Modeling realized volatility with implied volatility for the EUR/GBP exchange rate

Anna Rokicka and Janusz Kudła

Abstract

This paper concerns the application of implied volatility in modeling realized volatility in the daily, weekly and monthly horizon using high-frequency data for the EUR/GBP exchange rate. The EUR/GBP rate was chosen because it is under stress triggered by the uncertainty related to Brexit. The heterogeneous autoregression (HAR) model of realized volatility and its extensions were used for the study: the HAR model extended by the measurement error component and by the components of the asymmetry of returns. The introduction of implied volatility into the HAR forecast model improves the predictions in the periods considered irrespective of the period chosen. In particular, we find new evidence that implied volatility obtained from options with a weekly expiration date improves the quality of volatility forecasts. The best prediction for daily and weekly volatility provides the model with asymmetry of returns. For monthly volatility, the model corrected for measurement errors has the best fit.