Working Papers
Institutional Reforms and Child Health
Religious legal reforms shape child welfare in pluralistic societies. This paper examines Nigeria's 2000 Sharia Penal Code reform, adopted by 12 northern states with varying enforcement intensity. Using Demographic and Health Survey data from 1990 to 2018 and difference-in-differences estimation, I find the reform reduced child anthropometric measures by 0.23 to 0.27 standard deviations. Effects concentrated among non-Muslim minorities, who experienced declines of 0.76 standard deviations in height-for-age and 0.44 standard deviations in weight-for-age, approximately three times larger than Muslims. Muslims showed mixed effects: no significant changes in height-for-age but declines in weight-for-height. Enforcement intensity shaped outcomes for non-Muslims. A theoretical model and empirical analysis reveal two mechanisms. Among Muslims, stronger enforcement encouraged health investment but increased son-biased fertility. These opposing effects offset one another. Weak enforcement enabled (perceived) discriminatory service denial among non-Muslims, while strong enforcement protected access. These findings demonstrate how religious legal reforms create unintended spillovers that harm minority populations, with enforcement strength determining the scope for discrimination.
Feeding Fairly (With Teresa Molina-Millàn, Emily Ouma, and Nils Teufil)
This paper uses a clustered randomized controlled trial across 240 villages in rural Uganda to study whether informational or norm-based constraints limit maternal nutrition interventions, comparing women-only nutrition sessions, couple-based sessions, and a couple-based intervention that additionally addresses food allocation norms. Providing information to both spouses substantially increases nutrition knowledge for men and women, with comparable gains across the two couple-based arms. This, however, generates limited improvements in dietary outcomes. The combined nutrition and norms intervention reduces adherence to conservative food allocation norms, substantially increases women's and children's minimum dietary diversity, and raises animal-source food consumption for both groups. Comparable knowledge gains but divergent dietary outcomes across the two couple-based arms suggest norms as the binding constraint on dietary behavior. These gains occur without increases in total household food expenditure. They appear to operate through a repositioning of men as active participants in the nutrition domain, with reductions in eat-out expenditures, increases in pro-child allocation behavior, and a contraction of women's sole decision-making authority toward joint arrangements.
Social Norms Under Observation
Does gender matching in household surveys improve the truthfulness of responses, or does it elicit answers that conform to the perceived expectations of a same-gender enumerator? I provide causal evidence using randomized enumerator assignment to 2,400 respondents in rural Uganda. Enumerator effects are minimal for objective measures but substantial for sensitive topics like knowledge, household activities, and social norms. Gender matching also produces asymmetric effects, relative to unmatched pairs. Male-male pairs report more conservative attitudes and higher dietary taboo prevalence, while female-female pairs express more progressive views and greater involvement in household decision-making. An embedded priming experiment shows that response shifts occur specifically when the enumerator is female, regardless of respondent gender, consistent with a female interviewer effect. These findings challenge conventional gender matching protocols in household surveys.
Convergence or Reversal? (With Damien de Walque and Carolina Lopez)
The generational evolution of gender gaps in education in Sub-Saharan Africa remains poorly documented at the regional scale. Using micro DHS and MICS data from 35 countries (approximately 736,000 individuals), complemented by PASEC test scores for 14 francophone countries, we trace gender differences in educational attainment and learning by birth cohort. Attainment gaps narrow sharply across cohorts: the female disadvantage falls from 13--15 percentage points in primary and secondary attendance and 1.9 years of schooling among those born in the 1970s to far smaller values among the most recent cohorts, and is statistically insignificant across most of the youngest birth cohorts. Among those born after 1999, females exceed males in secondary attendance and years of education in a majority of countries. Convergence, however, is uneven. It partly reflects stagnation in boys' secondary attendance rather than female gains, learning gaps show no uniform direction across countries and often run counter to attainment, and wider attainment gaps persist among rural, poorer, and Muslim populations.
Works in Progress
Convergence or Reversal? The Evolution of the Gender Gap in Education in India (With Damien de Walque)
Sibling Gender and Child Labor: Evidence from Patrilineal and Matrilineal Societies (With Chantale Oweggi)
Social Constraints to Food Consumption: A Qualitative Study in Uganda. (With Esther Achandi and Emily Ouma)