st: Xtreg, fe robust cluster() and Xtregar

[May Ster <mayfrank@gmail.com>]

Dear all,

I wonder which is better between Xtregar, fe robust cluster() and
Xtregar under FE model if one finds autocorrelation or even assumes
they exist.

I tried running regressions and found substantive differences provided
by these two estimators. I thought if autocorrelation exists, it
shouldn't change the coefficient that largely i.e. only the standard
errors are effects?

And, since -Xtregar also works under the FE , it then should provide
close if not similar estimates to the -Xtreg, fe robust cluster()?
I don't understand why the estimates provided are different largely
and there are even the changes in signs.

I will be very grateful if anyone could help with this. Thank you very
much in advance,

May
*
*   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/