The article concludes there is no difference except the ones outlined, you pointed out that sub economies have differing results. To have no difference in population but significant difference in sub population thw variance needs to be opposite enough to cancel out significance just by how the math works.