Division of Care and Leave Arrangements in Gay Father Families in Sweden

This study analyses the division of parental leave and the income development in gay father families through surrogacy in Sweden, seen as one of the most family-friendly and egalitarian countries in the world. Based on longitudinal population register data, descriptive and bivariate regression models are estimated to analyse the parental leave uptake and income development of married partners becoming (first-time) parents in 2006–2015 (in total 53 couples). Retrospective in-depth interviews with 23 gay men in 12 couples, conducted in 2010 and 2018 are analysed thematically to study how fathers discussed and decided how to divide the leave. The process of establishing legal parenthood delays the fathers’ access to reimbursed parental leave. Despite this, the fathers’ earnings were not considerably affected by the addition of a child to the family. Once the fathers had access to reimbursed leave, they generally shared this equally, with a tendency for the genetic father to take leave first and for a slightly longer period. The fathers becoming parents via surrogacy arrangements are a well-off group, able to counter the negative financial consequences of becoming parents. Swedish family policies enable parents to share the leave equally. Less impacted by gender and parenthood norms creating difference between parents, gay father families are in a better position to realise ideals of shared care and sameness. The delayed access to reimbursed parental leave structures the ability of gay couples to become parents and contributes to class inequalities in the transition to parenthood.


Introduction
When a couple is expecting a baby, they make decisions on how to care for the infant. Cultural gender expectations, legal access to paid family leave, the financial situation, jobrelated demands and ambitions, and personal ideals and values may contribute to the couple's decision on how to divide the leave. Decisions like these have a short-term impact on the earnings of the parents and often have consequences for wages and career prospects in the years to come (Aisenbrey et al., 2009;Budig & England, 2001;Musick et al., 2020).
The research exploring work-care arrangements and the use of care leave in heterosexual couples show large gender differences, with women remaining primary care takers in most families (e.g. Hook, 2006;Reimer, 2020;Sundström & Duvander, 2002). Some research has explored these issues in lesbian couples, indicating differences where birth mothers do more childcare and take longer leaves than nonbirth mothers Malmquist, 2020;Moberg, 2016;Rudlende & Lima, 2018). In contrast, little is known about the leave arrangements of gay, male parenting couples. Gay fathers challenge hegemonic notions of masculinity as well as heteronormative meanings of parenthood (Moore & Stambolis-Ruhstorfer, 2013;Stacey, 2006). By learning more about their motives along with their division of care during and after infancy, we can gain important insights into issues at the heart of family political debates and family sociological theory (cf. Moore & Stambolis-Ruhstorfer, 2013).

3
This article explores the reasoning around taking care leave and the actual division of leave among gay fathers who have become parents through surrogacy. Gay fathers through surrogacy are an interesting population to study, as they are parents who raise a child from infancy, but where physiological factors (e.g. recovery from childbirth and breastfeeding) will not impact the choice of stay-at-home parent during the first months of a child's life. Institutional relations shape how gay men navigate the reproductive arena (Berkowitz, 2008). We focus on Sweden, known for its family and father-friendly policies. Although Sweden's policies are also seen as progressive to the extent that same-sex registered partnership and adoption have been legally recognised since 1995 and 2003, respectively, gay father families are not always covered by family leave policies during the early period of the child's life.
In this paper, we combine qualitative and quantitative data and analyses to study: (1) how parental leave is divided between the genetic father and spouse in married gay couples, (2) how gay fathers through surrogacy describe and justify their division of paid work and care during the first period of the child's life, and (3) how the parents' earnings and income develop as a result of these decisions. By comparing labour earnings with income development including social transfers, our analysis provides insight into the impact of family leave policies on remuneration and the financial consequences of the wait until legal parenthood is recognised and paid parental leave becomes available.
To the best of our knowledge, this is the first longitudinal analysis of gay fathers' labour market attachment and shortterm earnings and income development during the transition to parenthood. Given the scarcity of earlier research, the qualitative analysis provides a favourable backdrop that anchors the quantitative findings in the everyday lives and narratives of fathers. The different data and analyses inform one another and contribute to a better understanding of the first period as parents, thereby reducing the risk of unfounded conclusions (cf. Small, 2011). Worth noting is that gay fathers through surrogacy are a financially privileged group (Malmquist & Höjerström, 2020). Hence, the findings reported in this study are not representative of all gay fathers. In addition, this group is small, coming from a small country like Sweden (SOU 2016:11). In light of this, there are clear advantages to using qualitative data besides the quantitative evidence to illustrate and understand the patterns and relations observed.

Work-Care Arrangements-Theory and Earlier Research
Fatherhood and motherhood have different cultural meanings, and there are different expectations on mothers as child carers and parents, compared to fathers (Doucet, 2006;Grunow & Evertsson, 2019;Miller, 2011). Gender essentialist discourses portray mothers as innately care-giving, completed and fulfilled by nurturing children, and fathers as fumbling and reluctant care-figures in need of (mothers') guidance and support. Heteronormative ideals prevail, and heterosexual parenting continues to be privileged.
The question of how meanings and parenting ideals influence and inform the social lives of gay parents has received modest but growing interest in the literature (e.g. Berkowitz, 2008;Bos, 2010;Brown et al., 2009;Carroll, 2018;Vinjamuri, 2015). Gay men who become parents challenge hegemonic notions of masculinity and must deal with the idea that men in general, and gay men in particular, are not naturally predisposed to raising children (Stacey, 2006). As men, they are expected to prioritise work and the publicrather than the private-sphere. The unencumbered, ideal worker is a man (Acker, 1990), and masculine workplace norms can make it difficult for fathers to invest time in childcare and to take leave (Haas & Hwang, 2019;Hodges & Budig, 2010). Research on the division of work in gay compared to lesbian couples suggests that male same-sex couples spend more time in paid work than female same-sex couples do (Jaspers & Verbakel, 2013;van der Vleuten et al., 2021). Still, evidence from small-scale studies suggests that gay men may be balancing work and care with greater ease than heterosexual fathers (Richardson et al., 2012) and that they are more likely to prioritise their parenting role over the worker role (Bergman et al., 2010).
Although still a small field of study mainly focused on the USA, research on the division of paid labour, housework and childcare in gay father families has found that gay fathers on average have a more egalitarian division of work and care compared to heterosexual couples (Farr & Patterson, 2013;Goldberg et al., 2012a, b;Tornello et al., 2015b). In a small European survey study, gay fathers also showed higher levels of cooperation by being more likely to exchange ideas about their parenting with their partner than heterosexual fathers were (Neresheimer & Daum, 2020). This may partly be due to the need for gay fathers to create a new fatherhood role outside the normative constraints of heterosexual parenting.
Even though research suggests that specialisation in paid work and care is less common in gay couples than in heterosexual couples (cf. Becker, 1991), it may still play a role in gay father families. Gay fathers who work more hours outside the home than their partners tend to do less housework and participate less in childcare (Tornello et al., 2015b), and partners who take more responsibility for childcare also perform more feminine coded housework (Goldberg et al., 2012a, b). Studies of lesbian couples show that employment security, flexible work conditions and expected financial loss were important when deciding who would carry the child (Chabot & Ames, 2004;Goldberg, 2006;Mezey, 2008). The mother carrying the child also often took the first and the longest parental leave Rudlende & Lima, 2018). For men, genetic parenthood is less consequential for the division of leave than for women and genetic fatherhood does not seem to affect work-care arrangements among gay fathers through surrogacy in the USA (Tornello et al., 2015a).

Gay Men's Paths to Parenthood and Access to Parental Leave
Gay father families received the first legal safeguard in 2003, when Swedish law allowed same-sex couples to apply for adoption. Still, adopting jointly as a same-sex couple has proven to be difficult and only a few such adoptions have been completed (Malmquist & Spånberg Ekholm, 2020;Statistics Sweden, 2018). Shared parenting arrangements with lesbian women is also an established route to parenthood for gay male couples (Ryan-Flood, 2009;Zetterqvist Nelson, 2007). However, since lesbian couples gained access to tax funded reproductive treatment in 2005, fewer seem to choose shared parenting arrangements with men (Rozental & Malmquist, 2015).
Apart from adoption and shared parenting with a woman or a lesbian couple, a gay male couple may consider surrogacy as a path to parenthood. Swedish legislation does not allow this type of arrangement at Swedish fertility clinics (SOU 2016:11). However, people with financial means can pursue gestational surrogacy abroad, where the USA is an available destination for gay male couples (Malmquist & Höjerström, 2020). Previous research shows that gay fathers through surrogacy are often upper-middle-class men who are highly educated and financially secure (Bergman et al., 2010;Malmquist & Höjerström, 2020;Tornello et al., 2015a;Vinjamuri, 2015). This is unsurprising given that a recent study on gestational surrogacy in the USA reports that total costs exceed $100,000 (Lindheim et al., 2019). While only being an option for the financially privileged, the number of children born through surrogacy abroad has increased rapidly. Before 2007, no Swedish municipality had handled any children born through surrogacy, but one decade later it was estimated that around 50 children per year were born to Swedish couples through surrogacy arranged abroad (SOU 2016:11). The majority of these children are brought up in heterosexual parent families and only a fraction are children in gay father families.
Swedish law recognises the person who gives birth as the legal parent and custodian, even when the birth is the result of surrogacy abroad (Malmquist & Spånberg Ekholm, 2020). When the genetic father is registered as a legal parent, custody can be transferred to him. The genetic father's partner can apply for second-parent adoption, and once granted, the surrogate's legal bond to the child ceases.
Gay fathers through surrogacy can make use of parental leave once the genetic father is registered as a legal parent of the child (Malmquist & Spånberg Ekholm, 2020). As this process can be lengthy, the option of using parental leave is postponed, and parents may need to find other ways to finance their time at home with the infant. Once the genetic father's legal parenthood is established, he may sign over leave days to his spouse. At the time of observation in this study, the Swedish parental leave system allowed parents 480 days of job-protected parental leave per child. Of the 480 days, 90 were reimbursed at a low flat rate and 390 days at close to 80% of previous earnings, up to a ceiling.

Parental Leave and the Ideal of Sharing Care
Sweden is known for its family-friendly policies and a tradition of supporting female employment and shared care (Korpi et al., 2013;Lundqvist, 2015;Misra et al., 2007). When the Swedish Prime Minister Olof Palme gave a talk to a US women's organisation in 1970, he emphasised the importance of involving men in gender equality endeavours, not least for men's own sake (Klinth & Johansson, 2010). This is indicative of the Swedish ideal and discourses linked to the double emancipation of women and men, and the father as an equally important parent as the mother (Klinth, 2002). In practice, there is still room for improvement.
Introduced in 1974, the parental leave insurance was one of the reforms intended to facilitate the combination of work and care for both men and women (Klinth, 2002). While mothers still utilise the lion's share of parental leave in most heterosexual families, fathers' parental leave use has increased over time, especially since the mid-1990s. In 2016, 22% of all couples shared the parental leave equally (i.e. used at least 40% of the leave each) during the child's first 2 years (Swedish Social Insurance Agency, 2020). Fathers with a high level of education and high earnings use more leave compared to those with lower education and earnings (Duvander & Johansson, 2012;Ma et al., 2020). Higher education and earnings often come with a better bargaining position vis-a-vis the employer and education is also often linked to egalitarian ideals (Duvander, 2014;Evertsson, 2014). On average, older fathers take longer leave than younger fathers (Ma et al., 2020).
Turning to gay father families, to our knowledge, no previous research on the division of leave exists. Based on results from studies of heterosexual couples, we expect that relative earnings, age and education may influence how the spouses divide the leave. In addition, having a genetic link to the child may influence leave use, partly because the genetic father is the first to have access to parental leave, and partly due to ideas and negotiations 1 3 linked to biological relatedness. As gay father families through surrogacy do not have access to parental leave from the day the child is born, it is also important to study the fathers' earnings and income trajectories. This would give an indication of how the unpaid leave is divided and contribute to an understanding of priorities linked to continued employment and earnings.
The abovementioned research together with egalitarian ideals and the importance of both parents taking significant leave as a way to establish a close relationship with the child (e.g. Klinth, 2002), leads us to expect that parental leave will be shared fairly equally in gay father families in Sweden. Men are expected to prioritise work rather than care, but given that gay fathers through surrogacy have no presumed 'natural' carer (e.g. linked to breastfeeding), the incentive to share care equally may be more significant.
Apart from deciding the amount of time each parent will stay at home to care for the child, the fathers also need to decide who should be home first. Although earlier (mainly qualitative and USA) research suggests that genetic parenthood has little impact on work-care arrangements in gay father families, no previous studies have investigated whether genetic parenthood is related to the division of care leave or the order in which parents stay at home.

Methods
This study responds to three main research questions: (1) How do gay fathers through surrogacy divide paid parental leave between them? (2) How do the fathers themselves describe their division of leave to care for their child and the negotiations leading up to them? (3) How do short-term earnings and incomes develop during the transition to parenthood for these fathers? To respond to research questions 1 and 3 above, we draw on data from population registers, focusing on gay male couples who became parents of an infant in the period from 2006 to 2015. To address research question 2 and shed additional light on the reasoning and decisions that have contributed to the patterns observed in analyses linked to research questions 1 and 3, we rely on in-depth interviews with gay fathers, conducted in 2010 and 2018. The different data and methods used enables us to conduct a study that is clearly anchored in peoples' everyday lives and experiences, yet empirically verified and described in a way that has relevance beyond the limited group of informants. In other words, we use a complementary approach to enrich the analysis and allow one type of data to partly compensate for the weaknesses of the other (Small, 2011).

Quantitative Data, Variables and Analyses
The population register data used include couples in which one father is identified as the genetic father and the other father has adopted the child. The analyses draw on data from the Total Population Register, the Multigenerational Register and the Longitudinal Integration Database for Health Insurance and Labour Market Studies (LISA).

Inclusion Criteria
In the analyses, we focus on couples who became parents to their first child in the period from 2006 to 2015. We include male same-sex married couples and registered partners who have a child in their current relationship, where the child was born in the year of marriage or later. By narrowing down onto couples in which one father is registered as the genetic father and the other has adopted the child, we have most likely identified children born through surrogacy, because a jointly adopted child would have two adoptive fathers, and a child conceived in a shared parenting arrangement with a woman would be registered as the legal child of the birth mother, not as the adoptive child of the non-genetic father. Couples for which information exists for at least the years before the birth of the first child, the same year and the year after were included in the analyses. As parental leave periods are longer for couples with two children, we exclude those who have more than one child in the same year. Based on the abovementioned criteria, in total, 53 couples were included in the analysis. Although a small group, it comprises all families in the Swedish population who meet the inclusion criteria. We follow the individuals at most 3 years after the birth of the child, or until 2016. This means that data from the second and third year after birth is missing for couples whose children were born late in the period studied. The number of couples decreases to 43 in year 2 after the child's birth year and to 38 in year 3.

Variables
The variable parental leave is the sum of regular parental leave and temporary parental leave, measured as net days (individuals can take leave part-time). Temporary parental leave includes days used to take leave from paid work to care for a sick child. In addition to leave days, the adoptive father's share of the couple's total leave per year is presented. Two income variables are used. The first, earnings, includes income from employment and self-employment. The second, referred to as income, includes income from employment, self-employment and work-related social benefits. Hence, parental leave benefits and other work-related 1 3 social transfers are included in income. Both indicators are adjusted for inflation and correspond to the monetary value for 2014.

Descriptive Information, Quantitative Sample
The fathers were on average 37 years of age when the child was born (this is the same for both adoptive and genetic fathers). This means that they were at a stage in their career when they were established in the labour market and had a good income and education. In the couples studied, 60% of the genetic fathers (32/53) and 50% of the adoptive fathers (27/53) had post-secondary education of at least 3 years, as of the year before the child was born. The earnings of adoptive fathers and genetic fathers were very similar. The median incomes were SEK 434,000 (ca USD 47,000) for genetic fathers and SEK 443,000 (ca USD 48,000) for adoptive fathers the year before the child was born. The average earnings are high compared to men overall in this age group. In 2013, men 35-39 years of age earned on average close to SEK 350,000 (ca USD 41,000) (Statistics Sweden, 2014). In the register data, we have no information on inheritance, financial investments, wealth accumulation or the like, but it is not unlikely that resources like these enable some of the couples to have a child via surrogacy arrangements.

Analysis
The quantitative analyses are descriptive, showing the division of parental leave in the year of birth and onwards, the median number of days used by each father and the fathers' median earnings and income the years before and after birth. We also estimate simple bivariate regression models to check if factors such as relative earnings, age, the wait until adoption is finalised, and the birth month of the child are linked to the adoptive father's share of leave in the year after the birth year. The child's birth month is included as a continuous variable (centred on birth month, where 6 = 0), as is the elapsed time in months from the year and month the child was born until the adoptive father was legally recognised. Birth month is used to estimate the tendency to let the genetic father take leave first. If this was a common pattern, we should find a negative relationship between the birth month of the child and the adoptive father's share of the leave in year 1, suggesting that the later in the year the child is born, the lower the adoptive father's share of the leave will be in the succeeding year. Further, if it were common that couples wait until the adoptive father is legally recognised before he starts his leave, we should find a significant relationship between his share of the leave in year 1 and the length of time (in months) from the time of childbirth until the time of adoption. Finally, if relative earnings or age mattered for the division of leave, we would expect to find a significant relationship between the adoptive father's share of the household earnings/relative age the year before the child was born and his parental leave use in the year after the child was born.

Qualitative Data and Analyses
The interviews with gay men who had become fathers through surrogacy were part of a project on gay fathers' paths to parenthood and their experiences as parents of young children. 11 interviews were conducted with both partners and one with only one partner-a total of 23 interviewees. Dyadic interviews are common in family research (Morgan et al., 2013) and are often seen as an advantage when exploring topics that involves a dyadic decision (Kvalsvik & Øgaard, 2021). Dyadic interviews can also be an advantage when focus is on retrospective description and memories of, in our case, family leave periods and the negotiations and decisions that led up to them.

Data Collection
Interviewees were recruited through social media, where an advertisement was circulated in relevant communities, such as a Facebook group for gay fathers and the website for the Swedish national association for homosexual, bisexual, transsexual and queer rights (RFSL). The advert was further shared as open posts on the researchers' Facebook walls. Snowball recruitment included additional participants. The advert briefly explained the study and provided the contact details of the researchers.
Five of the interviews were conducted by the second author and the remaining seven by two master's students. The interviews followed a semi-structured interview guide covering topics such as the decision to parent, pathways to parenthood, relational aspects of parenting and encounters with society. The fathers were asked to tell their family story, describing their journey chronologically from when they started to plan for parenthood up until the situation at the time of the interview. Parental leave and care leave were topics that were brought up along the way. The interviewer made sure that important topics were covered by consulting the interview guide. Examples of questions asked were 'How did you divide the time with your child at home?' 'How come you chose to [do like this]?', and 'What was important when making this arrangement?'.
The interviews were audio-recorded and were between 45 and 107 min in length. Interviews were conducted faceto-face, and the interviewees themselves were invited to select the location and time for the interview. A few interviewees chose to meet at public cafés, but most invited the interviewer to their homes. The interviewees provided their written informed consent to participate. They were assured that they would remain anonymous, that they had the right to discontinue their participation at any time and had the right to refuse to answer questions. Interviews were transcribed verbatim, with pseudonyms replacing names.

Analysis
The interview data was analysed using thematic analysis as described by Clarke (2006, 2019). This means that we acknowledge meaning-making to be context-bound, positioned and situated, and consider themes as patterns of shared meanings (Braun & Clarke, 2006. The analysis was inductive and although the material was approached with certain questions in mind (linking the quantitative and qualitative data), the researcher entered the process without analytical preconceptions. More specifically, the second author read the entire dataset and marked all instances where parental leave arrangements or work-carearrangements were mentioned. This data was copied into a separate document, where detailed coding was carried out by the second author. Codes were kept close to the data in accordance with the inductive approach. Codes were also compared across interviews to establish concurrency. During the subsequent step, codes were sorted into candidate themes, each covering aspects mentioned as relevant for the arrangements of parental leave. In the next step, the selected parts of the transcripts were re-read, and relevant excerpts for each candidate theme were generated. The candidate themes were revised during this phase until the final themes were developed.

Participants
The 23 interviewees were between 33 and 51 years old, most of them in their forties. Nineteen resided in large Swedish cities and four in medium-sized cities. A majority (n = 19) had a university education and were employed in highincome professions (e.g. physician, lawyer, engineer). The remaining (n = 4) had exams equivalent to high school (with positions such as operator, inspector). The couples can be described as middle to upper-middle class. Most interviewees were born in Sweden, and most were white. Nineteen (10 couples) had sought gestational surrogacy arrangements through agencies in the USA. The remaining four (two couples) had children through traditional surrogacy and had conceived through self-insemination. Most couples had one child at the time of the interview, while a few couples had two or three children. The children were born in 2007-2018 and were 0-7 years old at the time of the interview, with a majority being 3 or younger.

Overlapping Samples
With the quantitative data covering all families with children born in 2006-2015 (excluding parents of twins), and the qualitative data covering families with children born in 2007-2018, most but not all interviewed fathers are likely part of the group of fathers included in the statistical analyses. Despite this overlap, we cannot-and have no desire to-link the fathers in the two samples.

The Division of Parental Leave and Income Development in Gay Couples Transitioning to Parenthood-Quantitative Evidence
The first research question refers to how the fathers divide parental leave between them. In Fig. 1, we see the median share of leave for adoptive fathers in couples who have taken any leave in a given year, where year 0 is the year of birth of the child. The graph shows that couples tend to share the leave equally from year 1 onwards, as the line representing the adoptive father's share is close to 50%. It is worth noting that couples may have had an additional child in the period, and this may influence and to some extent exaggerate the impression of equal sharing if the fathers take turns at being the genetic father and if the genetic father takes leave first. If we look at the 36 couples in the dataset that used some leave time in the third year since the birth year, 68% have more than one child aged 0-3 in the household. Table 1 shows the median number of leave days used by the genetic and the adoptive father. This table includes all fathers (not only those taking leave) and is partly indicative of the ways in which the leave uptake influences the fathers' time away from paid work. The rather low uptake of parental leave in the year the child is born is most likely a consequence of the time gap before legal parenthood is registered and the fathers have access to paid leave. In line with the conclusions based on Fig. 1, we see a tendency for higher leave use among genetic fathers in the year the child was born and in the first year thereafter, whereas the leave use for the adoptive father is slightly higher in year 2, even though fewer days are used. On average, adoptive fathers use 95 days of leave and genetic fathers 108 days during year 1 after the child's birth, that is, the year when both fathers take the most parental leave. Parental leave use is low in years 2 and 3, when most children are in preschool.
In order to get a better understanding of potential factors that may influence the division of leave and to provide a tentative answer to the second research question, we run separate and bivariate regression models to predict the adoptive fathers' share of leave in year 1, by including the birth month of the child, the time it takes for the adoption to be legally recognised, the adoptive father's earnings and relative age. If genetic parenthood directly or indirectly influences which parent takes the first and longest leave, we would expect to find a significant correlation between birth month and the adoptive father's leave use in year 1. In Table 2, we see a significant negative correlation between the month of birth and the adoptive father's share of the leave. For simplicity, we assume that the link between birth month and the father's share of leave is linear (the relationship may very well be curvilinear, but we refrain from specifying such a model due to the small N and to minimise the degrees of freedom used). The coefficient in Table 2 suggests that a one unit increase in birth month leads to a 2% decrease in the adoptive father's share of the leave. In other words, the adoptive father's share of the leave is smaller if the child is born in the later months of the previous year compared to the earlier months. This is further evidence of the tendency to let the genetic father take the first, slightly longer share of leave in year 1. In the interviews, we find clues to the fathers' reasoning in this respect.
The time until adoption is legally recognised varies considerably in the sample (from 2 to 26 months). For more than 70% of the couples, the adoption process is complete 9 months after birth. If the adoptive father waits until he is legally recognised before he takes leave, we can expect to find a significant negative correlation between the length of time from childbirth until adoption, and the adoptive father's share of leave in year 1. This assumption is not supported when we include an indicator of the length (in months) from childbirth until the adoption process is finalised (results available from the authors on request). In simple, bivariate regression models, we find no indication of the fathers' ages or their relative age difference being of importance for the division of leave. In addition, no significant correlation between the adoptive father's share of total household earnings and his share of parental leave in year 1 is observed. This may not come as a surprise, given that the average earnings difference between genetic and adoptive fathers is small in these couples. Even so, work-related factors may enter into the negotiations, as evidenced in the interviews with the fathers.
As indicated above, couples who become fathers through surrogacy arrangements cannot make use of parental leave before the genetic father is registered as a parent. This means that not only access to paid leave, but also the wait until the leave is available to them may influence their earnings and income development in the years around childbirth. In Fig. 2, we see the median earnings of adoptive and genetic fathers in the period from 2 years before, until 3 years after   the birth year of the child. To give an indication of the spread within the sample, we include dotted lines showing the 25th and 75th percentiles of the earnings distribution. This means that 50% of the genetic (green lines with boxes) and adoptive fathers' (blue lines with triangles) earnings fall within the dotted lines. The earnings of the genetic fathers are slightly lower than those of the adoptive fathers in year 0, whereas in years 1 and 2, the genetic fathers earn slightly more than adoptive fathers. The differences are small and insignificant. The difference in the income development from year 0 to years 1 and 2 is smoothed out somewhat, not least for adoptive fathers, when parental leave benefits and other workrelated social transfers are added to the yearly labour income (Fig. 3). This responds to the third research question and by and large, the figure suggests that the arrival of a child has a minor impact on the fathers' income development. However, it is worth noting that income varies considerably within the group we study here, and in the interviews, we learn more about some of the measures the fathers took to maintain a favourable financial situation in the period up until parental leave became available to them.

Reflections about Care-taking and Parental Leave Decisions: Qualitative Analysis of Interviews
Children typically start day care at approximately 1.5 years of age in Sweden. Prior to this, one parent stays at home to care for the child or both parents take turns staying home. Among the 12 couples represented in the interviews, most described an arrangement where the time at home had been shared equally (e.g. Mats and Hasse stayed home for 9 months each). Only a few had a semi-equal arrangement (e.g. Lars stayed home for 10 months and Jonas for 6 months) or an unequal arrangement (e.g. Anders stayed home on a full-time basis and Frank worked part-time for a period). Below, we provide a picture of the reasons and considerations that spurred the division of paid and unpaid leave, thereby contributing to answering the second research question, but also informing the third question (linked to income/earnings development). The interviewees described various concerns that influenced the arrangements they made for parental leave. The topics brought up have been arranged into the following themes: legal obstacles to paid parental leave, valuing equal sharing, meanings attributed to genetic parenthood, and work-related considerations.

Legal Obstacles to Paid Parental Leave
The interviewees described a time gap from the birth of their child until legal parenthood was established for the genetic father, who only at this time could access paid parental leave. Several interviewees described how one or both fathers had taken unpaid leave of absence to be able to stay home and care for the child during this period. As Karl reflects: You don't have access to paid parental leave until you're a legal parent. It took a long time before we had paid parental leave; first, we needed a custody decision showing that you are the parent, eh, so there were a couple of months when we had to live on our savings.
Alternatives to the use of savings or taking loans to finance leaves of absence from work were the use of saved vacation time or working part-time from home. As each family had to find their own way to manage the period until accessing paid parental leave, each parent's unique circumstances, including their employment position, negotiations with their employers, and willingness to stay at home, likely impacted the arrangement (cf. , as partly described in the theme on work-related reasons below. Based on the average earnings development described earlier (Fig. 3), the combination of part-time work and vacation seems to be a common way to finance leave, as average earnings were not affected much. A few couples also seemed unaware of the possibility to let the genetic father transfer parental leave to his spouse and explained that the non-genetic father only went on leave after the second-parent adoption was approved.

Valuing Equal Sharing
As indicated above and in the quantitative analyses, many of the interviewees had shared care leaves equally. The decision to do so was often depicted as self-evident, desirable or non-negotiable. The parental leave period was described as important for establishing a close bond with the child, and several interviewees described how both fathers shared a mutual interest in forming such bonds. Some also emphasised the importance of sharing equally to maintain equality in their adult relationship. Levi and Mikael depicted the need for equal sharing as a given: Levi: We have, like, not even had a discussion, none, I think neither of us would have accepted anything else. Mikael: No, rather, we wanted to be with this little cutie, and also, like, why wouldn't we share this equally?
For Levi and Mikael, sharing parental leave equally matched their otherwise equal share of duties in the household. Several interviewees also emphasised that the time at home had been important for them and described how they valued close parent-child relations over their career. Another interviewee, Karl, explicitly discussed parental leave arrangements in relation to equality values: Karl: I am very satisfied and proud, that we have always, eh, ourselves tried to see ourselves as equal parents […] Like we have always had this everysecond principle. Eh, for instance, that we put him to bed every second time, take every second diaper, every second feeding, every second… everything, like. Interviewer: Mm, now this time has passed in your lives, but what was your thinking when deciding on sharing parental leave? How did you solve that? Karl: Ehm… then we just thought that we should be at home about the same amount of time with him, and that's how we did it too.
Karl depicted equality as a central value in their relationship and emphasised how proud they were about their equal arrangement. By highlighting his pride, Karl adheres to and embraces the egalitarian parenthood ideal that has been central in Swedish family policy debates. Some interviewees stated that equality is easier to achieve in same-sex couples, while others claimed that achieving equal arrangements required active engagement on their part. The tendency to share parental leave fairly equally was also evident in the quantitative data. However, a few interviewees downplayed the importance of sharing equally and described how each partner might contribute to the family's wellbeing differently.

Genetic Parenthood
In the quantitative analyses, we saw tendencies for genetic fathers to take the first and slightly longer leave than adoptive fathers. In the interviews, several fathers raised the subject of genetic fatherhood in relation to their parental leave arrangements. Most often, however, they claimed that genetic fatherhood did not matter to how they arranged their leave. Some explained that they decided how to arrange parental leave before they decided or knew who the genetic father would be. Several also pointed out that biology, in a broader sense, matters less in male samesex couples compared to female same-sex or different-sex couples. As David reflects: That often happens if one breastfeeds: in that case, one parent, the mother of course, will gain a very tight bond, ehm, and in that case, it's maybe hard to let go of that, and it becomes this classic thing with "How should the father come in and be part of this" […] but for us, we could give the bottle every second time.
David here presents breastfeeding as having consequences way beyond the practicalities of feeding. It can be a desired practice ('hard to let go of') that influence bonding and that may shut the other parent out. Seen from this perspective, bottle-feeding gay fathers are, on the other hand, depicted as unaffected by biological differences and therefore free to arrange care in accordance with their own preferences. In this case, this led them to take turns in feeding, something that also has consequences for bonding.
A few interviewees mentioned that genetic parenthood factored into their leave arrangements. Lars, a non-genetic father, had been on parental leave from infancy until the child was 10 months old. He said that it had been beneficial for him to bond early with the child, which had felt like a form of compensation for not being a genetic parent. Another couple, Thomas and Fredrik, highlighted genetic fatherhood as their primary concern when deciding on their arrangement. Thomas explained: Thomas: We didn't even talk about it. It was just, it felt natural for us that the biological father [began]; also, this suited us better because I, I just got a job then too, and Fredrik wanted to take paternity leave and so, no it just, it just moved, it just evolved naturally. Interviewer: So, the biological parenthood was important for..? Thomas: No. And we have faced many questions about who's the biological, like, you notice that people are curious. This is quite interesting, that people find it so important who the biological parent is. And we don't find that important, because we're both dads.
In contrast to Lars' reflection, Thomas initially depicted it as natural that the genetic father (here labelled 'biological') would take the first part of parental leave, although he also highlighted other considerations related to work and personal preferences. But when the interviewer asks him about the importance of biological parenthood, the conversation takes another turn as Thomas downplays any importance of biological parenthood. Clearly the matter of genetic parenthood is surrounded by strong discourses, which the interviewee navigates. Possibly, the tendency seen in the quantitative data that genetic fathers took the first and slightly longer parental leave, might have been toned down by the interviewees, who preferred highlighting other reasons for their decisions.

Work-Related Considerations
Many interviewees mentioned work-related concerns when they decided how to arrange parental leave. Among those who shared parental leave equally, work concerns had often affected the decision of when each partner took their parental leave. Mikael and Levi's child was born in the summer. They decided that both fathers would stay home together during the first few months, before each father would take about 9 months of parental leave each. Mikael explained how they decided that he would take the first period: I needed a substitute for my job, so it was better for me to be [at home] for a coherent period from the beginning. And you [to Levi] did not need a substitute, so it made more sense to split it up a bit, so you could have the summer, then you worked, and then we switched sometime during spring.
According to the couple, their joint interest in staying home with their child had led them to share the leave equally. However, work-related concerns affected the arrangement at the next step. Rather than affecting each parent's amount of parental leave, it affected the order in which they stayed at home.
Most interviewees emphasised that their employers had approved leave of absence without any objections. However, for the few couples with a semi-equal or unequal parental leave arrangement, work-related concerns were often cited as the main factor for their decision. Some of them depicted their employer as unsupportive of their desire to take parental leave. For example, a couple with a semi-equal arrangement, Jonas and Lars, described how Jonas' boss made fun of men who took parental leave. Jonas further explained how the boss expected him to continue working from home during his parental leave. Another couple, Anders and Frank, had an unequal arrangement, where Anders had taken fulltime parental leave, while Frank had worked part-time for a period, then returned to work full-time. Anders described their situation: I've wanted to start working a little, so it would have been beneficial if Frank had been able to take paternity leave. Still, you [to Frank] have a job that doesn't really allow you to be on paternity leave, because no one would jump in to do your assignments.
[…] If he worked part-time, that would just burden the others who already have full-time jobs, or Frank would have to do it anyway, despite working part-time. So that is also part of it. It is not that simple to just work parttime.
Although Frank's work is depicted as a job where no one jumps in to do your assignments, parental leave rights are protected by law and employers cannot deny parental leave altogether. However, by referring to work demands, the couple suggests that the decision is out of their hands. While work demands seem to be a justifiable reason for sharing parental leave unequally, any individual preferences for such arrangements may be a less available line of argumentation in a context where equality is highly idealised. In other words, this may be a legitimate way to motivate an unequal division of leave in a country with a tradition of idealising both parents' shared care of the child.
Summing up, while those who shared parental leave equally often depicted their decision as ideal and self-evident, the fathers who had a less equal arrangement often depicted it as unfortunate and not infrequently out of their hands. Thus, the equal share of leave seemed to be idealised by most interviewees. This is also in line with findings from qualitative interviews with heterosexual couples as well as lesbian couples in Sweden (Alsarve et al., 2016;Malmquist, 2020). Although, in practice, couples do not always live according to the shared care and equal-carers ideal, they tend to consider it and refer to it.

Discussion
This paper examined a unique combination of population register data (based on 53 couples) and in-depth interviews (representing 12 couples) to study how Swedish gay fathers through surrogacy discuss, describe and divide leave during the first years of their child's life. By combining longitudinal analyses of the division of leave and the fathers' earnings and income development with retrospective interviews, we provide a more comprehensive picture of gay fathers' transition to parenthood and division of work and care than any previous study to date. This study responds to three main research questions; (1) How do gay fathers through surrogacy divide the time at home with the child? (2) How do the fathers describe their arrangements of leave to care for their child and the negotiations leading up to them? (3) How do short-term earnings and income develop during the transition to parenthood for these fathers? In the following, we respond to these questions in a contextualised and holistic manner, based on our complementary approach (cf. Small, 2011) and earlier research.
Situating the study in the sociohistorical setting of Sweden, the couples in focus find themselves in a context where equality and shared work-care divisions are the ideal and where the father's care is prioritised and considered important to the child. Still, the normative father is heterosexual, and the gay fathers' narratives provide evidence of the complexities that characterise the first period after the birth of a child, prior to the establishment of legal parenthood and access to paid parental leave. The obstacles the fathers must overcome to access parental leave insurance are also visible in the register data, where the fathers' median total use of parental leave in the year when the child is born is only 45 days (compared to over 200 days the year after). By taking leave of absence, working part-time, using saved vacation, or even taking loans, the fathers secured their ability to take the first stay-at-home period with the child. Still, it is a financially select and highly motivated group that can become parents this way and on average, their earnings trajectories seem to suffer little from the transition to parenthood. This suggests that the fathers have a plan for arranging the first period in the child's life before they undertake the process of becoming parents. The need to plan ahead and to free up (and afford) unpaid time for care contributes to inequalities in the ability to become a father. In contexts where paid parental leave is non-existent or shorter in duration, social class inequalities in the capacity to become parents are likely even greater.
The fathers' narratives are characterised by an expressed desire to share the first period at home with the child equally. Many describe an equal division of leave as natural and self-evident, and none of the men expressed plans to ease domestic responsibilities by, for example, hiring a nanny (cf. Berkowitz, 2020). This may be indicative of the ways in which dual-carer, dual-earner norms have permeated Swedish society (e.g. Misra et al., 2007), supported by campaigns highlighting a child's need for both parents, implicitly suggesting that such a need is fulfilled when they share parental leave (Alsarve et al., 2016;Klinth, 2002). Although normative discourse on equal sharing also shapes how parents talk about and describe the first period of the child's life in heterosexual and lesbian couples (e.g. Alsarve et al., 2019;Malmquist, 2020), the fathers in our study are better able to live up to this ideal, as evidenced in both the interviews and the analyses of population register data. Previous research suggests that birth mothers' relative use of parental leave days in the year of birth and the subsequent year is about 80% in heterosexual couples and 65% in lesbian couples . In our study, the genetic and adoptive fathers' respective share of parental leave use is about 50% in the first year after the birth year. Whereas the parental leave decisions of heterosexual couples are affected by gendered expectations and a wide earnings gap, the decisions of same-sex couples are not Moberg, 2016). An analysis of lesbian couples' division of parental leave shows that equal sharing is highly idealised, but while the birth mother's time at home is considered necessary (to recover from childbirth and breastfeeding), non-birth mothers' parental leave is more likely to be seen as optional and is often shortened for practical reasons (Malmquist, 2020). In gay father families, the ideal of sharing equally, a small within-couple earnings gap, and the non-presence of a birth parent are possible explanations for equal sharing. As Mikael, one of the informants said, 'why wouldn't we share this equally?' Being forerunners in many respects, gay fathers construct and exemplify a new fatherhood identity, signified by the importance of shared care and mutual closeness to the child.
Several fathers reported having an understanding employer, even though examples of managers hostile to parental leave surfaced. As men, gay fathers must navigate through expectations linked to masculinity and the ideal worker norm (Acker, 1990). Although it is in no way unusual for men to take leave in Sweden, it is not uncommon that employers expect fathers on leave to do some work from home or to stay in contact with the employer (Haas & Hwang, 2019). Such continuous work from home could partly explain why we see small earnings dips in the years the couple becomes parents. The fact that they (are known to) live with another man may also have contributed to their employers' understanding attitude and the fathers' generally successful leave negotiations. In the absence of a mother who is expected to take most of the leave, the employers' ability to question lengthy leaves are reduced. This suggests that it is not only the parents' own gender but the gender of the spouse that has implications for how employees are treated, a relationship that is not often acknowledged in family sociological research and theory.

Conclusion
Our study is the first to explore how gay fathers through surrogacy divide care leaves, how they negotiate the division of leave, and how this influences their earnings and income. Obstacles in accessing parental leave are experienced as a consequence of Swedish family law, which is not adapted to the social situation of fathers through surrogacy, where the establishment of legal parenthood and custody is a drawnout process. Despite a lack of access to paid parental leave for a period after birth, analyses of population register data suggest that most fathers are able to get through this period without any major decreases in earnings. Many of the interviewed fathers have flexible working conditions, which is common among the Swedish upper-middle class, including the ability to work from home, or to use saved vacation time or various forms of overtime compensation. Gay fathers through surrogacy embrace the equal sharing of parental leave, both in their expressed ideals and in practice, which is in line with Swedish egalitarian ideals, the dominant earnercarer ideology and the discourse on the child's right to both parents. Although a select and, in many ways, privileged group, gay fathers through surrogacy may be the most successful in fulfilling these ideals.

Limitations and Directions for Further Research
While being the first study to combine a longitudinal perspective on gay fathers' work-care arrangements with the stories and experiences expressed by fathers themselves, our research has limitations. The number of gay couples transitioning to parenthood is small in Sweden, and this prevents more sophisticated statistical analyses of the predictors of the division of leave and earnings development. In addition, although population register data is rich in many ways, it mainly includes yearly information. In many cases, monthly data on parental leave use and earnings development would have been preferable.
Another limitation concerns the interview sample, which was self-selected through responses to advertising. While most participants in the interviews are likely to be included in the register analyses, they were not selected as a randomised sample of the population. Nevertheless, the interviews provide a more in-depth understanding of how gay fathers through surrogacy reason about and realise family leave. Although dyadic interviews may be a disadvantage to some extent-as the partners may adjust their stories due to the presence of the other partner-we mainly see them as an advantage in our study, as joint decisions and retrospective descriptions can be corrected if memory fails one of the partners.
Further research is needed to explore care leave among gay fathers in broader contexts, including gay men who have sought other paths to parenthood than surrogacy, gay men who are less financially privileged and gay men in other countries. Such endeavours should be sensitive to the sociohistorical frameworks and institutions that shape how gay men navigate the reproductive arena and any family policies and insurances that come with legal parenthood (Berkowitz, 2008). A study that includes both men and women in different relationship forms and with different paths to parenthood would also be useful to illustrate the interplay between gender composition, gestation, genetic relatedness and income when arranging care leave.