Seasonal adjustment by optimally smoothed tim-variable parameter estimation

Abstract

Recursive methods of time-series analysis developed in recent years provide a natural approach to the estimation of models with time-variable parameters. This paper describes how fixed-interval recursive smoothing can be applied to the seasonal adjustment of time-series. A criterion is established to determine the degree of variability appropriate to the seasonal component, which permits the development of an automatic method for selection of the estimator parameters. The proposed method is easy to apply in such practical situations and an example is provided which demonstrates its successful use in the seasonal adjustment of economic time-series.