Estimation of Spearman-type measure of local dependence

Abstract

Measures of local dependence exhibit the behaviour of dependence between random variables along their support instead of resulting in a single numerical value like the global measures of dependence or association. The aim of this work is to study the Spearman-type measure of local dependence for two continuous random variables and a bivariate stationary process. We suggest estimators and study their assymptotic properties. Simulations are performed in both cases considering components with correlation 0.70. Some empirical illustrations are provided considering two indices of performance of large companies in Brazil and for financial time series (DAX and FTSE).