Testing heteroskedasticity for predictive regressions with nonstationary regressors

摘要

This paper proposes the Cramer-von Mises type test statistic for testing heteroskedasticity in predictive regression when regressors are nonstationary. A Monte Carlo simulation study is conducted to illustrate the finite sample performance and a real empirical example is examined. (c) 2021 Elsevier B.V. All rights reserved.