Seasonal heteroskedasticity exists in a number of monthly time series from major statistical agencies. Accounting for such variation in calendar month effects can be important in estimating seasonal and trend movements. In the context of seasonal adjustment, the standard procedure uses nonparametric (X-11) filters of different lengths in the signal extraction routine of X-12-ARIMA. This serves as a simple, pragmatic procedure that is, however, limited in its ability to adapt to different datasets. In this paper I extend the model-based methodology introduced recently by Proietti (2004) and Bell (2004). I discuss different forms of the seasonal specific model, showing examples of estimation and analysis of trend and seasonal components. A statistical test for seasonal heteroskedasticity is presented and applied to a number of Census Bureau series on housing starts and building permits. It is shown how seasonal noise can be separated from nonsystematic noise and included in the seasonal adjustment of a time series.
seasonal heteroskedasticity, time series, trends, unobserved components.
[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.
