Modelling Nigerian Exchange Rates with Asymmetric GARCH Models

Abstract

This work consider the estimation of some naira exchange rate returns by volatility models which include the asymmetric variants, with estimation performed under normally distributed assumption of Generalized Autoregressive Conditional Heteroscedastic (GARCH). The symmetric versions are Riskmetrics, ARCH and GARCH models. Initially, first order serial correlation was observed in the returns series, implying the dependencies of current returns on the immediate past. Of the asymmetric volatility models, the Exponential GARCH (EGARCH) and Asymmetric Power ARCH (APARCH) posed to perform better than the other symmetric forms in the predicting the volatility of naira exchange returns.