Seasonal Adjustment

[bdamasio]

Dear all,

Do you know how can I remove seasonality from a series imposing the restriction that yearly totals of the seasonally adjusted series to equal those of the original series?
I checked in X-13ARIMA-SEATS Reference Manual and it is possible (section 7.6 FORCE), but I am having problems in implementing it directly on EViews.


Thanks in advance.
Bruno.

[EViews Gareth]

You can probably do it via command using the spec= option to provide your own spec file.

[bdamasio]

Thank you Gareth.

I created a spec file with the contents:

force{ lambda = 0.0 mode = ratio rho = 0.85 round = no start = oct target = calendaradj type = regress usefcst = no print = (none saa) save = saa }

I gave the instruction "gcn.x13(sfile=‪C:\Users\"..."\example.spc) gcnsa" but it doesn't work.
The error message is "?C:\USERS\"..."
\EXAMPLE.SPC not found on disk.

Do you have any suggestion.

Thanks.

[EViews Gareth]

Did you give it the correct path?

[bdamasio]

Yes, I gave.
It works, I mean without any error message, if I do it manually.
I performed a seasonal adjustment with menus, then I created a text file with the name "dois".
I modified it by adding the "force" option, please find it below.
Then I gave the command: "trn.x13(spec=dois)" but the option "force" doesn't work.
The annual sums of the original series doesn't match with the deseasonalized ones.
Moreover, if I omit the force instruction it doesn't make any difference.
Can you help me please?
Do you have any suggestion to desasonalize a series in such a way that the annual sums of the desasonalized series equals the annual sums of the original series?


My text file:
series{
title = "trn"
start = 1999.1
period = 4
modelspan = (1999.01, 2015.01)
name = "TRN"
file = "C:\Users\bdamasio\AppData\ev_temp\EVX13TMP.DAT"
decimals = 1
}

x11{
save = (d11)
}

force {
mode=ratio
save= e6a
target=original
type=denton
}

[EViews Gareth]

Your spec file should not have a series statement.

[bdamasio]

I erased the series statement but the results are the same: the sums doesn't match and the desasonalized series doesn't change if I drop also the force statement.
Any suggestion Gareth?

My test file:

x11{
save = (d11)
}

force {
mode=ratio
save= e6a
target=original
type=denton
}


if I consider this one

"x11{
save = (d11)
}"

or this one

"x11{
save = (d11)
}

force {
}"

The deseasonalized series doesn't change.

[EViews Gareth]

Have the series open when issuing the command. Then you can see the X-13 output in the series. See if the X-13 output says anything about FORCE.

[bdamasio]

Please find attached the output.
It just reproduce the spec file, I think it doesn't explicitly refers the FORCE option.

[EViews Gareth]

Well that shows that X-13 is receiving your FORCE option ok. No idea what it is doing with it.

[bdamasio]

So do you have any tip or trick, for instance an alternative way or procedure, to help me solve the aforementioned issue?
It is possible to put it into practice throughout EViews 9?
Thank you.

[xprimexinverse]

Hi Bruno (bdamasio).

I've attached a program (x13force.prg) containing a subroutine that should do the trick.

You might need to make a few changes to the specification file (which I create as a text object). Depending on how you have EViews set up on your computer, you may not have to make any other changes to the program; say, with regard to locating the .saa file.

It ain't pretty, but it seems to get the job done.

Cheers,
Graeme

[bdamasio]

Dear xprimexinverse,
Thank you very much!
It did works!
Bruno.