Application of generalized autoregressive conditional heteroscedasticity for exchange rate volatility in Nigeria

Abstract

In this study, we want to examine impact of exchange rate volatility in Nigeria using an application of Generalized Autoregressive Conditional Heteroscedasticity. The naira exchange rate has exhibited presence of asymmetry in the currency confirmed by the non-zero leverage parameters, naira exchange rate. The results suggest the sum of ARCH and GARCH terms must be nonnegative and less than one. i.e. ARCH+GARCH < 1 existence of volatility clustering in the currency is highly volatile followed by calm period and responds significantly to shock. The Durbin-Watson statistics adjusted the data even better by detecting the presence of autocorrelation in the residual, leading to better estimates for AIC, SIC and HIC.

Keywords: Exchange Rate Volatility, Jaque-Bera, GARCH Model, Volatility Clustering