Page 4 of 9

Re: Threshold Structural VAR

Posted: Wed Oct 19, 2016 6:20 pm
by dakila
Can you run the example program? Could you post the workfile?

Re: Threshold Structural VAR

Posted: Sun Nov 13, 2016 1:03 am
by nasa
Hi,
How can I generate cumulative generalized impulse response function?
Thanks

Re: Threshold Structural VAR

Posted: Mon Nov 14, 2016 9:30 am
by dakila
Now you can do it. Check the add-in. I hope it will posted on the web soon.

Re: Threshold Structural VAR

Posted: Thu Nov 17, 2016 6:40 am
by doreenruba
dakila wrote:Can you run the example program? Could you post the workfile?


Attached. rgdp is the threshold variable. want to run a threshold SVAR.

Re: Threshold Structural VAR

Posted: Thu Nov 17, 2016 2:54 pm
by dakila
Doreenruba, you need to adjust sample size manually. Because the first observation (2000q1) is missing. So change sample 2000q1 2015q4 to 2000q2 2015q4.

Re: Threshold Structural VAR

Posted: Tue Nov 22, 2016 5:17 am
by doreenruba
dakila wrote:Doreenruba, you need to adjust sample size manually. Because the first observation (2000q1) is missing. So change sample 2000q1 2015q4 to 2000q2 2015q4.


Thanks. Help so far appreciated. I have run the command thsvar(girf=1, sample="2001q1 2015q4") 2 rgdp @ rgdp pinv agr reer tot using sample 2001q1 2015q4. I got the attached results. But i don't see IRs in graphs. Did I do the right thing? It's the first time I am using thsvar. What important things should i look out for?

Re: Threshold Structural VAR

Posted: Tue Nov 22, 2016 8:02 pm
by dakila
I think you forget to include the response variable for the impulse response analysis.
After the threshold variable (rgdp) type the response variable. For example: thsvar(girf=1, sample="2001q1 2015q4") 2 rgdp tot @ rgdp pinv agr reer tot
If you are not sure about the command then use the dialogue from the Add-ins menu. That would be easy to implement ThSVAR.

Re: Threshold Structural VAR

Posted: Wed Nov 23, 2016 7:55 am
by nasa
Hello,

What is the use of cumulative generalized IRFS and on what respects do they differ from the GIRFs?

Could you give me a bit explanation about this,please?

Thanks,

Re: Threshold Structural VAR

Posted: Wed Nov 23, 2016 3:30 pm
by dakila
The former accumulate shocks and the latter is not. If VAR is estimated with first order differrenced variables, cumulative GIRF is approprate. Because it would be easy to explain the IRF graph.
if you estimated VAR model then do not do cumulative GIRF. It does not make sense.

Re: Threshold Structural VAR

Posted: Fri Nov 25, 2016 9:24 am
by nasa
Hi,

I read the Balke's paper and noticed that he run the VAR in first difference and then generates the generalized impulse response functions (not CUMULATIVE generalized IRFs), If I am not wrong.

How do you see this thing,please?

Thanks

Re: Threshold Structural VAR

Posted: Mon Nov 28, 2016 8:17 am
by nasa
Hi,

What does C1 represents? is zat positive one standard deviation or positive 2 standard deviation? I hope it is not negative shock C1.

Thanks.

Re: Threshold Structural VAR

Posted: Mon Nov 28, 2016 3:11 pm
by dakila
I don't know. Maybe you need to update your Eviews.

Re: Threshold Structural VAR

Posted: Mon Jan 09, 2017 11:34 am
by jwojnilower
Hi,

Could you verify if the GIRF function in the ThSVAR add-in displays cumulative generalized impulse response functions or not? If not, is there a way to export the underlying values to Excel?

Thanks

Re: Threshold Structural VAR

Posted: Mon Jan 09, 2017 5:53 pm
by dakila
Did you update the thsvar add-in? There is option to implement the cumulative GIRF.

Re: Threshold Structural VAR

Posted: Tue Jan 10, 2017 8:28 am
by jwojnilower
I had not yet updated the Add-in, so thank you for the suggestion.

What syntax should we use in the command line to estimate cumulative generalized impulse functions? Previously it was girf=1. Is it not cgirf=1? Or something else?

Thanks again!