NEW CONTROL FUNCTION APPROACHES IN THRESHOLD REGRESSION WITH ENDOGENEITY

This paper studies control function (CF) approaches in endogenous threshold regression where the variable is allowed to be endogenous. We first use a simple example show that structural (STR) estimator of point Kourtellos, Stengos and Tan (2016, Econometric Theory 32, 827–860) inconsistent unless endogeneity level low compared effect. correct CF STR generate our using method extends two-stage least squares procedure Caner Hansen (2004, 20, 813–843). develop second which can treated as an extension classical approach linear regression. Both these embody effect information conditional variance beyond mean. Given estimates, we propose new estimates for slope parameters. The by-product approach, type employs generalized moment (GMM) procedures based on two sets conditions. Simulation studies, conjunction with limit theory, confidence interval together associated GMM parameter dominate other methods. further apply estimation methodology empirical application from international trade illustrate its usefulness practice.