The Labor Market Returns to Delaying Pregnancy (with Yana Gallen and Juanna S. Joensen)

For women, the birth of a new child is associated with large and persistent declines in earnings (Kleven, Landais, & Søgaard, 2019; Angelov, Johansson & Lindahl, 2016). These large costs may induce women to invest more in careers which are less affected by childbirth, or to time children to reduce their career impact (Adda, Dustmann & Stevens, 2017;  Park & Rim, 2020). The impact of children on women's careers is central to understanding the gender pay disparities more generally.

Causal research on the timing of children is challenging precisely because this timing is so important for careers and families---experiments that shift the timing of childbirth are rare. In this project, we will study the career and family outcomes of women who become pregnant while using long-acting reversible birth control, in particular, Intrauterine Devices (IUDs) and the birth control shot (DMPA). These methods of birth control are extremely effective, but not perfect---about 0.5% of women using a non-expired IUD will get pregnant in a year, resulting in a natural experiment in which women who had hoped to delay childbirth become pregnant earlier than they expected.

Our data come from several Swedish administrative registers. We link hospital records, medical prescriptions for birth control, out-patient doctor visits and abortion appointments, infant health records, tax records, and additional socioeconomic and labor market information for all women in Sweden born between 1965 and 1983.