Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary* Zsuzsanna Makay Abstract: This study focuses on changes in the length of paid parental leave in Austria, France and Hungary between the 1960s and the fi rst decade of the 2000s. Its aim is to analyse to what extent extensions and reductions of the duration of paid parental leave affect mothers’ labour market entry after childbirth. For each country, periods according to the different policy reforms are analysed and data from the Generations and Gender Survey are used to account for changes in mothers’ labour market entry. Scholars have argued that too long periods of paid parental leave might act as work-reducing policy in that they hinder women’s labour market entry while short leaves have positive effects on labour market participation and wages. This phenomenon is studied in three countries with very different conciliation policies. Results for Austria show that the introduction of more fl exibility into the leave legislation in 2008 increased mothers’ relative risks of labour market entry especially among higher educated women. In France, the extension of paid leave for second born children in 1994 reduced labour market entry. We found however, no signifi cant effect of the 1986 reform for third children and the 2004 reform introducing a paid leave of six months for fi rst-born children. In Hungary, labour market entry of mothers was infl uenced by structural changes which accompanied the political transition of 1989 and resulted in longer leave take up of mothers. Labour market entry before the transition was more intensive than ever since that. Keywords: Parental leave · Family policy · Labour market entry after childbirth · Policy reforms · Work family conciliation Comparative Population Studies Vol. 48 (2023): 425-456 (Date of release: 17.08.2023) Federal Institute for Population Research 2023 URL: www.comparativepopulationstudies.de DOI: https://doi.org/10.12765/CPoS-2023-17 URN: urn:nbn:de:bib-cpos-2023-17en8 * This article belongs to a special issue on “Demographic Developments in Eastern and Western Europe Before and After the Transformation of Socialist Countries”. http://www.comparativepopulationstudies.de https://doi.org/10.12765/CPoS-2023-17 • Zsuzsanna Makay426 1 Introduction Parental leave policies have contradictory effects on women’s employment. On the one hand, long and paid parental leaves coupled with employment security make it possible for women to combine childrearing with employment. On the other hand, they prevent women from achieving high-paying jobs, since employers are reluctant to hire women who may leave the labour market for a longer period of time (Mandel/ Semyonov 2006). Thus, generous welfare states, which make it possible for women to combine work and family life, can affect labour force participation of mothers in either a positive or a negative way, depending on the length of the leaves, the security of the employment and childcare arrangements. This so-called welfare state paradox has been demonstrated in different contexts and the importance of policies that “promote parental attachment to the labour force” (Misra et al. 2011: 154) has been emphasised by several studies using aggregated data (Misra et al. 2011; Mun/Jung 2018; Thévenon 2011). The objective of this study is to investigate the effect of various changes in the length of paid parental leave (PPL) since the 1960s on women’s employment after childbirth in Austria, France and Hungary using individual data on employment and childbirth histories. Several reforms of PPL took place in the last four decades in the three countries, and available data permits us to study how women’s post-birth employment changed with the subsequent reforms. We analyse in detail which factors infl uenced labour market entry after childbirth and how intensity of labour market entry has changed over time. Thus, we investigate how the extension of PPL (which may operate as a work-reducing policy) and shortenings of PPL (which may act as a work-facilitating policy, while both effects may highly depend on other factors) affect mothers’ employment after childbirth. This cannot be studied without contextualising each reform and taking into account other aspects that infl uence the conciliation of family life and paid work (Neyer/Andersson 2008). So, the political transition of 1989 in Hungary will be investigated, as well as other aspects of work- life balance policies (childcare availability, job security, etc.). The main contribution of this article is to show that changes in the length of parental leave by themselves have very small effect on the length of women’s non-employment after childbirth. Conciliation strategies are indeed developed in the context of social norms, personal desires and structural possibilities, namely labour market situation and the availability of childcare facilities. 2 Factors affecting labour force participation of mothers with young children The employment of mothers with young children has risen constantly over the last decades in the European countries. But it is highly dependent on the number and the age of the children and on structural factors, such as the availability of childcare, the prevalence of part-time work or leave policy benefi ts (Bernhardt 1993; Del Boca Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 427 2002; Hantrais 2004). Indeed, there are important differences in family policies and in the manner the different states support mothers’ labour market participation. In an article published a decade ago, Misra and colleagues (2011) spoke about the welfare-paradox concerning parental leave policies since these policies can help and prevent mothers’ employment and career advancement at the same time. Namely, generous welfare policies assuring long parental leaves for mothers may decrease their labour force participation and wages. Authors call these “work-reducing policies” since they encourage mothers to stay off work. Work-facilitating policies, on the contrary, encourage women to take up work and support this decision with the possibility to resort to different childcare arrangements (Misra et al. 2011). Too long parental leave arrangements seem to have other side-effects as well. They may be responsible for discrimination against women in the labour market in that they offer job-protection to mothers which is offered neither to childless women nor to men (Mandel/Semyonov 2006). They also imply nonwage costs on fi rms since they have to deal with replacements and the uncertainty about the return of their employee. This effect is amplifi ed in the case of job protection since the position must constantly be available to the mother on leave (Ruhm 1998). Thus, employers may prefer to hire their workforce among people not benefi tting from this protection and neglect mothers’ application or that of women of childbearing age. Generous welfare states may also create “sheltered” labour markets for mothers where their rights are protected and secured, most commonly in the public sector. In this regard, the State acts as an employer and guarantees access to this secured sector for mothers. This facilitates women’s labour force participation and the conciliation of work and childrearing. In the same context however, the private sector may be inaccessible for the protected group which creates labour market segregation with mothers having diffi culties of entering valuable elite positions. Indeed, these are not compatible with their family roles which in return affects their wages and career opportunities negatively. Moreover, it has negative implications for gender egalitarianism on the labour market and in the family sphere (Bianchi et al. 1999; Mandel/Semyonov 2006; Thévenon 2011) So, different welfare regimes promote mothers in the labour market differently. The length of paid parental leave is one aspect of mothers’ labour force participation, which should neither be too short nor too long in order to promote labour market participation and equal wages. Several scholars have established estimations concerning the “optimal” length of parental leave. Based on aggregated data in 16 European countries over the period 1970 to 2010 and taking into account labour force participation and wages, they came to the conclusion that the optimal length of paid parental leave is between 20 weeks and 30 weeks. Any extension of the leave only has minimal impact on participation rates, while it decreases wages and causes occupational segregation. The negative effects are especially visible among women in high-level occupations (Akgunduz/Plantenga 2013; Bianchi et al. 1999; Ruhm 1998). It cannot be denied however that longer leaves have been found to have a positive effect on the health and the development of children as well as on fertility (Lalive/Zweimüller 2009; Ruhm 2004, 2000). • Zsuzsanna Makay428 It is not only policies that shape mothers’ employment after childbirth, but also personal values and attitudes. According to socialisation theory, women’s employment decisions are driven by their preferences and feelings which are partly determined by socialisation (Blair-Loy 2003). As stated by Kremer, care is crucial for women since it shapes their identity. So, women’s decisions concerning work are often made in the context of care which means that if mothers do not work, this is because they do not want to; they prefer to care. However, “ideals of care” differ according to welfare states which are shaped by diverse cultural dimensions (Kremer 2007). At the same time, values may not only differ between countries, but also within countries, even though the main characteristics and preoccupations of women – namely the effort to conciliate work and family life – may be the same (Hakim 1999). Indeed, women see career and family as competing goals, which are not compatible since both require whole devotion. Qualitative interviews with highly educated mothers showed that they do not even challenge to combine the two spheres but rather resign to having children and giving up their career after childbirth (Blair-Loy 2003). The incompatibility of career and motherhood was also emphasised recently in a study conducted among Polish women, which found that even though women were very much motivated to work for pay, they found a professional career too demanding to be combined with childrearing (Matysiak/ Mynarska 2021). The tendency to challenge the conciliation of work and family may be lower in countries where the male-breadwinner model is prevalent and where social norms expect mothers to care for their children. Opinions concerning mothers and fathers respective roles are decisive from this point of view, and they are especially conservative in Hungary where 80 percent of individuals of reproductive age agree that “A pre-school child suffers, if his/her mother works”, while in Austria and France this is around 50 percent (Panova/Buber-Ennser 2016). The availability and the quality of childcare is an important factor of mothers’ labour market participation, since it decreases the opportunity costs of childrearing. However, studies analysing this effect among mothers of children below the age of three years show mixed results (Neuberger et al. 2022). Some studies found a positive correlation between the availability of childcare and labour market decisions (Brewer et al. 2022; Del Boca 2002), while in Germany for example, scholars came to the conclusion, that the increase of subsidised childcare after 2008 raised maternal participation rates only modestly. Furthermore, it could not be concluded that the lack of childcare accounted for low participation rates of mothers with children below three (Bick 2016). Another study, conducted in Germany as well, showed positive effects but only in West Germany and after the birth of a second child (Zoch/Hondralis 2017). In Japan, an important reform of childcare availability increased women’s labour market entry, but the effect was mainly visible among women who were strongly attached to their job and thus had the highest probability to use childcare (Yamaguchi et al. 2018; Zhang/Managi 2021). Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 429 Parental leave in the three countries of Austria, France and Hungary was modifi ed several times in the last decades. In the following, we look at differences and similarities between the three countries and examine PPL reforms. 3 Parental leave and women’s employment in Austria, France and Hungary Even though Austria and France both belong to the ‘conservative-corporatist’ welfare regime in Esping-Andersen’s typology of welfare regimes (Esping- Andersen 1990), they are very different in what concerns family policy, labour force participation of mothers or the availability of external childcare. Hungary, as a post-socialist country also differs in several aspects, despite the fact that in an attempt to incorporate post-communist countries into the classical typology of the welfare regimes, it has been demonstrated, that these countries mostly resemble the conservative-corporatist group, even though they are characterised by a lower level of governmental programmes and social assistance (Fenger 2007). Still, low female participation, reliance upon social contributions instead of taxes, moderate income redistributions, and rather high levels of unemployment are typical in Esping-Andersen’s conservative-corporatist group as well as in Central-European countries. While in Western Europe, it was mainly after the 1970s that women’s labour force participation increased rapidly, in Eastern Europe this happened earlier, since communist ideology encouraged women’s employment, and the labour market needed their workforce. Therefore, childcare facilities were developed and the family impact of childrearing on labour force participation during the following decades was smaller than in several Western European countries (Van der Lippe/ Van Dijk 2002). In Hungary, a fi rst paid childcare leave of three years was introduced in the 1960s, and women extensively made use of it throughout the following forty years. Indeed, the majority of mothers left the labour market after childbirth for several years to care at home for their children (Inglot 2008; LaFont 2001; Massey et al. 1995). This incorporated the idea into the Hungarian society that at least until the age of three years, mothers’ have to take care of their children themselves. This norm has persisted far beyond the political transition of 1989 and still, in 2022, prompts the great majority of women to take a leave of several years after childbirth (Hasková et al. 2012; Makay 2018). The historical background was different in Austria and France, where women entered the labour market in response to the development of the service sector and following major changes in women’s social status (Prskawetz et al. 2008). In France, parental leave benefi ts and childcare facilities were developed parallel to the rise in women’s participation, while in Austria, because of the lack of childcare arrangements, part-time work became a strategy of work-family conciliation. As a result, parental leave benefi ts and the labour force participation of mothers with young children vary considerably between the three countries. So do childcare policies, and this results in very different post-employment patterns. In 2005, only • Zsuzsanna Makay430 6.6 percent of children below the age of three years were enrolled in early childhood education and care services in Hungary, while this was 44 percent in France. In Austria, where we have data for the year 2010, 12.5 percent of children were in this situation.1 Different developments concerning parental leave benefi ts resulted in these contrasts, and we will look at these in more detail country by country. 3.1 Austria Austria has had a strongly familialistic family policy until the last decade, characterised by a gendered division of work and care, a male breadwinner model and a conservative ideology (Blum 2010; Neyer 2010; Valarino et al. 2018). The fi rst paid parental leave was introduced in 1961 and until 1990, its maximal duration was 12 months (Blum 2010). The benefi t was employment related and dependent on former income. In 1990, the duration of the leave was extended to 2 years, and fathers became eligible to take the leave. Because of strict eligibility criteria, however, very few fathers did so. In 1995, the leave was shortened again and a “fathers’ quota” was introduced: parents could only take the whole 24 months of the leave, if the father took at least six months. The major change occurred, however, not in fathers’ participation (which changed very little), but in work requirements needed for automatic renewal of parental leave eligibility which had an important effect on fertility timing (Lalive/Zweimüller 2009). In 2002, a universal childcare allowance was introduced which was independent of previous employment. It could be claimed for a maximum of 30 months, completed by six months for the second parent. Job protection however remained for 24 months only. In 2008, a more fl exible model was introduced with different choices concerning the length of paid leave and the amount received: 30+6 months for the second parent with 436 EUR/month, 20+4 months with 624 EUR/month or 15+3 months with 800 EUR/month. In 2009, the fi rst option was the most popular among Austrian families. In 2010, two additional short options were introduced with the aim to encourage women to enter the labour market sooner and thus reduce the opportunity costs of childbearing. The different options could also be combined with labour market participation and additional earnings, even though with limited amounts (Rille-Pfeiffer 2011). The reforms of the Austrian parental leave system during the fi rst decade of the 2000s are seen as a “muddled modernisation”, since the system was complemented with new elements which were added to the existing ones without major shift towards better work-life balance or gender equality (Blum 2010; Buber-Ennser 2015). Indeed, despite the reforms, the lack of substantial development of childcare facilities before the mid-2000s made it very diffi cult for women to conciliate employment with childrearing and left the gendered aspect of the system mostly unchanged (Auer/Welte 2009; Leibetseder 2013). The enrolment in external childcare of children below age three was only 11.8 percent in 20072 and important regional differences 1 OECD Family Database. 2 It increased rapidly afterwards: 23.3 percent in 2013 and 25.5 percent in 2017 (Buber-Ennser 2015; Mauerer 2018). Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 431 persisted with variable opening hours making it diffi cult for both parents to work full-time (Buber-Ennser 2015). Indeed, part-time work became one of the main conciliation strategies of work and family in Austria which is largely adopted among lower and higher educated mothers and is used as a long-time arrangement of conciliation (Riederer/Berghammer 2020). This is coupled among families with pre- school children with the return of a “modernised” male breadwinner model where the mother works part-time and the father full-time (Berghammer 2014). 3.2 France In France, women’s employment rates began to rise in the 1970s and increased steadily since then, reaching relatively high rates in the 2000s compared to other European countries (Fig. 1). Maternal employment rates were also relatively high: even with a child below the age of three, more than half of the mothers were in employment in 2005. Indeed, compared to the other two countries, France stands out with a benefi t system that supports mothers’ quick return to the labour market after childbirth, a widely available childcare system and social norms that accept external childcare for children from a very young age (Salles et al. 2010; Toulemon et al. 2008). Maternity and parental leave benefi ts were developed in the second half of the 20th century. A paid maternity leave of 16 weeks was introduced in 1978 and extended later to 26 weeks for third or higher order births (Odul-Asorey 2013). A paid parental leave of three years was introduced in 1985 for families having three children or more which was extended to second-born children in 1994 with relatively modest replacement rates, but with the possibility to take the leave on a part-time basis (Collombet 2016). This reform negatively affected labour market participation of mothers, since participation rates decreased to levels of the beginning of the 1980s. It was mainly low educated mothers and mothers in precarious working situations who resigned from paid work after a second birth. The reform also drastically increased leave take-up after third births, especially among women who left the labour market after a second birth. Besides, 98 percent of benefi ciaries were women, and interruptions negatively affected their income after their re-entry into the labour market (Afsa 1996; Lequien 2012; Piketty 2005). In 2004, a parental leave of six months was introduced for fi rst births, which resulted in shorter career interruptions, but its wage penalties for mothers are visible until two years after their return to the labour market (Joseph et al. 2012). Parental leave benefi ts in France were always related to employment with relatively strict requirements. For being eligible for paid parental leave during six months after a fi rst birth, mothers have to be on the labour market for two years before the birth, and requirements are not less strict for higher order births (even though they take into account eventual periods off employment because of a previous birth). At the same time, diverse forms of external childcare arrangements are available and several benefi ts enable families to combine childrearing and employment (Toulemon et al. 2008). • Zsuzsanna Makay432 3.3 Hungary In Hungary, as in other socialist countries, women began to enter the labour market in large numbers from the beginning of the 1960s. Socialist ideology saw labour force participation as gender egalitarianism and encouraged women to take up work outside of the family sphere. In addition, the massive industrialisation required a large number of workers, and men as well as women were mainly employed in heavy industry factories. Compared to Western countries, labour force participation of women was already high: 71 percent of women aged 18-54 were active in 1972 and 80 percent in 1987 (Abraham/Kezdi 2000). Most women were eligible for parental leave and parental leave benefi ts. In the middle of the 1960s, a fi rst paid parental leave of three years was introduced. The leave was very popular among the mothers who suffered from a “double shift”: after their long working-day, they had to do the vast majority of the child-related tasks themselves, since fathers’ involvement was not very common. Around 44.4 percent of mothers having a child younger than two years old were on childcare leave in 1970. In 1980, 68.8 percent were in this situation, while this number increased to 73.2 percent in 1987 (Központi Statisztikai Hivatal 1989). A second parental leave benefi t was introduced in 1984, which was income related and could be claimed for two years. Nurseries were developed for children under the age of three years, and kindergarten from the age of three was almost universally accessible on a full-time basis. After the change of the political regime in 1989, parental leave benefi ts remained unchanged, but the economy underwent drastic changes with privatisation, large drop in GDP, high unemployment rates and structural transformations (Kertesi/Köllő 2002). In addition, the number of places in nurseries for children under the age of three fell drastically: 64,000 places were available in 1980, 50,000 in 1990 and only 25,000 in 2000 (Központi Statisztikai Hivatal 2012). In 1995, the “Bokros Package” was introduced and meant important restrictions in public spending, the cut of social and family benefi ts, the introduction of tuition fees in higher education and other restrictive measures (Ilonszki/Kurtan 1996). Parental leave benefi ts became means-tested, their duration was shortened to a maximum of two years instead of three, and the more generous benefi t introduced in 1984 was abolished. Poverty rates increased drastically and birth rates dropped: In 1996, the total fertility rate decreased for the fi rst time to under 1.5 in Hungary. In 1998, a new government was elected and most of the restrictive measures were withdrawn. Paid parental leave was reformed: a fl at rate benefi t could be claimed during three years without previous employment requirements and a higher, income-related benefi t could be claimed until the second birthday of a child for women (and men) who worked prior to childbirth. This parental leave system persisted until 2008 . Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 433 4 Data and methods 4.1 Data We use the fi rst two waves of the Generations and Gender Survey (GGS Round 1): a longitudinal, representative survey on social and demographic topics conducted in the three countries in the 2000s. The same questionnaire was used during data collection in the three countries and data were harmonised and are freely available on the GGP website.3 In Austria, wave 1 was conducted between September 2008 and February 2009 among 5,000 men and women aged 18-45, while wave 2 was conducted four years later among the same respondents (age range 22-49). The sample was refreshed, but the new respondents are not in the harmonised GGS data fi le. 78 percent of fi rst-wave respondents answered at the second wave. In both waves, a computer- assisted (CAPI) face to face interview method was used.4 CAPI was also used in France, where the fi rst wave took place in 2005 among 10,079 persons aged 18-79 and the second wave in 2008 among the same persons, with a second response rate of 88 percent. In Hungary, the fi rst wave took place at the end of 2004 and the Fig. 1: Labour force participation rate of women aged 15-64 years, % 0 10 20 30 40 50 60 70 1960 1970 1980 1990 1995 2000 2005 Austria France Hungary Source: International Labour Organization. https://www.ilo.org/shinyapps/bulkexplorer38/ ?lang=en&segment=indicator&id=EAP_DWAP_SEX_AGE_RT_A 3 https://www.ggp-i.org/ 4 https://ggp.colectica.org/item/int.ggp/cddfada5-819d-4f02-8d62-df2e3c3a1220/2 https://www.ilo.org/shinyapps/bulkexplorer38/ https://www.ggp-i.org/ https://ggp.colectica.org/item/int.ggp/cddfada5-819d-4f02-8d62-df2e3c3a1220/2 • Zsuzsanna Makay434 fi rst months of 2005 among men and women aged 21-78 and the second wave at the end of 2008 and the fi rst months of 2009 among the same respondents. 13,540 people answered in the fi rst wave and 10,641 in the second. At both waves a paper and pencil questionnaire was used by the interviewer. The questionnaire included the complete fertility history of the respondents, including month and year of birth of each child ever born, and the information was updated in the second wave. Thus, births that occurred between the two waves are also recorded. Besides, during the second wave, retrospective employment histories were collected since the age of 16 of the respondents. Employment history was collected by asking respondents to take into account activities that lasted at least three months. A distinction was made between 12 different types of activities which we grouped into periods of “employment” (defi ned as employed and self- employed) and non-employment (student, helping family member, unemployed, retired, in military or social service, homemaker, maternity leave, parental leave, care leave, ill or disabled for a long time or permanently, other status).5 There are some statuses which could also have fi t into the “employed” category, such as helping family member in military service. These, however, concerned less than 2 percent of the women, and it seemed more relevant to take a strict defi nition of employment. We linked these employment and non-employment spells to the fertility histories of the respondents and thus constructed an event history database with all births where the exact birthdate of the child was known (month and year) and where the employment history of the mother included information of her pre- birth employment status and her activity after birth. The focus of the present study is on employment or non-employment after childbirth, that means, from the date of birth of the child until his/her age of 40 months. The cut-off period of 40 months was chosen, because in the three countries, the maximum length of paid (and unpaid) parental leave is 36 months.6 Running our analysis until a length of 40 months permits us to see how women using the maximal length of parental leave enter the labour market after a prolonged absence. Since we study leave take-up directly after childbirth (it may only include maternity leave for some respondents) only women are included in the analysis. Pre-birth employment is defi ned as employment (employed or self-employed) during the 12-months period before the birth of the child. Since only activity spells lasting at least three months were recorded, our criterion was to have an employment spell during the 12 months. This spell could begin earlier than 12 months before the birth or during this 12-month period, and end at any time before the birth of the child. In some cases where women worked during the whole pregnancy and took only a very short leave afterwards, we do not have any interruption in the 5 See the second wave questionnaire for more details http://www.ggp-i.org/sites/default/fi les/ questionnaires/GGS_Wave2_Questionnaire_V.2.0.pdf and the study profi le Gauthier et al. (2018). 6 Except for Hungary, where since 1993, women having at least three children can prolong the leave until the 8th birthday of the youngest child. http://www.ggp-i.org/sites/default/fi Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 435 employment spell. These women are treated as having taken only a three months maternity leave period after birth. Background variables include basic demographic information about the mother: her age, highest level of education (at the time of the fi rst interview), marital status (at the time of the current birth), pre-birth employment (if the mother worked or not), the birth order of the child. Detailed information on the sample and the distribution of the background variables can be found in table 1. Different periods were constructed in the three countries in order to analyse the effects of parental leave reforms on labour market entry after childbirth. In Austria, a paid parental leave of one year was available for employed mothers during our so-called period 1 (1981-1990). This was extended to 2 years afterwards (period 2: Tab. 1: Sample description Austria France Hungary N % N % N % Period1 Period 1 206 13.2 2439 48.4 283 13.2 Period 2 392 22.2 1029 19.7 668 31.4 Period 3 640 31.7 1203 21.2 414 18.7 Period 4 456 22.9 519 10.7 869 36.7 Period 5 183 10.0 - - Marital status at the time of birth Married 1241 68.0 3465 69.2 1913 85.3 Not married 636 32.0 1725 30.8 321 14.7 Highest level of education at the time of the interview Lower secondary or lower 254 16.4 1787 38.2 473 22.1 Upper secondary 1012 51.1 1957 37.2 1149 50.7 Post-secondary or higher 611 32.5 1446 24.6 612 27.2 Birth order First child 979 53.3 2462 46.7 1036 45.6 Second child 680 34.9 1750 33.1 831 37.7 Third or higher order child 218 11.7 978 20.2 367 16.7 Activity before / during the pregnancy Worked 775 56.9 3525 65.4 1023 45.1 Did not work 1102 43.1 1665 34.6 1211 54.9 Competing events (outcome) Entered work 492 26.8 3286 60.5 815 37.2 New childbirth 367 19.9 910 18.9 656 29.9 Censored 1018 53.4 994 20.6 763 32.9 Mean age (years) 27.2 27.9 26.3 N 1877 5190 2234 1 See the description of the periods after the table and table 2 for more details. Source: Generations and Gender Survey wave 1 and wave 2. Percentages are weighted. • Zsuzsanna Makay436 1991-1995). In 1996, the paid leave for mothers was shortened to 1.5 years, so period 3 stretches from 1996 to 2002. That year, the leave was extended to 30 months with the three months of “father’s quota” and the paid leave became universal, independent of previous employment (period 4: 2003-2007). In Austria, we have a fi fth period (2008-2012) to account for the introduction of the new fl exible form of parental leave. In France, period 1 goes from 1960 to 1985 where no paid leave was available after maternity leave. The introduction of a paid leave of three years after a third birth marks the second period (1986-1994). The leave was extended to second children (Period 3: 1995-2004). A paid parental leave of six months was introduced for fi rst-born children in 2005 (period 4: 2005-2008). In Hungary, the fi rst period (1965-1987) marks the socialist decades where two kinds of leave (of two and of three years) were available for previously working mothers. The years of the political-economic transition mark the next period where parental leave benefi ts remained unchanged but the labour market was highly affected by the political and economic transition and the number of nursery places decreased (1988-1994). The austerity years of the Bokros Package described above mark period 3 (1995-1998). During the last period (period 4: 1999-2008), parental leave was made universal and extended again to a maximum of three years. The different periods in the three countries are summarised in Table 2. 4.2 Methods We analyse labour force participation after childbirth and build event history models, where the dependent variable is the incidence of labour market entry. We measure time to entry on a discrete scale, starting our observation at the birth of the child (month and year) and ending when the mother enters labour market or 40 months after childbirth at the latest. Labour market entry after childbirth is likely to be infl uenced by further fertility plans (Fahlén 2013), since many women take advantage of parental leave benefi ts to have several children without entering or re-entering the labour market after birth. In addition, in the three countries, average birth intervals between fi rst and second, second and third births are below three years, thus our analysis has to take into account the possibility of a new childbirth during the observation period. We decided to treat a second childbirth during the observations time as a “competitive” event meaning that a mother can only experience either labour market entry or a new childbirth during the maximum observation period of 40 months, but not both events. Indeed, in the competitive risks models, after experiencing one event, the mother is no longer at risk of experiencing the other event. Thus, we construct cause-specifi c hazard models which show the instantaneous risk of failure from one of our main cause, namely labour market entry. For this purpose, we use the “stcrreg” command in Stata and show in the output the subhazard ratios (SHR) which are the exponentiated regression coeffi cients. We also compute cumulative incidence functions (CIF) from the subhazard ratios to visualise the results of the competing-risks regression (Cleves et al. 2016). Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 437 T a b . 2 : T im e p e ri o d s ta ke n in to a cc o u n t in t h e t h re e c o u n tr ie s P E R IO D 1 P E R IO D 2 P E R IO D 3 P E R IO D 4 P E R IO D 5 A U S T R IA 19 8 1 -1 9 9 0 19 9 1 -1 9 9 5 19 9 6 -2 0 0 2 2 0 0 3 -2 0 0 7 2 0 0 8 -2 0 12 P L 1 y e a r p a id p a re n ta l P a id p a re n ta l l e a v e P a id p a re n ta l l e a v e f o r P a id p a re n ta l l e a v e o f 3 0 F le x ib le p a id p a re n ta l le a v e f o r m o th e rs e x te n d e d t o 2 y e a rs m o th e rs s h o rt e n e d t o m o n th s + 6 m o n th s le a v e w it h 3 o p ti o n s 1. 5 y e a rs + 6 m o n th s "f a th e rs q u o ta " "f a th e rs q u o ta " E lig ib ili ty P re v io u sl y e m p lo y e d P re v io u sl y e m p lo y e d P re v io u sl y e m p lo y e d U n iv e rs a l U n iv e rs a l p a re n ts p a re n ts p a re n ts F R A N C E 19 6 0 -1 9 8 5 19 8 6 -1 9 9 4 19 9 5 -2 0 0 4 2 0 0 5 -2 0 0 8 P L N o p a id p a re n ta l l e a v e; P a id p a re n ta l l e a v e o f 3 P a id p a re n ta l l e a v e o f P a id p a re n ta l l e a v e o f 6 m a te rn it y l e a v e o f y e a rs a ft e r th ir d o r 3 y e a rs a ft e r m o n th s a ft e r fi rs t b ir th s 16 w e e ks h ig h e r o rd e r b ir th s se co n d b ir th s E lig ib ili ty P re v io u sl y e m p lo y e d P re v io u sl y e m p lo y e d P re v io u sl y e m p lo y e d P re v io u sl y e m p lo y e d m o th e rs p a re n ts p a re n ts p a re n ts H U N G A R Y 19 6 5 -1 9 8 7 19 8 8 -1 9 9 4 19 9 5 -1 9 9 8 19 9 9 -2 0 0 8 P L P a id p a re n ta l l e a v e o f P a id p a re n ta l l e a v e o f P a id p a re n ta l l e a v e o f P a id p a re n ta l l e a v e o f 3 y e a rs 3 y e a rs 2 y e a rs 3 y e a rs E lig ib ili ty P re v io u sl y e m p lo y e d P re v io u sl y e m p lo y e d M e a n s te st e d a n d U n iv e rs a l m o th e rs m o th e rs p re v io u sl y e m p lo y e d p a re n ts S o u rc e : O w n d e si g n • Zsuzsanna Makay438 A mother can have several children, and this had to be taken into account in the analysis. Since we observe mothers labour market behaviour via their children, we had to account for the fact that the observations (this means the children themselves) were not completely independent. For this reason, we added the vce (cluster) option in the regressions, which adjusted the standard errors of the estimated parameters to account for the possible correlations. Data are weighted using the variable bweight of the GGS dataset. 5 Hypothesis Labour market entry after the birth of a child is expected to be more intensive in France than in Austria and Hungary, since in the latter countries, a paid parental leave of at least two years can already be taken after the birth of the fi rst child, and childcare for children under the age of one year is less developed than in France (H1). In line with the literature, we expect that due to their work-reducing effect, parental leave extensions directly and negatively affect labour market participation. This may be visible in Austria after 1991, when paid parental leave was extended to two years instead of one year and in 2003-2007, when it was prolonged to 30+6 months (H2_at). And in Hungary after 1999, when PPL became a universal right (H2_hu). In France, PPL depends on the number of children, so the work-reducing effect may vary according to birth order, since the extension of 1986 concerned third born children and that of 1994 second born children. A work reducing effect may also be visible after 2005, when a six-month paid leave was introduced after a fi rst birth (H2_fr). In Austria, the introduction of fl exibility into the PPL system in 2008 may have facilitated the conciliation and was likely to accelerate labour market entry (H3). In Hungary, structural changes may be more important than in the other two countries. The political and economic transition coupled with high unemployment was likely to reduce labour market entry after 1989. Parental leave benefi ts remained unaffected by the transition, and a three-year paid leave could be claimed after a birth (H4). Background characteristics may infl uence mothers’ labour force participation in the three countries. We look at education, since higher educated women’s human capital is more negatively affected by prolonged inactivity (Bernhardt 1993). So, we assume that education is an important predictor of the intensity of labour market entry after the birth of a child (H5). Mothers who worked before the birth are in general more likely to return quickly to the labour market, since they are more attached to it. Moreover, in the three countries, job security is guaranteed during parental leave (only during 24 months in Austria), so these mothers do not have to spend time until fi nding a new job. They may enter labour market more quickly after childbirth than women who did not work before the birth (H6). Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 439 6 Results 6.1 Descriptive results The cumulative incidence functions show that mothers in France enter the labour market much quicker after childbirth than Austrian and Hungarian mothers. In France, 43 percent enter the labour market after three months, which in most cases marks the end of maternity leave, and half of them work when the child is six months old. Afterwards however, labour market entry slows down and the majority of women who stayed at home do not enter the labour market before the child turns three. 40 months after birth around one third of the French mothers are not working (Fig. 2). In both Austria and Hungary, labour market entry after childbirth is less intensive, and in both countries, participation is rare in the fi rst six months. After this, Austrian mothers slowly begin to take up work, but it is only after 16 months that the proportion of women already on the labour market reaches 10 percent. In Hungary, this happens much later, only at the second birthday of the child. In Austria, more than one fi fth of women enter the labour market around the second birthday of the child, and labour market entry increases in an accelerated manner afterwards. In Fig. 2: Intensity of labour market entry after childbirth in Austria, France and Hungary 0 6 12 18 24 30 36 42 Time since birth in months Austria France Hungary 1 .75 .5 .25 0 Cumulative incidence functions. Source: Own calculations based on the Generations and Gender Survey. • Zsuzsanna Makay440 Hungary, the labour market-entry is mainly concentrated around the third birthday of the child. At the end of the observation period, less than half of Hungarian and Austrian mothers have taken up work. As described above, we consider a next childbirth as a competing risk of entering employment. Thus, women who have a new child are no longer at risk of labour market-entry. Cumulative incidence rates for a subsequent childbirth show that in Austria and Hungary, almost 25 percent of women have a next child within 40 months without entering the labour market between the two births. In France, this proportion is 12 percent, since women are more likely to enter the labour market rapidly after childbirth. They may quit it again for a next birth (Fig. 3). 6.2 Results of the regression models Regression models were performed separately for the three countries in order to be able to incorporate the different time periods. All results are weighted. Austria Compared to the years 1981 to 1990, women who had a child between 2008 and 2012 were more than two times more likely to enter the labour market while controlling for background characteristics (Fig. 4). Indeed, their subhazard ratio is increased signifi cantly with relatively large confi dence intervals which is probably due to the lower sample size of this group. Labour market entry is not signifi cantly different at the other periods. Married women are less likely to take up a job than non-married women. Their subhazard ratio is decreased by 17 percent. Birth order does not have any effect on employment after birth in Austria. The risk of highly educated women is however increased by 44 percent compared to low educated. Having worked before the birth matters as well: those mothers who had a job see their risk increased by 2.8. In Austria, younger women are a bit more likely to take up work after childbirth: a one- year increase in age slightly reduces the relative risks. Fig. 3: Cumulative incidence of labour market entry and the birth of a next child 0 6 12 18 24 30 36 42 time_comp Austria Cum. Inc. Work Cum. Inc. Childbirth France Hungary 1 .75 .5 .25 0 1 .75 .5 .25 0 0 6 12 18 24 30 36 42 1 .75 .5 .25 0 time_comp 0 6 12 18 24 30 36 42 time_comp Source: Own calculations based on the Generations and Gender Survey. Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 441 When we look at the interaction of education and period, we see that in Austria, compared to medium educated women, only highly educated mothers see their labour market entry risks increased at period 5, while there is no signifi cant change among women with medium or lower education. We see this result in Figure 5, where only the two signifi cant interaction effects are presented. It becomes evident that the effect of education has changed its meaning over the last decades. Compared to the 1981-1990 period, where higher educated women entered the labour market less likely than medium educated ones (their risk was reduced by 52 percent), this changed radically and they had a 2.4 times higher risk to do so in 2008-2012. In the latter period, more than 75 percent of women entered the labour market in the 40 months after birth while there were only 25 percent to do so in the 1980s . Interaction effects of the periods and pre-birth employment show that the latter signifi cantly infl uences post-birth employment at every period (Fig. 6). In 1981- 1990, women who worked before the birth were four times more likely to enter employment after birth than those who were not working. This effect decreased somewhat in 1991-1995, where labour market entry of working mothers decreased a bit while that of non-working mothers did not change signifi cantly. Between 1996 and 2002, participation of women who have worked during the pregnancy increased Fig. 4: Subhazard ratios for labour market entry in Austria (with 95% CI) 1981-1990 (ref) 1991-1995 1996-2002 2003-2007 2008-2012 Not married (ref) Married Low (ref) Middle High First child (ref) Second child Third child or more No (ref) Yes Age^2 Period Marital status Level of education Birth order Worked before the birth 0 1 2 3 4 Subhazard ratios Red dots mark signifi cant effects (p<0.10). Source: Own calculations based on the Generations and Gender Survey. • Zsuzsanna Makay442 again to their level of the 1980s, but decreased thereafter between 2003 and 2007 where PL entitlement became universal. This change of the legislation negatively affected labour market entry of both working and non-working mothers. After the introduction of a more fl exible parental leave system in 2008 however, there was a sharp increase in job take-up among mothers who have worked before the birth of the child. Fig. 5: Competing risks regression of labour market entry in Austria according to highest level of education and periods 0 0 6 12 18 24 30 36 42 Time since birth in months 1981-1990 Middle educated (ref.) 1981-1990 Highly educated 2008-2012 Highly educated .25 .5 .75 1 Cumulated incidence functions based on the exponentiated regression coeffi cients (subhazard ratios) of the interaction terms. Only signifi cant results are shown. Source: Own calculations based the Generations and Gender Survey. Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 443 France In France, there was not much change in the relative labour market entry risks since 1960. Compared to the reference period of 1960-1985, there is only signifi cant change between 1986-1994 when relative risks were increased by 11 percent (Fig. 7). Marital status and age do not infl uence labour market entry, while the other variables are signifi cant. Relative risks decrease with the number of children. After a second birth, the risk is decreased by 23 percent, while it is 37 percent lower after a third or a higher order birth. Working status in the 12 months before the birth is a very important predictor of post-birth employment in France, since women who worked at the beginning of the pregnancy are almost nine times more likely to take up work after childbirth. Interaction effects show, however, that the importance of pre-birth employment has somewhat decreased over the last decades. While in 1960-1985, working women were 14 times more likely to enter the labour market after birth than inactive women, the difference of the risk decreased to 13 times in the 2005-2008 period. At the mean time, job-entry of women who did not work before childbirth already increased (fi gures available on request). Fig. 6: Competing-risks regression of labour market entry in Austria according to labour force participation before the birth and periods 0 .25 .5 .75 1 0 6 12 18 24 30 36 42 Time since birth in months 1981-1990 Didn't work (ref.) 1981-1990 Worked 1991-1995 Worked 1996-2002 Worked 2003-2007 Didn't work 2003-2007 Worked 2008-2012 Worked Cumulated incidence functions based on the exponentiated regression coeffi cients (subhazard ratios) of the interaction terms. Only signifi cant results are shown. Source: Own calculations based the Generations and Gender Survey. • Zsuzsanna Makay444 Interaction effects show that while highly educated mothers take-up a job more often after childbirth, their labour market entry risks have somewhat decreased between 2005-2008 compared to the fi rst two periods (1960-1980 and 1986-1994). Concerning very low and medium educated mothers, there is now signifi cant change over the last decades Figure 8. We look more attentively at interactions with birth order, since in France the length of PPL depends on the number of children (as mentioned in the country description). To see the effect of the extension of PPL to third birth in 1986, we compared these to third order births in the previous period. Results show that there is no signifi cant effect of the reform despite the fact that mothers entered the labour market somewhat more rapidly after a third birth in 1986-1994 than in the previous period (Fig. 9, Panel A). The extension of PPL to second-born children in 1994, however, reduced mothers’ labour market entry by 28 percent compared to second children born between 1986 and 1994, and the cumulative incidence curves show that labour-market entry was slowed down in an important manner during the whole observation period of 40 months after childbirth (Fig. 9, Panel B). Fig. 7: Subhazard ratios for labour market entry in France (with 95% CI) 1960-1985 (ref) 1986-1994 1995-2004 2005-2008 Not married (ref) Married Low (ref) Middle High First child (ref) Second child Third child or more No (ref) Yes Age^2 Period Marital status Level of education Birth order Worked during pregnancy 0 2 4 6 8 10 Subhazard ratios Red dots mark signifi cant effects (p<0.10). Source: Own calculations based on the Generations and Gender Survey. Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 445 Fig. 8: Competing risks regression of labour market entry in France according to highest level of education and periods 0 0 6 12 18 24 30 36 42 Time since birth in months 1960-1985 Low educated 1960-1980 Middle educated (ref.) 1960-1980 Highly educated 1986-1994 Highly educated 1995-2004 Highly educated 2005-2008 Highly educated .25 .5 .75 1 Cumulated incidence functions based on the exponentiated regression coeffi cients (subhazard ratios) of the interaction terms. Only signifi cant results are shown. Source: own calculations based the Generations and Gender Survey. Fig. 9: Competing-risks regression of labour market entry in France after a third, a second and a fi rst child at different periods Panel A 0 0 6 12 18 24 30 36 42 Time since birth in months Panel B 0 6 12 18 24 30 36 42 Time since birth in months 1986-1994 2nd child 1995-2004 2nd child Panel C 0 6 12 18 24 30 36 42 Time since birth in months .25 .5 .75 1 0 .25 .5 .75 1 0 .25 .5 .75 1 Cumulated incidence functions based on the exponentiated regression coeffi cients (subhazard ratios) of the interaction terms. Only the results of the middle panel are signifi cant. Source: Own calculations based the Generations and Gender Survey. • Zsuzsanna Makay446 The third reform of PPL was the introduction of a six-month leave for the fi rst child in 2004, and compared to the previous period, this reform did not affect labour market entry of fi rst-child mothers (Fig. 9, Panel C). Hungary Compared to the years of state socialism (1965-1987 in our analysis), labour market entry after childbirth was less common in Hungary at every period until the 2010s. (Fig. 10). Between 1988 and 1994 and 1995-1998, the risk was reduced by around 40 percent, while between 1999-2008 it was reduced by 30 percent compared to the years preceding the political transition. Labour market entry is also signifi cantly infl uenced by all covariates in Hungary except for marital status. Birth order has an unexpected effect. Compared to fi rst children, labour market entry is more common among women having a second child (and less common among those having a third or higher order birth). This may be the effect of a strong two-child norm in Hungary and of a norm of timing, according to which most second-born children arrive in the three years after the fi rst child. Our competitive-risks models took into account the fact that either of the two events (a Fig. 10: Subhazard ratios for labour market entry in Hungary (with 95% CI) 1965-1987 (ref) 1988-1994 1995-1998 1999-2008 Not married (ref) Married Low (ref) Middle High First child (ref) Second child Third child or more No (ref) Yes Age^2 Period Marital status Level of education Birth order Worked before/during pregnancy 0 1 2 3 4 Subhazard ratios Red dots mark signifi cant effects (p<0.10). Source: Own calculations based on the Generations and Gender Survey. Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 447 new childbirth or labour-market entry) can arrive but not both. Still, by looking only at the labour-market entry risks for women who did not have a next child within 40 months, those having had a second child enter the labour market more rapidly than those having had a fi rst child. This result is somewhat puzzling. We tend to think that mothers with a fi rst child stay at home and wait for the next pregnancy which has not yet occurred while those of a second child who are less likely to plan a third one may speed up labour market entry. Women who have worked in the 12 months before the birth of a child are more likely to enter the labour market: their risk is increased by 41 percent compared to women who have not worked. The effect is less pronounced than in France and in Austria. Women having medium or high education levels are more likely to work after childbirth. The risk for women with a middle level of education is two times higher, while in the case of highly educated women the risk is almost three times higher compared to low educated women. Age is signifi cant, but only has a slight effect . The interaction between the periods and highest level of education shows that in Hungary, education level always was an important predictor of labour market entry (Fig. 11). However, even by controlling for the other variables, it remains evident that Fig. 11: Competing-risks regression of labour market entry in Hungary according to highest level of education and periods Cumulated incidence functions based on the exponentiated regression coeffi cients (subhazard ratios) of the interaction terms. Only signifi cant results are shown. Source: Own calculations based the Generations and Gender Survey. 0 6 12 18 24 30 36 42 Time since birth in months 1965-1987 Low educated 1965-1987 Middle educated (ref.) 1988-1994 Low educated 1988-1994 Middle educated 1995-1998 Low educated 1995-1998 Middle educated 1999-2008 Low educated 1999-2008 Middle educated 0 .25 .5 .75 1 • Zsuzsanna Makay448 labour market entry has decreased since the political transition and this for both low educated women and for women who have completed secondary education, while there is no signifi cant effect for higher educated women. For middle educated women, it was mainly the years of transition (1988-1994) that negatively affected their labour market entry, which decreased by 46 percent. It slightly increased afterwards. Low educated women’s labour market entry was affected both by the transition and by the economic restrictions during the 1990s: Their job-entry was lower in the second half of the 1990s than during the transition years. No signifi cant effect was found among highly educated mothers . Pre-birth employment affects post-birth employment, and compared to the early years of 1965-1987, non-working mothers’ labour market entry has signifi cantly decreased at every period by around 55 percent. The same is true for mothers who worked before the pregnancy. Their probability of entering the labour market has decreased during the transition years and in the period of austerity between 1995 and 1998 (Fig. 12). 7 Conclusion and discussion Using the retrospective employment histories and fertility histories in the Generations and Gender survey, the aim of this article was to account for the effect of the extensions and reductions of paid parental leave periods on labour market entry after childbirth in Austria, France and Hungary. Scholars have indeed argued that too long periods of paid parental leave act as work-reducing policies in that Fig. 12: Competing-risks regression of labour market entry in Hungary according to labour force participation before the birth and periods Cumulated incidence functions based on the exponentiated regression coeffi cients (subhazard ratios) of the interaction terms. Source: Own calculations based the Generations and Gender Survey. Women who didn't work before/during the pregnancy 0 0 6 12 18 24 30 36 42 Time since birth in months Women who worked before/during the pregnancy 0 6 12 18 24 30 36 42 Time since birth in months 1965-1987 1988-1994 1995-1998 1999-2008 .25 .5 1 .75 0 .25 .5 1 .75 Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 449 they hinder women’s labour market entry, what contributes to wage penalty and to labour market segregation. Since the 1960s, paid parental leave was extended several times in the three countries, which differ considerably regarding work-family conciliation: Differences are visible in parental leave regulations, the social norms regarding mothers’ employment and the availability of external childcare. They are also diverse regarding their historical background, with Hungary having transitioned from a socialist country to a democratic regime during the observation period. Results have shown that labour market entry after childbirth is much more intensive in France than in the other two countries (H1 confi rmed) which was what we expected. Results concerning the changes of PPL and our other hypothesis are less obvious. In Austria the extension of PPL to 2 years in 1991 and to 30+6 months in 2003 did not signifi cantly affect labour market entry compared to the 1980s, which means that women’s time spent at home with their child after a birth did not signifi cantly increase (H2_At rejected). Indeed, in Austria, the one year PPL did not mean that mothers took up work after the fi rst birthday of their child in the 1980s, they already stayed at home longer (Leitner 2014). The fact that the results were not signifi cant in the regression models means that the extension of the PPL was in line with mothers’ desire to prolong childcare at home since leave preferences favour a long leave in Austria (Valarino et al. 2018). Our results are in contradiction with previous ones, since it had been shown that the extension negatively infl uenced the return to work. Our descriptive fi ndings showed the same result (available on request), while the effect disappeared in the regression model. This may be because of the different time frame and methodology: Our dataset compares the period 1981-1991 to the next one, instead of comparing the months before and after the reform. Besides, we control for several background variables, which were unavailable in the Austrian Social Security Database used by Lalive and Zweimüller who focused mainly on fi rst births (Lalive/Zweimüller 2009). In Hungary, the PPL reform of 1999 did not have a signifi cant negative effect on labour market entry, since subhazard ratios increased compared to the previous period (H2_HU rejected). With this reform, the three-year PPL was introduced again in Hungary, and the previous employment requirements were abolished. The explanation of the lack of a signifi cant effect may be similar to that in Austria: independently of the reforms, the majority of women already took a leave of around three years and so the reform did not introduce signifi cant changes. They mostly continued to stay at home for a longer period after childbirth. In France, PPL reforms were changed differently according to the number of children. Among the three reforms studied, the 1994 reform (which introduced a leave of three years for second-born children) negatively and signifi cantly affected labour market entry of mothers. Their subhazard ratios were reduced by 28 percent compared to the period 1960-1985 (H2_FR partly confi rmed). This confi rms the results of previous studies (Afsa 1996; Piketty 2005). The other two French reforms were not found to signifi cantly reduce mothers’ labour market entry. The reform of 2004 concerning fi rst children may be without effect, because of the relatively • Zsuzsanna Makay450 low income replacement of the benefi t and the possibility to take it on a part-time basis. Social norms that favour mothers’ quick return to the labour market and the availability of childcare facilities may lead mothers to resume work rapidly after a fi rst birth. The introduction of a more fl exible PPL system in Austria led to more intensive labour market entries after 2008, since subhazard ratios were increased by 2.4 compared to the reference period of 1981-1990, but also compared to the period 2003-2007 (H3_confi rmed). The development of childcare facilities after 2008 made it possible for women to enter the labour market more rapidly, and according to our results, this has mainly accelerated highly educated women’s entry into paid work. They may enter the labour market on a part-time basis, since this pattern became their main conciliation strategy over the last decades (Riederer/Berghammer 2020) we address the question of whether the parenthood effect on employment has declined. By following synthetic cohorts of mothers and childless women up to retirement age, we can study both the short-term and long-term consequences of having a child. We consider employment participation as well as working time and also perform analyses by educational level. Our study is based on the Austrian microcensus, conducted between 1986 and 2016, and uses descriptive methods, logistic regression models, and decomposition analysis. The results show that the increase in the proportion of part-time work has led to a declining work volume of mothers with young children, despite employment rates of mothers having increased across cohorts. Return to the workplace is progressively concentrated when the child is 3-5 years old, but the parenthood effect has become weaker only from the time children enter school. Part-time employment is primarily adopted (at least with younger children. In Hungary, labour market entry was less intensive after the political transition of 1989, a period characterised by high unemployment rates, the closure of nurseries and other economic constraints. PPL legislation did not change during the fi rst years of the transition, still mothers’ labour market entry was hindered (H4 confi rmed). Indeed, PPL was not able to compensate for the impact of economic and political crisis, since unemployment and inactivity were important predictors of poverty which considerably affected families with children (Gábos/Szívós 2002). So, even without modifi cation in the legislation, labour market entry of mothers was made signifi cantly more diffi cult by structural changes. Women of all levels of education were concerned and our results show that labour market entry since 1988 has never reached the levels of the years before the political transition. On the one hand, this may be due to the fact that the social norm of the caring mother is now defi nitely widespread in the society, and women do not see themselves entering the labour market before their child turns two or three years old. On the other hand, the post-socialist decades were characterised by the security of employment, which disappeared after the transition. Women did not have many personal choices, but the majority of them was in employment, and their return to the labour market was assured by the socialist ideology of full employment. This job security guaranteed by the state disappeared after the transition to market economy and made labour market-entry after childbirth more uncertain. Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 451 The importance of education and of a previous job for an accelerated labour market entry was confi rmed by our results (H5 and H6 confi rmed). In Austria, the importance of education has risen over the last decades, and it is mainly highly educated mothers who benefi t from fl exibility and are better able to conciliate work and family life, at least after the second birthday of the child. In France, it is mainly low educated women who exit the labour market for a longer time, and this negative effect has increased in 2005-2008. A previous job speeds up labour market entry, and this effect has increased in Austria over the last decades. Our study has some limitations. First, they rely on employment data collected retrospectively, so some of the information on the exact length of employment and non-employment periods may be biased. Second, childbirth and labour market participation around the time of birth is highly dependent on structural circumstances, both on the macro level (e.g. employment or unemployment rates) and the micro level (e.g. personal situation on the labour market, individual eligibility to paid parental leave or its replacement rates). Our data and the already high complexity of our analysis did not make it possible to account for additional structural factors, which would be worth to study in a more detailed manner. Finally, we could not include the very important changes that the three countries underwent over the last decades into our analysis, namely the educational expansion, the overall rise in women’s employment rates and the changing transformation of parenthood and gender attitudes (Goldscheider 2000). Still, our results show the important interrelation between PPL and other structural and social factors. Indeed, reforms of PPL do not seem to affect mothers’ labour market choices by themselves, neither in a negative way, nor in a positive way, if social norms are not ready to cope with new behaviour and if the lack of childcare facilities do not make it possible to better conciliate work and family life. This means, that in very conservative settings where mothers would like to care for their children themselves, even the effect of fl exible policies is limited on the short run. However, a slight increase towards earlier labour market entry may be achieved if mothers are attached to the labour market, which is mainly the case among highly educated mothers. On the long run, social norms may positively infl uence labour market participation during early childhood. In a less conservative setting, such as in France, where mothers are attached to the labour market and childcare facilities are available, small changes – e.g. the introduction of a leave of six months – do not reduce labour market participation. However, the introduction of a leave of three years decreased mothers paid work, which means that it acted as a work-reducing policy. In France, this negative effect mainly concerned low-educated mothers and contributed to their prolonged absence from paid work. These results show the importance of weighing up the pros and the cons before every reform regarding the length of PPL, which has to be done in line with other contextual and structural settings in order to avoid negative consequences. • Zsuzsanna Makay452 References Abraham, Arpad; Kezdi, Gabor 2000: Long-run Trends in Earnings and Employment in Hungary, 1972-1996. Budapest Working Papers on the Labour Market BWP-2000/2. Budapest: Hungarian Academy of Sciences, Institute of Economics, Labour Research Department. Afsa, Cédric 1996: L’activité féminine à l’épreuve de l’allocation parentale d’éducation. In: Recherches et Prévisions 46: 1-8. https://doi.org/10.3406/caf.1996.1753 Akgunduz, Yusuf Emre; Plantenga, Janneke 2013: Labour Market Effects of Parental Leave in Europe. In: Cambridge Journal of Economics 37,4: 845-862. https://doi.org/10.1093/cje/bes052 Auer, Manfred; Welte, Heike 2009: Work-family Reconciliation Policies without Equal Opportunities? The case of Austria. In: Community, Work & Family 12,4: 389-407. https://doi.org/10.1080/13668800802556455 Berghammer, Caroline 2014: The Return of the Male Breadwinner Model? Educational Effects on Parents’ Work Arrangements in Austria, 1980-2009. In: Work, Employment and Society 28,4: 611-632. https://doi.org/10.1177/0950017013500115 Bernhardt, Eva M. 1993: Fertility and Employment. In: European Sociological Review 9,1: 25-42. https://doi.org/10.1093/oxfordjournals.esr.a036659 Bianchi, Suzanne M.; Casper, Lynne M.; Peltola, Pia K. 1999: A Cross-National Look at Married Women’s Earnings Dependency. In: Gender Issues 17,3: 3-33. https://doi.org/10.1007/s12147-999-0001-0 Bick, Alexander 2016: The Quantitative Role of Child Care for Female Labor Force Participation and Fertility. In: Journal of the European Economic Association 14,3: 639-668. https://doi.org/10.1111/jeea.12143 Blair-Loy, Mary 2003: Competing Devotions: Career and Family among Women Executives. Cambridge: Harvard Univ. Press. Blum, Sonja 2010: Between Instrument Tinkering and Policy Renewal: Reforms of Parental Leave in Germany and Austria. In: German Policy Studies/Politikfeldanalyse 6,3: 83-118. Brewer, Mike et al. 2022: Does More Free Childcare Help Parents Work More? In: Labour Economics 74, 102100. https://doi.org/10.1016/j.labeco.2021.102100 Buber-Ennser, Isabella 2015: Childrearing in Austria: Work and Family Roles. In: Journal of Research in Gender Studies 5,2: 121-146. Cleves, Mario Alberto; Gould, William; Marchenko; Yulia V. 2016: An Introduction to Survival Snalysis Using Stata. Revised third edition. Texas: Stata Press. Collombet, Catherine 2016: Histoire des congés parentaux en France. Une lente sortie du modèle de rémunération de la mère au foyer. In: Revue des politiques sociales et familiales 122,1: 111-122. https://doi.org/10.3406/caf.2016.3168 Del Boca, Daniela 2002: The Effect of Child Care and Part Time Opportunities on Participation and Fertility Decisions in Italy. In: Journal of Population Economics 15,3: 549-573. https://doi.org/10.1007/s001480100089 Esping-Andersen, Gøsta 1990: The Three Worlds of Welfare Capitalism. Oxford: Polity Press. https://doi.org/10.3406/caf.1996.1753 https://doi.org/10.1093/cje/bes052 https://doi.org/10.1080/13668800802556455 https://doi.org/10.1177/0950017013500115 https://doi.org/10.1093/oxfordjournals.esr.a036659 https://doi.org/10.1007/s12147-999-0001-0 https://doi.org/10.1111/jeea.12143 https://doi.org/10.1016/j.labeco.2021.102100 https://doi.org/10.3406/caf.2016.3168 https://doi.org/10.1007/s001480100089 Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 453 Fahlén, Susanne 2013: Capabilities and Childbearing Intentions in Europe. The association between work-family reconciliation policies, economic uncertainties and women’s fertility plans. In: European Societies 15,5: 639-662. https://doi.org/10.1080/14616696.2013.798018 Fenger, Menno 2007: Welfare Regimes in Central and Eastern Europe: Incorporating Post-Communist Countries in a Welfare Regime Typology. In: Contemporary Issues and Ideas in Social Sciences 3,2: 1-30. Gábos, András; Szívós, Péter 2002: A Jövedelmi Szegénység Alakulása, a Gyermekes Családok Helyzete. In: TÁRKI Társadalmi Riport 2002. Gauthier, Anne H. et al. 2018: Generations and Gender Survey Study Profi le. In: Longitudinal and Life Course Studies 9,4: 456-465. https://doi.org/10.14301/llcs.v9i4.500 Goldscheider, Frances K. 2000: Men, Children and the Future of the Family in the Third Millennium. In: Futures 32,6: 525-538. https://doi.org/10.1016/S0016-3287(00)00005-7 Hakim, Catherine 1999: Models of the Family, Women’s Role and Social Policy: A New Perspective from Preference Theory. In: European Societies 1,1: 33-58. https://doi.org/10.1080/14616696.1999.10749924 Hantrais, Linda 2004: Family policy matters: Responding to Family Change in Europe. Bristol University Press, Policy Press. https://doi.org/10.2307/j.ctt1t893wm Hasková, Hana; Gy\Hory, Adrienn; Szikra, Dorottya 2012: How Did We Get the ‘Magic 3’? The Timing of Parental Leaves and Child Care Services in the Visegrád-Countries. Unpublished Manuscript, Budapest Institute, Budapest. Ilonszki, Gabriella; Kurtan, Sandor 1996: Hungary. In: European Journal of Political Research 30,3-4: 359-360. Inglot, Tomasz 2008: Policy Legacies and Welfare States under Communism: Cycles of Social Policy Expansion and Retrenchment in Czechoslovakia, Poland, and Hungary, 1945-1989. In: Inglot, Tomasz (Ed.): Welfare States in East Central Europe, 1919-2004. Cambridge University Press: 119-210. https://doi.org/10.1017/CBO9780511510175.005 Joseph, Olivier et al. 2012: Faut-Il Tourner La PAJE ? : L’impact de La Prestation d’Accueil Du Jeune Enfant Sur Le Parcours Professionnel Des Mères. In: Formation Emploi 118: 103-123. https://doi.org/10.4000/formationemploi.3642 Kertesi, Gábor; Köllő, János 2002: Economic Transformation and the Revaluation of Human Capital – Hungary, 1986-1999. In: de Grip, Andries; van Loo, Jasper; Mayhew, Ken (Eds.): The Economics of Skills Obsolescence. Emerald Group Publishing Limited: 235-273. https://doi.org/10.1016/S0147-9121(02)21013-4 Központi Statisztikai Hivatal 1989: N k a Mai Magyar Társadalomban. Budapest. Központi Statisztikai Hivatal 2012: Kisgyermekek Napközbeni Ellátása [https://www.ksh. hu/docs/hun/xftp/idoszaki/pdf/kisgyermnapkozbeni.pdf, 14/08/2023]. Kremer, Monique 2007: How Welfare States Care: Culture, Gender and Parenting in Europe. Amsterdam. Amsterdam University Press. https://doi.org/10.5117/9789053569757 LaFont, Suzanne 2001: One Step Forward, Two Steps Back: Women in the Post- Communist States. In: Communist and Post-Communist Studies 34,2: 203-220. https://doi.org/10.1016/S0967-067X(01)00006-X Lalive, Rafael; Zweimüller, Josef 2009: How Does Parental Leave Affect Fertility and Return to Work? Evidence from Two Natural Experiments. In: Quarterly Journal of Economics 124,3: 1363-1402. https://doi.org/10.1162/qjec.2009.124.3.1363 https://doi.org/10.1080/14616696.2013.798018 https://doi.org/10.14301/llcs.v9i4.500 https://doi.org/10.1016/S0016-3287 https://doi.org/10.1080/14616696.1999.10749924 https://doi.org/10.2307/j.ctt1t893wm https://doi.org/10.1017/CBO9780511510175.005 https://doi.org/10.4000/formationemploi.3642 https://doi.org/10.1016/S0147-9121 https://www.ksh.hu/docs/hun/xftp/idoszaki/pdf/kisgyermnapkozbeni.pdf https://www.ksh.hu/docs/hun/xftp/idoszaki/pdf/kisgyermnapkozbeni.pdf https://doi.org/10.5117/9789053569757 https://doi.org/10.1016/S0967-067X https://doi.org/10.1162/qjec.2009.124.3.1363 • Zsuzsanna Makay454 Leibetseder, Bettina 2013: Parental Leave Benefi t in Austria. Stratifi ed Take-up in a Conservative Country. In: International Review of Sociology 23,3: 542-563. https://doi.org/10.1080/03906701.2013.856160 Leitner, Sigrid 2014: Varieties of Familialism: Developing Care Policies in Conservative Welfare States. In: Sandermann, Philipp (Ed.): The End of Welfare as We Know It? Continuity and Change in Western Welfare State Settings and Practices. Opladen/ Berlin/Toronto: Barbara Budrich: 37-51. Lequien, Laurent 2012: Durée d’une interruption de carrière à la suite d’une naissance: Impact sur les salaires. L’exemple de la réforme de l’APE. In: Politiques sociales et familiales 108,1: 59-72. https://doi.org/10.3406/caf.2012.2687 Makay, Zsuzsanna 2018: The Family Support System and Female Employment. In: Monostori, Judit; Őri, Péter; Spéder, Zsolt (Eds.): Demographic Portrait of Hungary 2018. Budapest: HDRI: 85-105 [https://demografi a.hu/en/publicationsonline/index. php/demographicportrait/article/view/953/725, 01.06.2023]. Mandel, Hadas; Semyonov, Moshe 2006: A Welfare State Paradox: State Interventions and Women’s Employment Opportunities in 22 Countries. In: American Journal of Sociology 111,6: 1910-1949. https://doi.org/10.1086/499912 Massey, Garth; Hahn, Karen; Sekulić, Duško 1995: Women, Men and the ‘Second Shift’ in Socalist Yugoslavia. In: Gender & Society 9,3: 359-379. https://doi.org/10.1177/089124395009003006 Matysiak, Anna; Mynarska, Monika 2021: Motives for Combining Motherhood with Employment: Evidence for Medium and Highly Educated Polish Women Around the EU Accession. In: Central European Economic Journal 8,55: 63-78. https://doi.org/10.2478/ceej-2021-0005 Mauerer, Gerlinde 2018: Both Parents Working: Challenges and Strains in Managing the Reconciliation of Career and Family Life in Dual-Career Families. Empirical Evidence from Austria. In: Social Sciences 7,12: 269. 12. https://doi.org/10.3390/socsci7120269. Misra, Joya; Budig, Michelle; Boeckmann, Irene 2011: Work-Family Policies and the Effects of Children on Women’s Employment Hours and Wages. In: Community, Work & Family 14,2: 139-157. https://doi.org/10.1080/13668803.2011.571396 Mun, Eunmi; Jung, Jiwook 2018: Policy Generosity, Employer Heterogeneity, and Women’s Employment Opportunities: The Welfare State Paradox Reexamined. In: American Sociological Review 83,3: 508-535. https://doi.org/10.1177/0003122418772857 Neuberger, Franz; Rüttenauer, Tobias; Bujard, Martin 2022: Where Does Public Childcare Boost Female Labor Force Participation? Exploring Geographical Heterogeneity across Germany 2007-2017. In: Demographic Research 46: 693-722. https://doi.org/10.4054/DemRes.2022.46.24 Neyer, Gerda 2010: Familienpolitik in Österreich – Zwischen Beharrung und Veränderung. In: Revue d’Allemagne et des pays de langue allemande 42,1: 57-70. Neyer, Gerda; Andersson, Gunnar 2008: Consequences of Family Policies on Childbearing Behavior: Effects or Artifacts? In: Population and Development Review 34,4: 699-724. https://doi.org/10.1111/j.1728-4457.2008.00246.x Odul-Asorey, Isabel 2013: Congé Maternité, Droit des Femmes? In: Revue Des Droits de l’homme 3. https://doi.org/10.4000/revdh.204 Panova, Ralina; Buber-Ennser, Isabella 2016: Attitudes towards Parental Employment: A Ranking across Europe, Australia, and Japan. In: Journal of Research in Gender Studies 6,2:11-37. https://doi.org/10.1080/03906701.2013.856160 https://doi.org/10.3406/caf.2012.2687 https://demografia.hu/en/publicationsonline/index.php/demographicportrait/article/view/953/725 https://demografia.hu/en/publicationsonline/index.php/demographicportrait/article/view/953/725 https://doi.org/10.1086/499912 https://doi.org/10.1177/089124395009003006 https://doi.org/10.2478/ceej-2021-0005 https://doi.org/10.3390/socsci7120269 https://doi.org/10.1080/13668803.2011.571396 https://doi.org/10.1177/0003122418772857 https://doi.org/10.4054/DemRes.2022.46.24 https://doi.org/10.1111/j.1728-4457.2008.00246.x https://doi.org/10.4000/revdh.204 Paid Parental Leave Reforms and Mothers’ Employment in Austria, France and Hungary • 455 Piketty, Thomas 2005: Impact de l’allocation Parentale d’éducation Sur l’activité Féminine et La Fécondité En France. In: Histoires de Familles, Histoires Familiales. Les Résultats de l’enquête Famille de 1999. Ined Éditions: 79-96. Prskawetz, Alexia et al. 2008: Austria: Persistent Low Fertility since the Mid-1980s. In: Demographic Research 19: 293-360. https://doi.org/10.4054/DemRes.2008.19.12 Riederer, Bernhard; Berghammer, Caroline 2020: The Part-Time Revolution: Changes in the Parenthood Effect on Women’s Employment in Austria across the Birth Cohorts from 1940 to 1979. In: European Sociological Review 36,2: 284-302. https://doi.org/10.1093/esr/jcz058 Rille-Pfeiffer, Christiane 2011: Austria Country Note. In: Moss, Peter (Ed.): International Review of Leave Policies and Research 2011. Institute of Education University of London: 52-58. Ruhm, Christopher J. 1998: The Economic Consequences of Parental Leave Mandates: Lessons from Europe. In: The Quarterly Journal of Economics 113,1: 285-317. https://doi.org/10.1162/003355398555586 Ruhm, Christopher J. 2000: Parental Leave and Child Health. In: Journal of Health Economics 19,6: 931-960. https://doi.org/10.1016/S0167-6296(00)00047-3 Ruhm, Christopher J. 2004: Parental Employment and Child Cognitive Development. In: Journal of Human Resources XXXIX,1: 155-192. https://doi.org/10.3368/jhr.XXXIX.1.155 Salles, Anne; Rossier, Clémentine; Brachet, Sara 2010: Understanding the Long Term Effects of Family Policies on Fertility: The Diffusion of Different Family Models in France and Germany. In: Demographic Research 22: 1057-1096. https://doi.org/10.4054/DemRes.2010.22.34 Thévenon, Olivier 2011: Family Policies in OECD Countries: A Comparative Analysis. In: Population and Development Review 37,1: 57-87. https://doi.org/10.1111/j.1728-4457.2011.00390.x Toulemon, Laurent; Pailhé, Ariane; Rossier, Clémentine 2008: France: High and Stable Fertility. In: Demographic Research 19: 503-556. https://doi.org/10.4054/DemRes.2008.19.16 Valarino, Isabel et al. 2018: Exploring Leave Policy Preferences: A Comparison of Austria, Sweden, Switzerland, and the United States. In: Social Politics: International Studies in Gender, State & Society 25,1: 118-147. https://doi.org/10.1093/sp/jxx020 Van der Lippe, Tanja; Van Dijk, Liset 2002: Comparative Research on Women’s Employment. In: Annual Review of Sociology 28: 221-241. https://doi.org/10.1146/annurev.soc.28.110601.140833 Yamaguchi, Shintaro; Asai, Yukiko; Kambayashi, Ryo 2018: Effects of Subsidized Childcare on Mothers’ Labor Supply under a Rationing Mechanism. In: Labour Economics 55: 1-17. https://doi.org/10.1016/j.labeco.2018.09.002 Zhang, Chi; Managi, Shunsuke 2021: Childcare Availability and Maternal Employment: New Evidence from Japan. In: Economic Analysis and Policy 69: 83-105. https://doi.org/10.1016/j.eap.2020.11.001 Zoch, Gundula; Hondralis, Irina 2017: The Expansion of Low-Cost, State-Subsidized Childcare Availability and Mothers’ Return-to-Work Behaviour in East and West Germany. In: European Sociological Review 33,5: 693-707. https://doi.org/10.1093/esr/jcx068 https://doi.org/10.4054/DemRes.2008.19.12 https://doi.org/10.1093/esr/jcz058 https://doi.org/10.1162/003355398555586 https://doi.org/10.1016/S0167-6296 https://doi.org/10.3368/jhr.XXXIX.1.155 https://doi.org/10.4054/DemRes.2010.22.34 https://doi.org/10.1111/j.1728-4457.2011.00390.x https://doi.org/10.4054/DemRes.2008.19.16 https://doi.org/10.1093/sp/jxx020 https://doi.org/10.1146/annurev.soc.28.110601.140833 https://doi.org/10.1016/j.labeco.2018.09.002 https://doi.org/10.1016/j.eap.2020.11.001 https://doi.org/10.1093/esr/jcx068 • Zsuzsanna Makay456 Date of submission: 02.08.2022 Date of acceptance: 15.05.2023 Dr. Zsuzsanna Makay (). Hungarian Demographic Research Institute. Budapest, Hungary. E-mail: makay@demografi a.hu URL: https://www.demografi a.hu/en/staff-makay-zsuzsanna https://www.demografia.hu/en/staff-makay-zsuzsanna Published by Federal Institute for Population Research (BiB) 65180 Wiesbaden / Germany Managing Publisher Dr. Nikola Sander 2023 Editor Prof. Frans Willekens Managing Editor Dr. Katrin Schiefer Editorial Assistant Beatriz Feiler-Fuchs Wiebke Hamann Layout Beatriz Feiler-Fuchs E-mail: cpos@bib.bund.de Scientifi c Advisory Board Kieron Barclay (Stockholm) Karsten Hank (Cologne) Ridhi Kashyap (Oxford) Natalie Nitsche (Rostock) Alyson van Raalte (Rostock) Pia S. Schober (Tübingen) Rainer Wehrhahn (Kiel) Comparative Population Studies www.comparativepopulationstudies.de ISSN: 1869-8980 (Print) – 1869-8999 (Internet) Board of Reviewers Bruno Arpino (Barcelona) Laura Bernardi (Lausanne) Gabriele Doblhammer (Rostock) Anette Eva Fasang (Berlin) Michael Feldhaus (Oldenburg) Alexia Fürnkranz-Prskawetz (Vienna) Birgit Glorius (Chemnitz) Fanny Janssen (Groningen) Frank Kalter (Mannheim) Stefanie Kley (Hamburg) Bernhard Köppen (Koblenz) Anne-Kristin Kuhnt (Rostock) Hill Kulu (St Andrews) Nadja Milewski (Wiesbaden) Roland Rau (Rostock) Thorsten Schneider (Leipzig) Tomas Sobotka (Vienna) Jeroen J. A. Spijker (Barcelona) Heike Trappe (Rostock) Helga de Valk (The Hague) Sergi Vidal (Barcelona) Michael Wagner (Cologne) http://www.comparativepopulationstudies.de mailto:cpos@bib.bund.de