## Sign Restricted VAR

**Moderators:** EViews Gareth, EViews Moderator, EViews Esther

### Re: Sign Restricted VAR

Hi! I have two simple question. First, I'm estimating a BVAR model and I'd like to use sign restrictions too for my impulse responde functions. In this way, is it possible to do this with your add-in. Second, I don't get how to put the restrictions. Let's imagine I have a trivariate VAR (GDP, inflation and monetary policy interest rate). How can I define impact of monetary policy shock on the variables and, then, impact of a demand shock (let's imagine a shock of product). Is it possible?

### Re: Sign Restricted VAR

Did you read the instruction? It is located in C:\Users\...\Documents\EViews Addins\srvar. (srvar.pdf)

### Re: Sign Restricted VAR

Hello,

I have got a question regarding the IRF outputs. Can I somehow save the underlying data?

Because only the basic VAR and the graphs are getting saved into the workfile, I cannot obtain the underlying data of the IRFs. And exporting is only possible to a picture file.

Thanks

I have got a question regarding the IRF outputs. Can I somehow save the underlying data?

Because only the basic VAR and the graphs are getting saved into the workfile, I cannot obtain the underlying data of the IRFs. And exporting is only possible to a picture file.

Thanks

### Re: Sign Restricted VAR

I updated the srvar add-in. Please check the add-in webpage.

### Re: Sign Restricted VAR

Thank you so much for the add-in! I read the guide, but I am still confused about one thing.

How do I add this sign restriction vector?

(a11) 0 0 0

a(21) a(22) 0 0

a(31) a(32) a(33) 0

a(41) a(42) a(43) a(44)

Sorry if the question is really easy to solve! But I have been trying to understand it

How do I add this sign restriction vector?

(a11) 0 0 0

a(21) a(22) 0 0

a(31) a(32) a(33) 0

a(41) a(42) a(43) a(44)

Sorry if the question is really easy to solve! But I have been trying to understand it

### Re: Sign Restricted VAR

It is not possible to restrict the parameter or the matrix. The sign restriction vector is related the IRF.

### Re: Sign Restricted VAR

Dear dakila,

Thank you for posting this. Trying to specify a SVAR with 6 variables, I get the following error message

"6 is not a valid index for vector-series-coefficient SCALES01." Would you be so kind and help me in interpreting/solving this?

Also, is it possible to implement a combination of sign and zero restrictions using your add-in?

Kind regards and many thanks

Thank you for posting this. Trying to specify a SVAR with 6 variables, I get the following error message

"6 is not a valid index for vector-series-coefficient SCALES01." Would you be so kind and help me in interpreting/solving this?

Also, is it possible to implement a combination of sign and zero restrictions using your add-in?

Kind regards and many thanks

### Re: Sign Restricted VAR

Could you post the working file?

No. It is not possible.

Also, is it possible to implement a combination of sign and zero restrictions using your add-in?

No. It is not possible.

### Re: Sign Restricted VAR

I have a basic and very silly question about interpreting the impulse response graphs that come as the output from the srvar command. When doing a standard VAR, it gives us graphs which explicitly state the effect of endogenous x on y... however, srvar just gives graphs with "impulse response of X." In my data, I am using two variables, one which is a dependent and one which is independent but both are endogenous. Would I interpret the impulse function of Y graph as the effect of an innovation shock in X? And vice versa?

### Re: Sign Restricted VAR

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.

### Re: Sign Restricted VAR

dakila wrote: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.

### Who is online

Users browsing this forum: No registered users and 17 guests