I never trust a
book whose author fills its pages with promises about all the wonderful things
he plans to do later in the book. Too often, he never does them. In this book,
Riccardo Rebonato repeatedly promises to explain how behavioral psychology and
Bayesian probabilities might transform financial risk management. In the closing
chapter, he disappoints:
" ... I am well aware that what I
propose could be readily improved upon by a more systematic and "scientific"
approach ... It may all sound rather vague and abstract at this stage ... If
this book gives pause for thought and serves as a call to arms for related
disciplines (Bayesian analysis, decision theory, experimental psychology) to
take a look at what is happening in the field of risk management, it will have
more than fulfilled its purpose."
With this
recommendation—that perhaps others might like to recommend something—Rebonato
hastens to conclude his Plight of the Fortune Tellers.
The book is the
latest in an emerging genre where authors of technical finance books pitch
themselves as finance gurus to a different and technically less sophisticated
audience. Taleb defined the genre with his unfortunate Fooled By Randomness and Black Swan. This year, Haug (2007),
Bookstaber (2007) and
Rebonato all pitch in with their own offerings. Bookstaber's is saved by the
wonderful first hand accounts that populate his narrative. Haug's book and this
one by Rebanato have little to recommend them.
Rebonato is known
for writing a number of accessible, if verbose texts on fixed income financial
engineering. With this book, he focuses on one problem in financial risk
management, which I first pointed out in a couple blog postings a year earlier
(read the blog postings
here
and here).
It is the tendency of regulators and practitioners to calculate value-at-risk
and related metrics at extraordinarily high quantiles—say the 99.9%, 99.95% and
even 99.99% levels. I pointed out in my blog postings that doing so is not only
technically difficult but philosophically meaningless. Rebonato devotes his book
to the same conclusion.
Contents
1. Why This Book Matters
2. Thinking about Risk
3. Thinking about Probabilities
4. Making Choices
5. What Is Risk
Management For?
6. VaR & Co: How It
All Started
7. Looking Beneath the
Surface: Hidden Problems
8. Which Type
of Probability Matters in Risk Management?
9. The Promise of Economic Capital
10. What Can We Do Instead?
My recommendation
was that practitioners start calculating risk at the more reasonable 90% or 95%
quantiles. Rabonato takes a different tack, offering readers a lengthy introduction to behavioral
psychology and Bayesian statistics. Targeting laymen and non-quantitative
professionals, he also explains in simple language concepts such as
value-at-risk, economic capital, skewness and kurtosis. Everything culminates in
his closing chapter's vague non-recommendations.
Rebonato is the
"global head of market risk and global head of quantitative research and
quantitative analysis at the Royal Bank of Scotland." I am not sure what all
that means, but in the convoluted language of finance, it sounds like he plays a
role in
"institutional marketing." It would have been nice if he had drawn on his
experiences at the bank to illustrate discussions with real-world examples.
Instead, the book is populated with examples of coin tossing experiments and
dinner parties where guests bet their life savings on games involving 40%-60%
odds.
At least Haug
admits his book was an "ego trip." Still, some poor souls thought Taleb's Black Swan
profound. I feel badly for them, but I'm sure they will be floored by Rebonato's Plight of the Fortune Tellers. [October 17, 2007]
.jpg)
