Hi all,

This is my first post, so apologize in advance for any posting mistake I may make.

I am trying to run a difference-GMM model using xtabond2 based on the following columns (4) and (6) in Carol Newman, John Rand, Finn Tarp, Nguyen Thi Tue Anh, Exporting and Productivity: Learning from Vietnam, Journal of African Economies, Volume 26, Issue 1, 1 January 2017, Pages 67–92, https://doi.org/10.1093/jae/ejw021:

Array

My issue is with the use of instruments in their note, where they state that "...L.lnva [log value-added], L.export, lnlab [log employment] and lncap [log capital] are treated as endogenous. The third and fourth lags are used as instruments for L.lnva and lnlab in first differences, the fourth and fifth lags of lncap, and the fifth lag of L.export (and of HHI4 × L.export in column 6)...Selection controls and ownership dummies are treated as exogenous." Does that mean my model specification in xtabond2 should look like this for column (4):

Code:
 xtabond2 lnva L.export lnlab lncap L2.export L2.lnlabprod L2.lnlab L2.cap-lab L.lnva soe fdi yr* ind1*, robust gmm(lnva L.export lnlab lncap) iv(D.L3.lnva D.L4.lnva D.L3.lnlab D.L4.lnlab L4.lncap L5.lncap L5.export L2.export L2.lnlprod L2.lnlabor L2.caplab yr* ind1*) noleveleq small
where soe and fdi are ownership dummies for state-owned and foreign-invested enterprises, respectively, and yr* and ind1* are time and 4-digit industry dummies.

I haven't run xtabond2 before, and would truly appreciate your help on the matter.

Thank you for your time,
Phan