Time varying SVAR

[asif_1210]

Hi, can anyone please help me as I'm consistently getting msg of 'illegal date vector'. I've created the date vector using the suggested command.

[dakila]

if you are using the dialog interface (menu) do not use the command for date selection vector. Just put dates on the box. For example, enter 1975q1 1981q3 1996q1 on the box (date selection vector)

[asif_1210]

Hi, is there any method by which we can we can have data series of IRFs generated through TV SVAR?

[nadybe]

Hello Davaa,

I constantly have the "size don't match matrix function" error, what does it mean and how do i fix this issue?

Best regards

[adrangi]

Hi. I'm new to add-ins. I tried to download tv_svar and open the prog in a workfile on eviews 9.5. Ended up with three files, tvsavr.prg, tvsavr.install.prg and tvsavr.ex.prg. Do I need to install before running the program? How do I install it? Any tips are appreciated. Thanks.

[EViews Gareth]

http://www.eviews.com/help/helpintro.ht ... d-ins.html

[adrangi]

http://www.eviews.com/help/helpintro.ht ... d-ins.html

Thanks very much. Best, BA

[adrangi]

Hi. I read the pdf file for the TV-SVAR model. Does the "impulse variable" mean the endogenous variable that is being shocked? Don't we normally shock all endogenous variables and check the response by others in VARS? Any help is much appreciated. Thanks.

[nadybe]

The impulse variable is the variable that will affect all the other endogenous variables including itself, you can specify more than one impulse variable and see the impulses responses fonctions of the other endogenous variables as in the other Var model. If you need tips just message me i can help you.

[adrangi]

Hi. Thanks much. Clarifies it. Best, BA

[marti]

Hi,

In April there was a question regarding saving the results as series, so that they can be exported. As you replied it would be possible with the release of version 2.0. I would like to ask if you are going to include the residuals as series as well despite the graph of the standard dev. When can we expect the new version?

Thanks,
Marti

[dakila]

I will try to include the residuals. I hope it will be released in 2 weeks.

[ali_Economist]

Hi, I am trying to look at the impact of real rates on the unemployment and cpi, but is giving error Near Singular Matrix, I can,t log the series as these have negative values as well and also these are rates. Can you please help ?

Regards

[dakila]

you have the multicollinearity problem. the correlation coefficient is almost minus one between real rates and cpi

[ali_Economist]

Yes, you are absolutely right, actually the CPI and unemployment are dependent variables and real rates are independent variables, so in simple regression it should not be the case of multicolinarity. Can you please suggest me how I can solve this issue in TVSVAR ? Thanks again for help.

Regards

Ali