A reappraisal of the chaotic paradigm for energy commodity prices

Energy commodity prices have been examined over the last 20 years to detect the presence of chaos as an alternative to stochastic models, with contrasting results. In this paper, we wish to reassess the chaotic paradigm in the light of two new pieces of information with respect to the literature: the appearance of noise-aware estimation methods for the correlation entropy and the availability of longer time series. Our analysis shows that the literature has heavily underestimated the presence of noise, and that chaotic characteristics coexist with stochastic ones in the time series of prices. Through the recurrence quantification analysis, we have also observed the presence of intermittency, where periods of regular behaviour are replaced by periods of chaotic behaviour, which could explain the emergence of bubbles.