In the following sections, I first discuss compositional shifts in Sweden, such as changes in childhood living standards, educational attainment, and family demographic behaviour. Next, I conceptualize family life course change and, based on the discussion of compositional shifts in family formation, argue that life courses have become more standardized as delayed entry into parenthood within cohabitation has become a new dominant life course pattern. In a third section, I briefly sketch theories that link individuals’ social background, educational attainment, and childhood family structure to their family demographic behaviour at the micro-level. For example, numerous theoretical approaches associate high socio-economic background, educational attainment, and labour market attachment with delayed marriage and parenthood. Finally, I link compositional shifts to change in family life courses. Specifically, I argue that early family life courses have become more standardized as childhood standards of living as well as levels of educational attainment and labour market attachment have increased. However, I also highlight arguments that changes in childhood family structure may have polarized early family life courses, leading to lower levels of life course standardization.
The Swedish Context from 1960–2010
Swedish labour market institutions and family policies have traditionally aimed at securing individuals from unemployment as well as maintaining high female labour market participation and high fertility rates. However, Swedish society and its institutions have undergone great changes since the 1960s. An overview of education, family, and labour market statistics associated with these changes is displayed in Table 1. As in most European countries, the second half of the twentieth century is associated with a considerable expansion of the tertiary education system. Only four tertiary education institutions existed in Sweden in 1960. However, 17 new universities were founded following a reform in 1974 (Andersson et al. 2004). Although university enrolment rates grew relatively slowly at first, between 1980 and 2010 the percentage of the population with a tertiary education grew from 14.4 to 25.2%.
The Swedish labour market and labour market institutions were surprisingly stable between 1960 and 1990. Labour market participation rates remained high and increased, and unemployment remained under 3%. Further, real wages rose consistently and income inequality decreased. These developments have been largely attributed to highly centralised wage bargaining, high union density, and extensive activating labour market policies (Edin and Topel 1997; Van Winkle and Fasang 2017). However, the Swedish labour market was hit by a major recession in the 1990s, which increased pressure for labour market reforms. Unemployment and income inequality increased as a result of the recession. Nonetheless, absolute incomes increased from 157,800 Swedish Krona (SEK) in 1990 to 265,500 SEK in 2010, partially due to a more educated workforce.Footnote 1
Swedish family policy is perceived to be among the most family friendly. Since 1963 Swedish mothers benefited from more than 26 weeks of paid and job protected maternity leave. In 1974 maternity leave was replaced by an extensive parental leave system, which by 1980 was expanded to one year of paid and job protected leave to be split between mothers and fathers (Ifo-Insitute 2015). Further, a 10-day paternity leave was introduced in 1980. Since the mid-1990s, publicly subsidised early and preschool childcare became widely available (Garrouste 2010). Nonetheless, fertility and marriage became less common across the late twentieth century, accompanied by increases in divorce and nonmarital childbirth. The total fertility rate dropped from 2.17 in 1960 to 1.55 in 2000. Further, the crude marriage rate decreased from 6.7 in 1960 to 4.5 in 2000. Following the introduction of no-fault divorce legislation in 1974, the crude divorce rate doubled from 1.2 in 1960 to 2.4 in 1980. However, both marriage and divorce rates stabilized following the 1980s to roughly 4.5 and 2.4, respectively. Despite relatively stable marriage rates, nonmarital childbirths continued to increase. By the year 2000, over half of all live births occurred outside of marriage, although the vast majority to cohabiting couples.
Conceptualizing Early Family Life Course Standardization
Age 16–35 is thought to be the most demographically dense phase in the life course. Cook and Furstenburg (2002) demonstrated that despite cross-national differences in the timing and order of life course events, most individuals have completed full-time education, entered full-time employment, founded an independent household, and have married and entered parenthood by age 35. The timing of marriage and parenthood has traditionally been the focus of demographic and sociological research on early family life courses (Hogan and Astone 1986). However, more recent research has focused on other early life course events that are associated with marriage and parenthood, such as parental home leaving, cohabitation, and divorce (Buchmann and Kriesi 2011). This is important, because it has enabled scholars to study more comprehensive life course patterns across socio-historic contexts.
The extent that family life courses vary within populations is affected by three factors: variation in (1) the occurrence of family life course states, (2) the timing of transitions between two states, and (3) the order of states in the life course (Huinink 2013). For example, early family life courses in the early twentieth century are thought to have varied only slightly (Mayer 2004). Early family life courses consisted of two tightly coupled events, marriage and parenthood. However, early family life courses in the late twentieth century are thought to vary to a much greater extent. The number of family states has increased, and the timing and order of transitions has changed. For example, individuals tend to enter cohabitation at relatively young ages, while the transition to parenthood often occurs much later. Further, marriage no longer precedes parenthood, but is often its antecedent. Therefore, it is important to study family life courses holistically as “process outcomes” (Abbott 2005), because life courses vary in the occurrence, timing, and order of states and transitions as well as the duration spent within states. Process outcomes, such as family trajectories, can be conceived as the result of numerous “point-in-time outcomes”, such as the timing of first and subsequent births.
Recent literature contends that family life courses variation has increased, although the empirical evidence is mixed (Van Winkle 2018). Terms denoting this increase in variation are used ambiguously (Brückner and Mayer 2005). De-standardization is most common and signifies an increase in life course patterns that diverge from early marriage and parenthood (Brüderl 2004) or an increase in family formation dissimilarity relative to early marriage and parenthood (Hofäcker and Chaloupková 2014). However, an increase in family patterns that differ from trajectories of early marriage and parenthood is better described as de-traditionalization or de-institutionalization (Kohli 1985). Rather than a shift away from traditional life course patterns, Brückner and Mayer (2005) define de-standardization as a process in which patterns characterize smaller portions of a population or in which life course events occur at more dispersed ages. Therefore, standardization is a process where life course patterns characterize larger proportions of the population.
It is most commonly assumed that early family life courses have become less standardized across birth cohorts (H1) (see column 1 of Table 2). This is generally founded on evidence that early marriage followed by parenthood has become less common and has been replaced by diverse life course patterns (Corijn and Klijzing 2001). Next to lower rates of marriage and parenthood as well as postponed entry into marriage and parenthood, higher rates of cohabitation, divorce, and nonmarital childbirth are expected to decrease standardization levels.
Family life courses may become more standardized if change is a transitory process where one dominant life course patterns replaces another (Huinink 2013). Indeed, the compositional shifts discussed in the previous section suggest that a shift in family life courses has occurred. Rather than early marriage followed by parenthood, the majority of individuals now enter parenthood within cohabiting unions at later ages (Baizán et al. 2004; Holland 2013). Moreover, divorce and marriage rates in Sweden stabilized in the 1980s rather than a continuation of declining trends in marriage and increasing trends in divorce (Ohlsson-Wijk 2011). Therefore, I find it more plausible that early family life courses have become more standardized across birth cohorts (H2) following a relatively universal shift in the timing and ordering of entering nonmarital cohabitation, parenthood, and marriage.
Theories on Family Life Course Change
To develop hypotheses about which compositional shifts induced a standardization or de-standardization in early family life courses, it is necessary to review theories that link individual characteristics to family formation. Sociologists and demographers have developed numerous theoretical approaches to account for changes in the prevalence, timing, and ordering of family events. Generally, these can be subsumed under four broad theoretical frameworks: an ideational, a rational choice, a structural constraint, and a socialization framework. The theoretical approaches within these frameworks commonly link individuals’ childhood standard of living, educational attainment, labour market attachment, and childhood family structure to their future family demographic behaviour.
The second demographic transition (SDT) thesis, one of the most prominent explanations within the ideational framework, has a long history of relating cultural change to a de-standardization of family life courses (see Zaidi and Morgan 2017 for a review). The core argument of the SDT thesis postulates that as childhood standards of living increase, individual value orientations will shift to prioritize self-actualization rather than the long-term commitments of early marriage and parenthood (Lesthaeghe 2010, 2014). This in turn initiates an irreversible increase in the mean age at marriage and parenthood, a lower propensity to marry and enter parenthood, as well as a higher prevalence of singlehood and cohabitation, divorce, and nonmarital childbirth.
Other theoretical explanations within the ideational framework concentrate on the relationship between educational attainment, gender norms and family formation (Goldscheider 2000; e.g. McDonald 2000). Recently, Goldscheider et al. (2015) predicted that fertility rates and marital stability will increase when gender-egalitarian norms become more widespread and men’s gender roles adapt to women’s. Esping-Andersen and Billari (2015) maintain that educational attainment is central to the spread of gender egalitarianism and the recuperation of fertility.
Prominent theoretical explanations within the rational choice framework also link childhood living standards and educational attainment to family formation. Family demographic decisions are described as utility maximization processes, where the utility of an event is a function of (opportunity) costs and benefits (e.g. Becker 1974a; Becker et al. 1977; Becker and Tomes 1994). Opportunity costs partially result from foregone income, rise with increases in an individual’s educational attainment and labour market experience. Higher educational attainment and labour market attachment should then lead to delayed or forgone marriage and parenthood. However, the opportunity costs associated with marriage and especially parenthood are highly context-contingent. In the Swedish context, with gender-egalitarian parental leave schemes and widespread public childcare, the opportunity costs of childbearing are likely much lower than in contexts with lower levels of public family support.
Other rational choice theorists have used a market framework to link living standards and educational attainment to the timing of marriage and parenthood. For instance, Easterlin (1975, 1976) predicted that couples will enter parenthood only after they had achieved their aspired standard of living developed during childhood. Oppenheimer (1988) used partner markets to account for variation in the timing of marriage. Specifically, she maintains that individuals spend time on the marriage market to find an acceptable match, which is partially a function of socioeconomic background. Men and women use the information currently available, e.g. educational attainment and labour market experience, to estimate the potential socioeconomic attainment of a possible partner. Further, she suggests that premarital cohabitation represents a cost-effective strategy to prolong the search for a future partner.
The structural constraint framework links globalization and deindustrialization to change in family life courses (Mills and Blossfeld 2003, 2005, 2013). Importantly, this framework suggests that the mechanisms described by Easterlin (1975, 1976) and Oppenheimer (1988) will strengthen in times of increasing economic uncertainty. Young adults facing more precarious labour markets will tend to delay or forgo marriage and parenthood, and resort to other family life course models associated with fewer and shorter-term commitments, such as cohabitation. However, economic uncertainty will not affect all members equally, but will vary by socioeconomic background. Parents with high incomes and educations will invest in their children’s educational attainment and support their labour market entry to diminish the risk of downward social mobility (Breen and Goldthorpe 1997; see also Bernardi and Grätz 2015; Erola and Kilpi-Jakonen 2017), which may lead to delayed marriage and parenthood. In sum, the economic uncertainty approach relates socioeconomic background to the timing and propensity of marriage and parenthood, but also to the occurrence of premarital cohabitation.
Finally, McLanahan’s (2004) diverging destinies hypothesis has brought more attention to the relationship between socialization and family formation. She contends that the socioeconomic and family outcomes of children from single-parent households are becoming increasingly less favourable compared to the outcomes of children from two-parent households (see also Martin 2000; Raver et al. 2015; Schoon 2015). Further, she maintains that this difference cannot be explained by socioeconomic differences between single- and two-parent households alone. The transmission of parents’ values and preferences, their ability to supervise and control their children’s activities, and stress generated by family instability are seen as important mechanisms associated with the intergenerational transmission family demographic behaviour (see Teachman 2003). While children from single-parent families are more likely to enter parenthood early and become single parents themselves, the children of two-parent families will delay parenthood until marriage.
Compositional Shifts and Family Life Course Change
In the previous section, I described prominent theoretical approaches that link individual characteristics to the prevalence, timing, and ordering of family events. The theoretical frameworks and explanations discussed above are commonly depicted as separate approaches in a clear-cut manner; however, they may interact to generate population-level change in family life courses. In the following discussion, I will concentrate on how compositional shifts in childhood standards of living, educational attainment, labour market attachment, and childhood family structure across cohorts may stimulate change in family life courses through different theoretical pathways.
Higher childhood living standards and parental resources are of central importance to the ideational, rational choice, and structural constraints frameworks. In particular, the SDT thesis, Easterlin’s (1975, 1976) theory on fertility, Oppenheimer’s (1988) theory on marriage, and the structural constraints approach predict that higher parental education and incomes will induce more delayed entry into marriage and parenthood as well as lower rates of marriage and parenthood in times of economic uncertainty. Further, individuals from advantaged backgrounds will be more likely to postpone marriage and parenthood to pursue more individualistic life styles. An increased prevalence of premarital cohabitation may result from both a cost-effective strategy to prolong the search for an optimal match on the marriage market and as a new possible step in an individualistic life course.
A higher educated population has implications for gender-ideology theories within the ideational framework and utility theories within the rational choice framework. The opportunity costs related to marriage and parenthood increase as individuals, especially women, become more educated (Becker et al. 1977). Increased educational attainment is also associated with higher labour market attachment, which may additionally increase the opportunity costs of marriage and parenthood (e.g. Becker and Tomes 1994). A more educated population is more likely to hold gender-egalitarian norms, which increase gender equity within the private sphere (Esping-Andersen and Billari 2015; Goldscheider et al. 2015). As a result, marriage and fertility rates should stabilize, although individuals will likely continue to enter marriage and parenthood at later ages.
In sum, I expect that higher levels of parental income and education as well as educational attainment and labour market experience are associated with higher levels of early family life course standardization (H2A, H2B, and H2C). Increased standards of living, educational attainment, and labour market participation will delay entry into marriage and parenthood as well as increase cohabitation rates as individuals pursue more individualistic life styles, spend more time to establish themselves on the labour market and find an adequate partner, and fulfill their consumption aspirations to enter parenthood.
McLanahan’s (2004) diverging destinies framework stands out as the only approach that links a dramatic compositional shift, i.e. the number of single-parent households, with a de-standardization of family life courses. Specifically, a polarization in the transmission of family demographic behaviour between single-parent and two-parent families may lead to lower levels of life course standardization. While young adults from two-parent households enter parenthood within cohabitation at later ages, young adults from single-parent households continue to enter marriage and parenthood at early ages. This process may also be reinforced by the associations between childhood living standards and educational attainment discussed above: Single-parent households are on average less educated and have lower household incomes than two-parent families (Mood and Jonsson 2016). Further, children of single-parent families attain a lower level of education than children from two-parent families (Björklund et al. 2007). Both factors may additionally facilitate an earlier entry into marriage and parenthood (cf. Easterlin 1975, 1976; Oppenheimer 1988). In sum, higher levels of single-parent families during childhood are associated with lower levels of early family life course standardization (H1D).
All the theoretical perspectives discussed above have one thing in common: they expect stronger associations for women compared to men. The frameworks proposed by Becker (1973, 1981) as well as Esping-Andersen and Billari (2015) explicitly state that increases in women’s educational attainment is the impetus for change in family life courses. The increase in educational attainment and labour market participation across birth cohorts in Sweden has been most pronounced for women. With regard to McLanahan’s (2004) diverging destinies framework, women are at a higher risk of becoming single custodial parents through lone birth or divorce. Therefore, parents may be especially interested in investing in their daughter’s educational and labour market attainment as well as influencing her partner choice. I therefore expect that the associations between changing levels of educational attainment, labour market participation, childhood family structure, and parental resources with changing levels of early family life course standardization are stronger for women (H3).