The exact bias of s2 in linear panel regressions with spatial autocorrelation

We investigate the OLS-based estimator s2 of the disturbance variance in an error component linear panel regression model when the disturbances are homoskedastic, but spatially correlated. Although consistent (Song and Lee, Econ. Lett. 2008), s2 can be arbitrarily biased towards zero in finite samples.

Research Highlights► The OLS variance estimator in a spatial linear panel regression is consistent. ► It can nevertheless be arbitrarily biased towards zero in finite samples. ► Examples of the extent of the bias are provided.