On Jun 9, 2005, at 11:40 AM, wei song wrote:
To whom it may concern,
This is my fisrt try of SVAR. I hope there is someone can help me.
Your best bet is the Time Series [TS] manual:
<http://www.stata.com/bookstore/ts.html>; see -var intro-. If you are
new to working with VAR models, I'd recommend Walter Enders' "Applied
Time Series Analysis", 2nd edition (Wiley, 2004), particularly chapters
5 and 6. The review article by Stock & Watson (2001) mentioned in the
references of the [TS] manual for -var intro- and -var svar- might also
be helpful. Amisano & Giannini (1997) (also cited in the references)
is a more advanced (but commplete) reference for SVARs.
Given a two-equation model:
Yt=a10+a11*Tt+b11*Yt-1+b12*Tt-1
Tt=a20+a21*Yt+b21*Yt-1+b22*Tt-1
a11=0
Since this appears to be a simple (just-identified) recursive model,
Example 1 in [TS] -var svar- explains how to impose the Choleski
decomposition with the -svar- command. Basically, it looks something
like the following (untested):
mat A = (1,0\.,1)
mat B = (.,0\0,.)
svar Y T, lags(1) aeq(A) beq(B)
and the error term of the second equation=0
You've lost me here: the second equation is non-stochastic?
how to set the constraints when I run SVAR command? What do the matrix
A, B (short run constraints) and C (long run constraints) refer?
See above for A and B; since your model imposes only contemporaneous
constraints (at least as I understand it), there is no role for C.
Also, Stata is not capable of simultaneously using short-run and
long-run restrictions. This all is explained in the Time Series [TS]
manual, as well as in Amisano & Giannini (1997). You will want to read
carefully one or more of these sources before proceeding with any
(S)VAR estimation. Watson's (1994) Handbook chapter (see references)
also is recommended. Hope this helps.
-- Mike
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
