Automation and Family: Marriage, Divorce, and Fertility Rates (thesis)

Abstract

I exploit variation in the job composition of US local labor markets to determine how automation affects familial outcomes from 1980 to 2019 and look at employment, marriage, divorce, and fertility rates as outcomes. Following previous work, I use historical prevalence of routine jobs to measure exposure to automation. I find that individuals susceptible to automation are more likely to become unemployed from 2000 to 2010; however, from 2010 to 2019 these individuals are more likely to become employed and work comparatively more hours. Fertility rates correlate with marriage rates; the results show higher fertility and marriage rates from 1990 to 2000 in local labor markets susceptible to automation, but this trend shifts from 2000 to 2019 as fertility and marriage rates become relatively lower. Moreover, I find that women in areas most susceptible to female-specific automation are more likely to experience lower marriage rates and are simultaneously are more likely to leave the labor force. This finding is correlated with higher fertility rates for women but is not correlated with the fertility rates of men, which may suggest that many women susceptible to automation are becoming single mothers and dropping out of the labor force.