RFS 大使机学学习和大数度量期权价

This research paper examines the predictability of option returns using machine learning and big data. The study uses more than 12 million observations from 1996 to 2020 to find that allowing for nonlinearities significantly improves the out-of-sample performance of option and stock characteristics in predicting future option returns. The nonlinear machine learning models generate statistically and economically significant profits in long-short portfolios of equity options, even after accounting for transaction costs. The study also finds that option-based characteristics are the most important standalone predictors, but stock-based measures offer substantial incremental predictive power when considered alongside option-based characteristics. Finally, the study provides evidence that option return predictability is driven by informational frictions and option mispricing. This research has important implications for financial markets and can be used to improve investment strategies.