Family size matters: How low fertility affects the (re)production of social inequalities

Project Information

FAMSIZEMATTERS

Grant agreement ID: 681546

DOI

10.3030/681546

Project closed

EC signature date 16 March 2016

Start date 1 July 2016

End date 30 June 2022

Funded under

• EXCELLENT SCIENCE - European Research Council (ERC)

Total cost

€ 1 902 598,00

EU contribution

€ 1 902 598,00

Coordinated by

THE CHANCELLOR, MASTERS AND SCHOLARS OF THE UNIVERSITY OF OXFORD
United Kingdom

Periodic Reporting for period 4 - FAMSIZEMATTERS (Family size matters: How low fertility affects the (re)production of social inequalities)

Reporting period: 2021-01-01 to 2022-06-30

Summary of the context and overall objectives of the project

Socio-economic inequality, such a difference in educational attainment, jobs and income, is a phenomenon of all time. Some societies are more unequal than others and over time some societies become more or less equal. It is thought that the family plays in important role in reproducing inequalities from one generation to the next. If the family plays such a central role, what ten are the implications of the family itself changing? Over the last fifty years or more, most societies in Europe and East-Asia moved or started moving towards women having fewer children and/or fewer women having children. What does this change in family size mean for social inequalities? This project tries to answer this question through a series of studies on various aspects of family size and the well-being of adults and children. Many of these studies compare trends in different countries and across (birth) cohorts. Earlier studies have often only looked at the determinants of family size, rather than the consequences, while studies on social inequality have not taken into account changes in family size. This project connect these two issues. We ask three sets of related questions on the consequences of low fertility for inequalities in (1) children, (2) adults and (3) societies. With regard to children, we investigate multi-generational processes (ie the importance of grandparents), the changing role of coming from a large or small family, and the advantage or disadvantage of being an only-child. We also assess how the overall importance of the family for children’s education varies between countries and over time in the West and East-Asia. For parents with adult children, we study whether who your children are (instead of how many you have) is important for parents’ health and mortality, and we examine trends in childlessness and whether childlessness increases or decreases inequalities in income, wealth and well-being. We also study the effects on well-being of becoming a grandparent. Together these studies allow us to better understand under which conditions smaller families increase or decrease social inequalities.s.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

We have produced almost twenty papers, fourteen of which have been published as journal articles, working papers or as full manuscripts on SocArxiv. Moreover, we have finished one large data harmonization project (International Sibsize and Educational Attainment Database (ISEAD) https://dataverse.harvard.edu/dataverse/isead) and are close to finishing the second; in these projects we bring together more than a hundred separate surveys to create a new data base that allow us to compare the links between family size and educational attainment across birth cohorts and across countries. We also organized two workshops.

We have produced two systematic overviews of all available studies with regard to two questions that are important for the link between family size and inequalities. The first review brings together all published studies that have estimated a link between grandparents’ education and the educational attainment of their grandchildren (ie they ask whether children with highly educated grandparents reach higher levels of education themselves, taking into account the resources their parents have). We show systematically, for the first time, how strong the correlation is, how it varies between studies and between countries, and we show that after taking into account the middle generation the correlations between grandparents and grandchildren reduces more than two thirds. The second review is a meta-analysis where we have brought together all estimates of sibling correlation in educational attainment that we could find in the literature. Sibling correlations show how similar children from the same family are in terms of educational outcome. If families are more important in shaping education, no matter for what reason, siblings will be more alike. This study is the first to show the full range of sibling correlations across different studies, countries and time periods. Knowing this range helps us to interpret the results from individual studies that use sibling correlations and it tells us about countries differences in how significant the role of the family is.

Some of the main findings in our studies:
- Grandfather’s education is associated with grandchildren’s education even when taking into account resources of the parents (middle generation) but this association does not vary systemically by family size, nor is it stronger if the grandparents and grandchildren have more overlap in life time
- The range of published sibling correlations for education is 0.3-0.7 and shows only few systematic variations between countries. The correlation has not change substantially between older and younger cohorts. China is an important exception; here the sibling correlation was low in early (1940s) birth cohorts and in the next forty years increase to the higher end of the distribution.
- In the majority of countries we studied there is an increase in the disadvantage coming from a large family has for education, with only two exception (i.e. a reduction). The trend of increasing sibship size disadvantage was strongest amongst post-socialist and East Asian countries. Such a trend of growing sibship size disadvantage is not found in Nordic countries, while Anglo-Saxon countries showed a trend of decreasing sibship size disadvantage.
- Average sibship sizes have declined in virtually all countries. Fertility rates underestimate average sibship sizes. Average sibship sizes are socially stratified, with smaller sibship sizes among higher educated parents. Educational disparities in average sibship size have declined over time, indicating convergence in most countries under study. This convergence is taking place at the upper tail of the parity distribution, i.e. for large families, but not for only-child families.
- Childlessness first declines and then increase with human development in low income countries.
- The prevalence of only-children varies widely across the world and among rich countries too. In some countries only-children are rather rare, in other they are quite common. In many, but no means all countries, has the share of only-children increased. These differences in levels and trends are poorly understood so far. Only-children do better in terms of educational outcomes when there are more of them, partly because in those countries only-child families tend to have high socioeconomic status while in countries with few only-children they are more common in disadvantaged families.
- The often-found positive link between being a grandparent and well-being isn’t borne out when looking at the transition to grandparenthood, suggesting positive effects later on may be more to do with the interaction with older grandchildren than with the change in role and social status.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

We have gone beyond the state of the art in several ways. We have created two new unique data sets that allow us to study questions about family size and education across more cohorts and countries than ever before. Both the scope of our sibling correlations meta-analysis as well as our new original analysis of sibling correlations in education are unprecedented and allow for a more systematic comparison of cohorts and countries. We have introduced more rigorous methods - fixed effects; regression discontinuity designs combined with censored data; meta-analysis - and new approaches - agent-based simulation models calibrated to empirical data; Bayesian hierarchical growth curve modelling approach developed to assess accumulation of (dis)advantage throughout the life courses.