In this project I link divorce filings, which include granted and dismissed divorce cases, in Michigan from 1900 to 1920 to the full-count census to study the long-run effects of divorce. I use a leave-one-out instrumental variable approach at the county-year level to proxy for a court's leniency in granting divorces. Using this IV strategy, I find that women who marginally divorce are 40 percentage points less likely to be living with their former partner, are 15 percentage points more likely to be divorced, and 22 percentage points more likely to be remarried to a new partner. Women are also 9 percentage points more likely to enter the labor force and 7 percentage points more likely to live with parents. In contrast to existing literature, I find evidence that divorce was a positive shock to families. Daughters in adolescence at the time of the divorce filing earn an additional year of education and marry higher-earning partners. I also find spillover effects of divorce: young women whose neighbors marginally divorced are 6 percentage points more likely to enter the labor force and are 4 percentage points less likely to marry. This highlights how exposure to divorce may alter young women's preferences between work and family.

The 1922 Cable Act repealed automatic citizenship for foreign-born women marrying citizens after a specific date, effectively introducing naturalization requirements for first-generation female immigrants in the US. Using newly digitized marriage records and a regression discontinuity design around couples' marriage dates, I study the effects of naturalization requirements for women on their and their children's integration outcomes. Naturalization requirements reduced married women's probability of obtaining citizenship by 15-18 percentage points in the medium run but did not affect citizenship status in the long run. Naturalization requirements negatively affected women's integration, decreasing their probability of speaking English by 3 p.p. and increasing the likelihood that daughters are given names common among immigrants, but otherwise had no effects on women's labor force participation or employment. I also find some evidence that sons are less socially integrated: they are 3 p.p. less likely to volunteer for WWII and possibly less likely to marry natives. These results highlight the unique role that mothers' access to citizenship plays in creating a home environment for social integration.  

Women's increased participation in the workforce over the past century was the most significant change in the US labor market (Goldin, 2006). An often-cited but understudied reason for women's increased LFP was the elimination of the marriage bar, a policy prohibiting married women from working. We gathered new data from 1900-1940 to document the prevalence of the marriage bar in teaching across US cities. Using the 1880-1940 Census and a difference-in-differences design, we show that marriage bars decreased the share of married female teachers by 2.3 percentage points (15%). This decrease was offset by an increase in single female teachers. We also find suggestive evidence that marriage bars increased retention among single teachers and possibly delayed or deterred marriage, particularly for older women and those exposed to the policy longer. These findings highlight the role of discriminatory employment policies in affecting workforces and workers' family decisions. 

Abstract: The Special Supplemental Nutrition Program for Women, Infants, and Children, popularly known as WIC, serves about 40 percent of all infants in the US. The largest expense for the federal government in this program is infant formula. This paper explores the health effects of the most prominent cost-containment effort in the WIC program— the introduction of sole-source infant formula contracts awarded to manufacturers through a competitive bidding process—the “rebate system”—in the late 1980s. Exploiting the staggered adoption of this system across states, we find that sole-source contracting reduced the incidence of low birth weight by 1.4 percent and increased average birth weight by 5.7 grams. There are particularly large effects for Black infants, whose incidence of low birth weight decreased by 3.6 percent and whose average birth weight increased by 11 grams. Evidence from enrollment and per-participant expenditure data suggests these health improvements were driven by increased WIC enrollment rather than changes in the quality of the food package.