Welcome to read the abstracts of the symposium presentations. A few abstracts are not published at authors’ request.


Wednesday 14 June

1.1. Families and households
1.2. Life Course
1.3. International migration

Poster session

2.1. Families and households
2.2. Health, morbidity and mortality
2.3. Migration

Thursday 15 June

3.1. Families and households
3.2. Life Course
3.3. Migration

4.1. Fertility
4.2. Educational inequalities
4.3. Intergenerational relations in humans and non-human animals

5.1. Families and households
5.2. Ageing and intergenerational relations
5.3. Health, morbidity and mortality 

Friday 16 june

6.1. Fertility
6.2. Educational inequalities
6.3. Migration

7.1. Families and households
7.2. Fertility
7.3. Health, morbidity and mortality



1.1. Families and households


Does marriage and having children make life better in Europe?

By: Mare Ainsaar, Kadri Rootalu

The diversity of family types is increasing in Europe. Existing research provides inconsistent results on whether having children in a family influences life satisfaction, and even less is known about how different partnership and social support types influence life satisfaction.

We investigate the life satisfaction of officially registered and cohabiting persons with and without children in 24European countries European Social Survey 2012. The unique contribution of this work is the ability to simultaneously compare different family types and take the effect of children into account. In order to explain family type differences and control for an individual selectivity of persons into the different family types (see Kravdal, 2014), we added a set of individual and country indicators into the analyses and run multilevel models.

All results of this study lead us to the conclusion that although the diversity of different familyforms might increase, married couples still have several advantages in Europe. There are no doubts about the benefit of partnership on well-being, but the benefit of children on life satisfaction seems to depend on the economic situation of parents and to some extent on the wealth level and support of a country. Single parenthood is the least advantageous family form from the point of view of lifesatisfaction, but the large share of divorced and separated persons in this group shows that this is not a voluntary choice. Although in general life satisfaction levels follow the East–West divide, the life satisfaction of families within countries is more driven by the diversity of social support and a clear division of countries into Eastern and Western family types is not present.

His education, her education, and marriage stability in China

By: Ma Li

Background: Existing research shows that couple’s education parings play an important role in marriage stability. Normally, couples with equal levels of education have lower risk of divorce than educationally heterogamous couples (Jalovaara 2003, Kalmijn 1998, Lyngstad 2004). Among the latter, couples in which the woman has education advantage have substantially higher likelihood of divorce than other couples (Schwartz and Han 2014). So far, little relevant knowledge has been found for the context of China, a country that has experienced substantial socioeconomic changes in the past few decades, where hypergamy or “marrying up” has been a norm among women (Ji and Yeung 2014, Tian 2013).

Research questions: Based on existing literature for other societies, we pose the following questions for the context of China. 1. Do couples with equal levels of education have lower risk of divorce than educationally heterogamous couples? 2. Among educationally heterogamous couples, are couples in which the woman has education advantage more likely to divorce?

Data and methods: Data used for analyses come from the China Family Panel Studies (2010-2012 waves). The longitudinal nature of the data allows for applying event-history analysis to study the association between couple’s education parings and divorce. Our observation covers the period of 1970-2012.

Findings: Preliminary results show that in Chinese society a marriage in which the woman’s educational attainment is lower than that of the man is the most stable. In comparison, couples with equal educational levels and couples in which the woman enjoys education advantage are more likely to divorce, especially the latter. These patterns hold across our entire observation period from 1970 to 2012. During the two most recent decades, couples in which the woman had education advantage had especially higher likelihood of divorce.

Conclusions: The findings provide important implications for the progress of gender equality in htererosexual relationships for the context of China. (References omitted)

Cohort Trends in Divorce and Family Dissolution in Sweden, 1970-2005

By: Juho Härkönen

I analyzed recent cohort trends in divorce and family dissolution (defined here as the break-up of unions in which women had their first children) in Sweden for union cohorts 1970 to 2005. Using population register data on the complete Swedish population, I estimated life table estimates and discrete-time event history models. Both types of union instability—divorces and dissolution of families with common children—increased from the 1970 to the 1990 cohort, and decreased slightly thereafter. The increase was particularly pronounced for least educated women. The increase in union instability would have been even more pronounced without accompanying increases in the ages of marriage and first birth and educational expansion. Decreases in marital first childbearing, on the other hand, partly explained the increases in family instability. However, the decrease in divorce risks from the 1990 to the 2005 cohort could be observed even when accounting for age at marriage, education, and region of birth. The opening up of the educational gradient in divorce and family instability was not explained by these variables. Rather, increases in age at union formation and decreases in marital first births affected instability levels across the board.

Two decades of same-sex marriage in Sweden: A demographic account

By: Martin Kolk, Gunnar Andersson

In this study we examine period trends in same-sex marriage formation and divorce during 1995-2012 in Sweden. This period begins with the introduction of the registered partnership for same-sex couples and also covers the introduction of formal same-sex marriage in 2009. We use register data for the complete population of Sweden, and contrast patterns for male and female same-sex unions with those of opposite-sex unions. Our study shows that female same-sex union formation has increased rapidly over the period, while trends for male same-sex unions show less of increase. The introduction of same-sex marriage legislation in 2009 appears to have had no noticeable effect on the pace of formation of new same-sex unions. In contrast, legal changes that supported parental rights in same-sex unions may have fueled the formation of female same-sex marriages. Further, we find that divorce risks in the marital unions of two women are much higher than in other types of marriages. We present some evidence of a convergence in divorce patterns over time. At the end of our study period male same-sex unions have the same divorce risk levels as opposite-sex marriages, while the elevated risks of divorce in female same-sex unions appear to have stabilized at levels somewhat lower than those observed at the late 1990s.

Who did disabled people marry in a past society Cases from the Sundsvall region in Sweden in the nineteenth century

By: Helena Haage

To find a partner, marry and start a household of ones own were a common and an expected life path for nineteenth-century people. Only a few previous studies have been conducted on whether disabled people married and who they united. The aim of this paper is to explore with whom disabled men and women married. eside the possible bad health caused by the impairment itself, it is most likely that the appearance and behavior of disabled people would have labeled them deviant by prevailing attitudes and norms in the surrounding society. This would in turn have generated disadvantages for them in the marriage market, and perhaps also limited their choices of partners to marry In this study the ministers marks of impairment are used to recognize disability in nineteenth century parish registers from the Sundsvall region, which are digitized and stored at the Demographic Data ase (DD) at Ume University. The results show that a majority of the disabled people did not marry a disabled partner. It was neither common to contract a spouse characterized by other social disadvantages, such as having a criminal record. The spouses were also very similar in age at marriage. Disabled men married frequently women originating from the same socio-economic strata, while disabled women married downwards to some extent. This is one of the first studies to explore partner selection among a comparatively large amount of disabled people. Analyzing them and their spouses using demographic data, this study reflects issues of disabled peoples inclusion in social life historically.


1.2. Life Course


Early adulthood education and employment trajectories of young adults placed in out-of-home care as children

By: Antti Kääriälä, Pasi Haapakorva

Compared with peers, children placed in statutory out-of-home care (OoHC) are in a more vulnerable position in their early adulthood. For example, as young adults they have elevated risks to attain low education, experience unemployment, live on social assistance or disability pension, and suffer mental health problems. In this study, we explore educational and employment trajectories among young adults placed in OoHC as children, and investigate risk factors associated with typical trajectories.

We use individual level national birth cohort data, which comprises register-based follow-up information of all children born in Finland in 1987 (n=59 476, of whom 1900 in OoHC). We apply sequence analysis in exploring monthly education and employment information to construct the trajectories for the period 2005–2012. We estimate risk factors associated with typical pathways with multinomial logistic regression.

For the most part of the seven-year follow-up period, 49% of the young adults placed in OoHC were in stable education or employment trajectories. In contrast, 26% experienced trajectories characterized by alternate unemployment and social assistance periods. For 15% the main activity was child rearing. In our preliminary analysis, risk factors that predict trajectories outside education and employment involve biological parents’ low educational level, mental health problems in adolescence, and placement in OoHC at teenage. When writing this abstract, the analysis is still uncompleted.

Although majority of young adults placed in OoHC are in a stable course to working life through studies, a considerable number of them are outside education and employment. Childhood disadvantages, mental health problems, and placement-related factors are associated with long-lasting unemployment and life on social welfare. To prevent permanent social exclusion, many young adults with a history in OoHC need strong support well into adulthood. Our study shows the importance of examining longitudinal trajectories when assessing the social inclusion of disadvantaged young adults.

Children of broken families – how do they do in adulthood?

By: Amy Frølander

A register-based study of differences in educational level and family-status between young people from broken families and their peers from nuclear families is presented. Educational level of a person is strongly correlated to the educational level of their parents. Children with educated parents more often get secondary or tertiary education than children with parents with no education. But children from broken families more often end up with only primary education than their peers. This is also the case when the educational level of their parents is taken into account. Does the timing of a divorce during childhood have influence on the share of youngsters with only primary education? Do children from broken families form families as their peers from nuclear families? Do they cohabit, marry, get children and divorce as their peers from nuclear families? Do they form families earlier, later or maybe not at all? Is the risk of divorce inherited across generations? This study will be based on the Danish population registers from 1986-2017 and on the register on the highest education completed. Children born in selected years, with parents being alive at the 18.th birthday, are followed in the registers up to present day.

The effect of parental absence during childhood on social mobility in the Netherlands, 1860-1970

By: Matthias Rosenbaum-Feldbrügge

The demographic and social processes of the last 150 years have dramatically changed the proportion of children growing up in the presence of both biological parents. In the Netherlands for instance, this proportion increased constantly from the middle of the 19th century until the 1960s and fell again afterwards. Some authors therefore point to similarities between the experiences of children in present-day families and those born in the 19th century. Thus, analyzing the life courses of children born in the 19th and beginning of the 20th century helps to gain a deeper understanding about the long-term consequences of growing up in the absence of one biological parent. This is extremely valuable as later life data for individuals born in the last two decades are of course not yet available.

This article examines social mobility by looking at the impact of parental absence on the occupational careers of men being born in the Netherlands between 1860 and 1923 (there are not enough data for women). For these purposes the Historical Sample of the Netherlands (HSN) is exploited. It contains rich longitudinal information about around 37,000 research persons, their occupations as well as the kin and non-kin members present in the household. That makes it possible to follow the occupational careers of children compared to their parents and to analyze the role of the family and the household composition during childhood. It is expected that parental absence has a negative influence on the son’s occupational prospect due to the loss of parental investments and professional networks. Moreover, the absence of a parent might lead to early life stress which is also expected to exert a negative impact on the occupational career of the son. Regression techniques such as event history analysis are used to investigate the relationship between parental absence and occupational career.

Behind every successful (wo)man is a successful parent-in-law?

By: Sanna Kailaheimo, Elina Kilpi-Jakonen, Jani Erola, Antti O. Tanskanen

It is well known that children with high status parents attain higher socioeconomic status in adulthood compared to children with low status parents. However, there is a lack of studies investigating whether parents-in-law boost the socioeconomic success of their children-in-law. Using Finnish Census Panel data of over 50 000 individuals born between 1969 and 1975 who are followed from the age of 18 to the age of 35–40 we study whether the resources of the parents-in-law have an effect on status attainment. We find that the resources of the parents-in-law are positively associated with one’s own status. This holds even after taking into account the status of one’s own parents, the status of the partner, and both own and partner’s education. Interaction analysis suggest that the resources of the parents-in-law are most beneficial if own resources are high, the couple is married and their relationship has lasted for a long period of time. However, for women low resources of the parents-in-law are more adverse the longer the relationship has lasted for.


 1.3.International migration


The impact of migration on the population of Sweden

By: Lena Lundkvist, Andreas Raneke

In 1969, roughly 8 million people lived in Sweden, of whom nearly 500 000, or 6 percent, were born abroad. Since then, immigration to Sweden as well as emigration from here has increased. During the period 1970–2015, 2.8 million people have immigrated while at the same time, 1.6 million people have emigrated. Migration has thus contributed to roughly 1 million more people in Sweden. Besides contributing to an increased population themselves, immigrants also contribute to the population with the children they give birth to in Sweden.
This paper presents a hypothetical calculation of the population in Sweden 2015 if Sweden had completely closed borders since 1970. The number of births has been calculated using the observed age specific fertility rate and the number of deaths with the observed death risk respectively year. The results show that without migration, Sweden would have had a stagnating population. The population in 2015 would largely look like the population in 1969, that is, nearly 2 million fewer persons than today. The age structure would have been different. Without immigration there had in 2015 been more aged over 65 than aged less than 20. The gender distribution is affected only marginally by migration

Why do migrants go back and forth? circular migration between Finland and Sweden?

By: Rosa Weber, Jan Saarela

Circular migration in a setting of free mobility has received increasing attention among policy makers over the last two decades. However, to date we know relatively little about the mechanisms underlying circular migration. In this paper, we try to fill this gap in the literature by analysing circular migration between Finland and Sweden using detailed linked register data. The data set covers the years 1987-2005 and provides information on individuals in Finland and Sweden, thus allowing us to follow migrants across national borders. Finland and Sweden are both members of the common Nordic labour market, which means that migrants can move freely between the two countries. We use an event history framework and focus on the first four moves. Specifically, (1) we compare the likelihood of moving for the first, second, third or fourth time using a repeated risks model and (2) analyse the mechanisms underlying the decision to make a move. In this, we focus on human capital, demographic characteristics and a possible indicator of cultural affinity. Our study shows that each migration accentuates the selection mechanisms underlying the previous move. In this way, the mechanisms underlying the first migration are similar to those underlying move 3 and move 2 and 4 are more similar. This indicates that the decision to move is strongly influenced by the direction of the move and preferences or cultural affinity for Sweden or Finland may be important factors in the decisions to move again. Moreover, country-specific factors seem to be more important for the decision to move again in a setting of free mobility than for migration between Mexico and the US.

Fertility patterns and sex preferences for children in immigrant-native intermarriages in Sweden

By: Annika Elwert

A growing demographic literature focuses on migrant fertility and shows that immigrants tend to adapt to local fertility levels over time.  A different aspect of migrant fertility has gotten much less attention, namely sex preferences for children. This aspect of immigrant fertility can add knowledge on processes of socialization and adaption related to fertility behavior, given that there are strong sex preferences for boys in certain parts of the world, but relative preferences for girls in Sweden.
This study focuses on fertility patterns in intermarriages between immigrants and natives in Sweden. Intermarried migrants’ fertility is largely outside of the focus of research in Europe. Intermarried couples bring an interesting aspect to the study of (migrant) fertility as they highlight the necessity to study childbearing as a joint decision of couples, and give the opportunity of studying whose socialization is more important for fertility decisions, the native or the immigrant partner’s. Using Sweden’s longitudinal registers for the period 19902009, I apply event-history techniques and explore transitions to third births on sex preference patterns across various immigrant origins. Given a relatively strong two child norm in many Western societies, transitions to third children can inform about preferences, particularly if related to the sex composition of the previous two children. Results show a considerable amount of adaption in fertility behavior of intermarried immigrants across all origins, both in regard to fertility levels and in regard to sex preferences for children.

Public responses to depopulation, regional and social inequalities in Latvia

By: Juris Krumins

Background: Latvia has faced considerable social inequalities and depopulation since the beginning of 1990s. All regions, both native and foreign-born population are affected by depopulation. A need for revision of attitude towards socio-demographic issues appeared on societal arena.

Research question is to analyse public responses to depopulation, regional and social inequalities and policy initiatives to promote a socio-demographic development.

Data and methods: Current statistics, «Migration and Population Policy Survey – 2016» data and data on regional socio-economic conditions and policy initiatives have been analysed, particularly by use of geographically weighted regression models (to identify direct and spatially-distributed socio-economic effects and policy interventions). Content analysis of the parliamentarian (Saeima) debate corps and documents dealing with depopulation and population policy issues has been performed.

Findings: About 4/5 of Latvia’s population consider depopulation as threat for the growth of economy on regional and national scale. Two main policy actions, according to survey results, are following: the increase of fertility and the reduction of social inequality. Among three major contributors to solve forthcoming demographic problems only one relates to the government. Support from families/households and principle “Self comes first” are considered as important by respondents (about 50% of answers). As response to massive emigration during the last economic crisis population policy initiatives have intensified – Demography Affairs Council was established by Cabinet of ministers (2011); Demography Affairs Sub-Committee was established in Saeima (2014); the Expert cooperation platform “Demography Affairs Center” was initiated by Cabinet of ministers (2016).

Conclusions: In spite of many policy documents and action plans, supported by political parties and government, consequent and efficient population policy actions to terminate depopulation and noticeably reduce inequalities have not succeeded yet.


Poster session


Older adults’ living arrangements in Estonia and other European countries

By: Liili abuladze

There is a lack of comparative research of older people’s living arrangements’ developments over time involving European countries, especially including eastern parts of Europe. Previously, Estonia has not fit into one of the existing living arrangements’ country regime typologies, based on data collected in one point of time (Iacovou & Skew 2011). Among the older population Estonia showed one of the highest proportions of people in the EU living alone (Iacovou & Skew 2011), raising questions about the potential social support resources for older people.

Research questions: There are two research questions: 1) How have Estonian older adults’ living arrangements changed over timein comparison to other European countries? 2) How can Estonia be categorized in terms of country typologies of living arrangements?

Data and Methods: Data from the Estonian censuses (1989, 2000, 2011) is used. Data for Hungary, Austria, Spain and Greece from the international census harmonized database IPUMS is added for comparative purposes. This paper looks at the 50+ population. Mostly descriptive data exploration methods are used to present the change in basic household indicators.

Findings:Findings indicate that the share of older people living alone in Estonia has increased over time, includingamong the 80+ population. Besides Austria, Estonia indicates having one of the largest proportions of older adults living alone in all observed age groups. Among the studied countries, Estonia has the lowest proportion of older people living in extended families.

Conclusions: Large proportion of people living alone may indicate that Estonia belongs to a similar category of livingarrangements as other Nordic countries. However, further analysis including data from other European, especially Nordic countries, will give a better answer to this assumption.

Does the risk of hospitalization for ischemic heart disease rise already before widowhood?

By: Elina Einiö, Heta Moustgaard, Pekka Martikainen, Taina Leinonen

Background: The death of a spouse has been shown to increase mortality from various causes, including ischemic heart disease. It is unclear, however, whether cardiac problems are already on the rise before widowhood.

Methods: Using longitudinal register data of Finnish widows-to-be (N=19,185), we assessed the risk of hospitalization for ischemic heart disease 18 months before and after widowhood. Hospital admissions were derived from national hospital discharge registers between 1996-2002. Analyses utilized population-averaged and fixed-effects logistic models, the latter of which controlled for unobserved time-invariant characteristics, such as genetic susceptibility, personality, and behavioral and medical history.

Results: For men, fixed-effects model revealed that hospitalization for ischemic heart disease increased two-fold already 0-3 months prior to the death of a spouse (odds ratio (OR)=2.09, 95% confidence interval (CI): 1.22, 3.60), relative to the period of 15-18 months before widowhood. It stayed at a heightened level up to 6 months following bereavement (OR=2.15, 95% CI: 1.07, 4.30). Among women, the fixed-effects analysis detected no statistically significant increase in hospitalization for ischemic heart disease before widowhood.

Conclusions: These findings indicate that men are already vulnerable to cardiac problems before the death of a wife. Medical interventions and health counseling could be targeted to the husbands of terminally ill patients, in order to improve their cardiovascular health over the transition to widowhood.

Love and Population Structure – Changes in Regional  Sex Ratios in Finland 1976–2016

By: Aura Pasila

Recently in Finland, the uneven number and distribution of the sexes across the country has been widely discussed in publicity. The main concern is related to the accompanying difficulties in forming (heterosexual) romantic relationships and families. Population structure has been studied in detail by calculating sex ratios in different areas (Lainiala & Miettinen, 2013). In big cities there are more women, whereas in countryside, there are more men.

However, it should be noted that the regionally uneven sex ratio is not a new phenomenon in Finland, it has been present at least several decades. In this poster, the years 1976 and 2016 are compared to illustrate this. The distribution of the sexes is also presented in more detail by adding background variables, which can affect the availability of partners and the choice of a partner. These factors, in addition to the regional distribution, may lead to very different possibilities to find partner/partners.

The analysis will be descriptive. Datasets used cover the whole population and are derived from the census database and various register-based statistics. The sources of the data are the open StatFin database and internal databases of Statistics Finland.

Transition from abortion-based family planning to modern contraceptives: the case of Latvia

By: Luize Ratniece

Induced abortion has been a significant driver of the fertility reduction in Eastern Europe. Fertility reduction in the USSR was achieved with traditional methods, scarce and very low quality versions of modern methods, and an elevated number of induced abortions. Therefore, the contraceptive revolution in post-Soviet countries meant shifting the logic of induced abortion-based family planning towards one that reliably prevents unwanted pregnancies instead of aborting them.

This paper analyses the contraceptive revolution in Latvia, which emerged from the USSR with the described paradigm of family planning and an intention to lower the induced abortion rates.

I use two representative surveys of women at reproductive ages carried out sixteen years apart in order to assess the changes in family planning strategies. I use FFS data from 1995 and a representative survey on reproductive health commissioned in 2011. Division in three age groups disentangles age effects while a comparison between the two surveys reveal cohort effects.

This data reveal a profound shift in family planning patterns. Women responding to 2011 survey report less pregnancies and at later ages. Experience of induced abortions, especially among the youngest cohorts, has dropped dramatically. Preliminary regression results confirm the importance of sociodemographic variables and socio-sexual milieu in establishing a profile of women likely to have undergone an induced abortion. Education level, choice of contraceptives at sexual debut and age at sexual debut still are linked with lower probability to have undergone an induced abortion. Ethnicity, type of habitat and use of modern contraceptives are no longer indicators of lifestyle that is linked to induced abortion. Meanwhile having only basic education has become a hazard and religious practice have become a stratifier that makes recurring to induced abortion less likely.

Demographic overview of fertility in Finland – the case of the swedish speaking finns

By: Ana Vbrink

In this article, a demographic perspective is used to provide an overview of fertility changes through history in Finland. In focus are Swedish speaking Finns, who are compared with Finnish-speaking Finns and the total population in Finland. The first part follows the growth of the Swedish-speaking population in 1610-1980. It gives an historical background to movements in fertility and differences between the Swedish-speaking and Finnish-speaking population groups at the start of the demographic transition. The second part studies newer trends in fertility in 1980-2015. In focus are similarities and differences between Finnish-speaking, Swedishspeaking, and ethno-linguistically mixed families. The main data sources are vital statistics and population censuses, but longitudinal intergenerational samples based on the Finnish population register will also be used. Standard demographic methods are used for the purpose of exploring key factors that affect births in Finnishspeaking, Swedish-speaking and ethno-linguistically mixed families in different regions. In focus are women in reproductive age, and also young people living in cohabiting unions and marriages, as they are the greatest potential carriers of birth. Some interrelation with social stratification and age at marriage have been found. Demographic factors which had influence of the overall vitality of the Swedish speaking minority in Finland were changed throughout history. In an international perspective, the causes behind variation in fertility, such as differences across periods, regions, and parental economic status, educational attainment, profession and marital status, are of utmost concern. The current main factors behind regional variation in birth rates in Swedish speakers ought to be further explored. One such aspect is the scope and dynamics of cultural identity.

The relationship between online distance tertiary education and female students’ transition to parenthood in Sweden

By: Linus Andersson

New technology drives fertility change by altering the fundamental possibilities and restrictions to demographic behavior. Today, university institutions are transforming as distance education is enabled by online platforms. As fertility prior to educational completion is central to first birth postponement, qualitative chances in tertiary enrolment brought forth by distance education represent a key point for understanding the link between technological innovation and demographic behavior. The present study is the first to examine the relationship between online distance education and the transition to parenthood during enrolment in higher education. Event history analysis of register data of all women in Sweden born between 1968 and 1991 is used to examine the impact of distance and campus education on the risk of becoming a parent while enrolled in tertiary education from 2004 to 2012 (N = 938,768). Results indicate that the negative association between enrolment and first parity conception differ substantially between campus and distance enrolment. Compared to non-enrollment, the hazard ratio is 70 % lower during campus enrolment but only 43 % lower during distance enrolment. These findings are discussed in relation to educational expansion and fertility postponement, the impact of information communication technologies on family dynamics and possible policy implications for lifelong learning and fertility promotion.

 Selective moves drive spatial clustering of fertility

By: Rannveig Hart, Janna Bergsvik, Sara Cools

Numerous studies have shown that fertility behavior is spatially clustered. While this clustering could be due to neighbors influencing each other’s fertility behavior, a similar pattern would emerge if fertility plans and/orbehavior influence moving decisions. This study explores the latter mechanism empirically. To obtain exogenous variation in fertility behavior, we use the same-sexness of the two first children and twinning at second birth as Instrumental Variables for family size. Data on full residential and childbearing histories are drawn from Norwegian administrative registers. Individual neighborhoods are defined by way of time-varying GIS coordinates on place of residence. We follow the residential and birth histories of females aged 25 to 35 who gave birth to a second child between 2001 and 2008 (~ 125 037 individuals). Our main outcome variable is whether a family (mother) has moved between the year before the second birth, and six years after the second birth. We also explore tempo effects in moving by assessing the same outcome yearly until the second child issix years old. Our results show that family size increases indeed boosts residential moves. Thus, the spatialclustering of fertility is at least in part driven by self-selection into neighborhoods based on family size.

Progression to second birth starts to decline in Sweden

By:Vitor Miranda

In countries that have completed the demographic transition, most couples still tend to report that two children is the ideal number for a family. In Sweden, a country at advanced stages of the fertility transition, behavior and attitudes on this topic greatly match: the vast majority of women who have their first child also have a second one. This in large part related to the generous family public policies aimed at supporting the balance of work and family life. Nonetheless, this study shows that this historical pattern has changed in the younger generations. Using administrative data on the total population, the results show that the proportion of mothers who did not have a second child became increasingly more common among women born after 1960. Does this suggest a weakening of the two-child family norm? Should one expect the trend to extent to other countries in Europe and elsewhere? In order to get further insights on this trend, the results are also presented by selected socioeconomic and demographic variables.

With or without you – a qualitative study on how single parents by choice reorganize their lives to facilitate single parenthood

By: Dries Van Gasse, Dimitri Mortelmans

Single parenthood is often approached as a problematic situation. People become single parents by divorce, separation or bereavement and have to cope with this situation (Fisher & Low, 2015; Pai & Ha, 2012). These transitions to single parenthood provoke unforeseen struggles in reconciling work and family responsibilities. In order to maintain the household, single parents have to both work and take up all the household roles(Bakker & Karsten, 2013). Nevertheless, there are also single parents for whom single parenthood is a positive story. These people are single parent by choice (Hertz, 2006). As it is no longer necessary to have a partner to start a family, people are able to choose to become single parent by sperm donation (Zadeh, Freeman, & Golombok, 2013) or adoption (Pasch & Holley, 2015). Hertz (2006) states that single parents by choice remain single due to circumstances but bear a strong desire to become parent. Nevertheless, single parents by choice have the same task to combine their work responsibilities and their parental roles. In this paper, we explore how single parents by choice made their decision with regard to a single parent work-family situation and how they adapted their life to facilitate their single parenthood.

Research question: How do single parents by choice reorganize their lives to facilitate single parenthood?

Data and methods: We used qualitative in depth interviews with 20 single parents by choice to explore how they organized their work-family life before and after giving birth.

Findings (preliminary): Although Single parenthood by choice is an individual decision, Single parents by choice seem to prepare a social support network.

Conclusions (preliminary): Although Single parenthood by choice is a positive story, about becoming parent, they still struggle with work-life problems

Ethnically mixed marriages in Latvia

By: Denīze Ponomarjova, Kristīne Lece

Since the collapse of the Soviet Union in 1991 to present, the proportion of ethnically mixed marriages in Latvia has not changed – it slightly fluctuated around 30% of total marriages. The tendency of marriages within Latvian ethnic group also remained stable at the beginning of 21st century –in ~80% of all cases Latvians preferred homogeneous partnership. The reverse situation is for the numerically largest ethnic groups (Russians, Belarusians, Ukrainians) – the number of partnerships within the same ethnic group has declined.

The aim of the study is to provide an overview of the tendency of ethnically mixed marriages in Latvia since 1990s. Additionally, the length of marriages, the number of common children in these marriages and average age of married persons were calculated.

Data and methodology: based on the data obtained from the Central Statistical Bureau of Latvia (including data from Population and Housing Census 2011) as well as Population Register (Latvian Office of Citizenship and Migration Affair) share of ethnically mixed marriages by brides and bridegrooms were calculated. Due to the small number of population with particular ethnic background, data for the numerically largest ethnic groups – Latvians, Russians, Belarusians – were analysed.

To research the influence of ethnical aspect to the length of marriage, 84 276 couples which have registered their marriage from 2001 to 2008 are observed. The point of reference is the legal status of the couples on January 1st, 2017.

Results: the following changes in the preferences of partner for marriage have been observed: (1) the rise of proportion of Russian bridegrooms and brides who have established partnership with Latvians; (2) an almost double increase of proportion of Belarusian and Ukrainian bridegrooms and brides who have married a person from Latvian ethnic group.

In this relation, the share of still-married couples exceeds the average value in the groups where both spouses belong to smallest ethnic groups (not Latvians, Russians, Belarusians or Ukrainians) or both spouses are Latvian.

The role of linguistic similarity and workplace in the formation of inter-ethnic partnerships in Finland between 1999–2014

By: Leen Rahnu, Allan Puur, Tiit Tamarru

The formation of mixed ethnic unions hinges on many factors that relate to people and places of encounter. We ask whether the probability to start co-habiting or marital union with an immigrant is shaped, in addition to other factors, by linguistic similarity and working with immigrants at the same workplace. Our study is based on a longitudinal Finnish register data that cover all residents who ever lived in Finland during 1999-2014. nlike many register-based studies, we focus on natives instead of immigrants. We do so since register data remains blind in the true partnership status of immigrants. Many of them have left-behind family members at their origin county and, hence, it is difficult to estimate the at-risk immigrant population for the formation of mixed ethnic unions.

Our research population is formed by the adult single native Finns aged 18 and over. We will employ the event history models in order to find out whether linguistic similarity together with the higher share of migrants at workplace establishment increases the propensity of forming an inter-ethnic partnership for native Finns, both for men and for women. We study linguistic similarity by using the linguistic tree approach. We define workplace establishment based on address and link co-workers working at this address to each other. We expect that both closeness in linguistic tree and working together with migrants increase the probability to form mixed ethnic unions for both men and women. We know from previous research that native men are more prone to form mixed ethnic unions than native women. In this study, we are interested, how does workplace mediate such relationship.

Group affiliation of children of interethnic couples in Finland

By: Obucina Ognjen, Jan Saarela

The goal of this paper is to analyze the factors determining the ethno-linguistic affiliation of children who are born to interethnic couples in Finland between 1987 and 2011. Data are drawn from the Finnish register data. We have access to a sizeable sample of the total population that contains socio-demographic information on 20% of the Swedish-speaking population and 5% of the Finnish-speaking population in Finland. We look not only at the affiliation of the first child, but also at the combined affiliation of the first and the second child. We distinguish between Finnish-speaking man/Swedish-speaking woman and Swedish-speaking man/Finnish-speaking woman couples.

By means of logistic regression, we test four principal and mutually non-exclusive hypotheses: bargaining hypothesis, awareness hypothesis, cultural plurality hypothesis and cultural supply hypothesis. Descriptive analysis confirms previous findings that a majority of children of intermarriage in Finland are affiliated to the Swedish-speaking community, but also that this trend has been increasing since the 1990s. Whereas the bargaining hypothesis is given only a partial support, the other three hypotheses are largely supported in our analyses. The determinants of ethno-linguistic affiliation of children of intermarriage also depend on the combination of gender and ethnicity. Further, our results show that in a majority of cases, the affiliation of the second child is the same as that of the first child. There are some considerable difference though, as a mixed ethnic affiliation of children is much more common in the families in which the first child is Finnish-speaking.

Under the veil of multiculturalism: The minority religion’s community experience regarding places of worship in the Helsinki metropolitan Area

By: Hossam Hewidy, Emma Terämä, Kimmo Lapintie

The backlash against multiculturalism in Europe is obvious. Multicultural urban planning in the Nordic countries is positioned within two conceptions: 1) the academic discourse dominated by the Anglo-Saxon literature; and 2) the planning policies directed at functions rather than individuals, fuelled by biopolitics (Foucault). The former is confronted with a different immigrant composition to the European receiving countries. The latter colours these policies with a ‘veil of neutrality’ with respect to cultural diversity when in fact the needs of the individual should be addressed irrespective of the source of difference.
Religion, as a component of culture, is not reflected in current urban planning in Finland. This manifests as lacking support for certain services, such as minority worshipping spaces. This research project targets the Muslim community using worshipping spaces in the Helsinki Metropolitan Area. It examines a minority religion’s community experience regarding places of worship and their influence as destinations, in shaping the mobility routines of Muslims using them. Postcodes, used as the survey respondents’ identifiers, are used to create mobility patterns in a spatial analysis. It also examines whether the community, with its limited knowledge and resources, has unconsciously created a conflict zone and negatively impacted the surroundings of these worshipping places.
We suggest that a review of the worshipping spaces and personal mobility (reflecting personal activity, need and demand) around these spaces will highlight the need for cultural awareness in the design of public spaces and transit routes.

 The old and the new part of the Norwegian population

By: Kåre Vassenden

The background of the study is a wish to better understand the population growth as it has played out in Norway in recent decades. In particular, it would be useful to map the impact of immigration on the current resident population. Figures describing the population growth belong to the core of current population statistics. Population account tables provide figures for the total growth and for the weight of each factor. Statistics on residents by country of birth (and/or parents’ or grand parents’ country of birth) make us understand the impact of immigration to Norway.

However, these sources (and even the underlying individual level data) provide data either on the timing of the different events or a possible background from some untimed immigration – not both at the same time. In addition, in respect of immigration as a factor of population growth the Norwegian classifications of immigrant background provide either a skewed dichotomy or a continuum – not a neutral dichotomy.

For these reasons, variables have been developed that distinguish between those who have been resident before a chosen date and those who have entered the Norwegian population after the same date. The point is to split the population into and “old” and a “new” subpopulation. The babies born in the measuring period are distributed according to the percentage of grandparents (or the parents when grandparents are not found) included in the old population. When the background is mixed the value old/new is drawn. Such data makes it easy to study the direct and indirect effects of the immigration of a specified period.

Among the findings are different percentages of the new population dependent on the length of the measuring period. 9 per cent of the Norwegian population as of 1 January 2016 had its background from the immigration 2006-2015. In Oslo the percentage is 14.The main data sources are event history data sets.

From Germany to Finland in 150 years

By: Linus Garp, Margareta Carlsson

Based on the demography statistics in Sweden, we can describe how immigration changed in more than 150 years; which countries the foreign-born came from and where they have settled.

To be able to get a full picture of the persons living in the country population and housing censuses have been carried out from 1860 until the late 1960s when the demography statistics came to be based on the population register. This makes it possible to get a relatively detailed picture of the structure and size of the population during a long period.

The period 1860-1900 When the first population and housing census was conducted, the number of foreign-born persons was nearly 8000. This corresponded to 2 per thousand of the population. Germany was at that time the most common country of birth among immigrants.

The immigration increased in the late 1800s, but was still relatively small. The foreign-born population at this time accounted for 7 per thousand of the population. In the early 1900s, people came to seek refuge in Sweden. The most common country of birth in 1930 was Norway. The number of persons was almost 15,000, accounting for 1 percent of Sweden’s population.

After the Second World War In the late 1940s the immigration to Sweden increased and consisted mainly of workers from the neighboring Nordic countries. Almost four out of ten of the nearly 410,000 foreign-born persons living in Sweden in 1970, was born in Finland.

During the 1970s, the work-related immigration decreased and the immigration became instead characterized by persons coming to Sweden as a result of, among other things, war and political instability. The foreign-born population in 2015 accounted for nearly 1.7 million persons, representing 17 percent of the population. Finland was the most common country of birth.

Internal migration among refugees in Sweden

By: Andreas Raneke, Karin Lundström

The aim of this study is, by the means of administrative data, to describe internal migration patterns among refugees that immigrated during 2006–2010. Asylum seekers that are granted a residence permit can either choose to organize their own accommodation, be assigned to a municipality that arranges one or immigrate as quota refugees. Quota refugees and those with an assigned accommodation immigrated in a large extend to a local labour market without larger cities whereas those with an own accommodation mostly choose to immigrate to one of the major cities. Five years after immigration, most of the refugees with an assigned accommodation had moved from their initial local labour market to one of the major cities and Stockholm was the most common destination. The majority of refugees who had organized their own accommodation or immigrated as quota refugees lived in the same local labour market after five years as at immigration. The share of gainfully employed were similar for those with own or assigned accommodation but differs to some extent between movers and non-movers. About 50 per cent of these men who immigrated to a larger city and moved to one of the major cities were gainfully employed while the share was somewhat lower among those who did not move. Also, about half of these men who immigrated to a local labour market without a larger city were gainfully employed after five years, irrespective if there was a migration or not. The share was lower among those who moved to a larger city. Similar patterns could be seen among women, although with a lower share of gainfully employed. Statistical analyses show that differences in the percentage of gainfully employed persons cannot be explained by differences in the groups’ composition in terms of sex, age, marital status and level of education.

Representations of unaccompanied minors in Nordic and US media

By: Kathleen Mufti

This paper explores and compares media portrayals of unaccompanied minors in the US and Nordic regions between 2014 – 2016 by way of qualitative thematic coding, presenting accounts of how these children are positioned in such discourses in terms of their identity, belonging and security. Unaccompanied minors, simply defined as persons under the age of 18 who cross national borders without a parent or guardian, have been a topic of interest in the media and wider public discourse in both the United States and in the Nordic countries in the recent years. In 2015, Sweden reported 35,250 unaccompanied minors that were among the registered asylum seekers; Norway 5,050; Denmark 2,125; Finland 2,535; and Iceland 5. In the US, 39,970 unaccompanied children were apprehended in 2015.  Kaukko and Wernesjö (2016), unaccompanied minors tend to be categorized by others as traumatized, vulnerable individuals while rarely considering the children’s’ own identities. Preliminary findings of this study of media portrayals present unaccompanied children as, for example, trafficked and sexually abused victims; while at other times they are criminalized and portrayed as untrustworthy and a threat to societal security. This demographic trend poses challenges and potential social inequalities to this population in their new societies.

Disability and mobility: Migration propensities and patterns among disabled and non-disabled individuals in 19th-century Sweden

By: Lotta Vikström, Erling Häggström Lundevaller, Helena Haage, Glenn Sandström

Similar to today, migrations were part of young people’s lives in 19th-century society as they sought for work and income to establish themselves and form a family on their own. Historically, little is known about how disabilities shaped people’s lives and spatial mobility. Our previous studies show that disabled people ran higher mortality risks and lower marital chances than did their non-disabled peers. This analysis aims to obtain information on the migrations and destinations of disabled individuals as this would reflect their opportunities to participate in work, social life and society. Comparing them with non-disabled individuals, who resided in the same time-space area, helps us to identify the impact disability had on people’s spatial mobility and whether lock-in mechanisms were associated with particular disabilities. This would be evidenced by low levels of migration or even immobility as authorities rejected those who were unable to provide for themselves or did not have someone supporting them to leave their parish of residence in 19th-century Sweden. Sweden’s parish registers allow studies of migration across lifetime and document individuals’ demographic characteristics including impairments. The Demographic Data Base (DDB), Umeå University, has digitized these registers of the Sundsvall region, which benefits our statistical analyses of a dataset comprising 35,000 individuals. First, we employ Cox regression models to estimate their propensity to depart from the parish with specific regard to disability and gender. Second, we conduct sequence analysis to visualize migratory trajectories and destinations in a subset consisting of 8,874 individuals under observation from 15 years of age to 33. Focusing on migration, our life course findings will indicate disabled people’s opportunities in life that work to extend and differentiate the picture of a group confronting social inequalities in societies, both past and present.

Stable or changing differences in life expectancy between various country of birth groups in Sweden?

By: Örjan Hemström

The proportion foreign born in the Swedish population has increased from 6 to 17 percent between 1970 and 2015, and it is projected to become 22 percent in 2060. The main aim of this study is to analyze if there are stable or changing differences in life expectancy by country of birth group in Sweden. Another aim is to evaluate of any such differences could be used in mortality assumptions in the Swedish national population projections. In the population projections Statistics Sweden use seven country of birth groups: Sweden, other Nordic countries, non-Nordic European nion (E) countries, European countries outside E, non-European countries with high, intermediate and low Human Development Index as defined by nited Nations.

The Swedish population registry was used to estimate life expectancy at age 30 by sex and country of birth group in two periods, 2000 to 2004 and 2010 to 2014. The contribution of various age groups to the difference in life expectancy in the period 2010 to 2014 was calculated for native born compared with those born in other Nordic countries and native born comapared with those born in non-European copuntrries with high Human Development Index.

Findings were stable between periods. Those born in Nordic countries and in European countries outside E had lower life expectancy at age 30 than the native born. Those born in non-Nordic E countries and outside Europe had higher life expectancy than the native born. Mortality differences from nearly all age groups contributes to the life expectancy gap between country of birth groups.

There seems to be rather stable differences in life expectancy between country of birth groups in Sweden. The possible use of foreign born groups with somewhat lower  and higher mortality will still lead to small effects for population projections. This is because mortality assumptions mainly effects the elderly part of the population and the proportion of foreign born in Sweden is still relatively small in ages where most deaths occur

Lineage Extinction after four Generations: A study of 19th, 20th, and 21st century Northern Sweden

By: Martin Kolk, Martin Fieder

Knowledge on patterns of the extinction of family lines are of central importance for biological evolution as well as demographic, cultural and economic change. We investigated the proportion leaving no descendants in the Swedish coastal city of Skellefteå. This was done across four generations, observed 1885 to 2007. Our analysis show that 48% of the first generation went extinct, leaving no living descendants, by the end of the period. Extinction was driven primarily by low fertility and survival in the first generation, which strongly increased extinction risk in subsequent generations. The high scale of linage extinction can have cultural and genetic consequences at the population level. Individuals who have no descendants will not pass their genes to the next generation, alleles associated with less beneficial phenotypical traits, will decrease in frequency. When large shares of the population fail to reproduce over few generations, this have evolutionary as well as social consequences.

Regional inequalities in length of life: a Nordic comparison

By: Ben Wilson, Isaac Sasson, Sven Drefahl, Paul Henery, Caroline Uggla

There has been considerable research on the life expectancy of men and women in Finland and Sweden, but relatively little comparisons between the two countries. Similarly, research has examined a number of explanations for mortality inequality within the two countries, but there has been far less focus on regional mortality inequality.

In this study, we focus on four interrelated research questions: (1) What are the regional differences in male and female mortality in Finland and Sweden? – (2) What are the differences between the two countries? – (3) How have these differences within and between the two countries changed over time? – (4) How does the comparison change when we examine different measures of life expectancy, in particular life span variation, rather than average mortality? We use register-based data for the whole population of Finland and Sweden from 1990-2012, and our analysis builds upon recent research which shows that inequalities in mortality between population subgroups are not only revealed by comparisons of average life expectancy, but also by examining measures of within-group variation (like the Gini-coefficient). As such, we focus on life expectancy at ages 25 and 65, and associated measures of lifespan variation. Provisional findings show that there are important differences by sex, region, and country, but that measures of life span variation provide new insights about regional inequalities in the Nordic countries. Comprehensive results will be presented, and the implications of these results will be discussed.

Social mobility in times of austerity and times of growth: the impact on mortality risk

By: Carolina Uggla, Sunnee Billingsley

Background and research question: Unemployment and downward social mobility both are well established correlates to premature mortality. This relationship may be due to social selection (Claussen et al. 2005), whereby poor health or health behavior leads to downward mobility or unemployment, or it may reflect social causation, whereby individuals are exposed to the health risks and benefits associated with the culture and environment of a class (Bartley & Plewis 2007; Chandola et al. 2003). In a context of economic stability, upward and downward mobility are at least partially driven by individual level factors such as career commitment. Studying the effects of downward mobility and unemployment during economic recession provides a scenario in which poor health and health behavior are less likely to be the cause of downward mobility and unemployment than in times of economic stability or growth. Data and methods: Here we use Finnish mortality data  from the early 1990s and onwards, to contrast the effect of mobility and unemployment on mortality risks during times of recession and growth, which may help us understand the strength of the selection component. We perform survival analysis controlling for a range relevant time-varying covariates. Results: we will report on our first results for our sample of men and women age 25 and over. Conclusions and contribution: As part of the TITA project, we will discuss how our findings inform our understanding of inequality in mortality risk in times of austerity.

Administrative data usage for the determination of status in employment in register based population census in Latvia

By: Anna Klusa

Background: Population census is a unique source of socio-demographic and economic characteristics of population. The Central Statistical ureau of Latvia have started the scheduled research and analysis of the employed and unemployed persons to prepare for the next Population Census in 2021. Research question: Objective of this study is usage of data from administrative data sources that characterie economic activity and status in employment (within VIP.ADMIN project of Eurostat) of Latvian population for the next population census in Latvia. Data and methods: Information on economic activity of Latvian population from Labour Force Survey (LFS), samples of the 3rd and 4th quarters for year 2014 and of the 1st and 2nd quarters for year 2015 are used in quantitative analysis and compared with corresponding data from administrative data sources.  In total, data about 96 of employed persons from LFS were found in administrative data sources.  Evaluation of main problems on detecting the status in employment from administrative data sources is made. Findings, conclusions: State Revenue Service (SRS) contains data on employed persons as well. Approximately 80 of employed persons have just one job. As regards status in employment significant differences (24.7) between LFS and administrative data sources of SRS on taxpayers were observed among own-account workers. Only 30 of employers were found in the administrative data sources compared to the estimates based on LFS. Further investigations are needed.

Educational inequality by religion, gender & generation

By: Conrad Hackett, David McClendon, Michaela Potancokova, Marcin Stonawski, Vegard Skirbekk

This first of its kind study describes differences in educational attainment among major religious groups in 151 countries. We bring together census and survey data to estimate the educational composition and average years of schooling for six major religious groups – Christians, Muslims, Hindus, Buddhists, Jews, and the religiously unaffiliated. We also show how gender gaps in education vary by religion and how religious groups have made progress over generations toward educational and gender equality. We find large differences in educational attainment between Muslims and Christians in sub-Saharan Africa and consider explanations for this gap.

Jews are the most highly educated, with an average of more than 13 years of schooling (among those ages 25 and older). Christians, Buddhists and religiously unaffiliated people – who include atheists, agnostics and those who say their religion is nothing in particular – each have a higher number of years of schooling than the global average (7.7 years), while Muslims and Hindus have the fewest average years of schooling.

Parents’ ethno-linguistic affiliation and income poverty in households with children: Finland 1987-2011

By: Camilla Härtull

The child poverty rate has increased noticeably in Finland since the mid-1990s and in the past decade it has been around eleven per cent. However, little is known about whether there are differences between Swedish- and Finnish-speaking households with children. We study this issue by comparing endogamous Swedish, endogamous Finnish, and exogamous (Swedish/Finnish) families. Of particular interest is whether any effects related to parental ethno-linguistic affiliation have varied over time.

Detailed register-based data that cover the period 1987-2011 are utilised. The poverty measure is based on equalized taxable household income, and odds of poverty are estimated with logistic regression models. The analyses are restricted to areas with both Swedish- and Finnish-speaking settlement, and they are undertaken per region and calendar year. Single-parent and two-parent households are analyzed separately.

Poverty rates increased in all major types of households during the study period, but variation by ethno-linguistic affiliation was fairly modest. Swedish-speaking single-parent households experienced a smaller increase as compared with Finnish-speaking single-parent households, and were in 2011 less likely to be found in poverty. A similar pattern could be observed for two-parent households, although there was some variation across regions, and much of the difference could be explained by educational differences and the specific municipality of residence. Patterns observed for exogamous households were less clear.

Considering that the ethno-linguistic groups studied are autochthonous and equal, the small variation in poverty rates are evidence in favour of a well-functioning welfare state, although the widening difference over time require future scrutiny by both policy makers and researchers.

Reaching beyond primary education among children placed in out-of-home care: a comparative study

By: Antti Kääriälä, Marie Berlin

Previous studies have reported that children and adolescents who have been placed in statutory out-of-home care (OoHC) for protection of their safety and welfare face considerably high risks of achieving no further education beyond primary education. Our study adds to the literature by comparing the association between children’s exposure to placement in OoHC and lack of secondary education and across three Nordic countries, Denmark, Finland, and Sweden.

We use data from national registers for children born in the country in 1987, and follow the children until age 23. The datasets for Denmark (N = 52,784, of whom 2,634 in OoHC), Finland (N = 59,476, of whom 1,190 in OoHC), and Sweden (N = 99,864, of whom 2,921 in OoHC) cover the entire birth cohort. To estimate and compare country-specific risks, we calculate average marginal effects from binary logistic regression and adjust the effects for birth mother’s socio-economic background and mental health problems.

On basis of previous research, we expect that young adults placed in OoHC as children face higher risks of achieving no secondary education across the countries. Moreover, we expect that the risks are significantly higher for boys and for those placed as teenagers, while girls and children placed at younger age and in longterm care report more moderate results. Our preliminary results support these hypotheses. In addition, we discovered some country-specific variation in the strengths of the effects. For final results, we will run more specified analysis.

The cross-country comparison shows country-specific strengths and weaknesses, thus calling for attention to identify areas of improvement in individual countries. We discuss these results and recommend developing effective interventions to improve the educational attainment of children in OoHC, also after leaving care.

Disparities in children’s family experiences by maternal education: The case of Finland

By: Marika Jalovaara, Gunnar Andersson

A well-known argument is that the trends of increasing socioeconomic inequalities and those associated with the Second Demographic Transition of family demographic change produce even greater disparities in children’s resources: children born to less educated women are more likely to experience family changes and structures associated with the loss of resources while the opposite holds true for children born to highly educated parents. This study examines how children’s experiences of family dynamics vary by maternal education in Finland. We use Finnish register data on women born in 1969­–1993. Based on the women’s childbearing and union histories (until 2009), we provide life-table estimates of children’s experience of family disruption and family formation between ages 0 and 15, stratified by maternal education at childbirth.

We find huge socio-economic disparities in children’s experiences of family transitions and family life. Children of low educated mothers are almost twice as likely to be born to a cohabiting mother and four times as likely to be born to a lone mother than children of tertiary educated mothers, who in turn are particularly likely to be born in marriage. For children of low educated mothers, the likelihood of experiencing parental separation is remarkably high: 43% of children born in a union with a low-educated mother experience their parents’ separation by age 6 and two thirds have done so by age 15. Differences in union-dissolution rates are largest at young child ages. On average, children of low educated mother spend three times as much of their childhood years in single-parent families than children of highly educated mothers, who in turn spend much more childhood time with both their (cohabiting or married) parents (81% vs. 51% of the time at ages 0-15).

Our descriptive findings provide evidence on remarkable educational disparities in children’s family experiences. The socio-demographic inequalities among children in Nordic welfare states clearly deserve more scholarly attention.

How do family structure and family process matter? A comparative study on the impacts of military deployment and single parenthood on children’s psychological wellbeing

By: Ruoqing Wang-Cendejas

Children from single parent families have poorer outcomes, on average, than children from two-parent families. Yet the mechanisms associated with family structure and family process that produce divergent outcomes are less well understood. Based on data from the 2011-2015 National Health Interview Survey (n = 18,945), I leverage the case of military families with deployment and examine the impacts of parenting quality (process), economic capital (structure), and social capital (structure) on children’s psychological wellbeing. Using linear regressions, this study confirms that children from married military families with deployment and single-parent civilian families have lower levels of psychological wellbeing than children from two-parent families. After adjustments for parenting quality and socioeconomic positions, married families, military or civilian, deployed or not, enjoy advantages that translate into positive child outcomes. These results show that family processes partially explain the impacts of family structure. More importantly, marriage emerges as the primary axis of inequality.


2.1. Families and households


Female main earners – why are they a threat to union stability?

By: Janna Bergsvik

This study deals with the consequences of income differences within households for union stability. It is a widely found phenomenon that couples with female main earners have higher dissolution risks. The mechanisms behind are however not yet completely clear. Inspired by studies by Bertrand, Kamenica and Pan (2015) and Pierce, Dahl and Nielsen (2013), I want to examine whether gender identity norms on income provision in households form an important explanation for increased union instability in couples. sing Norwegian administrative data on income in married and cohabiting couples covering the years 2005 to 2015 and a regression discontinuity (RD) design, I will explore whether couples where females contribute slightly more to the household income than their male partners suffer from higher instability than couples with a female income share just below the income-equality threshold. Studying the impact of just overstepping this threshold may enhance present knowledge greatly. Apart from a small difference in the relative income contribution of each partner, couples close to the income equality point should not be systematically different from each other, enabling us to get around typical confounding mechanisms as one partners economic dependency or general financial strain.

First analysis of the data suggest that there is no selective sorting around the income equality threshold, meaning that Norwegian couples probably do not actively attempt to avoid a situation where one earns slightly more than the other, and that there is close to random assignment to so-called treatment and control groups.

Working time arrangements, time with spouse and subsequent divorce risk: a register-based follow-up study among Finnish employees

By: Jouko Nätti, Mia Hakovirta, Satu Ojala, Mia Tammelin, Timo Anttila, Tomi Oinas

Background: A considerable proportion of workers work outside the standard working hours in Europe (EWCS 2015). Earlier research has shown that non-standard working hours are stressful for the worker and can have a negative impact on the worker’s physical and psychological health and well-being. The studies have, however, shown mixed effects of non-standard working hours on family well-being. Some studies have reported that unsocial work schedules are significantly related to an increased risk of divorce (Jekielek 2003; Perry-Jenkins et al. 2007).

Aims: The aim of the paper is to examine to what extent respondent’s working time arrangements and time spent together with the spouse are separately and combined associated with the subsequent divorce risk during a 10-year follow-up time.

Methods: The analysis is based on the Finnish time use surveys (1979, 1987-1988, 1999-2000) merged with register-based follow-up data (1980–2011). The present study is restricted to 25 to 64-year-old employees with a spouse (n=7.167). The relative risk of divorce is examined by Cox proportional hazards analyses. The results are adjusted to background and work-related factors.

Results: Both non-day work and short spouse time were separately  associated with the increased subsequent divorce risk among employees (main effects). In addition, there was a combined effect: among non-day workers longer spouse time moderated the effects of non-day work on divorce risk compared to short spouse time.

Conclusions: Employees with non-day work could benefit from longer spouse time: employees should have more possibilities to influence their working time duration and timing.

Social class and socio-economic differences in first divorce in Lithuania

By: Aiva Jasilioniene, Aušra Maslauskaitė, Vlada Stankūnienė, Domantas Jasilionis

This study extends the prior research on first divorce differentials by education and economic activity status in Lithuania by examining divorce risk by social class. The dataset is based on the linkage between all records from the 2001 census and all first divorce records for the period between April 6, 2001 and December 31, 2003. The final dataset used for this analysis covers all married individuals between the exact ages 15 and 60, and include 3.18 million person-years of population exposure, and 41 thousand first divorces. The originally available ILO-based occupation groups have been allocated into six social classes using the Erikson Goldthorpe-Portocarero classification scheme. Economically inactive people and people with unknown economic activity status and occupation (included as additional categories of the occupational variable) were also considered in the statistical modeling.

Primary results based on Poisson regression models controlling for duration of marriage, marriage cohort, and age at first marriage suggest that there is no statistically significant difference in the risk of first divorce between upper and lower white collar employees (both sexes). However, self-employed males and females showed a higher propensity to divorce than those in the upper white collar category (reference). Divorce probability was lower among self-employed farmers and farm workers, unskilled manual workers, and skilled manual employees (males only). Additional control for education, place of residence, and ethnicity led to a reversal in the effect for self-employed male farmers and farm workers, now showing the highest divorce incidence. A slight advantage of lower divorce among unskilled workers found in the initial model disappeared for males. The excess in first divorce of self-employed females and males remained statistically significant. Further steps (currently in progress) will include examining divorce risk by occupation by the branch of industry.

Anticipating divorce: is knowing half the battle?

By: Gert Thielemans, Dimitri Mortelmans

In the ongoing debate on the relationship between female employment and divorce risks, a few scholars have directed attention towards reversed causality due to anticipation effects. Women’s anticipation of the negative consequences of a pending divorce by increasing employment hours leads to an overestimation of the employment-effect on divorce risks. However, the implications of the existence of anticipation also extend into the study of the consequences of divorce. Furthermore, while the concept of anticipation is frequently used in social research, interpretations of its meaning vary widely, uncovering the need for conceptual clarity.

Methodologically, grouping women who anticipated a divorce and those who didn’t is likely to underestimate the negative effects for the latter group. From a substantive point of view, women who anticipate divorce by increasing their employment have a head start in coping with negative consequences. This potentially leads to long-term differences in women’s well-being after divorce depending on whether or not they were able to anticipate.

After proposing a clearly defined definition and properties of what should be considered ‘anticipation’ and what should not, we study the existence of anticipation in the context of divorce and female employment by analysing longitudinal data from the Gender & Generations Programme (GGP) on several European countries. The main contributions of our research are:

– We provide conceptual clarity for the notion of ‘anticipation’, specify a definition and assign the concept with properties making it suitable for systematic research.
– Former research has limited itself to modelling anticipation in order to estimate the effect of female employment on divorce risks, rather than looking at the existence of anticipation surrounding divorce.
– We depart from the tradition of research on the gender divide by aiming to uncover the differences between women of different socio-economic backgrounds.


2.2. Health, morbidity and mortality


Disability, life and death in a past population: Life course analyses of disabled and non-disabled individuals in 19th-century Sweden

By: Lotta Vikström, Erling Häggström Lundevaller, Helena Haage, Sören Edvinsson

From contemporary disability studies, we know that impairments tend to jeopardize individuals’ health status and that disabled people run the risk of constituting the ‘otherness’ in society. In history, most disability studies concern institutionalized individuals for whom the sources tell little about their lives beyond institution; and if the sources do, they usually document only a few persons with disabilities. The research design and data of this paper enable us to detect how disability shaped human life and death in a past population comprising some 35,000 people in the 19th-century Sundsvall region, Sweden. First, we conduct event history analysis (Cox regression models) to investigate their premature mortality risks, and then examine their death causes. Comparisons are made across different disabilities, the genders and with a reference group of non-disabled cases. Second, we make sequence analysis to obtain knowledge on how individual life trajectories developed with regard to work and family formation and eventually ended in premature death. Sweden’s parish registers (digitized by the Demographic Data Base, Umeå University) work as key to identify people defined as disabled or not, whom we can observe across their lifetime. Our regression results show that disability significantly jeopardized their survival propensity, particularly of men, but with some variation by type of disability. Our trajectory findings reveal that disability implied difficulties for individuals in both the labor market and marriage market, which help explain the premature death of disabled people although they avoided infectious diseases compared to non-disabled.

Inequalities in sickness absence between Finnish and Swedish speakers in Finland: a register-based study

By: Jan Saarela, Kaarina Reini

Sickness absence is as an indicator of poor or reduced health. Periods of sick leave increase the risk of subsequent work absence, while long-term sickness absence increases the risk of adverse economic and social conditions. Previous research has shown that, in terms of both disability pension and mortality, Finnish speakers in Finland are less healthy than Swedish speakers. Many surveys indicate also that Finnish speakers have poorer self-reported health than Swedish speakers, while some studies even suggest there being differences in mental health and well-being between the two ethno-linguistic groups.

We examine whether Finnish speakers and Swedish speakers differ in health with regard to sickness absence, which is a previously unexplored objective health measure in this context. Based on previous research, we expect that, also when controlling for socioeconomic and demographic background variables, Finnish speakers are more likely to be on sick leave than Swedish speakers.

The individual-level data used come from the Finnish longitudinal population register. They cover the period 1987-2011, and consist of five per cent of all Finnish speakers and 20 per cent of all Swedish speakers. To estimate the likelihood of receiving sickness benefit in prime working ages, and account for multiple occurrences at the individual level over the study period, we estimate logistic regression models with generalized estimating equations.

We find that, even when we control for a number of socioeconomic and demographic background factors, Finnish-speaking men are on average 30 per cent more likely to receive sickness benefit than Swedish-speaking. In women, the difference is approximately 15 per cent.

The results corroborate previous research based on other objective health measures. Since sickness absence causes substantial costs for the society, these results can be utilised in policy making processes that aim to lower sickness absence rates and thus help in equalising health differentials.

Lengths and diagnosis of sickness absence spells as predictors of disability retirement in different socioeconomic group

By: Laura Salonen, Jenni Blomgren, Mikko Laaksonen, Mikko Niemelä

BACKGROUND: It is well known that especially long-term sickness absence and low socioeconomic status predict disability retirement among other factors. Low socioeconomic status is associated with longer sickness absence spells and certain types of diagnosis. Simultaneously the probability of disability retirement also depends on the diagnosis of sickness absence. However, less is known, how the length and the diagnosis of sickness absence predict disability retirement differently among socioeconomic groups.

RESERACH QUESTION: In this paper we are examining are there socioeconomic differences in the probability of transferring to disability retirement when focusing on the length and diagnosis of the preceding sickness absence.

DATA AND METHODS: We apply Cox proportional hazard regression model on multi register data from Insurance Institution of Finland, Finnish Centre for Pensions and Statistic Finland. Information on demographics and sickness absence are gathered from 2005 and each individual is followed  for 7 years, starting from 2007 and ending in 2014 (N=1,674,414) or until either receipt of a disability pension or death.

RESULTS: Our preliminary results show that there are socioeconomic differences in the probability of transferring to disability retirement depending on the lengths of sickness absence spells. The predictive value of the length of sickness absence also differs depending to the diagnosis of the sickness absence and the differences vary among the socioeconomic groups.

CONCLUSION: Our preliminary conclusion is that work disability seems to develop differently among  socioeconomic groups. Therefore, more attention should be paid on how socioeconomic status is associated with the prognosis of work disability. Knowledge on this relationship can help policy makers to recognise work disability risks earlier and to develop more effective actions on preventing prolonged work disabilities

Pre-retirement labour market position and main-cause mortality after retirement: A longitudinal study from Finland

By: Julia Klein, Jan Saarela

In the face of population ageing and strained government budgets, the discussion of whether working life ought to be prolonged or shortened in order to promote health among pensioners and consolidate the pension system. It is therefore important to gather evidence of the health effects that are linked to the ending of the economically active life. One aspect of this topic is how the transition from the labour market into retirement is related to health.

Different labour market states are associated with differing degrees of ill health, and death is used as a crude proxy for very ill health. To study this association, we quantify the effect of various pre-retirement positions on the labour market on main-cause mortality after having reached retirement age.

The data stem from a five per cent random sample of the Finnish population, covering the period 1987-2011. We follow 30,000 individuals from age 50. Hazard models are used to quantify the association between labour market status in ages 50-64 years and main-cause mortality in ages 65-70 years.

Preliminary results show that for both sexes, the risk of dying early is about twofold among disability pensioners as compared with people who were employed. Being unemployed at age 50+ is also associated with reduced life expectancy at age 65+.

The adverse health effects associated with unemployment can to a high extent be explained by socioeconomic and demographic variables, while the effect of being a disability pensioner is rather independent of socioeconomic background. The results can be used to help in equalizing the adverse health effects associated with certain labour market positions.

The effect of retirement on male mortality. Quasi-experimental evidence from Norway

By: Adrian Farner Rogne, Astri Syse

On average, people who retire earlier tend to die sooner, partly because poor health is an important reason for early retirement. A number of theories also predict that retirement is detrimental to health and increases mortality. Yet, there is apparently no consensus about the causal relationship between retirement and mortality.

We treat a Norwegian pension reform that lowered retirement eligibility age by three years in 1973 as a natural experiment. The reform affected virtually the entire elderly working population. This approach allows us to estimate the long-term effect of retirement age and the short-term effect of retirement eligibility among men, using instrumental variables and difference-in-differences methods.

The results show that earlier retirement or retirement eligibility does not affect male mortality to any substantively meaningful degree. Separate analyses by occupational groups indicate that the average effect estimates do not conceal heterogeneous effects between occupational groups. Finally, we argue that the 1973 reform serves as a critical case; a situation where we should have been able to identify an effect of retirement on mortality if there were one.

This study provides strong evidence that retirement does not affect male mortality.


2.3. Migration


Spatially explicit population projections: minding the urban-rural interface

By: Maija Mattinen, Antti Rehunen,Ville Helminen, Emma Terämä

Demographic change is an important driver which will influence society and environment. There is proven demand for spatially explicit population projection information including population structure. Local decision makers, planners, building industry and others can make use of this type of information.

Our research questions are the following: How does the use of a non-administrative spatial typology affect population foresight? Can existing estimations and projections be improved by using more accurate jump-off population structure?

We use a bespoke Finnish typology of urban-rural classification as a basis for sub-national, region-specific population projections, and implement our own R-script to account for the differences compared to the national statistical office-generated data. hilst the national statistical office (Statistics Finland) acknowledges and uses the urban-rural classification developed by the Finnish Environment Institute, they do not use this information as a basis for projections.

The urban-rural typology is based on the combination of population density and built-up area, and includes seven classes in total – of which regions studied here are allowed to contain 3-4 each. In addition, we have improved migration statistics based on household specific data obtained from the Finnish Building and Dwelling Register (maintained by Population Register Centre).

We find that via the inclusion of population density and built-environment based typology for urban, peri-urban and rural regions, we can better project the local urbanisation and rural decline. e can improve on the spatial understanding of migration events within regions, as well as allocate intra-regional migration (and international migration) better to locations where they are likely to belong.

Via improving on the spatial understanding of migration events, we can couple this to workflow migration based on economic trends etc.

Effects of extreme values on migration modelling and forecasting

By: Violeta Calian

Based on population register data and time series analysis, Statistic Iceland provides models and forecasts of the migration components needed for population projections. We have built dynamical (auto-regressive distributed lag) models in order to obtain point estimates and confidence intervals of the number of immigrants/emigrants of Icelandic and foreign citizenship as functions of several time series predictors: unemployment, change in GDP, the number of graduating students and dummy variables mirroring the EEA resizing  and the Icelandic economic boom.

The economic and social variables had strong fluctuations, especially in recent years. This is why, we applied a systematic analysis of the rare events in the employed time series: a) the outliers and the time scale of their effects, b) the  extreme values‘ probability distributions and impact. We used hypothesis testing, the distributions of the models‘ residuals and the past distribution of shocks for this purpose.

The main focus of this paper is to analyze these models focusing on the prediction uncertainty for migration forecasts as an informative measure, to discuss methods of validation for such measure and the appropriate methods for its dissemination. We adopt the strategy of separating the prediction uncertainty into components related to the statistical modeling errors and to the unpredictable shocks due to natural or political causes

Assessment of policy Implications of ethnic segregation through a socio-demographic decomposition of its dynamics

By: Timo Kauppinen

Selective intra-urban migration of different ethnic groups is often assumed to be the main micro-level mechanism (re)producing ethnic residential segregation, and it may be the politically most salient one, as it may indicate preferences for co-ethnic neighbours or constraints against spatial integration. However, also the contributions of positive natural change and international migration among ethnic minorities may be important, especially when larger-scale immigration is more recent. If development of ethnic segregation is driven to a considerable extent by these two processes, the implications for policy differ from those related to intra-urban migration. Implications of ethnic segregation also depend on its connection with socioeconomic segregation. This connection may change even if ethnic segregation itself does not change. Therefore, it is important to consider the role of social mobility in the development of the connection between ethnic and socioeconomic segregation. This paper presents analysis utilizing longitudinal individual-level register-based data on complete populations of the three largest urban regions in Finland, aiming to find out 1) To what extent are changes in ethnic segregation related to intra-urban residential mobility and to what extent to other demographic processes, and 2) How do migration flows and in situ social mobility affect the relationship between socioeconomic and ethnic segregation? The implications of the results in terms of challenges and opportunities in different policy sectors are assessed.

Residential segregation and local housing policies in the three largest urban regions in Finland

By: Jarkko Rasinkangas

Local housing policies are one important determinant of low-income households’ access to housing in a local urban context. The research that I would like to introduce at the Nordic Demographic Symposium is an analysis of local housing policies and practices in the three largest urban regions in Finland (the Helsinki, Tampere and Turku regions), from the perspective of access to housing and residential segregation among low-income households. Population forecasts indicate that growth will be the highest in these urban regions and is based a lot on migration, where immigrants have a key role. Segregation levels are known to differ between the regions and this may relate to the differences between them in the application of social mix policies, in income inequalities, and in immigration histories, for example. In the study, the focus is on local housing policies and housing allocation strategies for vulnerable groups. These are studied with data based on policy documents and interviews of the local key actors especially in the allocation of social rental housing. The key question is to study what kind of practices on access to housing the local gatekeepers and general guidelines have and what might be influence on the segregation. Document material and interviews will be carried out in the spring 2017. The first preliminary results might be possible to present at the time of the symposium.

This research is a part of a multi-disciplinary research consortium Urbanization, Mobilities and Immigration (URMI), funded by the Academy of Finland.


3.1. Families and households


The cohorts of convergence? Danish women and the changing paradigm of women’s labour market participation

By: Luize Ratniece

As the pioneering cohorts are ending their labour market careers, now is the time to map out the advancement that women have made towards convergence with men in labour market achievements. I explore the extent of convergence of the labour market trajectories and closure of the gender gap in Danish labour market for 1941-1980 cohorts using Danish register data.

The novelty of this article lies in a longitudinal approach and use of sequence analysis that includes the intensity of labour market attachment as a key aspect of analysis. Such design ensures the capacity to analyze the working lives as a dynamic phenomenon, capturing the labour market trajectories with greater precision than any cross-sectional data ever could. A comparison of four subsequent birth cohorts permits to follow the gradual change of milieu in the population for 26 consecutive years between 1986 and 2011. I apply sequence analysis to create two longitudinal trajectories, one of labour market related activity and the other of household composition. Once the clusters are created, I explore these sets of clusters descriptively and apply regression analysis to examine how household composition is related to labour market activity.

Labour market trajectories for both men and women are dominated by full-time work. Private sector jobs dominate the trajectories of men while public sector jobs dominate those of women, especially in the older age groups. The division between full-time and part-time work has a negligible long-term impact as a defining characteristic of the labour market trajectories.

While the strong labour-market attachment of Danish women is very clear, so is the fact that at least so far they have tended to build their careers in the more sheltered public sector. While even the youngest age group does not reach complete gender convergence, the 1971-1980 birth cohorts is the one with the most similar results for men and women

Back to work and stay-at-home mother. Maternal employment in Finland

By: Eva Österbacka, Tapio Räsänen

The labor force participation among Finnish women is comparatively high, almost as high as among Finnish men. However, a feature of the Finnish parental leave policy is that the maternal and family leave periods are relatively long and in addition, the child home care subsidy makes it possible for parents to stay home with the child(-ren) until the youngest child is three years of age. This family policy prolongs the parental breaks among women with children since the family leaves are predominantly used by mothers. Child home care subsidy is a distinguishing feature for Finland. The policy is not as generally used in any other Nordic country, where universal family policies are prevalent. However, the home care subsidy is not used by all mothers. The majority of mothers use the child home care allowance at least some time but there is a large variation in the durations.

In this paper, we investigate the impact of childbirth and family leave policies on mothers decision to return to the labor force. e scrutinize factors affecting the return to the labor force after childbirth, and especially characteristics of mothers (such as age, education, and previous work experience), family characteristics (such as fathers labor force participation and immigration status), effects of business cycle (such as unemployment rate), and policy effects (such as municipal supplements for home care allowance, municipal support for private care, and the speed premium implemented in 2005). e analyze the effects of covariates in several ways visually, with Kaplan-Meier curves, and with hazards models. Our data is a 60 random sample of Finnish mothers giving first birth between 1999 and 2009, about 11 000 mothers per year. Each cohort is followed up until 2013. The results will have both academic and policy relevance.

Fertility timing and men’s later-life earnings: evidence from siblings

By: Elina Einiö

Background: Young mothers are known to be vulnerable to interruptions in career building and to wage loss. However, few studies have investigated the effect of fatherhood timing on men’s later-life earnings. Research question: Using a large sample of Finnish siblings with children, the study investigated the subsequent disadvantage of early fatherhood on men’s later-life incomes. Data and Methods: The sibling model employed reduced the selection bias related to unobserved social and genetic characteristics common to siblings. Findings: The results show that fatherhood penalties to later-life income are substantial when the transition to parenthood occurs at a young age. However, delaying fatherhood further after the age of 30 does not further improve men’s later-life incomes. Conclusions: The study suggests that the increases in male earnings cease after the transition to fatherhood has been delayed past the critical period of career building.

The effect of children’s residence arrangements on the work-family balance among divorced families

By: Annelies Van den Eynde, Dimitri Mortelmans

Balancing work and family is a challenge for many men and women nowadays. It is the rising number of dualearner families and higher job demands that suggests that the pressure on the private organization of work and family has increased (del Carmen Huerta et al., 2011; Keene & Quadagno, 2004). This paper contributes to the literature of work-family balance by focusing on different family constellations than the dual-earner couples which dominate this study field (Gatrell, Burnett, Cooper, & Sparrow, 2013).

The current study examines the effect of children’s residence arrangements on the work-family balance of divorced parents. In divorced families, the children’s residence is formally divided between two separate households (Willekens, Vanderheyden, & Mortelmans, 2011). The classic model entailed a dominant stay with the mother, while now a shift is recognized towards more shared residence between the two parents (Sodermans, Vanassche, & Matthijs, 2011). Children’s presence or absence in the household of a divorced parent determines the available time and energy they can devote to either family or work. The study uses survey data collected in 2008 in Flanders, Belgium. The multi-actor study “Divorce in Flanders” enables us to compare different family constellations after the divorce such as single parents, parents who cohabiting with a new partner, remarried parents, and parents in a LAT relation. Also the form of residence arrangement of the child is available in the questionnaire and can take different forms such as living full-time with one parent, shared residence, and week mother/weekend father or week father/weekend mother.


3.2. Life Course


Life course perspective on economic shocks and income inequality through Age-Period-Cohort analysis

By: Esa Karonen, Mikko Niemelä

This research is part of TITA-consortium’s (Tackling Inequalities in the Times of Austerity) working project 1.1. This research paper strives to study intergenerational differences in equivalent income across cohorts of Finnish households by quantitative means using life course perspective. The main research questions are: how the intergenerational income trajectories will differ between generations and how the attainment of middleincome has changed over time?

The research utilizes cross-sectional data from the Finnish Centre of Statistics which is composed of two register based datasets: the income distribution (1987-2014) and household’s consumption (1966-1985) statistics.

This research utilizes new age-period-cohort models which are designed to detect how an outcome is explained by the position in the life cycle (age), date of birth (cohort) and the time of the statistical measurement (period). This is achieved by two different approaches: measuring relative and absolute changes in income distributions. First, Age-Period-Cohort (APC) “detrended” -approach focuses on how the effect of age, period and cohort fluctuate around a linear trend, which it absorbs. The second method, APCT (T for “trended”) is developed as an instrument to measure absolute declines and progressions.

The results will present how the different cohorts are holding up regarding income development and how period and age effects are affecting income trajectories when control variables are taken account. Central focus of the research is to observe the changes in relative and absolute income distribution, which will bring more holistic insight how the standard of living has changed over different generations. The results will contribute the questions regarding the income dynamics of the intergenerational differences and hopefully illustrate the effect of the welfare state on cohorts.

With these methods, this paper aims to report cohort changes in intergenerational level and find indications if there are differences at the level of income inequality.

Finnish life-courses in a comparative perspective

By: Kathrin Komp

Life-courses are of central importance to the social sciences. They describe how lives progress from the cradle to the grave, highlighting activities and events. As such, the life-course is an important concept for capturing mid- and long-term effect of, for example, job loss and the 2008 crisis. Recently, country-comparative lifecourse research bloomed thanks to the life-history data of the Survey of Health, Ageing and Retirement in Europe (SHARE). Unfortunately, Finland did not participate in previous waves of SHARE. Consequently, information on Finland is missing from many country-comparative life-course studies.

This presentation fills this lacunae. It uses data from the NoWork project to describe life-courses in Finland, explore social inequalities between them, and compare them to the life-courses in other European countries. The NoWork project conducted life-history interviews with 700+ Finns aged 50+ years in 2016/2017, using an excerpt from the SHARE life-history questionnaire. Sequence and cluster analyses are carried out with the NoWork data. Moreover, the result are compared to publications using SHARE data. Findings show that many Finnish lifecourses follow the tripartite life-course model, which splits lives into subsequent phases of education, work, and retirement. The prevalence of this model is higher than in most other European countries, while the variation among life-courses is lower than in many European countries. The reason for this pattern is the persistently high female labourforce partition rate in Finland. Findings indicate that the tripartite model is suitable for capturing Finnish life-courses. Moreover, they indicate that dependency ratios have a high explanatory power in Finland. Dependency ratios are measures for the progression of population ageing, which are calculated using the structure of the tripartite life-course model. With life-courses like the Finnish ones, these ratios can capture changes in population structures with high accuracy.

Organising demographic data sets of life course quality

By: Kåre Vassenden

Norwegian basic population statistics data are stored either as stock data files or as flow data files. These files are produced shortly after the reference point in time or the end of the reference period.

The events included in the statistics of one reference year are not supposed to be included in the statistics of another year, and no real events should fall outside the counting.

In general, the research community asks for event history data. Immediately, it is natural to think that each statistical file with flow data – or several annual files combined – directly provides event history data of the quality researchers want or need.

This is not the case, however. Correction data received after the extraction date are not necessarily included. The consistency of the events across the annual files is not checked. Consequently, one immigration event one year may be followed by a new immigration the next year. Before the turn of the century some delayed data were not included in the official statistical files. In some old data files the same event are sometimes found in two different annual files. In any case really old data may be of unsatisfying quality.

For all these reasons it is desirable to establish separate event history files for the demographic events. These files should be of “life course” quality – meaning that the data is of high internal and external consistency.

Demographic research papers using and presenting event history data abounds, but papers discussing the general data requirements are few and far between. An aim is to remedy this situation. The presentation will include both a discussion of useful concepts and principles and practical experiences from different projects. Most important so far have been the establishment of a data file of 12 million births, deaths, immigration and emigration events (for all the people in the Norwegian population register). A thoroughly revised file of citizenship allocation and changes since the 1970s is another example.

Indexes in statistical demography

By: Ene-Margit Tiit

When using register-based statistics sometimes problems connected with accuracy of registration occur. In register-based census such problems are the actual living place and actual partnership. Especially important is solving such problems in preparation of register-based census. In Estonia we have created the methodology of indexes where so-called signs of life have been used. This methodology has been applied since 2015 in practical calculation of population size and international migration.

As the next step also index of partnership between two persons h and j from the set of all possible pairs (i, j) of residents of the country has been created in the following way: P(h,j)=∑_(i=1)^m▒a_i  P(h,j;i), where different signs of partnership P(h,j;i),i=1,2,…,m are used, indicating the partnership between two persons. Such signs are, for instance, common children, common property (real estate, car), common duties etc. As sources of information for these signs different registers, but also banks and sources of so-called big data can be used. The weights a_i have been calculated using the test (learning) data. There are two possible approaches: either to use classical method of statistics (logistic regression) or calculate the weights a_i proportionally to the frequency (or chance) of occurring given sign among the pairs of persons forming actual couples. The main task here is finding enough signs of partnership for couples registered in different dwellings (with the aim of getting some bonuses), but de facto living together. Finding such couples is quite serious challenge in preparation of register-based census in Estonia. The testing of the methodology using Estonian data is in process. If it turns out to be useable and confident, it will be used in the forthcoming Estonian register-based census in 2021.

Nordic cooperation in microdata access

By: Claus-Göran Hjelm, Anton Örn Karlsson, Marianne Johnson

The Nordic National Statistical Institutes (NSI’s) have over the years collected and stored a huge amount of microdata for official statistical production. This microdata is also a valuable source for researchers in the Nordic countries. Although each of the Nordic NSI’s have a well formed process with regards to researcher’s access to microdata, no cross-national processes or protocols dealing with data requests involving two or more NSI’s have been available. Because of interest from the Nordic research community the Nordic NSI’s have started a two year project in January 2016 with two main goals: 1) To test and implement a common Nordic model of microdata access for researchers and 2) propose a common metadata framework for cross-Nordic research projects. The presentation will be aimed at describing the Nordic model for microdata access as well as the project, its background, main goals and its structure. It is foreseen that this project will increase the ease of applying for cross Nordic microdata access for research purposes and thus increase the satisfaction of users from the research community.


3.3. Migration


How many refugees’ families come to Norway?

By: Minja Tea Dzamarija

According to the Norwegian Directorate of Immigration, more than 31 000 people applied for asylum in Norway in 2015, mostly from Syria, Afghanistan and Iraq. Questions have been raised about the need to restrict the right to family immigration for refugees. In the wake of the refugee crisis, it has been argued that the record number of asylum seekers in Norway in 2015 will trigger a marked increase in family immigration. Research question and findings: Family immigration has always played a key role in the growth of immigrant groups. This paper examines immigrants who came to Norway in the period 1990–2015 as the family member of a refugee who arrived during the same period.  The main question is: how many of those refugees who have been granted residence in Norway have been able to get family members to Norway? What is the scope of family-related immigration to Norway? A total of 141 300 refugees generated 45 100 family members, that is an average of 0.32 family members per refugee during the period. This rate does not, however, take into account three important dimensions of family immigration, which we will examine in this paper:

Time aspect, i.e. the time elapsed from when a refugee is granted residence until a family member arrives through family immigration. Finding: As time passes, the number of family immigrants ac-cepted into Norway will increase as the reference person, in this case a refugee, fulfils the various requirements for family immigration.

What country they come from? There are major disparities here.Finding: Somalis and Iraqis have highest family immigration levels, Bosnians and Kosovan Albanians have the lowest.

Multi-stage family immigration: this type of immigration that takes place via a refugee, but indirectly through various stages has a certain inpact.

Finding: The rate increases from 0.32 family immigrants per refugee to 0.35.

Data and methods: Data used in this paper combines information from two data sources: Statistics Norway’s population database (demography) and The Alians Register (reason for immigration

Do institutional entry conditions matter for migrants’ labor market integration?

By: Yulia Kosyakova, Herbert Brücker, Agnese Romiti

We study  how the introduction of free movement of workers (FMW) for citizens from countries participating in the Common Single Market affected migrants’ labor market integration in Germany. We focus on two complex issues: Firstly, whether and how FMW impacted on the composition of the migrant population in terms of human capital characteristics such as qualification, language proficiency, cognitive abilities, and professional experience. Secondly, we analyse whether and how FMW affected migrants’ labor market integration as measured by the duration of getting the first job and its quality. Exploiting rich longitudinal data from the IABSOEP migration sample, we identify FMW effects on migrants’ outcomes by imitating a regression discontinuity design. Findings reveal that the introduction of the FMW indeed had a substantial impact on the composition of the migrant population in form of a higher influx of skilled and educated workforce. Findings regarding integration outcomes were ambiguous. While FMW improved migrants’ labor market integration when measured by the time to first job significantly, the results, however, also point to a potential increase of qualification mismatch in migrants’ first jobs after the introduction of FMW. Overall, our findings suggest that FMV increased welfare by (i) raising the scale of migration and the skill level of migrants, and (ii) reducing the waiting-time for labour market integration and, hence, the unemployment rate of new arrivals

Data on Immigrants permits of residence

By: Annemette Olsen Lindhart

Statistics Denmark has received data from the Danish Immigration Service on all permits of residents issued from 1997 and onwards on personal identification number and now receives data on issued permits quarterly. All immigrants who do not hold a Danish or Nordic citizenship are required to have a valid residence permit in order to be included in the Central Population Register. Statistics Denmark has accordingly matched all immigrants without a Danish/Nordic citizenship who have immigrated since 1997 with the data from the Danish Immigration Service. The data on residence permits subsequently covers 68 per cent of the immigrant population per 1 January 2016.

In order to get full coverage of immigrant’s reasons for immigrating to Denmark, Statistics Denmark will conduct an alternative method of imputation on immigrants. Since there is no data on residence permits on PIN earlier than 1997 there will be created an imputation method which is based on knowledge of types of residence permits issued by citizenship and year and combine it with the data on the immigrants in the Central Population Register on citizenship, year of immigration, age at immigration and marital status. The presentation will include a description of the method and of the data used for the imputation method including a mapping of the results on the Danish immigrant population as well as a discussion of the pros and cons of this method of imputation concerning reasons for migrating.

The study and the considerations on whether it will be possible to use this method to improve data on reasons for immigrating will be conducted in the spring 2017 and therefore there are no results in the abstract.



4.1. Fertility


Understanding the factors that influence childbearing over the life course: causal links between employment, partnership, socio-economic status and fertility

By: Ben Wilson

It is well known that childbearing is associated with other life course processes like partnership and employment. However, there is very little research that studies the causal links between these processes. We study a birth cohort of women who were born in the UK in 1970, and demonstrate a new approach for the analysis of causally interrelated family events. This approach uses the g-formula, an advanced method for causal inference, to examine the causal links between employment, partnership, socio-economic status and fertility, while also controlling for important factors like age and education. Our results show that a significant reduction in fertility rates would be caused by a reduction in marriage rates, and to a lesser extent by a rise in full-time employment immediately after giving birth. However, after taking account of causal mediation, we find no impact of parental socio-economic status on fertility under plausible scenarios of upward social mobility.

Couples’ employment dynamic during the great recession and their transition to parenthood

By: Chiara Ludovica Comolli

The aim of this paper is to investigate how various aspects of the economic uncertainty generated by the Great Recession combined and affected fertility behavior among American couples, focusing in particular on the transition to the first child.

The theoretical argument followed here points at the manifest growing difficulties in the labor market in the period 2008-2010 in the US, and at how different employment dynamics within the couple might have produced heterogeneous consequences in terms of childbearing decisions. The couple perspective is, with respect to the existing literature, the first important innovation of this paper. The second distinguishing feature of the paper is that it also addresses the interplay between individual and aggregate-level labor market conditions in shaping the probability of having children, testing the existence of a multiplicative or moderating effect of contextual conditions on top of individual level circumstances. Using PSID dataset I show that, especially at high rates of unemployment, dual earner couples are more likely to enter parenthood compared to any other couple tipology.

Economics and the timing of first birth after the financial crisis

By:Trude Lappegård, Lars Dommermuth

The aim of this paper is to study the influence of economic resources on timing of first birth after the financial crisis in 2007/08. We focus on Norway which is an interesting case, as the country is recognized for generous family policies and its robust economy where the financial crisis led only to a slight increase in unemployment rates. Nevertheless there has been a steep drop in fertility level since 2009. In the Norwegian context the recent changes with lower fertility levels and increased postponement in the transition to motherhood is somewhat puzzling. It is well known that women’s labor market participation, education and economic resources are important features for explaining variations in women’s timing of first birth. Whether and how the importance of these factors is changing over time is less noticed and documented. Using detailed information from administrative registers and event history models we explore whether the relationship between individual economic resources and macro-economic conditions and the timing of first birth has changed after the financial crisis. If we find a change in the influence of economic resources on the timing of first birth one interpretation would be that women’s fertility is very sensitive to economic factors which cannot be compensated by generous welfare policies

Fertility at the dawn of a new millennium: The interplay of family policy, economic crisis and childbearing in Iceland

By: Ari Jónsson

Background: In the first decade of the 21st century, two potentially influential events took place in Iceland in relation to subsequent fertility outcomes: a reform was enacted in the parental leave scheme between 2001 and 2003, and a deep economic crisis came ashore in late 2008.

Objective: The purpose of this study is to evaluate the effects these two events had on first-, second- and thirdbirth intensities.

Data and Method: y means of event history analysis we analyse individual longitudinal register data, consisting of the total female population of relevant ages, and present the findings as relative risks of giving birth during 1998-2013.

Findings: In terms of results, we find that after the parental leave reform was implemented the propensity to have a second and a third child increased constantly until 2010, while a declining trend in the age-standardied first-birth rate came to a halt and the first-birth intensities stabilied. After the onset of the economic crisis in 2008, a trend of decreasing first-birth intensities re-emerged. In 2011, three years into the crisis, we see a turnaround in second- and third-birth rates, which began to decline and continued to do so until the end of the study period.

Conclusions: The development in the standardied second- and third-birth rates indicates that the reform had a positive influence on continued childbearing. Albeit the observed changes in the parity specific rates occur at different time points, the economic crisis appears to be associated with a general decline in the birth intensities. Simultaneous decrements in parental leave benefits may be partially responsible for the post-crisis decline

The labor market and the transition to third birth

By: Lars Dommermuth, Trude Lappegård

The aim of this paper is to study the relationship between the labor market and the transition to third birth. The proportion of women having three or more children has decreased gradually during the past decades in Norway. At the same time as there has been an increase in generous family policies and a fairly stable economy. Traditionally, women working in the health and caring sector had a comparatively high fertility. At the same time many of these women worked part-time but over the last years there has both been a supply and a demand for more full-time work.

In this paper we ask whether the decrease in the proportion of women having three children or more can be related to changes in womens participation in the labor market. sing detailed information from administrative registers about labor market conditions, as well as socio-demographic characteristics, over the last 16 years, and applying event history models we estimate the probability of having a third child. If our result show that one driver behind lower third birth rates is that womens participation in the labor market has changed one could argue that generous family policies as a condition for fertility cannot be seen independent of labor market policies.


4.2. Educational inequalities


The unequal opportunity for skill acquisition over the Great Recession in Europe

By: Sara Ayllón, Natalia Nollenberger

This paper is the first to investigate to what extent the high levels of joblessness brought about by the Great Recession across Europe have translated into higher school attendance among youth. Using cross-sectional and longitudinal data from the EU-SILC on 28 countries, we establish a robust counter-cyclical relationship between rising unemployment rates and school enrollment. The same is true for transitions back to education. However, our analysis by subgroups reveals a worrisome trend by which youths belonging to most disadvantaged backgrounds (measured by low household income) became less likely to enroll in University studies. Our findings suggest that austerity measures and educational cutbacks imposed during the economic downturn have made the opportunities for skill acquisition more unequal according to socio-economic characteristics.

Something good out of the bad times? Inequalities in college enrolment during great recession in the USO

By: Heta Pöyliö

Previous literature has well demonstrated the positive relationship between recession – and more precisely unemployment – and educational enrolment; when labour market opportunities are few, the cost and risks of education diminishes resulting to an increase in enrolment. Even though this holds for the whole population, analysis with American Community Survey shows a steady increase in immediate college enrolment after high school over time, regardless of the impact of the recession on unemployment. However, recent studies have shown evidence of a decrease in parental investments in education, and in the financial support of students in higher education during the recession, leading to an increased importance of individual payments and limited educational opportunities for children with low parental resources. This research examines whether the socioeconomic inequalities have resulted to different outcomes in college enrolment during the Great Recession in the United States.

The paper focuses on the high school graduation cohorts 2004-2013, using data from the Transition into Adulthood Supplement of the Panel Study on Income Dynamics. The logistic regression results show increase in college enrolment during the recession for children from low-income families, even when taking into account parental education. This suggests that recession had a positive influence on educational inequalities in college enrolment, benefiting the disadvantaged families. Unfortunately, this positive trend in enrolment diminishes rather quickly when the economy starts recovering. This raises the question whether the measures taken during recession, that have – knowingly or not – promoted the disadvantaged families in acquiring more education, could be extended outside the recession periods.

Educational contraction in Finland? Cohort trends in educational attainment and educational inequality

By: Outi Sirniö, Juho Härkönen

Despite the educational expansion during earlier decades and the top-ranking PISA scores, the educational level of Finland seems to be ceasing, as members of recent birth cohorts no longer achieve higher education compared to those born earlier. This study analyzes whether the contraction shows signs of increasing social inequality. We focus on birth cohorts whose parents experienced rapid educational expansion but who themselves reach decreasing level of higher education. We use register-based longitudinal data set comprising the total population living Finland and study those born between 1972 and 1984.

We focus on enrollment to secondary- and tertiary-level institutions and completion of degrees from corresponding levels with an annually updated follow-up into early adulthood. Preliminary results show that the decline in higher-education attainment started from the cohorts born in the late 1970s. The educational contraction stems particularly from the declining proportion of those completing lower tertiary-level degrees. Spearman’s rank correlation coefficients and unidiff layer effect scores point to slightly strengthening intergenerational persistence across cohorts, the increase being stronger among women. Future work will concentrate on decomposing the association between parental and achieved education into the contribution of each of the educational transitions using sequential logit modeling.

Effect of school segregation on school results and transition from school to work

By: Larsson Ann-Charlott, Emma Snölilja, Marit Jorsäter

School segregation has become an important area for discussion and research. During the period 1988 – 2009 school segregation between immigrants and natives and between children of high/low educated parents has increased in Sweden . School segregation can be an effect of increasing gaps in the society in general and equal opportunities for persons with different background are an important area for further research and studies.

In this study we will focus on describing school results and transition from school to work for persons with different background  that have attended compulsory schools with different socioeconomic distribution of pupils.

The data material consists of persons in four different cohorts (born 1979, 1984, 1989 and 1994) registered in Sweden in 2014, who have been enrolled in compulsory and/or upper secondary school in Sweden. With the help of different registers at Statistics Sweden the persons will be followed through the school system and into the labor market. The main result consists of school results, highest completed level of education, enrolment in post-secondary education for older cohorts and labor market situation for older cohorts.

Variable is socioeconomic distribution of compulsory schools regarding parent’s level of education, immigrant background and income. The schools have been divided into four groups. Other individual background variables are for example country of birth, gender and school results.

Preliminary results indicates that school segregation have an effect on results from compulsory and upper secondary school. Later in life other factors like parents level of education are more important for enrolment in higher studies and highest completed level of education at the age of 25 years.

The future of Arctic population: factoring educational attainment in projections

By: Anastasia Emelyanova

This study goal is to project populations at the sub-national level stratified by age, sex, and education for 25 Arctic provinces in addition to all-nation patterns of 8 Arctic countries for the period 2015−2050, based on the multistate methodology from Lutz et al. (2014).

In the study, alternative scenarios for the future are developed for the projection components. In fertility, we assume setting on the UN 2030/2050 “medium” forecasts and secondly, Arctic-specific assumptions for the recent patterns in Arctic fertility. In mortality, three scenarios cover the dynamics of the Arctic forerunner in life expectancy (Faroe Islands); medium mortality assumptions on the basis of a global conditional convergence model of Wittgenstein Centre for Demography and Global Human Capital with the forerunner Japan; and the setting on the UN 2030/2050 “medium” forecasts in life expectancy. In migration, we condition population changes with the growing in-migration vs. constant out-migration; growing out-migration vs. constant in-migration; a levelled-off convergence scenario between them as well as no migration change scenario to see the effect of alternative migration assumptions on the future distribution of the population. In education, we define the future shares based on the backward and forward education attainment progression ratio.

Next, we introduce mortality and fertility differentials throughout levels of primary schooling to the tertiary degree completion depending on the national system of education (5 to 9 levels). Education is an important element to include since it will keep affecting population dynamics in the Arctic. Women with more education will likely have fewer children, highly educated men and women may have better health/life expectancy and lower mortality and a better chance of survival of their children. Migration is also affected allowing for easier integration of well-educated individuals, and impacting choice/format of migration trajectories (Lutz et al. 2014). Factoring education in the projections, our goal is in producing meaningful options for stakeholders that will support human capital development and are aware of its impact on the future population structure of the Arctic residents.

Reference: Lutz, W., Butz, W., and S. KC. 2014. World population and human capital in the 21st century. Oxford University Press


4.3. Intergenerational relations in humans and non-human animals


Grandmothering in long-lived Asian elephants

By: Mirkka Lahdenperä, Khyne U. Mar, Virpi Lummaa

Background: The prolonged post-reproductive lifespan shown by human women and females of a few whale species is an evolutionary puzzle, and the reasons behind it are currently hotly debated. One of the prevailing adaptive hypotheses, Grandmother Hypothesis, predicts that the help of older females towards their grandoffspring has led to evolution of menopause and subsequent post-reproductive lifespan. However, scattered evidence suggests that like the above species, several animals with synchronous reproductive and somatic senescence seem to benefit from the help of older females in a group, too. Because very few of these studies have shown measurable fitness effects of grandmothers and studies in long-lived terrestrial species are lacking, at present the strength of evidence for the Grandmother Hypothesis cannot be evaluated in a general framework. Research question: Here we investigate whether Asian elephant females exhibit menopause and long postreproductive lifespan comparable to human females and second, whether elephant grandmothers increase their inclusive fitness by enhancing their daughter’s reproductive rate and success.

Data and methods: In this study we use unique demographic records on semi-captive Asian elephants in Myanmar spanning nearly 100 years. Findings: We found that although some elephants can live relatively long after last birth, many Asian elephant females maintain reproductive capacity into old age without a clear-cut menopause. Second, we found, that irrespective of the reproductive status of the grandmother, her presence increased the grand-calf survival and shortened the daughters’ birth intervals.

Conclusions: Our results suggest that fitness-enhancing grandmother effects might be more widespread among social species than thus far appreciated, not limiting to humans and a few whale species with menopause and decades long post-reproductive lifespan. Consequently, our results imply that the beneficial grandmother effects are insufficient, alone, to select for menopause coupled with long post-reproductive lifespan across

Demography of grandmothers through time in Finland

By: Simon Chapman, Mirkka Lahdenperä, Jenni E. Pettay, Virpi Lummaa

The importance of grandmothers for their grandchildren is well-studied in biological, sociological and medical contexts, and an increasing number of studies also address the effects of grandchildren on their grandmothers’ wellbeing. Any such beneficial or negative effects are limited by the number of years grandmothers share with their grandchildren. Changing child mortality rates, longevity of grandparents and childbearing patterns may have profoundly altered the length of grandparenthood across the demographic transition, but this has received little scientific attention. Using a large genealogical dataset from Finland, we investigate changes in the shared time between grandchildren and grandmothers from the late 18th to mid-20th century. We show for the first time how this shared time in the grandmother-grandchild dyad has changed, as well as present other time trends relevant to this relationship, both from the grandmother and from the grandchild perspective

Is early parenthood costly or advantageous? Availability of alloparents and reproductive timing in contemporary Finland

By: Venla Berg, David W. Lawson, Anna Rotkirch

Individuals coming from lower socioeconomic backgrounds typically enter parenthood earlier. In Western societies, early parenthood is often considered to have negative repercussions, including reduced opportunity for skill acquisition and income generation. However, early parenthood may represent an adaptive strategy to low healthy life expectancy if it, for example, maximizes the availability of familial alloparents (i.e., people other than the mother or the father who could potentially care for the offspring). Empirical evidence for this hypothesis is scant. We test this hypothesis using highly reliable family network data.

Using FINNFAMILY, a register-based, representative, multigenerational dataset from contemporary Finland (N = 37,509), differences by socioeconomic background in numbers of alloparents (parents and siblings) alive and dead, the pace of alloparents dying, and their effects on reproductive timing were examined by linear and Cox regression.

Individuals from lower socioeconomic groups tended to come from larger families and therefore had more potential alloparents. However, this initial pool of alloparents deceased at a higher rate compared to individuals with an upper-middle-class background. Controlling for socioeconomic background, a higher number of alloparents and a faster rate of them dying were both associated with a younger age at first birth.

Humans are cooperative breeders and the presence of kin remains important for childbearing behaviour in contemporary high-income societies. Our results on a large, reliable dataset confirm that, in terms of available familial alloparents, postponing childbearing is more costly to individuals from lower socioeconomic backgrounds. Coming from a lower socioeconomic background also speeds up childbearing, optimising potential alloparental care.

Shared lifetime, multigenerational relations and educational achievements

By: Hannu Lehti, Antti O. Tanskanen, Jani Erola

In present-day western societies grandparents and grandchildren have longer years of shared lifetime than ever before. Here we investigate whether children with more shared lifetime with grandparents earn higher educational success compared to children with less shared lifetime, the topic that, to our knowledge, has not been previously studied. We use high quality Finnish Census Panel data (n = 53,956) and apply multilevel regression models with cousin fixed effects.

We found that children with maximum amount of shared lifetime (16 years) were 10% more often than those with minimum amount of shared lifetime (zero years) completing secondary school. There was a slight difference indicating that the shared lifetime with maternal grandparents benefited children more than that with paternal grandparents. In addition, shared lifetime had more beneficial effect among grandchildren who live in same municipality with grandparents compared to others. Finally, grandparents with higher education had stronger effect on child outcomes compared to grandparents with lower education. The study shows the benefits to measure grandparental effect by shared lifetime with grandchildren and the importance to investigate grandparental influence with cousin fixed effect method.

Reproductive competition between co-resident women in historical population

By: Jenni Pettay, Mirkka Lahdenperä, Anna Rotkirch, Virpi Lummaa

Costly reproductive competition among females is predicted to lead to strategies that reduce these costs, such as reproductive schedules. Simultaneous births of co-resident women in human families can reduce their infant survival, but whether such competition also affects overall birth rates and whether females time their pregnancies to avoid simultaneous births remain unknown, despite being key questions for understanding of how intra-female competition affects reproductive strategies.

Here, we used detailed parish registers to study female reproductive competition in historical Finnish joint-families where brothers stayed on their natal farms and sisters married out, and consequently unrelated daughters-in-law often co-resided and competed for household resources. We quantified the time-varying effects of having reproductive-aged competitor(s) on a woman’s interval from marriage to first childbirth, on age-specific fertility, and on birth scheduling. Contrary to our hypothesis, the presence of one or several potential female competitors did not lead to longer first birth intervals or lowered age-specific probability of reproduction. We also found no evidence that women would schedule their reproduction to avoid the known cost of simultaneous births on their offspring mortality risk; age-specific reproductive rates were unaltered by changes in the presence of other infants in the household. These results raise interesting questions regarding the evolution of fertility suppression in social mammals in different contexts, the costs and benefits of extended families for female reproductive success and strategies deployed, and the cultural practices that may help to avoid the negative outcomes of female reproductive competition in human families.


5.1. Families and households


Men’s childcare: a comparative study of fathers’ parental leave use in Finland and Sweden

By: Ann-Zofie Duvander, Eleonora Mussino, Jussi Tervola

Multiple countries, especially in the Nordic regime, promote gender equal use of parental leave through quotas and paternal leave. The motives to promote fathers’ participation in care are many. The most eminent may be gender equality per se. The support of paternal care can also be motivated through outcomes for the child and the couple. e investigate fathers’ leave use in two countries with different constructions of quotas for fathers, Finland and Sweden, with the aim to explain differences in use by variations in policy set-up. One important difference is for instance the larger flexibility in terms of use in Sweden and another the extensive use of cash for care after the parental leave in Finland. The study sets out to answer to the following research questions: hat factors influence fathers’ use of parental leave in the two countries Is the use similarly affected by factors like socio-economic status and work sector?

Does any of the leave system facilitate long leaves among subgroups of fathers or is it the same groups of fathers who use long leaves in both countries we use detailed longitudinal register microdata for first births between 1999 and 2009. These data include sociodemographic variables where parents and children are linked, as well as data on income and parental benefits on an annual basis. We use linear probability models to estimate the propensity to use parental leave among fathers in the two countries. Preliminary results indicate large differences also when factors such as income level of the household is considered. We make a contribution to the literature on fathers’ parental leave use, particularly regarding the importance of the policy context. Finland and Sweden are both part of the Nordic regime but differences in policy context are found to be important determinants of differences in parents’ sharing of childcare at an early age.

Occupations and fathers on extended parental leave, Sweden 2009

By: Helen Eriksson

The transition to parenthood is the main life-course event in which division of labor becomes specialized in a couple: the mother substitute family work for paid work while the father’s time allocations remain rather static. If the father takes an extended period of parental leave, the process of specialization may be challenged. This study utilizes a new and unique set of administrative register data that includes dated spells of parental leave uptake as well as characteristics of both mothers and fathers for 2,158 children born in Sweden in 2009. Unlike previous studies of parental leave uptake, these data allow us to construct an indicator for spells of extended parental leave for fathers who took more than two months of consecutive parental leave with a less than two years old child. Detailed information on the father’s occupation is utilized as the main explanatory variable for the possibility to take leave from the workplace for an extended period of time. We hypothesize that occupations may function as a resource in the negotiations to take leave so that more prestigious occupations allow for the father to negotiate an extended parental leave period with his employer.

Options and limitations: paths to child care choices

By: Tapio Räsänen, Maria Valaste, Eva Österbacka, Anita Haataja

Finland has a universal child care policy. After parental leave, when the child is at age of 9-10 months, families can choose between public and private child care as well as child home care and care leave from work if employed until the youngest child is three years of age. Parental leaves, including maternity, paternity and parental leaves are compensated either with earnings related or basic benefits. The home care period is compensated with a flat rate home care allowance. In addition to the flat rate, means tested supplement to low income families, supplements for siblings under 7 years of age if taken care of at home, and in some municipalities also local supplements can be added.  Public child care services are affordable and of high quality, but care fees versus allowances may create different disincentives in child care solutions and in labour market participation decisions as well. However, universal child care provides financial support for all types of child care independent from the parents’ labour market participation. Families have many options, but individual and family characteristics, such as earning possibilities, number of children as well as employment history and status, affect or limit the child care choices that families can take.

We explore different child care choices that families make based on mother’s and spouse’s activity before and after childbirth as well as number of children and spouse’s participation in child care. We analyze different child care paths that families take based on their activities related to childbirth. We utilize sequence analysis to exploratory analyze time structures of employment, unemployment, family leave and child care activities before, after and between childbirths.

Our data is a 50 % random sample of Finnish mothers giving first birth between 1999 and 2009, about 11 000 mothers per year. Each cohort is followed up until 2013.

The father’s quota: an earnings equalizer?

By: Rannveig Hart, Nina Drange, Synnøve Andersen

Parental leave quotas reserved for the father may change the distribution of paid and unpaid work in the family, but we lack empirical evidence of whether this in turn reduces the wage-gap between mothers and fathers. This paper uses an extension of the Norwegian father’s quota from 6 to 10 weeks to estimate the causal effect of paternity leave on three outcomes; the parental leave uptake by both mothers and fathers, the long-term earnings of both mothers and fathers, and the degree of specialization within couples. We identify causal effects by implementing a regression discontinuity design, using full population data from Norwegian administrative registers. We restrict our sample to parents of children in a two month window around the reform, and focus on families where the mother had earnings the year prior to the reform (N=16678). The results show that the reform significantly increased the number of leave days taken by fathers. We find no effects on earnings and couple specialization in the year following the reform in the full sample, nor in subsample analysis.

Trends over time in his and her earnings following parenthood in Sweden

By: Anna-Karin Nylin, Ann-Zofie Duvander, Sunnee Billingsley, Marie Evertsson

Swedish women have improved their labor market position over time, and women’s earnings today are a substantive part of family income. But women still take the greater part of domestic responsibilities. Becoming a parent is a critical turning point when mothers reduce their paid work hours, leading to widening earnings differentials between partners. Given significant changes in labor market opportunities and gender ideologies, this study asks whether men’s and women’s employment responses to childbirth have become more symmetrical over time. To date, little work has addressed trends over time within couples.

This study contributes to closing that gap, expanding our understanding of how couples negotiate roles following childbirth. Having data on the full population in the Swedish registers allow us to document fine-grained differences in the pace and magnitude of change over time. Further, we explore potential differences in how change has evolved among subgroups of the population, in particular, among couples with different socio-economic backgrounds. We construct a sample of couples who have had a first child together and map their earnings trajectories starting 2 years prior to birth and up to 8 years after. We include couples with a first birth from 1992 to 2007 so that we can follow first-time parents for up to 10 years (and a minimum of seven years for the most recent cohorts). Our main analyses censor couples who separate. We will test the sensitivity of our results to different window lengths around the time of first birth. In particular, we will compare results for couples based on a shorter time period following birth to minimize attrition via parental separation. Preliminary results are currently being prepared.


5.2. Ageing and intergenerational relations


Living with and living without spouse or partner in old ages: Evidence from Estonia compared to Belgium

By: Anne Herm, Michel Poulain

The cohorts entering the older age are larger in each living arrangement than were a decade ago but the relative increase is not the same in all living arrangements. The share has also changed in older age groups, and remarkable difference is observed between men and women. The common trend is a large increase in people in unmarried cohabitation compared with those living with a spouse. Yet, the attention should be paid to people who are living without spouse or partners at old age. We assume that widowhood as the reason for this living arrangement has decreasing importance compared with divorced and never married statuses.

With Estonian census data from 2000 and 2011 and Belgian population registry data extracted for the same years we analysed the living arrangements of people aged 50 years and over. More specifically, we studied people living with and without a spouse or partner considering their marital status, and how their share has changed during this period. In the case of Estonia, increasing trend of cohabitation involves all ages above 50, except oldest women, among the latter more are living with spouse. The proportion of divorced people among aged 50 and over is slightly increasing that could result in more people living without partner in old age. Comparatively in Belgium, the share of divorced and cohabiting among older population has strongly increased since turn of millennium. The change in never married people also has effect to the population living with and without spouse or partner at older ages. These trends have importance for implementation of policies towards care and well-being of the elderly

Everyday activity Limitations, death and social networks of older Estonians

By: Liili Abuladze, Luule Sakkeus

Cross-sectional analysis of the European older adults has shown that network size is slightly bigger for those with less severe limitations, but smallest for those with severe limitations compared to the non-limited population, indicating that there is a changing dynamic with regard to networks depending on the level of activity limitations (Abuladze & Sakkeus 2013). Some countries such as Switzerland, France and Belgium indicated having more diverse networks whereas Austria, Estonia, Hungary and Portugal had more familybased networks.

Research Question: Which type of social networks lead to worsening health outcomes?

We use data from the SHARE Wave 4 (2011) and 6 (2015) that collected information on social networks and how they changed in-between the waves. The Global Activity Limitations Index will be used as a measure of disability. Data from the Estonian death registry will be added to check for deaths occurring in-between the waves and after Wave 6. Descriptive and longitudinal analysis methods will be used.

Preliminary findings indicate that Estonia’s older people keep having the largest proportion of people with everyday activity limitations among SHARE countries also by Wave 6. Estonian and Italian older people are the only ones that have lost more social network members between the waves than they have added members.

Estonia indicates one of the worse health as well as social network outcomes among older people in Europe. Our analysis will explore the relationship between social networks and health outcomes in-depth by combining different datasets, and by focusing on Estonian older adults.

Eight-year institutional long-term care use in relation to proximity to dementia death: the influences of socio-demographic factors among Finnish men and women

By: Kaarina Korhonen, Elina Einiö, Taina Leinonen, Lasse Tarkiainen, Pekka Martikainen

Background: Dementia increases the use of institutional long-term care (LTC) for extended periods before death, but little is known about how socio-demographic characteristics of individuals affect the care use trajectories before a dementia death. Using register data, we assess how age, marital status and household income together with dementia combine to predict the use of institutional LTC up to eight years before death.

Data and methods: Individuals aged 70 years and over dying from dementia (n=20,645) and all other causes (n=82,699) in 2001–2007 and surviving to 2007 (n=60,208) were identified from an 11% random sample of the Finnish population using cause-of-death register. Marginal effects of age, marital status and household income on institutional LTC use were estimated in relation to the proximity to death or end of follow-up. Logistic regression with generalized estimating equations (GEE) was used in estimation.

Results: Generally, older age, not being married and a lower household income predicted a greater adjusted probability of institutional LTC use. Age was the strongest single predictor in all cause-of-death groups. The differences by all socio-demographic factors constantly increased up until a non-dementia death and the end of follow-up for the survivors, whereas before a dementia death, the differences substantially reduced or disappeared after an initial increase.

Conclusions: The results show that age, marital status and household income are important determinants of institutional LTC use among older men and women regardless of cause of death, but decreasingly so in the last years of life for people dying from dementia. These results have implications for targeting of interventions aimed at promoting independent living.

Inequalities among future pensioners: a dynamic microsimulation approach

By: Heikki Tikanmäki

In this study we analyse income distributions and life-courses of future pensioners. We partition the life-course of a birth cohort into working life, out of the labour force, unemployment and retirement. The results are shown separately for different educational groups and genders. We also compare life-courses of people with and without disability pension background.

We use the Finnish dynamic microsimulation model ELSI to analyse life-courses and inequalities among future pensioners. The model is based on administrative register data. The model runs on an individual level data with sample size of 32 per cent of the adult population of Finland. The adult life of each individual is simulated using one-year time steps. The simulation spans until the year 2085. Model ELSI has recently been used to evaluate the impacts of the 2017 pension reform in Finland (Tikanmäki & al. 2015; 2017).

Our main results state that the pension levels of the highest and lowest educational groups improve fairly modestly compared to other groups. The working lives of the least educated pensioners will be significantly shorter in the future than today. This is mainly due to the changes in the educational structure.


5.3. Health, morbidity and mortality


Disease and fertility: Evidence from the 1918 influenza pandemic in Sweden

By: Maryna Ivets, Nina Boberg-Fazlic, Martin Karlsson, Therese Nilsson

In this paper we study the effect of the 1918 influenza pandemic in Sweden on subsequent fertility and the mechanism though which fertility behavior changes. A historical dataset is compiled from multiple sources including midwife journals, reports of chief medical officers, church books and censuses. We explore the dynamic structure of the data by employing difference-in-difference method. Therefore, we aim at uncovering a causal relationship between flu morbidity/mortality and fertility. Our results suggest an immediate reduction in fertility driven by morbidity, and additional behavioral effects driven by mortality. We find some evidence of community rebuilding and replacement fertility, but the net long-term effect is fertility reduction. In districts highly affected by the flu there is also an improvement in parental quality: we observe a relative increase in births to married women and better-off city dwellers. Our findings help understand the link between mortality and fertility, one of the central relations in demography, and show that several factors — including disruptions to marriage and labor markets — contribute to fertility reduction in the long term. Our results are consistent with studies that find a positive fertility response following natural disasters, but with high-quality historical data we show that this effect is short-lived.

Increasing the mortality gap? Social inequality in mortality among elderly and adults in northern Sweden 1851–2013

By: Sören Edvinsson, Göran Broström

We investigate the development of social inequality in mortality among elderly (65-89 years) and adults (40-64 years) in northern Sweden during the mortality transition. Social class (based on occupational titles categorized in three groups) is among the strongest determinants for health in present-day societies. It is often assumed that such differences were present also in historical societies. Available studies have however shown more complex patterns. This study takes a unique long-term perspective on this issue. The study focuses on two main questions, the first relate to the long-term change in social differences in mortality. The second question is whether socio-economic position have less impact on the elderly population compared to population in working age. Furthermore we consider possible gender-specific patterns in this process.

The development of mortality in different social classes is analyzed according to both total mortality and major cause-of-death categories. For the later periods, we also compare the results from the class-based analysis with other measures of social position, in this case income and education. Focus is on mortality in the Skellefteå and Umeå regions in northern Sweden 1851-2013. The study is based on the historical population data from the Demographic Data Base, Umeå University and modern population register data from Statistics Sweden and National Board of Health and Welfare (FoB, LISA, Cause of Death Registers). Preliminary findings show a surprisingly late appearance of the modern social pattern in mortality, but with clear gender differences. Higher social class did not protect men in historical contexts, while this was the case among women. During the latest decades, higher social position is protective for women as well as for men.

Alcohol-related mortality among occupations with high all-cause mortality in 2001–2015

By: Hanna Rinne, Mikko Laaksonen, Veijo Notkola, Riikka Shemeikka

There are many occupational connected risk factors which have been found to be connected with an increased alcohol consumption. These include e.g. social norms supporting workplace drinking, easy access to alcohol, hazardous physical working conditions and mental or physical stress. It is, however, unclear, to what extent the high mortality in some occupations is linked with high alcohol consumption.

The main aim is to examine, which are the occupations with the highest all-cause mortality, and whether there are differences in the role of alcohol in explaining the excess mortality among occupations with high all-cause mortality.

We used longitudinal individual level register based data from the registers of Statistics Finland. Study population consisted of employees aged 30-64 at the end of the year 2000. The follow-up period was 2001-2015.Analysis methods included death rates, age standardized mortality ratios and confidence intervals.

The occupations with increased risk of death were mostly manual workers. The occupations with the highest risk of death were building construction labourers (SMR 180) among men and locomotive engine drivers (SMR 281) among women. The contribution of alcohol-related mortality to the all-cause mortality varied from 0 to 63% among men and from -11 to 120% among women.

Among men, the contribution of alcohol-related mortality to high all-cause mortality was over 50% among child-care workers, hospital, domestic and kitchen helpers as well as miners. The contribution was under 10% among hairdressers and mining and mineral-processing-plant operators.

Among women, the contribution of alcohol-related mortality was over 40% among chemical-processing-plant operators, mobile-plant operators and kitchen helpers and under 10% among newspaper deliverers and electrical- and electronic-equipment assemblers.

Preventions should be focused on occupations with high alcohol-related mortality to decrease health disparities between population groups. Examining other causes of death contributing to high mortality is needed.


6.1. Fertility



Education, gender, and cohort fertility in the Nordic countries

By: Marika Jalovaara, Gerda Neyer, Gunnar Andersson, Johan Dahlberg, Lars Dommermuth, Peter Fallesen, Trude Lappegård

Systematic comparisons of fertility developments based on education, gender and country context are rare. Using harmonized register data, we compare cohort total fertility (CTF) and ultimate childlessness by gender and educational attainment for cohorts born beginning in 1940 in four Nordic countries. CTF has remained fairly stable in all countries. Childlessness, which had been increasing, has plateaued except in Finland. Women’s negative educational gradient has vanished, while men’s positive gradient has persisted. The highest level of men’s childlessness appears among the less educated, revealing striking educational differences. Childlessness has increased among low-educated women but not among highly educated women. The educational gradient in women’s childlessness has shifted from positive to negative. Thus, we witness both a new gender similarity and widening social inequalities in childbearing in the Nordic welfare states. Low-educated citizens of both sexes have apparently become an increasingly marginalized segment with regard to childbearing

Socioeconomic resources and fertility in cohabiting unions

By: Anneli Miettinen

Traditionally childbearing in non-marital unions has been associated with a disadvantaged position in the society. However, in most western countries and in the Nordic countries in particular, unmarried cohabitation is now widely spread, not only as a short transitional period or a precursor to marriage in young ages, but as a lasting co-residential relationship which may also include children. With cohabiting unions becoming more marriage-like, it could be that the differences in socio-economic antecedents for childbearing between nonmarital and marital unions are becoming smaller. However, non-marital unions are more vulnerable than marriages and frequently end in dissolution. Prior research has not explicitly considered the associations between socioeconomic resources and union (in)stability and how this is linked with fertility outcomes. Cohabitors who are likely to separate are not likely to bear children and vice versa. Those more committed to union may have children first but marry soon afterwards which can bias the estimates for non-marital fertility. Marriage may still entail higher costs that keep some from marrying, even if with children. Lower social and legal barriers for dissolution may encourage using cohabitation as a screening process through which a more ‘eligible’ partner for childbearing is searched.

In this study, a multi-state modeling approach that allows for correlation between different states will be used to examine how socioeconomic resources of each partners are related to transitions to childbirth, dissolution of the cohabiting union or marriage among cohabiting couples. Data are drawn from 10% sample of Finnish population (Statistics Finland), and the study focuses on cohorts born in 1967-1990 and family formation events during 1987-2009. Data set contains information on educational attainment, income, occupational class and employment, as well as dating of births, cohabitations, and marriages to the precision of month.

Why has childlessness increased among highly educated men in Sweden?

By: Margarita Chudnovskaya

Men with higher education have typically been attractive as potential partners due to their access to the social and economic capital that education provides. As higher education has expanded, men have become a minority in higher education, and thus–following theories of the partner market–even more attractive as potential partners. However, the rate of childlessness among highly educated men has not decreased over time, but rather increased. This study investigates these developments in childlessness among highly educated men. Using Swedish register data and a cohort approach, I examine whether trends in childlessness can be explained by changes in men’s access to potential partners or changes in their employment conditions. Preliminary results on changes in men’s educational and employment profiles suggest that existing explanations (theories of partner market and economic stability) cannot explain trends in childlessness among highly educated men. I consider other possible explanations such as increasing opportunity costs of childbearing for men, and changes in underlying preferences for childbearing.


6.2. Educational inequalities


Childhood health conditions and dropout from post-compulsory education

By: Janne Mikkonen, Heta Moustgaard, Hanna Remes, Pekka Martikainen

Studies considering early-life health a determinant of educational outcomes have usually focused on the effects of self-rated health or birth-related characteristics. Some studies have also examined the effects of specific health conditions, but most of them have been conducted in the United States and none of them has simultaneously assessed the effects of several different health conditions or separated health-induced delays in educational careers from long-term effects.

Our goal is to estimate the short-term and the long-term contributions of different types of health conditions on dropout from post-compulsory upper secondary education at both the individual and the population level.

We had a 20% random sample of Finns born between 1988 and 1995 (n=103,575). We used robust Poisson regression model to calculate risk ratios between children with and without health conditions at ages 10-16 and population attributable fractions (PAF) to evaluate the contribution of health conditions at the population level.

Children having any condition, somatic condition, mental disorder, or injury at ages 10-16 had higher risk of dropout at ages 17, 19 and 21 after adjusting for a number of individual- and family-related factors. In each case, mental disorders were associated with the largest increases in risk. The effect of having any condition was similar for girls and boys at age 17 and stronger for girls at ages 19 and 21. Childhood health conditions could account for up to 30% of dropout at age 17, 23% at age 19, and 21% at age 21.

This study suggests that a major part of educational dropout might be attributable to childhood health conditions, especially mental disorders. Health conditions do not only predict delays in educational careers but might also hinder some young people from obtaining an upper secondary degree at all.

Reasons given for dropping out of upper secondary school and consequences on the labour market

By: Paula Kossack

On behalf of the Swedish Ministry of Education and Research, Statistics Sweden is conducting a study on young people who have not completed upper secondary school. Three subgroups are studied: young people who never registered for upper secondary school, young people who registered but dropped out, and young people who studied but did not complete their schooling.

The study aims to answer the following questions:  How large are the three subgroups as a percentage of a cohort of the same age? What characterises them in terms of social background variables? This part of the study is based on data from statistical registers.What reasons do the young people in the three subgroups give for not completing upper secondary school? In their experience, what has been done in society to help them complete school? This part of the study is based on survey data. How does the labour market situation look like for the three subgroups when they reach the age of 20 to 25 years? What percentage later completed their upper secondary education as adults? Are there any differences between the three groups, for example in terms of social background? This part is based on statistical registers and survey data.

Preliminary results show that there are some differences between the three groups with regard to their reasons for not completing upper secondary school. Our study confirms that completing upper secondary school is crucial for their situation on the labour market both in the short run and in the long run. With regard to social background variables, our study shows, for instance, that among immigrants, their age at the time of immigration is significant for their performance in school and, as a consequence of this, for their entry in the labour market

The role of social capital and geographical mobility in educational attainment in Germany and Finland

By: Patricia McMullin, Jani Erola, Elina Kilpi-Jakonen, Aleksi Karhula

Geographical mobility occurs often because the lack of economic resources in the location of origin are intended to be replaced by the economic resources available in the destination area. However, Geographical mobility is often accompanied by a loss of capital that remains in the location of origin, such as social capital. In families with socioeconomically advantaged parents movement may occur more often because of the occupational expertise of the parents. In this case intra-family social capital (in terms of parental support) can compensate for the loss of social capital that the child may face as well as for the loss of other sources of social closure in the community that result from family moves (Hagen et al. 1996). Another reason for geographical mobility is the search for a partner, in this case inter-generational success in geographically mobile families can also depend on their ability to compensate the lost social capital through re/partnering. We study the effects of geographical mobility according to children’s age and parental socio-economic and union status on children’s higher educational enrollment in Finland and West Germany.

We use high quality Finnish register data, including reliable annual indicators for employment, parental education and other familyrelated variables such as marital status. The results from these data will be compared to results using regional data from the German Socio-Economic Panel (SOEP).

Preliminary findings from the Finnish data indicate that even after controlling for parental education, divorce/separation, childhood household income, and parental unemployment, moving during childhood is associated with the risk of not attaining a secondary level degree and a lower probability of entry into higher education by the age of 22. This suggests that other forms of capital that a parent might gain from a move cannot fully compensate for the loss of social capital a child experiences after movement from one geographical location to another.


6.3. Migration


Mapping circular migration in Sweden

By: Linus Garp, Tomas Johansson

In the budget bill for 2016 the Swedish government have allocated funds to Statistics Sweden to produce annual statistics on circular migration. The background to this task is that the government in 2014 gave Statistics Sweden and the Swedish Migration Agency the task to undertake a mapping study of available statistical data on migration.

One suggestion in the study was to establish a new migration register, which would include all circular migrants. This register would fill a gap in Swedish migration statistics because today there are no published statistics that describe this group of migrants.

At the end of 2016, the register will be completed and the first round of the annual statistics will be published on Statistics Sweden´s website. This will probably be the first register in the world focusing on circular migration. At the conference we can present up-to-date statistics on circular migrants, as well as be able to explain how we created our circular migration register.

The presentation will focus on the developing of the register and on presenting the new statistics on circular migration.

Factors connected to moving from rural to areas with universities among adolescents during 2002 to 2008

By: Matti Saari

The biggest change in internal migration in recent decades in Finland has been the increase in the level of propensity of moving among adolescents at the end of 1990’s. A little earlier a strong recession hit Finland. In the countryside adolescents may have specially reacted to the bad economic conditions or poor educational possibilities by moving away. There are a few studies of all movers from rural to urban areas. This article examines the moves by adolescents from rural areas to larger urban areas and connected regional factors in the area of departure.

In this study, unemployment and economic growth are used as explanatory variables for the moves of adolescents. In addition, the study examines the connection between the share of industrial workers and highly educated persons and out-migration from rural areas to urban areas with universities. We use also as covariates of migration the type of the educational centre of the countryside and boy’s share of cohort among persons belonging to cohort preceding the target cohort.

The study analyses the migration flows of persons aged 20 to 24 in 2002 to 2008. The number of rural areas is 100.  The method of the study is regression analysis of panel data using within estimation.

Positively correlating factors for out-migration were the unemployment rate and the share of highly educated people. The type of educational centre of the countryside was also significantly correlated factor. As expected, unemployment and the share of highly educated people had a positive effect and the type of the educational centre of the countryside had a negative effect on migration.

The impact of children on emigration – A study of E-15 migrants in Sweden

By: Andreas Raneke

The freedom of movement within the European nion is one of the pillars in which the E is built and is encouraged as a mean to create a European citizenship. Even though economic reasons for moving within the union are most common, other reasons such as family or education is important in migration decisions. As immigration from member states of the nion have raised since the Swedish accession, many also returns to their country of origin and emigration rates from Sweden are high. Besides economic reasons for leaving Sweden, the family context also plays an important role in immigrants decision to emigrate. The aim of this study is to explore what impact family life and especially the presence of children might have on out-migration but also to look into socio-economic determinants of emigration. To analyze this, I use longitudinal population register data on E-15 migrants that immigrated to Sweden during 1998–2009 to apply an event-history analysis.

The results suggest that economic integration in Sweden plays an important part as unemployed and having a low income means higher emigration propensities. Having no partner or a partner not born in Sweden means a higher risk of leaving compared to those with a Swedish-born partner, but when information about children is included, having children or not seems to matter more than being in a relationship or not. A closer look at the country of birth of the children shows that having Swedish-born children inhibits emigration while having foreign born children increases the probability to emigrate. The results indicate that having a Swedish born child provides motives to stay on in Sweden.


7.1. Families and households

Are life-course trajectories of single fathers and single mothers shaped differently by educational attainment?

By: Klara Capkova, Marika Jalovaara

Research on the relationship between socio-economic inequalities in single parenthood has focused predominantly on single mothers, leaving single fathers aside. The recent increase in the single father category mirrors both legal and normative changes occurring in Western societies, where widowed fathers are being replaced by divorcees and unmarried single fathers. The demographic profile of single fathers is not changing only with regard to marital or partnership status – the educational and socio-economic composition of single father category has been changing as well. In our study use the life-course approach and address the link between socio-economic status and resources measured by educational attainment and single fatherhood and motherhood.

We focus on how and to what extent educational attainment shapes partnership and family trajectories that include single parenthood for men and women, and whether these trajectories differ for single fathers and single mothers. We use data on 11% sample of the total population from the Finnish longitudinal registers. The analysis relies on union (cohabitation and marriage) and childbearing histories, and prospective information on educational attainment. It focuses on birth cohorts 1969–1970, and covers the life-course histories of men and women between 1987 and 2009.

We employ sequence and cluster analysis to capture the partnership and family trajectories of single fathers and single mothers in contemporary Finland.  To our best knowledge, our study is the first application of sequence and cluster analysis in the research addressing the association between socio-economic indicators and single parenthood.

A comparison on the effects of paid child support on the economic well-being of fathers paying child support in Finland, the UK and the USA

By: Mia Hakovirta, Dan Meyer, Christine Skinner

Increased numbers of divorce, separation and non-marital childbearing over the past several decades has contributed to the rise of parents not living with their children in the same household. These non-resident parents are typically fathers who are sharing the economic responsibility of their children across households by paying child support. However, earlier research on child support has focused on parents with whom the children live and very little is known about the characteristics, living circumstances and child support payments of non-resident fathers.

This study compares the economic well-being of male child support payers in three countries: Finland, the UK and the USA. These countries organize their child support policy in different ways and the countries represent the different child support regime. This study uses LIS datasets from the year 2013 for analysis.

First, we compare the level of paid child support in different subgroups of child support payers. Second, we calculate the poverty rates of child support payers before and after paid child support presenting the percentages of those that fall into poverty by paying child support and how much the poverty gap increases due to the child support. Thirdly, we calculate the poverty rates of single mothers receiving child support to figure out whether it is single mothers that are drawn out of poverty by child support or child support payers pushed to into poverty by paying child support.

Preliminary results suggest that in all countries the largest group of child support payers are single males who do not live with children and do not have a new partner. They also contribute the most financial support for their children living in other household. In the US, the level of paid child support is much higher than in Finland and the UK in all subgroups. For the poverty effects, results show that very few child support payers fall into poverty because of the amount of child support they pay, but increase in poverty rates is clearly higher in the US than in the other countries. The analysis indicates that more single mothers are drawn out of poverty by the receipt of child support than there are child support payers pushed into poverty by child support payments.

Is the home-care-allowance disadvantaging single mothers? Labor market consequences of the cash-for-care benefit for single mothers in Finland

By: Kathrin Morosow, Marika Jalovaara

This paper examines the labor market consequences of the home-care-allowance for single mothers in Finland. The home-care-allowance, also called cash-for-care, is a benefit paid to parents who take care of their children at home instead of using public daycare. In light of the fact that Finland and the Nordic countries are characterized by high levels of gender equality and mother’s labor force participation, this policy is widely discussed and criticized for contradicting these aims. Previous research indicated for the home care allowance to reduce maternal labor force participation. Using Finnish register data, this paper assesses: (1) to what extent and when single (vs. partnered) mothers use the cash-for-care benefit in Finland, and the factors that influence the use. (2) Employment trajectories for single mothers compared to partnered mothers by different cash-forcare use lengths and considering the different life stages at which they become single mothers. This will be done by means of multichannel sequence analysis. (3) The consequences for labor market participation when single mothers use the home care allowance, and if that effect varies by length of the cash-for-care use. Here fixed-effects event-history analysis for repeated events will be used. Considering that single mothers face disadvantages in terms of fewer resources anyway, this paper contributes by answering the question whether single parents are disproportionally disadvantaged when using cash-for-care.


7.2. Fertility


Explaining sibling similarity in fertility – what more than socialization matters?

By: Johan Dahlberg

The intergenerational transmission of fertility has received much attention in demography. This has been done by estimating the correlation between parents’ and offsprings’ fertility. The alternative method that provides a more comprehensive account of the role of family background – sibling correlations – has not been introduced in this field of research until recently. Sibling correlations are broader measures of the impact of family and community influences on individual outcomes than intergenerational correlations.

I employ a data set with rich family information to explore what factors in addition to the traditional measures mother’s age at the birth of her first child and the number of siblings – can explain the intergenerational transfer of fertility. I estimate the overall importance of family background on entry into parenthood and completed fertility and whether it changed over time. Furthermore, by adding family-specific factor shared by siblings, I explore what factor explains siblings’ similarity in fertility.

Brother and sister correlations in age first birth and final family size were estimated using multi-level linear regression on Swedish longitudinal register data. First unconditional model that involves no covariates were estimated. Then family-specific factors were added, separately, aggregated in groups – socialization, social background and family complexity –and in the last step all level 2 measurements were included simultaneously.

The overall variation in fertility that can be explained by family of origin is approximately 15%-30% for women and 13%-20% for men. The overall importance of the family of origin has not changed over the approximately twenty-five birth cohorts that were studied (1940-67 for women, 1940-63 for men). The family-specific factors accounts for about 30 percent of siblings’ similarity in age at parenthood and about 25 percent of sibling similarity in the completed fertility. Social background matter the most, followed by socialization.

Estimating male fertility from vital registration data with missing values

By: Christian Dudel, Sebastian Klüsener

Comparative research on the fertility of men is rare, even though there is a rising interest in the subject as the involvement of men in family activities is increasing. One potential reason for this lack of research is that the age of the father is often missing in vital registration data.

We compare two imputation approaches which allow to estimate male age-specific fertility rates and related measures for data with missing values. The first approach is based on the age distribution of fathers for which the age is known. This approach is commonly employed in the literature. The second approach uses the age distribution of fathers conditioned on the age of the mother (conditional approach).

Based on Swedish register data, we conduct simulations in which we vary the overall proportion of missing values as well as selectivity of missing values, and we assess the bias of both methods.

Generally, the conditional approach outperforms the unconditional approach in the majority of simulations. Especially for the mean age at childbirth, the bias of the unconditional approach can be large in some cases. If the proportion of missing values is small and selectivity of missingness low, both approaches perform well.

A provisional conclusion of our preliminary findings is that the conditional approach seems to be preferable, as it works well under many conditions, even when the proportion of missing values is high.

Education and the second birth among men in Finland

By: Jessica Nisén, Pekka Martikainen, Mikko Myrskylä, Karri Silventoinen

The transition to fatherhood is selective on socioeconomic characteristics in many high-income countries in that more highly educated men are eventually more likely to become fathers. Less evidence exists on socioeconomic differences in men’s transitions to higher-order parities; previous studies have documented mainly positive effects of education on second-birth rates in men. Societal changes, such as the postponement of parenthood, men’s strengthening role as caregivers, and changes in union patterns, may affect the transition to the second birth differently across educational groups. This study aims to analyze educational differentials in the transition to the second birth among Finnish men born in the 1940s, 1950s and 1960s in order to shed light on trends in men’s fertility by educational level.

The study is based on Finnish register data with monthly records of live born children until 2012. The readily analyzed cohorts born in 1944-48 consisted of 14,982 men. The main explanatory variable was level of education at the age of 30-34. The statistical method was discrete time event history analysis: men were followed from the first birth until the second birth or censored at their 44th birthday.

Among men born in 1944-48, education speeded up the transition to the second birth, net of the age at first birth, age difference to and education of the mother of the first child, and background characteristics. The second-birth rates were higher particularly among those educated to the high level (10), (OR 1.43, CI 1.321.54), and slightly higher among those educated to the middle level (49), (1.05, CI 1.01-1.09), as compared to those educated to the low level (41).

Education may enhance second-birth rates of men because of several reasons, e.g. higher incomes, stronger inclination to follow the two-child norm and more stable unions. We further aim to replicate the analysis in cohorts born in the 1950s and 1960s to study potential cohort differences


7.3. Health, morbidity and mortality



Life expectancy by education in Sweden: recent trends and regional variation

By: Örjan Hemström, Jeroen de Munter

Disparities in mortality and longevity by educational level have been widely reported from a large number of countries and for both women and men. Many studies also report that such disparities increase over time.

The main aim of this study is to analyze recent changes in life expectancy by educational level in Sweden. A number of questions are raised. Have there been any changes in differences in life expectancy between educational groups in recent years? Are there regional variations in the size of the gap in life expectancy between educational groups?

The study uses administrative registry data for the Swedish born population 30 years and older for two 5-year periods, 2006–2010 and 2011–2015, to calculate sex- and group-specific life expectancy at age 30 with conventional life table technique. With the use of age-specific changes in mortality by educational level the contribution from various age groups to life expectancy change is analyzed.

Life expectancy at age 30 increased for women and men in all educational groups. The group with compulsory education had the smallest and those with a post-secondary education had the greatest increase. The gap in life expectancy between those two groups increased from 4.9 to 5.3 years among women and from 5.0 to 5.6 years among men. An age-specific analysis revealed a mortality increase in a number of age groups among both women and men with a compulsory education. There were relatively small educational differences in life expectancy in some counties in the south of Sweden and relatively large differences in some counties in mid-Sweden.

Differences in life expectancy between educational groups have increased among both women and men in recent years. There is regional variation in the size of differences in life expectancy between those with compulsory as compared with post-secondary education.

Education and the basis of self-rated health

By: Liina Junna

Self-rated health (SRH), a frequently used survey indicator of general health, is periodically utilised in the study of educational health disparities. It has been found associated with mortality and several other health indicators. Some previous studies imply that these associations may be strengthened by higher education. However, these studies disagree on the strength and the scope of the interaction effect.

This paper analyses 1) which aspects of health are included in dichotomised poor or very poor SRH ratings, and 2) whether education moderates the association between SRH and the selected indicators of health that represent five health dimensions identified in previous studies (clinical-, functional-, and mental health, health behaviours and bodily symptoms).

The analyses are conducted using logistic regression and nonlinear decomposition methods. The data is derived from the nationally representative Health 2000 data for the household and institution dwelling population over the age of 30 residing in mainland Finland.

Somatic complaints, functional health, mental health disorders, and some diseases were found associated with poor SRH. An interaction effect was found for cardiovascular disease, subjective functional limitations, and somatic complaints. Higher education strengthened the associations. However, the interaction effect was statistically significant only when comparing the highest educational group to the lowest. For the majority of the indicators, no interaction was found. Also, the decomposition showed that the selected health indicators and demographic variables explained virtually the entire gap between the groups in reporting poor SRH.

The paper adds to our knowledge of the basis of SRH ratings. No previous studies of educational SRH reporting heterogeneity have been conducted in Norther Europe.