Time-Varying and Quantile-Based Relationship among Geopolitical Risks, Oil and Gold Prices

Abstract

We examine the relationship among geopolitical risks (GPR), WTI oil, and gold prices utilizing the time-varying causality and quantile regression approaches. The sample period spans from January 1986 to January 2022, comprising 433 monthly observations and representing the longest common period of data availability. The results show that there is no causality between the pairs of GPR–WTI, and GPR–gold prices for the full sample period, while the causality between gold and WTI is unidirectional, running from gold to WTI. Using the rolling causality test, however, we find that the dynamic causal relations strengthen over time. The Granger causality from the gold prices to GPR and WTI is stronger than the other way around, suggesting that the gold market dominates the other two variables in terms of strength of the lead-lag structure of causality. Besides, we observe the strongest causation effects between GPR and WTI spot prices. Before 2009, the causal relationship between WTI and GPR is mostly unidirectional while also a bidirectional linkage emerges, coinciding with the crisis periods including the Dot-Com and 2007 US Subprime crises. During the causal periods, these variables respond negatively to changes in others. For the COVID19 period, the direction of causality considerably changes in favor of WTI for the GPR–WTI pair whereas it is unchanged for the WTI–gold pair. The results indicate that WTI has positive and negative predictive powers for GPR and gold while it receives negative and positive causation effects from GPR and gold during the pandemic, respectively.

Keywords

• time-varying causality
• GPR
• WTI
• gold
• quantile-regression