Performances of estimators of linear auto-correlated error model with exponential independent variable.

Abstract

The performances of five estimators of linear models with autocorrelated disturbance terms are compared when the independent variable is exponential. The results reveal that for both small and large samples, the Ordinary Least Squares (OLS) compares favourably with the Generalized least Squares (GLS) estimators in respect of bias property. On the basis of variance and root mean square error property, OLS compares favourably with maximum likelihood (ML) and Maximum Likelihood Grid (MLGRID) estimators for small autocorrelation coefficient of the error term ρ but it appears uniformly superior to Cochrane-Orcutt (COC) and Hildreth and LU (HILU) estimators especially when ρ is large.

Journal of the Nigerian Association of Mathematical Physics Vol. 9 2005: pp. 385-388