International Evidence on Persistence in Output in the Presence of an Episodic Change.

This paper uses some newly developed methods and techniques to examine the dynamic properties of international output in the presence of a structural break. The author provides statistical evidence to show that the unit root test results can, in some cases, be sensitive to whether a one-time structural break in the data is modeled exogenously or endogenously. However, in most cases the unit root test results remain robust to specification of the structural break exogenously or endogenously; moreover, he finds that the null hypothesis of a unit root in output can be rejected in favor of a "flexible" trend alternative for a number of countries such as Canada, Denmark, France, and the United States. Copyright 1992 by John Wiley & Sons, Ltd.