Buongiorno,

ecco l'annuncio e il link per il secondo seminario della serie.

Date: 23 July 2026, 2:00 PM

Title: A Spearman-like coefficient for upper-tail association analysis

Abstract: Quantifying association during extreme events is of paramount importance in fields such as finance, insurance, and environmental sciences. Classical correlation measures often fail to capture co-movements in the tails of distributions, while traditional boundary tail coefficients collapse under asymptotic independence, rendering them blind to sub asymptotic concentration. In this paper, we introduce a unified framework of Spearman-like, tail-sensitive indices constructed via the ranks and concomitants of the data to assess localized association conditional on large threshold exceedances. We establish two complementary families of measures: a discrete two-parameter family that encompasses Spearman's
 as a special case, and a continuous single-parameter family obtained as its asymptotic limit. This threshold-defined framework offers a flexible, interpretable diagnostic tool that continuously tracks tail-association paths and successfully discriminates among fundamentally distinct sub asymptotic structures where classical boundary indicators vanish. For both families, we investigate their theoretical properties - including their limiting behavior at the absolute boundary - propose empirical plug-in estimators, and derive their asymptotic distributions. The operational advantages of the indices are verified through analytical illustrations across diverse copula families and demonstrated via an empirical application to global catastrophic risk.
backup link: https://teams.microsoft.com/meet/324967652703789?p=0vt10YjhRZpVlDLRC1