SPRAWL:Auto Dependence book

Is increasing automobile usage a symptom of improved mobility or “auto dependency”? A book just published by the University of Colorado Press argues that “high levels of automobile dependence” are “inefficient and impractical.” The book also argues that land-use and transport policies, such as those adopted by major European cities, can reduce this unnecessary dependency. Those policies include heavy taxes on autos and fuels, heavy subsidies to transit, and subsidies or land-use laws aimed at increasing urban densities.

The real value of the book, “An International Sourcebook of Automobile Dependence in Cities 1960-1990,” is its huge collection of data from forty-six cities in Asia, Australia, Europe, and North America. Data include CBD, inner city and metropolitan area populations, jobs, and land areas; numbers of motor vehicles; miles driven; vehicle occupancies; average road speeds; mode split between transit, auto, and walk/cycle; transit ridership and services; and more. Moreover, the authors made an effort to collect all of these data from all forty-six cities for 1960, 1970, 1980, and 1990, so it is possible to look at trends instead of just one instant in time.

The data do show that, for any given year, population densities are greater, auto usage is lower, and transit usage is higher in European and Asian cities than in Australian and North American cities. But the time series data reveal that almost all cities are undergoing the following changes:

* Major reductions in population densities as urbanized areas increase, suburban populations grow, and populations in the central cities either decline or remain stagnant;

* Major dispersions of jobs as central business district jobs decline or grow very slowly while overall urban jobs grow rapidly;

* Major increases in per capita auto ownership and miles driven each year;

* Transit ridership in most cities is increasing since 1970, but transit’s share of travel is stagnant or declining. In a few cases where it is increasing, it is at the expense of walking/cycling, not of auto driving.

A close look at the data reveals some peculiarities. The total populations of most of the European urban areas in the book, including Amsterdam, Brussels, Copenhagen, Frankfort, Hamburg, London, Stockholm, and Vienna, declined by around 10% or more in the twenty to thirty years preceding 1990. Munich and Zurich also declined slightly after 1970. The only European urban area in the sample that did not decline was Paris. Collectively, these cities lost nearly 2.2 million people! Where did they all go? The book answers this question for Stockholm. Although the municipality of Stockholm lost 134,000 people after 1960, the COUNTY of Stockholm gained more than 400,000 people. The book includes a map showing that more than 90% of the county is urbanized. But it does not list the urbanized land area, transit usage, or other data for the county except for population, motor vehicle numbers, and miles driven. This greatly skews the data for Stockholm. Within the municipality, transit is holding its own with a 55 percent share of commuter travel. In the county, however, auto miles driven more than tripled between 1960 and 1990.

The book notes that “Stockholm is unique in the sample of cities in this book in recording a marginal decline in the absolute level of car use experienced in the Municipality area between 1980 and 1990. It is perhaps not coincidental that the Municipality of Stockholm is the only city in the sample to have experienced an increase in density between 1980 and 1990.” But auto driving in the county increased between 1980 and 1990, and its density probably declined — though we don’t know for sure because the book does not include urbanized land area for the county.

In short, in Stockholm as in other urban areas, people have escaped the congestion of the inner city and its immediate suburbs by moving out to more distant suburbs. The book misses this by failing to count the entire urbanized area. I suspect that the same thing happened to Amsterdam, Copenhagen, and many of the other European cities in the book.

The book’s authors, Australians named Jeffrey Kenworthy and Felix Laube, reach two major erroneous conclusions. First, they claim that “the cost of cars and the wealth of cities do not provide reliable or consistent evidence in explaining the degree of dependence on the automobile in different cities.” They specifically deny the conclusions of University of California economist Charles Lave that “other nations head down the same path (as the U.S.) as their incomes increased” and that “the desire for personal mobility seems to be unstoppable.”

However, the data they use to reach this conclusion is very limited. They do not have incomes per capita over time. Instead, they only show per capita gross regional product (GRP) for 1990, and only for some of the cities in their sample (all but one Canadian city and many smaller U.S. cities are left out). For this reduced set of cities they also show the average cost of driving in dollars per kilometer. Their data show that the cost of driving in European cities is 66% greater than in the US, and in wealthy Asian cities it is more than double that of the US/Australian costs are 28% greater than in the U.S. Kenworthy & Laube use per capita gross regional product to compare the wealth in various cities. Their data show that European cities have 18 percent higher per capita GRP than cities in the US. But tax rates in Europe are much higher than in the US, and thus personal take-home pay is probably lower in Europe. This, along with the 66% higher cost of driving in Europe, helps to explain lower auto usage. They base their case on comparisons of selected city pairs. For example: “New York has 36% less car use per capita than Houston, but is actually 10% higher in GRP per capita.” Based on such comparisons, they conclude that differences between cities “cannot be explained by simple economic factors alone.”

Of course, no one ever claimed that “simple economic factors alone” explained car usage. But they did claim that, as wealth increased, car usage would increase. Although the book does not include wealth data for any year but 1990, it is safe to say that wealth increased in every city in the sample, and the book does show that auto usage also increased in every city.

Confuse cause & effect

The book’s second conclusion is equally wrong: The authors say that “urban form, especially higher urban density, is a key and consistently strong factor in helping to explain the level of car use in a city and the level of transit use.” In other words, increasing the density of an urban area will increase transit ridership and reduce auto driving. Here again they confuse cause and effect. Cities that can’t afford autos will have higher densities. When people start to buy autos, densities decline. The densities are an effect of auto ownership, not the other way around. If density were enough to discourage auto usage, then noone in dense cities such as Amsterdam and Brussels in 1960 would have begun driving. But they did begin driving and the densities of those cities have declined significantly.

Differences in auto usage between cities can explained by two major sets of factors: economic and historical. They can be summed up the question: At what time in the city’s growth did people’s incomes become sufficient to own and drive autos — including enough to pay for any tax penalties that governments charge for auto driving? If a city is still new by the time income reaches that stage, then the city will be built to auto densities and transit will play little role. Except for New York and major European cities, most urban development falls in this category.

If the city is largely or fully formed by the time income reaches that stage, then it will be far denser than is comfortable for auto usage. New developments on the urban fringe will aim to accommodate the auto, but inner city densities will decline only slowly. For a time, the city will appear more transit- and pedestrian-friendly, but that will fade as people express their mobility by leaving. This historical difference explains why New Yorkers drive less than Houstonians along with most of the other differences between cities that the book’s authors can’t explain by “simple economic factors alone.”

Virtually all of the 46 cities in the book fit one of these two patterns. Even Paris, which alone among the European cities continues to grow in population, lost 33 percent of its density between 1960 and 1990, and per capita driving nearly tripled.

“Auto dependence”

Perhaps the most important bias in the book is the most subtle: the repeated use of the term “auto dependence,” as if the automobile cripples us rather than enables us to be more productive and enjoy more things in life. The text in the book fully supports smart-growth policies. Yet the data in the book are invaluable for smart-growth opponents. For those who are interested, I will send a data file in tab-delimited format listing some of the most important data in the book for thirty-five cities (I left out some of the Australian and poorer Asian cities along with a couple of Canadian cities). Jeffrey R. Kenworthy and Felix B. Laube, “An International Sourcebook of Automobile Dependence in Cities 1960-1990” (Boulder, CO: University of Colorado, 1999), 704 pp. $125, available for $99.68 incl US shipping from www.ecampus.com. (Contact Randal O’Toole, The Thoreau Institute rot@ti.org www.ti.org)