Granger Causality in Eviews

[chrisRock]

How can I interpret the granger casuality test done in Eviews?

For example -

Pairwise Granger Causality Tests
Date: 08/18/10 Time: 20:44
Sample: 1135 1891
Lags: 22
Null Hypothesis: F-Statistic Prob.
DL_K1M_1 does not Granger Cause CV 0.72815 0.8127
CV does not Granger Cause DL_K1M_1 0.61860 0.9128

What combination of F-stat and Prob value would reject the hypothesis of no causality?
What is the interpretation of these stats?

[dia]

hi,
i thik the simple way to interpret yours results is to compare the probability values to 5%, beyond 5%, you accept the hypothesis H0:not granger causality;
in your case, your 1rst result
DL_K1M_1 does not Granger Cause CV, the p-value is 0.8127 (81%), so higher than 5%, you msut accept the H0 and conclude that DL_K1M_1 does not Granger Cause CV!
hope that will help you,

[abhi82jeet]

Can anyone tell what is the role of the F-stats in accepting/rejecting the null hypothesis? What I mean is what is the appropriate value of F-stat to accept/reject the Ho? Like in the previous post, if the Probab value is beyond 5%, the H0 can be accepted but, without considering the F-stats???

Thanks in anticipation..

[sntgo11]

The Prob. or P-values are calculated based on the F-stat. So by considering the P-values you are implicitly considering the F-stat. You can also try to find critical values for your F-stats in order to see if you accept or reject Ho.