Skip to content

About PERC

All Areas of Focus

All Research

Driving and Density

  • Randal O’Toole

  • One technique proposed for
    combating “urban sprawl”
    is increasing population densities.
    Government planners are
    beginning to require new residential
    developments to house
    more people per acre. Even existing
    low-density suburbs are
    supposed to be rebuilt to higher

    The “smart growth” plan
    for Portland, Oregon, considered
    a model for anti-sprawl
    policies, calls for increasing the
    density of the entire urban area,
    suburbs and all, from under
    3,000 to nearly 5,000 people per
    square mile (Metro 1994). A report
    by Metro (1996, 20), Portland”s regional planning
    agency, says that “congestion signals positive
    urban development.”

    In reality, however, most people have always
    lived in fairly low densities. They don”t necessarily
    want to change that fact. Nor is increasing density
    likely to achieve the objectives of its proponents.

    It was not until 1920 that the Census Bureau
    counted more Americans living in urban areas than
    in rural areas (U.S. Census Bureau 1995). And by
    that year, many city dwellers were already moving to
    low-density suburbs (which census takers count as
    urban rather than rural). In 1990, the Census Bureau
    found that nearly half of all Americans lived in the
    suburbs. Half the remainder lived in rural areas (U.S.
    Census Bureau 1993). In other words, only a quarter
    of the nation lives in what the Census Bureau calls
    “central cities.”

    A careful scrutiny of census data (U.S. Census
    Bureau 1993) shows that only a small percentage of
    Americans live in truly high densities:

    Only about half of all
    Americans live in densities of
    2,000 per square mile or higher.
    To put this in perspective, Indianapolis
    and Tulsa have densities
    around this figure. Half of all
    Americans live in cities that are
    this dense or more, while half
    live in even less crowded areas.
    About 28 percent live in rural
    areas, which have between zero
    and 200 people per square mile.

    Only 3 percent of Americans
    live in densities like New
    York”s 20,000 people per square
    mile. (Manhattan”s density is
    50,000 per square mile.)

    Only about 18 percent of
    Americans live in cities with densities above 5,000
    per square mile-cities such as Seattle, Washington,
    D.C., Chicago, and Boston. About a third live in densities
    of 3,000 or more (Houston and Atlanta have
    densities of about 3,000). In other words, two-thirds
    of all Americans live in cities with densities of 3,000
    per square mile or less. Bozeman, Montana, has about
    2,300 people per square mile, while Little Rock has
    1,700; Colorado Springs, 1,500; Chattanooga, 1,300.

    All these numbers count only the central cities.
    Suburbs around these cities are typically much lower.

    Advocates of higher density say that it will improve
    urban livability. Because people will live more
    closely together, they will reduce auto usage, and more
    people will walk or ride transit (Sierra Club 1998).

    But Census Bureau (1990) surveys (which asked
    one out of six households how they got to work) indicate
    that hopes for reducing auto usage are unrealistic.
    Ninety percent of commuters drive to work until
    densities are above 5,000 per square mile. Even above that density, huge changes in density are needed to
    significantly change driving behavior.

    For example, quadrupling Seattle”s density from
    5,000 per square mile to New York”s 20,000 might cut
    per capita driving in half. But with four times as many
    people, twice as much traffic would be on the roads.

    It is not even likely that simply quadrupling density
    would lead Seattle to achieve New York”s low
    driving rates. Unlike Seattle, New York enjoys very
    high employment densities as well as a historically
    dense transit network . Most American cities have
    widely dispersed employment, with less than 10 percent
    of jobs located downtown. Mass transit is only efficient
    when employment densities
    are high.

    “Smart growth” is even more
    impractical when applied to entire
    urban areas-meaning the central
    cities and their suburbs. Only three of
    the nation”s 400 urban areas have
    densities greater than 5,000 per
    square mile. For the rest, even doubling
    density would reduce per capita
    auto driving by only about 5 percent.
    Twice as many people, each driving 95 percent as much translates to a 90 percent increase
    in traffic. Since “smart growth” calls for few to no new
    roads, this means far more congestion.

    Once we understand that the “smart growth” solutions
    will increase congestion, we can begin to
    understand their appeal to people who expect to benefit
    from more crowds. Transit agencies love density.
    The Metropolitan Council (1996, 54), which operates
    transit in Minnesota”s Twin Cities, says that “as
    traffic congestion builds, alternative travel modes will
    become more attractive.”

    Big city mayors and officials also love these policies,
    which will give them more control over the suburbs.
    Portland city councilor Charles Hales makes no
    bones about his dislike of the suburbs around his city.
    He has called the suburbs “trashy . . . godawful subdivisions”
    (quoted in Ehrenhalt 1997, 23). Officials
    and business executives tied to urban downtowns also
    like “smart growth.” It would reduce the low-congestion
    advantage that suburban shopping and office
    centers have over downtowns.

    Ironically, the push for congestion is at odds with
    worldwide trends toward lower densities. Today”s
    “smart-growth” policies echo those adopted by most European countries after World War II, when they
    decided to emphasize high-density housing and transit
    and heavily tax autos and gasoline. Yet today, European
    cities are losing population, their suburbs are
    growing rapidly, urban densities are falling, auto ownership
    and usage is rising, and transit usage is stagnant
    or declining (Kenworthy and Laube 1999).

    “In worldwide perspective, rapid growth of automobiles
    began in the United States because we were
    richer than other nations,” says University of California
    (Irvine) economist Charles Lave (1992, 11). “But
    other nations headed down the same path as their incomes
    increased.” Lave concludes that “the desire for
    personal mobility seems to be unstoppable.”

    Policies to increase density will be an urban disaster.
    In the years ahead, they will have
    the opposite of the intended effect,
    leading to an even more rapid movement
    away from the cities.

    Ehrenhalt, Alan. 1997. The Great
    Wall of Portland. Governing,
    May, 20Ð24.
    Lave, Charles. 1992. Cars and Demographics.
    Access 1: 4Ð11.
    Kenworthy, Jeffrey R., and Felix B. Laube. 1999. An
    International Sourcebook of Automobile Dependence
    in Cities
    , 1960-1990. Boulder, CO: University
    of Colorado Press.
    Metro. 1994. Regional 2040 Recommended Alternative
    Technical Appendix
    . Portland, OR, September 15.
    —. 1996. Regional Transportation Plan Update.
    Chapter 1. Portland, OR, March 22.
    Metropolitan Council. 1996. Regional Transportation
    . St. Paul, MN: Metropolitan Council.
    Sierra Club. 1998. The Dark Side of the American
    . San Francisco: Sierra Club.
    U.S. Census Bureau. 1990. Census Lookup. Current
    Level: State. Available:
    —. 1993. Land Area, Population, and Density for
    Places: 1990. Available:
    www/censusdata/places.html. (March 23).
    —. 1995. Urban and Rural Population: 1900 to
    1990. October. Available:

    Randal O’Toole is senior economist with the Thoreau Institute

    Written By
    Related Content