[Openspace] spatial 2-SLS in R

[Roger Bivand]

On Mon, 19 Nov 2007, German Muchnik Izon wrote:

> Hi,
>
> Does anybody what instruments R uses when you run stslsR (robust 2-SLS)? Are 
> the explanatory variables the default instruments? Can you choose the 
> instruments?

(The robustness is only in the eye of the beholder - if the model is 
misspecified, it really will not help you make a better judged call on the 
null hypothesis because the assumptions for making the "robust" adjustment 
will most likely not be met; however, I know that supervisors and 
referees often ask for robust SE, it's their prefered topping).

If you say:

library(spdep)
?stsls

you can see under "Details" that:

"     The fitting implementation fits a spatial lag model:


                        y = rho W y + X beta + e


      by using spatially lagged X variables as instruments for the
      spatially lagged dependent variable.
"

(Alternatively, the help page can be read in spdep.pdf on your chosen CRAN 
mirror).

So this implementation does not permit the user to change the instruments. 
It might however be sensible to extend the current function to take an 
optional formula interface that would permit the user to specify the 
instruments directly.

Roger

>
> Thank you,
> German.
> _______________________________________________
> Openspace mailing list
> Openspace at sal.uiuc.edu
> http://sal.uiuc.edu/mailman/listinfo/openspace
>
>

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no