European Societies (4) 2003: 353-375 © 2003 Taylor & Francis Ltd ISSN
1461-6696 print 1469-8307 online
R
Routledge
Taylor & Francis Croup
VARIETIES OF FAMI LI ALISM
The caring function of the family in comparative perspective
Sigrid Leitner
Department of Social Policy, University of Goettingen
ABSTRACT: This article focuses on the caring function of the family in child care and elderly care and follows a comparative as well as a gender perspective. It aims at developing a gender-sensitive theoretical concept of familialism which allows to identify real world variations of familialism. Four ideal types of familialism are distinguished and care policies in the EU member states are classified according to them. It can be shown that countries cluster differently for different care policies. In addition, structural gender effects of the four ideal types of familialisrryare reflected upon. The empirical analysis of paid parental leave in nine EU member states gives an example of gendered and de-gendered variants of familialism. Keywords: family policy; child care; elderly care; gender; welfare state typology
1 Introduction: a two-fold starting point
This article ties up to two fields of comparative social policy research: the area of family policy and the tradition of comparing welfare states in terms of typologies. In 1978, Sheila B. Kamerman and Alfred J. Kahn published their seminal work on family policy in fourteen countries. Since then, comparative literature on family policy has developed rapidly, especially within the last decade.1 Overall, the heterogeneity of research designs in the field is striking. Such a 'pluralism' owes much to diverging definitions of family policy. This article focuses on one specific function family policy might be aiming at: the caring function of the family, and it will mainly deal with policies which are directed towards supporting the family in its caring function.
1. For an overview of comparative research on family policies in Europe see Kaufmann (2002: 439-44).
DOI: 10.1080/1461669032000127642
353
EUROPEAN SOCIETIES
Choosing the caring function of the family as subject connects the field of family policy with the paper's second research context: welfare state typologies. In 1999, Gosta Esping-Andersen published a revision of his influential welfare state typology that distinguished 'three worlds of welfare capitalism' (Esping-Andersen 1990). One of the main innovations in the new book was the conceptualization of the social policy function of the family. Esping-Andersen identifies familialistic and de-familializing welfare regimes with regard to the extent to which families are held responsible for their members' welfare. The article draws from this distinction but argues that familialism should be defined in more detail. I suggest to use public policies which explicitly support the family in its caring function as an indicator for familialism and to emphasize the gender perspective of family policy. The overall aim is to develop a gender-sensitive theoretical concept of familialism which allows to identify real world variations of familialism. Thus, the article can be understood as a critical statement to Esping-Andersen's typology of welfare regimes.';
Section 2 will show that the caring function of the family has recently become prominent in social policy. While the need for care is rising, the availability of formal care provision as well as the reliability of family care is increasingly questioned. This leaves social policy with the problem of unmet care needs. Especially in times with tight social budgets, it seems to be a reasonable strategy for welfare states to strengthen the family in its basic caring role. The extent to which the caring function of the family is promoted determines whether a welfare regime is conceptualized as a familialistic or a de-familializing system. Section 3 of the paper identifies four types of policies according to the provision of care for the young, the disabled and the elderly: a de-familializing type, an implicitly familialistic, an explicitly familialistic, and an optionally familialistic type)1 It, secondly, focuses on an empirical analysis of two different fields of care policy: child care and elderly care. The EU member states' policies are compared and classified according to the four types of familialism. It will be shown that countries cluster differently for different care policies. Thereafter, in Section 4, it will be elaborated how familialisms differ according to their dimension of gender (inequality. A gendered model of familialism which is connected to the male breadwinner/female homemaker ideology and a de-gendered model of familialism which enables gender equality in the field of family care can be distinguished. Taking paid parental leave in nine European countries as an example, gender-sensitive criteria will be developed to engender familialism. The paper concludes with some reflections on the perspectives which the varieties of familialism offer for further research.
Varieties of familialism LE1TNER
2 The caring function of the family
With the development of the male breadwinner model and its structural embodiment in most welfare states after the Second World War,2 a family care regime was institutionalized that enabled (and at the same time envisioned) women to care for children, the aged and the handicapped at home (cf. Knijn/Kremer 1997: 329). With the increasing labour market attachment of women, the family's caring responsibility has come under pressure, especially in the field of child care. The employment rates of mothers with a child under 6 are rising rapidly:
While the employment rate of mothers is much lower than that for fathers (54%, on average for the countries shown, as compared with well over 90% for fathers), the gap has been closing quite rapidly, at around one percentage point per year over the past decade.
(OECD 2001: 133)
Combining employment and child care is difficult given nuclear family structures with both parents employed (at least part time) and inflexible working conditions. The mismatch between both 'worlds', family and employment, generates a need for support with regard to child care. Consequently, parents can be supported by institutional^ child care provision and/or by work-related policies that 'enable them to care for their young children without losing complete contact with the labour market' (OECD 1995: 171).Asa matter of fact, the old male breadwinner model which guaranteed the availability of the female family carer is no longer fijting the employment pattern of women nor the preferences of families for dual earner models (cf. OECD 2001: 135).
If we look at the care provided for tie elderly throughout Europe, the family is usually the primary care provider (cf. Helmer 1993: 442). But the need for elderly care is rising due to population ageing. Especially the 80+ population with its high disability rate is growing considerably (OECD 1996: 15). Today, a great deal of family care is provided by elderly wives and husbands, but there is a growth in the proportion of elderly people living alone, among whom are many of the very oldest. Alternatively, family care is provided by female relatives in the 45-64 age group. Unfortunately, 'the ratio of women in middle age to more elderly people has shrunk considerably and will continue to do so' (OECD 1996: 19). Whether future care needs can be met by family care thus depends on the availability of family carers in younger age groups. Those potential informal care givers will be confronted with the problem of combining work and care giving, especially if the rising labour market participation I 2. For variations of the breadwinner model in Europe see Lewis and Ostner (1995).
354
355
EUROPEAN SOCIETIES
Varieties of famiIialism
LEITNER
of women is considered. With rising levels of need for elderly-care and shrinking state budgets, paying for family care can be a suitable solution for the welfare state:
Provided that the level of financial support given is not high and does not approximate to market rates for paid care labour, then informal care represents a cheaper solution to long term care needs than any other.
(McLaughlin and Glendinning 1994: 53)
Moreover, payments for informal care givers are important not only in terms of resources but also with regard to 'the willingness and ability of family members to assume responsibility for relatives' (Hantrais and Letablier 1996: 165). Apart from payments for care, informal care givers can be supported by complementary home help services and work related policies like care leave regulations or working time reductions (cf. Dumon 1990: 185-199; 1991: 94-96).
During the last 15—20 years, European welfare states increasingly have taken the value of unpaid care work into consideration, at least partially. With regard to the care of young children, most European countries have installed or extended parental leave regulations, many in connection with cash benefits for the carer. A trend towards low level, conditional payments has emerged. In addition, the quality of social rights granted to the care provider has been improved, e.g., protective labcJur market legislation for carers and pension recognition of care periods. As far as caring for elderly, ill and disabled people is concerned, European welfare states increasingly subsidise private informal care arrangements (cf. Daly 1997: 138-142). Notwithstanding these general trends, welfare states have treated the family quite differently according to its social policy function of care provision. Within the 'welfare triangle' of state, market and private households, caring responsibilities can be distributed in various ways. Some welfare states explicitly attempted to strengthen social responsibility among family members, others strongly supported the extension of public care services, several states hoped for the market to provide suitable services, and others did not intervene at all or tried a mixed strategy.
With reference to Gosta Esping-Andersen (1999: 51) familialistic and de-familializing welfare states can be distinguished:
A familialistic system, [. . .], is one in which public policy assumes - indeed insists - that households must carry the principal responsibility for their members' welfare. A de-familializing regime is one which seeks to unburden the household and diminish individuals' welfare dependence on kinship.
In order to determine the degree of de-familialization, Esping-Andersen distinguishes between de-familialization through public social services
and de-familialization through market driven service provision. Whereas the Scandinavian welfare regimes are characterized by the first, the liberal welfare regimes rely on the latter. In this respect, only the conservative welfare regimes resemble familialistic systems since they lack both service provision by the state and the market. Two critical remarks should be made at this point: Firstly, market driven care provision makes de-familialization a class biased issue since either only the better-off can afford to be de-familializedor the quality of de-familialization varies considerably by income. Secondly, even in Scandinavian welfare regimes the family is the most important agent of care provision.
Undoubtedly, it is not sufficient to define familialistic welfare regimes by the absence of public and market provision of services. Therefore, a second kind of analytical differentiation is needed. Esping-Andersen determines the familialistic character of a welfare regime by indicators that measure the intensity of familial welfare responsibilities: the percentage of elderly living with children, the percentage of the unemployed youth living with parents, and the number of women's weekly unpaid hours spent on domestic obligations. Here, the Southern European countries (and Japan) stand out for their familialism and the Nordic welfare regimes partly confirm their de-familialized status. All in all, the indicators used seem to be inadequate for identifying variations of familialism since they measure the outcome of familialistic structures but do not say anything about these structures nor about the relations between structures and outcome. These outcome-indicators of familialism should, thus, be replaced by a more predicative instrument of measurement. If familialism means that public policy wants private households to secure the welfare of their members, the ways and means used to enforce this goal should be at the centre of analysis. Thus, service provision would be one part of the story, policies which support the caring function of the family would be the other.
So, we can distinguish between welfare regimes that rely on and actively support the family as the main source of care provision and welfare regimes that attempt to relieve the family from caring responsibilities. Whereas the first type of policy puts the family at the centre of care provision, the second one either socializes or 'marketizes' the caring function of the family through public social services or market-driven care provision. Since welfare states do not resemble unambiguously the categories of familialistic or de-familializing systems, I will distinguish in the following between single measures of family policy rather than between welfare states as such.
356
357
EUROPEAN SOCIETIES
Varieties of familialism
LEITNER
3 Familialization versus de-familialization
Familialistic policies not only oblige (and at the same time: enable) the family to meet the care needs of its members, they also enforce the dependence of people in need of care on their family. In this perspective, de-familialization means not only taking away care responsibilities from the family. De-familialization also reduces the extent to which the satisfaction of individual care needs is dependent on the individual's relation to the family (cf. McLaughlin and Glendinning 1994: 65). However, this paper focuses on the perspective of the care givers rather than on those in need of care. In the following subsections, the dimensions of familialization and de-familialization will be developed in more detail to distinguish ideal typical (3.1) as well as empirical (3.2) variations of familialism.
3.1 Care provision by the family or by the state/market
Policies which actively aim at strengthening the family in its caring function can be found in the field of child care as well as in the field of care for the elderly and handicapped. We can distinguish (1) time rights (such as parental leave and care leave), (2) direct and indirect transfers for caring (such as cash benefits and tax reductions) and (3) social rights attached to care giving like individual pension rights or the (partial) inclusion in other social security schemes or derived rights for non-employed wives (husbands). These familialistic policies can be contrasted with de-familializing policies which aim at unburdening the family in its caring function, like the public provision of child care and/or social services or the (public subsidy of) care provision through the market. If we suppose variation, i.e., strong and weak expressions within both kinds of family policies and if welfare states may contain combinations of familialistic and de-familializing family policies, the following matrix can be drawn (see Table 1).
Four ideal types of familialism emerge from this matrix. The explicit familialism not only strengthens the family in caring for children, the handicapped and the elderly through familialistic policies: It also lacks the provision of any alternative to family care; This lack in public and market
TABLE 1.  Combinations of strong/weak familialization and strong/weak de-familialization
Familialization	De-familialization	
	Strong	Weak
Strong	Optional familialism	Explicit familialism
Weak	De-familialism	Implicit familialism
driven care provision together with strong familialization explicitly enforces the caring function of the family.
Within the optional familialism services as well as supportive care policies are provided. Thus, the caring family is strengthened but is also given the option to be (partly) unburdened from caring responsibilities. In both the explicit and the optional familialism 'the right to time to care' (Knijn and Kremer 1997) is honoured by familialistic policies which enable the family to care. Yet, only in the optional familialism is the family's right to care not equated with the family's obligation to care.
The implicit familialism neither offers de-familialization nor actively supports the caring function of the family through any kind of familialistic policy. Nevertheless, the family will be the primary caretaker in these welfare regimes since there are no alternatives at hand. This type, therefore, relies implicitly upon the family when it comes to care issues.
Finally, de-familialism would be characterized by strong de-familialization due to the state or market provision of care services and weak familialization. Thus, family carers are (partly) unburdened but the family's right to care is not honoured.
The different types of familialism are analytically useful on three different levels: They could help to classify (1) the intentions of care policies, articulated for example by political parties or the government, as well as (2) the empirical effects of care policies on care arrangements. However, the focus of my analysis will be (3) on the structural implications of care policies. In the following subsection 3.2, the 15 EU member states will be grouped according to the above matrix. The aim is to demonstrate thatfwelfare states resemble different family policy profiles (and therefore cluster differently) for different care policies-lSince detailed and compre- ■ hensive comparative data on care policies is only selectively available, I will focus on central policies for child care and for elderly care. Firstly, cash benefits for child care in terms of parental leave regulations are contrasted |with institutional child care provision to determine country clusters in the field of child care. Secondly, direct transfers for the care of frail elderly people in terms of payments for care are contrasted with the provision of home help services in order to determine country clusters in the field of elderly care.
3.2 Varieties of familialism in child care and elderly care in Europe
In order to look at the empirical evidence of the above developed types of familialism in the field of child care, indicators for the dimensions of familialization and de-familialization have to be chosen. One of the main policies to support the caring function of the family with respect to child care is the
358
359
EUROPEAN SOCIETIES
provision of cash benefits during parental leave. Paid parental leave enables parents to suspend their labour market participation in order to care for their young children or simply provides a financial compensation for the carer. Thus, the existence (respectively: the absence) of regulations for paid parental leave will be taken as an indicator for the dimension of strong (respectively: weak) familiarization. This is a structural indicator since parental leave regulations provide (or deny) structural options for parents and their care decisions. Therefore, the indicator holds no information about the use of these structures or their importance for family care decisions.
Consequently, the indicator fox strong (respectively: weak) de-familial-ization has to correspond with the group of the cared for persons which are very young children. Thusf formal child care for those children will be focused on. More specifically, the percentage of children under 3 years of age who are in formal child care will indicate the dimension of strong and weak de-familialization. Though these data are about the usage of formal child care, they can be used as an indicator for the structure of child care services since the offer provided is completely used, i.e., the demand exceeds the offer.
The empirical comparison demonstrates that the EU member states differ considerably with respect to our two basic dimensions of familializ-ation and de-familialization (see Table 2). On the one hand, only five countries (Greece, Ireland, Portugal, Spain, United Kingdom) do not provide cash benefits for child care in terms of paid parental leave. The other ten EU members support the family's child care provision through paid parental leave.3 On the other hand, only two nations (Denmark and Sweden) provide a high degree of formal child care for children under three years of age (64 and 48%). Four nations range between 29 and 38% (France, Belgium, Ireland, United Kingdom), which suggests that there is a choice between family care and formal care for at least approximately one-third of the parents. Finland's score (22%) is a little lower but still differs significantly from the remaining eight EU member states who show a very poor coverage of formal child care (3-12%).
These differences between the EU member states leave us with four country clusters which represent four types of familialism in the field of child care: (1) the optional familialism with widespread formal child care and payments for child care within the family, (2) the explicit familialism with poor rates of formal child care but payments for child care within the family, (3) the implicit familialism with poor rates of formal child care as well as a lack of cash support for child care within the family, and (4) de-familialism with widespread formal child care but a lack of payments for child care (see Table 3).
3. The Netherlands does not have statutory regulations on cash benefits for child care but collectively bargained benefits are widespread.
Varieties of familialism
TABLE 2. Paid parental leave and formal child care for children under 3 years of member states
	Paid parental leave	% of children under 3 in formal child care
Denmark	Flat-rate payment for 10 weeks'3, thereafter	64
	reduced flat-rate for 1 year per parent3	
Sweden	75% of former wage for 1 yeare, thereafter	48
	flat-rate for 3 months	
Belgium	Flat-rate payment for 1 year3	30
France	Flat-rate payment for 3 yearsd	29
Finland	66% of former wage for 6,5 months,	22
	thereafter flat-rate payment0 for 2 years	
	and 2 months	
Germany	Flat-rate payment for 2 years'	10
Italy	30% of former wage for 6 months	6
The Netherlands	Varying payment (collectively bargained)	6
	for 6 months	
Austria	Flat-rate payment for 3 years	4
Luxembourg	Universal flat-rate payment for 2 years	2
Portugal	None	12
Spain	None	5
Greece	None	3
Ireland	None	38
UK	None	34
a The Belgian 'career break' is not directly comparable with parental leave regulations in other countries as
it does not particularly refer to child care but might be used for a child care break. Similarily, the extended
leave period in Denmark is part of a leave scheme that allows employees to take paid leave for training,
child care or other reasons, such as a sabbatical break from work.
b Mostly increased up to 100% of former wage through collective bargaining.
c Payable only if the child does not use public child care.
d Not available for the first child.
eThe benefit has a maximum ceiling.
'The benefit is income tested.
Sources: Data on formal child care comprises publicly and privately funded care arrangements and is from OECD (2001: 144), supplemented with Lohkamp-Himmighofen (1993: 343) for Luxembourg. Data on parental leave is from Bundesministerium für Familie, Senioren, Frauen und Jugend (1998) and OECD (1995: 175-8).
TABLE 3.  Classification of countries in the field of child care
	Formal child care	
	Widespread	Poor
Payment for child care	(1) Belgium, Denmark,	(2) Austria, Germany, Italy,
	(Finland)*, France, Sweden	Luxembourg, Netherlands
No payment for child care	(4) Ireland, United Kingdom	(3) Greece, Portugal, Spain
(1) Optional familialism, (2) explicit familialism, (3) implicit familialism, (4) de-familialism. * Finland is kind of a borderline case between optional and explicit familialism.
360
361
EUROPEAN SOCIETIES
Varieties of f am i I i a I i sm
LEITNER
The resulting country clusters differ from the 'three worlds approach' of Esping-Andersen. His de-familializing Scandinavian cluster is firstly classified as familialistic system (optional familialism) and secondly complemented by Belgium and France. The latter two countries were given a special status within the conservative welfare regime by Esping-Andersen himself because of their strong tradition of formal child care services. In my analysis this tradition makes them share one country cluster with the Scandinavian welfare states in the field of child care policies. Esping-Andersen's conservative cluster is further split into explicitly and implicitly familialistic systems with the Southern welfare states (except Italy) in the implicit familialism type. This is compatible with the special status of the Southern nations within his measurement of the family's responsibility for its members. Last not least, Ireland and the United Kingdom which are in the liberal cluster of Esping-Andersen, are found within the de-familialism type.
Using a modified analytical framework enabled us to identify varieties of familialism and led to country clusters in the field of child care that seem to catch similarities and differences between the EU member states more adequately than before. But\ child care policies for very young children are only one of the various policies that aim at supporting the family in its caring function.! Because of the heterogeneity of care policies it is very likely that countries will cluster differently for separate policies.
Moving on to the field of family care for frail elderly people, strong (respectively: weak) familialization will be indicated by the availability (respectively: the lack) of direct transfers for elderly care through cash payments to the family. The strength or weakness of de-familialization/ could be assessed by the percentage of elderly people in need of care who live in institutional care arrangements.. Unfortunately, the total numbers of elderly people in need of care are widely unknown. If we estimate that the percentage of elderly people who need care in relation to the total number of elderly people is within the same range in all European countries, we could also use the percentage of elderly people who live in institutional care arrangements as a whole. Unfortunately, this indicator shows hardly any variation between the EU member states,4 although the extent of formal care provision for the elderly differs considerably within Europe. But the general trend in care policy leads towards mobile care services to enable elderly people to stay in their familiar surroundings as long as possible. At the same time, home help services unburden the family from its caring responsibilities. Thus, the percentage*of people aged over 65 who receive home help services seems to be a useful indicator
4. The percentages are very small anyway (Helmer 1993: 445) since usually only the frailest are cared for in institutional settings.
for de-familialization.5 Again, both indicators are structural: benefits for elderly care provide structural options for (potential) family carers but cannot tell us anything about the use of these structures or their importance for family care decisions. Data on the usage of home help services are also data about the supply of these services since the supply provided is completely used.
The empirical evidence across the EU member states (see Table 4) shows variation on both dimensions of familialism. On the one hand, only five nations (Greece, Italy,6 The Netherlands, Portugal, Spain) do not provide any kind of payments for elderly care within the family. Within the remaining ten welfare states payments for elderly care vary considerably: some do have cash benefits paid to the carer, others pay the carer only indirectly via the cared for person, the level of the benefit varies, and so on. The heterogeneity of regulations is striking. Together with the lack of coherent and comprehensive comparative data on this very young policy, this makes it difficult to classify countries in more detail.7 On the other hand, the percentage of elderly over 65 receiving home help services is rather restricted in most cases8 (1-9%). Only Finland, Denmark (24 and 22%) and - with a small gap - also Sweden (16%) differ considerably.
However, the three Scandinavian countries show a considerable degree of home help coverage as well as direct transfers for elderly care and thus in the field of elderly care resemble the optional familialism (1). The J countries with poor home help coverage differ firstly according to the existence or the absence of payments for elderly care. Five of them do not (Greece, The Netherlands, Portugal, Spain) or hardly (Italy) provide any such payments. They therefore resemble the implicit familialism (3) of elderly care. In the remaining countries the combination of direct transfers for elderly care with a poor home help coverage points out the explicit familialism (2) in the field of elderly care (see Table 5).
Comparing these results to Esping-Andersen's clusters, the Scandinavian countries are once again classified as familialistic systems (optional familialism). The conservative cluster is again split into explicit and implicit familialism with Austria, Belgium, France, Germany and Luxembourg resembling the first and the Southern European nations together with The Netherlands resembling the latter. (Note that Italy changed
5. As a matter of fact home help services not only relieve the family from caring duties but also may help the family to maintain its caring function. So, to some unknown extent this indicator represents also familialization, since home help services are not in all cases simply a substitute for family care but are also an incentive for family care.
6. Italy has established pilot projects in four regions but there is no national programme yet.
7. For a typology of payments for care see Ungerson (1997: 366-73).
8. For Luxembourg the data sources provide no account, but it can be assumed that Luxembourg has a rather small rate (cf. Dienel 1993: 287).
362
363
EUROPEAN SOCIETIES
TABLE 4.  Payment for elderly care and home help services for the elderly in the EU member states
Payment for the care of frail elderly people
% of elderly over 65 receiving home help services
Finland
Denmark
Sweden
UK
France
Belgium
Austria
Ireland Germany
Luxembourg
The Netherlands
Spain
Greece
Italy
Portugal
A home care allowance is paid to the carer.
Caring relatives are paid by the community.
Cash allowances are made to elderly people to enable them to pay a family carer.
A flat-rate allowance is paid to the carer. An income-tested dependency allowance can be received by frail elderly people. It is paid directly to the elderly person and can be spent either for professional or informal care, e.g. for compensating a family member (but not the spouse). A career break for caring can be taken and is compensated by a flat-rate payment.
A care allowance is paid to elderly persons requiring care, graduated according to the degree of care needed. The care allowance can be spent for compensating a family member for supplying help.
A means-tested allowance is paid to the carer.
A care allowance is paid to the person receiving family care, graded according to the degree of care needed. The care giver is not paid directly but (hopefully) by the cared for person. Frail elderly being cared for at home, receive either professional care or a care allowance to remunerate the carer. ■
None None None
Four regions adopted payment for care: a means-tested subsidy to people opting to keep a dependent elderly person at home with them.; None
24 22 16
no account
Sources-. Data on home help services are from Esping-Andersen (1999: 71), supplemented by Anttonen and Sipilä (1996: 92) for Greece. Data on payment for care are from European Commission (2000), Lewis (1998) and OECD (1996).
Varieties of familialism
TABLE 5.  Classification of countries in the field of elderly care
LEITNER
Home help coverage
Widespread
Poor
Payment for elderly care No payment for elderly care
(1) Denmark, Finland, Sweden
(2) Austria, Belgium, France, Germany, Ireland, Luxembourg, United Kingdom
(3) Greece, Italy, The Netherlands, Portugal, Spain
(1) Optional familialism, (2) explicit familialism, (3) implicit familialism.
from explicit familialism in child care to implicit familialism in elderly care and The Netherlands vice versa, whereas Belgium and France changed from optional familialism in child care to explicit familialism in elderly care.) Esping-Andersen's liberal cluster is found on the weak side of de-familialization and on the strong side of familialism this time. Thus, Ireland and the United Kingdom changed from de-familialism in child care to the explicit familialism in elderly care.
As this first approach to conceptualize empirical variants of familialism demonstrates, countries cluster differently for different care policies. Since there are still other care policies to consider, it will be difficult if not impossible to subsume each welfare state under one exclusive category of familialism. Countries will also cluster differently at different points in time which would demand a historical approach to catch the dynamics of national variants of familialism. Including the gender dimension into the analysis makes things even more complicated.
4 Gendered versus de-gendered familialism
Since family care is usually provided by women (wives, daughters, daughters-in-law), supporting the caring function of the family represents a gendered subject per se. In order to complete the theoretical considerations about the varieties of familialism, the question of gendered care giving has to be conceptionalized.
4.1 Gender and the caring function of the family
Re-examining the four types of familialism identified in Section 3.1, questions of gender equality may be raised. As a matter of fact, the implicit familialism with its weak direct support for the family's caring function and
364
365
EUROPEAN SOCIETIES
its lack in service provision does not directly intervene in gender relations. It resembles a laissez-faire model of family policy where the state seeks to exercise no influence on the family at all (cf. Harding 1996: 183-6). Nevertheless, this type affects gender relations since it simply reproduces and thus confirms the status quo of gendered care provision within the family.
Similarily, de-familializing care policies per se do not directly intervene in gender relations.9 But, since they relieve the family from care provision, they provide options for family carers. De-familializing care policies thus weaken breadwinner models in which one person is employed and the other person is available for family care and they represent an important structural condition for gender equality in the labour market. The effect will be most widespread in the field of child care where formal care arrangements enable parents (mostly: mothers) to participate in the labour market. In the field of elderly care the provision of formal care wilrtiave a similar effect only for the younger age groups of carers for whom labour market participation would still be an option. Both,; de-familialism and the 4 optional familialism; provide de-familializing care policies. Whereas de-familialism provides family carers exactly with the option of labour market participation, the optional familialism additionally supports family care directly as an alternative to formal care arrangements. The optional familialism, therefore, provides a better opportunity to choose between family care and formal care than de-familialism which does not actively enable family care.10 ,
In contrast to de-familializing care policies, familialistic care policies i directly regulate gender relations. On the one hand, they seem to automatically enforce traditional gender roles since they aim at maintaining and strengthening the family's caring function. As family care is in most cases women's work, these policies will not only support the family as such in its caring function but will also strengthen the caring role of women and, thus, reproduce the gendered division of family care. On the other hand, this does not necessarily have to be the case. To avoid the dilemma11 of, validating
9. They may have direct effects on gendered care provision within the sector of formal care though. If formal care is mostly provided by female carers, the gender division of care giving is enforced.
10. Similarily, Hochschild distinguishes between a 'cold modern' and a 'warm modern' ideal model of care. The first one is 'represented by impersonal institutional care in year-round ten-hour day care and old-age homes' whereas in the second one 'institutions provide some care for young and elderly' while the family provides care as well (1995: 332).
11. This was called die contemporary variant of the Wollstonecraft dilemma: 'We are torn between wanting to validate and support, through some form of income maintenance provision, the caring work for which women still take the responsibility in the 'private' sphere and to liberate them from this responsibility so that they can achieve economic and political autonomy in the public sphere.' (Lister 1994 cited in Knijn/Kremer 1997: 350)
Varieties of fpmilialism LEITNER
family care and strengthening the independence of family carers while at the same time engendering family care, familialistic policies could also contain incentives to ensure that care provision is shared on equal terms among male and female family members. Therefore, the explicit as well as the optional familialism hold both options: a gendered or a de-gendered variant. In the following I will focus on these two types of familialism since they are the interesting cases with regard to the political regulation of gender relations.
To assess the extent of gender (inequality in familialistic care policies, an analytical framework is needed which connects the different kinds of gender discrimination to social policy. For this purpose I will draw from earlier works (cf. Leitner 1999: 19-36; 2001) and distinguish three forms of sex differentiation:
1) The distinction between man and woman due to biological characteristics
This biological differentiation uses biologically determined differences to legitimize different treatment of men and women. It is institutionalized within social policy by implementing regulations that refer explicitly to biological sex differences. The discriminating mechanism attached to biological sex differentiation is characterized by the implicit connection of biological sex differences with a gendered division of labour: if, for example, benefits for child care or elderly care are available only for women, all caring men are excluded from the provision of benefits. Thus, familialistic policies discriminate if they explicitly assign family care to women (or men) only.
2) The distinction between man and woman due to the gendered division of labour
This differentiation due to the gendered division of labour results from the distinction between a "public sphere of paid employment (in which men usually are) and a private sphere of unpaid work (for women). It is implemented within social policy by separate schemes for employment on the one hand and family work on the other. The differentiation is non-discriminatory if both kinds of schemes provide equal protection with regard to benefit levels and quality of social rights. Thus, familialistic policies discriminate due to the gendered division of labour if family care is devalued in relation to employment. Furthermore, the gendered division of labour per se restricts the options of men and women to move between the public and the private sphere. Thus, familialistic policies which differentiate along the gendered division of labour will be even more discriminating if no choices are provided to (re-)move from family care to employment.
366
367
EUROPEAN SOCIETIES
3) The distinction between 'male' and 'female' sexuality The distinction between 'male' and 'female' sexual desire - men desire women and women desire men as sexual partners - reflects the norm of heterosexuality. Heterosexuality is implemented within social policy by regulations that refer to (married) couples. Familialistic policies discriminate against alternative family care arrangements if they do not provide comparable benefits.
Ideal-typically, gendered familialism shows familialistic policies that assign family care to women (or men) only, devalue family care in relation to employment, do not provide choices to (re-)move from family care to employment, and focus on (married) heterosexual couple families. Contrarily, de-gendered familialism shows familialistic policies that do riot refer to biological sex differences, validate family care, enable financial independence of the carer, provide choices to move between family care and employment, and provide comparable benefits for different family care arrangements.
As has been argued above, the explicit as well as the optional familialism have the potential to aim either at a gendered or at a de-gendered model of family care; In addition, it will be important to consider not only direct but also indirect structures of gender discrimination. For example, the gendered variant of the optional familialism will be less restrictive than the gendered variant of the explicit familialism since the optional familialism provides also de-familializing policies which hold the option of alternatives to family care. Moreover, the overall character - gendered or de-gendered - of a welfare regime has to be taken into consideration. E.g. a 'male breadwinner model' (cf. Lewis 1992; Lewis and Ostner 1995) indirectly supports the family's caring function through derived social rights (e.g., sickness insurance, survivor benefits) or tax benefits for non-working wives. It can, therefore, either strengthen a gendered form of familialism or be a hindrance for a de-gendered model of familialism. Contrarily, dual earner models with individualized social rights would strengthen a de-gendered model of familialism.
In the following sub-section, I will focus onparental leave policies and their attached cash benefits for child care. The aim is to provide an example for a gender sensitive analysis of familialism and to classify the different policies according to the gendered or de-gendered model of gender relations. Only those nine countries are part of the analysis which resemble either the optional or the explicit type of familialism in the field of child care (see Section 3.2).12
12. The Netherlands will be skipped from the analysis because it has no statutory provision of parental benefits.
Varieties of familialism LEITNER
4.2 Gendered and de-gendered paid parental leave policies in Europe
Following the above-presented dimensions of gender discrimination in family care, the extent of gender (in)equality of the nine paid parental leave policies will be assessed (see Table 6).
Discrimination due to biological sex differentiation
None of the policies discriminate by sex. All of the countries allow for shared parenting since parental leave periods are transferable between the parents. Thus, "nocpuntry has explicit gender norms. But only four countries 'punish' the unequal engagement of parents. Belgium and Denmark grant parental leave periods as individual rights for mothers and fathers. If one parent does not make use of his/her right, it lapses and cannot be passed on to the other parent. Similarily, Sweden reserves one month of the whole leave period for each parent. The rest of the parental leave period can - but does not have to - be shared among the parents, Austria presents a special case in this respect: the maximum duration of the parental benefit amounts to three years per child, but is cut to thirty months if only one of the parents qualifies as primary caretaker, i.e., only the other parent is entitled to the remaining six months of the benefit.
Discrimination due to the devaluation of family care Parental benefits are generally low. Only Denmark and Sweden - and to some extent also Finland - provide wage replacement rates that allow a certain extent of financial independence. With the exemption of Italy which has a very low wage replacement rate (30%), all the other countries pay a flat-rate between €271 and €456. Therefore, six of the countries provide benefits below the poverty line which devalues family care in relation to employment and enforces the financial dependence of family carers on a breadwinner (or on welfare programmes). Finland presents a special case since it has a two-fold regulation: a rather short period of the leave is paid rather well (66% of the former wage) whereas the longer period of the leave has a low flat-rate payment attached.
Connected to the question of financial independence is the duration of paid parental leave in relation to parental leave as a whole. In all countries but Germany and Italy, parental benefits are paid at least throughout the whole leave period. In Germany two out of three years and in Italy six months out of three years are covered by parental benefits which decreases the financial independence of parents who take the whole leave period. Due to the lack of formal child care in both countries this concerns
368
369
EUROPEAN SOCIETIES
especially (single) parents who can not rely on a breadwinner. Thus, the question of financial independence is also closely related to the question of equal treatment of different family care arrangements. Insufficient benefits discriminate against one parent families as well as low income families.;
Discrimination due to labour market disincentives for carers In contrast, in Austria the duration of the parental benefit exceeds the leave period which leads to the exlusion from protective labour market legislation attached to parental leave if the carer does not return to his/ her job after two years. This is clearly a disincentive for the labour market participation of carers. Similarily, Luxembourg provides a universal parental benefit for two years, but the eligibility of parental leave is very limited. Finally, Sweden provides a low flat-rate parental benefit during three months after the end of the parental leave period.
The labour market attachment of family carers is further affected by the general duration of parental leave. Like other times of absence from the labour market, 'long leave periods may create difficulties in returning to the job if there have been significant changes in the technological and organizational context of the firm in the meantime' (OECD 1995: 188). Only Belgium, Denmark and Sweden show relatively short leave periods. Carers can take a maximum leave of one year in these countries. The other six countries provide parental leave between two ■ and three years and, therefore, set labour market disincentives for carers.
The comparison shows that Denmark and Sweden have the most de-gendered profiles in paid parental leave throughout the European Union: short leave periods, wage replacement rates that allow for financial independence - though on a rather low level - and incentives to share child care among the parents characterize their policies. Belgium, too, approaches the de-gendered model of familialism but fails with regard to the insufficient level of the parental benefit.
On the contrary, France, Germany, Italy and Luxembourg clearly resemble a gendered familialism: long leave periods together with insufficient parental benefits and a lack of incentives for shared parenting structurally enforce the caring role as well as the financial dependence of mothers. Austria comes very close to this group of countries, but differs with respect to its incentive for shared parenting. Finland, finally, shows gendered elements due to a rather long leave period and the lack of incentives for shared parenting. But, it provides a comparably high parental benefit for the first (shorter) part of the leave.
Table 7 shows that only two of the optional familialistic countries have de-gendered paid parental leave policies and not all of the explicit
Varieties of familialism LEITNER
TABLE 6.   Parental leave and parental benefits in nine European countries
	Parental leave		Parental benefit	
	Duration	Transferability	Duration	Monthly amount
Belgium	1 year per parent	individual right	throughout the	Flat-rate (€271)
			whole leave period	
Denmark	10 weeks	Family right	Throughout the	Flat-ratea for 10
			whole leave period	weeks (minimum:
				€1.513)
	plus 1 year per	Individual right		Reduced flat-rate
	parent			thereafter (€909)
Sweden	1 year	Family right and	1 year	75% of wage
		individual rightd		
			plus 3 months	Flat-rate
				thereafter(€199)
Austria	2 years	Family right and	3 years'3	Flat-rate (€436)
		individual right3		
Finland	6.5 months	Family right	Throughout the	66% of wageb for
	plus 2 years, 2		whole leave period	6,5 months (min.:
	months			€252)
				Flat-ratecthereafter
				(€252)
France	3 years'	Family right	Throughout the	Flat-rate (€456)
			whole leave period	
Italy	3 years	Family right	6 months	30% of wage
Germany	3 years	Family right	2 years	Flat-rate (income
				tested, maximum:
				€307)
Luxembourg	2 yearss	Family right	2 years	Flat-rate (€408)
a Mostly increased up to 100% of former wage through collective bargaining.
b Parents not gainfully employed receive a flat-rate benefit.
cThe benefit is only paid if the child does not use public child care.
d One month is reserved for each parent.
e If the parental leave is not shared, the parental benefit is paid for 30 months only. f Parental leave is not available for the first child. 8The availability of parental leave is very limited.
Sources: Bundesministerium für Familie, Senioren, Frauen und Jugend (1998) and OECD (1995: 175-8).
TABLE 7.  Gendered and de-gendered variations of familialism
	Gendered familialism	Mixed cases	De-gendered familialism
Optional familialism	France	Finland, Belgium	Denmark, Sweden
Explicit familialism	Germany, Italy,	Austria	
	Luxembourg		
370
371
EUROPEAN SOCIETIES
familialistic countries rely on a completely gendered model of familialism. Due to relatively widespread formal child care in France and Belgium, family carers are provided with an alternative to gendered family care. But if parents opt for family care, it will be gendered, in France probably more than in Belgium. In Finland, the alternative to family care is more restricted since the coverage with formal child care is not as well developed as in the other optional familialistic countries (see Section 3.2). Thus^ parents have less choice between employment or gendered family care! Finally, the incentive for shared parenting distinguishes Austria from a gendered familialism. But, the de-gendering effects of this incentive have to be questioned since the insufficient parental benefit is a structural incentive for female child care at home; It is very likely that gendered and de-gendered country profiles of familialism differ across care policies and over time.
5 Summary and research perspectives
This article focuses on the caring function of the family and (re)defines the concept of familialism by drawing from Esping-Andersen's distinction between familialistic and de-familializing welfare regimes and by introducing a more predictable indicator for familialization: policies that actively support the family in its caring functions Contrasting familialization and de-familialization leads to four ideal types of familialism: optional, explicit, and implicit familialism, and de-familialism. In a second step, empirical indicators for familialization and de-familialization in the fields of child care and elderly care are defined. The empirical analysis of the fifteen EU member states leads to country clusters that differ considerably - and for good reasons - from the 'three worlds' typology of Esping-Andersen. Moreover, countries cluster differently for different care policies. Both results call for a more differentiated approach in comparative welfare state research.
The article further develops a concept to include the gender dimension into the analysis of familialism. Applying gender sensitive criteria to paid parental leave policies in nine European countries leads to an assessment of the extent of structural gender (in)equality. The analysis shows gendered and de-gendered variants of familialism among the countries. More conceptual work will have to be done to analyse also other care policies with regard to their extent of structural gender (in)equality. A more comprehensive approach would also have to include other analytical levels like the political intentions and the empirical outcomes of care policies.
Last but not least, a dynamic comparative perspective would be most interesting to analyse the development of familialism over time. Further
Varieties of familialism LEITNER
research would be especially beneficial in those countries which historically move between different types of familialism.
Acknowledgements
I am grateful for comments and suggestions by the participants of the fellows' colloquium at the Centre for European and North American Studies in Goettingen who discussed an early version of this paper. I also thank Ilona Ostner, Klaus Stolz and two anonymous referees for their constructive criticism.
References
Anttonen, Anneli and Sipilä, Jorma (1996) 'European social care services: is it possible to identify models?', Journal of European Social Policy 6(2): 87-100.
Bundesministerium für Familie, Senioren, Frauen und Jugend (ed.) (1998) Ubersicht über die gesetzlichen Maßnahmen in den EU-Ländern bei Erziehung von Kleinkindern, Stuttgart: Kohlhammer.
Daly, Mary (1997) 'Welfare states under pressure: cash benefits in european welfare states over the last ten years'', Journal of European Social Policy 7(2): 129^-6.
Dienel, Christiane (1993) 'Luxemburg', in Erika Neubauer, Christiane Dienel and Marlene Lohkamp-Himminghofen (eds), Zwölf Wege der Familienpolitik in der Europäischen Gemeinschaft. Eigenständige Systeme und vergleichbare Qualitäten? Länderberichte. Schriftenreihe des Bundesministeriums für Familie und Senioren, Band 22.2, Stuttgart: Kohlhammer, pp. 269-89.
Dumon, Wilfried (1990) Families and Policies. Evolutions and Trend in 1989-1990, Brussels: Commission of the European Communities.
Dumon, Wilfried (1991) Nationale Familienpolitik in den EG-Ländern in 1991, Brüssel: Kommission der Europäischen Gemeinschaften.
Esping-Andersen, Costa (1990) The Three Worlds of Welfare Capitalism, Cambridge: Polity Press.
Esping-Andersen, Gosta (1999) Social Foundations of Postindustrial Economies, New York: Oxford University Press.
European Commission (2000) Sicherung bei Pflegebedürftigkeit, Luxembourg: Amt für amtliche Veröffentlichungen der EG.
Hantrais, Linda and Letablier, Marie-Therese (1996) Families and Family Policies in Europe, New York: Longman.
Harding, Lorraine Fox (1996) Family, State and Social Policy, London: Macmillan.
Helmer, Elvira (1993) ,Unterstützung häuslicher Pflege', in Erika Neubauer, Christiane Dienel and Marlene Lohkamp-Himminghofen (eds), Zwölf Wege der Familienpolitik in der Europäischen Gemeinschaft. Eigenständige Systeme und vergleichbare Qualitäten? Schriftenreihe des Bundesministerium.s für Familie und Senioren, Band 22.1, Stuttgart: Kohlhammer, pp. 437-60.
372
373
EUROPEAN SOCIETIES
Hochschild, Arlie Russell (1995) 'The culture of politics: traditional, postmodern, cold-modern, and warm-modern ideals of care', Social Politics 2(3): 331-46.
Kamerman, Sheila B. and Kahn, Alfred J. (eds) (1978) Family Policy. Government and Families in Fourteen Countries, New York: Columbia University Press.
Kaufmann, Franz-Xaver (2002) 'Politics and policies towards the family in Europe: a framework and an inquiry into their differences and convergences', in Franz-Xaver Kaufmann, Anton Kuijsten, Hans-Joachim Schulze and Klaus Peter Strohmeier (eds), Family Life and Family Policies in Europe. Volume 2: Problems and Issues in Comparative Perspective, Oxford: University Press, pp. 419-90.
Knijn, Trudie and Kremer, Monique (1997) 'Gender and the caring dimension of welfare states: toward inclusive citizenship', Social Politics 4(3): 328-61.
Leitner, Sigrid (1999) Manner und Frauen im Wohlfahrtsstaat. Zur strukturellen Umsetzung von Geschlechterkonstruktionen in sozialen Sicherungssystemen, Frankfurt am Main: Peter Lang.
Leitner, Sigrid (2001) 'Das Splitting der Geschlechtskategorie in die Dimensionen gender und sex: Überlegungen zur geschlechtsspezifischen Sozialpolitikanalyse', Zeitschrift für Frauenforschung & Geschlechterstudien 19(1+2): 217-30. (English version: Implementing the Philosophical Differentiation between Gender and Sex. A Gainful Feminist Approach towards Structural Analyses of Social Policy. Paper presented at the Intrnational and Interdisciplinary Conference 'Gender and Markets in the Reconstruction of European Welfare States', 8-10 July 1999, University of Bremen.)
Lewis, Jane (1992) 'Gender and the development of welfare regimes'', Journal of European Social Policy 2(3): 159-73.
Lewis, Jane (ed.) (1998) Gender, Social Care and Welfare State Restructuring in Europe, Aldershot: Ashgate.
Lewis, Jane and Ostner, Ilona (1995) 'Gender and the evolution of European social policies', in Paul Pierson and Stephan Leibfried (eds), European Social Policy. Between Fragmentation and Integration, Washington, DC: The Brookings Institution, pp. 159-93.
Lohkamp-Himmighofen, Marlene (1993) 'Ansätze zur Förderung der Vereinbarkeit von Familie und Beruf, in Erika Neubauer, Christiane Dienel and Marlene Lohkamp-Himminghofen (eds), Zwölf Wege der Familienpolitik in der Europäischen Gemeinschaft. Eigenständige Systeme und vergleichbare Qualitäten? Schriftenreihe des Bundesministeriums für Familie und Senioren, Band 22.1, Stuttgart: Kohlhammer, pp. 313-67.
McLaughlin, Eithne and Glendinning, Caroline (1994) 'Paying for care in Europe: is there a feminist approach?', in Linda Hantrais and Steen Mangen (eds), Family Policy and the Welfare of Women. Cross-National Research Papers. Third Series: Concepts and Contexts in International Comparisons, Loughborough: University of Technology, pp. 52-69.
OECD (1995) Employment Outlook, Paris: OECD.
OECD (1996) Caring for Frail Elderly People. Policies in Evolution, Paris: OECD. OECD (2001) Employment Outlook, Paris: OECD.
Varieties of familialism LEITNER
Ungerson, Claire (1997) Social politics and the commodification of care, Social Politics 4(3): 362-81.
Sigrid Leitner is assistant professor at the Department of Social Policy, University of Goettingen, Goettingen, Germany. She works on comparative social policy, pension policies, family policies, and gender discrimination in social policy. Recent publication (with Ilona Ostner and Margit Schratzenstaller (eds)): 'Wohlfahrtsstaat und Geschlechterverhältnis im Umbruch. Was kommt nach dem Ernährermodell?' (Opladen: Leske+Budrich 2003).
Address for correspondence: Department of Social Policy, Georg-August-University, Platz der Goettinger Sieben 3, D 37073 Goettingen, Germany. E-mail: sleitne@uni-goettingen.de
374
375