Unit Root Tests are Useful for Selecting Forecasting Models

Francis Diebold () and Lutz Kilian ()

New York University, Leonard N. Stern School Finance Department Working Paper Seires from New York University, Leonard N. Stern School of Business-

Abstract: We study the usefulness of unit root tests as diagnostic tools for selecting forecasting models. Difference stationary and trend stationary models of economic and financial time series often imply very different predictions, so deciding which model to use is tremendously important for applied forecasters. We consider three strategies: always difference the data, never difference, or use a unit-root pretest. We characterize the predictive loss of these strategies for the canonical AR(1) process with trend, focusing on the effects of sample size, forecast horizon, and degree of persistence. We show that pretesting routinely improves forecast accuracy relative to forecasts from models in differences, and we give conditions under which pretesting is likely to improve forecast accuracy relative to forecasts from models in levels.

Date: 1999-01-18
View list of references View citations in EconPapers

Downloads: (external link)
http://www.stern.nyu.edu/fin/workpapers/papers99/wpa99063.pdf (application/pdf)

Related works:
Working Paper: Unit Root Tests Are Useful for Selecting Forecasting Models (1999)
Journal Article: Unit-Root Tests Are Useful for Selecting Forecasting Models (2000)
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: http://EconPapers.repec.org/RePEc:fth:nystfi:99-063

Access Statistics for this paper

More papers in New York University, Leonard N. Stern School Finance Department Working Paper Seires from New York University, Leonard N. Stern School of Business-
Address: U.S.A.; New York University, Leonard N. Stern School of Business, Department of Economics . 44 West 4th Street. New York, New York 10012-1126
Contact information at EDIRC.
Series data maintained by Thomas Krichel ().