Autosimilarty, Long Memory and Chaos: Evidence from the Tunisian Market

Abstract

Fractal Finance came to the rescue of the classical models unable to explain financial anomalies and of linear models inadequate to characterize complex processes. The characterization of financial series is still topical. The calculation of the Hurst exponent, the fractal dimension, the Lyapunov exponent, the window of Theiler and the realization of the determinism test, have allowed us to understand the dynamics of the Tunisian indexes returns. Clearly, findings show that the returns are, on the one hand, nonlinear, follow alpha-stable laws, have a long memory and on the other hand, are not chaotic. Thus, the hypothesis of a Brownian fractal motion is privileged.