Exploring Sprawl:

Findings of a Comprehensive Review of the Literature

Related to “Sprawl”


What Do We Really Know?





Nancy Frank




Sammis White

Zhong-Ren Peng

Kirk Harris

Welford Sanders



Department of Urban Planning

University of Wisconsin—Milwaukee


Paper presented at the Association of Collegiate Schools of Planning, Atlanta, Georgia, November 2-5, 2000.



            As anti-sprawl and smart growth initiatives become more ubiquitous, planning academics will find public constituencies expecting us to be able to provide clear and concise statements about “what we know” about sprawl.  Since the topic crosses normal sub-disciplinary boundaries, planning academics find themselves unprepared to answer the kinds of questions that policy makers ask as they prepare to respond to concerns about sprawl.  Anticipating this challenge, the faculty of the Department of Urban Planning at the University of Wisconsin—Milwaukee undertook a series of seminars to develop our own internal capacity to respond to these sorts of questions.  This paper is intended to summarize “what we know.”  Each section of the paper will report findings, dissenting views, and specific research questions for future research.  When possible, research strategies that might answer the question will also be proposed. 



In 1956, William Whyte published Suburban Sprawl, recognizing that American cities were changing.  Ever since, scholars and public commentators have disagreed about whether sprawl is an evil scourge upon the land or simply a different pattern of development that is neither good nor bad.

In recent years, concern about sprawl appears to have accelerated in tandem with the march of new subdivisions and strip malls across the landscape.  Numerous compilations of the “facts” about sprawl have been produced by groups as diverse as the Sierra Club and the Cato Institute.  Frequently, the political and ideological biases are so obvious that the validity of the “facts” presented is a trifle doubtful.  The findings produced by these groups range from characterizing sprawl, on the one hand, as the root cause of many of our most pressing social problems to characterizing attempts to control sprawl as a threat to national economic survival.

Acceleration in the passage of state mandates for local comprehensive planning and other land use measures, often under the banner of “Smart Growth,” has followed in the wake of increasing public concern about urban sprawl.  “In 1999 alone, approximately 1000 land-use reform bills have been introduced in state legislatures, with 200 of them enacted into law” (American Planning Association 2000).

In 1999, in light of the ongoing debates about sprawl, the faculty of the Department of Urban Planning at the University of Wisconsin—Milwaukee recognized a need to become better informed about sprawl.  Individual faculty members already possessed slices of knowledge related to our own individual specializations, whether in economic development (e.g., jobs-housing mismatch issues), transportation (e.g., the growth in vehicle miles of travel and associated congestion), environmental planning (e.g., air quality, water quality, and habitat loss), or urban design (e.g., New Urbanism). 

Yet none of us possessed a sufficient breadth of knowledge about sprawl to be able to answer some basic questions.  Our goal was to conduct a thorough and neutral review of the evidence to answer four questions.

·        What are the causes of sprawl?

·        Is sprawl a “problem” or not?

·        If sprawl is a problem (that is, if the costs of sprawl are greater than its benefits), how serious a problem is it, relative to other problems facing communities and the nation?

·        What are the comparative advantages and disadvantages of various proposed solutions to solve the sprawl problem?

            With the support of a small grant from the Herzfeld Foundation, the faculty undertook a year-long exploration of sprawl.  Through a series of seminars, faculty members briefed one another on their particular corners of expertise related to sprawl.  In addition, guest experts were invited to offer briefings as needed to supplement the faculty’s own expertise (see Figure 1 for a list of topics and speakers). 

This paper summarizes the results of these seminars and offers a framework for pursuing research and policy development related to sprawl.  We will be preparing a subsequent draft of this paper, in a somewhat more condensed and readable form, to be distributed to local officials in the Milwaukee metropolitan area.

Given the limited time and resources devoted to the effort, the results are (regrettably) more limited than we might have hoped.  In the best of all possible worlds, we would have first compiled a truly comprehensive review, like Burchell’s (1998) review of the “costs of sprawl” literature, and then produced our readable summary from that.  What we offer here is more limited, but we hope at least as useful for its purposes.

            The topics that we included in the seminar series seemed to us to reflect the most commonly debated topics related to sprawl, especially relating to questions concerning the consequences of sprawl and the effectiveness of proposed alternatives. 

For each topic, the summaries are organized into five sections: Question, Bottom Line, Discussion, Uncertainties and Alternative Viewpoints, and Future Research.  The “Bottom Line” is intended to communicate in a few sentences the main “known facts” in answer to the question.  The “Discussion” elaborates on the evidence supporting the bottom line.  Since, for most of the questions, definitive answers are not yet possible, we briefly comment on uncertainties and alternative viewpoints.  Finally, we offer some suggestions for research that would help policy makers better understand the risks and benefits of sprawl and anti-sprawl remedies.

Question: What is Urban Sprawl?

Many discussions of sprawl begin with the admission that sprawl has not been well-defined (see, for example, Weitz and Moore 1998).  At the same time, the term “sprawl” has become value-loaded, so much so that we debated whether to use the term at all. 

Bottom Line: Lack of a clear definition of sprawl prevents research that could tell us whether sprawl is getting worse and whether specific policies are effective in curbing sprawl.  In addition, although the term “sprawl” has value-loaded, negative connotations, no other term quite captures the combination of land use patterns that sprawl represents. 



Many definitions of sprawl observe that sprawl is characterized by land development patterns that are inefficient and more costly than traditional development patterns. The costs that are most often mentioned are the costs of building and maintaining infrastructure, pollution, congestion, and destruction of resources, including farmland and habitat.

This defines sprawl in terms of consequences that some believe to be caused by sprawl.  Such definitions are useful if ones purpose is to persuade, but are not very useful if ones purpose is to understand.  Since we don’t yet know if sprawl causes the many ills that have been attributed to it, and since one purpose of research is to determine whether sprawl causes such ills, sprawl should not be defined in terms of outcomes.  Instead, sprawl should be defined in terms of the way land is used. 

Sprawl is frequently described as a development pattern that is low density and auto-dominant.   Ewing (1997) reviews 17 studies concerning sprawl and identifies four characteristics defining sprawl: low-density, strip development, scattered development, and leapfrog development.  He also notes that Gordon and Richardson (1997), in their criticism of planners who promote “compact cities,” suggest that sprawl is low density, dispersed, decentralized, polycentric, and suburban.  Yet Ewing finds fault with each of these characteristics as defining aspects of sprawl.  Ewing would define sprawl primarily in terms of measures of accessibility and the functionality of open space (Ewing 1997).

Many of the characteristics associated with sprawl relate to different facets of density.  Density has been measured in dozens of different ways (Churchman 1999).  Density means different things at different scales of analysis, for example.  An appropriate measure of density at the neighborhood scale is inadequate to measure density in a polycentric metropolitan area with large tracts of land reserved for public use or future development. Consequently, density is so inconsistently measured, across time and across jurisdictions, that this apparently simple measure can result in unreliable results.

Some of the possible measure of density and sprawl include (Burchell 1998; Wheeler 1998):

·        (de)concentration—measured in terms of the percentage of population or employment that is located in the central city of a metropolitan area (Burchell 1998);

·        lot sizes;

·        floor-to-area ratios—measuring the intensity of development on a parcel of land, especially for commercial and industrial development;

·        continuity of development—whether new development occurs adjacent to existing development or “leapfrogs” over open spaces, creating a patchwork of sparse development;

·        compact development—whether the buildings in a subdivision or commercial area are clustered into nodes of activity or are more spread out.

While concentration, lot sizes, and floor-to-area ratios can be easily measured, continuity and compactness are not so easily measured.  In addition, regional variations create inconsistencies in the interpretation of these measures in different parts of the country and over different periods of time.  For instance, lot sizes are actually smaller in  (sprawling?) West Coast suburbs than in other parts of the country (Ewing 1997).

            Another set of characteristics concerns the pattern of roads found in a neighborhood or an entire city.  Finely grained, grid-style road networks are associated with traditional development.  Sprawl, in contrast, has fewer route options and greater reliance on arterial streets and cul-de-sac street patterns.  While measuring the street pattern for research analysis is possible (see, for example, Friedman, Gordon, and Peers 1992), such measurements are more complex and require more effort than measuring density.


Uncertainties and Alternative Viewpoints

            Churchman discusses the subjective aspects of the concept of density.  A similar observation could be made about sprawl.  In our seminar discussions, we considered whether sprawl should be defined and operationalized in terms of the subjective experience of sprawl.  An advantage of this approach is that the measurement of sprawl would more clearly correspond to what the public perceives as undesirable about certain development patterns. 

We do not yet know whether subjective definitions are consistent across different cultural contexts.  For example, image of urban sprawl existing in the minds of residents of Tokyo or Berlin might be quite different than the image obtained from residents of Los Angeles or Phoenix, which might be different still from that of residents of Milwaukee. 


Future Research:

            A satisfactory sprawl index needs to be developed that measures density in several ways (e.g., metropolitan concentration, neighborhood compactness, continuity of development) and captures elements of sprawl that area not directly related to density (e.g., street and road patterns).  To the extent possible, the index should be comprised of elements for which data is likely to be available, or easily collected, in most metropolitan areas.

            Since virtually all major cities in the U.S. have become polycentric rather than monocentric, measures might consider the proportion of all employment or housing in subcenters as an indicator of sprawl, as well as other attributes of subcenters (e.g., distance or travel time between subcenters).

Finally, researchers should consider whether varying definitions of sprawl are needed to characterize sprawl in large metropolitan areas, in smaller cities, in rural villages, and in the countryside.

Question: What Causes Sprawl?

            Understanding the causes of sprawl allows one to better understand some of the challenges involved in using public policy to shift patterns of land development to be more compact, dense, and concentrated.

Bottom Line: Sprawl has multiple causes.  Some of the most powerful causes—the rent gradient, demographic changes, growing affluence, and the economics of land assembly—are largely beyond the reach of public policy.  Others, such as the costs of transportation, the mortgage interest deduction, and land use regulations, could be modified through public policy changes, but the political feasibility of getting such policies enacted is quite slim.  The recent public reactions against fuel taxes in both the U.S. and Europe underscore the political barriers to policy changes in these areas.



A remarkable degree of agreement appears to exist concerning the causes of sprawl.  Most treatments of the subject identify a combination of economic forces and social and technological changes that have interacted to produce sprawl.  The principle points of contention relate to the emphasis placed on market failure versus policy failure in producing “excessive” sprawl. 

The summary presented at our seminar by Richard Green, professor of real estate and urban land economics at the University of Wisconsin—Madison, is substantially similar to a summary by Anas, Boyce, Hopkins, Knaap (1999) and other sources.  Green (2000) identifies nine causes of sprawl.  The list is weighted toward those causes that seem to be predominant today, rather than some of the causes that were more important in the 1950s and 1960s.

·        Rent Gradient—Where land is less expensive (such as on the periphery of developed areas), people will consume more of it for each house or shop they build (Green 2000).   This effect can be observed across historical and geographic contexts.

·        Demographic Changes—Household size has declined steadily since World War II.  As a result, “even if population had stayed constant, and housing unit density had remained unchanged, the amount of land required to house the population would have increased by 40 percent between 1940 and now” (Green 2000).  Since population doubled in that period as well, demographic changes alone (population growth and household size) account for almost a tripling of the amount of land devoted to housing between 1940 and now.  In comparison, each household uses about twice as much land now than in 1940.  In proportionate terms, demographic changes—rather than land use changes—are the larger factor.  But in an additive sense, tripling the number of households at a time when households are using twice as much land as previously, creates the conditions that we call sprawl.

·        Growing Affluence—Affluent people buy more land and build larger homes.  “Across all income groups, households are materially better off now than they were in 1945 (or in 1965 for that matter)” (Green 2000).  This affluence has not been evenly distributed, and has come about through a substantial increase in the number of hours of work required to attain this standard of living.  Nonetheless, “the number of households earning in excess of $100,000 in real income has increased six-fold over the last ten years” (Green 2000).  Because of this affluence, Americans have done in greater numbers what people with increasing affluence have always done—they have consumed more.

·        Transportation—Before the automobile, the ability of cities to spread out in response to the rent gradient was limited.  Because the costs of traveling farther out from the center would quickly exceed the economic benefits, cities remained more compact.  The automobile did not repeal the law of the rent gradient, but it did expand the geographic radius within which consumers could seek to substitute transportation costs for land costs. 

Consumers are individually better off by avoiding high land costs in the centers of cities by spending just a bit more on transportation (in time and fuel).  If transportation is fast and cheap, people can save more by moving farther, and with the money they save, they can afford to buy even more land and buy an even larger house.  This effect—and the radius people are willing to travel—increases when transportation costs are kept artificially low through subsidies (see Transportation section of this paper).

·        Differences in Government Services and Attitudes—Local units of government compete with each other for residents by offering packages of services (including public schools) and taxes that will attract people and businesses to their communities.   This empirical fact puts older cities at a disadvantage relative to newer cities arising on the periphery.  Older cities have older infrastructure, that is more expensive to operate and maintain.  “Second, and perhaps more important, newer cities have used land use regulation to prevent the development of low-cost housing.  Old cities, on the other hand, have old housing, which tends to be of lower quality than newer housing and which also tends to be affordable” (Green 2000).  As a result, low-income people, who require a disproportionate share of government services, are concentrated in (older) central cities.  After spending disproportionately more on infrastructure maintenance and services for the poor, older cities are less able to offer the sorts of public services sought by middle-class homebuyers.  Since the middle-class home buyer can move to newer cities that offer the public services and low taxes buyers seek, the middle-class flees to the suburbs.

·        Racial Discrimination and Segregation—Discrimination and segregation in housing markets has been a serious and tragic cause of sprawl.  In a recent survey, 50 percent of white survey respondents reported that they would be unwilling to move into a neighborhood that was 21 percent black (Harris 2000).  Similar attitudes have operated throughout the post-World War II period, resulting in the development of land at the periphery that would not have been if people’s location decisions were race neutral (Green 2000).

·        Land Assembly—Obtaining a parcel of land to develop is inherently easier at the periphery than in the center of a city.  At the periphery, large parcels are available that can be purchased from a single owner.  In the city a developer must negotiate with several owners to assemble a large enough site.  In addition, one owner may “hold out,” demanding a premium through their monopoly over the last parcel.  This inherent advantage to “greenfield” development is compounded when central city sites are likely to be environmentally contaminated and require expensive clean-up.

·        Federal Tax Policy—The mortgage interest deduction encourages higher-income home buyers to purchase larger, more expensive homes, generally on larger lots.  The mortgage interest deduction does little to increase the percentage of households that are homeowners, because at lower income levels, the tax benefits are too insignificant to serve as an inducement to home-buying.

·        Land Use Regulation—Setbacks, street widths, minimum parking requirements, and the like virtually legislate sprawl in many newer communities.  Some regulations are enshrined in the canon of city planning and city engineering.  Suburban land use regulations were designed to avoid the problems of congestion experienced in older cities, and from which suburban residents had fled.  Overcrowding, lack of adequate parking facilities in commercial areas, gridlock on city streets, inability of modern public works and emergency equipment to easily access older neighborhoods—all were to be prevented by requiring more spread-out development patterns.  These policies are supported by residents, who generally prefer densities the same as or less than where they currently reside (Brown, Fuguitt, Heaton, and Waseem 1997).

In addition, in many communities land use regulations that promote low-density living are adopted to keep housing prices high, sometimes justified by the “need” to assure that the property taxes generated from housing is sufficient to support the services its residents will require.  Whether intended or not, these regulations prevent low-income (often minority) households from settling in the community (Pendall 2000) and simultaneously legislate sprawl. 


Uncertainties and Alternative Viewpoints:

A surprising level of consensus appears to exist concerning the causes of sprawl.  Some sources mention additional causes not cited by here.  These include reduced financial support for public transit while simultaneously investing in roads; other federal policies (especially federal housing policy during the 1950s and 1960s favoring single-family construction in the suburbs), deindustrialization of central cities, changes in retailing (first to enclosed malls and later to “big box” discount chains), and reliance on “trickle down” to provide low-income housing. 


Future Research:

            More research is needed assessing the magnitude of market failure in the production of sprawl. 
Question: What is the impact of sprawl on economic efficiency?

Given the strong influence of market forces in creating sprawl, economists would expect sprawl to produce economic efficiency in the absence of market failures.  In contrast, sprawl is often defined as an inefficient form of land development.  Critics of sprawl contend that, though sprawl improves private economic gain (or welfare) for some, it imposes costs on others, for a net loss to society.  If this is true, government intervention in the market to prevent sprawl can be justified by welfare economics.

Bottom Line:            A 1993 Brookings Institution study found that urban form had only “modest effects on infrastructure costs” (Hayward 1998, 10, quoting the Brookings Institution report).  Even Burchell’s work shows overall cost differentials to be rather modest, in the range of 10% higher costs for conventional compared to compact development.  Because the benefits of sprawl have not been carefully measured, little research has been conducted on the overall economic efficiency of sprawl.



A growing literature on the costs of sprawl argues that sprawl is not economically efficient because it creates a host of private, public, and social costs that are not adequately captured through market processes.  Countering this view have been a number of antagonists, attempting to de-bunk the “sprawl costs all” rhetoric (see, for example, Gordon and Richardson 1997).  Finally, new research casts doubt on the notion that suburban location is beneficial to the private sector in all metropolitan areas.

The “sprawl costs all” literature provides numerous analyses of the increased infrastructure costs associated with sprawling development compared to infill or contiguous and compact development.  Among the more rigorous studies are those by Burchell.  For example, Burchell and Shad (n.d.) synthesized the results from four major studies of the infrastructure costs associated with sprawl and found that the costs of compact growth are substantially lower than the costs of conventional suburban development.  For roads, compact development costs just 75% of conventional development, and for utilities, the costs of compact development are 85% of conventional development costs.  Savings for schools and other costs are much less significant, ranging from 0-5%.  Overall, Burchell’s work finds sprawling development to cost about 10 percent more than more compact development. 

This result is supported by a cost/benefit assessment conducted by Persky and Wiewel (2000).  They conclude that after public costs and private benefits are taken into account, they find a very modest net public cost.  Their research suggests that deconcentration of urban areas is beneficial to the private sector in terms of increased efficiencies.  In Chicago, the nation’s third largest metro area, private business owners enjoy distinct benefits by locating in the suburbs.  In one example, a large manufacturing facility experiences an estimated total private benefit of $2.6 million annually (Persky and Wiewel 2000).  Most of the benefits result from lower wages in the suburban location, but also contributing are lower land and construction costs and lower property taxes. These results are consistent with our understanding of the causes of sprawl.  If sprawl is primarily a consequence of market forces, households and firms that locate in the suburbs, almost by definition, should be garnering advantages.  But this may not always be the case.

Surprisingly, White and Dent (2000) found that Milwaukee offers contrary evidence.  At least for some industries, suburban location offers no economic benefits in terms of  wages, land costs, and building costs.  White and Dent found that because of recent suburban labor shortages, wages in some occupations and industries are actually higher in the suburbs.  And in the smaller Milwaukee market, city and suburban land costs and building costs are similar.  Suburban tax rates may be lower, but this is not universally the case.  Thus, private benefits do not appear to justify the suburban location.  Other factors, such as fear of crime and its cost, lack of information, and lack of available building sites in the city, undoubtedly play a role since employers in Milwaukee have been locating facilities in the suburbs rather than in the city. 

Nonetheless, labor costs are beginning to turn the tide in some locations.  We may well find that with growing labor shortages, more employers will find it advantageous to locate in the central city, should space be available. 

Additional social costs of sprawl not included in the work described above include traffic congestion, environmental impacts, and equity costs.  These are discussed separately in the following sections.


Uncertainties and Alternative Viewpoints

Current snapshots of the efficiency of compact cities compared to sprawl are incomplete and inadequate.  New research will begin to answer these questions, but the complexity of these problems suggests that definitive answers are unlikely to available for at least the next decade.

Much higher costs of sprawl have been cited in some studies, especially by advocates of more compact development.  For example, the Sierra Club presents graphs showing that the costs of conventional development are two to five times that of compact development (Sierra Club 2000).  An introductory caption states, “Sprawl-like development can use many more resources -- five times more pipe and wire, five times as much heating and cooling energy -- than urban living. Sprawl also costs us 35 times as much land, and it requires 15 times as much pavement as compact urban living,” but no sources for these estimates are provided.


Future Research

Research is underway (e.g. Anas, Boyce, Hopkins, and Knaap 1999) that attempts to measure whether laissez-faire land development policies produce a more efficient city form than policies intended to promote a more compact city.  Results from research such as this may add considerably to what is now known about the accuracy of claims and counter-claims regarding the economic efficiency of sprawl.  The creation of models of urban dynamics that are capable of representing the economic, social, and land use results of alternative public policies would provide a stronger basis for reaching conclusions about the costs of sprawl. 

In addition, as the contrasting results from the Chicago (Persky and Weiwel 2000) and Milwaukee (White and Dent 2000) work demonstrates, research on costs and benefits needs to recognize the potential for variation between metropolitan areas related to differences in transportation costs (congestion and transit service), city/suburb wage differentials, and tax differentials. 

This also points to a need for longitudinal research.  For example, as the White and Dent study of Milwaukee suggests, given current economic and transportation conditions, private firms may begin to recognize the benefits of locating in the central city.  Subsequent changes in the regional economy and infrastructure, however, could reverse this (apparent) trend.
Question: What are the impacts of sprawl on transportation costs?

The enormous influence of the automobile is widely acknowledged as a cause of sprawl.  Less agreement exists, however, about the magnitude of subsidies to automobile transportation.  In addition, regardless of the size of the subsidy, observers disagree about the effects of sprawl on congestion.  One hypothesis suggests that, although automobile travel initially makes peripheral locations more attractive, once those peripheral areas become built up, congestion and travel times increase, and transportation costs actually rise in comparison to what they would have been if the city had developed in a more compact fashion. 

Bottom Line: Automobile transportation is subsidized, and that subsidy clearly facilitates sprawl.  What is less clear is whether the costs imposed through automobile use and the sprawl it produces exceed the benefits of automobile travel.



Automobile use is subsidized in the U.S. because users do not pay the costs imposed on others in the form of congestion, environmental damage (see following section), and other social costs.   The magnitude of the subsidy is not yet clear.  With autos as dominant, public authorities expand road and parking facilities.  As land uses become more spread out, the practicality of transit, walking, and biking is reduced. 

Congestion is difficult to measure, and reliable data on the effects of sprawl on congestion are not yet available.  However, on indicator that congestion may be increasing is person-hours of delay per driver, which has increased 54 percent for metropolitan areas from 1982 to 1996 (U.S. Department of Transportation 1999).  According to Burchell, et al., “both sides” in the congestion debate agree that sprawl and congestion are related, but they disagree about whether congestion is getting worse or better. 

Economists estimate that the costs of congestion are, on average, equivalent to $0.10 to $0.20 per vehicle mile (Straszheim 1979; Levinson, et al., 1996; Chicago Fed Letter 1999).  This is a total for all travel, however, and not specifically that fraction of congestion that might be attributable to sprawl.  Even within cities, people are driving more than in earlier periods, so that a substantial proportion of the congestion costs cited here cannot be laid at sprawl’s doorstep.

Even if congestion is worse, it may not have had any effect on the amount of time people spend commuting. Because of residential location decisions of commuters (Levinson and Kumar 1994) and increased speeds for many suburban commuters, commuting times have remained remarkably stable over the past century.  Between 1980 and 1990, commuting travel times increased by an almost imperceptible amount, from 21.7 minutes one-way to 22.4 minutes, or roughly 40 seconds (Pisarski 1996).  Some see this as a benefit of sprawl, because without sprawl, commuting times might have increased dramatically.  In addition, commuting times in central cities are not different from commuting times in suburbs.  Travel times for central city residents also remained relatively stable from 1980 to 1990, despite funding cuts in many transit systems, perhaps reflecting (or necessitating) the mode shift to personal vehicles in cities as well as suburbs (Pisarski 1996). 

Uncertainties and Alternative Viewpoints

Uncertainies are too numerous to catalogue here.  For a full review of the studies of sprawl and the social costs of transportation, Burchell, et al., (1998) and Murphy and DeLucchi (1998) offer good summaries. 

Although most studies find that suburbanization reduces commuting times, and that commuting time has remained constant over the last several decades—even the century—Vincent et al. (1994) report a 14% increase in travel times to work

None of these studies take the crucial step of attempting to quantify the proportion of all auto travel that can be attributed to sprawl.  Certainly, transit and other non-auto modes capture a larger share of  all travel in more densely settled areas.  In ultra-dense New York City, only 32 percent of trips to work are by car.   More typical are Milwaukee, Wisconsin, and Portland, Oregon, with about 80 percent of work trips being taken by car, and most of those with a single occupant.  Because cars capture a substantial proportion travel, especially for non-work trips, in even dense cities served by transit, only a small portion of the social costs of automobile transportation can be attributed to sprawl. 


Future Research

            Because suburban locations initially reduce congestion costs, but congestion rises as the suburban location develops, longitudinal analyses and simulations are needed to examine the cumulative costs over time for compact versus business-as-usual development patterns.
Question: What are the environmental impacts of sprawl?

Environmental groups have been at the forefront of the anti-sprawl movement.  Environmentalists cite sprawl as a significant factor in energy use and environmental degradation related to air quality and greenhouse gas emissions, surface water and groundwater quality, flooding risk, and habitat and wetland destruction. 

Bottom Line: An estimate of the environmental costs of sprawl may be in the range of one cent per vehicle mile of travel.  Of the total environmental costs of transportation (4-8 cents per gallon), less than 10 percent can be attributed to sprawl; most of the aggregate environmental costs of transportation would remain even if cities were more compact and transit-oriented.  The environmental costs of sprawl related to water quality, flooding, and habitat loss have not been quantified reliably.



Surprisingly little research has been done on the share of environmental harm that can be attributed to sprawling land development patterns.  Almost all of the research that has been conducted considers the only environmental impacts of transportation, in an effort to identify costs per vehicle mile of travel (VMT). 

If one can estimate the total environmental harm of personal vehicle travel, one may then estimate the proportion of that travel (and the related environmental damage) attributable to sprawl.

·        Ninety percent of the environmental costs of transportation are related to the impacts to air quality; impacts to water and habitat combined have been estimated at less than 10 percent of the environmental costs of transportation (Hanson 1992);

·        Estimates of the “total” environmental costs of transportation have been estimated to be equivalent to $0.01 to $0.61 per vehicle mile of travel.

·        (DeLulcchi 1995; Hanson 1992; DeCicco and Morris 1998; Litman 1995);

·        “Best guess” estimates of the environmental costs of transportation are in the range of four to eight cents per vehicle mile of travel, or about the same as the direct fuel costs of transportation.  Assuming a gas price of $1.50 per gallon and fuel efficiency of 25 miles per gallon, direct fuel costs are six cents per mile.

·        Of the total environmental costs of transportation, less than 10 percent can be attributed to sprawl; most of the aggregate environmental costs of transportation would remain even if cities were more compact and transit-oriented;

§        Portland’s LUTRAQ study found that a compact development alternative would reduce VMT and air emissions by about 6 percent (Crane 2000; Ewing 1997);

§        Increasing average gross density from 3,600 to 5,400 per square mile (a 50 percent increase in density) would reduce VMT by less than 3 percent (Schimek 1996).


Uncertainties and Alternative Viewpoints

The studies cited above may grossly underestimate the environmental costs of sprawl.  Not all of the costs operate through sprawl’s impacts on transportation choices.  Non-economic costs (premature death and illness from pollution, damage to wildlife, etc.) are not easily estimated in dollar terms.  Current research probably underestimates the impacts of sprawl on water quality and habitat. 

Another view argues that examining the transportation costs of sprawl without considering the benefits people derive from auto travel is incomplete and misleading (Richardson and Gordon 1999).  “Hogarty (1998, 1999a, and 1999b) estimates the net social benefits of personal vehicle travel in the United States to be about $2 trillion per year,” (Richardson and Gordon 1999, 12) or about $1.33 per passenger-vehicle mile (vehicle-mile figures from Bureau of Transportation Statistics 1999). 

Another cost of sprawl frequently cited, and related to open space losses, is the loss of farmland to suburban development.  The literature on this topic is vast and complex, yielding inconsistent conclusions.  Our review of this literature is not yet complete.


Future Research

More research is needed on the environmental and economic costs associated with stormwater management, flooding, and urban runoff and the fraction of these costs that are attributable to sprawl.  In addition, more research is needed assessing the costs of sprawl in relation to habitat destruction, loss of open space, and the like. 

Research is needed concerning the most cost-effective measures for preventing environmental damage associated with sprawl.  For example, Landis (1995) shows that a compact cities strategy would consume three times more environmentally sensitive land than a policy that was specifically targeted at maximum protection of sensitive lands.  Research is also needed on the political feasibility of anti-sprawl policies compared to expanded environmental regulations.

Finally, even if targeted policies would be more cost-effective than anti-sprawl policies in reducing environmental damage, other benefits derived from anti-sprawl policies might make such policies more socially efficient.
Question: Are the costs and benefits of sprawl fairly distributed?

Sprawl might (or might not) be economically efficient.  If the benefits of sprawl flow to one group of citizens, while the costs are borne disproportionately by another group, the inequity that results should be included among the costs of sprawl.

Bottom Line: Sprawl exacerbates underlying inequities in U.S. society. This effect is especially striking during the recent period of strong economic growth.  The spatial isolation of central city workers from suburban opportunity imposes costs on individuals, businesses, and on society.  The least advantaged citizens bear a disproportionate share of the burden because:

·        Even while central city unemployment rates remain high, suburban jobs go unfilled;

·        Central city residents experience much higher costs for commuting, traveling longer (though not necessarily farther) than suburban residents.



The combined effects of suburbanization and racial discrimination has resulted in increasing income polarization between central cities and suburbs, in which central city residents have become poorer and jobs have become less and less accessible.  According to a Russell Sage Foundation report (O’Connor, Tilly, and Bobo 2000), in most metropolitan areas, two-thirds of employers are located outside the central city.  Moreover, black central city residents are unable to take advantage of suburban opportunities for a variety of reasons.

·        Low-density (exclusionary) zoning and the resulting lack of affordable housing in suburbs serves as a barrier, preventing minorities from following jobs to the suburbs (Pendall 2000);

·        Racial discrimination in housing prevents even middle-class minorities, especially African-Americans, from locating in the suburbs.  “Housing discrimination segregates blacks in areas comparatively distant from employment” (HUD 1994);

·        Experiences of hostility from white residents discourages minorities from pursuing housing and job opportunities in the suburbs (O’Connor, Tilly, and Bobo 2000).

More spread out metro areas require central city residents to travel longer to work,  reducing access to jobs and resulting in unemployment and higher transportation costs (Khattak, Amerlunck and Quercia 2000; Persky and Weiwel 2000).

Finally, as noted in the section below on Growth Management, requiring cities to be more compact, thereby increasing the numbers of middle-class and white residents, might increase the base of public support for city services, including schools, public safety, and recreational facilities.


Uncertainties and Alternative Viewpoints

No one has seriously challenged the interpretation that the combined effects of deconcentration and hypersegregation have made minorities worse off (see, for example, Hawkins and Ihrke 1999).  The point of contention appears to be whether sprawl is the cause of hypersegregation and whether reversing sprawl would produce better outcomes for minorities.

Further Research

Further research is needed to determine whether compact metropolitan areas truly offer greater access to economic opportunity for minorities and the poor, all else being equal.  The alternative hypothesis might be that compact cities can also be hypersegregated and that affluent residents would shift local spending and programming to address their own interests, so that minorities would be no better off in compact cities than in sprawling ones.  However, if compact cities can improve opportunity for the worst-off in our cities, this is a benefit of sprawl reduction that should be given substantial weight.

Question: How promising are urban design solutions as a measure to reduce sprawl?

            New Urbanism, Traditional Neighborhood Development (TNDs), Transit-Oriented Development (TODs), cluster development, and Open Space Subdivisions are widely recommended as strategies to reduce the amount of land used for development.  Cluster development and Open Space Subdivisions are designed for rural areas.  Typically, the strategy is to maintain gross densities, but reduce the visual and environmental impact of the subdivision on the land by clustering buildings to preserve some of the land as open space.  New Urbanism, TNDs and TODs are designed for urban areas, whether a rural village, a large suburban development, or infill development in urban and suburban sites.  In theory, TNDs are intended to increase gross densities in urbanized areas and promote compact development.

            The promise of TNDs to solve the problems cited in the literature on sprawl depends on two things:

·        The willingness of a large segment of homebuyers to accept living in denser neighborhoods and more compact communities; and

·        The actual effectiveness of TNDs in reducing the social costs of development.

Bottom Line: Consumers prefer larger lots, but when forced to choose they will prefer a larger home on a smaller lot, rather than a smaller home on a larger lot.  Yet, as practiced up to now, New Urbanism has not delivered on its promises.  Only weak evidence is available regarding costs and benefits associated with New Urbanist developments.



Recent experiments with neo-traditional developments offer mixed results.  In some cases—for example, Celebration, Florida, and Kentlands in Maryland—demand for housing units appears to have outstripped supply, leading to rapidly escalating selling prices in these developments, and research on four New Urbanist developments found that buyers are willing to pay a premium of $20,000 on average to live in these communities (Eppli and Tu 1999).  In other cases, such as Middleton Hills, Wisconsin, the absorption rate has lagged behind expectations, despite a torrid housing market.

Commercial ventures within new urbanist developments have enjoyed less success.  According to Kerry Vandell, professor of real estate and urban land developmnet at UW-Madison, “mom and pop” retail has been almost wholly a failure in TNDs.  The only exceptions appear to be high-end specialty shopping, where service, custom products, and high prices allow smaller retailers to survive.  In short, new urbanist residents shop where they can take advantage of economies of scale, and those opportunities have not been included within many new urbanist developments.

Aside from mixed evidence regarding the marketability of new urbanist housing and commercial development, even if a substantial market exists, will new urbanist-style developments reverse the ills associated with sprawl?  The answer to this question is a clear “it depends.”  First, it depends on the location of the development in relation to the rest of the metro area and the quality of transit service provided.  Celebration and Kentlands, two of the best-known neotraditional developments, represent precisely the sort of leapfrog or discontinuous development that anti-sprawl policies are intended to avoid.  Many of the development projects that purport to be New Urbanist are not easily accessible to transit (Hall 2000). 

Second, it depends on the use of land within the development.  In some new urbanist communities, despite the narrower streets recommended by designers to make communities more pedestrian-oriented, the amount of paved surface is at least as much as in conventional developments because of the addition of paved alleyways (Southworth 1997; Vandell 2000).  Gross densities of these subdivisions are the same or even lower than in conventional subdivisions.  “Granny flat” units that were “designed in” to promote affordability and higher density tend to be used as family surge space rather than as an additional dwelling unit. Generous allotments of “public” open space make these developments livable, but also reduce gross densities.  Finally, evidence is weak at best regarding the influence of New Urbanism on travel behavior (Crane 2000).


Uncertainties and Alternative Viewpoints

The new traditional neighborhood developments that have been constructed have a number of shortcomings, including—in addition to the issues cited above—lack of affordability and an overly “planned,” sterile character.  Some question, however, whether judging the merits of New Urbanism on the basis of these early efforts is a fair assessment of the potential of New Urbanist principles if applied on a broader scale.  Given the financing and regulatory hurdles that many of these developments have had to overcome, perhaps these developments have been judged too harshly.  If implemented as part of an integrated regional plan, many of the shortcomings of new urbanism might be avoided.


Further Research

The poor state of our understanding of the real, on-the-ground impacts of new urbanism and related design interventions reflects some of the research challenges alluded to throughout this paper.  For example, if many of the new urbanist developments that have been studied are located in areas not contiguous to existing urban areas, these developments represent a different pattern of sprawl, rather than an example of compact development.

Question: What Have We Learned about the Costs and Benefits of Growth Management?

Two questions need to be addressed assessing growth management as a tool to reduce sprawl: (1) Are growth management strategies effective in reducing sprawl and (2) Does growth management create other costs, such as increased housing costs or reduced economic growth, that are greater than the benefits produced by reducing sprawl?

Bottom Line: Containment policies, such as those imposed in Portland and several other localities, appear to have been effective in promoting more compact development.  In addition, other indicators often associated with sprawl, such as vehicle-miles of travel, appear to have moved in beneficial directions in growth managed communities.  In short, if implementation hurdles can be overcome, growth management can work.  Evidence regarding costs, or unanticipated consequences, remains spotty.



Much has been written about growth management, but few conclusions emerge regarding the overall costs and benefits of growth management or even the effectiveness of different approaches in actually reducing sprawl.  “There have been few efforts to evaluate the implementation and performance of growth-management programs in individual communities, and they have been largely qualitative in approach” (Nelson and Peterman 2000, 283). 

Evidence is beginning to mount, however, that the growth boundary approach taken in Oregon and other communities has, in fact, limited development outside the growth boundary and concentrated development within designated areas.  “Boulder County, Portland, and Napa County have been successful in preventing farmland and open space from being developed, although the magnitude of this impact is not clear” (Staley, Edgens, and Mildner, n.d., 41)  In addition, growth boundaries appear to be effective in increasing densities.  Densities in Portland increased from 5 units per acre on average to 8 units from 1994-1997, even though densities were one-third lower than allowed (targeted) by land use policies (Staley, Edgens, and Mildner, n.d., 14).  The share of metropolitan area new housing occurring in the central city increased from 6.5% in 1988 to 18.1% in 1998 (Nelson 2000).

Metropolitan areas in growth management states (Oregon and Florida) have experienced measurably better results in relation to vehicle miles of travel, commute times, air quality, and per capita energy consumption compared to Atlanta, Georgia (a “laissez faire” state), even though these metro areas experienced similar population growth (Nelson 2000, 3; Nelson 1999).   Moreover, the Portland area continued to experience strong economic growth (Nelson 2000). 

The one gray cloud in an otherwise rosy economic picture has been the steep increase in housing prices in Portland.  Housing prices increased 62% from 1990-91 to 1995-96, compared to a 19% increase in Atlanta, an area experiencing similar population growth pressures during the period.  This increase in housing prices was mitigated by a rise in personal income—more than 72% in the decade between 1987 and 1997.  Nonetheless, Atlanta also had steeply rising personal income (60%), but housing prices that remained more moderate. 

Inner city neighborhoods gentrified, with costs increasing by 75-100 percent (Staley, Edgens, and Mildner n.d.).  Consequently, despite rising average incomes, Portland’s poor may have experienced steep increases in housing costs that were not accompanied by similar increases in income. 


Uncertainties and Alternative Viewpoints:

Although the comparisons Nelson has offered between Atlanta and Portland are tremendously intriguing, many questions remain unanswered concerning the public policy efforts that led to this result.  While the Urban Growth Boundary (UGB) and related land use control measures are probably the most visible policy differences between Oregon and Georgia, other policy differences might have been necessary for the growth management controls to have the effects observed.  Understanding these interactions will be essential if Portland’s success is to be replicated in other cities.


Future Research

Debates concerning the effectiveness of growth management are likely to continue because of the substantial obstacles to conducting good evaluations of growth management policies (Nelson 2000). 

Quasi-experimental (before-and-after) analysis is hindered by insufficient controls and the fact that only over time can growth management effects be detected.  Longitudinal analysis, such as Knaap’s and Nelson’s of Oregon (1992), may only be applicable to single states over time.  Cross-sectional analysis is often problematic because of the difficulty of acquiring data comparable to many locations.

Nelson’s comparisons of Portland and Atlanta are provocative, even exciting.  For a social researcher to find such dramatic reversals on such a large set of indicators is notable.

            Two important questions need to be examined in relation to this data, however.  First, are the changes in Portland attributable to the Urban Growth Boundary or to other policies or economic factors?  Second, can slower-growth metropolitan areas (such as  Rustbelt cities)  sustain regional economic growth if growth controls increase the costs of development?  Portland’s remarkable rate of growth generated the political will to pursue growth control.  Because of strong demand in Portland’s economy, it may have been able to absorb some of the economic costs in ways that a slow-growing regional economy could not. 
Question: Would metropolitan or regional government be more effective in controlling the impulses driving sprawl than the hundreds of independent (fragmented) local units of government characterizing most U.S. metropolitan areas?

            Various forms of metropolitan or regional governance have been suggested as antidotes to sprawl by creating a more unified policy toward the development of metropolitan areas (Rusk 1993; Orfield 1996).  Portland, Minneapolis, and Toronto are frequently cited as examples of the sorts of metro area governance solutions that are needed to curb sprawl.

Bottom Line: If a metropolitan governance structure can be formed (which is unlikely) and if the government has the will to promote compact development (which many do not), such governance structures may be effective in stemming sprawl.  Conversely, without some form of super-local governance—whether at the state or metro/regional level, achieving regulatory control over sprawl is unlikely to occur.  Nonetheless, care must be taken to avoid creating inefficiencies associated with large, consolidated local governments.



            Metropolitan government can be achieved through annexation (e.g., Indianapolis), consolidation (e.g., Toronto), or superimposing a regional governance structure that supercedes local authority in relation to land use and/or infrastructure decisions (e.g., Portland). 

            Annexations and consolidation reduce the number of local governments.  As a result, the expanded local government has more control over development at the edges of the metro area.  By reducing the number of local governments, consolidation also reduces consumer choice.  Research generally supports the Tiebout hypothesis, that consumers decide where to live based upon a package of goods and services, including tax rates and public services that a city can offer.  Since this is the case, diversity of governments in metropolitan areas creates a free market rather than a monopoly in which a single government may provide poor and inefficient government services and high cost to taxpayers.  In contrast to this perspective, some critics of suburbanization argue that suburbs impose costs on central cities, and propose the creation of metropolitan governments both to reduce intra-metropolitan inequalities by creating a unified taxing and service jurisdiction and to create a metropolitan-wide land use policy and regulatory structure.

            Hawkins and Ihrke (1999, 117-120) reviewed studies in an attempt to determine whether “larger and fewer governments saved money, were more efficient, solved problems that crossed municipal boundaries, and helped cities.”  They concluded that 16 of the 25 reports they reviewed offered evidence challenging these claims.  In Hawkins comments at our seminars, he also asserted the view that the current political trend is toward more fragmentation of governments in metropolitan areas rather than less.  He was skeptical of the political feasibility of consolidation.

            In short, metropolitan governance may be both expensive and ineffective.  But without some supra-local authority, metropolitan areas will have a difficult time controlling their patterns of growth. 

Uncertainties and Alternative Viewpoints

            A metropolitan or regional governance organization with more authority limited to planning and land use may be politically feasible and avoid the costs associated with consolidation.  Portland’s Metro may offer a model of the sorts of authority needed by a regional planning agency to have any significant impact on metropolitan form.


Future Research

            Comparison of the effectiveness of different forms of regional governance in stemming sprawl would provide a greater understanding of what works.  Freilich (1999) lists the following cities as examples of metro areas with regional goernance for planning: Portland (of course); Central Puget Sound area; San Diego; Minneapolis/St. Paul; “Research Triangle” area of North Carolina; a four county area in Baltimore; Reno-Washoe County, Nevado; Lexington-Fayette County, Kentucky; Palm Beach-Sarasota County, Florida; Dane County, Wisconsin; and the Windham area of Vermont.  Pursuing comparisons for even this short list might be informative.
Question: Can investment in central city revitalization be effective in retaining residents and attracting new residents?

            Major U.S. cities have lost population, as a percentage of metro area population, for at least the last two decades.  The causes of this trend have been summarized above.  Moreover, for at least the last two decades, because of the continuing decline of central cities, urban redevelopment policies have been judged a failure by many observers.

Bottom Line: This topic will be the focus of the Department of Urban Planning’s seminar series in 2000-2001.  We are unable to offer any firm conclusions at this time.  One thing is clear, however.  The demand for central city land surely depends on the supply of land at the city’s edge.  Urban redevelopment has a better chance of success if land policy at the edge discourages location there.  A limiting factor may be that constraining too large a proportion of new development to the central city may weaken demand as some residents and employers relocate to other metro areas posing fewer obstacles to development.



            Central cities are handicapped as they compete for residents because of:

·        Non-competitive tax rates and service delivery in central cities compared to suburbs, including schools and public safety;

·        Aging housing stock, with many housing units that do not offer the size and amenities new homebuyers seek;

·        Limited supply of large, clean sites for locating job-creating industry resulting in poor access to jobs

·        Poor access to retail facilities and services sought by middle-class homebuyers—a classic chicken/egg problem: if the middle-class could be enticed to live in the central city, retailing would follow, but enticing the middle class without retailing is doubly difficult.

Certainly the results of urban revitalization programs through the 1980s and 1990s has been mixed.  Some observers attribute the poor performance of urban programs during the previous two decades to inadequate investments and underfunding of programs, compounded in some cases by fraud and abuse by government agencies.  Other observers view investment in central city revitalization as a poor risk as long as private investment continues to leak out to the suburban areas.

Strategies for overcoming these competitive disadvantages include:

·        Locating new housing adjacent to public amenities, such as water, views, or rapid transit, that will compete effectively with suburban amenities by offering an alternative rather than reproducing suburban amenities.

·        Promoting infill projects with middle-income and upscale housing near central city activity centers.

·        Creating middle-income single-family housing comparable in interior amenities to suburban housing (in terms of bathrooms, closets, kitchens, quality of construction, etc.) but more affordably priced and conveniently located near downtown or other job centers in the central city.

·        Assuring that retail and recreational facilities appealing to middle-class tastes are available near new housing by revitalizing downtown and neighborhood commercial areas.

·        Publicizing successful revitalization efforts that are attracting suburban residents back to the city.

·        Continuing to improve school quality and safety.


Uncertainties and Alternative Viewpoints

            One of the questions that has not been adequately addressed is the level of subsidy that can be justified to produce middle- and upper-income housing and commercial areas in the city.  After years of neglecting low-income neighborhoods, offering substantial subsidies (for example, through TIF financing) to middle-class developments needs to be scrutinized on equity grounds. 


Future Research

            Nelson and Dawkins (2000) suggest the need to examine the impact of containment policies on central city revitalization.  Future research needs to explore the interaction between local/regional containment policies and the effectiveness of urban redevelopment programs at the federal and state level. 

            The potential success of central city revitalization may depend on how easy or difficult it is to continue moving outward, which will depend on a combination of regulatory and market factors. 
Question: Might changes in the way metropolitan areas structure taxes and fees act to reduce sprawl?

            As noted above, central cities are at a disadvantage compared to newer, suburban communities in competing for residents.  Higher taxes and inadequate public services have been cited as important barriers to retaining residents in cities.  This section considers three kinds of taxes or fees that have been mentioned as potential tools for combating sprawl: tax base sharing, split-rate property taxes, and impact fees.

Bottom Line: Tax base sharing has been a limited success in the Minneapolis metropolitan area and is unlikely to be replicated in other metro areas.  Split-rate taxation, in which land is taxed more heavily than improvements, shows promise but more research is needed.  Impact fees do not appear to have had any effect on the spatial form (compactness, contiguousness, etc.) of communities and regions in which they have been imposed, but the potential of impact fees to do so has probably not been tapped.


Summary Answer:

            Within the United States, states and local governments use a variety of mechanisms to raise revenues.  At the local level, property taxes and (in some places) sales taxes are among the most prominent revenue sources.  In some fast growing suburban communities, impact fees have been used to assure that new development pays the “full costs” of providing public services to the new residents. 

A.  Tax Base Sharing          

Because property taxes and local sales taxes are collected at the local level and—except for the programs discussed here—retained within the local unit, communities with higher valuations on homes or that are home to regional facilities that generate property or sales taxes usually have an advantage over other communities in the metropolitan area.

One means of leveling the playing field is to redistribute the tax base.  Tax base sharing is unlike state revenue sharing, in which state-collected revenues are returned to local governments according to some formula, with central cities usually receiving a higher per capita share.  Tax base sharing does not transfer revenues.  Rather, it merely creates a tax base pool upon which all communities levy their usual property tax rate.

The only example of tax base sharing in the U.S. within an urban metropolitan area is in Minneapolis-St. Paul.  Known as the Fiscal Disparities Program, its objectives are to improve equity in the distribution of fiscal resources and to promote better regional planning objectives.  The program began operation in 1974 and continues today despite legal challenges.  As a strategy for improving equity, tax base sharing appears to have succeeded.  Minneapolis and, especially, St. Paul have had a larger tax base to draw upon than if tax base sharing had not been implemented.

As a strategy for influencing land use, tax base sharing does not appear to have had significant impacts, however.  The Minneapolis-St. Paul metro area has experienced substantial suburban growth, similar to other areas that have not used tax base sharing.  The tax redistribution appears to be too small to be very influential in housing or business location decisions and has been dwarfed by other forms of revenue sharing from the state government.

In addition, tax base sharing is unlikely to be adopted elsewhere.  Even in Minnesota, too many communities would object if tax base sharing were to be enacted today, rather than 25 years ago.  The original tax base sharing legislation was engineered by rural legislators before the advent of open meeting laws and was forced upon the Twin Cities.  Tax base sharing simply fails the political feasibility test.

B.  Split-rate property taxes

The split-rate property tax (also known as land value tax policy) has been proposed as a way of increasing the incentive of landowners to develop land more intensively.  Under normal property tax policies, because land and improvements are taxed at the same rate, but land is valued much lower in property assessments, landowners may acquire more land than is needed for their land improvements.  As a result, more land is consumed per home or shop than is really needed.  Split-tax rates place a higher tax rate on land in order to minimize the amount of land consumed and maximize the improvements placed on the land.  Some research has found that split-tax rates in 15 Pennsylvania cities produced more building permits in urban areas and more intense land use as intended (Hartzok 1997).

C.  Impact fees

Impact fees are levied on new development to pay for the estimated impact that new residents will have on the provision of services in the community.  New students require more school buildings; increased traffic due to increased population requires street widening projects, left turn lanes, and other improvements; more demands on public facilities means that facility expansions will be needed sooner than if the development had not taken place.  Impact fees are intended to cover the incremental increases associated with each new home and required to provide schools, roads, water, sewer, parks, libraries, city administrative buildings, police services, and fire protection.  In addition, some economists consider the value of open space to be poorly reflected in the market and recommend the imposition of development fees to account for the benefits of open space in real estate development transactions (Brueckner 2000).

Impact fees might limit sprawling growth by increasing the costs of development in far-flung areas.  By making housing more expensive, development should be shifted to those (more centrally located) communities that have lower fees because they already have the infrastructure capacity to serve new development.

Impact fees have not been adopted everywhere.  In California, where impact fees are quite common, the highest impact fees, such as those in California, average about $25,000 per lot.  In Wisconsin, where state law limits the costs that can be included, fees tend to range between $2000 to $4000 per lot.  One might expect that impact fees of $25,000 might influence development patterns.  However, if fees are almost universally imposed in both developed and developing communities, as in California, fees will have little effect on where people locate.


Uncertainties and Alternative Viewpoints

Economists generally prefer taxes and fees as a means of correcting market deficiencies, if any are identified in relation to a perceived social problem.  Surprisingly little work explores the potential for land-based taxes as a means of intensifying the use of land. 

State constitutional provisions may need to be changed to permit split-tax rates and impact fees that are geographically targeted rather than based on service cost increases—especially if calculated as average costs. 

Further Research

Little research has been conducted on the impacts of taxes and fees in stemming sprawl. 



Question: How promising is urban modeling as a tool for analyzing the land use impacts of local policymaking?

            Improvements in spatial and economic modeling offer promising new policy-making tools.  If modeling tools could accurately predict the likely impacts of policy decisions, local officials would have a stronger basis for making decisions that create short-term costs and conflicts.

Bottom Line: Models could provide an important source of new information that would influence policy-making. Currently, too little research has been conducted to validate modeling tools.  Nonetheless, the heuristic value of these tools in generating discussions about the connections between land use, transportation, economic development, and environmental impacts needs to be recognized.  At the same time, poor models can lead to increased conflict and, ultimately, poor decisions.



            Models have been developed that are intended to project land use and transportation interactions in order to examine how land use changes may change transportation demand (and how transportation changes may influence land development).  While these models attain satisfactory levels of internal consistency, little research has been done to validate the models in terms of their ability to predict what happens in the real world.  Only two of the more than a dozen models examined by Wegener (1994) had been subjected to a validation process.

            Models like those proposed by Anas, et al., propose to integrate information about urban economies, transportation, and land use policy.  The objectives of these models are exciting. 


Uncertainties and Alternative Viewpoints

            Given the long time scales involved in urban modeling, achieving external validity may be an impossible goal.             


Further Research

            As more and more models and visioning tools become available, planners and local officials are going to need information to make them better-informed consumers of the information produced by these models.  Planners need to provide appropriate data as inputs to the models.  Because of the complexity of these models, the ultimate end-users of model results will need understandable explanations about the interpretation of a model’s results and its limitations.

            Model results will be ignored if users do not trust the results.

Overall Conclusions

            The purpose of this overview of the “known facts” about sprawl was to enable a more informed discussion among policy makers about the causes, consequences and potential solutions to a widely perceived problem with current patterns of metropolitan development.  In many areas, evidence remains too inconsistent to draw any firm conclusions, and this is especially true in relation to potential solutions to the problem.

This paper began by identifying 4 types of questions that policy makers ask as they search for the right policies for their states and communities.

What are the causes of sprawl?  Although a clear definition of sprawl eludes researchers, most observers agree that it is a ubiquitous phenomenon, occurring throughout the United States and in recent decades in Europe and Asia as well. The causes of sprawl appear to be found in the interaction between general economic laws of land markets, increasing affluence, decreasing transportation costs, and demographic changes, especially decreasing household size. 

Is sprawl a “problem” or not?  A review of the research concerning the costs and benefits of sprawl appear to warrant the conclusion that sprawl is a problem, that it generates more costs to society—and distributes those costs unfairly—than are warranted by the benefits it creates.

If sprawl is a problem, how serious a problem is it, relative to other problems facing communities and the nation?  Much disagreement persists about the magnitude of the net loss to social welfare created by sprawl.  Given the enormous racial and economic disparities in the U.S., the uneven distribution of costs and benefits related to sprawl may be a more serious cost than the additional road miles or incremental increases in air pollution that it creates.

Our review of the literature, while certainly less than comprehensive, uncovered no examples of research comparing the net costs of sprawl (seriousness) to other problems competing for public policy attention and action.  We cannot say whether the risks from unabated sprawl are greater or less than other risks facing cities.  In part, the answer to this question cannot be separated from some rough estimate of the cost and effort that would be required to stem sprawl, as well as the positive and negative side effects that controlling sprawl might bring. 

For example, reversing the trend toward sprawl might improve central city educational systems by increasing the political constituency concerned about the quality city schools.  (Nelson makes a similar observation relative to environmental justice.)  On the other hand, increased housing costs might generate a host of negative impacts for the poor, including increased homelessness and related health problems.

What are the comparative advantages and disadvantages of various proposed solutions to solve the sprawl problem?  Some of the strategies considered in this paper appear unlikely to be up to the task of preventing sprawl without creating serious negative side effects.  Metropolitan government, annexation, tax base sharing (as implemented in Minnesota), and (at least in the near term) higher fuel taxes are probably not politically feasible in most metropolitan areas, at least not in a form that could be effective.

Other strategies appear more promising, among them growth management (Smart Growth), leveraging private investments in the central city through public/private partnerships, and (in the right context) urban design strategies.

Our review of anti-sprawl strategies has made more visible a distinction between three kinds of approaches.  One approach, which I will call the economists’ approach, proposes to attack sprawl by attacking the root cause—market failure—through changes in taxes and fees especially on automobile use and land development.  Brueckner (2000) is representative of this point of view.  “Because of the secondary role of market failure, a draconian [regulatory] attack on urban sprawl is probably not warranted.  By greatly restricting urban expansion, such an attack might needlessly limit the consumption of housing space, depressing the standard of living of American consumers.”  The tools proposed by economists—increased fuel taxes, congestion tolls, and development fees—are appealing in their simplicity, but their political feasibility has yet to be demonstrated. 

Another approach is land use regulation.  To many, this is a more direct approach than the economists’ reliance on taxes to modify consumer behavior.  Through land use regulation, communities can identify areas “appropriate” for development and areas that are not, rather than merely shaping consumer preferences.  Land use regulation, especially the use of Urban Growth Boundaries, is more appealing because it appears to offer more direct control.  Unfortunately, these measures are also difficult to implement politically, especially in areas with well-organized property rights interests or weak long-term economic growth prospects.

Finally, a third approach focuses on the negative consequences of sprawl, without necessarily addressing the underlying causes of sprawl or regulating land use per se.  Peiser (1989), in his oft-cited study of the relationship between leapfrog development and density, asserts the following: “How to reduce the negative aspects of sprawl—rather than sprawl itself—should be the thrust of public policy with respect to fringe growth.” For example, purchase of development rights for open space, tighter fuel economy standards for cars, and increased investment in workforce and business development in the central city might solve many of the most critical problems associated with sprawl.  Some of these strategies, such as purchase of development rights, can be pursued by local governments, though state and federal support would be helpful.  Others, such as fuel economy standards, require federal policy action to implement.

Given the evidence that the costs of sprawl probably do exceed its benefits, and given the evidence of sprawl’s role in accentuating existing inequities in American society, public policy ought to pursue at least one of these sets of strategies.  It is premature to recommend one strategy over another.  Communities will need, at least in the near future, to find their own way. 


Figure 1.  Topics and Speakers

All speakers are from the University of Wisconsin—Milwaukee, unless otherwise noted.

Definitions of Sprawl

Sammis White

Causes of Sprawl

Richard Green, University of Wisconsin—Madison

Economic Productivity

Sammis White


Kirk Harris

Infrastructure Costs

Zhong-Ren Peng

Environmental and Health Impacts

Nancy Frank

Growth Management

Stuart Meck, American Planning Association

Urban Design Solutions

Kerry Vandell, University of Wisconsin—Madison

Edward Beimborn


Metropolitan Government


Central City Revitalization

Welford Sanders

Farmland Preservation

Brad Steinke, Director of Community Development, Mequon, Wisconsin


Tax Base Sharing

Sammis White

Annexation Reforms

George Hall, Wisconsin Department of Administration


Urban Modeling for Planning

Alan Horowitz


Additional sources: 


Horridge 1994, see Web of Science printout, p. 39




American Planning Association. 2000.  Growing Smart index. http://www.planning.org/plnginfo/GROWSMAR/gsindex.html.

Anas, Alex, David Boyce, Lewis Hopkins, and Gerrit Knaap.  1999.  Project Summary: Infrastructure and Metropolitan Development.  National Science Foundation, Urban Research Incentive Grant.

Brown, D.L., G.V. Fuguitt, T.B. Heaton, and S. Waseem.  1997.  Continuities in size of place preferences in the United States, 1972-1992.  Rural Sociology 62(4): 408-428.

Brueckner, J.K.  2000. Urban sprawl: Diagnosis and remedies.  International Regional Science Review 23(2): 160-171.

Burchell, Robert, et al.,  1998.  The Costs of Sprawl Revisited.  Washington, D.C.: National Academy Press.

Burchell, Robert, and Naveed A. Shad.  N.d.  The benefits and costs of growth.  Milwaukee, WI: Milwaukee Metropolitan Sewerage District Synposium.

Chicago Fed Letter.  1999.  Should governments try to control suburban growth?  Chicago: Federal Reserve of Chicago.

Churchman, Azra.  1999.  Disentangling the concept of density.  Journal of Planning Literature 13(4): 389-411.

Crane, Randall.  2000.  The influence of urban form on travel: An interpretive review.  Journal of Planning Literature 15(1): 3-23.

Downs, Anthony. 1999.  Stuck in traffic: Coping with Peak-hour Traffic Congestion.  Washington, D.C.: Brookings Institution; Cambridge, MA: Lincoln Institute of Land Policy.

Eppli, and Tu.  1999.  Valuing the New Urbanism: The Impact of the New Urbanism on Prices of Single-Family Homes.  Washington, D.C.: Urban Land Institute.

Ewing, Reid. 1997.  Is Los Angeles-style sprawl desirable?  Journal of the American Planning Association 63(1): 107-126.

Friedman, Bruce, Stephen P. Gordon, and John B. Peers. 1992.  The effect of neotraditional design on travel characteristics.  In Compendium of Technical Papers, Institute of Transportation Engineers, 1992 District 6 Annual Meeting.  Anchorage: Institution of Transportation Engineers, Alaska Section.

Gordon, Peter, and Harry W. Richardson. 1997.  Are compact cities a desirable planning goal?  Journal of the American Planning Association 63(1): 95-106.

Green, Richard K.  2000.  Causes of sprawl.  Illinois Real Estate Newsletter.  Winter issue.

Hall, Peter.  2000.  Urban Renaissance/New Urbanism: Two sides of the same coin.  Journal of the American Planning Association 66(4): 359-360.

Hartzok, A.  1997.  Pennsylvania’s success with local property tax reform: The split rate tax.  American Journal of Economics and Sociology 56(2): 205-213.

Hawkins, Brett W., and Douglas M. Ihrke.  1999.  Reexamining the suburban exploitation thesis in American metropolitan areas.  Publius 29(3): 109-121.

Hayward, Steven.  1998.  Legends of the sprawl.  Policy Review September-October (91): 1-18.

Hogarty, Thomas F. 1998.  The social benefits of personal vehicle travel.  Transportation Quarterly 52(2): 5-9.

Hogarty, Thomas F. 1999a. Taxing away the benefits of personal vehicle travel.  Transportation Quarterly 53(3): 5-8.

Hogarty, Thomas F. 1999b.  The untold benefits of roads and travel.  Consumers’ Research 82(7): 10-14.

Khattak, Asad J., Virginia Amerlynck, and Roberto G. Quercia. 2000.  Are travel times and distances to work greater for residents of poor urban neighborhoods?”  Transportation Research Board, 79th Annual Meeting, Washington, D.C.  http://www.unc.eedu/~khattak/wktt.htm.

Knaap, Garret, and Arthur C. Nelson. 1992.  The Regulated Landscape: Lessons on State Land Use Planning from Oregon.  Cambridge: Lincoln Institute of Land Policy.

Landis, John D.  1995.  Imagining land use futures.  Journal of the American Planning Association 61(4): 438-457.

Levinson, D., D. Gillen, and A. Kanafani.  1996.  The Social Costs of Intercity Transportation: A Review and Comparison of Air and Highway.  Berkeley, CA: University of California at Berkeley, Institute of Transportation Studies.

Levinson, D.M., and A. Kumar.  1994.  The rational locator—Why travel-times have remained stable.  Journal of the American Planning Association 60(3): 319-332.

Murphy, James J., and Mark A. DeLucchi.  1998.  A review of the literature on the social cost of motor vehicle use in the United States.  Journal of Transportation and Statistics.  January: 1-42.

Murphy, James J., and Mark A. DeLucchi. 1998.  A review of the literature on the social cost of motor vehicle use in the United States.  Institute of Transportation Studies.  University of California, Davis.

Nelson, Arthur C.  1999.  Comparing states with and without growth management—Analysis baseed on indicators with policy implications.  Land Use Policy 16(2): 121-127.

Nelson, Arthur C. 2000.  Smart growth or business-as-usual? Which approach improves quality of life and central city vitality?  In Susan Wachter, J. Leo Penne, and Arthur C. Nelson, eds., Bridging the Divide: Making Regions Work for Everyone—Shaping the Federal Agenda.  Washington, D.C.: U.S. Department of Housing and Urban Development.

Nelson, Arthur C., and Casey J. Dawkins.  2000.  The containment of urban America or how to really revitalize central cities. 

Nelson, Arthur C., and D.R. Peterman.  2000.  Does growth management matter? The effect of growth management on economic performance.  Journal of Planning Education and Research 19(3): 277-285.

O’Connor, Alice, Chris Tilly, and Lawrence Bobo.  2000.  Urban Inequality: Evidence from Four Cities.  Russell Sage Foundation.

Orfield, Myron.  1996.  Metropolitics: A Regional Agenda for Community and Stability.  Washington, D.C.: Brookings Institution Press; Cambridge, MA: Lincoln Institute of Land Policy.

Pendall, Rolf.  2000.  Local land use regulation and the chain of exclusion.  Journal of the American Planning Association 66(2): 125-142.

Persky, Joseph, and Wim Wiewel. 2000.  When Corporations Leave Town.  Detroit: Wayne State University Press.

Pisarski, Alan E.  1996.  Commuting in America II.  Lansdowne, VA: Eno Transportation Foundation.

Richardson, Harry W., and Peter Gordon.  1999.  Is sprawl inevitable? Lessons from abroad.  Paper presented at the Association of Collegiate Schools of Planning conference, Chicago.

Rusk, David.  1993.  Cities Without Suburbs.  Washington, D.C.: Woodrow Wilson Center Press; Baltimore, Md.: Johns Hopkins University Press.

Schimek, Paul.  1996.  Household motor vehicle ownership and use: How much does residential density matter?  Unpublished manuscript.

Sierra Club.  2000.  The costs of sprawl. http://www.sierraclub.org/sprawl/community/costs.asp

Staley, Sameul R., Jefferson G. Edgens, and Gerard C.S. Mildner.  N.d.  A line in the land: Urban-growth boundaries, smart growth, and housing affordability.  Reason Public Policy Institute, Policy Study No. 263.

Straszheim, Mahlon R. 1979.  "Assessing the Social Costs of Urban Transportation Technologies," in Peter Mieszkowski and Straszheim, Current Issues in Urban Economics.

Southworth, Michael. 1997.  Walkable suburbs? An evaluation of neotraditional communities at the urban edge. Journal of the American Planning Association 63(1): 28-44.

U.S. Department of Transportation, Bureau of Transportation Statistics.  1999.  National Transportation Statistics 99. 

U.S. Department of Transportation, Bureau of Transportation Statistics.  1999.  Pocket Guide to Transportation.  BTS99-06R, http://www.bts.gov/pg.pdf. 

Vandell, Kerry.  2000.  New Urbanism.  Presentation at the University of Wisconsin—Milwaukee, School of Architecture and Urban Planning, Exploring Sprawl seminars, March 9.

Wegener, Michael.  1994.  Operational urban models: State of the art.  Journal of the American Planning Association 60(1): 17-29.

Weitz, J. and Moore, T. 1998.  Development inside urban growth boundaries—Oregon’s empirical evidence of contiguous urban form.  Journal of the American Planning Association 64(4): 424-440.

Wheeler, Steven. 1998.  [cite information missing]

White, Sammis, and Brian Dent.  2000.  Benchmarking study.  Unpublished response from the author.