Deterministic Shock in A Stochastic Model- IRF

[tomnilson40]

Hi there,

I have been trying to see the impact of a deterministic shock in a stochastic model and in almost all of the IRF's there are some fluctuations before they get to their new steady state (see the attachment as an example). My question is if this makes sense.

My understanding is that the IRF would only get to the new steady state SMOOTHLY if the model is set up in a deterministic framework as well! (NOT SURE though!)

Any help is greatly appreciated.

Cheers
Tom

[jpfeifer]

Could you provide me with the mod-file to see what is actually going on?

[tomnilson40]

Thanks for that! I just sent it to you in a PM.

[jpfeifer]

Your model is not really stochastic. You solve it under perfect foresight. The IRFs you get simply reflect the endogenous dynamic of your model. If you simulate more periods, the IRFS at the end are rather smooth. I don't know that is going on at the beginning, but overshooting of IRFs is not uncommon. Consider for example a positive and persistent TFP shock. In this case the household will initially work more because it pays off and will save into capital. Later on, when TFP is back to normal, the household still has high savings. It will draw down the capital stock and work less than before the impulse (take a vacation). This will result in an initial over- and then undershooting. You need to better understand the economic mechanisms of your model to understand if there is a problem.