This paper examines the consequences of model misspecification using a panel data model with spatially autocorrelated disturbances. The performance of several maximum likelihood estimators assuming different specifications for this model are compared using Monte Carlo experiments. These include...