A Measure of Distance for the Unit Root Hypothesis

Patrick Marsh ()

Discussion Papers from Department of Economics, University of York

Abstract: This paper proposes and analyses a measure of distance for the unit root hypothesis tested against stochastic stationarity. It applies over a family of distributions, for any sample size, for any specification of deterministic components and under additional autocorrelation, here parameterised by a finite order moving-average. The measure is shown to obey a set of inequalities involving the measures of distance of Gibbs and Su (2002) which are also extended to include power. It is also shown to be a convex function of both the degree of a time polynomial regressors and the moving average parameters. Thus it is minimisable with respect to either. Implicitly, therefore, we find that linear trends and innovations having a moving average negative unit root will necessarily make power small. In the context of the Nelson and Plosser (1982) data, the distance is used to measure the impact that specification of the deterministic trend has on our ability to make unit root inferences. For certain series it highlights how imposition of a linear trend can lead to estimated models indistinguishable from unit root processes while freely estimating the degree of the trend yields a model very different in character.

New Economics Papers: this item is included in nep-ecm and nep-ets
References: Add references at CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
https://www.york.ac.uk/media/economics/documents/discussionpapers/2005/0502.pdf Main text (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:yor:yorken:05/02

Access Statistics for this paper

More papers in Discussion Papers from Department of Economics, University of York Department of Economics and Related Studies, University of York, York, YO10 5DD, United Kingdom. Contact information at EDIRC.
Bibliographic data for series maintained by Paul Hodgson ().