Dear list members,

I have a special request estimating a dynamic spatial durbin model.
I would like to estimate something of the form :

y_{t} = alpha + tau y_{t-1} + psi W y_{t-1} + rho W y_t + epsilon_t

As you can see, [beta y] must equal 0 for this model to have any sense.
Which part of the function could I modify to estimate something like this ?

Gratefully
Best Regards

Lisa