Testing the Random Walk Hypothesis on Thinly-Traded Markets: The Case of Four African Stock Markets

This paper investigates the random walk behaviour of stock returns on four African stock markets taking into account the thin-trading effect. Worthy noting is the way returns are calculated using trade-to-trade approach and adjusted to account for the thin-trading effect. The adjustment was done by dividing each trade-to-trade return with the number of days between trades. A test was performed to find if this adjustment method is justifiable. The findings were that there is a positive relationship between the absolute trade-to-trade returns and the number of days between trades. The adjusted stock returns were also tested for normality, which was rejected on all the four markets. in testing if stock returns follow a random walk, two simple traditional testing methods, that is, the serial correlation test and the runs test were used. The findings were that almost half of the stocks on each of the four markets showed significant serial correlation. There was therefore not enough evidence to accept the hypothesis of a random walk.