st: xtabond2 with static model?

[Agnes Sadowski <agness@uni-bonn.de>]

Dear stata users,

I have the following questions and would be very grateful if someone could help me:

I have a panel dataset and want to do a regression were one of my explanatory variables is probably endogenous, more specific there could be a problem of reverse causality.
Since there are no appropriate instruments, I use a within estimation (fixed effects).

But if I understood correctly, this wipes out only an endogeneity problem caused by omitted variables rather than the reverse causality problem (since the causality problem could also come from the times series, but fixed effects only eliminate cross section heterogeneity?!)

Therefore, I thought to do an estimation with Arrelano Bond, more specific xtabond2. So far, I have only seen this estimator in context of dynamic panels. I also did my estimation with the lagged dependent variable and indeed the lagged variable is highly significant and my variable of interest is less so. Does this mean that my model was wrongly specified before (since it was static) or could it be since xtabond2 also accounts for endogeneity in other variables and therefore, my other endogenous variable is not significant anymore?

Furthermore I was wondering if I could use xtabond2 with my static model- at least technically it is possible with xtabond2 to do an estimation without the lagged dependant variable. When I ran this estimation, all my estimates are as expected (right signs, hoghly significant) and there is no autokorrelation of 2nd order  - everything seems plausible, but is this a reliable estimation? Does it make sense to use xtabond2 without any lags?


I would be really happy if someone could help me, I am pretty new to stata and panel data analysis, but I need to solve this issue for my thesis. 

Thanks in advance!!!!

Anna
*
*   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/