Page 1 of 1

Panel data correlation

Posted: Sun Jun 02, 2019 10:11 am
by wstrybol
Dear All

I want to study the effect on ROA for a pharmacy that has an online pharmacy department compared to pharmacies that don't.

This is my regression:

Return on assets = b0 + b1*ln(total assets) + b2*ln(firmage) + b3*leverage + b4*current ratio + b5*dummy online pharmacy + b6*dummy online pharmacy*ln(total assets) + e

However, the correlation between 'dummy online pharmacy' and 'dummy online pharmacy*ln(total assets)' is around .995...
Is this a problem (and do I need to let the last variable out (dummy*ln total assets)?) or is this a normal thing? All other variable are not really correlated (around .3 max).

Above that, can I check VIFs when using panel data in Eviews 9?

I'm working with panel data that runs from 2009-2017

Thx in advance :-)