The purpose of this article is, firstly, to compare and classify the different signal extraction methods empirically according to their optimality. Secondly, to apply these methods to anti-cyclical economic policy. For this purpose this article is built from five sections. In Sect. 1 we propose a criterion of optimality and a calculation algorithm which will make it possible to calculate the said criterion for all the statistical series and to compare the results. Section 2 analyzes the following three points. Firstly we present the database that will be used by the previous algorithm. Secondly we make the comparisons between the series after having calculated the different standard deviations of the residuals estimated by all the competing methods. After this the algorithm determines the optimal method within the meaning of the criterion that we have defined. Thirdly we propose a principal component analysis, in order to study the different correspondences and similarities between all the methods. This approach, as we will see, will lead to the designation of groups of methods that have correlations or contradictions. This last result will allow a classification of the groups of methods according to their performance and their degree of similarity. The Sect. 3 proposes to study the usefulness of these methods for short-term economic policy in general. This section ends by underlining their necessity for the Moroccan case. Section 4 analyse and discuss the result of different tests used. Section 5 summarizes the results and important conclusions.