A Graphical Approach to Causality in Econometrics
Kevin Hoover - AE
Building on the work of Sprites, Glymour, and Scheines and of Pearl, we show how causal relationships among economic variables can be represented graphically. Graphical representations can be related to classic identification issues in econometrics. We show how search algorithms can uncover evidence for the underlying causal structure. These techniques are illustrated with applications to the identification of the contemporary causal ordering of structural autoregressions.
For information contact Jacopo Mazza or Adrian de Groot Ruiz.