Dear All,
As we seem to be in a question-answering mood this afternoon, I am re-sending an earlier message: I am running a panel on STATA (this is a 1997-2006 sample of 980 firms), but...
1) I seem to be loosing many, many values when I do the log and lag transformations. Does any of those procedures (especially the lag one) may cause problems with panel data in STATA?
2) I run some base regressions before running a the Arellano-Bond (AB) GMM estimator, my aim. Is the xtreg specification
-xtreg g_loanst g_loanst_l rrc rrc_l assets capt liqu rrc_assets rrc_liqu rrc_capt rrc_l_assets_l rr c_l_liqu_l rrc_l_capt_l tbf_l tde_l reer gdp cpi
where a post l (_l) indicates a lag generated by a formula like
-gen g_loansc_l = g_loansc[_n-1]
and pre c or g (g-, c-) a difference generated by a formula like
-g_loanst = log_loanst-log_loanst[_n-1]
equal to the AB formulation below?
-xtabond g_loanst l(0/0).(assets capt liqu gdp reer cpi) l(0/1).(rrc rrc_assets rrc_liqu rrc_capt tde tbf), lags(1)
3) Finally, when I run the AB specification above, I get the error message below
-onestep-naive VCE is not symmetric
Any obvious reason for that?
Best,
Lucio.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
