Having children costs women's earnings, but the impact varies from country to country

The second half of the 20th century saw dramatic improvements in educational attainment and labour market outcomes for women in the Western world. However, progress in gender equality slowed around the turn of the century (Goldin 2012, Blau and Kahn 2018) and significant inequalities remain in all countries. Lower labour earnings for women arise from lower labour force participation, lower hourly wage rates, and lower hours worked. Along all these margins, the average woman fares worse than the average man in any country.

Recent research has highlighted the importance of children in understanding the persistence of gender inequality in labour market outcomes. Angelov et al. (2016) and Kleven et al. (2019), using data from Scandinavian countries, show that women experience a “’child penalty’ in earnings compared to men. This penalty persists up to 20 years after the birth of the first child and explains most of the remaining gender inequality.

Our latest research suggests that child penalties are a pervasive phenomenon across all developed countries. In a recent paper (Kleven et al. 2019a), we provide cross-country evidence for child penalties in earnings for six countries that span a wide range of policies and norms: two Scandinavian countries (Denmark and Sweden), two German-speaking countries (Germany and Austria), and two English-speaking countries (the UK and US). Using comparable data and a similar methodology across all six countries, the analysis reveals some striking similarities in the qualitative effects of children, but also some sharp differences in the magnitude of the effects.

Figure 1 shows the evolution of total labour market earnings of men and women over time, relative to the year of birth of their first child in Denmark and Sweden, for both mothers and fathers. In the years up to the birth of the first child, female and male earnings follow almost the same trend, but in the years just after the first birth the earnings of women drop by 30% in Denmark and by 60% in Sweden. While the labour market effect of having a first child is large on mothers, the impact of fathers is either non-existent or small. In Denmark, fathers’ earnings remain essentially unchanged, while a small dip after the birth of the first child in earnings is observed for Swedish fathers. This could reflect more generous parental leave and/or a higher incentive in the Swedish system for fathers to spend time when the child is still very small. However, the dip in Swedish fathers’ earnings is comparably small and fades away quickly with the age of the first child.

The bottom line is that despite the fact that Sweden and Denmark have implemented progressive family policies over the past decades, a substantial penalty remains on the labour market outcomes of mothers relative to fathers in both countries. In these countries, the child penalty on earnings for women amounts to around 20-25% ten years after the birth of the first child.

Figure 1 The child penalty in Denmark and Sweden

Notes: This figure plots our estimates for the child penalty in total labour earnings of women and men in Denmark and Sweden. The first child is born at event time zero. Our measure of earnings includes all labour income in a calendar year but excludes transfer payments. Our estimation methodology controls for the underlying wage growth in the economy and the effect of individuals getting older and more experienced over time.

Figures 2 plots the child penalties in the UK and the US, while Figure 3 plots the penalties for Austria and Germany. The general pattern remains the same – women experience a large, immediate, and persistent drop in earnings after the birth of their first child, while men are essentially unaffected.

Despite these similarities, the graphs also reveal some striking differences. In particular, the size of the long-run child penalty (defined as the average penalty from event time 5 to 10) differs substantially across countries. The Scandinavian countries feature long-run penalties of 21-26%, the English-speaking countries feature penalties of 31-44%, while the German-speaking countries feature penalties as high as 51-61%.

Figure 2 The child penalty in the UK and US

Figure 3 The child penalty in Austria and Germany

Explanations

These differences in child penalties are quite striking for countries that share a similar level of development.

One set of explanations for the differences in child penalties focuses on government policies. These include taxes, transfers, and family policies such as parental leave and child care provision that directly affect mothers’ incentive to work. There is a voluminous literature on the impact of such policies on female labour supply and gender gaps (for a review, see Olivetti and Petrongolo 2017). In ongoing work, we focus more closely on Austria, which has an exceptionally large long-run child penalty of more than 50%. We consider the impacts of parental leave and public child care provision; if these policies are the key to solving the remaining gender inequality, Austria would be the country where such policies should have the biggest effects. Our results so far suggest these policies have little or no effect on child penalties in the long run, while they are somewhat important in the short run.

If policies cannot explain the large differences in long-run child penalties across countries, then what can? A natural candidate revolves around gender norms and culture, but it is hard to provide conclusive evidence on the importance of such mechanisms. Kleven et al. (2019b) find that women in Denmark are more negatively affected by the arrival of a child if their mothers worked little compared to their fathers. This suggests that the intergenerational transmission of gender norms plays an important role in understanding the persistence of gender differences in earnings.

We test the potential role of gender norms by looking at survey evidence in our set of six countries. The International Social Survey Programme (ISSP) reports a variable that captures elicited gender norms. In this survey, participants were asked whether women with children under school age or in school should work outside the home (full-time or part-time) or stay at home. Figure 4 plots our estimate long-run child penalties in earnings against the fraction of respondents who think women should stay at home. The correlation between child penalties and gender norms is quite striking – the countries that feature larger child penalties are also characterised by much more gender-conservative views.

Figure 4 Long-run child penalties and gender norms