Empowered mothers and co-resident grandmothers: Two fundamental roles of women impacting child health outcomes in Punjab, Pakistan.

We show that i) empowered mothers and ii) coresident grandmothers each benefit children's nutritional health measured by height-for-age z-scores (HAZ) and weight-for-age z-scores (WAZ) for age groups 5 years and less. First, using cross-sectional data from the Pakistan Demographic and Health Survey (PDHS) for the year 2017-18, we estimate the impact of empowered mothers on child health outcomes using an instrumental variable approach to correct for endogeneity. Empowerment is measured by two indices: as a sum of the questions that gauge both attitudinal and behavioral dimensions of female agency and also and using multiple correspondence analysis (MCA) for these same questions. Second, we use a fuzzy regression discontinuity design (FRDD) to measure the causal impact of coresident grandmothers on the health outcomes of the children using multiple rounds of the Multiple Indicator Cluster Survey (MICS) from the years 2008, 2011, 2014 and 2018. The difference between the actual ages of the grandmother from the Potential Retirement Eligibility Criteria (PREC) has been used to correct for potential endogeneity. The results show that on average, the weight for age z-scores (WFA) for children under five increases by 0.28 SD with a one-index point increase in mother's empowerment. Similarly, on average, WFA increases by 0.098 SD when grandmothers are present in a household. Finally, we explore heterogeneity in the average effects stated above based upon the gender of the child as well as the wealth and geographic location of the household. The benefits of mothers' empowerment are largely driven by improvements in girls' nutrition as well as children living in rural areas while the presence of grandmothers primarily improves the nutrition of boys, children in rural areas, and children belonging to poor families.