Trust in forecasts? Correlates with ridership forecast accuracy for fixed-guideway transit projects

The accuracy of ridership forecasts for fixed-guideway transit projects in the United States has improved in recent decades. A better understanding of the causes for this improvement can help decision makers, project evaluators, and other forecast users identify ridership forecasts that are most likely to be reliable. The analysis in this paper applies a series of linear regression models to evaluate the relationship between ridership forecast accuracy for 67 New Starts projects completed between 1983 and 2012 and four types of project characteristics: time between forecast and observation, local experience with the project mode, physical characteristics, and financial characteristics. The results indicate that local experience and financial characteristics (including the share of a project's costs funded by federal grants) are not significantly related to forecast accuracy, but there are differences by project mode, where forecasts for commuter rail projects are less accurate than those for other modes. The time until ridership observation does relate to forecast accuracy. However, not all of this elapsed time is important. The length of time required for project planning and development does not have a significant relationship with forecast error, nor does the total time between forecast preparation and ridership observation. However, the length of time required to construct the project is significantly associated with the accuracy of the ridership forecast. These results can help planners, policy makers, and other decision makers make judgments about the degree of trust they should place in transit ridership forecasts.