Confusing R2 number

#1
Hello,

I'm currently working on my own leading indicator index on inflation and want to find economic time-series that have the highest r2 to the CPI index with a lag. My question is related to the very low r2 number of 0.03 for ISM Manufacturing Prices Index when regressed against CPI. I look at the two times series together in a chart and they seem to follow each other reasonably close and qualitatively I assume manufacturing prices would influence the overall CPI.

I normalized the data by transforming them both into log year over year percent changes . If I only normalize the CPI index and keep the manufacturing index as is, the r2 is 0.82 but the Breusch Pagan test shows Heteroscedasticity.

I'm hoping someone can shed some light on this as I am a novice when it comes to statistics and am still learning but this result has confused me.

Thank you
Matt
 

Attachments