by Lance Williams, California Watch
When a route is proposed for the California bullet train, opponents mobilize.
In the Central Valley, Kings County farmers are protesting a line that would run through prime agricultural land. In Southern California, the city of Palmdale has sued to force the state High-Speed Rail Authority to build a line to serve northern Los Angeles County’s Antelope Valley.
But the bitterest dispute over the $45 billion bullet train’s route may be playing out on the San Francisco Peninsula. There, a coalition of environmentalists and local cities has gone to court twice to challenge the project.
At issue is the rail authority’s plan to connect San Francisco with the Central Valley by running trains down the Peninsula and over the Pacheco Pass, south of San Jose.
Opponents, saying they fear blight and sprawl, contend that it would be better to route the trains along the I-580 corridor in the East Bay and then over the Altamont Pass.
But in the latest lawsuit, opponents also are challenging the financial planning that underlies the entire project.
If the opponents prevail, the rail authority not only might face redoing the required environmental impact report for the project, it also could be required to re-examine the fundamental economic assumptions on which the bullet train plan is based.
Plaintiffs in the lawsuit include the cities of Palo Alto and Menlo Park, the Planning and Conservation League environmental group and the nonprofit California Rail Foundation. They won an earlier lawsuit that required the rail authority to rewrite its environmental study for the Peninsula segment of the project.
Stuart M. Flashman, lawyer for the plaintiffs, says the latest lawsuit, filed last year, cites multiple legal arguments for challenging the Peninsula route. He says the lawsuit’s critique of the rail authority’s ridership studies “is a biggie” in terms of potential impact. A Sacramento judge may rule on key aspects of the lawsuit next month.
The $2 million studies that are under dispute were commissioned by the rail authority in 2005. A consulting firm, Cambridge Systematics, polled California travelers and collected other data to create a computer model that would predict demand for the bullet train. Outside rail experts were supposed to subject the model to peer review to ensure it was valid.
When it was finished, the model predicted robust demand for bullet trains in California. At one point, the authority said as many as 100 million passengers would be riding the California bullet train by 2035; since then, projections have been scaled back, with the authority predicting 50 million passengers per year by 2030.
The results produced by the computer model also seemed to settle the argument about where the rail line should link the Bay Area with the Central Valley. The forecast showed the line would attract millions more riders if it were built over the Pacheco Pass. The rail authority had chosen that route, saying the Altamont route would face insurmountable political and environmental opposition because it would require building another bridge across San Francisco Bay.
In 2008, bullet train proponents used the projected ridership figures from the Cambridge Systematics study in their successful campaign to persuade voters to authorize issuing $9.95 billion in state rail bonds. Based on the projected ridership, the rail authority said the system would not need operating subsidies and in fact would turn a profit of $1 billion per year.
The rail authority also cited the study’s conclusions in applications for federal stimulus funds, activists say.
But in 2010, a Palo Alto citizens group called Californians Advocating Responsible Rail Design said it had discovered serious flaws in the computer model. The analysis was primarily the work of Elizabeth Alexis, the group’s co-founder.
A money manager with training in econometric studies, Alexis said she found bias in the surveys used to predict demand for high-speed rail: 96 percent of the commuters queried about whether they would prefer to get to work by bullet train already were train riders, for example. That’s a tiny fraction of California commuters, and ones who were predisposed to rail travel.
Alexis said she also found that the consultants had privately made “drastic” changes to key assumptions in the computer model used to produce its results. The changes were made after the model had been subjected to peer review. The outside experts who were supposed to provide guidance on whether the study was valid never saw the final version, she claimed.
Some of the model’s projections were counterintuitive: According to the model, bullet train stations in Gilroy, south of San Jose, and Merced, in the Central Valley, would each have more than triple the ridership of the station in San Francisco and more than 15 times the ridership of the station in Los Angeles.
Compared with a 2000 ridership forecast done for the rail authority, the computer model gave a boost to ridership projections for the Pacheco Pass route, while deflating projected ridership on the Altamont Pass route, said Richard Tolmach, president of the California Rail Foundation.
He said he suspected consultants had “gamed” the model to produce desired results.
“I think they were tweaking it to pump up the numbers for Pacheco Pass,” he said.
For an independent review of the computer model, the plaintiffs in the lawsuit retained their own expert, Norman Marshall of Vermont-based Smart Mobility Inc.
In his report, he said he found grave problems, calling the ridership and revenue forecasts in the rail authority’s study invalid because of technical mistakes, court records show. Marshall also called the consultants’ description of their methodology “incoherent.”
UC Berkeley’s Institute of Transportation Studies also looked at the issue. That study, commissioned by the state Senate, declared that the rail authority’s model was “flawed at key decision-making junctures,” as institute Director Samer Madanat put it. Like the citizens group, the UC study found survey bias, saying that in polling on whether business travelers would switch to bullet trains, too many air travelers and not enough motorists were surveyed.
The model also exaggerated the importance of frequent service in predicting ridership, the institute said. In the end, the study was too unreliable to predict whether high-speed rail would “incur significant revenue shortfalls,” the institute said.
The consultants disputed the criticism, contending that their methodology was sound and the forecasts valid.
“We did not change key parameters to accord with our a priori assumptions,” wrote Lance Neumann, Cambridge Systematics' board chairman. Roelof van Ark, chief executive of the rail authority, wrote that the model was valid, “a sound tool for use in high-speed rail planning.” He said there was “no foundation” for the institute’s conclusion that the model was useless for predicting the bullet train’s bottom line.
But state Sen. Alan Lowenthal, D-Long Beach, chairman of the Senate committee that commissioned the report, told the Los Angeles Times that the Berkeley institute’s conclusions were “very damning.”
In court, the rail authority has continued to defend the accuracy of its computer model. The authority says it is legally entitled to rely on the model to predict bullet train ridership. Those conclusions are a component of the environmental studies that must be completed for the project to proceed. The plaintiffs are asking Sacramento County Superior Court Judge Michael Kenny to order the rail authority to redo its environmental studies.
Lance Williams is an investigative reporter for California Watch, a project of the non-profit Center for Investigative Reporting. Find more California Watch reporting here.