8 minute read

By Lukas Kleine-Rueschkamp (Lukas.KLEINE-RUESCHKAMP@oecd.org) & Cem Özgüzel (Cem.OZGUZEL@oecd.org), Centre for Entrepreneurship, SMEs, Regions and Cities (OECD)

Migration has risen to the top of the policy agenda in recent years, and not just at the national or international level. Many regions in OECD countries face major demographic change and a declining supply of labour as their population age. To alleviate these challenges and reap the benefits of migration for local and regional development, the integration of migrants is crucial, making it one of the most pressing policy challenges in OECD countries. While the patterns of migration and the size of migrant communities differ from country to country, subnational differences within countries also tend to be significant.

To support effective migration policy design, policymakers need to understand the different types of challenges and difficulties migrants face. Foremost, this requires comprehensive and detailed data on migrants, in particular their geographic distribution across different regions and cities within OECD countries. As documented by previous research, the foreign-born population (i.e., migrant population) differs from the native-born population in terms of where they choose to live, with migrants being more geographically concentrated in specific areas.

Existing subnational datasets on migrants and previous OECD analysis are limited to large administrative regions such as states in the US or federal states in Germany, called TL2 regions. ^{1 2} Such data may provide only crude information on migration as it could miss important differences within those regions, especially in very large and populous regions. For example, some regions in OECD countries such as Germany and the US have populations of 10 million inhabitants or more, and extend for thousands of square kilometres. As a result, data for such large territories and populations often obscure potentially interesting and meaningful intra-regional discrepancies. Another drawback of such regional data is that it lacks the granularity to examine important migration trends in rural, urban or metropolitan areas across the OECD.

Constructing the OECD Municipal Migration Database (MMD)

As part of the OECD project The Contribution of Migration to Regional Development (OECD, forthcoming), the OECD has engaged in an extensive data collection effort. The resulting novel dataset (Astruc-Le Souder et al., forthcoming) offers unprecedentedly detailed information on the subnational geography of migration in OECD countries. Based on data from continuous population registers as well as censuses, the database contains population statistics for 22 OECD member countries between 2000 and 2020, primarily at the level of municipalities with a few exceptions such as Germany (districts / Kreise) and the United States and Canada (Census tracts /Census subdivisions). ³ The main characteristics available at the municipal level include country of origin, age and sex. The data have been collected using Application Programming Interfaces (APIs) when possible (for 19 countries) or directly from the national statistical institutes. Data harmonisation ensures consistency of the data over time despite changes to municipal boundaries. This project will be officially launched in January 2022. Soon after, the OECD Municipal Migration Database (MMD) will be made public and will then be updated and refined on an ongoing basis.

The Municipal Migration Database (MMD) offers new opportunities to examine the spatial settlement patterns of migrants at a highly granular level, going beyond large regions. While the level of granularity depends on the number of municipalities or census tracts in each country, the new dataset provides additional insightful spatial information, as it can be aggregated to larger geographic levels. For example, the new dataset enables consistent international comparisons of migration trends across small regions (i.e. TL3), metropolitan areas or by Degree of Urbanisation. As a result, it is now possible to measure the concentration of migrants within regions and cities of all sizes, and it can also be used to analyse possible intra-metropolitan patterns of segregation. Moreover, the consistency of the dataset (with respect to municipal boundaries) facilitates tracking changes in settlement patterns of migrants over time.

Migrants concentrate more in metropolitan areas and cities than native-born

Within OECD countries, significant spatial differences in migration exist. Figure 1 displays the share of migrants among the local population across OECD countries with granular population data on foreign-born individuals. The MMD reveals clear geographic differences, especially in countries with detailed geographic breakdown (i.e. information on small administrative units), such as France, Spain and Italy. In France, migrants are particularly concentrated in and around large cities. In Spain, the data show the large concentration in municipalities that surround the major cities such as Madrid, Barcelona and Valencia, as well as in the communities along the Mediterranean coast.

Figure 1: Share of foreign-born population in municipalities and census tracts in the OECD, 2020

Population share of foreign-born across municipalities and census tracts, 2020 or latest available year

Note: The maps show the population share of foreign-born individuals across municipalities or other granular administrative units in OECD countries. Data are for 2020 or the latest available year. The underlying sample covers the entire local resident population.

Source: based on data from Astruc – Le Souder et al. (forthcoming).

In OECD countries, the migrant population share has increased in recent years, reaching 12% in 2019. Using the novel granular data on migration contained in the MMD, OECD analysis shows that migrants are significantly more concentrated in specific types of regions than the native-born population. More than half of the foreign-born population (53%) live in large metropolitan regions (small regions that include a metropolitan area of 1.5 million inhabitants), compared to only 40% of natives (Figure 1). ⁴ Less than a fifth of migrants (19%) reside in non-metropolitan regions, compared to almost 30% of the native-born population. The difference in the location of migrants and natives is particularly striking in regions near a metropolitan area and remote regions, where only 6% and 3% of migrants live, respectively. Among the native-born population, those regions account instead for 12% (regions near a metropolitan area) and 5% (remote regions) of the entire population.

Figure 2: Distribution of the foreign- and native-born population by type of TL3 region, 2020

Distribution of foreign- and native-born population by type of OECD (TL3) region, 2020 or latest available year

Note: Footnote 5 describes and explains the classification of small (TL3) regions by their access to metropolitan areas. The underlying sample covers the entire local resident population. Data are for 2020 or the latest available year.

Source: Author’s elaboration based on data from Astruc – Le Souder et al. (forthcoming).

Differences in urbanisation offer another insightful perspective on the geography of migrants and the changes over time. Using the Degree of Urbanisation methodology ⁵ to distinguish different types of settlement for European countries and the novel granular migration dataset shows that cities – defined as local units above 50 000 inhabitants with a population density of over 1 500 inhabitants per square kilometre – have significantly higher migrant population shares compared with other areas in almost all OECD countries with available data. ⁶ For example, in Austria, Belgium, Australia and France, migrants made up at least twice as share of the population in cities than in towns and semi-dense areas or rural areas in 2019. The spatial differences are particularly striking in Belgium and the Netherlands, where migrants account for 33% (Belgium) and 17% (Netherlands) of the population in cities but only 12% (Belgium) and 7% (Netherlands) in towns and semi-dense areas, with rural areas reporting even lower migrant population shares. However, in various other OECD countries, the migrant community is more equally spread out along the urban-rural continuum. In Italy, differences between cities (12%), towns and semi-dense areas (11.1%) and rural areas (9.5%) are relatively small. Additionally, cities and towns and semi-dense areas have relatively similar migrant population shares in both Ireland and Spain.

Figure 3: Share of migrants across OECD countries by Degree of Urbanisation, 2020

Foreign-born population share by Degree of Urbanisation, 2020 or latest available year

Note: 2020 or latest available year Data for the United Kingdom are limited to England and Wales. The underlying sample covers the entire local resident population.

Source: Author’s elaboration based on data from Astruc – Le Souder et al. (forthcoming).

More effective subnational migration policies with more granular data

To manage the integration of migrants successfully, policy makers in OECD countries require more detailed and informative data on migrants and migration flows. The need for detailed, targeted, data is particularly important for regional development policies because migration and its potential economic or demographic benefits differ widely within countries, which could weaken the effectiveness of policies implemented at the national level policies. From a policy perspective, understanding the spatial distribution of migrants is the first step to adopting tailored and targeted policies to fit local conditions and challenges. The new OECD dataset presented in this article offers a novel source of subnational data on migration. It not only entails unprecedentedly detailed geographic information for 22 OECD countries but also supports policy design by enabling analysis of how migration differs across cities, metropolitan and rural areas.

References

• Astruc-Le Souder, M. et al. (forthcoming), “Going granular: a municipal migration database”, OECD Regional Development Working Papers.
• European Commission, and Statistical Office of the European Union (2021), Applying the Degree of Urbanisation — A methodological manual to define cities, towns and rural areas for international comparisons, http://dx.doi.org/10.2785/706535.
• European Union et al. (2021), Applying the Degree of Urbanisation, http://dx.doi.org/10.2785/706535.
• OECD (forthcoming), Contribution of Migration to Regional Development, OECD Publishing, Paris.
• OECD et al. (2021), Applying the Degree of Urbanisation: A Methodological Manual to Define Cities, Towns and Rural Areas for International Comparisons, https://doi.org/10.1787/4bc1c502-en.
• OECD-EU (2020), Cities in the World: A New Perspective on Urbanisation, OECD Publishing, Paris, https://doi.org/10.1787/d0efcbda-en.

1. Regions within the 37 OECD countries are classified on two territorial levels reflecting the administrative organisation of countries. The 398 OECD “Territorial Level 2” (TL2) regions are those at the highest subnational administrative level, for example, the federal states in Germany. For more, see: OECD (2020), OECD Territorial Grids, http://stats.oecd.org/wbos/fileview2.aspx?IDFile=cebce94d-9474-4ffc-b72a-d731fbdb75b9.↩
2. See, for example, Diaz Ramirez, M., et al. (2018), “The integration of migrants in OECD regions: A first assessment”, OECD Regional Development Working Papers, No. 2018/01, OECD Publishing, Paris, https://doi.org/10.1787/fb089d9a-en.↩
3. The dataset covers the following countries: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Korea, Luxembourg, Mexico, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, the United Kingdom (England and Wales) and the United States of America.↩
4. To assess differences in socio-economic trends in regions, OECD groups small regions (TL3) based on the presence/absence of metropolitan areas and the extent to which the latter is accessible by the population living in each region. According to this typology, TL3 regions are classified as metropolitan if more than half of their population lives in a Functional Urban Area (FUA) of at least 250 000 inhabitants and as non-metropolitan otherwise. A metropolitan region becomes a large metropolitan region if the FUA accounting for more than half of the regional population has over 1.5 million inhabitants. In turn, the typology further classifies non-metropolitan regions based on the size of the FUA that is most accessible to the regional population.↩
5. The Degree of Urbanisation is a methodology to classify cities, towns & semi-dense areas, and rural areas for international comparative purposes. The method proposes three types of areas reflecting the urban-rural continuum instead of the traditional urban–rural dichotomy. The methodology was developed jointly by the OECD and other international organizations, and has been endorsed at the UN Statistical Commission as the recommended method to make international statistical comparisons between cities, urban and rural areas. Two recent global application of the definition are made in the OECD-EU publication Cities in World. A new perspective on urbanisation (OECD-EU, 2020) and Applying the Degree of Urbanisation: A Methodological Manual to Define Cities, Towns and Rural Areas for International Comparisons (OECD et al., 2021).↩
6. Figure.2 combines data provided by Eurostat for European countries with the new granular migration dataset for non-European countries. For the latter, local areas such as municipalities can be categorised by the degree of urbanisation using grid level information on population size and density of those areas (European Commission, and Statistical Office of the European Union, 2021). ↩