Weaknesses of the Baxter-King Filter: Is a Pattern-Based Filter an Alternative? / Schwächen des Baxter-King Filters: Ist ein musterbasierter Filter eine Alternative?

Summary The Baxter-King filter shows some weaknesses, particularly with regard to monthly time series. This procedure involves not only a loss of data for the border areas of time series, but suppresses inadequately high frequency components and shows as a low-pass filter only the performance of ordinary moving averages. Another important finding is that the maximal lag Κ not only determines the degree of approximation to an ideal filter, as Baxter and King argue, but also the ability to extract a slow moving component. A pattern-based filter, whose weights are generated by a trigonometric function as well, does not possess such weaknesses, and is thus a real alternative to the Baxter-King filter.