Resumo

PARISI, Franco; ESPINOSA, Christian  e  PARISI, Antonino. Pruebas de comportamiento caótico en índices bursátiles americanos. El trimestre econ [online]. 2007, vol.74, n.296, pp.901-927.  Epub 20-Nov-2020. ISSN 2448-718X.

This article validates the chaotic behavior in the Argentinean, Brazilian, Canadian, Chilean, American, Peruvian and Mexican Stock Markets using the Merval, Bovespa, S&P TSX Composite, IPSA, IGPA, S&P 500, Dow Jones Industrials, Nasdaq, IGBVL and IPC Stock Indexes respectively. The results of different techniques and methods like: Graphic Analysis, Recurrence Analysis, Temporal Space Entropy, Hurst Coefficient, Lyapunov Exponential and Correlation Dimension support the hypothesis that the stock markets behave in a chaotic way and rejected the hypothesis of randomness. Our conclusion validates the use of prediction techniques in those stock markets. It’s remarkable the result of the Hurst Coefficient Technique, that in average was of 0.75 for the indexes of this study which would justify the use of Arfima models among others for the prediction of such series.

Palavras-chave : teoría de caos; análisis de recurrencia; entropía de espacio temporal; coeficiente de Hurst; exponente de Lyapunov; dimensión de correlación; prueba BDS.