Territorial typologies manual - introduction
Purpose of this publication
Reliable and comparable datasets for different territorial typologies can only be produced on the basis of coherent building blocks. Location is a key attribute to virtually all official statistics: it provides the structure for collecting, processing, storing, analysing and aggregating data. The framework provided by a specific geographic feature, such as a national border, or proximity to a coastline, is often the only feature shared by different datasets. Moreover, location is a concept that most people are comfortable with, as statistics on specific areas help people to understand the relevance of particular statistics.
Eurostat’s Methodological manual on territorial typologies has been principally designed as a guide for data suppliers within the European Union (EU) so that they have the necessary information to ensure coherency within their data collections. It may also be of interest to users of subnational statistics so they may better understand and interpret the wide range of official statistics that are available at a subnational level for the EU.
The decision to make this publication reflects an important milestone concerning legislative developments: an amending Regulation (EU) 2017/2391 of the European Parliament and of the Council was adopted on 12 December 2017 as regards territorial typologies (Tercet), followed on 18 January 2018 by a consolidated and amended version of Regulation (EC) No 1059/2003 of the European Parliament and of the Council on the establishment of a common classification of territorial units for statistics (NUTS). Prior to the Tercet initiative, these territorial typologies and their related methodologies did not have any legal basis and they were, as such, not formally recognised by the European statistical system (ESS). These issues were subsequently addressed with subnational statistics now having a legal basis that is developed around a set of impartial and transparent territorial typologies. The main objectives of Tercet included, among others:
- establishing a legal recognition of territorial typologies for the purpose of European statistics by laying down core definitions and statistical criteria;
- integrating territorial typologies into the NUTS Regulation so that specific types of territory could be referred to in thematic statistical regulations or policy initiatives, without the need to (re-)define terminology such as cities, urban or coastal areas;
- ensuring methodological transparency and stability, by clearly promoting how to update the typologies.
To mark the end of this period of legislative developments, Eurostat has compiled this exhaustive guide providing information on the latest territorial typologies used for official statistics within the EU. The publication is structured following these legislative developments:
- it starts with this introduction, providing an overview of the typologies presented, details concerning the basic building blocks that are used to construct these typologies, and some background as to the various uses that can be made of subnational statistics;
- this is followed by an explanation of cluster types (Part A) — which are related groups of 1 km² grid cells that share the same characteristics in terms of their population density;
- the main body of the manual contains detailed explanations for the two main types of territorial typologies that are covered in Tercet-related legislation:
- local typologies (Part B) — which presents a range of typologies that are based on data for local administrative units (LAUs);
- regional typologies (Part C) — which presents a range of typologies that are based on regional data classified according to NUTS level 3;
- the last part presents a number of other regional typologies that are not currently covered by legislation (Part D), where further development work and/or agreement with other European Commission services is required before a legislative basis might be established; these typologies are also based on regional data classified according to NUTS level 3.
Overview of typologies
Most economic, social and environmental situations and developments have a specific territorial dimension — they are located in a fixed place — dependent, to some degree, upon a range of territorial resources, for example, transport or communications networks, access to services, as well as natural and human resources. Such geospatial diversity makes analysing different territories a complex task. In order to cast some light on territorial patterns, Eurostat has expanded its range of statistics that are published for territorial typologies.
A broad range of territorial typologies were integrated into the NUTS Regulation in December 2017, underlining the importance of subnational statistics as an instrument for targeted policymaking and a tool for understanding and quantifying the impact of policy decisions in specific territories. The Regulation provides a legal basis for the use of 1 km² grid cells, local administrative units (LAUs) and NUTS, as well as providing information as to how each of these concepts relates to establishing a complementary set of territorial typologies. By integrating these territorial typologies into a single legal text, it is hoped that they will be applied consistently and harmoniously, making it possible for them to be cross-referenced from other acts and programmes.
There are a number of key terms which appear in several places throughout this Methodological manual on territorial typologies. A short definition of these key terms is provided here before going into more detail later in this introduction (see Overview of typologies and Building blocks for typologies for more detailed definitions/descriptions).
Clusters: groups of grid cells that conform to a particular criterion in relation to their population density. For example, similar grid cells may be grouped into rural grid cells, urban clusters (moderate-density clusters) or urban centres (high-density clusters).
Local administrative units (LAUs): a system for dividing up a territory for the purpose of developing statistics at a local level. These units are usually low level administrative divisions within a country, ranked below a province, region, or state.
Classification of territorial units for statistics (NUTS): a territorial classification that subdivides the territory of the EU into three different hierarchical levels — NUTS level 1, level 2 and level 3 respectively — detailing larger to smaller territorial units. The NUTS is based on Regulation (EC) No 1059/2003 of the European Parliament and of the Council of 26 May 2003 on the establishment of a common classification of territorial units for statistics (NUTS), which is regularly updated. A consolidated version of the latest (amended) legislative text is available here.
Population grid: a lattice composed of 1 km² grid cells overlaying a particular territory, for which information is collected relating to the number of inhabitants. These grids are a powerful tool to describe the spatial distribution of population which may be used to analyse the interrelationships between human activities and the environment. The grid is stable over time, not dependant on changes in administrative boundaries and may be used for spatial aggregations to various territories of interest.
Region: a term with two distinct meanings in a statistical context. For the purpose of this Methodological manual on territorial typologies a region refers to a geographical area at a subnational level based on NUTS. Although not used in this publication, a region may also refer to a supranational level, as in a region of the world (for example, Latin America, south-east Asia, or indeed the EU).
Figure 1 presents an overview of the main territorial typologies that have been developed, often in conjunction with other European Commission services and/or other international organisations. At the most basic level, these concepts can be split into three different groups, covering grid typologies, local typologies and regional typologies.
Grid typologies: Eurostat collects population statistics based on 1 km² grid cells. These very detailed statistics are used to establish various cluster types — namely, urban centres, urban clusters and rural grid cells.
Local typologies: based on statistics for local administrative units (LAUs) which generally comprise municipalities or communes across the EU. Statistics for LAUs may be used to establish local typologies including the degree of urbanisation (cities; towns and suburbs; rural areas); functional urban areas (cities and their surrounding commuting zones); coastal areas (coastal and non-coastal areas).
Regional typologies: statistics that are grouped according to the classification of territorial units for statistics (NUTS); they provide information at a relatively aggregated level of detail, with data presented for NUTS level 1, level 2 and level 3 regions respectively, detailing larger to smaller territorial units. Only the most detailed statistics at NUTS level 3 are used as building blocks to establish the urban-rural typology (predominantly urban regions, intermediate regions and predominantly rural regions), the metropolitan typology (metropolitan and non-metropolitan regions), the coastal typology (coastal and non-coastal regions), each of which has a legislative basis. There are three other regional typologies (covered in this manual) for which there is, at present, no legal basis: the border typology (border and non-border regions), the island typology (island and non-island regions) and the mountain typology (mountain and non-mountain regions). The EU also has a number of outermost regions (defined by Article 349 of the Treaty of the Functioning of the European Union (TFEU)) which are geographically distant from the European continent. As their collective legal definition is provided for by the TFEU, this typology is not specifically covered in this manual. That said, the outermost regions form an integral part of the EU and as such they are included — as individual regional entities — within each regional typology.
The three different types of territorial typologies — grid, local and regional — are closely interlinked, as they are based on the same basic building blocks, namely, classifying population grid cells to different cluster types and then aggregating this information either by LAU or by region to produce statistics for a wide variety of different typologies. Figure 2 presents an example for how urban areas in the EU are defined at three different — but coherent — levels:
- at an initial level, urban centres (or high-density clusters) are identified as groups of grid cells with a population density of at least 1 500 inhabitants/km² and collectively a population of at least 50 000 inhabitants;
- these urban centres may be superimposed onto LAUs to identify cities (LAUs where at least 50 % of the population lives in an urban centre) and commuting zones (LAUs surrounding a city characterised by at least 15 % of their population commuting to work in the city); the term functional urban area is used to describe this wider aggregate that consists of a city and its surrounding commuting zone(s);
- functional urban areas may then be superimposed onto NUTS level 3 regions to identify metropolitan and non-metropolitan regions, defined as a NUTS level 3 region or groups of NUTS level 3 regions where at least 50 % of the population lives inside a functional urban area that is composed of more than 250 000 inhabitants.
This brief overview concludes with a set of 11 maps for Spain and Portugal that show the various territorial typologies that are presented within the main body of this publication; they underline the broad range of potential analyses that may be carried out when exploiting Eurostat’s subnational statistics. The first of these (Map 1) shows that the vast majority of the population in mainland Spain and Portugal is concentrated in areas that are close to the coastline, with relatively high population density in north-west and north-east Spain and northern Portugal. The principal urban centres shown in Map 2 — composed of high-density clusters — include Madrid, Barcelona and Valencia in Spain as well as Lisbon and Porto in Portugal.
The following two maps show local typologies. Map 3 confirms that large parts of the interior of Spain and Portugal are rural areas that have relatively low levels of population density. Commuting zones surrounding cities in Spain and Portugal are often relatively small with the exception of both capital cities (see Map 4). Indeed, many Iberian cities are characterised by highly concentrated city centres that remain relatively compact, with relatively clear boundaries dividing cities and sparsely-populated, surrounding areas.
There are two different territorial typologies that concern maritime areas, one based on LAUs and the other based on NUTS level 3 regions. Map 5 shows the first of these detailing coastal and non-coastal areas in Spain and Portugal and Map 8 shows the other, detailing coastal and non-coastal regions.
Maps 6 to 11 present information for regional typologies based on NUTS level 3 regions. The first of these shows the urban-rural typology, with the vast majority of regions in Portugal characterised as being predominantly rural. By contrast, relatively few regions in Spain are considered to be predominantly rural, with intermediate regions accounting for a majority of the NUTS level 3 regions in Spain. Map 7 shows that the vast majority of metropolitan regions in Spain and Portugal are located around their coastlines, with a few notable exceptions in Spain.
The final map for regional typologies shows the second typology related to the sea, namely, the coastal typology. This may be used to identify regions by individual sea basins (for example, regions that border the Mediterranean Sea or the North-East Atlantic Ocean), as well as outermost regions (which form an integral part of the EU). There are two such outermost regions in Portugal (Região Autónoma dos Açores and Região Autónoma da Madeira) and a single outermost region in Spain (Canarias; note this is composed of seven NUTS level 3 regions).
This final group of maps presenting the various territorial typologies shows three more regional typologies; note however that contrary to the typologies shown in Maps 6-8, the statistics collected for the typologies that are shown in Maps 9-11 do not have any legal basis. Map 9 identifies the (common) inland border regions of Spain and Portugal, as well as Spanish regions that share a land border with France.
Island regions are defined as NUTS level 3 regions that consist entirely of territories that are at least 1 km from the mainland without a fixed link (bridge, tunnel or dyke), with a minimum surface of 1 km² and a resident population of more than 50 inhabitants. Map 10 shows the island regions of Spain and Portugal (which include the outermost regions, as shown by the insets of this map).
Map 11 presents the mountain typology, which is based on regions characterised by having more than half of their population and/or surface area in mountain areas. Topographic mountain areas are defined using a range of criteria linked to height/altitude: they include all areas that are above 2 500 m, but also include areas that are as low as 300 m if these are characterised by a sufficiently undulating landscape (for example, the Scottish lochs or Norwegian fjords).
Building blocks for typologies
As noted above, the territorial typologies that have been developed by Eurostat can be split into three principal groups, covering grid-based typologies, local typologies based on LAUs and regional typologies based on NUTS level 3. This section provides a more detailed explanation for the three basic building blocks that underpin the various typologies, namely:
- the population grid;
- local administrative units (LAUs);
- the classification of territorial units for statistics (NUTS).
A population grid is composed of (usually square) grid cells containing population counts for each cell. Eurostat gives preference to the use of a 1 km² square grid that is overlaid across the EU territory.
Why use grid statistics?
Making use of a geographical grid for displaying population densities is not a new idea: indeed, examples exist from the late 18th century when this approach was adopted in statistical atlases. However, it is only relatively recently that it has become possible to consider the creation of population grids over very large areas, such as the EU, in a harmonised way. With the introduction of new technologies and increased computing power, a growing number of national statistical authorities developed the ability to produce statistics for very small areas. For example, during a census, most statistical authorities are now able to capture data based on geocoded data collection points such as addresses or buildings that is usually much more detailed than any other official data they publish.
Traditionally, official statistics have been reported in accordance with a hierarchical system of administrative units ranging from local administrative units (LAUs), through regions and countries to supranational aggregates covering, for example, the whole of the EU. While these systems provide high-quality data for accounting purposes and for the respective territorial authorities at each level of the hierarchy, they are not so suitable for studying the causes and effects of many socioeconomic and environmental phenomena which are often independent of administrative boundaries, for example, commuting, leisure activities, flooding or weather events. When studying such phenomena, a system of grids with equal-size grid cells has many advantages:
- grid cells all have the same size allowing for easy comparison;
- grids are stable over time;
- grids integrate easily with other (scientific) data;
- grid systems can be constructed/assembled to form areas that match a specific purpose or study area.
The detailed nature of population grid statistics is generally considered an advantage over more traditional statistics that are based on larger administrative or statistical areas. For example, the results presented for the urban-rural typology (in Map 6 above) do not show any variation in the level of population density for urban-rural regions around the Spanish capital of Madrid. By contrast, the information shown for the population grid (in Map 1 above) contrasts extremely dispersed populations for most areas around the Spanish capital with distinct population concentrations, such as the cities of Guadalajara, Segovia or Toledo, thereby providing a more realistic portrait of where people actually live.
Explanation of the population grid
Population grids are a powerful tool for describing the spatial distribution of a population and are particularly useful for analysing socioeconomic phenomena that are independent of administrative boundaries. This next section provides an example of how population grids work in practice.
A population grid is composed of a set of equally-sized cells that is overlaid across the territory. For European data, a 1 km² grid represents a good compromise between analytical capacity and data protection. There are three methodological solutions foreseen for establishing the total number of inhabitants living in each of these 1 km² grid cells.
Method 1: aggregation method
The aggregation method for producing population grid data is based on aggregating geocoded micro data (it is also called the ‘bottom-up approach’). This is the preferred method for producing population grid data, for example, aggregating a geocoded point-based data source, such as an address.
Population grid cells are referenced according to their coordinates which makes it possible to pinpoint them on a map. The example of Figure 3 shows a large number of point-based data plots — the individual blue circles — that have been overlaid onto a statistical grid; the plots represent the population at their usual place of residence.
In the example, the geocoded references in each grid cell relate to the lower left corner of each cell and provide information concerning their relative position in relation to the origin (as measured in a northerly and easterly direction): for example, cell B2 concerns grid reference N4626E5034, while the adjacent cell C2 is referenced as N4626E5035. These references indicate that grid cell N4626E5034 identifies a 1 km² cell that has coordinates for its lower left corner of Y=4 626 km (north), X=5 034 km (east) in relation to the origin (0 km, 0 km). Note that a new coding system will be introduced when the grid is revised as part of the 2021 census exercise.
This point-based information in each cell may be aggregated to produce a population count for each cell, as shown in Figure 4. For example, cell B2 has a total population count of 58 inhabitants, while in the adjacent cell (C2) the count is 52 inhabitants. The population density of each grid cell may subsequently be assigned to a population density class, as shown by the coloured shading in Figure 4: no inhabitants per km² (white background); 1-4 inhabitants per km² (light yellow shade); 5-19 inhabitants per km² (medium yellow shade); 20-199 inhabitants per km² (darker yellow shade).
Method 2: disaggregation method
In the absence of geocoded micro data, alternative approaches may be employed to produce data for the grid. The first of these approaches — the disaggregation method (also called the ‘top-down approach’) — uses population statistics for LAUs in combination with auxiliary spatial data. The total population count for an LAU may be disaggregated using data on land use and/or land cover to estimate the number of inhabitants that are living in each 1 km² grid cell; this may be done, for example, through the visual inspection of satellite images overlaid on the grid to determine if there are any buildings in each grid cell. Such an approach may have some limitations: for example, it is difficult to define the actual height of buildings from a satellite image and, as a consequence, the number of dwellings for each building’s footprint. This has been identified as a shortcoming, insofar as models using this approach tend to systematically underestimate the population living in densely populated areas (where there tend to be higher buildings that may be composed of multiple dwellings) while overestimating the population of thinly populated areas (that are generally characterised by one or two-storey dwellings). For more information, see:
- Spatially disaggregated population estimates in the absence of national population and housing census data, N. A. Wardrop, W. C. Jochem, T. J. Bird, H. R. Chamberlain, D. Clarke, D. Kerr, L. Bengtsson, S. Juran, V. Seaman, and A. J. Tatem;
- A high-resolution population grid map for Europe, F. Batista e Silva , J. Gallego and C. Lavalle;
- Spatial disaggregation of population data onto Urban Footprint data, S. M. Starmans;
- A volumetric approach to spatial population disaggregation using a raster build-up layer, land use/land cover databases (SIOSE) and LIDAR remote sensing data, F. J. Goerlich.
Method 3: hybrid method
The third and final method for producing population grid statistics is a hybrid method based on combining the aggregation and disaggregation techniques; this method provides a compromise between accuracy and the availability of data. Hybrid solutions may refer to using different source data to establish a geocoded framework, for example, combining geospatial, administrative and statistical sources. Note for the 2021 census exercise it is likely that the second and third methods will become largely obsolete, as data providers are generally expected to switch to aggregation methods (at least for total population measures).
GEOSTAT was launched by Eurostat in cooperation with the European Forum for Geography and Statistics (EFGS) in early 2010. It is a long-term programme designed to set up and promote the use of geospatial statistics including grid-based statistics through developing a methodology for official geospatial statistics in the EU, both for individual EU Member States and the EU as a whole. The initiative also aims to develop a set of common guidelines for the collection and production of population grid statistics.
In part to meet their requirements in relation to Regulation (EC) 763/2008 on the population and housing census, public administrations developed their geospatial statistics, collecting information either for LAUs or at an even more detailed level. Indeed, the census acted as a stimulus to trigger a range of initiatives for establishing geocoded building, address and population registers. As geocoded data sources with sufficient accuracy and reliability were made available, many of these were subsequently used as inputs for developing the GEOSTAT 2011 population grid as part of the GEOSTAT programme, which contains information for 29 EU and European Free Trade Association (EFTA) countries. As a result, the aggregation method was used to derive population grid statistics for 62 % of the census population in 2011, while the corresponding share for the GEOSTAT 2006 population grid was 30 % coverage. The geospatial framework for GEOSTAT 2011 is a standardised 1 km² grid that follows INSPIRE specifications and is based on the adoption of the ETRS89 Lambert Azimuthal Equal Area coordinate reference system. The grid is currently used in various statistical production processes, including grid-based typologies, local typologies and regional typologies. The underlying data can be downloaded and used free of charge for non-commercial purposes. While the GEOSTAT population grid covers continental Europe, as well as the Açores, Canarias and Madeira, it does not extend to the remaining outermost regions of the EU. The Joint Research Centre (JRC) of the European Commission’s produced a global population grid that was used to determine regional typologies for the remaining outermost regions.
One negative effect of developing grid-based statistics that have a much greater level of geographical detail is that there are increased concerns around data confidentiality and/or the risk of disclosure. Moreover, when introducing supplementary variables linked to the population (such as analyses by sex, by age or by type of housing) these issues may become even greater. The GEOSTAT 2011 population grid only contains information for the total number of inhabitants at their place of usual residence. This statistic was usually considered as non-sensitive by national statistical authorities which, as a result, did not apply any data protection methods for confidentiality issues; note however that national laws may require NSIs to protect the identification of individual citizens. For those countries that set confidentiality thresholds for GEOSTAT 2011 the minimum number of inhabitants per grid cells was 3 to 10 individuals; under this threshold the population count was suppressed.
The results from the GEOSTAT 2011 exercise are shown in Map 12; it presents the number of inhabitants per 1 km² grid cell and uses the same population density classes as Figure 4 in order to classify the grid cells from sparsely populated (a light orange shade) to densely populated (a dark orange shade).
Beyond GEOSTAT 2011
Broadening the methodology of the GEOSTAT 2011 population grid, GEOSTAT subsequently proposed a generic national (point-based) geocoding infrastructure for statistics (GEOSTAT2), building on registers for national addresses, buildings and/or dwellings. One of the principal drivers behind this initiative was the goal of delivering a fully geocoded population census in 2021 and the geocoding of other social and economic statistics using the census infrastructure as a geocoding frame.
In the next iteration of GEOSTAT, on-going work is focused on developing and testing a European version of the Global Statistical Geospatial Framework (GSGF) for the European statistical system (ESS). One of the principal drivers behind this initiative is the goal of delivering a fully geocoded population census in 2021. The GSGF is expected to provide a full methodology for the capture and production of harmonised European geospatial statistics and its full integration of geospatial information into statistical production processes. For more information, see this file.
Does the population grid change over time?
Population grids essentially remain stable over time unlike systems that are based on administrative boundaries (which generally change each year). For example, the population grid that is currently in use by Eurostat is GEOSTAT 2011, while the next major update of the grid is expected to relate to 2021 (when the next census takes place).
As there have been only two population grids produced at a pan-European level — GEOSTAT 2006 and GEOSTAT 2011 — it is only possible to compare these for an analysis of changes to the population grid. Note that some of the differences between the results for these two grids may be linked to changes in the production methodology and in particular the fact that there was less disaggregated information used in 2011 (compared with 2006). Furthermore, GEOSTAT 2006 covered the territory of 26 EU Member States (no information for Croatia or Cyprus) and the four EFTA countries and was based on a total population of 502 6 million inhabitants, while GEOSTAT 2011 was extended to include information for Croatia and Cyprus, with 514.9 million inhabitants spread across 1.95 million unique grid cells, while there were 2.47 million uninhabited grid cells.
For the GEOSTAT 2011 population grid, there were 56 208 grid cells with only one inhabitant while at the other end of the scale the highest number of inhabitants per 1 km² grid cell — some 53 119 people — was recorded for Barcelona (Spain). There were almost 466 million persons living in grid cells that were characterised by 150 or more inhabitants. As such, around 90 % of the population was living in approximately 10 % of the grid cells. There were 132 million people (or 26 % of the total population) living in the most densely populated areas of the covered countries, characterised by at least 5 000 inhabitants per km², these grid cells covered just 0.35 % of the total area of the population grid.
While the underlying building blocks — the GEOSTAT 2011 population grid — remain stable, it is important to note that this does not preclude changes to statistics that are based on the population grid. For example, each time there are changes to the boundaries of LAUs (usually an annual exercise) or to the boundaries of NUTS regions (usually every three years) then changes to these classifications should be reflected in the statistics that are produced for territorial typologies. Taking the example of the schematic overview defining urban areas in the EU (as shown in Figure 2 above), any modification to the boundaries of LAUs or regions would require the underlying information — that derived from grid-based data — to be reassessed in relation to such boundary changes.
Note Eurostat are discussing post-2021 census developments with national statistical authorities. It is possible that from the mid-2020s onwards, the ESS will agree to produce annual counts of populations (based on usual place of residence) for a 1 km² grid, with data to be made available within 12 months of the end of the reference period.
Which territorial typologies are impacted by changes to the population grid?
Those territorial typologies impacted by changes to the population grid can be identified by referring to Figure 1 above. It shows that population grid statistics for 1 km² cells are used directly to classify groups of grid cells into the following cluster types: urban clusters, urban centres and rural grid cells (see Chapter 1 for more details on cluster types).
This link to cluster types is particularly important insofar as these statistics are themselves used as building blocks for developing basic territorial typologies such as the degree of urbanisation (see Chapter 2) or the urban-rural typology (see Chapter 5). Information for urban clusters and urban centres is also used as a building block for developing urban typologies, with these statistics forming the basis — as already shown in Figure 2 — for data on cities and their commuter zones (see Chapter 3) as well as statistics on metropolitan regions (see Chapter 6).
Statistics Explained glossary entry: population grid cell
Detailed methodology: GEOSTAT 2011
Dedicated section: Gisco
Other information sources: European Forum for Geography and Statistics (EFGS) and INSPIRE — infrastructure for spatial information in Europe
Visualisation tools: Eurostat publishes data on the GEOSTAT 2011 population grid through the Statistical atlas (select Background maps and then GEOSTAT population grid, 2011).
Download data: the GEOSTAT datasets can be accessed and used for non-commercial purposes. The data are available on Eurostat’s website, here. GEOSTAT data are provided as shapefiles, geospatial vector data which is quasi-standard in the world of geographic information systems (GIS); indeed, almost any commercial or open source GIS software should be able to process shapefiles. Note that there are quite specific rules concerning the licensing conditions for these datasets that govern access, conditions and restrictions of use.
Local administrative units
Local administrative units (LAUs) are used to divide up the territory of the EU for the purpose of providing statistics at a local level. They are low level administrative divisions of a country below that of a province, region or state. Not all countries classify their locally governed areas in the same way and LAUs may refer to a range of different administrative units, including municipalities, communes, parishes or wards.
What are local administrative units?
Administrative divisions are generally the oldest nomenclatures for territorial units provided for by law, as they delineate local authorities with representative bodies. These administrative units are — depending upon the degree or centralisation/autonomy of political systems — charged with fulfilling the needs of local communities, for example, socioeconomic development, spatial planning, utilities, culture or the environment.
Such administrative divisions usually exist at different hierarchical levels (although some levels may not exist in smaller EU Member States), ranging from regional and/or county/state administrations, through districts and/or local councils, down to municipalities/communes. It is this collection of units at the bottom of the administrative hierarchy that is used to define LAUs. LAUs implement policies and are considered as appropriate building blocks for constructing local level typologies, such as statistics for the degree of urbanisation, functional urban areas and coastal areas.
Up until 2016, there were two different levels of LAU:
- LAU level 1 (formerly NUTS level 4) which was defined for most, but not all of the EU Member States;
- LAU level 2 (formerly NUTS level 5) which consisted of municipalities/communes or equivalent units across all EU Member States.
Since 2017, only one level of LAU has been kept. It is important to note that existing administrative units within the EU Member States constitute the first criterion used to define LAUs. This means there are considerable differences across the EU Member States between the naming conventions and concepts used. Indeed, Regulation (EC) No 1059/2003 on the establishment of a common classification of territorial units for statistics (NUTS) defines an administrative unit as: a geographical area with an administrative authority that has the power to take administrative or policy decisions for that area within the legal and institutional framework of the Member State. The distinct administrative divisions used for each Member State are defined within Annex III.
|Bulgaria||Населени места (Naseleni mesta)|
|Ireland||Counties, County boroughs|
|Cyprus||Δήμοι, κοινότητες (Dimoi, koinotites)|
|Latvia||Republikas pilsētas, novadi|
|Romania||Municipii, Orașe and Comune|
To give an idea of the variations that may exist between these national concepts for LAUs, in 2016 there were 35 442 LAUs identified in France, compared with 11 135 in Germany, 7 983 in Italy and just 415 in the United Kingdom; these differences reflect, to a large degree, the organisation of local government/representation in each EU Member State.
Furthermore, there are often sizeable differences between LAUs within the same EU Member State, for example, in terms of their number of inhabitants or the area that they cover. In 2016, there was a single person living in the French commune of Rochefourchat in the south-east of France, while there were as many as 2.2 million inhabitants living within the commune of Paris. In the United Kingdom, the boundaries of the City of London delineated an area of just 3.15 km² which can be contrasted against an area of 7 763 km² for Caithness & Sutherland (in the north of Scotland).
Encoding administrative divisions within the national territory is an essential task of national statistical systems, assigning an alphanumerical code to the various levels. This makes it easier for national statistical authorities to provide a wide range of subnational statistics, often at a highly disaggregated level of detail, in an attempt to meet the growing need for socioeconomic information at a local level. As well as being a basis for statistical analysis in their own right, LAUs are also used as one of the principal building blocks to produce data for regions and for other territorial typologies.
Why does the list of local administrative units change?
To meet the increased demand for statistics at a local level, Eurostat maintains a list of LAUs. This tracks any changes that take place: some EU Member States make frequent changes to their LAUs while others almost never change them. Article 4 of the NUTS Regulation (EC) No 1059/2003 provides details of how this system should be managed:
- during the first six months of each calendar year, the Member States should provide details of any changes to local administrative units with reference to 31 December of the previous year;
- Eurostat is responsible for amending, on an annual basis, the complete list of LAUs on the basis of changes to administrative units that have been communicated to it by the Member States;
- Eurostat should publish the revised list of LAUs by the end of each calendar year.
Guidelines for developing LAU lists
- An LAU code is the key used for correspondence with all related territorial typologies.
- There are no coding conventions at an EU level for LAUs, national codes are employed.
- In the case of LAU closures, old codes must not be re-used or maintained.
- LAU codes should be provided as alphanumeric strings rather than numbers.
To amend the LAU list on an annual basis, Eurostat introduced a new transmission format in 2018. As well as detailing changes to LAUs it also seeks to integrate information on changes to territorial typologies for LAU-based classifications — the degree of urbanisation, functional urban areas and coastal areas — which are directly impacted by any LAU boundary changes. As such, the LAU list is managed together with local typologies in order to align correspondence tables at the same time. This single procedure makes it possible for Eurostat to publish annual updates for territorial typologies together with the annual LAU list at the end of each year.
Which typologies are impacted by changes to LAUs?
At a local level (for LAUs), the following typologies are based on LAUs:
- degree of urbanisation — cities, towns and suburbs, rural areas (see Chapter 2);
- functional urban areas — cities and their surrounding commuting zones (see Chapter 3);
- coastal and non-coastal areas (see Chapter 4).
As such, any changes made to LAU boundaries need to be checked to see if they impact on these local level typologies and, where necessary, the correspondence tables for these typologies should be updated. Concerning the datasets compiled using these, any modifications to the typologies caused by changes to the list of LAUs can be implemented in one of two different ways: applying the specific methodology for each data collection to the new LAU boundaries; or applying a simpler approach that does not use geographical information systems to estimate the resulting statistics based on changes to LAU boundaries. The first approach is more labour intensive, while the second is particularly suitable if boundary changes for LAUs are relatively small or consist principally of merging LAUs. A practical example of how this may be done is presented for the degree of urbanisation in Chapter 2 (under the heading, Changes over time that impact on the classification).
As well as providing basic statistics in their own right (for population and area) and serving as the building block for local level typologies, LAUs are also used as a building blocks for regions, as described in Article 4 of the NUTS Regulation: In each Member State, local administrative units (LAU) shall subdivide NUTS level 3 into one or two further levels of territorial units. Map 13 provides an example for part of Czechia showing how its LAUs (identified by the grey borders) are aggregated to NUTS level 3 regions (identified by the light blue borders); note the interesting case of the capital city of Praha that is both an LAU and a NUTS level 3 (and indeed a NUTS level 2) region.
Statistics Explained glossary entry: Local administrative unit (LAU)
Dedicated section: Local administrative units (LAUs)
Download data: the NUTS Regulation requires EU Member States to send lists of their LAUs to Eurostat each year. This information may be supplemented by additional administrative data for the population and the total area of each LAU. These lists are published each year on the Eurostat website.
NUTS, the classification of territorial units for statistics, is a geographical classification subdividing the territory of the EU into regions at three different levels — NUTS level 1, level 2 and level 3 (moving from larger to smaller territorial units). The legal basis for NUTS is provided for Regulation (EC) No 1059/2003, hereafter referred to as the NUTS Regulation. A consolidated version (including subsequent amendments) is available here.
What is the NUTS classification?
The NUTS classification is a hierarchical system for dividing up the territory of the EU for the purpose of:
- the collection, development and harmonisation of EU regional statistics;
- socioeconomic analyses of the regions;
- NUTS level 1: major socioeconomic regions;
- NUTS level 2: basic regions for the application of regional policies;
- NUTS level 3: small regions for specific diagnoses;
- framing EU regional policies;
The NUTS classification is set out in Annex I of Regulation (EC) No 1059/2003. It is a hierarchical classification that ascribes a specific code and name to each territorial unit and subdivides the EU Member States into NUTS level 1 territorial units, each of which is subdivided into NUTS level 2 territorial units, these in turn each being subdivided into NUTS level 3 territorial units. Note that a particular territorial unit may be classified at several NUTS levels — for example, the German capital city of Berlin is coded as DE3 (NUTS 1), DE30 (NUTS 2) and DE300 (NUTS 3), all of which cover the same area.
The diagram below shows the hierarchical structure of NUTS, moving from the national territory of Germany (DE) through progressively more detailed levels of NUTS. At NUTS level 1, the German regions are aligned with the Länder, for example, Baden-Württemberg (DE1) and Bayern (DE2). Each NUTS level 1 region is subsequently subdivided into NUTS level 2 regions, for example, Bayern is split into Oberbayern (DE21), Niederbayern (DE22), Oberpfalz (DE23), Oberfranken (DE24), Mittelfranken (DE25), Unterfranken (DE26) and Schwaben (DE27). In a similar vein, NUTS level 2 regions may be subdivided into the most disaggregated regional units, as defined by NUTS level 3, for example, some of the 14 different level 3 regions within Schwaben include Ostallgäu (DE27B), Unterallgäu (DE27C) and Oberallgäu (DE27C).
The current version of the NUTS classification is NUTS 2016. It covers 104 regions at NUTS level 1, 281 regions at NUTS level 2 and 1 348 regions at NUTS level 3. The amendment introducing NUTS 2016 came into force on 19 December 2016 and applies to the transmission of data (to Eurostat) as of 1 January 2018 onwards.
Note: the NUTS classification is defined only for the EU Member States. Eurostat, in agreement with the countries concerned, also has a coding of statistical regions for countries that do not belong to the EU, but which are:
- candidate countries awaiting accession to the EU; or
- potential candidates; or
- EFTA countries.
History of the NUTS classification
At the beginning of the 1970’s, Eurostat set-up the NUTS classification as a single, coherent system for dividing up the EU’s territory in order to produce regional statistics. For around 30 years, the implementation and updating of the NUTS classification was managed under a series of “gentlemen’s agreements” between the EU Member States and Eurostat. Work on Regulation (EC) No 1059/2003, to give NUTS a legal status, started in spring 2000; it was adopted in May 2003 and entered into force in July 2003.
The NUTS Regulation specifies that there should be stability in the classification to ensure that data refers to the same regional unit (considered crucial for time series statistics). However, sometimes national interests require changes to the classification of a territory and when this happens, the EU Member State concerned informs the European Commission about the changes. The European Commission, in turn, amends the classification at the end of each predefined period of stability.
Principles and characteristics used in the NUTS classification
The development of the NUTS classification is based on three underlying principles.
Principle 1: population thresholds
The NUTS Regulation defines minimum and maximum population thresholds for the size of NUTS regions (see Table 1); for the purpose of the Regulation, the population of each territorial unit consists of those persons who have their usual place of residence in the area concerned.
If the total population of an EU Member State is below the minimum threshold for a given NUTS level, then the whole of that Member State shall be covered by a single territorial unit for the level in question. For example, Cyprus and Luxembourg are both covered by single territorial units at each NUTS level (1, 2 and 3).
For NUTS regions that are based on administrative levels, it is sufficient if the average size of the corresponding regions lies within the thresholds; in case of regions not based on administrative levels, each individual region should do so. However, exceptions do exist in case of geographical, socioeconomic, historical, cultural or environmental circumstances. For example, in 2016 the population of 30 NUTS level 1 regions was below the minimum threshold, including, among others: seven EU Member States, eight German Länder, the Portuguese autonomous regions of Madeira and Açores, the Spanish and Finnish island regions of Canarias and Åland, as well as the French Départements d’outre-mer. As a result, despite the aim of ensuring that regions of comparable size all appear at the same NUTS level, each level may contain regions which differ greatly in terms of their total population size.
Principle 2: NUTS favours administrative divisions
As noted above, for practical reasons the NUTS classification generally mirrors the territorial administrative divisions of each EU Member State. In doing so, this supports the availability of data and the capacity to implement policy developments.
Principle 3: regular and extraordinary amendments
The NUTS classification can be amended: the Regulation specifies under regular circumstances the classification should remain (unchanged) for a period of at least three years. Note however that additional amendments to the NUTS classification may take place for exceptional circumstances, for example, when new Member States join the EU, or if there is a substantial reorganisation of the administrative structure of an EU Member State; at the time of writing this has only happened once, in 2014 for Portugal. In the case of either regular or extraordinary amendments to the NUTS classification, the Member State concerned should replace its historical data by time series according to their new regional classification within a period of two years.
Which typologies are impacted by changes to NUTS?
As such, any changes made to NUTS level 3 boundaries need to be checked to see if they impact on the regional typologies (applying again any rules for determining classifications to the new NUTS boundaries), with updates to the NUTS classification reflected in correspondence tables for each of the regional typologies.
At a regional level (for NUTS level 3 regions), the following typologies have a legal basis:
- the urban-rural typology — predominantly urban regions, intermediate regions, rural regions (see Chapter 5);
- the metropolitan typology — metropolitan regions and non- metropolitan regions (see Chapter 6);
- the coastal typology — coastal and non-coastal regions (see Chapter 7).
Note there are three additional regional typologies (also based on NUTS level 3 regions) for which there is (currently) no legal basis:
- the border typology — border and non-border regions (see Chapter 8);
- the island typology — island and non-island regions (see Chapter 9);
- the mountain typology — mountain and non-mountain regions (see Chapter 10).
Statistics Explained glossary entry: NUTS
Legislation: on the Eurostat website
Dedicated section: NUTS
Visualisation tools: Eurostat publishes data in the form of maps that are based on NUTS regions through Regions and cities illustrated (RCI). The different levels of the NUTS classification may be viewed through the Statistical atlas; the example below shows NUTS level 3 regions principally in Germany.
Maps: in *.PDF format presenting the different NUTS levels are available on Eurostat’s website.
Database: Eurostat’s website provides regional statistics by NUTS for 16 separate domains covering a wide range of socioeconomic data. These statistics are available for the following areas — demography, education, health, the labour market, labour costs, poverty and social exclusion, crime, economic accounts, structural business statistics, business demography, tourism, the digital economy and society, science and technology, transport, agriculture, the environment and energy. The data may be found: here.
Using data on territorial typologies
Eurostat publishes EU statistics at a regional level for many statistical domains: these statistics are widely used in the context of EU regional policy. Through the Tercet initiative, the European Commission has defined territorial typologies in cooperation with the OECD, establishing legal recognition for these typologies by integrating them into the NUTS Regulation and its implementing provisions, thereby promoting a set of harmonised definitions that are based on methodological transparency, core definitions, and established criteria for creating and updating each typology (as required). The Tercet initiative therefore aims to improve the comparability and stability of these territorial typologies and has been designed to impact on the compilation and dissemination of EU subnational statistics. In turn, this has made it possible for those developing thematic statistical and policy-based regulations to refer directly to the territorial typologies when they instigate new areas for collecting or analysing subnational statistics.
The integration of a broad range of territorial typologies into the NUTS Regulation in December 2017, underlines the importance of subnational statistics as an instrument for targeted policymaking and a tool for understanding and quantifying the impact of policy decisions for specific types of territories. As shown in Maps 1-8 these cover several different territorial typologies for which data are now available across the EU at different levels — grid typologies, local typologies and regional typologies. The availability of these typologies and related data have in turn stimulated policymakers to ask questions such as: does it make sense to have the same policy target for pollution in a city centre as in an area of natural beauty? or does it make sense to have the same policy target for educational attainment in a capital city as in a remote, sparsely-populated rural area?
Analyses such as these have led to a territorial dimension being introduced into a range of EU policy areas and their related statistics. Grouping different types of regions and/or areas according to territorial types can help in understanding common patterns, for example, urban areas/regions generally perform better in economic terms and may act as hubs for innovation and education; at the same time, they may also be characterised by a range of different challenges such as congestion, pollution or housing problems.
Indeed, while some of the most pressing challenges facing the EU — for example, globalisation, climate change or poverty and social exclusion — have traditionally been approached through broad sectoral policies, often implemented across the EU, policymakers have more recently analysed spatial developments for these challenges at a much more disaggregated level of detail between different types of territory both within and across EU Member States; more details are provided below.
The EU’s cohesion policy invests in measures to support growth and jobs and promotes territorial cooperation; it is behind thousands of projects that have taken place all over the EU. It aims to reduce the disparities that exist between EU regions, promoting a balanced and sustainable pattern of territorial development, by supporting job creation, business competitiveness, economic growth, sustainable development, and an overall improvement in the quality of life. The bulk of cohesion policy funding is concentrated on less developed EU regions in order to help them to catch-up with other regions and to reduce the economic, social and territorial disparities that exist across the EU.
The EU’s cohesion policy is established on the basis of seven-year funding periods. The current period covers 2014-2020, for which expenditure of EUR 352 billion has been allocated for measures in the EU Member States, equivalent to almost one third of the total EU budget.
Cohesion policy is delivered through three main funds (the European Regional Development Fund (ERDF), the Cohesion Fund and the European Social Fund (ESF): the NUTS classification defines the regional boundaries that are used to determine geographic eligibility for two of these funds. For the programming period 2014-2020, eligibility for the European Regional Development Fund (ERDF) and the European Social Fund (ESF) was calculated on the basis of regional GDP per inhabitant (in PPS and averaged over the period 2007-2009), with NUTS level 2 regions ranked and split into three groups:
- less developed regions (where GDP per inhabitant was less than 75 % of the EU-27 average);
- transition regions (where GDP per inhabitant was between 75 % and 90 % of the EU-27 average); and
- more developed regions (where GDP per inhabitant was more than 90 % of the EU-27 average).
The European Commission’s cohesion policy for 2014-2020 emphasised territorial development strategies focusing on urban, rural and coastal areas. The principles for cohesion policy were set out in a common strategic framework (Regulation (EU) No 1303/2013) stressing that the promotion of smart, sustainable and inclusive growth must reflect the role of cities, urban, rural and coastal areas and take urban-rural linkages into account. An early example of this approach is the use that was made of the degree of urbanisation typology in Regulation (EU) No 522/2014 to define eligibility for ERDF support to carry out innovative actions in cities or in towns and suburbs.
The Europe 2020 strategy is the EU’s agenda for growth and jobs: it emphasises smart, sustainable and inclusive growth as a way to overcome the structural weaknesses in the EU’s economy, improving its competitiveness and productivity and underpinning a sustainable social market economy. As the period covered by the strategy (2010-2020) passed, there was a switch in policy focus towards a more integrated territorial approach that sought to understand more clearly the uneven socioeconomic developments experienced both within and across EU Member States, for example, differences between urban and rural areas or differences between capital city metropolitan regions and smaller metropolitan regions.
Although the Europe 2020 strategy does not specifically refer to regional policy, the European Commission has underlined that it may be neither realistic nor desirable that all regions seek to attain the same national targets. Rather, it was considered important for the EU Member States to take account of their different needs and to draw up national and regional programmes that reflect local specificities so as to promote smart, sustainable and inclusive growth. Highlighting these regional and territorial aspects, there have been a number of calls to align regional funding more closely with the Europe 2020 strategy and to monitor in more detail the performance of EU regions with respect to Europe 2020 targets. This approach was also supported by the findings of the mid-term review of the Europe 2020 strategy, which noted that there was growing evidence of regional divergence in several EU Member States. More practically, the Directorate-General for Regional and Urban Policy has increased its efforts to align the various dimensions of regional funding more closely to the Europe 2020 targets.
Sustainable development goals
Sustainable development may be defined as economic growth and social progress that meets the needs of present generations without jeopardising future generations. It provides a comprehensive approach bringing together economic, social and environmental considerations in ways that mutually reinforce each other.
The United Nations (UN’s) 2030 Agenda for Sustainable Development, adopted by world leaders in 2015, represents a global sustainable development framework based around 17 Sustainable Development Goals (SDGs) and 169 specific targets. It is a commitment to eradicate poverty and achieve sustainable development by 2030 worldwide, ensuring that no one is left behind.
European policymakers recognise that coherent and integrated regional policy should form an essential part of the EU’s implementation strategy for the 2030 Agenda, whereby SDG indicators have to capture problems at a scale where they occur (the regional, sub-regional and city-level). The EU is fully committed to be at the forefront of implementing the UN’s 2030 Agenda. In November 2016, the European Commission outlined its strategic approach in a Communication, Next steps for a sustainable European future: European action for sustainability (COM(2016) 739 final).
The EU’s Urban Agenda
The EU’s Urban Agenda is an integrated and coordinated approach designed to deal with the urban dimension of EU and national policies. By focusing on concrete issues through dedicated partnerships, the Urban Agenda seeks to improve the quality of life in urban areas. In 2016, EU ministers responsible for urban matters agreed the Pact of Amsterdam which underlies the Urban Agenda. It is based on the principles of subsidiarity and proportionality, focusing on three key pillars of EU policymaking: better regulation, better funding and better knowledge.
Through a series of dedicated partnerships which involve — on a voluntary and equal basis — cities, EU Member States, the European Commission and stakeholders such as businesses or non-governmental organisations (NGOs), work programmes and actions are designed to successfully tackle the principal challenges that are facing cities as well as contributing towards smart, sustainable and inclusive growth. For more information, see the website of the European Commission.
Urban development in the EU
The various dimensions of urban life — economic, social, cultural and environmental — are closely inter-related. Successful urban developments are often based on coordinated/integrated approaches that seek to balance these dimensions through a range of policy measures such as increasing education opportunities, urban renewal, preventing crime, encouraging social inclusion or encouraging environmental protection. As such, urban development policy has the potential to play an important role in promoting the Europe 2020 strategy and delivering smart, sustainable and inclusive growth.
One important change in European policymaking for the 2014-2020 funding period is recognition of the important role that may be played by the urban dimension of regional policy, in particular concerning measures that are designed to assist in the fight against poverty and social exclusion. Indeed, the EU has put the urban dimension at the heart of cohesion policy, with at least half of the resources foreseen under the ERDF being invested in urban areas. The European Commission estimates that during this six-year period some EUR 10 billion from the ERDF will be allocated to sustainable urban development, covering around 750 different cities.
A number of commentators and stakeholders have argued that cities need to be more involved in the conception and implementation of EU policies, as, despite their economic weight, there is no explicit urban dimension to the Europe 2020 strategy or its targets, although three flagship projects — the digital agenda, the innovation union and youth on the move — address particular urban challenges.
Rural development in the EU
There are also considerable differences between EU Member States as regards their urban-rural territorial divisions. Some Member States — for example, Ireland, Sweden or Finland — are very rural in character. By contrast, the Benelux Member States and Malta have a high degree of urbanisation. Equally, within individual Member States there can be a wide range of different typologies, for example, the densely-populated, urbanised areas of Nordrhein-Westfalen in western Germany may be contrasted with the sparsely-populated, largely rural areas of Brandenburg or Mecklenburg-Vorpommern in north-eastern Germany.
EU rural development policy is designed to help rural areas and regions meet a wide range of economic, social and environmental challenges; it complements the system of direct payments to farmers and measures to manage agricultural markets. The European agricultural fund for rural development (EAFRD) provides finance for the EU’s rural development policy, which is used to promote sustainable rural development and to contribute towards the goals of the Europe 2020 strategy for smart, sustainable and inclusive growth.
For the period 2014-2020, the EAFRD has been allocated EUR 99.6 billion. The EAFRD is intended to help develop farming and rural areas, by providing a competitive and innovative stimulus, at the same time as seeking to protect biodiversity and the natural environment. As with other structural and investment funds, from 2014 onwards, rural development policy is based on the development of multiannual partnership and operational programmes which are designed at a national/regional level by individual EU Member States.
- Regulation (EC) No 1059/2003 - Regulation of the European Parliament and of the Council of 26 May 2003 on the establishment of a common classification of territorial units for statistics (NUTS)
- Regulation (EU) 2017/2391 - amending Regulation (EC) No 1059/2003 as regards the territorial typologies (Tercet)
- Regulation (EC) 763/2008 - Regulation of the European Parliament and of the Council on population and housing censuses
- Regulation (EU) No 1303/2013 - Regulation of the European Parliament and of the Council laying down common provisions on the European Regional Development Fund, the European Social Fund, the Cohesion Fund, the European Agricultural Fund for Rural Development and the European Maritime and Fisheries Fund and laying down general provisions on the European Regional Development Fund, the European Social Fund, the Cohesion Fund and the European Maritime and Fisheries Fund and repealing Council Regulation (EC) No 1083/2006
- Regulation (EU) No 522/2014 - Commission Delegated Regulation supplementing Regulation (EU) No 1301/2013 of the European Parliament and of the Council with regard to the detailed rules concerning the principles for the selection and management of innovative actions in the area of sustainable urban development to be supported by the European Regional Development Fund