Allowing for Jump Measurements in Volatility: A High-Frequency Financial Data Analysis of Individual Stocks
Bulletin of Economic Research, Vol. 68, Issue 2, pp. 124-132, 2016
Posted: 1 Jul 2015 Last revised: 13 Apr 2016
Date Written: June 29, 2015
Abstract
Following recent advances in the non-parametric realized volatility approach, we separately measure the discontinuous jump part of the quadratic variation process for individual stocks and incorporate it into heterogeneous autoregressive volatility models. We analyze the distributional properties of the jump measures vis-à-vis the corresponding realized volatility ones, and compare them to those of aggregate US market index series. We also demonstrate important gains in the forecasting accuracy of high-frequency volatility models.
Keywords: Volatility jumps, realized volatility, high-frequency data, HAR-RV model
JEL Classification: C1, C14, C53, C58, G1
Suggested Citation: Suggested Citation
Vassilios G. Papavassiliou (Contact Author)
University College Dublin (UCD) - Michael Smurfit Graduate School of Business ( email )
Blackrock, Co. Dublin
Ireland
UCD Geary Institute for Public Policy ( email )
University College Dublin
Belfield, Dublin Dublin 4
Ireland