Econometric Tools for Analysis of Derivatives Data

Northwestern University, Evanston IL

Investigators

Abstract

This project develops new statistical techniques for extracting real-time economic information from the prices of derivatives securities. The basic idea is that option prices include information about market conditions, the general pricing of risk, and the specific risk of a crash. The new method will be valuable to a wide audience, ranging from researchers to policy makers and regulators interested in the real time assessment of market and economic conditions, including the probability and pricing of 'tail' events. The work advances science in these fields and also promotes the national prosperity by providing new techniques for gathering information useful to businesses and policymakers from existing data. The team develops new inference techniques to ensure satisfactory performance based on a parametric model for option prices with minimal restrictions on the dynamics driving the underlying asset prices. The estimated system embeds valuable information about market conditions, the pricing of risk in general and crash risk in particular. The team will apply these methods to predicting future equity returns, extracting spot volatility, pricing volatility risk and volatility derivatives, and documenting global linkages in market 'fears'. Formal inference and specification testing for option pricing is difficult. Complications arise from latent and persistent state variables, no-arbitrage constraints, nonlinear pricing, unbalanced panels, sizeable measurement errors and time-varying risk premiums. The PIs develop procedures inspired by methods for nonparametric volatility estimation from high-frequency time series data. The extension to a parametric cross-sectional setting provides the tools for model evaluation, and they will obtain joint asymptotic distributional results for high frequency and option implied volatility estimators.