Accuracy and efficiency in the U.S. Department of Energy's short-term supply forecasts

Abstract

Listen

Translate article

One-step-ahead forecasts of quarterly crude oil, natural gas, electricity, and coal supplies are evaluated under two general approaches: accuracy-based measures and classification- or directional-based measures. Results suggest the U.S. Department of Energy (DOE) supply forecasts for U.S. domestic energy products are generally more accurate than a naïve alternative. There is only limited evidence of bias and inefficiency in the forecasts; although there is some evidence of error repetition. Directional forecasts for supply changes are statistically better than random, but they generally do not outperform a naïve forecast.