http://esp.sagepub.com
Journal of European Social Policy
DOI: 10.1177/0958928707087589
2008; 18; 115Journal of European Social Policy
Dorota Szelewa and Michal P. Polakowski
Who cares? Changing patterns of childcare in Central and Eastern Europe
http://esp.sagepub.com/cgi/content/abstract/18/2/115
The online version of this article can be found at:
Published by:
http://www.sagepublications.com
can be found at:Journal of European Social PolicyAdditional services and information for
http://esp.sagepub.com/cgi/alertsEmail Alerts:
http://esp.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.co.uk/journalsPermissions.navPermissions:
http://esp.sagepub.com/cgi/content/refs/18/2/115Citations
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
ESPAnet/JESP Doctoral Researcher Prize Essay
Who cares? Changing patterns of childcare in Central and
Eastern Europe
Dorota Szelewa, European University Insitute, Florence, Italy
Michal P. Polakowski*, Maastricht Graduate School of Governance,
The Netherlands
Summary This article compares childcare provisions in the new member countries of the EU. It takes
into account two pillars of childcare policy: publicly provided childcare services and parental leave
provisions. In the analysis, the fuzzy set ideal types approach is utilized. In contrast to the studies
conducted so far, this article stops treating the region of Cental and Eastern Europe as a monolith
and demonstrates the existence of cross-country variation of childcare policies within the region.
Furthermore, the difference is systematized by identifying four clusters of childcare policy. These are:
explicit familialism, implicit familialism, female mobilizing and comprehensive support types. The
countries are clustered as follows: the Czech Republic, Slovakia and Slovenia in the explicit familism
policy model; Estonia and Latvia in the female mobilizing type policy; Lithuania and Hungary pursuing
the childcare policies typical of the comprehensive support model; and finally the childcare
policy in Poland resembles characteristics of the implicit familism model.
Key words Central and Eastern Europe, childcare, fuzzy sets, gender regimes, transition
Introduction
Childcare policies have of late been perceived as one
of the main instruments of the state to influence
gender relations and women’s economic autonomy
(Daly, 1994; Jenson and Sineau, 2001). While family
policy and gender regimes in the Western world have
already received a lot of attention, the situation in
the newly democratized post-communist countries
has, to date, remained under-analysed. The development
of childcare policies in the region is, therefore,
often presented in a one-sided way. It is argued that
under socialism women were encouraged to join the
labour force by special incentives which included
publicly provided, affordable childcare services.
However, after the collapse of communism, the governments
in the newly democratized states assumed
that these societies would opt for a male-breadwinner
model. As a result, authorities started to close
many childcare centres, especially nurseries, and
withdrew financial support (Ferge, 1992; Sainsbury,
1994).
These processes were indeed put into practice,
and no doubt these countries have had a lot in
common. Thus, post-communist family policies
have been treated as a monolith and the common
trend of refamilialization has been emphasized
(Hantrais, 2004; Pascall and Lewis, 2004). Studies
which have started to deal with the possible diversity
of policies in these states are still quite rare (for
exceptions see: Fodor et al., 2002; Saxonberg and
Sirovatka, 2006), and have typically focused on
* Author to whom correspondence should be sent: Dorota Szelewa, Department of Political and Social Sciences, European
University Institute, Badia Fiesolana, Via dei Roccettini 9, I-50014 San Domenico di Fiesole (FI), Italy. [email:
dorota.szelewa@eui.eu]
Journal of European Social Policy 0958-9287; Vol 18(2): 115–131; 087589 Copyright © SAGE Publications 2008, Los Angeles, London,
New Delhi and Singapore, DOI: 10.1177/0958928707087589 http://esp.sagepub.com
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
parental leave provision (omitting the question of
childcare services) at the same time dealing with two
or three case studies. Additionally, the comparisons
are often not systematized enough and do not cover
the differences between these countries.
This article aims to further develop the picture of
childcare policy developments in this region of
Central and Eastern Europe (CEE) by way of a
precise and systematic comparative description of
the differences between these countries. This task
first required gathering the various data sets which
were previously unknown, especially in relation to
the early stage of the post-communist transformation.
Second, we present a more nuanced picture of
the post-communist world, providing a systematic
description of differences between these countries
with a more dynamic, historical perspective. Third,
a new methodological approach (fuzzy set ideal
types analysis) has been applied, allowing for a
comparison of many cases without losing focus on
their complexity. This method provided findings
which reveal the existence of different versions of
familialist, as well as non-familialist, childcare policies
in the post-communist world. The article also
contributes to the works relating to childcare policies
in this part of Europe, as we take into account
changes which occurred throughout the whole
period of transformation. Finally, the analysis
covers both the schemes providing cash benefits, as
well as childcare services.
We decided to compare childcare policies in
the eight new European Union member states:
the Czech Republic, Estonia, Hungary, Latvia,
Lithuania, Poland, Slovakia and Slovenia during the
period 1989–2004. This is the first time such a large
number of post-communist countries have been
chosen for a dynamic comparative analysis of childcare
policies. All these countries were facing similar
challenges within similar time frames; each of them
was regarded as successful in their ‘unification with
the West’, and they gained EU membership at the
same time. Thus, on the basis of these similarities
one might expect them to have pursued, and to be
pursuing, similar types of policies.
This article takes the fuzzy set ideal types
approach to comparative studies as it facilitated
delineation of differences in childcare policies, and it
is especially useful for medium-n comparisons. The
fuzzy set approach also allows for tracing the
dynamics of change through the years and opens the
way for new combinations of policies. Instead of
checking whether particular mixes of policies
conform to the ideal type of an already existing
typology which was developed in the West, the
approach actually allows for discovering new policy
mixes. In this sense the whole analysis has a strong
inductive element while being both very precise and
systematic at the same time. In using this method, it
was possible to overcome the problem of applying
the Western analytical framework. With its defined
concepts and welfare policy models, these conceptdriven
analyses might take some important elements
out of the Eastern European picture.
The work follows in several steps. First, some
points on the theoretical discussion concerning the
comparative approach to childcare policy are presented,
followed by the choice of the policy dimensions
used for comparison. The second section of
the article describes the fuzzy set ideal types method.
Third, some space is devoted to the justification of
the sets’ calibration. Fourth, we present the empirical
findings, highlight the regularities, and cluster
the countries.
Theoretical discussion
Comparative works on childcare have been involved
in the debate between the mainstream research and
the feminist work on welfare states and social policy
in general (Esping-Andersen, 1990; Lewis, 1992;
Orloff, 1993). For example, many studies which
have already taken the feminist critique into account
when considering mainstream research on welfare
state and family policy have based their analyses on
the notion of defamilialization (Esping-Andersen,
1999; Hantrais, 2004; Leitner and Lessenich, 2005).
The central question of this line of research is how
different mixes of family policy measures influence
the strength and kind of intra-family dependencies.
The most obvious example of defamilializing policies
is publicly provided, accessible and cheap care
services. Among others, they enable women
(primary care providers) to join the labour force,
and therefore facilitate access to commodification
(Orloff, 1993).
As defamilializing policies shift the responsibility
for care away from the family, familialist policy
measures do just the opposite. Mahon (2002)
describes the ‘new familialism’ with regard to the
institutional evolution of family policies in France
116 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
and Finland as a response to defamilialization.
Similarly, Hantrais (2004) talks about the neofamilialist
policies in Western Europe and refamilialization
in post-communist countries. Giving less
priority to publicly provided childcare services is
presented in such countries as offering ‘more choice’
or ‘the rights to care’. Here, an example of familialist
policies may be parental leave with a long duration
of payment.
However, as some other studies have demonstrated,
one might as well talk about different
‘faces’ or ‘varieties’ of familialism (Haney, 2003;
Leitner, 2003), to a large part depending on the
degree and kind of state activity (or the availability
of services provided by the market). Leitner (2003)
differentiates between four types of policy mixes
based on the presence of familialistic and defamilializing
elements within them: ‘implicit’, ‘explicit’
and optional ‘familialism’ and ‘defamilialization’.
Only in the case of the latter do care services relax
the family responsibility for care as the primary
family policy measure. Acknowledging the ‘varieties
of familialism’ has important consequences
for grasping the developments in childcare policies
in Eastern Europe. As mentioned above, this part
of Europe is very often perceived as a monolith,
and the authors underline the common trend of
refamilialization (Hantrais, 2004; Pascall and
Lewis, 2004).
The common tendency towards reforming childcare
policies in the familialist direction is by no means
present in the post-communist countries. Legacies of
state socialism included high rates of female economic
activity, usually highly developed childcare
services and quite generous systems of state support
for maternity and family. At the beginning of transformation,
with the withdrawal of the state from
social policies, defamilializing policies started to be
cut, especially the provision of public childcare services.
Due to budget constraints, decisionmakers were
faced with important trade-offs and dilemmas. Cuts
in family policy were often the solution to these
dilemmas, justified by slogans emphasizing that
maternity and rearing children are a woman’s role
(Domanski, 1992).
The new stream of research on this region has
begun to acknowledge the gender perspective in
analysing economic transformation in Eastern
Europe, and especially the dynamics of childcare
policy reforms (Fodor et al., 2002; Haney, 2002;
Aidukaite
.
, 2004). Still, systematic research and
especially clustering countries was difficult due to
the lack of stability of the policy outcomes and their
temporary character (Deacon, 1993; Ksiezopolski,
1999). This was also one of the reasons why in 1996
Esping-Andersen rejected the possibility of using his
regime types for a comparative analysis of the
Eastern European countries since they were ‘a
virtual laboratory of experimentation’ (EspingAndersen,
1996: 267).
Haney (2002), for example, defined the features
of the Hungarian welfare state as liberal, but
without taking into account the generous parental
leave provision and well-developed childcare services.
There were similar problems with classification
of Polish childcare policies. The latter have usually
been described as conservative (Heinen, 1997),
mainly due to the role and position of the Catholic
Church in this country. Still, the development of
pronatalist policies reflecting the religious orientation
of political elites could barely be observed in
Poland, in contrast to the case of the secular Czech
Republic (Saxonberg and Sirovatka, 2006). Finally,
less attention has been paid to Baltic post-communist
countries. One of the most recent studies concluded
that the childcare policies in Lithuania have
been less generous than in Estonia and Latvia
(Aidukaite
.
, 2004).
Our study aims to tackle these analytical and
empirical obstacles. First, this article includes consideration
of childcare services as the family policy
dimension that has been underexplored in previous
research. Moreover, the possible diversity among
such a number of countries has never been studied
before. Using a broad range of empirical data
makes the comparison even more reliable and
shows that different mixes of childcare policies are
present in these countries. Second, in order to overcome
the problem of the temporal instability of
these countries’ policy output, the analysis attempts
to present synchronic and diachronic perspectives
simultaneously. In other words, this article presents
a comparison between the countries, but also
between different points in time. Thus, it is possible
to present the dynamics of changes, policy trends
and the diversity among the countries. The next
section describes the dimensions which constitute
childcare policies and provides a basis for
distinguishing between different combinations of
policies.
Who cares? Changing patterns of childcare in Central and Eastern Europe 117
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
Childcare policy – dimensions of
comparison
Following feminist scholarship on welfare policies,
but also other comparative studies of family policies
(Kamerman and Kahn, 1981; Zimmermann, 1995;
Gauthier, 1996; Harding, 1996), we include parental
leave provisions and day-care arrangements in the
analysis. These two elements have been recognized as
the most important pillars of childcare policies.
Accordingly, Daly and Rake argue that ‘the two measures
most telling of how public policies treat care are
parental leave, as distinct from maternity, and public
childcare facilities’ (Daly and Rake, 2004: 51).
Further, we decided to express two kinds of childcare
policies through four dimensions: (a) the extensiveness
of childcare services; (b) the quality of
childcare services; (c) the generosity of parental
leave; and (d) its universality. Below we provide a
more detailed description of these four dimensions.
Extensiveness of childcare services. The extensiveness
of childcare services reflects the role of the
state in the provision of services aimed at families
with small children (at pre-primary school age).
This aspect also provides information on the locus
of the service: if the extensiveness of the publicly
provided childcare is high, it means that the care is
provided by the state; if it is low, then the importance
of other forms of care (provided by the family
or private agencies) increases. This is why the
dimension of extensiveness is of great relevance to
the discussion about the interplay between public
(state-provided) and private (mostly home-provided)
services (cf. Lewis, 1992; Orloff, 1993). The
exclusion of for-profit care seems reasonable, since
this type of service is marginal in the region
(Rostgaard, 2004). The extensiveness of the childcare
is operationalized here through the net enrolment
rate in pre-primary education.
Quality of childcare services. The next dimension of
childcare policy is the quality of the services provided
by publicly run kindergartens. The higher the quality
of the kindergartens, the more attention the state pays
to the issue of institutional support for parents. The
commitment of the welfare state to the provision of
high-quality institutional childcare should be perceived
as an indication of the incentivization of female
employment, in addition to the promotion of gender
equality (among partners), social cohesion and equalizing
life chances of children.
As Kamerman points out, ‘there is no agreed
definition of – or standards concerning – quality of
ECEC programs cross-nationally and little systematic
attention to this subject in the literature’ (2003:
10). This article focuses on the pupil–teacher ratio
as it is the only available measure for the comparison
conducted here. Moreover, it is argued that this
measure of quality has a serious impact on the educational
outcomes of pupils (Kvist, 1999). Finally,
this transparent measure can be observed by the
parents, and in this way may determine the childcare
attendance of their children (especially those
aged 3–6) (Gornick and Meyers, 2003).
Generosity of parental leave provisions. The function
of parental leave may vary according to its
duration and generosity. On the one hand, some
leave is necessary to recover from the delivery and
for the first period of breastfeeding. On the other
hand, extensive periods of leave can contribute to
the economic ‘deactivation’ of women.
The generosity of parental leave provisions is here
expressed by the length of paid leave and the level of
payment. The two elements are the basis for the
index of generosity. It is measured by the number of
paid weeks’ leave multiplied by the level of payment
expressed as the replacement rate of the average
monthly net wage. We also include the number of
weeks available for using maternity leave plus the
whole duration of paid parental leave.
Importantly, in the post-communist countries
this leave is divided into two parts which we will
call maternity leave and extended leave. Maternity
leave varies up to a dozen weeks, is directly connected
with the birth of a child, and usually cannot
be shared with the father. Then parental leave
follows, which can last for three years. For the
purpose of comparison, families with one child are
considered. Finally, wherever more than one system
exists, the scores are added, presenting the overall
effort of the state. See Table 1 for examples of the
calculation.
Universality of parental leave provisions. Finally,
the universality of parental leave benefits is taken
into account. The benefits for persons taking care of
a small child are important for the assessment of the
state policies towards parenthood. The analysis of
the universality of parental leave benefit reflects the
approach of the state towards this problem: if benefits
are easily available, then the state commitment
to the issue is considered as high.
118 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
The analysis of the universality of benefits should
reflect the conditions under which the benefit is
granted to the eligible individuals. Usually this is
based on the differentiation between whether the
benefit is universal (on the basis of the citizenship or
residence), selective (insurance-based) or residual
(available only to those deemed in need after passing
the means or income test) (Clasen and van Oorshot,
2002). In the case of maternity and extended
parental leave benefits, the situation is complex as
there might be different principles applied to each
benefit. This can give different combinations; for
instance, universal access to maternity benefit can
be combined with an insurance-based access to
extended parental leave.
Further, the dimensions are used as a basis for
translating data into the fuzzy set scores. The
following section explains this methodological
approach.
Fuzzy set ideal types analysis – an
introduction
Fruitful research on the welfare state (including childcare
policy) should be sensitive to the variation among
cases and driven by clear conceptual and theoretical
guidelines. The complexity of childcare policy, methodologically
speaking, refers to the fact that one has to
analyse many dimensions (in our case four) of this
policy simultaneously. However, this task becomes
harder when the number of cases grows. The conceptual
solution to studying complexity has been proposed
by Lazarsfeld (1937) who developed the idea of
viewing different instances (types) of the same phenomenon
as configurations of attributes constituting
this phenomenon. In effect, each type has its own specific
combination of attributes.
A more formal and systematic application of the
configurational approach, the fuzzy set theory, has
been proposed by Ragin (2000). This article utilizes
the extension of this approach, namely fuzzy set
ideal types analysis. Based on the fuzzy set theory, it
allows for a careful comparison of a greater number
of cases without losing the emphasis on the complexity
of childcare policy. As shown by Kvist
(1999; 2006), this type of comparative approach
can be successfully used for studying complexity,
diversity and change of social policy.
The approach introduces the notion of fuzzy sets.
A fuzzy set is a representation of an empirical
phenomenon which is guided by rules stemming
from theoretical and substantive knowledge. Fuzzy
sets should not be seen as conventional variables as
they comprise qualitative anchors (defining boundaries
of a set) which affect the membership assessment.
The qualitative anchors are established in the
process of calibration. Fuzzy set ranges from 0 (fully
out of the set) to 1 (fully in the set) and these values
– with 0.5 as a cross-over point – constitute qualitative
boundaries of a phenomenon. Calibrating
the set means establishing points for translating
empirical values into fuzzy set scores. For example,
a theoretical concept of childcare quality is operationalized
by the pupil–staff ratio. The set of childcare
quality is then calibrated according to some
theoretical premises (relevance for child development,
relating to cognitive skills, etc.) and empirical
knowledge (some systems of childcare identified
previously as having high/low quality). The ‘real’,
empirical values – like, for example, six children per
teacher – is translated into a particular fuzzy set
score (in our analysis it is 1, thus fully in the set of
childcare quality).
Furthermore, the fundamental assumption of the
fuzzy set approach is that empirical phenomena
have a complex character. This configurational
character stems from the presumption that theoretical
constructs can rarely be described by one aspect
(characteristic), and only multidimensional treatment
of cases can provide their sufficient reflection
in the process of research. To give an example, the
analysis of a component of childcare policy, such as
parental leave, should focus on at least two features:
universality and generosity. These two dimensions
represent the constituting aspects of a given phenomenon
(here: parental leave). Together, the two
dimensions give four combinations: low universality
and high generosity, low universality and low generosity,
high universality and high generosity, high
universality and low generosity. The configurations
of the extreme values of sets (0 and 1) are treated as
Weberian ideal types. This example illustrates how
different configurations of sets and their extreme
values (0 and 1) yield different parental leave types.
All logically possible configurations (in the above
example, four) create so-called ‘property space’.
One can imagine the two-dimensional property
space as a square, three-dimensional as a cube and
so on. The property space is delineated by the lines
connecting its corners (i.e. configurations). If one
Who cares? Changing patterns of childcare in Central and Eastern Europe 119
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
treats corners of the property space (configuration)
as a Weberian ideal type, empirical cases will be
located within this space. Thus, empirical cases will
somehow conform to ideal types. The great benefit
of the fuzzy set ideal types approach is that it
enables the researcher to define what ideal type a
given case conforms most closely to. In this way, it
is a great step forward in modelling or typologizing
an empirical reality. We have derived the relevant
dimensions of childcare policy from the theory, and
then we view their configurations as ideal types.
Finally, thanks to the operations on sets, we are able
to determine to which ideal type of childcare policy
a ‘real policy’ conforms most closely to.
There are some rules which govern fuzzy sets and
their interpretation. The operations used in this
study are: logical negation as well as and. As already
said, since the phenomenon is present in a fuzzy set
to a different extent, it can, logically, be also nonpresent
to a different degree. The fuzzy membership
value of a case in set not-A is equal to 1-set A
(Ragin, 2000: 172). For example, if childcare policy
membership in the set of generosity G (set of generosity
described by letter G, where the capital letter
stands for the set) is 0.6, then this case scores 0.4 in
the set of non-generosity g (described by the letter g,
where the lower-case letter stands for the negation
of the set of generosity).
Logical and (here represented by the sign ∗) is
used when two or more sets are intersected; that is,
when a phenomenon is defined by two or more
dimensions and seen as a combination. As mentioned,
the configurations of extreme values of sets
are treated as Weberian ideal types. For example,
the two dimensions of universality and generosity
will give the following combinations: u∗g, u∗G,
U∗g, U∗G or in plain English low universality and
low generosity, low universality and high generosity,
etc. It is the minimum rule which governs this
operation. It means that the membership value of
such sets is equal to the lowest score achieved by any
of them (the degree they occur in set U and in set G
– their ‘common terrain’). Thus, if one takes into
account the combination of sets U∗G, and the case
scores 0.6 and 0.8 respectively, the membership of
this case in the intersection of sets U∗G is 0.6.
These operations on fuzzy sets allow for proper
and fast assessment of the cases’ membership in a
complex combination of sets. This is done by the
evaluation of a membership degree in the all logically
possible combinations of full membership and nonmembership
in the sets which create the property
space. Logically, the number of combinations equals
2k
where k stands for a number of dimensions. In our
case it means there are 16 possible combinations. The
idea of comparison is to contrast empirical evidence
with each combination representing an ideal type and
find the instance (the particular combination) with
the highest membership. The membership in the combination
also ranges from 0 to 1. If it reaches 1, the
characteristics of a studied policy are very close to the
ideal type. As this is a rather rare situation, this ideal
typical combination of policies serves here more as an
analytical tool. These properties of the approach offer
a great advantage when studying the dynamics of
childcare policy change: one studies the conformity of
policies to a variety of ideal-types repeatedly in subsequent
points of time. This way, one may trace
changes both within one type of policy, and shifts
from one policy type to another.
To sum up, the fuzzy set ideal types approach
offers great advantages when studying childcare
policy. It allows for transparent and systematic
measurement of policy, driven by clear conceptual
and theoretical guidelines. When applied to many
points of time (in this study, eight), it gives a strong
basis for describing trends. Finally, used in a crosscountry
comparison, it allows for clustering countries.
All three features of this approach are used in
this study.
The analysis of childcare policy will proceed as
follows. The four constituting dimensions of childcare
policy have already been identified. Next, the
calibration of sets according to theoretical and substantive
knowledge takes place, which serves as a
basis for translating data into fuzzy sets scores.
Calibration of the sets – defining
boundaries
This section defines and argues concretely what it
means for a policy to be extensive, of good quality,
generous, and accessible. Table 1 presents the calibration
for all the four dimensions.
Extensiveness of childcare services. The upper cutoff
point (fully in the set) is established at the level of
80 percent. This level is found to be typical of the
Scandinavian, universalist type of childcare policy,
characterized by almost full coverage of children in
the pre-primary education age (3–6) (Kvist, 1999).
120 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
Further, as Hantrais (2004) argues, such a high level
is justified when a significant percentage of women
are active in the labour market. As far as the crossover
point is concerned, it has been established at 50
percent of the net enrolment rate. Comparative
studies indicate that levels similar to this reflect
modest coverage (Daly and Rake, 2004). Finally, the
lower qualitative breakpoint for the extensiveness set
equals 20 percent of children enrolled in public preprimary
education. This level refers to very marginal
engagement of public authorities in the childcare
policies (Gornick and Meyers, 2003).
Quality of childcare services. The higher qualitative
point is set for the ratio 6 (six pupils per one teacher).
This number reflects the European standards for very
high-quality childcare, typical for the Scandinavian
countries (OECD, 2001; Gornick and Meyers, 2003).
The lower anchor is established at the pupil–teacher
ratio of 12. This level refers to the target set up by EU
experts, below which the quality should not drop (EC,
1996). Following OECD suggestions, the cross-over
point will be established at the ratio of nine pupils for
one teacher (OECD, 2001).
Generosity of parental leave provisions. Some
guidance can be taken from the comparative
research on parental leave (Gornick and Meyers,
2003; Daly and Rake, 2004). In sum, 12–16 weeks’
leave with some payment was usually considered a
minimum, with 14 weeks being an internationally
recognized standard.1
Second, the authors agree that
parental leave of one year means a relatively moderate
to high level of support. Third, the possibility to
extend the leave up to two or three years (with some
payment) has been considered as very generous.
Accordingly, we decided to establish the lower cutoff
point at the level of 14 weeks and 0.3 of the
average net wage. The cross-over point corresponds
to the moderate state support as discussed by Gornick
and Meyers (2003); that is, one year with the lower
support (0.3 of the average net wage) accompanied
by the system of generous maternity leave provision
(26 weeks and full wage replacement).
The communist countries were the pioneers in
introducing paid parental leave (the extended
leave), already available in the late 1960s and
1970s. For these reasons the upper cut-off point for
the analysis, expressing possibly the most generous
combination of the parental leave provision, comprises
the high standard of maternity leave (52
weeks paid in full) plus the three-year-long
extended leave with the flat-rate payment at the
level of 0.3 of the average net wage. All these provisions
are also available to a different degree. This
is tackled by the dimension of universality.
Universality of parental leave provisions. The following
approach is proposed here: one takes into
account the coverage and the conditions which are to
be met by the person eligible for the benefit as the
basis of the universality assessment. In other words,
which social groups are eligible for the parental leave
benefit is analysed as well as which situations are
taken into account as the basis for the entitlement.
The upper qualitative point is set for the situation
where every citizen (resident) is entitled to some basic
benefit and where access to it is not conditional.
The lower cut-off point is characterized by the
benefit which is provided by the state, but it is
means-tested (income-tested). Finally, the cross-over
point should reflect the situation where there is a
clear link between receiving the benefit and contribution.
Thus, in effect one has a triad of welfare provision:
it can be universal (fully in), contributory
(cross-over) and finally selective (fully out) (cf. Kvist,
1999). These three broad clusters were divided into
smaller parts in order to allow for more finely
grained analysis, encompassing the combinations of
principles and taking into account different scopes of
application within each cluster. Table 1 gathers the
calibration for these four dimensions.
Analysis of the empirical material,
dimension by dimension
This part of the article analyses the data alongside
the analytical anchors which were established
through calibration. We study the dynamics of
changes in each of the four dimensions. This manner
of presenting our findings allows for an insight into
the empirical results and serves as a convenient
starting point for the configurational analysis,
where the article clusters the countries around some
ideal types. Afterwards each country is considered
separately in order to observe the most important
changes in each of these dimensions.
As Table 2 shows, the overall balance of changes
in the extensiveness of childcare services for children
at the age of 3 to 6 has been positive for all the countries,
apart from Slovakia. Naturally, the countries
started from different initial points and can generally
be divided into two groups. The first group, with
Who cares? Changing patterns of childcare in Central and Eastern Europe 121
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
quite strong membership, comprised such countries
as the Czech Republic, Slovakia and Hungary.
Estonia, with a modest membership in the set, lay
somewhere in the middle, while the other Baltic
counties, Slovenia and Poland (with the weakest
membership score: mostly but not fully out), were at
the other end. Not all the differences held for the
next 15 years. While Hungary kept full membership
for the whole period, the Czech Republic experienced
some decline at the beginning of the 1990s.
Simultaneously, Slovakia fell even below the crossover
point in the mid-1990s and never returned to
the 1989 level. The first half of the 1990s can be
characterized by an overall decline which was
strongest in Lithuania, Latvia and Poland. After that
period, several countries noted a considerable
improvement, the greatest being observed in cases
like that of Latvia, where at the beginning of the
122 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
Table 1 Specification of empirical indicators and translation of data to fuzzy score ranges and verbal qualifiers:
childcare policy
Mostly More or Mostly
but not less fully Neither in More or but not
Fully out fully out out nor out less in fully in Fully in
Empirical indicator 0 0.01–0.24 0.25–0.49 0.50 0.51–0.75 0.76–0.99 1.00
Extensiveness of <20% 20.10– 35.10– 50% 50.10– 65.10– >80%
the family policy 35.00 49.90 65.00 79.90
measured as the
enrolment rate of
children in
kindergartens
Quality of the >12.00 10.51– 9.10– 9.00 7.51– 6.10– <6.00
family policy 11.99 10.50 8.99 7.50
measured as a
child-to-staff ratio
in the kindergartens
Generosity of <4.20 4.30– 21.60– 41.60 41.70– 69.60– >98.80
maternity-related 21.50 41.50 69.50 98.70
benefits measured
by the index based
on weighted
replacement rate
and duration
of benefitsa
Universality of Means-test, Insurance, Insurance, Insurance, Universal, Universal, Universal,
parenthood-related Means-testb
Means-test Insurance Insurance Means-test Insurance Universal
benefits ( maternity (no conditions)
and parental
benefits) measured
as the combination
of principles
guiding the access
(residual, selective
and universal)
Notes:
a
Calculated according to the formula given in the text: the number of weeks paid maternity leave times the replacement
rate plus the same but in relation to the extended leave (childcare leave, etc.). For the lower cut-off point the
index is 14 × 0.3 = 4.2, for the upper cut-off point it is (52 × 1) + (156 × 0.3) = 98.8.
b
First: maternity leave; second: parental leave (extended, child care, etc.).
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
1990s enrolment rates fell below the level of 20
percent (fully out of the set), but then rose up to
almost 80 percent in 2002 and 2004 (and reached
full set membership). Similarly, the enrolment rates
rose in Estonia, though the situation there in 1989
was already better than in other Baltic countries.
The dynamics of changes within the dimension of
the quality (see Table 3) of childcare services is
hardly regular or linear. The clearest case of low
quality is Poland, with its quality of services always
below the threshold level, but the Czech Republic
can also be characterized by relatively stable, low
standards. The group of Baltic countries, however,
are the leaders in this respect and this has also been
almost constant. Slovakia and Slovenia had more
irregularities and shifts. While in 1993 in Slovenia
there were fewer than five pupils per teacher in
kindergartens, from 1999 on, this country scores 0.
Hungary, though typically the leading case in the
remaining three dimensions, here stays below the
cut-off point of 12 pupils per teacher, and only
almost achieves the cross-over point in 2004 (10
pupils per teacher). In general, the quality of childcare
is not high for this group of countries.
Expressing the generosity (see Table 4) of the
parental leave provisions was a more complex
Who cares? Changing patterns of childcare in Central and Eastern Europe 123
Journal of European Social Policy 2008 18 (2)
Table 2 Extensiveness of childcare services: fuzzy set membership
Country/year 1989 1991 1993 1995 1997 1999 2002 2004
Czech Rep. 1.00 0.76 0.87 0.93 0.97 1.00 1.00 1.00
Estonia 0.56 0.52 0.41 0.62 0.81 0.90 1.00 1.00
Hungary 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Latvia 0.34 0.00 0.00 0.18 0.32 0.56 0.94 0.98
Lithuania 0.34 0.20 0.00 0.00 0.08 0.25 0.32 0.58
Poland 0.21 0.08 0.07 0.13 0.19 0.24 0.28 0.33
Slovakia 0.93 0.81 0.58 0.43 0.63 0.74 0.77 0.71
Slovenia 0.40 0.39 0.52 0.63 0.66 0.75 0.61 0.54
Source: Data sources available from the authors.
Table 3 Quality of childcare services: fuzzy set membership
Country/year 1989 1991 1993 1995 1997 1999 2002 2004
Czech Rep. 0.00 0.28 0.00 0.08 0.11 0.00 0.00 0.00
Estonia 0.71 0.94 0.84 0.89 0.94 0.71 0.71 0.71
Hungary 0.08 0.02 0.00 0.00 0.00 0.00 0.49 0.49
Latvia 0.71 0.72 0.82 0.73 0.73 0.74 0.61 0.23
Lithuania 0.71 0.99 1.00 0.84 0.84 0.84 0.71 0.71
Poland 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Slovakia 0.00 0.25 0.80 0.30 0.30 0.48 0.48 0.00
Slovenia 0.55 0.46 1.00 0.25 0.52 0.00 0.00 0.00
Source: Data sources available from the authors.
Table 4 Generosity of parental leave provision: fuzzy set membership
Country/year 1989 1991 1993 1995 1997 1999 2002 2004
Czech Rep. 0.62 0.67 0.57 0.60 0.79 0.70 0.78 0.87
Estonia 0.29 0.59 0.47 0.35 0.35 0.42 0.43 0.41
Hungary 1.00 1.00 1.00 1.00 0.94 1.00 1.00 1.00
Latvia 0.29 0.44 0.56 0.41 0.37 0.34 0.32 0.30
Lithuania 0.29 0.54 0.89 0.67 0.72 0.70 0.69 0.71
Poland 0.26 0.49 0.48 0.43 0.46 0.49 0.49 0.52
Slovakia 0.62 0.63 0.62 0.55 0.66 0.65 0.59 0.60
Slovenia 0.15 0.15 0.73 0.74 0.72 0.74 0.69 0.68
Source: Data sources available from the authors.
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
task, since a composite index needed to be developed
on the level of ‘raw’ data. There was also a
greater variety in the scores. Only one country,
Hungary, reached full membership. About four
decades ago, Hungary already had the most generous
family support systems in the world (Szikra,
2005). The earnings-related payment (childcare
fee) lasts for two years, which is accompanied by
the scheme for non-insured people who take care
of a child under three years of age. These provisions
were seriously cut in 1996 for a couple of
years: the insurance-based extended leave was
abolished, and the access to the flat-rate scheme
was restricted by the introduction of the principle
of income-testing. Another country which has
scored relatively highly was the Czech Republic,
which already had quite a generous system in the
1990s. Importantly, this country introduced a universal
parental leave scheme in the middle of the
1990s, which was also extended up to four years,
making it the longest extended leave in the world.
Slovakia, however, with the similar level of payments,
kept the duration of parental leave at the
level of three years.
A considerable group of countries has improved
its situation since 1989. For example, the Baltic
states, after achieving independence, introduced
extended (parental) leave for up to three years, as
well as the possibility of extending the insurancebased
leave by up to one year in the case of Lithuania
and Estonia (Aidukaite
.
, 2004). The differences in
overall scores for the three countries are also driven
by the levels of payments for parental leave – the
replacement rate has never been more than about
5–10 percent in Latvia, while in Lithuania it ranged
from 10–15 percent. The latter country also has two
more weeks of maternity leave.
Slovenia improved its scheme in 1993. Though it
included only two years of paid parental leave, the
payments for the non-insured accounted for
around 30 percent of the net average wage. Only
recently, when the level of payment was established
as a flat-rate benefit (previously it was
related to the minimum wage), the replacement
rate became less generous (around 20%).
However, the overall score is relatively high thanks
to 260–365 days of maternity leave with a full
wage replacement. Finally, in Poland the rise in
the value of parental leave benefit has been quite
constant.
In the case of universality (see Table 5) of
parental leave, the overall tendency has been to
open access to the benefit to all families with
small children. As described above, in many
countries there is more than one system; for
instance, for the insured and the non-insured. As
mentioned above, Hungary had a period of
means-testing in the case of extended leave in the
second half of the 1990s. The countries that
introduced parental leave at the beginning of the
1990s (such as Latvia and Slovenia) made it a universal
right. In contrast, Estonia made the insurance-based
benefit available for more and more
groups over the years until the universal childcare
allowance was introduced in 2002. In Poland the
situation did not change from 1983, when the
parental benefit was introduced – it remained
income-tested. Poland also has no universal
system of financial support for childcare; maternity
benefit is paid only under the condition of
previous employment. The next section takes the
analysis a step further to make some more general
conclusions about the changes in the combinations
reflecting policy types.
124 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
Table 5 Universality of parental leave provision: fuzzy set membership
Country/year 1989 1991 1993 1995 1997 1999 2002 2004
Czech Rep. 0.40 0.40 0.40 0.40 1.00 1.00 1.00 1.00
Estonia 0.30 0.55 0.80 0.80 0.75 0.85 0.85 1.00
Hungary 0.85 1.00 1.00 1.00 0.30 1.00 1.00 1.00
Latvia 0.30 0.40 1.00 1.00 1.00 1.00 1.00 1.00
Lithuania 0.35 0.35 0.40 1.00 1.00 1.00 1.00 1.00
Poland 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20
Slovakia 0.40 0.40 0.40 0.40 1.00 1.00 1.00 1.00
Slovenia 0.40 0.40 1.00 1.00 1.00 1.00 1.00 1.00
Source: Data sources available from the authors.
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
Childcare policy – configurational
analysis
It should be recalled that it is not enough to take
into account one dimension of policy to define its
location in a given cluster because members of the
cluster may be very similar in all but one respect.
Therefore the policies should be analysed as configurations
of aspects.
In this article the four dimensions of childcare
policy are taken into account. This gives 16 possible
combinations. Membership of each combination of
sets is assessed and the highest value is chosen. Table
6 presents the highest scores of the childcare policy
dimensions for each country in selected years. It
should be noted that when the article refers to
Estonia, Lithuania and Latvia during the period
1989–91, these countries were republics of the Soviet
Union, just as the Czech Republic and Slovakia were
Czechoslovakia until 1993, and Slovenia belonged to
Yugoslavia (1989–91).
As far as the highest scores are concerned, the policies
conformed to 13 ideal types. The highest scores
for each country and year are depicted in Table 6.
The results in Table 6 should be read as
follows. Combinations of letters illustrate the
policy type, while the number in parentheses
describes conformity to the ideal type (the higher
the number, the closer the combination is to the
ideal type). For example, Hungarian policy in
1989 belonged to the type E∗q∗G∗U, which
means that the policy was characterized by high
extensiveness (E) and low quality (q) of childcare
services and the generous (G) and universal (U)
nature of parental leave benefits. The high
number reflects the close conformity of the policy
to this ideal type. As mentioned above, the fuzzy
sets approach is sensible to two modes of change:
from one type to another (difference in kind), and
change in the conformity to the ideal type (difference
in degree). The latter can be illustrated by
the change of score for the same combination
(e.g. Hungary scored 0.85 in 1989 and 0.88 in
1991 for the combination E∗q∗G∗U, which
shows more conformity to the ideal type in 1991).
The qualitative change takes place when policy
shifts to a different type (as represented by
another combination).
Some combinations occurred more often than
others, some of them occurred only once. The
advantage of the methodological approach used
here is that instead of trying to fit the policies to preexisting
policy types, we can actually read them out
of the table presented, which means that the new
policy types may emerge. In this way the article provides
answers to the theoretical problems of applying
the concepts invented in the West to the cases of
post-communist countries. In order to make the
conclusions from the comparison even more systematic,
four clusters are distinguished and labelled.
The main argument of this article centres on the
issue of care arrangement by answering two main
Who cares? Changing patterns of childcare in Central and Eastern Europe 125
Journal of European Social Policy 2008 18 (2)
Table 6 Fuzzy membership scores of childcare policy in ideal types
Year 1989 1991 1993 1995 1997 1999 2002 2004
Czech Rep. E*q*G*u E*q*G*u E*q*G*u E*q*G*u E*q*G*U E*q*G*U E*q*G*U E*q*G*U
(0.60) (0.06) (0.57) (0.60) (0.79) (0.70) (0.78) (0.87)
Estonia E*Q*g*u E*Q*G*U e*Q*g*U E*Q*g*U E*Q*g*U E*Q*g*U E*Q*g*U E*Q*g*U
(0.56) (0.52) (0.53) (0.62) (0.65) (0.58) (0.57) (0.59)
Hungary E*q*G*U E*q*G*U E*q*G*U E*q*G*U E*q*G*u E*q*G*U E*q*G*U E*q*G*U
(0.85) (0.88) (0.87) (0.87) (0.70) (1.00) (0.51) (0.51)
Latvia e*Q*g*u E*Q*g*u e*Q*G*U e*Q*g*U e*Q*g*U E*Q*g*U E*Q*g*U E*q*g*U
(0.66) (0.56) (0.82) (0.59) (0.63) (0.56) (0.61) (0.70)
Lithuania e*Q*g*u E*Q*G*u e*Q*G*u e*Q*G*U e*Q*G*U e*Q*G*U e*Q*G*U E*Q*G*U
(0.71) (0.54) (0.60) (0.67) (0.72) (0.70) (0.68) (0.58)
Poland e*q*g*u E*q*g*u e*q*g*u e*q*g*u e*q*g*u e*q*g*u e*q*g*u e*q*G*u
(0.74) (0.51) (0.52) (0.57) (0.54) (0.51) (0.51) (0.58)
Slovakia E*q*G*u E*q*G*u E*Q*G*u e*q*G*u E*q*G*U E*q*G*U E*q*G*U E*q*G*U
(0.60) (0.60) (0.60) (0.55) (0.63) (0.52) (0.52) (0.60)
Slovenia e*Q*g*u E*q*g*u E*q*G*U E*q*G*U E*Q*G*U E*q*G*U E*q*G*U E*q*G*U
(0.55) (0.54) (0.52) (0.63) (0.52) (0.74) (0.61) (0.54)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
questions: first, where the locus of responsibility for
care is; and second, how this facilitates women’s
access to commodification. The first dimension,
therefore, asks about the degree of defamilialization:
how much the ‘caring function of family’
(Leitner, 2003) is encouraged by the policy measures.
By adding the second point of reference we
intend to move on from a simple defamilialization
approach and underline how different state policies
facilitate access to paid work. In this way, while the
content of the policies we took into consideration
would allow for describing the combinations in the
language of defamilialization, we decided to put
more emphasis on the more ultimate aim and the
potential result of childcare policies, which is how
the state treats the problem of female access to commodification
in situations where women are still
perceived as the primary carers.
This article demonstrates that instead of a unified
tendency towards familialization of policies, many of
the post-communist countries followed different
paths of familialization while some of them strengthened
the defamilializing components of their policies.
Our classification partially overlaps with some
of the most recent theoretical accounts of childcare
policy analyses (Leitner, 2003) including the works
on CEE (Saxonberg and Sirovatka, 2006). What is
innovative in our case is that we tend to think about
defamilialization versus familialism (and furthermore,
its different versions) also as different types of
state policy towards the labour force. This is especially
important in the context of Eastern Europe,
where after 1989 unemployment was a new phenomenon.
As many authors argue, refamilializing
policies served as an important instrument of dealing
with the unemployment problem; burdened with
care, women were less able to join (or re-enter) the
labour force. For all the reasons stated above, we
decided to distinguish between four policy-types:
implicit and explicit familialism, comprehensive
support and female mobilizing respectively.
There are two groups of countries where the mix of
childcare policies locates the responsibility for care
within the family. The first of them can be clustered
under the label of explicit familialism (E∗q∗G∗U), and
comprises the Czech Republic, Slovakia and Slovenia.
In this model the state pursues some more active
policies to support the traditional family model. Thus,
the periods of paid parental leave are longer (usually
two to four years), and the same is true for the length
of parental leave payments. Due to the promotion of
private care, the state does not subsidize any form of
childcare centres. Women are perceived through their
role as carers; there are incentives for exercising the
task of care at home (long periods of paid leave), and
even if maternity leave is over, the support lasts for a
long time (extended leave). The ideal combination for
this type of policies would be: low scores on the extensiveness
and quality of childcare services, and high on
both generosity and universality (e∗q∗G∗U). The
combination closest to this, which repeated several
times, is the same albeit for a high score on extensiveness
of childcare services. Low quality of childcare
accompanied by extensive support for domestic
care during the first years of a child’s life suggests
shifting the locus of care to family with less concern
about the commodification of potential earners (that
is to say, women).
The second familialistic combination is implicit
familialism. This model only truly happened to be
the case in Poland, and is represented by a low score
in all the four dimensions (e∗q∗g∗u). The policies
are residual and formally neutral, with the assumption
that family should not be interrupted in its task
of educating children. They neither suggest the locus
of responsibility for care, nor do they explicitly
mobilize women to join the workforce. However,
due to gender inequalities in the labour market, the
lack of affordable and available childcare (provided
by the market, for instance) leaves the sphere of care
almost solely to families.
The third group of cases comprises the countries
with low generosity of parental leave provisions
with the rest of the dimensions scoring highly
(E∗Q∗g∗U). In this cluster, not only is the rate of
children in the childcare services high, but so is the
quality of services. Parents might even have the universal
right for using the parental leave, but since
the provision is not on a very generous level, the
incentive to do so might not be encouraging enough
for parents (i.e. women) to resign from paid employment.
This is the case of some of the scores of two
Baltic countries – Estonia and Latvia – but at the end
of the period under review. The label for this type is
female-mobilizing, due to the small effect of
parental leave support on locating the responsibility
for care within the family and for more concern with
supporting the access to commodification.
Finally, when policy scores high on all the dimensions
(E∗Q∗G∗U), the family is receiving support on
126 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
a couple of fronts. The patterns of combining paid
employment and childcare can be chosen by the
families individually. However, due to the availability
of high-quality childcare services, and when
there is a need for two incomes within the family,
this combination of the policies generally supports
families with dual earners. This is very rare in the
case of this group of countries and only Lithuania
and Hungary approach this model. Here the families
confronted with the ‘childcare problem’ have
more alternatives at hand, since the policies are
more diversified and rather multipurpose. As a
result, we decided to label this combination of policies
as comprehensive support. In other words, families
and women are both paid and relieved in their
care responsibilities, and this also leaves some space
for mobilizing two earners within the family.
Though the previous cluster (female mobilizing)
also supports the dual-earner family, the mobilizing
function there is underlined by the lack of generous
incentives for providing the childcare at home.
Below, each of the countries is briefly described
with reference to the dynamics of policy transformation
and its consolidation into a particular model.
Childcare policies – analysis of the
countries’ trajectories
In the period 1989–95, the Czech Republic
remained in the cluster characterized by generous
parental leave benefits (but of narrow scope) as well
as the high extensiveness and low quality of the
childcare (female mobilizing). The membership
scores in this period were moderate. The childcare
policy in 1997–2004 shifted to the cluster sharing
the characteristics of the previous one, except for
high universality (explicit familialism). As for the
membership strength, it grew from 1999 and at the
end of the period the Czech family policy conformed
closely to this ideal type.
The policy pursued by the Slovak Republic after
the dissolution of Czechoslovakia was to some
extent similar to the Czech one. Thus, at the beginning
of the analysed period the countries conformed
exactly to the same ideal type (female mobilizing).
In 1993 and 1995 their paths diverged to come
together in the period 1997–2004 when the childcare
policy of Slovakia shifted to the explicit familialism
model. Even though the conformity to the
ideal types was weak (which may lead to ambiguities
in interpretation), the policy pursued in this
country resembled the broad characteristics of the
specified model.
The childcare policy of Slovenia was also changing:
from 1989–91 it was typical for the implicit
familialism ideal type. From 1995 on, the policy was
more typical for the explicit familialism model.
From 1999, however, the policy conformity to this
ideal type decreased.
As far as the Polish childcare policy is concerned,
one may observe the following trends. For most of
the time Poland belonged to one cluster, characterized
by low extensiveness and quality of services
combined with not generous and hardly accessible
benefits (e*q*g*u). However, Poland’s membership
in this was generally low from 1993. Though in
2004 the policy shifted to the cluster with high generosity
(e∗q∗g∗u), the developments observed
through the whole period allow for placing this type
of policy in the implicit familialism cluster.
In Estonia the changes in childcare policy were
more frequent and profound. In 1991 Estonia’s
policy belonged to the cluster of high extensiveness
and quality of the childcare combined with low generosity
and universality of benefits (E∗Q∗g∗u), and
then moved to the type of high extensiveness,
quality, generosity and universality (E∗Q∗G∗U).
From 1995 on, Estonia clustered in the type characterized
by high extensiveness and quality, universal
but not generous benefits (E∗Q∗g∗U). In the case of
Estonia, the membership values were small, so it is
hard to classify the early trends. Since 1995, the
policy resembled a female mobilizing type of policy.
The second Baltic country, Latvia, pursued a
childcare policy characterized by the lack of continuity.
The period 1989–91 featured high extensiveness
and high quality of services, but not generous
and not universal benefits (E∗Q∗g∗u). In 1993, it
belonged to the cluster which differed from the previous
one in high generosity (E∗Q∗G∗u). In two
subsequent periods, Latvian childcare policy moved
to the type characterized by low extensiveness and
high quality of services and neither generous nor
universal benefits (e∗Q∗g∗u). In the period from
1999–2002, the country changed its position again.
As far as the Latvian membership values are concerned,
childcare policy conformed to the ideal
types to different degrees: in 1989, 1993 and 2004
the policy was mostly but not fully in, while in the
remaining years it was more or less in. All these
Who cares? Changing patterns of childcare in Central and Eastern Europe 127
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
128 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
Table7ClassificationofchildcarepoliciesinCentralandEasternEurope
Year19891991199319951997199920022004
CzechFemaleFemaleFemaleFemaleExplicitExplicitExplicitExplicit
mobilizingmobilizingmobilizingmobilizingfamilialismfamilialismfamilialismfamilialism
EstoniaE*Q*g*uE*Q*G*Ue*Q*g*UFemaleFemaleFemaleFemaleFemale
mobilizingmobilizingmobilizingmobilizingmobilizing
HungaryExplicitExplicitExplicitExplicitE*q*G*uExplicitComprehensiveComprehensive
familialismfamilialismfamilialismfamilialismfamilialismsupportsupport
Latviae*Q*g*uE*Q*g*ue*Q*G*Ue*Q*g*Ue*Q*g*UFemaleFemaleFemale
mobilizingmobilizingmobilizing
Lithuaniae*Q*g*uE*Q*G*ue*Q*G*uComprehensiveComprehensiveComprehensiveComprehensiveComprehensive
supportsupportsupportsupportsupport
PolandImplicitImplicitImplicitImplicitImplicitImplicitImplicitImplicit
familialismfamilialismfamilialismfamilialismfamilialismfamilialismfamilialismfamilialism
SlovakiaFemaleFemaleE*Q*G*ue*q*G*uExplicitExplicitExplicitExplicit
mobilizingmobilizingfamilialismfamilialismfamilialismfamilialism
SloveniaImplicitImplicitE*q*G*UExplicitExplicitExplicitExplicitExplicit
familialismfamilialismfamilialismfamilialismfamilialismfamilialismfamilialism
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
complexities seem to allow for placing the Latvian
childcare policy in the female mobilizing type, but
only in the period 1999–2004.
The last Baltic country, Lithuania, also resembles
the pattern typical for this group. Thus, the childcare
policy was changing clusters frequently. The
conformity of the childcare policy to the ideal types
varied, but generally it was modest. Here again
it is possible to define a type of policy only for the
second half of the 1990s. Lithuania pursued the
variant of a comprehensive support model of childcare
policy.
Meanwhile, Hungary remained in the policy type
featuring encompassing childcare of low quality
combined with the generous and accessible parental
leave benefits for the whole of the analysed period
(except for 1997 due to cuts in family policies). The
membership scores remained high and stable until
1995 and in 1999 (fully in, the only such case). In
1997 and in 2002–04 the scores were lower,
especially in the latter cases, where the values were
close to the cross-over point. This raises ambiguity,
and it seems more relevant to talk about the explicit
familialism model in 1989–97; and then from
2002, the comprehensive support type of childcare
policy. This choice stems from the significant
increase in the quality of childcare, moving
Hungary towards this cluster. Importantly, both
Hungary and Lithuania conform very weakly to the
comprehensive support type.
Table 7 shows the occurrence of cases as attached
to particular childcare policy types. As mentioned,
to some of the combinations which occurred only
accidentally, none of the labels is attached.
Conclusions
This analysis began with the theoretical discussion
concerning the possibility and necessity of comparing
childcare policies in the new EU member states.
After identifying the main issues with the research
which had been conducted so far, we argued that the
fuzzy set approach to comparative studies is useful
for solving the problems with systematic comparisons
of post-communist social policies. After
describing and justifying the calibration of the sets,
we presented our empirical findings and then followed
this by an analysis of dynamics of change.
This has served as a basis for clustering the countries
under suitable analytical terms.
As was mentioned in the theoretical part of this
article, previous studies dealing with welfare policies
in CEE either treated this group of countries as
monolithic, or applied Western frameworks for
comparison, which often led to contradictory
results. The great advantage of the fuzzy set ideal
type approach is that it allows for an analysis which
is non-biased by existing frameworks as it opens
the way for discovering new types of policies. As a
result, it was possible to present the scope of diversity
among this group of countries. One of the most
general observations is, for instance, the difference
between Poland and Hungary. In Poland the provision
of public childcare services is very poor, and
the state also does not offer much financial help for
parents caring for the small child. This can be contrasted
to the case of Hungary which has good
availability to childcare centres of relatively high
quality accompanied by the generous schemes of
parental leave.
In order to encompass the differences in the broad
picture, this article identified four models of childcare
policy. These were: explicit familialism,
implicit familialism, female-mobilizing and comprehensive
support. The countries were clustered as
follows: the Czech Republic, Slovakia and Slovenia
comprised the group of countries under the label of
the explicit familialism model. Estonia and Latvia
represented the female mobilizing policy type.
Lithuania and Hungary pursued the childcare policy
typical for the comprehensive support model. The
childcare policy in Poland resembled the implicit
familialism model.
Furthermore, the policies are marked by many
qualitative and quantitative changes. This is especially
visible in the case of the states that emerged
after the collapse of the Soviet block (especially in the
case of the Baltic states). Accordingly, in the initial
period of transition the policies can be characterized
by many shifts, resulting from various adjustments
and ‘trial and error’ strategies (Offe and Preuss,
1998). The crucial observation here is that for the
later stage of transition, one may talk about the consolidation
of policies – somewhere in the middle of
the period in which changes started to be less rapid
or the policies stabilized. It refers not only to the classification
of the policy to one policy type, but to
some extent also to the degree of conformity to this
one ideal type. Three out of the eight countries not
only remained in one policy type for the last three to
Who cares? Changing patterns of childcare in Central and Eastern Europe 129
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
four points in time, but also improved their conformity
to that particular ideal type.
The countries differed also in their tendencies
towards hybridization. While in the case of the
Czech Republic, for example, the degree of conformity
to the explicit familialism type rose from
0.79 in 1997 to 0.87 in 2004, for Slovenia it was
0.74 in 1999 and only 0.54 in 2004. While the
Slovenian childcare policy – as well as in some
other cases – may still be prone to change, the
overall tendency towards stability is promising.
This observed stability, in turn, opens up the possibility
of analysing welfare policies of the region
and deriving more time-enduring generalizations.
This article takes a first step in this process and
contributes to the debate by demonstrating
a variety of policy models in CEE, contrary to
conventional expectations.
Acknowledgements
We would like to thank to Jon Kvist and Olli
Kangas who generously shared their experience
with us. Anonymous referees provided us with comments
and hints which improved the quality of this
article significantly. As usual, all errors remain ours.
Note
1 Council Directive 92/85/EEC of 19 October 1992,
accessed on 5 June 2007 at: [http://europa.eu.int/eur-
lex/lex/LexUriServ/LexUriServ.do?uri=CELEX:31992
L0085:EN:HTML].
References
Aidukaite
.
, J. (2004) The Emergence of the Post-Socialist
Welfare State: the Case of the Baltic States: Estonia,
Latvia and Lithuania. Stockholm: Södertörns
Högskola.
Clasen, J. and van Oorshot, W. (2002) ‘Changing Principles
in European Social Security’, European Journal of Social
Security 4 (2): 89–114.
Daly, M. (1994) ‘Comparing Welfare States: Towards a
Gender Friendly Approach’, in D. Sainsbury (ed.)
Gendering Welfare States. London, Thousand Oaks,
CA and New Delhi: Sage.
Daly, M. and Rake, K. (2004) Gender and the Welfare
State. Cambridge: Polity Press.
Deacon, B. (1993) ‘Developments in Eastern European Social
Policy’, in C. Jones (ed.) New Perspectives on the Welfare
State in Europe. London and New York: Routledge.
Domanski, H. (1992) Zadowolony Niewolnik? Studium
O Nierownosciach Miedzy Mezczyznami I Kobietami
W Polsce. Warsaw: Instytut Filozofii I Socjologii, Polska
Akademia Nauk.
EC (1996) Quality Targets in Services for Young Children:
Proposals for a Ten Year Action Programme.
Luxembourg: European Commission.
Esping-Andersen, G. (1990) The Three Worlds of Welfare
Capitalism. London: Polity Press.
Esping-Andersen, G. (ed.) (1996) Welfare States in
Transition. London, Thousand Oaks, CA and New
Dellhi: Sage.
Esping-Andersen, G. (1999) Social Foundations of
Postindustrial Economies. Oxford: Oxford University
Press.
Ferge, Z. (1992) ‘Social Policy Regimes and Social
Structure. Hypothesis about the Prospects of Social
Policy in Central-Eastern Europe’, in Z. Ferge and I. E.
Kolberg (eds) Social Policy in a Changing Europe.
Boulder, CO: Campus and Westview.
Fodor, E., Glass, C., Kawachi, J. and Popescu, L. (2002)
‘Family Policies and Gender in Hungary, Poland and
Romania’, Communist and Post-Communist Studies
35: 475–90.
Gauthier, A. H. (1996) The State and the Family: a
Comparative Analysis of Family Policies in
Industrialized Countries. Oxford: Clarendon Press.
Gornick, J. C. and Meyers, M. K. (2003) Families that
Work. Policies for Reconciling Parenthood and
Employment. New York: Russell Sage Foundation.
Haney, L. (2002) Inventing the Needy: Gender and the
Politics of Welfare in Hungary. Berkeley: University of
California Press.
Haney, L. (2003) ‘Welfare Reform with a Familial Face.
Families of a New World’, in L. Haney and L. Pollard
(eds) Gender, Politics, and State Development in a
Global Context, pp. 159–78. New York and London:
Routledge.
Hantrais, L. (2004) Family Policy Matters: Responding to
Family Change in Europe. Bristol: Policy.
Harding, L. F. (1996) Family, State and Social Policy.
Basingstoke and London: Macmillan.
Heinen, J. (1997) ‘Public/Private. Gender, Social and
Political Citizenship in Eastern Europe’, Theory and
Society 26 (4): 577–97.
Jenson, J. and Sineau, M. (2001) Who Cares? Women’s
Work, Childcare, and Welfare State Redesign. Toronto:
University of Toronto Press.
Kamerman, S. (2003) ‘Welfare States, Family Policies, and
Early Childhood Education, Care and Family Support’,
Consultation Meeting on Family Support Policy in
Central and Eastern Europe. Budapest: UNESCO and
Council of Europe.
Kamerman, S. and Kahn, A. J. (1981) childcare, Family
Benefits, and Working Parents: a Study in Comparative
Policy. New York: Columbia University Press.
Ksiezopolski, M. (1999) Polityka Spoleczna. Wybrane
Problemy Porownan Miedzynarodowych. Katowice:
Slask.
Kvist, J. (1999) ‘Welfare Reform in the Nordic Countries
in the 1990s: Using Fuzzy-set Theory to Assess
Conformity to Ideal Types’, Journal of European Social
Policy 9 (3): 231–52.
Kvist, J. (2006) ‘Diversity, Ideal Types and Fuzzy Sets in
Comparative Welfare State Research’, in B. Rihoux and
130 Szelewa and Polakowski
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from
H. Grimm (eds) Innovative Comparative Methods for
Policy Analysis: Beyond the Quantitative–Qualitative
Divide, pp. 167–84. New York: Springer.
Lazarsfeld, P. (1937) ‘Some Remarks on Typological
Procedures in Social Sciences’, Zeitschrift Fur
Socialforschung 6: 119–39.
Leitner, S. (2003) ‘Varieties of Familialism. The Caring
Function of the Family in Comparative Perspective’,
European Societies 5 (4): 353–75.
Leitner, S. and Lessenich, S. (2005) ‘(In-)Dependence and
Dependent Variable: Conceptualising and Measuring
“De-familisation”. Exploring the Dynamics of Reform:
the Dependent Variable Problem in Comparative
Welfare State Analysis’, paper presented at the conference
on Exploring the Dynamics of Reform (May),
University of Stirling, UK.
Lewis, J. (1992) ‘Gender and the Development of Welfare
State Regimes’, Journal of European Social Policy 2 (3):
159–73.
Mahon, R. (2002) ‘Childcare: Toward What Kind of
“Social Europe?”’, Social Politics 9 (3): 343–79.
OECD (2001) Starting Strong: Early Childhood Education
and Care. Paris: OECD.
Offe, C. and Preuss, U. K. (1998) Institutional Design in
Post-Communist Societies: Rebuilding the Ship at Sea.
Cambridge and New York: Cambridge University Press.
Orloff, A. S. (1993) ‘Gender and the Social Rights of
Citizenship: the Comparative Analysis of Gender
Relations and Welfare States’, American Sociological
Review 58: 303–28.
Pascall, G. and Lewis, J. (2004) ‘Emerging Gender
Regimes and Policies for Gender Equality in a Wider
Europe’, Journal of European Social Policy 33 (3):
373–94.
Ragin, C. (2000) Fuzzy-set Social Science. Chicago, IL and
London: The University of Chicago Press.
Rostgaard, T. (2004) Family Support Policy in Central and
Eastern Europe – a Decade and a Half of Transition.
Paris: UNESCO.
Sainsbury, D. (1994) ‘Women’s and Men’s Social
Rights: Gendering Dimensions of Welfare States’,
in D. Sainsbury (ed.) Gendering Welfare States,
pp. 150–69. London, Thousand Oaks, CA and New
Delhi: Sage.
Saxonberg, S. and Sirovatka, T. (2006) ‘Failing Family
Policies in Eastern Europe’, Journal of Comparative
Policy Analysis 8 (2): 185–202.
Szikra, D. (2005) ‘Family and Child Support in a
Postcommunist Society: Origins of the Mixed
Hungarian Welfare Capitalism’, in M. Cain, N. Gelazis
and T. Inglot (eds) Fighting Poverty and Reforming
Social Security: What can Post-Soviet States Learn
from the New Democracies of Central Europe?
Washington, DC: Woodrow Wilson International
Center.
Zimmermann, S. L. (1995) Understanding Family Policy.
London, Thousand Oaks, CA and New Delhi: Sage.
Who cares? Changing patterns of childcare in Central and Eastern Europe 131
Journal of European Social Policy 2008 18 (2)
at Masarykova Univerzita on February 16, 2009http://esp.sagepub.comDownloaded from