​Subjective Expectations and Demand for Contraception (with Grant Miller and Aureo de Paula). Policy Briefing. Minor Revisions requested by the Journal of Econometrics.

​

Abstract:

One-quarter of married, fertile-age women in Sub-Saharan Africa report not wanting a pregnancy and yet do not practice contraception.  We collect detailed data on the subjective beliefs of married, adult women not wanting a pregnancy and estimate a structural model of contraceptive choices.  Both our structural model and a validation exercise using an exogenous shock to beliefs show that correcting women's beliefs about pregnancy risk absent contraception can increase use considerably.  Our structural estimates further indicate that costly interventions like eliminating supply constraints would only modestly increase contraceptive use, while confirming the importance of partners' preferences highlighted in related literature.

​

Saving Neonatal Lives at Scale: Lessons for Targeting (with Hans H. Sievertsen and Mahesh C. Puri). This supersedes a paper circulated under the title "Saving Neonatal Lives for a Quarter".

​

Abstract:

Neonatal mortality contributes an increasing share of under-5 mortality. Experimental estimates of a low-cost preventive measure (chlorhexidine cord care) vary widely across settings, leading to external validity concerns. We provide the first, quasi-experimental, estimates of the effect of a nationwide roll-out in Nepal and apply recently developed machine-learning techniques (ML) to analyze treatment effect heterogeneity. We find that the program decreases neonatal mortality by 36% and that a simple targeting policy leveraging heterogeneous treatment effects improves neonatal survival relative to WHO recommendations. Heterogeneous treatment effects extrapolated from our ML analysis are broadly in line with experimental findings across five countries.​​

​

Women's (and Men's) Probabilistic Beliefs about Contraception and Contraceptive Use in a High-Fertility Environment (with Grant Miller, Bintu Ibrahim Abba and Aureo de Paula). Draft available upon request.

​

Abstract:

Over one in five married women in Sub-Saharan Africa report not wishing to become pregnant, but also not using any form of contraception---often not because of inadequate supply or cost of contraceptives. In this paper, we use detailed new data from a large, diverse sample of women and their husbands across Nigeria (the most populous country in Sub-Saharan Africa) to study the accuracy of probabilistic beliefs about pregnancy risk and contraception and their relationship to actual contraceptive choices. We document systematically mistaken beliefs held by respondents, and we find that two are also strongly related to actual contraceptive choices: women’s underestimation of pregnancy risk absent contraception and women’s mistaken beliefs about their partners’ approval of contraception. Partner approval may not easily be amenable to change, but importantly, our results suggest that credible, contextually-tailored information about pregnancy risk absent contraception could align women’s contraceptive choices more closely with their fertility desires and therefore reduce unwanted pregnancies.

​

Parental Minimum Wages, Children's Education, and Racial Inequality (with Luyang Chen and Hans H. Sievertsen).

​

Abstract

Parental income inequality contributes to disparities in educational outcomes. One of the most widely implemented strategies to address income inequality is the establishment of minimum wage mandates. Yet we do not know whether the children of those who benefit from these mandates gain more education. In this paper, we study the effect of the 1966 Fair Labor Standards Act (FLSA), which successfully reduced racial wage inequality, on the long-term education outcomes of teenage children whose parents worked in industries covered by the 1966 FLSA. Exploiting variation in exposure across cohorts and parental occupation, we find that the new minimum wage mandate improved second-generation completed education. This effect is larger for black children, contributing to reduced racial inequality in completed education. The evidence points to 1966 FLSA improving the long-term education outcomes of teenage children by relaxing the budget constraint of households headed by individuals working in industries covered by this reform. Indeed, we show that second-generation education effects are larger for groups who experience larger first-generation wage increases, and find suggestive evidence, among young black men, of reduced teenage labor force participation and of a reduction in school drop-out due to financial difficulties.