helo,
I wonder the diffence between --xtscc ..,fe-- and --xtpsce ...dummies,
correlation(ar1)-- ? It seems that they make the same hypotesis :
heteroskedastic, autocorrelated up to some lag and possibly correlated
between the groups (panels)
but the results are quite different :
. xtpcse lm_dolc lpib_dolc lagsum new_* , correlation(ar1) rhotype(regress)
(note: estimates of rho outside [-1,1] bounded to be in the range [-1,1])
Prais-Winsten regression, correlated panels corrected standard errors
(PCSEs)
Group variable: newid Number of obs
= 727
Time variable: obs Number of groups
= 14
Panels: correlated (unbalanced) Obs per group: min
= 51
Autocorrelation: common AR(1) avg =
51.92857
Sigma computed by casewise selection max
= 52
Estimated covariances = 105 R-squared =
0.9620
Estimated autocorrelations = 1 Wald chi2(15) =
6052.70
Estimated coefficients = 16 Prob > chi2 =
0.0000
------------------------------------------------------------------------------
| Panel-corrected
lm_dolc | Coef. Std. Err. z P>|z| [95% Conf.
Interval]
-------------+----------------------------------------------------------------
lpib_dolc | .4707492 .046833 10.05 0.000 .3789581
.5625402
lagsum | .0183111 .0812503 0.23 0.822 -.1409366
.1775587
new_1 | 2.308959 .23099 10.00 0.000 1.856227 2.761691
new_2 | 2.831064 .2640616 10.72 0.000 2.313513 3.348615
new_3 | 2.042727 .217553 9.39 0.000 1.61633 2.469123
new_4 | 3.725258 .2971162 12.54 0.000 3.142921 4.307595
new_5 | 3.437277 .3115808 11.03 0.000 2.82659 4.047964
new_6 | 2.439881 .2631095 9.27 0.000 1.924195 2.955566
new_7 | 3.052517 .2688442 11.35 0.000 2.525592 3.579442
new_8 | (dropped)
new_9 | 2.703009 .2567918 10.53 0.000 2.199707 3.206312
new_10 | 2.110368 .2099764 10.05 0.000 1.698822 2.521914
new_11 | 2.074326 .2524244 8.22 0.000 1.579583 2.569069
new_12 | 2.520334 .2575537 9.79 0.000 2.015538 3.02513
new_13 | 3.940914 .2955988 13.33 0.000 3.361551 4.520277
new_14 | 5.234917 .4166243 12.57 0.000 4.418348 6.051486
_cons | -.063312 1.005865 -0.06 0.950 -2.034771
1.908146
-------------+----------------------------------------------------------------
rho | .9187073
. xtscc lm_dolc lpib_dolc lagsum , fe lag(1)
Regression with Driscoll-Kraay standard errors Number of obs
= 727
Method: Fixed-effects regression Number of groups
= 14
Group variable (i): newid F( 2, 13) =
202.94
maximum lag: 1 Prob > F =
0.0000
within R-squared =
0.6606
------------------------------------------------------------------------------
| Drisc/Kraay
lm_dolc | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+----------------------------------------------------------------
lpib_dolc | .5669392 .0394445 14.37 0.000 .4817244
.6521539
lagsum | -.0010265 .1381459 -0.01 0.994 -.2994727
.2974196
_cons | 1.781487 1.483853 1.20 0.251 -1.424182
4.987155
Could you help me ?
Best
Ghislain
*
* 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/
