The Evolution of Trading Rules in an Artificial Stock Market

This paper applies evolutionary modeling to expectation formation of an asset's price. As a first step, I consider a population of n investors each of whom takes on one of two possible cultural variants. Every individual is a potential role model for all other individuals and can pass on their variant with a certain probability determined by the relative return to being that type. Different types of traders employ different 'models' which forecast future price and dividend movements. With the two basic types being traders who follow the fundamentals suggested by the CAPM model and those who follow technical trading rules (such as, sell if the price is above it's 50 day moving average). I show that given these two types of simple traders, prices can fluctuate between periods of low volume and volatility and periods of high volume and volatility. Results indicate that, given a random walk fundamental valuation, as the random fluctuations increase in magnitude, technical trading can become more profitable than fundamental trading and for a period dominate the market.