Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Questions concerning the xttest2

From   "Vasilis Sarafidis" <>
To   <>
Subject   st: RE: Questions concerning the xttest2
Date   Tue, 6 Mar 2007 08:58:46 +1100

I don't know what the problem is, however even if you make it work, you
need to be very cautious with interpreting your results about
cross-sectional dependence.  The xttest2 command is based on the LM
test, which is asymptotically valid for N fixed T large.  If N is of
similar size with (large) T, the LM test may be subject to large bias.
If N and T are of similar size and on top of that T is small (like in
your case) the bias will be horrendous. You may be better off by using
the xtcsd command (ssc instal xtcsd) and use the 'pesaran' option.



-----Original Message-----
[] On Behalf Of ????
Sent: Monday, 5 March 2007 5:55 PM
Subject: st: Questions concerning the xttest2

Dear users,

I tried to use the xttest2 to check for cross-sectional correlation.
Yet, I obtain the error message:
"Correlation matrix of residuals is singular.
not possible with test"

The problems seems to be, that the number of households in my samples is
bigger than the number of periodes. But the nember of countries in my
samples is smaller than the number of periods. (N=15, T=17)

Waht is the cause of the problem?
In that case,  how can I test for contemporaneous correlation?
And  if PCSE analysis ignoring the error message, any problem?

Any help would be greatly appreciated.
Thank you.

Youngmi Kim
*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index