Families & households abstracts

Strand organiser: Berkay Ozcan, London School of Economics

Family processes of immigrant and ethnic groups - Monday 8 September 4.45pm

Lone motherhood among immigrants in Germany: Characteristics, entry, and exit Nadja Milewski, University of Rostock, Laura Bernardi, University of Lausanne

Our paper investigates patterns and determinants of lone motherhood among international migrants living in Germany. Migrant status is a risk factor for child poverty as it is lone motherhood. However, compared to the non-migrant population, migrant unions are generally more stable, the number of children in immigrant families is higher, and extra-marital childbearing is less frequent among immigrants. Hence, we expect that immigrants have lower risks of experiencing lone motherhood. Yet, we do not know much about whether their lone motherhood experience (timing and type of entry into lone parenthood and the duration in the status of lone parenthood) differs from that of non- migrants. We used data from the German Socio-Economic Panel (SOEP) and constructed the partnership careers of about 8500 mothers who were born between 1946 and 1990 and who had at least one child under age 18. About 25% of the sampled women were either immigrants or descendants of immigrants. The prevalence of ever experiencing lone motherhood was about 18% among immigrants and about 24% among non-migrant mothers. We applied event-history techniques to the transition into and out off lone-motherhood status. When controlling for the union status at birth, the age at first birth and other socio-demographic characteristics of the mother as well as the age and number of children affected, immigrants showed a significantly lower risk of becoming a lone mother. Yet, also the risk of the exit from lone motherhood was lower as compared to non-migrants. These risk differentials could not be explained by control variables. A comparison of the characteristics of lone migrant mothers and those without migration background showed that immigrant lone mothers are significantly more often out of employment, they receive more often social benefits, and have a larger number of children.

Email: nadja.milewski@uni-rostock.de

Partnership Dynamics among Immigrants and Their Descendants in Europe
H. Kulu, T Hannemann, University of Liverpool

This study investigates union formation and dissolution among immigrants and their descendants from a comparative perspective. Although there is a growing literature on the dynamics of immigrant fertility and mixed marriages, partnership trajectories among immigrants and ethnic minorities are little studied from a comparative perspective. We use longitudinal data from six European countries (UK, France, Sweden, Switzerland, Estonia, Spain) and apply Poisson regression to model count data in the form of disaggregated exposure-occurrence tables. We will first investigate partnership formation by distinguishing between cohabitation and direct marriage; we will then study cohabitation outcomes (marriage versus separation) and marital separations. We will contrast partnership trajectories of various immigrant groups and compare these with those of the ‘native’ populations. Our preliminary analysis shows that immigrants from non-European countries are more likely to follow the ‘traditional’ partnership trajectories than natives. Specifically, they have higher marriage rates, lower (premarital) cohabitation levels and are less likely to separate. These differences largely persist after individuals’ socio-demographic characteristics are controlled. Still, in comparison to ‘natives’, ‘modern’ family formation patterns dominate among some non-European immigrant groups (e.g., Caribbeans in the UK and Latin Americans in Spain). This result may be due to specific patterns in the region of origin (Caribbean countries) or the prevalence of traditional patterns in the destination country (Spain). The study will deepen our understanding of how contextual factors shape family trajectories of immigrants and their descendants.

Email: hill.kulu@liverpool.ac.uk

Conceptualising demographic capital in relation to migration and employment outcomes in London
Ibrahim Sirkeci, Regent's Centre for Transnational Studies, Regent's University London, Sinan Zeyneloglu, Regent's Centre for Transnational Studies, Regent's University London and Department of City and Regional Planning, Gaziantep University, Turkey

Development literature refers to the concepts of social capital and human capital as key factors influencing educational and labour market outcomes while individual components of demographic background are frequently referred to without much attempt to conceptualise as a construct. Demographic capital is discussed in this study as a conceptual attempt to supplement the measures and debates revolving around the concepts of social capital and human capital. Demographic capital as a conceptual construct is based on selected demographic characteristics including marital status, age, sex, children in the family, size and composition of the household, migration status. We have also included education, religion, and ethnicity in our models. We have used the 2001 UK Population Census. We have restricted our analysis to London to test the model with diverse populations and demographic characteristics. We have run two logistic regression models. The first model included the demographic characteristics above while we have added ethnicity and religion in the second model. The total sample for London –including inner and outer London- comprised of 219,027 individuals in the SARs data. We have reduced the data to focus on individuals within working ages while also excluding those in full time education. Since we are concerned with household characteristics, non-household population is also excluded. Thus final sample included 110,014 individuals. Based on regression models, our findings show that while education appears to be the primary determinant of the occupational outcomes, components of demographic capital such as family type account for a third of occupational skill level variation.

Email: sirkecii@regents.ac.uk

Partner ethnicity and ethnic minority socio-economic occupation
Greta Morando, ISER, University of Essex

High rates of unemployment, inactivity, and concentration in certain occupations are among the most relevant issues concerning ethnic minorities in the UK. In this context it is important to understand what role integration plays in reducing disparities in the labour market between ethnic minority and white majority populations. We shed some light on this by considering the having of a partner in the white majority group as a proxy indicator of integration of ethnic minorities into British society. Using data from the first wave of Understanding Society (2009/2010) we compare the socio-economic positions of ethnic minorities in co- and inter-ethnic partnerships. We implement propensity score matching techniques to account for selection bias. We hypothesise the existence of several channels –social networks and language fluency- through which having a partner in the white majority population can help a member of the minority population to integrate into the society, and to subsequently affect their labour market outcomes. We find that after accounting for self-selection into an inter-ethnic partnership, having a white majority group partner increases the probability of being in certain socio-economic positions. This is in particular the case for women.

Email: gmoran@essex.ac.uk

Families: Gender equality within the family - Tuesday 9 September 9.00am

Proximity of Couples to Parents: Influences of Gender, Labour Market and Family Tak Wing Chan, John Ermisch, University of Oxford

We use household survey data from the UK to study how close ‘middle-aged’ men and women in partnerships live to their own parents and their partners’ parents. We find a slight tendency for couples to live closer to the woman’s parents than the man’s. This tendency is more pronounced among couples in which neither partner has a degree and in which there is a child. In other respects, proximity to parents is gender neutral, with the two partners having equal influence on intergenerational proximity. Better educated couples live farther from their parents. And although certain family characteristics matter, intergenerational proximity is primarily driven by factors affecting mobility over long distances, which are mainly associated with the labour market, as opposed to gender or family circumstances.

Email: tw.chan@sociology.ox.ac.uk

Exploring gender, work and families – Domestic labour, paid work and parental gender attitudes
Lauren Bird, Department of Epidemiology & Public Health, University College London, Amanda Sacker, Anne McMunn, International Centre For Lifecourse Studies In Society and Health, Department of Epidemiology & Public Health, University College London

On-going changes in the division of labour in families have occurred within the wider context of societal changes in gender attitudes and normative beliefs. The diversification of modern family work and lifestyle patterns exemplifies a socio-cultural shift away from strongly gendered work and family roles. However, change is neither consistent within individuals nor universal in populations, and changes in behaviour and attitudes are not necessarily correlated as may be expected. This study uses longitudinal data from over 10,000 two-parent families in the Millennium Cohort Study, a nationally representative cohort of children born in the UK during 2000-2002. This research focuses on associations between divisions of domestic labour, paid labour and attitudes to maternal employment, and their socio-economic patterning. Using multivariate logistic and linear regression, strong associations were found between parents’ paid labour, gender attitudes and domestic labour. For both men and women, fewer hours spent in paid work were associated with increased participation in domestic labour. Less favourable attitudes to maternal employment were associated with greater responsibility for mothers to perform domestic tasks. Social class was marginally associated with divisions of domestic labour, in particular, small employers and self-employed were associated with less equality and semi-routine and routine employment with more equality. Contrary to some previous research, neither education nor income was strongly associated with mothers’ perception of how domestic labour is divided. This research points to strong correlations between gendered attitudes and behaviour as well as less socio-economic patterning in gender divisions of labour than was hypothesised.

Email: lauren.bird.12@ucl.ac.uk

Father involvement, family time and social class in the UK
Ursula Henz, Killian Mullan, London School of Economics

Fathers in modern families are no longer viewed simply as the breadwinner. Today’s fathers are expected to be more involved in the care and nurture of their children and in shared family time. In tandem with changing norms and expectations, research shows that fathers are now more engaged in childcare and spending more time with children in general. However, evidence suggests that these changes are unevenly spread across socio-economic groups. A substantial literature shows that highly educated fathers spend more time with children, and there is some evidence from the UK that time spent with children is greatest among fathers in professional/managerial occupations. Social class, however, has not been a primary concern in previous research, focused on parental employment patterns, and merits a closer examination, especially in the UK context where traditional gender norms are known to be strongest in working class families. Using data from couples in the United Kingdom Time Use Survey 2000-01, we tested the impact of social class and education on varied aspects of father’s time with children, with and without mothers, including time in different childcare activities and in shared family activities. We found that fathers in intermediate and routine/manual occupations spent less time with children on weekends only. However, social class was unrelated to father’s engagement in physical or interactive childcare, or to shared family activities, on either weekdays or weekend days. Results are discussed in light of debates about the positive impact of father involvement for both gender equity and child wellbeing.

Email: u.henz@lse.ac.uk

Gender Inequality in the Division of Housework over the Life Course: a European Comparative Perspective
Tine Kil, Karel Neels, University of Antwerp

Today paid work is more equally divided in European families than a few decades ago. The evolution is only partly compensated by a more equal gender distribution of unpaid work. This tension between public and private gender inequality creates role conflicts and may cause women to postpone or renounce family formation. Therefore this study aims to examine how gender inequality in the division of housework varies across different stages of the life course and whether this gender inequality varies between different contexts. Using data from the European Social Survey (2010) a sample of 24045 hetero-sexual couples from 24 countries was selected. Using multilevel analysis we examined how the distribution of domestic work over the life course is affected by (1) time availability, relative resources and gender ideology and (2) the cultural and institutional context. We also examined (3) the influence of context variables on the extent to which these individual factors play a role. The results show that housework is least equally shared among couples with children. A progressive gender ideology has a small positive influence on gender equality for couples with young children, but this effect depends on the context as cross-level-interactions suggest that they better succeed in implementing their progressive ideas in a country with a progressive national gender culture and more full time child care. This leads to the conclusion that context plays a role in averting the domination of emerging parenting practices and ideas over gender ideology that seems to occur at the birth of children.

Email: tine.kil@uantwerpen.be

Families: Cohabitation versus marriage - constraints and contexts - Tuesday 9 September 4.45pm

How Does Cohabitation Fit into the Family Life Course? An exploration of ideals and norms in Europe
Jennifer A. Holland, University of Southampton, Helga de Valk, Netherlands Interdisciplinary Demographic Institute and Free University Brussels

Whether life courses are shaped by social norms or are primarily governed by individual preferences and opportunities is a matter of scholarly debate. While norms are particularly important for sanctioning behaviors, previously sanctioned behaviors may become normatively acceptable within permissive social contexts or once a behavior becomes widespread. Still, these behaviors may continue to be socially regulated. Under these circumstances, ideals or “what is believed to be best for individuals/society,” may give meaning to a behavior and guide individuals’ actions. In this paper we use the concept of ideals to discuss variation in the acceptance of non-marital cohabitation across European countries. While the incidence and prevalence of cohabitation is rising, the meaning of cohabitation varies across and within populations throughout Europe. Using data from the European Social Survey (Round 3, 2006/07), we build a typology of the normative context of non-marital cohabitation across the 23 survey countries. Then we investigate individual-level reports of the ideal age for cohabitation. We explore how cohabitation ideals are shaped by individual-level characteristics, country-level normative context and by additional country-level characteristics using two-level multinomial logistic regression (MLwiN; Rasbash, Steele, Browne, & Goldstein, 2009). We demonstrate that norms and ideals tap into two different concepts and explore what each reveals about the role and timing of cohabitation for individuals living in European societies.

Email: j.a.holland@soton.ac.uk

The link between the divorce revolution and the cohabitation boom
Brienna Perelli-Harris, Ann Berrington, Paulina Galezewska, Jennifer Holland, Nora Sanchez Gassen, University of Southampton

Over the past decades, divorce has increased dramatically throughout Europe. The increase in divorce has fundamentally altered the institution of marriage, from a life-long union to one that has the potential to dissolve. At the same time, cohabitation emerged as a way for a couple to live together without having to marry. In this paper, we investigate whether the relationship between divorce and cohabitation is direct, with increases in divorce facilitating increases in cohabitation. This relationship may be evident on the micro-level, with individuals who experience divorce choosing to avoid marriage in favor of cohabitation. Or the change may occur on the macro-level, with the implementation of new divorce legislation, increases in the social acceptability of divorce, and shifts in social norms about the institution of marriage and perception of cohabitation. Here we explore explanations for the link between cohabitation and divorce by investigating potential causal pathways. We draw on new sources of quantitative and qualitative data to provide evidence. Using harmonized partnership histories from across Europe, we describe the emergence of both behaviors and examine how they are linked. Using focus group data from 8 European countries, we illustrate how divorce has changed social norms relating to marriage and cohabitation. We expect that the explanations and pathways will differ across countries and that some relationships will be more evident than others. Our investigation will provide insights into why and how cohabitation developed, but also how country-specific factors may have led to different developments.

Email: B.G.Perelli-Harris@soton.ac.uk

Alternatives to age in analysing early adult transitions
Eva Beaujouan, Vienna Institute of Demography, Mike Murphy, London School of Economics, Karel Neels, University of Antwerp, Maire NI Bhrolchain, University of Southampton

By long established custom, variation in the rates of demographic events in young adulthood – e.g. first birth, partnership and, marriage—is described and analysed with reference to chronological age. Other dimensions of personal time have also been in use—e.g. births have been analysed by duration of marriage, and second and later births by duration since previous birth. But chronological age remains central to the delineation of individual level heterogeneity in transition rates in demography. While age is clearly a powerful differentiator of transition rates among young adults, the behavioural origins of this differentiation remain obscure. Furthermore, the question as to what gives rise to age variation is rarely examined explicitly in a systematic way. In this study, we look at alternative dimensions of personal time – time enrolled in education, age at completing education, and duration since completing education—and assess how far they account for individual heterogeneity in rates of selected early adult transitions by comparison with chronological age, in three European countries, the UK, France and Belgium. We find that time enrolled in education and duration since first leaving full time education are more closely connected with e.g. first birth rates than is age. We examine in addition whether the timetable of educational enrolment has always had its current demographic significance or whether there have been changes over time in that respect.

Email: mnb2@soton.ac.uk

Individual strategies in the face of the realities of the marriage market. Japan as a case study
Ekaterina Hertog, Oxford University, Eunmi Mun, Amherst College, Mary C. Brinton, Harvard University

Very low fertility rates have come to characterize large parts of the postindustrial world, including Eastern Europe, Southern Europe, and East Asia. An extensive demographic literature has focused on the cultural and economic reasons behind the large variation in fertility rates throughout Europe, and a parallel though less extensive literature has examined the reasons for fertility decline throughout East Asia. Late marriage and the incidence of non-marriage are particularly important in explaining East Asia’s recent very low fertility rates. Strong social norms dictating that childbearing occur within marriage mean that age at marriage and the propensity to marry hold considerable significance for explaining low fertility. This paper will use Japan as a case study to throw light on the factors that may delay marriage formation or even make individuals exit the marriage market. Using process data from a large Japanese marriage agency we use fixed-effects models to analyse how quality and quantity of the date offers an individual has received in the recent past as well as time spent actively searching for a marriage partner in the marriage agency affect their behaviour. We find strong gendered patterns with men being rather broad-minded about potential marriage partners from the start of partner search and women making adjustments in the types of partners they are willing to accept depending on their successes and failures with previous dates. We believe that a better understanding of how individuals are affected by their marriage market experiences allows us to theorize a core mechanism behind reduced fertility and thereby can potentially contribute to more adequate demographic modelling.

Email: ekaterina.hertog@sociology.ox.ac.uk

"Til Work Do Us Part" A Cross-National Analysis of Globalization and the Role of Work in Non-cohabiting Marriages
Giulia Ferrari, Ross Macmillan, Dondena Centre for Research on Social Dynamics, Bocconi University

From numerous vantage points, cohabitation has been deemed the quintessential feature of conjugal life and a fundamental feature of the personal and social goods that come with marriage. At the same time, there is almost no research on non-cohabiting marriages. Using data from the IPUMS-International program, we investigate the extend, the heterogeneity, and the political economic context of non-cohabiting marriages. Specifically, we estimate multilevel models that examine the direct and conditioning effects of globalization on the effects of work on the probability of couples living together. Results indicate that globalization has a complex relationship with non-cohabiting marriages and has an important role in shaping labor effects, particularly for low-skill labor that is not dependent upon geography. Implications are discussed.

Email: giulia.ferrari@unibocconi.it

Family instability and child outcomes - Wednesday 10 September 9.00am

Family change, employment change, and home moves in the first five years of life: a comparison of US and UK
Ludovica Gambaro, Heather Joshi, Institute of Education, Anthony Buttaro, Mary Clare Lennon, City University New York

This paper compares the first five years of life of two cohorts born around 2000 in the US and UK, the Fragile Families Study (FFS) and the Millennium Cohort Study (MCS). We compare the stability of their family lives in terms of family structure, parental employment and residential mobility. In all three dimensions there is more instability for American families, particularly residential mobility. We seek to explain the greater number of residential moves in US in terms of family and employment change, housing tenure and other control variables which are reasonably comparable across contexts. In both countries we find that family instability is associated with higher residential mobility, but differences emerge in relation to housing tenure. Social housing in UK appears to reduce residential mobility to a far greater extent than public housing in US. We then estimate at least one aspect of the impact on children of the three dimensions of instability, taken singly and together, with and without allowing for other controls. Preliminary results indicate that there are adverse associations of residential mobility and family change with child behaviour, which are accounted for, in slightly different ways and degrees in the two countries, by confounding and mediating variables.

Email: heather.joshi@ioe.ac.uk

The effects of parental separation on children’s health outcomes
Alice Goisis, Berkay Ozcan, University College London

Over the last few decades, most industrialized countries have experienced a significant increase in marital instability and divorce. Shifts in family structure and stability generated a wealth of research aiming to assess the consequences of parental separation on children’s wellbeing and development. To date, this literature has largely focused on children’s educational outcomes and attainment, which might have important consequences for subsequent wellbeing. Conversely, with the exception of a few U.S. studies, considerably less attention has been given to investigate the consequences of parental separation on children’s physical health. This is a shortcoming since parental separation, by changing the amount and type of resources available in the family, could affect children’s health trajectories negatively. In this study, we aim to expand the focus of existing research on the consequences of parental separation to include its effects on children’s physical health. We use data from the U.K. Millennium Cohort Study, which provides information on children’s BMI, height and obesity at 3, 5, 7 and 11 years of age. By exploiting the longitudinal nature of the MCS, we use a fixed effects regression model to control for time-invariant family (and child) characteristics, which might influence both the propensity to parental separation and child health. Our contribution also involves analysing the process of separation more carefully than previously done by comparing the short and medium terms effects of parental separation and by considering the pre-separation period.

Email: a.goisis@lse.ac.uk

Household membership and child health in Botswana: do non-parental members help or hinder?
Oleosi Ntshebe, Victoria Hosegood, Andrew Channon, University of Southampton

Evidence in high income countries show striking differences in health outcomes between children living in one parent and two parent families. However, living experiences are different in low and middle income countries (LMICs). Families and households in LMICs are larger, and more complex in terms of membership. Living with other non-parental members provides several opportunities to the biological parents for child care: a safety net if the parent cannot afford child care alone, and access to socio- economic resources from other adult household members. On the other hand, whether non-parental members help or hinder depends on the specific knowledge, levels of investment, and effectiveness afforded. The influence of living with either parents or non-parental members is also assumed to vary by household wealth. This study uses data from the nationally representative 2007 Botswana Family Health Survey (BFHS) to investigate household membership and diarrhoea prevalence among children less than five years old. Multivariate models serve to control for confounding variables associated with diarrhoea prevalence. Preliminary results indicate that diarrhoea is lower in households where there are other household members aside from the parents, irrespective of whether the household has one or both parents living there. However, the effect depends on who the other household members are and their relationship to the child. Across households, those which are richer, and regardless of who is in the household, do better than those which are poorer. The analysis underscores the importance for diarrheal prevalence of non-parental adult household members.

Email: on1g11@soton.ac.uk

Children’s experiences of family dispersal in rural South Africa: The influence of union stability, mortality and migration of parents
Victoria Hosegood 1,2,3, Gabriela Mejia-Pailles 2,3, Ann Berrington 1,2, 1 Department of Social Statistics and Demography, School of Social Sciences, University of Southampton 2 ESRC Centre for Population Change, University of Southampton 3 Africa Centre for Health & Population Studies, University of KwaZulu-Natal

High levels of residential separation (family dispersal) of children from parents, particularly fathers, in South Africa have endured despite major political and economic changes in recent decades. The circumstances in which children live apart from one or both parents are diverse and often dynamic. Characterising children as co-resident with either one or both parents masks variation in their social and residential trajectories and their parents over space and time. These trajectories have potentially different consequences for child health and wellbeing. In this paper, we use data from a demographic surveillance system in rural KwaZulu Natal to investigate the experiences of children and parents that determine whether at the age of 5 years the child is living with one, both or neither biological parents. The experiences of particular interest in this study are: a) in- and out-migration events in which the child, parent(s), family or the whole household participate; b) changes in parental union status or type; and c) the death of one or both parents. In addition to then identify retrospective events that influence the family dispersal, we also describe the heterogeneity in children’s living arrangements and consider the utility of characterising children on the basis of parental presence in contexts of high levels of migration and orphaning. Understanding heterogeneity in children’s living arrangements in different settings is important for comparative family-based research given that concepts such as dispersed families and living apart together are widely used in descriptions of contemporary family life.

Email: g.mejia-pailles@soton.ac.uk

Family processes and education - Wednesday 10 September 11.00am

Further Education Aspiration and Attainment: The role of Sibling configurations Feifei Bu, ISER, University of Essex

Previous studies have found that firstborn children enjoy a distinct advantage over their later-born counterparts in terms of educational attainment. This paper advances the state of knowledge in this area in two ways. First, it analyses the role of young people’s aspirations, estimating the effects of sibling configurations on adolescents’ educational aspirations, and the importance of these aspirations on later attainment. Second, it employs multilevel modelling techniques, using household-based data (derived from the British House Panel Study) which include information on multiple children living in the same families. The paper finds that firstborn children have higher aspirations, and that these aspirations play a significant role in determining later levels of attainment. We also demonstrate a significant positive effect of age spacing on educational attainment.

Email: fbu@essex.ac.uk

Maternal employment and adolescent education: a case-study of urban Ghana Philippa Waterhouse, University of Southampton

Since the Educational for All campaign and the Millennium Declaration, significant progress has been made in increasing school enrollment in Sub-Saharan Africa. Despite this achievement, there is concern rapid expansion has put pressure on school systems at the cost of quality. Focusing on the West African country of Ghana, whereas the proportion of children age 6-17 years having ever attended school rose by 10% between 1991 and 2006, basic school completion has remained constant at approximately 50%. The percentage of children completing basic education at correct-age-for grade has declined over this period suggesting possible rising levels of late enrolment, temporary drop-out and grade repetition. Consequently, research needs to take a wider definition of educational access and extend focus beyond initial engagement and consider appropriate progression. This paper examines the associations between basic school completion and secondary school enrolment with maternal employment in urban Ghana using the fifth round of the Ghana Living Standards Survey. Although studies of children’s education frequently incorporate maternal characteristics, such as educational levels, their labour force participation is a relatively neglected consideration. By 17 years, 57% of the sample had completed basic education. Controlling for maternal education, it was found adolescents with mother engaged in agricultural self-employment or family work had lower odds of completing basic education. Engagement of adolescents age 15-17 years in senior secondary schooling is low, with only a third of this age-group being enrolled. Those whose mothers were engaged in paid employment had lower odds of not being in senior secondary education.

Email: pjw2v07@soton.ac.uk

Which is the steeper cost? The effects on partnership formation of increasingly skewed levels of education between genders in historically endogamous ethnic groups in Denmark.
Ask Foldspang Neve, Department of Sociology, University of Oxford

As has been the case in many Western countries, the last two decades have witnessed a dramatic rise of the levels of education among women, especially those of ethnic minorities, creating very different levels of education among the two genders in several minority groups. Whereas both ethnicity and education have been studied extensively as assortative factors in partnering processes. Individually, these factors are often considered among the strongest predictors of partnership behavior as class background has decreased in importance. However, little research has been conducted with the potential to investigate the behavior of individuals faced with few chances of partnering both ethnically endogamously and educationally homogamously, as is the case of many ethnic minority women. This study applies a Cox proportional hazards model with discrete time intervals. The dependent variables are combinations of ethnic and educational partnership formation, leading to a controlled comparison. Ethnicity is measured as partnering within one’s own group; with the native population; or with a third group. Educational sorting is measured as partnering downwards, in one’s own group, or upwards. Finally, non-partnering is its own category. The study uses micro-level registry data of the entire adult population of Denmark in a 15-year observation window. Individual observations are person-years for the population at risk. All individuals not married or cohabiting are considered at risk. Preliminary results show that age of partnership formation has risen faster in ethnic minority groups than the educational expansion alone would predict. Ethnicity remains a strong barrier to inter-group partnership formation.

Email: ask.neve@nuffield.ox.ac.uk

Union formation and educational differentials in micro- and macro-level economic context in France (1989-2008)
Jorik Vergauwen, Karel Neels, University of Antwerp

The short-term impacts of economic recession, rising economic uncertainty and its educational underlying driver on fertility are well documented in the literature. Whereas postponement of union formation has been suggested as one of the main pathways through which economic conditions affect fertility, some papers have directly addressed the micro and macro-level economic conditions on union formation. Union formation (especially marriage) hazards are theorized to decrease due to a lack of financial and social long-term prospects. School enrolment, postponement of transitions and flexible partnership forms are identified as coping strategies to uncertainty in literature. This contribution aims to examine the entry into a first unmarried cohabiting union and marriage after cohabitation among different educational levels in relation to their employment status and aggregate-level economic context. The analyses use union and employment histories (1989-2008) of male and female respondents between the age of 16 and 39 from the French Harmonized Histories and Generations and Gender Survey Wave 2. To test our research hypotheses the analyses draw on discrete-time event history methods. We find that employment is particularly an important prerequisite for union formation among the higher educated. An indication that first co-residence becomes a strategy to deal with economic uncertainty for lowly educated men is found as well. In correspondence, susceptibility to aggregate-level economic context with regard to entry into a first cohabiting union prevails amidst highly educated men. Our results furthermore suggest that these effects might be rather similar for the transition from cohabitation to marriage.

Email: jorik.vergauwen@uantwerpen.be