Bay Area transit planners want to convert all of the region's carpool lanes into toll lanes. They say that toll lanes will not only ease traffic congestion, but also generate much-needed funds for road and mass transportation projects. They estimate that turning the Bay Area's 500 miles of carpool lanes into toll lanes, along with adding 300 miles of new toll lanes, will generate $6.1 billion in profit over the next 25 years. So far, the plan, which is now in the state legislature, has generated little opposition. But one UC Berkeley engineering professor who has studied the effectiveness of carpool and toll lanes says the idea is wildly optimistic, and he predicts it will ultimately waste huge amounts of taxpayer money.
For years, professor Pravin Varaiya of UC Berkeley's electrical engineering and computer sciences department has pored over extensive amounts of CalTrans data. He concludes that the new toll lanes will lose money for two main reasons. In less-congested areas, not enough people will use them. And on the Bay Area's more-congested freeways, heavy demand from carpoolers won't leave enough room for those single-occupancy vehicles that would pay the new toll. "I'm willing to bet that the toll lanes won't be able recover operating costs, let alone the capital costs," he said. "They'll lose a lot of money."
Varaiya based his prediction on a 2007 study of a plan to convert a fourteen-mile section of Interstate 680 between Sunol and Milpitas into toll lanes, and on the experiences of other jurisdictions that have turned carpool lanes into toll lanes. They include Interstate 15 and State Route 91 in Southern California, and Interstate 394 in Minnesota. The only one of those three that has turned out to be profitable is SR 91, Varaiya said, and that's because it defines a carpool as containing at least three occupants. As a result, most of the cars in the toll lanes are toll-paying single-occupancy vehicles. "Basically, everybody pays," he explained. By contrast, most of the planned toll lanes in the Bay Area, including I-680, will only require two occupants per car to ride free, thereby limiting the toll revenues.
The toll lanes plan is being pushed by the Metropolitan Transportation Commission, a state agency that oversees all transportation funding in the Bay Area. The MTC views the plan as an essential way to raise money for transportation infrastructure and operating costs, especially when state funding is tight. Varaiya said he has invited MTC planners to come debate him on campus, but they have declined. MTC spokesman Doug Kimsey did not return two phone calls seeking a response to Varaiya's assertions. The plan is being carried in the California Legislature by Assemblyman Alberto Torrico of Newark. After being told of Varaiya's contentions, Torrico's spokesman Jeff Barbosa said the assemblyman would call to discuss the issue, but he did not.
The MTC wants to operate the toll lanes 24 hours a day, seven days a week. The plan is to charge single-occupancy vehicles a toll to use carpool lanes by employing the FasTrak technology now used on Bay Area bridges. The amount of the toll would be determined by how crowded the toll lane is. Paradoxically, the more crowded the lane, the higher the toll would be. The MTC hopes that such tolls would persuade single-occupancy cars to get out of the lane when demand is high, thereby improving traffic flow for carpoolers.
In 2007, Varaiya raised some eyebrows when he published a detailed report on the effectiveness of carpool lanes throughout California. After feeding reams of CalTrans data into computer models, he concluded that carpools lanes don't actually reduce traffic for anyone other than carpoolers. In fact, he argues, carpool lanes haven't even accomplished their main mission — getting more people to stop driving alone. Census data and other studies show that the percentage of people who carpool has remained flat for the past decade, he said.
The researcher came to his conclusions after studying data collected from sensors that CalTrans installs in freeway pavement. The sensors show how many cars pass through a given lane and how fast they are traveling at any given time. Using more than 86,000 data samples taken from 780 miles of carpool lanes during six months in 2005, Varaiya concluded that because carpool lanes are substantially underused in many areas, they actually make traffic far worse in the regular lanes than it otherwise would be. Meanwhile, the carpool lanes usually run slow for two other reasons. First, some motorists choose to drive well below the speed limit because they're uncomfortable driving past backed-up traffic. Second, single-occupant vehicles often lurch into the carpool lane for brief stretches to avoid bad traffic in the regular lanes.
So if the goal of carpool lanes is to increase the overall flow of traffic, Varaiya believes these factors make carpool lanes ineffective. And if the goal is to reduce the number of vehicles on the road, thereby easing traffic and cutting back on emissions of greenhouse gases and other air pollutants, then carpools are a failure because their use has remained flat. The engineer believes most carpool lanes should be eliminated and that the best way to regulate heavy traffic is through the extensive use of metering lights. He points to the success of metering lights in managing traffic on the Bay Bridge.
Varaiya is not without detractors, however, and his research is not without controversy. His research conclusions so defied the conventional wisdom that one of his colleagues, UC Berkeley engineering professor Michael Cassidy, decided to videotape Bay Area carpool lanes and bottlenecks and then study the tapes closely. In a paper released last year, he concluded that carpool lanes were effective because they limit lane changing on congested freeways.
One of the main causes of heavy traffic is that too many cars change lanes too often, slowing everyone else down. With carpool lanes, fewer opportunities for cars to change lanes means improved traffic flow. "If there is less lane changing, you could get higher flows, but we didn't expect to find that," Cassidy said in an interview. "But when we did find that, we said, 'Oh, that makes sense.'" And what about toll lanes? Cassidy said he has not studied the issue, and declined to comment on it.
So is Varaiya right? Will toll lanes turn out to be a big waste of taxpayer money? Well, even Cassidy doesn't advocate expanding carpool lanes because he acknowledges that when they're significantly underutilized they slow down other traffic. Yet the MTC and Torrico want to nearly double the amount of carpool lanes in the Bay Area and then turn them into toll lanes.
Varaiya believes that motorists will revolt at the 24/7 toll lane plan. He thinks that few people will be willing to pay to ride carpool lanes during noncommute hours and on weekends. And that they'll get even angrier when the existence of the toll lane slows down traffic in the regular lanes during those hours. That's exactly what happened on Interstate 394 in Minnesota, he said, and it forced transit authorities to roll back the toll lanes to commute hours only. Once such lanes are only open for part of the day, he says that they don't bring in enough money to pay for themselves.
Clearly, Varaiya is raising some serious concerns that need to be addressed. The MTC and Torrico should slow down their toll lane plan and take a closer look. It would be a huge mistake to waste massive amounts of scarce transportation funds on what could end up being a money-losing 800-mile network of toll lanes.
However, at this point, it looks like the MTC plans to plow ahead. Varaiya believes the toll lanes are being pushed by road builders and traffic consultants, and that there's simply too much money to be made. The MTC estimates that it will spend $7.6 billion over the next 25 years to convert carpool lanes, build new toll lanes, and run the system. "It's a boondoggle," he said. "And it's totally irresponsible."
Seven Days - January 16, 3:41 PM
Seven Days - January 16, 7:54 AM
Seven Days - January 12, 12:40 PM
Seven Days - January 11, 4:53 PM
Seven Days - January 10, 4:38 PM