Changes in commuting to work times over the 1990 to 2000 period

Travel times to work from the 2000 Census show an increase in average commuting times that is difficult to reconcile with the viewpoint expressed in earlier literature that suburbanization has provided a solution by acting as a traffic “safety valve”, preventing a “traffic doomsday” from occurring in the face of urban growth. We examine commuting times to work using US tract-level Census data for the years 1990 and 2000. A spatial econometric modeling approach that allows us to distinguish between commuting time congestion spillover impacts arising from shared roadways is developed. We compare the influence of 1990 and 2000 tract-level characteristics of residents that give rise to long commute times (over 45 min).