A1719
Title: Models for realised volatility
Authors: Dario Palumbo - Ca' Foscari University of Venice (Italy) [presenting]
Andrew Harvey - University of Cambridge (United Kingdom)
Abstract: A statistical framework is set for modeling realised volatility (RV) using DCS/GAS. It shows how a preliminary analysis on FTSE, based on fitting a linear Gaussian model to logRV confirms a two component specification and at the same time reveals a weekly pattern in RV. It also yields an interesting comparison with the HAR model, which is a simple way of accounting for long memory in volatility. Fitting the two component specification with leverage and a day of the week component is then carried out directly on RV with a Generalised Beta of the second kind (GB2) conditional distribution - equivalent to the estimation of a model for logRV with an Exponential Generalised Beta of the second kind (EGB2) conditional distribution, of which the normal distribution is a limiting case. The preliminary analysis of logRV also indicates heteroscedasticity in the residuals. The relationship between the GB2 and EGB2 distributions suggests that this heteroscedasticity may be due to a dynamic tail index in the GB2 model, and the DCS model is extended to allow for this possibility. Ultimately the forecasting power of the DCS model is compared with the HAR revealing similar forecasting performance besides its higher descriptive power.
