I have a question concerning the Sargan test of over-identifying
restrictions after using xtabond. If I use the recommended one step
specification I have to reject H0 of the sargan test with p-value 0,
however if I use the two step option my p- value for the same test turns
out to equal 1 suggesting not to reject.
My panel size is not overwhelmingly large, 33 observations times 141 groups.
Which result should I rely on?
Thanks for your advise