VAR, error correction and pretest forecasts at long horizons
Article Abstract:
Long-horizon interval and point forecasting is a very difficult activity as a typical four-year forecast is usually constructed based on 20 to 40 years' worth of information. The difficulty often lies in the presence of persistence in the form of unit or near-unit autoregressive roots. Even by using efficient pre-testing or model-selection procedures, these difficulties are hard to overcome. As such, long-horizon and point forecasting sometimes leads to forecasts that have large asymptotic root mean squared errors and very wide prediction intervals.
Publication Name: Oxford Bulletin of Economics & Statistics
Subject: Economics
ISSN: 0305-9049
Year: 1996
User Contributions:
Comment about this article or add new information about this topic:
Specifying and testing output and pricing structures of alternative macroeconomic models
Article Abstract:
A study has been conducted to develop a framework which can be used to assist the economic modeller in making a more informed choice between the various economic models. Findings indicated that the two other important macroeconomic models, namely the imperfect substitutes model and the model with traded and non-traded goods, are special cases of the more general three-goods model. A simple two-stage test has been developed to help the macroeconomic modeller to choose the particular model structure on the basis of economic data.
Publication Name: Oxford Bulletin of Economics & Statistics
Subject: Economics
ISSN: 0305-9049
Year: 1996
User Contributions:
Comment about this article or add new information about this topic:
VAR, error correction and pretest forecasts at long horizons
Article Abstract:
An analysis is presented of methods of estimating errors in interval forecasts and pretest forecasts for long horizon forecasting in multivariate systems. Strategies include the use of vector autoregression (VAR), vector error correction model (VECM), and unit root and cointegration pretests. Numerical data suggest that, even for 10% or 20% sample size, the distortions and biases of VAR and VECM point forecasts and coverage rates, respectively, are strong.
Publication Name: Oxford Bulletin of Economics & Statistics
Subject: Economics
ISSN: 0305-9049
Year: 1996
User Contributions:
Comment about this article or add new information about this topic:
- Abstracts: The timing and value of forecast and recommendation revisions. Value creation and corporate diversification: the case of Sears, Roebuck and Co
- Abstracts: Core labour standards and export performance in developing countries. The government procurement agreement: implications of economic theory
- Abstracts: Mergers in open economies. Competition and trade in services: the airlines' global alliances
- Abstracts: Local economic development process and government policy-making: an international perspective. "Successful partners": a promotion strategy for Scarborough businesses and the city of Scarborough
- Abstracts: Help wanted: hiring a CEO for an economic development organization. A comparison of rural and urban economic development programs in Pennsylvania