Abstract
We analyze the high-frequency dynamics of S&P 500 equity-index option prices by constructing an assortment of implied volatility measures. This allows us to infer the underlying fine structure behind the innovations in the latent state variables driving the evolution of the volatility surface. In particular, we focus attention on implied volatilities covering a wide range of moneyness (strike/underlying stock price), which load differentially on the different latent state variables. We conduct a similar analysis for high-frequency observations on the VIX volatility index as well as on futures written on it. We find that the innovations over small time scales in the risk-neutral intensity of the negative jumps in the S&P 500 index, which is the dominant component of the short-maturity out-of-the-money put implied volatility dynamics, are best described via non-Gaussian shocks, i.e., jumps. On the other hand, the innovations over small time scales of the diffusive volatility, which is the dominant component in the short-maturity at-the-money option implied volatility dynamics, are best modeled as Gaussian with occasional jumps.
Original language | English (US) |
---|---|
Pages (from-to) | 532-546 |
Number of pages | 15 |
Journal | Journal of Econometrics |
Volume | 187 |
Issue number | 2 |
DOIs | |
State | Published - Aug 1 2015 |
Funding
Andersen gratefully acknowledges support from CREATES, Center for Research in Econometric Analysis of Time Series ( DNRF78 ), funded by the Danish National Research Foundation . Todorov’s work was partially supported by National Science Foundation grant SES-0957330 . We are also grateful for support from a grant by the CME Group. We thank Makoto Takahashi for providing assistance with collecting the VIX futures data.
Keywords
- High-frequency data
- Implied volatility
- Jump activity
- KolmogorovSmirnov test
- Stable process
- Stochastic volatility
- VIX index
ASJC Scopus subject areas
- Economics and Econometrics