Hi everyone,

I am looking for your suggestion about how to deal with endogeneity bias which comes from including spatially lagged dependent variable (Wy) in the spatial model?
The XSML command is used to estimate spatial econometric models by Quasi-Maximum likelihood estimation. However, I am not sure that whether this estimation method can solve the endogeneity problem or not? Another way to deal with this problem is two-stage least square (2sls), but I am still not sure that this way can address endogeneity bias and how to select the proper instrumental variables. Please consider my queations. Thanks in advance.

Paul.