Full Citation
Title: Estimating Heterogeneous Treatment Effects with Instrumental Forests
Citation Type: Dissertation/Thesis
Publication Year: 2020
ISBN:
ISSN:
DOI:
NSFID:
PMCID:
PMID:
Abstract: It is of importance to understand the long-run effect of immigration for efficient policy making, especially since the recent rise in immigration in the United States. The present study investigates the effect of historical immigration during the Age of Mass Migration on the support of the Republican party and the assimilation of recent foreign borns on the labor market. To identify the causal effect, we construct a "leave-out" version of the shift-share instrument and implement a two-stage least squares regression analysis. Then, we allow for a more flexible specification through instrumental forests. The forests show that the average conditional local average treatment effect of historical immigration is insignificant when nonlinearities in the data are accounted for. Potential effect modifiers are identified in a data-driven fashion. The effect of historical immigration is found to be significant for some subgroups in the sample. The results suggest a persistence of the effect of historical immigration through a two-sided assimilation process as well as through the establishment of migrant communities. Although immigrants appear to assimilate well within those communities, our findings might suggest an increased segmentation between natives and foreign borns. This has important implications for policy making.
User Submitted?: No
Authors: Hof, Nadja van't
Institution: Erasmus University Rotterdam
Department: Erasmus School of Economics
Advisor:
Degree:
Publisher Location: Rotterdam
Pages: 1-70
Data Collections: IPUMS NHGIS
Topics: Labor Force and Occupational Structure, Migration and Immigration
Countries: