by Is ethnic segregation positive? In many European countries over the past decades, there has been increasing frustration over what is perceived as a lack of social integration of immigrants and children of foreign-born parents. One factor that researchers and politicians have highlighted as a possible explanation is what is called ‘neighbourhood effects,’ which, simply put, refers to how the area where one grows up and/or lives affects one’s life. There has often been interest in seeing if the population composition in specific neighbourhoods (‘immigrant-dense areas’) indicates the presence of segregation that may reinforce inequalities between foreign-born and native-born individuals in terms of education and employment. In a recently published report
Topics:
Lars Pålsson Syll considers the following as important: Politics & Society
This could be interesting, too:
Lars Pålsson Syll writes Problemen med Riksbankens oberoende
Lars Pålsson Syll writes The Berlin Wall
Lars Pålsson Syll writes Trump wins
Lars Pålsson Syll writes Ojämlikheten i siffror
Is ethnic segregation positive?
In many European countries over the past decades, there has been increasing frustration over what is perceived as a lack of social integration of immigrants and children of foreign-born parents. One factor that researchers and politicians have highlighted as a possible explanation is what is called ‘neighbourhood effects,’ which, simply put, refers to how the area where one grows up and/or lives affects one’s life. There has often been interest in seeing if the population composition in specific neighbourhoods (‘immigrant-dense areas’) indicates the presence of segregation that may reinforce inequalities between foreign-born and native-born individuals in terms of education and employment.
In a recently published report from the Expert Group on Public Economics (ESO) — Good Neighbours: an ESO report on the significance of neighbourhoods for integration — the authors show that ethnically segregated housing does not ‘per se’ have to be something negative. This finding has received some attention in the media and is portrayed as something exciting and new because the prevailing image of ‘immigrant-dense’ areas is often that a significant part of the problem there is that if people with the same background choose to live in the same area, it hinders their opportunities to learn Swedish, obtain a good education, and secure employment.
Upon closer reflection, however, it is obvious that no one could reasonably dispute that if one grows up and lives in a neighbourhood characterized by good role models, well-functioning networks, and highly educated and successful individuals, then the chances increase that one – regardless of ethnicity – will acquire a good education and job.
The interesting question, therefore, is not whether segregation can be positive — we already know that – but whether the existing segregation we have in Sweden today is positive. In the policy area, this question becomes important because many politicians want to limit the right to free choice of residence, arguing that families who do not settle in immigrant-dense areas or who leave these ‘vulnerable’ areas have better opportunities for good education, language skills, and jobs.
Research conducted in Sweden has (unsurprisingly) found that where we live matters and that for socioeconomically disadvantaged groups and immigrants, there is a high degree of correlation between where one lives and the extent to which one has higher education and stable employment. We know that ethnic segregation has increased over the past three decades and that there is a strong connection — especially in some metropolitan regions — between ethnic and economic segregation. However, we have significantly less knowledge about the mechanisms behind these correlations.
What one – especially in quantitatively oriented research – does is select some measurable factors (‘variables’) which are put into so-called regression models, trying to measure to what extent these factors covary (‘correlate’). This is often done — as in the ESO report — without a thorough examination of to what extent these measurable factors (education, income, place of residence, etc.) really represent the mechanisms we ultimately are (should be) trying to identify. Naturally, this is troubling if we want to use policy measures to reverse what most perceive as a negative development. Without identifying the mechanisms that create societal problems, it is difficult to choose which measures can be expected to have the greatest effects.
What we need is causal knowledge about the mechanisms behind the observed relationships. To provide us with useful knowledge, also requires using other types of tools in the toolbox — such as e.g. attitude surveys — that can complement the statistical data analyses. This is important, especially in the case of neighbourhood effects, where we are often not interested in stating that these on average exhibit a certain pattern but want to know why the effects for specific groups in specific contexts can sometimes be positive and in other cases and for other groups can be negative. Focusing only on statistical results in evaluation contexts often does more harm than good.
The type of mathematical-statistical investigations that the ESO report is a clear example of have often received an undeservedly high status within social sciences in recent times. Why? Fundamentally since they give the impression of delivering rigorous and precise answers to the questions researchers ask. The price for the sought-after precision, however, is in most cases far too high. By adapting the questions in such a way that one can get precise answers, we get answers to questions we are not really interested in.
Another telling example of this is today’s hype among quantitatively oriented social researchers to engage in so-called randomized control trials (RCT). The design becomes the main focus and just by getting more or less ingenious experiments in place, one believes they can draw far-reaching conclusions about both causality and the ability to generalize experimental outcomes to larger populations. Unfortunately, this often means that this type of research gets a negative shift away from interesting and important problems to instead prioritize method selection. Design and research planning are important, but the credibility of research fundamentally lies in being able to answer relevant questions we as both citizens and researchers want answers to. Instead of accepting that a lower degree of certainty is inevitable, one engages in model constructions that enable certain knowledge. If the goal is knowledge about the real world, however, the value of these is at best unclear.
The focus on analyzing social problems using statistical models has largely created a new incomprehensibility. The narrow focus on building mathematical-statistical models and theories, where precision, elegance, and refinement take precedence over relevance, has led to a lack of overview and that quantitatively oriented social research today contributes more to the development of applied mathematics than to the social sciences.
The ESO report Good Neighbours, unfortunately, contributes no more than marginally to an accumulation of knowledge that can form the basis for evidence-based policy measures in the area of neighbourhood effects and immigration. What emerges in the report mostly confirms things we already knew, and most of the truly important policy questions we ask are not answered beyond rather meaningless statements of the type “segregation can be good if… but can also be bad if…”. How this is supposed to help us solve the important integration problems in our societies is hard to see. One might also ask whether the enormous resources that are now annually spent on this type of research could be used in a more policy-effective way.