FOREIGN EXCHANGE MARKET EFFICIENCY IN SELECTED SUB-SAHARAN AFRICAN COUNTRIES: A TEST FOR MARTINGALE DIFFERENCE HYPOTHESIS WITH STRUCTURAL BREAKS

Not much attention has been given to the efficiency of forex markets in African countries, as most existing studies have focused on a single country and only one study seem to have tested the forex market efficiency of a group of sub-Saharan African countries comprising Nigeria, Ghana, Uganda, South Africa and Zambia. Therefore, this study constructs a larger sample consisting of foreign exchange (forex) markets of 10 countries in sub-Saharan Africa by testing for martingale difference hypothesis with structural breaks. It uses data on the average official exchange rate of currencies of the concerned countries to the US dollar from November 1995 to October 2015(i.e. 1995M11-2015M10). The study accounts for the break period/point in each return series to determine when the structural change occurred in the forex market by employing the Perron's unit root test that endogenously determines the most significant break period in the return series in an Innovative Outlier (10) model. Due to the tendency for structural breaks to make the test for martingale difference hypothesis misleading, the study divides the full sample period into before structural break (pre-break period) and after the structural break (post-break period). The pre-break period spans from November 1995 to the month before the structural break while the post-break period extends from the month after the structural break to October 2015. The empirical results reveal that the forex markets of the selected countries are not weak form efficient in the full sample period except for Burundi and Ghana. Dividing the full sample period into pre-break and post-break periods, it discovers that the forex markets of Burundi, Ghana, Mauritius, and Mozambique are weak form efficient in both periods while the forex markets of Gambia and Madagascar are weak form inefficient in both periods. However, the forex markets of Mauritania, Sierra Leone, Uganda, and Zambia show inconsistent results in the pre-break period compared to the post-break period.