Abstrak RSS

Predictive Powers Of Implied Volatility And Volatilities From Time Series Models Of Options

Predictive Powers Of Implied Volatility And Volatilities From Time Series Models Of Options
Nury Effendi
Universitas Padjadjaran, Economic Journal FE-Unpad Vol. 22, No. 1, March 2007: 30 - 42 ISSN : 0854-1493
Bahasa Inggris
Universitas Padjadjaran, Economic Journal FE-Unpad Vol. 22, No. 1, March 2007: 30 - 42 ISSN : 0854-1493
, ,

Black and Scholes option pricing model states that (the call) option value is a function offive, readily-observable (or at least computable) parameters – the current stock price, the option’s exercise price and term to maturity, the risk-free interest rate, and the (stock’s) instantaneous price volatility. Volatility is, therefore, one of the most important factors when pricing options. It is widely believed that an option’s implied volatility represents the market’s best forecast of future volatility. However, previous work on predictive power of implied volatility shows mixed results. Several studies find that implied volatility is found to be a poor forecast for the future volatility. Other studies find that implied volatility outperformed volatilities from time-series models although implied volatility is also a biased estimate of future volatility. However, most of the previous work suffers from not having enough data and degrees of freedom and does not explicitly consider option’s expiration days. This study uses several years of daily observations on S&P 500 options on futures and breaks up the data according to the length of time to expiration. The study finds that the relative predictive power of implied volatility versus volatilities from time-series models probably depend on how far away the option is from expiration. For options which are relatively close to expiration, implied volatility seems to be a better estimate of future volatility. But for options with a longer period of expiration, volatilities from time-series model may outperform the implied volatility. For options in the middle range, implied volatility and volatilities from time-series models do not der by much in predictive power.

Download: .Full Papers