Essays on gender and children

This thesis investigates the causes of childhood outcomes. In Chapter 1, written with Elisabetta De Cao, we show that unemployment causes child neglect in the United States, during the period from 2004 to 2012. We identify this effect using a Bartik instrument, which we create by combining national-level unemployment rates across industries with differences in the initial industrial structure across counties. We find evidence that parents in the United States lack the social and private safety nets needed to provide for their children's basic needs in the event of a job loss. In Chapter 2, written with Sonia Bhalotra, we find that fathers work less after having a first born son rather than a daughter in the United Kingdom, during the period from 2009 to 2016. The effect occurs on both the extensive and intensive margins, and in the first year of the child's life. Our identification strategy exploits the randomness of the gender of the first child. We find evidence that degree educated fathers and their partners both take time off work to increase the total parental time invested in sons, because boys' development is less advanced than girls' development at a young age. We find evidence that non-degree educated fathers take time off work to reduce the burden of childcare on their partners, which helps their partners to return to work sooner after childbirth. Finally, in Chapter 3, I find that female leaders elected to the State Legislative Assembly improve child learning in their districts in rural India, during the period from 2008 to 2014. The effect occurs for both reading and mathematics, and the gender gap in mathematics narrows. To identify this effect, I instrument for the fraction of seats won by a woman in a district with the fraction of seats won by a woman in a close election against a man. The most plausible explanation for the effects is that female leaders try to improve educational outcomes around the time of an election, in the hope of re-election.