Three Essays on Gender and Development

This dissertation consists of three essays on gender and development. The first essay of this dissertation is the first to evaluate the gendered effects of child grants on patterns of time allocation across SNA (System of National Accounts) production work, household maintenance, care work, leisure, self-care, and other non-work activities. SNA production work includes paid market work, subsistence and informal work, job search, and other production activities which standard labor market indicators generally fail to capture. In this essay, I use data from the 2010 South Africa Time Use Survey on grant-eligible single parents aged 20-54 years to estimate a system of equations describing the time allocation of single parents. I address the endogeneity of the key grant receipt parameter using a probit model with an originally-constructed instrumental variable, regional median travel time to the welfare office. I find that single fathers living in grant recipient households reduce SNA production work by 22.5 percent (61.5 minutes per day) and single mothers by 61.5 percent (116.3 minutes per day). Single parents primarily redistribute their reduced SNA production work time to household maintenance and care work. Single fathers increase their time in household maintenance and care work by 72.2 percent (81.8 minutes per day) and single mothers by 62.8 percent (142.1minutes per day), respectively. This rise in household maintenance and care work leads to an overall increase in total work time, especially of single mothers. Single mothers living in grant recipient households increase their total work time by 5.4 percent, which is an increase of 25.8 minutes per day. A series of robustness checks confirms the results. The second essay of this dissertation focuses on measuring the impact of unpaid elder caregiving in the US on labor force participation. The essay also explores the association between unpaid elder caregiving and time use of the caregivers. The need for better understanding and measuring unpaid eldercare has become an urgent and pressing issue given the trend towards population aging. This study takes into account the diversity of eldercare arrangements and focuses on those who provide unpaid eldercare on a frequent basis (daily or several times a week). Using 2011-17 American Time Use Survey data for a subsample of individuals aged 25-61-years old, the essay examines the effect of frequent eldercare provision on labor force participation using a bivariate probit IV approach. The study also examines the time allocation pattern of frequent eldercare providers using multivariate regression method. The findings suggest that frequent eldercare provision reduces labor force participation, especially among male caregivers. Additionally, frequent eldercare provision is associated with more time spent on domestic chores and significantly less time on market work and self-care. Robustness and sensitivity checks confirm these findings. Finally, the third essay of this dissertation evaluates the relationship between sex-based attitudes of mothers and stunting among children in India. India is in a major malnutrition crisis, topping the list of countries with the largest number of stunted children. In this essay, I examine the relationship between three distinct attitudes of mothers (pro-boy, egalitarian, and pro-girl) and stunting among boys and girls of age 0-14 years in India using the Indian Human Development Survey (IHDS) 2004-05. I estimate a probit model to assess the association between mothers’ attitudes and childhood stunting. The findings suggest that having egalitarian attitudes is not associated with stunting among their children. However, mothers with pro-girl attitudes are 8 and 10 percentage points less likely to observe stunting among their girls and boys, respectively. Additional analysis by wealth categories shows that stunting among girls reduces by 15 percentage points when they have mothers with pro-girl attitudes, and they live in wealthy households. Robustness tests conducted with “severely stunted” as the dependent variable confirm the findings.