Interplay Multifractal Dynamics Among Metal Commodities and US-EPU

Abstract

This research employs the Multifractal Detrended Fluctuations Analysis (MF-DFA) and the Multifractal Detrended Cross-Correlation Analysis (MF-DCCA) to study the components of the observed multifractality considering eight metals commodities time series (gold, silver, platinum, palladium, copper, lead, nickle and tin) and US-EPU time series. For both multifractal methods, we analyze the generalized Hurst exponent 𝑕(𝑞) and the Rényi exponent 𝜏 (𝑞), and evaluate their statistical properties, which allow us to examine separately the small scale contributing (primarily via the negative moments 𝑞) and the large scale (via the positive moments 𝑞). Also, we calculate the multifractal spectrum 𝑓(𝛼) and perform a fourth-degree polynomial regression fit to evaluate the complexity parameters that describe the degree of multifractality, considering the volatility time series and the volatility time series pairs&n&t&l& &r&0 > 0.5) and these volatility time series pairs (𝛼𝑥𝑦(0) > 0.5) are characterized by overall persistent behaviour. We discover that Gold seems to be more related with the other commodities than EPU, which is relevant for the different market agents, mainly investors, which could retrieve important information from these results, for example for diversification purposes.

Keywords: Metals commodities, US-EPU index; Multifractality; Cross-correlation; Complexity.