Abstract

Following Flores and Flores-Lagunes (2009) I will analyze the causal effect of a treatment on an outcome, and reveal the mechanisms or channels through which the treatment works. In this training I will introduce the concept of net and mechanism average treatment effects (NATE and MATE, respectively), which provide an intuitive decomposition of the total average treatment effect (ATE) that enables learning about how the treatment affects the outcome. I will then derive informative non-parametric bounds for these two effects allowing for heterogeneous effects, without requiring the use of an instrumental variable or having an outcome with bounded support. Weak monotonicity of mean potential outcomes within or across subpopulations defined by the potential values of the mechanism variable, under each treatment arm, will be employed. Reference: Flores C., and A. Flores-Lagunes (2009) “Nonparametric Partial Identification of Causal Net and Mechanism Average Treatment Effects”, IZA Discussion paper, 2009. Dr. German Blanco is assistant professor in the Department of Economics at the Illinois State University and a visiting scholar at LISER. His main research interests are in Labor Economics with an emphasis on applied econometrics for the evaluation of public programmes. His published articles on programme evaluation appear in the Journal of Human Resources, Journal of Productivity Analysis, and the American Economic Review Papers and Proceedings.