http://www.hoover.org/publications/digest/13866677.html" onclick="window.open(this.href);return false;
Could such a great drain of liquidity happen again? Many financial experts dismiss the idea as mere doom mongering. A full-scale war, they say, is one of those “ten-sigma” events, events so rare that they lie outside the realm of professional risk management. A ten-dollar oil-price hike in 2007 is a risk to which a probability can be attached; big war belongs in the realm of Frank Knight’s kind of uncertainty, like an asteroid hitting the earth or a global influenza pandemic—you just can’t price it in. This line of argument recalls the philosophical point (usually associated with David Hume) about the color of swans: just because all the swans you’ve ever seen have been white doesn’t mean there’s no such thing as a black swan. By the same token, just because today’s top traders have never seen a massive liquidity crisis doesn’t mean those crises never happen. Even if those traders have survived in the bear pit long enough to have firsthand memories of 1987, their memories are simply too short. And some of their models have even shorter memories. One of the biggest defects in modern risk management is the dangerously short time horizon used in many models—such as the historical simulation models used for calculating value at risk. A number of widely used models rely on as little as the past three years of data.
