San Diego’s public transit authority has had some problems lately.
Voice of San Diego has detailed that Measure A, a half-cent sales tax SANDAG proposed last November, would have raised far less than the promised $18 billion , and it appears that agency officials knew it and promoted the figure anyway . Measure A was ultimately rejected. Making matters worse, the agency’s existing sales tax, TransNet, has a $17 billion shortfall thanks to the same forecasting error and escalating prices for the projects  voters approved in 2004. The agency is scrambling to fill its funding hole.
In such an environment, it is important to build every existing project as inexpensively as possible.
Unfortunately, that’s not what’s happening with one of SANDAG’s highest-profile projects, the Mid-Coast Corridor Transit Project, which will extend the trolley’s Blue Line north to the UCSD campus. The official cost of the project is high by any standard – $2.1 billion for 11 miles – but it’s especially high for a light-rail project.
SANDAG expects the project to attract 35,000 weekday riders by 2030, for a total of $60,000 per rider. That’s a very high per-rider cost: The initial trolley lines cost about $16,000 per rider by 2000 , or $22,000 in today’s money. In Los Angeles, the Expo Line has cost about $60,000 per rider, but projections for future increases in ridership by 2030 reduce that to just $40,000 per rider.
Worse, the high costs of the Mid-Coast extension come following what’s already been a major cost overrun. In 2010, the project was expected to cost $1.24 billion ; the current budget of $2.1 billion is 69 percent higher. That’s because of both additional scope, such as more parking at stations, and higher costs for the same items, such as the same length of bridge that’s now more expensive.
Contrary to what you might expect, large cost increases are no longer normal for American light-rail projects: Researchers at George Mason University found in 2009  that since the 1980s and early 1990s, forecasts had steadily improved, and by 2005 they were on average correct.
In this decade, a cost overrun as large as that of the Mid-Coast Extension suggests something is seriously wrong with the agency’s planning and design.
This contributes to a political climate in which the public lacks trust in SANDAG’s projections. The environmental and transit activist group Cleveland National Forest Foundation  accuses SANDAG of failing to promote mass transit in the region. Activist Harry Jensen, a frequent critic of the Mid-Coast project, said in a recent email to SANDAG board members that the region should “suspend all construction work on the Mid-Coast Trolley project.”
Cleveland National Forest Foundation Director Duncan McFetridge criticizes the project because the land use patterns between downtown and UCSD are auto-oriented and hostile to mass transit. UCSD and downtown are dense and walkable and can generate ridership, but the areas in between cannot. The group is instead pushing an alternative Mid-Coast alignment, using the Coaster tracks and a proposed tunnel  under UCSD, which would have the added benefit of speeding up intercity trains between San Diego and Los Angeles.
There is a third way.
SANDAG could consider a serious change to the project that would reduce its cost and take advantage of both existing light-rail lines and the Coaster rail line. This is a German invention called the tram-train. It is used mainly in small cities like Karlsruhe in southwestern Germany; but there are also under-construction lines in the Paris suburbs.
The tram-train uses light-rail vehicles, like those used on the San Diego trolley, that can drive on mainline rail tracks between urban areas (like the Coaster) and on city streets in light-rail mode within city centers (like the trolley downtown).
The Mid-Coast extension runs alongside the tracks used by the Coaster commuter trains and the Amtrak Pacific Surfliner, except at the northern end, where it is planned to climb Miramar Hill to serve UCSD; the mainline tracks go around the hill instead. Right now, the plan is to build dedicated tracks for the trolley; if it were a tram-train, it could instead use the existing tracks shared with regular trains and only use new tracks to climb the hill to UCSD.
Tram-trains also have a history in the U.S., by another name: the interurban. In the first half of the 20th century, private companies built long commuter lines and even intercity rail lines. They were mostly separate from the conventional rail network, and were electrified rather than steam-powered, but ran like regular railroads between cities; within cities they ran on city streets, like streetcars. At the beginning of the 20th century, there was some track-sharing between urban transit and commuter rail in New York, but federal regulations outlawed it in the 1910s on safety grounds.
Today, some former interurbans have been turned over to light rail, which runs on city streets in the center and on dedicated railroad lines farther out. Unfortunately, the need to construct new dedicated tracks for every light-rail line, rather than sharing regular train tracks, leads to high construction costs. Using the existing rail tracks is cheaper, and San Diego should investigate if it is possible to modify the Mid-Coast extension to do so.
There is already one place in the U.S. using a proper tram-train using mainline rail: the New Jersey River Line  uses city streets in Trenton and Camden and mainline track between them. The regulations for this are complicated, and River Line has strict time separation, with freight trains running only at night. But European regulations are friendlier to tram-trains. An ongoing overhaul of American passenger rail regulations is planned to permit lightly modified European trains on American tracks, and it is likely that sharing tracks between intercity passenger and freight trains and light rail will soon be permitted.
Public transit consultant Jarrett Walker discussed tram-trains in 2009 . He said that the technology is useful when there is a dominant destination that is a little too far away from the train tracks. In the case of San Diego, this is UCSD. Walker cautioned that street-running slows down the trains, since they have to obey street traffic rules; in this case this is not a problem, since the segment through UCSD is planned to be elevated rather than on the street.
SANDAG should study the tram-train option as soon as possible. It is likely to be substantially cheaper than the current proposal to build new light-rail tracks alongside the railroad, and the sooner the change happens, the less it would cost and the less it would delay the project. The agency is facing a revenue shortfall and justified public mistrust, and needs to find places to cut costs, preferably without cutting useful scope. The tram-train permits cutting scope that, while useful, is primarily there because of past federal regulations that are already being reformed.
Alon Levy is a Paris-based mathematician and public transportation policy writer.