The Effect of Children on Earnings Using Exogenous Variation in Family Size: Swedish Evidence

This paper takes advantage of an exogeneous variation in the sex composition of previous children, to study the effect of an additional child on women’s earnings. I use OLS and IV as well as quantile regression to analyze the impact of an increase in family size on labour force participation and level of earnings from 1980-2005 Swedish register data. The IV technique produces estimates that are not systematically different from those from OLS, at the expense of a low precision. Including men in the analysis shows that fathers’ labour force outcomes are less likely to be affected by an increase in family size compared to mothers. My findings indicate that having an additional child has a stronger negative impact on earnings than on labour force participation. However, there is evidence of catching-up effect over time, as women tend to recover gradually from the negative earnings effect. Using different time perspective, the results remain stable with respect to the rapid expansion of the Swedish family policies. The quantile regression approach suggests that other mechanisms than childbearing lie behind the large wage gap at the top of the wage distribution, often referred to, in Sweden, as the glass ceiling pattern.