Abstract:
This paper proposes the Shannon entropy as an appropriate one-dimensional measure of behavioural trading patterns in financial markets. The concept is applied to the illustrative example of algorithmic vs. non-algorithmic trading and empirical data from Deutsche Börse's electronic cash equity trading system, Xetra. The results reveal pronounced differences between algorithmic and non-algorithmic traders. In particular, trading patterns of algorithmic traders exhibit a medium degree of regularity while non-algorithmic trading tends towards either very regular or very irregular trading patterns.