Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: SUR system - Compare intercepts from several regression

From   Adrian Stork <>
Subject   st: SUR system - Compare intercepts from several regression
Date   Tue, 29 Oct 2013 20:01:10 +0100

Dear Stata friends,

Sorry for asking but I've got the following problem(s):
Starting with the actual issue I simply want to compare intercept
values from 50 regressions with each other. I found the following
giving me almost the solution. It works perfectly fine for the
coefficients but I don't know how to apply this test on the intercept.
I tried

suest regprice regmpg
test [regprice_mean]intercept=[regmpg_mean]intercept but that doesn't work.

In addition when I looked at the suest command in the description, I
saw that it refers to "Seemingly unrelated estimation" what is
actually what I wanted in the first place (running "sureg" over all 50
dependent variables and compare intercepts) I received the error
message "insufficient observations". Of couse, I know that I have
quite a few missing values (i.e. the first dependent variable is
complete with monthly 250 observations while the second one has merely
50 monthly observations, etc.). Still, when I running the SUR system
with the first and second dependent variable (on independent variables
with complete 250 observations), I get the error message "insufficient
obs". When I'm running it with suest: everthing fine and I get a test
statistic for the Null.
So, in a nutshell, how can I test the intercepts with suest and is
there a way to make "sureg" run although  I've got quite some missing
values in the dependent variables?

Best regards,
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index