This chapter discusses the policy challenges associated with urban sprawl. It distinguishes urban sprawl from other forms of urbanisation and explains why a new perspective on the phenomenon is needed. It then describes how the new perspective on urban sprawl developed in this report can help policy makers address the challenges associated with it. To this end, the report’s contributions on the definition and measurement of sprawl, the analysis of its causes and consequences, and the identification of policy actions to address the latter are highlighted. Last, the chapter navigates the reader through the different sections of the report.
Rethinking Urban Sprawl
Moving Towards Sustainable Cities
Report
Table of contents
Share
Facebook
Twitter
LinkedIn
Chapter 1. The policy challenge of urban sprawl
Abstract
1.1. Urban sprawl is different from urbanisation
The impact of cities extends well beyond their administrative and physical boundaries. Economic activities taking place within cities play a critical role for economic growth and employment, but they are also responsible for an important part of environmental and health problems facing the world today. Taking policy action at the urban level is not only pivotal for tackling major environmental issues, such as climate change and exposure to outdoor air pollution, but also for addressing social exclusion and pressures on public finance.
Cities are hubs for knowledge and innovation, and the loci where the cross-fertilization of ideas is crystallised into massive job creation. Only during the past fifteen years, the two hundred largest metropolitan areas of the OECD countries have generated more than 60% of jobs and economic growth in these countries (OECD, 2016). OECD (2006) identifies several mechanisms through which city size may translate into higher output per capita and productivity. First, it stresses the importance of agglomeration economies, i.e. advantages generated by the spatial clustering of firms. Second, it highlights the role of specialisation advantages and more efficient matching, which emerge due to larger and more diverse labour markets. The spatial concentration of firms and people also facilitates knowledge spill-overs, innovation and economies of scale.
The role of cities for economic growth, employment and the environment is expected to become even more important in the future. Current trends show that urbanisation is a global and momentous process that is unlikely to decelerate in the decades to come (United Nations, 2008; United Nations, Department of Economic and Social Affairs, Population Division, 2014). Recent forecasts estimate that 70% of the world population – a number that rises to 86% for OECD countries – will be living in urban areas by 2050. In the same time horizon, global urban land area is projected to increase from 603 000 km2 in 2000 to over 3 million km2 in 2050, hosting 6.5 billion people (Angel et al., 2011). It is, thus, of primary importance that urban development occurs in a way that stimulates growth and employment on the one hand, and has the minimum possible environmental impact on the other.
Many of the environmental, economic and social challenges faced by cities can be attributed to some extent to certain characteristics of urban development patterns. For instance, low population density and dispersion of key points of economic activity tend to promote car dependency. In turn, this translates into higher levels of car use and more emissions of air pollutants and greenhouse gases. Providing sparsely populated areas with public services is also more costly, which makes them more prone to the generation of public budget deficits. In addition, low-density development can negatively impact housing affordability and facilitate income-based residential segregation, thereby discouraging social inclusion. Fragmentation (i.e. discontinuity) of urban development can also have similar consequences, as it increases the need for travelling and the costs of providing public infrastructure.
Fragmentation of urban fabric and different manifestations of low population density are features of a specific pattern of urban development: urban sprawl. The phenomenon has been fuelled by the growth of population and income, and a sharp reduction of real transport costs since the middle of the 20th century. The fall in transport costs stemmed from the technological progress in car manufacturing, the massive investment in road infrastructure, and the persistent underpricing of car use in many countries and urban areas (e.g. absence of road pricing, implicit parking subsidies, low end-prices of motor fuels). Urban sprawl has also been encouraged by a shift of preferences towards living in larger dwellings, situated far from environmentally degraded and often expensive city centres. For this reason, urban sprawl was initially considered to be the natural outcome of a desirable and efficient transformation process, rather than a misallocation of resources. This way, it went on for decades to become, in many parts of the world, the dominant form of urban development, possessing a momentum that is hard to reverse.
A common misperception is that it is not a particular form of urban development that is responsible for various environmental problems emerging in cities, but instead that these problems stem from urbanisation per se. This misperception is largely driven by the statistical correlation between city size and indicators of environmental pressures. However, such statistical relationships are often found by models failing to control for differences in urban form across cities or over time, so they should be interpreted with caution. In fact, such models often cannot disentangle the effects of city size from the effects of urban form. Attributing the costs of particular urban forms to urbanisation can be problematic when it comes to decision making, since the net benefits of urbanisation would be underestimated. Urbanisation is inextricably connected with modern economic growth and in order for its benefits to be calculated correctly, the exact urban development pattern, i.e. the distribution of population and structures across urban space, must be known in detail.
1.2. Why is a new perspective on urban sprawl needed?
Urban sprawl is still widely regarded as an elusive concept, even though the term has already been in use for about 80 years (Brueckner, 2000; Nechyba and Walsh, 2004). The phenomenon has been defined in numerous ways, and different disciplines have a different understanding of how sprawl manifests itself. Apart from the frequent confusion between urban sprawl and urbanisation delineated above, a common problem occurring with many definitions of urban sprawl is that the underlying phenomenon is confounded with its causes and consequences. This is problematic as it allows for subjective claims for or against urban sprawl. More importantly, it hinders policy makers from identifying the exact cause of various problems occurring at the city level and the right course of action to tackle them. For example, when car dependency is considered as part of the definition of urban sprawl, the latter will be blamed for the environmental consequences of car dependency. In turn, policy makers may direct efforts only to policies promoting the reduction of car use, while completely neglecting the influence of urban form on car dependency.
The confusion over urban sprawl is not limited to the conceptual level: it is also manifested in the measurement of the phenomenon. Approaches to measure urban sprawl differ in a number of aspects, but two of them are of particular importance: i) whether urban sprawl is considered a uni- or multidimensional phenomenon, and ii) what urban form characteristics are used to describe the phenomenon. The most common unidimensional measure of urban sprawl is the average population density in an urban area. As sprawl is assumed to decrease with average population density, the inverse of it, i.e. land uptake per capita, is often used instead. The problem with unidimensional approaches to measuring urban sprawl is that they are too simplistic to describe such a complex phenomenon. Multidimensional approaches are much better suited to capture its different aspects and manifestations. However, the majority of these approaches fail to include important measures of the distribution of population and urban fabric over space. Therefore, they do not provide a complete description of urban sprawl and are of limited use when it comes to investigating the relationship of the phenomenon with environmental, economic and social outcomes and alerting governments about possible needs for policy action.
The lack of a clear and neutral definition of urban sprawl and the disagreement over its measurement has hampered the identification and analysis of its consequences. Despite urban sprawl being intuitively associated with a number of environmental, economic and social problems, many of these relationships are not adequately substantiated by economic theory or empirical evidence. Empirical analysis of the effects of urban sprawl on the environment is particularly scarce, mainly due to the absence of sufficiently long time series of good indicators of urban sprawl for a large number of cities. The absence of empirical evidence for many of the environmental, economic and social pressures attributed to urban sprawl hinders conducting a proper cost-benefit analysis of the phenomenon and thus identifying whether and when policy intervention is desirable.
Various cities have developed policies to tackle urban sprawl and its potential consequences. However, some of these policies may have important side-effects which are not always considered prior to policy implementation. For example, stringent urban containment policies may cause leapfrog development outside of the regulated area. The latter can severely undermine the effectiveness of these policies in inhibiting urban encroachment on e.g. farmland and forestland and in preventing environmental and economic losses. At the same time, certain land-use policies, such as rigid building height restrictions, may themselves encourage urban sprawl. Once again, the confusion over the definition of urban sprawl and the lack of good indicators to measure it has not allowed a proper evaluation of the effects of different policies on the phenomenon. In any case, it is important that policy makers gain a better understanding of the possible implications of different policy options for urban sprawl, the environment, economic growth and social cohesion.
1.3. How does the new perspective presented in this report help address the policy challenge of urban sprawl?
The new perspective on urban sprawl presented in this report aims to help policy makers:
Better understand the complex phenomenon of urban sprawl, by defining it in a neutral way without references to its causes or consequences;
Monitor urban sprawl, by providing a set of indicators that reflect the multidimensional nature of the phenomenon and accounts for the distribution of people and urban fabric over space;
Make comparisons with development patterns in other cities and over time, by measuring these indicators for more than 1 100 urban areas and in 3 different time points;
Obtain an objective assessment of the environmental and economic consequences of urban sprawl through an extensive review of existing theoretical and empirical work;
Develop possible courses of action to address the environmental and economic consequences of urban sprawl by providing a menu of policy options grounded in economic theory and empirical evidence.
The following paragraphs explain in more detail these contributions of the report.
The definition of urban sprawl presented in 1.1 reflects its multidimensional nature and is based exclusively on the characteristics of this development pattern. The causes and consequences of urban sprawl do not constitute part of its definition and, thus, potential tautologies that could hamper policy analysis are avoided. Urban sprawl can manifest itself through various distributional patterns of population and urban fabric over space. Understanding this basic feature of the phenomenon can shift policy makers’ attention away from average population density and towards other characteristics of urban development that determine its environmental, economic and social effects. In line with this, the report highlights several cases in which a considerable fraction of the urban fabric exhibits very low population density levels, while, at the same time, the average population density of the entire urban area is high.
Indicators of urban sprawl are required to enable monitoring the phenomenon, undertaking comparative analyses with other cities and over time, and evaluating its causes and consequences. As sprawl is a multidimensional phenomenon, different indicators are needed to measure its different dimensions. Such dimensions comprise different aspects of the distribution of population density, the fragmentation of urban fabric, and the distribution of urban population between high and low density areas. In addition to the widely used indicator of average urban population density, six indicators are developed in this report:
Variation of urban population density: the degree to which population density varies across a city;
Land-to-density allocation: the share of urban footprint in which population density lies below a predefined threshold;
Population-to-density allocation: the share of population living in areas where urban population density is below a predefined threshold;
Polycentricity: the number of high population density peaks in an urban area;
Fragmentation: the number of urban fabric fragments per km2 of built-up area; and
Decentralisation: the percentage of urban population residing outside areas of peak density.
It is important to note that the size of the artificial area per se is neither considered a dimension of urban sprawl nor among the determinants of the phenomenon. Therefore, it may well be that cities of larger urban footprint are less sprawled than smaller ones.
Some earlier studies of urban sprawl that acknowledged its multidimensional nature constructed composite indicators to measure it. Such indicators essentially combine different dimensions of a concept in a single measure. Despite their usefulness in measuring the evolution of the phenomenon over time in a certain geographical context, composite indicators have a number of drawbacks. The first disadvantage of composite indicators is that they are much more difficult to interpret than their components. Their second disadvantage is that they are not very informative for policy development and evaluation. The reason for this is that they do not enable pinpointing the exact dimension(s) of urban sprawl that are problematic and need to be targeted by policy action. Furthermore, weighting methods used to construct composite indicators often seem ad-hoc. Discouraged by these drawbacks, this report abstains from constructing a composite indicator of urban sprawl. On the contrary, it relies on the set of indicators mentioned above to provide a comprehensive assessment of different dimensions of urban sprawl and facilitate the identification of problematic development patterns in different urban settings.
The different dimensions of urban sprawl have been quantified for an unprecedented combination of geographical and time coverage. The seven urban sprawl indicators have been computed for more than 1 100 urban areas in 29 OECD countries and for three time points: 1990, 2000 and 2014. The analysis has been enabled by a unique combination of satellite, administrative and other GIS data sources offering a granular level of spatiotemporal information. The calculation of these indicators for such a geographical and time coverage allows monitoring the evolution of urban sprawl in a 25-year period and making comparisons of the phenomenon across cities and countries. It also enables conducting retrospective analyses of the causes and effects of different dimensions of urban sprawl in a cross-country context.
The indicators are used to assess the current state of urban sprawl in the examined cities and countries, as well as the trends of the phenomenon over time. The analysis reveals that there is great diversity in the way that different dimensions of urban sprawl have evolved across OECD cities. A close look at the distributional characteristics of urban population density – referring here only to population density in functional urban areas (the OECD’s economic definition of a city) – reveals interesting patterns. In some OECD countries, average urban population density has substantially decreased, driven primarily by rapid suburbanisation. In contrast, a process of densification is observed in the urban areas of other OECD countries, which is, however, far from homogeneous. In some, low-density areas have grown faster than high-density ones, implying that suburbanisation has co-evolved with densification. Focusing on the other dimensions of urban sprawl, cities in most OECD countries have become more fragmented, but also more centralised, i.e. a larger share of their population lives in urban centres (areas of peak density).
The report assesses the drivers and the environmental and economic consequences of urban sprawl drawing upon a voluminous empirical literature, which spans several scientific disciplines, including economics, biology, and environmental and agricultural sciences. Causes and consequences of different dimensions of urban sprawl which are substantiated by empirical evidence are highlighted. However, the report also mentions cases in which a link is likely to exist, despite not being adequately supported by empirical evidence. Furthermore, the review distinguishes the consequences of urban sprawl from those of other forms of urban expansion, such as infill development. Missing this semantic detail may redirect the interest of policy makers from interventions aiming to discontinue problematic development patterns to others aiming to curb urbanisation.
Urban sprawl is caused by various demographic, economic, geographic, social and technological factors. As demand for floor space is very sensitive to income changes, the increase of real incomes in the past decades has significantly contributed to urban sprawl. Demand for floor space particularly increased in suburban areas where land prices were lower than those closer to urban centres. Rising real incomes induced low-density development in suburban areas and fragmentation of the urban fabric. Another important driver of the phenomenon is individual preferences for more space, comfort and privacy, which have also encouraged low-density and fragmented development. Such preferences are not only reflected in residential location choices, but also in transport mode choices: car use is partially driven by similar considerations. Technological progress in car production has further facilitated urban sprawl, as it has allowed covering longer distances more safely, comfortably, reliably and at substantially lower costs. Urban development pressures combined with physical conditions that hamper contiguous development also contribute to fragmentation of urban development.
The factors listed above are important determinants of urban sprawl, but certain policies have also implicitly encouraged the phenomenon. Stringent maximum density (e.g. building height) restrictions have induced low-density urban development. When combined with urban containment policies, they have often also led to leapfrog development in remote areas. The long-lasting underpricing of car use – stemming from a lack of road pricing, low on-street parking prices and in some cases low motor fuel taxes – and massive investments in road infrastructure have also encouraged households to move away from urban centres and their job location and form low-density and fragmented urban structures.
Higher emissions from road transport and loss of open space and environmental amenities in suburban areas are relatively well-established environmental consequences of urban sprawl. Fragmented and low-density development induces car dependency, which in turn translates to higher emissions of greenhouse gases and air pollutants. Furthermore, low density implies higher losses of open space, as more land needs to be developed to house a given population. Urban sprawl has also been claimed to harm biodiversity more than other forms of urban development. However, different dimensions of urban sprawl may have different and, in some cases, opposite effects on biodiversity. Low-density development is generally harmful to biodiversity, as larger urban footprints translate in higher losses of periurban biodiversity. In contrast, higher fragmentation of urban development is harmful to biodiversity only to the extent that it leads to higher fragmentation of natural habitats. This depends on the initial development pattern and on where new development takes place. Urban sprawl has also been suggested to have other consequences for the environment and health, including degraded water quality and higher obesity rates. However, such claims are not equally substantiated by empirical evidence.
The economic and social consequences of urban sprawl include higher costs to provide public services to low-density areas, as well as significant losses of time and productivity due to longer commuting distances. Furthermore, the increased costs of providing public services often cause pressures on local public finance. Among the dimensions of urban sprawl, costs of providing public services are heavily influenced by the share of population residing in low-density areas (e.g. below 3 500 inhabitants per km2). The higher that share is, the more costly it becomes for governments to provide those services. Commuting distances and infrastructure requirements and maintenance costs also increase with fragmentation, but the impact of low density on the costs and profitability of public investments is generally higher. From a social perspective, low-density development has been associated with social inequality and segregation, as the regulatory mechanisms that are used to maintain it may severely affect housing affordability.
The multiple private benefits of urban sprawl should be weighed against the social (private and external) costs entailed by its consequences. Market forces fail to take these external costs into account and, thus, policy intervention is necessary in many urban settings to address environmental problems and restore economic efficiency. As explained above, the new perspective of sprawl presented in this report provides the necessary tools and information to policy makers to help them evaluate the state of the phenomenon and determine whether policy action needs to be taken. To help policy makers further with developing possible courses of action, the report also provides a menu of policy options that can steer urban development to more sustainable patterns. Beyond this framework, customised and more elaborate solutions can be developed with the use of a case-study approach. Such an approach allows utilising detailed land-use data, elaborate socioeconomic and institutional information, and state-of-the-art modelling techniques, and can provide useful policy insights for a specific urban area. A case-study approach goes beyond the scope of this report, but it could be pursued as a separate stream of future work.
Policy action needs to focus on promoting socially desirable levels of population density and fragmentation through policies that often bear fruit in a relatively longer time horizon. In line with that objective, the report examines a series of relevant policy options: relaxation of stringent maximum density restrictions, adjustment of urban containment policies to deter leapfrog development, reforms of property taxation to encourage infill development and incentives for developers to provide public infrastructure for new constructions.
At the same time, urban policies need to directly address the car dependency caused by sprawl and mitigate its environmental and economic consequences. With regard to this objective, the report discusses a menu of possible policy interventions: the introduction of road pricing mechanisms; the abolishment of minimum parking requirements; the pricing of on-street parking at its marginal social cost; the increase of motor fuel taxes where the latter do not fully account for the external costs of fuel consumption; and the shift of focus from investments in road network expansions to the development of infrastructure for public transport, walking and cycling.
New land-use and transport policies and related reforms will interact with the existing policy framework, so governments should carefully consider the possible outcomes of such interactions before implementing them. The relevant framework contains policies set by national or local authorities that affect urban development despite pursuing other objectives, as well as land-use and transport policies under the jurisdiction of local authorities in neighbouring urban areas that may give rise to policy competition.
1.4. Navigation through the report
In the remainder of this report, the new perspective on urban sprawl is developed in two parts. The first part (Chapter 2 Chapter 3) explains in detail how the definition of urban sprawl provided here differs from earlier definitions of the phenomenon and describes the indicators used to measure it. It also presents a comparative analysis of the current state and trends of urban sprawl across 29 countries and over a period of 40 years. The second part (Chapter 4 Chapter 5) provides a critical assessment of the causes and consequences of the phenomenon and a detailed discussion of their implications for policy-making.
Chapter 2 develops a conceptual framework in which the various characteristics of urban sprawl are systematically categorised. The chapter discusses how the definition of urban sprawl reflects its multidimensional nature and the critical differences of the phenomenon from other urban development patterns. The various ways through which sprawl manifests itself are explained in detail and depicted schematically.
Chapter 3 operationalises the definition of urban sprawl. The sprawl dimensions described in Chapter 2 are represented by indicators which can be computed by widely available granular data on land cover, and enable both an analysis of the current state of the phenomenon and its evolution over time. The chapter elaborates on the methodology employed to compute the indicators and on the global datasets used to that end. The indicators are computed for more than 1 100 cities in 29 OECD countries and for three time points spanning the period from 1990 to 2014. The results of this work are used in a cross-country analysis, in which country rankings are provided for the state and evolution of each sprawl indicator. This analysis provides useful insights into the current state and trends of different dimensions of urban sprawl. The chapter also provides a comprehensive summary of the most interesting findings for each country.
Chapter 4 discusses in detail the causes and consequences of urban sprawl. The analysed drivers of urban sprawl include individual preferences, expectations of property owners, technological progress, physical geography, and certain policy instruments. The chapter also discusses a wide array of impacts ranging from the more well-established consequences, such as car dependency, emissions and inefficient provision of public services, to more disputable propositions, including the impact of sprawl on water quality and biodiversity. It also covers a series of hypotheses that have been formed inconsistently, have been falsified or have received limited empirical support.
Chapter 5 focuses on policies to steer urban development to more sustainable pathways. The chapter discusses in detail concrete policy actions that can help control urban sprawl and mitigate its environmental consequences. It also offers some general directions for addressing concerns over urban sprawl and promoting more sustainable urban development patterns. Finally, it highlights the importance of considering interactions between policy instruments and following an integrated approach to designing and implementing urban policies.
References
Angel, S. et al. (2011), Making Room for a Planet of Cities, Policy Focus Report/Code PF027, Lincoln Institute of Land Policy, Cambridge, MA, Available online at: www.lincolninst.edu/sites/default/files/pubfiles/making-room-for-a-planet-of-cities-full_0.pdf.
Brueckner, J.K. (2000), “Urban sprawl: Diagnosis and remedies”, International Regional Science Review, Vol. 23/2, pp. 160-171.
Nechyba, T.J. and R.P. Walsh (2004), “Urban sprawl”, The Journal of Economic Perspectives, Vol. 18/4, pp. 177-200.
OECD (2016), Making Cities Work for All: Data and Actions for Inclusive Growth, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264263260-en.
OECD (2006), Competitive Cities in the Global Economy, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264027091-en.
United Nations (2008), State of the World’s Cities 2008/2009: Harmonious Cities, United Nations Human Settlements Programme, Nairobi, Kenya.
United Nations, Department of Economic and Social Affairs, Population Division (2014), World Urbanization Prospects: The 2014 Revision, Highlights, (ST/ESA/SER.A/352), New York, NY, https://esa.un.org/unpd/wup/Publications/Files/WUP2014-Highlights.pdf.