Sunday, June 20, 2010

Fat tails


Reading Niall Ferguson's narration on the history of finance, I saw many almost con-like well-planned inventions that have advanced the system of money to more sophisticated levels, and at the same time devastating moments of human spontaneity that shook our world with bubble bursts and crises. The stock markets, for example, are mirrors of an amplified tendency of overreacting. When prices start to go up, people rush to go in and buy more as if possessed by a collective euphoria - what Alan Greenspan called "irrational exuberance." But if any bad news surfaces, people can flip overnight from greed to fear, selling and withdrawing, causing a dramatic plunge on a global scale.

In statistics, the graph of a "normal distribution" looks like a classic bell curve, with higher probability clustered around the mean and fewer instances towards the extremes. Many natural and physical phenomena, such as human heights and laser light intensity, seem to follow this principle. But the movements of stock market prices are more results of human emotional volatility than rational science of "normal." Prices can surge up steeply one day, and drop with extreme abruptness the next. Statisticians call this distribution with higher likelihood at the extremes "fat tails."

Fat tails imply risks. Things can go extraordinarily well, or terribly wrong. And it's hard to predict. Today you have a winner, and tomorrow you could have a crisis. Impulsive decisions and mood swings push things to extremes, jumping inconsistently between one end to an other. It is almost impossible to understand or follow or react. That's why the rocket science of the Black-Scholes pricing model did not succeed. Maybe the only way to deal with subjective irrationality is guesswork, which by definition gives you 50% chance.

Some say stress or anxiety is the source of poor decisions. People under stress may swing between the poles of mania and depression, suffer from perceptual narrowing that prevents them from seeing the big picture, dramatize trivial happenings that should be expected normally, or even distort reality through denial and fabrication. Is there a way to pull the fat tails back to the mean? I would say: "Calm down."

No comments: