Applied Volatility and Correlation Modelling Using Excel推荐.pdfVIP

10 Applied Volatility and Correlation Modelling Using Excel∗ ´ ´ FREDERICK BOURGOIN ABSTRACT This chapter implements a range of univariate and multivariate models within Microsoft Excel. This is extremely useful as a large proportion of finance professionals, students and researchers that are familiar with this package. We show how to generate one-step ahead forecasts of volatility using the JP Morgan RiskMetrics model, JP Morgan RiskMetrics model with an optimal decay, a GARCH model with and without a variance reduction technique and finally using the GJR model to account for asymmetric reaction to news. A comparison of forecasts is made and some useful insights to the efficiency of the models is highlighted. In the second part of this chapter, we model the time-varying correlation using different models. As with the univariate approach this includes the JP Morgan RiskMetrics model with and without optimal decay, a GARCH model with and without variance reduction and finally the so-called “Fast GARCH” model of which the author has previously made significant contributions to the literature. 10.1 INTRODUCTION The practicability of multivariate GARCH models has been the subject of several articles in the past few years, particularly regarding the feasibility of large size problems (Athayde, 2001; Bourgoin, 2000, 2002; Engle and Mezrich, 1995, 1996; Ding and Engle, 1994; Ledoit, 2001; Ledoit et al., 2001). Some models work well only on specific data, like the orthogonal GARCH (Alexander and Chibumba, 1997) with yield curve or term structure of implied volatilities (see Bourgoin (2000) for the reasons). Because all these models requ