A simple nonparametric test for independence

Bruce Mizrach (Department of Economics, Rutgers University)

A stationary stochastic process is defined to be locally independent if it eventually becomes independent of pastrealizations. I develop a simple nonparametric test for this condition. Size and power comparisons favor this statistic over the one proposed by Brock, Dechert and Scheinkman (1987) in samples under 250 observations.

MoreLess

Year of publication:	1995 ; This draft: January 1995
Authors:	Mizrach, Bruce Marshall
Publisher:	New Brunswick, NJ : Rutgers University, Department of Economics
Subject:	nonlinear dependence \| U-statistics \| Nichtparametrisches Verfahren \| Nonparametric statistics \| Theorie \| Theory \| Statistische Methodenlehre \| Statistical theory \| Nichtlineare Regression \| Nonlinear regression