VIF analysis in Panel data

[khor1984]

HI,

I tried to test multicolinearity for my panel data. Unfortunately, i am not able to find the button for VIF in EVIEW 7.

May i know if the function for VIF analysis is not available in panel data analysis? If it is so, is there any other test that can use to test multicolinearity?

Thanks..

[startz]

Out of curiuosity, why do you want to test for multicollinearity?

[EViews Gareth]

VIF is not available for panels, nor is any other multicolinearity statistic.

[khor1984]

The reason i test for multicollinearity is to check whether there is correlation between my variables.. is there any test that i can use for testing multicollinearity in panel data? or, we are assuming no multicollinearity in panel data?

[startz]

The reason i test for multicollinearity is to check whether there is correlation between my variables.. is there any test that i can use for testing multicollinearity in panel data? or, we are assuming no multicollinearity in panel data?

Why do you care whether there is correlation between your variables? There almost certainly is correlation, but that's what OLS is built to sort out.

[yasir]

I just want to say something on Gareth's comment. You said none of the tests for multicollinearity is included in Eviews. But I have been able to use the VIF for panel data. I estimate an equation, give the equation a name and then type "eq01.varinf" and it gives me the varinace inflation factor both centered and uncentered (if constant is included). Isn't that a test for multicollinearity?

[EViews Gareth]

ok, VIFs shouldn't be available for panel equations

[Cabinho]

This is probably not set in stone but can somebody explain what is the rule of thumb for uncentered and centered VIF in cross sectional data?

[franciscoA]

Can you explain a little bit this sentence. Is multcollinearity irrelevant in panel equations?

ok, VIFs shouldn't be available for panel equations

[EViews Gareth]

EViews doesn't offer VIF analysis for panel data. That's all that statement meant.

[franciscoA]

Thank you. Can you explain why? Is Eviews planning to include this test or any other tests related to multicollinearity for panel in the future? What would you recommend if I want to consider multicollinearity in the analysis I'm performing using EViews?

EViews doesn't offer VIF analysis for panel data. That's all that statement meant.

[EViews Gareth]

There was no particular reason why. The computation of VIF in panels are slightly different, that's all.

[franciscoA]

No plans for the future?

There was no particular reason why. The computation of VIF in panels are slightly different, that's all.

[EViews Gareth]

Not currently.

[saizal87]

You can actually test for multicollinearity based on VIF on panel data. lets say the name of your equation is eq01, so type "eq01.varinf" and then click enter. then you will get centered (with constant) vif and uncentered (without constant) vif. Keep in mind, if your equation dont have constant, then you will only get the uncentered.