A1817
Title: Realized volatility based forecasting models: Exploiting the errors
Authors: Tim Bollerslev - Duke University (United States) [presenting]
Andrew J Patton - Duke University (United States)
Rogier Quaedvlieg - Maastricht University (Netherlands)
Abstract: We propose a new family of easy-to-implement realized volatility based forecasting models. The models exploit the asymptotic theory for high-frequency realized volatility estimation to improve the accuracy of the forecasts. By allowing the parameters of the models to vary explicitly with the (estimated) degree of measurement error, the models exhibit stronger persistence, and in turn generate more responsive forecasts, when the measurement error is relatively low. We document significant improvements in the accuracy of the resulting volatility forecasts compared to the forecasts obtained from some of the most popular existing realized volatility based forecasting models. We also discuss multivariate extensions of the new class of models and various applications thereof, including portfolio allocation decisions and systematic risk measurement.
