Re: Singular Data

On Sat, 29 Sep 2007 13:23:18 -0700, nick <kwstaskoukou@xxxxxxxxxxx>
wrote:

On 29 , 04:54, Richard Ulrich <Rich.Ulr...@xxxxxxxxxxx> wrote:
On Fri, 28 Sep 2007 13:52:33 -0700, nick <kwstaskou...@xxxxxxxxxxx>
wrote:

Hi,
I am using GMM (Generalized Method of Moments) to estimate a dynamic
panel in TSP. However, I receive the following error message
"At least one coefficient in the table above could not be estimated
due to singularity of the data or derivatives"
Any ideas?

idea 1: You have a singularity in the data, but you
don't know what that means.

idea 2: You know that you don't have a singularity,
and this is a question about peculiarities of TSP.

We don't see a lot of TSP questions here, so I don't
know if there are resident experts in it or not.
But someone can probably say something about
where to seek TSP answers, if they know that is the
question.

Are you at (1) or (2)?

--
Rich Ulrich, wpi...@xxxxxxxxxxx://www.pitt.edu/~wpilib/index.html

Well,
I know that my data are singular. However, there has to be a way to
overcome the problem of singularity. Maybe some kind of
transformation. Or maybe I could add some "noise" to my data.
What do you think?

If you have a singularity in the data, then there really is no cure
because the data doesn't have the information you need.

If that's not the problem, then

(a) if the problem is linear you may have a specification error,
something equivalent to the dummy variable trap.

(b) if the problem is nonlinear, try different starting values.

Two alternative:

(1) email TSP, they're quite responsive.
(2) read the file into EViews and see if you have the same problem

-*** Startz
.