Changes in the division of labor within highly educated German couples when the first child is born

When becoming parents for the first time, German couples often adapt their division of paid and unpaid work, creating a more gender-specific allocation. Using longitudinal data from the qualitative event-centered project “Household division of domestic labor as a process”, we compare theoretically-postulated mechanisms of change in the division of work within couples with explanations given by the couples interviewed themselves. Our qualitative analysis demonstrates that economic and gender norm theories are quite successful at predicting changes towards a more traditional specialization when couples become parents for the first time, while they are less helpful in explaining the persistence of equal arrangements in the domestic division of work, or the change towards more equal arrangements. The interviews also show that the explanations which differentiate – within unpaid work – between childcare and housework are a better predictor of the realities of the arrangements. Furthermore, the causal order of the decisions suggested by the theories differs from the couples’ actual decision making processes: when facing the transition to parenthood, they decide first upon the division of childcare, and then of labor market activity and housework. Zusammenfassung: Wenn Paare in Deutschland Eltern werden, verändern sie häufig ihre Aufteilung von bezahlter und unbezahlter Arbeit in Richtung eines geschlechtsspezifischen Arrangements. Auf Basis von qualitativen, ereigniszentrierten Längsschnittdaten des Projektes “Innerfamiliale Arbeitsteilung als Prozess” vergleicht der vorliegende Artikel theoretische Annahmen zu Veränderungen der Arbeitsteilung in Paarbeziehungen mit den Erklärungen, die hochgebildete Paare selbst geben. Unsere qualitative Analyse zeigt, dass sowohl ökonomische als auch Gender-Theorien relativ erfolgreich erklären, warum Paare sich für eine eher traditionelle Spezialisierung der Arbeiten entscheiden, wenn sie den Übergang zur Erstelternschaft erleben. Diese Theorien sind hingegen weniger erfolgreich darin, zu erklären, warum Paare egalitäre Arrangements beibehalten oder wählen. Anhand der qualitativen Interviews wird ersichtlich, dass Erklärungsmuster, die im Bereich der unbezahlten Arbeit zwischen Kinderbetreuung und Hausarbeit unterscheiden, besser die Realität der Paare erfassen. Darüber hinaus ist die von den Theorien vorgeschlagene kausale Reihenfolge der Entscheidungen eine andere als die der Paare: im Übergang zur Elternschaft entscheiden werdende Eltern zunächst über die Kinderbetreuung und dann über Erwerbstätigkeit und Hausarbeit. 1 Note: The authors would like to thank the German Research Foundation (Deutsche Forschungsgemeinschaft, DFG) for funding the data collection and analysis. We would also like to thank the anonymous reviewers for their input and William Tayler for proof-reading. A. Dechant & H.-P. Blossfeld: Changes in the division of labor within highly educated German couples 374

mogamous couples in Germany has risen (Blossfeld et al. 2001: 61), and changes in these couples' division of labor are most likely, it is particularly important to explore how highly educated couples respond to the birth of a child in their organization of work within the household.
The aim of this paper, therefore, is to analyze the division of labor within highly qualified West German dual earner couples during their transition to parenthood. We are particularly interested in assessing and explaining if, how, and why highly qualified couples change their division of labor with the first birth, how they perceive this change, and how they negotiate their living arrangements in this life course transition. As the couples' own interpretation is at the center of our attention, a longitudinal qualitative approach was selected which allows for the identification of explanations (Flick et al. 2008) about the reasons for decisions about the division of labor between men and women before and after the birth of a child, which can then be related to potential theoretical explanations.
The paper is structured as follows: We shall first discuss relevant theories regarding changes in the division of work within couples which occur due to the birth of the first child. Secondly, we give a description of our qualitative longitudinal study and our applied methods of analysis. Thirdly, we shall compare the couples' situations before and after the birth of their first child in terms of paid and unpaid work. We will, in particular, contrast couples' plans for their division of labor after the child's birth with the actual arrangements at the first and second interview. The theories discussed in the first section guide the analysis and presentation of the interviewees' argumentations and rationales. Finally, we will summarize our results and draw some more general conclusions regarding the existing theoretical approaches.

Theoretical frameworks
When couples become parents, their division of labor may (a) persist, (b) become more specialized, or (c) move towards a more egalitarian arrangement. Specialization means that one of the partners concentrates on paid work, while the other focuses on domestic work. Change towards a more egalitarian pattern means that domestic labor and paid work are shared more equally among the partners. Different theoretical approaches aim to explain couples' division of labor; we will focus on economic (Becker 1998;Ott 1992;Coverman 1985) and gender approaches (Bielby/Bielby 1989;van Berkel/de Graaf 1999;Brines 1994), and will deduce expectations regarding the division of labor of couples who are transitioning to parenthood.

Economic approaches
Economic theories assume that the division of labor within couples is based on economically rational decisions. According to the new home economics, partners are expected to maximize their joint household utility function by specializing in different work spheres (Becker 1998). The specialization is based on relative human capital investments which result in different productivities and incomes. A complete specialization is the most effi-cient solution for most couples (Becker 1998). Therefore, the highly educated couples interviewed here should have a specialized division of household labor before the birth of the first child, based on income differences, as specialization is more efficient and resulting returns are shared between the spouses. Since only couples consisting of two employed or self-employed partners were interviewed, specialization regarding household labor is expected. When both partners have the same income potential, it is theoretically unclear which partner should specialize in housework. Since boys and girls are socialized differently, women tend to have jobs more easily reconcilable with household obligations (Becker 1998), and should therefore be more likely to specialize in housework. Thus, we expect the women interviewed to take responsibility for domestic labor if they have an income lower than or equal to that of theur partner. The amount of domestic work increases with the birth of a child, and this reinforces specialization (Gjerdingen/Center 2005). Due to the different socialization processes, the women interviewed should be even more likely to specialize in housework and childcare after the birth of their first child.
In contrast to new home economics, Ott's bargaining approach assumes that each partner maximizes his/her own utility function, as the altruistic division of specialization rewards is unrealistic (Ott 1992). Since the division of paid and unpaid work influences their utility functions, the partners bargain for their division of labor. They do this on the basis of relative resources, as bargaining power depends on human capital accumulation. Housework is considered less attractive than paid work due to its lower transferability to a new partnership and its lower prestige; by the same token, paid work is positively associated with bargaining power, as it can easily be transferred to another relationship. Regarding the division of labor within the couples interviewed, it can be assumed that if one partner has fewer resources, and thus less bargaining power, he or she does more housework. With the birth of the first child, a reinforcement of this specialization is expected, as the additional task of childcare increases the amount of unpaid work. In couples with equal resources, both partners are expected to share the housework equally -an arrangement that should persist after the birth of their first child. Nevertheless, obligatory maternity leave might influence the bargaining power of employed women, even though income losses are compensated. The time spent out of the labor force might decrease women's market-specific human capital and thus affect their bargaining power and the division of labor.
Unlike the previous economic theories, the time availability approach takes into account the fact that a person spending a lot of time in the labor market has less time available for -and should consequently perform less -housework (England/Farkas 1986). Coverman (1985) specifies this assumption by adding three ideas: first, housework and childcare are socially primarily assigned to women; second, if women devote time to employment, the demands on men to participate in childcare and housework will increase; and third, the time men spend performing domestic work is also dependent upon their capability to react to these demands. Women's employment situation and children living in the household affect the demand, while men's working hours influence their capability. According to the demand/response capability approach, we therefore expect that, among the couples interviewed, women will generally assume more responsibility for housework than men. Men with partners in full-time employment should take care of a greater part of the domestic work than men with partners in part-time employment. The transition to parenthood increases the amount of domestic labor, and thus the demand for men to engage in housework, especially if their partners are active in the labor market. If women reduce their working hours -as they are obligated to due to maternity leave -the demand for men's participation in housework decreases. Among couples in which the women return to the labor market after maternity leave, the men are expected to take over more domestic work upon their partner's reentry.

Gender approaches
In contrast to the economic perspective, sociological, norm-based theories explain the gendered division of work by gender role expectations. In everyday life, women and men display their own gender identity, for instance by doing (women) or not doing (men) housework (West/Zimmerman 1987). Bielby and Bielby's (1989) identity formation approach furthermore assumes that everybody has both a work and a family identity; time spent in the labor market or in the family strengthens the corresponding identity. Gender roles affect how men and women form their identities differently: having children has a greater impact on a woman's family identity than a man's, as her family identity is built upon caring, whereas his is consistent with being a successful provider. We thus predict for the couples interviewed, that -regardless of their resource constellation -women will do more household chores than men after the birth of the first child. Following the child's birth, men who are successful providers build not only a strong work identity but also a strong family identity, as good providers are also good fathers. Women, conversely, perceive a conflict between work and family identities; if they pursue their caregiver role they only invest in their family identity, and not in their work identity, and vice versa. Consequently, the birth of the first child will reinforce traditional gender roles. Women's identity trade-off may result in less involvement in the labor market in order to take care of the baby and household chores. Brines's (1994) dependency approach combines core ideas from economic theories and doing gender assumptions: if one spouse has to rely on the other's income, she or he exchanges unpaid household labor against income. As actors display gender in everyday actions like performing paid or unpaid work, Brines (1994) anticipates a u-shaped relationship between a woman's share of the household income and her proportion of housework within couples. With an increasing income, a woman can reduce her share of the housework until she earns as much as her partner. Once her income exceeds his, she will take over more chores in order to fulfill gender roles in this area, as the couple act in opposition to them in economic terms. For the couples in our study, we therefore expect that if a woman earns either less or more than her partner, she will do more housework to compensate either for economic dependency or for the violation of gender norms. Brines (1994) does not include childcare in her arguments, but as the gendered expectations with regard to childcare are even more pronounced than they are regarding chores, we assume this tendency will increase with the birth of a child. If both partners have equal incomes, they should tend to share the housework fairly equally after becoming parents.
The egalitarian values approach (van Berkel/de Graaf 1999) draws on the cultural dimension of education as the central mechanism determining the division of work in the family. Education is linked to democratic values; with increasing levels of education, gender equality ideas take on greater significance. The egalitarian values approach hypothesizes that the division of housework is most equal when both partners have high levels of educational attainment, whereas it is specialized (with men taking on a smaller share) when both partners have a low educational level (van Berkel/de Graaf 1999). Partners with different levels of education will most likely practice a division of labor that lies in between the two extremes. As a certain gender-specific asymmetry is to be found in most societies, van Berkel and de Graaf (1999: 790ff.) suggest that the man's education has a greater impact. For the highly educated couples studied here, we expect the division of household labor to be equal both before and after the birth of the first child, as the transition to parenthood affects neither their educational levels nor the associated ideas of gender equality. For couples with different educational levels, we predict that the woman will do more domestic work than her partner, both before and after the transition to parenthood, since the division of labor should lie in between the extremes of equality and total specialization. Table 1 shows the anticipated outcomes of the division of labor in our sample, comparing the time before and after the birth of the first child, for each of the theoretical approaches.

Data and method of analysis
To address the question as to whether and why highly qualified parents practice a division of labor that is consistent with the middle class family model after the birth of their first child, we use data from the DFG-funded qualitative study "Household division of domestic labor as a process". This qualitative longitudinal event-centered survey, conducted in 2006 and 2007, aimed to explore the reasons for changes in the division of labor when highly educated couples become first-time parents. The issues and questions for the semistructured qualitative interviews were partially derived from the theories discussed above, in order to analyze their capability to capture the on-going processes. Before and after the birth of their first child, the spouses were interviewed separately about their current, past, and future division of paid work, housework, and childcare (for more information on the design see Schulz et al. 2008).
The composition of the sample was affected by a number of theoretical assumptions. The couples had to be expecting their first child at the time of the first interview, they had to share one household, both partners had to be active in the labor market, and both partners should to have similar educational levels. They agreed to one interview during the pregnancy and another after the child's birth. Initially, the recruitment of the couples was focused on Northern Bavaria. Due to a slightly insufficient number of respondents from this region, we also had to include a few couples from South West Germany. In order to recruit parents-to-be, the researchers went to places where they might be found: prenatal classes, a bazaar for baby clothes, and an information evening for future parents. Potential interviewees were personally informed about the study, and leaflets with information and contact details were distributed. Furthermore, midwives were contacted and asked to circulate the information leaflets. Potential interviewees received information about the selection criteria, and the study's focus on parents' plans for their future as a family, and their expectations from parenthood. We were able to conduct 56 interviews with 14 couples that mostly fit our sample criteria. At the time of the first interview, the women were between the fourth and ninth month of pregnancy. The second interview took place about six to twelve months after the birth, i.e. after the end of the mandatory maternity leave and within the period of partly-remunerated parental leave. The longitudinal, eventcentered design allows for a comparison of the couples' planned and realized arrangements, as well as for changes in their arrangements before and after the birth of their first child. The analysis in this paper focuses on highly educated, educationally homogamous couples with similar resources: In 10 out of 14 couples, both partners have similar levels of educational attainment (Table 2) levels which, due to sampling, are above the German average (Authoring Group Educational Reporting 2013). In most couples interviewed, the men have higher incomes than their partners (Table 3).  Statistical representativeness was not the goal of our approach, since our qualitative study was the result of an earlier quantitative longitudinal study on couples (Grunow et al. 2012). However, this quantitative data set left some interpretive puzzles. The aim of our inquiry was to obtain a data set of event-oriented interpretations, allowing us to analyze the motivations of a specific subpopulation. On this basis, explanatory generalizations are possible for this group (Mayring 2007). Two qualitative methods were used for the analysis: qualitative content analysis (Mayring 2008) and the confrontation of hypotheses from elaborated theories with qualitative data (Hopf 1993). We used the qualitative data analysis tool MAXQDA to organize, categorize, and analyze the transcribed interviews. The first step of the analysis was to code the transcripts in order to structure the data. The coding categories for the content analysis were primarily derived from theoretical expectations. Additionally, inductive categories were generated from the data (Mayring 2008). Based on the summarizing content analysis, each couple's prenatal and postnatal division of labor was analyzed with regard to changes, aiming to understand them. To make use of the longitudinal design, the data from before and after the transition to parenthood are compared for each interviewee and within each couple.
We used the theoretical assumptions described above to analyze the couples' arrangements and the interviewees' explanations for their actual and planned division of labor. These hypotheses focus on the individual level, therefore enabling comparisons with the individual cases (Hopf 1993). The birth of the first child changes the previous employment arrangements: All women take some time off work; some of them stayed at home for the two months of obligatory maternity leave, others are still in parental leave at the second interview. The couples C and I keep their full-time/full-time arrangement. Couple H convert the male full-time/ female part-time arrangement into a part-time/part-time arrangement. In other couples, the men work full-time while the women work part-time; they used to have full-time/full-time (A, F, K, M), male full-time/female part-time (B, D, J), or male part-time/female full-time (E) arrangements. Finally, the couples G, L, and N have a male full-time/female nonemployment arrangement. The part-time working women's working hours range from approximately five hours per month to 32 hours per week.

Housework.
Before the child's birth, there are couples with equal divisions of housework in the sample (A, C, F, G, K, N), meaning that both partners perform approximately the same amount of housework. They normally do not share every single task equally; each partner in most couples is responsible for certain tasks, and only some tasks are undertaken by both partners. If housework is not shared equally, the women assume a greater share -irrespective of the working arrangement.
With the birth of the child, the division of housework becomes more specialized, with an increase in the woman's share: At the time of the second interview, only the couples C and H share housework equally. One of them retain their equal division of housework; the other one have moved from their partly specialized division, with a greater female share, into an equal division. The other couples interviewed practice a (partly) specialized division, with the woman doing more housework. Most of the couples had a short period after the child's birth during which the men took over almost all of the housework as the women were recovering from the delivery.
Childcare. The couples C and H share childcare equally at the time of the second interview. The other couples interviewed have a partial or complete specialization of childcare in which the women have a greater share. Some men assume responsibility for more childcare in the very first days after childbirth.
Overall division of labor. Taking paid and unpaid work into consideration, six couples have an equal division of labor before the birth of the first child (A, C, F, G, K, N); the others practice a partly specialized arrangement: the women do more housework, while either the men have more working hours (B, D, H, J), both partners have similar working hours (I, L, M), or the man has fewer working hours (E).
About one year after the birth of their child, most of the couples have changed their arrangements. Two couples (C, H) divide paid work, housework, and childcare equally, while the other couples have specialized arrangements: the men work full-time and the women do more housework and childcare, whilst still working full-time (B, I, J), parttime (E, F, K), in marginal employment (A, D, M), or are not employed (G, L, N). The changes from the first to the second interview are documented in Table 4; it seems as if the theories expecting specializations to result from the transition to parenthood describe the actual changes as they were reported, even if the motivations stated by the interviewees did not necessarily mirror theoretical expectation

Explanation of the developments
Do the interviewees make use of the theoretical hypotheses when explaining the actual changes we observed? The following section shows the interviewees' rationales for their plans, and contrasts them within each couple, within the group of women, within the group of men, using the longitudinal design with the couples' plans at the time of the interview. Since the theories discussed above were guiding principles of the data analysis, the couple's explanations are ordered accordingly: first, we analyze arguments that are economic in nature, and then we focus on gender arguments.
Economic and resource arguments. The interviewees apply financial reasoning to two decisions: when they explain why and which partner is intending to stay at home for an extended period after the birth of their child, and when they explain why and when the partner who will leave the labor market will return to work. Furthermore, they speak about available time as a resource when explaining the planned division of household labor.
Interviewees who decided on a specialized arrangement, in which one partner does not adapt his/her labor market activity, while the other partner interrupts his/her employment or self-employment after childbirth, explain this decision with income differences which favor the partner who intends to continue his or her employment. This explanation fits expectations arising from the new home economics (Becker 1998). The planned specializations are (with the exception of couple E) characterized by the fact that the man continues to work full-time and the woman stays at home to care for the child. These plans reflect the fact that the interviewees consider it normal for at least one parent-to-be to care for the child and change his or her labor market situation accordingly. The following quote taken from the interview during pregnancy illustrates how interviewees use economic reasoning to explain their plans. Woman D, who has a much lower income from part-time self-employment than her full-time working partner, says: Well, parental leave [for her partner] would not work for us, for financial reasons, because I don't earn as much as [husband's name], and … we wouldn't be able to manage financially with just my earnings. (D, woman, 1 st interview: 360) guments of the interviewees explaining their planned specialization of the divisional arrangement widely reflect their economic situation, even if the income differences are only small in some couples. There is one exception to this economic rationality: woman K argues with economic impossibilities when explaining why her partner could not take parental leave, even though her monthly earnings are several hundred euros more than her partner's. Her economic argument is conditional on the fact that she will take parental leave.
Beside the couples planning a specialized arrangement after the child's birth, man H, who was actually planning an equalization of their division of labor, also uses financial arguments -in a different way. The man earns almost twice as much as his partner; according to the arguments of the other interviewees, and in line with economic theories, the man should continue to work full-time while his partner interrupts her work, at least for maternity leave. Contrary to this assumption, the couple plan for the man to take parental leave, and reduce his working hours by 60%, while the woman is to return to her unaltered part-time job after maternity leave. This change allows both partners to care for the child and participate in housework equally. The man explains that the couple can afford for him to reduce his working hours, and thus that their financial situation does not limit their possible options for the future arrangement, but gives them the freedom to decide.
The income differential is mostly used by couples envisioning a specialization after childbirth. For the most part, both partners use this argument, which normally reflects the couple's actual economic situation. There are some couples for whom only the women make use of the actual income difference to explain their planned specialization. Women who plan to reduce or interrupt their employment without explaining this plan economically have in common that they and their partners have very similar incomes and that their partners also do not use economic arguments. Furthermore, women and men who are not planning a lengthy absence from the labor market (for either partner) do not use financial differences as explanations. There is no clear pattern of using or not using economic arguments to explain which partner reduces or interrupts paid work regarding the interviewees' educational levels.
Regardless of referring or not referring to economic reasons, most couples report the successful realization of their planned arrangements at the second interview. Some couples who used the partners' income differential to explain their plans describe minor changes compared to their expectations. Couple A, for example, planned for the woman to stay at home with their child for at least one year while the man continued to work fulltime; the woman considered reentering the labor market in marginal employment after six months, at the earliest, and thought that it would be difficult to negotiate her desire for reduced working hours with her employer. At the second interview, however, the couple report that the woman had returned to marginal employment after three months, earlier than expected. The couple decided to expand her parental leave to three years, during which time she plans to stay in marginal employment. When asked if she could imagine working part-time instead of marginal hours, she explains that she does not want to be responsible for the income: I don't want that responsibility, if I were to go back to work, well, part-time -earning more money, then we'd just end up planning our finances around that. No, I don't think I'd want to go that far. (A, woman, 2 nd interview: 309) This quote shows that woman A thinks that her partner should be the one who is responsible for the basic income. This is a new development when compared to the first interview, where her employment was quite important to her; however, economic rationalities alone do not fully explain this change, since the couple would certainly benefit from her income, having decided to become homeowners.
Couple E -who planned to have the man stay at home after the birth of the child, while the woman was to return to her employed and self-employed work, due to the huge income difference between the partners -were not able to realize their plan for any length of time. At the time of the second interview, the man is working full-time, while the woman is on parental leave from her employer, working self-employed part-time, and handling most of the childcare and housework. The couple did adhere to the planned arrangement directly after the birth; the man searched for a new job, and when he was offered a full-time job with approximately his partner's remuneration, he accepted it and the couple subsequently changed their arrangements. During pregnancy, the man explained that he had tried to find a part-time job, since this best fit to the couple's ideal of sharing paid and unpaid work equally. Both partners say that the man could not find part-time work and, as they were afraid of the negative consequences of a longer interruption, they went for the full-time offer. Thus, the couple decided in favor of his labor market human capital, and against the logic of the income difference and a further specialization of their then-living arrangement.
The second aspect the interviewees looked to explain economically is the timing of the reentry to the labor market of the person who left it to care for the child in the first place. The specific rationale is that the couple cannot afford to have one partner not contributing to the household income for longer than a certain period, since two incomes are necessary to maintain their standard of living or pay off any debts; this explanation fits the expectations deriving from economic theories. Woman F, for example, says that the couple could not maintain their standard of living without her returning to her job after two to three months of parental leave. She explains that she did not mind reducing their standard of living, but that her partner would not accept any changes, and that she was therefore planning to return earlier than she would have wished. Her partner says that, for him, not only is the standard of living important, but that he is also afraid of negative long-term consequences for his partner's future self-employment status if she were to take more parental leave. Additionally, he argues that the couple has to consider the economic risk that arises from the fact that both partners are self-employed. A short period of parental leave for his partner is a good way to reduce this risk for him. The statements of both partners reveal a process of bargaining over future arrangements, in which the woman's return to paid work is not fully her own decision, but also a result of her partner's arguments.
Some women want to earn their own income, and do not even entertain the possibility of not returning to employment. Therefore, they and their partners plan the woman's return to paid work within two to twelve months after childbirth; most also plan for her not to reduce her working hours substantially upon return. This explanation of the planned timing of reentry fits to the bargaining approach (Ott 1992). Man B explains the couple's plans for the woman to take up her self-employed work some weeks after childbirth with the significance she assigns her job: She told me: "I want to keep working. I don't want to be a typical housewife, demoted to stove, kitchen, and children." (B, man, 1st interview: 21) As exemplified in this quote, the impetus for the planned arrangement comes from the woman in all these couples. The men, who never considered reducing or interrupting their own employment, do not have any concrete ideas concerning their partner's future employment. One exception is man L. He wanted his partner to stay at home for at least three years, whereas she wanted to return to her job immediately after maternity leave. The couple then agreed for her to take one year off as a compromise between the different plans.
Women who plan on taking some time off paid work speak more about the planned timing of reentry than their partners do, and they use more often economic arguments for the timing than the men interviewed. Some couples have very different ideas concerning the timing of the reentry, and bargain over the woman's future career; this bargaining is not always in line with the theoretically expected idea that paid work is preferable to unpaid work. The women who explain the timing of their reentry to the labor market with recourse to financial arguments are all highly educated; none of the women interviewed with an intermediate or low educational level argues this way. Three of the four women who earn as much as their partners use financial aspects to explain the timing of their planned return to work. The men who explain the timing of their partner's reentry with economic arguments say that two incomes are necessary to maintain their standard of living or service their debts. If they refer to the importance of their partner's career as a personal source of income or as part of their identity, they always do so with reference to their partner's own thoughts on the matter, and to conversations with their partner about the future -they do not say that their partners' labor market activity is important for their own appreciation of their partners.
At the time of the second interview, all the women who applied economic reasoning to their plans to return to the workplace have returned to their jobs; most of them did this in the way they planned to. Some report minor changes in the timing or in the number of working hours; one example of these changes would be the offer of part-time work for woman C, who said at the time of the first interview that there were no options for parttime work at her company in her position. Since her employer created the possibility to work part-time, the woman took the chance to return -after two months' maternity leave -to her job part-time, increasing to full time after six months. When the interviewees describe the developments which have transpired since the first interview, they often use the same economic rationales they did in the first interview. The couple who planned an equalization of their arrangement did not utilize economic arguments like the other couples, but mention an unexpected form of bargaining: The man's willingness to take leave was a necessary precondition for the woman's decision to have a child at all.
Most interviewees explain the planned division of household labor with reference to the available time both partners will have after childbirth, resulting from the planned division of paid work. Interviewees whose plans include specialization often say that the partner who takes parental leave has more available time at home, and is therefore expected to assume responsibility for most household labor. This explanation is found more often in the interviews with the partner who plans to take time off work than in the interviews with the partner who plans to continue paid employment. Man E -who plans to take parental leave -expects to be responsible for more housework than he was at the time of the first interview; in contrast to the women interviewed who plan to stay at home after childbirth, he still expects his partner to handle a substantial part of the domestic labor, since his primary concern is childcare. He plans to concentrate on housework when the child is asleep, for example, whereas the women who are planning a specialized division think that they can easily combine childcare and household labor. They argue that they are at home, have the time, and want their partners to be able to spend time with the child after paid work or at the weekends, rather than doing housework. These women think that it is more important that their partners spend time with the child than that they do household labor. Most of the men whose partners are planning to take parental leave do not expect any change in the division of housework. These men often do not spend much time on housework either, as their partners do more, or because they outsource parts of the routine domestic work. Like their partners, the men expect to spend time with the child when they have free time. Conversely, a few men anticipate that their partners will do more housework when they are taking parental leave. Man A explains that the couple spoke about this aspect of the future division of labor, and that he and his partner agreed that she has to do most of the housework when she decides to take parental leave. This explanation reveals another form of unexpected bargaining: the couples bargain over who is allowed to stay at home. This is of particular importance to some women, who actively want to stay at home with their child, and rate this opportunity much higher than their employment. In order to stay at home, these women accept doing more housework, even if they had not done so before the birth of their child.
Couple H, who expect to equalize their division of labor, say that the man will do more domestic labor, since he will have the time to do so while caring for the child. This is exemplary of the way availability of time is an argument for all couples, irrespective of their planned division of labor. This rationale is also not linked to the couples' educational levels.
At the time of the second interview, most interviewees who referred to available time as an explanation for their future housework arrangements share the domestic labor in the way they had planned. An exception is couple E, who differed from their plan for paid work; due to the unforeseen work situation, it is not the man but the woman who handles most of the domestic labor; both partners explain this responsibility for unpaid work with reference to the different availabilities of time resulting from the unplanned work arrangement.
Gendered explanations. The decision as to which partner takes parental leave is not only influenced by the economic factors discussed above. Gendered aspects are also important for the interviewees; they often imply gendered identities when explaining which partner should be primarily responsible for childcare, which is often also related to the division of paid work. However, the expectation that the person who stays at home with the child should do most of the housework seems not to be gendered.
The interviewees refer to gendered identities that are partially in line with the identity formation model in their discussions of their plans after childbirth. Many women speak about a desire to care for the baby, about maternal feelings, and about the intention to spend time with the child; therefore, they plan to take parental leave, interrupt their employment, and stay at home for some time. Some women's wish to spend time with their child is so important to them that they hinder their partners from taking parental leave. They argue with their ability to breast-feed, which they consider as very important for the child's wellbeing, and it is non-negotiable for them that they intend to do what is best for their child. This argumentation is exemplified in the following quote from woman A, who was asked if the couple had spoken about her partner taking parental leave -something she does not want: [A]s far as I'm concerned this wasn't an issue, since I like being at home -although he would have liked to stay at home too, and I think he wouldn't, and he said he wouldn't have any problem with being a stay-at-home dad -but then of course, with the breastfeeding alone, there's really no other way. (A, woman, 1 st interview: 531) Her partner says that he could have envisaged taking parental leave; however, since staying at home with the child is so important for his partner, he did not argue with her. He explains that she has specific ideas about mothering, which include her staying at home and doing most of the childcare and housework. The idea that the mother should be the child's primary care-giver -for some time, at least -is more often expressed by women than by men. Women expressing strongly gendered explanations mostly earn less than their partners, and more often have an intermediate or low educational background; however, two women who earn approximately as much as their partners and are highly educated also express strongly gendered explanations. They all have in common that they are either relatively young, compared to the other interviewees, or had problems conceiving. The men who think that their partners should care for the child as part of their stereotypical maternal role are not highly educated and have a higher income than their partners.
The other women interviewed seem to have no need to focus solely on their family identity. Women B, C, E, H, I, J, K, and L have a strong desire to reenter the labor market in the future, as their paid work is an important part of their identity. They describe their own careers as reconcilable with actively caring for their child. However, only women C and E plan to work full-time in the first year after their child's birth. The other women are planning, or considering, a reduction of their working hours, or already work part-time. The reduction of working hours is sometimes relatively small, as in couple I: the woman wants to reduce her working hours by approximately 4 hours a week. She does not explain this reduction, simply stating that it is important not to return to work with unreduced hours. The partners of the women who plan to return to work within the first year after childbirth support them in this decision, with the exception of man L: he thinks that it would be best for the child if the mother were to stay at home for at least three years to take care for the child. The women who have a strong identification with their careers and plan to return to paid work within two to twelve months are highly educated, as are all their partners -again with the exception of man L, who has an intermediate educational attainment.
The interviewees also associate biological sex with specific roles or identities for men: neither the men nor the women interviewed question the male partner's labor market activity. This is also true for both partners of couple E, who are planning for the man to stay at home after the child's birth. They think that being active in the labor market is important for an individual's self-worth, and describe the planned specialization as resulting from their income differential; their ideal arrangement would involve both partners having equal working hours. Often both partners in the couples planning for the man to re-main in continuous full-time employment express the idea that it is impossible for the man to take parental leave due to his job. Men B and L, for example, say that their jobs require certain technological knowledge, and that they cannot take parental leave since the technology changes so rapidly. Their female partners make the same arguments, contributing to the association between maleness and labor market activity after childbirth.
Besides being active in the labor market, most interviewees expect the men to be active fathers who spend time with their children, and are involved in everyday childcare. For most of the interviewees, however, this does not necessarily mean that they intend for the man to spend as much time with the child as the woman, since providing for the family is still an important part of his role. Couples E and H diverge from this pattern, saying that it is just as important for the man's identity as a parent to care for the child as it is for the woman's: […] actually, my husband has always said himself that, if we had a baby, he would want to look after it too […] and as for how we see our roles […] we've always had an attitude of equality -it's not like I think, for example, that mothers are better at raising children than fathers. (H, woman, 1 st interview: 426) Almost all of the interviewees who explain their plans with recourse to gendered identities, or with the rejection of gendered identities, describe in the second interview that they had been able to realize their ideas, albeit often with some minor changes. The couple who made major changes were couple E, from an arrangement in which the man stayed at home to care for the child and the household while the woman was working full-time to an arrangement in which the man is working full-time, and doing much less unpaid work, while the woman is working part-time during parental leave, and is responsible for most household labor and childcare. The man stated in the first interview that it is hard for him to imagine not being employed; he admits in the second interview that he enjoys going to work, even knowing that this results in a double burden for his partner. The woman confirms this double burden. However, in the first interview, she also thought that being the sole breadwinner might be a burden too; adding that she knew it would be difficult for her partner to be out of work. The change in the arrangement could not be explained solely with economic mechanisms, but can also be understood as resulting from the selfperceptions of the interviewees.
In the second interview, most interviewees describe their new identity as a mother or father as being very similar to what they had anticipated. Some underestimated the influence which the transition to parenthood would have on them and their ideas regarding the division of paid work, for example. One example is woman A, who is in marginal employment, and who does not envisage working more than that in the future; as described above, she explains her feelings by stating that the couple's financial plans would be altered if she were to work part-time, adding that she cannot imagine being responsible for their income. She also makes reference to her maternal feelings: Woman A: I don't want that responsibility, if I were to go back to work, well, part-time -earning more money, then we'd just end up planning our finances around that. No, I don't think I'd want to go that far.
Interviewer: OK -responsibility for what, exactly?
Woman A: For earning money; I'm already contributing as it is, and if my input were to become indispensable, well, I wouldn't want that -I'm too much of a mother for that. (A, woman, 2 nd interview: 309-313) From a strictly economic point of view, the woman's position is hard to understand, as the couple had decided to buy a house, and her extra income would help to pay for any liabilities; this discrepancy can be best explained with recourse to gendered identities.
Couples, but especially women, who neither planned to and nor actually did stay at home with their child for a longer period report that their social surroundings confronted them with the idea that mothers should stay at home with their child since it is not good for the child if the mother is active in the labor market. Woman E describes this when she speaks about the responses of others to the changes in their arrangements: Well, it was really difficult for me, after I returned to my job […] to apply for parental leave. And also, like […] you have to take a lot of crap, you know, going back to work as a mother and all, and you prepare yourself for that, and then you go and do it, and then […] Well, it feels a bit like they're saying "now you see why this isn't going to work".
[…] Like that: I felt like people were thinking, "I see, Ms.
[Name] thought she could just shoulder it all", or, 'They think they can just reverse their roles -well, it doesn't work like that". (E, woman, 2 nd interview: 1076-1085) This quote shows that the woman does not want to conform to the gendered identities mothers and fathers are confronted with, and that she struggles with them. Negative reactions are only experienced by women; none of the men are criticized as negatively affecting his child's well-being, regardless of how much time he planned to and actually does spend with his child.
The gendered identities we describe here were not important for the explanations of the current situation during pregnancy in most couples. In line with the egalitarian values approach (van Berkel/de Graaf 1999), the highly educated interviewees expressed the ideal that both partners should share paid as well as unpaid work, even if they did not in practice . Sharing paid and unpaid work equally was, for most interviewees, not part of their planned post-birth arrangements. Most interviewees did not question the idea that their division of labor would become specialized with the birth of the child, and that the woman would be the one to take on most of the housework and childcare, while the man would be responsible for most paid work. Some -like woman A, who refused at the time of first interview to take on more housework simply based on her gender -argue with gendered identities regarding the future division of work.
There are also several couples -C, E, and H -who also speak explicitly about equality concerning their post-birth arrangements. The couples C and H plan an equal division of paid and unpaid labor after at least six months, and justify this plan partly with the concept of equality. Couple C plan an equal division, based on the woman's strong work identity and the man's wish to be an active father. Woman C does not think that she, as a woman, should be the one to stay at home, and her partner agrees with that. The couple plans to achieve equality after six months of her taking parental leave, when both partners will work full-time and outsource large shares of household labor and childcare. This is different to the motivations of couple H: this couple also plan an equal arrangement, but the rationale behind it is that both partners want the man to be an active, caring father who spends more time with his child than other fathers do; at the same time, the woman wants to continue her career. Both partners say that they believe that men and women are equal-ly able to take care for children, and that they believe in gender equality. Both partners plan to have an equal arrangement after maternity leave, when the man will take parental leave and reduce his working hours to take more responsibility for childcare and housework.
In contrast to the couples planning equal arrangements, couple E speak about gender equality, but do not plan accordingly. Both partners say that their ideal arrangement would be that they both work part-time with equal hours, sharing childcare and housework equally. As the man explains, they had to decide against this ideal, and instead for a specialization: I have to add that this decision is, of course, a decision born out of need, out of necessity, financial necessity. Our ideal plan would have been for both of us to work part-time. (E, man, 1 st interview: 125) The idea of a gender-equal division of paid and unpaid work is discussed by the same interviewees in the second interview in a very different way. The interviewees in couples C and H enjoy their division of labor, which is as they planned it. The only aspect that couple H changed was that they outsourced household labor for a time. Woman H says in the second interview that the willingness of her partner to share paid and unpaid work equally was one important precondition for her decision to have children. Both partners in couple E -who did not plan to put their ideal of equality into practice, and whose arrangements did not conform to their plans -refer to their ideal of equality in the second interview, but explain that it was and is impossible, due to economic and occupational restrictions.
In the couples' plans for the division of housework, gender did not play a role. As discussed earlier, available time was an important factor. However, there was a difference between the women and the men planning to stay at home after childbirth: the men thought that the amount of housework they would be able to do would be dependent on the child's needs, while the women thought that they could easily take responsibility for the majority of both household labor and childcare. This is not in line with the economic dependency approach (Brines 1994), which would expect gendered rationales for the division of household labor: partners with very similar incomes share household labor the most equally, while in all other constellations the woman does more. The couples who decide that the woman will reduce or interrupt her paid work do not employ arguments of economic dependency when explaining that she will do more of the housework after childbirth.
Only one couple, couple E, planned for the man to become economically dependent on his partner. In the second interview, when this plan is no longer in practice, both partners say that the man had taken over a greater share of chores while the woman was the sole breadwinner. It had not been problematic for the male partner to take over more housework, even if it was -as expected -a problem for him to have no income of his own. The woman also does not report that it was problematic for her, or that she felt that she was more responsible for housework during the time of her sole providership.

Discussion
Previous research has shown that couples alter their division of labor upon the transition to parenthood in such a way that the women take over a greater share of the domestic tasks (Baxter et al. 2008;Cooke 2007;Gjerdingen/Center 2005;Huinink/Reichart 2008;Kühhirt 2012). We examined how predominantly highly educated German couples change their division of labor upon the birth of their first child based on a qualitative, longitudinal, and event-centered study. The reason for studying the transition to parenthood in a qualitative way was to explain why previous research has found mixed evidence for the different theoretical mechanisms.
About half of the couples interviewed displayed an equal division of paid and unpaid work before the birth of the first child; the other half practiced a partly specialized arrangement, with the women doing more housework. About six to twelve months after the birth of the child, this changed: two couples share paid work, housework, and childcare equally, while the others have a specialized arrangement in which the men focus on breadwinning and the women on housework and childcare, even if the majority of these women work part-time. The couples explained their plans for the period after the child's birth -and the realization of these plans -with a mixture of economic and gendered arguments that often go hand in hand. In line with previous quantitative and qualitative studies, we could show that none of the theories is able to capture the complexity of the decisions concerning the division of labor during the transition to parenthood.
The assumption of complete specialization proposed by the new home economics (Becker 1998) does not coincide with the ideas of the couples interviewed, with both partners often wanting to be active in the labor market, and all couples wanting to share childcare. Additionally, some couples do not really have a choice concerning their future division of labor, since there are job-related aspects influencing the decision. Thus, even if the couples explain their plans for the division of labor after childbirth with income differences, as the new home economics suggests, they do not plan a complete specialization, since childcare, at least, seems to have a utility of its own. Some couples even made their decision with disregard for economic rationalities, since this produced a higher utility at least for one partner; in these decisions, norms and values were more important than income differences.
The weighing up and discussion of the different ideas described by the intervieweesand the results of comparisons of the interviews within the individual couples -were partly expected by bargaining approaches (Ott 1992). The negotiations described in the interviews deviated in two aspects from theoretical expectations: childcare as an element of domestic labor is not seen as an unfavorable task, but in fact as favorable, and one's own career is not always preferred over other activities. Some couples bargained about which partner should be allowed to stay at home, since both would have liked to take parental leave in order to care for the child, or since their income differences would have suggested another decision. Thus, the bargaining approach could benefit from including childcare as a favorable activity, even if this is not transferable to other relationships.
The available time was, for some couples, a resource used to explain the anticipated and realized division of housework, as expected by Coverman (1985): The partner who has fewer working hours is responsible for more domestic tasks. The demand/response capability approach suggests an order to the decisions: the time spent at work determines how much household labor a person can do. Nevertheless, this is not the causal order the interviewees with specialization plans and realizations suggest: they decide first upon who is to care for the child, and whether this person is to be active in the labor market or not. Then, the person who has more time at home to take care for the child is supposed to do most of the housework. The causal order for the couple who altered their division of labor towards a more equal arrangement is also different from the theoretical assumption: it was not the woman's return to work that demanded a more equal arrangement, but the very ideal of equality that demanded both the woman's reentry and the man's reduction of his working hours. Thus, the idea of an order in the decisions is important to an understanding of couples' divisions of labor; however, when a new field of work -here, childcare -is added, the order of decisions seems to be affected, and the new field is the first to be decided upon.
The interviews showed the significance of gendered ideas, and that gendered explanations are often connected to economic rationales. The related theories discussed above helped in understanding the interviewees' explanations, without being sufficient to understand the ongoing processes. As expected, the interviewees make reference to different identities, and the women in particular anticipate that their family identity will be more important to them after the birth of their first child (Bielby/Bielby 1989). The interviews also show that the plans and realizations include complementary identities for men and women: women care for the child and do most of the housework, while men earn the greater part of the family income. Contrary to the assumptions of the identity formation approach, the men and women interviewed expected, in the first interview, the men to be active fathers who spend time with their child. The women who decided on an equal division of paid and unpaid work should feel a burden, as they have to balance their work and family identities; however, none of them expresses that feeling. When the concept of identities is included in the analysis, decisions and explanations that contradict the economic theories' logic of rationality become understandable.
The dependency approach (Brines 1994) argued that, in contrast to the economic theories, women should take over more domestic labor not only if they are economically dependent, but also when their partners are economically dependent on them. The sample only included one couple in which the man was economically dependent for some time; during this time, he did most of the housework, which is not in line with the theoretical assumption.
Since most of the couples are highly educated, it can be assumed that most of them have egalitarian values that influence their division of labor (van Berkel/de Graaf 1999). Indeed, the couples interviewed stated that they had egalitarian gender roles before the birth of the child. Contrary to the theoretical assumption of constant attitudes, many couples applied a different interpretative frame for the time after the child was born: clearly gendered ideals shaped their plans, and were still dominant at the time of the second interview, as the couples perceive themselves no longer just as men and women, but as (future) fathers and mothers. This change is more in line with the expectations from the identity formation model. Three couples continue to express a desire for egalitarian gender roles, with two of them putting these into practice.
The qualitative approach allowed us to conclude that the main explaining factor in labor division is the couples' ideals concerning parenthood, and to show that economic and gender rationales are interconnected, and in what way (Perry-Jenkins et al. 2013). Our study focused on highly educated, South German couples in their transition to parenthood, at a time before the new parental leave legislation was enacted. Therefore, we can only draw conclusions on this basis. Our findings suggest that carrying out qualitative research focusing on crucial events in couples' life courses could be of great benefit for future theoretical discussions.