The cornerstone of modern finance is the efficient market hypothesis. Under this hypothesis all information available about a financial asset is immediately incorporated into its price dynamics by fully rational investors. In contrast to this hypothesis many studies have pointed out behavioral biases in investors. Recently it has become possible to access databases that track the trading decisions of investors. Studies of such databases have shown that investors acting in a financial market are highly heterogeneous among them, and that heterogeneity is a common characteristic of many financial markets. The article describes an empirical study of the daily trading decisions of all Finnish investors investing Nokia stock over a time period of 15 years. The investigation is performed by adapting and using methods and tools in network science. By investigating daily trading decisions, and by constructing the time-evolution of statistically validated networks of investors, clusters of investors—and their time evolution— which are characterized by similar trading profiles are detected. These clusters are performing distinct trading decisions on time scales ranging from several months to twelve years. These empirical observations show the presence of an ecology of groups of investors characterized by different attributes and by various investment styles over many years. Some of the detected clusters present a persistent over-expression of specific investor categories. The study shows that the logarithm of the ratio of pairs of statistically validated trading decisions is different for different values of the market volatility. These findings suggest that an ecology of investors is present in financial markets and that groups of traders are always competing, adopting, using and eventually discarding new investment strategies. This adaptation process is observed over a multiplicity of time scales, and is compatible with several conclusions of behavioral finance and with the assumptions of the so-called adaptive market hypothesis.
|Numero di pagine||12|
|Stato di pubblicazione||Published - 2018|
All Science Journal Classification (ASJC) codes