On the Asymptotic Relative Efficiency of Gaussian and Least Squares Estimators for Vector ARMA Models

This paper is concerned with the asymptotic relative efficiency of the Gaussian and least squares estimators when employed to estimate the parameters of vector ARMA models presented in echelon canonical form. The relative efficiency is assessed via the variance-covariance matrices of the limiting normal distributions of the two estimators. Situations under which substantial loss or gain in efficiency could be exprected are discussed and illustrated with some numerical examples.