Interpretation of shocks in IRFs

[bmicallef]

Hi,

I have estimated a small scale NKM with bayesian inference. Suppose I only have two shocks, a monetary shock and a technology shock, with the standard errors of the shocks (estimated with inv_gamma_pdf) being 0.55 and 1.7, respectively.

In the shock block, I wrote:

shocks;
var em; stderr 0.55;
var ez; stderr 1.7;
end;

How should I interpret the results of the IRF? Let's take the monetary shock. What is the correct interpretation, that say GDP declined by X% following a one standard deviation shock or a 1% shock? If the former is the correct interpretation, what should I do to shock the model by 1%, since I am interested in this interpretation?

Thanks for your help.

[p.gelain]

The correct interpretation is

a one standard deviation shock

If you want 1% shock you may use the "relative_irf" command

Alternative, but I am not sure, you can try dividing alla the standard deviation for the shock you want the 1%. Namely if you are interested to em


shocks;
var em; stderr 0.55/0.55;
var ez; stderr 1.7/0.55;
end;