New Unit Root Tests in the Nonlinear ESTAR Framework: The Movement and Volatility Characteristics of Crude oil and Copper Prices.

This paper proposes new wild bootstrap GLS-detrended unit root tests in the exponential smooth transition autoregressive framework, which could handle both nonlinear movement and time-varying volatility for financial time series. We derive the asymptotic distributions of the proposed tests and explore the finite sample properties. Simulation results show that the size and power performance of the proposed tests are better than the conventional tests. An application on crude oil and copper prices further underlines our test's priority, and our proposed tests do not reject the unit root null of crude oil and copper prices. [ABSTRACT FROM AUTHOR]