A review of the Granger-causality fallacy

By Mariusz Maziarz. Journal of Philosophical Economics, 2015

150 citations recorded by the source corpus.

Methods used to infer causal relations from data rather than knowledge of mechanisms are most helpful and exploited only if the theoretical background is insufficient or experimentation impossible. The review of literature shows that when an investigator has no prior knowledge of the researched phenomenon, no result of the Grangercausality test has any epistemic utility due to different possible interpretations. (1) Rejecting the null in one of the tests can be interpreted as either a true causal relation, opposite direction of the true causation, instant causality, time series cointegration, not frequent enough sampling, etc. (2) Bi-directional Granger causality can be read either as instant causality or common cause fallacy. (3) Non-rejection of both nulls possibly means either indirect or nonlinear causality, or no causal relation.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/46d7d2b630aa52c4e1ad1fb4676595c9958882dd

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Air transport and economic growth: a review of the impact mechanism and causal relationships (2020, background): “If both A and B are driven by a common third process with different lags, one might still fail to reject the alternative hypothesis of Granger causality and so forth to produce misguiding results (Maziarz, 2015).”
Granger Causality: A Review and Recent Advances (2021, background): “The above requirements were discussed in Granger’s original and followup papers (Granger, 1969, 1980, 2001) and extensively by other authors (Stokes and Purdon, 2017; Maziarz, 2015); see also the recent review by Glymour et al. (2019).”