Essays in empirical microeconomics

This thesis consists of three essays addressing open empirical questions in applied microeconomics. Chapter 1 attempts to quantify the impact of climate change on Indian agriculture. I use historical data on past yearly weather fluctuations and crop yields to measure the effect of these weather fluctuations on output, then use climate change prediction models to derive projections of the impact of future climate change on future productivity. I find that even moderate climate change could be seriously detrimental to productivity, with a consensus prediction for warming over the period 2010-2039 reducing productivity 4.5 to 9 percent. Chapter 2 provides a new tool for analysis of distributional, or quantile, effects in regression discontinuity (RD) models. RD has become increasingly popular over the last decade as a method of obtaining quasi experimental estimates of mean treatment effects. This paper extends the methodology to the measurement of quantile treatment effects. I provide simulation evidence on the effectiveness of the estimator and an empirical application to returns to compulsory schooling in the United Kingdom. Chapter 3, written jointly with Esther Duflo and Michael Greenstone, examines the impact of a water and sanitation intervention in Orissa, India, on health outcomes, in particular the monthly incidence of severe cases of diarrhea and malaria. The design of the intervention, in particular the fact that the water system is activated suddenly, unpredictably and simultaneously for all households in a given village, allow us to overcome several empirical challenges that have impeded credible estimation in the past. We find large effects: the arrival of services appears to reduce severe cases of diarrhea by as much as forty percent, with similar effects on severe cases of malaria. Furthermore, these effects appear to be persistent, as they continue to be apparent in the data after three and even five years.