Page 1 of 2

Panel SVAR

Posted: Thu Jul 26, 2018 5:00 pm
by dakila
This thread is about the psvar add-in that estimates Panel SVAR of Pedroni (2013).

Re: Panel SVAR

Posted: Fri Aug 03, 2018 2:00 am
by dakila
The psvar add-in is updated. It now includes variance decomposition analysis.

Re: Panel SVAR

Posted: Fri Aug 03, 2018 8:56 am
by nachito_mun
Many thanks for this add-in, Dakila! Is there any tutorial or help file to understand how to run it and also the output it produces?

Re: Panel SVAR

Posted: Tue Aug 07, 2018 11:04 pm
by jiangziyaok
dakila wrote:The psvar add-in is updated. It now includes variance decomposition analysis.

Dear dakila, when we use the variance decomposition analysis,it shows a error messages: matrix sent a function which operates on a vector.
can you give us some suggests?

Re: Panel SVAR

Posted: Wed Aug 08, 2018 12:08 am
by dakila
Could you send the command or description of the date file or data file?

Re: Panel SVAR

Posted: Wed Aug 08, 2018 11:02 am
by pmchileshe
Hi Dakila,
I am also interested in the information on how the Panel SVAR addin can be used. I have been checking all manuals and there is nothing.
Please help

Re: Panel SVAR

Posted: Wed Aug 08, 2018 4:17 pm
by dakila
The add-in document (psvar.pdf) and example(psvar_ex1.prg) are located in folder C:\Users\...\Documents\EViews Addins\Psvar.

Re: Panel SVAR

Posted: Mon Aug 27, 2018 2:54 am
by chirs003
I encountered the same problem as jiangziyaok. When I added the variance decomposition option, it showed a error messages: matrix sent a function which operates on a vector. My data is in the attachment section.
In addition, I don't understand what e1, e2...en in the impulse response diagram is?
Can you give us some suggests?

Re: Panel SVAR

Posted: Mon Aug 27, 2018 5:43 am
by dakila
What is the version of eviews you use? It works for my eviews.
For example, the following command is working:
psvar(ident=1, horinzon=10, vd=1) 5 @ dlntfp dlnhc

You do not need to take the first difference. The add-in automatically will take the first difference.

Re: Panel SVAR

Posted: Mon Aug 27, 2018 11:16 pm
by chirs003
dakila wrote:What is the version of eviews you use? It works for my eviews.
For example, the following command is working:
psvar(ident=1, horinzon=10, vd=1) 5 @ dlntfp dlnhc

You do not need to take the first difference. The add-in automatically will take the first difference.

I use the eviews10. I also use your code "psvar(ident=1, horinzon=10, vd=1) 5 @ dlntfp dlnhc". However, i still shows: matrix sent a function which operates on a vector.
hc.jpg
hc.jpg (71.18 KiB) Viewed 11055 times

Maybe it's time to change my eviews version. QAQ.

Re: Panel SVAR

Posted: Mon Aug 27, 2018 11:23 pm
by chirs003
When i click OK. It shows
HC2.png
HC2.png (59.24 KiB) Viewed 11055 times

Perhaps the result of the variance decomposition is in the above picture, but I don't know that it is the end result.
Also, can we do Panel Granger causality test by psvar´╝č

Re: Panel SVAR

Posted: Tue Aug 28, 2018 12:25 am
by dakila
Try to update the eviews 10 and try update the psvar add-in to the version 2.0. It is not possible to do Granger causality test using the psvar add-in.

Re: Panel SVAR

Posted: Fri Feb 15, 2019 2:50 pm
by Niqu
dakila wrote:Try to update the eviews 10 and try update the psvar add-in to the version 2.0. It is not possible to do Granger causality test using the psvar add-in.

@dakila, I'm trying ro run the program but I get the following error: "Error 29 in encrypted program". I'm using Eviews 10+, university edition, and it's up to date. Do you have any idea how to get around with this error? Thanks!
LE: it seems that the error pop ups because the panel is unbalanced. After I make it balanced loosing too many observations, the error dissapears. My question still stands: do you have any idea how to get around with this error?

Re: Panel SVAR

Posted: Sun Feb 17, 2019 7:16 am
by dakila
I don't know. maybe I can figure out if you send me the data.

Re: Panel SVAR

Posted: Sun Feb 17, 2019 11:35 am
by Niqu
dakila wrote:I don't know. maybe I can figure out if you send me the data.

Hi, Dakila,

I do not know if you received the database in PM, and to be sure, I attach it here too.
Thanks!