st: RE: How to use lags (for each group in panel) with stata commandXTPMG? (by Blackburn & Frank 2007)

["Nick Cox" <n.j.cox@durham.ac.uk>]

I think people need to see exact code to comment helpfully. For example,
it is not clear to me what you mean by string constraints (possibly,
constraints on string variables) and in any case what have they to do
with fitting a numeric model? 

I don't think that the authors of -xtpmg- are members of the list, so
good advice on its specifics may not be forthcoming unless you address
them. 

The FAQ advises, and with good reason, at

<http://www.stata.com/support/faqs/res/statalist.html#stata> 

Say exactly what you typed and exactly what Stata typed (or did) in
response. N.B. exactly! 

Nick 
n.j.cox@durham.ac.uk 

Hannu Antero Karhunen

I have a problem concerning lags with XTPMG command in stata (-findit
xtpmg). 
How can I incorporate right amount of lags to variables for each
individual panel member? What is the most efficient way to do this?

Time series command's (L.L2. etc.) affect all panel groups equally
(regions in my study), but I need to find a way to have different lag
order for each region when running xtpmg.

I already tried  to use maximum lag order  (same four lags for each
variable per region using time series command) and then utilize 
constrain() command to restrict all extra lags to zero (according of BIC
information criterion). But when I define constrains to xtpmg, program
only recognize the first constrain (it seems that normal string
definitions do not work for some reason inside constrain command..).
So I don't really know how to proceed if I cannot use constrain command
to equal extra lags to zero...

If someone has used XTPMG, it would be most helpful to hear how lags
(defined with information criterion for each group in the panel)  are
used with xtpmg.

*
*   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/