Pseudo-Panel Data

[Jeta Statovci]

I am working with repeated cross-sectional data, 2000-2005, and want to use a pseudo-panel approach. I want to create clusters with individuals from the same region (variable reg).
Can someone help me with the commands to create clusters, and how to make state recognize the clusters when running the regression with other variables as well?

[Carlo Lazzaro]

Jeta:
assuming that what follows makes sense from a methodological point of view, you may want to try:

egen new_panelid=group(panelid reg)

[Jeta Statovci]

Thank you Carlo