Wednesday, August 31, 2011

From Correlation to Granger Causality

A rough draft of a new paper so don't cite it but comments are welcome. I have posted a couple of sections already on the blog. It's written for a workshop where I meant to present on "From correlation to causality" and I think I need to talk about both major approaches - instrumental variables and Granger causality. Then I do a Granger causality analysis of a bunch of variables with 150 year time series from Sweden to illustrate the points but also to contribute to the literature on energy and growth. The results support causality from energy to output but that relationship appears to have weakened or reversed in the second half of the 20th Century. Or maybe it is just harder to find anything significant in a shorter time series. Energy prices have a more significant impact on GDP than energy quantity in a model of GDP, energy price, and energy quantity, but again the relationship seems to weaken.

No comments:

Post a Comment