Torben G. Andersen and Oleg Bondarenko. In recent years, and particularly following the ``flash crash'' on May 6, 2010, warning signals for impending market stress have been in high demand, yet only the VPIN metric of Easley, Lopez de Prado and O'Hara (ELO) has claimed significant success. In addition, ELO deem VPIN a useful predictor of short-term volatility.
VPIN involves decomposing volume into active buys and sells. We utilize quotes and trades to construct an accurate trade classification measure for the E-mini S&P 500 futures. Against this benchmark, the ELO Bulk Volume Classification (BVC) scheme is inferior to a standard tick rule. Moreover, VPIN predicts volatility solely because increasing volatility induces systematic classification errors into the BVC procedure. Finally, all predictive power of VPIN for future volatility vanishes once we control for standard realized volatility measures obtained from high-frequency return observations. We conclude VPIN is unsuitable for capturing order flow toxicity or signaling ensuing market turbulence.
Back to Workshop IV: Forensic Analysis of Financial Data