How to interpret Variation Inflation factor of $1$, when tolerance is equal to $1$ and there are only two predictors.

52 Views Asked by At

So, I need to determine multicollinearity of predictors, but I have only two. So, if VIF = $\frac{1}{1-R_j^2}$, then in case there are no other predictors VIF will always equal $1$? So, maybe it's not even possible to gauge multicollinearity if there are only two predictors?

1

There are 1 best solutions below

0
On

Let's say you have $Y$ and two predictors $X_1$ and $X_2$, then the $R_j^2$ in the VIF index is $R_1^2$ obtained from the model $X_1 = \alpha_0 + \alpha_1X_2 + \xi$. That is, unless $X_1$ and $X_2$ are uncorrelated, $R^2_1 > 0$, hence VIF $>1$.