ECON 627Econometrics of High Dimensional Data

Estimation of econometric models using automatic model selection and dimension reduction techniques, with emphasis on estimation of the causal effects of economic policies. Methods may include Lasso and related estimators, principal components, regression trees or forests, and double/debiased machine learning. These methods may be used with conditional ignorability assumptions in linear models, directed acyclic graphs, or structural equation models to estimate heterogeneous treatment effects with economic data.