Using data from the 1990 and 2017 European Values Studies, this study investigates the link between values and living arrangements among young adults from France, Germany and the Netherlands. Researching the diversity in transitions between living arrangements of young adults gained momentum in the 1990s. The logic to link these transitions to values can be traced back to the Second Demographic Transition argument that states that demographic transitions follow generational changes in values that reflect increasing orientations to autonomy. Almost thirty years later, we are able to update this research by looking to what extent today’s generation of young adults – which is in fact the generation of children of the young adults in 1999 – compare with their parents as far as the values profiles of living arrangements is concerned. The results show that to a large extent, the values profiles of living arrangements are similar across the two generations of young adults or even became more articulated when values were related to topics that can be linked to the religious sphere.
The diversity in transitions between living arrangements of young adults that we experience today is the outcome of a process that started after the Second World War. Scholars started to recognize the complexity of transitions in the late 1980s and the early 1990s (Bumpass, 1990; Rindfuss & VandenHeuvel, 1990; Lesthaeghe, Moors & Halman, 1992; Lesthaeghe & Moors, 1996). The traditional sequence of experiences by which young adults leave the parental home after getting married and then become parents themselves was clearly no longer the only pathway that was followed. Intermediate states such as premarital cohabitation and parenthood outside of marriages had emerged and, in some cases, consolidated in end-states. Today hardly anyone is surprised by the finding that diversity in pathways leaving the parental home has become the rule.
Up until the 90s, explanations for the processes involved were predominantly linked to socio-economic theories of demographic transitions. With the rise of large-scale survey research since the 1980s – of which the European Values Studies (EVS) played a pioneering role – opportunities to study the role of cultural factors in shaping the living arrangement transitions sharply improved. Upon my initial study in the 1990s, Ron Lesthaeghe invited me to join him in a trip to ‘Mister European Values Studies’; the trip was to Tilburg, and ‘Mister EVS’ appeared to be Loek Halman. Present chapter is an update of our initial work that was the product of that first meeting three decades ago. Also present study departs from the idea that cultural factors alongside socio-economic dynamics define demographic transitions: this thesis is at the hard of the the Second Demographic Transition (SDT) theory. In a nutshell, this theory claims that generational changes in demographic transitions follow generational changes in values that reflect increasingly orientation towards autonomy.
In this present chapter, I look back at these alleged cultural profiles of young adults in 1990 and make a comparison with the cultural profiles of young adults in 2017 using data from the EVS in the Netherlands, France, and Germany. Young adults are selected within the age groups of 20-24 and 25-29. Given that there is 27 years in between the two selected EVS waves, the young adults from 2017 can be considered to be the generation representing the children of the generation of young adults from 1990. This opens up a rather interesting perspective since I will investigate how the linkage between values and demographic transitions of the 1990 generation is reconfirmed in the generation of their ‘children’ in 2017, thereby bridging generations.
This chapter is primarily an exploratory study. The starting point is to find out to what extent the linkage between living arrangements and values of young adults that was documented in the 1990s (Lesthaeghe, Moors & Halman, 1992) is transmitted to the next generation of young adults in 2017. In this section, I briefly present the original theoretical perspective that yielded the argument that cultural factors are essential factors in understanding early adulthood transitions. Subsequently, I reflect on how the processes initiating this relationship of cultural and demographic transition in the 1990s may still play a role in shaping this relationship almost three decades later.
SDT was first introduced in the late 1980s (Lesthaeghe & Van de Kaa, 1986; see also: Batool & Morgan, 2017) and focuses on providing a framework to ‘predict’ demographic changes towards low fertility and increasing diversity of union formation. The SDT perspective is inclusive in the sense that it recognizes the contribution of socio-economic types of explanations as provided by, for instance, neo-classic economic theory (e.g., Becker, 1981) or social-deprivation type of theories (e.g., Easterlin, Macdonald & Macunovich, 1990). What SDT emphasizes is that values reflecting autonomy and self-actualization constitute independent prime drivers of demographic changes. Essential assumption is that these values themselves are rooted within the – proverbially -- generational DNA. Applied to this study, this implies that researching value differences between different living arrangements of young adults should reflect differences in autonomy and self-actualization. A drawback of this study is that values are measured after the transition is made and not prior to the transition. This is inevitable with cross-sectional data. However, using panel data it has been demonstrated that values do predict future demographic transitions and that demographic transitions in turn reinforce values consistency with the choices made (Moors, 2000; 2003a). As such, cross-sectional findings provide a kind of footprint of the reciprocal connection of values and transitions.
In this study, I compare differences in values of young adults in four different types of living arrangement, namely living with parents, single, cohabiting or married. I define two age-groups, namely 20-24 and 25-29 years old. For the youngest age group, the three states of living independently from one’s parents indicate early transitions. For the older group these states include both earlier (before the age of 25) and later (at the age of 25 or later) transitions. Especially, as far as the situation of married persons is concerned, I expect that early marriages will differ in values compared to the other living arrangements. Hence, this expectation presents the need to check interaction effects of living arrangements and age groups on values.
SDT has gained increasing attraction during the 1990s (Billari & Liefbroer, 2004) and continued to do so this century (Zaidi & Morgan, 2017). It was not left uncriticized (see Zaidi & Morgan, 2017 for an overview) in that period. One of the critical questions was whether the observed changes are truly complete and irreversible. This critical question directly links to a key research goal of this study: are the initial findings of the relationship between values and living arrangements observed amongst young adults in 1990 also observed among the young adults of 2017? Also here, interaction effects need to be studied, i.e. the interaction between year of survey and living arrangements in the effect on values.
The data used in this research stem from the European Values Surveys of 1990 and 2017. In the reference study (Lesthaeghe, Moors & Halman, 1992) the data from the Netherlands, Germany, Belgium and France were selected because these countries exhibit similar developments and could for that reason be pooled. In 2017, however, no data were collected for Belgium, implying that I only focus on the Netherlands, Germany and France. All the analyses are pooled data of the three countries to overcome small sample sizes within each living arrangements state and survey year.
As indicated before, our analysis is restricted to include only respondents in the age range of 20 thru 29 and respondents that could be allocated to one of four living arrangements: living with parents, living in a single household, cohabitation and marriage. Sample size information split by living arrangement and survey wave is presented in Table 18.1. From that table it can be read that the increasing popularity of cohabitation and singlehood at the expense of marriage has continued to increase between 1990 and 2017.
Table 18.1 Living arrangements (LA) by year of survey (YE)
single living with parent
single not living with parents
with partner cohabiting
with partner married
When analysing secondary data, researchers are always constrained by the initial design of the survey. In this, the questionnaires of the EVS have changed in time. A particular drawback relevant to this study is that the EVS only allows to snapshot current state of living arrangements without differentiating into previous states. The limitation present is that categories of living arrangements are more heterogenous in composition than is preferred. For instance, in the case of marriage, it might be relevant to know whether marriage was preceded by periods of cohabitation or living in a single household. This problem is partially covered by distinguishing between two age groups, i.e. respectively the 20-24 and 25-29 year olds. In case of the youngest age group, the category of married respondents is much more likely to be homogenous than in the case of the oldest age group. After all, immediate marriages, i.e. people who leave the parental home for marriage, correlates with early marriages, i.e. marriage at younger ages of young adulthood.
Instead of focusing on scales, I make use of the individual items. The main reason is that if a systematic pattern would be observed across several items, the mechanism that links values to living arrangements is indicative of an important claim of the SDT, namely that living arrangements and values are systematically and consistently linked to one another. The items selected are in the domains of religion, marriage, children, morality and post-materialism. The latter is included because SDT explicitly refers to the concept of post-materialism (Lesthaeghe & Moors, 1996). The other items hold a normative component towards individual behaviour in the religious, public and personal sphere. SDT claims that values that reflect autonomy or self-actualization, will result in choosing ‘new’ intermediate and/or end-states after leaving the parental home: cohabitation and/or living single.
I used LatentGold (version 6, see https://www.statisticalinnovations.com) to run ordered logit regression analyses. Deviation coding was used. Effect sizes are log odds of the given category relative to the average log odds. Positive estimates thus indicate higher than average scoring and negative estimates lower than average. First, I estimated the main effect models only in which each item is regressed on living arrangement (LA), age group (YA), and year of survey (YE). Gender, education, occupational status, and country are included as additional covariates (control variables). Second, I added the two interaction effects of interest to this study, i.e. interaction of LA with YA and interaction of LA with YE. Results are presented in figures to facilitate interpretations. Details about operationalizations and full tables with estimated effect sizes are available on request.
In this section, I present results grouped by theme: religion, marriage, children, morality and post-materialism. It is important to realize that the effects presented in all the figures are significant net effects controlling for all other covariates.
The religious foundation of marriage as opposed to cohabitation is clearly visible in all four religious items (see Figure 18.1). The category of single persons falls in between both. The largest difference between married and cohabiting people is observed in religious practice but also the believe in Christian religious symbols, i.e. God, Heaven and Hell, is clearly profiled. The finding is consistent with the SDT argument that secularization in religious beliefs is associated with a lower likelihood of marriage.
In Figure 18.2, I bring together the results of two items expressing traditional views on what is important to marriage, i.e. faithfulness and children; and two items that reveal weakening ties, i.e. evaluating marriage as outdated and justifying divorce. The general picture is that the profile of married persons fits the expectation that they would score highest on faithfulness and importance of children to marriage and lowest on finding marriage outdated – in this case obviously reinforced by the very fact that they are married – as well as lowest in justifying divorce. Although cohabiting people take an opposite view, it is the category of single households that contrasts most with married people on the marriage issues of faithfulness and children. Their autonomy desire is highest on these matters.
The items selected from the questionnaire of what people find important for children reflect the autonomy versus conformity distinction that plays a major role in the SDT theory. Typical of the category of married respondents (Figure 18.3) is that they rate good manners and thrift higher than cohabiting and single living persons. At the same time married people score less on the qualities of imagination and independence. In the latter case differences are small though. The general picture again confirms what was expected from the SDT framework.
The morality questionnaire gauges what actions can be justified and hence indicates the amount of autonomy people should have in making their own choices. Two items indicate civil morality: ‘claiming state benefits one is not entitled to’ and ‘cheating on tax’. Justification of ‘homosexuality’ and ‘prostitution’ represent sexual morality. Ethical morality is linked to issues of interference in the lives of people, i.e. justification of ‘abortion’ and ‘euthanasia’ (Figure 18.4). On almost all morality issues, single and cohabiting people are at the same high level of justifying the acts. This is somewhat less the case with the issue of ‘claiming state benefits’ that is less justified amongst cohabitants compared to singles. This is in sharp contrast with married people who clearly tend much more not to justify all listed behaviours.
The original post-materialism index (Inglehart, 1990) contrasts a post-materialist item preference of ‘giving people more say’ and ‘freedom of speech’ with the materialist combination of preferring ‘fight rising prices’ and ‘maintaining order in the nation’. The post-materialism theory perfectly fits in the SDT framework since it claims that post-materialist preferences gradually replace materialist preferences by the succession of generation: younger generations freed from materialistic concerns shift their priorities to autonomy and self-expression. The same mechanism is defined within the SDT framework: young adults in the 1980s and 1990s have shifted towards more autonomous pathways out of the parental home compared to their own parents. In later research on post-materialism, it was found that the index could be unfolded into two subindices (Moors, 2003b) of which the subindex ‘giving people more say versus maintaining order’ was the most clearly linked to generations. This was much less the case with the subindex ‘freedom of speech versus fight rising prices’. In Figure 18.5, I present the results from the full index and the ‘giving people more say versus maintaining order in the nation’ subscale. The relationship of living arrangements with the second subdimension was not significant in any of our models. As such this is an important finding since it does indicate that the autonomy-autocracy distinction reflected in the ‘more say – maintaining order’ contrast triggers choices regarding living arrangements.
Results in Figure 18.5 are perfectly in line with expectations: married people score lowest on post-materialism and the ‘more say – maintaining order’ subindex in contrast with cohabiting and single household persons. I like to underscore the significance of this finding since it has been demonstrated that the generational profile of post-materialism (Inglehart, 1990) and the ‘more say – maintaining order’ subindex (Moors, 2003b) are indeed generational profiles and not profiles that change with age.
One of the most important puzzles of this article is whether values differences in living arrangements in 1990 could be reproduced in 2017. For nine out of the selected 20 items, I found significant interaction effects between year of survey and living arrangements. The first conclusion is that in more than half of the items the results presented in the previous sections apply for both survey years indicating consolidation of the associations between values and living arrangements.
When significant interaction effects were found, I ran separate analyses for each survey year using the same model as the pooled main effect approach in previous sections. This way, estimates of living arrangements are estimates relative to the specific average for each survey year. In four of the nine cases, I found that the 2017 differences were no longer significant: ‘importance of thrift for children’, ‘justification of cheating on tax’, and both post-materialism indices. As far as the post-materialism indicators is concerned, the differences between living arrangements were still in line with the pooled result picture, but no longer significant. This was not the case with the other two items. For five items I observed that the 2017 data showed an increasing differentiation in values amongst the categories of living arrangements as is shown in Figure 18.6.
Three religious items, i.e. believe in God, Heaven and Hell, and two morality items, i.e. justification of homosexuality and abortion, reveal the same pattern by which the married persons in 2017 even more strongly separate from cohabiting and single living persons than in 1990.
The motivation to estimate the interaction effect between age groups and living arrangements on values flowed from the expectation that, for both age groups, the composition of the living arrangements might differ in the amount of heterogeneity. This expectation results from the diagnosis that there is no information on the history of living arrangements available in the data – only information of current status is present. In merely five out of twenty items, I found significant interaction effects. How the categories of living arrangements differed across age groups on these items is presented in Figure 18.7.
In three of the five items it is especially the category of married people that differs in values depending on the age group they belong, i.e. believe in God and justification of abortion and euthanasia. Each time the difference between married people and other categories becomes smaller in the 25-29 age group compared to the 20-24 group. This is in line with the argument that in the age group of 20-24, the proportion of immediate marriages (without prior transition to intermediate states) is most likely higher than in the 25-29 group that combines immediate marriages with later marriages. Typical of later marriages is that intermediate states between leaving the parental home and the marriage are more likely to have occurred. I do not want to overly underscore this finding since for most items, I did not find significant interaction effects of age groups with living arrangements.
A unique finding, not replicated in any of the other items, is how the cohabiting people and those who live with their parents position themselves on the topic of faithfulness in marriage. In the 20-24 age group, they align in between married and single persons. If one, however, is (still) living with their parents at the age of 25 or later, than faithfulness is found even slightly higher than married people. Cohabitants move in the other direction with cohabitants in the 25-29 group almost equally low on faithfulness than their single living counterparts.
Finally, the significant interaction effect between age groups and living arrangement in the case of justification of ‘claiming state benefits’ is due to how the two categories without a partner differ. In the youngest age group, they share the same average level of justification; however in the oldest age group, the single persons are much more willing to justify non-entitled claims of state benefits whereas the single persons (still) living in the parental home are much less willing to do so.
Nowadays, most people would probably ask ‘why would choices regarding living arrangements not be driven by values people hold?’. Of course, the same people would recognize the socio-economic forces and restrains involved in the process. More than three decades ago, however, the question was rather ‘to what extent do values play a role in socio-demographic transitions?’. At the very heart of the Second Demographic Transition framework lies the proposition that values do have their own and independent role to play and that this mechanism would continue to be observed. This research contributes to this line of thinking by studying whether patterns of association between values and living arrangements among young adults in the 1990s re-emerged almost three decades later in the generation of 2017. My conclusion is: it does to a large extent. In case of the religious-morality domain the profiling of living arrangement became even somewhat more pronounced. In a few items that were somewhat more linked to the political-materialist domain the differences between living arrangements decreased and became not significant.
Of course, this study has limitations that are linked to doing secondary data analysis on cross-sectional data. The EVS was not particularly designed to study transitions in living arrangements among young adults. It does not use a life-history design which limited this research to focusing on current states at the time of the survey. Questionnaires themselves changed over time. Constructing comparable measures of occupational status of the household or even parenthood in the 1990 and 2017 dataset proved to be impossible. Limitations of the cross-sectional design in putting the research questions to the test, have been repeatedly discussed in the past (Zaidi & Morgan, 2017) but also recognized by the adherents of the SDT framework (Lesthaeghe & Moors, 2002). My longitudinal repeated cross-sectional research at least provided some evidence that the pattern of associations of values and living arrangements is there to stay. And let’s be honest, no one of us really ever saw a dinosaur; however, in the footprints of their existence, provided by anthropological excavations, we are capable of creating a reasonable accurate picture.
Billari F.C, & Liefbroer AC. (2004). Is the Second Demographic Transition a Useful Concept for Demography? Introduction to a Debate. Vienna yearbook of population research. 2004:1–3 https://ideas.repec.org/a/ vid/yearbk/v2y2004i1p1-3.html
Becker G. (1981). A Treatise on the Family. Cambridge: Harvard University Press.
Bumpass L. L. (1990). What’s Happening to the Family? Interactions Between Demographic and Institutional Changes. Demography, 27(4): 483-498.
Easterlin R. A., Macdonald C., & Macunovich D.J. (1990). How Have American Baby Boomers Fared? Earnings and Economic Well-being of Young Adults, 1964-1987. Journal of Population Economics, 3(4): 277-290.
Inglehart R. (1990). Culture Shift in Advanced Industrial Society. Princeton: Princeton University Press.
Lesthaeghe, R., Moors, G. & Halman, L. (1992). Living Arrangements and Values Among Young Adults in the Netherlands, Belgium, France and Germany, 1990. IPD-Working Paper 1992-3. https://www.vub.ac.be/ demography/wp-content/uploads/2016/02/WP-IPD-1992-3.pdf
Lesthaeghe, R., & Moors, G. (1996). Living Arrangements, Socio-economic Position and Values Among Young Adults – a Pattern Description for France, Belgium, Germany and the Netherlands. In: D. Coleman, (Ed.), Europe’s Population in the 1990s (pp. 163-221). Oxford: Oxford University Press.
Lesthaeghe R., & Van de Kaa, D. (1986). Twee Demografische Transities? In R. Lesthaeghe and D. van de Kaa (Eds.), Groei of Krimp? (pp. 9-24). Deventer: Van Loghum Slaterus.
Moors, G. B. D. (2000). Values and Living Arrangements: a Recursive Relationship. In L. J. Waite (Ed.). The Ties That Bind. Perspectives on Marriage and Cohabitation. Hawthorne: Aldine de Gruyter.
Moors, G. B. D. (2003a). Estimating the Reciprocal Effect of Gender Role Attitudes and Family Formation: A Log-linear Path Model with Latent Variables. European Journal of Population. 19(2): 199-221.
Moors, G. B. D. (2003b). The Two Faces of (Post)materialism: A Decomposition Approach. International Journal of Public Opinion Research. 15(4): 396-412.
Rindfuss, R., & VandenHeuvel, A. (1990). Cohabitation: A Precursor to Marriage or an Alternative to Being Single? Population and Development Review, 16(4): 703-726.
Zaidi, B. & Morgan, S.P. (2017). The Second Demographic Transition Theory: a Review and Appraisal. Annual Review of Sociology, 43: 473-492.