Date:         Sun, 30 Oct 2011 16:44:19 -0500
Reply-To:     Gabe <gzs872@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Gabe <gzs872@GMAIL.COM>
Subject:      Seemingly Unrelated Regression
Content-Type: text/plain; charset=ISO-8859-1

Hello List,

I have a statistics and a related SAS question. I have two equations
(Y1i=c1+b11X1i+b12X2i+u1i) and (Y2j=c2+b21X2j+b22X2j+u2j) i and j are not
equal. This is to say that the two estimations can be done from two
different tables and they don't have any common observations.  So for
example Y1 and Y2 are firm's output and grouped by different type of
ownership where one is domestic owned firm and the other is foreign owned
firm (there are no firms owned by both domestic and foreign person). The
exogenous variables (the Xs) are industry variables. Despite the difference
in ownership, these firms could belong to the same industry and have same
Xs. I want to know the interdependence between the two firms an same
industry. I have two sets of observations i.e. one for domestic owners and
the other for foreign owned firms. And the number of observations is
different. I wanted to do SURE for these two equations. I have done SURE
when the observation comes from the subject (or a firm) but the subject is
different in this case. How can I do SURE in this case? And what is the SAS
proc in this case?

Thank you in advance,
Gabe