Abstract
Monthly consumption forecasts for U.S. oil, natural gas, and coal are made using state space and multiple regression applied to the same data. These forecasts are compared with actual consumption for a test period. The forecasts made using state space are preferred to those made using multiple regression models for both expost and exante cases. The state space forecasts track data cycles better than do the regression forecasts. Average absolute forecast errors are less for the state space models than they are for the multiple regression models.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Aoki, M., 1987, State space modeling of time series: New York, Springer-Verlag, 314 p.
Bae, K., 1994, Energy consumption forecasting-Econometric model versus state space: Tucson, University of Arizona, Ph.D. dissertation, 179 p.
Box, G.E.P., and Jenkins, G.M., 1970, Time series analysis-Forecasting and control: San Francisco, Holden-Day, 553 p.
Business Forecast Systems, Inc., 1990, Forecast Master Plus, short run forecasting package: Business Forecast Systems, Inc., 68 Leonard St., Belmont, Mass.
Energy Information Administration (EIA), 1990, Annual outlook for U.S. coal.
Goodrich, R.L., 1992, Applied statistical forecasting: Belmont, Mass., Business Forecast Systems, Inc., 265 p.
Harvey, A.C., 1989, Forecasting, structural time series models, and the Kalman filter: New York, Cambridge University Press, 554 p.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Bae, K., Harris, D. A comparison of state space and multiple regression for monthly forecasts: U.S. Fuel consumption. Nat Resour Res 4, 325–339 (1995). https://doi.org/10.1007/BF02263380
Received:
Revised:
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF02263380