An empirical investigation of the random walk hypothesis of stock prices on the Nairobi Stock Exchange

Abstract

Stock exchange markets play a critical role in their economies. They facilitate the movement of capital, often aggregating resources of small individual savers into sufficiently large capital sums that can be successfully invested by commercial companies and at the same time provide opportunities for investors to generate returns. Fluctuation in this markets influence personal and corporate financial live and economic health of a country. The debate on random walk hypothesis has been pointed out as dealing with whether or not security price fully reflect historical prices or returns information.

This study empirically investigates whether or not stock prices at Nairobi stock exchange follow a random walk model. Previous studies have been inconclusive and produced varying and conflicting findings. The key question investigate is whether successive share price returns on the Nairobi stock exchange independent random variables so that the price return cannot be predicted from historical price returns. This study employed serial correlation tests and runs tests to analyze daily price returns for eighteen companies whose stocks constituted the NSE 20 share over the period July 2008 to June 2011.

The findings suggest that random walk model cannot be a good description of successive price returns at the Nairobi stock exchange. This is analogous to the earlier evidence by Parkinson 1984 that random walk model does not apply on Nairobi stock exchange. The findings are contrary to earlier evidence (Dickson and Muragu; 1994, Githiga; 2008 and Anyumba; 2010) that random walk model was a good description of successive price returns at the NSE. The results obtained are contrary to the hypothesis that successive stock prices are independent random variables and also not consisted with efficient market hypothesis.