Forecasting exchange rate volatility using autoregressive random variance model

Abstract

Recently, as an alternative to the GARCH model, the autoregressive random variance (ARV) model has been gaining popularity in the modelling of changing volatility, mainly because of the capability in capturing the stochastic nature of volatility. This article highlights the ARV model as an alternative to the GARCH model in modelling volatility. The main focus is to compare the two models in forecasting exchange rate volatility. Although the two approaches generally give close forecasting performance, the ARV method provides a notable improvement in Canadian/ Dollar and Australian/Dollar. The outstanding performance seems to be related to the 'volatility of volatility', i.e. the volatility changes from day to day.