Dynamic complexity and causality of crude oil and major stock markets

Abstract

This paper investigates the causality and complexity evolution of two major crude oil returns and eight representative stock returns globally. Utilizing multiple methodologies originated from information theory and physics, i.e., the Kernel method Granger Causality Index and the Transfer Entropy, this study aims to provide new insights into the interrelationship between oil and equity markets. The study also applies the Sample Entropy to quantify the dynamic complexity of each individual market. Overall, nonlinear bidirectional causal relationships are found to be present and the corresponding information transfers between oil and stock markets are reported. Applying rolling windows and sub samples, information transfers are found to be generally time-varying, with the most common pattern featuring more significant and stronger information flows during the period of financial crisis. Moreover, the WTI and the Brent crude oil prices are not playing identical roles in their interactions with multiple stock markets either statically or dynamically. In future research, this study can be extended to consider delayed information transfers and to incorporate sectoral stock price information.