A Note on the Cliff and Ord Test for Spatial Correlation in Panel Models