The Failure of Risk Management, 2nd Edition

CHAPTER 9Bears, Swans and Other Obstacles to Improved Risk Management

The perfect is the enemy of the good.

—VOLTAIRE

The truth is much too complicated to allow anything but approximations.

—JON VON NEUMANN

Even if every argument I made up to this point were accepted by all managers, there would still be some serious conceptual obstacles to overcome from some corners. Risk is often perceived as beyond measurable, and any attempt to manage risk may not be considered feasible. Even the simple quantitative model I proposed in chapter 4 to replace the risk matrix will be seen as overly complex or (ironically, compared to the alternative) too simple to capture risks. Most of these objections to risk management boil down to some fundamentally different ideas about basic concepts such as the nature of models, how to assess the relative performance of models, and basic concepts of measurement and probability. We need to address these obstacles to make progress.

I argue that the probabilities and consequences of events can be measured in a meaningful way. In one of my previous books, How to Measure Anything, I argue that measurement is simply a quantitatively-expressed reduction in uncertainty based on observation. The objective of measurement is to improve (even just slightly) our current knowledge about an unknown quantity that is relevant to some decision. By this standard, the objective of quantitative assessments of risk is to improve on the unaided intuition of managers, not ...