CMStatistics 2021: Start Registration
View Submission - CFE
Title: On the ability of risk-neutral densities to forecast the direction of change Authors:  Maria Magdalena Vich Llompart - Washington College (United States) [presenting]
Antoni Vaello Sebastia - University of Balearic Islands (Spain)
Abstract: Option-implied measures have been proven to be better forecasters of future realized measures than their historical counterparts. However, when it comes to the forecasting ability of option-implied Risk-Neutral densities (RND) there is no consensus in the literature. Therefore, being the prediction of a point estimate challenging, the focus is on the ability of the option-implied Risk-Neutral densities in predicting the future direction of the market instead. Risk-Neutral distributions are estimated from information extracted from option prices (for a 30-days horizon) using non-parametric techniques. Using a probit regression analysis, the aim is to study whether a daily change in the probability of a positive return estimated by the option-implied RNDs helps predict the sign of the future return 1, 2, 10 and 30 days ahead. We provide evidence for 11 international financial markets and we find that, overall, results are positive and significant.