Hallo,
Is there someone who can help me in interpreting correctly the constant term in a xtreg procedure with fixed effects?
I have estimated the following model:
xtreg ros lros, fe
I have obtained the following results:
Fixed-effects (within) regression Number of obs = 18102
Group variable: codice_i Number of groups = 2586
R-sq: within = 0.1347 Obs per group: min = 7
between = 0.9609 avg = 7.0
overall = 0.6225 max = 7
F(1,15515) = 2415.35
corr(u_i, Xb) = 0.7916 Prob > F = 0.0000
------------------------------------------------------------------------------
ros | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
ros |
L1. | .3752229 .0076348 49.15 0.000 .3602578 .390188
_cons | 6.288536 .0885829 70.99 0.000 6.114903 6.462169
-------------+----------------------------------------------------------------
sigma_u | 3.7527417
sigma_e | 4.0508966
rho | .46184853 (fraction of variance due to u_i)
------------------------------------------------------------------------------
F test that all u_i=0: F(2585, 15515) = 2.24 Prob > F = 0.0000
I thought that the constant term (_cons=6.288536) was the average of the individual fixed effects.
Instead, if i compute summary statistics of the fixed effects obtained by using the following command:
summarize fe
Ithe results are:
Variable | Obs Mean Std. Dev. Min Max
-------------+--------------------------------------------------------
fe | 18102 .0010154 3.752186 -20.95681 17.61897
How is possible that the mean value of the fixed effects (.0010154) is so different with respect to the constant term in the xtreg output?
Which is the right interpretation of the constant term in the xtreg estimation with fixed effects?
Thank you for the attention
Eleonora Bartoloni
*
* 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/
