In Brief:
The future is either going to be very bleak, surprisingly OK, or, in all likelihood, an unpredictable mixture of the two.
That’s the hazy prognosis that many transit agencies are making these days when it comes to the future of ridership. Early predictions of a relatively quick rebound from the steep ridership losses of the pandemic have since given way to murky projections of a “new normal” for many transit agencies — with many predicting much lower ridership long term.
Planners at Bay Area Rapid Transit in San Francisco, which is anticipating one of the most challenging budget gaps in the country, are projecting upside, downside and “base case” scenarios for ridership over the next decade — with even the rosiest predictions only hitting about 80 percent of the pre-pandemic forecast.
The Washington Metropolitan Area Transit Authority is anticipating no more than 75 percent of pre-pandemic ridership in fiscal year 2025. After the Metropolitan Transportation Authority (MTA) in New York hired the consulting firm McKinsey & Company to predict ridership in November 2020, ridership figures closely tracked the firm’s most optimistic scenario at first, but dropped off after the omicron variant began to spread. The firm revised its predictions last summer, with MTA ridership now projected to be somewhere between 73 percent and 88 percent of pre-pandemic levels by the end of 2026.
Transit agencies try to predict future ridership for all types of scenarios — not just when making their yearly budgets but also when considering the impact of service changes or fare increases, and when applying for funding to build out new bus and train lines. It’s a practice with a wide range of methodologies, from sophisticated formulas to rough rules of thumb. While the upheavals of the pandemic haven’t fundamentally altered the way predictions are made, experts say, they have highlighted a need for new data on how people live and get around their cities.
How to Predict the Future
There are two schools of thinking about ridership projections, says Carole Turley Voulgaris, an assistant professor of urban planning at the Harvard Graduate School of Design. On one side there’s a “naive, technocratic” belief that the future of transit ridership can be predicted objectively, without bias, and more or less accurately. On the other side is the view that “a skilled modeler can produce almost any result they want,” and that modelers often produce forecasts in order to justify a policy decision that’s already been made, she says.
“The truth, maybe, lies somewhere between them. It kind of just depends on how cynical you want to be,” Voulgaris says.
Most projections rely on a variety of data inputs, incorporating things like population demographics, household incomes, rates of car ownership, land-use patterns, building permits and other factors. The inputs are evolving.
Some things, like the price of gas, now seem to have less of an impact on transit ridership than they did before the pandemic, says Matthew Dickens, director of policy development and research at the American Public Transportation Association, which monitors weekly ridership trends through a partnership with Transit App.
Other inputs, like office occupancy rates, are becoming much more important. There’s also a chicken-and-egg dynamic with ridership and service levels, Dickens says, with likely declines in ridership if budget-strapped agencies are forced to make service cuts.
Flavia Tsang, who works as a travel modeler for the Metropolitan Transportation Commission (MTC) in the San Francisco Bay Area, says the pandemic has ushered in all sorts of changes in social behavior, and it’s not yet clear which changes are permanent and which are temporary.
To get more information about how travel demand is shifting, MTC is planning to survey riders of all 24 transit operators in the Bay Area about how and when they use buses, trains and ferries, Tsang says. In the past those comprehensive surveys have taken half a decade or more, she says, but the commission is hoping to finish its next survey in under a year.
“We don’t have a lot of data about this new travel behavior,” Tsang says. “We need the data to inform our models so our models can inform policy.”
Are Ridership Predictions Accurate?
Long-range traffic and transit projections are often required for federal grant programs, which are critical funding for most big projects. But projections made in decades past have often proven to be wildly off-base. Tsang says that one of her mentors worked as a modeler for half a century, and most of the projections he made in the 1970s were wrong in part because they didn’t anticipate the big increase in women participating in the labor force.
“I take that as a good reminder that there are probably some unknown unknowns out there that could significantly change people’s travel behavior,” Tsang says.
Ridership projections are still imperfect, but they’re becoming more accurate. Overall, long-range projections used in federal grant applications have been about 25 percent more optimistic than actual ridership turns out to be, says Jawad Hoque, who recently wrote a dissertation on the accuracy of ridership projections and who now works in transportation modeling. But the projections have tended to be much more accurate since 2000.
Starting around 2006, transit agencies that receive federal New Starts grants, one of the main sources of transit funding in the U.S., have been required to produce before-and-after studies comparing their projections to actual ridership after a new service starts. That has helped modelers get a better gut sense of when a projection is reasonable versus when it’s obviously off-base, says Voulgaris.
In some ways, the roughly sketched scenarios that many transit agencies are projecting in their budget documents have an advantage over highly complex processes that are used to generate a single prediction of the most likely future scenario, she says. Simple rules of thumb that can be used to predict a range of outcomes are more useful in an environment with so much uncertainty.
“The days of pretending like we know what’s going to happen in 50 years, and planning for that one thing, probably need to be over,” Voulgaris says. “The focus of our planning should be: Let’s create a system that will serve our needs under a wide variety of possible alternatives.”
Related Articles
