The Joint Cross Section of Option and Stock Returns Predictability with Big Data and Machine Learning
76 Pages Posted: 2 Mar 2021 Last revised: 15 Jun 2021
Date Written: December 11, 2020
Using large set of stock and option specific characteristics, and machine learning, we first provide a comprehensive analysis of what drives expected delta hedged and delta-neutral straddle returns of equity options. In contrast to the previous literature, we find that option, rather than stock, characteristics are important predictors of option returns. Second, option, rather than stock, characteristics are also important predictors of stock returns. Moreover, stock characteristics alone fail to predict cross-section of stock returns in recent data. In terms of variables importance for stock return predictability, most of machine learning methods we use outweigh option illiquidity. Consequently, stock long-short portfolio strategies formed conditioning on options illiquidity outperform in out-of-sample those of machine learning portfolios, and the market overall. The results are consistent with substantial increase in options trading activity compared to stocks, and that options markets leading stock markets in recent years.
Keywords: Machine learning, Option pricing, Stock return predictability
JEL Classification: G10, G12, G13, G14
Suggested Citation: Suggested Citation