Search This Blog

Wednesday, June 9, 2010

The Evolutionary Algorithm & Complexity Economics

I just read this book - The Origins of Wealth by Eric Beinhocker. Thanks to Steve, CEO of an up and coming company, for bringing it to my attention. The books contents relate to my last article on genetic evolution and risk management albeit treating it in much more depth (and richness) as only a 450 page book can. I now feel vindicated in a sense that my ideas were not full tangent and there are more people seeing this as an entirely conceivable approach.

The book in itself  deals with complexity economics - a view that is in stark contrast with the static equilibrium views propounded in most contemporary economic texts and presents the economy as an dynamic evolutionary system that mirrors biological evolution. The key take away is - presentation of the economy as a combination of physical technologies (PT), social technologies (ST) and business plans (likened to biological DNA) that evolve to adapt to fitness funtions driven by participant choices. In this manner the economy evolves in a fashion very similar to biological evolution and the most efficient algorithm to evolve happens to be - as anyone can by now guess - the evolutionary algorithm driven by continuous micro risk taking. Apparently this happens to be the most efficient algorithm for searching roughly correlated  large design spaces (something the economic landscape subscribes to) for reaching the best possible fit.

Eric relates the economic system to a thermodynamic system. He deems wealth creation to be a consequence of  irreversible processes that reduce the local entropy of the economy.Wealth in essence is treated as a fit indicator and wealth creation is determined by how closely a business is able to match the fitness function. For the financial practiioners Eric's got some news - markets are not random {yes thats right NOT random). Financial markets are dynamic systems that are complex enough to give the appearance of randomness. This is evident in the applicability of power laws to return distributions and temporal variations that resemble punctuated equilibrium rather than covariance stationary processes. This part is identical to Nassim Taleb's ideas on turkey economics and Mandelbrot's fractals and invariance of time scales.

Overall it was an excellent read - covers a lot of ground on bleeding edge modern and future thinking. I recommend this book to any one with an interest in strategy, finance or economics.