Testing for contemporaneous correlation of disturbances in seemingly unrelated regressions with serial dependence

This paper develops a multivariate independent (MI) statistic which can conveniently and powerfully test the independence between the disturbances of a seemingly unrelated regressions (SURE) model. The implementation of our test is straightforward and can be carried out by standard statistics packages. As compared to the results in Tables 2 and 3 of Dufour and Khalaf [Journal of Econometrics 106 (2002) 143], our MI test performs reasonably well, even though the sample size is only 25