Alexander's
Market Models is a
practitioner-oriented text on various aspects of, and applications for,
volatility and correlation modeling in financial markets. The book is divided
into three parts:
Volatility and Correlation Analysis
Modeling the Market Risk of
Portfolios
Statistical Models for Financial
Markets
The book reads like a "stream of consciousness."
Some of the many topics discussed are:
adjusting delta hedges to reflect a
correlation between the value of an underlier and its implied volatility;
a comparison of uniformly-weighted
and exponentially-weighted moving average models for volatility;
a survey of popular univariate GARCH
models;
modeling returns with non-normal
distributions;
use of principal component analysis
with value-at-risk;
an introduction to cointegration;
modeling high-frequency data.
There is no unifying plot or direction to all this.
It is as if the author has much to say but nothing specific to say, so the
discourse meanders from one topic to another. The only clear
organization is in the second part, which is focused on modeling techniques that
are applicable to portfolio analyses—value-at-risk and portfolio selection. The
first and third parts might be described as each offering a sampling of topics
from financial econometrics. GARCH models are described in Part 1, but more
fundamental ARMA models are only introduced in Part 3. The notion of
stationarity is invoked throughout the text, but isn't defined until Part 3.
Alexander tries to
keep discussions as intuitive as possible. Formulas are used sparingly. There
are plenty of charts of financial time series and other graphics. Technical
concepts—like covariances, volatility and copulas—are introduced with
qualitative descriptions rather than formal mathematical definitions.
Contents
Volatility and Correlation Analysis
1. Understanding Volatility and
Correlation
2. Implied Volatility and Correlation
3. Moving Average Models
4. GARCH Models
5. Forecasting Volatility and
Correlation
Modelling the Market Risk of
Portfolios
6. Principal Component Analysis
7. Covariance Matrices
8. Risk Measurement in Factor Models
9. Value-at-Risk
10. Modelling Non-normal Returns
Statistical Models for Financial
Markets
11. Time Series Models
12. Cointegration
13. Forecasting High-Frequency Data
The author spends considerable effort explaining basic concepts and then assumes
familiarity with more advanced topics. In the first chapter, she carefully
introduces the notion of correlation, but the discussion assumes familiarity
with regression and Student's t-statistic. Later, she carefully explains
what it means for an option to be in-, at-, or out-of-the-money. She then
assumes familiarity with binomial option pricing models. She carefully explains
the difference between historical and implied volatility, but assumes
familiarity with stochastic differentials and Ito's lemma. Given the largely
intuitive nature of the exposition, and the plethora of topics covered, this
isn't too significant a problem. Topics do not build one upon the other. If you
are reading a section and are confused, skip forward a few pages and start
afresh with a new topic.
The book is practitioner oriented. Compared to a financial
econometrics text like Gourier and Jasiak (2001),
which is rigorous but not as in touch with financial practice, this book is not
rigorous but more in touch with financial practice. The choice of topics is, for the
most part, relevant for traders, financial engineers or risk managers.
There are numerous examples based on actual financial data. The book is an easy read. You can flip through it, skipping topics that
don't interest you and focusing on those that do.
Who is this book
suitable for? I recommend it for experienced practitioners who are interested in
current research in applied financial econometrics. Readers with modest
quantitative skills can use the book to learn the "buzz words" and gain an
intuitive understanding of what they mean. More quantitative readers can use the
book somewhat as a survey of relevant literature—and follow up on cited
references for topics they find interesting. Note that many of the reference
require far greater technical knowledge than this book offers. To strengthen
your own background, see books on time
series analysis.
.gif)
.gif)
Volatility and Correlation Analysis