What Is a Forecast for?

Abstract

Problem, research strategy, and findings: The forecasts transit agencies submit in support of applications for federal New Starts funding have historically overestimated ridership, as have ridership forecasts for rail projects in several countries and contexts. Forecast accuracy for New Starts projects has improved over time. Understanding the motivations of forecasters to produce accurate or biased forecasts can help forecast users determine whether to trust new forecasts. For this study I interviewed 13 transit professionals who have helped prepare or evaluate applications for federal New Starts funds. This sample includes interviewees who have had varying levels of involvement in all 82 New Starts projects that opened between 1976 and 2016. I recruited interviewees through a snowball sampling method; my interviews focus on the interviewees’ perspectives on how New Starts project evaluation and ridership forecasting has changed over time. Interview results suggest that ridership forecasters’ motivations to produce accurate forecasts may have increased with increased transparency, increased influence on local decision making, and decreased influence on external (federal) funding.

Takeaway for practice: Planners can evaluate the likely trustworthiness of forecasts based on transparency, internal influence, and external influence. If forecast users cannot easily determine a forecast’s key inputs and assumptions, if the forecaster has been tasked with producing a forecast to justify a predetermined action, and if an unfavorable forecast would circumvent decisions by the forecaster’s immediate client, forecasts should viewed with skepticism. Planners should seek to alter conditions that may create these conflicts of interest. Forecasters seem to be willing and able to improve forecast accuracy when the demand for accurate forecasts increases.

Keywords:

• ethics
• forecasting
• public transportation
• transit ridership

ACKNOWLEDGMENTS

I thank Editor Ann Forsyth, Sandra Rosenbloom, Carol Ann Litster, and three anonymous reviewers for their thorough and helpful comments on early drafts of this article.

RESEARCH SUPPORT

This research was funded by grants from the University of California Center on Economic Competitiveness in Transportation, the University of California Institute of Transportation Studies Mobility Research Program, and the UCLA Graduate Division.

SUPPLEMENTAL MATERIAL

Supplemental data for this article can be found on the publisher’s website.

NOTE

NOTE

Notes

1 Boyle’s (2006) findings on the use of four-step travel demand models would also apply to activity-based travel demand models, which have become more common since the publication of his findings because activity-based models are even more resource intensive than four-step travel demand models.

Additional information

Notes on contributors

Carole Turley Voulgaris

CAROLE TURLEY VOULGARIS ([email protected]) is an assistant professor of urban planning at the Harvard Graduate School of Design.