Wednesday, January 28, 2009

Tail You Lose (Big Time)

In one article in their special report on the future of finance, The Economist, describe why the quants' model failed to foresee the recent crisis as follows:
In markets extreme events are surprisingly common—their tails are “fat”. Benoît Mandelbrot, the mathematician who invented fractal theory, calculated that if the Dow Jones Industrial Average followed a normal distribution, it should have moved by more than 3.4% on 58 days between 1916 and 2003; in fact it did so 1,001 times. It should have moved by more than 4.5% on six days; it did so on 366. It should have moved by more than 7% only once in every 300,000 years; in the 20th century it did so 48 times.
And the Catch-22 is
On the one hand, you cannot observe the tails of the VAR curve by studying extreme events, because extreme events are rare by definition. On the other you cannot deduce very much about the frequency of rare extreme events from the shape of the curve in the middle.
Modern finance may well be making the tails fatter, says Daron Acemoglu, an economist at MIT. When you trade away all sorts of specific risk, in foreign exchange, interest rates and so forth, you make your portfolio seem safer.
To make the point clear, look at the graph on that article.

One of my old teachers, when discussing economic model for developing countries, once said that sometimes the real story lies in the outliers. I think he made a good point, and I should add the modern advanced finance too.


  1. Rizal, thanks for the link. I'm thinking that making a distribution of monthly returns, weekly returns, or daily returns, or even half-day returns or hourly returns would produce different VaR. For example daily returns, Dow open at 8100 up to 8200 in midday and then close at 8000, the 100 loss doesn't have to be observed. we can pick the 100 profit up to midday to be observed and put in the distribution. same as weekly returns, Monday start at 8100, Wed goes to 8300, Fri down to 8000. we can choose the 200 up to Wed to be taken as sample. Some traders don't trade all day/week long i think. Any suggestion for a new model?

  2. Anymatters, finance is the subject that I know next to nothing, thus I am afraid I can be of no help here. But I am sure some of our cafe regulars might have something to say --or perhaps even yourself do -- on this intriguing yet recently relevant stuff.

    Please feel free to chip in, folks.