The year 2003 has
witnessed the release of some groundbreaking books on market risk management
offering a level of sophistication unseen in earlier books. Allen (2003)
delivers a "from the trenches" look at the practical work of market risk
management on an active derivatives trading floor. My own book, Holton (2003),
is the first rigorous book on value-at-risk, disclosing for the first time in
book form such cutting-edge techniques as quadratic VaR and variance reduction
for Monte Carlo VaR. With this book, Christoffersen takes a close look at
elements of VaR measures, especially inference procedures
that capture conditional heteroskedasticity in both standard deviations and
correlations.
Don't be fooled by
the title of the book. A better title would be Elements of Value-at-Risk,
because that is precisely what Christofferseen delivers. While the book is not
as comprehensive as Holton (2003),
it delves more deeply into selected topics. Three in particular are
implementing multi-dimensional GARCH for VaR
applying extreme value theory to VaR, and
backtesting.
It is a short
book. You can easily read it in a couple sittings. Discussions are mostly
theoretical with few examples. However, there are plenty of practical exercises
that get you analyzing historical market data that is provided on an
accompanying disk.
Contents
1. Risk Management and Financial
Returns
2. Volatility Modeling
3. Correlation Modeling
4. Modeling
the Conditional Distribution
5.
Simulation-Based Methods
6. Option
Pricing
7. Modeling Option Risk
8. Backtesting and Stress Testing
If you are new to
VaR, this should not be the first book you read—but it definitely can be the
second. It does not discus mapping procedures at all. Its coverage of
transformation procedures is cursory. Its real strength is the time series
analysis that supports inference procedures. While various techniques are
discussed, the focus is on a multi-dimensional GARCH model just published by
Engle and Shepherd in the past couple years. My only complaint with the
discussion is that it would be nice if it provided some empirical evidence for
how well this new model performs in practice.
This book will
appeal to practitioners and researchers in financial risk management—especially
those interested in the quantitative aspects of market risk management. It
offers perspectives that are well presented and unavailable anywhere else.
Unfortunately, the book was published by a house
suspected of sabotaging its own books on value-at-risk to benefit a certain
competing author. The book's sales have been pitiful.
.gif)
implementing multi-dimensional GARCH for VaR