Re-examining Granger Causality from Causal Bayesian Networks Perspective

TL;DR

This paper re-examines Granger causality by integrating causal Bayesian networks and Reichenbach's principles, providing a more rigorous causal interpretation and demonstrating its effectiveness through theoretical analysis and simulations.

Contribution

It introduces a novel framework that endows Granger causality with proper causal interpretation using CBNs and RCCPs, addressing longstanding criticisms.

Findings

GC can be interpreted causally under certain assumptions

The reformulation aligns GC with causal Bayesian network principles

Simulation results support the theoretical claims

Abstract

Characterizing cause-effect relationships in complex systems could be critical to understanding these systems. For many, Granger causality (GC) remains a computational tool of choice to identify causal relations in time series data. Like other causal discovery tools, GC has limitations and has been criticized as a non-causal framework. Here, we addressed one of the recurring criticisms of GC by endowing it with proper causal interpretation. This was achieved by analyzing GC from Reichenbach's Common Cause Principles (RCCPs) and causal Bayesian networks (CBNs) lenses. We showed theoretically and graphically that this reformulation endowed GC with a proper causal interpretation under certain assumptions and achieved satisfactory results on simulation.