Inframarginal treatment effects: Estimating the welfare losses from groundwater depletion in India

Seminar | September 28 | 12:10-1:30 p.m. | 248 Giannini Hall

 John Loeser/Job Market Presentation, ARE

 Agricultural & Resource Economics

Abstract
Market failures in the management of groundwater, a common pool resource, have contributed to groundwater depletion in arid districts in India. An estimate of the effect of groundwater depletion on welfare is necessary to determine the optimal policy response. This paper does so by defining and estimating the Inframarginal Treatment Effect (ITE), the effect on welfare implied by an observed change in adoption of treatment (in this paper, irrigation) caused by a shift in potential utility under treatment (in this paper, groundwater depletion). I derive sufficient conditions for the ITE to be well defined and policy invariant. I show that a Local Average ITE can be estimated from the difference between two linear IV estimators, the first of which uses an instrument satisfying a non-standard exclusion restriction. This approach extends conventional revealed preference approaches by allowing estimation of welfare effects without an observable price of treatment. In my setting, I construct instruments for costs of irrigation using hydrogeological characteristics, and for potential revenue under irrigation using climate and soil characteristics. The estimated Local Average ITE implies a one standard deviation decrease in access to groundwater would cause welfare losses equal to 12.4% of agricultural revenue, while the estimated LATE implies observed agricultural revenue would fall by just 8.9%.

 CA, lazo@berkeley.edu, 510-642.3345