2.1 Introduction

The family of origin plays an important role in the demographic choices that young adults make. There is a large body of literature linking childhood socio-economic conditions and living arrangements to the timing and type of their demographic choices (e.g. Barber 2001; Kiernan and Hobcraft 1997; McLanahan and Percheski 2008; Sigle-Rushton et al. 2005). Previous studies have shown that young adults from advantaged or high-status families delay their first co-residential union, their first marriage, and the birth of the first child compared to young adults from disadvantaged or low-status families (e.g. Axinn and Thornton 1992; Barber 2001; Dahlberg 2015; South 2001; Wiik 2009). With regard to young adults’ demographic choices, the focus in this chapter is on union formation and dissolution, which is called union dynamics from here onwards. Starting a co-residential union is one of the demographic choices that most young adults make, but the timing, the type of union chosen and also the risk to dissolve a union are socially stratified. It is important to examine to what extent family (dis)advantage influences the union formation and dissolution process of young adults, since it can have potential negative consequences for their subsequent life course. People who enter a co-residential union at an early age have, for example, a higher risk to dissolve this union (Berrington and Diamond 1999). Moreover, previous research shows potential negative consequences of unmarried cohabitation as well; cohabiters enjoy lower health quality, receive fewer social provisions and are also less committed to their relationship, which results in a higher risk to dissolve a union (e.g. Soons and Kalmijn 2009). Finally, existing research also shows that people who dissolve a union can experience many negative consequences, such as lower well-being, economic hardship, and loss of emotional support (Amato 2000). The first research question this chapter attempts to answer is:

To what extent is there a link between parental socio-economic status and union formation and dissolution?

Answering this first research question will increase our understanding of how social inequalities in the family domain are produced and reproduced, providing fresh insights into one of the key questions in social science.

A limitation of most existing research is that it has mostly examined the link between family (dis)advantage and union dynamics within a single societal context. Should we expect that the effect is universal and replicates across countries? Authors have argued that effects differ across countries due to cultural, economic and institutional differences between countries. For example, in societies where the family is more central, the effect of family (dis)advantage can be expected to be stronger than in societies in which individualism plays a greater role (Inglehart 2006). A similar kind of expectation can also be formulated concerning the economic development of a country: in economically weaker societies, family (dis)advantage is expected to have a stronger effect on demographic choices of young adults (Schneider and Hastings 2015). This is because young people in these countries depend more on their parents and their resources. Finally, welfare arrangements may also play a role in explaining cross-national variation. If a society has a non-generous welfare regime, young people are generally more dependent on their parents and their resources, so the influence of family (dis)advantage can be expected to be stronger in these societies (Esping-Andersen 1990; Bäckman 2008).

Whereas the existing literature acknowledges the presence of cross-national differences in the role of demographic choices in producing and reproducing social inequalities, empirical evidence supporting and explaining these differences is scarce. This chapter starts from the assumptions of the Second Demographic Transition (SDT) theory as a key explanation for this cross-national variation (the choice of this theoretical orientation will be extensively elaborated in the next section). Thus, the second research question this chapter addresses is:

To what extent does cross-national variation exist in the link between parental socio-economic status and union formation and dissolution and how can this cross-national variation be explained?

In conclusion, this chapter combines the family context with the societal context, which makes this chapter innovative and relevant. If we understand why family (dis)advantage is more important in some countries than in others, for example, due to differences in cultural norms and values, we have unraveled one piece of the bigger question why the level of social inequality differs considerably across countries. First, this chapter provides a theoretical discussion on mechanisms explaining the association between parental socio-economic status (SES) and union dynamics (union formation and dissolution). Second, based on the SDT theory, this chapter addresses why cross-national variation can be expected with regard to the link between parental SES and union dynamics and how this variation, theoretically, can be explained. Third, a collection of integrated results is shown from existing studies which are recently conducted by the author of this chapter within the Context of Opportunities (CONOPP) project.

2.2 Family of Origin and Adult Family Dynamics

Previous studies have shown that young adults from advantaged families delay their first co-residential union and their first marriage compared to young adults from disadvantaged families (e.g. Axinn and Thornton 1992; Barber 2001; Dahlberg 2015; South 2001; Wiik 2009). But why is a higher parental socio-economic status linked to delays in the demographic choices of their offspring? A first explanation is that the delay is in fact determined by the educational level and enrollment of young adults themselves. Higher-SES parents are likely to have higher educational aspirations for their children than lower-SES parents. As a result, children of advantaged families are motivated to invest more energy and time in their educational career, which often leads them to delay romantic unions or parenthood at young ages (e.g. Axinn and Thornton 1992; South 2001). However, whereas young adults’ own educational achievement acts as an important mediator between parental SES and demographic transitions, many existing studies indicate that there still remains a significant impact of parental SES (e.g. Dahlberg 2015; Mooyaart and Liefbroer 2016; Wiik 2009). Men and women with higher status parents tend to delay demographic transitions to later ages, even if one takes their level of education and actual enrollment in education into account. Explanations for this remaining link between family (dis)advantage and demographic choices are higher standards regarding their future partner or higher consumption aspirations among young adults from advantaged backgrounds (Axinn and Thornton 1992; Easterlin 1980; Oppenheimer 1988). A second explanation could be that high-SES parents socialize their children to enter a romantic union or a marriage at a later age than lower-SES parents (Wiik 2009). Parents want to have a say in the union formation process, since it is one of the most serious decisions young adults face and which can have enduring negative consequences on the further life course if young adults form a union at an early age. Previous research shows, for example, that those who start a co-residential union young have a higher risk to dissolve that union compared with late starters (Lyngstad 2006). A possible explanation why high-SES parents are more successful in persuading their offspring to avoid early entry into a union can be that these parents are more aware of the potential negative consequences of choices made in the early life-course (Wiik 2009).

Parental status influences not only the timing of demographic events, but also the type of demographic choices (e.g. cohabitation, marriage, union dissolution). Previous research shows that parental SES is differently related to cohabitation versus marriage as first union type (Wiik 2009). Given that marriage is less easily reversible than unmarried cohabitation, parents might want to have a higher stake in the timing of their children’s marriage than their cohabitation. Moreover, young adults from high-status families more often choose for cohabitation instead of marriage as first union compared to young adults from low-status families (Wiik 2009; Mooyaart and Liefbroer 2016). A higher level of education has been associated with more liberal attitudes and values with regard to the choice to cohabit (Billari and Liefbroer 2010). High-SES parents are likely to transmit these values to their children, which can result in a higher probability to cohabit for young adults from advantaged backgrounds.

Another demographic choice, namely union dissolution, is also associated with parental SES. A couple of studies have found that adults from high-status families have a higher risk to dissolve their union compared to adults from low-status families (e.g. Lyngstad 2004, 2006; Todesco 2013). This association has been theorized as reflecting class-related socio-cultural factors or financial support from the parents. Sociocultural factors or values linked to higher-class backgrounds are related to a “bourgeois culture” in which divorce is more accepted (Hoem and Hoem 1992; Lyngstad 2006). Just as education is positively related to the approval of cohabitation, it is also shown to be positively related to the approval of divorce (Rijken and Liefbroer 2012). Moreover, high educated parents are in a better situation to financially support their children if they dissolve their union, which could be another explanation for the impact of parental SES on union dissolution.

2.3 Cross-National Variation Explained Through the Second Demographic Transition (SDT) Theory

A limitation of most existing research is that it has mostly examined the link between family (dis)advantage and union dynamics within a single societal context, while this link can be expected to vary between countries, due to cultural, economic and institutional differences between them. In this chapter, the focus will be on the Second Demographic Transition (SDT) theory, first proposed by Lesthaeghe and Van de Kaa in 1986, as a key explanation for this expected cross-national variation. It is called the Second Demographic Transition to mark the distinction with the First Demographic Transition. During the first demographic transition, which began in the early 1800s and continued into the early 1900s in Western industrialized countries, mortality and fertility declined mainly due to industrialization and in particular associated with social and economic development, modernization, improvements in food supply and sanitation.

Since the 1960s/1970s, the SDT started and primary trends of this second transition include delays in fertility and marriage and increases in cohabitation, divorce and non-marital childbearing (McLanahan 2004; Van de Kaa 1987; Zaidi and Morgan 2017). The SDT theory has often been used to describe and explain cross-national variation in family and living arrangements (Lesthaeghe 2010; Sobotka 2008; Van de Kaa 2001). According to SDT theory, the major demographic changes across Europe and North-America (e.g., decline in marriage rate, growth of cohabitation, and postponement of union formation) in the twentieth century are the result of changes in values and attitudes (Lappegård et al. 2014; Lesthaeghe 2010; Lesthaeghe and van de Kaa 1986). Improved living standards, weakened normative regulation, and increased female autonomy have resulted in an increasing demand for self-development, autonomy and individualism (Lesthaeghe 2010; Sobotka 2008; Van de Kaa 1987, 2001). These value changes manifested themselves in various demographic changes, like increased acceptance of cohabitation, below-replacement fertility and rising divorce rates. Moreover, due to these value changes, important socializing institutions, such as the church and the family, have lost some of their grip on their members and wider society (Lesthaeghe 2010; Sobotka 2008). Processes of individualization and secularization imply that individuals enjoy more freedom of choice and attach greater importance to self-fulfillment, self-development and autonomy (Lesthaeghe 2010). Due to this focus on autonomy, young adults may have become less responsive to their parents’ preferences and less dependent on their parents’ resources. It can, therefore, be expected that the impact of family (dis)advantage on their offspring’s union dynamics is weaker in more secularized and individualized societies.

The SDT theory argues that all countries will experience the consequences of growing individualization, secularization and the weakening of family ties, but starting at different points in time and with different speeds of diffusion. Because of these differences in the onset and speed of diffusion of these demographic and value-related changes, countries vary in the extent to which SDT-related values and behaviors have been adopted at a given point in time (Lappegård et al. 2014; Sobotka 2008). Earlier research shows that Sweden and Norway are SDT-forerunners (e.g., high cohabitation and divorce rates and high level of individualistic values), followed by Western, Eastern and Southern European countries (Lesthaeghe 2010; Sobotka 2008). Figures 2.1 and 2.2 show two SDT-indicators for 25 European countries from the ESS (2006), both related to unmarried cohabitation. The proportion of adults who cohabit as their first co-residential union is used as an institutional indicator (Fig. 2.1), while the proportion of people who disapprove of unmarried cohabitation is used as an attitudinal indicator (Fig. 2.2). For both figures, the SDT pattern is clearly visible. In Northern European countries the cohabitation rate is highest (more than 0.80, thus over 80%), followed by Western, Eastern and Southern European countries. Moreover, in Northern European countries, few people disapprove of unmarried cohabitation (less than 10%), while especially in Eastern European countries this proportion is still above 30%.

Fig. 2.1
figure 1

The proportion of adults who cohabit as their first co-residential union, born between 1960 and 1980. (Source: European Social Survey, 3rd wave (2006), own calculation. Cartography: Peter Ekamper/NIDI)

Fig. 2.2
figure 2

The proportion of adults who (strongly) disapprove of unmarried cohabitation. (Source: European Social Survey, 3rd wave (2006), own calculation. Cartography: Peter Ekamper/NIDI)

Both figures show considerable cross-national variation with regard to the demographic changes that all Western countries have experienced. Some countries are further advanced in these demographic changes, as suggested by the SDT theory, than other countries. Because of these country differences with regard to the SDT, the general cross-national hypothesis examined in this chapter is that the impact of family (dis)advantage on young adults’ union dynamics is weaker in countries that are further advanced in the SDT than in countries that are less advanced in the SDT. In more SDT-advanced countries, processes of individualization have progressed, making family ties less important. In countries where the SDT and related individualization processes are more advanced, young adults can become more detached from their disadvantaged family background, and more focused on developing themselves and making their own choices, while in countries where the SDT and individualization are less advanced, young adults are still very dependent on their parents and their preferences and resources. Therefore, a weaker link between family (dis)advantage and union dynamics can be expected in countries where the SDT is more advanced.

2.4 Results on Family (Dis)Advantage, Union Dynamics and Cross-National Differences

To answer the research questions, two large-scale and cross-national comparative datasets are used, namely the third wave of the European Social Survey (ESS 2006) and the first wave of the Generations and Gender Programme (GGP, see for more information Fokkema et al. 2016).

2.4.1 Family (Dis)Advantage and Union Formation

As already shown in Fig. 2.1, large country differences exist in whether the first co-residential union of young adults is a cohabitation or not. In Norway, Sweden and Finland, more than 80% of the adults cohabit as their first co-residential union. In many Eastern European countries (e.g. Russia, Poland, Romania) and also in Portugal, this percentage of cohabitation as first union is between 10% and 30%.

Also, the age at which young adults enter their first co-residential union varies substantially across countries. Table 2.1 shows the median age at which men and women enter their first co-residential union. The median age of entering a first union is 25 years for women in Ireland, while it is just below 21 years for women in Bulgaria. This difference in median age of more than 4 years is also visible among men; the highest median age for men is for Ireland (27.5 years), while the median age for men in Russia is just above 23 years.

Table 2.1 Median age of entering a first co-residential union for men and women for 25 European countries (from Brons et al. 2017)

Brons et al. (2017) examined to what extent the timing and type of union formation depends on parental SES. They focused on the link between parental SES, measured by an index based on information about parental education and occupation, and the timing of first union formation from a cross-national comparative perspective by using the European Social Survey (3rd wave, 2006). Meta-analytical tools were used to first analyze whether there is cross-national variation and if so, in a second step test whether this cross-national variation can be explained by specific country-level indicators.

Figure 2.3 shows the results of a meta-analysis in which for 25 European countries the total effect of parental SES on the timing of first union for women is estimated (see Brons et al. 2017 for results among men). The dotted line represents the overall effect of parental SES on the rate of entry into a first union for all European countries and this overall effect is negative (b = −.171, p < .01), meaning that the higher the SES of parents, the later women enter their first co-residential union. This finding is in line with previous research (e.g. South 2001; Wiik 2009). The delaying effect of parental SES for women is observed in almost all countries, but substantial between-country heterogeneity is also found regarding the link between parental SES and first union formation. The strongest delaying effect is found for Bulgaria (b = −.41, p < .01), while there is no significant delaying effect found for Finland (b = −.05, p = ns). The I2 in Fig. 2.3 shows a percentage of 62.8%, which is above 50%, thus substantial heterogeneity across countries exists (Higgins et al. 2003).

Fig. 2.3
figure 3

Total effect of parental SES on the timing of first union for women in 25 European (and EU associated) countries. Meta-analysis of estimates from discrete-time logistic models (from Brons et al. 2017). (Source: European Social Survey, 3rd wave 2006)

The next step was explaining this cross-national variation in the link between parental SES and union formation. Based on the SDT theory, Brons et al. (2017) hypothesized that the impact of parental SES on union formation will be weaker in countries that are more advanced in the SDT.

An SDT-indicator which might be a possible explanation for the cross-national variation is the country-specific prevalence of cohabitation. Figure 2.4 shows that this SDT-indicator indeed explains some of the cross-national variation in the effects investigated. Differences in the timing of first union between young adults from advantaged and disadvantaged backgrounds are smaller in countries were cohabitation is more common.

Fig. 2.4
figure 4

Association between the total effect of parental SES on the timing of first union for women and the percentage of men and women in a country who cohabit as their first union (from Brons et al. 2017). Note: AT Austria, BE Belgium, BG Bulgaria, CH Switzerland, CY Cyprus, DE Germany, DK Denmark, EE Estonia, ES Spain, FI Finland, FR France, GB United Kingdom, HU Hungary, IE Ireland, LV Latvia, NL Netherlands, NO Norway, PL Poland, PT Portugal, RO Romania, RU Russia, SE Sweden, SI Slovenia, SK Slovakia, UA Ukraine. (Source: European Social Survey, third wave 2006)

Brons et al. (2017) also analyzed, more specifically, how parental SES is related to the entry into cohabitation versus marriage as first co-residential union and to what extent this link varies across countries. Results show that the delaying impact of parental SES on timing of first union is stronger for marriage than for cohabitation. This is in line with the idea that marriage is a stronger commitment than cohabitation, which implies that parents want to be more involved in the marriage than in the cohabitation process. Moreover, between-country heterogeneity in the impact of parental SES is lower for cohabitation than for marriage as first union, but for both union types cross-national variation is found regarding the impact of parental SES.

As mentioned earlier, an individual’s own educational attainment is suggested to be an important mediator in the link between parental SES and the timing of first union. Brons et al. (2017) also examined the role of individual’s own educational attainment and the results still show an overall significant, though somewhat reduced, delaying effect of parental SES on the timing of first union. However, once own education and enrollment are included as mediators, the variation across countries completely disappears (see Brons et al. 2017). Thus, country differences in achieved educational level are also an important explanation for the cross-national variation in the link between parental SES and first union formation.

2.4.2 Family (Dis)Advantage and Union Dissolution

Parental SES affects the union formation process of young adults, but does the socio-economic status of parents also have an impact on the union dissolution process? Regarding union dissolution, many previous studies focused on the intergenerational transmission of divorce, both in single countries and from a cross-national comparative perspective. However, only a few studies analyzed the link between parental SES and the risk to dissolve a union and these studies were all conducted in single countries. Therefore, Brons and Härkönen (2018) focused on the link between parental SES and union dissolution, or more specifically, the risk to dissolve a childbearing union in 17 different European countries using data from the first wave of the GGP. Parental SES was measured by parental education. They studied the dissolution of first childbearing unions, since this indicator is a better measure of family instability than divorce, given the high cohabitation rates in the countries analyzed. Also, in this study, meta-analytical tools were used to first analyze whether there is cross-national variation and if so, whether this cross-national variation can be explained by a country-level indicator.

Brons and Härkönen (2018), first of all, show that already with regard to the prevalence of union dissolution, large country differences are found. The percentage of people that dissolved their childbearing union was highest in Estonia (29.8%) and Russia (28.9%), while in Georgia, Italy and Bulgaria, this percentage was below 10% (see Brons and Härkönen 2018 for percentages for all the countries).

Figure 2.5 shows the results of a meta-analysis in which for 17 European countries the net effect of parental education on union dissolution was estimated (Brons and Härkönen 2018). The dotted line shows that overall, the higher the educational level of parents, the higher the risk to dissolve a childbearing union (OR = 1.12, p < .01). The net effect of parental education means that next to some important mediators, namely individuals’ own educational level, parental separation and the timing of union formation, parental education still influences union dissolution. This positive association between parental education and union dissolution is observed in most of the countries, but substantial between-country heterogeneity is also found (I2 = 74.3%). This is in line with earlier research on the link between own educational attainment and family dissolution, although the variation found appears less strong than the one between own education and family dissolution. Although the size of the relationship between parental education and family dissolution varies considerably across countries, it is generally positive, whereas the educational gradient of family dissolution shows more variation both in size and in sign. The strongest positive effect of parental education is found for Poland and Georgia (OR = 1.25, p < .01), while no significant effect of parental education is found for United Kingdom, France and Estonia.

Fig. 2.5
figure 5

The net association between parental education and family dissolution. Meta-analysis with discrete-time event-history models – odds ratios and 95% confidence intervals are presented (from Brons and Härkönen 2018). Note: Controlled for gender, year childbearing union started, duration, duration squared, parental separation, own education, age at family formation, and married at family formation. (Source: Generations and Gender Programme (GGP), first wave)

Since cross-national variation is observed in the link between parental SES and union dissolution, the next step would be to analyze whether this cross-national variation can be explained by the SDT. Brons and Härkönen (2018) expected that the link between parental SES and union dissolution is weaker in countries that are more advanced in the SDT. An SDT-indicator which might be a possible explanation for the cross-national variation is the crude divorce rate. As expected, Fig. 2.6 shows that the strength of the link between parental SES and union dissolution is indeed weaker in countries where the divorce rate is higher (so in countries that are further advanced in the SDT). Another country-level indicator, namely the generosity of the welfare state, could not explain the cross-national variation in the link between parental SES and union dissolution.

Fig. 2.6
figure 6

The association between the net effect of parental education and union dissolution, and the average crude divorce rate (without Russia as influential case); b = −0.053; p = .033 (from Brons and Härkönen 2018). Note: 1 = old union cohort (1970–1987), 2 = young union cohort (1988–2013). Note: AT Austria, BE Belgium, BG Bulgaria, CZ Czech Republic, EE Estonia, FR France, GE Georgia, HU Hungary, IT Italy, LT Lithuania, NL Netherlands, NO Norway, PL Poland, RO Romania, RU Russia, SE Sweden, UK United Kingdom. (Source: Generations and Gender Programme (GGP), first wave)

2.5 Conclusions and Discussion

Research on the social stratification of union formation and dissolution focused largely on individuals’ own educational attainment and enrollment as important determinants (e.g. Blossfeld and Huinink 1991; Härkönen and Dronkers 2006). The focus of this chapter was on another important determinant of union formation and dissolution, namely parental socio-economic status and analyzed the relationship between parental SES and union dynamics from a cross-national comparative perspective. With regard to the first research question, it can be concluded that next to individuals’ own education, also parental SES is important in explaining the processes of union formation and union dissolution. Young adults from disadvantaged backgrounds, for example, enter their first co-residential union at an earlier age than those from advantaged backgrounds. This delaying impact of parental SES on the timing of union formation is stronger for young adults who marry as their first union than for young adults who cohabit as their first union. Moreover, research presented in this chapter shows that parental SES not only influences union formation, but also union dissolution. Individuals from advantaged backgrounds have a higher risk to dissolve their union than those from disadvantaged backgrounds.

As already mentioned, demographic choices made during young adulthood can have potentially negative consequences for their subsequent life course. Young adults who enter a co-residential union at an early age have, for example, a higher risk to dissolve this union (Berrington and Diamond 1999). This chapter shows that demographic choices that may result in negative consequences later in life are not always more common among young adults from disadvantaged family backgrounds. Rather, the results regarding union dissolution show that adults with high educated parents do not have a lower, but a higher risk to dissolve a union. Previous research shows that people who dissolve their union can experience several negative consequences later in life, such as lower well-being and economic hardship (Amato 2000).

With regard to individuals’ own educational attainment as important mediator, the results show that parental SES still has an influence on union formation and dissolution after taking into account this mediator. Individuals’ own educational attainment only partly explains the link between family (dis)advantage and union dynamics. Thus, next to the intergenerational transmission of education, there are more reasons why children from disadvantaged and advantaged backgrounds behave differently on the partner market. But although the results of this chapter show that family (dis)advantage is an important determinant of union formation and dissolution processes, we still do not know what the exact mechanisms are that play a role in the link between family (dis)advantage and demographic choices. Previous research suggests some important mechanisms that might explain this link, like differences in the socialization of norms and values, in family attitudes or in availability of resources. Do high status parents socialize their children to start, for example, a romantic union at a later age? Do high-SES parents have more liberal attitudes towards union dissolution, which they transmit to their children and result in a higher probability to dissolve a union for young adults from advantaged backgrounds? Or does it have to do with the (financial) resources that parents have and transmit to their offspring? Future empirical endeavors have the important task of answering many follow-up questions related to potential mechanisms explaining the link between family (dis)advantage and union dynamics. Unfortunately, to date, available data is limited in providing country-level information to test specific mechanisms (e.g. information on the transmission of norms, or values and attitudes). Prominent cross-national data platforms such as ESS and GGP may consider (in their future data collection activities) the addition of data facilitating testing of mechanisms related to, for example, the transmission of norms and values regarding cohabitation and union dissolution.

Another innovative aspect of this chapter is the cross-national comparative perspective in the link between family (dis)advantage and union formation and dissolution processes. As already mentioned, it can be expected that this link varies across countries, due to economic, cultural and institutional differences between countries. For example, in more individualistic countries or economically well-developed countries, it can be expected that family (dis)advantage plays a less important role, since young adults are often less dependent on their parents and their resources. This chapter shows that it is indeed important to take into account in which country young adults live when analyzing the link between family (dis)advantage and union dynamics, since this link varies considerably across countries. This finding of cross-national variation is related to the second research question of this chapter, in which the focus is not only on whether there are differences between countries, but also on how this cross-national variation can be explained. In this chapter, the SDT theory is used as the major source of explanation to explain cross-national variation in the relationship between family background and union formation and dissolution processes. Results show that the included SDT-indicators could indeed explain (part of) the considerable cross-national variation in the link between family (dis)advantage and union dynamics. The country-specific cohabitation rate explains part of the cross-national variation in the link between parental SES and union formation and the crude divorce rate explains the cross-national variation in the link between parental education and union dissolution.

However, one of the conclusions of this chapter is that the SDT theory is not the complete explanation for the cross-national variation in the link between family (dis)advantage and union dynamics. On the one hand, the SDT offers a good explanation for part of the analyzed relationships, on the other hand it shows that it is more complicated than just focusing on the country-specific demographic and value changes, resulting from processes of individualization of secularization. Next to these country-level indicators that focus more on the cultural change in norms and values of people in a country, institutional and also economic country-level indicators might also play a role. Next to the cohabitation rate, for example, the educational expansion of a country also offers an important explanation for the link between parental SES and union formation. The SDT theory already suggests that demographic changes are driven by both cultural (values) as well as structural factors (such as the rise of higher education) (Lappegård et al. 2014). More specifically, Lesthaeghe (2010) highlighted change in the educational composition of western societies as a major contributor to the SDT process, but this has not been analyzed yet. Moreover, Mills and Blossfeld (2013) argue, for example, that the degree of economic uncertainty that young adults face when they make demographic choices, such as those related to union formation and dissolution, is also important. It can be expected that the lower the degree of uncertainty, the less young adults depend on their parents. This level of dependence on the family of origin and the uncertainty that young adults face, are linked to the country-specific culture, but next to this also to the economic possibilities and institutional support from the state. In general, SDT critiques (e.g. Mills and Blossfeld 2013; Zaidi and Morgan 2017) have argued that the SDT-theory has ignored the role of domestic path-dependent institutions, like the welfare regime, the employment systems and the educational system. It is likely that cross-national differences in family patterns are at least partially accounted for by differences in these path-dependent institutions. In countries with social-democratic welfare regimes, young adults make the transition to partnership easier than in countries with conservative welfare regimes. Moreover, educational systems differ in the amount of time spent in schools and the link to the labor market. All these factors influence the degree to which young adults face uncertainty and exacerbate inequality by offering more opportunities to young adults from advantaged backgrounds.

The last issue to keep in mind regarding the SDT theory is that this theory is a developmental theory, so an important question is also to know what is happening over time, in addition to the country differences found in this chapter (Thornton 2013). Next to the country differences, it can also be expected that the impact of family (dis)advantage on union dynamics varies between birth cohorts. The change over time in the impact of parental SES on union dynamics is also analyzed within the CONOPP project, but the impact of parental SES did not change that much over time. Moreover, most country-level indicators did not go that far back in time, which makes it harder to analyze the change over time. In general, the results suggest that the differences between countries in the link between family (dis)advantage and union dynamics seem stronger than the differences in this link within countries over time. However, it would be interesting for future research to also include the temporal dimension, next to the cross-national dimension, especially if more data over a larger time span become available. There is a clear need for more longitudinal analyses on this topic, so that it is possible to better disentangle the temporal and cross-national dimensions.