Acceso usuarios
The martingale hypothesis is tested for two Bulgarian stock price indices and eight stock prices using finite-sample variance ratio tests in a rolling window. The data cover the period beginning in October 2000 and ending in August 2012 and are corrected to remove the effects of infrequent trading. The rolling window captures short-lived predictability and tracks the evolution of stock market predictability. There are successive periods when returns are predictable and then not predictable. This is consistent with the adaptive markets hypothesis, not the efficient markets hypothesis. Overall, returns are more predictable in times of crisis.
