This paper by M. Hashem Pesaran focuses on testing the hypothesis that errors in a panel data model are weakly cross-sectionally dependent (CD). The paper introduces the exponent of cross-sectional dependence, which measures the degree of cross-sectional dependence. It argues that in large panels, the null hypothesis of weak dependence is more appropriate than the null of independence, which can be restrictive. Using Monte Carlo experiments, the paper demonstrates that the CD test has the correct size for values of the cross-sectional exponent in the range [0, 1/4], regardless of the combination of N and T, and as long as there are no major asymmetries in the error distribution. The paper also discusses the small sample properties of the CD test and its power characteristics. The CD test is shown to be consistent for values of the exponent greater than 1/2, with power rising with N√T. The paper concludes by discussing the application of the CD test to heterogeneous dynamic panels.This paper by M. Hashem Pesaran focuses on testing the hypothesis that errors in a panel data model are weakly cross-sectionally dependent (CD). The paper introduces the exponent of cross-sectional dependence, which measures the degree of cross-sectional dependence. It argues that in large panels, the null hypothesis of weak dependence is more appropriate than the null of independence, which can be restrictive. Using Monte Carlo experiments, the paper demonstrates that the CD test has the correct size for values of the cross-sectional exponent in the range [0, 1/4], regardless of the combination of N and T, and as long as there are no major asymmetries in the error distribution. The paper also discusses the small sample properties of the CD test and its power characteristics. The CD test is shown to be consistent for values of the exponent greater than 1/2, with power rising with N√T. The paper concludes by discussing the application of the CD test to heterogeneous dynamic panels.