Interpretation of p-values in Granger Test for panel data

[Andrea Arancio]

I'm sorry for a question which might sound naive, but I want ot be 100% sure.

I've been using the command

vargranger

if in the table I read a p-value < 0.05 I can reject H0 and say that the excluded variable could cause the variable under equation right?
p<0.05 = THERE IS Granger Causality, am I right?

Now the problem replicates on panel data where I have repeated measures for countries over time.
In this case I set the time this way

xtset Country Time

and then I perform the Dumitrescu & Hurlin test, using

xtgcause

Am I right to say that also in this case a p-value < 0.05 for Z-bar indicates granger causality?

Among the other things, this tutorial confuses me, as it seems from the last line of text, that a p-value > 0.05 indicates Granger Causality. Which is the correct way to read this?
http://www.eviews.com/help/helpintro...y_Testing.html
And also this one seems to contrast what I just wrote: http://www.eviews.com/help/helpintro...Causality.html


Thanks a lot!
Andrea

[Andrea Arancio]

Can someone help me here? It's really important..