Geeks like to read – and not only programming books. Most of us read incessantly. Whether it’s popular science, sci-fi or fantasy, a good thriller or an occasional popular history book or biography, it’s a rare geek who isn’t in love with books. And I am no exception, although I have to confess I am rather an extreme case since my love of books and eclectic tastes borders on the “gentle madness” aka “bibliomania.” 

What I am going to do in this regular column is feed my habit by highlighting some of the books I am reading, and (mostly) enjoying. (I will only rarely write negative reviews; it’s a rare book that I “do not put down gently but throw across the room with great force” after all.)

Finally, since I remain involved with Apress (www.apress.com), the publishing company for IT professionals I helped found, is there any potential for a conflict of interest? I don’t think so nor do the kind people at SYS-CON; the kind of books I will be reviewing are the books I read “non-professionally” –for fun – in my spare time. It is only these books, rather than professional books, that I will review.

Title: A Demon of Our Own Design: Markets, Hedge Funds, and the Perils of Financial Invocation
Author: by Richard Bookstaber
Publisher: John Wiley 2007, ISBN 978-0-471-22727-4.
Price: $27.95

Talk about prescience, this amazingly relevant book, 10 years in the making, was actually published in April of 2007 – four months before the market meltdown that was caused by subprime mortgage-backed instruments going very, very south. Yet, in this extremely interesting, extraordinarily well-written book, Bookstaber essentially predicted this, and also (alas) he predicts more crises to come. The first paragraph of the book sets the stage:

“While it is not strictly true that I caused the two great financial crisis of the late 20th century – the 1987 stock market crash and the Long-Term Capital Management (hedge fund debacle)…”

Who is Bookstaber? He was one of the early “quants” on Wall Street. These are the physicists, mathematicians, statisticians, and mathematical who revolutionized trading on Wall Street by trying to search out hidden patterns and values. The foremost among them, the great mathematician James Simon, made a mere $1.7 billion last year by apparently doing exactly this.

The idea, roughly speaking, of most of these strategies is: can you find two (seemingly) equivalent securities that should trade in parallel but are currently not? Buy one and sell short the other, wait (hope) for a regression to the mean and make a bundle. Only you won’t make a bundle unless you use a lot of leverage because the anomalies are small at best. And leverage, as the old Wall Street adage has it, cuts both ways. If something really strange happens (a “black swan event,” in the words of best-selling author Nicolas Taleb whose very interesting book I’ll review soon) that makes the divergence greater than historical norms, your billions upon billions of leveraged investments often leads to catastrophic losses. And they did. What’s worse, there is a Heisenberg quality to this kind of historical/statistical arbitrage. As Bookstaber puts it: “Predicated on their conviction that the relationship had long-term stability, they would take positions based on the assumption that it would return, or converge, back to its historical value. What they did not appreciate was that they had changed history. There had never been someone trading hundreds of billions of dollars in the middle of this relationship before.”

Of course, it isn’t only about statistical arbitrage affected by the size of the investments made by people playing the game. As Bookstaber often points out, there is a certain amount of stupidity, cupidity, and outright deceit at work here. For example, the Enron-related transactions of Citigroup were described by the then head of risk management at CitiBank as: “[our] accounting is aggressive and a franchise risk to us if there is publicity,” which Bookstaber translates beautifully as: “we’re making this up and if anyone finds out, we’re in trouble.” Of course, Enron was found out and billions upon billions were lost and some of the people who went beyond merely aggressive accounting actually went to jail. (He has a wonderful suggestion about how to handle financial reporting by the way – one that will appeal to programmers: replace a lot of the hocus pocus in financial reports by requiring that the raw data be made available in basically XML form, and thus let smart data miners make sense of it without it being filtered through high-priced accounting firms with potential stakes in the matter.)

As this book is just filled with interesting information and great anecdotes, I could go on for a long time describing its virtues. For example, there’s a wonderful discussion of what portfolio insurance is and how it (inevitably with hindsight) led to the crash of 1987, for example. In the end, however, the whole point of this book can be summed up by the following passage:

“The danger to the system is the system…despite all the risks we can control, the greatest ones remain beyond our control. Those are the risks we do not see, things beyond the veil…”

Combine this inescapable fact with the amazing leverage that modern hedge funds can use, throw in the usual amounts of human stupidty and cupidity, the speed with which things happen in our modern era, the interconnections of everything with everything and Bookstaber has convinced me at least that more and more disasters await the financial system.

Before I leave this review though I do want to point out some parts of the book that will be especially interesting to programmers. Here is an amusingly scary one that occurred in the past and perhaps is not so relevant anymore; however, it remains of vital importance to everything we do. He describes how one early crisis (1995) was caused by the use of a now exotic programming language called APL (“A Programming Language”) created by Turing award-winner Ken Iverson. APL was an amazingly fun language for math types. You could invert a matrix with a single key stroke and write the most amazing one-liners. But let’s forget about the maintenance problems of a language that makes PERL seem like COBOL in it lucidity; the real problem was that APL was interpreted in an era when JITs were still a decade away and computers were hundreds of times slower. But many financial calculations require zillions of iterative calculations to compute prospective values, and APL, since it was not compiled in those days, simply couldn’t do them. So oddball situations weren’t well modeled in the financial models built with APL, and, once upon a time, this failure to use a compiled language led to a loss of between 100 and 250 million dollars to UBS – and this was back when a quarter of a billion dollars was real money.  

But more important is his stressing that the worst nightmare of financial machinations is our worse nightmare as well: the effects of tight coupling on systems. OOP was designed to solve this problem but it hasn’t proven the panacea its early proponents had hoped for. And, as anyone who was on the Vista team can testify (for example), tight coupling remains the biggest problem we face for large software projects.

In sum, get this book. Read it and learn. It’s beautifully written and full of insights that apply not only to financial management but to “life, the universe, and everything.”