Journal of Social Service Research

ISSN: 0148-8376 (Print) 1540-7314 (Online) Journal homepage: https://www.tandfonline.com/loi/wssr20

Re/Familialization or De-Familialization:
Differences in the Effects of Family Policies in
Terms of Stratification

Kyo-seong Kim & Eun Jung Hwang

To cite this article: Kyo-seong Kim & Eun Jung Hwang (2019): Re/Familialization or De-
Familialization: Differences in the Effects of Family Policies in Terms of Stratification, Journal of
Social Service Research, DOI: 10.1080/01488376.2019.1693477

To link to this article: https://doi.org/10.1080/01488376.2019.1693477

Published online: 25 Nov 2019.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=wssr20
JOURNAL OF SOCIAL SERVICE RESEARCH
https://doi.org/10.1080/01488376.2019.1693477

Re/Familialization or De-Familialization: Differences in the Effects of Family

Policies in Terms of Stratification
Kyo-seong Kim and Eun Jung Hwang
College of Social Science, School of Social Welfare, Chung-Ang University, Seoul, Korea

ABSTRACT KEYWORDS
This study examined the re/familialization and de-familialization effects of family policy types Family policy; re/
(in-cash or in-kind), and analyzed differences in the effects in terms of stratification. The familialization; de-
family policies of 27 Organization for Economic Co-operation and Development (OECD) familialization; stratifica-
tion; HGLM
countries were classified into in-cash and in-kind policies, and the effects of such policies on
women with children in each country were investigated. Data were obtained from the
International Social Survey Program (ISSP) and the OECD Social Expenditure Database
(SOCX) and analyzed using a hierarchical general linear model (HGLM). Only the in-kind pol-
icy had a statistically significant direct effect: lower levels of in-kind policy were associated
with higher levels of re/familialization (women staying at home). There was also a statistic-
ally significant interaction effect between in-cash policy and income level: women from
lower income groups were more likely to stay at home. Follow-up studies are needed on
the typology of family policies by public and private sector status.

Introduction familialization or de-familialization. Such policies

Family policies have been developed in many assume that parents with children have “freedom
welfare states to achieve various objectives, such of choice” e.g., the ability to choose whether to
receive allowances to take care of their children
as to: increase female employment, combat child
themselves at home, or use childcare facilities
poverty, and promote fertility rate. Rather than
that allow them to work. However, is this a real
fitting a single type, family policies can instead be
choice when not all members of society share
defined according to their specific objectives,
equal conditions or resources? Socio-economic
range, and type of benefits. One simple classifica-
status, such as the parents’ income level, may
tion scheme is to classify policies in accordance
affect the choice of re/familialization or de-famili-
with “in-cash” or “in-kind” benefits (OECD, alization. For instance, home childcare allowance
2011). This typology is based on the kind of bene- can also channel low-income households into re/
fit and the policy objectives; each policy has differ- familialization (Morel, 2007), as parental leave
ent effects. Generally, an in-cash policy has the benefits may encourage low-income families to
effect of re/familialization—where childcare is pro- leave the labor market as well (Hegewisch &
vided at home by the parents. In contrast, an in- Gornick, 2011). Therefore, the belief that family
kind (goods or services) policy has the effect of policies affect all income levels equally—in terms
de-familialization—where childcare is provided by of re/familialization or de-familialization
public childcare facilities for mitigating the burden choices—might be a fiction. Low-income people
of families’ childcare. In practice, family polices with fewer resources are more likely to face
have multi-faceted goals and different effects. restrictions in freedom of choice. This can lead
Many welfare states have implemented both to a paradoxical result where family policies gen-
in-cash and in-kind policies to support members erate social inequality, creating a freedom of
of society so that they can choose between re/ choice “trap”.

CONTACT Kyo-seong Kim [email protected] College of Social Science, School of Social Welfare, Chung-Ang University, 84, Heukseok-ro,
Dongjak-gu, Seoul 06974, Korea.
ß 2019 Taylor & Francis Group, LLC
2 K.-S. KIM AND E. J. HWANG

Few studies have considered the potential for to examine the specific categories. Family policies
differential re/familialization and de-familialization are categorized by type of benefit: in-cash or in-
effects in terms of stratification. Studies that have kind (OECD, 2011). In-cash policies are cash
mentioned differences in the effects of family poli- benefits related to childcare provided directly or
cies in terms of stratification have reported such indirectly to parents, including maternity allow-
differences as incidental findings, e.g., associated ances, child allowances, family allowances, home
with individual countries rather than universal childcare allowances, maternity leave benefits,
(Morel, 2007; Rainer et al., 2014; Szab
o-Morvai, parental leave benefits, single parental family
2014). Unlike these studies, this study was con- allowance, and other tax-benefits. In-kind or ser-
ducted to verify differences in the re/familializa- vice policies are all forms of services provided for
tion or de-familialization effects of family policies childcare, including childcare facilities, after-
focusing on stratification. In particular, we investi- school programs, and home help services. In par-
gated whether cash benefits that encourage women ticular, this classification has the advantage of
to stay home and care for their children are con- explaining specific effects of family policy, and is
centrated in low-income families. If the re/famili- borrowed from empirical studies that compare
alization effect of in-cash policies is biased, it the various effects of family policies across coun-
follows that freedom of choice is limited for low- tries (Luci & Thevenon, 2011). The present study
income households. The study also differs in that also adopted this family policy classification.
it accumulates comparative experimental results Family policy needs to be multi-faceted to
moving beyond assertive claims and case studies. achieve various socio-economic objectives. While
Ultimately, this study aims to contribute to the there are differences in degree, supporting child-
establishment of future universal family policies care and promoting female employment are com-
and detailed strategies. Research questions: mon goals for family policies in most countries.
Encouraging parents’ participation in the labor
1. How do in-cash and in-kind family policies
market and covering childcare needs became core
affect re/familialization or de-familialization?
tasks of family policies (Liljestrom, 1978); conse-
2. How do re/familialization or de-familialization
quently the main purpose of family policy devel-
effects of in-cash and in-kind policies differ in
oped into supporting both paid work and unpaid
terms of stratification?
domestic work. Therefore, this study focused on
the effects of de-familialization that supports
Concepts and Effects of Family Policies parents’ right to engage in paid work, and re/
Family policy is not a single concept, but rather a familialization that supports parents’ right to stay
range of concepts that involve a perspective for home to provide their own childcare.
considering policy in relation to families. The Re/familialization is defined as caring for a
term “family policies” is used to describe public dependent family member who needs help at home
policies that are designed to affect the situation of without any paid work being performed. That
families with children—or individuals in their means allowing parents to care for their children at
family roles—and which have clear, though pos- home by decommodification of their labor force
sibly unintended, consequences for such families (Daly, 2000). On the other hand, de-familialization
(Kammerman, 2010). Kammerman’s definition is defined as the degree to which individual adults
sets the policy target to “families with children,” can uphold a socially acceptable standard of living,
clarifies goals that are intended for “explicit” independent of family relationships, either through
results, and uses the term “public policies” to rep- paid work or social security provision. In other
resent the responsibility of the state. Therefore, words, parents can perform paid work, have leisure
her definition was adopted in this study since it time, and be reeducated to mitigate the burden of
corresponds to the characteristics of family policy. childcare (Lister, 1994). Extending the concept, de-
To more clearly understand the theoretical and familialization refers to the degree to which a wel-
abstract concepts of family policy, it is necessary fare state or market provisions ease the burden of
 JOURNAL OF SOCIAL SERVICE RESEARCH 3

caring responsibilities of families (Esping- age (Gustafsson & Stafford, 1992). An expansion
Andersen, 1999). of childcare services for children aged between 0
Within the same family policy category, policy and 4 years in Quebec had a positive effect on
goals are set differently according to type of benefit the labor supply of married women (Baker,
and have different effects in practice. The category Gruber, & Milligan, 2008). Increasing the provi-
of in-cash policy—including home childcare allow- sion of childcare facilities also increased the labor
ance, child allowance, and parental leave benefit— force participation rate in German women with
has the re/familialization effect that reduces the children aged between 0 and 2 years (Bick, 2011).
need for income activities in the labor market by the In a comparison of childcare services across some
level of benefits. Previous studies have suggested OECD members, countries with a higher avail-
that home childcare allowance reduces women’s ability of decent quality childcare service exhib-
labor supply. Kosonen (2014) found that the ited high parental—in particular, maternal—labor
Finnish home childcare allowance reduced women’s force participation rates as well (Vuri, 2016).
labor supply by 3 percentage points with each EUR
100 increase in home childcare allowance. This
Different Effects of Family Policies in Terms of
reduction in women’s labor market participation as Stratification
a result of cash benefits was also found in Norway
(Morosow, 2017; Naz, 2004; Rønsen, 2009; Schøne, There may be differences in re/familialization
2004). Moreover, it has been analyzed by several and de-familialization effects in terms of stratifi-
studies, and the finding that child allowance reduces cation. If a certain income group has only one
incentives for women to remain in the labor force is effect of either re/familialization or de-familializa-
consistent (Alz
ua, Cruces, & Ripani, 2010; Elborgh- tion, it means that the family policy may have a
Woytek et al., 2013; Ferragina, 2017; Jaumotte, mechanism that restricts freedom of choice even
2003; Kinoshita & Guo, 2015; Rainer et al., 2014; though the policy was universally applied across
Schirle, 2015). Gehringer and Klasen (2015) found all income groups. The group with limited choice
that child allowance had a negative effect on the is likely to be below-income families with fewer
labor supply of women and non-working-age men. socio-economic resources, which results in a
Women with universal child allowance tended to paradox whereby family policies themselves gen-
stay at home longer after childbirth, and thus their erate social inequality.
labor supply decreased in Spain (Gonz
alez, 2011). Differences in family policy effects in terms of
Parental leave benefits also have the characteristic of stratification have been empirically examined
re/familialization by de-commodifying the labor through cases in individual countries. In the case
force of parents who are commodified in the labor of cash benefits, studies have shown that home
market, so that they can take care of their children childcare allowance generates a strong stratifica-
within the family. Generous parental leave benefit tion effect. The expansion of a home childcare
helps parents to stay longer at home. Some studies allowance lowers the relative value of the wage
have reported that parental leave benefit also had a that women can earn through labor market par-
negative impact on women’s labor supply (Laroque ticipation, and thus increases the possibility that
& Salanie, 2005). low-income women will abandon their low-wage
On the other hand, the public childcare service work. Most parents have the opportunity to
is a representative policy that promotes de-famili- choose whether to use home childcare allowance;
alization by reducing parents’ need to care for however, this choice is affected by several factors
their children at home (Lohmann & Zagel, 2015). such as educational level, labor market attach-
The de-familialization effect of public childcare ment, and family structure (Morosow, 2017). As
service has been reported by many studies with a result, the use of home childcare allowance is
consistent overall results (Esping-Andersen, 1999; concentrated in less-educated low-income moth-
Leitner, 2003; Lister, 2000). In Sweden, high qual- ers with many children (Repo, 2010). Home
ity public childcare encouraged labor market par- childcare allowance that was provided under the
ticipation in women with children under school condition that the recipient does not use
4 K.-S. KIM AND E. J. HWANG

childcare facilities also lowered low-income wom- allowance until 2010, the child allowance system
en’s labor market participation rate in France and showed a decrease in the regular employment rate
Finland (Fagnani, 1998; Haataja & Valaste, 012). of women (Kinoshita & Guo). In addition, universal
Other studies found that in France the type of child allowance in combination with labor market
childcare differed remarkably according to house- situations might prevent low-income women from
hold or women’s income (Morel, 2007; Morgan, entering the labor market. Rainer et al. (2014)
2002). The stratification effects were: middle- reported that German child allowances have a ten-
income households used the public childcare ser- dency to reduce employment, especially in low-
vice; high-income households employed nannies income women. Studies also show that generous
individually or used private childcare services; child allowances can have a negative impact on the
low-income households took care of their chil- entry of low-income women into the labor market
dren directly through home childcare allowances (Elborgh-Woytek et al., 2013; Jaumotte, 2003).
(Morel, 2007). In sum, the effects of home child- On the other hand, the results of studies on the
care allowances on re/familialization differ in stratification effects of in-kind policies regarding
terms of stratification (Leira, 2002; Rissanen & childcare service are inconsistent. Some studies have
Knudsen, 2001; Rønsen & Sundstrom, 2002; reported that childcare services increased women’s
Sipila & Korpinen, 1998; Vikat, 2004). participation in the labor market, while other studies
Studies have also shown that the effects of paren- have shown that the inhibition effects of paid work
tal leave on re/familialization also differ in terms of were greater for lower-educated women (Blau &
stratification. First, in terms of length of leave, when Currie, 2003). In Germany, labor market participa-
less-educated women used parental leave the rate of tion of higher-educated mothers disproportionately
reinstatement decreased two years after childbirth in increased as a consequence of childcare services,
Germany (Hegewisch & Gornick, 2011). In contrast,
compared to less-educated and low-income mothers
higher-educated mothers’ labor market participation
who preferred to rely on home childcare allowances
was disproportionately increased as a consequence
(Kennett & Lendvai-Bainton, 2017; Morgan, 2013).
of parental leave expansion, compared to less-edu-
Moreover, a childcare service may have a different
cated mothers (Kennett & Lendvai-Bainton, 2017).
effect on de-familialization in terms of stratification
In addition, in terms of income replacement rates,
when provided by the market rather than the public
the labor market exclusion of low-income women
sector. Low-income women may prefer re/familiali-
can be explained by the “wealth effect” of a cash
zation if they have to pay additional charges to use
benefit: accumulated monetary reserves allowed for
childcare services. In this context, de-familialization
these mothers to choose to stay home instead of
undertaking a full-time job (Szab
o-Morvai, 2014). through national schemes can only guarantee a wom-
The rate of reinstatement after parental leave was an’s right of access to paid work. The Swedish gov-
only high for higher-educated women with higher ernment tried to prevent the stratification effect in
incomes (Kennett & Lendvai-Bainton, 2017). advance by maintaining the principle of maximum
At first glance, child allowance appears to be a public childcare and minimum marketization
neutral policy; however, depending on how the pro- (Naumann, 2011).
gram is designed, it might also have re/familializa-
tion effects that differ in terms of stratification. In Research Design
particular, the introduction of low-income-based
Units of Analysis and Data Sources
child allowances with means testing can have a
negative impact on the labor market participation of The unit of analysis for this study was divided into
low-income women, similar to home childcare national and individual levels. The national-level
allowances. As the income of a worker increases, the unit included family policies in 27 OECD countries
benefit for the child is reduced, which lowers the including Australia, Austria, Belgium, Canada,
incentive for a second wage earner to enter the labor Chile, Czech Republic, Denmark, Finland, France,
market. In fact, in an empirical analysis of Japan, Germany, Hungary, Iceland, Ireland, Israel, Japan,
which had implemented income-based child Korea, Latvia, Mexico, Netherlands, Norway,
 JOURNAL OF SOCIAL SERVICE RESEARCH 5

Portugal, Slovakia, Slovenia, Spain, Sweden, women’s re/familialization or de-familialization.

Switzerland, and United States. The unit at the Such factors included age, education level, income
individual-level was 20- to 50-year-old women level, attitude to gender role, and number of chil-
with children under school age. Age was limited to dren under school age. The age variable was the
50 years or less to include only the family policy- current age reported by respondents. Education
target population. Women with children under level was the respondents’ final level standardized
school age were included because they are more into three categories, based on the different aca-
acutely affected by family policies and childcare demic systems in each country. Respondents
choices than women of school-age children. In all, whose final level was “no school” or “elementary
2,173 cases were used in the analysis. school graduate” were classified as “below middle
Data sources included the Family and Changing school”; graduates of middle school, high school
Gender Roles module from the 2012 International and vocational school were classified as “high
Social Survey Program (ISSP) and the OECD school graduate”; and graduates of college, uni-
Family Database. The ISSP is a collection of indi- versity and graduate school were classified as
vidual-level survey data from a nationally represen- “university graduate or above.” Income level was
tative sample of standardized questionnaires measured by estimating the interval of household
administered in 40 countries.1 The survey covers income to which a respondent belonged based on
most welfare states that implement family policies, the median household income of all respondents.
and includes various items that reflect individual The OECD. (2017) defined low income as less
and household characteristics. Therefore, it is con- than 50% of the median income, and high income
sistent with the purpose of this study for examin- as more than 150% of the median income, with
ing the effects of national and individual level middle income defined as falling between those
factors on re/familialization and de-familialization. two groups (more than 50% and less than 150% of
In addition, the OECD Family Database is an median income). Household income was used as
appropriate source for measuring family policy lev- the variable because cases with no occupation
els because it provides national family policy would be excluded from the analysis using indi-
expenditures for both in-cash and in-kind policies. vidual income. Both income and educational level
were treated as continuous variables in the
HGLM analysis.
Dependent Variables
Attitude to gender role was measured by an
The dependent variable was re/familialization or de- average value of the answers to the following
familialization. Re/familialization was operationalized four items: (i) a preschool child is likely to suf-
as “women with children under school age who stay fer if his or her mother works; (ii) family life
at home.” De-familialization was operationalized as suffers when the woman has a full-time job; (iii)
“women with children under school age who per- it’s all right for women to get a job, but what
form paid work.” The ISSP survey questions con- most women really want is to stay home and
cerning “whether to perform paid work” provided raise their children; (iv) all a man has to do is
the re/familialization and de-familialization meas- make money and a woman’s job is to take care
ures. Re/familialization included cases where of her family. Lower values meant a stronger
respondents had previously performed paid work but traditional attitude to gender roles, while higher
were not currently working, or did not perform paid values meant a more egalitarian attitude. The
work at all; de-familialization included cases where number of children under school age was the
respondents were currently performing paid work. total number of children under school age in
the respondent’s household.
Individual-Level Independent Variables
National-Level Independent Variables
Individual-level variables representing individual
and household characteristics were selected by refer- National-level variables were family policies such
ring to previous studies on the factors affecting as in-cash policy and service policy. The family
6 K.-S. KIM AND E. J. HWANG

Table 1. Major variables and data sources.

Variables Measurements Data
Dependent Re/familialization Not performing paid work ISSP (2012)
variables (Individual-level) De-familialization Performing paid work now
Independent National-level In-cash policy Percentage of in-cash policy expenditures to GDP OECD (2009–2011)
variables In-kind policy Percentage of in-kind policy expenditures to GDP
Individual-level Age Actual age ISSP (2012)
Level of education Final level of education
Attitude to gender role Average value of scale of attitude to gender role
Level of income Level of household income
Number of children Number of children under school age
Interaction In-cash policy  Level of Interaction term of in-cash policy and level of income ISSP (2012),
income Interaction term of in-kind policy and level of income OECD (2009–2011)
In-kind policy  Level
of education

policy variable was operationalized to the level of country. As the data used in this study had a
family-related expenditure to GDP. “Expenditure” hierarchical structure, the hierarchical linear
can be seen as a useful indicator of a country’s model (HLM) was used to overcome several stat-
welfare efforts, despite some limitations (Castles, istical problems. If the hierarchical structure is
2004). Several quantitative studies on family pol- ignored, there is a risk of ecological fallacy and
icy have used family-related expenditures as a overestimation of statistical significance
measure of policy level (Esping-Andersen, 1999; (Kerlinger & Lee, 2000). However, the dependent
Gornick & Meyers, 2003). Specifically in-cash variable, which is the choice of re/familialization
policy was measured as the percentage of cash or de-familialization, had a binary feature which
benefit expenditures relative to GDP, including does not meet HLM data assumptions that the
child allowances, home childcare allowances, relationship between independent variables is lin-
maternity leave benefits, and parental leave ben- ear and normally distributed. Therefore, a hier-
efits. In-kind policy was also measured as the archical generalized linear model (HGLM), an
percentage of social service expenditures relative expanded form of HLM, was adopted for this
to GDP, including direct financing or subsidiza- study. HLM 7.03 software was used to run
tion of childcare, public childcare supports, and the models.
home help services. As dependent variable, re/ The unconditional model included only an
familialization or de-familialization was meas- intercept without any independent variables.
ured in 2012, the average of expenditure over Level
1 Model : re=familiarizationij
three years from 2009 to 2011 was used to pro- ¼ b0j þ cij ðcij  N ð0, r2 ÞÞ
vide a reasonable causal relationship.
Level
2 Model : b0j ¼ c00 þ l0j ðl0j  N ð0, sÞÞ
In addition, interaction terms were estab-
lished for individual-level variables and In the conditional model, individual variables
national-level variables to examine differences such as age, income level, education level, atti-
in family policy effects in terms of stratifica- tude to gender role, and spouse’s employment
tion, which is this study’s core research ques- status were added to the within-group model at
tion. These included the interaction between level-1. National variables including in-cash
in-cash policy and income level, and the inter- policy and service policy were also added to the
action between in-kind policy and income level between-groups comparison at level-2.
(See Table 1).
Level
1 Model : re=familiarizationij
¼ b0j þ b1j ðageÞ þ b2j ðeducationÞ
Analysis Methods
þ b3j ðincomeÞ þ b4j ðgenderroleÞ
A descriptive analysis was performed using SPSS
Statistics, version 20.0, to examine individual þ b5j ðchildrenÞ þ cij
characteristics and family policies level by
 JOURNAL OF SOCIAL SERVICE RESEARCH 7

Figure 1. The ratio of re/familialization and de-familialization across countries. Source: ISSP (2012). Family and Changing Gender
Roles module

Level
2 Model : b0j ¼ c00 þ c01 ðcashpolicyÞ followed by the Czech Republic (61%), Korea
þ c02 ðservicepolicyÞ þ l0j (61%), Slovakia (53%), and Chile (52%). In these
b1j ¼ c10 countries, more women with a child under school
b2j ¼ c20 age were staying at home than performing paid
b3j ¼ c30 work. In the case of de-familialization, Norway
b4j ¼ c40
had the highest rate with 95%, followed by
b5j ¼ c50
Portugal (83%), Iceland (83%), Netherlands
Finally, a model was set up to examine inter- (81%), Denmark (79%), Finland (78%), and
action terms between the income level and each Belgium (77%). In these countries, more women
family policy by adding the in-cash policy and the with a child under school age were engaged in
in-kind policy to the regression coefficient, b2. paid work than were those who stayed at home.
re=familiarizationij ¼ b0j þ b1j ðageÞ þ b2j ðeducationÞ
Frequency analyses were conducted to examine
the overall characteristics of women with a child
þ b3j ðincomeÞ þ b4j ðgenderroleÞ
under school age (N ¼ 2,173). As shown in Table
þ b5j ðchildrenÞ þ cij
2, 52.6% were high school graduates and 42.4%
Level
2 Model : b0j ¼ c00 þ c01 ðcashpolicyÞ
were university graduates or above. Most women
þ c02 ðservicepolicyÞ þ l0j
were in the middle-income level, 1,428 (65.7%);
b1j ¼ c10
b2j ¼ c20 12.1% and 22.2%, respectively, were in the low-
b3j ¼ c30 þ c31 ðcashpolicyÞ þ c32 ðservicepolicyÞ þ l0j income and high-income groups. The average age
b4j ¼ c40 was 33.98 years, and the average level of attitude
b5j ¼ c50 to gender role was 3.35 points. The average num-
ber of children under school age was 1.39.
Table 3 shows the level of family policies by
Results and Discussion type of benefit across countries. Overall, the aver-
age OECD public spending on family policy in
General Characteristics
cash was 1.21% and service was 1.01% of GDP.
Figure 1 shows the ratio of re/familialization and Ireland had the highest cash spending (2.87%)
de-familialization in women with children under followed by Hungary (2.20%), Austria (2.13%),
school age across various countries. Hungary had Australia (1.83%), Belgium (1.80%), and Czech
the highest rate of re/familialization (63%) Republic (1.80%). Denmark had the highest
8 K.-S. KIM AND E. J. HWANG

Table 2. Characteristics of individual-level variables.

Frequency (%) Mean (SD)
Level of education Below middle school 108 (5.0) Age 33.98 (6.0)
High school graduate 1,144 (52.6)
University graduate or above 921 (42.4) Attitude to gender role 3.35 (1.0)
Level of Income Low income 263 (12.1)
Middle income 1,428 (65.7) Number of children 1.39 (0.6)
High income 482 (22.2)
Source: ISSP (2012). Family and Changing Gender Roles module.

Table 3. The level of family policies across countries. Table 4. HGLM analysis of the effects of re/familialization and
Countries In-cash In-kind Countries In-cash In-kind de-familialization: The unconditional model.
Ireland 2.87 0.83 Germany 1.23 0.90 Coef. S.E. t-value Odds R.
Hungary 2.20 1.13 Latvia 1.23 0.27 Fixed Effects
Austria 2.13 0.57 Switzerland 1.13 0.30 Intercept
0.684 0.142
4.808 0.505
Australia 1.83 0.80 Canada 1.03 0.20 SD Variance df v2
Belgium 1.80 1.03 Israel 1.03 0.93 Component
Czech Republic 1.80 0.53 Portugal 0.90 0.47 Random Effects
France 1.60 1.30 Japan 0.73 0.50 Variance for level 1 1.000
Denmark 1.57 2.27 Chile 0.70 0.77 Variance for level 2 0.680 0.463 26 184.269
Finland 1.57 1.53 Netherlands 0.70 0.83
Slovenia 1.57 0.50 Spain 0.57 0.80 Note. p < .05; p < .01; p < .001.
Slovakia 1.57 0.40 Mexico 0.40 0.70
Sweden 1.43 2.00 United States 0.10 0.63
Iceland 1.37 2.23 Korea 0.00 0.60 higher than that of performing paid work. The
Norway 1.33 1.73 Average 1.21 1.01
variance for level 2 was also statistically signifi-
Source: OECD Family database (2009–2011).
cant (0.463, p < .001), thus the effect of re/fami-
service spending (2.27%) followed by Iceland lialization was explained by differences
(2.23%), Sweden (2.00%), Norway (1.73%), and across countries.
Finland (1.53%). High levels of service policy In the conditional model, Model 1 examined
were highly correlated with high rates of de-fami- the influence of policies on the re/familialization
lialization. This result suggests that the in-kind or of women with a child, controlled for individual
service policy and de-familialization effects may characteristics. The individual-level and national-
be related. level variables that were predicted to influence
the re/familialization of women with a child
were entered.
HGLM Analysis on the Effects of Family Policies Table 5 shows the result of the multilevel mod-
The HGLM process is divided into an uncondi- els. The individual-level factors, age, education
tional model and a conditional model. In the level, income level, attitude to gender role and
unconditional model, significant differences number of children under school age significantly
between level 1 and level 2 variances were exam- influenced the re/familialization effects in Model
ined and the level 2 variance was derived. In the 1. Income level had a negative effect on the re/
conditional model, the impact of each variable on familialization of women with a child. Previous
the dependent variable was examined by entering studies reported that lower household or individ-
explanatory variables (Raudenbush & Bryk, ual women’s income was associated with wom-
2002). We assumed that the level 1 variance was en’s lower participation in the labor market
“1” for convenience because the level 1 variance (Chang & Lee, 2012). Age and education level
is heteroscedastic in the case of a binary depend- also had negative effects. Our result supports pre-
ent variable. Only the change in level 2 variance vious reports that women with higher educational
was explained by entering predicted variables in attainment tend to maintain labor market partici-
each model. pation (Light & Ureta, 1990). Women with a
Table 4 shows the results from the uncondi- more egalitarian attitude toward gender roles
tional model. The re/familialization effect was were also more likely to be working full-time
statistically significant. The likelihood of women than staying at home. This is also consistent with
with a child staying at home was 0.51 times the previous findings that gender role attitudes
 JOURNAL OF SOCIAL SERVICE RESEARCH 9

Table 5. HGLM analysis of the effects of re/familialization and de-familialization: The conditional model.
Model 1 Model 2
Coef. S.E. Odds R. Coef. S.E. Odds R.
Fixed Effects
Intercept
0.751 0.129 0.472
0.765 0.129 0.465
Level 1:
Age
0.023 0.009 0.978
0.023 0.009 0.977
Level of education
0.259 0.095 0.772
0.271 0.106 0.762
Attitude to gender role
0.491 0.061 0.612
0.483 0.062 0.617
Level of income
0.790 0.095 0.454
0.880 0.100 0.415
Number of children 0.371 0.083 1.449 0.383 0.084 1.467
Level 2:
In
cash policy 0.028 0.201 1.029 0.002 0.203 1.002
In
kind policy
0.864 0.236 0.422
0.891 0.238 0.410
Interaction terms:
In-cash policy  Level of income
0.388 0.129 0.679
In-kind policy  Level of income
0.193 0.227 0.824
SD Variance component SD Variance component
Random Effects
Variance for level 1 1.000 1.000
Variance for level 2 0.595 0.354 0.596 0.356
Note. p < .05; p < .01; p < .001.

may influence decisions to participate in the women’s labor market participation (Leitner,
labor market after childbirth (Farre & Vella, 2003; Lister, 2000). Therefore, if increasing wom-
2013). Our analysis showed that the number of en’s labor market participation is the most
children under school age had a positive effect important goal of family policies, it is the in-kind
on re/familialization, which is also consistent policy that needs to be strengthened rather than
with previous studies that showed that the num- the in-cash policy. This would allow de-familiali-
ber of children had an impact on women’s labor zation effects that would increase women’s labor
market participation (Cha & Won, 2014). market participation.
Turning to national-level effects, only the in- In Model 1, the general re/familialization and
kind or service policy had a statistically signifi- de-familialization effects were examined across
cant effect; the in-cash policy had no effect on income levels. Service policies had a universal de-
re/familialization or de-familialization. Although familialization effect regardless of income bracket,
some studies have reported that conditional cash but the in-cash policy had no significant effect.
benefits (e.g., cash benefits provided under the The introduction of individual-level variables and
condition that women stay home and care for national-level variables reduced the between-
their children) may encourage women’s re/famili- groups variance by 23.44 percent from 0.463 to
alization (Leira, 2002; Vikat, 2004), our results 0.354. This suggests that re/familialization could
showed no effect of re/familialization associated be explained by the various individual and
with the in-cash policy. Child allowance had neu- national-level independent variables that enter
tral characteristics because it was provided uni- the conditional model.
versally, regardless of whether the parents While it is clear that the in-cash policy did not
worked. By contrast, in our results the effect of show a universal effect in Model 1, it is not clear
the in-kind policy on de-familialization was whether there may be differential effects on re/
highly significant (
0.864, p < .01). Lower levels familialization in terms of income level.
of service policy were associated with higher lev- Therefore, in the conditional model, interaction
els of re/familialization (women staying home). terms were added to investigate differences in the
Similarly, higher levels of service policy were effects of family policies in terms of stratification;
associated with an increased likelihood that the interaction terms (i) in-cash policy and
women with a child under school age would per- income level and (ii) service policy and income
form paid work. This second result agrees with level were entered in Model 2. The re/familializa-
earlier studies reporting that the availability of tion effect of the in-cash policy was found to be
quality public childcare services increased dependent on income level (
0.388, p < .01);
10 K.-S. KIM AND E. J. HWANG

Figure 2. The effect of interaction between in-cash policy and level of income on re/familialization. Note. Standardized variables
are those that are both centered around zero and are scaled to have a standard deviation of 1 (Dawson, 2010).

with the in-cash policy re/familialization effect middle-income, and high-income groups show
being stronger in lower income groups. Women the same de-familialization effect of in-kind poli-
in households with relatively low-income were cies, an effect that does not differ in terms of
more likely to be receiving cash benefits and car- stratification. It is possible that the re/familializa-
ing for their children at home. This is in line tion effects driven by strengthening in-cash policy
with previous reports that child allowance, home are concentrated in low-income women.
childcare allowance, and parental leave benefits Model 2 added cross-level interactions, reduc-
hinder low-income women’s participation in the ing the between-groups variance by 23.14 percent
labor market (Fagnani, 1998; Haataja & Valaste, from 0.463 to 0.356. This suggests that the effect
2012). Our results showed that freedom of of re/familialization may be explained by the
choice, which was intended by family policies, various individual-level and national-level inde-
was not possible in all households but rather pendent variables and interaction terms that were
determined by income level. entered into the conditional model.
On the other hand, the service policy and To better clarify the interaction between in-
income level interaction term was not significant, cash policy and income level, Figure 2 shows the
indicating that there was no difference in the results of calculating the effects of in-cash policy
effect of in-kind policy in terms of stratification. levels on re/familialization in terms of low and
This means that women with children under high income. Dawson (2010) provides a variety
school age perform paid work by using public of worksheets that can be used to easily interpret
childcare facilities regardless of household interactions based on the interaction effect ana-
income. Previous study results present two com- lysis described by Aiken and West (1991) and
peting arguments: (i) in-kind policy has the same Dawson and Richter (2006). We analyzed the
hierarchical effect as in-cash policy (Blau & interaction effect using a worksheet for GLM
Currie, 2003), (ii) when the universal service models. The re/familialization effect of in-cash
benefit is provided by the state, it has no hier- policies increased from 1.19 to 5.25 for the low-
archical effect (Lloyd & Penn, 2010). Our results income group, but decreased from 0.22 to 0.06
support the latter argument. Since it is possible for the high-income group. This suggests that in-
for all income levels to enter the labor market by cash policy levels differentially influence the like-
childcare facilities being used, the service policy lihood of women staying home: under in-cash
has a universal effect. When the results of policies, lower-income women are more likely
Models 1 and 2 are integrated, low-income, than higher-income women to stay home. The
 JOURNAL OF SOCIAL SERVICE RESEARCH 11

re/familialization effect of in-cash policies effect of universal de-familialization that results

increases sharply when the income level is low, from in-kind policies. A public childcare service
but increases relatively little when the income is an important foundation of family policies that
level is high. can foster equitable opportunities by supporting
parents’ de-familialization regardless of income
Conclusions level. Women with a child can perform paid
work in the labor market irrespective of socio-
The results of this study have several implica- economic status if public childcare service is pro-
tions. First, it is necessary to reestablish the focus vided by the government without additional cost.
of family policies by considering differences in It is important that a childcare service is univer-
the in-cash re/familialization effects in terms of sally provided in the public arena. If a childcare
stratification. As we have continued to empha- service is based on marketization, there may be
size, if re-familialization is truly a freedom of stratification effects because the choice of child-
choice, it should be chosen according to the indi- care service may be limited by income level and
vidual’s preference, regardless of income level. women’s participation in the labor market (Lloyd
However, there is a strong effect of in-cash policy & Penn, 2010). The target population for child-
that leaves women with preschoolers staying at care services needs to include parents and chil-
home in the lower income bracket. What has dren from all income levels to assure
made the choice of low-income women focus on universal effects.
re-familialization? Low-income women with a Finally, the most important consideration for
child may choose re/familialization even though redesigning family policies and achieving policy
they do not really want to stay at home, because goals is to find a balance between in-cash and in-
the cash benefits may offset most of what they kind service policies. This means that in-cash
might earn from paid work. The incentives for benefits and service benefits should be developed
re/familialization are low in middle- and high- simultaneously and be mutually complementary.
income women because their wages are higher If the overall childcare infrastructure is well
than the cash allowance. In addition, this cash established, the negative impact of hierarchical
allowance increases their disposable income, differences within in-cash policies can be mini-
allowing them the option to use private childcare mized. For example, if accessibility to high-qual-
while continuing to perform paid work. The ity public childcare service increases, it will
expansion of in-cash policies fosters a choice that mitigate the re/familialization incentives that
is driven by income level rather than a true affect low-income parents, allowing them to base
desire to stay home. Only low-income women their childcare decisions solely on their prefer-
with low opportunity in the labor market experi- ence. For maximum effectiveness, policies to
ence the in-cash re/familialization effect (Kremer, mitigate re/familialization inequality through
2002). Therefore, establishing a forum for social high-quality childcare (e.g., as found in Nordic
discussion regarding the potential benefits and countries) should be developed simultaneously
drawbacks of expanding an in-cash policy is rec- with program changes designed to strengthen in-
ommended. If complete freedom of choice is not cash policies.
possible in itself, then in-cash policy with stratifi- This study differs from earlier research in that
cation will only limit the choice of low-income it examines the overall effects of family policies,
people. Furthermore, considering that these and analyses the differences in these effects in
stratification phenomena are closely related to a terms of stratification. However, this study has
poor labor market environment, efforts to limitations in that it cannot investigate specific
improve labor conditions—such as the reduction policy characteristics because the cash benefit
of night work or expansion of flexible work data are limited to child allowance, home child-
hours—would also be beneficial. care allowance, and parental leave benefits.
Second, the continual expansion of accessible Similarly, the separate roles played by public and
high-quality childcare services will enhance the private sector childcare providers could not be
12 K.-S. KIM AND E. J. HWANG

strictly considered. Follow-up studies are needed Dawson, J. F. (2010). Interpreting interaction effects.
on the typology of family policies by public and Retrieved from http://www.jeremydawson.co.uk/slopes.
htm
private sector status.
Dawson, J. F., & Richter, A. W. (2006). Probing three-way
interactions in moderated multiple regression:
Development and application of a slope difference test.
Funding Information
Journal of Applied Psychology, 91(4), 917–926. doi:10.
This research was supported by the Chung-Ang 1037/0021-9010.91.4.917
University Graduate Research Scholarship Elborgh-Woytek, K., Newiak, M., Kochhar, K., Fabrizio, S.,
Kpodar, K., Wingender, P., … Schwartz, G. (2013).
in 2017.
Women, work and the economy: Macroeconomic gains
from gender equity. (IMF Staff Discussion Notes, SDN/
Note 13/10). International Monetary Fund. Retrieved from
https://www.imf.org//media/Websites/IMF/imported-full-
1. Sampling procedures differ for the individual countries text-pdf/external/pubs/ft/sdn/2013/_sdn1310.ashx doi:10.
as partly simple or partly multi-stage stratified random 5089/9781475566567.006
sampling (ISSP, 2012). Please refer to the ISSP data Esping-Andersen, G. (1999). Social foundations of postindus-
archive for sample sizes and survey times by country, trial economies. New York, NY: Oxford University Press.
https://dbk.gesis.org/dbksearch. Fagnani, J. (1998). Recent changes in family policy in
France: Political trade-offs and economic constraints. In
E. Drew, R. Emerek, & E. Mahonin (Eds.), Women, work
References and the family in Europe (pp. 58–65). London; New
York, NY: Routledge.
Aiken, L. S., & West, S. G. (1991). Multiple regression:
Farre, L., & Vella, F. (2013). The intergenerational transmis-
Testing and interpreting interactions. London: Sage.
sion of gender role attitudes and its implications for
Alz

ua, M. L., Cruces, G., & Ripani, L. (2010). Welfare pro-
female labour force participation. Economica, 80(318),
grams and labor supply in developing countries.
219–247. doi:10.1111/ecca.12008
Experimental evidence from Latin America. (CEDLAS, Ferragina, E. (2017). Does family policy influence women’s
Working Papers 0095, CEDLAS, Universidad Nacional de employment? Reviewing the evidence in the Field.
La Plata). Retrieved from http://www.cedlas.econo.unlp. Political Studies Review, November, 1–16. doi:10.1177/
edu.ar/wp/wp-content/uploads/doc_cedlas95.pdf 1478929917736438
Baker, M., Gruber, J., & Milligan, K. (2008). Universal Gehringer, A., & Klasen, S. (2015). Labor force participation
childcare, maternal labor supply and family well-being. of women in the EU—What role do family policies play?
Journal of Political Economy, 116(4), 709–745. cege Discussion Papers, No. 242. G€ ottingen, Germany:
Bick, A. (2011). The quantitative role of child care for female University of G€ottingen, Center for European,
labor force participation and fertility (MPRA Paper Governance and Economic Development Research (cege).
31713). Munich: University Library of Munich. Gonz
alez, L. (2011). The effects of a universal child benefit.
Blau, D., & Currie, J. (2003). Preschool, day care and after (Barcelona GSE Working Papers 574). Barcelona
school care: Who’s minding the kids? (NBER Working Graduate School of Economics. Retrieved from http://
Papers 10670). Cambridge, MA: National Bureau of blogs.barcelonagse.eu/file/2709/download?token=uFJ8LbdX
Economic Research, Inc. Gornick, J., & Meyers, M. (2003). Families that work:
Castles, F. G. (2004). The future of the welfare state: Crisis Policies for reconciling parenthood and employment. New
myths and crisis realities. New York: Oxford University York, NY: Russell Sage Foundation.
Press. Gustafsson, S., & Stafford, F. (1992). Child care subsidies
Cha, Y., & Won, S. (2014). Antecedents for labor force par- and labor supply in Sweden. The Journal of Human
ticipation of mothers: Focusing on the influence of eco- Resources, 27(1), 204–230. doi:10.2307/145917
nomic crisis. (in Korean). The Korea Association for Haataja, A., & Valaste, M. (2012, September). Cutting child
Policy Studies, 23(3), 303–327. home care allowance in Finland? Paper presented at the
Chang, J., & Lee, B. (2012). Increasing inequality and policy ESPAnet Anniversary Conference, Edinburgh.
options. (in Korean). Democratic Society and Policy Hegewisch, A., & Gornick, J. C. (2011). The impact of
Studies, 23, 71–109. work-family policies on women’s employment: A review
Daly, M. (2000). The gender division of welfare: The impact of research from OECD countries. Community, Work &
of the British and German welfare states. New York, NY: Family, 14(2), 119–138. doi:10.1080/13668803.2011.
Cambridge University Press. 571395
 JOURNAL OF SOCIAL SERVICE RESEARCH 13

ISSP. (2012). Family and changing gender roles IV. Policy Research, 17(1), 42–48. doi:10.1111/j.1744-540X.
Retrieved from https://www.gesis.org/issp/modules/issp- 2010.00600.x
modules-by-topic/family-and-changing-gender-roles/2012 Lohmann, H., & Zagel, H. (2015). Family policy in com-
Jaumotte, F. (2003). Labour force participation of women. parative perspective: The concepts and measurement of
OECD Economic Studies, 37, 51–107. doi:10.1787/eco_ familization and defamilization. Journal of European
studies-v2003-art9-en Social Policy, 26(1), 48–65. doi:10.1177/
Kammerman, S. (2010). Child, family and the state: The 0958928715621712
relationship between family policy and social protection Luci, A., & Thevenon, O. (2011). The impact of family pol-
policy. In S. Kamerman, S. Phipps & A. Ben-Arieh icy packages on fertility trends in developed countries.
(Eds.), From child welfare to child well-being (pp. Retrieved from https://hal.archives-ouvertes.fr/hal-
429–437). New York, NY: Springer. 00657603
Kennett, P., & Lendvai-Bainton, N. (2017). Handbook of Morel, N. (2007). From subsidiarity to ‘free choice’: Child-
European social policy. Cheltenham, UK: Edward Elgar and elder-care policy reform in France, Belgium,
Publishing. Germany and the Netherlands. Social Policy &
Kerlinger, F. N., & Lee, H. B. (2000). Foundations of behav- Administration, 41(6), 618–637. doi:10.1111/j.1467-9515.
ioral research (4th ed.). Orlando, FL: Harcourt College 2007.00575.x
Publishers. Morgan, L. (2002). Does anyone have a “libre choix”?
Kinoshita, Y., & Guo, F. (2015). What can boost female Subversive liberalism and the politics of French child
labor force participation in Asia (IMF Working Paper, care policy. In S. Michel & R. Mahon (Eds.), Child care
WP/15/56). doi:10.5089/9781498329750.001. Retrieved policy at the crossroads (pp. 143–167) . London:
from https://www.imf.org/~/media/Websites/IMF/ Routledge.
Morgan, K. J. (2013). Path shifting of the welfare state:
imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1556.
Electoral competition and the expansion of work-family
ashx
policies in western Europe. World Politics, 65(1), 73–115.
Kosonen, T. (2014). To work or not to work? The effect of
doi:10.1017/S0043887112000251
childcare subsidies on the labour supply of parents. The
Morosow, K. (2017). Is the home-care-allowance disadvan-
B.E. Journal of Economic Analysis & Policy, 14(3),
taging single parents? Labor market outcome consequences
817–848. doi:10.1515/bejeap-2013-0073
of the Cash-for-Care benefit for single parents in Finland.
Kremer, M. (2002). The illusion of free choice: Ideals of
Poster session 1 presented at Giornate di Studio sulla
care and child care policy in the Flemish and Dutch wel-
Popolazione. University of Florence.
fare states. In S. Michel & R. Mahon (Eds.), Child care at
Naumann, I. (2011). Towards the marketization of early
the crossroads (pp. 113–142). New York, NY: Routledge.
childhood education and care? Recent developments in
Laroque, G., & Salanie, B. (2005). Does fertility respond to
Sweden and the United Kingdom. Nordic Journal of
financial incentives? (CEPR Discussion Paper 5007).
Social Research, 2, 1–17. doi:10.15845/njsr.v2i0.109
Leira, A. (2002). Working parents and the welfare state. New
Naz, G. (2004). The impact of cash-benefit reform on
York, NY: Cambridge University Press. parents? Labour force participation. Journal of
Leitner, S. (2003). Varieties of familialism: The caring Population Economics, 17(2), 369–383. doi:10.1007/
function of the family in comparative perspective. s00148-003-0157-y
European Societies, 5(4), 353–375. doi:10.1080/ OECD. (2009–2011). OECD Social Expenditure Database.
1461669032000127642 Retrieved from http://stats.oecd.org/Index.
Light, A., & Ureta, M. (1990). Gender differences in wages aspx?DataSetCode=SOCX_DET
and job turnover among continuously employed workers. OECD. (2011). Doing better for families. Paris: OECD.
The American Economic Review, 80(2), 293–297. OECD. (2017). Compare your income—Methodology and
Liljestrom, R. (1978). Sweden. In S. Kamerman & A. Khan conceptual issues. Paris: OECD.
(Eds.), Family policy: Government and families in four- Rainer, H., Bauernschuster, S., Danzer, N., Fichtl, A.,
teen countries (pp.19–48). New York, NY: Columbia Hener, T., Holzner, C., & Reinkowski, J. (2014). Child
University Press. benefit and child allowances in Germany: Their impact on
Lister, R. (1994). She has other duties: Women, citizenship family policy goals (CESifo D ICE Report 1/2014, 37-45).
and social security. In S. Baldwin & J. Falkingham (Eds.), Munich, Germany.
Social security and social change: New challenges (pp. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear
31–44). Hemel Hempstead: Harvester Wheatsheaf. models: Applications and data analysis methods (2nd ed.).
Lister, R. (2000). Dilemmas in engendering citizenship. In Thousand Oaks, CA: Sage Publications.
B. Hobson (Ed.), Gender and citizenship in transition (pp. Repo, K. (2010). Finnish child home care allowance—Users’
33–83). New York, NY: Routledge. perspective and perceptions. In J. Sipil€a, K. Repo, & T.
Lloyd, E., & Penn, H. (2010). Why do childcare markets Rissanen (Eds.), Cash for childcare: The consequences for
fail? Comparing England and the Netherlands. Public caring mothers (pp. 46–64). Cheltenham: Edward.
14 K.-S. KIM AND E. J. HWANG

Rissanen, T., & Knudsen, C. (2001). The child home care Sipila, J., & Korpinen, J. (1998). Cash versus child care serv-
allowance and women’s labour force participation in ices in Finland. Social Policy and Administration, 32(3),
Finland, 1985-1998: A comparison with Norway. Oslo: 263–277. doi:10.1111/1467-9515.00102
Norweign Social Research. Szab
o-Morvai, A.
(2014). Who benefits from child benefits?
Rønsen, M. (2009). Long-term effects of cash for childcare The labor supply effects of maternal cash benefit.
on mothers’ labour supply. LABOUR, 23(3), 507–533. Retrieved from http://hetfa.eu/wp-content/uploads/2013/
doi:10.1111/j.1467-9914.2009.00456.x 12/Szabo-Morvai_June2013_2.pdf
Rønsen, M., & Sundstrom, M. (2002). Family policy and Vikat, A. (2004). Women’s labor force attachment and child-
after-birth employment among new mothers–A comparison bearing in Finland. (MPIDR Working Papers WP-2004-
of Finland, Norway and Sweden. European Journal of 001). Rostock, Germany: Max Planck Institute for
Population/ Revue Europenne de Dmographie, 18(2), 121–152. Demographic Research.
Schirle, T. (2015). The effect of universal child benefits on labor Vuri, D. (2016). Do childcare policies increase maternal
supply. Canadian Journal of Economics/Revue Canadienne employment? Subsidized childcare fosters maternal
D’
economique, 48(2), 437–463. doi:10.1111/caje.12132 employment, but employment status, childcare quality,
Schøne, P. (2004). Labour supply effects of a cash-for-care and availability matter. IZA World of Labor, 2016, 241.
subsidy. Journal of Population Economics, 17(4), 703–727. doi:10.15185/izawol.241