Accuracy of past projections of US energy consumption

• Published in: Energy policy Vol. 33; no. 8; pp. 979 - 993
• Main Authors: O’Neill, Brian C., Desai, Mausami
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01.05.2005; Elsevier; Elsevier Science Ltd
• Series: Energy Policy
• Subjects: Accuracy; Applied sciences; Consumption; Economic data; Energy; Energy consumption; Energy economics; Energy policy; Exact sciences and technology; Forecasts; General, economic and professional studies; Policy studies; Projections; Studies; U.S.A; Uncertainty; United States > US; USA; Consumption; Uncertainty; Projections; Energy; Energy consumption; Evaluation; Environmental policy; Priority; Gross domestic product; Energy intensity
• ISSN: 0301-4215; 1873-6777
• DOI: 10.1016/j.enpol.2003.10.020

Energy forecasts play a key role in development of energy and environmental policy. Evaluations of the accuracy of past projections can provide insight into the uncertainty that may be associated with current forecasts. They can also be used to identify sources of inaccuracies, and potentially lead to improvements in projections over time. Here we assess the accuracy of projections of US energy consumption produced by the Energy Information Administration over the period 1982–2000. We find that energy consumption projections have tended to underestimate future consumption. Projections 10–13 years into the future have had an average error of about 4%, and about half that for shorter time horizons. These errors mask much larger, offsetting errors in the projection of GDP and energy intensity (EI). GDP projections have consistently been too high, and EI projection consistently too low, by more than 15% for projections of 10 years or more. Further work on the source of these sizable inaccuracies should be a high priority. Finally, we find no evidence of improvement in projections of consumption, GDP, or EI since 1982.