Fat tails & why CAPM is bunk
Mandelbrot does a great job illustrating how traditional Gaussian models of financial markets don’t apply, and that in reality markets follow classical fractal models (chaotic).
He takes two approaches:
1. Statistical: For example the 1987 market collapse represents a change equivalent to 22 standard deviations. The odds of that happening are 10^50. In fact there were many swings greater than 5 standard deviations - way more than modern finance theory allows.
2. Graphical: He uses graphs and charts extensively to clearly illustrate what a proper Gaussian market would look like and what actual markets look like.
Although it has become widely accepted that markets are not a truly ‘random walk’, much of modern finance theory is based on this critical assumption. Everything from options pricing through to the capital asset pricing model (CAPM).
Most corporate investment decisions today are made from CAPM calculations that depend on Beta, an improper measure of risk. Mandelbrot’s book is a call to action of the finance research establishment to abandon the normal distribution and rebuild their theories with a more accurate depiction of the behavior of markets.
Well-written and well-worth a read for all involved professionally or personally in financial markets - they are far riskier than we have imagined.
