Moment-based estimation of nonlinear models

Orientador(a): Marcelo Medeiros

The aim of this paper is to propose a nonlinear estimator and compare its performance with the the two-stage nonlinear least-squares (NL2SLS) estimator and the nonlinear limited information maximum likelihood (LIML) estimator. In this paper we use simulated data to obtain estimations of the parameters for logistic and exponential functions. Our results show that usually LIML estimator performs better than the created estimator and the NL2SLS estimator for parameters of these functions. Results for a application with real data show that proposed estimators also provide estimations in line with economic theory.