The influence of metropolitan spatial structure on commuting time

Abstract

The monocentric city ultimately becomes inefficient as urban growth continues because of increasing congestion close to the CBD. Although the exhaustion of CBD agglomeration economies is another major reason for the breakdown of monocentricity, these congestion costs are sufficient grounds alone for the emergence of the policentric metropolitan region. Theoretical models of the policentric city usualIy imply shorter commuting times because of the assumption that households in choosing residential locations cluster around employment subcenters to minimize commuting trips. However, in a world of multiple economic sectors that are not uniformly dispersed, heterogeneous labor, and residential neighborhood differentiation, workers may commute across metropolitan areas so that average trip times could become longer rather than shorter. Which of these outcomes is more prevslent is an empirical question that has not received much attention. Apart from the number of major employment centers, commuting times are also influenced by city size and metropolitan density. However, at the level of descriptive analysis the relationship between city size, population density, and travel time is not clear-cut. For example, Table 1 shows these data for the 19 urbanized areas with population densities greater than 3500 persons per square mile and for the 20 largest urbanized areas (7 of the cities are members of both groups). New York is an outlier with the longest average commute (32.3 minutes) combined with a very low value for vehicle miles traveled (VMT) per capita, suggesting severe congestion. The eight