Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | "D. Demetriou" <dd358@cam.ac.uk> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: xtoverid' command and 'sigma_u=0 |
Date | 28 Mar 2012 16:57:12 +0100 |
Hi everyone,I am basically trying to use the 'xtoverid' command after performing an ' xi: xtreg, re vce(cluster) ' regression on panel data.
The problem here is that sigma_u (and hence rho subsequently) equals to zero (sigma_u = 0). Why might be I getting this?? The interesting thing here is that when I add a 'time dummy' (equals 1 for each period except the reference period) sigma_u is no longer zero.
Sigma_u=0 disables the 'xtoverid' command which I need to decide on FE Vs RE.
Here is the estimation table:Random-effects GLS regression Number of obs = 90 Group variable: region Number of groups = 9
R-sq: within = 0.7575 Obs per group: min = 10 between = 0.4357 avg = 10.0 overall = 0.7426 max = 10Wald chi2(5) = 493.00 corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000
(Std. Err. adjusted for 9 clusters in region) ------------------------------------------------------------------------------ | Robust publicprov1 | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+----------------------------------------------------------------
hownership | -.0731302 .0143642 -5.09 0.000 -.1012835 -.0449769 incvars | -.1222154 .0138148 -8.85 0.000 -.1492918 -.095139 popabove60 | -.0387703 .0197494 -1.96 0.050 -.0774783 -.0000622 ILO | -.2344766 .0418934 -5.60 0.000 -.3165863 -.152367 wave | -.0623215 .0242992 -2.56 0.010 -.109947 -.0146959_cons | 30.88966 1.00746 30.66 0.000 28.91508 32.86425 -------------+----------------------------------------------------------------
sigma_u | 0 sigma_e | .29067783rho | 0 (fraction of variance due to u_i) ------------------------------------------------------------------------------
* * 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/