How to report F-test (footnote below -xtreg,fe-) results of clustered robust standard errors fixed-effect model

[Rosamond Waston]

Dear all,

I would like to ask how can I report the F-test results for the clustered robust standard errors fixed-effect model in stata? My case is following

First, I run the fixed-effect model:

xtset time country
xtreg v1 v3 v4 v5 v6 v7 v8, fe

where the footnotes of the fixed-effect model above told me that there is no unobserved heterogeneity (suggesting pooled OLS model):

F test that all u_i=0: F(19, 7768) = 0.84 Prob > F = 0.6605

Because there is heteroskedasticity existing in my model, it suggests me use clustered robust standard errors fixed effect model. However, the stata does not report the F-test results for unobserved heterogeneity automatically for the clustered robust standard errors (like below). Thus, i would like to ask how can report the F-test results of unobserved heterogeneity for clustered robust se fixed effect models (it does matter because I wish to motivate why I use pooled ols instead of fe model). Thanks in advance.

xtreg v1 v3 v4 v5 v6 v7 v8, fe vce(cluster time) nonest

Fixed-effects (within) regression Number of obs = 7794
Group variable: v2 Number of groups = 20

R-sq: within = 0.5410 Obs per group: min = 271
between = 0.6358 avg = 389.7
overall = 0.5411 max = 498

F(6,497) = 274.89
corr(u_i, Xb) = 0.0094 Prob > F = 0.0000

(Std. Err. adjusted for 498 clusters in v12)
------------------------------------------------------------------------------
| Robust
v1 | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
v3 | .9697055 .02715 35.72 0.000 .9163626 1.023048
v4 | .2430544 .0350753 6.93 0.000 .1741402 .3119686
v5 | .1299614 .0738581 1.76 0.079 -.0151513 .2750741
v6 | .0821983 .0320707 2.56 0.011 .0191874 .1452093
v7 | .1026056 .0371527 2.76 0.006 .0296099 .1756013
v8 | -.1126672 .0256944 -4.38 0.000 -.1631502 -.0621843
_cons | .0019904 .0009872 2.02 0.044 .0000507 .00393
-------------+----------------------------------------------------------------
sigma_u | .00143895
sigma_e | .03197364
rho | .00202129 (fraction of variance due to u_i)
------------------------------------------------------------------------------

.
end of do-file

[Carlo Lazzaro]

Rosamond:
you are dealing with a T>N panel dataset.
Therefore, you should switch from -xtreg,fe- to -xtregar,fe-, provided that -fe- is the way to go.

[Rosamond Waston]

Rosamond:
you are dealing with a T>N panel dataset.
Therefore, you should switch from -xtreg,fe- to -xtregar,fe-, provided that -fe- is the way to go.

Many thanks for your response. May I ask what does it mean for T and N? Does T mean the number of time periods?N means the number of observations?

[Carlo Lazzaro]

Correct.

[Maxence Morlet]

@rosamond: How many countries & time periods do you have?

[Carlo Lazzaro]

Maxence:
as per #1:
N=20
T (average)=389.7
NxT=7794