Lewis Dijkstra (p)

Michiel Nicolaas Daams

This article presents a new methodology, the degree of urbanisation, and its application to the globe. The degree of urbanisation relies on a population grid to classify local units into three classes: cities, towns & suburbs, and rural areas. These three classes can be further disaggregated into cities, towns, suburbs, villages, dispersed rural areas and mostly uninhabited areas. The population share in rural areas as defined by the degree of urbanisation is similar to the share reported based on national definitions in most countries in the Americas, Europe and Oceania, but radically different in many African and Asian countries. A possible explanation for these differences is that in Africa and Asia smaller settlements are considered rural, while they are classified as urban in the rest of the world. An analysis of the national definitions reported to the United Nations and attempts to replicate these nationally defined shares using density and population size criteria indicate that differences in national definitions (in theory and in practice) make them unsuitable for international comparisons. The global share of population in cities over 300,000 inhabitants, however, is identical between the national definitions and this new method. This implies there is a broader consensus on what is a city than on what is urban and rural. This new definition has been applied to two new global population grids with different methods. In countries where population data is available for very small spatial units, the results are almost identical. In countries where population data is only available for very large units, the results vary substantially. The discrepancies are particularly wide in several African countries. Nevertheless, both grids classify a substantially larger share of the global population as living in urban areas (defined as settlements of 5,000 inhabitants or more) than what is reported by the UN based on heterogeneous national definitions and conventions.

Jean-Philippe Aurambout (p), Andrius Kučas , Carlo Lavalle

Paolo Veneri

The degree of urbanization (DEGURBA) is a methodology proposed by the European Commission (adopted by Eurostat) which combines 1 km resolution grids of population and built-up to classify human settlements. A recently proposed refined methodology now combines several population and density thresholds to identify six classes of human settlements: Cities, Towns, Suburbs, Villages, Dispersed rural areas and Mostly uninhabited areas.
This paper describes the results of spatial analysis, conducted at the European Union (EU28) level, using the refined DEGURBA for a set of 11 territorial indicators computed by the LUISA Territorial Modelling Platform and the European Environmental Agency (EEA). These indicators took into consideration elements such as population, roads and accessibility, green infrastructure as well as landscape morphology. The analysis revealed that the refined DEGURBA provided, at country level, values significantly different for the majority of variables considered, confirming they indeed characterize different types of territorial developments. We then provided a detailed characterisation of European member states “through the lens” of the level 2 DEGURBA.
The refined DEGURBA indicates that 37% of the EU28 population lives in cities, 21% in Towns, 12% in Suburbs, 12% in Villages, 14% in Dispersed rural areas and 4% in Mostly unhabituated areas. Urban areas concentrate 72% of the population over only 3% of the EU territory but contain 17% of all roads (uniformly distributed between cities, Towns and Suburbs). In addition, the refined DEGURBA identified Slovakia and Slovenia as having the lowest proportion of population in cities (14%). Belgium presented the highest proportion of it population living in Suburbs (28%) while Slovenia had the highest proportion of population in Dispersed rural areas (33%) and Ireland had the highest proportion of population in Mostly uninhabited areas (19%). From this new classification we found that across Europe locations identified as Suburbs differed from Towns in that, they present lower population density, higher UGI per hectare (ha), lower percentage built-up areas, shorter distance to access sub-regional, regional but not to local services, more road and public transport stops per capita. Location classified as Villages differ from dispersed rural areas in that they presented a much higher density and percentage built-up area slightly better access to services, lower length of road per capita and more public transport stops per ha.
These first insights indicate that the refined DEGURBA provides a valuable new lens through which we can better understand urbanity in Europe and globally.

Ana Moreno Monroy, Paolo Veneri (p), Marcello Schiavina

Jean-Philippe Aurambout

This paper proposes a method to define the boundaries of urban economic agglomerations – or functional urban areas (FUAs) – in the whole world and assesses patterns of population growth within such agglomerations. The method builds on the EC-OECD definition of FUAs, which are composed of urban centres least 50 thousand people surrounded by interconnected commuting zones. The estimation of commuting zone borders relies on the use of population gridded data with regular cells of one-km2 across the globe. Commuting zones surrounding urban centres are identified by relying on the estimated probability that a one square-kilometre cell with at least 300 inhabitants is part of a FUA. Such probabilities are the outcome of a logit model that relies on characteristics at both cell and country level. For this classification problem, a set of about one thousand FUAs in 31 countries is used to construct a training set of about half a million points. Results show that around 54% of the world population live in urban agglomerations, out of which 12% live in their commuting zones. Both these shares increase with level of development, with the proportion of suburban population in developed countries disproportionally overcoming that of the rest of the world. Since 1990, the increase in population in commuting zones as a percentage of the total FUAs’ population has been fastest in Europe and Latin America.

Michiel Nicolaas Daams (p), Philip McCann, Paolo Veneri, Dennis Schoenmaker, Richard Barkham

Laura de Domonicis

This paper measures potential agglomeration economies within 17 US cities. Our analysis extends other studies, which tend to consider differentials in productivity or wages at the level of cities, by adopting an intra-metropolitan scale. Intra-metropolitan areas that potentially harbor agglomeration economies are scrutinized based on fine-scale densities of economic production and real estate investments. Indeed, urban economic theory suggest that if firms gain productivity from density this is reflected in land values – and, by consequence, in property yields, reflecting the relative price at which investors buy commercial buildings and associated rental income flows. To classify areas where agglomeration economies may be present, we apply a sophisticated clustering model on the locations of investment-grade properties, offices, that transacted during 2003-2015 at a combined value of one trillion US dollars. The resultant clusters are jointly analyzed with information on local employment density. Regression analysis is then used to estimate whether yields vary between office buildings in buoyant or less buoyant clusters and lesser agglomerated city subareas. If so, this may provide evidence of local agglomeration economies. This study’s findings combines thinking in urban economics and real estate economics, and special attention is paid to the consistency of our comparative analysis across cities of different shape and density. Our findings may be relevant to policy makers and market actors who seek to identify the economic development potential of cities.