Every year, thousands of holiday travelers endure the soul-numbing BART trip to and from SFO. Does it really have to be this way?
As BART is about to open its extension to Berryessa, it is worth reflecting on an earlier major expansion to the system, which linked Millbrae and San Francisco International Airport. That project was planned poorly, and now, almost fifteen years later, public transportation between SFO and the greater region still doesn't work. So what went wrong, and how can we avoid making the same mistakes again?
When BART started, the southern terminus on the Western side of the Bay was Daly City. BART reached it from Downtown San Francisco through an old alignment of the Southern Pacific rail line now used by Caltrain. Since 1907, Caltrain has opted for a more direct route, known as the Bayshore Cutoff, because the old alignment took the long way around San Bruno Mountain, and even by the standards of those times, it was too slow. BART repurposed the entire old alignment in its extension to SFO and Millbrae:
Instead of investing in Caltrain in the 1990s, which would involve electrification and the Downtown Extension (both considered even back then), the region chose to build BART the long way. As a result, trips between Millbrae and the Embarcadero take 32 minutes.
Local rail activists Clem Tillier and Richard Mlynarik have developed sample schedules for an electrified Caltrain service, finding that trips from Millbrae to Transbay Terminal would take 24 minutes on the local train and 19 on the express (such as the Baby Bullet).
Millbrae itself comes from BART's lofty ambitions to take over Caltrain, dating back to the original system plans from the early 1960s, which went as far as Palo Alto, along the old way rather than along the Bayshore Cutoff. In the 1990s, a local coalition opposed to the way BART was constructing the SFO extension, called Coalition for a One-Stop Terminal, or COST, suggested BART only go to the airport and not to Millbrae, and connect to Caltrain at San Bruno, where the two lines would intersect.
Instead, BART built two branches, one to Millbrae and one to the airport, and opened a separate San Bruno station from the Caltrain station, too far for a transfer. Because BART is underground in that area, opening an infill station to connect to Caltrain is prohibitively expensive.
The distribution of branches ensures that there is only one service to each terminal, coming every 15 minutes. Both BART and Caltrain are scheduled railroads, but the transfer at Millbrae is not timed, which makes transferring an arduous and time-consuming endeavor.
The Millbrae end is bad by itself, but the airport end is hardly better. BART wanted to extend all the way to the terminal, to avoid a forced transfer to a people-mover, such as the JFK AirTrain in New York City, which connects subway and commuter rail hub Jamaica Station with JFK in about 15 minutes. But that’s exactly what ended up happening.
SFO's multi-terminal layout forced BART to make a choice, and it only extended to the International Terminal. Passengers at other terminals still have to connect to the AirTrain. The door-to-door travel time can take long enough that many people prefer to use road access, often riding a shared van or a ridesharing service like Uber or Lyft. In 2006, before Uber, only 7% of SFO passengers used rail, compared with 8% at JFK, 10% at Atlanta, and an impressive 28% at Frankfurt.
If things are bad in San Francisco proper, they are utterly outrageous on the Peninsula. BART's SFO extension, partly funded by SamTrans, replaced shuttle buses from Millbrae. During weekdays in the daytime, there is no direct service from Millbrae to SFO, forcing passengers from Silicon Valley to transfer from Caltrain to two separate BART trains, and if they're flying domestically, then again to the AirTrain. For passengers taking Caltrain to the airport, the connection is worse than useless.
As a result of the poor layout of the extension, ridership is low. Total ridership at the four stations comprising the extension—South San Francisco, San Bruno, SFO, and Millbrae—is just 21,500 per weekday. (Colma was added separately.) The project cost $1.6 billion, or $2.3 billion in today's money, $107,000 per rider. The original ridership projections estimated 34,000 boardings at these four stations, and were made in 1993, before the tech boom.
Most of the shortfall came from Millbrae, where the projections said there would be 16,500 daily riders where there are fewer than 7,000 today.
The people involved in this disaster never faced any consequences. Quentin Kopp, then a State Senator, played a key role in the two-terminal layout of the extension, which raised its cost by $500 million; he subsequently played a key role in the planning for California High-Speed Rail last decade.
With this track record, it is difficult to trust many of the decisions the region’s power brokers make regarding public transportation.
Those decision-makers are still making important decisions—about future BART extensions, about investment in Caltrain, and about high-speed rail in the Bay Area. And as long as they remain shielded from criticism, they are likely to make the same mistakes again. It’s time to start holding them accountable.
The conversation is not over: Judge Quentin L. Kopp has written a passionate defense of his work building the BART to SFO extension. Then, Alon Levy comes back with a strong rebuttal and explains why he still disagrees with Kopp.
Alon Levy is a mathematician with a strong interest in urbanism and mass transit, and currently works as a freelance writer. He contributes to the Bay City Beacon as a weekly transit columnist for Pedestrian Observations. You can find more of his writing supporting walkability and good transit.
Join us to celebrate the Beacon's first anniversary! There'll be food, drink, and great political gossip. RSVP to the Beacon Birthday Bash today!