A study of the relationship between volatility premium and option returns over different time horizons: an ex-post and ex-ante empirical analysis using bid-ask data

Abstract

There are three distinct avenues of empirical research relating to option returns. (1) attempts to explain option returns; (2) analysis of models forecasting option implied volatility (IV) versus alternative forecasts of futures realized volatility (RV); and (3) estimation of the economic benefit of volatility forecasting. This study shows that the three apparently disparate fields of research are closely related since option returns are positively related to volatility spread, and asset returns are negatively related to volatility shock. We show that IV outperforms, and indeed subsumes, a subset of time-series historical volatility (TS-HV) forecasts in predicting RV, although the finding that TS-HV does not provide incremental information in forecasting RV, the use of the alternative predictor can enhance the economic profit to option traders. The study also shows that option horizons significantly affect the impact of option mispricing and market direction on option returns. We provide incremental evidence that puts are more expensive than calls and reinforce the argument that pricing asymmetry can be attributed to the greater skewness of put returns due to a negative return-volatility relationship.