Estimates of the contribution of rising educational participation to partnership postponement: a model-based decomposition for the U.K., France and Belgium

Michael Murphy, London School of Economics and Political Science (LSE)
Karel Neels, Universiteit Antwerpen
Maire Ni Bhrolchain, University of Southampton
Eva Beaujouan, Wittgenstein Centre (IIASA, VID/ÖAW, WU)

Postponement of partnership is a major recent trend in Europe and other developed countries that has contributed to low period fertility since the early 1970s. Recent findings suggest that delayed childbearing is closely linked to educational participation but the role of educational participation on partnership formation is unclear. In this paper, we draw on data from three countries, United Kingdom, France and Belgium, that have high quality data stretching back to at least the 1970s on age leaving education. There is generally parallel relationship between age at leaving full-time education and age at partnership during the latter part of the twentieth century, although this appears to be weakening as the gap between ages of finishing education and forming first partnership is shrinking in the latest period. The patterns are similar; in particular, trends for United Kingdom and France over the period 1970 to 2000 are almost identical for age of first partnership. However, educational participation is very different, with much higher levels being observed in France than in the other countries. We find that the proportion of partnerships starting within education has been increasing, although the levels in the United Kingdom and Belgium are lower than those in France. In addition, the rises in the United Kingdom and Belgium are considerably less than in France, where recently about one quarter of first partnerships occur while the woman is in full-time education. Using generalized additive models (GAM), we analyse variation in first partnership rates by age at leaving education and duration since leaving education between 1970 and around 2000. Direct and indirect standardization by age at leaving education and duration since leaving education are used to decompose variation of first partnership by age in terms of composition effects and rate effects in the countries considered.