Abstract
The article studies the conceptions of causality encountered in macroeconometrics. From an empiricist standpoint, it is natural to privilege measurements of phenomenal associations, spread over a period of time, between series of data. This method has made a major comeback since the 1970s, whereas these years seem to be moving away from the established approach to research in causality in this area. The authors seek to establish in particular the type of causality underpinning the approaches put forward by Clive Granger and Christopher Sims. They show that this definition of causality has distinctive advantages for the study of phenomena in which forecasting, different sets of information, interaction between the behaviour of developers of modelization tools and that of the other agents play a determining role.
Get full access to this article
View all access options for this article.