In this paper we study the possible "chaotic" nature of some energy and commodity futures time series (like heating oil and natural gas, among the others). In particular the sensitive dependence on initial conditions (the so called "butterfly effect", which represents one of the characterizing properties of a chaotic system) is investigated estimating the Kolmogorov entropy, in addition to the maximum Lyapunov exponent. The results obtained with these two methods are consistent and should indicate the presence of butterfly effect. Nevertheless, this phenomenon - which is usually showed by deterministic systems - is not here completely deterministic. In fact, using a test introduced by Kaplan and Glass, we prove that, for all the series analyzed here, the stochastic component and the deterministic one turn up to be approximately in the same proportions. The presence of butterfly effect in energy futures markets is a controversial matter, and the evaluations obtained here confirm the findings of some authors cited in this paper. Thus, we can say with reasonable certainty that in energy futures markets we cannot talk about deterministic butterfly effect.
↧