Realized Range-Based Estimation of Integrated Variance

2006-06-04

AI Summary

This study addresses the downward bias and inefficiency of conventional realized variance estimators under discrete observations by proposing a novel high-low range–based estimator for the quadratic variation of continuous semimartingales. The method replaces squared returns with normalized squared price ranges, yielding a consistent and asymptotically mixed normal estimator that effectively corrects bias induced by non-trading periods. Leveraging probabilistic limit theory, continuous semimartingale modeling, and high-low price statistics, the approach achieves an 80% reduction in theoretical variance compared to traditional estimators. Empirical analysis using TAQ data demonstrates that the proposed estimator substantially outperforms existing benchmarks in terms of estimation accuracy, efficiency, and robustness.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Problem

• quadratic variation
• realized variance
• range-based estimation
• downward bias
• discrete sampling

Innovation

• realized range-based variance
• quadratic variation
• mixed Gaussian limit
• high-low range
• semimartingale