Elaborating on the Granger Test

After reading a post by our beloved classmate Tamara Nelson, I decided to read and research the foundations of the Granger Test.

The Granger Test is an statistical Hypothesis test ( which means we deal with hypothesis) to determine whether the data of a time series can provide an accurate prediction about another.

The idea of Granger Testing relies on the assumption that the cause has to occur before the effect. For example, lets assume that we have the variables Xt and Yt presented on a time series and that we want to forecast Yt+1 using the data we have on Xt. If the test shows that there is a correlation in the results forecasted in Yt based on Xt, we can conclude that X contains information that will affect Y, implying, but not necessarily proving causation.

Granger called the signals and information contained in Xt that affected Yt "Granger causes", as they contained, as shown in a regression, an statistical significance in proving the effects of X in Y mathematically speaking. 

What do you guys think about this kind of testing ?

-Xavier