Job Market Paper
Information and Access in School Choice Systems: Evidence from New York City. [paper]
Abstract: Disadvantaged urban students attend lower-quality schools, on average, than their more advantaged urban peers. This paper asks how information about school quality affects this gap. Specifically, I examine the effects of New York City’s introduction of a letter-grade system rating the quality of its high schools. The ratings shifted Black and Hispanic students’ choices more than those of white and Asian students, narrowing racial gaps both in enrollment at high-quality schools and in academic achievement. Using a structural model of school choice and surveys of families, I find that race differences in the response to quality information stem in part from different beliefs and preferences. The model estimates suggest that Black and Hispanic students have less accurate perceptions of school quality, making them more receptive to the grade-based scoring system. In addition, white and Asian students are less influenced by information on school quality because they have strong preferences for other school attributes. Simulations suggest that better quality information narrows racial gaps in choice and achievement. A system that releases coarse quality ratings for high-quality or oversubscribed schools increases test scores among lower achieving students more than perfect information by reducing the competition for high-quality schools from higher achieving students.
Abstract: We investigate whether the introduction of the right to unilateral, no-fault, divorce for women has an impact on domestic abuse, investments in children’s human capital, women’s labor force participation, and other proxies of women’s agency in the context of the Egyptian Khul reform of 2000. We employ a difference in differences design, comparing mothers of children older than the age cutoffs used to assign the children’s custody to the mother, to mothers of younger children, before and after the reform. The first group of women is less affected by the legislative change in terms of being able to make credible divorce threats because it faces higher divorce costs, including the loss of alimony and the marital house. Results suggest that the introduction of Khul decreased domestic abuse and increased investments into children’s education while we do not find significant effects on labor force participation.
Abstract: Why aren’t workplaces better designed for women? We show that changing the priorities of those who set workplace policies can create female-friendly jobs. Starting in 2015, Brazil’s largest trade union federation made women central to its bargaining agenda. Using a difference-in-differences design that exploits variation in affiliation to the federation, we find that “bargaining for women” increases female-centric amenities in collective bargaining agreements, which are then reflected in practice. These changes lead women to queue for jobs at treated establishments and separate from them less—both revealed preference measures of firm value. We find no evidence that these gains come at the expense of employment, wages, or firm profits. Our results suggest that changing institutional priorities can narrow the gender compensation gap.
Abstract: Differences in school choice by race contribute to school segregation and unequal access to effective schools. Conditional on test score and district of residence, Black and Hispanic families consistently choose schools with fewer white and Asian students, lower average achievement, and lower value-added. This paper combines unique survey data and administrative data from NYC to study what drives these disparities and shows that attending a more integrated middle school can mitigate them. Our extensive post-application survey with guardians of high school applicants reveals that information gaps and homophily in school preferences explain cross-race differences in choice. Attending a more integrated middle school affects information and social preferences, reducing racial gaps in school choice. Instrumental variable estimates show that middle school students exposed to more peers from a different race apply to and enroll in high schools that are also more diverse. These effects are consistent across racial groups, particularly benefiting Black and Hispanic students who enroll in higher value-added high schools. Crucially, we show that these results cannot be explained by changes in middle school test scores but rather by shifts in the known school options and preferences for peer diversity.