----- Original Message -----
From: "Clive Nicholas" <[email protected]>
To: <[email protected]>
Sent: Thursday, November 20, 2003 9:05 PM
Subject: st: -tsset- and -xtpcse-
> Statapeeps,
>
> I wondered if any -xtperts- could help with this. :-) My cross-sectional
> dataset has n=3452, with n=6 per group (except for one group where n=7).
>
> After successfully fitting several cross-sectional FGLS regressions (both
> for fixed and random effects), I thought I'd try and run an -xtpcse-
> regression, since even the most controlled FGLS model tested 'positive'
> for serial correlation.
>
> Unfortunately, Stata refused, because:
>
> no time periods are common to all panels, cannot estimate disturbance
> covariance matrix using casewise inclusion
> r(459);
>
> Indeed it wasn't (as it was generated by [g time = _n], but that's because
> I can't use a time variable for -tsset- because Stata says:
>
> repeated time values within panel
> r(451);
>
> Naturally, a panel variable has also been included in my -tsset-. Since
> it's necessary for Stata to recognise that my dataset is a time-series
> dataset before running -xtpcse-(otherwise: "time variable not set, use
> -tsset varname ...- r(111);"), what's the solution? (I didn't pick up
> anything in -help- that answers my question!)
>
> Yours,
> CLIVE NICHOLAS,
try:
bysort group: gen time = _n
This will generate a time variable for each group. -xtreg, fe or re- do not
explicitly use the time variable so your incorrectly created time variable to
show up before.
Hope this helps,
Scott
*
* 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/
