Cross market and cross security interdependencies in capital markets complex network based elucidation and influence relationships analysis in financial time series

Abstract

Financial risk management and portfolio management are the primary functions of newlineinvestment science. Investment decisions require identification of the sources of risks newlineand their measurement. Strategies need to be developed to address the elements of newlinerisk so as to have a minimal impact on the investable portfolio classes. The global newlinefinancial crisis in 2008 exposed the inadequacy of the earlier measures and models of newlinerisk assessment in financial markets specifically in capturing and assessing the newlinesystem-level risk in financial systems in general, and equity markets, in particular. newlineThe extant studies uniformly report that traditional financial models and measures are newlineunable to capture such system-wide risk because of their reductionist approach; thus a newlinewholistic approach to risk management is the need of the hour. newlineThe network approach derived from the mathematical principles of graph theory newlineserves to be an alternative strategy to model the system level interdependencies newlinequantitatively. This study investigates: a) the cross-market interdependencies amongst newlinethe global market indices and the Asian regional market indices, b) the cross-security newlineinterdependencies in the Indian capital market. newlineTwofold investigative approaches are utilized for analyzing the complex newlineinterdependency structure. The first strategy involves an examination of the complete newlinenetwork by deploying network topological measures and subsequently exploring newlinethese metrics at the different multi-scalar levels over a temporal axis. The second newlineapproach consists of the filtration of the network with appropriate link reduction newlinemethods leading to a reduced construction that is information-rich and noise-poor. newlineThe cross-market interdependencies among the global markets are studied on the newlineweekly return data of 43 global market indices for fourteen years duration (2002- newline2016). The study uses weighted network modeling approach and threshold filtering newlinemethodology.