Abstract:
Volatility modelling is critical to modern financial risk management as it provide an accurate measure of the risk faced by investor. Vast amount of research and effort has devoted to measure the volatility of prices in the past century, yet there is no consensus to how the volatility should be modelled and measured. Amongst the enormous varieties of models, the most widely recognised and applied is the ARCH/GARCH family pioneered by Engle in the late 1980s. Due to the size and the potential impact of the result, highly accurate measures are desired. Consider a moderate portfolio of a billion dollar, a 1% error in the estimation of the volatility can result in a risk measure differ by one million dollar. In this paper we propose a flexible class of distribution known as the scale normal mixture for GARCH in which the volatility can be measured accurately and can be applied over a large spectrum of assets with the potential to explain the micro-market structure or driving force of volatility.