EViews.com

I feel like I must be missing something obvious. But is it possible to specify impulse response shocks other than 1 S.D. shocks when using a SVAR (structural decomposition) in Eviews 7.0? That is, I want to specify a 1% shock to government spending and trace its effect on several variables (GDP, Inflation, Fed Funds, etc) but my identification scheme is more complex than a Cholesky decomposition.

Anyway, It makes it easier and more straight forward to think about the multiplier when talking about shocks in percentage terms so I was hoping there was an easy way to do this.

If I understand your question, unfortunately not using the single radio button.

You can, however, always use the estimated factorization and the user-specified impulse response option to generate the desired impulses from innovations, or you can provide user-specified shocks to the residuals.

I guess my question was: How do I combine my structural decomposition with a user specified shock to the residual?

*edit* or I guess, is such a thing possible with the graphical interface?

Do you want the shock to the residual or to the innovation from the structural decomposition?

The residual, you don't really need to do any work as the user-specified shocks are residual shocks. To do a shock to the innovation, you'll need to extract the A and B decomposition matrices from the VAR object, and then feed your innovation shock through them to get residual s hocks.

Hi all,

Gentlemen, I have a question that can seem to be primitive.
What is the difference between static forecast and impulse response?

Dear EViews Glenn,
Dear EViews Users,

if I get it right, I should follow this sequence in order to induce simultaneous shocks to the innovations in an SVAR model:

- estimate the VAR model
- estimate the structural factorization
- extract the A and B matrices, and calculate (A^-1)*B
- use the calculated elements to induce the shock

Say that I have a VAR with three endogenous variables, of which the latter two have an immediate effect on the first one. My A-matrix hence contains non-zero elements for a11,a12,a13, a22 and a33. My B-matrix is restriced to a diagonal matrix. The resulting matrix M=(A^-1)*B consequently has the same "shape" as my A matrix. To induce a simultaneous shock to the innovations (the latter two variables), I specify a user-specified shock which has the form of a 3x1 matrix, with the following elements:

m12+m13
m22
m33

Can you, or anyone else, confirm that this is the right way to induce a simultaneous shock to the innovations in an SVAR model in EViews?

Thanks!

Maarten

Dear Eviews users/ Gareth/ Glenn,

Following the discussion on this topic, can someone please explain how to do the "use the calculated elements and induce a shock" part (also referred to as "feed your innovation shock through them" part) after I have obtained my VAR residuals, then estimated the structural factorization and obtained my two matrices- A and B and also calculated (A^-1)*B. I have a 4-variable VAR with budget deficit, GDP, Inflation and Interest rates. I want to find the effect of a -1% shock to the structural innovation of the budget deficit on all other variables. (shock to the innovation of only 1 variable) In desperate need of help! My lecturers said they are unable to help since this is not covered in our material for a Bachelors degree.

Thank you,

Elena