Matrices before VAR representation

Hi all,

I would like to obtain such matrices as alpha0, alpha1 and alpha2 as below;
alpha0Z(+1)+alpha1Z+alpha2*Z(-1)=0,
where Z is the vector of endogenous variables. This is the linearlized forward looking model before backward (Blanchard-Kahn) transformation (in some cases, this may include Z(-2) etc).

I know that I can obtain the matrix, ghx, contained in dr_, in backward state space form as below:
Z=ghx*X(-1),
where X is the vector of state variables. However, I would like to obtain the matrices in dynamic system in forward looking state space form as well.

Thanks in advance,
Ippei

Dear Ippei,

these matrices aren’t made available. But they exist in dr1.m.

They are submatrices of jacobia_, around line 47.

jacobia_ is organized as

[alpha2 alpha1 alpha0] in your notation

In each of alpha0, alpha1, alpha2 the variables are arranged in alphabetical order.

The corresponding variable and lag of each column of jacobia_ is given in iy_

The first row of iy_ represent maximum lag

Kind regards

Michel

Dear Michel,

Thank you very much for your reply and detailed information. I really appreciate it!

Cheers,
Ippei