Neighbourhood Concentration and Representation of Non-European Migrants: New Results from Norway

In a special issue of the European Journal of Population, Andersson et al. (2018) presented a comparative study of segregation patterns of non-European migrants in Belgium, Denmark, the Netherlands and Sweden. In all countries, individual-level, geo-coded, register data were used to compute segregation measures based on the k-nearest neighbours approach. These individualized scalable neighbourhoods provide comparable neighbourhood definitions and allow for an analysis of segregation patterns at different scales. One important finding of this study was that small-scale segregation patterns, based on neighbourhoods encompassing the 200 nearest neighbours, were remarkably similar across the different countries, with dissimilarity indices (DI) in the 47.5–51.2% range. At larger scales, differences were much more marked. For neighbourhoods encompassing the nearest 51,200 neighbours, the DI in Belgium was 40.6% compared to only 25.3% in Denmark. These findings point to the relevance of scale and suggest that factors influencing segregation at the local level can be different from those that influence segregation at larger scales. A tentative interpretation of these patterns was that similarities in small-scale segregation patterns might be related to the influence of ethnic preferences, whereas differences in housing policies, housing market structure and settlement policies are candidate explanations for a more pronounced large-scale segregation in Belgium.

This research note updates the results from this earlier study by adding results from Norway. In terms of its welfare state structure, Norway is similar to Denmark and Sweden, but differs from these countries in important areas such as settlement policies for refugees and asylum seekers, settlement patterns, de-centralization policies, geography and housing market structure. Thus, extending the analysis of segregation patterns to include Norway will make it possible to evaluate, first, if small-scale segregation in Norway falls within the narrow interval identified for Belgium, Denmark, the Netherlands, and Sweden. And, second, if Norway, with its particular combination of policies, has lower segregation levels and a more even representation of non-European immigrants.

According to Musterd and Ostendorf (1998), ethnic segregation levels have historically been lower in Europe than in the USA. However, studies of ethnic segregation in Europe have shown large differences between countries. Regarding the countries considered in this research note, Musterd (2005) found relatively low segregation levels in Oslo. Higher levels were found in Belgian cities and for Iranians in Stockholm. Segregation levels for different groups in different Dutch cities varied widely. Musterd and Van Kempen (2009) show relatively high segregation levels for some groups in Dutch and Belgian cities, compared to two Swedish cities. Arbaci (2007) emphasizes the role of welfare regimes in producing patterns of ethnic segregation. Comparing Nordic capitals with similar welfare arrangements, Skifter Andersen et al. (2016) found the highest segregation levels for non-European migrants in Stockholm and the lowest in Helsinki, with Oslo and Copenhagen in between. They emphasize the role of housing markets and housing policies.

Norway is a Nordic welfare state, similar to Denmark and Sweden with regard to policies in areas such as health care, social services, and education, but there are marked differences in some areas that are salient to residential segregation. Immigration has historically been higher in Sweden than in Denmark and Norway. Norway is generally considered to have taken an intermediate position in immigration policy, more restrictive than Sweden and less restrictive than Denmark (Brochmann 2017). Norway’s settlement policies for refugees and asylum seekers have also differed substantively from those of Sweden, but resemble those of Denmark. While Swedish policies emphasize voluntary settlement, asylum seekers and refugees in Norway are for the most part settled in municipalities through a system of agreements between the municipalities and the central authorities (Brochmann 2017; Directorate of Integration and Diversity 2010; Ministry of Justice and Public Security 2016). While they may move freely after this initial settlement, many choose to move to the Oslo region or other major cities (Stambøl 2013). This policy leads us to expect a more dispersed non-European immigrant population in Norway than in Sweden. Further, Norway and Sweden are both larger in geographical terms and have a lower population density than the other countries. The total land area of mainland Norway is approximately eight times that of Denmark and the Netherlands, and ten times that of Belgium, while Sweden is 1.4 times the size of Norway (CIA 2016). While the population density in central areas of Norway and Sweden is high, large areas are sparsely populated or uninhabited. Also, settlement is less centralized in Norway than in Sweden, as a smaller proportion of the Norwegian population resides in urban areas (The World Bank 2018). De-centralization has historically been an important political goal in Norway. Finally, housing policies and the housing market structures in Norway are quite different from those of Sweden, Denmark, the Netherlands and Belgium. In Norway, 77% of households own their dwelling, and public housing only comprises 4% of the housing stock (Statistics Norway na. a, na. b, na. c).

The data used here are based on population register data for the entire Norwegian population registered as resident on 1 January 2011, provided by Statistics Norway. The results are compared to corresponding figures from Sweden, Denmark, the Netherlands and Belgium, as described in Andersson et al. (2018). To facilitate comparative analyses, we have aimed to make the data as similar as possible across countries. The process of harmonizing the national data sets is documented in Nielsen et al. (2017) and was the aim of the ResSegr project.¹

The k-nearest neighbours approach to measuring segregation is well suited for comparative analyses, as it provides a comparable definition of a neighbourhood; the k-nearest neighbours of each individual. This partially circumvents the Modifiable Areal Unit Problem (Hennerdal and Nielsen 2017) by allowing for a comparison of residential patterns that do not rely on administrative borders. Further, the neighbourhoods are scalable, allowing us to study segregation at both the macro-level (k = 51,200), the micro-level (k = 200) and at intermediate levels. However, a drawback of this method is that the geographical size of each neighbourhood is determined by the local population density. Thus, the geographical area that is considered a “neighbourhood” is highly variable and may become very large at high k values, particularly in less densely populated areas in Norway and Sweden.

Andersson et al. (2018) found that small-scale segregation patterns of non-European migrants are similar across Belgium, Denmark, the Netherlands and Sweden, indicating a striking consistency in small-scale segregation levels across contexts. The segregation patterns presented here for Norway contradict this consistency and illustrate the diversity of segregation patterns, also at the small scale. We find overall segregation levels, as measured by the DI, to be lower in Norway at the neighbourhood scales of 200 and 1600 nearest neighbours than in the four countries studied by Andersson et al. (2018), and we find non-European immigrants to be more evenly represented across the country here than in the other national contexts. Concentration patterns in Norway are similar to those found in Denmark, but with higher concentration levels in the most immigrant-dense neighbourhoods at the macro-scale.

There are several candidate explanations for these patterns. One is the high prevalence of owner-occupied housing, which may contribute to a relatively even representation of non-European immigrants in Norway. Our findings are in line with those of Skifter Andersen et al. (2016) for Scandinavian capitals, which would support this idea. However, since we consider entire countries, and not just urban segregation, we believe that settlement policies for refugees and asylum seekers may be central to explaining why Norway displays a relatively even representation and low concentration levels. These policies may work in tandem with housing policies, policies aimed at maintaining the rural population, and a high rural employment rate, as well as universal social and welfare policies, making it relatively more attractive to remain in rural areas and small towns after initial settlement. Finally, a distinct pattern of macro-level segregation in Oslo between the east and the west (cf. Wessel 2017) may explain the higher concentration at the top of the distribution at high k levels in Norway compared to Denmark. Unfortunately, our data do not allow us to test these different explanations directly.

In sum, our results emphasize the highly variable and context-dependent nature of segregation patterns, the importance of comparable measures of segregation in comparative research, and the relevance of geographical scale in the study of segregation. Also, our results are consistent with the notion put forth by Andersson et al. (2018); non-European migrants are not only concentrated in migrant-dense areas. To the contrary, they are represented more evenly across the country in Norway than in Belgium, Denmark, the Netherlands and Sweden, especially at smaller geographical scales.