THE CHANGING EFFICIENCY OF AFRICAN STOCK MARKETS

Abstract

This paper classifies formal African stock markets into four categories and discuses the principal characteristics of the seven markets covered in this study: South Africa, Egypt, Morocco, Nigeria, Zimbabwe, Mauritius and Kenya. Using a GARCH approach with time-varying parameters, a test of evolving efficiency (TEE) is implemented for periods starting in the early 1990s and ending in June 2001. This test detects changes in weak form efficiency through time. The TEE finds that the Johannesburg stock market is weak form efficient throughout the period, and three stock markets become weak form efficient towards the end of the period: Egypt and Morocco from 1999 and Nigeria from early 2001. These contrast with the Kenya and Zimbabwe stock markets which show no tendency towards weak form efficiency and the Mauritius market which displays a slow tendency to eliminate inefficiency. The paper relates weak form efficiency to stock market turnover, capitalisation and institutional characteristics of markets. JEL Classification: G14, G15, O16 Keywords: African Stock Markets, efficiency, GARCH WITH THE POSSIBLE EXCEPTION of the JSE Securities Exchange (JSE), there have been relatively few studies of the weak form efficiency of African stock markets. While the evidence for the JSE is mixed, the few studies that have been carried out for other markets find, not surprisingly, that most are inefficient (in a financial sense, meaning that stock prices do not reflect all available information, and that stocks are not therefore being appropriately priced at their equilibrium values). A variety of empirical tests can be used to assess market efficiency. Thompson and Ward (1995) reviewed a wide range of literature covering empirical tests of the efficiency of the JSE and noted that, with different methodologies for testing efficiency giving different results, no clear conclusion is possible. Jefferis and Okeahalam (1999a) applied unit root tests to stock price indices to assess the efficiency of the stock markets in South Africa, Botswana and Zimbabwe over the period 1989-96, and find that the South African and Zimbabwean markets were efficient during this period, although Botswana was not, at least during the early part of the period. However the unit root test of market efficiency is not a powerful one, and subsequent analysis using different tests provided contrasting results. Jefferis and Okeahalam (1999b) used an event study of the same three markets to test the response of individual stock prices to information announcements, by evaluating the speed and efficiency with which information is incorporated into market prices. They found that the Botswana and Zimbabwe markets are inefficient, while the JSE is weak form efficient. This corresponds with the findings of Smith et al (2002), who tested whether eight African stock markets follow a random walk using multiple variance ratio tests. Of the eight markets (South Africa, Egypt, Kenya, Morocco, Nigeria, Zimbabwe, Botswana and