Three seemingly unrelated essays in development economics

Mitigating risk, new technology adoption and improving health and education are key mechanisms for lifting households out of poverty in rural areas of developing countries. In this dissertation we use micro-level analysis to investigate these issues in the contexts of Vietnam, India and Kenya. Although all three chapters look at diverse topics in different settings, at the heart of each analysis is an attempt to address issues essential to improving the well-being of households in developing countries.;In Chapter 2 we draw on the full-insurance literature to examine the ability of households in Vietnam to smooth consumption when faced with idiosyncratic income shocks. Given the potential vulnerability of households in developing countries to weather shocks, illness and other sources of income variability, it is important to understand the extent to which households can cope with an uncertain income flow. We modify the standard approach taken in the full-insurance literature by focusing on smoothing of quantities, as opposed to expenditures. This is an important distinction because quantities, not expenditures, are what really matter for household welfare. In addition we allow for a more general characterization of preferences. We develop a simple model that leads to an estimation equation which we then test using panel data from Vietnam. We still reject full insurance across all goods, although the degree of insurance varies across goods. In particular we find that households are better able to smooth normal goods, such as rice, as opposed to more luxury goods such as meats.;Chapter 3 looks at the factors that drove technology adoption of high yielding variety seeds in India during the Green Revolution. We test alternative models of technology adoption using household level panel data from a nationally representative sample of rural Indian households from 1968-1971, years that correspond with the onset of the Green Revolution. The "price model" emphasizes input availability and price as the key determinants of the scale of adoption, while the "learning model" focuses on learning and experience. Using decision rules derived from these two alternative models we find that although both models capture certain aspects of the adoption decision, they each disregard important components of the alternative model. We then propose a third model that combines aspects of these two approaches and use it to characterize the decision both on the scale of adoption as well as whether or not to adopt.;Chapter 4 examines the long run health and education impacts of a deworming intervention in primary schools in Western Kenya. We collect a panel dataset of Kenyan youth from 1998 to 2005 to examine the medium to long term impacts of the intervention on health and education outcomes. Our results suggest that deworming treatment does have some medium-run effects, particularly on health. We find positive impacts of deworming on height, weight and a subjective measure of general health. These results seem to be largely driven by benefits to females, students in lower grades in 1998, and students living in higher infection areas, particularly high schistosomiasis areas. We find some impacts of deworming on education outcomes, most notably on dropout rates. Overall, our results suggest that a larger scale deworming project targeted at specific vulnerable sub-groups of the population may have fairly substantial health and education benefits.