SABL: A Resistant Seasonal Adjustment Procedure With Graphical Methods for Interpretation and Diagnosis

Abstract

SABL (Seasonal Adjustment-Bell Laboratories) is a new seasonal adjustment procedure for monthly time series with a yearly penodicity. From a class of transformations of the data, SABL selects the one that minimizes the lack of stability of the seasonal component. It then decomposes the transformed time series into additive trend, seasonal, and irregular components using resistantsmoothing techniques. The smoothers consist of an initial resistant smooth using moving medians, followed by one or more linear smooths. Graphical displays are provided as diagnostic tools in assessing the nature and adequacy of the decomposition into trend, seasonal, and irregular components.