Childbearing during Recessions in the Nordic Countries

Martin Kolk, Stockholm University
Lars Dommermuth, Research Dep., Statistics Norway
Chiara Comolli, Stockholm University
Gunnar Andersson, Stockholm University
Trude Lappegard, University of Oslo
Peter Fallesen, Stockholm University

Empirical research shows that birth rates decline during economic downturns and upsurge during growth periods, however, many studies also shows that the fertility response varies significantly across recession episodes, countries, parities, age and social groups. This study contributes to the literature on the cyclicality of fertility by investigating childbearing risk in a comparative analysis of the five Nordic countries. Within this context, we harmonize high-quality register data from Denmark, Finland, Iceland, Norway and Sweden with the objective of comparing the impact on birth risk of the two most recent recessions that hit those economies: the 1990s and 2010s crises. We thus investigate 5x2 country-period cases of childbearing dynamics following the recession episodes. We use the harmonized longitudinal data to analyze the fertility history of women and examine the effect of economic downturns by age, cohort, and parity. For the analyses, we follow a previously tested method using event-history techniques to present parity-specific indices of childbearing risks during a given calendar year relative to a baseline year. To our knowledge, this is the first study to do this in a comparative framework.

Long abstract

The pro-cyclicality of fertility rates to business cycles has been extensively investigated, especially in the aftermath of the recent economic and financial crisis of 2008-2009. Empirical research shows that birth rates decline during economic downturns and upsurge during growth periods (among others Sobotka et al. 2011; Goldstein 2013; Neels et al. 2013; Comolli 2017). However, many studies also show that the fertility reaction varies significantly across recession episodes, countries, parities, age and social groups (Kreyenfeld et al. 2012; Neels et al. 2013). We contribute to the literature on the cyclicality of fertility by investigating the fluctuations in childbearing risk in a cross-country comparison of the Nordic countries and, within this context, we compare the impact on childbearing risk at different parities of two main recession episodes: the 1990s and 2010s crises. We harmonize high-quality register data from the five countries and use those longitudinal data to analyze the fertility history of women and examine the effect of economic downturns by age, cohort, and parity. To our knowledge, this has not been done within a comparative framework, as this literature seems to rely on aggregate date. The Nordic countries are universally recognized for their generous family policies and protective welfare systems, and for the solidity of their economies. Despite that, all five Nordic countries in the last decades have been shaken by two waves of economic turmoil - first in the 1990s and later during the Great Recession - and, with varying degrees, birth rates have fluctuated accordingly. Preliminary descriptive trends suggest that the Nordic countries share some similarities in their fertility reaction to the business cycles while still displaying a quite large variability, especially when comparing the two recession episodes. It is thus relevant to produce a cross-country comparison informative on the context specific childbearing response, across time and space. Some studies address the topic of fertility response to the business cycles separately in each of the five Nordic country but there are only a few comparative studies on Europe which include the Nordic countries. Neels et al. (2013) show that the deterioration of the economic context between 1970 and 2005 in 14 European countries including four Nordic countries (not Iceland) significantly slowed down first births, especially among men and women below the age of 30 and among the higher educated. Goldstein et al. (2013) do not find a significant fertility response to unemployment rates in the Nordic cluster (excluding Iceland) during the recent Great Recession, except for the early-twenties age group of women. However, further studies (Comolli 2017) suggest that when considering the more recent years of the crisis in Europe, the impact on fertility is larger than previously thought.

The first major contribution of this paper is, as mentioned, its cross-country comparative character and the use of high quality harmonized register data for five Nordic countries: Denmark, Finland, Iceland, Norway and Sweden. The latter offer full, accurate and reliable coverage of the local populations and their vital events. The second important contribution of this study is that we compare childbearing response to two main episodes of economic downturn that recently affected the Nordic economies: the 1990s crisis and the Great Recession of the years 2010s. For the analyses, we follow a previously tested approach (Hoem 1991, 1993; Andersson 1999, 2002, 2004; Kravdal 2002; Andersson and Kolk 2016) using event-history techniques to present parity-specific indices of childbearing risks, namely time series of relative risks of the propensity of women to give birth during a given calendar year relative to a baseline year. By applying event-history techniques to the combined population-register data of five countries, we are able to get an accurate picture of similarities or differences in childbearing response to the two episodes of recession in the five countries considered. We investigate thus 5x2 country-period cases of childbearing dynamics following the crises. In addition, in our calcula­tions, we standardize for age, parity, and the durations since any last previous birth, which means that we control for the effect of compositional changes among women over the categories of these variables. We focus explicitly on the period effect of the two recessions by presenting separate time series of relative risks for women in two ten-year periods: 1990-2000 vs. 2005-2015. Each period starts with a reference year (1990 and 2005 respectively) as baseline year from which the deviation from the relative risk is measured in the following years. In summary, we compare across countries and periods the relative risks of childbearing for each parity progression separately, based on the number of births and the appropriate population under risk of giving birth to a child of the actual order (Andersson 2004).

Presented in Session 1152: Fertility