Filippo Celata1, Cary Y. Hendrickson1, 4, Federico Martellozzo2, Martin Rossi1, Luca Simeone3
1University of Rome La Sapienza, 2University of Firenze, 3Aalborg University Copenhagen, 4Food and Agriculture Organization of the United Nations
The tendency of migrants to concentrate spatially, particularly in cities, is a recurrent outcome of migration. A neighbourhood named “Chinatown” was first founded in San Francisco in 1840, and in many other cities worldwide afterward. Then came “little Italy”, “little Ireland”, “little Germany”, an so on. What we define as immigrant urban “hot-spots” – and others define as immigrant or “ethnic” neighbourhoods, or “ethnic ecnclaves” – are indeed crucial viewpoints to observe the patterns of integration or non-integration of foreigners in the host countries, as well as to understand more generally how place, culture and social capital influence social behaviour. As such, the study of ethnic neighbourhoods or enclaves has provoked a lively debate.
 For a review, see: Celata F. (2015) Immigrant Enclaves. In Wherry F.F. and Golson, J.G. (eds), Encyclopedia of Economics and Society, London: SAGE, 911-913.
The spatial concentration of foreigners is associated with the tendency of migrants to create social and economic spaces that are geographically and functionally isolated. This tendency have been observed since the beginning of contemporary migration and may be the result of spatial segregation or social exclusion. At the same time, foreigners may concentrate because they benefit from proximity with members of the same community: ethnic neighbourhoods, according to this view, facilitate the access of migrants to socioeconomic networks bounded by co-ethnicity and to a variety of “ethnic resources” such as interpersonal relationships, human capital externalities, and community-based associations that facilitate access to information, labor and credit, while fostering intracommunity solidarity. Immigrant hot-spots may therefore be interpreted as enclosures, socio-spatial traps or “ghettos” that prevent the assimilation of foreigners, or as places where social capital and solidarity are facilitated, leading to better socioeconomic well being and even to an easier – although slower – integration of migrants. Both outcomes are indeed possible. The issue should be therefore the object of carefull and case by case examination.
In this frame, the identification of immigrant hot-spots, and the availability of comparable information about them, is still largely an open issue, but also a crucial prerequisite to any further observation, analysis or intervention. Difficulties in this regard are related to the availability of accurate and detailed information about the spatial distribution of foreigners at the geographical scale that is sufficient to identify the specific local areas where immigrants concentrate, and other recurrent problems in the analysis of spatial data such as the modifiable area unit problem. To this end, we used a dataset produced by the Italian Joint Research Centre of the European Union, and derived from a spatial disaggregation of statistics of the 2011 Census, collected from National Statistical Institutes. The results of the spatial processing of the original data is a uniform grid showing the concentration of migrants in cells of 100 by 100 meters in all cities of seven EU Member States – Spain, Germany, Italy, Netherlands, UK, Ireland, France. The dataset permitted to solve most of the above mentioned problems and to allow the full comparability of results, in order to facilitate further investigation and policy interventions, as well to provide some interesting new evidence.
Immigrant hot-spots are mostly associated to specific neighbourhoods within cities. In this research, we defined immigrant hot-spots as those neighbourhoods with the highest concentration of foreigners. Comparable statistical data about urban neighbourhoods are basically non-existent, insofar as even a common definition of what a neighbourhood is, and how it should be delimited, is lacking, also within individual countries. A partition of cities into neighbourhoods may be provided by some municipalities, especially in big cities, while statistical data about them may be obtained by aggregating census tracks. These partitions may be, consequently, very different in terms of shape and extent, both within and between cities; their geometry may not perfectly match that of census tracks – which may also vary from one census to another; and the data provided about them is often poor. The dataset we used provided a unique opportunity to avoid the above mentioned difficulties and to obtain solid, rich, and homogenous data.
The methods we adopted were intended, as already mentioned, to guarantee the full comparability of results. To this end, the 100×100 meters grid was aggregated into square kilometers: an extent that was supposed to correspond to that of a neighbourhood, and that was chosen also for the sake of clarity and simplicity. Furthermore, we decided to not overimpose above the 100×100 meters grid a predefined square kilometers partition but, instead, we considered any potential alternative aggregation of 100×100 cells in order to identify those square kilometers that, in any local area, show the highest concentration of foreigners. This method and this partition was intended, in other words, to avoid the so called “modifiable area unit problem” (MAUP): a common issue in the analysis of spatial data. The MAUP is due to the fact that any predefined partition or geographical discontinuity is artificial, more or the less arbitrary, and modifiable. Such partition, however, heavily influences the result of the analysis and, consequently, the interpretation of results. The MAUP can be first due to a “scale problem”, ie. to the spatial resolution of data. For example, statistical relations are stronger the lower the degree of spatial resolution because variance is lower: the more we aggregate data, consequently, the stronger they correlate. Secondly, the MAUP can be due to a “zoning problem”, which is related to the specific geometry of the partition: for any given number of zones, even if they pertain to the same geographical scale, results are influenced by their shape: as this shape changes, results may vary dramatically. In our case, the square kilometer grid permitted to aggregate data according to an homogenous scale, and to adopt a partition which is uniform and regular. In order to avoid the MAUP completely, we tested all the alternative 100×100 cells aggregations and ranked the results in order to identify and to delimit, for each country, those non-overlapping square kilometers with the highest number of foreigners. To this end we converted the grid data into a raster geodataset and ran a focal statistic analysis to attribute to any 100×100 meters cell the sum of the foreign population in the surrounding 100 cells.
The analysis we conducted for each of those zones is, at this stage, exploratory and descriptive. We first attributed a name to the identified zones using those partitions of cities into neighbourhoods that we mentioned above, and produced a basemap using satellite imagery provided by Esri and Open Street Map geodata. We extracted from the D4I dataset the composition of each zone in terms of nationality of the population. Some preliminary results are illustrated below.
The main immigrant hotspots in Europe
Below we show the square kilometre that, in each of the countries we had data about, hosts the highest number of foreigners.