Long-Memory Models in Testing the Efficiency Market Hypothesis of the Algerian Exchange Market

Abstract The purpose of this study is to examine the Efficiency Market Hypothesis (EMH) from the perspective of the Algerian exchange rate market. We apply different tests of dependence, long memory, volatility clustering and unit root tests over the three main Algerian exchange rate returns series vis–à-vis the US Dollar, the Euro, and the British Pound. Empirical findings suggest that combined Autoregressive Moving Average (ARMA)-Fractionally Integrated Generalized Autoregressive Conditional Heteroskedastic (FIGARCH) models were the most appropriate to represent the behavior of exchange rate returns. We also compare the predictive qualities of the estimated models and the Random Walk (RW) in terms of out-of-sample forecasting. The results are held to imply the rejection of the EMH in the Algerian exchange rate market. Therefore, the exchange rates fluctuations can be predicted, which may help public authorities intervene in the exchange market and assess the consequences of different economic policies.