However, even a casual look at air pollution data suggests that this line of reasoning is simplistic. Nitrogen dioxide, ozone, carbon monoxide, and lead — pollutants often associated with automobile use — have fallen consistently since the 1970s[93] and smog has also become less of a problem.

The real story of the past several years has been the increasing complexity of metropolitan areas.

Moreover, and perhaps more importantly, metropolitan areas with the lowest population densities have the fewest air pollution problems. Economist Randal O’Toole, executive director of the Oregon-based Thoreau Institute, examined the relationship between population density, automobile use, and air pollution.[94] He found that metropolitan areas with the highest population densities also had the highest smog rating (see Chart 13).[95] The 234 metropolitan areas that the Environmental Protection Agency (EPA) considered smog-free had an average density of 1,505 people per square mile. Similar results were found when central cities were analyzed.

O’Toole also found that population density — or compactness — had little relationship to the automobile’s share of commuter trips.[96] In other words, more densely populated areas did not reduce residents’ automobile use. "Autos hold more than 75% of the market [in commuter trips] in every area except New York and more than 90% in the vast majority of areas, including Los Angeles and Miami, the two densest areas," noted O’Toole[97]. Ironically, the number of vehicle miles traveled increases with density.[98] So the idea that low-density residential areas contribute more to pollution than high-density areas do is not supported by the data.

In fact, a policy strategy that attempts to increase population density could lead to more congestion unless road capacity is increased. Thus, an increase in density risks increasing air pollution and smog, potentially putting urban areas into "non-attainment," or polluted, status with the EPA.

Higher density development may affect the number of miles people travel in automobiles, but the effects are not large enough to offset the congestion costs it creates. Proponents of high-density "compact development" argue that a doubling of density could result in a 25 to 30% reduction in vehicle miles traveled.[99] Yet, as O’Toole pointed out, this reduction in traveling distance is overwhelmed by the increase in the number of people making trips.

As an example, let’s say a commuter-based bedroom community of 7,000 people registers 2,000 automobile trips. The regional planners somehow raise the population of the community to 14,000, which doubles the population density. This also doubles the number of commuter trips to 4,000. If the city has an effective bus, jitney, or taxi system, the number of automobile trips might fall by 30%, or 600 trips. But the community still would have to accommodate 1,400 additional automobile trips.[100] If the fictitious community did not also increase its road capacity — as many planners recommend — congestion would also increase. Compact city development, then, becomes congestion-inducing development.

Few communities have the kinds of mass transit systems in place necessary to accommodate a significant share of current commuting trips. Indeed, most urban mass transit systems are neither efficient nor cost-effective under current policies. Light rail systems are too inflexible and costly to be effective mass transit alternatives.[101] Until cities deregulate their transit industry or institute "curb rights"[102] — reducing burdensome licensing and inspection systems for taxi, van, and bus services — cost-effective mass transit is unlikely to emerge. Without this deregulation, continued suburbanization and the decentralization of employment may well shorten commute times.[103]

The real story of the past several years has been the increasing complexity of metropolitan areas. Traditionally, urban development has been characterized by a large central city that serves as the economic, political, and cultural hub of the region. Since the suburbanization of people and decentralization of employment, a new regional urban form has developed where a number of different urban centers emerge within a region.[104]

This is clearly happening in Michigan. Suburban, or "collar," counties experienced the largest increases in population density. Central city counties are losing population and density. These trends are implicit in the decline in Detroit’s population and the rise of suburban cities such as Sterling Heights, Livonia, and Warren. But the trend is broader than the rise of individual suburban cities.

To the extent that Michigan’s suburbanization and population decentralization results in the emergence of new population and employment centers, environmental concerns may be mitigated even further. Reid Ewing, an architect of Florida’s statewide growth management plan, recently admitted that when multiple employment and population centers are considered within a region, the environmental benefits are superior to those achieved by policies that attempt to maintain central city dominance.[105]