keywords: Exchange rate volatility, heteroscedasticity, GARCH model, stationarity
This study examines error of estimate under different distributions with the generalized autoregressive conditional heteroscedasticity. The naira exchange rate in Nigeria has exhibited the features of continuous depreciation and instability, this instability and continuous depreciation of the naira in the foreign exchange market has resulted in declines in the standard of living of the populace. Empirical results further reveal that the GARCH model achieved under three different error distributions for all the currencies, are then compared using the minimum value of the Akaike information criterion (AIC), Hannan-Quinn criterion (HIC), and Schwarz information criterion (SIC). The result shows that PARCH (1, 1) model with asymmetric order 1 under the Generalized error distribution (GED) proved to be the best model which reduces the persistence of volatility clustering in the Nigeria exchange market.