Page 1 of 1
GMM and 2SLS
Posted: Mon Jan 27, 2014 7:05 pm
by ev2889
I am using the 2-stage least squares method to control for endogeneity problem. Because GMM can also resolve this problem and control for heteroskedasticity and serial correlation, I re-run the data using GMM. But it turns out the 2SLS and GMM results are exactly the same. Am I setting certain GMM features incorrectly? Please let me know.
EV
Re: GMM and 2SLS
Posted: Mon Jan 27, 2014 7:58 pm
by EViews Gareth
Which GMM weighting matrix are you using?
Re: GMM and 2SLS
Posted: Tue Jan 28, 2014 1:25 am
by ev2889
I chose 2SLS as weighting matrix. Is it the reason why I got the same results??
EV
Re: GMM and 2SLS
Posted: Tue Jan 28, 2014 8:04 am
by EViews Gareth
Yes.
Re: GMM and 2SLS
Posted: Tue Jan 28, 2014 6:15 pm
by ev2889
Are there any details about choosing the weighting matrix in GMM? I could not find any in Illustrated EViews Manual!
Thanks
Re: GMM and 2SLS
Posted: Tue Jan 28, 2014 6:19 pm
by EViews Gareth
User Guide II has technical details. For econometric advice, consult your econometric textbook.