Measuring association and dependence between random vectors

Measures of association are suggested between two random vectors. The measures are copula-based and therefore invariant with respect to the univariate marginal distributions. The measures are able to capture positive as well as negative association. In case the random vectors are just random variables, the measures reduce to Kendall’s tau or Spearman’s rho. Nonparametric estimators, based on ranks, for the measures are derived. Their large-sample asymptotics are derived and their small-sample behavior is investigated by simulation. The measures are applied to characterize strength and direction of association of northern and southern European bond markets during the recent Euro crisis as well as association of stock markets with bond markets.