A set of variables zt is said to be caused by xt in Granger’s sense, if the information on past and present xt helps to improve the forecast of zt. Causality in Granger’s sense, or Granger causality, does not imply true causality: the former refers to the concept of predictability, and can be tested using statistical tools; the latter refers to the concept of cause and effect, and is often postulated or implied by economic theory.