Endogeneity in stochastic frontier models: Copula approach without external instruments

- We propose a Copula approach for estimating endogenous stochastic frontier models.
- We discuss the model identification strategy.
- Maximum likelihood estimation procedure is proposed.
- Monte Carlo results show that the proposed estimator performs well in finite sample.

This papers considers an alternative estimation procedures for estimating stochastic frontier models with endogenous regressors when no external instruments are available. The approach we propose is based on copula function to directly model the correlation between the endogenous regressors and the composed errors. Estimation of model parameters is done using maximum likelihood. Monte Carlo simulations are used to assess and compare the finite sample performances of the proposed estimation procedures.