Economics, Econometrics and Finance
4 years ago

Nonparametric test for Granger causality may need re-interpretation, study suggests.

A note on the Hiemstra-Jones test for Granger non-causality
Cees Diks, Valentyn Panchenko

Paper Summary

Related papers

Economics, Econometrics and Finance
4 years ago

Innovative Test Reveals Impact of Money on Economic Predictions

Monetary Models
Econophysics
1 view
Theme background image
Engineering
5 years ago

New study reveals groundbreaking insights into predicting causal relationships accurately.

System Identification
Regularization Methods
Multicollinearity
Theme background image
Computer Science
8 years ago

Unstable data leads to false causality in Granger test results.

Neural Networks

The Hiemstra-Jones test for Granger causality has a consistency problem. The relationship tested is not what it should be for Granger non-causality. When using processes that should show no causality, the rejection rate can increase with sample size. This means that evidence for nonlinear Granger causality needs to be re-evaluated.