A New Approximate GLS Estimator for the Regression Model with MA(1) Disturbances.

This paper introduces a new approximate GLS estimator for the regression model with MA(1) disturbances, which outperforms both the P. Balestra (1980) and the C. Y. Park and R. G. Heikes (1983) approximations and which is almost as efficient as the exact GLS estimator. The method used for comparison is the relative efficiencies of the various estimators vis-a-vis the exact GLS estimator in the context of the "intercept only" regression model. Copyright 1987 by Blackwell Publishing Ltd and the Board of Trustees of the Bulletin of Economic Research