Multifractal behavior of commodity markets: Fuel versus non-fuel products

Abstract

We investigate multifractal properties of commodity time series using multifractal detrended fluctuation analysis (MF-DFA). We find that agricultural and energy-related commodities exhibit very similar behavior, while the multifractal behavior of daily and monthly commodity series is rather different. Daily series demonstrate overall uncorrelated behavior, lower degree of multifractality and the dominance of small fluctuations. On the other hand, monthly commodity series show overall persistent behavior, higher degree of multifractality and the dominance of large fluctuations. After shuffling the series, we find that the multifractality is due to a broad probability density function for daily commodities series, while for monthly commodities series multifractality is caused by both a broad probability density function and long term correlations.