Relevance of the disposition effect on the options market: New evidence

Abstract

A moneyness‐based propensity to sell (MPS) measure, at the aggregate level, determines the propensity of option holders to exercise their winning relative to losing positions. Using data on individual stock and S&P 500 Index options, we find that the MPS measure has significant predictive power over the cross section of delta‐hedged option returns. We test the disposition effect in the options market based on a long–short strategy that exploits price distortions induced by the disposition bias. More pronounced evidence of the disposition bias is found for individual at‐the‐money call options than put options where the significance of abnormal returns remains robust across different subsamples even after we control for the portfolio option greeks and market‐based risk factors. The profitability of the long–short strategy is related to limit‐to‐arbitrage proxies suggesting that behavioral explanations help explain the positive relation between the MPS measure and delta‐hedged option returns.

Universal Arbitrage-Free Estimation of State Price Density

“…Although the dataset does not have a large coverage of call prices as high-frequency data, we argue that our dataset is efficient enough to cover the call price surface. Compare this with Chiang et al (2016),who also follows Constantinides et al (2013) and obtains a total 404,822 observations on S&P 500 options between 1996 and 2011, so our dataset is relatively large. Furthermore, to compare our no arbitrage filter with Zhang and Xiang (2008), we plot the time value of call option price on 04/11/2003 in Figure A.9 and check whether the dataset shows no arbitrage.…”

“…This is a reasonable number compares withChiang et al (2016) who get total 404,822 obervations using the same filters on S&P 500 options between 1996 and 2011.10 We use the Python library GridSearchCV and Cvxopt to program the code and the laptop conducts the codes is an Intel Core i7 and 2.9 GHz.11 The reason why we use relative distance is addressed in Section 5.…”