st: weak IV tests using xtabond2

[David Parsley <David.Parsley@owen.vanderbilt.edu>]

I am using xtabond2 (stata version 11.2) and would like to examine 1st stage regressions and test statistics. However, xtabond2 doesn't seem to report first stage regressions. Can someone point me to an explanation (online somewhere?) of how I can produce exactly the 1st stage regressions xtabond2 is performing?  Then, I guess I could use something like xi: ivreg2?

I am also looking for a (presumably) standard weak IV test as output from xtabond2 and seem not to be able to find it.  I have read the "How to do xtabond2:..." from the Stata Journal (2009) as well as other general references, but have struck out so far.

I have a panel with > 1000 cross-sections, and potentially 100 or so observations, but with missing values some members of the panel have very few observations (< 10). My STATA command looks something like:

xi:xtabond2 y x1 x2 x3 x4 l.y i.year, orthogonal  twostep robust iv(x5 x6 x7 x8) gmm(y  x3 x4)  lag(3 4))

where x5-x8 are my exogenous instruments, and I limit the lags to mitigate instrument proliferation.

David Parsley

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/