Inside the Black Box: The Simple Truth About Quantitative Trading

11.2. QUANTS CAUSE MORE MARKET VOLATILITY BY UNDERESTIMATING RISK

This criticism contains components of truth and of falsehood. Many managers, quants included, are subject to a fundamental type of model risk we discussed in the last chapter, namely asking the wrong questions and using the wrong techniques. Techniques such as VaR, for example, make numerous wrong assumptions about the market in an effort to distill the concept of risk down to a single number, which is a goal that seems mostly pointless. Furthermore, as illustrated by the August 2007 quant liquidation crisis, quants have underestimated the downside risk of being involved in large-scale, crowded trading strategies. This, too, stems from a fundamental flaw of quantitative trading. Computers can be given a problem that is badly framed or makes too many assumptions, and they can come up with an answer that is both highly precise and entirely wrong. For example, I can drum up a model of my wealth that assumes that this book will sell 50 million copies, that I will receive 50 percent of the proceeds, and that I can then invest the proceeds into a vehicle that will earn 100 percent per year, compounded, forever. With this model I can get precise answers to the question of my earnings as far into the future as I want. However, all my assumptions are highly suspect, at best.

The computer's job is not to judge my assumptions, so this kind of error is ultimately attributable to my poor judgment. Similarly, some quants can ...