Identification and Estimation of Spatial Causal Effects

• Activity Number: 525
• Type: Topic Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Peter M. Aronow* and Cyrus Samii
• Companies: Yale University and New York University
• Keywords: causal inference ; randomized experiments ; field experiments ; spatial estimation ; identification

Abstract

We derive identification results and semiparametric estimators for causal inference for experimental treatments administered on K-dimensional spaces. We parametrize exposures as functions of distance from points of treatment, and define our inferential targets as potential outcomes that marginalize over both regime and space. We propose a class of Kriging-type estimators that are consistent under mild smoothness conditions. We apply our approach to data from a canonical field experiment on the causal effects of deworming treatments on health and academic outcomes.