Estimating the Risk Neutral Probability Density Functions Natural Spline versus Hypergeometric Approach Using European Style Options

This paper examines the ability of two recent approaches to estimate implied risk neutral probability density functions (RNDs) - the smoothed implied volatility smile method (SML) and the density functionals based on the confluent hypergeometric functions (DFCH) from the prices of European-style options. A Monte Carlo experiment is conducted to compare the capability of the two techniques to recover simulated distributions based on Heston's (1993) stochastic volatility model. The paper investigates the accuracy and stability of the two methods via two categories of estimated summary statistics. We find that while the SML method outperforms the DFCH method for the summary statistics which are sensitive to the tails of the distribution, the DFCH method dominates the SML method for the summary statistics that are less sensitive to outliers. Due to the lack of observations in the tails when estimating RNDs, we feel that the most appropriate measures for comparing the two methods are the ones less sensitive to extreme values. In this sense, the DFCH method seems to be more appealing. We also apply the two methods via an empirical application.