Approaches to forecasing option volatility

Abstract

The article investigates a new approach to the idea of volatility. In spite of the well-known assumption that option volatility in future will be exactly the same as today, the author puts forward a method, which links the change in volatility to change of only one parameter, i.e. the price of basic asset. The idea that the price of basic asset is a ‘guide’ for option volatility does not need any proof, as like terminal contracts options are estimated proceeding from their basic asset. This method can help estimate future volatility for one (or even more steps) ahead. Like any other forecast method it builds up the error as the number of steps in the future increases, however the simplicity of its use and low resource-intensiveness make it a worthy alternative to the method accepted now, which shows volatility while presenting prospects of the current option position. To forecast volatility for one step ahead we used the following basic statistic methods and economic models: the method of linear regression, Newton-Rafson method for finding option strikes for the set deltas, the method of spline-interpolation, the model of calculating ‘option smile’ Vanna-Volga.