The accuracy of toll road traffic forecasts: An econometric evaluation James Odeck a,b,⇑ , Morten Welde a a Department of Civil and Environmental Engineering, NTNU – Norwegian University of Science and Technology, Trondheim, Norway b Department of Economics and Logistics, Molde University College, Molde, Norway article info Article history: Received 28 May 2016 Received in revised form 29 April 2017 Accepted 2 May 2017 Available online 11 May 2017 Keywords: Econometrics Toll traffic forecasts Accuracy Biasedness Efficiency Norway abstract This study assesses the accuracy of toll road traffic forecasts in Norway. This study con- tributes to extant literature regarding transportation economics and policy because few studies have analyzed the accuracy of toll road traffic forecasts and current studies on this topic do not use a succinct econometric framework to infer the forecast’s bias and effi- ciency, which are determinants of the accuracy of forecasts. The data for this study include 68 toll road projects where forecasts and actual out-turns were available. All projects that are included in this study were opened for traffic during 1975–2014. Transport models were made mandatory for forecasting toll traffic in 2006. The results of this study reveal the following: (i) Norwegian toll road traffic forecasts are underestimated but are close to accurate because the mean percentage error is a mere 4%. This result sharply contrasts international studies that resulted in large overestimations at more than À20%. (ii) The accuracy of forecasts has not improved since transport models became mandatory. (iii) The Norwegian toll road traffic forecasts are unbiased, which implies that they do not tend to be one-sided. (iv) Norwegian forecasts are efficient, which implies that available infor- mation is used adequately when making forecasts. Our conclusion is that the toll traffic forecasts in Norway perform fairly well compared to the forecasts of other countries. In addition, we recommend that the assessment of forecasts in the transportation sector should be based on succinct econometric frameworks; otherwise, the conclusions may lack important information regarding the bias and efficiency of the forecasts. Ó 2017 Elsevier Ltd. All rights reserved. 1. Introduction Traffic forecasts are a major determinant of transportation policies. Before regulating authorities nominate any trans- portation development for approval, they require an assessment of the potential impacts, particularly regarding future traffic volumes, which determine the financial and/or socio-economic merits of the project. Governments rely heavily on these assessments when making policy decisions. The importance of accurate and unbiased traffic forecasts is obvious; over- or underestimating traffic may severely bias the estimates on which government decisions are based. Overestimating traffic may overestimate the social surplus derived from the cost-benefit analyses that is used in the appraisal of potential projects. If the government had known the actual traffic volume and the actual net present value, it may have opted for other more viable options or implemented the same project in a different form. Conversely, underestimation may represent a case of http://dx.doi.org/10.1016/j.tra.2017.05.001 0965-8564/Ó 2017 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. Tel.: +47 480 25 940. E-mail addresses: james.odeck@vegvesen.no, james.odeck@ntnu.no (J. Odeck). Transportation Research Part A 101 (2017) 73–85 Contents lists available at ScienceDirect Transportation Research Part A journal homepage: www.elsevier.com/locate/tra