Clark Williams-Derry points to methodological flaws in the Urban Mobility Report that could drastically change how we measure congestion in cities:
Consider two hypothetical cities, Sprawlville and Compact City. In Sprawlville, people travel a long way to work -- an average of 20 miles door to door. In free-flowing traffic, the trip would take 20 minutes, but it takes 10 extra minutes during rush hour, for a total commute of 30 minutes. In Compact City, people don't have to travel as far: it's just 10 miles from home to work on average; the trip takes 10 minutes off-peak, and 10 extra minutes during rush hour, for a total of 20 minutes.
In this example, congestion slows commutes by the same amount -- 10 minutes -- in both cities. Sprawlville residents wind up with longer total commutes, since residents travel longer distances. Yet the "Time Travel Index" shows that Compact City has a worse rush hour!! That's because the Time Travel Index shows a 2:1 ratio (i.e., 20 minutes vs. 10 minutes) for rush hour vs. off-peak travel in Compact City, and a 3:2 ratio (i.e., 30 minutes vs. 20 minutes) in Sprawlville.