Title: Volatility estimation and sampling efficiency: An intrinsic time approach
Authors: Yifan Li - The University of Manchester (United Kingdom) [presenting]
Ingmar Nolte - Lancaster University (United Kingdom)
Sandra Nolte - Lancaster University (United Kingdom)
Abstract: The concept of intrinsic time sampling (ITS) is developed for a semimartingale, which samples the semimartingale whenever its path triggers some homogenous stopping rule. Based on the ITS scheme, we propose the intrinsic time volatility (ITV) estimator, which is a class of consistent and jump-robust integrated variance (IV) estimators. When the path of the semimartingale is available, we show that the ITS based realized variance and the ITV estimators have smaller asymptotic variances relative to their calendar time-based counterparts under a common sampling frequency. This is driven by the fact that the ITS scheme contains more information about the integrated variance, a concept which we formalize as the sampling efficiency of sampling schemes. We derive explicit bias-correction to the ITV estimators when the semimartingale is observed discretely in time, which provides a simple and effective variance reduction technique for IV estimation using sparsely sampled data.