Author: Gwilym Pryce (on behalf of the Life at the Frontier research team)
Project: Life at the Frontier: The Impact of Social Frontiers on the Social Mobility and Integration of Migrants
Key findings:
- Social frontiers arise when neighbouring communities have abrupt, rather than gradual, social boundaries, created by a lack of residential mixing at the border.
- Our research has confirmed that social frontiers are not just a UK/US phenomenon: we find strong evidence of social frontiers between migrants and non-migrants in Norwegian and Swedish cities.
- We find evidence in the UK context of a strong association between social frontiers, higher crime rates, and higher concentrations of offenders.
- Our research on Dutch data reveals a strong association between social frontiers and residents’ moving decisions. We find that the effect is asymmetric: native Dutch households near the frontier are more likely to move, whereas non-Western migrant households are less likely to move, compared to equivalent households without exposure to social frontiers.
- Our research on Oslo finds that house prices are lower near social frontiers.
- Our qualitative research suggests that the meaning and impact of social frontiers is likely to be highly contextual, dependent on local ethno-social, political, and economic contexts.
- Our theoretical and methodological work suggests that more work is needed to develop methods and data linkage that will help us understand the causal impact of social frontiers and how they differ from other neighbourhood effects.
Introduction
There is a long history of research exploring the prevalence and impacts of residential segregation and neighbourhood effects. Cutler and Glaeser (1997 p.827), for example, found that ethnic minorities in more segregated areas in the US had significantly worse outcomes than their compatriots in less segregated areas. Based on their findings, a decrease in segregation by one standard deviation would eliminate one-third of the ethnic differences in education, unemployment and income. More recently, Raj Chetty’s research on the US has found that areas with high levels of ethnic segregation tend to have lower rates of social mobility (see the previous article on Refugee and Migrant Integration).
Alongside the social and economic geography literatures and geography, there is a sizable amount of epidemiological literature exploring the effect of ethnic density. These studies have shown that for ethnic minorities, living in areas with a high concentration of their own group can confer significant reductions in the risk of poor mental health. Becares et al.’s (2018) systematic review of 41 ethnic density studies, for example, finds a large reduction in the relative odds of psychotic experiences and suicidal ideation for each ten percentage-point increase in own ethnic density.
How can we reconcile these findings from different literatures and what do they mean for the integration of migrants? To understand how these different effects interact, it is important to address a potentially important omission from both these literatures: namely the failure to consider what happens where neighbourhoods border. For example, if there are gradual transitions in ethnic mix between neighbouring communities, but with high levels of ethnic mixing at the border, it may mean that ethnic minorities can benefit from the emotional and social support from their own group, without being cut off from the majority group, enabling them access wider social networks that lead to job opportunities and provide access to positive role models. However, if these same neighbourhoods had abrupt transitions in the ethnic mix creating a clearly defined “social frontier” then the outcome may be quite different.
An important part of our NordForsk research has been to explore what abrupt social boundaries between migrant groups might imply for the lives of residents. Drawing on insights from human geography, sociology and social psychology our task has been to think through the various causal pathways through which social frontiers may impact neighbourhood conflict and individual life outcomes, with a view to providing a rigorous framework for future empirical research in this area.
Our research suggests that an important feature of social frontiers is that they are likely to inhibit opportunities for positive direct contact between groups. This is significant because we know from the social psychology literature that positive direct contact with members of an outgroup decreases prejudice and hostility towards that group (Allport, 1954) via three explanatory pathways: increased knowledge, increased empathy, and decreased anxiety (Pettigrew & Tropp, 2008). However, opportunities to build shared experience and understanding between groups via positive direct contact are less possible in neighbourhoods divided by social frontiers which, over time, may increase social tension and conflict.
The abrupt and clearly defined nature of separation implied by social frontiers may also be important. We know from the human geography and political geography literatures that rival social groups often exhibit territorial behaviours when there are clearly demarcated spatial territories to defend. Therefore, because social frontiers represent such a distinct spatial location for the boundaries between groups, they have the potential to heighten territoriality and sour any contact that does take place. This is significant because “negative contact predicts increased prejudice more than positive contact predicts reduced prejudice” (Barlow et al. 2012, p.1629).
Do social frontiers exist in Nordic countries?
Research by Dean et al. (2019) and Smith et al. (2022) has confirmed the existence of social frontiers between migrants and native-born households in the UK (specifically, Sheffield and Rotherham). Our ongoing NordForsk research exploring the nature of social frontiers in UK towns and cities has found that they are a common phenomenon, although we do not yet know whether they are becoming more or less prevalent. Previous research has also found evidence of social frontiers in US cities (Legewie and Schaeffer 2016; Legewie 2018), although the focus of these studies has been on racial divides rather than migrant/native frontiers.
The question, then, is whether such frontiers exist in Nordic countries? Our analysis of the geographical distribution of migrants and non-migrants in Norway and Sweden has confirmed that social frontiers do indeed exist in Nordic towns and cities. See, for example, the map of Stavanger in Norway where the black lines indicate the existence of frontiers between migrants and non-migrants. Moreover, results using data as far back as 1990 show that the existence of these social frontiers has been very stable over time throughout the country.
Social frontiers: Impact on offender location and crime
One of the most important questions with respect to social frontiers is whether they have an impact on crime. We noted above that sharp boundaries between contrasting groups may help ferment social tensions by reducing opportunities for contact and by encouraging territorial attitudes. Our previous empirical research on this phenomenon (Dean et al. 2019) found evidence of a strong association between the location of social frontiers between migrants and non-migrants and higher crime rates in the city of Sheffield in the UK. However, this study relied on crime data that did not have precise geographical coordinates of where offences occurred. Also, the paper did not explore the location of those committing offences so we could not establish, for example, whether social frontiers were making locals more likely to commit offences, or whether they had become places that attracted criminal activities that would otherwise have taken place elsewhere.
In our NordForsk project, we have worked closely with South Yorkshire Police to explore whether the association between social frontiers and crime still holds when we use crime data with more reliable information on the geographical location of the crimes. Our analysis finds that this is indeed the case. We found that the results of Dean et al. (2018) were confirmed when more spatially precise crime data was used: frontier zones tend to have elevated crime rates. We also extended the research to the neighbouring local authority of Rotherham and found similar effects. Moreover, we found that it was not only social frontiers between migrants and non-migrants that are associated with higher crime rates. A similar relationship holds for social frontiers between Muslims and non-Muslims, and between White British and non-White British populations (Smith et al. 2022).
Our collaborative work with South Yorkshire Police also enabled us to identify the residential addresses of known offenders and thus explore the relationship between social frontiers and the location of offenders. We found that there was a higher concentration of offenders living near social frontiers. However, at this stage we are unable to identify the precise causal mechanism. For example, we do not know whether social frontiers are attracting existing offenders to locate nearby, or whether social frontiers increase the likelihood of locals becoming offenders or foster the creation of local crime networks. Legewie and Schaeffer (2016) argue that because social frontiers will tend to lie at the periphery of communities and therefore away from the mechanisms of social control which tend to be strongest near the geographical centre of a community. Lower social control and collective efficacy near social frontiers will tend to attract criminals and also leave the emergence of criminal behaviour unchecked.
Impact of social frontiers on household mobility
One of the potential interpretations of social frontiers is that they represent a defensive boundary between established native households and incoming migrants: a moving defensive “wall” in the dynamic process of “native flight”. Community boundaries could therefore be interpreted as attempts by the majority group to contain the residential spread, and perceived threat, of ethnic minorities (Alba 2005: 27). As such, one side may see the frontier as something they want to preserve as a way of maintaining social distance with respect to the outgroup, whereas the other side may view it as an unwelcome barrier to their progress and integration into wider society. The implication is that social frontiers will have asymmetrical impacts for the groups on each side: native-born residents near a social frontier may be more likely to move out than natives not living close to a frontier. They view the social frontier as a defence against the threat of incoming migrants. In contrast, migrants may view the frontier as a place of opportunity and settlement, an expression of their desire to integrate with the native population. As a result, migrants near social frontiers may be less likely to move than migrants living further away.
So what does the data say? Working with colleagues from Delft Technical University, we developed a statistical model of moving behaviour to estimate the impact of social frontiers between natives and non-Western migrants in Rotterdam on moving decisions. On the migrant side of the frontier, we found that households of all ethnicities in the 28-37 age group tend to have reduced odds of moving, compared to non-frontier parts of the city.
However, we found that the opposite was true on the Dutch native side of the frontier where households tend to have an increased propensity to move away. One implication of these findings is that the existence of social frontiers may indicate resistance to migrant integration, not from migrants, but from the host population.
Impact of social frontiers on house prices
One possible reason why social frontiers form is that they reflect a relative shortage of households willing to live near the outgroup. Consider the following thought experiment: suppose all households, both native and migrant, are averse to living in areas where they will be in the minority among their immediate neighbours. Other things being equal, we would expect this to cause them to locate to areas where they will be surrounded by a concentration of their own group (Schelling 1971). A by-product of this will be the prevalence of social frontiers where there is minimal residential mixing along neighbourhood boundaries because there is no group of households with a preference towards living near the outgroup (Piekut and Pryce 2022).
This has implications for house prices and rents. Because no one wants to live at the frontier, we would expect house prices to be lower near to social frontiers. More generally, if the supply of houses at the frontier is large relative to demand, house prices will be lower there. Put another way, households will only locate to the frontier if the discomfort of locating there is sufficiently offset by lower housing costs.
Now consider another thought experiment. Suppose that, rather than both groups having homogenous preferences, there are in fact a variety of preferences among households with respect to living near outgroups, with some households very happy to live near the outgroup. In this scenario, we would only expect social frontiers to emerge as a result of arbitrary or external factors (e.g. historical planning decisions or refugee allocation policies). Moreover, because there is no shortage of people willing to live near the frontier, we would not expect house prices to be any different near social frontiers than anywhere else, assuming other factors are constant, such as the size and quality of housing.
Note, however, under this scenario, and in the absence of external constraints (such as aparthied laws), social frontiers would not survive for very long. The relative abundance of households willing to live near outgroups would likely lead to residential mixing at the boundary, and all social frontiers disappearing over time. There would be no aversion to residential mixing, and so the locations of migrants and natives would become intermingled and residential integration would be achieved.
The implication of all this is that, in the absence of external barriers to residential mixing, social frontiers which persist over time will likely reflect an underlying shortage of households willing to live at the boundary between groups, relative to the number of dwellings there. As a result, we would expect house prices to be lower near social frontiers. If there is no aversion to living near other ethnic/migrant groups, then we would not expect house prices to be lower, but then we would also not expect social frontiers to emerge and persist.
To explore these ideas, we have developed a detailed model of the factors that determine geographical variation in house prices and used this model to estimate whether or not house prices near social frontiers are noticeably lower (Myatt et al. 2022). The model controls for a wide range of factors that influence house value, including dwelling characteristics (house type, size, age), demographics, and the socio-economic characteristics of the neighbourhood. The results suggest that holding these other factors constant, social frontiers do seem to have an effect; house prices do indeed tend to be lower near social frontiers. Moreover, the longer and the more pronounced the frontier, the greater the house price discount.
Further research is needed, however, to establish the precise causal mechanism at work. For example, it may be that social frontiers generate higher rates of crime and it is this that drives down house prices. Alternatively, it may be that migrants also have lower socio-economic status and it is because wealthy households are averse to living near poorer ones that house prices are lower, rather than the fact that these poorer households have a different ethnicity or country of birth.
Qualitative research
As well as developing statistical analysis of the impacts of social frontiers, we are in the process of conducting in-depth interviews with residents, community workers and local experts in the cities of Rotherham (UK) and Malmö (Sweden). Our goal in conducting these interviews is to understand in more depth how locals recognise the existence of social frontiers, how they perceive them, what their experiences are of living near them, and what they think the impacts are. Although we are still analysing the data, our initial findings suggest that single-dimensional frontiers such as the divides between foreign-born and native-born residents, may be too simplistic. There are often multiple overlapping factors that give rise to geographical divisions, such as social class, education, housing and subtle cultural differences between multiple ethnic groups that are not captured by a simple native/migrant divide.
Future work
The story of social frontiers and their impact remains a new and relatively unexplored field. There is still much that we do not know about the nature, prevalence, morphology and impacts of social frontiers. It would take many decades of research to fully understand how all these factors change over time and vary across different cultural, social, economic and institutional contexts. Nevertheless, in the remaining 12 months of our NordForsk project and associated PhD research and research grant applications over the next four years, we plan to extend this pioneering field of enquiry in a number of ways:
- We will develop aggregate measures of social frontiers that allow us to compare the prevalence of social frontiers across different towns, cities and regions, and how they change over time.
- We will explore the associations between these aggregate measures and various outcome measures such as crime, wellbeing, educational attainment, individual income and social mobility.
- We will develop new methods of estimating social frontiers that lead to more precise estimates of their geographical location and that allow for the detection of frontiers in multiple dimensions (ethnicity, country of birth, race, religion, housing wealth, income and social class), not just the migrant/native divide.
- We will conduct validation exercises to check whether our estimates of frontier locations resonate with locals’ perceptions of where community boundaries lie.
- Explore the relationship between residential segregation and labour market integration – particularly the importance of social frontiers for work income.
- Draw on qualitative research to explore the lifestyle expectations, aspirations and desires of immigrants and natives in social frontier areas.
Policy implications
Our initial findings appear to confirm the relationship between social frontiers and higher crime rates, greater concentration of offenders, lower house prices and impacts on geographical mobility. Although these statistical associations are strong, it is difficult at this stage in our work to be sure of the causal effect. Nevertheless, we believe our emerging body of research provides sufficient de facto evidence of the potential importance of social frontiers to warrant policy interest. Our key recommendation at this stage is for policy makers to develop randomised control trials that explore the effectiveness of interventions in reducing the prevalence of social frontiers and, by extension, the effect on crime, house prices, moving decisions and social mobility. The Ukrainian refugee crisis potentially offers an important opportunity to implement such trials with a view to helping policy-makers improve the life outcomes of migrants and refugees.
References
- Bécares, L., Dewey, M. E., & Das-Munshi, J. (2018). Ethnic density effects for adult mental health. Psychological medicine, 48(12), 2054-2072.
- Cutler, D. M., Glaeser, E. L., & Vigdor, J. L. (1999). The rise and decline of the American ghetto. Journal of Political Economy, 107(3), 455-506.
- Dean N, Dong G, Piekut A, Pryce G (2019) Frontiers in residential segregation: Understanding neighbourhood boundaries and their impacts. Tijdschr Voor Econ Soc Geogr 110(3), 271–288
- Frontiers Exploring the asymmetric impacts of social frontiers on household mobility in Rotterdam, NordForsk Life at the Frontier Working Paper (in preparation).
- Galster G, Sharkey P (2017) Spatial foundations of inequality: a conceptual model and empirical overview. Russell Sage Found J Soc Sci 3(2):1–33
- Iyer, A. and Pryce, G. (2022) Theorising the Causal Impacts of Social Frontiers, NordForsk Life at the Frontier Working Paper (in preparation).
- Olner, D. Pryce, G. van Ham, M. and Janssen, H. (2022) The Conflicting Geographies of Social Frontiers, NordForsk Life at the Frontier Working Paper (in preparation).
- Piekut, A. and Pryce, G. (2022) Life at the Social Frontier: Re-conceptualisation, synthesis, and future directions, NordForsk Life at the Frontier Working Paper (in preparation).
- Schelling, T. C. (1971). Dynamic models of segregation. Journal of mathematical sociology, 1(2), 143-186.
- Smith, J., Pryce, G. and Zhang, M.L. (2022) Impact of social frontiers on offender location and crime in South Yorkshire, NordForsk Life at the Frontier Working Paper (in preparation).