Multivariate Linear and Non-Linear Causality Tests with Applications

Abstract

The traditional linear Granger test has been widely used to examine the linearcausality between any pair of time series. Hiemstra and Jones (1994) developeda nonlinear Granger causality test to investigate the nonlinear causality betweenstock prices and trading volume. In this thesis, we extend their work by developingboth linear and non-linear causality tests in multivariate settings instead of inpairwise context. We then apply the tests to identify the linear and non-linearmultivariate causality relationships among the indices of the Chinese segmentedstock markets.