Abstract
A novel spatial contagion measure is proposed that is based on the calculation of suitable conditional Spearman’s correlations extracted from the financial time series of interest. Algorithms for the numerical estimation of this measure are illustrated, together with a simulation study showing its features in relations with popular families of copulas. Finally, two applications are presented about the use of spatial contagion measure for determining (asymmetric) linkages in the financial systems, and creating clusters of financial time series.
In particular, contrarily to previous approaches in the literature, such clusters identify which time series increase their (positive) associative when the market is under distress. The presented methodology is also expected to be useful to select a diversified portfolio of asset returns.