Page 1 of 2

multiple dependent variables

Posted: Mon Apr 25, 2011 4:49 pm
by gregjackson05
Hi
Eviews 6; I'm trying to run multiple regressions at once. I have about 1000 mutual fund companies and about 3 years of monthly returns.
I have tried OPEN - as SYSTEM but this only seems to allow me to run 30 - 50 regressions.
Is there a way to run all at once??

Cheers Greg

Re: multiple dependent variables

Posted: Mon Apr 25, 2011 4:57 pm
by EViews Gareth
Do you need to estimate them simultaneously from a theoretical point of view, or from a easy-of-use point of view? If the latter, you're probably best off writing a simple program that will estimate them all.

Re: multiple dependent variables

Posted: Mon Apr 25, 2011 5:30 pm
by gregjackson05
Hi Gareth,
Yeah I just want the quickest option, for both ends (sorting output and results).
Simple program sounds good, but I would not know where to start??

Greg

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 11:18 am
by gregjackson05
Hi any help on top of the intro to programming would be appreciated.
I'm a little lost with the examples that use 'nrnd'.
My example is no doubt typical - list of companies and returns; regressed on a couple of constant independent variables.
how do I create my dependent variable???

Thanks Greg

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 11:20 am
by EViews Gareth
I'm not sure I understand your question. Surely you already have dependent variables and do not need to create them at all?

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 11:29 am
by gregjackson05
yes I have a list of companies and their returns that are my dependents.
I'm struggling for the terminology to explain.....
lots of dependent variables - 3 constant independents.

Greg

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 11:38 am
by EViews Gareth
So just write a loop that loops through your dependent variables, regressing each one. Something like:

Code: Select all

for %dep dep1 dep2 dep3 dep4 equation eq{%dep}.ls {%dep} c x1 x2 x3 next
Where dep1, dep2, dep3 and dep4 are your dependent variables, and x1 x2 x3 are your independent.

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 12:10 pm
by gregjackson05
Thanks, can I just cut and paste my company names into %dep ??

Greg

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 12:33 pm
by EViews Gareth
Sure.

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 2:14 pm
by gregjackson05
I've tried to copy your instructions but getting "flow of control statement..."
I appreciate any help.

Greg

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 2:26 pm
by EViews Gareth
EViews programming code needs to be written in EViews programs. Click on File->New Program.

Perhaps you should read the programming chapter of the EViews Command and Programming Reference.

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 2:58 pm
by gregjackson05
nevermind got it!

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 2:59 pm
by gregjackson05
sorry didn't see your last post, yeah that was my mistake

Thanks Greg

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 4:00 pm
by gregjackson05
Hi, just when I thought I was there now I'm faced with maximum line length error. What is max. number of dependent variables allowed? I have tried reducing my dependent name size to shorten command line??

Cheers Greg

Re: multiple dependent variables

Posted: Tue Apr 26, 2011 4:08 pm
by EViews Gareth
Run it through more than one loop then.

Code: Select all

for %dep dep1 dep2 dep3 dep4 equation eq{%dep}.ls {%dep} c x1 x2 x3 next for %dep dep100 dep101 dep102 dep103 equation eq{%dep}.ls {%dep} c x1 x2 x3 next