What scale matters? exploring the relationships between individuals' social position, neighbourhood context and the scale of neighbourhood

Over the past few years, neighbourhood effects research has received significant attention from the academic world, not only in the US, where that attention has a longer tradition, but also in Western Europe. There is also substantial interest among policy makers. Most policy makers intend to reduce concentrations of poverty by enhancing the social mix of neighbourhoods. Avoiding high immigrant concentrations in particular neighbourhoods is another issue that fuels political debate and policy intervention in many Western European countries, Scandinavian countries included. However, there are clear gaps in the understanding of the relationship between neighbourhood composition and social outcomes. One of these gaps regards the scale of the neighbourhood; if there would be neighbourhood effects, what scale is it relevant to consider? Is mix good or bad for the social prospects of individuals at a level that is very local, for example a few neighbouring streets, or could mix be helpful at a somewhat higher scale? This article will focus on this issue, applying individual longitudinal data in multi-level models for the entire active population of the three largest metropolitan areas in Sweden. We will explore the degree to which the social and ethnic composition of geographical districts, at a variety of scales (measured at time t), are statistically related to individual employment and earnings for adult metropolitan residents at time t+1, controlling for relevant personal and household characteristics.