Interdependencies and Causalities in Coupled Financial Networks
I. Vodenska, H. Aoyama, Y. Fujiwara, H. Iyetomi, Y. Arai, Interdependencies and causalities in complex financial networks, PLoS ONE 11(3): e0150994. doi:10.1371/journal.pone.0150994 (2016)
30 Pages Posted: 8 Aug 2014 Last revised: 11 May 2016
Date Written: July 15, 2014
We explore the foreign exchange and stock market networks for 48 countries from 1999 to 2012 and propose a model, based on complex Hilbert principal component analysis, for extracting significant lead-lag relationships between these markets. The global set of countries, including large and small countries in Europe, the Americas, Asia, and the Middle East, is contrasted with the limited scopes of targets, e.g., G5, G7 or the emerging Asian countries, adopted by previous works. We construct a coupled synchronization network, perform community analysis, and identify formation of four distinct network communities that are relatively stable over time. In addition to investigating the entire period, we divide the time period into into "mild crisis," (1999-2002), "calm," (2003-2006) and "severe crisis" (2007-2012) sub-periods and find that the severe crisis period behavior dominates the dynamics in the foreign exchange-equity synchronization network. We observe that in general the foreign exchange market has predictive power for the global stock market performances. In addition, the United States, German and Mexican markets have forecasting power for the performances of other global equity markets.
Keywords: Complex Principal Component Analysis, Hilbert Transform, Financial Networks, Foreign Exchange, Equity Markets
JEL Classification: C21, C22, F31, G15
Suggested Citation: Suggested Citation