MODELLING EXCHANGE RATE VOLATILITY OF THE GHANA CEDI TO THE US DOLLAR USING GARCH MODELS.

Abstract

The  study examines exchange rate volatility with GARCH models using monthly exchange rate data from January 1990 to November 2013. Simple rate of returns is employed to model the exchange rate volatility of Ghana Cedi-United States Dollar. The models included both symmetric and asymmetric models that capture the most common stylized facts about returns such as volatility persistence and leverage effect. The result identified EGARCH (2, 2) as the overall best fitted model. This model has the least AIC of -6.28 and SIC of -6.16.  Diagnostic test of the models residuals with the Ljung-Box test, the ARCH-LM test and the ACF plots revealed that the models are free from higher order autocorrelation and conditional heteroscedasticity separately. Our results also revealed persistence of volatility and the non-existence of leverage effects as shown by the asymmetric models.

Keywords: Leptokurtic, volatility persistence, leverage effect.