- Granger causality is based on the idea that the cause happens before the effect, so a variable can be said to Granger-cause another variable if it can be shown, usually through a series of F-tests, that lags of the first variable help predict the second variable. - There can be no causality between two variables, one-way causality where one variable Granger-causes the other, or two-way/feedback causality where each variable Granger-causes the other. - Granger causality measures precedence and predictive power but does not necessarily imply true causality in the philosophical sense. It is a statistical notion of causality, not a factual one.