A Copula Nonlinear Granger Causality

被引：18

作者：

Kim, Jong-Min ^{[1
]}

Lee, Namgil ^{[2
]}

Hwang, Sun Young ^{[3
]}

机构：

[1] Univ Minnesota, Div Sci & Math, Stat Discipline, Morris, MN 56267 USA

[2] Kangwon Natl Univ, Dept Informat Stat, Chunchon, Gangwon, South Korea

[3] Sookmyung Womens Univ, Dept Stat, Seoul, South Korea

来源：

ECONOMIC MODELLING | 2020年 / 88卷

基金：

新加坡国家研究基金会;

关键词：

Copula; Granger causality; Directional dependence; Permutation test; BETA REGRESSION; GUIDELINES;

D O I：

10.1016/j.econmod.2019.09.052

中图分类号：

F [经济];

学科分类号：

02 ;

摘要：

We propose a new copula nonlinear Granger causality test that is more robust than the current available linear and nonlinear Granger causality tests when there exists an asymmetric and nonlinear directional dependence. To perform the statistical test of the copula nonlinear causality, the Gaussian Copula Marginal Regression (GCMR) model and copula directional dependence (Kim and Hwang, 2017) are employed in this paper. By using GCMR and two-sample permutation test with rank sum statistic for the copula nonlinear Granger causality, we can confirm that the result of the proposed copula nonlinear Granger causality test is a reliable test through the simulated data and real data both for small and large sample sizes.

引用

页码：420 / 430

页数：11

共 25 条

[1]

[Anonymous], 1992, WORKING PAPER

[2]

[Anonymous], [No title captured]

[3]

Cherubini U., 2011, Dynamic copula methods in finance

[4]

Cribari-Neto F, 2010, J STAT SOFTW, V34, P1