Key points are not available for this paper at this time.
Abstract The main purpose of this study is to analyse the magnitude and the nature of the revisions that the time varying seasonal filters of the X‐II and X‐II‐ARIMA methods introduce in the current seasonally adjusted series The total revision is measured by the mean absolute difference of the transfer functions corresponding to the forecasting and the concurrent seasonal filters with respect to the central‘final’seasonal filter. To take into consideration the fact that the spectrum of a typical economic time series peaks at the low and seasonal frequencies, the revision measures are calculated for selected frequency intervals associated to the trend‐cycle, seasonal variations and the irregular component.
Estela Bee Dagum (Thu,) studied this question.