24 October 2010

Market (Mis)Behavior ~ Far From Random Walk...

Thanks to the FT's Lex column for noting Benoît Mandelbrot's inconvenient truth, that market behavior is far from a random walk...
"The work of the mathematician who disproved the precepts of the efficient markets model half a century ago, well before financiers had started to bet huge sums on products derived from that model, suddenly regained attention once those bets failed in 2007 and 2008. His insights should have been devastating. The efficient markets hypothesis, and with it modern portfolio theory and the Black-Scholes model for pricing options, all assume that markets reflect all known information and follow a “random walk”, like coin tosses or Brownian motion. That implies that returns should follow the “bell curve” distribution often found in the natural world. But, as the charts show, extreme outliers in currency and stock markets are far more common than the coin-toss model would predict. These outliers make up the bulk of long-run returns. [...] His idea that markets could only be modelled with complex mathematical techniques that do not yet exist. In the absence of more research, his ideas imply an imprecise approach to risk management. Markets do indeed behave as if they are efficient for long periods. Investors can be excused for ignoring Mandelbrot’s ideas but in future they must accept that risk cannot be measured precisely and that “fully invested” will mean holding a higher proportion of cash. Academic economists’ refusal to acknowledge him was scandalous. He believed this was because his ideas meant “a great amount of work, trouble and effort”, while the efficient markets literature promised “capital on which one could live for a while”. Several economists won Nobel prizes by living on that capital: assuming market efficiency."

No comments: