Page 1 of 1

Impact of each Variable on Coefficient of Determination

Posted: Fri Jun 13, 2014 5:24 am
by Economists
Hello all,

I'd like to show in a basic OLS framework the effect of every single variable on the coefficient of determination.

Let's say we've got an adjusted R-squared of 50%, and a 2 exogenous variables:
What I would like to create would be something like that:

exogenous variable (1) explains 30% of total 50% R-squared
exogenous variable (2) explains 20% of total 50% R-squared.

I haven't found anything like that in the eviews guides, nor in this forum.

Can this be implemented, or programmed in any way?
Any suggestion would be very helpful und very much appreciated.

Best regards
Economists

My Eviews Version is Eviews 8

Re: Impact of each Variable on Coefficient of Determination

Posted: Fri Jun 13, 2014 5:57 am
by startz
There is no unique way to do this. Nothing to do with EViews. The issue is that covariance between the variables means there is not a unique decomposition.