Abstract
In this paper, we propose an alternative Lagrange multiplier test for volatility interactions or causality in conditional variance in the multivariate GARCH models with constant conditional correlations. Although a similar test has recently been suggested by the authors, the test necessitates estimation of the constant conditional correlation GARCH model. Our new test, on the other hand, can be computed only through univariate GARCH estimations. In addition, a robust version of the new test is provided. Finite sample properties of the new test are investigated through Monte Carlo simulations. The results show that the new test has reasonable size and power properties under the normally and leptokurtotically distributed innovations as well as under changing conditional correlations. Usefulness of the new test is illustrated by empirical examples.