Page **1** of **1**

### SVAR Blanchard and Perotti 2002

Posted: **Fri May 06, 2011 3:24 am**

by **roxanac**

Hi ,

I am trying to estimate a 5 variable SVAR following Blanchard and Perotti (2002) (AN EMPIRICAL CHARACTERIZATION OF THE DYNAMIC

EFFECTS OF CHANGES IN GOVERNMENT SPENDING AND TAXES ON OUTPUT ) and Perotti (2004).

Do you happen to now if I can do it in Eviews 7 ?Is there an example on how I should proceed ?

Thank you very much

Roxana

### Re: SVAR Blanchard and Perotti 2002

Posted: **Mon May 09, 2011 4:18 am**

by **trubador**

If you are familiar with the theoretical aspects of SVAR modeling, then you will find it pretty straightforward to do it in EViews. Please refer to manual for a detailed explanation on how you can estimate a Structural VAR model in EViews.

### Re: SVAR Blanchard and Perotti 2002

Posted: **Fri May 20, 2011 12:14 am**

by **roxanac**

Hi,

I managed to do it by setting an AB model by imposing short run restrictions .Howhever my IRF when choosing Analytic and Structural Decomposition are flat.

I can not figure out what I did wrong .

Can you please advise me ?

Thank you,

Roxana

### Re: SVAR Blanchard and Perotti 2002

Posted: **Sat May 21, 2011 6:52 am**

by **izabelab**

Hi Roxana. I think the answer to your question is that the initial data included are I(0). For stationary VARs, the impulse responses should die out to zero and the accumulated responses should asymptote to some (non-zero) constant. I am trying to replicate the same model on Romanian data. Are you somehow doing the same thing?

### Re: SVAR Blanchard and Perotti 2002

Posted: **Thu Jul 12, 2012 1:25 pm**

by **msh855**

Hi all,

I am doing the same. However, not sure about a few steps.

first, in the identification restrictions do you use the coefficients that BP provide for c1 c2, or you leave Eviews to estimate them? I am asking because BP estimated c1 and c2 using IV regressions. And from what I checked from their code, they did this before imposing restrictions on the matrix.

Secondly, how do I perform the re-scaling for getting the dollar unit increase impulse instead of 1% deviation and therefor to the get the short-run multipliers?

Could you please anyone here help me on this? I am not that strong using Eviews and actually just now start playing around.

Many thanks

### Re: SVAR Blanchard and Perotti 2002

Posted: **Thu Jul 26, 2012 11:59 pm**

by **icprag**

Hi! i am trying to use BP identification and i am dealing with the same problem as Roxana did, my IRF are flat. Could you please anyone help me?

Kindest regards,

ip

### Re: SVAR Blanchard and Perotti 2002

Posted: **Wed Mar 27, 2013 11:36 pm**

by **lawrence**

Dear,

Doing the SVAR BP for 3 variables is easy and I will upload an example. We should remember since the E(uu')=I, then the diagonal of matrix B will be na's and they are st. errors of vars and if we compare this with RATS , E(uu') D ( cov-matrix is a diagonal) and I will add this first. But my problem is if we perform a 4-svar model following BP i.e., if we put order them as tax , ge, gdp then consumption(cons) for example. In this case, E(uu')=1 but there is a cov b/w gdp and cons.For moderators, please respond to this question. How can we deal with the 4-svar

Thanks in advance

Lawrence

### Re: SVAR Blanchard and Perotti 2002

Posted: **Fri Mar 29, 2013 7:42 am**

by **lawrence**

Dear Eviews followers,

Any one can answer this please . I am waiting since it is very important to me in a time wise

Regards,

Lawrence

### Re: SVAR Blanchard and Perotti 2002

Posted: **Mon Apr 14, 2014 3:02 pm**

by **felix.casares**

Hi.

please, anyone has de data file of the paper AN EMPIRICAL CHARACTERIZATION OF THE DYNAMIC EFFECTS OF CHANGES IN GOVERNMENT SPENDING AND TAXES ON OUTPUT.

i need to replicate in eviews but i dont have the matrix restriction.