Preface

Applied Quantitative Methods for Trading and Investment is intended as a quantitative finance textbook very much geared towards applied quantitative financial analysis, with detailed empirical examples, software applications, screen dumps, etc. Examples on the accompanying CD-Rom detail the data, software and techniques used, so that contrary to what frequently happens with most textbook examples, they clarify the analysis by being reasonably easily reproducible by the reader.
We expect this book to have a wide spectrum of uses and be adopted by financial market practitioners and in universities. For the former readership, it will be of interest to quantitative researchers involved in investment and/or risk management, to fund managers and quantitative proprietary traders, and also to sophisticated private investors who will learn how to use techniques generally employed by market professionals in large institutions to manage their own money. For the latter, it will be relevant for students on MSc, MBA and PhD programmes in Finance where a quantitative techniques unit is part of the course, and to students in scientific disciplines wishing to work in the field of quantitative finance.
Despite the large number of publications in the field of computational finance in recent years, most of these have been geared towards derivatives pricing and/or risk management. 1 In the field of financial econometrics, most books have been subject specific,2 with very few truly comprehensive publications.3 Even then, these books on financial econometrics have been in reality mostly theoretical, with empirical applications essentially focused on validating or invalidating economic and financial theories through econometric and statistical methods.
What distinguishes this book from others is that it focuses on a wide spectrum of methods for modelling financial markets in the context of practical financial applications. On top of "traditional" financial econometrics, the methods used also include technical analysis systems and many nonparametric tools from the fields of data mining and artificial intelligence. Although we do not pretend to have covered all possible methodologies, we believe that the wide breadth of potential methods retained in this manual is highly desirable and one of its strengths. At the same time, we have been careful to present even the most advanced techniques in a way that is accessible to most potential readers, making sure that those interested in the practical utilisation of such methods could skip the more theoretical developments without hindering comprehension, and concentrate on the relevant practical application: in this respect, the accompanying CD-Rom should prove an invaluable asset.
An applied book of this nature, with its extensive range of methodologies and applications covered, could only benefit from being a collaborative effort of several people with the appropriate experience in their field. In order to retain the practitioner's perspective while ensuring the methodological soundness and, should we say, academic respectability of the selected applications at the same time, we have assembled a small team of quantitative market professionals, fund managers and proprietary traders, and academics who have taught applied quantitative methods in finance at the postgraduate level in their respective institutions and also worked as scientific consultants to asset management firms. As mentioned above, the range of applications and techniques applied is quite large. The different applications cover foreign exchange trading models with three chapters, one using technical analysis, one advanced regression methods including nonparametric Neural Network Regression (NNR) models and one a volatility filter based system relying on Markov switching regimes; one chapter on equity statistical arbitrage and portfolio immunisation based on cointegration; two chapters on stock portfolio optimisation, one using Kalman filtering techniques in the presence of time-varying betas and the other using matrix algebra and Excel Solver to derive an optimal emerging stock market portfolio; one chapter on yield curve modelling through the use of affine models; one chapter on credit classification with decision trees, rule induction and neural network classification models; two chapters on volatility modelling and trading, one using Excel to compute both univariate and multivariate GARCH volatility and correlation in the stock market, the other using straddle strategies based on GARCH and Recurrent Network Regression (RNR) to build a forex volatility trading model; one chapter on Value at Risk (VaR) and option pricing in the presence of stochastic volatility; one chapter on the information contained in derivatives prices through the use of risk-neutral density functions and, finally, one chapter on weather risk management when confronted with missing temperature data. The first part of the book is concerned with applications relying upon advanced modelling techniques. The applications include currencies, equities, volatility, the term structure of interest rates and credit classification. The second part of the book includes three chapters where the applications on equities, VaR, option pricing and currency trading employ similar methodologies, namely Kalman filter and regime switching. In the final part of the book there are five chapters where a variety of financial applications ranging from technical trading to missing data analysis are predominantly implemented using Excel. In the following we provide further details on each chapter included in the book.
1. "Applications of Advanced Regression Analysis for Trading and Investment" by C. L. Dunis and M. Williams: this chapter examines the use of regression models in trading and investment with an application to EUR/USD exchange rate forecasting and trading models. In particular, NNR models are benchmarked against some other traditional regression based and alternative forecasting techniques to ascertain their potential added value as a forecasting and quantitative trading tool. In addition to evaluating the various models out of sample from May 2000 to July 2001 using traditional forecasting accuracy measures, such as root-mean-squared errors, models are also assessed using financial criteria, such as risk adjusted measures of return. Transaction costs are also taken into account. Overall, it is concluded that regression models, and in particular NNR models, do have the ability to forecast EUR/USD returns for the period investigated, and add value as a forecasting and quantitative trading tool.
2. "Using Cointegration to Hedge and Trade International Equities" by A. N. Burgess: this chapter analyses how to hedge and trade a portfolio of international equities, applying the econometric concept of cointegration. The concepts are illustrated with respect to a particular set of data, namely the 50 equities which constituted the STOXX 50 index as of 4 July 2002. The daily closing prices of these equities are investigated over a period from 14 September 1998 to 3 July 2002 - the longest period over which continuous data is available across the whole set of stocks in this particular universe. Despite some spurious effects due to the non synchronous closing times of the markets on which these equities trade, the data are deemed suitable for illustration purposes. Overall, depending on the particular task in hand, it is shown that the techniques applied can be successfully used to identify potential hedges for a given equity position and/or to identify potential trades which might be taken from a statistical arbitrage perspective.