Microstructural biases in empirical tests of option pricing models

Abstract

Option pricing, Microstructure, Jump-diffusion, Risk-neutral moments, G13,

Robust estimation of risk‐neutral moments

“…For example, at a spacing just above 4.5 dollar between strikes, risk‐neutral skewness takes a value smaller than –0.4, compared to the true value of 0. Dennis and Mayhew () make a similar finding about the domain width, that is, how far option data extends into the tails. In their normally distributed example of a stock with a volatility of 20% the skewness estimates are only unbiased if the strike prices extend about 20% into both tails.…”

“…In theory, the estimation of risk‐neutral moments based on the theorem of Breeden and Litzenberger () is an easy exercise given that a continuum of option prices is available. However, the practical estimation of risk‐neutral densities from empirical option data is subject to many biases from discrete option prices and microstructural noise (e.g., Bliss & Panigirtzoglou, ; Dennis & Mayhew, ). These biases may also be reflected in the results of empirical studies.…”

“…Thus, the differences in results could be driven by the difference in the estimation method. The main two studies analyzing the efficiency of risk‐neutral moment estimators, Dennis and Mayhew () and Bliss and Panigirtzoglou (), rely on early implementations of risk‐neutral moments and only provide comparisons between a few select estimators. For example, Dennis and Mayhew () do not use any interpolation or extrapolation, whereas the study of Bliss and Panigirtzoglou () was published before the seminal work of Bakshi et al () and does not incorporate the use of the central moment formulas of Bakshi et al ().…”

“…The main two studies analyzing the efficiency of risk‐neutral moment estimators, Dennis and Mayhew () and Bliss and Panigirtzoglou (), rely on early implementations of risk‐neutral moments and only provide comparisons between a few select estimators. For example, Dennis and Mayhew () do not use any interpolation or extrapolation, whereas the study of Bliss and Panigirtzoglou () was published before the seminal work of Bakshi et al () and does not incorporate the use of the central moment formulas of Bakshi et al (). We fill this gap in the literature by providing a comprehensive overview over the efficiency of the central moments of Bakshi et al (), quantile moments, the simple VIX (SVIX) of Martin (), and the rare disaster index (RIX) of Gao, Gao, and Song () and Gao, Lu, and Song ().…”

Option Market Characteristics and Price Monotonicity Violations

“…Most of the studies claim that the violations are caused by market microstructural noises and overlook other possibilities. Dennis and Mayhew (2009) suggest that much of the violations are influenced by microstructural biases. Lin et al (2011) andP erignon (2006) attribute the violations to tactical trading effects and microstructure factors.…”