الخصائص العشوائية لمؤشرات الأسعار و إمكانيات التنبؤ بالعوائد في أسواق الأوراق المالية الخليجية

الملخص EN

This paper examines the behavioral characteristics of stock price and the predictability of returns in all of GCC Stock Markets.

The data consists of daily stock prices indices from January 2003 to October 2012.

The individual stochastic investigation is conducted by means of the unit root tests and test for autocorrelation of returns.

The Autoregressive models of returns are also applied to examine their predictability.

The results show that GCC Stock Markets have a high extent of risk associated with increased returns, characterized by nonlinear dynamics.

As to the unit roots tests, it seems that most of stock market indices under scrutiny (except Kuwait, Qatar and Mascot) contain stochastic trends, thus are non-stationary in level, but they tend to be stationary in first difference, indicating that the majority of them are only weak form efficient in level.

Besides these tests, it was found significant autocorrelation coefficients among the daily return series in most of stock markets (except Dubai), suggesting that they do not follow a random walk.

In the same context, it was found a strong relationship among actual and past returns in most stock markets (unless Dubai).

This evidence implies that returns may be predictable in short term on the basis of past information for returns.

Key Words: Random Walk Hypothesis, Weak Form of the Efficient Market Hypothesis, Return's Predictability, Unit Root Test, Autocorrelation Test, AR Models, GCC Stock Markets.