Working Paper: NBER ID: w18046
Authors: Torben G. Andersen; Nicola Fusari; Viktor Todorov
Abstract: We develop a new parametric estimation procedure for option panels observed with error which relies on asymptotic approximations assuming an ever increasing set of observed option prices in the moneyness- maturity (cross-sectional) dimension, but with a fixed time span. We develop consistent estimators of the parameter vector and the dynamic realization of the state vector that governs the option price dynamics. The estimators converge stably to a mixed-Gaussian law and we develop feasible estimators for the limiting variance. We provide semiparametric tests for the option price dynamics based on the distance between the spot volatility extracted from the options and the one obtained nonparametrically from high-frequency data on the underlying asset. We further construct new formal tests of the model fit for specific regions of the volatility surface and for the stability of the risk-neutral dynamics over a given period of time. A large-scale Monte Carlo study indicates the inference procedures work well for empirically realistic specifications and sample sizes. In an empirical application to S&P 500 index options we extend the popular double-jump stochastic volatility model to allow for time-varying jump risk premia and a flexible relation between risk premia and the level of risk. Both extensions lead to an improved characterization of observed option prices.
Keywords: No keywords provided
JEL Codes: C51; C52; C58; G12; G13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| parametric estimation procedure (C51) | consistent estimation of risk-neutral density (C51) |
| parametric estimation procedure (C51) | realized trajectory of state vector governing option price dynamics (G13) |
| observation of option prices (G13) | convergence to mixed-Gaussian law (C46) |
| estimation method (C51) | understanding of option pricing dynamics (G13) |
| penalized nonlinear least squares approach (C51) | efficient estimator under homoskedasticity of option errors (C51) |
| double-jump stochastic volatility model (C58) | improved characterization of observed option prices (G13) |
| observation errors (C20) | estimation of parameters governing risk-neutral distribution (C13) |
| observation errors (C20) | recovery of volatility state from option surface (C58) |