Granger causality

[haraldweber]

Hello everybody,
I have got a question regarding the Granger causality test for a VECM in Eviews. I have learned from the literature that there can be spurious Granger causality in case one uses the Granger causality test for a model with more than two variables which contain unit roots. However, this literature is about VAR models. Does this problem also exist when using the VEC model?
Moreover, according to the literature there do exist complex procedures to avoid the problem of spurious Granger causality due to unit root variables in a multivariate VAR (VEC?) model. Does Eviews by default use such a procedure or will results obtained from Eviews be biased for multivariate models containing unit root variables?
Thank you very much in advance!
Kind regards