Using granger causality in VAR(lags =2), i find that x granger cause y, but the sum of coefficents for lagged x, namely x_1, x_2, is insignificant by t statistic. How can we explain this result. Thank you in advance!