讨论稿No.E2015007:Modeling Long Memory Volatility Using Realized Measures of Volatility: A Realized HAR GARCH Model

发布日期:2015-10-13 02:27    来源:北京大学国家发展研究院

Modeling Long Memory Volatility Using Realized Measures of Volatility: A Realized HAR GARCH Model

Zhuo Huang   Hao Liu   Tianyi Wang
No.E2015007                           October 2015

 

Abstract

Long memory is an important feature of the volatility of financial returns. We document that the recently developed Realized GARCH model (Hansen et al. 2012) is insufficient for capturing the long memory of underlying volatility. We develop a parsimonious variant of the Realized GARCH model by introducing the HAR specification of Corsi (2009) into the volatility dynamics. A comparison of the theoretical and sample autocorrelation functions shows that the new model specification better captures the long memory dynamics of volatility. We calculate the multi-period out-of-sample volatility forecasts for several return series and find that the new model is a significant improvement over the classic Realized GARCH model.

Keywords: Realized GARCH, HAR, Long Memory, Realized Kernel

No.E2015007