st: Independent variable lag length selection

[Clive Gilbert <clivegilbert@gmail.com>]

Dear all

I am running several time series panel data regressions to measure
change in 5 dependent variables representing important institutions in
the political economies of 14 countries in relation to a number of
independent variables. In order to capture the effects, I’m using
lagged independent variables. However, because of the number of
possible permutations on variables, I’m conscious of generating
results which seem to fit my theory but are otherwise misleading.

Are there any techniques which can help justify lag length selection
for independent variables on methodological grounds?

Many thanks,

Clive

*
*   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/