Examining the asymmetric information flow between pairs of gold, silver, and oil: a transfer entropy approach

The present study examines the asymmetric information flow between bivariate pairs of gold, silver, and oil daily returns for the period: 1 September 2000 to 5 May 2022 with a special emphasis on crisis periods. We use Shannon and Rényi entropy transfer techniques instead of the commonly used Granger causality approach to get robust estimates while allowing for nonlinear, non-parametric, and asymmetric relationships in bivariate returns series. The study’s findings support a unidirectional information flow from silver to gold, but not the other way around. Similarly, over the whole sample, a unidirectional information flow from oil to gold is observed but during the financial crisis of 2008–09, the flow of information is reversed from gold to oil. Furthermore, during COVID, there is no statistically significant information transfer between “gold and oil.” Throughout the study period, a bidirectional information flow between “silver and oil” is observed. During the financial crisis, however, there is no statistically significant information flow between “silver and oil.” As a result, portfolios based on “gold and oil” or “silver and oil” would be risky. Policymakers and investors should recognize dynamic interrelationships across market conditions and leverage the value of time-sensitive information flow for risk assessment and portfolios.