Research

Health Economics & Policy; Labor Economics; Economics of Education; Economics of Public Policy; Business Research

Job market paper

Breastfeeding and Child Development Outcomes across Early Childhood and Adolescence: Doubly Robust Estimation with Machine Learning

Released as IZA Discussion Paper #17080 

Presentations: SEA Annual Meetings 2023 


⇰ Using data from the Panel Study of Income Dynamics, we estimate the impact of breastfeeding initiation and duration on multiple cognitive, health, and behavioral outcomes at two points in time, spanning early childhood through adolescence. To mitigate the potential bias from misspecification, we employ the doubly robust (DR) estimation method, which addresses any misspecification in either the treatment (i.e., the propensity score) or outcome models and adjusts for selection effects. Rather than adopting a probit/logit model to estimate the propensity score (PS), our novel approach is to use and evaluate a battery of supervised multiple machine learning (ML) algorithms to improve estimates of the PS. We demonstrate that the gradient boosting machine (GBM) algorithm removes bias more effectively and minimizes other prediction errors compared to logit and probit models as well as alternative ML algorithms. Across all outcomes, our DR-GBM estimation generally yields lower estimates than OLS, DR, and PSM using standard and alternative ML algorithms and even sibling fixed effects estimates. We find that breastfeeding is significantly linked to multiple improved cognitive outcomes at an early age. However, the impact reduces with time as children grow older. In contrast, we find mixed evidence regarding the impact of breastfeeding on non-cognitive (health and behavioral) outcomes, with the effect being most pronounced in adolescence. Our results also suggest that breastfed children of minority mothers receive higher cognitive benefits compared to those of non-Hispanic white mothers. Furthermore, children of mothers with at least some post-high school education experience relatively greater benefits from breastfeeding. Regarding breastfeeding duration, results suggest minimal marginal benefits beyond 12 months for cognitive outcomes and 6 months for non-cognitive outcomes. 

Working papers

Explaining the Rural-Urban Gaps in Women’s Nutritional Outcomes in Bangladesh: Evidence from Counterfactual Decomposition Methods

Presentations:  SEA Annual Meetings 2023


⇰ The rural-urban disparity in women’s health outcomes has been largely overlooked in the literature, particularly in the context of developing countries. Applying Oaxaca-Blinder and quantile regression-based counterfactual decomposition methods to the data from the latest five rounds (2004-2018) of Bangladesh Demographic Health Surveys (DHS),  I identify sources of the sizeable differences in body mass index (BMI), overweight, and obesity across rural and urban areas among women of childbearing age. I consider socio-economic, healthcare access, health and fertility behaviors, and infrastructure-related variables as potential determinants of this rural-urban gap, including differing levels and returns to these factors across areas. Additionally, my empirical strategy allows me to investigate contributors to the rural-urban gap across the distribution of nutritional outcomes. Estimates reveal that disparities in education, healthcare access, infrastructural developments, and inequitable distribution of wealth explain a major fraction of the rural-urban gap in nutritional outcomes. I also find that sizeable gaps are attributable to differential health-related returns to education, institutional births, and wealth status. These findings suggest that experiencing unbalanced economic growth between rural and urban areas drives health disparities in developing countries, highlighting the need for institutionally different policies and strategies in rural and urban settings.

The Long-Term Impacts of Breastfeeding on Educational Attainment of Adults

⇰  Using data from the National Longitudinal Study of Adolescent Health, we examine the impact of breastfeeding on the likelihood of obtaining a college degree and attending graduate school. Using a doubly robust estimation technique, we find that having been breastfed is linked to an increased probability of obtaining a college degree, but its association with graduate school attendance is largely insignificant. The effect on college degree attainment is primarily concentrated among males, blacks, and whites. Investigating potential channels, we find that these effects may be operating through the impact of breastfeeding on adolescents’ cognitive ability.

Selected Publications


Pre-PhD Research:

Work in progress