Toronto politicians have long described transit as a “decongestant” because buses, streetcars and subways get people out of their cars. But a new study on driving patterns in U.S. cities, released recently by a pair of University of Toronto economists, shows that adding transit service doesn’t relieve traffic gridlock in the long term.

The research, by Gilles Duranton and Matthew Turner, expands on a long-observed phenomenon: that adding highway capacity doesn’t solve traffic congestion and the additional lanes just fill up over time. The study shows the results hold true for all urban roads. “If you increase the kilometres of roadway in a city by 10 per cent, you’ll increase the vehicle kilometres traveled per year by 10 per cent as well.” The main reason: both individuals and truckers change their driving behaviour, triggering even more traffic. “Build it,” says Turner, “and they will drive.”

Adding transit doesn’t short-circuit traffic build-up, Turner concludes. But he notes the study shouldn’t be interpreted as a case against transit because the provision of additional buses, in particular, allows more trips on existing transportation infrastructure at little extra cost.

Why is all of this important? Because North Americans spent a lot of unproductive time in their cars, and public concern about climate change is putting pressure on policy-makers to find ways to reduce vehicle emissions.

The study notes that the average American household spent 161-person minutes a day in a passenger vehicle in 2001, and that the amount of time spent on routine household travel had increased by 10 per cent in just six years.

“Multiplying by the number of households in the U.S. and any reasonable dollar value of time, we see that society allocated billions of dollars more to traffic congestion in 2001 than in 1995. That Americans rank commuting among their least enjoyable activities confirms our suspicion that the costs of congestion are large.”

Duranton and Turner also calculate the benefits drivers gain from the new roads and conclude that they are well below the cost of providing the roads in the U.S. “An average extension of the interstate network does not result in sufficient travel time improvements to justify its cost,” they say. Extending subways and streetcar routes are similarly expensive and result in little or no overall gain for the average commuter.

So what is the best public policy response to gridlock? The authors suggest that the City of London may have had the right idea, using peak-hour traffic tolls to dissuade drivers from taking their cars to work. “These findings strengthen the case for congestion pricing as a policy response to traffic congestion,” the report concludes.

The study was published by the National Bureau of Economic Research.