Re: st: xtpcse - "no time periods are common to all panels"

[Gordon Hughes <G.A.Hughes@ed.ac.uk>]

Frequently and I suspect that you are misinterpreting the 
message.  What it means is there is no set of years with data for all 
120 countries.  The fact that "many" countries have data is not 
sufficient because, given your description, it is necessary that 
every country should have data for at least some of the 3 years that 
represents the minimum length of data across panel units.  Probably, 
even one year is not enough, because the routine is trying to 
calculate an m x m covariance matrix for the errors across panel units.

The simple solution is to use the "pairwise" option which relies upon 
the pairwise covariances between panel units.  Even then you should 
consider dropping the panel unit(s) with very small numbers of observations.

Gordon Hughes
g.a.hughes@ed.ac.uk

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


I have panel data for about 120 countries, unbalanced (the number of 
years ranges from 3 to 20). I set it up with "xtset panel year." But 
when I try to run -xtpcse- I get this error message: "no time periods 
are common to all panels, cannot estimate disturbance covariance 
matrix using casewise inclusion." But I know that that the panels do 
include common years. Most of the panels include the same 15 - 16 
years. Does anyone have experience with this problem?

Thanks,

Wayne


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