Modeling and Forecasting of Exchange Rate Volatility of Ethiopia

No Thumbnail Available



Journal Title

Journal ISSN

Volume Title


Addis Ababa University


Modeling and forecasting of volatile data have become the area of interest in _nancial time series.This study was conducted to apply a hybrid model between Autoregressive Integrated Moving Average (ARIMA) model and Generalized Autoregressive Conditional Hetero- cedasticity (GARCH) family model. Symmetric (GARCH) and asymmetric(EGARCH) models are used in this study.The time series data used in this study consist of average monthly Ethiopian Birr/ USA Dollar exchange rate from January 2001 to February 2020 obtained from National bank of Ethiopia.The data were converted to returns to enhance their statistical properties and the returns was used to _t a mean and the variance equa- tion.The parameters for ARIMA models were estimated using Ordinary Least Squares Estimation (OLS) method. For hybrid ARIMA-GARCH and ARIMA-EGARCH, the pa- rameters were estimated by using Maximum Likelihood Estimation (MLE).The ARCH LM test indicate that presence of conditional heteroscedasticity in the ARIMA model. The performance of modeling and forecasting of hybrid ARIMA- GARCH type models have been investigated based on forecasting performance criteria such as MSE, MAE and MAPE tests. The data are divided into two parts where 89% of the data is used as in- sample((build the model) period taken from January 2001 until January 2018. while the rest of data (11%) were used for the out-sample period taken from February 2018 until February 2020.The modeling performance of the hybrid models are evaluated using AIC. Results showed that,hybrid ARIMA (3,1,3)-EGARCH(3,1) model was found to be perform better in modeling and forecasting the volatility of monthly exchange rate return compared to hybrid ARIMA(3,1,3)-GARCH (1,1) model. The processes of modeling and forecasting was done by using Eviews statistical software.



Modeling, Forecasting, Exchange Rate, Volatility, Ethiopia