Re: st: SVAR estimation with Stata

[Syed Basher <syed.basher@yahoo.com>]

Hi,
I think your B matrix is not correctly specified. Replace "1" with "." in your B matrix, it should work. Have a look at the the Stata manual (var svar), it is useful.

--
Syed Abul Basher
Qatar Central Bank
P.O. Box 1234
Doha, Qatar.
http://www.syedbasher.org/


----- Original Message -----
> From: "ting.wang@yale.edu" <ting.wang@yale.edu>
> To: statalist@hsphsun2.harvard.edu
> Cc: 
> Sent: Tuesday, May 3, 2011 5:22 AM
> Subject: Re: st: SVAR estimation with Stata
> 
> Dear all,
> 
> I am running Structural VAR model with Stata. Suppose the contemporaneous
> coefficient matrix is G0, that is,
> 
> e(t)=G0*u(t),
> 
> where e(t) is the structural error and u(t) is the error of the reduced form
> VAR.
> 
> My inquiry is about how to set up the A and B matrix in Stata based on G0. Is B
> matrix supposed to be an identity matrix? How do we involve lags in structural
> VAR estimation? If G0 is considered to be instantaneous, does it really matter
> how many lags to include in the model?
> 
> Here is the codes I used in Stata. It does not converge when we carry out a
> maximum likelihood estimation. Iteration only goes on forever, which forces me
> to limit the maximum number of iterations.
> 
> matrix A (exactly the G0 matrix)
> =(1,.,0,0,.,0,.\.,1,.,.,0,0,0\0,0,1,.,.,0,0\0,0,0,1,.,0,0\0,0,0,0,1,0,0\0,0,0,0,.,1,0\.,.,.,.,.,.,1)
> matrix B
> =(1,0,0,0,0,0,0\0,1,0,0,0,0,0\0,0,1,0,0,0,0\0,0,0,1,0,0,0\0,0,0,0,1,0,0\0,0,0,0,0,1,0\0,0,0,0,0,0,1)
> 
> svar `varlist', lags(1/6) aeq(A) beq(B)
> 
> I hope someone can help with this. Thank you very much. I really appreciate it.
> 
> *
> *  For searches and help try:
> *  http://www.stata.com/help.cgi?search
> *  http://www.stata.com/support/statalist/faq
> *  http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/