Re: [Gretl-users] X-12-Arima

On Mon, 27 Sep 2010, John C Frain wrote: > I have attached a pdf file which, I hope, clarifies my proposals for > the gretl x12arima interface. Thanks for the suggestions, John. > Basically there are three proposals > > 1. An amendment to the graphics output of the x12arima interface > in gretl to improve these graphs when there are missing values. This one is based on a misunderstanding. Under no circumstances should you enter -99999 in a gretl data series when you mean "not available": use the gretl missing value string, "NA", or, when editing a series via the GUI "spreadsheet", leave the value blank. Gretl will take care of passing the correct "missing" code to X-12-ARIMA, and the graphs will be OK. > 2. A greater emphasis on the .err file to ensure that the user > is aquainted with possible problems in the seasonal adjustment > process. I take your point on this, but I don't think it is a good idea to pop up a window showing the x12a ".err" file in all cases: this file is always created by x12a, but in most cases it contains nothing of any interest. If we can figure out a good test for the cases where it actually contains a warning, then I agree we should show its content in those cases. > 3. The inclusion of a trading day option in the gretl interface > to x12arima. OK, I've now figured out what this means. The "trading day effect" is the effect whereby a given month (or quarter) may contain more or fewer "trading days" (i.e. weekdays, as opposed to weekends) from one year to another, depending on how weekends fall in relation to the calendar. This is obviously of some potential relevance for seasonal adjustment, particularly for monthly data (it's likely to wash out for quarterly data), where the series in question are flows that typically cumulate on weekdays only. I'll see about adding an option for that. Allin _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users