Page 1 of 1
Near-VAR
Posted: Thu Sep 10, 2009 12:34 pm
by raphael
Dear Tom, is there a difference between the method from Enders in its Programming Mannual, which is to estimate each equation using LINREG(define = equation x) then use SUR and finally GROUP the equations, and the one in the User's Guide, p. 343 ?
Thanks
Re: Near-VAR
Posted: Thu Sep 10, 2009 2:52 pm
by TomDoan
Yes. For a near-VAR, the estimates from a SUR are different for the equations that have "extra" variables than they are for the same equation done one LINREG at a time. They usually aren't that much different, but they *are* different. The equations with the subset of the RHS variables end up being the same both ways.