For decades, American urban areas have grown in land area much more than they have grown in population. This geographic expansion is often attributed to increasing dependence on the automobile and construction of the interstate highway (freeway) system. A relatively new school of urban planners, the "new urbanists," blame the expanding urban area for a number of problems, including increased traffic congestion, higher air pollution, the decline of central cities, and a reduction in valuable agricultural land. (New urbanist policies also go by the label "smart growth"). Moreover, new urbanists believe that more spacious urban areas typical of the United States are inherently inefficient relative to more compact cities, exhibiting higher costs for infrastructure and public services.
Recently, the issue of "urban sprawl" received top billing at a White House event at which President Bill Clinton and Vice President Al Gore announced their Livable Communities Initiative, which, it was promised, would reduce traffic congestion, promote cleaner air, preserve open spaces, and retard urban sprawl. To achieve these objectives, Clinton and Gore propose to provide the suburbs with additional funds for mass transit and loans to buy land for parks and greenbelts. Their initiative also would assign to the Department of Housing and Urban Development the responsibility for encouraging and financing "smart growth" strategies to encourage "compact development" and regional cooperation.
Establishment of urban growth boundaries (UGB).
Channeling urban development toward "infill" (undeveloped areas within the urban growth boundary).
"Transit oriented development" along urban rail corridors, higher population density, and higher employment density.
Little if any expansion of street or highway capacity.
Retail developments less oriented toward the automobile (smaller stores with less parking located generally in town centers rather than suburbs).
The new urbanists believe that these strategies will produce a more compact city in which automobile dependency, traffic congestion, and air pollution are reduced. New urbanist concepts have been incorporated into a number of state laws and regional planning policies. In the United States, the most advanced model of new urbanist policies can be found in Portland, Oregon, where a long-range plan has been adopted by an elected regional government.1 This plan involves an urban growth boundary;2 concentrated employment and high-density housing patterns, such as town houses and apartments; significant expansion of the light rail system; and little street or highway expansion.
New urbanist policies, especially as adopted in Portland, have evoked considerable interest among legislators, local officials, and civic leaders around the world. There are, however, difficulties with new urbanism, both in terms of analysis and in terms of policies.
Higher concentrations of urban residential and employment density will produce higher concentrations of automobile traffic (and air pollution). This is already evident. Contrary to new urbanist claims, traffic congestion is already worse in urban areas with higher densities.
Urban areas with higher levels of traffic congestion, as measured by the federal government's Roadway Congestion Index, have higher population densities (see Chart 1).3 This is to be expected, since higher density means less road space on which to accommodate the high volume of private vehicle traffic.
Transit-oriented development increases traffic congestion. Except in a very few centers, such as Midtown Manhattan and Chicago's Loop,4 a majority of trips are by automobile. The overwhelming majority of travel to proposed transit-oriented developments--which include high-density housing, retail, and employment located around transit stations, especially rail--will be by automobile (new employment centers attract from six to 100 times as many automobile commuters as transit commuters). The higher concentrations of employment and residences therefore must bring an increase in automobile trips in the area. This will strain road space, slowing traffic and increasing pollution as a consequence.
Higher levels of air pollution are associated with higher densities, not lower densities. Generally, the greater the intensity of air pollution, the higher the population density (see Chart 2).5 As transit-oriented development increases traffic, it will reduce speeds and increase pollution, because higher pollution is associated with slower, more congested traffic. To the extent that new urbanist policies are implemented, air pollution is likely to be increased relative to levels that would be experienced in less dense environments.6
Expanding urban areas do not threaten agricultural production. Since 1950, U.S. agricultural acreage has fallen by 15 percent, while production has risen by more than 105 percent (see Chart 3). The area required for agricultural production has declined, quite independently of urban expansion.
Between 1960 and 1990, the area taken out of agricultural production was greater than that of Texas and more than eight times the area consumed by expanding urban areas (see Chart 4). At current rates of urban expansion, it would take more than 250 years to urbanize the amount of agricultural land taken out of production between 1960 and 1990.7
Urban expansion is far too complex to be blamed simply on the automobile and interstate highways. First of all, urban interstates largely were not open until the early 1960s (the Interstate Highway Act was enacted in 1956). Yet the suburbs already were gaining population at the expense of the central cities.
During the 1950s, the major central cities that did not expand by annexation lost approximately 5.0 percent of their population. Similar rates of pre-interstate urban population loss occurred in the 1960s (7.2 percent) and 1980s (5.7 percent).8 Only during the 1970s was the rate significantly higher, at 14.6 percent. Other factors, such as escalating crime rates, the urban riots of the 1960s, and declining educational performance in central city school districts, probably were much more responsible for flight from the central cities.
Indeed, the 1970s, during which urban flight was the greatest, followed closely on the urban unrest of the 1960s and was also a period of particular deterioration with respect to the crime rate and educational performance. Additional contributing factors included higher central city taxes, lower quality central city services, and increasing affluence, which allowed people the option of living in larger houses on larger lots.
Despite the popular misconception, public service costs tend to be lower where population densities are lower.9 There are a number of reasons why the reality differs from the theory on urban costs. For example, the larger, more dense local government units tend to have larger bureaucracies, and their political processes are more susceptible to special-interest control. Both of these factors tend to increase costs.10
Increasing density and growth restrictions are likely to have a negative impact on economic growth in metropolitan areas adopting new urbanist policies. For example, even Portland's new urbanist regional government (Metro) found that higher densities and lower automobile usage rates appear to be associated with "higher housing prices and reduced housing output."11
As a result of higher housing prices, new urbanist policies are likely to make the American dream of home ownership more elusive. By limiting housing output, they are likely to limit job creation in construction trades and allied fields. Further, discouraging construction of additional suburban shopping centers can be expected to raise the cost of living while retarding job growth even more. Broad implementation of new urbanist policies could well bring to the United States the economic stagnation that afflicts Europe, where minimal job creation and high unemployment are associated with a high-cost and less competitive economy.
Portland's new urbanist policies will not deliver lower levels of traffic congestion and air pollution. Portland's regional government, Metro, has stated that "[W]ith respect to density and road per capita mileage it (Los Angeles) displays an investment pattern we desire to replicate."12 In fact, Portland is well on the way to replicating the traffic congestion problems of Los Angeles.
Traffic congestion in Portland already is approaching that of the New York metropolitan area-which is 15 times larger-and Portland projections indicate that, even after building five additional light rail lines,13 traffic volumes will rise by more than 50 percent by 2015. It is estimated that Portland's Roadway Congestion Index will rise to 1.62 from its current 1.16 (see Chart 5). This would represent a worse level of traffic congestion than is currently experienced by Los Angeles (which has the highest Roadway Congestion Index in the nation).
Portland seems to have chosen a future with two million cars in 500 square miles instead of 600 square miles. It can be expected that air pollution will be greater as a result.14
European cities are suburbanizing, despite their higher population densities, more comprehensive transit systems, higher gasoline prices, lower income,15 and more focused cities.16 Like their American counterparts, many European central cities have lost population.
No freeways enter the central city of Paris, which has one of the world's most intensive rail transit systems. Yet Paris's central city population loss and suburban population explosion mirror those of Philadelphia, a metropolitan area that has experienced similar overall growth (see Chart 6). At the same time, both traffic congestion and air pollution are severe. Average automobile travel speed in the city of Paris is 12.5 miles per hour.17
Inner London and Manhattan (inner New York) lost similar percentages of population over a period of 40 years until 1990-1991 (25 percent and 24 percent, respectively).
The cities of Copenhagen, Liverpool, Manchester, and Glasgow lost approximately 40 percent of their population in the past 40 years. By comparison, Detroit and Cleveland lost 45 percent, Newark lost 39 percent, and Washington lost 32 percent. In each of these European and American cities, all growth was suburban growth.
The central city of Stockholm has lost 16 percent of its population since 1950, with all growth occurring in the suburbs.
While San Francisco's population was rising by 1 percent from 1970 to 1990, Toronto's fell by 8 percent and Montreal's fell by 20 percent.
Tokyo's population has fallen by more than two million since 1960, with all population growth occurring in the suburbs.
Central area populations have fallen in virtually all cities in the developed world.18 In most cases, the declines are masked by population added through annexation or consolidation. In fact, central area depopulation and suburban expansion have been occurring for some time. Inner London began losing population between 1901 and 1911, while Manhattan began losing population between 1910 and 1920. Central area depopulation was first noted in Philadelphia between 1820 and 1830, as people moved to the suburbs.19
The depopulation of central cities in Europe and other developed nations is particularly notable because these cities generally did not face important factors that contributed to the depopulation of U.S. central cities, such as high crime rates, urban riots, forced busing, falling education standards, freeways, and home mortgage tax deductions. In addition, Europe's much stronger land use policies, higher suburban land costs, and overall higher cost structure might have been expected to forestall suburbanization.
Europe's comparatively high public transit market share has led to the mistaken impression that transit is gaining at the expense of the automobile. This is not the case. European automobile use has grown at three times the U.S. rate since 1970, largely as a result of increasing affluence.
In recent decades, transit market shares have dropped from even higher levels in Europe as increased affluence has made the automobile affordable for more people. In Europe (as in the United States), urban rail's record in attracting people away from automobiles has been insignificant: No such transfer has taken place.20 Europe's trend toward higher automobile dependency and lower transit market shares is following U.S. trends by a decade or two, just as its rising affluence has followed U.S. trends.
By imposing urban growth boundaries, new urbanists hope to force higher densities and infill development. No material increase in density is likely to occur, except where the urban growth boundaries encompass wide expanses of undeveloped land (as was the case in Portland when its urban growth boundary was established).
Even Portland's draconian policies are projected to increase densities to a level less than that of Los Angeles. Portland will continue to have densities barely one-quarter those of Paris, which is highly automobile dependent except in the inner city. While new urbanist policies may produce small reductions in average automobile miles traveled per capita, the increasing traffic congestion is likely to generate a more than compensating increase in the average hours per capita traveled by automobile. This will increase air pollution and retard the quality of life by reducing leisure time.
Queen Elizabeth I established an urban growth boundary in London in 1580.21 Development continued outside the urban growth boundaries.
King Louis XIII established an urban growth boundary in Paris in 1638. It failed to contain development, as did subsequent urban growth boundaries established by Louis XIV and Louis XV.22
London imposed an urban growth boundary by purchasing a "Green Belt" surrounding the city in the 1930s. Since that time, London's population density inside the Green Belt has fallen as 1.5 million people have left the city. Inner London's population dropped 43 percent, while that of outer London (the pre-1940 suburbs inside the Green Belt) rose 12 percent. Population in the surrounding counties increased 273 percent23 as development "leap-frogged" across the urban growth boundary to exurban areas beyond the Green Belt (see Chart 7). The 1931 census indicated that 19 percent of the population was outside what was to become the Green Belt. The 1991 census showed that more than one-half of the population was in the outer counties.
Despite all the criticism, America's spacious urban areas provide significant advantages. Their very geographical expansion has provided a safety valve that has kept travel times relatively stable.24
Average peak hour commuting time fell approximately 6 percent from 1969 to 1995 (from 22.0 minutes to 20.7 minutes).25
The automobile has improved travel times. According to the United States Department of Transportation, one of the most important reasons that average commuting time has not increased materially over the past 25 years is that people have abandoned transit services for automobiles, which are considerably faster.26 The average transit commute trip takes approximately 80 percent longer than the average automobile commuter trip (see Chart 8).27
The flexibility of the automobile has improved the efficiency of labor markets, making a much larger market of employers and employees conveniently accessible to one another.
The competition provided by large suburban shopping malls and retailers has lowered consumer prices.
spacious urban area, with its increased retail competition and more
efficient labor markets, has helped to create a comparatively
low-cost economy in the United States. It is likely that these
advantages have contributed to America's unparalleled standard
This is not to suggest that traffic congestion is not a problem. But today's urban motorist experiences much greater mobility and speed than can be provided by any practical alternatives. The question is not how governments are going to force people out of their cars, but whether capacity will be provided for the traffic growth that will occur regardless of which measures are adopted. Unless the automobile is accommodated, traffic can and will get much worse. Few places in the United States experience the intractable traffic congestion that is a day-to-day occurrence in the largest centers of Europe, despite higher densities, rail transit, and strong land use controls.
The fundamental problem with the new urbanism is that, despite aggressive planning policies, it is incapable of either increasing densities or materially improving the match between origins and destinations sufficiently to make alternatives to the automobile viable. Much stronger land use policies and much higher densities in suburban Stockholm failed to produce the anticipated reliance on rail transit, as automobile use continued to increase substantially.29 It is "neither certain nor self evident" that new urbanist policies, if they were to occur, would reduce traffic congestion.30
The new urbanist city would be only marginally more dense than today's spacious city, and travel patterns would be little different. The overwhelming majority of travel would continue to be by automobile. Even more than today, American urban areas would remain far below the "critical mass" that would generate significant ridership and too dense to avoid intractable traffic congestion. As a result, consistent with the plans of Portland, the higher density would worsen traffic congestion. The simple fact is that more cars in a more compact area mean more traffic and more air pollution, not less.
A more appropriate term than "new urbanism" might be "new suburbanism." At most, new urbanist policies will produce small enclaves of somewhat higher density surrounded by a sea of low-density suburbs. New urbanist policies could hasten the coming of a new suburbanization, with a much less dense urban sprawl than already has been experienced. More people are likely to choose to live outside the urban growth boundary, in smaller communities which gradually will become larger and more urban. More businesses are likely to locate outside major urban areas. Residents inside urban growth boundaries will make longer journeys to shop at the new, larger retail establishments in exurban areas.
New urbanist policies are being proposed at the very time that information technology (such as the Internet) threatens to make urban centers less important. Already, major urban centers have few advantages over medium and smaller sized urban areas. Generally, these smaller areas have virtually everything that major centers have except for international airports.
Previous generations of urban planners imposed their visions of a better city through policies such as urban renewal and construction of high-rise public housing. These planners believed in their theories just as devoutly as do today's new urbanists. It is not impossible that to analysts a quarter-century from now, the new urbanism will seem every bit as anti-city as any of the failed policies of the past.
Wendell Cox is Principal, the Wendell Cox Consultancy, St. Louis, Missouri.
2. The urban growth boundary requirement was imposed by state law in the 1970s. At that point, the urban growth boundary was established well outside the limits of development. In recent years, development has approached the urban growth boundary.
6. Because of the continuing improvement in air pollution that is attributable to improved vehicle emission technology, aggregate levels of air pollution could be reduced from present levels even with the higher concentrations of automobile traffic that would result from new urbanist policies.
7. Typical of the invalid data on which new urbanist proposals are based, President Clinton indicated that "farmland and open spaces are disappearing at a truly alarming rate. In fact, across the country, we lose about 7,000 acres every single day." This might be alarming if it were true. Such a rate would consume an area the size of Ohio every decade. In fact, however, the President's figure is off by a factor of nearly three. See "Remarks by the President and Vice-President on Announcement of Lands Legacy Initiative," January 12, 1999; available on the Internet at www.whitehouse.gov/WH/New/html/19990112-1036.html.
9. For example, see Helen F. Ladd, "Population Growth, Density and the Costs of Providing Public Services," Urban Studies, Vol. 2 (1992), pp. 273-295, and Wendell Cox, Local and Regional Governance in the Greater Toronto Area: A Review of the Alternatives, City of Toronto, 1997.
14. Vice President Gore has provided another example of the invalid data used to promote the new urbanist agenda. In a September 2, 1998, speech to the Brookings Institution, he indicated that "a new light rail system has attracted 40% of all commuters" in Portland. In fact, Census Bureau data indicate that only 5.4 percent of commuters used transit and that the vast majority of these commuters were on buses, not light rail. During the 1980s, when light rail was opened, transit's work trip market share dropped by one-third in Portland.
28. According to the latest OECD data, the United States had the highest gross domestic product per capita of any major nation (on a purchasing parity power basis, which measures cost of living). One small nation was higher: Luxembourg, with 418,000 people (1996), would rank 94th if it were a U.S. metropolitan area, just ahead of Modesto, California.
29. Sir Peter Hall, in Cities in Civilization (New York: Pantheon, 1998), pp. 842-887, describes the resistance of Stockholm area residents to planning dictates which required that suburban development be on rail lines and at higher housing densities. In recent years, most new housing has been single-family detached, and automobile dependency has increased.