Predictable dynamics in implied volatility smirk slope : evidence from the S&P 500 options

Abstract

This study aims to investigate whether there are predictable patterns in the dynamics of implied volatility smirk slopes extracted from the intraday market prices of S&P 500 index options. I compare forecasts obtained from a short memory ARMA model and a long memory ARFIMA model within an out-of-sample context over various forecasting horizons. I find that implied volatility smirk slopes can be statistically forecasted and there is no statistically significant difference among competing models. Furthermore, I investigate whether these implied volatility smirk slopes have predictive power for future index returns. I find that slope measures have predictive ability up to 20 minutes.