Relationship may often be correlational between two variables. The word impact somehow implies a causality, i.e., one variable causing the other to happen or change in a certain manner. For a variable to be the cause of another variable three conditions must be met: (i) cause and effect are correlated, (ii) cause occurs before effect in time sequence, and (iii) the observed relationship between the cause and effect variables cannot be explained by a third variable.
Two variables can in general possess a relationship, however if one variable "impacts" the other then there is a sense of time ordering. If one variable assumes a numerical value before the second variable then it can have an "impact" on the second variable.