interpreting VAR and generalized impulse responses

[loooooo]

Hi all,

Suppose two variables, X and Y, both of which are I(1) variables but not cointegrated. Hence, I difference them, and specify a VAR(p) model, where p can be chosen by AIC or SIC, if you'd like.
Since I'm interested in the interaction between them, I'd ignore responses of each to a shock to itself - I only look at the two off-diagonal pictures in the GIRF figure that Eviews generates.
The GIRF reports that X and Y are contemporaneously associated ((+)-ive), with X affecting Y through the first lag ((+)-ive) while no lagged effect of Y on X.
The contemporaneous relationship is not too hard to interpret I think, but how would you interpret the response of Y to yesterday's shock to X (or a shock to X affecting the next day's Y value - whichever way you'd like)?
Does this mean a lead of X over Y, despite the contemporaneous association?

Thanks,
Jason