|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: FGLS-SUR model estimation in STATA
GLS is a least squares procedure in which the variance-covariance matrix
of the residuals is not proportional to the identity matrix. My
interpretation is that FGLS refers to any estimation procedure in which an
initial procedure is used to generate a consistent estimate of the error
covariance matrix, which is then plugged in to a GLS procedure. Because
of this, I believe that any SUR procedure that does not ask you to somehow
input information about the error covariance matrix must estimate this
matrix itself, and so has to performing FGLS.
Kyle
On Thu, 7 Dec 2006, Jeremy Whitridge Cheesman wrote:
Hi,
I have a cross-sectional dataset that pools multiple observations
household volumetric demand for piped versus well water given changes in
piped and well water prices. I have multiple observations for each
household and different numbers of observations between households. I have
socio-economic information for each household that is invariant across
observations. I want to estimate a system of demand equations for both
private and well water controlling for error term correlation across the
equations and using the water prices and socio-economic data as
explanatory variabels.
Its my first time dealing with this sort of data. My understanding is that
I need to implement a Seemingly Unrelated Regression model with FGLS to
get efficient estimators. I have found how to implement SUR in STATA but I
don�t know whether this is a FGLS estimation or not. Can someone help? If
SUR is not FGLS, how do I go about implementing it in STATA?
Thanks,
Jeremy
*
* 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/
