Re: st: Fixed Effects controlling for Heteroskedasticity andAutocorrelation

["G. Chidambaran Iyer" <cgiyer@igidr.ac.in>]

Dear David,
Thank you very much for your prompt reply and the method
suggested. I have one more query: Wouldn't the command xtpcse do
the job for me if i add cross section dummies and then control for
heteroskedasticity and autocorrelation?

I tried matching an areg command on a panel data set I knew had fixed
effects with heteroskedasticity and then tried the same with the command xtpcse
after controlling for heteroskedaticity.

The results for both came the same, prompting me to think about the option
outlined above. I have pasted these results below

. areg logsa age roybysa rdbysa exbysa imcapbysa horisa bacsa forsa demsa
hhi loglab logcappim logrm , absorb(id1) robust

Regression with robust  standard errors              Number of obs = 2460
                                                       F( 13,  1917) =  475.53
                                                       Prob > F      =  0.0000
                                                       R-squared     =  0.9862
                                                       Adj R-squared =  0.9823
                                                       Root MSE      =  .19968

------------------------------------------------------------------------------
             |               Robust
       logsa |      Coef.   Std. Err.      t    P>|t|     [95% Conf.Interval]
-------------+----------------------------------------------------------------
         age |   .0076688   .0091298     0.84   0.401    -.0102365    .0255742
     roybysa |   1.274503   .6611799     1.93   0.054     -.022204    2.571211
      rdbysa |  -11.04995   2.600459    -4.25   0.000    -16.14997   -5.949919
      exbysa |   .1112171     .07028     1.58   0.114    -.0266162    .2490503
   imcapbysa |   -.099638   .0305532    -3.26   0.001     -.159559    -.039717
      horisa |   .2535539   .4686134     0.54   0.589    -.6654918      1.1726
       bacsa |   25.32033   14.98044     1.69   0.091     -4.05935        54.7
       forsa |   49.32972   19.06817     2.59   0.010     11.93318    86.72626
       demsa |  -.0000313   .0000352    -0.89   0.374    -.0001003    .0000377
         hhi |   .0004407   .0006658     0.66   0.508    -.0008651    .0017465
      loglab |   .1828649   .0229375     7.97   0.000     .1378798    .2278501
   logcappim |   .0296622   .0181171     1.64   0.102    -.0058691    .0651934
       logrm |   .9036091   .0278091    32.49   0.000     .8490699    .9581483
       _cons |  -1.134341   .3125901    -3.63   0.000    -1.747394   -.5212888
-------------+----------------------------------------------------------------
         id1 |   absorbed                                     (530 categories)


. xtpcse logsa firmdum* age roybysa rdbysa exbysa imcapbysa horisa bacsa
forsa demsa hhi loglab logcappim logrm , hetonly

Linear regression, heteroskedastic panels corrected standard errors

Group variable:   id1                           Number of obs      =      2460
Time variable:    year                          Number of groups   =       530
Panels:           heteroskedastic (unbalanced)  Obs per group: min =         1
Autocorrelation:  no autocorrelation                           avg =  4.641509
                                                               max =        16
Estimated covariances      =       530          R-squared          =    0.9862
Estimated autocorrelations =         0          Wald chi2(477)     =  1.41e+07
Estimated coefficients     =       543          Prob > chi2        =    0.0000

------------------------------------------------------------------------------
             |            Het-corrected
             |      Coef.   Std. Err.      z    P>|z|     [95% Conf.
Interval]
-------------+----------------------------------------------------------------
         age |   .0076688   .0073141     1.05   0.294    -.0066665    .0220041
     roybysa |   1.274503   .3309436     3.85   0.000     .6258659    1.923141
      rdbysa |  -11.04995   1.833549    -6.03   0.000    -14.64364   -7.456256
      exbysa |   .1112171   .0537815     2.07   0.039     .0058073    .2166268
   imcapbysa |   -.099638   .0303635    -3.28   0.001    -.1591494   -.0401266
      horisa |   .2535539     .37504     0.68   0.499     -.481511    .9886187
       bacsa |   25.32033   13.04146     1.94   0.052    -.2404652    50.88112
       forsa |   49.32972   17.76468     2.78   0.005     14.51159    84.14784
       demsa |  -.0000313   .0000298    -1.05   0.294    -.0000897    .0000272
         hhi |   .0004407   .0005643     0.78   0.435    -.0006653    .0015467
      loglab |   .1828649   .0148773    12.29   0.000      .153706    .2120239
   logcappim |   .0296622   .0141798     2.09   0.036     .0018702    .0574541
       logrm |   .9036091   .0179724    50.28   0.000     .8683837    .9388344
       _cons |  -1.504451   .5533657    -2.72   0.007    -2.589027   -.4198737
------------------------------------------------------------------------------


Would be grateful for your comments on using the xtpcse command for my
purpose. Thanking you for your time.

Sincerely

Chidambaran

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/