Hi there,

the SVAR analysis is a powerful tool and Jack's SVAR package already in it's present state is magnificent!
At the moment I'm working through it step by step. My ultimate goal is to perform a 'Structural vector autoregressions with
nonnormal residuals' by Lütkepohl (see attached file). Unfortunately his examples are SVECMs which makes the stuff even more complicated.
However, while learning from the 'SVAR code' and with Lütkepohl's book 'New Introduction to Multiple Time Series Analysis' I came across the following questions:

a) In his book on page 367 (section 9.1.3) the Rd matrix (following the notation of Jack) seems to me wrong.
b) Is it in order to use 'abs(nullspace(R))~R'invpd(R*R')*d' instead of the things done in imp2exp (and btw. exp stands for expansion, or, and imp for ?)

<hansl>
nulldata 10
include SVAR.gfn
#example
matrix R_Luedkepohl={1,0,0,0,0,0,0,0,0;0,0,1,0,0,0,0,0,0;0,0,0,0,1,0,0,0,0;0,0,0,0,0,1,0,0,0;0,1,0,0,0,0,1,0,0;0,0,0,0,0,0,0,0,1}
eval R_Luedkepohl
matrix R={1,0,0,0,0,0,0,0,0;0,0,1,0,0,0,0,0,0;0,0,0,0,1,0,0,0,0;0,0,0,0,0,1,0,0,0;0,0,0,0,0,0,1,0,0;0,0,0,0,0,0,0,0,1}
eval R
matrix R_orth={0,0,0;1,0,0;0,0,0;0,1,0;0,0,0;0,0,0;0,0,0;0,0,1;0,0,0}
eval R_orth
eval R*R_orth
matrix d ={1;0;1;0;0;1}
matrix Rd=R~d
eval imp2exp(Rd)
matrix s = R'invpd(R*R')*d
eval abs(nullspace(R))~s
<hansl>

Cheers
Leon