We investigate frequency domain properties, revealed by the squared gain and phase delay functions, of short and moderate-length linear seasonal adjustment filters of the ARIMAmodel- based signal extraction method of SEATS. X-11/12-ARIMA filters are also considered to a limited extent. The focus is on the one-sided (concurrent) and symmetric (central) filters associated with the Box and Jenkins’s airline model for monthly time series of lengths 49 and 109. A Digital Signal Processing perspective on filter properties, favoring interpretability of the filter output, is presented. We show that important features of the finite filters actually used are often not visible in the diagnostics of the infinite filters, such as the Wiener-Kolmogorov symmetric filter gain functions provided by SEATS. For comparing competing adjustments, properties of concurrent filters, especially their phase delays, can be more important than properties of symmetric filters. Our phase delay results illustrate that adjusters who favor smoother seasonal adjustments, or who favor trends for their greater smoothness, must usually reckon with greater delay of turning point and business cycle information for the most recent months. Trend filters are considered briefly. New analytical results are obtained for transfer functions and phase delays.
Gain function; phase delay function; seasonal adjustment; trend estimation;
ARIMA-model-based signal extraction; X-11-ARIMA; X-12-ARIMA.
[PDF] or 
denotes a file in Adobe’s Portable Document Format.
To view the file, you will need the
Adobe® Acrobat® Reader
available free from Adobe.
This symbol
indicates a link to a non-government web site. Our linking to these sites does not
constitute an endorsement of any products, services or the information found on them.
Once you link to another site you are subject to the policies of the new site.
