SOME ESTIMATION METHODS FOR FIXED AND RANDOM PANEL DATA MODELS WITH SERIALLY CORRELATED ERRORS WITH APPLICATION

Abstract

This paper includes the study of panel data models with fixed and random parameters when the errors are serially correlated of the first order, the parameters of these models were estimated by two methods, the feasible lest square method (FGLS) and the mean group method (MG), efficiency comparison of these estimators were made when the regression parameters were fixed and random. Real data were applied to evaluate both the (FGLS) method performance and the (MG) method on the two models, and the (MAPE) scale was used to compare the efficiency of the estimators. The results showed that for the panel data model with fixed parameter, the (FGLS) method is better in in estimating the parameters then the (MG) method. As for the panel data model with random parameters, the results showed that the (FGLS) method in the case of the (Swamy) model is better in estimating than the (MG) method.