This dissertation consists of three empirical studies. The first study considers how expansion of communication-intensive economic activity in mid-20 th century India led to increased bilingualism and the consolidation of spoken languages. It relates to a worldwide phenomenon: A small set of existing languages has substantially expanded their share of the world population in recent centuries. I argue one important cause is that productivity is higher in certain activities, such as factory work and urban living, when participants can communicate with each other. I model the relationships between factory employment and language acquisition and explore the implications using a new panel of Indian districts for 1931 and 1961. Instrumental variables estimates show manufacturing employment growth strongly encouraged bilingualism, particularly among minority language speakers. In turn, bilingualism among mother tongue speakers of a language leads to its relative decline.
The second study explores deindustrialization in colonial India, long described by historians but for which evidence is thin. Previously unexploited data on prices enables a price-dual analysis based on a three-sector neo-Ricardian model of deindustrialization. The study assesses how much of India's deindustrialization was due to local supply-side forces and how much to world price shocks. The analysis pushes the roots of deindustrialization to the mid-18 th century, where political instability and unusually frequent drought depressed agricultural productivity. Deindustrialization accelerated in the early 19 th century as textile prices fell steeply following the industrial revolution and prices for India's export commodities rose. Local supply side forces dominated early, and world price shocks later.
The third study concerns health service delivery in developing countries, where distortions in both government and private fee-for-service provision are severe. Cambodia recently tried a new approach: It contracted out management of government health services to NGOs in five districts that had been randomly made eligible. The contracts specified targets for maternal and child health service improvement. Targeted outcomes are estimated to have improved dramatically while changes in non-targeted outcomes were small. Public health funding was increased, but the program led to roughly offsetting reductions in private expenditure as residents in treated districts switched from unlicensed drug sellers to government clinics.
My Research