Econometrics: Treatment effect estimation with noisy conditioning variables
Published: 8 February 2022
18 February. Professor Kenichi Nagasawa, University of Warwick
Professor Kenichi Nagasawa, University of Warwick
'Treatment effect estimation with noisy conditioning variables'
February 18, 3.00pm-4.00pm
Hybrid: Room 305, University of Glasgow and Zoom online seminar
Abstract
When estimating causal effects, controlling for confounding factors is crucial, but these characteristics may not be observed. A widely adopted approach is to use proxy variables in place of the unobserved ideal controls. However, this approach generally suffers from measurement error bias. In this paper, I develop a new identification strategy that addresses this issue. I use proxy variables to construct a random variable conditional on which treatment variables become exogenous. The key idea is that, under appropriate conditions, there exists a one-to-one mapping between the distribution of unobserved confounding factors and the distribution of proxies. To satisfy overlap/support conditions, I use an additional variable, termed excluded variable, which satisfies certain exclusion restrictions and relevance conditions. I also establish asymptotic distributional results for flexible parametric and nonparametric estimators of the average structural function.
Biography
Kenichi Nagasawa is an Assistant Professor of Economics at the University of Warwick. He has studied econometric theory with focus on bootstrap inference in non-standard settings and causal inference under measurement errors. He earned a PhD in Economics and an MA in Statistics in 2019 from the University of Michigan.
Further information: business-events@glasgow.ac.uk
First published: 8 February 2022
<< 2022