[ARTICLE] This study explores a new method (particle MCMC) for estimating complex option pricing models that involve hidden factors (latent state variables). They show this method leads to different risk measure estimates compared to previous methods, and that the specific options chosen can significantly impact the results.

by Jeroen Rombouts (ESSEC Business School), Yuguo Liu, Kris Jacobs, Arnaud Dufays

The cross-section of options holds great promise for identifying return distributions and risk premia, but estimating dynamic option valuation models with latent state variables is challenging when using large option panels. We propose a particle MCMC framework with a novel filtering approach and illustrate our method by estimating workhorse index option pricing models. Estimates of the variance risk premium, variance mean reversion, and higher moments differ from the literature. We show that these differences are due to the composition of the option sample. Restrictions on the option sample's maturity dimension have the strongest impact on parameter inference and option fit in these models.

[Please read the research paper here]