Do we conduct a cross sectional dependence test for panel data with N being listed firms and T being financial years. Is it only conducted when the data has geographies involved? Please share a reference for either answer.