# Re: st: Stata7 and Stata8 are doing two very different things.

 From TEWODAJ MOGUES To Mark Schaffer Subject Re: st: Stata7 and Stata8 are doing two very different things. Date Wed, 12 May 2004 23:05:15 -0500

```Hi Mark,

Yes, let's use abdata as an example! Run the following regression, once using Stata7, and once using Stata8, and you'll see that the results differ:

xtabond n w L.k , lags(1) pre(ys , lag(0,1)) twostep noconstant

Running it in Stata7, you get the output:
==================================================================================

Arellano-Bond dynamic panel data                Number of obs      =       751
Group variable (i): id                          Number of groups   =       140

Wald chi2(.)       =         .

Time variable (t): year                         min number of obs  =         5
max number of obs  =         7
mean number of obs =  5.364286

Two-step results
------------------------------------------------------------------------------
n            |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
n            |
LD |   .3966709   .0447336     8.87   0.000     .3089946    .4843471
ys           |
D1 |   .8290601   .0596861    13.89   0.000     .7120775    .9460427
w            |
D1 |   -.527372   .0399636   -13.20   0.000    -.6056992   -.4490449
k            |
LD |   .1368371   .0324049     4.22   0.000     .0733248    .2003495
------------------------------------------------------------------------------
Warning: Arellano and Bond recommend using one-step results for
inference on coefficients

Sargan test of over-identifying restrictions:
chi2(33) =    47.09      Prob > chi2 = 0.0531

Arellano-Bond test that average autocovariance in residuals of order 1 is 0:
H0: no autocorrelation   z =  -2.17   Pr > z = 0.0301
Arellano-Bond test that average autocovariance in residuals of order 2 is 0:
H0: no autocorrelation   z =  -0.58   Pr > z = 0.5602

======================================================================

But running the same command in Stata8 you get a different output:

=====================================================================
Arellano-Bond dynamic panel-data estimation     Number of obs      =       751
Group variable (i): id                          Number of groups   =       140

Wald chi2(.)       =         .

Time variable (t): year                         Obs per group: min =         5
avg =  5.364286
max =         7

Two-step results
------------------------------------------------------------------------------
D.n          |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
n            |
LD |     .36652   .0428952     8.54   0.000     .2824469    .4505931
ys           |
D1 |   .8148182   .0547207    14.89   0.000     .7075676    .9220689
w            |
D1 |  -.5323957   .0380125   -14.01   0.000    -.6068989   -.4578926
k            |
LD |   .1355049   .0309182     4.38   0.000     .0749063    .1961034
------------------------------------------------------------------------------
Warning: Arellano and Bond recommend using one-step results for
inference on coefficients

Sargan test of over-identifying restrictions:
chi2(40) =    51.84      Prob > chi2 = 0.0994

Arellano-Bond test that average autocovariance in residuals of order 1 is 0:
H0: no autocorrelation   z =  -1.98   Pr > z = 0.0480
Arellano-Bond test that average autocovariance in residuals of order 2 is 0:
H0: no autocorrelation   z =  -0.50   Pr > z = 0.6173

====================================================================

Does that happen for you too? If yes, I wonder whether similar problems exist for other regression types...
Tewodaj
~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
Tewodaj Mogues
Dept. of Agricultural and Applied Economics
427 Lorch St. #317, Taylor Hall

----- Original Message -----
From: Mark Schaffer <M.E.Schaffer@hw.ac.uk>
Date: Wednesday, May 12, 2004 6:39 pm
Subject: Re: st: Stata7 and Stata8 are doing two very different things.

> Tewodaj,
>
> Quoting TEWODAJ MOGUES <tmogues@students.wisc.edu>:
>
> > Hi Stata list users,
> >
> > If any of you have access to both Stata7 and Stata8, I would be
> > curious to know if you have any insight why these produce different
> > results for the same commands. I haven't tried out a load of
> > different commands to see if there is always discrepancy, but given
> > my interest in dynamic panel data modelling, I tried for example
> > this and I get wildly different results:
> >
> > xtabond y x1 x2 L.x3 , lags(1) pre(x4 x5 , lag(0,1))  twostep
> > noconstant
> >
> > Please try it out on your own data, using any variables for the y
> > and the x's, and see the different output.
>
> I just tried with a nonsense regression using the abdata dataset,
> and got
> the same results in Stata 7 and Stata 8.
>
> Have you tried checking whether xtabond2 agrees with xtabond?
>
> --Mark
>
> > You don't have to know
> > much about xtabond to check whether the results are different (since
> > they should not be). After knowing what's going on here, the obvious
> > next step is to find out which of the two STata versions is "doing
> > the right thing".
> >
> > Thanks,
> > Tewodaj
> >
> > ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
> > Tewodaj Mogues
> > Dept. of Agricultural and Applied Economics
> > University of Wisconsin - Madison
> > 427 Lorch St. #317, Taylor Hall
> >
> > *
> > *   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/
> >
>
>
>
> Prof. Mark Schaffer
> Director, CERT
> Department of Economics
> School of Management & Languages
> Heriot-Watt University, Edinburgh EH14 4AS
> tel +44-131-451-3494 / fax +44-131-451-3008
> email: m.e.schaffer@hw.ac.uk
> web: http://www.sml.hw.ac.uk/ecomes
> ________________________________________________________________
>
> DISCLAIMER:
>
> This e-mail and any files transmitted with it are confidential
> and intended solely for the use of the individual or entity to
> whom it is addressed.  If you are not the intended recipient
> you are prohibited from using any of the information contained
> in this e-mail.  In such a case, please destroy all copies in
> Watt University does not accept liability or responsibility
> for changes made to this e-mail after it was sent, or for
> viruses transmitted through this e-mail.  Opinions, comments,
> conclusions and other information in this e-mail that do not
> relate to the official business of Heriot Watt University are
> not endorsed by it.
> ________________________________________________________________
>

*
*   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/
```