I am estimating a model using Arellano Bond estimator (xtabond2). Everything
works fine. I fail to reject the null hypothesis for the overidentifying
restrictions test. Howerever I wonder if there is a routine or a way to get
what would be the first stage regressions.
My question is motivated beacause, in general the Arellano Bond estimator is
a procedure that tells how to select instruments for the endogenous
variables, an in doing so biud a huge instrument matrix. I wonder how
powerful the OIR test might be.Specifically, if there are not enough
relevant instruments in the "first stage", the OIR may not be powerful. Am I
wrong?
