• Uniquely written from a practitioner's perspective, this title is sympathetic to the needs of the busy practitioner and is designed to provide rapid and succinct access to useful statistical methods in one handy volume

• The use of practical examples and accessible panels will allow the market risk manager to quickly and easily implement, evaluate and extend a wide variety of statistical modelling tools and techniques for more accurate market risk assessment

• This timely release illustrates the value to be gained from the statistical analysis of market risk data providing a valuable competitive edge in these times of increased regulation

• Key topics such as extreme value theory, volatility modelling, principle components, confidence intervals and fitting probability distributions to real data are covered in sufficient detail so that these methods can be integrated into your own risk management systems

CONTENTS

About the Author
Preface
1. Introduction to Market Risk Management
2. Random Variables and Probability
3. Describing Risk Factors and Portfolios
4. Displaying Risk Factors and Portfolios
5. The Normal Distribution
6. The Method of Maximum Likelihood
7. Fitting Probability Distributions to Risk Factors and Portfolios
8. Principle Component Analysis
9. Modelling Volatility
10. Extreme Value Theory
11. Methods of Simulation
12. Hypothesis Testing
13. Statistical Tests for Market Risk Management
14. Confidence Intervals
15. The Theory and Practice of Market Risk Management
Glossary
Bibliography
Index

Reviewed in September 2003 Revue bancaire et financiere Bank- en Financiewezen by Jan Annaert, University of Ghent

Since the publication of JP Morgan's technical handbook RiskMetrics nearly a decade ago, the modelling of market risks has enjoyed a period of prolific growth. This is a field in which various methods from econometry, statistics and numerical analysis converge. In view of the often highly technical nature of these methods, it is difficult for a risk manager who is starting out to obtain a good overview of the entire field. And every once in a while even the seasoned user will want to look up one method or another, or test statistics, again, just to be on the safe side.

With this work, Nigel Da Costa Lewis (holder of qualifications in statistics, economics, finance and information technology, and also active in the financial world) wanted to create a concise reference book, introducing the basic terms and techniques using a step-by-step approach.

It starts by defining chance variables and probability distributions. In view of the ultimate objective (measuring market risk), attention is of course paid to the relationship between different chance variables and value at risk. It then examines how portfolios and risk factors can be described (descriptive statistics) and visualised (including histograms and plots). Considering the central position occupied by the normal distribution, this subject receives an entire chapter, in which it is specified that financial reality does sometimes differ from this distribution. From Chapter 6 the main focus is on the statistical estimation of parameters, with maximum likelihood estimation, the choice of different distributions, principal components analysis and volatility modelling. One chapter is also dedicated to extreme values theory. Finally, simulation techniques, the creation of statistical tests and the calculation of reliability intervals are discussed.

Needless to say, this book has not become a handbook for market risk management - this was not the intention. However, the author has succeeded in providing an introduction to the most important basic techniques, covering over 200 pages, illustrating each of these techniques with the help of practical examples. The wording is very clear and quite informal throughout: the reader will not find any inferences or evidence of theorems or results, even though at the end of each chapter there is a reference to works providing more in-depth information. The way in which the author draws attention to the possible problems with estimators of location, dispersal and form for the various distributions is very informative. In many instances, alternatives for the standard estimators are provided and discussed. In addition, for a whole range of distributions the analytical expression for these parameters is given, which is very useful in itself.

The fluency with which this book has been written and its compactness mean that it will prove to be a very handy reference book for the experienced risk manager and, at the same time, will provide the newcomer with a thorough grounding in how to measure market risks.

Dr Nigel Da Costa Lewis has many years work experience as a quantitative analyst and statistician in the City of London, on Wall Street and in academia. His work in market risk management dates back to the early 1990s where he developed stress-testing methodologies for portfolios of derivative securities for Legal & General Investment Management Limited. He now specialises in the application of computational intensive quantitative methods to problems in finance. His experience includes the application of neural networks to trading, Bayesian Belief Networks models for risk management and the application of classical and Bayesian statistical methods to market and credit risk. Nigel has an award-winning PhD in statistics from the University of Cambridge, and four Master's degrees, all from the University of London: statistics, finance, economics and advanced computer science.