The Efficient Markets Hypothesis is one of the fundamental concepts of modern financial theory. It states that the random behavior of prices stems from the unpredictable information flow. One of its paradigms defines the random returns distribution as normal. In this article we evaluate the hypothesis that the stochastic return generating process in the emerging markets of Asia and Latin America follows a non-normal alpha stable process. We estimate the distribution parameters through maximum likelihood estimation and simulation and present evidence that stock market returns would be better described by the alpha stable or fractal distribution. These distributions accommodate large and frequent price fluctuations as well as asymmetry, when the likelihood of large losses is greater than what would be warranted under a normal distribution. Inputs to usual financial models, such as Markowitz' mean-variance optimizers and the Black and Scholes option pricing model, may be better estimated by means of an alpha stable distribution.
fractals; efficient market hypothesis; alpha-stable distributions; extreme values; emerging markets
