multiple systems gmm

[danish90]

I've having some difficulties as to how to carry out an estimation and would greatly appreciate any help in this matter.
I would like to estimate a model by using multiple systems by GMM. I am not too sure I am using the correct terminology so here is what I wish to estimate:
dec1 = constant + c(1)*market + c(2)*TERM + c(3)*DEF
dec2 = constant + c(1)*market + c(2)*Term + c(3)*DEF
dec3 = constant + c(1)*market + c(2)*Term + c(3)*DEF
dec4 = constant + c(1)*market + c(2)*Term + c(3)*DEF
The constant and coefficients c(1), c(2), c(3) must be constrained to be the same for all systems of equations. Dec1, dec2, dec3, dec4 are different portfolios. TERM is the term spread, and DEF is the default spread.
I hope I am making myself clear as to what I would like help with, otherwise please let me know and I will try to elaborate.
Thanks

[EViews Gareth]

What are your instruments?

[danish90]

For my instruments I would use my independent variables, so Def, Term and market. The article from where I'm getting this method to estimate the model sets the moment conditions as on page 11 (506) of this http://docentes.fe.unl.pt/~psc/Multifac ... sICAPM.pdf
Sorry for not writing it out, but it looked too messy when I tried.

[EViews Gareth]

Create a new System object (Object->New Object), and write out the specification exactly as you have above. Then add a line:

Code: Select all

@inst market term def

Then hit estimate and choose a GMM method.


Note I didn't look at the paper :D

[danish90]

Thanks for your help :) I did what you said, and it seems to working out for me.