Quantitative analysis of non-exchangeability in bivariate copulas: Sharp bounds, statistical tests and mixing constructions
Manuel Úbeda-Flores
Abstract
A bivariate random vector $(X,Y)$ is exchangeable if $(X,Y)$ and $(Y,X)$ share the same distribution, which in copula terms amounts to $C(u,v)=C(v,u)$. Building on the axiomatic framework of [F. Durante, E.P. Klement, C. Sempi, M. Úbeda-Flores (2010). Measures of non-exchangeability for bivariate random vectors. Statistical Papers 51(3), 687--699], we develop three original contributions. We derive sharp upper bounds on the non-exchangeability measure $μ_p(C)$ in terms of the Schweizer and Wolff dependence measure and Spearman's $ρ$. We prove the exact scaling identity $μ_p(αC+(1-α)C^t)=|2α-1|\,μ_p(C)$ for all $p\in[1,+\infty]$, enabling explicit prescription of any target degree of non-exchangeability. Finally, we propose and analyse a nonparametric permutation test for $H_0:C=C^t$, prove its consistency, and validate its finite-sample performance via Monte Carlo simulation.