The Role of Risk Aversion and Technical Trading in the Behaviour of Financial Markets

Abstract

In this paper we show how agent based social simulation helps us to improve some of the traditional models and theories in financial economics. In particular, we explore the links between the micro-behaviour of investors and the aggregated behaviour of Stock Markets. First, we build an agent based model of an artificial financial market, populated only with rational investors. We observe that the statistical features of this market are in agreement with the theoretical markets suggested by mainstream financial economics, but far away from the features shown by real financial markets, like the Spanish Ibex-35, the Spanish Stock Market main Index. In order to fill the gap, we introduce heterogeneity in the model. We add psychological investors, as suggested by Kahnemen and Tversky (1979), and we are able to reproduce non-normality, excess kurtosis, excess volatility, and volatility clustering. Then, we introduce technical traders, and we also get from the model higher levels of excess volatility and unit roots. In other words, psychological dealers seem to be responsible for volatility clustering, whereas technical traders trend to introduce unit roots into the process. All these “financial patterns” are a common feature not only for Spanish Ibex-35, also the most important stock markets. We conclude that agent based social simulation helps us to fill the gap between economic theory and real markets, as we explain the statistical features of financial time series from the bottom-up.