NATIONAL BUREAU OF ECONOMIC RESEARCH
NATIONAL BUREAU OF ECONOMIC RESEARCH

A Comparison of Linear and Nonlinear Univariate Models for Forecasting Macroeconomic Time Series

James H. Stock, Mark W. Watson

NBER Working Paper No. 6607
Issued in June 1998
NBER Program(s):   EFG   ME

A forecasting comparison is undertaken in which 49 univariate forecasting methods, plus various forecast pooling procedures, are used to forecast 215 U.S. monthly macroeconomic time series at three forecasting horizons over the period 1959 - 1996. All forecasts simulate real time implementation, that is, they are fully recursive. The forecasting methods are based on four classes of models: autoregressions (with and without unit root pretests), exponential smoothing, artificial neural networks, and smooth transition autoregressions. The best overall performance of a single method is achieved by autoregressions with unit root pretests, but this performance can be improved when it is combined with the forecasts from other methods.

download in pdf format
   (1721 K)

email paper

This paper is available as PDF (1721 K) or via email.

Machine-readable bibliographic record - MARC, RIS, BibTeX

Document Object Identifier (DOI): 10.3386/w6607

Published: Published as "Evidence on Structural Instability in Macroeconomic Time Series Relations", JBES, Vol. 14, no. 1 (January 1996): 11-30.

Users who downloaded this paper also downloaded these:
Stock and Watson w7023 Forecasting Inflation
Stock and Watson w6528 Business Cycle Fluctuations in U.S. Macroeconomic Time Series
Stock and Watson w8180 Forecasting Output and Inflation: The Role of Asset Prices
Stock and Watson w6702 Diffusion Indexes
Diebold w6290 The Past, Present, and Future of Macroeconomic Forecasting
 
Publications
Activities
Meetings
NBER Videos
Data
People
About

Support
National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email: info@nber.org

Contact Us