Saturday, 27 January 2007

Chaotic Systems

Regular readers will have concluded by now that we are not convinced as to the universal validity of neoclassical economics. The acceptance of the rationality of human beings simply does not match up with market realities.
We intend producing a series of posts in the coming months focusing on behavioural economics and how psychology drives the dynamics of all financial markets.
Firstly, we need to have a brief overview of chaotic systems.
A structured system is regarded as chaotic when the processes observed appear to be random. Chaos generally occurs via feedback loops and all complex systems are made up of a multiplicity of such loops. One must utilise non-linear mathematics to evaluate these systems as traditional economics does not accommodate the destabilising impacts of the feedback loops.
Many structures, from markets to climate, have more than one stable equilibrium position (known as bifurcations). The reason for bifurcations is that there is a sudden dramatic change in the dominance of different feedback loops. An example related to climate would be the Ice Ages. Ervin Laszlo separated out the factors that might influence feedback loops in economic systems into three categories - technical innovation, conflicts and social/ economic disturbances.
Benoit Mandelbrot introduced the idea of fractals whereby a system may continually repeat it's patterns on a range of different scales - this is called self-simulation. An example would be the coast of an island - Mandelbrot proved that such a coast is infinitely long depending upon one's perspective ie the closer one looks, the longer the coast becomes.
The behaviour of chaotic systems is difficult to analyse but it is analysable. Random walk theorists are simply wrong in their assessment of randomness. It is merely a question of developing suitable trading models in combination with holistic experience. The best traders maintain a consistent edge.
In future posts, we will demonstrate how personality disorders and styles may be built in to models for trading financial markets.