Abstract

My dissertation is comprised of three independent empirical chapters. Below is a brief description of each.

The first Chapter is titled "Infant Mortality Rates in India: District-Level Variations and Correlations". This paper examines the correlates of infant mortality in India using district-level data from the 1991 and 2001 Census of India. In our empirical work, the explanatory variables used are male and female literacy, male and female labor force participation, the level of poverty, urbanization and other socio-economic variables. We use quantile regression analysis to determine which of these factors impact infant mortality. Quantile regression is preferred over OLS because it allows us to estimate models for the conditional median function, and the full range of other conditional quintile functions and therefore provides a more complete statistical analysis of the stochastic relationship among random variables. The analysis brings out the powerful influence of women's characteristics on infant mortality, especially literacy and labor force participation. Increases in both of these variables significantly reduce child mortality at the district level. Improvements in male laborers in non-agricultural work and reductions in poverty also reduce child mortality, but their quantitative impact is weak in comparison. Further the non-parametric analysis reinforces the results found in the parametric section. They indicate that the action or the impact of the covariates is strongest in the districts which lie in the center of the conditional distribution, rather than those at the extreme. This analysis allows us to determine in which districts the impact of additional target policies would yield the greatest reduction in infant mortality.

The second paper is titled "Same-sex siblings and their affect on mothers labor supply in South Africa". This study aims to look at the labor-supply consequences of childbearing for women in South Africa. The data used for the study is the 10% household sample from the 2001 census. The covariate of interest in the labor supply model is the indictor More than two children. Demographic variables include mother's age, age of the mother at first birth, years of schooling, indicators for race, and an urban dummy. Labor-supply variables include hours worked per week, worked for pay and total income. Unlike previous studies which restrict their sample to include only female household heads or spouse of male household heads, this study expands the sample size to include all childbearing women in the household aged 15-35. The IV estimates that exploit the fertility consequences of sibling sex do not confirm the OLS estimates showing that more children lead to lowering of female labor supply. While OLS estimates exaggerate the causal effect of children, children seem to have a smaller effect on the labor supply of college-educated women. I find that labor market outcomes of childbearing are more severe for married women in South Africa.

The third paper is on "The effect of Immigration on Ethnic Composition and Occupational Reallocation". We aim to assess whether the impact of immigration is mitigated by occupational transition of natives. Being over represented among HSDs, we focus on the labor market outcomes for Black workers. We use data from the 5% public use sample of the census (1980, 1990 and 2000) as well as the 1% sample of the population from the American Community Survey (2005, 2006 and 2007) to estimate the effect of occupational reallocation on the wages of Black workers as well as the effect of immigration on reallocation. A shift-share analysis reveals that occupational transitions caused wages for Blacks to rise by 46% more than they would have with a static occupational distribution. However, we find that these occupational shifts were due to crowding out effect of Hispanics on Black occupations: a 10 percentage point increase in the share of workers in an occupation who are Hispanics leads to a 5 percentage point decrease in the share of Black workers in that occupation. This is significantly large to explain substantially, occupations that declined in importance for Blacks during the period of study. We find a strong correlation between importance of occupations to Hispanic and Blacks, suggesting that most occupational transition for these two groups has not only been driven by outside factors such as trade and technological change, but that these shocks are affecting the two groups similarly. (Abstract shortened by UMI.)