Modeling exchange volatility in Egypt using GARCH models

In this study, we consider the generalized autoregressive conditional heteroscedastic approach in modeling real effective exchange rate in Egypt using monthly data from 1994 to 2009. Various GARCH extensions are performed here. The main results show that real effective exchange rate volatility may have different behaviors based on measures enable to determine it. More importantly, when we take into account volatility clustering (i.e. Standard GARCH), we observe a quite persistence implying a mean reverting variance process. However, when we consider the leverage effect (i.e. Exponential GARCH), we notice a tendency to a long memory which can be itself a source of an explosive process.