Kaeck, Andreas and Alexander, Carol (2013) Continuous-time VIX dynamics: on the role of stochastic volatility of volatility. International Review of Financial Analysis, 28. pp. 46-56. ISSN 10575219

Preview

PDF - Accepted Version Download (810kB) | Preview

Abstract

This paper examines the ability of several different continuous-time one- and two-factor jump-diffusion models to capture the dynamics of the VIX volatility index for the period between 1990 and 2010. For the one-factor models we study affine and non-affine specifications, possibly augmented with jumps. Jumps in one-factor models occur frequently, but add surprisingly little to the ability of the models to explain the dynamic of the VIX. We present a stochastic volatility of volatility model that can explain all the time-series characteristics of the VIX studied in this paper. Extensions demonstrate that sudden jumps in the VIX are more likely during tranquil periods and the days when jumps occur coincide with major political or economic events. Using several statistical and operational metrics we find that non-affine one-factor models outperform their affine counterparts and modeling the log of the index is superior to modeling the VIX level directly.

Item Type:	Article
Schools and Departments:	School of Business, Management and Economics > Business and Management
Subjects:	H Social Sciences > HG Finance
Depositing User:	Carol Alexander
Date Deposited:	02 Jun 2013 10:52
Last Modified:	17 Mar 2017 07:02
URI:	http://srodev.sussex.ac.uk/id/eprint/45210

View download statistics for this item

📧 Request an update