Bookmark and Share

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


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

st: Doing a likelihood ratio test when estimating a system of equations


From   Alex Olssen <[email protected]>
To   [email protected]
Subject   st: Doing a likelihood ratio test when estimating a system of equations
Date   Wed, 9 Mar 2011 17:26:17 +1300

Dear Statalist,

I am estimating a system of share equations and I want to use a
likelihood-ratio test to see if certain linear restrictions on
coefficients are rejected or not.

Can I do this after -nlsur-?  With the -ifgnls- option the estimates
are equivalent to maximum-likelihood estimates, so will I be able to
do this?  Or is the likelihood-ratio test only implementable following
estimation that has used Stata's -ml-?
In [u] 20.11.3 it says likelihood-ratio tests can be estimated with
-estimates store- following an -ml maximize-.

Also an observation:

I tried to estimate my linear share system using -nlsur- in the hopes
of being able to easily perform a likelihood-ratio test afterwards.
I was surprised that while -sureg- managed to estimate the system
-nlsur- told me I had more parameters than observations.
I have 35 time periods and observations on 4 shares.  Using -sureg-
Stata seems to estimate OLS equation by equation where each equation
has 35 observations and 12 parameters to estimate.
Is -nlsur- trying to estimate 36 parameters with 35 observations?  If
so is there an obvious reason why this is happening?  I specified the
same linear model in both situations.  Is it that -sureg- knows to do
OLS equation by equation but with -nlsur- I am telling it not to do
that?

Any help would be appreciated.

Alex Olssen
Motu Economic Research
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


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