Modeling and Forecasting of British Pound/US Dollar Exchange Rate: An Empirical Analysis

Abstract

The aim of this paper is to develop and examine the characteristics of volatility of exchange rate on British pound/US dollar, using symmetric and asymmetric GARCH(p,q) models. Given that there are ARCH effects on exchange rate returns, we estimated ARCH(q), GARCH(p,q), and EGARCH(p,q) including these effects on mean equation. These models were estimated with maximum likelihood method using the following distributions: normal, t-Student, and generalized error distribution. The log-likelihood function was maximized using Marquardt’s algorithm (1963) in order to search for optimal parameters. The results showed that ARIMA(0,0,1)-EGARCH(1,1) model with t-Student distribution is the best in order to describe exchange rate returns and also captures the leverage effect. Finally, for the forecasting of ARIMA(0,0,1)-EGARCH(1,1) model, both the dynamic and static procedures are used. The static procedure provides better results on the forecasting rather than the dynamic.