Correlation Matrix and VIF

[Aulia Irfan Mufti]

Hello everyone, sorry to bother you all.
I got some question about multicollinearity.

Since my model using Fixed Effect Model, I have test the multicollinearity using syntax pwcorr for all the independent variable. There is no a single one variable that exceeds the 0.9 limit. But when I did Vif, Uncentered. It turns out that some of the variable have multicollinearity problem since its vif score exceeds the 10 limit. Those variables basically an logharitm natural variable.

So is it possible that something like this happened ? Or should I just ignore it.?

Regards,
Aulia

[Joro Kolev]

Just ignore it and proceed to whatever estimation you have in mind.

Here is a discussion of the "(non-)problem of multicollinearity":

https://www.statalist.org/forums/for...ical-variables

[Carlo Lazzaro]

Aulia:
see also Chapter 23 of https://www.hup.harvard.edu/catalog....40&content=toc.

[Aulia Irfan Mufti]

Dear Carlo and Joro, Thanks a lot. That helped me.