I am trying to detect multi-collinearity in my regression model.I ran the Belsley, Kuh, Welsch Variance-decomposition and i saw that for k=607 there is collinearity between variables x1 and x5 where x5 is an interaction variable between x1 and x4.
Isn't it normal to have collinearity between a variable and the interaction of this variable with another variable? I am not sure whether i should eliminate the interaction variable or keep it in my model. I appreciate any advice related to this matter