This paper presents a variance decomposition method - factor analysis with Procrustes rotation - that is capable of separating the global, regional and idiosyncratic components of various financial market indicators. The method is applied to indicators of five key financial markets: sovereign CDS spreads, stock indices, exchange rates, EMBI Global bond spreads and 10-year reference yields of domestic government bond markets. The results support the finding of the literature of a significant global component in most markets, but also point out the importance of regional correlations. Based on the method two practical applications are proposed: one, which is useful in the daily monitoring of financial markets to identify magnitudes of risk premium shocks of global, regional and country-specific origins; and another one, which gauges channels of risk propagation from the eurozone periphery.
JEL: G15, C38, E44.
Keywords: variance decomposition, factor analysis, Procrustes rotation, spillover, cross-country correlations, cross-asset correlations.