Yesterday, I described how an effective private transit system could be designed for all but the smallest urban areas by simply scaling up the taxicab association model to buses. Rather than use this model, many of America’s subsidized transit agencies have tried a different sort of scaling upwards: from buses to rail. This has led to numerous planning disasters.
Too many public officials imagine that running a rail transit system is just the same as running a bus system, only with bigger vehicles. Jonathan Richmond’s paper (since expanded to a book), The Mythical Conception of Rail Transit in Los Angeles quotes the Los Angeles Times: “One of the arguments made most often for the rail line is that it will be cheaper to operate because a single driver on a train can carry up to five times as many passengers as a bus.”
These are, as they say, famous last words. The reality is that a bus system does not scale up to a train system: they are quite different beasts, and running a train system not only requires a completely different set of skills, it entails a much higher risk — a risk that transit agency officials can ignore because they simply impose it on taxpayers and, in the worst cases, transit riders.
The first big problem is in construction. As I showed yesterday, “constructing” a bus system does not take a lot of capital outlay, and transit agencies can get most of their capital requirements from federal and state grants. This means the only thing a bus operator needs to worry about is this year’s revenue vs. this year’s costs.
Planning and constructing a rail line, however, takes 10 or more years and generally requires that the transit agency go heavily into debt by selling 30- and sometimes 40-year bonds. To plan the construction, the agency needs to accurately forecast future costs. To pay back the bonds, the agency needs to accurately forecast future revenues. But no one can accurately forecast the future, so even if the forecasts are not deliberately optimistic, chances are they will be wrong.
Denver’s RTD included something like half a billion dollars of “contingency” funds in their projected costs of the FasTracks rail system. That wasn’t enough and now they are something like a billion dollars over budget.
In forecasting future sales tax revenues, RTD figured that, in some years, tax revenues have grown by 12 percent, while in other years the growth has been zero. So it assumed revenues would increase an average of 6 percent per year forever. I looked at their detailed financial analysis and realized that a recession at any time during the 8-year construction period would sink them because the revenue shortfall would eat away their reserves and force them to dramatically cut existing transit services just to keep from defaulting on their debt. Of course, the economy rarely goes 8 years without a recession.
As it turned out, they didn’t even get that far: even before construction has begun, they are saying their revenue forecasts were $670 million too high. Cal Marsella, RTD’s general manager, was hired specifically because he did not believe that transit agencies should get into the train business. He puts on a good front today, but I bet he privately regrets giving into the train community. (Of course, he gets paid something like $270,000 a year, so maybe he doesn’t care.)
Once construction is complete, the problems are not over. Far from saving labor, a rail system requires a transit agency to maintain tracks, electrical facilities, stations, and — if it is an elevated or subway — elevators and often escalators. The most common public announcement on the Washington Metro system is a list of all of the elevators that are out for the day and a promise to provide shuttle-bus service from nearby stations for disabled passengers.
The whole “save-labor-on-drivers” argument is deceptive anyway because transit agencies nearly always supplement rail service with feeder buses at every suburban station. Transit-choice patrons prefer to drive to the park-and-ride or kiss-and-ride stations, but the agencies have to provide the feeder buses so transit-dependent people can get from the stations to their homes. The result is often a reduction in labor-efficiency.
For example, in 1992, before it began planning or building light rail, Salt Lake City’s transit agency carried 32,000 trips and 159,000 passenger miles per employee. Today, its supposedly successful light-rail and bus system together carry less than 24,000 trips and 100,000 passenger miles per employee.
After 30 years or so, just when a transit agency has almost repaid the debt it incurred to build the rail line, comes the time when most of the rail system needs to be replaced. Roadbed, rails, wires, signals, even large parts of the station facilities get pretty worn out after about three decades. Most often, the agency will want to invest in new technology as well. Replacement will take years and cost a significant fraction of the initial construction cost, so the entire cycle begins again.
This explains why so many rail transit projects turn into disasters. Of course, they always have an excuse. Buffalo and San Jose say they planned their rail systems when their regions booming; who knew they were going to bust? Cal Marsella says over and over that no one could have predicted that cement and steel prices would go up.
All the excuses come down to this: no one can predict the future. With a bus system, you don’t have to. With a train system, your guaranteed-to-be-wrong predictions put both taxpayers and transit riders at risk.