In conventional causality testing based on asymptotic distribution theory, there is a high risk of wrongly rejecting the true null of no causality especially when the sample size is as small as typically seen in the literature. In this study, we offer a formal diagnosis of the existing...