Statistical Methods in Finance 2018

Dec 17 - 20, 2018


Mesoscopic Financial Network: Sectoral Co-movements and Core-periphery Structure

by Anirban Chakraborti

"Network analysis has become a primary tool in fields as diverse as systems biology, ecology, epidemiology, sociology, economics and finance. We present the work that demonstrates the existence of an empirical linkage between nominal financial networks and the underlying economic fundamentals, across countries [1]. To study the topology of the return correlation network at mesoscopic level, we constructed correlation matrices from sectoral indices for 27 countries, and applied two commonly used clustering algorithms, viz., minimum spanning tree (MST) and multi-dimensional scaling (MDS), to group sectors based on their co-movements. The influence of the sectors in the mesoscopic network was found using the eigenvector centrality (EVC). We proposed a method to find a binary characterization of the 'core-periphery' structure by using a modification of the EVC. We showed that those sectors identified as core by the centrality measure, also constitute the backbone of the MST and cluster very closely in the MDS maps, thereby confirming the robustness of our method. We also studied the sectoral dynamics, and found that the core-periphery structure had at least two sectors in the core for all countries. But the core-periphery structure does change over time; so we compared the structures for the periods of market turbulence (2008-09) as well as periods of relative calmness (2012-13 and 2015-16). For most of the countries, Industries (ID) and Finance (FN) remained as core sectors. To establish the connection between the financial network and the underlying production process, we regressed the EVC on macro-variables: market capitalization, revenue and employment, all aggregated at the sectoral level. The results were reasonably robust with varying degrees of prosperity. We also present some related work and extensions of these analyses at the micro level (stock wise) [2] and macro level (index wise) [3].

[1] K. Sharma, B. Gopalakrishnan, A. S. Chakrabarti and A. Chakraborti, Scientific Reports, 7 (2017) 8055.
[2] K. Sharma, S. Shah, A.S. Chakrabarti and A. Chakraborti, Proc. of Economic Foundations for Social Complexity Science, (2017) 211-238.
[3] K. Sharma, A.S. Chakrabarti and A Chakraborti, "Multi-layered network structure: Relationship between financial and macroeconomic dynamics", to be published in Eds. F. Abergel et al., New Perspectives and Challenges in Econophysics and Sociophysics (Springer, Milan); available at arXiv:1805.06829 (2018). "