Spatial Econometrics - Bias?


New Member
Dear all,

I seem to be lost with the following problem in spatial econometrics:

Assume the true model is a spatial error model (SEM). It is well known that using OLS for estimation is inefficient but still unbiased.
Now, when I compare my OLS vs. SEM estimates in R, say, they are quite different.
But, knowing that OLS is unbiased, why not use the OLS estimates and just correct the standard error?
I would be very grateful for any hints that help me understand why OLS and SEM estimates differ (and not just their standard errors).

Thanks a lot!