2020/11/13 by Francis J. DiTraglia, Camilo García-Jimeno, DiTraglia, Francis J. +5 · 1 citation
Economics, Econometrics and Finance · Mathematics · Social Sciences · #Advanced Causal Inference Techniques #Econometrics (econ.EM) #FOS: Computer and information sciences #FOS: Economics and business #Labor market dynamics and wage inequality #Methodology (stat.ME) #Migration and Labor Dynamics
paper · pdf · doi:10.48550/arxiv.2011.07051
openalex publication_date 2020/11/13 · openalex created_date 2023/01/08 · openalex updated_date 2026/07/28
This paper shows how to use a randomized saturation experimental design to\nidentify and estimate causal effects in the presence of spillovers--one\nperson's treatment may affect another's outcome--and one-sided\nnon-compliance--subjects can only be offered treatment, not compelled to take\nit up. Two distinct causal effects are of interest in this setting: direct\neffects quantify how a person's own treatment changes her outcome, while\nindirect effects quantify how her peers' treatments change her outcome. We\nconsider the case in which spillovers occur within known groups, and take-up\ndecisions are invariant to peers' realized offers. In this setting we point\nidentify the effects of treatment-on-the-treated, both direct and indirect, in\na flexible random coefficients model that allows for heterogeneous treatment\neffects and endogenous selection into treatment. We go on to propose a feasible\nestimator that is consistent and asymptotically normal as the number and size\nof groups increases. We apply our estimator to data from a large-scale job\nplacement services experiment, and find negative indirect treatment effects on\nthe likelihood of employment for those willing to take up the program. These\nnegative spillovers are offset by positive direct treatment effects from own\ntake-up.\n