INTERNATIONAL
JOURNAL OF
SOCIAL WELFARE
ISSN 1369-6866
© 2006 The Author(s)
Journal compilation © 2007 Blackwell Publishing Ltd and the International Journal of Social Welfare.
Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA 307
DOI: 10.1111/j.1468-2397.2006.00480.x
Int J Soc Welfare 2007: 16: 307–313
Guo J, Gilbert N. Welfare state regimes and family policy: a
longitudinal analysis
Int J Soc Welfare 2007: 16: 307–313 © 2006 The Author(s),
Journal compilation © 2007 Blackwell Publishing Ltd and the
International Journal of Social Welfare.
The primary aim of the article is to access the extent to which
welfare state regimes support distinct family policies across
OECD (Organisation for Economic Co-operation and Development)
countries and over time. Focusing in particular on the
work of Esping-Andersen we address the following questions:
(i) What is the relation between de-commodification and defamilialisation?
(ii) Are there distinct welfare state regimes that
support different levels of family policy? (iii) To what extent
have levels of support for family policy differed over the last
decades toward divergence or convergence among OECD
welfare states? Using data from the OECD Social Expenditure
database (1980–2001), we find that the degree of defamilialisation
through welfare state regimes parallels the
de-commodification scores. However, the findings indicate a
degree of instability in the relationship between levels of
support for family policy and welfare state regimes over the
last decades. We conclude by discussing how the findings
speak to the changing character of welfare regimes and the
implications of a longitudinal perspective.
Jing Guo, Neil Gilbert
School of Social Welfare, University of California, Berkeley
Blackwell Publishing LtdOxford, UKIJSWInternational Journal of Social Welfare1369-6866© Blackwell Publishing Ltd and the International Journal of Social Welfare 2007161Original ArticleWelfare state regimes and family policyGuo & Gilbert
Welfare state regimes and family
policy: a longitudinal analysis
Key words: de-familialisation, welfare regimes, de-commodification,
family policy, social expenditure
Neil Gilbert, School of Social Welfare, University of California,
120 Haviland Hall, Berkeley, CA 94720, USA
E-mail: berkeley1@msn.com
Accepted for publication August, 2006
Among comparative studies of modern welfare states, Gosta
Esping-Andersen’s (1990) path-breaking analysis of the
Three Worlds of Welfare Capitalism has framed the field
of inquiry over the last decade.1 Using de-commodification
of labour as one of the central characteristics to classify
social welfare policy in 18 advanced industrial countries,
he identified three ideal-type welfare states: liberal, conservative
and social democratic regimes. De-commodification
represents the extent to which social welfare policies allow
individuals and families to uphold a normal and socially
acceptable standard of living regardless of their performance
in the labour market. Esping-Andersen operationally
defined this concept by developing an index based on data
from 1980 that measure such factors as the stringency
of eligibility rules and the rate of income replacement
for pension, unemployment and sickness benefits. Liberal
regimes are typically marked by social policies that
engender the lowest levels of de-commodification; conservative
regimes (sometimes identified as continental
European or Christian democratic regimes) score in the
middle; and social democratic regimes rank highest on
the de-commodification index.
Although Arts and Gelissen (2002) show that
researchers have found the three-welfare-regime typology
to be a convincing tool for describing and explaining
cross-national differences, this typology has also been
criticised on several levels. For example, some find that
welfare capitalism changed dramatically in the 1990s and
that the three regimes identified by Esping-Andersen’s
analysis of data from 1980 have begun to converge (Ferge,
1996; Gilbert 2002; Rojas, 2005). Others argue that the
three regimes ignore the distinctive characteristics of
welfare systems in Mediterranean countries (Ferrera, 1996;
Leibfried, 1993).2 Van Voorhis (2003) reveals methodological
flaws in the design and calculation of the decommodification
index. And various scholars argue that
the three-welfare-regimes typology is too narrowly based on
income transfer programmes (such as pension, unemployment
and sickness programmes) for male breadwinner
workers and lacks a gender lens for examining how different
welfare regimes affect women, mothers and family life
(O’Connor, 1993; Orloff, 1993; Sainsbury, 1994, 2001).
Indeed, the feminist critique of mainstream welfare
state theories and the significant change in family
1 For example, see Arts and Gelissen’s (2002) review of empirical
studies using Esping-Andersen’s (1990) typology as an
organising principle for comparisons.
2 For instance, TheThreeWorlds ofWelfare Capitalism classification
did not include countries such as Spain, Portugal and Greece,
while Italy was grouped among the conservative welfare regimes.
Guo & Gilbert
© 2006 The Author(s)
308 Journal compilation © 2007 Blackwell Publishing Ltd and the International Journal of Social Welfare
structure and gender roles over the last several decades
have opened new perspectives on comparative analyses
of welfare state policy. Dissatisfied with the prevailing
three-welfare-regimes typology, Orloff (1993) and other
feminists argue that de-commodification pays little attention
to women’s relationships with the welfare state. They
have conducted analyses of gender-related welfare state
issues that attempt to combine mainstream welfare state
studies and feminist studies (Sainsbury, 1994, 2001). In
particular, feminist studies have contributed to welfare
state research by documenting the inequalities between
men and women embedded in work and social welfare
policies; they have pointed out the underlying ideologies
of familial and gender roles and expanded the analytic
focus of welfare state analyses on policies designed to
harmonise work and care (Gustafsson, 1994; Sainsbury,
1994; Siaroff, 1994). Social measures to reconcile work
and care have varied enormously among countries (Daly
& Lewis, 1998). For example, in Scandinavian countries,
the extensive provision of childcare was a part of explicit
social policy designed to create a dual-breadwinner model
and full employment. In contrast, public childcare provision
in the UK has been primarily for children who
are in poverty or at risk, and in the USA childcare has
remained essentially in the private sphere with limited
public support (although this has been increasing in recent
years). Knijn and Kremer’s (1997) comparative analysis
of the UK, Denmark and The Netherlands reveals
different patterns of organising care in welfare states.
Examining maternity leave policies and institutional
childcare, they found that, due to the focus on care as
a right of citizenship, the Danish welfare state came
closest to creating a system of gender equality.
With gender concerns at the core of their research,
feminist scholars have created alternative typologies of
welfare states based on the extent to which they allow
women to form autonomous households. Lewis (1992)
analyses the ‘male-breadwinner/female-caregiver’ and
‘dual-breadwinner’ models, which reflect the degree to
which social policies enhance women’s independence and
minimise the tension between work and family responsibilities.
In a similar vein, Sainsbury (1994) differentiates
the male-breadwinner model from the individual model
under which women qualify for pensions and other benefits
based on their own status rather than as their husbands’
dependents. From a broader perspective, Kamerman
(1995) distinguishes alternative approaches that guide
the development of family policies, which include a
targeted focus on poor families, a categorical emphasis
on lone-mothers, universal support for young children
and a combination of labour market and family supports.
Welfare regimes revisited
The movement of gender-related issues to the forefront
of comparative welfare state research gives rise to the
question: How well does the prevailing model of welfare
state regimes, based on the degree of de-commodification,
help to account for alternative approaches to family
policy? Addressing this question, Esping-Andersen (1999)
re-examined the welfare regimes through the analytical
lens of family-related policy. In doing this he introduced
the concept of de-familialisation, which is defined as
‘the degree to which households’ welfare and caring
responsibilities are relaxed either via welfare state provision
or via market provision’ (Esping-Andersen, 1999: 51).
This concept is operationally defined with several
empirical measures of social policies that promote defamilialisation
of family care and responsibility, such as
public expenditure on family services and the percentage
of children under 3 in public childcare.3 In comparison
with policies that de-commodifiy labour by reducing
workers’ dependence on the market, policies that promote
de-familialisation reduce the individual’s dependence
on kinship.
Employing these measures in a cross-sectional analysis
of data on family policy, Esping-Andersen finds a
general consistency among the three standard welfare
regimes and the degree of de-familialisation. That is, in
social democratic welfare state regimes, family policies
emphasise a strong government role, stressing the
goal of gender equity and child wellbeing. Due to these
generous family policies, social democratic regimes
exhibit the highest degree of de-familialisation. In
continental European countries, family policies stress
the role of traditional family and discourage female labour
force participation. Continental family policies parallel
the conservative welfare regimes that emphasise status
and class with underlying male breadwinner model
assumptions. These countries show a modest level of defamilialisation.
The liberal model marked by the AngloAmerican
approach to family policy has the lowest level of
de-familialisation through public welfare intervention
among the standard welfare state regimes originally
delineated in Three Worlds of Welfare Capitalism. In
addition to these three regimes identified in the 1990
study, Esping-Andersen (1999) introduced a fourth regime
– Southern Europe – in his analysis of the relationship
between family policy and welfare regimes. On the
measure of public spending on family services as a
percentage of gross domestic product (GDP), the
3 In addition to social policies that support de-familialisation
of welfare and caring responsibilities, Esping-Andersen (1990)
also presents measures of activities that reflect the degree to
which families and the market may assume more or less welfare
and care responsibility, such as the percentage of unemployed
youth living with parents, the average hours of women’s
unpaid work and the cost of private day care. Since our concern
in this article is the extent to which family policy efforts
parallel the three-welfare-regime model, the analytical focus
here will be on the degree of de-familialisation through social
welfare programmes rather than family or market provisions.
Welfare state regimes and family policy
© 2006 The Author(s)
Journal compilation © 2007 Blackwell Publishing Ltd and the International Journal of Social Welfare 309
southern European regime falls at the very bottom of
the list, with a level of de-familialisation below that of
the liberal regimes. Beyond the general parallel between
welfare state regimes and family policy efforts, EspingAndersen
suggests that the level of de-familialisation
based on public provision showed such a huge gap
between the social democratic regimes and all the
others, as to form a bimodal distribution.
A caveat on comparative data
It has long been recognised that, in addition to the
cheques written directly by government, comprehensive
measures of generosity and welfare effort should include
other sources of social expenditure that promote individual
wellbeing. As early as the 1930s, special tax deductions
and exemptions were identified as a form of government
transfer in Arthur Pigou’s classic text, Economics of
Welfare. However, it was not until the mid-1970s that
data on tax expenditures became available and were
introduced as a regular component of the president’s
budget in the USA, which was among the first countries
to collect this information [for a discussion of this
development in the USA, see Stanley Surrey (1974)].
As Gilbert and Moon (1988) pointed out, in the 1980s
the conventional categories of gross social expenditure
are at best a crude metric for comparisons among different
countries, which does not account for the true cost and
value of all social welfare benefits. Developing an
alternative index of welfare effort, they demonstrated
the significant shifts in the rankings among countries
that occurred when the analysis controlled for tax
burden, tax expenditures and need.
More recently, as additional data have become available,
Adema and his colleagues (1996) have developed a new
ledger for social accounting that controls for the effects
of costs and benefits from various sources – introducing
the most rigorous and comprehensive measures of social
spending to date. Their net total expenditure index
represents the cumulative value of benefits distributed
through direct public expenditures, tax expenditures
and publicly mandated private expenditures, reduced by
direct and indirect taxes on these benefits, and adds in
the value of voluntary private expenditures, reduced by
direct and indirect taxes.
Even this highly sophisticated measure, however, does
not take account of the effects of government deficit
spending, which can finance current benefits through the
creation of debt that must be discharged at some point
in the future. Moreover, measures of both gross and
net social expenditure, which compare spending among
countries as a per cent of their GDP, yield a different set of
results than when this spending is calculated according
to a per capita basis.All this is to say that, whether focused
on the total spending or spending directed mainly on
family policies, at their best, comparative data on social
expenditures offer only rough estimates of the welfare
efforts being studied.
In this article we are seeking to analyse the degree
to which Esping-Andereson’s findings concerning the
relationship between welfare regimes and de-familialisation
hold stable over time. While not immune to criticism
about the comparative measures of social expenditure
employed, his work is widely regarded as authoritative. To
examine the pattern of de-familialisation among welfare
regimes from a longitudinal perspective, we employ
measures of direct public social expenditure that closely
replicate Esping-Anderson’s operational definition of
de-familialisation, recognising that this direct public
social expenditure approach is not without deficiencies.
Refocusing the analytic frame: a longitudinal
perspective
While cross-sectional analysis reveals a relationship
among levels of de-familialisation and the welfare state
regimes, the snapshot in time afforded by this type of
analysis captures a restricted picture, which can sometimes
be misleading and offers no indication of any temporal
trend that might influence an interpretation of the pattern
shown in the data. Taking the typical social democratic
country of Sweden as an example, Figure 1 shows the
trend of total public expenditure on family from 1980
to 2001.
Noticeably, public spending on family provisions
reached its peak in 1992 and has gradually decreased
since then. Based on cross-sectional analysis using data
from the year 1992, Esping-Andersen (1999) concludes
that the social democratic regimes have considerably
higher public spending on family services than other
welfare regimes. It is possible that the disparity in public
spending on family benefits among welfare regimes varies
considerably over time and what appears as a bi-modal
distribution in 1992 is significantly different from the
relationship among welfare regimes and family policy in
1980 and 2000. In that case, the extent to which regime
types can be used to explain levels of de-familialisation
is seriously open to question.
How well does the relationship between welfare regimes
and de-familialisation hold over time? Does it continue
Figure 1. Public expenditure on family as percentage of GDP in
Sweden (1980–2001).
Guo & Gilbert
© 2006 The Author(s)
310 Journal compilation © 2007 Blackwell Publishing Ltd and the International Journal of Social Welfare
to form a bi-modal distribution with the Scandinavian
welfare states on the high end of the scale of defamilialisation
as Esping-Andersen’s (1999) previous
analysis suggests? If the pattern over time is not stable,
do the fluctuations suggest a trend toward divergence or
convergence? In addressing these questions we analyse
the levels of family policy among different welfare regimes
over three time periods – 1980, 1990 and 2001. Table 1
shows the updated classification of welfare regimes and
the de-commodification score for each country.
Along with Esping-Andersen’s (1999) regime classification,
our analysis employs measures of family policy
that closely replicate his operational definition of defamilialisation
through welfare state provisions,
which focused on three key indicators – family service
expenditure, public childcare coverage of children
under 3 years of age and home-help service coverage
for seniors.4 He also suggests that cash transfers are
important. Following in this vein, we define defamilialisation
through public provision by calculating
the overall welfare spending on family benefits for each
country. Table 2 illustrates the main domains and subcategories
that form the operational definition of total
social expenditure on family policy, which encompasses
measures of both cash benefits and family services.
Family cash benefits include family allowances for
children, and maternity and parental leave, which are
widely considered to be two essential components of
family policy. In addition, the measures of family services,
which include day care, personal and home-help services,
capture the main indicators suggested by EspingAndersen
(1999) – public childcare coverage of young
children and home-help service coverage for seniors.
Since the data on total public expenditure on family
policy go beyond measures of family service expenditure,
we analysed the relationship between levels of defamilialisation
represented by these two indices and the
de-commodification scores. In order to compare the
results with those of Esping-Andersen’s (1999) study, we
used data on de-familialisation for the same year, 1992,
as reported in his analysis. (The de-commodification
index was constructed by Esping-Andersen in an earlier
study and based on data from 1980.)
As shown in Table 3, both the total public spending
on family (both cash and in-kind) and public spending
on family services in 1992 have statistically significant
relationships to the de-commodification scores. The
total public expenditure on family had a slightly higher
correlation (0.710) with the de-commodification scores
than the measures of spending only on family services
(0.691). This finding demonstrates a clear-cut parallel
between countries with different degrees of decommodification
and their total levels of expenditure on
family policies in 1992. Countries with higher decommodification
scores, which imply more generous
spending on key work-related welfare programmes such
as pensions and unemployment benefits, are more likely
4 The data used in our analyses are from the OECD Social
Expenditure database (1980–2001), which was released in 2004.
This is one of the latest, most reliable and internationally
comparable datasets. It provides an opportunity for exploring the
trends in aggregate social expenditure and analysing the
changes in any specific programme, such as family programmes.
Table 1. Classification of 17 OECD countries (de-commodification score).
Social-democratic regimes Continental Europe Liberal regimes Southern Europe
Denmark (38.1) Austria (31.1) Australia (13) Italy (24.1)
Finland (29.2) Belgium (32.4) Canada (22) Portugal*
Norway (38.3) France (27.5) Ireland (23.3) Spain*
Sweden (39.1) Germany (27.7) UK (23.4)
Netherlands (32.4) US (13.8)
Note: De-commodification scores are from Esping-Andersen (1990). * Portugal and Spain were not included in the original de-commodification
analysis (1990), but they are included in the examination of de-familialisation (Esping-Andersen, 1999).
Table 2. Social expenditure on family policy.
Family cash benefits Family cash benefits
Family allowances Family allowances for children
Family support benefits
Maternity and parental leave Maternity and parental leave
Other cash benefits Lone parent cash benefits
Family other cash benefits
Benefits in kind Family services
Day care/home-help services Family day care
Personal services
Household services
Other benefits in kind Household other benefits in kind
Source: The OECD Social Expenditure database (1980–2001).
Table 3. Correlation between de-commodifcation and de-familialisation
(1992).
Public expenditure on
family service (1992)
Total public expenditure
on family (1992)
De-commodification
score (1980)
0.691* 0.710*
(p = 0.004, n = 15) (p = 0.003, n = 15)
Welfare state regimes and family policy
© 2006 The Author(s)
Journal compilation © 2007 Blackwell Publishing Ltd and the International Journal of Social Welfare 311
to spend a higher percentage of their GDP supporting
families with children as well.
The correlation based on 1992 measures of defamilialisation
suggests, but does not confirm, that there
is a pattern of distinct welfare state regimes that support
different levels of family policy; it tells us nothing
about the extent to which these regimes form a stable
pattern of relationship with public expenditure on family
policy over time. To address this question we use
analysis of variance (ANOVA) to statistically test for
significant variations in public expenditure on family
among welfare regimes over time. Table 4 shows the
average public expenditure on family as a per cent of
GDP for each welfare regime. We include the year 1992
to examine the results in comparison with those of
Esping-Andersen’s study. Indeed, as shown in Table 5,
the average rate of total family expenditure in the social
democratic regime was significantly higher than that
of the other three regimes. However, these data also
show that, in addition to spending a significantly lower
percentage of GDP on family policies than social democratic
regimes do, the southern European countries also
spent significantly less than the continental European
countries. Hence, beyond the large divide that existed
between the social democratic regimes and the others
in 1992, the southern European countries were the
laggards even within the other regimes that had lower
levels of de-familialisation.
But how stable is this pattern of relationship over
time? Do social democratic regimes always show higher
public spending on family policies? Are the differences
among all the regimes large enough to be statistically
significant? Do the differences vary over time and, if
so, in what direction? In response to these questions,
Table 6 shows the ANOVA results regarding the variance
in total public expenditure on family benefits among
welfare regimes for the periods of 1980, 1990 and 2001.
The findings for 1980 show a bi-modal distribution
in which total family expenditure in social democratic
regimes and continental European countries is significantly
higher than in liberal regimes and southern European
countries. However, there is no significant difference
between total public spending on family benefits in social
democratic regimes and continental countries. Indeed,
at that time, family expenditure in continental European
countries was slightly higher than in social democratic
regimes. Overall, there are statistically significant
differences in the average levels of family expenditure
among regimes in four of the six possible comparisons
– which suggests that the regime classification is useful
in distinguishing the degree to which social policies of
different regimes promote de-familialisation in 1980.
The results of the 1990 comparison are similar to the
analysis for 1992. However, in this period the bi-modal
distribution shifts with a sharp division between the social
democratic countries on the high side and the other
three regimes with significantly lower average levels of
spending on family benefits. Again, the results show
statistically significant differences in family expenditure
among the regimes in four of the six possible comparisons.
But in 1990 these differences are not among all the
same regimes as in 1980, which indicates some instability
over time.
Table 4. Total family expenditure as percentage of GDP (mean and standard deviation).
1980 1990 1992 2001
Social democratic regimes 2.62 (0.64) 3.46 (0.74) 4.00 (0.70) 3.23 (0.39)
Continental Europe 2.70 (0.37) 2.22 (0.52) 2.32 (0.56) 2.22 (0.73)
Liberal regimes 1.19 (0.64) 1.23 (0.66) 1.43 (0.81) 1.59 (1.00)
Southern Europe 0.79 (0.31) 0.76 (0.38) 0.74 (0.30) 0.88 (0.34)
Table 5. Comparison of public expenditure on family between welfare
regimes, 1992.
(I) Regime (J) Regime Total family expenditure
difference (I–J)
Social democratic Continental Europe *
Liberal *
Southern Europe *
Continental Europe Liberal
Southern Europe *
Liberal Southern Europe
Total family expenditure is calculated as a percentage of GDP.
* The mean difference is significant at the 0.05 level.
Table 6. Comparison of total public expenditure on family between
welfare regimes (1980, 1990 and 2001).
(I) Regime (J) Regime Total family expenditure
difference (I–J)
1980 1990 2001
Social democratic Continental Europe Ns * ns
Liberal ** ** **
Southern Europe ** ** **
Continental Liberal ** ns ns
Europe
Liberal
Southern Europe ** ** ns
Southern Europe Ns ns ns
Total family expenditure is calculated as a percentage of GDP.
The mean difference is significant ** at the 0.05 level; * at the 0.10
level. ns, non-significant differences.
Guo & Gilbert
© 2006 The Author(s)
312 Journal compilation © 2007 Blackwell Publishing Ltd and the International Journal of Social Welfare
A heightened pattern of instability appears in the
analysis of differences in total public family expenditures
among welfare state regimes in 2001; here the differences
between social democratic regimes and continental
European countries, as well as those between the
continental and southern European regimes, no longer
show up as significant. In 2001, the variations in family
expenditure are statistically significant in only two of the
six possible comparisons – with the social democratic
regimes showing a significantly higher percentage of
spending than the liberal and southern European
countries. The shifting pattern of public expenditure on
family benefits from 1980 to 2001 is graphically
illustrated in Figure 2.
Conclusion: welfare regimes over time
These findings from a longitudinal perspective on the
extent to which welfare state regimes incorporate distinct
levels of de-familialisation (operationally defined by
public family expenditures) lend themselves to several
interpretations.
First, while the social democratic countries generally
have higher levels of public family spending than the
other regimes, our results only partially confirm EspingAndersen’s
conclusion about a bi-modal distribution,
based on his cross-sectional analysis of spending in 1992.
We found that, although such a distribution did exist
with statistically significant differences between the
social democratic and the other countries in 1990, it
was not present in 1980 or 2001, suggesting apparent
changes in levels of de-familialisation among welfare
regimes over time. In 1980, social democratic regimes
and continental European countries spent almost the
same percentage of GDP on family benefits. And in
2001, again the difference in average spending between
social democratic countries and continental European
countries was not significant.
Second, over the decades from 1980 to 2001 there
was a fair amount of fluctuation in the extent to which
average public expenditure on family benefits was
statistically differentiated among all four welfare state
regimes. Although in both 1980 and 1990 comparisons
of average public expenditures on family among the
four clusters of countries revealed significant difference
in two thirds of the cases, the significant differences did
not appear among all of the same regimes during these
two periods of time. Moreover, by 2001 the findings
show significant differences in average spending among
the four regimes in only one-third of the possible
comparisons.
Lastly, examining the patterns of public expenditure on
family benefits over time reveals that the differentiations
among the four regimes increased to its maximum level
between 1980 and 1990 and then narrowed between
1990 and 2001 – as the average public spending on
family benefits in social democratic regimes declined
while expenditure increased or remained constant in the
other regimes. Between 1980 and 1990 the average
spending diverged as different levels of de-familialisation
among the four regimes became more distinct. After
1990 this trend was reversed and the average levels of
spending began to converge.
Various reasons have been posited for the apparent
movement toward convergence among welfare regimes,
including the combined impact of demographic and
market forces. In recent times, social welfare systems
throughout the advanced industrialised nations have
experienced increasing social and fiscal demands from
the demographic pressures of rapidly ageing populations.
And during the same period policy makers have come
under the competitive discipline of well-integrated global
markets, which has intensified pressures to scale back
social benefits which are seen as hindering a country’s
capacity to keep production costs low and attract foreign
investment (Standing, 1999). In response to these and other
social forces, some analysts find a growing convergence
among policy reforms in welfare regimes toward emphasis
on privatisation, targeting of benefits, promotion of work
and individual responsibility (Ferge, 1996; Gilbert, 2002;
Rojas, 2005). From another perspective, the welfare
regimes of western Europe have been seen as increasingly
influenced by developments at the supranational level
of the European Union. In this regard, there appears
to be a growing interest in the question of European
convergence, particularly in terms of social policies,
such as childcare (e.g. Mahon, 2002; Randall, 2000).
It is always difficult to clearly identify a change process
as it unfolds. Although our findings point towards a
convergence in public spending on family benefits in
recent years, we would like to close with a word of
caution about the limitations of this analysis. The study
concentrates on direct expenditures to capture the ‘how
much’ dimension of social transfers to families. This
approach does not take into account transfers through tax
expenditure, mandated benefits and voluntary contributions,
which vary in significance among different countries
(Adema, 1999; Gilbert & Gilbert, 1989; Hacker, 2002;
Howard, 1997). Moreover, studies of welfare state
Figure 2. Public expenditure on family as percentage of GDP (1980,
1990 and 2001).
Welfare state regimes and family policy
© 2006 The Author(s)
Journal compilation © 2007 Blackwell Publishing Ltd and the International Journal of Social Welfare 313
classifications have argued that the use of aggregate
public expenditure data provides only a partial picture,
which excludes the substance of social provisions and
the ways these benefits, especially service benefits, are
delivered, i.e. to measure and understand the ‘how
much’ dimension does not tell us precisely how the
funds are being spent (Bonoli, 1997). Hence, these
findings should be seen as suggestive of a trend in
family policy that awaits further verification through
more detailed analyses.
References
Adema W (1999). Net social expenditures: labour market and
social. Occasional Papers No. 32. Paris, OECD.
Adema W, Einerhand M, Eklind B, Lotz J, Pearson M (1996).
Net Public Social Expenditure: Labour Market and Social
Policy. Occasional Papers No. 19 Paris, OECD.
Arts W, Gelissen J (2002). Three worlds of welfare capitalism
or more? A state-of-the-art report. Journal of European
Social Policy 12: 137–158.
Bonoli G (1997). Classifying welfare states: A two-dimension
approach. Journal of Social Policy 26: 351–372.
Daly M, Lewis J (1998). Introduction: Conceptualising social
care in the context of welfare state restructuring. In: Lewis J,
ed. Gender, social care and welfare state restructuring in
Europe, pp. 1–24. Brookfield, Ashgate.
Esping-Andersen G (1990). The three worlds of welfare capitalism.
Princeton, NJ, Princeton University Press.
Esping-Andersen G (1999). Social foundations of postindustrial
economies. Oxford, Oxford University Press.
Ferge Z (1996). The change of the welfare paradigm – the
individualization of the social. Paper presented at the Annual
Conference of the British Social Policy Association, Sheffield,
July 16–18.
Ferrera M (1996). The ‘southern model’ of welfare in social
Europe. Journal of European Social Policy 6: 17–37.
Gilbert N (2002). Transformation of the welfare state: the silent
surrender of public responsibility. New York, Oxford University
Press.
Gilbert N, Gilbert B (1989). The enabling state. New York,
Oxford University Press.
Gilbert N, Moon A (1998). Analyzing Welfare Effort: An Appraisal
of Comparative Methods. Journal of Policy Analysis and
Management 7: 326–340.
Gustafsson S (1994). Childcare and types of welfare states. In:
Sainsbury D, ed. Gendering welfare states, pp. 45–61. London,
Sage.
Hacker J (2002). The divided welfare state. Cambridge, Cambridge
University Press.
Howard C (1997). The hidden welfare state: tax expenditures and
social policy in the United States. Princeton, NJ, Princeton
University Press.
Kamerman SB (1995). Gender role and family structure changes
in the advanced industrialized west: Implications for social
policy. In: McFate K, Lawson R., Wilson WJ, eds. Poverty,
inequality, and the future of social policy: Western states in
the new world order, pp. 231–256. New York, Russell Sage
Foundation.
Knijn T, Kremer M (1997). Gender and the caring dimension
of welfare states: Toward inclusive citizenship. Social Politics
(Fall): 328–361.
Leibfried S (1993). Towards a European welfare state? In: Jones P,
ed. New perspectives on the welfare state in Europe, pp. 135–
156. London, Routledge.
Lewis J (1992). Gender and the development of welfare regimes.
Journal of European Social Policy 3: 159–173.
Mahon R (2002). Child care: Toward what kind of social Europe?
Social Politics (Fall): 343–379.
O’Connor JS (1993). Gender, class and citizenship in the
comparative analysis of welfare state regimes: Theoretical
and methodological issues. British Journal of Sociology 44:
501–518.
Orloff AS (1993). Gender and the social rights of citizenship:
The comparative analysis of state policies and gender relations.
American Sociological Review 58: 303–328.
Pigou AC (1960). The Economics of Welfare. New York, St.
Martin’s Press.
Randall V (2000). Childcare policy in the European states: Limits
to convergence. Journal of European Public Policy 7: 346–368.
Rojas M (2005). Sweden after the Swedish model: from tutorial
state to enabling state. Stockholm, Timbro Publishers.
Sainsbury D (1994). Women’s and men’s social rights: Gendering
dimensions of welfare states. In: Sainsbury D, ed. Gendering
welfare states. London, Sage.
Sainsbury D (2001). Gendering dimensions of welfare states. In:
Fink J, Lewis G, Clarke J, eds. Rethinking European welfare:
Transformations of Europe and Social Policy, pp. 114–139.
Thousand Oaks, CA, Sage.
Siaroff A (1994). Work, welfare and gender equality: A new
typology. In: Sainsbury D, ed. Gendering welfare states,
pp. 82–100. London, Sage.
Standing G (1999). Global labour flexibility: Seeking distributive
justice. New York, St. Martin’s Press.
Surrey S (1974). Pathways to Tax Reform. Cambridge, MA,
Harvard University Press.
Van Voorhis R (2003). Different types of welfare states? A
methodological deconstruction of comparative research.
Journal of Sociology and Social Welfare 29: 3–18.