Peter Lomas
2020-Mar-29 09:49 UTC
[R] GMM twoStep vs Stata's ivreg2 gmm2s or ivregress coefficients
Hi R-Help, I'm replicating IV-GMM models from Stata in R, but get small differences in coefficients (eg 100.0 vs 100.3); I'm hoping somebody can share anything they've learned attempting the same thing. The linear models match exactly, so there isn't a data problem. With both help files I've matched the arguments I think I can (Bartlett, bandwidth, errors), and experimented with others, but can't close the gap. Suggestions appreciated if people have knowledge of Stata's methods in R- I'm not sure if this is something that should be expected from numerical solvers. Thanks again Peter [[alternative HTML version deleted]]