Abstract:
Modeling and forecasting exchange rate volatility has important implications in a range of areas in macroeconomics and finance. A number of models have been developed in empirical finance literature to investigate this volatility across different regions and countries. This research work considers the autoregressive conditional heteroscedastic and the generalized autoregressive conditional heteroscedastic approaches in modeling daily exchange volatility between CEMAC CFA Franc and United States Dollar (XAF/USD) from 1st January 2010 to 4th January 2018. Both the symmetric (ARCH and GARCH) and the asymmetric (APARCH, GJR-GARCH and EGARCH) GARCH families of models have been taken into consideration to capture some stylized facts about exchange rate returns such as volatility clustering and leverage effect. All models are estimated using the maximum likelihood method under the assumption of several distributions of the innovation terms such as: Normal (Gaussian), Student-t and skew student-t distributions. Evaluating the models using standard information criteria (AIC and BIC) showed that the conditional volatility models are best estimated under the student-t distribution with EGARCH (1,1) being the best fitted model. In accordance with the estimated models there is empirical evidence at some point that negative and positive shocks imply a different next period volatility of the daily XAF/USD exchange rate return. Finally, the research work concludes that the exchange rates volatility can be adequately modeled by the class of GARCH models.
