Hi All,

I am using granger causality test on two variable A and B (Both becomes stationary after taking first lag) in eviews.

I used both the variables in test as it is and find out that p-values for both cases ( A does not granger cause B and B does not granger cause A) are very low in the initial lags. (there are 520 observations in both the series)

After taking 28 lags, p-value for first case (A does not granger cause B) becomes greater than 0.5 and Hence I can accept the null hypothesis. However, p-value for second scenario (B does not granger cause A) never crossed the threshold of 0.5 (till 170 lag).

Based on the above analysis, can I conclude that :

There are enough statistical evidence to state that B always causes A.
Similary, A causes B statement only holds upto a certain limit.

Please correct me, if my interpretation in incorrect somewhere.

Thanks in advance