interpretation of sigma_u, sigma_e and rho in xtreg, re

[Laura Hess]

Hello,
I am currently analysing some paneldata. The hausman test showed that I need to use the random effects model for this data set.
Following a part of my output:

sigma_u | 0
sigma_e | .16219692
rho | 0

How can I interpret sigma_u, sigma_e and rho?
Are these results a problem for my estimate?

Thank you in advance for your response!

[GUIRA Asmo]

hello,
I need your help. what command to use to know if these effects are random or fixed

[Carlo Lazzaro]

Laura:
welcome to this forum.
Please show what you typed and what Stata gave you back (as per FAQ).
Your results, wherever they come from, deny the evidence of any panel-wise effect.

[Carlo Lazzaro]

GUIRA:
go -hausman- if you're using default standard errors.

[Laura Hess]

This was my hausman test:

hausman fixed random

---- Coefficients ----
| (b) (B) (b-B) sqrt(diag(V_b-V_B))
| fixed random Difference Std. err.
-------------+----------------------------------------------------------------
M. | .5398418 .5396112 .0002306 .0001368
------------------------------------------------------------------------------
b = Consistent under H0 and Ha; obtained from xtreg.
B = Inconsistent under Ha, efficient under H0; obtained from xtreg.

Test of H0: Difference in coefficients not systematic

chi2(1) = (b-B)'[(V_b-V_B)^(-1)](b-B)
= 2.84
Prob > chi2 = 0.0918


and this is my regression:

xtreg R. M., re

Random-effects GLS regression Number of obs = 171,396
Group variable: id Number of groups = 529

R-squared: Obs per group:
Within = 0.5785 min = 324
Between = 0.9729 avg = 324.0
Overall = 0.5813 max = 324

Wald chi2(1) = 237907.07
corr(u_i, X) = 0 (assumed) Prob > chi2 = 0.0000

------------------------------------------------------------------------------
R. | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
M. | .5396112 .0011063 487.76 0.000 .5374428 .5417795
_cons | .0002957 .000412 0.72 0.473 -.0005117 .0011031
-------------+----------------------------------------------------------------
sigma_u | 0
sigma_e | .16219692
rho | 0 (fraction of variance due to u_i)
------------------------------------------------------------------------------

[Carlo Lazzaro]

Laura:
while -hausman- confirms -re-, you do not seem to have a panel-wise effect (sigmau=0).
In all likelihood, your model is ill-specified, since you have one predictor only.

[Laura Hess]

okay, thank you very much!

[Laura Hess]

apart from that: how can these three key figures be interpreted in general?

[Carlo Lazzaro]

Laura:
see -xtreg- entry in Stata .pdf manual and/or any decent textbook on panel data econometrics therein included.