A Comparison between Conservative Welfare States and Korean Childcare Policy, 1993-2003

Claus Offe (1972, cited in Knijn & Ostner, 2002) was among the first to conceptualize commodification and decommodification based on Polany's concept. Offe pointed out that the unequal power relationship between the labor force and capital triggered the need for state-led decommodification. Based on this concept, Esping-Andersen (1990, p. 37) defined decommodification as “the degree to which individuals or families can uphold a socially acceptable standard of living independently of market participation.” Despite various critiques, the concept of decommodification seems to be a useful tool to explain the pre-1980s welfare state characterized by the prevalence of male breadwinner households and the stability of male status in the labor market. However, in current conditions where the male breadwinner's status is being threatened, skepticism arises in its ability to explain and classify welfare states that do not take into account the relationship with unpaid work.

The stability of the male breadwinner in the labor market was a precondition for the welfare state to concentrate its efforts on the decommodification of male breadwinners. Industry based on semi-skilled manufacturing guaranteed the stable employment and wage for male breadwinners, and the decommodification of women was secured by male breadwinners (Knijin & Ostner, 2002, p. 148). However, the weakening of male status as main breadwinners resulting from labor market flexibility and the increase in the need for female labor market participation called for an unprecedented role of the welfare state. It is at this very point that the concept of defamilization emerged, which linked unpaid work and the role of welfare states.

Esping-Andersen (1999) acknowledged the importance of unpaid work in his classification and suggested defamilization as one of the criteria for welfare state classification. According to Esping-Andersen (1999), defamilization refers to the degree to which households' welfare and caring responsibilities are relaxed-either via welfare state provision, or via market provision (p. 51). Esping-Andersen's definition of defamilization discusses defamilization as a precondition for a person's (specifically women's) decommodification. That is, because women hold the traditional care responsibilities within the family, these responsibilities need to be relieved in order to include women in the concept of decommodification. Only then could women be commodified in the labor market, and thus be subjected to decommodification. Esping-Andersen's discussion of defamilization is expected to play a role in including women in the discussion of welfare states based on decommodification. It is a meaningful advancement that Esping-Andersen included unpaid work in the welfare state analysis.

His explanation of defamilization, however, includes a marked limitation. The major criticism lies in the argument that he does not recognize defamilization and decommodification to hold an equal status. Esping-Andersen does not consider the independent social value of unpaid work (care work) in his discussion of defamilization. In other words, he is less concerned about the provision of appropriate level of care after defamilization to those family members in need.² This is contradictory to his discussion of decommodification. Whereas decommodification is defined as assuring an appropriate level of income despite no paid work, the definition of defamilization does not entail a provision of care to members in need when care is not available by the traditional care provider. Significance is only ascribed to lessening the care responsibilities of the traditional care providers.

If so, how should defamilization be redefined as an independent concept? If we agree that like paid work, unpaid work itself holds an important social value, defamilization and decommodification need to be understood in a parallel framework rather than from a standpoint of defamilization supplementing decommodification. From the parallel framework, defamilization in this paper can be redefined as the degree to which the society can provide family members with the appropriate degree of care even when in-family care is unavailable. We will attempt to redefine defamilization through Esping-Andersen's (1990) true decommodification. If true decommodification refers to receiving an appropriate level of income at times of no labor market participation due to caretaking, education, or leisure activities, true defamilization may consist of schemes that permit family members to receive an appropriate level of care while parents are pursuing activities other than caretaking, such as working, receiving education, involvement in social activities or leisure³. In addition, just as decommodification includes the possibility of familization of one's labor, a true concept of defamilization implies the possibility of commodification of one's labor. When applied to this paper, defamilization can be understood in line with commodification, and decommodification in line with familization.

Why compare childcare policy between Korea and the conservative welfare states focusing on defamilization and familization? The answer to this question can be found in the process of welfare state restructuring. An impending task of the traditional welfare states was the settlement between capital and labor (Hobson, Lewis, and Siim, 2002). On the other hand, the core task of welfare states under restructuring is the new negotiation between the traditional welfare state based on the male breadwinner and women. In fact, unemployment and pension payments are decreasing, but childcare policy-related expenditures are continuously on the rise (Daly & Lewis, 2000, cited in Hobson, et al., 2002). It was on this ground that Esping-Andersen (1999) pointed to the family as the core of welfare state restructuring.

If female labor market participation is inevitable and people's welfare relies on women's labor market participation, it is only natural that welfare state responsibilities should expand to include defamilization and familization as well as decommodification and commodification. In fact, the process of commodification and decommodification of women employs a conjunction of defamilization and familization, different from that of men who are regarded as exempt from care responsibility.

This is more so because social risks such as child poverty are emerging in association with caregiving (Taylor-Gooby, 2004). LIS data shows that the child poverty related to caregiving is more evident among unskilled female workers in conservative welfare states in which balancing work and family life is hardly feasible (Cantillon, et al., 2001, 447, cited in Taylor-Gooby, 2006). Considering that the tasks related to caregiving comprise one of the major social risks in current welfare states, it would be essential to analyze childcare policies between Korea and the conservative welfare states based on defamilization and familization, both of which are at the core of childcare policy. This analysis would render a discussion of similarities and dissimilarities in childcare policy in these countries.

Although familization and defamilization are the main foci of this paper, an integrative discussion of the following concepts is critical: familization, defamilization, commodification, and decommodification. As previously mentioned, familization and defamilization policies have direct and interdependent relationship with those of commodification and decommodification. For example, childcare policy, a form of defamilization policy, serves as a major precondition for female commodification and decommodification (Rauch, 2007), and parental leave, a representative familization policy, assumes decommodification of the commodified parents' labor.

Figure 1

An integrated understanding of defamilization, familization, decommodification, and commodification

Although limitations exist in schematic diagramming, Figure 1 describes a basic integrated understanding of decommodification, commodification, defamilization, and familization. The starting point is defamilization. Defamilization of care work is rendered through childcare policy. According to Esping-Andersen, defamilization is a precondition for decommodification. Defamilization, however, does not guarantee commodification. Policies based on defamilization are geared toward decreasing the amount of caretaking in the family, but do not necessarily intervene in regard to what care providers will do after defamilization. Defamilization can take caretakers to a noncommodification state by socializing a portion of caretaking responsibilities in the family.

Even if persons become commodified by having the caretaking pressure relieved a through defamilization policy such as childcare, this does not ensure decommodification. Commodified persons can either be decommodified or can be excluded from decommodification. In particular, when we consider the fact that welfare states' response toward social risk in the labor market is institutionalized centering around regular workers, it is quite possible that a person commodified through a flexible labor market would be excluded from decommodification. In fact, 76.7 percent and 76.1 percent of Korean irregular (temporary, part-time, etc.) workers in 2007⁴ were left out of employment insurance and national pension, respectively (Korea Labour and Society, 2007). Furthermore, the rate of irregular workers among women is higher than that of men, therefore, those excluded from decommodification may well mostly be women.

Even if persons are decommodified, this too does not ensure familization. The relationship between decommodification and familization is different according to gender. Parental and maternity leave involve decommodifying labor and familizing it for a certain period of time and these are closely related to the familization of women. In reality, this is not easily applicable to men because in welfare states, most men do not choose decommodification for familization. These gender differences are evident in Belgium. Among those who take the “career break” in Belgium, men utilize the break for education, job training, or early retirement, whereas women use it mainly for child rearing (Devisscher, 2004). In other words, women become familized through decommodification whereas men, even if they become decommodified, are likely to remain non-familized. Even with those men who choose decommodification for familization, it is unclear whether they take the care responsibility within the family. According to an informal source of information, many men in Korea who took parental leave were involved in activities other than childcare.

Finally, workers familized during a certain period of time through the decommodification policy become defamilized through the childcare policy, and are given opportunities to re-enter commodification. Defamilization, commodification, decommodification, and familization are closely interconnected; however, one does not guarantee another. It is based on this notion that we predict the process of familization and defamilization in Korea would be dissimilar to those in conservative welfare states.

France, Belgium, Austria, and Germany are generally categorized as typical conservative welfare states⁵. Conservative welfare states share attributes such as a social insurance system based on labor market status, horizontal equity, the principle of subsidiarity, and strong familism (Daly, 2001; Morel, 2007). However, there is no single agreed definition of the conservative welfare model. Scruggs and Allen (2006), who reconstructed Esping-Andersen's decommodification index, maintained that the three welfare regimes do not exist. Pointing out the inconsistencies found in seven out of eighteen countries, Daly (2000) also questioned the legitimacy of the welfare regime classification by Esping-Andersen. Particularly in the realm of childcare policy, the focal point of this study, the claim of universal characteristics among conservative welfare states is controversial. Although Esping-Andersen (1990) identified France, Germany, and Austria's conservative welfare systems as the prototype of familism, the argument of similarity among conservative welfare states does not seem so lucid.

Haas (2003) categorized most European countries (with the exception of the Netherlands) as the family-centered care model. Esping-Andersen (1990) identified the low level of social services as a property of family-centered conservative welfare states. These arguments imply that childcare policies in conservative welfare states emphasize within-family care responsibilities. Compared to other European countries, therefore, we may argue that childcare policies in the conservative states are directed at familization rather than defamilization.

This argument, however, is weak to explain the similarity among conservative welfare states, because the rate of childcare in Belgium and France is considerably higher compared to other countries (Morel, 2007). In 2003, the rate of childcare for children between zero and two years of age in Belgium and France was slightly lower than in Sweden and Denmark, but higher than that of Finland and Norway (Blome & Muller, 2007). In addition, Belgium and France have a higher childcare rate for children between three and six compared to four Northern European countries, although the difference is not prominent. In terms of defamilization of childcare, France and Belgium share more similarities with northern European countries than with other conservative welfare states.

The utmost difference in the degree of defamilization in child care lies in the countries' perspective on women and care work. In the Netherlands, childcare responsibility lies solely with mothers and no official state support exists (Anttonen, 2006). In Germany and Austria, most of the responsibility for child rearing is laid on the family (mostly mothers), although the state supports the informal care given by the family. Childcare responsibility in Belgium and France, on the other hand, is imposed on both the state and the family (Morel, 2007; Taylor-Gooby, 2004). The difference gap widens once the social entitlement of women is considered. The social entitlement of women is based on their status as family caregivers in Germany, Austria, and the Netherlands (until recently), whereas in Belgium and France it is based on both family caregivers and workers (Morgan & Zippel, 2003; Sainsbury, 1999; Taylor-Gooby, 2004). The reason for such gap can be traced to unique historical experiences of each country. For instance, a relatively delayed industrial revolution in France leading to the significance of the family farm played a role in sustaining women as workers (Morgan, 2002). Before the Second World War women were considered paid workers in family enterprises and agriculture (Lewis, 1993). Women's status as agricultural workers created a favorable foundation for defamilization policies to expand in France relative to other countries. The concurrent expansion of both defamilization and familization policies in France and Belgium was grounded in the recognition that childcare responsibilities should be shared between the state and family.

Religion is also one of the important factors in explaining the differences as the attributes of family policy are closely interrelated with the degree of religion's political power. States with strong Catholic political power that supports traditional families are more likely to base their women's social entitlement on motherhood. In much of Europe, Christian democratic parties predicated welfare states on the principle of subsidiarity and aimed to preserve the patriarchal family (Morgan, 2002, p. 274). For example, although the religious party in Germany and Austria had contributed to the advancement of social rights, it did not expand to include the social responsibility of child rearing. On the other hand, in France, the secular political power took the leadership after an intense competition between the religious and the secular political power, and the role of religion in welfare expansion was very limited (Morgan, 2002). Welfare expansion in France was relatively untied to the principle of subsidiarity and therefore resulted in a more active role of the state regarding childcare.

In Austria and Germany, where the strong political power of Catholics promoted traditional families, familization policy that supports in-family care became the basic principle of childcare policies. On the other hand, Belgium and France were relatively free from religious influence and therefore development of a defamilization policy that reinforced direct responsibilities of the state was presumably possible. Based on this discussion, we may conclude that family policies in Austria and Germany (specifically, childcare policies) were based on familization whereas France and Belgium aimed at both familization and defamilization. This may explain the deferred commodification and decommodification of women in Austria and Germany. In contrast, France and Belgium maintained a relatively higher level of defamilization policy that increased the possibility of women's commodification and decommodification. It is therefore difficult to conform to Esping-Andersen and others' suggestion that childcare policies in conservative states share similarities based on the family-centered model.

Referring to the above argument, childcare policies in Korea seem to be grounded in familization. Traditional Confucianism that promote gender division of labor still remaining a dominant social ideology, imposing care responsibilities on women, and a low public childcare rate are all signs of familization-based policies. In reality, however, neither familization nor defamilization is receiving adequate amount of support from the state. Similar to the Netherlands before the 1990s, Korea emphasizes family responsibilities for childcare but does not appropriately subsidize either defamilization or familization in regard to childcare.

According to Table 1, France has the highest rate (56.4 percent) of childcare for children between ages of zero and two, followed by Belgium. Germany and Austria have the lowest rate, and the Netherlands and Korea are placed in between. The childcare rate for children between three and six years indicates that most conservative countries utilize the care. Since the differences in childcare rate are minimal for children three and up, the following discussion will center around the younger group.

Table 1

Defamilization: 1993, 1998, 2003

The childcare rates in France and Belgium, which indicate the level of defamilization, are comparable to the Nordic welfare states. A close examination, however, reveals distinctive features from that of Nordic welfare states. Unlike Nordic welfare states, the social entitlement of women in France and Belgium is based on both their status as a family caregiver and a worker (Morel, 2007). At the same time, they are promoting defamilization through expanding childcare programs. This policy may seem contradictory, but coexistence of disparate principles is not unseen in French history, and family policy is not an exception. In fact, during the formative years of welfare states after the Second World War, France took in both the Bismarckian corporatism and the Beveridge principle of universal rights (Palier & Bonoli, 1995, as cited in Kim, Ahn, Chung, & Hong, 2006, p. 275-276). Such experience is well applied to childcare policies, and is justified on the ground that it gives parents the right to choose what works best for their family regarding childcare (Sabatineli, 2006). After the 1980s, however, the trend in childcare policy shifted to include individual care systems such as direct care by children's own parents rather than reinforcing public childcare facilities. The government expenditure on daycare services between 1994 and 2001 decreased from 16 percent to 8 percent, whereas the assistance to individual forms of care⁷ increased from 78 to 84 percent (Leprince, 2003, as cited in Morel, 2007). This shift led to classification of childcare preferences among different groups: the low income groups preferred childcare allowance, the middle class daycare programs, and the upper class individual care of their children (Morel, 2007). This trend is also manifested in Belgium, which implies that expansion in childcare facilities and assistance for various childcare systems are not necessarily mutually exclusive.

The low rate of childcare in Germany and Austria shows properties of a typical conservative family-centered family policy (Daly, 2001). The low level of defamilization in Germany, in particular, may be associated with historical strategies of the German women's movement. During the period of welfare state restructuring, German feminists sought for caregiver-parity, and chose childcare allowance and longer maternity leave over defamilization of childcare (Sainsbury, 1999). In fact, from the beginning (1970s) the German women's movement did not consider childcare as a main issue, and as a result the issue faded away in the later movement (Naumann, 2006). Similarly, societal interest in public childcare programs was low in Austria (Morgan & Zippel, 2003), and women were considered as the main caretakers (Stell & Duncan, 2001).

Compared to the four countries discussed above, the changes in the Netherlands are dramatic. In 1993, the childcare rate for the zero to two age bracket was the lowest among conservative welfare states. By 2003, however, the rate had increased by 33.0 percent. This change is unforeseen in a country dominated by the notion that Childcare facilities contradict the meaning of a happy family life and deny the specifically female talents of raising children (Bussemarker & Kerbergen, 1994, p. 23). A unique trait held by the Netherlands amidst this expansion of defamilization is the decreased role of the state and increased responsibility of parents and employers (Knijn & Ostner, 2002).

Korea shares similar properties of conservative welfare states, characterized by underdeveloped social services and women's low labor market participation. What is noticeable, however, is that Korea seems to be undergoing many changes similar to the Netherlands. The childcare budget in Korea has recently increased six times during six years (2002-2008) (MHWFA, 2008). In 2003, the childcare rate for children between zero and two was higher than that of Germany and Austria.

In 2003, France showed the highest rate of expenditure on childcare as a percentage of GDP, followed by the Netherlands and Belgium. Korea had the lowest rate, and Austria and Germany were placed in between. Intriguingly, the expenditure rate of Belgium, a country with a reputation for strong public support of childcare, is lower than that of the Netherlands. The expenditure in Korea amounts to only 20 percent of that in Germany, which showed the lowest rate among conservative states.

According to Table 2, France has the largest number of weeks of maternity and parental leave (172 weeks), followed by Austria, Germany, the Netherlands, and Belgium. Korea marked 59 weeks, which is longer than that in the Netherlands and Belgium. The maternity and parental leave expenditure as a percentage of GDP in Belgium, Germany, France, and Austria is about 0.2 percent to 0.3 percent. The expenditure as GDP percentage was reported as zero for the Netherlands since it has institutionalized the unpaid leave. On the other hand, Korea showed 0.006 percent despite the flat rate payment in the year 2003.

Table 2

Familization: 1993, 1998, 2003

The following is a discussion of characteristics regarding the program in each country. Contrary to other countries, parental leave in the Netherlands was designed to redistribute paid and unpaid work as well as to enable women's return to the labor market (Bruning & Plantenga, 1999). It was designed to share responsibilities of care work between mother and father, and to ensure both familization and commodification. This can be seen through the fact that the use of parental leave was only recently possible in part-time work⁸. The institutional property of Dutch parental leave confirms that although parents' caretaking is widely supported, the main responsibility lies with the parents. This suggests that familization of parents' labor is of individual choice, and the state is not responsible for their choices. This is contradictory to the state's active support of childcare programs, which leads to defamilization of parents' labor (Knijn & Ostner, 2002). That is, aside from social rhetoric on childcare, parents' labor market participation in the Netherlands is given priority over childcare within the family due to socioeconomic needs.

Paternal leave in France is the very example that shows an ambiguous assumption on the gender issue (Faganai, 1999). The parental leave, CPE, was institutionalized in 1977 as part of an employment policy; it is an unpaid leave and pays a child rearing allowance (APE) until the child reaches 36 months old for those who qualify (not participating in the labor market). However, APE that functioned as reducing the labor force went through reform in 2001 to allow working for the first two months in order to assist women's employment (Palier & Madnid, 2004). In 2006, the program provided a short leave period as well as high payment as an effort to motivate women's rapid return to the labor market (Leitner, 2006). From its introduction in 1985 based on women's familization, APE shifted to implementing means-tested payment and promoting women's return to labor market in the 2000s, which implies similar characteristics of a liberal social security system that reinforces work incentive (Palier & Madnid, 2004).

Belgium shares similar characteristics with France, which developed contradicting policies (Marques-Pereira & Paye, 2003). The parental leave in Belgium, however, is distinct not only from France but the rest of the conservative welfare states. Career break, another form of parental leave, can be used for re-training and leisure in addition to childcare⁹ (Devisscher, 2004). A formal purpose of the leave policy was job rotation (Marques-Pereira & Paye, 2003). Until 2001, workers substituting for those taking parental leave were limited to the unemployed (Leitner, 2006). The goal of career break is a way of responding to economic crisis rather than to familize childcare (Morel, 2007).

While the major purpose of leave policy in the Netherlands, Belgium, and France is closely related to labor market policy, policies in Germany and Austria emphasize reinforcing women's familization of childcare. However, the two countries have distinctive characteristics. Erziehungsurlaub, the parental leave in Germany, was introduced in 1986 by the conservative Christian Democratic Party and the Liberal Party as a means to stem back childcare program expansions (Scheiwe, 2000). Erziehungsgeld (parental leave benefits) was not designed to sustain the worker's wage but to recognize caretaking within the household (Pettinger, 1999). But through the reform in parental leave, this shifted toward familization based on commodification of parents. In 2000, the right to work part-time during parental leave was institutionalized to assure mother's employment (Leitner, 2006). The reform in 2007 shifted the purpose of benefits from means-tested compensation of care work to income-related benefits designed to sustain one's wage during the leave. With these reforms the properties of parental leave changed from supporting the traditional family role to supporting both familization and decommodification. On the contrary, the parental leave system in Austria reinforces the family care system. The reform in 2002 expanded its eligibility from those with prior employment status to all parents, weakening the association with parents' labor market status (Leitner, 2006).

Parental leave in Korea was institutionalized in 1987 with the enactment of the Equal Gender Treatment Act. However, the real impetus behind the policy was to supplement the decreased labor force of cheap unmarried female workers with married female workers (Kim, 1991). Eligibility for parental leave in Korea is limited to those covered under employment insurance, which does not presuppose withdrawal from the labor market. In addition, it does not share a similar goal with Belgium of job rotation, and does not indicate strong motivation to reinforce familization as shown in Austria. The program itself exhibits a typical form of decommodification and familization based on commodification of parents. However, parental and maternity leave expenditure as percentage of GDP in Korea is 0.006 percent; it covers a very limited group of people and the level of payment is also very low (18 percent of average wage). Among those who gave birth in 2003 (based on number of births) only 6.5 percent and 1.4 percent used maternity leave and parental leave, respectively.

Family allowance as cash benefits could increase familization of parents' labor. Austria's family allowance expenditure as percentage of GDP is the highest (2.2 percent) among countries included in the analysis. Korea currently has no institutionalized family allowance. It is noticeable that the expenditure levels have decreased in all comparing countries except for Austria. In the cases of the Netherlands and Belgium, the expenditure decreased by 36.4 percent and 25.0 percent, respectively, in the past ten years. These figures reflect that policies in conservative welfare countries are shifting toward supporting defamilization. On the contrary, Austria exhibits reinforced traditional family roles even after 2000 (Leitner, 2006) which may be related to an increase in family allowance.

The cluster coefficient of 3.0 was utilized for the cluster analysis. According to Figure 2, ten cases from France in 1993, Belgium, Germany and the Netherlands create one cluster (cluster coefficient=2.952). Three cases from Austria (cluster coefficient=2.068), two cases from France (1998, 2003, cluster coefficient=0.771) and Korea (cluster coefficient=1.062) formed other separate clusters.

Detailed examination of the result indicates that France became an independent cluster after 1998, whereas Belgium, the Netherlands, Germany, and France still comprised a cluster in 1993. The main reason for France's location in a separate cluster from Belgium in 1998 and 2003 may be traced to the rapid increase in childcare rate (zero to two age bracket) and childcare expenditure after 1998. During the analysis period, Austria formed a cluster of its own different from other conservative states. Of note is that Germany and Austria, which are said to share common characteristics in general, fell in different clusters. This distinction may result from the Austrian government's stronger support of familization policy.

In the past decade, the Netherlands showed a marked improvement in the defamilization index, taking the lead from Germany and Austria in all three defamilization indices in the early 2000s. In 2003, it led all conservative countries, with the exception of France, in childcare expenditure. These changes, however, are not enough to place the country's childcare policy in a cluster independent from neighboring conservative welfare states. Korea shares similar characteristics with conservative welfare states in its socioeconomic condition and values; however, with the low level of policy index in familization and defamilization, it was placed in a different cluster. This shows that shared social values and circumstances do not necessarily result in shared characteristics in policy.

Figure 2

Result from cluster analysis: Conservative welfare states and Korea

Figure 3 shows results from multidimensional scaling and cluster analysis. The Kruskal's stress from multidimensional scaling was 0.13372, indicating an acceptable level of fit, and it explained 91.37 percent of the variance (R²=91.370). The axes in multidimensional scaling can be set variably, and this study set two axes: defamilization and familization. The top left axis refers to the support level for defamilization policy, and the lower left axis indicates the familization policy. Moving toward the top left indicates a higher support for defamilization, and to the bottom left direction indicates support for familization.

An examination of family policy changes based on multidimensional scaling and cluster analysis results suggests that all countries slightly shifted toward defamilization after the year 1998. Support for most familization programs is stagnating or decreasing while there is an increase in childcare rate and the expenditure on childcare, both of which are defamilization indicators. In Austria, familization support decreased between 1993 and 1998, but began to increase between 1998 and 2003.

In the larger picture, except for France, there are no evident qualitative changes in family policy in conservative states nor in Korea. This is supported by the fact that most countries included in the study stayed in the corresponding cluster in the past ten years. In the comparative analysis between conservative welfare states, Leitner (2006) maintained that prior similarities between Germany and Austria, and France and Belgium are shifting toward common traits shared between Austria and Belgium, and Germany and France. An analysis of family policy based on defamilization and familization, however, yields no such trend during the period up to 2003.

Figure 3

Results from multidimensional scaling: Conservative welfare states and Korea