Abstract
"In this article, we propose a methodology for measuring the information flows that underpin option price movements and for analyzing the distribution of these flows. We develop a framework in which information flows can be measured in terms of the relative entropy between the risk-neutral distributions obtained from implied volatility data at different dates. We set up a numerical methodology to compute such quantities using an empirical market dataset that corresponds to options written on the S&P 500 index. This methodology uses Normal Inverse Gaussian distributions for the log-return of the index. We apply our method to six years of daily data, from 2015 to 2021, and find that options with short maturities capture a greater share of new information. We also use a mixture of two exponential distributions to analyze the distribution of the information flows obtained. In this mixture, one component corresponds to frequent small values and the other to less frequent high values."