Granger Causalty question

[Georgios]

Hi guys ,

I am currently working on my dissertation ''correlation of oil prices with agricultural commodities''.
I started with ADF tests, continued with cointegration(Joahansen) then with Vector error correction and everything seemed ok for the moment.
However when I ran the gRanger causalty test it showed me ridiculous results.It always shows that the agri products granger cause the oil and not the opposite.I've tried it thousand times with several different values for lags and even changed the to a longer run of data.However the result is always the same.In some occasions the oil seems to granger cause the agri products for 10% significance but still the agri products granger cause oil for 1% significance.

I have found in the literature that as we all expected the agri prices are moved by the movements of oil but apparently my results do not agree and I am desperate at the moment.

Any ideas/suggestions????