Multivariate correlation analysis of agricultural futures and spot markets based on multifractal statistical methods

Agricultural commodity futures are the earliest listed futures in the world. The rapid development of their markets has greatly affected the world agricultural production and circulation. Thereby, the volatility characteristics of agricultural futures markets and the cross-correlations between the futures and spot markets have attracted extensive attention from investors, regulators and researchers. Two new fractal statistical methods are firstly developed in this paper. Then, from the perspective of system theory, some multifractal analysis methods, including the two newly developed methods, are used to study the autocorrelation, cross-correlation and coupling correlation for the return series of five agricultural futures and their corresponding spot varieties in CBOT. The empirical results show that the autocorrelations and cross-correlations of both the futures and spot systems are multifractal and present different dynamic fluctuation characteristics at different time scales. The long-term co-movement between the two systems is not strong and the overall risk spillover effect is not obvious. In addition, the coupling cross-correlations are found between the corresponding components, but the strengths are distinct due to the different influences of long-range correlations and fat-tailed distributions.