Siliverstovs and
Herzer (2006) mention that the results of Granger causality test may not be valid if a
model suffers from the omission of important independent variable. Therefore, only
the results of the multivariate framework are discussed in this conclusion.