Re: st: XTPCSE drops constant. Why?

[David Jacobs <jacobs.184@sociology.osu.edu>]

If the t1 through t10 variables are period specific dummies, my best 
guess would be that the machine is keeping all of of them (save those 
where that year isn't present in this data) and dropping the 
intercept.  Suppose you specified a cross-sectional model with four 
out four regional dummies included and put "noconstant" at the end of 
the equation.  The machine probably is doing that for you in this 
instance.  Drop one of the dummies it includes in run below and see 
if the constant returns.

Dave

At 03:13 PM 3/14/2008, you wrote:

> All,
>
> It is a Friday evening so stay with me if this is a ridiculous 
> question. I am running a time effects panel data model with panel 
> corrected standard errors to account for auto-correlation. I get 
> significant results with a high r-squared and the expected sign on 
> the coefficients. But Stata automatically drops the constant term in 
> some of the models. Why is it doing this? And is it a problem?
>
> If its any help the model is meant to be showing the elasticity 
> between mortality and income, income inequality and female 
> education. Any help would be greatly appreciated.
>
> The commands and results are pasted below:
>
> xtpcse lngdata74 lngdata36 lngdata38 lngdata99 t1-t10, 
> correlation(ar1) independent
>
> Number of gaps in sample:  5
> (note: computations for rho restarted at each gap)
> (note: estimates of rho outside [-1,1] bounded to be in the range [-1,1])
>
> Prais-Winsten regression, independent panels corrected standard errors
>
> Group variable:   unit_id                       Number of obs      = 175
> Time variable:    time5                         Number of groups   = 96
> Panels:           independent (unbalanced)      Obs per group: min = 1
> Autocorrelation:  common AR(1)                                 avg = 1.822917
>                                                               max = 3
> Estimated covariances      =         1          R-squared          = 0.9293
> Estimated autocorrelations =         1          Wald chi2(6)       = 16876.21
> Estimated coefficients     =         6          Prob > chi2        = 0.0000
>
> ------------------------------------------------------------------------------
>             |           Indep-corrected
>   lngdata74 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
> -------------+----------------------------------------------------------------
>   lngdata36 |  -.6876405   .0415146   -16.56   0.000    -.7690077 -.6062733
>   lngdata38 |   .7374432     .12041     6.12   0.000      .501444 .9734425
>   lngdata99 |  -.9479843   .2120751    -4.47   0.000    -1.363644 -.5323246
>          t1 |  (dropped)
>          t2 |  (dropped)
>          t3 |  (dropped)
>          t4 |  (dropped)
>          t5 |  (dropped)
>          t6 |  (dropped)
>          t7 |   10.74782    .934501    11.50   0.000     8.916227 12.5794
>          t8 |   10.74098   .9369609    11.46   0.000     8.904567 12.57739
>          t9 |   10.74641   .9380019    11.46   0.000     8.907963 12.58486
>         t10 |  (dropped)
>       _cons |  (dropped)
> -------------+----------------------------------------------------------------
>         rho |   .5561749
>
>
>
>
> *
> *   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/

*
*   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/