st: AW: Autocorrelation after xtdpdsys

["Dithmer, Jan" <jdithme@food-econ.uni-kiel.de>]

Dear Angela,

as is stated in the help file of the program - help xtdpdsys - 
artests(#) specifies the maximum order of the autocorrelation test to be calculated
It does not correct for anything, it is just a test for autocorrelation of different orders.
It looks fine in your case, as 1st order autocorrelation may be present but 2nd order must not,
for the model to be consistent. By the way, your number of instruments is very high. You may
want to cross-check your results when limiting the lags used for instrumentation and should also
test the validity of the employed instruments.

Best, Jan

-----Ursprüngliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von Angela C
Gesendet: Friday, February 01, 2013 2:20 AM
An: statalist@hsphsun2.harvard.edu
Betreff: st: Autocorrelation after xtdpdsys

I am using xtdpdsys to estimate a panel with N=27 and T=17.

the results come out as follows

xtdpdsys Deb Gr INT, lags(1) vce(robust) artests(2)

System dynamic panel-data estimation         Number of obs         =       430
Group variable: CTY                          Number of groups      =        27
Time variable: Year
                                             Obs per group:    min =        14
                                                               avg =  15.92593
                                                               max =        16

Number of instruments =    138               Wald chi2(3)          =    677.64
                                             Prob > chi2           =    0.0000
One-step results
------------------------------------------------------------------------------
             |               Robust
   Deb |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------
-------------+------
   Deb |
         L1. |   1.091666   .0627754    17.39   0.000     .9686287    1.214704
             |
      Gr |  -104.3667   16.53084    -6.31   0.000    -136.7665   -71.96683
         INT |   -1.38898   .4177452    -3.32   0.001    -2.207746   -.5702147
       _cons |   2.437429   2.130174     1.14   0.253    -1.737636    6.612494
------------------------------------------------------------------------------
Instruments for differenced equation
        GMM-type: L(2/.).Deb
        Standard: D.Gr D.INT
Instruments for level equation
        GMM-type: LD.Deb
        Standard: _cons



I then run the Arellano-Bond postestimation test and stata returns:

Arellano-Bond test for zero autocorrelation in first-differenced errors
  +-----------------------+
  |Order |  z     Prob > z|
  |------+----------------|
  |   1  |-3.4443  0.0006 |
  |   2  |-.37263  0.7094 |
  +-----------------------+
   H0: no autocorrelation

According to my model, the 2nd order autocorrelation appearing in the test has been corrected by the programme (I set artests(2))?

If it was not corrected, is there a way to correct it afterwards and take new, corrected coefficients?

Thank you.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/