Re: [Gretl-users] Magnitude of VAR shocks
On Tue, 23 Apr 2013, Henrique Andrade wrote:
Em 23 de abril de 2013, Gabriela Nodari escreveu:
I was wondering if there is the possibility to change the magnitude of a
> given structural shock within a SVAR framework. By default, Gretl calculate
> IRFs based on a one-standard deviation shock. Can I change this magnitude?
[...]
If I am correct, all that you need to do is multiply (or divide) the
irf
function by some scalar (in my example I used x1 and x2).I hope this helps
you :-)
Exactly. As Henrique said, you can use the output of the function irf() if
you use the built-in var command, or the "IRFs" bundle key if you use the
SVAR package. Then, it's just a matter of normalising the scale of the
IRFs to a suitable value. The example below (which assumes you use SVAR,
but can easily adapted to the other case) normalises the IRFs so that the
diagonal element all start from one, so that the instantaneous response of
each variable to its 'own' shock is normalised to 1:
<hansl>
Mod = SVAR_setup("plain", X, Z, 3)
SVAR_estimate(&Mod)
matrix IrfMat = Mod["IRFs"]
scalar n = Mod["n"]
matrix I0 = diag(mshape(IrfMat[1,],n,n))'
matrix norm = I0 ** ones(1,n)
print norm
matrix IrfMatNorm = IrfMat ./ norm
print IrfMat IrfMatNorm
</hansl>
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------