Two-Stage Estimation of Unobserved Effects Panel Stochastic Frontier Models with Endogenous Regressors
11 Pages Posted: 14 Mar 2019
Date Written: January 30, 2019
This paper proposes an alternative estimation procedure for estimating the unobserved effects panel stochastic frontier models with endogenous regressors. Specifically, two-stage estimation method is used, where in the first stage, the frontier parameters are estimated based on GMM procedure, which is robust to the distributional assumptions of the composed error; whilst in the second stage, the remaining parameters are estimated by using MLE technique. Monte Carlo simulations indicate that the proposed estimation approach behaves well in finite samples.
JEL Classification: C13, C23, C36
Suggested Citation: Suggested Citation