How accurate are national road traffic growth-rate forecasts?—The case of Norway

This paper evaluates the accuracy of the annual national road traffic growth-rate forecasts that are prepared by the Norwegian road authorities. The rationale for the study is the fact that national and regional traffic growth-rate forecasts are the basis for policy formation, and inaccurate forecasts may lead to inappropriate policies. The data correspond to the period of 1996–2008 whereas the forecasting models were revised in 2001. The results reveal the following: (1) traffic growth-rate forecasts were more inaccurate in the period before the revision as compared to the period after the revision and underestimation is most common; (2) the naïve growth-rate forecasts perform better than the official forecasts, but GDP growth-rate forecasts perform worse than the official forecasts; (3) the growth-rate forecasts are unbiased and efficient at the national level, but there are variations among regions and time periods. Overall, the Norwegian growth-rate forecast appears to perform fairly well in the short run, but the small inaccuracies that were observed may cause a problem in the long run, particularly in relation to the evaluation of the long-term effects of investments. We therefore advice that models need to be continually revised to accommodate more recent data such as international traffic and immigration which seem to greatly impact on the accuracy of models.

► The accuracy of national/regional road traffic growth-rate forecasts is evaluated.
► Traffic growth-rate forecasts are found to be inaccurate and are underestimated.
► The accuracy of forecasts improved after the forecasting models were revised.
► Naive forecast performs better than the official forecasts.
► The accumulative effects of the inaccuracies observed are a problem for long-term policy making.