EViews.com

Hi,

I am trying to forecast copper demand for China using a linear regression model. However the problem seems to be that the model is consistently underestimating the demand and the difference with the actual demand appears to be widening over the years. How can I try to improve my model?

I have already tried using usual transformations of variables like growth rates, log, indices etc. The series that I am using is stationary.

Any leads would be much appreciated.

Thanks,
Sreeja

Sreeja,
it's hard to advise on your query with so scant details.
First off, what is your regression model? Simple or multiple OLS?

Kind regards,
Carlo