Forecasting Exchange Rate Volatility: The Superior Performance of Conditional Combinations of Time Series and Option Implied Forecasts

Abstract

This paper shows that combinations of option implied and time series volatility forecasts that are conditional on current information are statistically superior to individual models, (unconditional) combinations, and hybrid forecasts. Hence, it finds empirical evidence that both, combining individual forecasts, and taking into account the conditional expected performance of each model given current information, are important to improve out-of-sample forecasting performance. The method used in this paper extends the application of conditional predictive ability tests to select forecast combinations. We show that this method works well in practice by applying it to volatility forecasts for the Mexican Peso-US Dollar exchange rate, where the actual value is taken to be the realized volatility measured using intra-day observations.