Impulse response function

[msh855]

Hi all,

I get an Oscillating IRF around the zero axis until they die out.

Does anyone know why this may happen ?

I have accounted for seasonality, but do not really understand the source of such unusual fluctuating IRF

Many thanks

[startz]

An oscillating IRF can occur anytime you have two or more lags, even for a single equation. Essentially, part of the solution to a seond order difference equation with comple roots can involve a sine wave.

[msh855]

Startz, Thank you so much ! That was indeed helpful !

However, for my case i cannot consider it as something normal, since the impulses here describe the fiscal multipliers. I.e dt/dy <=0 and dg/dy >=0
Where t=taxes, g=government spending, y=gdp and d()/d() as usual denote derivatives.

And indeed I had observed myself that the "oscillation" was a function of lags.

Do you recommend / suggest / advice something about this ? cause the maths make sense but not really the economics.

I have applied this VAR specification for a couple of countries but only on one occasion i found such oscillating impulse response.

[startz]

My only suggestion is to look at the confidence intervals to be sure whether there is really string evidence of the oscillation.

[msh855]

Thanks, Startz. I will try it.

[mcm]

Hello,
I'm estimating a SVAR with long-run restrictions and want to simulate disinflation's shocks. So,
1- how can'i simulate explicitly a negative shock on Eviews? i think, by default, Eviews simulates positive shocks.
2- On the other hand, how can'i simulate a 1% shock, a 2% shock, and so on.....to then obtain IRF from a 1% shock, 2% shock, 3% shock.....

Thanks!!
As my english is not good i try in french!

En fait je travaille sur les politiques de désinflation, alors j'ai des préoccupations de deux ordres:
1- Comment simuler un choc négatif sur Eviews? car j'ai l'impression que par défaut, les IRF sont issues de chocs positifs?
2- Comment simsuler des chocs de 1%, 2%, 3% et ainsi de suite sur Eviews? j'ai l'impression que par defaut les IRF sont issues de chocs de 1% uniquement.
Merci davance!!