CFE 2019: Start Registration
View Submission - CMStatistics
Title: Identifying marginal treatment effects in the presence of sample selection Authors:  Otavio Bartalotti - Iowa State University (United States) [presenting]
Desire Kedagni - Iowa State University (United States)
Abstract: Two identification results are developed for the marginal treatment effect (MTE) when there is sample selection. We show that the MTE is partially identified for individuals who are always selected regardless of treatment, and we derive sharp bounds on this parameter under various assumptions. The first identification result combines the standard MTE assumptions with monotonicity of the sample selection variable with respect to the treatment, while the second uses an additional (possibly invalid) instrument. Both results rely on a mixture reformulation of the problem. In the first approach, the mixture weights are identified. We therefore extend a previous trimming procedure to the MTE context. The second identification result relies on the mixture weights varying with the additional instrument, while the mixture component distributions do not.