, dedomestication. The new concept is then used in a comparative quantitative analysis of childcare provisions in 15 welfare states, aiming to offer a novel and comprehensive view on the level of care policy within these nations. The article concludes with a discussion that focuses, on the one hand, on the methodological limitations of the analysis and, on the other hand, on its research and policy implications. 2. Decommodification and defamilisation: different but similar Inspired by Marx and Polanyi who associated the process of human commodification with capitalism, Esping-Andersen (1990, pp. 3, 21-23, 37) presented his concept of decommodification as the basic criterion for the existence and scope of social rights in a society. According to him, the concept expresses: [...] the degree to which social policies permit people to make and maintain their living at a socially acceptable level independent of market forces, without having to sell their labour power on the labour market. Decommodification thus represents the degree to which citizens' status as "pure commodities" is diminished by existing social rights. Esping-Andersen emphasizes the word "degree"; he does not presume to find a total eradication of labour as a commodity in any welfare regime. According to Esping-Andersen (2000, pp. 353, 357), the concept of decommodification "captures one important dimension of freedom and constraint in the everyday life of advanced capitalism" because by decreasing market dependency, decommodifying social policy "creates greater space for individuals to control their lives". In a way, decommodification is a process whereby welfare states reduce citizens' economic dependency on the market, replacing it with a dependency on welfare benefits. Operationally, Esping-Andersen (1990, p. 54) chose three central social insurance programmes - old-age pensions, sickness benefits, and unemployment benefits - and analysed the breadth of each of them in 18 Western welfare states with four indicators, weighted by take-up rates[l]. Adding the final scores of the three benefit systems together gave him the total decommodification score of each individual welfare state. The international decommodification league table was then divided into three parts, which is how The Three Worlds of Welfare Capitalism came into being. Nevertheless, decommodification was soon heavily questioned. Particularly from a feminist perspective, decommodification measured "wrong" issues; it was not dependency on the labour market but dependency on the family that was the primary problem that social policy should solve. Feminist welfare research (Langan and Ostner, 1991; O'Connor, 1993; Orloff, 1993; Bussemaker and van Kersbergen, 1994; Daly, 1994; Hobson, 1994; Sainsbury, 1996) argued that patriarchal family structures cause many women to be economically dependent on their husbands (or fathers or brothers) and the solution is not to be found from decommodification but from the opposite direction, from commodification of women, and thus welfare states should actively support women's entry to the labour force. Instead of "opting out of work" women need support from social policy to "opt out of family" and to "opt into work". One of the alternative concepts that was soon developed within feminist scholarship as a more gender-sensitive or woman-friendly benchmark for comparative welfare state research was defamilisation[2]: [T]he dimension of decommodification needs also to be complemented by that of what we might call "defamilialisation", if it is to provide a rounded measure of economic independence. Welfare regimes might then also be characterised 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 (Lister, 1994, p. 37). Just before the publication of The Three Worlds, Lister (1990) had analysed women's economic dependence on their male partners and welfare benefits, highlighting how these "private" and "public" dependences limit women's social citizenship. Defamilisation, As Esping-Andersen's decommodification did not recognise these dependences, Lister (^domestication developed her own alternative concept that nonetheless bears a close resemblance to , decommodification in its formulation. In defamilisation, Esping-Andersen's "market ana care participation" as the source of dependency is replaced by "family relationships". Unlike Esping-Andersen, Lister defines her concept as applying solely to individuals and not to families. She does not mention women in her definition of defamilisation (that is restated 427 in Lister (1995, p. 29,2003, p. 172)), but it is evident that the "individual adults" in need of - "a socially acceptable standard of living" and independence from families, whom she has in mind, are not men in the first place. When looking for "independence from family relationships", it is actually independence from male economic domination that is in question. Paid work and welfare benefits are the two suggested routes to this independence. All in all, even though Lister's wording is close to the definition of decommodification, the perspective of her notion is profoundly different. Esping-Andersen's concept promises individuals and families freedom from the market, whereas Lister's concept calls for freedom from families for (female) individuals, sponsored not only by welfare states but also by the market. The market turns from a source of dependency to a source of independence, while the family turns from a receiver of independence into a source of dependency. Defamilisation and decommodification locate the source of dependence in opposite directions but they both look for solutions from the welfare state, even though for Lister, social policies manifest themselves only as a secondary and supplementary source of freedom. However, there is an even more significant parity between the two concepts than just seeing the welfare state as a liberating force; at the heart of both decommodification and defamilisation is economic independence. Esping-Andersen is looking for economic independence from the market, Lister from the family, but both aim to promote the financial autonomy of citizens. Decommodification may have been created with the male worker in mind, while defamilisation was launched specifically to serve women's interests, but in their concentration on economic independence, the two concepts are identical. 3. Another version of defamilisation: bringing care in From the perspective of social care research, the problem with both Esping-Andersen's decommodification and Lister's defamilisation is that they direct their attention towards welfare benefits, not towards care services. In Anttonen's (1990) terms, both focus on "the social insurance state" and disregard "the social service state". This goes also to Lister's concept. Even though she has otherwise written extensively on care (Lister, 1995, 2003), her defamilisation concept is centred on social security, income protection, and labour market entrance, not on care. In the same year as the launch of the defamilisation concept, Bussemaker and van Kersberger (1994, p. 24) argued that although economic independence covers many interests of women, it is not sufficient to cover all of them. Moreover, Lewis (1997, pp. 173-4) stated expressly that the search for gender-centred welfare policies is a more complicated business since, in addition to women's right to engage in paid work, it is also necessary to consider their right to engage in unpaid work as well as their right to do the opposite, not to perform unpaid work - and that even their right not to engage in paid work needs to be recognised. This comes close to Knijn and Kremer (1997) who have argued that citizens' rights to receive care (in the case of parents, for their children) need to be supplemented by a right to have time for care. The income-focused concept of defamilisation is unable to cover any of these care-related rights. However, there exists another definition of defamilisation. In the very same year that Lister introduced her concept, McLaughlin and Glendinning (1994, p. 65) used the same term ("de-familisation", to be exact) but with a different meaning, to represent "those provisions and practices which vary the extent to which well-being is dependent on 'our' relation to the (patriarchal) family". This version of the same concept is not limited to economic independence. McLaughlin and Glendinning (1994, p. 64) aimed expressly to move beyond the early 1990s' feminist critique of Esping-Andersen and to develop instead a conceptualisation which relates to the terms and conditions under which people engage - and do not engage - in caring relationships. It is noteworthy that they extended their concept to cover the experiences of people who receive care, in addition to those who provide care. They hoped that: [...] a more developed notion of the concept of de-familisation might provide us with [...] a set of criteria with which to judge the political merits of various political developments in this field from the perspectives of those receiving and those providing care" (McLaughlin and Glendinning, 1994, p. 64). As can be seen immediately, this version of the concept, focusing on care, is very different from the one launched by Lister, focusing on economic independence. In research that has been published during the last 15 years, these two competing definitions of the term have nevertheless often become mixed up with each other. For example, when replying to the feminist criticism on his decommodification, Esping-Andersen (1999, p. 51) adopted the concept of defamilisation (without any reference to its original creators) and denned it as "the degree to which households' welfare and caring responsibilities are relaxed either via welfare provision, or via market provision". Overall, defamilisation has become one of the most regularly used concepts in recent welfare state literature, but authors very rarely make any distinction between the two different original meanings of the concept, one focusing on economic independence and the other one looking for criteria for the evaluation of care policies. Nevertheless, the concept has received serious attention from Sigrid Leitner. In her comparative analysis of care policies of EU member states, she follows the concept of McLaughlin and Glendinning and defines defamilisation as "taking away care responsibilities from the family" or "unburdening the family in its caring function". She even takes into account the care receiver's perspective and states that defamilisation "reduces the extent to which the satisfaction of individual care needs is dependent on the individual's relation to the family" (Leitner, 2003, p. 358). In another article, written with Lessenich, Leitner places the concept under systematic scrutiny, criticising Esping-Andersen and the dominant EU labour market policy discourse (which aims to enable participation of women in the labour market by freeing them from family care responsibilities) for "not taking de-familization seriously", that is, of not recognising its complexity and many-sidedness (Leitner and Lessenich, 2007). Interestingly, even though they do not make any reference to Lister or to Lister's original definition of defamilisation, Leitner and Lessenich (2007, p. 252) argue that the main problem of "the mainstream conceptualizations of de-familization" is that these define the independence of the care giver solely as financial independence, "ignoring the social and emotional aspects of the care giver's 'liberation' from his or her care responsibilities". Their argument is that, in addition to the economic independence that women (or men) may (or may not) receive through paid work when (part of) their care responsibilities are taken over by the state or the market, care relationships have social and emotional dimensions that need to be taken into account. Defamilisation is not only about economic independence but also about social independence and, furthermore, it is a mistake to focus entirely on the care giver and fail to acknowledge the needs of the care receiver. They come up with an extended and complex version of the defamilisation concept where economic defamilisation is joined by social defamilisation and where defamilisation becomes almost a synonym for independence. 4. From defamilisation to dedomestication Leitner and Lessenich's contribution that is based on McLaughlin and Glendinning's original concept shows a way forward, indicating how care and both parties of the care relationship can be brought into the conceptual framework. Their criticism of the way mainstream discussions observe only the economic dimension of defamilisation, and how defamilisation as such does not necessarily guarantee labour market participation or economic independence highlights the need for a concept that focuses on care. This is actually what McLaughlin and Glendinning were initially aiming at with their notion. However, their version of defamilisation has received considerably less attention than the version that focuses on women's economic independence, and even where both meanings of the (almost) same concept have stayed alive, they have become largely mixed up with each other, causing conceptual confusion. Leitner and Lessenich's suggestion is to highlight that defamilisation has two sides - the economic one and the social/emotional one - but making the defamilisation concept even more complex does not really promote conceptual clarity[3]. Another difficulty with the concept of defamilisation is that it implies "freedom from family" and this has disturbing connotations in the area of care, since family care is and will continue to be of vital importance and highly valued by many of its givers and receivers (O'Connor et al., 1999, p. 32; Leitner and Lessenich, 2007, p. 252). All in all, it might be better to let defamilisation refer solely to economic independence, and develop a new notion to describe independence from familial care relationships. This paper suggests dedomestication as such a concept. This is understood as the degree to which social care policies make it possible for people to participate in society and social life outside their homes and families. Dedomestication refers thus to freedom from a confinement to the domestic sphere or, using the term of Lewis (1997), to citizens' right to limit their engagement in unpaid informal caring. From the care receiver's perspective, dedomestication is about not being dependent on informal care from close persons, in other words, about citizens' right to receive formal care (Knijn and Kremer, 1997). Whereas decommodification is a counter force to the process of commodification, which turns people into commodities that are sold and bought in the labour market, dedomestication represents a counteraction to the strong and gendered societal pressures that separate the domestic sphere and the "public" sphere from each other. These forces are pushing care-giving to the "private" sphere of the family and, furthermore, isolating both care givers and care receivers from the public sphere and from participation in social life. The concept of dedomestication is introduced as a criterion to evaluate the role of social care policies in making possible a life outside Defamilisation, dedomestication and care 429 IJSSP the domestic sphere for both care givers (through bringing freedom from "compulsory 2 ^ 7/g altruism") and care receivers (through offering alternatives to "enforced dependency" on ' family members). On the other hand, welfare states have increasingly adopted care policies that support the domestication of care, materialising people's right to engage in caring for those close to them (Knijn and Kremer, 1997). Social policy measures such as parental leave benefits 430 or carers' allowances increase the opportunities of people to offer informal care to their - family members within the domestic sphere. Such policies have been criticised for their gendered consequences and negative side-effects, in particular for weakening the position of women in the sphere of paid work, and for institutionalising the gendered division of caring labour. On the other hand, many women's organisations have been supporting such policies as a way to make women's informal care-giving work visible in society and to recognise its value (Lister, 1990, 1995). Domesticating policies are contested but they do offer time for care (Knijn and Kremer, 1997) and the right to engage in caring (Lewis, 1997). All in all, this paper builds on McLaughlin and Glendinning's original version of def amilisation and Leitner and Lessenich's discussion of the concept, but it suggests that it would be better to limit the concept of defamilisation to the meanings that Lister gave it in 1994. Defamilisation works well as a conceptual tool in studying economic independence from other family members and the role and design of social policy benefit systems in this respect. However, using the same concept in regard to care (and to "social independence") is not seen here as an effective conceptual strategy. The different meanings given to defamilisation in recent welfare state literature, used often in a mixed way without adequate distinction, bring conceptual confusion and limit its usefulness as an analytical concept in the area of care. Furthermore, it is problematic for a concept that implies liberation from the family to serve as a core concept in discussing care, including family care. As a consequence, the concept of dedomestication has been introduced and suggested as an approach for analysing and comparing the outcomes of care policies of different welfare states. The final section of the paper describes an experiment to apply the conceptual lens of dedomestication to a comparative study of childcare service provisions of different nations. The fact that only childcare is scrutinised here does not imply that the approach would be unsuitable for the study of care services for older or disabled people. On the contrary, the perspective of dedomestication would have direct relevance in these two areas for both care givers and care receivers, whereas in the case of childcare "confinement to the domestic sphere" is a risk that is certainly more relevant to care givers (parents) than to care receivers (children). The choice to focus on childcare here is only due to the availability of international data, as explained below. 5. Data and method The lack of comparable and reliable data has been the Achilles' heel of comparative social care research. Care services are a complex and ever-changing formation of dozens of different kinds of activities that have various labels in different countries, and that are divided by many administrative and functional boundaries - like the one between social care and early education - which are positioned dissimilarly in different societies. Capturing this multitude and complexity in figures in an internationally standardised way has proved to be difficult. Without reliable international databases, comparative care research has lagged methodologically Defamilisation, behind comparative social security research. dedomestication However, the situation has been changing gradually. OECD and Eurostat have invested considerable resources in the construction of comparative databases that aim a care to cover care provisions. Childcare in particular has recently become an object for intensive statistics-building, due to its heightened position on the political agenda, being considered as part of both "the social investment state" and "the reconciliation of 431 work and family" (OECD, 2006, 2007a). This paper uses childcare statistics that were gathered and harmonised by OECD and included in the OECD Family Database (OECD, 2007b, c, 2009) and in the publication Starting Strong II (OECD, 2006). However, it is necessary to remain cautious; there are still real risks of statistical errors and inconsistent categorisations. On the other hand, by using rough values instead of the original exact figures an occasional flaw in the original data becomes less harmful. This paper follows the example and procedures of Esping-Andersen's decommodification score, aiming to construct a new tool, a dedomestication index, for comparative study of social care systems. This index indicates the strength of welfare states in offering opportunities for participation in paid work and social life, in this case, for parents of young children. Like the decommodification score, the dedomestication index includes four main factors: (1) time replacement rate; (2) availability; (3) affordability; and (4) quality. These four indicators have been chosen in order to capture essential dimensions of care service systems that are of critical importance to parents[4]. Furthermore, as when measuring decommodification of social security provisions, the summary value of the four indicators is finally multiplied by take-up rates. Time replacement rate is self-evidently the functional equivalent of replacement rate of social security benefits. It refers to the temporal coverage of the concerned care service and is here operationalised as average hours of attendance per week. Like replacement rate in the decommodification score, this indicator is given double weight in the dedomestication index as it really makes a difference whether the average weekly hours of service use are 17 (as in The Netherlands) or 40 (as in Portugal). Availability is understood here as the breadth of universal access to childcare, and countries are rated according to whether they offer a right to childcare for all families or not. Without such a right, access to services will depend more or less on economic resources of families and on a number of other factors, including sheer good luck. Affordability is another major factor determining the use and the usefulness of childcare provisions of a country. Affordability is measured here as the proportion of net costs (after child care benefits, tax benefits, and other benefits that reduce the costs of childcare services) of family income. Even if families have universal access to full-day childcare services but the user charges are too high for them, the provisions remain worthless. The same goes for the last of the four indicators, quality. If the quality of childcare services is unacceptably low, families will not use them if they have any other way to organise the care of their children. IJSSP From many possible yardsticks of quality of childcare, child-staff ratios are used here as y/g there are data available on them[5]. ' The study compares dedomestication within childcare service provisions for children aged under three in 15 countries. The selection of countries as well as the selection of the service field is based solely on the availability of data. The OECD Family Database includes information about childcare services altogether from almost 432 40 countries, but only from 15 countries are there data on each of the four chosen - indicators plus take-up rates. On the other hand, the database includes several variables concerning services for the over three-year-old age group but unfortunately not all necessary factors are covered, which is why the analysis is here limited to childcare services offered for the 0-2-year-old age group. 6. Findings 6.1 Indicator values The four indicators of the dedomestication index focus on different features of childcare systems and the variations in their values show how individual welfare states are stronger in some respects and weaker in others (Table I). Concerning time replacement rate, that is, the temporal coverage of the service, Canada, Denmark, Finland, Norway, and Portugal provide the largest service and, at the other extreme, Australia, Austria, Ireland, The Netherlands, and the UK the most limited service. On the other hand, Finland is the only country of the 15 welfare states to offer unlimited access to all families with young children. In Belgium, Hungary, Norway, Portugal, and Sweden, childcare services are most affordable, while the other end of the spectrum is represented by Austria, Canada, Ireland, the UK, and the USA. Concerning the quality of care services for children under three years, measured by the child-staff ratio, the weakest performances come Time replacement Take-up Dedomestication Table I. Dedomestication index and indicator values, childcare services for children under three-year old in 15 countries, ca 2005 rate Availability Affordability Quality Sum rate index Denmark 6 2 2 3 13 70 9.1 Portugal 6 1 3 1 11 44 4.8 Sweden 4 2 3 2 11 44 4.8 The Netherlands 2 1 2 3 8 54 4.3 Belgium 4 2 3 1 10 42 4.2 France 4 1 2 2 9 43 3.9 Norway 6 1 3 1 11 35 3.9 Finland 6 3 2 3 14 25 3.5 United States 6 1 1 2 10 31 3.1 UK 2 1 1 3 7 40 2.8 Canada 6 2 1 1 10 24 2.4 Australia 2 1 2 2 7 25 1.8 Ireland 2 1 1 3 7 25 1.8 Hungary 4 2 3 2 11 10 1.1 Austria 2 1 1 1 5 10 0.5 Mean 4.1 1.5 2.0 2.0 9.6 34.8 3.5 Note: For the scoring procedure, see the Appendix Source: OECD (2006, 2007b, c, 2009) from Austria, Belgium, Canada, Norway, and Portugal, whereas there are considerably more staff working with children in formal day-care in Denmark, Finland, Ireland, The Netherlands, and the UK. In order to get a more comprehensive evaluation on the overall level of childcare provisions in these countries, the values of the four indicators were counted together. As a result, distinctive variations emerge between the 15 welfare states. Finland and Denmark receive the highest total values, being very close to the maximum value possible. They are followed by a large group of countries that receive almost identical values to each other including Belgium, Canada, France, Hungary, Norway, Portugal, Sweden, and the USA. The remaining group of countries rate less well. Australia, Austria, Ireland, The Netherlands, and the UK are the five welfare states that receive the lowest summed values of the four indicators. However, the summed values do not yet represent the final dedomestication index. They offer an evaluation of the institutional settings of childcare provisions but they do not yet show the actual outcomes before take-up rates are taken into account. Take-up rate measures actual service use, that is, how large part of families really uses the service. In that sense, it measures the factual contributions of welfare states, not just official principles of policy. Take-up rates reflect social practice that is influenced by a number of issues including normative conceptions of child care and gender roles, as well as the demand for workforce in the labour market. It is a simplification to think that take-up rates are identical to coverage rates, that a low take-up rate simply reflects the institutional provisions (Leitner and Lessenich, 2007). For example, in Finland all families with under-school-age children are entitled to formal childcare but many families do not use this right. In the case of childcare for children under three years, take-up rates vary significantly among the 15 countries under study and, interestingly, this variation does not correspond closely to the variation of the four previously mentioned indicators (Table I). Denmark does have the highest take-up rate but it is followed, more surprisingly, by The Netherlands and Portugal. Sweden, France, and Belgium follow, which is less unexpected, but the UK rates here considerably better than with the four previous indicators. The take-up rate in Finland is noticeably low compared to its indicator ratings. Take-up rates in all English-speaking countries are rather high (24-40%) while they received only low indicator values. The lowest take-up rates are in Austria and Hungary, and Austria also had the lowest indicator values. All in all, compared with the four earlier indicators, take-up rates show a largely dissimilar picture of the state of childcare policies within the 15 studied countries. Only by putting both parts of the jigsaw puzzle together, on the one hand the four indicators reflecting the institutional settings of childcare policy and on the other hand the take-up rates that show the actual service use, does the picture become complete. The final dedomestication index organises the countries into a ranking order, condensing all five analysed dimensions into a single figure. Defamilisation, dedomestication and care 433 6.2 Dedomestication index values The dedomestication index describes the actual capacity of social care policies of welfare states to support people's opportunities to participate in society and social life outside their homes and families, in practice, by taking over some of their care responsibilities. In this case, it is parents, especially mothers, of young children whose participation in society - in particular their commodification within the labour market and hence their economic independence - is facilitated by the service provisions. The dedomestication index aims to offer a compact tool for an interpretation of the basic qualities of a multi-dimensional phenomenon, that is, the social care service provisions of a welfare state. When applied to an analysis of childcare provisions for children under three years, the results show that Denmark forms "a welfare regime of its own", and is overwhelmingly superior to any other welfare state in liberating the parents of young children so they can have a life also outside the walls of the family home (Table I). The childcare service provisions of no other country even come close to those of Denmark, which has a total index value of 9.1. This is not such a surprise because Denmark rated extremely well on every institutional indicator and also had the highest take-up rate. It is nevertheless striking that Denmark's lead over all other countries proves to be so wide. It is a matter of discretion where the boundary lines between other groups of countries are drawn but the seven following countries can be said to be close to each other, all having dedomestication index values between 3.5 and 4.8. This "cluster" includes three Nordic nations: Sweden, Norway, and Finland. Finland is an interesting case because it is at the level of Denmark concerning institutional settings but drops considerably in the list as a result of a low take-up rate, which is mainly due to the popularity of a child home care allowance scheme (Kröger et al, 2003). However, this group also encompasses countries from both Southern Europe and Continental Western Europe. In comparative welfare studies these countries are usually placed in the "Latin rim" or the "conservative-corporatist" regimes and are not supposed to be so active in the building of their formal care services as the "social democratic regime" countries. On the other hand, a number of earlier comparative studies of social care regimes (Anttonen and Sipilä, 1996) have emphasised that France and Belgium have developed broad formal childcare services that are universally available even for very young children. It is more surprising that The Netherlands has also joined the same group of nations, as it mainly offers part-time childcare services. The time replacement rate and availability of Dutch childcare provisions are actually very low but the country fares better on affordability and quality. It has the second highest take-up rate among the 15 countries, and this elevates the position of The Netherlands in the list. Portugal is probably an even more surprising entry in this group of welfare states, receiving actually the second highest dedomestication index value, along with Sweden. Being a Southern European nation whose social policies began to be expanded only in the 1980s, Portugal's position in the list is an extraordinary achievement. When looking more closely at the indicator values, it can be seen that families do not have a legal right to childcare services in Portugal, and also the quality of its day-care services is not high, if measured with the child-staff ratio. However, the temporal coverage of formal childcare is the highest of all countries and also the affordability is on a good level. When this is complemented with a take-up rate that is at the level of Sweden, Belgium, and France, Portugal's position seems justified. All English-speaking countries in the study are situated in the remaining group of countries. Not just the UK and the USA but also Canada, Australia, and Ireland are placed in the latter part of the list. Dedomestication index values of the USA and the UK are actually not much lower than that of Finland, but on the other hand it is probable that their take-up rates also include non-publicly provided and funded childcare services, which means that their index ratings may represent overestimations. On the other hand, North American countries seem to differ from other English-speaking nations in offering Defamilisation, full-day coverage, while Australian, British, and Irish formal provisions are mostly very dedomestication part time. There are also other variations within this group of countries; Ireland for , example having a much lower take-up rate than others, but the availability and Slid care affordability of childcare services are low in all English-speaking countries included in the study. However, the lowest dedomestication index values do not come from within 435 "the liberal regime" but from Central Europe. The neighbouring countries, Austria and - Hungary, are placed at the bottom when the actual contributions of welfare states in supporting parents in participating in the wider society are assessed. Austria received the minimum values for every institutional indicator and its take-up rate was also the lowest, 10 per cent. Hungary shared the lowest take-up rate with Austria but otherwise all of its indicator values were considerably higher. For affordability, Hungary even received a full score. However, because its take-up rate was so much lower than those of the other countries, the final position of Hungary nevertheless remained second lowest. 7. Conclusions Conceptually, this paper has aimed to draft an alternative approach for the comparative study of social care service provisions, being influenced by both mainstream and feminist welfare state discussions, in particular by the concepts of decommodification and defamilisation. However, both of these concepts focus originally on economic independence, which reduces their applicability for the study of care services. As a way forward, based on the work of McLaughlin and Glendinning and, more recently, Leitner, the concept of dedomestication was drafted to measure the degree to which social care policies make it possible for people to participate in society outside their homes and families, including the sphere of paid work as well as social life in general. Empirically, this paper represents an experiment to follow the procedures through which the decommodification score was created by Esping-Andersen around 20 years ago, trying to apply these to a comparative study of childcare service provisions of different welfare states. Like decommodification, dedomestication focuses on outputs of welfare states, not on their inputs. Social expenditure on childcare services is therefore not included in the index; the focus is instead on trying to measure the different dimensions of the existing care service provisions. The indicators that were deemed most essential and chosen to be included in the dedomestication index were time replacement rate, availability, affordability, and quality of the childcare services, multiplied in the end by their take-up rates. When combined, these factors constitute "the dedomestication index" that measures the capacity of welfare states to help people participate in society by taking over some of their caring responsibilities. The operationalisation of these indicators was strongly limited by the availability of comparative international data. For example, the only comparable data that were available on the quality of childcare were child-staff ratios. The data used in the study was originally gathered and harmonised by OECD but, as the numerous footnotes in the original data source publications show, the data still include a number of limitations in its comparability, being collected from various European and national sources. As social care in its complexity and changeability is an extremely challenging field for those who are trying to build international statistical databases, the current situation with the data remains far from perfect but in any case represents a step forward. The end result of this empirical experiment is another ranking order of welfare states, based on the level of dedomestication of their formal childcare provisions for children under three years. It can be questioned if such league tables have any theoretical or policy significance. Compared with studies that use a single variable (such as the level of social expenditure), the dedomestication index nevertheless covers several dimensions of social care provisions and offers a more comprehensive view on the state of childcare policies in 15 different welfare states. The analysis is based on a novel conceptual approach and methodologically a new tool is developed for comparative research. The empirical results are partly different from earlier regime categorisations, showing that the actual capacity of the welfare state to promote social participation of parents of young children is on a rather high level also in Portugal and The Netherlands, in addition to the Nordic countries and France and Belgium. Based on the findings, Nordic and Western European countries seem to melt into one, large regime that is followed by "the liberal regime" where childcare policies are weaker. However, the weakest overall performances come from a "Central European regime", represented here by Austria and Hungary. It is nevertheless good to remember that many nations from Europe and elsewhere (including all welfare states from East Asia and Latin America) could unfortunately not be included in this analysis, due to lack of data. In any case, the findings highlight the position of Denmark as the pioneer of childcare policy. Policywise, the analysis shows that if welfare states wish to support their citizens' participation in paid employment and in society, they need to work simultaneously with several issues. It is not enough to extend access to childcare services if their quality and affordability remain on a low level. On the other hand, if the opening hours of childcare facilities are limited, not even a service of the highest quality that is provided free of charge can bring considerable dedomestication. Childcare services are an aggregate of several issues and this aggregate does not deliver results before each of these issues becomes tackled by care policy. In order to have positive outcomes, policy makers do need to address availability and affordability of formal childcare but at the same time, they should not forget or compromise quality or opening hours. This article does not look at the level of dedomestication among care provisions for children over three years or within care services for older or disabled people, simply because comparable international data on all the chosen indicators is not yet available in these care service fields. It is hoped that the situation will change in the future, making it possible to compare dedomestication levels within these care services and, furthermore, from the perspective of care receivers - a standpoint which has not been included in this empirical analysis. On the other hand, dedomestication among different kinds of care services could also be analysed by comparative qualitative studies. Finally, it is necessary to stress that dedomestication is not claimed here to be the only appropriate yardstick for the evaluation of social care policies. The above-mentioned quote from Lewis (1997, pp. 173-4) states expressly that there is no single policy criterion - and thus, no single valid ranking order of welfare states - that is alone sufficient for the comparative study of social policy. In reference to her four fold approach, decommodification focuses on people's right not to engage in paid work, and commodification promotes the right of women and men to participate in paid work. Domestication offers people the right to engage in unpaid care-giving, whereas dedomestication supports people's right to limit their engagement in unpaid care work. Notes 1. For the latter two benefit systems these indicators included replacement rates, number of weeks of employment required prior to qualification, number of waiting days, and number of weeks the benefit could be maintained. In the case of old-age pensions, the indicators were slightly different. For each indicator and each country, Esping-Andersen gave a value 1 (low), 2 (medium), or 3 (high decommodification). In order to "take into account the singular importance of replacement rates for people's welfare-work choices", he gave replacement rates extra weight by multiplying them by 2. After this scoring procedure, the values of the four indicators were counted together and weighted by take-up rates. (Esping-Andersen, 1990, p. 54.) 2. The original concept launched by Lister (1994) was "defamilialisation". Over the years, different writers have spelled the term in a number of ways. Here the shortest version "defamilisation" is used with the exception of direct quotes where the form used by the original author is displayed. 3. For example, Leitner and Lessenich (2007, p. 251) state that the economic defamilisation of a parent is best granted by payments for childcare, that is, by a policy instrument supporting not defamilisation but (social) familisation of care, which just adds to the conceptual confusion. 4. These indicators have been emphasised already by many writers. For example, Lister (1995, p. 20) has highlighted that "the availability of affordable, good-quality child care is key to the achievement of the first aim (that parents and other carers can take paid work)". As well, Leitner and Lessenich (2007, p. 250) state that "the provision of child care services will contribute to the social de-familization of the parent, especially if they are low-cost, high-quality, and easily available". 5. For a detailed description of these four indicators and of the procedures for how their values were determined, see Appendix and Appendix Table. The dedomestication index comes rather close to Rauch's (2007) "defamilialization score" that is said to measure "the capacity of service systems to defamilialize care". This score is constructed by calculating first a standardised indicator full-time equivalency (FTE) value both for childcare and elderly care service fields and then multiplying these values by coverage levels, ending up with "FTE coverage levels". However, his score does not include indicators on the availability, affordability, or quality of care services. Defamilisation, dedomestication and care 437 References Alber, J. (1995), "A framework for the comparative study of social services", Journal of European Social Policy, Vol. 5 No. 2, pp. 131-49. Anttonen, A. (1990), "The feminization of the Scandinavian welfare state", in Simonen, L. (Ed.), Finnish Debates on Women's Studies: Women's Worlds, Realities and Choices, University of Tampere, Tampere, pp. 3-25. Anttonen, A. and Sipila, J. (1996), "European social care services: is it possible to identify models?"', Journal of European Social Policy, Vol. 6 No. 2, pp. 87-100. Borchorst, A. (1990), "Political motherhood and child care policies: a comparative approach to Britain and Scandinavia", in Ungerson, C. (Ed.), Gender and Caring: Work and Welfare in Britain and Scandinavia, Harvester Wheatsheaf, Hemel Hempstead, pp. 160-78. Bussemaker, J. and van Kersbergen, K. (1994), "Gender and welfare states: some theoretical reflections", in Sainsbury, D. (Ed.), Gendering Welfare States, Sage, London, pp. 8-25. Daly, M. (1994), "Comparing welfare states: towards a gender friendly approach", in Sainsbury, D. (Ed.), Gendering Welfare States, Sage, London, pp. 101-17. Esping-Andersen, G. (1990), The Three Worlds of Welfare Capitalism, Polity Press, Cambridge, MA. Esping-Andersen, G. (1999), Social Foundations of Postindustrial Economies, Oxford University Press, Oxford. Esping-Andersen, G. (2000), "Multi-dimensional decommodification: a reply to Graham Room", Policy & Politics, Vol. 28 No. 3, pp. 353-9. Hobson, B. (1994), "Solo mothers, social policy regimes, and the logics of gender", in Sainsbury, D. (Ed.), Gendering Welfare States, Sage, London, pp. 170-87. Kamerman, S.B. and Kahn, AJ. (1981), Child Care, Family Benefits and Working Parents: A Study in Comparative Policy, Columbia University Press, New York, NY. Kautto, M. (2002), "Investing in services in West European welfare states"', Journal of European Social Policy, Vol. 12 No. 1, pp. 53-65. Knijn, T. and Kremer, M. (1997), "Gender and the caring dimension of welfare states: toward inclusive citizenship", Social Politics, Vol. 4 No. 3, pp. 328-61. Kröger, T. (2001), Comparative Research on Social Care: The State of the Art, European Communities, Luxembourg. Kröger, T., Anttonen, A. and Sipilä, J. (2003), "Social care in Finland: weak and strong universalism", in Anttonen, A., Baidock, J. and Sipilä, J. (Eds), The Young, the Old and the State: Social Care Systems in Five Industrial Nations, Edward Elgar, Cheltenham, pp. 25-54. Langan, M. and Ostner, I. (1991), "Gender and welfare: towards a comparative framework", in Room, G. (Ed.), Towards a European Welfare State?, SAUS, Bristol, pp. 127-50. Leitner, S. (2003), "Varieties of familialism: the caring function of the family in comparative perspective", European Societies, Vol. 5 No. 4, pp. 353-76. Leitner, S. and Lessenich, S. (2007), "Dependence as dependent variable: conceptualizing and measuring 'de-familization'", in Clasen, J. and Siegel, N.A. (Eds), Investigating Welfare State Change: The "Dependent Variable Problem" in Comparative Analysis, Edward Elgar, Cheltenham, pp. 244-60. Lewis, J. (1992), "Gender and the development of welfare regimes"', Journal of European Social Policy, Vol. 2 No. 3, pp. 159-73. Lewis, J. (1997), "Gender and welfare regimes: further thoughts", Social Politics, Vol. 4 No. 2, pp. 160-77. Lewis, J. and Hobson, B. (1997), "Introduction", in Lewis, J. (Ed.), Lone Mothers in European Welfare Regimes: Shifting Policy Logics, Jessica Kingsley, London, pp. 1-20. Lister, R. (1990), "Women, economic dependency and citizenship" Journal of Social Policy, Vol. 19 No. 4, pp. 445-67. Lister, R. (1994), '"She has other duties': women, citizenship and social security'", in Baldwin, S. and Falkingham, J. (Eds), Social Security and Social Change: New Challenges, Harvester Wheatsheaf, Hemel Hempstead, pp. 31-44. Lister, R. (1995), "Dilemmas in engendering citizenship", Economy and Society, Vol. 24 No. 1, pp. 1-40. Lister, R. (2003), Citizenship: Feminist Perspectives, 2nd ed., Palgrave Macmillan, Basingstoke. McLaughlin, E. and Glendinning, C. (1994), "Paying for care in Europe: is there a feminist approach?", in Hantrais, L. and Mangen, S. (Eds), Family Policy and the Welfare of Women, University of Loughborough, Loughborough, pp. 52-69. O'Connor, J. (1993), "Gender, class and citizenship in the comparative analysis of welfare state regimes: theoretical and methodological issues", British Journal of Sociology, Vol. 44 No. 3, pp. 501-18. O'Connor, J.S., Orloff, A.S. and Shaver, S. (1999), States, Markets, Families: Gender, Liberalism and Social Policy in Australia, Canada, Great Britain and the United States, Cambridge University Press, Cambridge, MA. OECD (2006), Starting Strong II: Early Childhood Education and Care, OECD, Paris. OECD (2007a), Babies and Bosses: Reconciling Work and Family Life - A Synthesis of Findings for OECD Countries, OECD, Paris. OECD (2007b), OECD Family Database. PF12: Childcare Support, available at: www.oecd.org/ dataoecd/52/ll/42004407.pdf OECD (2007c), OECD Family Database. PF14: Quality of Childcare and Early Education Services, available at: www.oecd.org/dataoecd/45/30/37864559.pdf (accessed 13 July 2009). OECD (2009), OECD Family Database. PF11: Enrolment in Day-care andPre-schools, available at: www.oecd.org/dataoecd/46/13/37864698.pdf (accessed 13 July 2009). Orloff, A.S. (1993), "Gender and the social rights of citizenship: the comparative analysis of gender relations and welfare states", American Sociological Review, Vol. 58 No. 3, pp. 303-28. Rauch, D. (2007), "Is there really a Scandinavian social service model? A comparison of childcare and elderlycare in six European countries", Acta Sociologica, Vol. 50 No. 3, pp. 249-69. Sainsbury, D. (1996), Gender, Equality and Welfare States, Cambridge University Press, Cambridge, MA. Defamilisation, dedomestication and care 439 Appendix Procedures for how indicator values and the dedomestication index for childcare service provisions for children under three years were determined. Indicator 1 (time replacement rate): Countries were organised into a ranking order based on average hours of attendance per week, countries 1-5 providing the largest number of hours were given value 3, countries 6-10 were given value 2, and countries 11-15 were given value 1. After this, all values were multiplied by 2 (due to the importance of this indicator). Indicator 2 (availability): Those countries that offered a legal right to childcare for all under 3-year-old children were given value 3, countries where this right was limited to some age or user groups or to certain regions were given value 2, countries that offered no right for the service were given value 1. Indicator 3 (affordability): Countries were organised into a ranking order based on childcare costs for dual earner and sole parent families (childcare costs as percentages of family net income of dual earner families were added to childcare costs of sole parent families and the sum was divided by 2), countries 1-5 where the cost level was lowest were given value 3, countries 6-10 were given value 2, and countries 11-15 were given value 1. Indicator 4 (quality): Countries were organised into a ranking order based on child-staff ratios in formal care provisions for children under three years (if staff levels were different for 0-, 1-, and two year old, these were added together and the sum was divided by 3), countries 1-5 having the lowest child-staff ratio were given value 3, countries 6-10 were given value 2, and countries 11-15 were given value 1. The dedomestication index was considered as the sum of the values of indicators 1-4 multiplied by the take-up rate (and divided by 100) (Table AI). IJSSP 31,7/8 440 Table AI. Data used in determining indicator values for the dedomestication index for childcare provisions for children under three years in 15 countries, ca 2005 Average hours Childcare cost Childcare cost of attendance for dual earner for sole parent Child-staff per week 2006 Right for family, % of family, % of ratio in formal (time childcare family net family net childcare, 0-2 replacement place 2005 income 2004 income 2004 year-old 2005 Country rate) (availability) (affordability) (affordability) (quality) Australia 18 No 9.7 6.6 5.0 Austria 23 No 14.9 9.3 8.7 Belgium 30 Only over 2.5 year old 4.2 3.5 7.0 Canada 32 Only in Quebec 22.0 44.5 6.5 Denmark 34 Only in 87% of municipalities 7.8 8.5 3.3 Finland 35 Yes 7.2 4.1 4.0 France 30 No 11.3 8.8 5.0 Hungary 29 Only working parents of over 0.5 year old 6.5 0.0 6.0 Ireland 25 No 29.2 51.7 4.0 The Netherlands 17 No 11.5 3.0 4.3 Norway 31 No 7.7 -5.8 8.0 Portugal 40 No 4.2 2.0 11.0 Sweden 29 Only working/ studying parents of over 1 year old 6.2 4.8 5.5 UK 18 No 32.7 14.4 3.3 USA 31 No 19.4 6.2 5.0 Mean 28 - 13.0 10.8 5.8 Source: OECD (2006, 2007b, c, 2009) About the author Teppo Kröger is Professor of Social Policy in the Department of Social Sciences and Philosophy at the University of Jyväskylä, Finland. His research interests include comparative welfare studies, childhood, ageing, disability, and care. Teppo Kröger can be contacted at: teppo.kroger@jyu.fi To purchase reprints of this article please e-mail: reprints@emeraldinsight.com Or visit our web site for further details: www.emeraldinsight.com/reprints Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.