HOW TO TRADE USING THE (SHANNONIAN) TRANSFER ENTROPY? AN APPLICATION FOR CENTRAL AND EASTERN EUROPEAN MARKETS

Transfer entropies have been frequently used in the quantification of statistical coherence between various time series with prominent applications in financial markets. The authors consider that (Shannonian) transfer entropy may be used to describe the functional linkages between financial markets, to detect various types of asymmetries in the interaction between two systems and, consequently, to distinguish between driving and responding forces in such interactions. This last property has relevance in trading diversified portfolios on multiple markets. The authors illustrate such application for a group of six Central and Eastern European markets. The results show that for a time span between 2001 and 2012, these markets are dominated from the viewpoint of net information flows by WIG and S&P 500 indexes. The article reports the performances for individual portfolios which mimic the indexes' structure as well as for a global portfolio with weights based on individual net information flows.