Page 1 of 1
2sls by hand vs. 2sls EViews
Posted: Fri Aug 10, 2018 8:25 am
by joff
Attached is my replication of the example from Wooldridge textbook. I am puzzled because I get different coefficient estimates (EViews vs. by hand). If anybody has an explanation for this I would be grateful. I am sure it is something trivial but I just dont see it.
Re: 2sls by hand vs. 2sls EViews
Posted: Fri Aug 10, 2018 8:37 am
by EViews Gareth
Make sure you're using the same sample in each estimation. LWAGE has NA values so observations are dropped from the TSLS equation, but is not included in the 1stage regression, so observations are not being dropped.
Re: 2sls by hand vs. 2sls EViews
Posted: Sun Aug 12, 2018 6:07 am
by joff
I was right that it is trivial:) Thank you.