Each IRF graph plots the response of a variable to an orthogonalized impulse, correct? but the labels above each graph only refer to variables, e.g. "order1, ffr, inflation".
I understand that in this case the 3rd variable ("ffr") is used to label the 3rd orthogonalized impulse, which makes sense when the structure matrix is triangular since then it has a 1 in the (3,3) entry.
But what if you use a non-triangular structure matrix, e.g. A1 = (.,.,1 \ 1,0,0 \ 1,.,0)?
How does stata label the orthogonalized shocks in this case? It would be misleading to label the 3rd orthogonalized shock as "ffr" since it does not contain "ffr" as a component.
↧
Re: Structural vector autoregression models
↧