Has the accuracy of energy demand projections in the OECD countries improved since the 1970s?

Since the 1970s almost all OECD countries have published projections or forecasts of future energy consumption. By now, three decades later, the actual values of energy consumption are available for the same number of countries and thus a considerable amount of empirical data is available concerning formal hypothesis testing, e.g. whether there have been improvements in the forecasting accuracy during this period. Using data for sixteen OECD countries the empirical evidence weakly favors the hypothesis that these countries have made some advances in forecasting accuracy concerning projections at the aggregate level of energy consumption and to a lesser degree at sectoral levels. Also, in accordance with a priori expectations, the forecasting failure is increasing with the length of the forecasting horizon.