This aerial view of Hanoi, Vietnam, clearly shows areas of decreasing density between the city and the countryside, making it hard to define the limits of the "urban" area.
Follow the author on Twitter:
@cd_planner
Anyone reading this blog is likely to have heard the statistic that ‘over half of the world’s population now lives in urban areas’. This has been the standard opening line of reports and presentations about urbanization since this milestone was supposedly reached in 2008. But what does it really mean?
In everyday usage, terms related to human settlements have vague, shifting meanings. What one person might describe as a small ‘city’ might be a ‘town’ or ‘village’ for someone else; one person’s ‘megacity’ might be a cluster of cities from a different perspective. Similarly, we can usually identify areas that are clearly within a city and others that are outside it, but there is usually a peri-urban area of intermediate density that usually lies between the two, making it hard to define a clear city limit. Formal administrative boundaries may have historic or political meaning, but are rarely aligned with the physical or economic extents of the urban area.
What exactly is a city? It depends who you ask
It turns out there is no standard international definition of an ‘urban’ area or ‘urban’ population. Each country has its own definition, and collects data accordingly. The statistic that 50% of the world’s population is urban is arrived at simply by adding up these incomparable, and sometimes conflicting, definitions.
A useful list compiled by the UN Population Division gathers the definitions of ‘urban’ population used in censuses in 232 countries. The criteria used by countries to decide whether to define a place as ‘urban’ include population size, population density, type of economic activity, physical characteristics, level of infrastructure, or a combination of these or other criteria. Some simply list their urban areas by name.
Each definition may be well-suited to its own national context, but the differences present a problem when trying to compare urbanization across countries. As Uchida and Nelson observe while attempting to address this issue, “[t]he matter is analogous to measurements of global poverty and comparisons of poverty levels across countries… A uniform definition—like the $1 or $2 per day index used in evaluating poverty levels across countries—makes analysis possible.”
So how can we come up with an urban equivalent of the dollar-a-day benchmark? Let’s look at how some countries use specific thresholds for defining urban areas, and how these can be adapted globally.
Using population and density thresholds to define ‘urban’ areas
101 countries use minimum population thresholds as a means of identifying settlements as ‘urban’, either as the sole criteria or together with others. The most frequently used threshold values are 2,000 inhabitants (used by 23 countries), and 5,000 inhabitants (used by 21 countries), as shown in the figure below. The average of all these thresholds was just under 5,000 inhabitants. (It may strike you how low these figures are. The most frequently used minimum population, 2,000 people, could easily be accommodated within a single large office building.)
Only 9 countries use minimum population density thresholds, of which only one, Germany, uses it as a sole criterion for defining urban areas. The lowest density threshold used is 150 persons per sq. km. (Germany), and the highest is 1,500 (China and Seychelles).
The density thresholds also do not take into account the variation in the size of the areas over which the density is being averaged, which would result from varying sizes of administrative units. Even within the same country, two identical settlements may be treated differently, depending on whether they fall within large or small administrative units.
Past attempts to create a global ‘urban’ definition
Others have already recognized this issue, and attempted to craft a standard ‘urban’ definition:
Using satellites and modeling techniques to map cities
Reliable commuting data, necessary to implement either approach described above, do not exist at the global scale. However, global maps of built-up areas and populations of increasing resolution and sophistication are now available. The European Commission’s Joint Research Centre is developing a Global Human Settlements Layer (GHSL), while the German space agency DLR is developing a layer called Global Urban Footprint. Meanwhile, the WorldPop project uses these and many other layers to create high-resolution population distribution maps (see the recent blog on WorldPop by my colleague Tatiana Peralta Quiros).
In the World Bank’s global Urban and Social Development unit, we are starting to experiment with ways in which we can use these layers to create standard, comparable definitions of urban areas, building on the OECD’s ‘core’ and ‘hinterland’ approach mentioned above.
Our approach will use pixels of standard size as the units of analysis, rather than administrative boundaries, which vary in size. The thresholds that we will have to define include minimum core size, core density, hinterland density, total population, and the maximum distance between pixels for them to be considered part of the same whole. These thresholds will need to be modifiable to suit different needs. They may be used to define the ‘degree of urbanization’ of a place, rather than implying a strict urban/ non-urban distinction.
Needless to say, a city is more than its physical form. It is a historical and cultural artifact, a social and political network, and an economy. But arriving at some common understanding of urban areas, using physical and demographic criteria, can at least ensure that when we compare urban areas or trends in urbanization between countries, we are comparing like with like. With the potential introduction of an ‘urban’ Sustainable Development Goal, along with other indicators and standards for cities, it is becoming more important than ever to develop a common vocabulary on urbanization.
Anyone reading this blog is likely to have heard the statistic that ‘over half of the world’s population now lives in urban areas’. This has been the standard opening line of reports and presentations about urbanization since this milestone was supposedly reached in 2008. But what does it really mean?
In everyday usage, terms related to human settlements have vague, shifting meanings. What one person might describe as a small ‘city’ might be a ‘town’ or ‘village’ for someone else; one person’s ‘megacity’ might be a cluster of cities from a different perspective. Similarly, we can usually identify areas that are clearly within a city and others that are outside it, but there is usually a peri-urban area of intermediate density that usually lies between the two, making it hard to define a clear city limit. Formal administrative boundaries may have historic or political meaning, but are rarely aligned with the physical or economic extents of the urban area.
What exactly is a city? It depends who you ask
It turns out there is no standard international definition of an ‘urban’ area or ‘urban’ population. Each country has its own definition, and collects data accordingly. The statistic that 50% of the world’s population is urban is arrived at simply by adding up these incomparable, and sometimes conflicting, definitions.
A useful list compiled by the UN Population Division gathers the definitions of ‘urban’ population used in censuses in 232 countries. The criteria used by countries to decide whether to define a place as ‘urban’ include population size, population density, type of economic activity, physical characteristics, level of infrastructure, or a combination of these or other criteria. Some simply list their urban areas by name.
Each definition may be well-suited to its own national context, but the differences present a problem when trying to compare urbanization across countries. As Uchida and Nelson observe while attempting to address this issue, “[t]he matter is analogous to measurements of global poverty and comparisons of poverty levels across countries… A uniform definition—like the $1 or $2 per day index used in evaluating poverty levels across countries—makes analysis possible.”
So how can we come up with an urban equivalent of the dollar-a-day benchmark? Let’s look at how some countries use specific thresholds for defining urban areas, and how these can be adapted globally.
Using population and density thresholds to define ‘urban’ areas
101 countries use minimum population thresholds as a means of identifying settlements as ‘urban’, either as the sole criteria or together with others. The most frequently used threshold values are 2,000 inhabitants (used by 23 countries), and 5,000 inhabitants (used by 21 countries), as shown in the figure below. The average of all these thresholds was just under 5,000 inhabitants. (It may strike you how low these figures are. The most frequently used minimum population, 2,000 people, could easily be accommodated within a single large office building.)
Only 9 countries use minimum population density thresholds, of which only one, Germany, uses it as a sole criterion for defining urban areas. The lowest density threshold used is 150 persons per sq. km. (Germany), and the highest is 1,500 (China and Seychelles).
The density thresholds also do not take into account the variation in the size of the areas over which the density is being averaged, which would result from varying sizes of administrative units. Even within the same country, two identical settlements may be treated differently, depending on whether they fall within large or small administrative units.
Past attempts to create a global ‘urban’ definition
Others have already recognized this issue, and attempted to craft a standard ‘urban’ definition:
- The WDR 2009 approach: One such attempt is the approach outlined by Chomitz et al and elaborated on by Uchida and Nelson, which identifies all settlements above a certain minimum population size and minimum population density that are within a certain travel time by road. This approach was used in the World Bank’s World Development Report 2009.
- The OECD approach: A similar but more elaborate approach is that of the Organisation for Economic Co-Operation and Development (OECD). The OECD methodology consists of three main steps: identifying contiguous or highly interconnected densely inhabited urban cores; grouping these into functional areas; and defining the commuting shed or ‘hinterland’ of the functional urban area. The OECD uses population size cutoffs (50,000 or 100,000 people, depending on the country) as well as population density cutoffs (1,000 or 1,500 people per sq. km.) to define the urban cores, and then selects those areas from which more than 15% of workers commute to the core as hinterlands
Using satellites and modeling techniques to map cities
Reliable commuting data, necessary to implement either approach described above, do not exist at the global scale. However, global maps of built-up areas and populations of increasing resolution and sophistication are now available. The European Commission’s Joint Research Centre is developing a Global Human Settlements Layer (GHSL), while the German space agency DLR is developing a layer called Global Urban Footprint. Meanwhile, the WorldPop project uses these and many other layers to create high-resolution population distribution maps (see the recent blog on WorldPop by my colleague Tatiana Peralta Quiros).
In the World Bank’s global Urban and Social Development unit, we are starting to experiment with ways in which we can use these layers to create standard, comparable definitions of urban areas, building on the OECD’s ‘core’ and ‘hinterland’ approach mentioned above.
Our approach will use pixels of standard size as the units of analysis, rather than administrative boundaries, which vary in size. The thresholds that we will have to define include minimum core size, core density, hinterland density, total population, and the maximum distance between pixels for them to be considered part of the same whole. These thresholds will need to be modifiable to suit different needs. They may be used to define the ‘degree of urbanization’ of a place, rather than implying a strict urban/ non-urban distinction.
Needless to say, a city is more than its physical form. It is a historical and cultural artifact, a social and political network, and an economy. But arriving at some common understanding of urban areas, using physical and demographic criteria, can at least ensure that when we compare urban areas or trends in urbanization between countries, we are comparing like with like. With the potential introduction of an ‘urban’ Sustainable Development Goal, along with other indicators and standards for cities, it is becoming more important than ever to develop a common vocabulary on urbanization.
Join the Conversation