Online Program Home
My Program

Abstract Details

Activity Number: 411 - Nonparametric Testing
Type: Contributed
Date/Time: Tuesday, July 30, 2019 : 2:00 PM to 3:50 PM
Sponsor: Section on Nonparametric Statistics
Abstract #304620
Title: A Consistent Nonparametric Test for Endogeneity
Author(s): Seolah Kim*
Companies: University of California, Riverside
Keywords: Consistent test; Kernel estimation; Conditional moment test; Local alternatives

I construct a consistent nonparametric test for endogeneity based on conditional moment. In constructing a new test for endogeneity, I use the error terms from a reduced-form equation under the assumption of exogeneity of Z as well as the error terms from a structural equation. The advantage of this test is that its asymptotic distribution is standard normal under the null. I also propose a resampling bootstrap test and it performs better than the asymptotic test in Monte Carlo simulations. In simulations, my test statistic performed better both in estimating size and power compared to Hausman and Blundell and Horowitz (2007) test with a bounded support. In addition, I apply this test to empirical data from Autor, Dorn, and Hanson (2013). In such cases where we may have misspecified the model regarding the functional form or we add other exogenous variables, the test results between the parametric test and the nonparametric results can be contradicting to each other. The current test statistic can provide a generic approach to test endogeneity in other econometric problems in future research.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program