EViews.com

Hi all,

I am working on batch mode in Eviews 6. I have 9 cross identifiers in a panel data and I am trying to make some groups. But when I run the following commands to group cross section identifiers such as:

sample smpl1 if id01="_TMC" or id01="_TMW" or id01="_TME"
sample smpl2 if id01="_AMC" or id01="_AMW" or id01="_AME"
for %condition smpl1 smpl2
smpl @first+!i-1 @first+!i+!window-2 if {%condition}

The answer I get is "Error in sample, id01 is not defined in sample smpl1.."

The commands are similar to what Gareth recommended to another user in this forum. Any help?

Thank you in advance

Arturo

I'll ask the obvious question first, do you have a series called ID01?

No, id01 it is not a series. I do have 9 cross identifiers and a set of series associated with them, but I want to group the cross identifiers and their associated series into 3 groups.
Arturo

I think you're going to have to flesh out your problem a bit more clearly. A thorough description of your data structure, and what you're trying to do.

Sure.
I have 9 cross identifiers (_tmc _tmw _tme _amc _amw _ame _cmc _cmw _cme) in a pool object call "regions" that represent geographic regions, and 10 time series (x1?, x2?,..,x10?) representing economic characteristics of the regions (income, population, education, etc). The periodicity is by quarters.

I want to do panel regressions (say x1? on x2? x3? or some other specification)for each group, where group 1 would comprise _tmc _tmw _tme; group 2 _amc _amw _ame; and group 3 _cmc _cmw _cme.

If I run the command

regions.ls x1? c x2? x3?

The program execute the regression including the 9 cross identifiers. How could I introduce commands to execute regressions for each group?

I hope this could clarify my question

Thank you
Arturo

Since you're dealing with a pool rather than a panel, the easiest way to do it is to simply define the pool to have only the cross-sections you want to include.

Thus where as your pool object currently has 9 cross-sections, simply remove 6 of them.

It might be best to actually create three different pool objects, one for each set of 3 cross-sections you want to estimate.

Thank you