Granger Causality Result in Eviews

[Kalec]

Hi

I have done Granger causality test in Eviews, but I don't know how to interpret the result. I checked the User Guide with no luck.
I was thinking to check critical value for the F-test? but I cannot find the values.

If I am on the right track, can anyone please let me know how I can obtain the critical values ?

Please help.

Here are my results:

ScreenHunter_08 Aug. 17 19.46.jpg (40.82 KiB) Viewed 45167 times

ScreenHunter_09 Aug. 17 19.53.jpg (40.44 KiB) Viewed 45169 times

Thanks again
Kalec

[theologos]

Hello,

Read carefully the manual and you will find the answer! Try to think if the critical values are of interest whenever the p-values are reported. The p-value does the job.

[Kalec]

Thanks theologos

In the guide, I can see an example on Granger- causality and it does indicate whether to reject the H0, but it does not give reason for that. (unless I am reading the wrong part?)

ScreenHunter_10 Aug. 17 20.54.jpg (54.79 KiB) Viewed 45150 times

In relation to the probability, does it mean that whenever there is a probability figure in the result, it could be relied on?
So, rejection for probability < 0.05 for 95% confident level?

Thanks again

Kalec

[theologos]

Hi,

Yes correct. We fail to reject the null hypothesis whenever the p-value is greater than the 0.05 (given that this is the selected level of significance).

Regards

[Kalec]

Thanks so much theologos

Kalec

[ekta1]

Hello Sir/Madam,

Please, the Pairwise Granger test does not ask for the order of integration of the data to be specified. Is the order of integration of the data part of the computation done by the test or not?

Thank you,
Emml.

[EViews Gareth]

No.

[ekta1]

Hello,

So does it mean the order of integration of the data is not necessary in a Pairwise Granger test? Some empirical models requires the specification of both lag order (k) and order of integration (d) for test of causality, example Seabra and Flach (2005:2).

Why the difference?

Thanks, Emml.

[ekta1]

Hello,

Please if the p-value is more than 5% but the F-statistic exceeds the F-critical value, should the null be accepted or rejected?

thanks.

[joshimrunal]

Thank you for your kind cooperation