Thursday, July 5, 2012

Coleman, Quantitative Risk Management

The title of Thomas S. Coleman’s book—Quantitative Risk Management: A Practical Guide to Financial Risk (Wiley, 2012) probably conjures up an image of pages of complex formulas, a book that only a true quant would ever want to read. That image would be wrong. Although the second half of the book deals with risk measurement and has its fair share of math, Coleman is writing for both those who manage risk and those who measure risk. He wants to bring managers and quants together so that each understands in plain language what the other is doing and how they both contribute to the firm’s profit and financial stability.

He argues that “managers cannot delegate their responsibilities to manage risk, and there should no more be a separate risk management department than there should be a separate profit management department.” Risk measurement may well be the mandate of a separate department, but managing risk “must be treated as a core competence of a financial firm and of those charged with managing the firm.” (p. 5)

Since Coleman is writing for practitioners, he refuses to get tripped up over theoretical niceties. For instance, he maintains that “we can often ignore any distinction between frequency-type (objective) and belief-type (subjective) probability. … The future is uncertain, subject to randomness that is not simply replication of a repeated game. But we have to make decisions, and probability theory is such a useful set of tools that we have to use it. The utility of treating frequency-type and belief-type probabilities as often interchangeable outweighs any problems involved in doing so.” (pp. 58-59)

Probability, however, is not intuitive. “The fact is that dealing with probability and randomness is hard and sometimes just plain weird.”(p. 24) There’s the classic birthday problem: What is the probability that if you enter a room with 20 people, two of them will share the same birthday? Over 41%. “And it only takes 56 people to raise the probability to more than 99 percent.” (p. 25)

Or there’s the matter of clustering, or bunching, in a random sequence that often appears to our intuition to be nonrandom. “The random shuffle on an iPod has actually been adjusted so it appears to us as more random.” (p. 25) Similarly, Bill Miller’s run at Legg Mason could easily have happened purely as a matter of luck.

In the financial world both risk and luck play critical roles. Risk, “the interaction of the uncertainty of future outcomes with the benefits and costs of those outcomes,” can be studied and modified. Luck, “the irreducible chanciness of life,” cannot be controlled, but it can be managed. (p. 64)

Coleman devotes a chapter to the “fiendishly difficult” business of managing risk, which involves managing people, processes, and projects. And, the author adds, learning more about measuring risk.

Risk comes in many stripes. Within firms (idiosyncratic risk) managers have to defend against fraud. They must also try to prevent normal business activity from going wrong. Coleman briefly describes some of best known cases of trading losses (he, of course, finished his book before the J.P.Morgan loss) and offers suggestions for combating financial disasters.

Coleman rounds out the first part of his book with an excellent chapter on practical risk techniques and a few pages on the uses and limitations of quantitative techniques. He focuses on volatility and VaR to measure the size of risk and marginal contribution and best hedges to understand the composition of risk.

And then it’s on to part two, measuring risk, which fleshes out in over 300 pages the material introduced in his chapter on practical risk techniques as well as credit risk and liquidity and operational risk.

Quantitative Risk Management is an exceedingly well crafted book that gets “the right balance between vagueness and precision.” For, as Coleman explains, “Using and understanding quantitative tools requires a careful balance between too much vagueness (where we can’t say anything useful) and false precision (where we can be very precise about things with no connection to the real world).” (pp. 142-43) Everyone who manages risk, at whatever level, could profit from reading this book.

No comments:

Post a Comment