Doktorarbeit

Drivers and future of the fertility decline in the 2010s: an analysis of Finland and other Nordic countries

Abstract

The Nordic countries have maintained relatively high and stable cohort fertility over the past three decades, which has inspired fertility theories and cemented the idea that the generous welfare system of these countries promotes fertility. However, the narrative changed in the 2010s, as these countries’ total fertility rates (TFR) declined strongly and reached all-time low levels. These declines were unexpected, remain poorly understood, and challenge the understanding of fertility patterns in the developed world. It remains unclear how this fertility decline relates to fertility timing, union patterns and the field of study – all of which are important factors in explaining fertility patterns that could generally help illuminate the mechanisms behind the decline.
This thesis analysed fertility dynamics in Finland and the other Nordic countries, focusing particularly on the decrease in period fertility in the 2010s and its consequences for completed cohort fertility by using existing methods and developing a new forecasting approach. The study used harmonized data across the Nordic countries from the Human Fertility Database (HFD) to
compare age, parity, tempo, and quantum drivers of the declines, and register data from Statistics Finland both to assess the importance of changes in union patterns for the fertility decline and to identify variation in fertility declines by field of education.  
Decomposition of the period fertility decline between 2010 and 2018 showed that a fall in fertility occurred at nearly all ages below 40, and that the decline in first births contributed the most to the overall decline in fertility among all Nordic countries. The fertility decline in the age group 30–39 is a departure from the long-term trend of increasing fertility, suggesting that not
only fertility postponement is driving the fertility decline. Tempo adjustments to the TFR and cohort fertility forecasts both indicate that quantum change is part of the decline.  
The forecasts indicated that cohort fertility is likely to decline from the long-lasting level of 2 children to around 1.8 children on average for late 1980s cohorts. Here, Finland diverges from the other Nordic countries, as its expected cohort fertility is much lower (below 1.6). In turn, Sweden and Demark are on a trajectory of weaker declines than those observed in Finland, Norway, and Iceland. The new non-parametric approach that was developed in this study assessed potential recuperation patterns and yielded the weakest declines of all methods; nevertheless, it still showed that, particularly in Finland, Norway and Iceland, cohort fertility is likely to decline even if higher age fertility were to begin to increase.  
Using an incidence-based multistate Markov model, trends in age-specific transition probabilities across states of single life, cohabitation, marriage, and first birth among childless men and women showed that after 2010, first-childbearing decreased in unions, more (cohabiting) unions were dissolved, and marriage and cohabitation formation decreased. Counterfactual simulations revealed that the decline in fertility within unions mattered more (three-quarter) than changes in union dynamics (one-quarter) for the total decline in first births. First births declined more strongly across the lower social strata, but, across all strata, decreasing first-childbearing in unions explained most of the total decline. Trends in total fertility and first births in the 2010s across 153 fields of education showed diverging patterns in the already prevailing large differences between fields of education. Weaker declines (around -20% and less) were typically observed in fields with initially higher levels (health and teaching) and stronger declines (around -40% and more) in fields with initially lower levels (ICT, arts and humanities). Regression analyses indicated that the strength of the declines was associated with characteristics reflecting uncertainty (higher unemployment, lower income, and a lower share of work in the public sector) within the fields – together, these uncertainty measures explained one fourth of the decline in TFR and two fifths of the decline in first births.
The findings highlight the need to revise the conceptualization of the Nordic model of high and stable fertility. The decline in the 2010s was primarily accounted for by childless couples postponing or forgoing childbearing rather than by parents having smaller families. New theories increasingly highlight the role of perceived uncertainty in explaining fertility changes in the 2010s, but the findings from this study indicate that objective uncertainty also seems to be fuelling the fertility decline.