[Openspace] spatial Durbin model: Selecting the lag of selected explanatory variables

[Ram Pandit]

Hello everyone,

I am trying to compare spatial models (Lag,Error,Spatial Durbin). In the
case of Spatial durbin model, i can generate and test the lags of all the
explanatory variables.  My problem is: i have a variable with  square and
cubic terms and i think it is reasonable just to test the lag term of the
LEVEL form not of the squared and cubic. Is it methodologically appropriate
to select the lags in this way? if so, please suggest how?. I am using
latest version of R.  Also, can you specify the lags for theoretically
important variables not all? if we can, please suggest how in R.

Thank you for your help in advance.

Ram