This chapter focuses on the patterns of population distribution and service availability across OECD countries. It first discusses differences in service provision across the rural-urban continuum and the relationship with demographic trends such as urbanisation. Next, it explains how grid-based population data are used to identify settlements (cities, towns and villages) and introduces several reachability indicators derived from measures of driving accessibility – access to a city and the presence of other larger settlements nearby – along with measures of public transport accessibility. In addition, the chapter discusses data and methods to quantify the prevalence or number of public and private sector services in towns and villages.
Getting to Services in Towns and Villages
Preparing Regions for Demographic Change
OECD Rural Studies
1. The geography of services and accessibility
Abstract
Introduction
Regions vary in the extent to which their inhabitants cluster in settlements of different sizes, from villages to cities. Many factors affect the concentration of the population within a country, including the distribution of economic activities within the country and the presence of public services or amenities and vice-versa. In addition, the number and variety of services within regions depend on the relative sizes and travel times between settlements, such as the time from a town to a city. They also depend on the cost of providing those services, relative, for example, to the national average, as well as policy priorities relating to equitable outcomes in quality and access to services.
In this context, a key issue for policy makers, particularly with respect to public services, is determining the appropriate levels of and access to services for settlements of different sizes and locations, especially those whose populations are shrinking and ageing. Consistent measurement methods for settlements – considering their population density, area and contiguity – are pivotal.
This study is one of the first to use such measures across OECD countries.1 It identifies settlements based on the degree of urbanisation (DEGURBA) definition, with a particular focus on smaller settlements like towns and villages (OECD et al., 2021[1]).
This introductory chapter presents many of the technical concepts used in subsequent analysis. The findings can help policy makers understand the interaction between services and geography to address territorial inequalities and promote well-being in all places. Informed by ongoing dialogue with national statistical agencies, several boxes highlight new applications of DEGURBA and measurement considerations related to the three topics considered in this report: the geography of population, services and transport accessibility.
Services and population change along the rural-urban continuum
A better understanding of the geography of service provision in networks of settlements is integral to targeted public interventions. The patterns and problems of provision vary strongly across places, especially by population size, density and growth (Jacobs‐Crisioni, Kompil and Dijkstra, 2023[2]; Cattaneo, Nelson and McMenomy, 2021[3]). In rural areas, certain services are often absent or lacking in variety, and distances to access services are typically longer, even after accounting for lower congestion than in urban areas, where services and public transport access are generally more plentiful.
However, despite these very clear spatial factors and differences, much of the literature to date on accessibility to services has emphasised non-spatial aspects of service provision such as facility-to-resident ratios, costs, usage rates and survey data on transport accessibility (Milstein, Castelli and Gutacker, 2023[4]; Llena-Nozal, Fernández and Kups, 2022[5]; OECD/WHO, 2018[6]; Eurofound, 2022[7]; OECD, 2021[8]; Ward and Ozdemir, 2012[9]). In part, this reflects challenges associated with acquiring and consistently analysing geolocation data. Advances in geographic information system (GIS) data and routing computation have more recently led to more sophisticated quantitative spatial analysis.2
Spatial access issues will likely increase in importance as the configuration of population and demographics continue to change. For instance, ongoing population declines in most countries will mainly be concentrated in smaller, more remote settlements (OECD, 2023[10]), with potentially significant implications on service delivery costs (OECD/EC-JRC, 2021[11]).
The geography of service provision
Residential choices are intertwined with services, as people who can choose where to live typically consider accessibility to jobs, services and amenities. Densely populated areas and cities generally benefit from economies of scale and agglomerations. Even when the number of service providers per capita in rural areas is comparable to that in more populated places, rural areas have fewer providers and thus less variety than urban areas. The costs of service provision per capita has been estimated to be higher in rural areas than in more densely populated areas across European countries (OECD/EC-JRC, 2021[11]).
Moreover, service providers in rural areas tend to be more general and less specialised. Thus, people in rural areas sometimes must travel long distances – to a larger town or city – to access more specialised services (OECD, 2021[8]). For example, cities usually have a high enough demand for specialised medical services both inside and outside of hospitals, as more people with a wider range of medical needs can sustain a greater variety of facilities, doctors and other professionals. Many cross-country studies of spatial accessibility focus on healthcare services and find that longer travel times lead to negative health outcomes (Kelly et al., 2016[12]; Pathman, Ricketts III and Konrad, 2006[13]).
Availability and ease of access to high-quality services in urban and rural areas can be challenging for different reasons. For example, although cities have more service locations and greater variety than towns or villages, the large numbers of people in urban areas can lead to longer wait times and reduced access. High land prices and lack of space also limit service capacity. In addition to capacity constraints, certain services also have substantial neighbourhood-level variation in availability and quality. While more people in cities are close to some services, the time costs of congestion can restrict physical access to services in cities. Public transport provision also differs considerably between urban and rural contexts. Finally, service provision is related to a variety of individual socio-economic characteristics, so spatial differences in the composition of people also matter (Bastiaanssen and Breedijk, 2022[14]). Lack of access to transport and services has been found to be particularly acute for individuals with low incomes (Baptista and Marlier, 2020[15]) and access to transport has a greater impact on women, who, for cultural and logistical reasons, tend to use public transit more than men (World Bank, 2020[16]).
Survey data can provide relevant measures of individual experiences in service access and quality across countries and geographies. Figure 1.1 shows the average shares of people reporting dissatisfaction with access to quality healthcare, education, public transport and roads across OECD countries, based on data from the 2022 wave of the Gallup global survey (Gallup, 2022[17]). Panel A exhibits a clear rural-urban continuum for public opinions about healthcare, with cities having the smallest share of dissatisfied people.
A rural-urban gradient is evident for both roads and public transport (Figure 1.1, Panels B and C). More rural residents report dissatisfaction with public transport than with roads, while the opposite is true of city residents. This gradient is not visible for the educational system (Figure 1.1, Panel D); however, if anything, there is a U-shape, with towns and semi-dense areas exhibiting the smallest share of dissatisfied people.
Healthcare and public transport systems may be more accessible in urban areas because economies of scale make it easier to sustain large infrastructures (e.g. specialised medical services or multimodal transport networks), leading to greater satisfaction in urban compared to rural areas. On the other hand, schools and roads are potentially more affected by congestion in densely populated places. In rural areas, students must travel much further to schools. Rural schools often benefit from strong community engagement (e.g. parents participating in extracurricular and fundraising activities) but rural schools are usually smaller and may thus have limited course offerings.
Figure 1.1. Dissatisfaction with services, by type and DEGURBA
Demographic patterns and trends affecting service provision
This report requires settlements (cities, towns and villages) to be defined consistently across countries. The DEGURBA definition has three categories: i) urban areas (cities); ii) towns or semi-dense areas; and iii) rural areas. The next level of DEGURBA, Level 2, differentiates smaller settlements, such as towns and villages, from their surrounding areas (Box 1.1). It thus provides a means to examine critical issues such as differences in population growth in towns and villages from small and large cities and the impact of the proximity of smaller settlements to any city (Chapter 4).
Box 1.1. Settlements and degree of urbanisation (DEGURBA)
The DEGURBA definition identifies settlements from clusters of adjacent 1 square kilometre (km2) grid cells with medium or high population density. Such clusters meet the criteria for settlements if their total population is also above a certain threshold (see below). The DEGURBA definition also allows the use of built-up areas in addition to population, to avoid the identification of multiple urban centres for a single city (see Box 1.2). However, with DEGURBA, settlements such as cities are defined by their population density, not including the surrounding commuting areas.
Table 1.1 shows the mapping of Level 1 definitions for local area units and Level 2 definitions for grid-based DEGURBA classifications. The Level 2 definition of DEGURBA distinguishes towns and villages, which are settlements, from suburbs and dispersed rural areas, which are not. The minimum population thresholds are shown in the right-most column: villages have at least 500 residents while cities start at 50 000 residents. This report uses the original DEGURBA definition, which defines towns as having at least 5 000 residents. The definition of semi-dense towns is currently being revised, as described in more detail in Annex 1.C.
Table 1.1. DEGURBA definitions
Source: UNSD (2020[18]), “A recommendation on the method to delineate cities, urban and rural areas”, https://unstats.un.org/unsd/statcom/51st-session/documents/BG-Item3j-Recommendation-E.pdf
As detailed in Box 1.3, many OECD countries are already using DEGURBA definitions for their own spatial analyses and for certain international comparisons.
More than half the total population of OECD countries live in settlements (cities, towns or villages, though mainly cities) (Figure 1.2). Korea is the most urban OECD economy, with more than 75% of its population living in cities. Australia and Canada are also relatively urban. The population split between settlement types is more balanced in countries such as Czechia and Hungary, with a higher proportion of the population living in towns and villages compared to cities. Sixty-five percent of people living in settlements in OECD countries live in cities, 25% in towns and 10% in villages.
In the past decades, the population in OECD countries has steadily gravitated towards large, densely populated regions and cities (OECD, 2023[10]). The share of the OECD population living in cities increased by around 3.5 percentage points from 2000 to 2020 (45.2% in 2000, 48.8% in 2020) (OECD, 2022[19]). This trend is consistent with the evidence that, as countries develop, they have larger urban population shares (OECD/EC, 2020[20]). The literature points to the advantages of agglomeration as a primary reason for the increasing geographic concentration of people, including economic opportunities and amenities (Combes and Gobillon, 2015[21]).
Urbanisation, defined as the increasing spatial concentration of populations in metropolitan regions, is projected to continue over the next two decades, in part because of negative population growth in OECD countries: even if the population in metropolitan regions remains roughly unchanged, non-metropolitan areas are expected to lose around 2.5% of their population over that period (OECD, 2022[19]). Within metropolitan areas, an increasing share of the OECD population is expected to move into the largest cities and their commuting zones by 2030, while the population in smaller functional urban areas (FUAs) is expected to shrink (OECD, 2022[19]).
The population of OECD countries is also increasingly ageing (Burgalassi and Matsumoto, 2024[22]). Although ageing will occur in all types of regions over the next 2 decades, non-metropolitan regions will be most impacted, as existing gaps in elderly dependency rates (around 20% in metropolitan regions versus. 22% elsewhere) are expected to increase, particularly in countries where non-metropolitan regions already have relatively high elderly dependency rates, such as Japan, Korea and Lithuania.
Figure 1.2. Share of population inside and outside of settlements, by settlement type
Note: Only those countries with detailed data on the location of services are included (e.g. Chile, Colombia, Japan, Mexico, the United Kingdom and others are omitted). Average includes all countries listed in Annex Table 1.B.1, weighted by population. Countries are listed in descending order by their percent of population in cities.
Source: Based on sources in Annex Table 1.A.3.
Older people in smaller settlements tend to rely particularly strongly on local services. They are often less mobile than working-age residents, which makes obtaining services in other settlements more cumbersome. They also tend to use public services such as healthcare more intensively than younger people and are more reliant on physical services compared to online services. Lower shares of working‑age people in rural areas can also impact the scope of sustaining or expanding local services. Similarly, lower shares of young people can present access and cost challenges in education provision (OECD/EC-JRC, 2021[11]). Chapter 2 examines the ways that service provision varies with the characteristics of settlements, including size and territorial attributes such as remoteness.
Definitions and measurement
The main measures used in the analysis in this report are the number of points of interest (POIs) within a settlement, the time it takes to reach a larger settlement from a smaller one and the population that lives within a certain travel time of a given settlement. To calculate these, this report makes use of three types of data: population grids, driving (or public transit) times and the location of services. The sample of countries covers most of Europe plus Australia, Canada, New Zealand, Korea and the United States, as well as many OECD countries and some accession countries with data on the location of services.
Population data are typically based on the country’s most recent census, converted into 1 km2 grid cells. The source data for population is GEOSTAT-2011 and 2021 for European countries, a national 2021 population grid for Korea, a national 2016 population grid for New Zealand and GHS-POP 2015 and 2020 for Australia, Canada and the United States (Annex Table 1.A.3). Data on built-up areas from GHS-BUILT, derived from satellite data, are used in the DEGURBA algorithm for Australia, Canada, New Zealand and the United States. Updated population grids for Australia (2021), Korea (2023) and New Zealand (2023) became available after the analysis was completed.
Accurate and up-to-date travel and accessibility indicators are available from a combination of road transport network data, public transit schedules and driving time computations. Finally, data on the location and scope of services have been collected with the help of national statistical agencies.
What is a settlement?
Some classifications define settlements via their service provision. For example, France delimits its “living area” classification, bassins de vie, based on the provision of services in municipalities (INSEE, 2024[23]).3 This report takes a different approach. It defines settlements by their resident population and then examines the services that are located there. Comprehensive data make it possible to examine the overlap between the location of people (based on place of residence) and the location of services across many OECD countries.
For this report, settlements are clusters or agglomerations of people. This requires internationally consistent definitions of settlements such as cities, towns and villages to compare their functioning across countries. This report identifies settlements using the DEGURBA definition, summarised in Box 1.1. The definition was co-developed by the European Commission and the OECD along with four other international organisations and endorsed at the United Nations Statistical Commission in 2020 as the recommended method for international statistical comparisons between cities and other settlements along the rural-urban continuum. The method, aggregating population from granular grid cells, ensures that settlements are defined in a consistent manner across countries, the absence of which was previously an impediment to this type of analysis (OECD et al., 2021[1]). These common definitions, described later in more detail, facilitate comparisons and enable a more nuanced analysis of the roles of settlements in service provision. Nevertheless, measurement issues – summarised in Box 1.2 – affect the computation and interpretation of DEGURBA within and across countries.
Table 1.2 shows the median population of each type of settlement in the OECD countries. Gridded data on the resident population are needed to map settlements and measure their access to services and amenities. The population grid approach ensures that settlements of different sizes as well as the indicators of accessibility to services and amenities, are broadly comparable across countries.
Table 1.2. Population sizes by type of settlement
1. Larger cities are comprised of mid-size and large cites and include all those with population above 250 000 inhabitants.
Source: Based on sources in Annex Table 1.A.2.
Building blocks: Grid-level population
Population data such as those derived from national censuses show that most people in OECD regions live in settlements. What data are needed to identify settlements? In some cases, national statistical agencies overlay “grids” of small polygons on their countries’ maps and use geocoded address data to report the number of people residing in each 1 km2 grid cell. In other cases, satellite data on the location of structures are used to impute the granular distribution of population from census data reported for statistical units such as municipalities. Such granular data on the location of population and structures are used to identify clusters of densely populated grid cells or buildings that can be classified as settlements.
Population data can be obtained for all countries from estimated grids provided by the Global Human Settlement Layer (GHSL) but some countries have higher-quality official grids. When available, data come from official national population grids at the 1-km2 detail, including the GEOSTAT gridded population estimates for European Union (EU) countries. Otherwise, the analysis uses data from the 2021 release of the GHSL gridded population estimates (GHS-POP) with 2019 reference year at 1-km2 detail produced by the European Commission Joint Research Centre (JRC). National population grids cover the entire territory of a country and aggregate georeferenced microdata into each 1 km2 grid cell, i.e. a bottom-up approach. Instead, the population grids in GHSL downscale census or administrative units to grid cells using the distribution and density of built-up area as mapped in the GHSL global layer (OECD et al., 2021[1]).
Box 1.2. Applying and using DEGURBA
Applying DEGURBA
The JRC) provides tools that researchers can use to derive DEGURBA definitions from a population grid, whether from national census data, the GHSL project (https://ghsl.jrc.ec.europa.eu/tools.php) or another source. The definitions of settlements and area typologies are straightforward. Nevertheless, certain issues lead to statistical differences between the intended definition and the actual computation and dissemination of DEGURBA.
Measurement considerations and caveats
Some standard conventions for measuring population may affect the interpretation of DEGURBA. The DEGURBA definitions depend crucially on the quality of the underlying population census data. Consequentially, there is an assumption that population is counted only in the case of primary residence, which can lead to some undercounting in areas where tourists and second residences are common. Residents of military sites are typically not counted in census tallies and such data may be suppressed due to national security considerations.
Many countries, including New Zealand and Türkiye, restrict the release of data in certain areas due to confidentiality considerations for regions or settlements with small populations. Türkiye has considered adding statistical noise to their geographically detailed data to facilitate public dissemination. Similar approaches have already been implemented in New Zealand and the United States, enabling some (previously restricted) data to be released at small geographic scales.
Other issues arise when classifying grid cells: for example, some grid cells are less than 1 km2 due to waterways, steep slopes or parks. Should these be classified and given the same weight as grid cells with 1 km2 that consist entirely of land? The DEGURBA definition of urban centres is being modified to exclude cells that face a body of water (shores, beaches, etc.) from the surrounding cells that must meet population density thresholds. This increases the extent to which urban centres can include areas along the shores of rivers, lakes and seas.
Built-up areas can be counted as urban in countries that tend to have sprawling cities. Without such an option, many cities that are considered single urban areas (e.g. Houston) appear as multiple urban centres that apply DEGURBA to the population alone. This happens because highways, railways, shopping centres, office parks and factories typically have almost no residential population. To address this issue, at least half of built-up cells can be counted as urban even if they have no population. The exact threshold (whether 50% or less) used for built-up areas should consider the source of satellite data and the resulting settlements identified, as including built-up areas can dramatically increase the size and reach of urban areas.
Source: U.S. Census Bureau (2023[24]), “Census Bureau releases 2020 Census DHC Noisy Measurement File”, https://www.census.gov/newsroom/press-releases/2023/2020-census-dhc-noisy-measurement-file.html
Box 1.3. Innovative uses of DEGURBA statistical information in OECD countries
Countries use DEGURBA not only for international comparisons but also for summarising economic and social conditions in their own urban, semi-dense and rural areas. A wide range of applications have emerged across countries, many of which have implemented DEGURBA within their national statistical offices (NSOs). In some cases, DEGURBA implementation has been part of a broad United Nations initiative to track Sustainable Development Goals (SDGs) for international comparability, particularly across cities.
Dialogues with these statistical agencies reveal innovative uses of the DEGURBA definitions. Some relevant and diverse projects are described below:
In New Zealand, natural disaster response uses DEGURBA to identify population clusters. This helps emergency responders prioritise time-sensitive search and rescue efforts and target preventive and post-event relief measures to maximise effectiveness (e.g. Cyclone Gabrielle, COVID-19).
South Korea’s NSO (KOSTAT) has a publicly accessible SGIS open platform (https://sgis.kostat.go.kr/view/urban/main). It maps all DEGURBA settlements and provides statistics on their size (population and land area). Other features include visualisations of population and density changes over time and additional layers showing the structure of households and the locations of businesses and public services.
In the European Union, employment rates and other economic data are now tabulated by DEGURBA on a quarterly or annual basis. This includes SDGs such as “People at risk of poverty or social exclusion” or those in “Households with very low work intensity”.
Colombia uses DEGURBA definitions for SDG measurement and monitoring, including access to adequate housing, public transport, open space and the relationship between land consumption and population.
Mexico uses business register data along with DEGURBA-based population measures to identify the geography of services and industrial production.
In Chile, demographic statistics are tabulated by DEGURBA, providing insights into the age and gender structure of semi-dense suburbs and other types of areas.
In Brazil, population and land use data integration reveals changes in built-up areas relative to population demands. Such systems can help track the preservation of protected areas and habitats over time.
Note: All examples draw upon material presented at the OECD/EC hybrid workshop on “Using DEGURBA around the World”, held on 26 June 2023 in Paris, France.
Sources: New Zealand: Dragonfly (2023[25]), “Interactive map shows community impacts of adverse weather”, https://www.dragonfly.co.nz/news/2023-05-02-cyclone-gabrielle-impact-map.html and ArcGIS (ArcGIS, 2023[26]), Cyclone Gabrielle GIS Story, https://storymaps.arcgis.com/stories/b51c6b5ba14d4ea18fc8350580983fe5; Europe: Eurostat (2022[27]), “Urban-rural Europe - Labour market”, https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Urban-rural_Europe_-_labour_market; Mexico INEGI (2023[28]), Directorio Estadístico Nacional de Unidades Económicas, https://www.inegi.org.mx/app/mapa/denue/.
Settlement position and reachability
In addition to the definition of settlements and measurement of population and services within them, other territorial characteristics of settlements are important for understanding geographic patterns of service provision. These territorial characteristics referred to as “reachability” relate to a settlement’s proximity and connections to other settlements nearby.
A settlement’s reachability depends on its accessibility (how far and well-connected it is) from other settlements. Identifying reachable settlements is key to developing a classification of a settlement hierarchy showing the centrality or remoteness of settlements. A settlement that is larger and easier to access will tend to be more central in providing services both to its residents and to others nearby. Geographic centrality is related to economic principles such as agglomeration benefits along with network science phenomena such as preferential attachment (i.e. the more connected a node is, the more likely it is to receive new links). According to the central place theories of Christaller (1933[29]) and Lösch (1940[30]), consumers are willing to travel a maximum distance or time to acquire particular goods and services. At the same time, these goods or services will be available only once a market reaches a minimum size in terms of population or income. Larger settlements will have a greater number of services along with more specialised varieties.
Several technical concepts, summarised in Box 1.4 reflect the spatial relationships between population, mobility and points of interest. This report uses accessibility as its primary concept because travel time is crucially important in assessing the extent to which people can physically reach services in their local area. It also uses driving time isochrones (representing equal travel times) to identify settlements that are reachable from other settlements. Travel time is preferable to distance since terrain, road quality and connectedness affect realised travel time, which are more relevant when people consider how to access services. Most chapters in this report use driving times as a benchmark because driving is a common form of transportation in both urban and rural areas. Furthermore, driving does not depend on transit stops and connections in the same way that train networks depend on them and, although also network-dependent, bus travel times are highly correlated with driving times. Chapter 3 explores the interaction between service provision and access via public transit, including bus, train and ferry connections.
The driving time data yield two measures of settlement reachability: access to a city, applicable to smaller settlements (i.e. villages and towns) and regional centres, applicable to all types of settlements.
Box 1.4. Transport accessibility measures
For all measures below, the main inputs are the number of destinations (or number of people) and the distance or travel time between them. Travel time measures depend not only on distance but on available and, in some cases, preferred modes of transport.
Proximity is defined as the total number of destinations available within a given distance from a given location (regardless of the travel time required):
In essence, proximity is expressed as the ratio of points of interest (number) to a measure of distance, such as kilometres.
Accessibility refers to the total number of opportunities (people, services, jobs) that can be reached from a location by driving, cycling, walking or taking public transport within a given amount of time (e.g. hours):
Accessibility adds a dimension of travel time to proximity, assuming a certain type or combination of transport modes.
Measures of accessibility reflect the availability of opportunities in the location’s surroundings and the characteristics of the transport network connecting that location to other places.
The transport performance ratio captures how well the network connects residents of a given area to nearby opportunities:
It is constructed as the ratio of accessibility (the total number of destinations or people reachable from a given location by a given transport mode within a given amount of time) to proximity (the total number of destinations or people within a given absolute distance from the location).
The population ratio represents the effective share of people with access to a town or village. It scales the number of accessible people within a given travel time to those living within a given distance from the settlement. For example, a ratio of 0.4 indicates that fewer than half of people within a 60-kilometre radius are accessible in an hour via public transport.
Reachability refers to the two main accessibility-based indicators in this report. All settlements (including cities) are classified by whether they are the largest within a certain driving time (i.e. regional centre or not). Smaller settlements are further characterised by their time (or access) to a city, measured from the small settlement’s centroid to the nearest city border, identified from the DEGURBA definitions.
Source: ITF (2019[31]), “Benchmarking Accessibility in Cities: Measuring the Impact of Proximity and Transport Performance”, https://doi.org/10.1787/4b1f722b-en.
Access to a city
The “Access to a city” measure indicates whether any part of a city is reachable from the centre of each smaller settlement. For example, a town close to a larger city may have fewer services for its own population, as residents are more likely to use some services from the city nearby. For villages and towns, the classification for “Access to a city” records whether these smaller settlements have any cities nearby. Thirty-minute drive-time isochrones(representing equal travel times from a central point) form the basis of the access to a city criterion. Instead of working with city centroid isochrones, these calculations use isochrones from the centroid of smaller settlements.4 Unlike large cities, there is not much difference between the centroid and edges of smaller settlements. This means that a small settlement has “Access to a city” if its residents can reach any part of a city, including the city’s (closest or outermost) border, within 30 minutes.
The measure of “Time to a city” uses isochrones to determine whether a city is reachable from each town or village. This definition uses travel time rather than distance, which is important because it accounts for differences in road quality and terrain that impact accessibility by car.5 To this end, the work leverages the Mapbox Isochrone Application Programming Interface and TomTom road network data to compute areas that are reachable within a specified amount of time from the population-weighted centroid of each settlement and returns the reachable regions as isochrones (i.e. contours of polygons representing equal travel times). Traffic information is only available for some countries. This option might be a useful extension, especially when considering accessibility to cities, but it presents other conceptual challenges, including variations by time of day. Thus, the driving times in this report do not account for traffic conditions.
As shown in Table 1.3, some of the subsequent analysis splits settlements into two consolidated groups depending on their accessibility to cities:
A settlement has “Access to a city” if located within a 30-minute drive from the boundary of any city (including multiple cities) or has “No access to a city” if no city can be reached within that time.
Settlements with “Access to a city” are further divided into two mutually exclusive categories:
Those “Close to a larger city”. Cities are categorised using a 250 000 population threshold and settlements close to smaller and larger cities are classified as close to larger cities.
Those “Close to a smaller city”. Time to a small city is relevant for smaller settlements such as towns or villages, whereas small cities might be classified by their access to a larger city.
Subsequent chapters of the report use this classification to study relationships between reachability and service provision.
Table 1.3. Time to cities classification
Note: Small cities have between 50 000 and 250 000 inhabitants. The categories “Farther from a city” and “Remote” are often considered together as having “No access to a city” in later analysis.
1. Larger cities include mid-size and large cities, with more than 250 000 inhabitants.
Figure 1.3 shows a map of all smaller settlements (i.e. villages and towns) in the United States by their “Time to cities” classification. In the eastern half of the country, many smaller settlements are clustered around cities (dark purple and blue dots). Settlements that are green dots are 30-60 minutes from a city, while “Remote” settlements (lighter yellow dots) are located more than an hour from a city. In the west of the country (e.g. Montana and Utah), a larger fraction of towns and villages are classified as “Remote”.
Figure 1.3. United States’ towns and villages by time to cities
Source: Based on Mapbox (2024[32]), Mapbox Isochrone API, https://docs.mapbox.com/api/navigation/isochrone/.
Regional centres
A regional centre (RC) is the largest settlement within a certain driving time (Jacobs‐Crisioni, Kompil and Dijkstra, 2023[2]). The notion of an RC captures the centrality of a settlement in relation to its surrounding territory. RCs are likely to be more prominent in the provision of services for their own populations and for larger catchment areas. They are defined independently from a settlement’s absolute population size or DEGURBA. Only accessibility to other settlements and their relative sizes are considered. RCs can be villages, towns or cities but cities are much more likely to be RCs since they tend to be the largest settlements within a given area.
Since there is often a home bias for services (especially for the services analysed: education, financial, health), the RC definition considers settlements only within national borders. In other words, a settlement close to a national border can be classified as a RC for a certain time threshold, even if a larger settlement is reachable with the same time threshold on the other side. For analysis purposes, RCs are defined by country because such centres are typically more important for the country’s residents even when there is another closer city in a neighbouring country. This is particularly true of educational and health services, where place of residence within administrative boundaries nearly always determines typical access to the services (OECD, 2021[8]).
Four different time thresholds were considered to explore the properties of RCs: 15, 30, 45 and 60 minutes. For each time threshold, a settlement is classified as RC if no larger settlement is within the isochrone for that time threshold. In other words, a 30-minute RC has no larger settlement within a half-hour drive from its centroid; however, non-RCs may have more than one 30-minute RC within a half-hour drive if these RCs are more than 30 minutes from each other. It follows that every settlement that is a 60-minute RC is also a 45-, 30- and 15-minute RC. Figure 1.4 illustrates the RCs by time threshold in Korea and Table 1.4 shows the share of cities, towns and villages that are classified as RCs using 15-minute time increments.
Figure 1.4. Regional centres in Korea
Source: Based on Mapbox (2024[32]), Mapbox Isochrone API, https://docs.mapbox.com/api/navigation/isochrone/.
Table 1.4. Percentage of settlements that are classified as regional centres in OECD countries
Notes: Small cities have between 50 000 and 250 000 inhabitants. Simple average of settlements, by DEGURBA type, pooled across countries.
1. Larger cities (including mid-size cities) have more than 250 000 inhabitants.
Source: Based on sources in Annex Table 1.A.2 and Mapbox (2024[32]), Mapbox Isochrone API, https://docs.mapbox.com/api/navigation/isochrone/.
A settlement with the largest population within a certain driving time (e.g. 30 minutes) is classified as a 30‑minute RC. This report applies a 30-minute threshold (corresponding to the highlighted column), which means 17% of all towns and 4% of all villages are classified as RCs. The 30-minute travel threshold is close to the average amount of time workers spend getting to work, which could entail travelling to a city or a larger town near where they reside. In 2019, average commuting times in the EU member states were 25 minutes, ranging from 20 to 30 minutes, depending on the country (Eurostat, 2023[33]). Workers living in rural areas typically have shorter commutes than those living in more densely populated areas like cities. In the United States, the average one-way commute was 28 minutes while it was slightly lower in Canada (24 minutes in 2021) (U.S. Census Bureau, 2021[34]; Statistics Canada, 2023[35]). Korea has a longer average commute: around 36 minutes each way (Min-sik, 2023[36]).
RCs can be villages, towns or cities but cities and even towns are much more likely to be RCs than villages. In densely populated countries like Belgium, Germany and the Netherlands, it is more likely that villages are within 30 minutes of a larger settlement, such as a town or city (Figure 1.5), so less than 1% of villages are RCs. On the other hand, in countries with sparsely populated areas (e.g. Australia, Canada, Finland, Norway), 17-24% of villages are RCs. Looking at RCs, only a small share of them are villages, despite a high prevalence of villages in all countries.
Travel time thresholds simplify the distance decay function that is often used in central place theory models (Christaller, 1933[29]). The assumptions underlying RCs are twofold: first, if there is an acceptable maximum travel time to a certain service for most customers, there would likely be a service provider within that radius; second, service providers will most likely establish themselves within the local largest settlement of a region. Since cities are larger than towns and villages, only towns or villages without access to a city can be RCs.
For smaller settlements (villages and towns), there is a clear overlap between “Access to a city” and RC classifications:
Villages and towns with “Access to a city” within 30 minutes (and not across a national border) are not 30-minute RCs because they are smaller than the nearby city or cities. Thus, any village or town that is a RC must have no access to a city.
Villages and towns with “No access to a city” can be RCs or not. For example, a village far from a city may be close to a town or a larger village that qualifies as the 30-minute RC.
Figure 1.5. Share of villages that are 30-minute regional centres
Source: Based on sources in Annex Table 1.A.2
and Mapbox (2024[32]), Mapbox Isochrone API, https://docs.mapbox.com/api/navigation/isochrone/.
Box 1.5 discusses an alternative way of characterising land use in the context of the rural-urban continuum.
Box 1.5. Comparing DEGURBA with spatial signatures
Spatial signatures are a new territorial classification based on both urban form – the appearance and spatial configuration of places – and function – the activities and opportunities available in those places (Arribas-Bel and Fleischmann, 2022[37]). Machine learning algorithms applied to land use data identify 16 distinct signature classes, ranging from “Wild countryside” to “Hyper-concentrated urbanity”. While DEGURBA uses 1 km2 population grid cells as inputs, signatures are delineated by granular irregular units called “enclosed tessellation cells”, based on building footprints and edges such as streets and rivers. Data on economic activity (employment, business locations) and other functions (residential/ commercial) help inform the signature classes.
Figure 1.6 shows an example of how DEGURBA classifications correspond to spatial signatures in Great Britain (the five “Urbanity” classes are aggregated into a single category). For instance, most British village grid cells are classified as “Countryside agriculture” within the signatures, while a third are classified as “Urban buffer” (which lies between “Disconnected suburbia” and “Residential quarters” in terms of urbanity). Towns are more frequently tagged as “Accessible suburbia” or “Sprawl”, while “Wild countryside” is a common signature for other DEGURBA areas, namely semi-dense and rural areas outside of settlements.
Figure 1.6. Distribution of spatial signatures in DEGURBA settlements and villages
Source: Arribas-Bel, D. and M. Fleischmann (2022[37]), “Spatial signatures - Understanding (urban) spaces through form and function”, https://doi.org/10.1016/j.habitatint.2022.102641.
DEGURBA and spatial signatures are distinct landscape and built environment classifications – with some complementarities. DEGURBA is more useful for delineating settlements, while signatures provide nuanced characterisations of land use within settlements. Overlaying the two definitions yields new insights into the typologies of smaller settlements, cities and non-settlement areas of different countries and regions.
Sources: Arribas-Bel, D. and M. Fleischmann (2022[37]), “Spatial signatures - Understanding (urban) spaces through form and function”, https://doi.org/10.1016/j.habitatint.2022.102641; Fleischmann, M. and D. Arribas-Bel (2022[38]), “Geographical characterisation of British urban form and function using the spatial signatures framework”, https://doi.org/10.1038/s41597-022-01640-8.
Measuring service provision in settlements
This report focuses on services that are important for people’s quality of life and, moreover, clearly defined and measurable across a large set of countries (Annex Table 1.A.1). Precise location information for services is available for 31 OECD countries. Five services are referenced throughout the report (banks, hospitals, pharmacies, schools and higher education institutions such as universities) based on their relevance to well-being and the availability of data. Due to data limitations, this analysis only considers the physical presence of services (without considering their size, quality or other characteristics). The physical location of service points indicates where services are provided; however, no information was collected regarding the consumer or user base. For example, information on hospital patient registers or population assignments to school districts was not collected. In addition, data on most services’ physical or economic size were not typically available, nor on the cost or (subjective) quality. Box 1.6 discusses how available data on services relate to ideal accessibility measures and several case studies are presented in Chapter 2 for the purpose of model validation.
Most service locations are identified from high-quality, official national sources such as health and education registries. In some cases, such data are publicly available through open data platforms but statistical agencies facilitate the search process. A few countries provided data directly to the OECD. Compared to user-sourced or private sector data, official data are more likely to be complete and thoroughly cover services in smaller and larger settlements. Most official data list precise locations (street addresses or latitude/longitude). Annex Table 1.A.1 has the full list of service data availability by country and Annex Table 1.A.2 has more details about the data sources.
The initial analysis of services looks at whether settlements have any service of a given type. The presence of a service is especially relevant for scarce (uncommon) services like hospitals, banks and even cinemas. The total number of service locations is more relevant for ubiquitous (common) services like schools and pharmacies, where multiple locations often serve the population of a single settlement.
Box 1.6. Measuring services
What type of services count? Common definitions are needed
Comparable data are the cornerstone of any analysis of service provision. There are many impediments to these comparisons, as institutional and cultural backdrops can differ markedly across countries, along with assumptions about what constitutes the physical presence of a service. For instance, this study assumes that automated teller machines (ATMs) do not substitute for physical bank branches but it does not distinguish between pharmacies with limited business hours and those open around the clock.
Location data do not typically record a location’s capacity, prices or specific qualities
Average service sizes or capacities vary across settlements. For example, a single school location may serve a small number of students in a village, whereas schools in cities typically serve a larger number of students at each location. Differences in service sizes are not captured by data that count the number of sites or locations. Even data on sizes may not measure normal usage: whether the facility is typically underutilised (which may mean its full capacity is not normally needed) or oversubscribed (some clients who need its services may lack access) can affect the extent to which a service is fulfilling its mission. The time dimension of service locations is also very difficult to track with public data.
Price is another important aspect of access, especially in relation to socio-economic status, yet it is still under-reported on a large scale. Quality, which is particularly difficult to measure, is related to the functions of services. In some cases, quality can be inferred from objective public accountability data (e.g. student test scores, performance outcome measurements) or from user-provided surveys or reviews. Data collection on quality is limited across service types and places, thus it is omitted from this cross-country study.
There are also differences in the breadth or variety of services offered within locations. Some hospitals specialise in treating children, others have more general services, while a small share specialises in advanced treatments for specific conditions. Some studies, such as those conducted by the Australian Bureau of Infrastructure and Transport Research Economics (2019[39]), have attempted to classify services according to a hierarchical rank, finding that large cities tend to have multiple hospitals and health facilities with more diversified and advanced services than smaller places. Places with multiple service options are more likely to rank favourably on price, quality, convenience and variety.
Other factors, such as information and physical mobility, also matter. For example, the Atlanta Federal Reserve Atlanta (United States) has considered car ownership when assessing driving accessibility to financial services. Undeniably, digital access and postal delivery are becoming increasingly important with the advent of online access and shopping options. Nevertheless, many services like education and healthcare are not equally effective when provided on line in a virtual format compared to in-person service provision. Overall, the presence of a physical location is still an important proxy for access.
Sources: For hospital services in Australia: Australian Government (2019[39]), An Introduction to Where Australians Live, https://www.bitre.gov.au/sites/default/files/An-introduction-to-where-Australians-live-BITRE-Information-Sheet-96.pdf; For access to banks in Atlanta, Georgia (United States): Haspel, M. (2023[40]), “Banking deserts and banking droughts: A deeper dive”, https://33n.atlantaregional.com/data-diversions/banking-deserts-and-banking-droughts-a-deeper-dive.
Overview of chapters
Chapter 2 analyses how service prevalence varies according to settlement characteristics. It uses detailed data from 30 OECD countries to investigate the location of public and private sector services, building a statistical model that relates population to the prevalence of services across space. The existence of at least one service location is assessed for uncommon services like universities, whereas for common services like schools, the total number of locations is assessed.
Chapter 3 investigates the role of public transportation in providing access to services. It examines whether local centres with good public transport service have better service provision for public and commercial amenities like hospitals, universities, banks and pharmacies. The analysis is based on five European countries/regions with available and relatively complete data on population, transport and amenities datasets. For given travel time thresholds (45 minutes), the analysis evaluates the population who can reach the settlement, illuminating the interaction between public transport connections and service provision. The analysis also compares accessibility to settlements via public (multimodal) versus private (car) transport.
Chapter 4 investigates the settlement characteristics associated with population growth over time, building on the other information gathered in Chapter 2 – including services and settlement reachability. It focuses on the population growth patterns of mid‑size settlements (towns and cities with fewer than 250 000 inhabitants), in which nearly a third of the OECD population resides.
Annex 1.A. Data sources
The types of services and brief definitions are listed in Annex Table 1.A.1 below, along with country coverage. Additional information about data sources follows.
Annex Table 1.A.1. Service definitions and country coverage
1. Data on cinemas, food stores, restaurants and bars are used only for analysis of a sample of countries in Chapter 3.
Data sources by country and type of service are listed in Annex Table 1.A.2. Australia and New Zealand have a variety of public data sources. For the European Union, two types of GIS databases (GISCO and ESPON) cover private and public sector establishments. In Canada, Statistics Canada maintains databases of healthcare, commercial and educational facilities. Data from Korea come from an official database on points of interest, provided in a confidential manner to the OECD from Korea’s Ministry of Land, Infrastructure and Transport. For the United States, most data come from the Homeland Infrastructure Foundation Level.
In Australia, the Royal Flying Doctor Service and School of the Air provide important access to health and educational services in remote areas but are not counted as belonging to particular settlements.6
Annex Table 1.A.2. Service data sources by country
Annex Table 1.A.3. Population data sources by country
Note: The built-up area 1 km2 grids are based on GHS-BUILT (2014), derived from satellite data.
Sources: GEOSTAT (Eurostat, 2011; 2021[41]) data and GHS-POP data from https://human-settlement.emergency.copernicus.eu/download.php
Annex 1.B. Population by settlement type
Annex Table 1.B.1. Population by settlement type and country totals (including non-settlement)
Source: Based on sources in Annex Table 1.A.3.
Annex 1.C. Changing the definition of semi‑dense towns
The Level 2 DEGURBA definition distinguishes between dense and semi-dense towns. The existence of semi-dense towns reduces the number of entirely suburban (or peri-urban) areas without any towns. However, the DEGURBA steering group, during its testing and consultation with countries, identified some issues with the definition of semi-dense towns.
In the original DEGURBA definition, semi-dense towns and expanses of suburban grid cells both use the same population density rule; differences between the two relied on grid cell clustering. Consequently, the definition identified semi-dense towns within some suburban areas while simultaneously missing clusters of population in other, similar-looking areas. Moreover, semi-dense towns tended to be larger than dense towns and also lacked a clearly identifiable suburban fringe. As a result, changes to the DEGURBA manual are underway to more clearly distinguish semi-dense towns from swaths of suburban areas.
The revision to DEGURBA makes three main adjustments. First, it increases the density threshold for semi-dense towns from 300 to 900 inhabitants per km2. Second, it reduces the minimum population threshold to 2 500 inhabitants. Finally, it implements a technical change by relaxing contiguity thresholds. The first change lowers the number of towns, while the other two changes increase the number. Overall, the new definition still has a non-negligible number of “disconnected suburbs”, large swaths of suburbs with no town. Many of these are in North America, where sprawling suburbs are relatively common.
Annex Table 1.C.1 below compares the definition’s two vintages. Although the 2024 changes have been adopted, the report uses the 2021 definition because all analysis was done before the revised definition was fully introduced.
Annex Table 1.C.1. Semi-dense town definitions
References
[26] ArcGIS (2023), Cyclone Gabrielle GIS Story, https://storymaps.arcgis.com/stories/b51c6b5ba14d4ea18fc8350580983fe5.
[37] Arribas-Bel, D. and M. Fleischmann (2022), “Spatial signatures - Understanding (urban) spaces through form and function”, Habitat International, https://doi.org/10.1016/j.habitatint.2022.102641.
[49] Australian Children (2021), About the School of the Air, https://australian-children.com/school-of-the-air/about-school-of-the-air/.
[39] Australian Government (2019), An Introduction to Where Australians Live, Australian Bureau of Infrastructure, Transport and Regional Economics, https://www.bitre.gov.au/sites/default/files/An-introduction-to-where-Australians-live-BITRE-Information-Sheet-96.pdf.
[15] Baptista, I. and E. Marlier (2020), Access to Essential Services for People on Low Incomes in Europe - An Analysis of Policies in 35 Countries, European Social Policy Network (ESPN), Brussels, https://ec.europa.eu/social/main.jsp?catId=738&langId=en&pubId=8340&furtherPubs=yes.
[14] Bastiaanssen, J. and M. Breedijk (2022), “Access for all? An analysis of the (in)accessibility of facilities and jobs in the Netherlands”.
[22] Burgalassi, D. and T. Matsumoto (2024), Demographic change in cities.
[3] Cattaneo, A., A. Nelson and T. McMenomy (2021), “Global mapping of urban-rural catchment areas reveals unequal access to services”, Proceedings of the National Academy of Sciences, Vol. 118/2, https://doi.org/10.1073/pnas.2011990118.
[29] Christaller, W. (1933), Central Places in Southern Germany.
[21] Combes, P. and L. Gobillon (2015), “The empirics of agglomeration economies”, in Handbook of Regional and Urban Economics, Elsevier, https://doi.org/10.1016/b978-0-444-59517-1.00005-2.
[25] Dragonfly (2023), “Interactive map shows community impacts of adverse weather”, https://www.dragonfly.co.nz/news/2023-05-02-cyclone-gabrielle-impact-map.html.
[7] Eurofound (2022), Access to Essential Services for People on Low Incomes: Energy, Public Transport and Digital Communications, Publication Office of the European Union, Luxembourg, https://www.eurofound.europa.eu/publications/customised-report/2022/access-to-essential-services-for-people-on-low-incomes-energy-public-transport-and-digital.
[33] Eurostat (2023), Main Place of Work and Commuting Time, https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Main_place_of_work_and_commuting_time_-_statistics#Commuting_time_for_majority_of_European_workers_is_less_than_30_minutes.
[27] Eurostat (2022), “Urban-rural Europe - Labour market”, Eurostat Statistics Explained, https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Urban-rural_Europe_-_labour_market.
[41] Eurostat (2011; 2021), GEOSTAT Population Grid, https://ec.europa.eu/eurostat/web/gisco/geodata/population-distribution/geostat.
[38] Fleischmann, M. and D. Arribas-Bel (2022), “Geographical characterisation of British urban form and function using the spatial signatures framework”, Scientific Data, https://doi.org/10.1038/s41597-022-01640-8.
[17] Gallup (2022), Homepage, https://www.gallup.com/home.aspx.
[40] Haspel, M. (2023), “Banking deserts and banking droughts: A deeper dive”, https://33n.atlantaregional.com/data-diversions/banking-deserts-and-banking-droughts-a-deeper-dive.
[28] INEGI (2023), Directorio Estadístico Nacional de Unidades Económicas, National Institute of Statistics and Geography, Mexico, https://www.inegi.org.mx/app/mapa/denue/.
[23] INSEE (2024), Base des bassins de vie 2022, Institut national de la statistique et des études économiques, https://www.insee.fr/fr/information/6676988.
[31] ITF (2019), “Benchmarking Accessibility in Cities: Measuring the Impact of Proximity and Transport Performance”, International Transport Forum Policy Papers, No. 68, OECD Publishing, Paris, https://doi.org/10.1787/4b1f722b-en.
[2] Jacobs‐Crisioni, C., M. Kompil and L. Dijkstra (2023), “Big in the neighbourhood: Identifying local and regional centres through their network position”, Papers in Regional Science, Vol. 102/2, pp. 421-458, https://doi.org/10.1111/pirs.12727.
[45] Kaufmann, T. et al. (2023), “Comparing access to urban parks across six OECD countries”, OECD Regional Development Papers, No. 45, OECD Publishing, Paris, https://doi.org/10.1787/4d17ce4c-en.
[12] Kelly, C. et al. (2016), “Are differences in travel time or distance to healthcare for adults in global north countries associated with an impact on health outcomes? A systematic review”, BMJ Open, Vol. 6/11, https://doi.org/10.1136/bmjopen-2016-013059.
[44] Lee, J. and C. Lubienski (2017), “The impact of school closures on equity of access in Chicago”, Education and Urban Society, Vol. 49/1, https://doi.org/10.1177/0013124516630601.
[5] Llena-Nozal, A., R. Fernández and S. Kups (2022), “Provision of social services in EU countries: Reform of the national framework for the provision of social services in Spain”, OECD Social, Employment and Migration Working Papers, No. 276, OECD Publishing, Paris, https://doi.org/10.1787/ba4fbaf2-en.
[30] Lösch, A. (1940), The Economics of Location.
[43] Luo, W. and F. Wang (2003), “Measures of spatial accessibility to health care in a GIS environment: Synthesis and a case study in the Chicago region”, Environment and Planning B Planning and Design, Vol. 30/6, pp. 865-884, https://doi.org/10.1068/b29120.
[32] Mapbox (2024), Mapbox Isochrone API, https://docs.mapbox.com/api/navigation/isochrone/.
[42] McGrail, M. and J. Humphreys (2009), “Measuring spatial accessibility to primary care in rural areas: Improving the effectiveness of the two-step catchment area method”, Applied Geography, Vol. 29/4, pp. 533-541, https://doi.org/10.1016/j.apgeog.2008.12.003.
[4] Milstein, R., A. Castelli and N. Gutacker (2023), “Strategies to attract primary care physicians to rural, undeserved areas. A review of policies in nine high-income countries”.
[36] Min-sik, Y. (2023), “Daily commute in greater Seoul takes 83 minutes: Report”, The Korea Herald, https://www.koreaherald.com/view.php?ud=20231221000657.
[48] OECD (2023), “Geospatial modelling in support of Latvia’s school network reorganisation initiative”, OECD Education Policy Perspectives, No. 87, OECD Publishing, Paris, https://doi.org/10.1787/1729b571-en.
[10] OECD (2023), OECD Regional Outlook 2023: The Longstanding Geography of Inequalities, OECD Publishing, Paris, https://doi.org/10.1787/92cd40a0-en.
[19] OECD (2022), OECD Regions and Cities at a Glance 2022, OECD Publishing, Paris, https://doi.org/10.1787/14108660-en.
[8] OECD (2021), Delivering Quality Education and Health Care to All: Preparing Regions for Demographic Change, OECD Rural Studies, OECD Publishing, Paris, https://doi.org/10.1787/83025c02-en.
[20] OECD/EC (2020), Cities in the World: A New Perspective on Urbanisation, OECD Urban Studies, OECD Publishing, Paris, https://doi.org/10.1787/d0efcbda-en.
[1] OECD et al. (2021), Applying the Degree of Urbanisation: A Methodological Manual to Define Cities, Towns and Rural Areas for International Comparisons, OECD Regional Development Studies, OECD Publishing, Paris/European Union, Brussels, https://doi.org/10.1787/4bc1c502-en.
[11] OECD/EC-JRC (2021), Access and Cost of Education and Health Services: Preparing Regions for Demographic Change, OECD Rural Studies, OECD Publishing, Paris, https://doi.org/10.1787/4ab69cf3-en.
[6] OECD/WHO (2018), Health at a Glance: Asia/Pacific 2018: Measuring Progress towards Universal Health Coverage, OECD Publishing, Paris, https://doi.org/10.1787/health_glance_ap-2018-en.
[13] Pathman, D., T. Ricketts III and T. Konrad (2006), “How adults’ access to outpatient physician services relates to the local supply of primary care physicians in the rural southeast”, Health Services Research, Vol. 41/1, pp. 79-102, https://doi.org/10.1111/j.1475-6773.2005.00454.x.
[47] Royal Flying Doctor Service (2024), Homepage, https://www.flyingdoctor.org.au/.
[35] Statistics Canada (2023), “Quality of employment in Canada: Commuting time, 2011 to 2022”, https://www150.statcan.gc.ca/n1/pub/14-28-0001/2023001/article/00003-eng.htm.
[46] Stats NZ (2021), Urban Rural New Zealand, Statistics New Zealand.
[24] U.S. Census Bureau (2023), “Census Bureau releases 2020 Census DHC Noisy Measurement File”, United States Census Bureau, https://www.census.gov/newsroom/press-releases/2023/2020-census-dhc-noisy-measurement-file.html.
[34] U.S. Census Bureau (2021), “Press release: Travel time to work in the United States”, United States Census Bureau.
[18] UNSD (2020), “A recommendation on the method to delineate cities, urban and rural areas”, Background document, Prepared by the European Commission – Eurostat and DG for Regional and Urban Policy, ILO, FAO, OECD, UN-Habitat, World Bank, United Nations Statistics Division, https://unstats.un.org/unsd/statcom/51st-session/documents/BG-Item3j-Recommendation-E.pdf.
[9] Ward, T. and E. Ozdemir (2012), “Disparities in access to essential services”, Social Situation Observatory Income Distribution and Living Conditions: Research Note No. 8/2012, European Commission, https://ec.europa.eu/social/BlobServlet?docId=9806&Iangld=en.
[16] World Bank (2020), Handbook for Gender-Inclusive Urban Planning and Design.
Notes
← 1. A few studies investigate spatial access to parks and green spaces in a cross-country context (Kaufmann et al., 2023[45]) whereas studies of educational services are often conducted at the national level (OECD, 2023[48]).
← 2. The most common approaches are distance/time calculations, two-step floating catchment areas and gravity-based models (McGrail and Humphreys, 2009[42]; Lee and Lubienski, 2017[44]; Luo and Wang, 2003[43])
← 3. Another example is New Zealand’s 1991 urban area classification, which used cultural, recreational and business services as inputs to define minor urban areas. Settlements with fewer than 10 000 residents needed to provide a variety of services (schools, banks, shops, sports facilities, etc.) to be classified as urban rather than rural (Stats NZ, 2021[46]).
← 4. Outside of Europe, the “Access to a city” classification is computed by considering only the cities within national borders (e.g. Canada and United States) whereas in Europe, small settlements close to a national border can be classified as having access to a city even if the city is on the other side of the border.
← 5. Mapbox data were made available through the OECD’s participation in the Development Data Partnership (https://datapartnership.org/).
← 6. See Royal Flying Doctor Service (2024[47]) and Australian Children (2021[49]) respectively.