Volatility Estimation: Numerical Methods

These notes have been written with the precisely purpose of summarizing the more often encountered and implemented volatility estimation techniques, to describe the realized volatility surface and its term structure, for example in developing Option Pricing libraries. The common and accepted assumptions behind the random fashion, that each quoted and traded asset follows, there are stochastic differential equations (SDEs) characterized by two main terms: one is the drift and the other one is the diffusion term or volatility. If the drift term is set uniquely by the definition of the martingale measure, imposing the drift’s value under such risk neutral measure, to be equal to the free interest rate; on the other side, the diffusion term or volatility is not estimated or defined uniquely. Indeed, the latter is estimated involving several different approaches, that over the time have been developed, trying to catch a better fit with the observed options’ quotation.