[Gretl-users] Estimate dynamic panel regressions