3. "Modelling the Term Structure of Interest Rates: An Application of Gaussian Affine
Models to the German Yield Curve" by N. Cassola and J. B. Luis: this chapter shows
that a two-factor constant volatility model describes quite well the dynamics and the
shape of the German yield curve between 1986 and 1998. The analysis supports the
expectations theory with constant term premiums and thus the term premium structure
can be calculated and short term interest rate expectations derived from the adjusted
forward rate curve. The analysis is carried out in Matlab and the authors include all
of the files with which to reproduce the analysis. Their findings will be of interest
to risk managers analysing the shape of the yield curve under different scenarios and
also to policy makers in assessing the impact of fiscal and monetary policy.
4. "Forecasting and Trading Currency Volatility: An Application of Recurrent Neural Regression and Model Combination" by C. L. Dunis and X. Huang: this chapter examines the use of nonparametric Neural Network Regression (NNR) and Recurrent Neural Network (RNN) regression models for forecasting and trading currency volatility, with an application to the GBP/USD and USD/JPY exchange rates. The results of the NNR and RNN models are benchmarked against the simpler GARCH alternative and implied volatility. Two simple model combinations are also analysed. Alternative FX volatility forecasting models are tested out of sample over the period April 1999 May 2000, not only in terms of forecasting accuracy, but also in terms of trading efficiency: in order to do so, a realistic volatility trading strategy is implemented, using FX option straddles once mispriced options have been identified. Allowing for transaction costs, most trading strategies retained produce positive returns. RNN models appear as the best single modelling approach, yet model combination which has the best overall performance in terms of forecasting accuracy fails to improve the RNN-based volatility trading results.
5. "Implementing Neural Networks, Classification Trees, and Rule Induction Classification Techniques: An Application to Credit Risk" by G. T. Albanis: this chapter shows how to implement several classification tools for data mining applications in finance. Two freely available softwares on classification neural networks and decision trees, respectively, and one commercial software for constructing decision trees and rule induction classifiers are demonstrated, using two datasets that are available in the public domain. The first dataset is known as the Australian credit approval dataset. The application consists of constructing a classification rule for assessing the quality of credit card applicants. The second dataset is known as the German credit dataset. The aim in this application is to construct a classification rule for assessing the credit quality of German borrowers. Beyond these examples, the methods demonstrated in this chapter can be applied to many other quantitative trading and investment problems, such as the determination of outperforming/underperforming stocks, bond rating, etc.
6. "Switching Regime Volatility: An Empirical Evaluation" by B. B. Roche and M. Rockinger: this chapter describes in a pedagogical fashion, using daily observations of the USD/DEM exchange rate from October 1995 to October 1998, how to estimate a univariate switching model for daily foreign exchange returns which are assumed to be drawn in a Markovian way from alternative Gaussian distributions with different means and variances. The application shows that the USD/DEM exchange rate can be modelled as a mixture of normal distributions with changes in volatility, but not in mean, where regimes with high and low volatility alternate. The usefulness of this methodology is demonstrated in a real life application, i.e. through the profit performance comparison of simple hedging strategies.
7. "Quantitative Equity Investment Management with Time Varying Factor Sensitivities" by Y. Bentz: this chapter describes three methods used in modern equity investment management for estimating time-varying factor sensitivities. Factor models enable investment managers, quantitative traders and risk managers to model co-movements among assets in an efficient way by concentrating the correlation structure into a small number of factors. Unfortunately, the correlation structure is not constant but evolves in time and so do the factor sensitivities. As a result, the sensitivity estimates have to be constantly updated in order to keep up with the changes. The first method, based on rolling regressions, is the most popular but also the least accurate. The second method is based on a weighted regression approach which overcomes some of the limitations of the first method by giving more importance to recent observations. Finally, a Kalman filter-based stochastic parameter regression model is shown to optimally estimate nonstationary factor exposures.
8. "Stochastic Volatility Models: A Survey with Applications to Option Pricing and Value at Risk" by M. Billio and D. Sartore: this chapter analyses the impact on Value at Risk and option pricing of the presence of stochastic volatility, using data for the FTSE100 stock index. Given the time varying volatility exhibited by most financial data, there has been a growing interest in time series models of changing variance in recent years and the literature on stochastic volatility models has expanded greatly: for these models, volatility depends on some unobserved components or a latent structure. This chapter discusses some of the most important ideas, focusing on the simplest forms of the techniques and models available. It considers some motivations for stochastic volatility models: empirical stylised facts, pricing of contingent assets and risk evaluation, and distinguishes between models with continuous and discrete volatility, the latter depending on a hidden Markov chain. A stochastic volatility estimation program is presented and several applications to option pricing and risk evaluation are discussed.