One of the main justifications for building a light rail line is the hope that it will reduce traffic congestion in a corridor, presumably by drawing commuters out of their cars and onto the train. When we last looked at this assumption, about a year ago, we found cautious support for the decongestive value of light rail corridors in Denver. While traffic continued to rise in these corridors, it rose even more in nearby areas without the rail.
A similar new study of British light rail systems comes away far less hopeful. In an upcoming issue of the Journal of Transport Geography, planners Shin Lee and Martyn Senior of Cardiff University found that the evidence for light rail reducing car use is unclear. Lee and Senior discovered that car ownership and car commute share often continue to rise in these corridors, and that ridership growth is often the result of travelers shifting over from buses — not cars.
For their study, Lee and Senior look at four light rail systems completed from 1991 to 2001: the Manchester Metrolink (built in two phases), the South Yorkshire Supertram, and Midland Metro, and the Croydon Tramlink. To determine the impact these lines had on their respective corridors, the researchers located nearby "control" areas to represent travel behavior that might have occurred if the systems hadn't been built. The control and light rail areas had similar car ownership levels before 1991, similar rail commuter shares, and a similar distance to nearby city centers.
First the researchers looked at whether the light rail systems reduced local desire to own a car. In a few words, they did not. Household multi-car ownership increased in three of the four corridors — and increased by more than the control areas in two of them. In Midland, for instance, two-car ownership rose 7.3 percent, against a 6 percent rise in a nearby control. Meanwhile, household no-car ownership showed a similar effect in the other direction.
Of course car ownership is not the same as car use, so next the researchers looked at whether the light rail systems had an impact on commute mode. Across the board, the systems increased the share of rail travelers in these corridors, in one case nearly 6 percent. However, much of this expanded rail patronage seemed to come at the expense of bus ridership. In the South Yorkshire corridor, for instance, bus share declined more than 10 percent, against only a 6.5 percent dip in a control area.
The light rail systems had the most encouraging impact when it came to commuters going into the city center. All the systems showed considerable increases in the share of rail riders heading downtown — increasing more than 20 percent, in the case of Croydon — while the control areas (obviously) showed no similar gain. Once again, though, the corresponding reduction in bus ridership seemed to explain most of the rise. Sticking with Croydon, the researchers found an 18 percent dip in bus commute share into the city center, against a nearly 3 percent increase in a control area.
All told, the researchers had a hard time concluding that the light rail systems, taken together, produced much of a shift away from car commuting. In some corridors car share declined, but it didn't always decline as much as the control areas — pointing to a general trend in the region. In the case of Midland, car commute share actually increased 5 percent into the city center, a figure that exceeded the increase in its control area.
In the end, the researchers caution against expecting major long-term reductions in road congestion after the creation of a light rail system. (They explain this unsettling observation by pointing to latent demand — the standard economic belief that any drop in road congestion merely encourages non-drivers to start driving). They conclude:
Growing rail shares in the light rail corridors have mainly come from buses and the evidence for light rail reducing car use is less clear. This latter finding is of particular significance, given that a major justification for investment in light rail rather than bus schemes is their presumed ability to bring about major modal shift by attracting substantial numbers of car users.
The work makes a good contribution to our understanding of urban transit, but it has some severe caveats. The study's biggest limitation is that it focused on the simple existence of light rail — not the quality of its service, or for that matter other factors like density or development or demographics. In the case of the Midland corridor, for instance, all of the findings are compromised by the fact that a major road was built parallel to the light rail line; a city that creates an incentive to drive shouldn't be surprised when people do.
With that in mind, the work still underscores some important lessons. For starters, it offers a sound piece of advice: cities considering a light rail system should strongly consider whether improving the local bus system would be cheaper and just as effective. It also provides yet another reminder of the irrational love people have for their cars; getting city residents to give up driving often requires more than just offering them a ride.