COMPROBACIÓN DEL COMPORTAMIENTO CAÓTICO EN BOLSA DE VALORES DE COLOMBIA

The efficient stock market hypothesis is one of the basic assumptions of asset pricing models such as the Capital Asset Pricing Model and Arbitrage Pricing Theory and states that financial asset prices can not be predicted since they behave randomly. Conversely, the fractal market hypothesis states that pri-ces have chaotic structure and could be predicted with nonlinear models, rejecting the efficient market hypothesis and invalidating the assumptions of asset pricing models. This paper seeks to evaluate the chaotic behavior in the Colombian stock market in order to reject or accept the efficient market hypothe-sis, using tools such as price charts, recurrence plots, correlation dimensión, Hurst coefficient, Lyapunov exponent and BDS test. The results show that the Colombian financial assets reveal chaotical behavior for “bull” market subperiods and randomized for “bear” market subperiods, supporting the fractal market hypothesis. These findings may support the use of nonlinear models for predicting prices in “bull” market episodes and reject the Colombian stock market efficiency.