This paper considers the statistical analysis of large panel data sets where even afterconditioning on common observed effects the cross section units might remain dependentlydistributed. This could arise when the cross section units are subject to unobserved commoneffects and/or if there are...