Yet more evidence: “If you build it they will drive”

By Eric Sundquist

There’s new evidence, from academia and a prominent real-world case, that ever-expanding highway capacity is a futile strategy for reducing congestion.

First, the specific example: Eight years ago Southern Californians famously endured “Carmageddon,” a temporary closure on Interstate 405 in the Sepulveda Pass area due to a road-widening project. The closure itself turned out to be not such a big deal, sort of a minor case of “disappearing traffic.” But the billion-dollar-plus investment appears not to be delivering on the promise that it would “reduce existing and forecasted traffic congestion.”

Crosstown, a data-analysis project at the University of Southern California, looked at vehicle speeds on the 405 over five years, capturing the last year before the new lanes opened and the period since. Crosstown found that outbound evening travel speeds dropped dramatically even in the first year after the project opened and have stayed about flat since then. Speeds were better during other parts of the day, but mostly below pre-project speeds.

One example does not confirm the Fundamental Law of Road Congestion, but the 405 case is consistent with that theory, as is a new comprehensive study of induced demand, by Kent Hymel of Cal State Northridge. Hymel addresses a methodological issue plaguing some previous induced demand studies—that while wider roads may induce more driving, transportation agencies add lanes where they expect more traffic “naturally.” Using a model with an instrumental variable to tease out this issue of causation, Hymel finds that:

  • A 1 percent increase in lane-miles induced a 1 percent increase in vehicle-miles traveled.
  • Over a five-year period, the induced travel essentially fills up new capacity, bringing speeds back to a pre-project level.