The empirical analysis for fractal features and long-run memory mechanism in petroleum pricing systems

The fractal behaviour for petroleum pricing is investigated in several international systems. Based on the time series of Brent & WTI crude oil and Rotterdam & Singapore Leaded gasoline prices (daily spot), this paper analyses the fractal features in the systems under study by using Rescaled Range analysis (R/S analysis). This paper also estimates the Hurst exponents; thus, long-term memory effects are evaluated by these systems. By tracing the evolutionary tracking of H(τ) vs. τ, three phases are divided by different system dynamic behaviours. Furthermore, V statistical method is used to obtain the lengths of non-periodic cycles of the long-term memory. Finally, this paper presents several conclusions and suggestions.