Files
Abstract
This paper discusses the maximum likelihood estimation of the parameters of a subsystem of nonlinear equations under partial disregard of structural information. It provides a generalization of Godfrey and Wickens' (1982) limited information estimator to subsystems of nonlinear equations and Amemiya's (1975) nonlinear limited information estimator to subsystems containing more than one equation. Some attention is devoted to the asymptotic properties of the estimator and different estimators for the asymptotic covariance matrix are discussed. The estimation method is illustrated by means of an example dealing with the estimation of an aggregate two-level CES production function for The Netherlands' enterprises sector.