Data Type for Sample Inputs in autoarma

For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.

Moderators: EViews Gareth, EViews Moderator

MarMal
Posts: 2
Joined: Thu Oct 19, 2023 6:04 pm

Data Type for Sample Inputs in autoarma

Postby MarMal » Sun Oct 22, 2023 2:52 pm

Hello,

I would like to ask a question regarding the variable type needed for sample inputs in autoarma.

My question is as follows. I am working on a forecast series which requires the ARIMA'ing of a number of indicator series. These indicator series are updated on a monthly basis. Currently, I use the following specification (which works):

series_name.autoarma(smpl="1924M01 2023M09", forclen="2023M11") series_name_f c ' Just an example

However, as there are quite a few series involved, it would be a bit of a hassle for me to modify the endpoint in the autoarma sample and the forecast length manually each time there is a data update, and it may also lead to mistakes in case there have been updates I am not aware of (I cannot run this in a loop because they have different starting dates / data release dates).

As a result, I would appreciate it if anybody could tell me if there is a way to input the autoarma sample (the smpl= portion) with the likes of @ifirst and @ilast, and to input 'forclen' as a scalar (as it is the same across all series). This would mean that the autoarma automatically updates as the data itself updates. When I tried @ifirst, however, I got the following error:

"Invalid specification for forecast length or endpoint in
"series_name.AUTOARMA(SMPL=@IFIRST(series_name)
@ILAST(series_name), FORCLEN="2023M11") series_name_F C"
on line xxx."

Ditto when I tried to input 'forclen' as a pre-defined scalar.

Thank you in advance for the help! And one other small clarification, please - should there accidentally be any NAs in-between the first and last dates with data in a series, would @ilast stop at the first NA or look for the last point with data, regardless of how many NAs come before that? Thank you!

EViews Gareth
Fe ddaethom, fe welon, fe amcangyfrifon
Posts: 13323
Joined: Tue Sep 16, 2008 5:38 pm

Re: Data Type for Sample Inputs in autoarma

Postby EViews Gareth » Mon Oct 23, 2023 9:32 am

Store the sample end points into string variables, and then use the string variables in the autoarma command.

Code: Select all

%first = @ilast(series_name)
%last = @ilast(series_name)
%smpl = %first + " " + %last
series_name.autoarma(smpl=%smpl)
Follow us on Twitter @IHSEViews

MarMal
Posts: 2
Joined: Thu Oct 19, 2023 6:04 pm

Re: Data Type for Sample Inputs in autoarma

Postby MarMal » Mon Oct 23, 2023 4:03 pm

Dear Gareth,

Thank you very much for the quick response. I believe @ifirst and @ilast generate scalars and not strings, but guided by your suggestion I ended up using the following, and it works:

smpl relevant_sample
%forclen = @right(@pagesmpl,7) ' As relevant_sample is in 'YYYY[M]MM' format
smpl @all

%first = series_name.@first
%last = series_name.@last
%smpl = %first + " " + %last
series_name.autoarma(smpl=%smpl, forclen=%forclen) series_name_f c

Now I can actually run this within a loop. I also verified manually that @last picks up the date with the latest datapoint in the series regardless of whether there are any NAs in-between. Thank you very much!


Return to “Estimation”

Who is online

Users browsing this forum: No registered users and 28 guests