San Francisco’s Van Ness BRT has hit multiple hiccups. But all the delays obscure the major problem with Van Ness BRT - it costs far more than it should. In fact, it might cost more than light rail.
San Francisco is building Bus Rapid Transit, with rapid routes and bus lanes, or at least, it is trying to. Muni has announced that bus lanes on Van Ness Avenue would be delayed by two years; BRT service is now planned to open in 2020.
The delays on Van Ness are a symptom of the poor state of Bay Area transit investment, with frequent timetable slips and cost overruns. But it's also important to take a step back and evaluate the project from the start.
For just two miles of bus lanes, the construction cost of the Van Ness BRT project is $190 million. While the cost is within the range of recent American light rail projects built on city streets, and high by the standards of European light rail, this is not normal for Bus Rapid Transit.
The Orange Line in Los Angeles, a BRT system formed by paving over a former Pacific Electric light rail right-of-way, cost $324 million ($423 million in today's money) or only $23.5 million per mile.
Why is the cost so high for Van Ness? In the most expensive cases, there are multiple factors, some of which are unclear. But one factor is ancillary street reconstruction costs, unrelated to the core BRT project. The Examiner reports that Van Ness BRT is part of a bigger sewer and street light replacement project, totaling $316 million.
While $190 million is just the BRT share, the connection to street lights and sewers should suggest the scope is too broad: sewers go underground, so most likely the bus lane construction is planned to include some digging. Even within the $190 million, a large fraction is likely to be street improvements and not bus lanes.
Streetscaping is a common feature of light rail plans. The arrival of a high-investment capital project gives the city grounds to bundle many other partially related plans into the same package. Nice, in the French Riviera, built a tramway with extensive urban reconfiguration. Discounting the tramway, the reconfiguration alone accounted for 30% of the project's cost.
This is not supposed to happen with Bus Rapid Transit. The main selling point of BRT is that it's cheap. In 2010, then-Federal Transit Administration chief Peter Rogoff said, “paint is cheap, rails systems are extremely expensive.” Earlier, Jaime Lerner, who practically invented modern BRT when he was the mayor of Curitiba and has spent many years promoting his system's success around the world, made the same pitch: “If you want creativity, cut one zero from your budget. If you want sustainability, cut two zeroes from your budget.”
The main elements of Van Ness BRT are not expensive. The Van Ness BRT plan involves center lanes rather than curb lanes, which are also present at the low-cost systems of Curitiba and Bogota. SFMTA is planning for passengers to board to the center bus lane from raised curbs; this has precedent in Curitiba as well as on some European bus lanes, such as Odengatan, in Stockholm.
Odengatan in particular is 100 feet wide, which is narrower than Van Ness; it should be easier to equip Van Ness with physically-separated lanes and boarding islands. There is no reason for costs to be so high.
The problem is that once cities think of BRT as a major innovative solution, they can't resist adding frills. Worse, when they do try economizing, they often don't economize on street reconfiguration (which is genuinely unnecessary) but on visible passenger amenities like bus shelters. This creates the impression that inexpensive construction is necessarily low-quality and encourages passengers to advocate for higher-end options.
Although the top BRT systems in the world are in middle-income countries—Brazil, Colombia, Mexico, and China—it’s perhaps Europe that offers the way forward. BRT as a distinct concept of separately-branded bus routes barely exists in Europe. Instead, there are system-wide improvements.
San Francisco is ahead of the rest of the United States in this regard: Muni has off-board fare collection on all buses, whereas other North American cities only have it on specially-branded lines.
The next step involves treating bus lanes as less special. This does not mean compromising design standards: buses on the most congested corridors should run in dedicated lanes, with raised curbs separating them from car traffic when possible. But this does mean treating bus lanes as an ordinary feature of major streets rather than as a special BRT treatment. If bus lanes are special, then it's hard to resist the temptation to bundle them into high-cost street reconstruction projects, which remove the only advantage of BRT over light rail—namely, that it's inexpensive to build.
The red painted lanes have a design compromise (they're not physically separated from car traffic) but aren't as expensive, because there's little infrastructure involved in building them. Even on Geary, the costs per mile are not as high as on Van Ness.
It is possible to build transit without breaking the bank. It is even possible to build high-quality transit with few design compromises without breaking the bank. But SFMTA would need to be concerned with construction costs and avoid unnecessary scope. Paint is cheap, slightly raising curbs between lanes is cheap, and even boarding islands for stations are not expensive.
It's everything else that's making Van Ness BRT cost more than light rail. The delay overall is bad, since it suggests problems with the planning process, but the city should take advantage of this to reduce construction costs. $190 million should be building bus lanes on many streets, rather than on just two miles of Van Ness.
Sign up for the Bay City Beacon weekly newsletter! It's a free way of getting the best of political gossip and cutting-edge culture in your inbox every Friday.
