Sign Restricted VAR

[Pmat]

No. You cannot interpret like that. First define the shock by the sign restriction, then interpret impulse response function of Y and X as the effect of that shock.

That makes perfect sense. Thanks for the clarification, it is all clear now.

I am sorry if I’m also being silly, but I have the same question and I did not understand the answer. For example, in a 3 variable VAR if you set +3 as a constraint, what exactly are you constraining? I think you are constraining the RESPONSE of the third variable to be positive. But the response to what shock?? If, let’s say, I want to constrain the response of the THIRD variable to a shock on the SECOND variable to be positive, how do I use the addin to perform this? Again, sorry if this is a silly question, but I’ve read the instructions and it is not clear to me. Thank you very much.

[gvera]

How do I write a zero restriction in the command line of srvar? Thank you.

[dakila]

For example:

Code: Select all

srvar(method=0, zvar=gdpc1) 12 rest @ gdpc1 gdpdef cprindex fedfunds bognonbr totresns

[Est123]

Hello!

Is it possible to use the srvar add-in with panel data?

[dakila]

No

[atbarbar]

Unfortunately the documentation just does not help to understand how to place constraints on multiple variables. A concrete example. I have a simple 2 variables VAR with two endogenous variables, y1 and y2 with respective shocks e1 and e2. I want the impulse response of y1 to a shock from e1 to be positive. I want the impulse response of y1 to a shock from e2 to be positive. I want the response of y2 to a shock from e1 to be negative and the response of y2 to a shock from e2 to be positive. Question:
1) Can I do it with srvar?
2) If I can, show me the vector necessary to implement that
3) Irrespective of whether I can do it or not with srvar, more in general how can I get the historical decomposition of a sign restricted var estimated with srvar? The output from srvar stores a VAR object but that is the simple unconstrained reduced form so the historical decomposition with that object is not what I want.

Thanks.

[dakila]

You can't identify multiple shocks for the srvar add-in.

[atbarbar]

Can you identify multiple shocks with the ARW addin?

[dakila]

No.

[atbarbar]

Ok, having cleared the fact that there is no way in Eviews to impose sign restrictions on the impulse responses of multiple shocks, you have not answered yet my last question above, question 3: suppose that I just want to impose sign restrictions on the impulse response of one shock. That can be done in Eviews. How do I get the historical decomposition from such structural identification?

[Thor]

Hi. I'm trying to use the Add-in in Eviews 14 but the code doesnt seem to work: the message I get is "Error 127 in encrypted program". Cant see what Im doing wrong (and I have used the code on an earlier version of Eviews).

[dakila]

Can you run an example file provided by the add-in (srvar_ex1)?