Statistics and Its Interface

Volume 1 (2008)

Number 2

Impact of overnight information on MEM volatility prediction

Pages: 297 – 306

DOI: https://dx.doi.org/10.4310/SII.2008.v1.n2.a7

Authors

C. F. Chu (Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong)

K. P. Lam (Department of Systems Engineering and Engineering Management, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong)

Abstract

Overnight return in stock market is one kind of information that can reflect the volatility of the corresponding financial instrument. However, some volatility estimators, either based on range-based or high-frequency data, do not include this information in their formulations. In this study, we investigate the impact of overnight return on Engle’s Multiplicative Error Model (MEM). Garman’s and Hansen’s whole-day-based estimators are studied to demonstrate the effects under minimum-variance situations. Besides, a general framework for incorporating overnight information is proposed and the results are discussed. Our findings demonstrate that overnight return gives a non-monotonic influence and it does contain useful information for predicting the CBOE volatility indexes under specific combinations.

Keywords

volatility forecast, multiplicative error model, MEM, overnight return

Published 1 January 2008