Stock market indices: testing for mean revertion and bias of Nairobi Stock Exchange indices

Abstract

This study uses the Pearson correlation to test whether the Nairobi Stock Market is mean reverting and the implications of such mean reversion for security valuation. To carry out this test the study first determines the bias of the Nairobi stock Indices and goes,further to test the mean reversion of the market. The study was wholly based on secondary data filed with NSE secretariat The result shows a short run bias of the NSE 20 Share Index that disappears in the long run. The graph (l of page 26) demonstrates this result. This indicates that the Geometric mean, the method used to compute the NSE 20 Index is a better measure of performance in the long -run. The second test result indicated that the NSE market is mean reverting, that is returns are negatively correlated. The findings were therefore in line with Cooper (1996) findings , that when the market is mean reverting, the arithmetic average is not necessarily superior as a forecast of long-term future returns contrary to Kibbet (2006), Odera (2000) findings which concluded that the Geometric indices are unsuitable for long term price movement. It would however not be overlooked that the analyses were based on a very small stock market whose comparison was made with the world's major established stock markets. The other limitation was that the study was based on the Kenyan case only and we would not conclude as to whether the mean reversion is due largely to country specific factors. Conclusion; NSE market is mean reverting and in such a market the geometric mean is an ideal method of constructing the market Index. However the use of multiple indices may be recommended to address the shortcomings ofthe NSE 20 Index in the short run period.