Help with SVAR modelling

[Oscar Craner]

Hi, I'm trying to follow the empirical strategy of this paper on Stata.

I have the datasets but I'm struggling as I'm confused as to what a recursive restriction is and what the B matrix for this specific SVAR model is.

Any help would be appreciated, thanks.

[Luis Pecht]

Oscar, welcome to this forum,

VAR technique has two-fold objetives; prediction and policy evalution. In the former, recursive restriction allows for "casual inference" (ideally) interpretation of the coefficientes and this restriction can be imposed in the SVAR:

 
. matrix A = (1, 0, 0\., 1, 0\., ., 1)
. matrix B = (., 0, 0\0, ., 0\0, 0, .)

. svar inflation unrate ffr, lags(1/2) aeq(A) beq(B)

or, in the varlist ordering, using

varbasic varlist, lags(#) step(#)

or, after, in the IRF or LPIRF( Stata18) commands..


I strongly recommend "John D. Levendis, Time Series Econometrics - Learning Through Replication, 2ed." ( Chapter 11), which is 100% Stata coded .

LP

[Oscar Craner]

Ok, thank you!