Mean error of prediction for a method of empirical growth extrapolation

The objective of this paper is to formulate a standard set of stochastic assumptions for a prediction method which consists in a linear extrapolation of the mean empirical growth. The author shows how to derive formulas for the mean error of prediction (the ex ante prediction error). These formulas are then compared to the prediction errors of the following methods: the status quo method, the mean extrapolation method and the extrapolation of the linear trend function estimated by the least-squares method. This paper shows that the extrapolation of the mean empirical growth is more efficient than the status quo method and under some assumptions (that are defined in this article) is more efficient than the mean extrapolation method or the extrapolation of the linear trend function estimated by the least-squares method.