A Geographical Path to Integration? Using Sequence Analysis to Explore the Interplay between Regional Context and Socioeconomic Integration Among Refugees in Sweden

Louisa Vogiazides, Stockholm University
Hernan Mondani, Stockholm University

Migrant integration is at the forefront of political and academic debates in many immigrant-receiving countries. A rich neighbourhood effects literature analyses the significance of residential environment for the socioeconomic integration of international migrants. A parallel strand of research explores the associations between immigrants’ initial region of residence and their subsequent socioeconomic integration. The literature usually measures residential context at a single point in time. Less attention has been paid to individuals’ geographical trajectories in their entirety and their relationship with socioeconomic integration. Furthermore, existing research often focuses on a single dimension of geographical context, such as the labour market situation. Using Swedish longitudinal register data, we apply sequence analysis in order to identify refugees’ typical geographical trajectories across regions with different levels of population density and labour market conditions. We then estimate regression models to assess how the identified trajectories influence refugees’ subsequent employment.

A geographical path to integration?Using sequence analysis to explore the interplay between regional context andsocioeconomic integration among refugees in Sweden

 

Background and aim

The issue of migrant integrationis at the forefront of political and academic debates in manyimmigrant-receiving countries. A rich ‘neighbourhood effects’ literatureanalyzes the significance of the residential environment on the socioeconomicintegration of international migrants. Many European studies have found thatresidence in deprived neighbourhoods hampers immigrants’ employment and incomeprospects (e.g. Musterd et al. 2008). A parallel strand of research exploresthe importance of the first region of settlement for immigrants’ futureintegration outcomes. In Sweden, researchers have studied the effects of arefugee dispersal scheme – the ‘Sweden-wide’ strategy –, implemented between1986 and 1994, finding that the refugees who were settled in municipalitieswith poor labour market conditions experienced long-term losses in terms ofemployment and income (Edin et al. 2004).

Existing studies focus on theeffects of immigrants’ residential context at a single point in time. Littleattention has been paid to the interconnection between the long-termgeographical trajectories of immigrants in the host country and theirsocioeconomic integration. To address this gap, this paper explores how thedifferent geographical trajectories followed by refugees in Sweden correspondto different levels of socioeconomic integration.

Using rich Swedish longitudinalregister data, we apply sequence analysis in order to identify refugees’typical geographical trajectories across regions with different levels ofpopulation density, labour market conditions and housing market structure.Next, we estimate regression models to assess how the identified trajectoriesinfluence refugees’ employment and income outcomes.

Data and methods

This paper is based on registerdata from the GeoStar database, a full-population database compiled byStatistics Sweden containing longitudinal, annually updated andindividual-level data for the period 1990 to 2012. It includes a wide range ofvariables, such as demographic, socioeconomic and housing characteristics.

Our dataset consists of refugeeswho arrived in Sweden between 1990 and 2004 (N = 76,028). The individuals arefollowed over an eight-year period beginning the year after their arrival toSweden.
In the study, regions correspond to so-called Labour Market Areas (LMAs) whichare annually constructed by Statistics Sweden based on commuting zones. Weclassify regions according to their level of population density (includingmetropolitan regions, large city regions and rural region) as well as theirrate of (immigrant) unemployment and housing market structure.

In our analyses, we willdistinguish between refugees who were assigned to a municipality by the Swedishauthorities (either as part of the Sweden-wide strategy in the early 1990s or,for more recently arrived refugees, through the ‘Anläggningsboende’ (ABO)system) and those who arranged for their own housing (Eget boende –EBO). Weexpect that individuals who did not choose their first municipality ofresidence may be more inclined to relocate.

Sequence analysis is awell-suited technique to describe time ordered series of states (Billari 2001).In our study, the states are the residence in one type of geographical region:e.g. metropolitan region, large city region or sparsely populated region. Thetransitions (or the lack of them) between regional contexts over timeconstitute a sequence. Sequence analysis provides a systematic tool to definesimilarity scores between groups of sequences, and use those scores to clusterthe sequences into a typology of trajectories.

Preliminary results

Our preliminary results showthat the vast majority of refugees stay in the same type of geographical regionthroughout the 8-year period. This applies to both the 2000-2004 and the1990-1994 cohorts (see Figs. 1 and 2 respectively for the sequence clusters). Asmall share of refugees experience a transition between different types ofregional contexts. For both cohorts, the most common transition involves a movefrom a large city region to a metropolitan region. These movements take placeroughly between two and five years after arrival to Sweden.

At the same time, there arequalitatively significant period effects. A larger share of the 2000-2004refugee cohort began their residential trajectories in a metropolitan  region(50.8%) compared to the 1990-1994 cohort (36.7%) which arrived during theimplementation of the refugee dispersal scheme.

The next step will be toclassify regions according to their labour and housing market characteristics.Finally, we will study the interplay between geographical trajectories andintegration, which is of high relevance to policymakers in the current contextof high refugee inflows.

migseq_8y_F9208_RepSeq_5cl1990-1994.pngFigure 1.Representative sequences for five sequence clusters. 1990-1994.

migseq_8y_F9208_RepSeq_5cl2000-2004.pngFigure 2.Representative sequences for five sequence clusters. 2000-2004.

 

References

BILLARI,F. C. (2001). ‘Sequence Analysis in Demographic Research’ Canadian Studiesin Population, 28(2): 439-458.

EDIN,P-A, FREDRIKSSON, P & ÅSLUND, O (2004) ‘Settlement Policies and theEconomic Success of Immigrants’ Journal of Population Economics, 17(1):133-155.

MUSTERD, S., R. ANDERSSON, G.GALSTER & KAUPPINEN, T. (2008) ‘Are immigrants’ earnings influenced by thecharacteristics of their neighbours?’ Environment and Planning A, 40:785–805.

 

Presented in Session 1232: Posters