Fertility decisions in transition: young adults’ perceptions on fertility three decades apart in Spain

ABSTRACT The aim of this unique study is to analyze the generational change in perceptions regarding fertility decisions between young Spanish adults in the mid-1980s and young Spanish adults in the early 2010s and to test the extent to which the theoretical perspectives on fertility are supported by individual narratives over time. For nearly thirty years, Spain has had low fertility levels that contrast with the persistent ideal family size of two children. However, the literature has barely explored individuals’ perceptions of fertility decision-making over time. This paper seizes the exceptional opportunity to analyze 97 in-depth personal interviews of highly educated young adults in urban Spain from two independent studies conducted in 1985 and 2012. The results indicate that although the value change associated with the second demographic transition remains over time, the changes in age norms and partnership formation norms underlie the main differences between the two generations. Economic uncertainty and the lack of support for work-life balance stand out as persistent structural factors that influence the fertility decisions in both samples. In addition, the respondents’ narratives confirm that the transition of gender-role norms towards greater egalitarianism remains unfinished. By assigning value to the perceptions of individuals, this study contributes to the understanding of how the normative framework and meaning of childbearing have changed under the rubric of the macrolevel theoretical perspectives that explain fertility decline.


Introduction
During the last four decades, low fertility has become a common characteristic for many postindustrial societies. From the previous research, we know that Western countries have gradually experienced important normative and cultural changes around family formation, especially regarding the postponement of marriage and childbearing. This set of value changes are known as the second demographic transition (SDT) (Lesthaeghe & Van de Kaa, 1986;Van de Kaa, 1987). Such changes are not independent from social and economic changes. Economic theories suggest that the incorporation of women into the labor market and the opportunity cost of having a child as a result is a reason for the postponing and lowering of fertility behavior. In addition, the role of macro-economic conditions regarding labor-market stability and the lack of institutional support for dual-earner couples cannot be neglected as influencing factors in generating perceived economic (in)security in making fertility decisions (Comolli, 2017;Kreyenfeld, Andersson, & Pailhé, 2012). In addition, and closely linked to the two above-mentioned factors, the research has emphasized the increasing importance of gender egalitarianism in explaining low fertility, both at the individual levelgender-role attitudes and behaviorsand at the institutional levellabor market institutions and family policies (Esping-Andersen & Billari, 2015;Goldscheider, Bernhardt, & Lappegård, 2015;McDonald, 2000aMcDonald, , 2000b. These three factorsthe value changes associated with the SDT, economic uncertainty, and gender-role conflictsare closely connected to one another. However, it remains unclear how these macro-level perspectives can or cannot be accommodated at the micro level in explaining changes in individuals' perceptions of fertility. Despite the abundant scientific literature on this topic, the evolution of individuals' perceptions regarding fertility decisions has not been compared over time, particularly from the time of emerging new family behaviors, i.e. the 1970s and 1980s, until today (the 2010s). How have young adults' perceptions about fertility evolved across generations? To what extent are the theoretical arguments connected with the normative changes identified in individuals' narratives? Using a unique research design and qualitative data, this study contributes to filling this gap by exploring whether a generational change occurs in the perceptions of fertility decisions between young adults in Spain in the mid-1980s and in the early 2010s. It leverages the alluring opportunity to compare the narratives from nearly 100 in-depth personal interviews of respondents from two generations approximately thirty years apart. The present qualitative study illustrates individuals' perceptions and mechanisms of fertility decision-making, which cannot be revealed using survey data. Re-visiting interview materials from the 1980s thirty years later -by applying qualitative secondary analysis (QSA)-allows them to be compared with contemporary interview data from a new perspective (Tate & Happ, 2018). The subsequent implication of this research is a contribution to theory validation at the micro level from the less frequently employed perspective of individual narratives.
The samples purposely comprise highly educated individuals for three reasons. First, there is a positive relationship between higher education and the fulfillment of fertility aspirations (Toulemon & Testa, 2005). Second, there is a closer link between higher education and gender egalitarianism (Goldscheider et al., 2015). Third, it has been suggested that highly educated individuals are the forerunners of fertility change (Sobotka, 2008).
Spain is a suitable context in which to explore these questions since it became a lowfertility country more than three decades ago (Billari & Kohler, 2004). However, in Spain and in other low-fertility settings, the two-child norm as the ideal family size for couples persists over time (Hilevych & Rusterholz, 2018;Sobotka & Beaujouan, 2014). This mismatch between fertility levels and couples' ideal fertility is occurring in a context in which gender egalitarian ideology has substantially increased (Arpino, Esping-Andersen, & Pessin, 2015) and in which the levels of economic uncertainty have remained high over time (Adserà, 2011a(Adserà, , 2011bDe la Rica & Iza, 2005).

Changes in perceptions on fertility and theoretical perspectives
With the social, economic and political changes associated with the passage of time, the mechanisms underlying fertility decisions are also expected to have evolved as well as the theoretical perspectives that are linked to them. Lesthaeghe and Van de Kaa (1986) and Van de Kaa (1987) identified the SDT with a set of value changes related to family formation. These changes included the increase and social acceptance of divorce, nonmarital cohabitation, births out of wedlock, new family models, a progressive secularization process, an increase in individualistic values and a greater importance attributed to self-fulfillment. Some technological and legal changes related to contraceptive measures, assisted reproduction, and abortion have contributed to this value change. In addition, the SDT cannot be understood as being separate from the structural change imposed by the educational expansion and the incorporation of women in the labor market. Thus, the SDT is used to understand the postponement of family formation and the subsequent patterns of low fertility in European societies. This process of the destandardization of the life course (Lesthaeghe & Moors, 2000, p. 153) originated as a Northern and Western European phenomenon during the 1960s that did not arrive in Spain and other Southern European countries until at least two decades later (Surkyn & Lesthaeghe, 2004). Although most of these value and behavioral changes have remained (Lesthaeghe & Surkyn, 2008;Sobotka, Zeman, & Kantorová, 2003), the dynamic nature of a 'transition' is meant to be ephemeral. Thus, by examining the perceptions of two generations on fertility, this study hypothesizes that the presence of the SDT's associated values in the fertility reasoning of the younger generation will be weaker than in the older generation given that the societal normative framework for having children will no longer be the same.
Since the 1960s, the SDT has become a key concept for European demographers; however, despite its general good reception, time has uncovered some competing reactions. Some scholars have pointed out that the SDT is ambiguous and not strictly 'demographic' (Coleman, 2004) and that, as a 'transition,' the starting and endpoints have never been defined (Bernhardt, 2004;Cliquet, 1991). Its main criticism may be that the concept itself implies a convergence of behaviors across countries that have not been observed (Sobotka, 2008). Several studies have aimed to test the SDT in country contexts different from Nordic countries. When finding empirical evidence of the value change associated with the SDT, authors have also acknowledged the different pace and process of transition in different European regions (Hoem, Kostova, Jasilioniene, & Mureşan, 2009;Kertzer, White, Bernardi, & Gabrielli, 2009;Sobotka et al., 2003;Surkyn & Lesthaeghe, 2004). However, in non-EU western countries such as East and Southeast Asia (Atoh, Kandiah, & Ivanov, 2004), Russia (Perelli-Harris & Gerber, 2011), and the U.S. (Raley, 2001), scholars have found partial or no evidence of the SDT.
One of the conclusions of Surkyn and Lesthaeghe (2004) in their work about the spread of the SDT through Europe is that the value and behavioral changes related to family formation are necessary but not sufficient to explain the demographic changes. Economic and political changes are also influencing factors that were not under the original scope of the SDT (Bernhardt, 2004;Liefbroer & Fokkema, 2008). The study of the relationship between economic uncertainty and fertility has a long tradition in explaining the postponement of having children (Sobotka et al., 2003). Many studies have found that unemployment, especially male unemployment, has a postponement effect on fertility decisions (Cazzola, Pasquini, & Angeli, 2016;Kravdal, 2002;Pailhé & Solaz, 2012;Schmitt, 2012;Vignoli, Drefahl, & De Santis, 2012). Evidence for fertility postponement has also been found among highly-educated women experiencing economic uncertainty (Kreyenfeld, 2010). Moreover, in contexts with a weak welfare state, such as the Spanish context (Esping-Andersen, 1999), economic uncertainty goes hand-in-hand with the lack of state support for childbearing, for example, the availability of public daycare, family policies, work-life balance, and labor-market flexibility (Domínguez-Folgueras, Jurado-Guerrero, & Botía-Morillas, 2018;Lapuerta, Baizán, & González, 2011), which lowers young adults' fertility intentions (Brinton, Bueno, Olah, & Hellum, 2018;Bueno & Brinton, 2019).
Although financial and economic conditions have traditionally affected fertility, the interest in this relationship in the literature was renewed with the onset of the Great Recession in 2008 (Comolli, 2017;Kreyenfeld et al., 2012;Sobotka, Skirbekk, & Philipov, 2011). For this analysis, it is acknowledged that the interviewees from both generations were experiencing times of economic uncertainty due to the oil crisis of the 1970s and the Great Recession of 2008. It is, therefore, hypothesized that the narratives from both sets of interviewees refer to economic uncertainty as a reason for limiting or postponing their fertility intentions.
Moreover, unstable economies and weak welfare states often coexist with low levels of gender equity both in labor-market institutions and family policies (Kreyenfeld et al., 2012); accordingly, these two factorseconomic uncertainty and gender equityare highly interrelated.
The gender dimension is not only closely linked to economic uncertainty. The changes associated with the SDT are also changes that are closely linked with a shift in gender-role norms, attitudes, and behaviors. However, some authors have identified the lack of a gender perspective as one of the shortcomings of the SDT perspective (Bernhardt, 2004;Kertzer et al., 2009). The shift in gender roles has recently paved the way for a new theoretical approach that places gender equity at the center of the fertility debate and posits that it is the keystone of a couple's fulfillment of their fertility aspirations (Esping-Andersen & Billari, 2015;Goldscheider et al., 2015;McDonald, 2000aMcDonald, , 2000b. According to this theoretical approach, the end of the prevalence of the male-breadwinner-femalecaregiver model led to the onset of a new, and not yet resolved, gender-role conflict, which is based on the inability of social institutions and individual behaviors to adapt to the new dual-earner model for families. At the societal level, in McDonald's (2000b) terms, there is an incoherence between the high level of gender equity in social institutions addressed to individuals (i.e. access to education and the labor market) and the low gender equity level in the social institutions that address families such as labor market characteristics (i.e. work hour flexibility, the gender wage gap) and state intervention that supports families (i.e. the tax system, parental and childcare leave policies). At the individual level, the mismatch occurs between the public and private spheres. Although women have been widely incorporated into the labor market, men do not equally share in housework and childcare tasks. This fact causes a downward trend in fertility since women still play a double role in the work-family conflict. In this regard, Bernhardt (2004) points out that the SDT framework should incorporate the gender inequality dimensions since the opportunity cost of having children and the value change regarding the family are not the same for women and men. In this analysis, it is hypothesized that the respondents from the older generation (interviewed in 1985) will express fewer constraints on fertility decisions related to gender inequality and the work-life balance conflict than the respondents from the younger generation (interviewed in 2012). The secondary role in the labor market for women from the eighties compared with the larger presence of dual-earner couples in the younger cohort can explain such a difference.
The three theoretical perspectives have points of connection that represent the evolution of the theoretical thinking in relation to fertility. This connection represents the transition itself. Therefore, the approach of this study is to examine the extent to which this shift in the theories that explain low fertility is also observed in the changing perceptions and motivations that surround fertility decisions over time as reported by young adults in Spain. What changes have happened in perceptions of fertility? What are the possible mechanisms underlying these changes in the specific context of urban Spain?

The Spanish context
Spain became a low-fertility country in the mid-eighties ( Figure 1). The total fertility rates (TFR) were 1.64 in 1985, and they crossed the barrier of 1.5 children per woman in 1987. In 2012, the TFR decreased to 1.32 children per woman (1.27 for native-born women). Simultaneously, the mean age at first child had uninterruptedly increased during the period and was six years apart between the time of our interviews in 1985, when it was 25.8 years old, and 2012, when it was 30.9 years old. Therefore, Spain has been identified as having the lowest-low-fertility levels and the latest-late mean age at first child (Billari & Kohler, 2004).   The continuous postponement of the first child and the fall in the birth rates of the second and third orders are the two main demographic explanations of the transition of Spain to a low-fertility country (Alberdi, 1995;Bernardi & Requena, 2003;Delgado & Livi-Bacci, 1992;Ortega-Osona & Kohler, 2001). Over the years, it has been shown that the increases in non-marital cohabitation and divorce rates have not decreased the fertility levels as they were expected to in the past (Billari & Kohler, 2004). The low fertility levels during the 1990s were expected to be transitory due to the postponement of fertility. However, time has shown that it was not only a matter of timing (Castro-Martín & Martín-García, 2016). Indeed, the postponement of fertility must be understood around a number of social and economic changes. The transition to democracy in Spain after Franco's death in 1975 brought radical transformations that led Spain towards the modernization of the economy and society. Important changes in family-related policies were achieved: the use of contraceptives was approved in 1978; the divorce law was instituted in 1981; and abortion was legalized in 1985. By then, the most traditional national-Catholic discourse regarding partnership formation and childbearing that persisted during most of the XXth century (Sánchez-Domínguez & Lundgren, 2015) lost part of its legitimacy, and young adults were more open to breaking social norms and adopting new behaviors. Thus, in light of the changes that occurred in family formation patterns, the SDT perspective was an appropriate framework to explain the fertility decline after 1975 (Alberdi, 1995;Bernardi & Requena, 2003;Devolder & Cabré, 2009).
Since that time, high unemployment has also characterized Spain and is a feature that has become a structural component of the Spanish labor market over time. The unemployment rates for individuals aged from 25 to 34 years old in 1985 and 2012 were 21.6% and 27.6%, respectively (according to OECD data). Because of high levels of economic uncertainty, young adults' transition to adulthood in Spain has been identified as one of the latest in the world, which results in delays in leaving the parental home, economic independence, partnership formation, and the arrival of the first child (Alberdi, 1995;Baizán, Aassve, & Billari, 2003). This chain of delays in life-course events is one of the leading explanations for the high levels of childlessness among Spanish women. The recent studies forecast that one in four women born in 1975 will reach the age of fifty without having had a child (Esteve, Devolder, & Domingo, 2016). Among partnered individuals, economic uncertainty is reported to be one of the main reason explaining unmet fertility intentions in Spain (Bueno & Brinton, 2019;Esteve & Treviño, 2019). The previous studies have shown how the fertility decline was more pronounced in countries that were more affected by the Great Recession such as those in Southern Europe (Sobotka et al., 2011).
The transition to democracy has also brought changes in gender roles because of women's educational expansion and greater labor force participation. However, although Spain's gender-role attitudes have rapidly evolved into higher levels of egalitarianism (Arpino et al., 2015), the pace of change is slower with respect to behaviors (Abril et al., 2015), which results in a non-resolved gender-role conflict. In this sense, the greater level of gender-egalitarianism within the couples has been identified as a positive effect for reproductive decisions in the Spanish case (Castro-Martín & Martín-García, 2016) following the path paved by other European countries (Neyer, Lappegård, & Vignoli, 2013).

Two qualitative studies
This work comprises a unique research design by analyzing two independent qualitative studies on family formation conducted in 1985 and 2012. The studies were conducted under different purposes. The 1985 study 1 aimed to understand the emergence of nonmarital cohabitation in Spain while the focus of the 2012 study 2 was to explore the link between gender equity and low fertility. However, the design of both studies shares a good number of commonalities that make their comparison feasible. Both studies were conducted in the metropolitan areas of Madrid and Barcelona, with men and women of similar ages and in different stages of the family formation process. In both studies, the interviewees provided details regarding their economic and job characteristics, partnership formation, fertility intentions, housework and childcare division of labor, and genderrole attitudes. In addition, a questionnaire that collected the sociodemographic characteristics of both partners was completed in both cases. While the exact wording of the questions asked differed in some cases, the general topics coincide. For example, regarding fertility decisions -the main question analyzed-the interviewer in 1985 would ask 'How has the process been toward considering having a baby?', while in the 2012 study, the mirror question was 'What are the conditions for deciding to have your first/next child?.'

Sample characteristics
Both samples comprise men and women in stable unions who hold a post-secondary education degree. 3 The total sample in the 1985 study comprised 144 interviews from which a subsample of highly educated respondents was selected for this analysis. The total sample in the 2012 study comprised more than 400 interviews in five different countries from which a subsample of partnered respondents in Spain was selected for this analysis. By focusing only on highly educated individuals, the social class heterogeneity of the qualitative comparative analysis is also minimized. The sample from 1985 (cohorts born between 1944 and 1964) consists of 44 respondents whereas the sample from 2012 (cohorts born between 1976 and 1987) consists of 53 respondents. Table 1 summarizes the basic sociodemographic characteristics of both samples. The interviewees from 2012 are equally distributed by gender. In 1985, there were twice as many males as females given that higher education among women was less common. The age range covers 20to 40-year-olds for the 1985 data and 24-to 35-year-olds for the 2012 data; the mean ages are similar at 29.3 and 30.3 years, respectively. The 1980s study allowed for respondents who had two children; therefore, a quarter of its sample are parents of two children whereas the 2012 study was designed for respondents who had one or no children.

Data collection
In both cases, the recruitment was conducted by using the snowball sampling method, which began with a large number of individuals as 'seeds' who were subsequently questioned for up to two referrals. The interviews were conducted face-to-face in Spanish or Catalan and lasted from 60 to 120 minutes. All interviews were voice-recorded and were subsequently transcribed in full by a native speaker. In 1985, the majority of the interviews were conducted by a male interviewer for the male respondents and by a female interviewer for the female respondents (interviewers were members of the 1985 research team). In 2012, one female interviewer, the author, collected all of the interviews. One of the strengths of both studies is the shared status of the interviewees and interviewers (who were also highly educated, native-born Spanish within the same age range as the respondents), which enhanced the mutual trust and made it easy to establish rapport. These commonalities may have facilitated greater sharing on the part of the participants (Bhopal, 2010).

Data analysis
The methodological approach consists in comparing the two studies through qualitative data analysis (QDA) to the 2012 data, and qualitative secondary analysis (QSA) of the 1985 data. In a first coding stage, descriptive codes were established by topic. Thematic coding for the 2012 study was performed by using the qualitative software Dedoose. The 1985 study's materials were already coded by topic. In a second stage, I inductively coded and wrote extensive detailed memos regarding the respondents' perceptions related to fertility decisions for both samples. I primarily drew on fertility intentions and the conditions and obstacles considered in the decision-making process. These questions were asked retrospectively for the respondents who were already parents. Other respondents' characteristics were also considered such as the respondents' and their partners' sociodemographic and labor characteristics and their household division of labor. Before presenting the findings, several limitations of the analysis must be acknowledged. First, the 1985 sample was purposely designed to over-represent cohabiters. The overrepresentation of cohabiters in the older sample might imply a bias toward more progressive respondents compared to their 1980s' peers and, consequently, for the interpretation of results in this study. Second, given the timing of the educational expansion in Spain, specifically, the 1980s and 1990s, the highly-educated respondentsparticularly the womenfrom 1985 represent a more selective sample than in 2012. This implies that the respondents from the 1985 sample might belong to a higher position in the socioeconomic scale compared with the 2012 sample. Finally, the views of men and women on the gender equity level within the couple may differ depending on which partner is asked. In general, men tend to overestimate their contribution at home (Kamo, 2000;Press & Townsley, 1998). This suggests that our male respondents may underestimate the gender-role conflict in their relationships as an obstacle to or condition of having children.

Findings
The results compare the respondents' narratives about fertility decisions through the lens of how related perceptions have changed. Table 2 provides a summary of the obtained results. I first grouped the interviewees' narratives that differ from one another in the two studies and that illustrate the changes posited by the SDT to suggest how some reasoning that was transitioning during the 1980s achieved the status of transitioned by disappearing in the 2010s. I then presented the individual perceptions that show how economic uncertainty and a lack of work flexibility remained largely unchanged over the thirty years between the two studies. Finally, I emphasize how the gender equity conflict remains in transition as after having been initiated in the 1980s, it continues to be unfinished in the 2010s.

Transitioners and transitioned
By comparing the narratives from both generations, two important normative changes are observed in relation to the age norms and partnership formation norms.
As discussed above, the delay in life-cycle transitions (i.e. adulthood, partnership formation, and childbearing) has also generated a shift in the age norms related to what is considered to be the appropriate age to have children. Some respondents from the 1985 study manifested an explicit interest in enjoying their period of being childfree to focus on personal projects or career development or to simply enjoy their lives as a childless couple. These respondents represent the values of individualism and argue for their right to choose their life decisions, which is one feature that characterizes the SDT. Guillermo, for example, was 24 years old in 1985. He and his partner were not planning to have a child until they were in their 30s.
What conditions do you need in order to have a child? Mainly . . . comfort. Having a child would limit us too much, our relationship but also many other aspects of our life, at work, socializing with others. We don't have a deadline, maybe in five years, but first, we want to live some years without children. Also, because we don't want to live for a very long time in the same place or work at the same company . . . and in this sense, a child cramps your style.
The respondents in the 2012 sample also reported the idea of 'enjoying childlessness' before having kids; however, the meaning and age references are different for them. For many respondents, having enjoyed childlessness is a reason for deciding to have a child rather than a reason to postpone parenthood as indicated in the 1985 sample. Cristian was 33 years old, and his partner was 31. They had a one-year-old daughter in 2012.
. . . [I don't miss] going to parties, going out at night, and these types of things . . . We used to do it, but we were already tired of it [by the time we had our daughter] to be honest. We have already spent many years doing it, and we got tired of it.
Guillermo and Cristian's ages are nine years apart -24 and 33 years old, respectivelybut although Guillermo's narrative reflects the necessity of justifying his fertility postponement during his 20s, Cristian's narrative suggests that he and his partner were simply ready to have a child in their early 30s because they had already enjoyed their childlessness during their 20s. This means that the SDT postmaterialist values remain over time; however, due to the shift in age norms, Cristian's generation did not feel the same social pressure to have children in their 20s as Guillermo's generation felt. A second example that illustrates the shift in age norms is social and family pressure. Some interviewees in 1985 express how their parenthood decisions are influenced by the pressure that they feel to have children at a certain age. This is the case of Maria Jesús, a 25-year-old married childless respondent in 1985.

I think [I will have] one [child] and let's see. I don't really like them [children]
, this is true . . . I will have it because . . . it seems that at a certain age, you should have a kid, or I don't know why . . . . But at a certain time, there is this need . . . .
Although the entire sample from 2012 reported their intention to have children someday, they felt relaxed in terms of social or family pressure. Although they declare that their parents may feel sad if they end up not having children, they also state that their decision would be fully respected by their family. For them, the concept of pressure as originating from family or society does not appear in their narratives. In contrast, pressure appears to be related to the feeling of 'running out of time' when, for some respondents, mostly women but also men, the biological clock begins to tick.
The shift in age norms for family formation illustrates the different narratives across the two generations. Some respondents from the older generation initiated fertility postponement for individualistic reasons while feeling family and social pressures for not having children. The newer generation adopted the new age norms after having had sufficient time to enjoy childlessness, and their pressure originates from biological age rather than from family or society.
The second set of reasons expressed by the respondents in 1985 that was missing in the 2012 interviewees' narratives address changes linked to partnership formation. An initial difference is the progressive loss of the legitimation of the institution of marriage.
The number of marriages in Spain has continuously decreased since the 1980s. This process was accompanied by a secularization process. In 1996, 76.7% of marriages were Catholic compared with 37.4% in 2012. The norm for the respondents in 1985 was that marriage should precede the birth of the first child as a necessary condition, even for cohabiting couples. This is the case for Esther, a 23-year-old childless woman who cohabited with her partner, a 27-year-old. Her words also reflect the younger age norms for parenthood in the eighties.
We have concluded that if we are going to get married, it will be by the time we want to have a kid. We know for sure that we want to have children, but I still see myself as very young. I tell him 'let's wait until I am 25 or so', and he respects it . . . . . . then, yes, we will get married because of the child and because I will make my parents happy.
Non-marital cohabitation is now well-accepted in Spanish society. Nearly 40% of the babies in Spain in 2012 were born from non-married mothers whereas this figure was only 4% in 1980(Castro-Martín & Seiz-Puyuelo, 2014. With almost no exceptions, the respondents in 2012 did not consider marriage a condition for having children. Many of the respondents who were already parents were not married or got married after having a baby. For many individuals, marriage means an administrative procedure that provides several practical advantages such as benefits from employers or better conditions for paying taxes.
The second difference in partnership formation and childbearing that emerges in the comparison of the two samples is the presence of numerous shotgun marriages in 1985, that is, hurried marriages that followed an unexpected pregnancy. The normalization of the use of contraceptives and the legitimation of abortion have contributed to individuals' right to choose the timing of their parenthood. There were also several cases of unexpected pregnancies in 2012; however, in the majority of them, the partners were in their early thirties and in a stable relationship. Thus, the implications of an unexpected pregnancy for the two cohorts were diametrically different. Some respondents in 1985 explained their experience as regretting having had to give up many things to start a family. Felix is one example. He and his partner were 24 years old in 1985, and she got pregnant when they were 18. His statement also reflects the individualistic values described above.
Not at the moment. No because she wants to continue studying in the university, she just has started so a child would interrupt [her studies]. I am very happy with my kid, but we also have to live our own life. And having another kid means, at least for another four years, to be very focused on the child without being able to live, socialize with others. This comparison suggests that many of the obstacles and conditions experienced by the 1985 respondents that support the SDT rationale became less significant among the 2012 respondents. The changes in the age and partnership formation norms are the reasons behind this intergenerational change. The 1985 respondents largely represented the SDTthe transitionerswhereas I argue that the 2012 respondents overcame these changes. However, this does not mean that postmaterialist values are no longer important for the younger generation; nevertheless, they are less relevant in explaining fertility behavior due to changes in societal norms.

Economic uncertainty remains unchanged
Despite the passage of three decades, the narratives from both samples shared some perceptions related to fertility. Both studies were conducted in times of economic crisis and a lack of policies that facilitate work-life balance, which were both reflected in the interviewees' narratives.
First, achieving a certain economic stability stood out as the most common obstacle for fertility alluded to by the respondents from both generations. Having a child is perceived as costly and, in some cases, a 'luxury.' Moreover, although the interviewees had relatively high human capital (at least a post-secondary education), the subjective perception of economic insecurity among them was also present in many of their narratives, even when both partners were employed. Perceived economic insecurity affects not only the timing of childbearing but also the number of intended children. In the 1985 sample, Felipe represents this perception. He was a married 37-year-old father of one child.
Yes, we want to have a second child, but so far, the decision depends mainly on the job situation we have gone through, which has influenced us in not having another kid yet, because [having another kid] would have brought economic problems to the family, so it stopped us.
Unemployment in the eighties sample affected 8 out of the 43 couples, and in 2012, it affected 18 out of the 53 couples. Men were more affected in 1985 whereas women were more affected in 2012. Moreover, in 2012, there were more couples in which the female partner out-earns the male partner or is the sole provider. Ruth and her partner, who are childless, are one of these couples. They were 26 and 28 years old, respectively, in 2012 and wanted to become parents.
Right now? My partner's job [is stopping us from having a child]. His salary is too low. I have a contract for four years, and [the salary] is more than what I used to earn before, but it is still peanuts. If he could have a more or less average salary, then maybe . . . [we would go for it]. But he now has a temporary contract, which implies a certain instability . . . yeah . . . now is not a good time.
Second, regardless of the time point, the insufficient intervention of the State in supporting families is a shared perception among our interviewees. Young couples must make arrangements in their daily life to be able to combine work and family, which becomes particularly difficult during the first three years of the child's life. In the eighties, the respondents who had more conservative values opted for a male-breadwinner and female-caregiver model whereas the couples who did not assume these roles postponed their parenthood or dealt with the work-family conflict. In 1985, Menchu was a 24-yearold, married, childless respondent. She envisioned children only when at least one of them had a certain degree of flexibility in their working hours.
Yes, but in a few years . . . My plan is, well, besides finishing my studies and having our own apartment, I would like that at least one of us have a flexible schedule to take care of the child . . . . . . because what I don't see [as] logical is that . . . ok, after the three months of maternity leave, you then have to put your child in daycare at 6 am. It's inhuman! But, well, it's what most people have to do, right? And maybe it's going to be the same for us . . . because if you start working at 8 am and you live in the suburbs, you have to leave home two hours earlier.
Thirty years later, this type of perception remains present in young adults' voices. During the previous three decades, improvements in parental leave and public childcare availability were limited. Paternity leave has consisted of 8 weeks since 2019 whereas it was only two days in 1985 and two weeks in 2012. Maternity leave of 16 weeks has remained stable since 1980. The Spanish parental leave system increases gender inequality by preserving the gender differences associated with gender-role specialization (Lapuerta et al., 2011). After paid allowances for parents, the childcare leave system is generous in its length; however, it does not offer wage replacement. The respondents from the 2012 sample expressed the difficulty of taking unpaid childcare leave and living with only one salary. Paradoxically, the limited availability of public daycare and the high cost of private daycare lead individuals to think in terms similar to Alberto. In 2012, he was 30 years old and his partner was 29. They were parents of a seven-month-old baby.
. . . what they [the government] should do is to allow people to have a child in an easier way. I think that reducing working hours is very important so that the couple can find balanced childcare. But what has to happen above all is the development of public [measures], like public daycare, because there aren't any, and it is so difficult to find a [daycare] vacancy. Look, we have two jobs, but our income is low, and it is not enough to pay 400€ [per month], the cost right now for private daycare, at least here in Madrid, but we don't qualify for aid, and we won't get a spot in public [daycare]. Then, I think that measures focused on this are necessary, and many. Alberto's perception was similar to that of several couples as a situation that may occur for them or around them. Unlike in the 1985 sample, all couples in the 2012 sample were dual-earner couples, the majority of whom did not consider the possibility of one of the partners assuming the full-time role of caregiver. An alternative arrangement for them was to obtain the support of the family network, which in most cases includes grandparents. The comparison of the two generations suggests that despite some of the progress that has occurred in Spanish family and labor-market policies during the last three decades, these changes have not fulfilled young couples' needs for fertility decisions.
However, in the 1980s, partnership formation and transition to parenthood was not as disassociated in time as it might be in the 2010s. Hence, couples that were living independently in the 1985 sample might have already overcome the difficulties of economic uncertainty to get married and have the first child. This condition implies that the comparison between the two generations might be biased by the normative changes relative to family formation.

Gender in transition
As mentioned above, the changes in gender roles, which are closely linked to the value changes associated with the SDT, have not been completely incorporated into individuals' behaviors and public institutions. Although gender egalitarianism has substantially advanced since 1985; it can still be considered to be a process in transition. Two aspects of this unfinished gender-role transition stood out among the respondents as obstacles and conditions for fertility: the changes in females' labor force participation and the persisting gender inequality within the household. These two reasons are present in the narratives of many respondents from 2012; however, they also began to appear in the respondents from 1985 who held a more progressive ideology and who coincided with the couples who were in stable cohabiting relationships at that time (Alabart, Cabré, Domingo, Fabré, & Stolcke, 1988). Thus, the cohabiters from 1985 may be considered to be the forerunners of the value change.
First, females' greater labor force participation coupled with economic uncertainty are the determining factors of the postponement of childbearing and fertility decline in Spain. Two aspects from the eighties' materials are of great interest in this regard. On the one hand, more traditional couples considered females' labor force participation to be a 'hobby' or a small contribution, to which they referred as 'help' to complement the household income (Alabart et al., 1988). For these respondents, female labor force participation was not a priority, and it was not among the reasons that explained their fertility decisions. On the other hand, there were a few progressive couples for whom having a(nother) child was subject to the female partner's job stability. María Luisa was a 21-year-old childless interviewee who was cohabiting with her partner. She had a number of plans for her future, among which becoming a mother would occur after fulfilling other personal goals.
Not now, but in the distant future. I need to do many things until I can say 'Well, this is the right moment to have a child' because a child is a lot . . . you need certain conditions . . . . And, right now, I don't have enough time, and I don't consider myself mature enough . . . I think, at least not in 10 years . . . now I want to finish my studies, I want to study abroad . . . . I want to work and have a good job. And of course, if you have a child, you cannot focus on jobs that require many hours.
However, María Luisa was part of a minority in the sample from 1985. In the comparison of both studies, it becomes clear that placing females' job and career development before parenthood is a very common perception for the respondents in 2012. The labor force participation of women from 25 to 54 years old increased from 35.1% in 1985 to 81.1% in 2012 (OECD Stats); however, job precariousness and unemployment persisted and affected more females than males, which is also reflected in our samples. Manuel and his partner, a childless couple aged 30 and 29 years in 2012, represent one case.
In our case, the most important factor is that my partner wants it [have a child]. I have told her that whenever she wants, whenever she feels ready and considers that it is the right moment. Economically, I think we can afford it now, but it will depend on when she feels fulfilled in other aspects. She wants to wait. Similar to other couples from 2012, Manuel and his partner were also very egalitarian in their distribution of domestic work. However, traditional gender roles were deeply rooted in 1980s Spain. Many women from the 1985 study had a secondary earner role, and the majority of them were responsible for most of the housework and childcare (Alabart et al., 1988). Some females from 1985 quit their jobs to become mothers. This was the case of Mariví, a 31-year-old biologist mother of two children in 1985. In light of the unequal distribution of unpaid work at home, Mariví rejected her husband's suggestion of having a third child.
My husband wouldn't mind [having another child] (She laughs). He thinks three is ok. But of course, he doesn't take care of them! Me? Three? No, no, no, no. Don't even talk about it! I mean, it is great to dedicate my life to them, and I love them [my children], but they drive you crazy, they exhaust you, precisely because I am devoted to them . . . . . . Here is the thing, if I spend half of the day outside and then I spend a little time with them in the evening, then I would probably have a different view. But when it is night, I am fed up with them, and now, I don't see myself with another child . . . forget it! In addition to stay-at-home women such as Mariví, other women from the 1985 sample, in addition to having a job, assumed their roles as primary housekeepers and caregivers with little negotiation with their partners. These women carried the second shift (Hochschild, 1989). The respondents who reported a more equal distribution of labor at home coincided with the cohabiters, i.e. the precursors of the change.
Although more than half of the 2012 sample reported having a relatively genderegalitarian distribution of chores, the reported conflicts in this regard were also a constant in their narratives. Daniel was a 2012 respondent who performed egalitarian behaviors with a passive attitude, which indicated that his partner managed the logistics and decisions of the household. He was 35, and she was 34, and they were the parents of a two-year-old daughter. He knew that to have a second child, he would need to be more involved in the housework and childcare.
Well, the truth is that it [the housework distribution] is a little unbalanced . . . She handles the cooking and meals, and I do more cleaning. And with respect to our daughter . . . we started a dynamic, and now it is complicated to change. And this is the kind of thing that we discuss, and it generates a conflict in order to have a second child . . . . . . we will have to be much more involved to better distribute the tasks [for two kids].
Despite being thirty years apart, the narratives of these two generations corroborate that the gender transition has advanced; however, it is unfinished. These results support the extensive literature that in recent years has focused on the central role of gender equity in the recovery of fertility in very low fertility countries such as Spain (Esping-Andersen & Billari, 2015;Goldscheider et al., 2015;McDonald, 2000aMcDonald, , 2000b.

Discussion
This paper compares the perceptions on fertility decisions in two generations approximately thirty years apart. It is the result of the thrilling opportunity to analyze two independent qualitative studies conducted in 1985 and 2012 in Spain on urban, nativeborn, highly educated young adults. Using a unique research approach, the paper analyzes whether there is a generational change in the perceptions related to fertility decision-making and whether theoretical perspectives are represented in individuals' narratives over time. Therefore, as lenses related to the SDT, economic uncertainty and the gender equity perspective were applied to evaluate the steadiness of fertility perceptions across two generations, both assessed during young adulthood.
Are the current young adults' perceptions of fertility decisions similar to the perceptions of young adults' thirty years earlier? To what extent can the theoretical arguments be identified in the individual narratives? The results show that the postulates of the SDT are, to a substantial extent, present in the reasoning that drove fertility decisions in the 1980s. As Lesthaeghe indicated in 1995, 'the reversibility of these changes seems highly improbable' (Lesthaeghe, 1995, p. 58). However, many of the fertility constraints of the past are no longer present in the narratives of the younger generation. This variation can be explained by the changes in the age norms for childbearing and the partnership formation norms. These changes have allowed for the fact that, although some of the perceptions related to individualism and postmodern values still prevail (i.e. developing careers, enjoying childlessness), they do not exert the same social and normative pressures on young adults at the same ages as in the past. The first hypothesis that the presence of the SDT's associated values in the fertility perceptions of the younger generation will be weaker than in the older generation is confirmed. In addition, the shift in the normative changes related to age and partnership formation have a further derived implication that must be highlighted, that is, the change in perceptions and meaning regarding childlessness and childbearing. The results suggest that perceptions of childlessness are not as stable as might have been expected, which is proved by the rising levels of childlessness in recent times. In this regard, today, the idea of childbearing is more linked to the biological meaning of fertility and less to the social position that the prospect of having children held in the past for the family economy. This point is fundamental for the understanding of the overall meaning of having children over time.
Furthermore, the results show how, despite the passage of time, other factors related to economic uncertainty and the institutional context remain as obstacles to fertility. The interviewees from both studies indicated how persistent economic uncertainty remains an invariable influence on the timing of fertility and, ultimately, on the intended number of desired children. As a consequence, the second hypothesis based on how the narratives from both sets of interviewees refer to economic uncertainty as a reason for limiting or postponing their fertility intentions is also confirmed. These results lead us to consider the mechanisms of how couples cope with economic insecurity differently across generations while the support of the state remains scarce. In this regard, it is necessary to highlight the possible changes that might have arisen over time in the fundamental role of kinship in providing intergenerational social support for families, especially in urban contexts as in the case of this study.
However, the transition in gender-role norms, which is closely linked to the value change of the SDT, exhibited a different pace. An important part of young adults' perceptions on their fertility intentions in the 2012 study is linked to the gender-role conflict derived from the unfinished transition of gender-role norms. Only the respondents in the 1985 study who held more gender-egalitarian values were already experiencing a similar gender-role conflict. The third hypothesis proposing that the respondents from the older generation will express fewer constraints on fertility decisions related to gender inequality than the younger generation due to the secondary role in the labor market for women from the eighties compared with the larger presence of dual-earner couples in the younger cohort is also confirmed. Accordingly, regarding the question of whether the fertility perceptions of young adults have changed across generations, the answer is affirmative, although they have not changed entirely. There is a strong and complex interrelationship among the value change, the economic and institutional contexts, and the unfinished gender-role transition in explaining fertility decisions. Public expenditure in the welfare system, including family policies, is not a government's priority during periods of economic recession, which contributes to a persistent lack of protection in the labor market and a gender-unequal parental leave system. This situation contributes to a lack of labor market flexibility in helping parents face the question of work-life balance, which consequently perpetuates gender inequality in the division of productive and reproductive tasks within households. The deregulation of the labor market and the persistent lack of work-life balance are obstructing the gender-role transition. As a consequence, young, highly educated adults find it difficult to adjust their lives to competent career prospects prior to making fertility decisions.
Despite having selected samples that originated independently for different purposesthe study of cohabitation and low fertility-the richness of the data allowed us to identify how perceptions regarding fertility decisions have transitioned across generations. The implications derived from the limitations of data comparability due to the specificities of the sample might lead one to think that the magnitude of such transition in fertility perceptions could be greater if the cohabiters had not been overrepresented in the 1985 study. The original data could be available upon request at brinton@wjh.harvard.edu. 3. The total sample in the 1985 study was comprised of 144 interviews, from which a subsample of highly educated respondents was selected for this analysis. The total sample in the 2012 study was comprised of more than 400 interviews in five different countries, from which a subsample of partnered respondents in Spain was selected for this analysis.