The effect of uncertainty on the precious metals market: new insights from transfer entropy and neural network VAR

Abstract

This study employs a fresh perspective to investigate the causal relationship between uncertainty, measured via the two proxies of Economic Policy Uncertainty (EPU) and the Chicago Board Options Exchange Volatility Index (VIX), and precious metals prices. In this paper, we use two state-of-the-art methodologies, namely Multilayer Perceptron Neural Network Non-linear Granger causality and Transfer Entropy, to examine the data for four kinds of metals: gold, silver, palladium and platinum. We determine that gold is still the dominant ‘safe-haven’ asset to hedge against uncertainty. Interestingly, these precious metals were also observed to effect EPU and VIX, while they are immune to EPU shocks but not to those from the VIX. Moreover, this paper also contributes to understanding the specific lag period of precious metals sending shocks to EPU, whereas these shocks immediately happen with equity volatility. This paper not only contributes new quantitative methods augmented with neural networks and econophysics to a growing body of literature but also provide insights to shock transmission mechanisms in commodity markets, particularly those for precious metals.