Testing Weak-Form Efficiency of the Russian Stock Market

The existence of weak-form efficiency in the Russian stock market is examined for the period 1st September 1995 to 1st May 2001 using daily, weekly and monthly Russian Trading System index time series. Several different approaches are used to assess the predictability of the RTS index time series. Unit root, autocorrelation and variance ratio tests are conducted for the null hypothesis of a random walk model. The results support the null hypothesis for the monthly data only. Further analysis is performed for the daily and weekly data. Linear and non-linear modelling of the serial dependence is conducted using ARIMA and GARCH models estimated on the in-sample period 1st September 1995 to 1st January 2001. Forecasts based on the best fitting models are performed for the out-of-sample period 2nd January 2001 to 1st May 2001. Comparisons of the forecasts reveal that none of the models outperforms the others, and the most accurate forecasts are obtained for just the first out-of-sample observation. Whilst our research results provide some limited evidence of short-term market predictability on the RTS, there is insufficient evidence to suggest that it would lead to a profitable trading rule, once transaction costs and risk are taken into account.