Water, Sanitation, and Development: Household Preferences and Long-Term Impacts

Abstract

This dissertation seeks to measure the long-term impacts of sanitation improvements in India and demand for water reliability improvements in Jordan. Throughout, I use program evaluation and nonmarket valuation tools to contribute to water and sanitation topics in the environmental economics, development economics, and public health literatures. In the first chapter, I use data gathered in Odisha, India to provide the first long-term follow-up of a sanitation intervention. In 2005, villages in Odisha were randomized to receive an intensive information campaign that highlighted the external community benefits of latrine adoption. Using difference-in-differences analysis, I find that communities that received the information campaign continue to have higher rates of latrine adoption for five years, at which point the rate of adoption slows considerably. In exploring the dynamics of latrine ownership, I find that households receiving the information campaign were less likely to abandon latrines. However, out of households that abandoned latrines, these treatment households were less likely to re-adopt latrines. This paper is the first to explore adoption and abandonment behaviors over such a long time period. The second chapter uses the same experimental design as Chapter One to explore the impacts of having a household latrine in one's early childhood on long-term cognitive development. While both development economics and public health literatures have documented large child health gains from improved sanitation, this paper is one of the first to explore how improved sanitation contributes to long-run human capital development. Using treatment assignment as an instrument for latrine adoption, I find that children who belonged to a household with a latrine between the ages of 0 and 5 score significantly higher on a cognitive test designed to measure analytic ability ten years later. I find that this effect is largely concentrated in female children. I explore multiple mechanisms examining the difference in effect size between genders and provide suggestive evidence that since girls receive fewer early-life health investments, a concave human capital development curve would imply that the impact of having a latrine is larger for girls. In Chapter Three, I transition from measuring the impacts of water and sanitation improvements to characterizing demand for improvements to water reliability. I use two nonmarket valuation techniques--contingent valuation and averting expenditures--to estimate demand for water reliability improvements in urban Jordan. Traditionally, averting expenditures (a revealed preference measure) have been considered a lower bound for demand relative to contingent valuation (a stated preference measure) estimates. I develop a theoretical model that demonstrates that this relationship critically depends on household perceptions. In our setting, this insight is important, because households have relatively low confidence in both the reliability and quality of existing water supplies, even though water quality tests suggest that utility water is safe to drink from a microbial perspective. Averting expenditures, which reach 4% of monthly expenditures on average, thus include substantial purchases of non-network water sourced from water shops or tankers, as well as costs in terms of water collection time, storage and in-home treatment. In contrast, the contingent valuation responses, while correlated with coping costs, reveal low willingness to pay for increases in water reliability from the utility network. I attribute this departure from the traditional relationship between averting expenditures and contingent valuation to the lack of household confidence in the quality of utility-provided water. This study thus adds to previous evidence in the literature, which points to the importance of consumer perceptions in determining demand for environmental improvements.