Modern, high dimensional data has renewed investigation on instrumental variables (IV) analysis, primarily focusing on estimation of effects of endogenous variables and putting little attention towards specification tests. This paper studies in high dimensions the Durbin-Wu-Hausman (DWH) test, a popular specification test for endogeneity in IV regression. We show, surprisingly, that the DWH test maintains its size in high dimensions, but at an expense of power. We propose a new test that remedies this issue and has better power than the DWH test. Simulation studies reveal that our test achieves near-oracle performance to detect endogeneity.
JEL classification: C12; C36Abstract This note summarizes the supplementary materials to the paper "Testing Endogeneity with High Dimensional Covariates". In Section 1, we present extended simulation studies for the low dimensional setting. In Section 2, we show that the DWH test fails in the presence of Invalid IVs. In Section 3, we discuss both method and theory for endogeneity test in high dimension with invalid IVs. In Section 4, we present technical proofs for Theorems 1, 2, 3, 4 and 5 and the proofs of technical lemmas.
Simulation for Low DimensionsFor the low dimensional case, we generate data from models the same models as the high dimensional simulations, except we have p z = 9 instruments, p x = 5 covariates, and n = 1000 samples. The parameters of the models are: β = 1, φ = (0.6, 0.7, 0.8, 0.9, 1.0) ∈ R 5 and ψ = (1.1, 1.2, 1.3, 1.4, 1.5) ∈ R 5 . We see that the three comparators, the regular DWH