Page 1 of 1

mistake SVARPatterns published

Posted: Wed Jan 22, 2014 6:09 am
by ecofin
hi
there is mistake in the publication date of the add-in SVARPatterns published in 15 jan 2014 not 15 jan 2015 (see Add-in Objects).

Re: mistake SVARPatterns published

Posted: Wed Jan 22, 2014 12:58 pm
by terrya
Also I haven't been able to get the add-in to work on a model that I've estimated already in RATS. It appears to give no output. My short and long patterns were identical with those I used in RATS. Or, perhaps, I've misunderstood the instructions. Is there an example that could be used to check the user's implementation?

Re: mistake SVARPatterns published

Posted: Wed Jan 22, 2014 2:49 pm
by trubador
There are too many potential cases, so it would be more efficient to see your model and workfile.

Re: mistake SVARPatterns published

Posted: Wed Jan 22, 2014 3:08 pm
by terrya
Here is the workfile.

Re: mistake SVARPatterns published

Posted: Thu Jan 23, 2014 12:51 am
by trubador
If you run the following command:

Code: Select all

bjvar.svarpatterns(shortmat=short,longmat=long,factormat=fmat,estimate)
You'll get the following factor matrix:

Code: Select all

0.436310 0.000000 0.000000 0.000000 0.000000 0.033299 0.657013 0.000000 8.08E-17 1.07E-16 0.269449 0.221643 2.010689 0.000000 6.97E-18 0.245102 0.043551 0.077802 0.377272 -0.361487 2.118307 2.644614 -0.267502 9.575080 8.140603

Re: mistake SVARPatterns published

Posted: Thu Jan 23, 2014 1:43 am
by terrya
I did that but didn't get the factor matrix. Perhaps there's another problem, but I'll check if I can replicate your result.

Re: mistake SVARPatterns published

Posted: Thu Jan 23, 2014 1:51 am
by terrya
Hi

I reinstalled the add-in and replicated your result. There must have been something dodgy about the original installation.

Thanks for your help.

Re: mistake SVARPatterns published

Posted: Thu Jan 23, 2014 2:26 am
by trubador
Actually, it was my mistake. EViews has nice string equality functions, but I did not use them. I made the necessary changes, and I am glad to hear that it is working.