Nonparametric regression and causality testing: A Monte-Carlo study

In this paper we propose a new procedure for causality testing using nonparametric additive models. We argue that the major advantage of our proposed method is that it can be used if the underlying data generation process (DGP) is either linear or nonlinear. Our results show that the nonparametric testing procedure provides a more robust rest of causality. Furthermore, we show that the loss of power associated with the nonparametric procedure is minimal if the true DCP is linear.