Forecasting inaccuracies: a result of unexpected events, optimism bias, technical problems, or strategic misrepresentation?

Abstract

Based on the results from a questionnaire survey and qualitative interviews among different actors involved in traffic forecasting, this paper discusses what evidence can be found in support of competing explanations of forecasting errors. There are indications that technical problems and manipulation, and to a lesser extent optimism bias, may be part of the explanation of observed systematic biases in forecasting. In addition, unexpected events can render the forecasts erroneous, and many respondents and interviewees consider it to be simply not possible to make precise predictions about the future. The results give rise to some critical reflections about the reliability of project evaluations based on traffic forecasts susceptible to several systematic as well as random sources of error.

Keywords

• Forecasting
• Inaccuracy
• Bias
• Modeling