In less than a century Western European societies have undergone a rapid demographic change, from ethnic homogeneity towards increasing diversity. This also applies to Sweden where one reaction attributed to this change has been the emergence of anti-immigrant radical right wing party politics. In this project we investigate a sample of politically active individuals - those who become candidates in municipal elections in Sweden - to learn about (1) How ethnic preferences shape residential choices and (2) Whether growth in the number of radical right wing party candidates can be attributed to exposure to diversifying neighbourhoods.
The first study focuses on residential mobility and the underlying question of how residential choices of the majority group affect residential segregation patterns. Over the years this question has generated large volume of literature, most notably the debate about the role and extent of white flight and white avoidance processes. Recent studies have highlighted the importance of looking beyond reactions to local neighbourhood conditions, because extra-local characteristics may reduce mobility by limiting the availability of attractive relocation choices. This issue can be addressed with Swedish register data, where we can use 100x100 meter grid data to construct neighbourhoods around each party candidate, and investigate the effect of local conditions at various geographic scales. The reason why we focus on political party candidates is that we assume that radical right wing party candidates are particularly sensitive to diversity related conditions in the neighbourhood. With discrete choice conditional logistic modelling we can compare their residential choices to choices of other party candidates to learn about the role of ethnic preferences on residential mobility choices.
In the second study we examine whether experience of living in a diversifying neighbourhood is associated with higher probability of joining the ranks of radical right wing party. We use population register panel data with information on whether individuals run for municipal elections in 2006 and 2010 and use first difference regression model which looks at the effect of change in neighbourhood conditions on change in individual running behaviour.