TY  - JOUR
AU  - Diebold,Francis X.
AU  - Kilian,Lutz
TI  - Unit Root Tests Are Useful for Selecting Forecasting Models
JF  - National Bureau of Economic Research Working Paper Series
VL  - No. 6928
PY  - 1999
Y2  - February 1999
UR  - http://www.nber.org/papers/w6928
L1  - http://www.nber.org/papers/w6928.pdf
N1  - Author contact info:
Francis X. Diebold
Department of Economics
University of Pennsylvania
3718 Locust Walk
Philadelphia, PA  19104-6297
Tel: 215/898-1507
Fax: 212/573-4217
E-Mail: fdiebold@sas.upenn.edu

Lutz Kilian
Dept.of Economics
University of Michigan
Ann Arbor,  MI 48109-1220
E-Mail: lkilian@umich.edu
AB  - We study the usefulness of 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.  Forecasters face three choices: 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.
ER  -